path: root/trunk/2.6.18/10002_xen-3.3.0.patch

diff -rpuN linux-2.6.18.8/arch/i386/boot/Makefile linux-2.6.18-xen-3.3.0/arch/i386/boot/Makefile
--- linux-2.6.18.8/arch/i386/boot/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/boot/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -26,7 +26,7 @@ SVGA_MODE := -DSVGA_MODE=NORMAL_VGA
 #RAMDISK := -DRAMDISK=512
 
 targets		:= vmlinux.bin bootsect bootsect.o \
-		   setup setup.o zImage bzImage
+		   setup setup.o zImage bzImage vmlinuz vmlinux-stripped
 subdir- 	:= compressed
 
 hostprogs-y	:= tools/build
@@ -133,5 +133,13 @@ zlilo: $(BOOTIMAGE)
 	cp System.map $(INSTALL_PATH)/
 	if [ -x /sbin/lilo ]; then /sbin/lilo; else /etc/lilo/install; fi
 
+$(obj)/vmlinuz: $(obj)/vmlinux-stripped FORCE
+	$(call if_changed,gzip)
+	@echo 'Kernel: $@ is ready' ' (#'`cat .version`')'
+
+$(obj)/vmlinux-stripped: OBJCOPYFLAGS := -g --strip-unneeded
+$(obj)/vmlinux-stripped: vmlinux FORCE
+	$(call if_changed,objcopy)
+
 install:
 	sh $(srctree)/$(src)/install.sh $(KERNELRELEASE) $(BOOTIMAGE) System.map "$(INSTALL_PATH)"
diff -rpuN linux-2.6.18.8/arch/i386/Kconfig linux-2.6.18-xen-3.3.0/arch/i386/Kconfig
--- linux-2.6.18.8/arch/i386/Kconfig	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/Kconfig	2008-08-21 11:36:07.000000000 +0200
@@ -16,6 +16,7 @@ config X86_32
 
 config GENERIC_TIME
 	bool
+	depends on !X86_XEN
 	default y
 
 config LOCKDEP_SUPPORT
@@ -103,6 +104,17 @@ config X86_PC
 	help
 	  Choose this option if your computer is a standard PC or compatible.
 
+config X86_XEN
+	bool "Xen-compatible"
+	select XEN
+	select X86_PAE
+	select X86_UP_APIC if !SMP && XEN_PRIVILEGED_GUEST
+	select X86_UP_IOAPIC if !SMP && XEN_PRIVILEGED_GUEST
+	select SWIOTLB
+	help
+	  Choose this option if you plan to run this kernel on top of the
+	  Xen Hypervisor.
+
 config X86_ELAN
 	bool "AMD Elan"
 	help
@@ -213,6 +225,7 @@ source "arch/i386/Kconfig.cpu"
 
 config HPET_TIMER
 	bool "HPET Timer Support"
+	depends on !X86_XEN
 	help
 	  This enables the use of the HPET for the kernel's internal timer.
 	  HPET is the next generation timer replacing legacy 8254s.
@@ -263,7 +276,7 @@ source "kernel/Kconfig.preempt"
 
 config X86_UP_APIC
 	bool "Local APIC support on uniprocessors"
-	depends on !SMP && !(X86_VISWS || X86_VOYAGER)
+	depends on !SMP && !(X86_VISWS || X86_VOYAGER || XEN_UNPRIVILEGED_GUEST)
 	help
 	  A local APIC (Advanced Programmable Interrupt Controller) is an
 	  integrated interrupt controller in the CPU. If you have a single-CPU
@@ -288,12 +301,12 @@ config X86_UP_IOAPIC
 
 config X86_LOCAL_APIC
 	bool
-	depends on X86_UP_APIC || ((X86_VISWS || SMP) && !X86_VOYAGER)
+	depends on X86_UP_APIC || ((X86_VISWS || SMP) && !(X86_VOYAGER || XEN_UNPRIVILEGED_GUEST))
 	default y
 
 config X86_IO_APIC
 	bool
-	depends on X86_UP_IOAPIC || (SMP && !(X86_VISWS || X86_VOYAGER))
+	depends on X86_UP_IOAPIC || (SMP && !(X86_VISWS || X86_VOYAGER || XEN_UNPRIVILEGED_GUEST))
 	default y
 
 config X86_VISWS_APIC
@@ -303,7 +316,7 @@ config X86_VISWS_APIC
 
 config X86_MCE
 	bool "Machine Check Exception"
-	depends on !X86_VOYAGER
+	depends on !(X86_VOYAGER || X86_XEN)
 	---help---
 	  Machine Check Exception support allows the processor to notify the
 	  kernel if it detects a problem (e.g. overheating, component failure).
@@ -384,7 +397,7 @@ config I8K
 
 config X86_REBOOTFIXUPS
 	bool "Enable X86 board specific fixups for reboot"
-	depends on X86
+	depends on !X86_XEN
 	default n
 	---help---
 	  This enables chipset and/or board specific fixups to be done
@@ -402,6 +415,7 @@ config X86_REBOOTFIXUPS
 
 config MICROCODE
 	tristate "/dev/cpu/microcode - Intel IA32 CPU microcode support"
+	depends on !XEN_UNPRIVILEGED_GUEST
 	---help---
 	  If you say Y here and also to "/dev file system support" in the
 	  'File systems' section, you will be able to update the microcode on
@@ -434,6 +448,10 @@ config X86_CPUID
 	  with major 203 and minors 0 to 31 for /dev/cpu/0/cpuid to
 	  /dev/cpu/31/cpuid.
 
+config SWIOTLB
+	bool
+	default n
+
 source "drivers/firmware/Kconfig"
 
 choice
@@ -616,6 +634,7 @@ config HIGHPTE
 
 config MATH_EMULATION
 	bool "Math emulation"
+	depends on !X86_XEN
 	---help---
 	  Linux can emulate a math coprocessor (used for floating point
 	  operations) if you don't have one. 486DX and Pentium processors have
@@ -641,6 +660,7 @@ config MATH_EMULATION
 
 config MTRR
 	bool "MTRR (Memory Type Range Register) support"
+	depends on !XEN_UNPRIVILEGED_GUEST
 	---help---
 	  On Intel P6 family processors (Pentium Pro, Pentium II and later)
 	  the Memory Type Range Registers (MTRRs) may be used to control
@@ -675,7 +695,7 @@ config MTRR
 
 config EFI
 	bool "Boot from EFI support"
-	depends on ACPI
+	depends on ACPI && !X86_XEN
 	default n
 	---help---
 	This enables the the kernel to boot on EFI platforms using
@@ -693,7 +713,7 @@ config EFI
 
 config IRQBALANCE
  	bool "Enable kernel irq balancing"
-	depends on SMP && X86_IO_APIC
+	depends on SMP && X86_IO_APIC && !X86_XEN
 	default y
 	help
  	  The default yes will allow the kernel to do irq load balancing.
@@ -741,7 +761,7 @@ source kernel/Kconfig.hz
 
 config KEXEC
 	bool "kexec system call (EXPERIMENTAL)"
-	depends on EXPERIMENTAL
+	depends on EXPERIMENTAL && !XEN_UNPRIVILEGED_GUEST
 	help
 	  kexec is a system call that implements the ability to shutdown your
 	  current kernel, and to start another kernel.  It is like a reboot
@@ -760,6 +780,7 @@ config CRASH_DUMP
 	bool "kernel crash dumps (EXPERIMENTAL)"
 	depends on EXPERIMENTAL
 	depends on HIGHMEM
+	depends on !XEN
 	help
 	  Generate crash dump after being started by kexec.
 
@@ -793,6 +814,7 @@ config HOTPLUG_CPU
 
 config COMPAT_VDSO
 	bool "Compat VDSO support"
+	depends on !X86_XEN
 	default y
 	help
 	  Map the VDSO to the predictable old-style address too.
@@ -810,18 +832,18 @@ config ARCH_ENABLE_MEMORY_HOTPLUG
 	depends on HIGHMEM
 
 menu "Power management options (ACPI, APM)"
-	depends on !X86_VOYAGER
+	depends on !(X86_VOYAGER || XEN_UNPRIVILEGED_GUEST)
 
-source kernel/power/Kconfig
+source "kernel/power/Kconfig"
 
 source "drivers/acpi/Kconfig"
 
 menu "APM (Advanced Power Management) BIOS Support"
-depends on PM && !X86_VISWS
+depends on PM && !(X86_VISWS || X86_XEN)
 
 config APM
 	tristate "APM (Advanced Power Management) BIOS support"
-	depends on PM
+	depends on PM && PM_LEGACY
 	---help---
 	  APM is a BIOS specification for saving power using several different
 	  techniques. This is mostly useful for battery powered laptops with
@@ -1006,6 +1028,7 @@ choice
 
 config PCI_GOBIOS
 	bool "BIOS"
+	depends on !X86_XEN
 
 config PCI_GOMMCONFIG
 	bool "MMConfig"
@@ -1013,6 +1036,13 @@ config PCI_GOMMCONFIG
 config PCI_GODIRECT
 	bool "Direct"
 
+config PCI_GOXEN_FE
+	bool "Xen PCI Frontend"
+	depends on X86_XEN
+	help
+	  The PCI device frontend driver allows the kernel to import arbitrary
+	  PCI devices from a PCI backend to support PCI driver domains.
+
 config PCI_GOANY
 	bool "Any"
 
@@ -1020,7 +1050,7 @@ endchoice
 
 config PCI_BIOS
 	bool
-	depends on !X86_VISWS && PCI && (PCI_GOBIOS || PCI_GOANY)
+	depends on !(X86_VISWS || X86_XEN) && PCI && (PCI_GOBIOS || PCI_GOANY)
 	default y
 
 config PCI_DIRECT
@@ -1033,6 +1063,19 @@ config PCI_MMCONFIG
 	depends on PCI && ACPI && (PCI_GOMMCONFIG || PCI_GOANY)
 	default y
 
+config XEN_PCIDEV_FRONTEND
+	bool
+	depends on PCI && X86_XEN && (PCI_GOXEN_FE || PCI_GOANY)
+	select HOTPLUG
+	default y
+
+config XEN_PCIDEV_FE_DEBUG
+	bool "Xen PCI Frontend Debugging"
+	depends on XEN_PCIDEV_FRONTEND
+	default n
+	help
+	  Enables some debug statements within the PCI Frontend.
+
 source "drivers/pci/pcie/Kconfig"
 
 source "drivers/pci/Kconfig"
@@ -1043,7 +1086,7 @@ config ISA_DMA_API
 
 config ISA
 	bool "ISA support"
-	depends on !(X86_VOYAGER || X86_VISWS)
+	depends on !(X86_VOYAGER || X86_VISWS || X86_XEN)
 	help
 	  Find out whether you have ISA slots on your motherboard.  ISA is the
 	  name of a bus system, i.e. the way the CPU talks to the other stuff
@@ -1070,7 +1113,7 @@ config EISA
 source "drivers/eisa/Kconfig"
 
 config MCA
-	bool "MCA support" if !(X86_VISWS || X86_VOYAGER)
+	bool "MCA support" if !(X86_VISWS || X86_VOYAGER || X86_XEN)
 	default y if X86_VOYAGER
 	help
 	  MicroChannel Architecture is found in some IBM PS/2 machines and
@@ -1146,6 +1189,8 @@ source "security/Kconfig"
 
 source "crypto/Kconfig"
 
+source "drivers/xen/Kconfig"
+
 source "lib/Kconfig"
 
 #
@@ -1171,7 +1216,7 @@ config X86_SMP
 
 config X86_HT
 	bool
-	depends on SMP && !(X86_VISWS || X86_VOYAGER)
+	depends on SMP && !(X86_VISWS || X86_VOYAGER || X86_XEN)
 	default y
 
 config X86_BIOS_REBOOT
@@ -1182,6 +1227,17 @@ config X86_BIOS_REBOOT
 config X86_TRAMPOLINE
 	bool
 	depends on X86_SMP || (X86_VOYAGER && SMP)
+	depends on !XEN
+	default y
+
+config X86_NO_TSS
+	bool
+	depends on X86_XEN
+	default y
+
+config X86_NO_IDT
+	bool
+	depends on X86_XEN
 	default y
 
 config KTIME_SCALAR
diff -rpuN linux-2.6.18.8/arch/i386/Kconfig.cpu linux-2.6.18-xen-3.3.0/arch/i386/Kconfig.cpu
--- linux-2.6.18.8/arch/i386/Kconfig.cpu	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/Kconfig.cpu	2008-08-21 11:36:07.000000000 +0200
@@ -252,7 +252,7 @@ config X86_PPRO_FENCE
 
 config X86_F00F_BUG
 	bool
-	depends on M586MMX || M586TSC || M586 || M486 || M386
+	depends on (M586MMX || M586TSC || M586 || M486 || M386) && !X86_NO_IDT
 	default y
 
 config X86_WP_WORKS_OK
@@ -312,5 +312,5 @@ config X86_OOSTORE
 
 config X86_TSC
 	bool
-	depends on (MWINCHIP3D || MWINCHIP2 || MCRUSOE || MEFFICEON || MCYRIXIII || MK7 || MK6 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || M586MMX || M586TSC || MK8 || MVIAC3_2 || MGEODEGX1 || MGEODE_LX) && !X86_NUMAQ
+	depends on (MWINCHIP3D || MWINCHIP2 || MCRUSOE || MEFFICEON || MCYRIXIII || MK7 || MK6 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || M586MMX || M586TSC || MK8 || MVIAC3_2 || MGEODEGX1 || MGEODE_LX) && !X86_NUMAQ && !X86_XEN
 	default y
diff -rpuN linux-2.6.18.8/arch/i386/Kconfig.debug linux-2.6.18-xen-3.3.0/arch/i386/Kconfig.debug
--- linux-2.6.18.8/arch/i386/Kconfig.debug	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/Kconfig.debug	2008-08-21 11:36:07.000000000 +0200
@@ -79,6 +79,7 @@ config X86_MPPARSE
 config DOUBLEFAULT
 	default y
 	bool "Enable doublefault exception handler" if EMBEDDED
+	depends on !X86_NO_TSS
 	help
           This option allows trapping of rare doublefault exceptions that
           would otherwise cause a system to silently reboot. Disabling this
diff -rpuN linux-2.6.18.8/arch/i386/kernel/acpi/boot-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/acpi/boot-xen.c
--- linux-2.6.18.8/arch/i386/kernel/acpi/boot-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/acpi/boot-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1168 @@
+/*
+ *  boot.c - Architecture-Specific Low-Level ACPI Boot Support
+ *
+ *  Copyright (C) 2001, 2002 Paul Diefenbaugh <paul.s.diefenbaugh@intel.com>
+ *  Copyright (C) 2001 Jun Nakajima <jun.nakajima@intel.com>
+ *
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ */
+
+#include <linux/init.h>
+#include <linux/acpi.h>
+#include <linux/efi.h>
+#include <linux/module.h>
+#include <linux/dmi.h>
+#include <linux/irq.h>
+
+#include <asm/pgtable.h>
+#include <asm/io_apic.h>
+#include <asm/apic.h>
+#include <asm/io.h>
+#include <asm/mpspec.h>
+
+#ifdef	CONFIG_X86_64
+
+extern void __init clustered_apic_check(void);
+
+extern int gsi_irq_sharing(int gsi);
+#include <asm/proto.h>
+
+static inline int acpi_madt_oem_check(char *oem_id, char *oem_table_id) { return 0; }
+
+
+#else				/* X86 */
+
+#ifdef	CONFIG_X86_LOCAL_APIC
+#include <mach_apic.h>
+#include <mach_mpparse.h>
+#endif				/* CONFIG_X86_LOCAL_APIC */
+
+static inline int gsi_irq_sharing(int gsi) { return gsi; }
+
+#endif				/* X86 */
+
+#define BAD_MADT_ENTRY(entry, end) (					    \
+		(!entry) || (unsigned long)entry + sizeof(*entry) > end ||  \
+		((acpi_table_entry_header *)entry)->length < sizeof(*entry))
+
+#define PREFIX			"ACPI: "
+
+int acpi_noirq __initdata;	/* skip ACPI IRQ initialization */
+int acpi_pci_disabled __initdata;	/* skip ACPI PCI scan and IRQ initialization */
+int acpi_ht __initdata = 1;	/* enable HT */
+
+int acpi_lapic;
+int acpi_ioapic;
+int acpi_strict;
+EXPORT_SYMBOL(acpi_strict);
+
+acpi_interrupt_flags acpi_sci_flags __initdata;
+int acpi_sci_override_gsi __initdata;
+int acpi_skip_timer_override __initdata;
+
+#ifdef CONFIG_X86_LOCAL_APIC
+static u64 acpi_lapic_addr __initdata = APIC_DEFAULT_PHYS_BASE;
+#endif
+
+#ifndef __HAVE_ARCH_CMPXCHG
+#warning ACPI uses CMPXCHG, i486 and later hardware
+#endif
+
+#define MAX_MADT_ENTRIES	256
+u8 x86_acpiid_to_apicid[MAX_MADT_ENTRIES] =
+    {[0 ... MAX_MADT_ENTRIES - 1] = 0xff };
+EXPORT_SYMBOL(x86_acpiid_to_apicid);
+
+/* --------------------------------------------------------------------------
+                              Boot-time Configuration
+   -------------------------------------------------------------------------- */
+
+/*
+ * The default interrupt routing model is PIC (8259).  This gets
+ * overriden if IOAPICs are enumerated (below).
+ */
+enum acpi_irq_model_id acpi_irq_model = ACPI_IRQ_MODEL_PIC;
+
+#if defined(CONFIG_X86_64) && !defined(CONFIG_XEN)
+
+/* rely on all ACPI tables being in the direct mapping */
+char *__acpi_map_table(unsigned long phys_addr, unsigned long size)
+{
+	if (!phys_addr || !size)
+		return NULL;
+
+	if (phys_addr+size <= (end_pfn_map << PAGE_SHIFT) + PAGE_SIZE)
+		return __va(phys_addr);
+
+	return NULL;
+}
+
+#else
+
+/*
+ * Temporarily use the virtual area starting from FIX_IO_APIC_BASE_END,
+ * to map the target physical address. The problem is that set_fixmap()
+ * provides a single page, and it is possible that the page is not
+ * sufficient.
+ * By using this area, we can map up to MAX_IO_APICS pages temporarily,
+ * i.e. until the next __va_range() call.
+ *
+ * Important Safety Note:  The fixed I/O APIC page numbers are *subtracted*
+ * from the fixed base.  That's why we start at FIX_IO_APIC_BASE_END and
+ * count idx down while incrementing the phys address.
+ */
+char *__acpi_map_table(unsigned long phys, unsigned long size)
+{
+	unsigned long base, offset, mapped_size;
+	int idx;
+
+#ifndef CONFIG_XEN
+	if (phys + size < 8 * 1024 * 1024)
+		return __va(phys);
+#endif
+
+	offset = phys & (PAGE_SIZE - 1);
+	mapped_size = PAGE_SIZE - offset;
+	set_fixmap(FIX_ACPI_END, phys);
+	base = fix_to_virt(FIX_ACPI_END);
+
+	/*
+	 * Most cases can be covered by the below.
+	 */
+	idx = FIX_ACPI_END;
+	while (mapped_size < size) {
+		if (--idx < FIX_ACPI_BEGIN)
+			return NULL;	/* cannot handle this */
+		phys += PAGE_SIZE;
+		set_fixmap(idx, phys);
+		mapped_size += PAGE_SIZE;
+	}
+
+	return ((unsigned char *)base + offset);
+}
+#endif
+
+#ifdef CONFIG_PCI_MMCONFIG
+/* The physical address of the MMCONFIG aperture.  Set from ACPI tables. */
+struct acpi_table_mcfg_config *pci_mmcfg_config;
+int pci_mmcfg_config_num;
+
+int __init acpi_parse_mcfg(unsigned long phys_addr, unsigned long size)
+{
+	struct acpi_table_mcfg *mcfg;
+	unsigned long i;
+	int config_size;
+
+	if (!phys_addr || !size)
+		return -EINVAL;
+
+	mcfg = (struct acpi_table_mcfg *)__acpi_map_table(phys_addr, size);
+	if (!mcfg) {
+		printk(KERN_WARNING PREFIX "Unable to map MCFG\n");
+		return -ENODEV;
+	}
+
+	/* how many config structures do we have */
+	pci_mmcfg_config_num = 0;
+	i = size - sizeof(struct acpi_table_mcfg);
+	while (i >= sizeof(struct acpi_table_mcfg_config)) {
+		++pci_mmcfg_config_num;
+		i -= sizeof(struct acpi_table_mcfg_config);
+	};
+	if (pci_mmcfg_config_num == 0) {
+		printk(KERN_ERR PREFIX "MMCONFIG has no entries\n");
+		return -ENODEV;
+	}
+
+	config_size = pci_mmcfg_config_num * sizeof(*pci_mmcfg_config);
+	pci_mmcfg_config = kmalloc(config_size, GFP_KERNEL);
+	if (!pci_mmcfg_config) {
+		printk(KERN_WARNING PREFIX
+		       "No memory for MCFG config tables\n");
+		return -ENOMEM;
+	}
+
+	memcpy(pci_mmcfg_config, &mcfg->config, config_size);
+	for (i = 0; i < pci_mmcfg_config_num; ++i) {
+		if (mcfg->config[i].base_reserved) {
+			printk(KERN_ERR PREFIX
+			       "MMCONFIG not in low 4GB of memory\n");
+			kfree(pci_mmcfg_config);
+			pci_mmcfg_config_num = 0;
+			return -ENODEV;
+		}
+	}
+
+	return 0;
+}
+#endif				/* CONFIG_PCI_MMCONFIG */
+
+#ifdef CONFIG_X86_LOCAL_APIC
+static int __init acpi_parse_madt(unsigned long phys_addr, unsigned long size)
+{
+	struct acpi_table_madt *madt = NULL;
+
+	if (!phys_addr || !size || !cpu_has_apic)
+		return -EINVAL;
+
+	madt = (struct acpi_table_madt *)__acpi_map_table(phys_addr, size);
+	if (!madt) {
+		printk(KERN_WARNING PREFIX "Unable to map MADT\n");
+		return -ENODEV;
+	}
+
+	if (madt->lapic_address) {
+		acpi_lapic_addr = (u64) madt->lapic_address;
+
+		printk(KERN_DEBUG PREFIX "Local APIC address 0x%08x\n",
+		       madt->lapic_address);
+	}
+
+	acpi_madt_oem_check(madt->header.oem_id, madt->header.oem_table_id);
+
+	return 0;
+}
+
+static int __init
+acpi_parse_lapic(acpi_table_entry_header * header, const unsigned long end)
+{
+	struct acpi_table_lapic *processor = NULL;
+
+	processor = (struct acpi_table_lapic *)header;
+
+	if (BAD_MADT_ENTRY(processor, end))
+		return -EINVAL;
+
+	acpi_table_print_madt_entry(header);
+
+	/* Record local apic id only when enabled */
+	if (processor->flags.enabled)
+		x86_acpiid_to_apicid[processor->acpi_id] = processor->id;
+
+	/*
+	 * We need to register disabled CPU as well to permit
+	 * counting disabled CPUs. This allows us to size
+	 * cpus_possible_map more accurately, to permit
+	 * to not preallocating memory for all NR_CPUS
+	 * when we use CPU hotplug.
+	 */
+	mp_register_lapic(processor->id,	/* APIC ID */
+			  processor->flags.enabled);	/* Enabled? */
+
+	return 0;
+}
+
+static int __init
+acpi_parse_lapic_addr_ovr(acpi_table_entry_header * header,
+			  const unsigned long end)
+{
+	struct acpi_table_lapic_addr_ovr *lapic_addr_ovr = NULL;
+
+	lapic_addr_ovr = (struct acpi_table_lapic_addr_ovr *)header;
+
+	if (BAD_MADT_ENTRY(lapic_addr_ovr, end))
+		return -EINVAL;
+
+	acpi_lapic_addr = lapic_addr_ovr->address;
+
+	return 0;
+}
+
+static int __init
+acpi_parse_lapic_nmi(acpi_table_entry_header * header, const unsigned long end)
+{
+	struct acpi_table_lapic_nmi *lapic_nmi = NULL;
+
+	lapic_nmi = (struct acpi_table_lapic_nmi *)header;
+
+	if (BAD_MADT_ENTRY(lapic_nmi, end))
+		return -EINVAL;
+
+	acpi_table_print_madt_entry(header);
+
+	if (lapic_nmi->lint != 1)
+		printk(KERN_WARNING PREFIX "NMI not connected to LINT 1!\n");
+
+	return 0;
+}
+
+#endif				/*CONFIG_X86_LOCAL_APIC */
+
+#ifdef CONFIG_X86_IO_APIC
+
+static int __init
+acpi_parse_ioapic(acpi_table_entry_header * header, const unsigned long end)
+{
+	struct acpi_table_ioapic *ioapic = NULL;
+
+	ioapic = (struct acpi_table_ioapic *)header;
+
+	if (BAD_MADT_ENTRY(ioapic, end))
+		return -EINVAL;
+
+	acpi_table_print_madt_entry(header);
+
+	mp_register_ioapic(ioapic->id,
+			   ioapic->address, ioapic->global_irq_base);
+
+	return 0;
+}
+
+/*
+ * Parse Interrupt Source Override for the ACPI SCI
+ */
+static void acpi_sci_ioapic_setup(u32 gsi, u16 polarity, u16 trigger)
+{
+	if (trigger == 0)	/* compatible SCI trigger is level */
+		trigger = 3;
+
+	if (polarity == 0)	/* compatible SCI polarity is low */
+		polarity = 3;
+
+	/* Command-line over-ride via acpi_sci= */
+	if (acpi_sci_flags.trigger)
+		trigger = acpi_sci_flags.trigger;
+
+	if (acpi_sci_flags.polarity)
+		polarity = acpi_sci_flags.polarity;
+
+	/*
+	 * mp_config_acpi_legacy_irqs() already setup IRQs < 16
+	 * If GSI is < 16, this will update its flags,
+	 * else it will create a new mp_irqs[] entry.
+	 */
+	mp_override_legacy_irq(gsi, polarity, trigger, gsi);
+
+	/*
+	 * stash over-ride to indicate we've been here
+	 * and for later update of acpi_fadt
+	 */
+	acpi_sci_override_gsi = gsi;
+	return;
+}
+
+static int __init
+acpi_parse_int_src_ovr(acpi_table_entry_header * header,
+		       const unsigned long end)
+{
+	struct acpi_table_int_src_ovr *intsrc = NULL;
+
+	intsrc = (struct acpi_table_int_src_ovr *)header;
+
+	if (BAD_MADT_ENTRY(intsrc, end))
+		return -EINVAL;
+
+	acpi_table_print_madt_entry(header);
+
+	if (intsrc->bus_irq == acpi_fadt.sci_int) {
+		acpi_sci_ioapic_setup(intsrc->global_irq,
+				      intsrc->flags.polarity,
+				      intsrc->flags.trigger);
+		return 0;
+	}
+
+	if (acpi_skip_timer_override &&
+	    intsrc->bus_irq == 0 && intsrc->global_irq == 2) {
+		printk(PREFIX "BIOS IRQ0 pin2 override ignored.\n");
+		return 0;
+	}
+
+	mp_override_legacy_irq(intsrc->bus_irq,
+			       intsrc->flags.polarity,
+			       intsrc->flags.trigger, intsrc->global_irq);
+
+	return 0;
+}
+
+static int __init
+acpi_parse_nmi_src(acpi_table_entry_header * header, const unsigned long end)
+{
+	struct acpi_table_nmi_src *nmi_src = NULL;
+
+	nmi_src = (struct acpi_table_nmi_src *)header;
+
+	if (BAD_MADT_ENTRY(nmi_src, end))
+		return -EINVAL;
+
+	acpi_table_print_madt_entry(header);
+
+	/* TBD: Support nimsrc entries? */
+
+	return 0;
+}
+
+#endif				/* CONFIG_X86_IO_APIC */
+
+/*
+ * acpi_pic_sci_set_trigger()
+ * 
+ * use ELCR to set PIC-mode trigger type for SCI
+ *
+ * If a PIC-mode SCI is not recognized or gives spurious IRQ7's
+ * it may require Edge Trigger -- use "acpi_sci=edge"
+ *
+ * Port 0x4d0-4d1 are ECLR1 and ECLR2, the Edge/Level Control Registers
+ * for the 8259 PIC.  bit[n] = 1 means irq[n] is Level, otherwise Edge.
+ * ECLR1 is IRQ's 0-7 (IRQ 0, 1, 2 must be 0)
+ * ECLR2 is IRQ's 8-15 (IRQ 8, 13 must be 0)
+ */
+
+void __init acpi_pic_sci_set_trigger(unsigned int irq, u16 trigger)
+{
+	unsigned int mask = 1 << irq;
+	unsigned int old, new;
+
+	/* Real old ELCR mask */
+	old = inb(0x4d0) | (inb(0x4d1) << 8);
+
+	/*
+	 * If we use ACPI to set PCI irq's, then we should clear ELCR
+	 * since we will set it correctly as we enable the PCI irq
+	 * routing.
+	 */
+	new = acpi_noirq ? old : 0;
+
+	/*
+	 * Update SCI information in the ELCR, it isn't in the PCI
+	 * routing tables..
+	 */
+	switch (trigger) {
+	case 1:		/* Edge - clear */
+		new &= ~mask;
+		break;
+	case 3:		/* Level - set */
+		new |= mask;
+		break;
+	}
+
+	if (old == new)
+		return;
+
+	printk(PREFIX "setting ELCR to %04x (from %04x)\n", new, old);
+	outb(new, 0x4d0);
+	outb(new >> 8, 0x4d1);
+}
+
+int acpi_gsi_to_irq(u32 gsi, unsigned int *irq)
+{
+#ifdef CONFIG_X86_IO_APIC
+	if (use_pci_vector() && !platform_legacy_irq(gsi))
+		*irq = IO_APIC_VECTOR(gsi);
+	else
+#endif
+		*irq = gsi_irq_sharing(gsi);
+	return 0;
+}
+
+/*
+ * success: return IRQ number (>=0)
+ * failure: return < 0
+ */
+int acpi_register_gsi(u32 gsi, int triggering, int polarity)
+{
+	unsigned int irq;
+	unsigned int plat_gsi = gsi;
+
+#ifdef CONFIG_PCI
+	/*
+	 * Make sure all (legacy) PCI IRQs are set as level-triggered.
+	 */
+	if (acpi_irq_model == ACPI_IRQ_MODEL_PIC) {
+		extern void eisa_set_level_irq(unsigned int irq);
+
+		if (triggering == ACPI_LEVEL_SENSITIVE)
+			eisa_set_level_irq(gsi);
+	}
+#endif
+
+#ifdef CONFIG_X86_IO_APIC
+	if (acpi_irq_model == ACPI_IRQ_MODEL_IOAPIC) {
+		plat_gsi = mp_register_gsi(gsi, triggering, polarity);
+	}
+#endif
+	acpi_gsi_to_irq(plat_gsi, &irq);
+	return irq;
+}
+
+EXPORT_SYMBOL(acpi_register_gsi);
+
+/*
+ *  ACPI based hotplug support for CPU
+ */
+#ifdef CONFIG_ACPI_HOTPLUG_CPU
+int acpi_map_lsapic(acpi_handle handle, int *pcpu)
+{
+	/* TBD */
+	return -EINVAL;
+}
+
+EXPORT_SYMBOL(acpi_map_lsapic);
+
+int acpi_unmap_lsapic(int cpu)
+{
+	/* TBD */
+	return -EINVAL;
+}
+
+EXPORT_SYMBOL(acpi_unmap_lsapic);
+#endif				/* CONFIG_ACPI_HOTPLUG_CPU */
+
+int acpi_register_ioapic(acpi_handle handle, u64 phys_addr, u32 gsi_base)
+{
+	/* TBD */
+	return -EINVAL;
+}
+
+EXPORT_SYMBOL(acpi_register_ioapic);
+
+int acpi_unregister_ioapic(acpi_handle handle, u32 gsi_base)
+{
+	/* TBD */
+	return -EINVAL;
+}
+
+EXPORT_SYMBOL(acpi_unregister_ioapic);
+
+static unsigned long __init
+acpi_scan_rsdp(unsigned long start, unsigned long length)
+{
+	unsigned long offset = 0;
+	unsigned long sig_len = sizeof("RSD PTR ") - 1;
+	unsigned long vstart = (unsigned long)isa_bus_to_virt(start);
+
+	/*
+	 * Scan all 16-byte boundaries of the physical memory region for the
+	 * RSDP signature.
+	 */
+	for (offset = 0; offset < length; offset += 16) {
+		if (strncmp((char *)(vstart + offset), "RSD PTR ", sig_len))
+			continue;
+		return (start + offset);
+	}
+
+	return 0;
+}
+
+static int __init acpi_parse_sbf(unsigned long phys_addr, unsigned long size)
+{
+	struct acpi_table_sbf *sb;
+
+	if (!phys_addr || !size)
+		return -EINVAL;
+
+	sb = (struct acpi_table_sbf *)__acpi_map_table(phys_addr, size);
+	if (!sb) {
+		printk(KERN_WARNING PREFIX "Unable to map SBF\n");
+		return -ENODEV;
+	}
+
+	sbf_port = sb->sbf_cmos;	/* Save CMOS port */
+
+	return 0;
+}
+
+#ifdef CONFIG_HPET_TIMER
+
+static int __init acpi_parse_hpet(unsigned long phys, unsigned long size)
+{
+	struct acpi_table_hpet *hpet_tbl;
+
+	if (!phys || !size)
+		return -EINVAL;
+
+	hpet_tbl = (struct acpi_table_hpet *)__acpi_map_table(phys, size);
+	if (!hpet_tbl) {
+		printk(KERN_WARNING PREFIX "Unable to map HPET\n");
+		return -ENODEV;
+	}
+
+	if (hpet_tbl->addr.space_id != ACPI_SPACE_MEM) {
+		printk(KERN_WARNING PREFIX "HPET timers must be located in "
+		       "memory.\n");
+		return -1;
+	}
+#ifdef	CONFIG_X86_64
+	vxtime.hpet_address = hpet_tbl->addr.addrl |
+	    ((long)hpet_tbl->addr.addrh << 32);
+
+	printk(KERN_INFO PREFIX "HPET id: %#x base: %#lx\n",
+	       hpet_tbl->id, vxtime.hpet_address);
+#else				/* X86 */
+	{
+		extern unsigned long hpet_address;
+
+		hpet_address = hpet_tbl->addr.addrl;
+		printk(KERN_INFO PREFIX "HPET id: %#x base: %#lx\n",
+		       hpet_tbl->id, hpet_address);
+	}
+#endif				/* X86 */
+
+	return 0;
+}
+#else
+#define	acpi_parse_hpet	NULL
+#endif
+
+#ifdef CONFIG_X86_PM_TIMER
+extern u32 pmtmr_ioport;
+#endif
+
+static int __init acpi_parse_fadt(unsigned long phys, unsigned long size)
+{
+	struct fadt_descriptor *fadt = NULL;
+
+	fadt = (struct fadt_descriptor *)__acpi_map_table(phys, size);
+	if (!fadt) {
+		printk(KERN_WARNING PREFIX "Unable to map FADT\n");
+		return 0;
+	}
+	/* initialize sci_int early for INT_SRC_OVR MADT parsing */
+	acpi_fadt.sci_int = fadt->sci_int;
+
+	/* initialize rev and apic_phys_dest_mode for x86_64 genapic */
+	acpi_fadt.revision = fadt->revision;
+	acpi_fadt.force_apic_physical_destination_mode =
+	    fadt->force_apic_physical_destination_mode;
+
+#if defined(CONFIG_X86_PM_TIMER) && !defined(CONFIG_XEN)
+	/* detect the location of the ACPI PM Timer */
+	if (fadt->revision >= FADT2_REVISION_ID) {
+		/* FADT rev. 2 */
+		if (fadt->xpm_tmr_blk.address_space_id !=
+		    ACPI_ADR_SPACE_SYSTEM_IO)
+			return 0;
+
+		pmtmr_ioport = fadt->xpm_tmr_blk.address;
+		/*
+		 * "X" fields are optional extensions to the original V1.0
+		 * fields, so we must selectively expand V1.0 fields if the
+		 * corresponding X field is zero.
+	 	 */
+		if (!pmtmr_ioport)
+			pmtmr_ioport = fadt->V1_pm_tmr_blk;
+	} else {
+		/* FADT rev. 1 */
+		pmtmr_ioport = fadt->V1_pm_tmr_blk;
+	}
+	if (pmtmr_ioport)
+		printk(KERN_INFO PREFIX "PM-Timer IO Port: %#x\n",
+		       pmtmr_ioport);
+#endif
+	return 0;
+}
+
+unsigned long __init acpi_find_rsdp(void)
+{
+	unsigned long rsdp_phys = 0;
+
+	if (efi_enabled) {
+		if (efi.acpi20 != EFI_INVALID_TABLE_ADDR)
+			return efi.acpi20;
+		else if (efi.acpi != EFI_INVALID_TABLE_ADDR)
+			return efi.acpi;
+	}
+	/*
+	 * Scan memory looking for the RSDP signature. First search EBDA (low
+	 * memory) paragraphs and then search upper memory (E0000-FFFFF).
+	 */
+	rsdp_phys = acpi_scan_rsdp(0, 0x400);
+	if (!rsdp_phys)
+		rsdp_phys = acpi_scan_rsdp(0xE0000, 0x20000);
+
+	return rsdp_phys;
+}
+
+#ifdef	CONFIG_X86_LOCAL_APIC
+/*
+ * Parse LAPIC entries in MADT
+ * returns 0 on success, < 0 on error
+ */
+static int __init acpi_parse_madt_lapic_entries(void)
+{
+	int count;
+
+	if (!cpu_has_apic)
+		return -ENODEV;
+
+	/* 
+	 * Note that the LAPIC address is obtained from the MADT (32-bit value)
+	 * and (optionally) overriden by a LAPIC_ADDR_OVR entry (64-bit value).
+	 */
+
+	count =
+	    acpi_table_parse_madt(ACPI_MADT_LAPIC_ADDR_OVR,
+				  acpi_parse_lapic_addr_ovr, 0);
+	if (count < 0) {
+		printk(KERN_ERR PREFIX
+		       "Error parsing LAPIC address override entry\n");
+		return count;
+	}
+
+	mp_register_lapic_address(acpi_lapic_addr);
+
+	count = acpi_table_parse_madt(ACPI_MADT_LAPIC, acpi_parse_lapic,
+				      MAX_APICS);
+	if (!count) {
+		printk(KERN_ERR PREFIX "No LAPIC entries present\n");
+		/* TBD: Cleanup to allow fallback to MPS */
+		return -ENODEV;
+	} else if (count < 0) {
+		printk(KERN_ERR PREFIX "Error parsing LAPIC entry\n");
+		/* TBD: Cleanup to allow fallback to MPS */
+		return count;
+	}
+
+	count =
+	    acpi_table_parse_madt(ACPI_MADT_LAPIC_NMI, acpi_parse_lapic_nmi, 0);
+	if (count < 0) {
+		printk(KERN_ERR PREFIX "Error parsing LAPIC NMI entry\n");
+		/* TBD: Cleanup to allow fallback to MPS */
+		return count;
+	}
+	return 0;
+}
+#endif				/* CONFIG_X86_LOCAL_APIC */
+
+#ifdef	CONFIG_X86_IO_APIC
+/*
+ * Parse IOAPIC related entries in MADT
+ * returns 0 on success, < 0 on error
+ */
+static int __init acpi_parse_madt_ioapic_entries(void)
+{
+	int count;
+
+	/*
+	 * ACPI interpreter is required to complete interrupt setup,
+	 * so if it is off, don't enumerate the io-apics with ACPI.
+	 * If MPS is present, it will handle them,
+	 * otherwise the system will stay in PIC mode
+	 */
+	if (acpi_disabled || acpi_noirq) {
+		return -ENODEV;
+	}
+
+	if (!cpu_has_apic) 
+		return -ENODEV;
+
+	/*
+	 * if "noapic" boot option, don't look for IO-APICs
+	 */
+	if (skip_ioapic_setup) {
+		printk(KERN_INFO PREFIX "Skipping IOAPIC probe "
+		       "due to 'noapic' option.\n");
+		return -ENODEV;
+	}
+
+	count =
+	    acpi_table_parse_madt(ACPI_MADT_IOAPIC, acpi_parse_ioapic,
+				  MAX_IO_APICS);
+	if (!count) {
+		printk(KERN_ERR PREFIX "No IOAPIC entries present\n");
+		return -ENODEV;
+	} else if (count < 0) {
+		printk(KERN_ERR PREFIX "Error parsing IOAPIC entry\n");
+		return count;
+	}
+
+	count =
+	    acpi_table_parse_madt(ACPI_MADT_INT_SRC_OVR, acpi_parse_int_src_ovr,
+				  NR_IRQ_VECTORS);
+	if (count < 0) {
+		printk(KERN_ERR PREFIX
+		       "Error parsing interrupt source overrides entry\n");
+		/* TBD: Cleanup to allow fallback to MPS */
+		return count;
+	}
+
+	/*
+	 * If BIOS did not supply an INT_SRC_OVR for the SCI
+	 * pretend we got one so we can set the SCI flags.
+	 */
+	if (!acpi_sci_override_gsi)
+		acpi_sci_ioapic_setup(acpi_fadt.sci_int, 0, 0);
+
+	/* Fill in identity legacy mapings where no override */
+	mp_config_acpi_legacy_irqs();
+
+	count =
+	    acpi_table_parse_madt(ACPI_MADT_NMI_SRC, acpi_parse_nmi_src,
+				  NR_IRQ_VECTORS);
+	if (count < 0) {
+		printk(KERN_ERR PREFIX "Error parsing NMI SRC entry\n");
+		/* TBD: Cleanup to allow fallback to MPS */
+		return count;
+	}
+
+	return 0;
+}
+#else
+static inline int acpi_parse_madt_ioapic_entries(void)
+{
+	return -1;
+}
+#endif	/* !CONFIG_X86_IO_APIC */
+
+static void __init acpi_process_madt(void)
+{
+#ifdef CONFIG_X86_LOCAL_APIC
+	int count, error;
+
+	count = acpi_table_parse(ACPI_APIC, acpi_parse_madt);
+	if (count >= 1) {
+
+		/*
+		 * Parse MADT LAPIC entries
+		 */
+		error = acpi_parse_madt_lapic_entries();
+		if (!error) {
+			acpi_lapic = 1;
+
+#ifdef CONFIG_X86_GENERICARCH
+			generic_bigsmp_probe();
+#endif
+			/*
+			 * Parse MADT IO-APIC entries
+			 */
+			error = acpi_parse_madt_ioapic_entries();
+			if (!error) {
+				acpi_irq_model = ACPI_IRQ_MODEL_IOAPIC;
+				acpi_irq_balance_set(NULL);
+				acpi_ioapic = 1;
+
+				smp_found_config = 1;
+				clustered_apic_check();
+			}
+		}
+		if (error == -EINVAL) {
+			/*
+			 * Dell Precision Workstation 410, 610 come here.
+			 */
+			printk(KERN_ERR PREFIX
+			       "Invalid BIOS MADT, disabling ACPI\n");
+			disable_acpi();
+		}
+	}
+#endif
+	return;
+}
+
+extern int acpi_force;
+
+#ifdef __i386__
+
+static int __init disable_acpi_irq(struct dmi_system_id *d)
+{
+	if (!acpi_force) {
+		printk(KERN_NOTICE "%s detected: force use of acpi=noirq\n",
+		       d->ident);
+		acpi_noirq_set();
+	}
+	return 0;
+}
+
+static int __init disable_acpi_pci(struct dmi_system_id *d)
+{
+	if (!acpi_force) {
+		printk(KERN_NOTICE "%s detected: force use of pci=noacpi\n",
+		       d->ident);
+		acpi_disable_pci();
+	}
+	return 0;
+}
+
+static int __init dmi_disable_acpi(struct dmi_system_id *d)
+{
+	if (!acpi_force) {
+		printk(KERN_NOTICE "%s detected: acpi off\n", d->ident);
+		disable_acpi();
+	} else {
+		printk(KERN_NOTICE
+		       "Warning: DMI blacklist says broken, but acpi forced\n");
+	}
+	return 0;
+}
+
+/*
+ * Limit ACPI to CPU enumeration for HT
+ */
+static int __init force_acpi_ht(struct dmi_system_id *d)
+{
+	if (!acpi_force) {
+		printk(KERN_NOTICE "%s detected: force use of acpi=ht\n",
+		       d->ident);
+		disable_acpi();
+		acpi_ht = 1;
+	} else {
+		printk(KERN_NOTICE
+		       "Warning: acpi=force overrules DMI blacklist: acpi=ht\n");
+	}
+	return 0;
+}
+
+/*
+ * If your system is blacklisted here, but you find that acpi=force
+ * works for you, please contact acpi-devel@sourceforge.net
+ */
+static struct dmi_system_id __initdata acpi_dmi_table[] = {
+	/*
+	 * Boxes that need ACPI disabled
+	 */
+	{
+	 .callback = dmi_disable_acpi,
+	 .ident = "IBM Thinkpad",
+	 .matches = {
+		     DMI_MATCH(DMI_BOARD_VENDOR, "IBM"),
+		     DMI_MATCH(DMI_BOARD_NAME, "2629H1G"),
+		     },
+	 },
+
+	/*
+	 * Boxes that need acpi=ht
+	 */
+	{
+	 .callback = force_acpi_ht,
+	 .ident = "FSC Primergy T850",
+	 .matches = {
+		     DMI_MATCH(DMI_SYS_VENDOR, "FUJITSU SIEMENS"),
+		     DMI_MATCH(DMI_PRODUCT_NAME, "PRIMERGY T850"),
+		     },
+	 },
+	{
+	 .callback = force_acpi_ht,
+	 .ident = "DELL GX240",
+	 .matches = {
+		     DMI_MATCH(DMI_BOARD_VENDOR, "Dell Computer Corporation"),
+		     DMI_MATCH(DMI_BOARD_NAME, "OptiPlex GX240"),
+		     },
+	 },
+	{
+	 .callback = force_acpi_ht,
+	 .ident = "HP VISUALIZE NT Workstation",
+	 .matches = {
+		     DMI_MATCH(DMI_BOARD_VENDOR, "Hewlett-Packard"),
+		     DMI_MATCH(DMI_PRODUCT_NAME, "HP VISUALIZE NT Workstation"),
+		     },
+	 },
+	{
+	 .callback = force_acpi_ht,
+	 .ident = "Compaq Workstation W8000",
+	 .matches = {
+		     DMI_MATCH(DMI_SYS_VENDOR, "Compaq"),
+		     DMI_MATCH(DMI_PRODUCT_NAME, "Workstation W8000"),
+		     },
+	 },
+	{
+	 .callback = force_acpi_ht,
+	 .ident = "ASUS P4B266",
+	 .matches = {
+		     DMI_MATCH(DMI_BOARD_VENDOR, "ASUSTeK Computer INC."),
+		     DMI_MATCH(DMI_BOARD_NAME, "P4B266"),
+		     },
+	 },
+	{
+	 .callback = force_acpi_ht,
+	 .ident = "ASUS P2B-DS",
+	 .matches = {
+		     DMI_MATCH(DMI_BOARD_VENDOR, "ASUSTeK Computer INC."),
+		     DMI_MATCH(DMI_BOARD_NAME, "P2B-DS"),
+		     },
+	 },
+	{
+	 .callback = force_acpi_ht,
+	 .ident = "ASUS CUR-DLS",
+	 .matches = {
+		     DMI_MATCH(DMI_BOARD_VENDOR, "ASUSTeK Computer INC."),
+		     DMI_MATCH(DMI_BOARD_NAME, "CUR-DLS"),
+		     },
+	 },
+	{
+	 .callback = force_acpi_ht,
+	 .ident = "ABIT i440BX-W83977",
+	 .matches = {
+		     DMI_MATCH(DMI_BOARD_VENDOR, "ABIT <http://www.abit.com>"),
+		     DMI_MATCH(DMI_BOARD_NAME, "i440BX-W83977 (BP6)"),
+		     },
+	 },
+	{
+	 .callback = force_acpi_ht,
+	 .ident = "IBM Bladecenter",
+	 .matches = {
+		     DMI_MATCH(DMI_BOARD_VENDOR, "IBM"),
+		     DMI_MATCH(DMI_BOARD_NAME, "IBM eServer BladeCenter HS20"),
+		     },
+	 },
+	{
+	 .callback = force_acpi_ht,
+	 .ident = "IBM eServer xSeries 360",
+	 .matches = {
+		     DMI_MATCH(DMI_BOARD_VENDOR, "IBM"),
+		     DMI_MATCH(DMI_BOARD_NAME, "eServer xSeries 360"),
+		     },
+	 },
+	{
+	 .callback = force_acpi_ht,
+	 .ident = "IBM eserver xSeries 330",
+	 .matches = {
+		     DMI_MATCH(DMI_BOARD_VENDOR, "IBM"),
+		     DMI_MATCH(DMI_BOARD_NAME, "eserver xSeries 330"),
+		     },
+	 },
+	{
+	 .callback = force_acpi_ht,
+	 .ident = "IBM eserver xSeries 440",
+	 .matches = {
+		     DMI_MATCH(DMI_BOARD_VENDOR, "IBM"),
+		     DMI_MATCH(DMI_PRODUCT_NAME, "eserver xSeries 440"),
+		     },
+	 },
+
+	/*
+	 * Boxes that need ACPI PCI IRQ routing disabled
+	 */
+	{
+	 .callback = disable_acpi_irq,
+	 .ident = "ASUS A7V",
+	 .matches = {
+		     DMI_MATCH(DMI_BOARD_VENDOR, "ASUSTeK Computer INC"),
+		     DMI_MATCH(DMI_BOARD_NAME, "<A7V>"),
+		     /* newer BIOS, Revision 1011, does work */
+		     DMI_MATCH(DMI_BIOS_VERSION,
+			       "ASUS A7V ACPI BIOS Revision 1007"),
+		     },
+	 },
+
+	/*
+	 * Boxes that need ACPI PCI IRQ routing and PCI scan disabled
+	 */
+	{			/* _BBN 0 bug */
+	 .callback = disable_acpi_pci,
+	 .ident = "ASUS PR-DLS",
+	 .matches = {
+		     DMI_MATCH(DMI_BOARD_VENDOR, "ASUSTeK Computer INC."),
+		     DMI_MATCH(DMI_BOARD_NAME, "PR-DLS"),
+		     DMI_MATCH(DMI_BIOS_VERSION,
+			       "ASUS PR-DLS ACPI BIOS Revision 1010"),
+		     DMI_MATCH(DMI_BIOS_DATE, "03/21/2003")
+		     },
+	 },
+	{
+	 .callback = disable_acpi_pci,
+	 .ident = "Acer TravelMate 36x Laptop",
+	 .matches = {
+		     DMI_MATCH(DMI_SYS_VENDOR, "Acer"),
+		     DMI_MATCH(DMI_PRODUCT_NAME, "TravelMate 360"),
+		     },
+	 },
+	{}
+};
+
+#endif				/* __i386__ */
+
+/*
+ * acpi_boot_table_init() and acpi_boot_init()
+ *  called from setup_arch(), always.
+ *	1. checksums all tables
+ *	2. enumerates lapics
+ *	3. enumerates io-apics
+ *
+ * acpi_table_init() is separate to allow reading SRAT without
+ * other side effects.
+ *
+ * side effects of acpi_boot_init:
+ *	acpi_lapic = 1 if LAPIC found
+ *	acpi_ioapic = 1 if IOAPIC found
+ *	if (acpi_lapic && acpi_ioapic) smp_found_config = 1;
+ *	if acpi_blacklisted() acpi_disabled = 1;
+ *	acpi_irq_model=...
+ *	...
+ *
+ * return value: (currently ignored)
+ *	0: success
+ *	!0: failure
+ */
+
+int __init acpi_boot_table_init(void)
+{
+	int error;
+
+#ifdef __i386__
+	dmi_check_system(acpi_dmi_table);
+#endif
+
+	/*
+	 * If acpi_disabled, bail out
+	 * One exception: acpi=ht continues far enough to enumerate LAPICs
+	 */
+	if (acpi_disabled && !acpi_ht)
+		return 1;
+
+	/* 
+	 * Initialize the ACPI boot-time table parser.
+	 */
+	error = acpi_table_init();
+	if (error) {
+		disable_acpi();
+		return error;
+	}
+
+	acpi_table_parse(ACPI_BOOT, acpi_parse_sbf);
+
+	/*
+	 * blacklist may disable ACPI entirely
+	 */
+	error = acpi_blacklisted();
+	if (error) {
+		if (acpi_force) {
+			printk(KERN_WARNING PREFIX "acpi=force override\n");
+		} else {
+			printk(KERN_WARNING PREFIX "Disabling ACPI support\n");
+			disable_acpi();
+			return error;
+		}
+	}
+
+	return 0;
+}
+
+int __init acpi_boot_init(void)
+{
+	/*
+	 * If acpi_disabled, bail out
+	 * One exception: acpi=ht continues far enough to enumerate LAPICs
+	 */
+	if (acpi_disabled && !acpi_ht)
+		return 1;
+
+	acpi_table_parse(ACPI_BOOT, acpi_parse_sbf);
+
+	/*
+	 * set sci_int and PM timer address
+	 */
+	acpi_table_parse(ACPI_FADT, acpi_parse_fadt);
+
+	/*
+	 * Process the Multiple APIC Description Table (MADT), if present
+	 */
+	acpi_process_madt();
+
+	acpi_table_parse(ACPI_HPET, acpi_parse_hpet);
+
+	return 0;
+}
diff -rpuN linux-2.6.18.8/arch/i386/kernel/acpi/Makefile linux-2.6.18-xen-3.3.0/arch/i386/kernel/acpi/Makefile
--- linux-2.6.18.8/arch/i386/kernel/acpi/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/acpi/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -4,5 +4,9 @@ obj-$(CONFIG_ACPI_SLEEP)	+= sleep.o wake
 
 ifneq ($(CONFIG_ACPI_PROCESSOR),)
 obj-y				+= cstate.o processor.o
+ifneq ($(CONFIG_PROCESSOR_EXTERNAL_CONTROL),)
+obj-$(CONFIG_XEN)		+= processor_extcntl_xen.o
+endif
 endif
 
+disabled-obj-$(CONFIG_XEN)	:= cstate.o wakeup.o
diff -rpuN linux-2.6.18.8/arch/i386/kernel/acpi/processor.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/acpi/processor.c
--- linux-2.6.18.8/arch/i386/kernel/acpi/processor.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/acpi/processor.c	2008-08-21 11:36:07.000000000 +0200
@@ -62,7 +62,18 @@ static void init_intel_pdc(struct acpi_p
 /* Initialize _PDC data based on the CPU vendor */
 void arch_acpi_processor_init_pdc(struct acpi_processor *pr)
 {
+#ifdef CONFIG_XEN
+	/* 
+	 * As a work-around, just use cpu0's cpuinfo for all processors.
+	 * Further work is required to expose xen hypervisor interface of
+	 * getting physical cpuinfo to dom0 kernel and then
+	 * arch_acpi_processor_init_pdc can set _PDC parameters according
+	 * to Xen's phys information.
+	 */
+	unsigned int cpu = 0;
+#else
 	unsigned int cpu = pr->id;
+#endif /* CONFIG_XEN */
 	struct cpuinfo_x86 *c = cpu_data + cpu;
 
 	pr->pdc = NULL;
diff -rpuN linux-2.6.18.8/arch/i386/kernel/acpi/processor_extcntl_xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/acpi/processor_extcntl_xen.c
--- linux-2.6.18.8/arch/i386/kernel/acpi/processor_extcntl_xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/acpi/processor_extcntl_xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,229 @@
+/*
+ * processor_extcntl_xen.c - interface to notify Xen
+ *
+ *  Copyright (C) 2008, Intel corporation
+ *
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or (at
+ *  your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful, but
+ *  WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ *  General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License along
+ *  with this program; if not, write to the Free Software Foundation, Inc.,
+ *  59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
+ *
+ */
+
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/types.h>
+#include <linux/acpi.h>
+#include <linux/pm.h>
+#include <linux/cpu.h>
+
+#include <linux/cpufreq.h>
+#include <acpi/processor.h>
+#include <asm/hypercall.h>
+
+static int xen_processor_pmbits;
+
+static int xen_cx_notifier(struct acpi_processor *pr, int action)
+{
+	int ret, count = 0, i;
+	xen_platform_op_t op = {
+		.cmd			= XENPF_set_processor_pminfo,
+		.interface_version	= XENPF_INTERFACE_VERSION,
+		.u.set_pminfo.id	= pr->acpi_id,
+		.u.set_pminfo.type	= XEN_PM_CX,
+	};
+	struct xen_processor_cx *data, *buf;
+	struct acpi_processor_cx *cx;
+
+	if (action == PROCESSOR_PM_CHANGE)
+		return -EINVAL;
+
+	/* Convert to Xen defined structure and hypercall */
+	buf = kzalloc(pr->power.count * sizeof(struct xen_processor_cx),
+			GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	data = buf;
+	for (i = 1; i <= pr->power.count; i++) {
+		cx = &pr->power.states[i];
+		/* Skip invalid cstate entry */
+		if (!cx->valid)
+			continue;
+
+		data->type = cx->type;
+		data->latency = cx->latency;
+		data->power = cx->power;
+		data->reg.space_id = cx->reg.space_id;
+		data->reg.bit_width = cx->reg.bit_width;
+		data->reg.bit_offset = cx->reg.bit_offset;
+		data->reg.access_size = cx->reg.reserved;
+		data->reg.address = cx->reg.address;
+
+		/* Get dependency relationships */
+		if (cx->csd_count) {
+			printk("Wow! _CSD is found. Not support for now!\n");
+			kfree(buf);
+			return -EINVAL;
+		} else {
+			data->dpcnt = 0;
+			set_xen_guest_handle(data->dp, NULL);
+		}
+
+		data++;
+		count++;
+	}
+
+	if (!count) {
+		printk("No available Cx info for cpu %d\n", pr->acpi_id);
+		kfree(buf);
+		return -EINVAL;
+	}
+
+	op.u.set_pminfo.power.count = count;
+	op.u.set_pminfo.power.flags.bm_control = pr->flags.bm_control;
+	op.u.set_pminfo.power.flags.bm_check = pr->flags.bm_check;
+	op.u.set_pminfo.power.flags.has_cst = pr->flags.has_cst;
+	op.u.set_pminfo.power.flags.power_setup_done = pr->flags.power_setup_done;
+
+	set_xen_guest_handle(op.u.set_pminfo.power.states, buf);
+	ret = HYPERVISOR_platform_op(&op);
+	kfree(buf);
+	return ret;
+}
+
+static void convert_pct_reg(struct xen_pct_register *xpct,
+	struct acpi_pct_register *apct)
+{
+	xpct->descriptor = apct->descriptor;
+	xpct->length     = apct->length;
+	xpct->space_id   = apct->space_id;
+	xpct->bit_width  = apct->bit_width;
+	xpct->bit_offset = apct->bit_offset;
+	xpct->reserved   = apct->reserved;
+	xpct->address    = apct->address;
+}
+
+static void convert_pss_states(struct xen_processor_px *xpss, 
+	struct acpi_processor_px *apss, int state_count)
+{
+	int i;
+	for(i=0; i<state_count; i++) {
+		xpss->core_frequency     = apss->core_frequency;
+		xpss->power              = apss->power;
+		xpss->transition_latency = apss->transition_latency;
+		xpss->bus_master_latency = apss->bus_master_latency;
+		xpss->control            = apss->control;
+		xpss->status             = apss->status;
+		xpss++;
+		apss++;
+	}
+}
+
+static void convert_psd_pack(struct xen_psd_package *xpsd,
+	struct acpi_psd_package *apsd)
+{
+	xpsd->num_entries    = apsd->num_entries;
+	xpsd->revision       = apsd->revision;
+	xpsd->domain         = apsd->domain;
+	xpsd->coord_type     = apsd->coord_type;
+	xpsd->num_processors = apsd->num_processors;
+}
+
+static int xen_px_notifier(struct acpi_processor *pr, int action)
+{
+	int ret;
+	xen_platform_op_t op = {
+		.cmd			= XENPF_set_processor_pminfo,
+		.interface_version	= XENPF_INTERFACE_VERSION,
+		.u.set_pminfo.id	= pr->acpi_id,
+		.u.set_pminfo.type	= XEN_PM_PX,
+	};
+	struct xen_processor_performance *perf;
+	struct xen_processor_px *states = NULL;
+	struct acpi_processor_performance *px;
+	struct acpi_psd_package *pdomain;
+
+	/* leave dynamic ppc handle in the future */
+	if (action == PROCESSOR_PM_CHANGE)
+		return 0;
+
+	perf = &op.u.set_pminfo.perf;
+	px = pr->performance;
+
+	perf->flags = XEN_PX_PPC | 
+		      XEN_PX_PCT | 
+		      XEN_PX_PSS | 
+		      XEN_PX_PSD;
+
+	/* ppc */
+	perf->ppc = pr->performance_platform_limit;
+
+	/* pct */
+	convert_pct_reg(&perf->control_register, &px->control_register);
+	convert_pct_reg(&perf->status_register, &px->status_register);
+
+	/* pss */
+	perf->state_count = px->state_count;
+	states = kzalloc(px->state_count*sizeof(xen_processor_px_t),GFP_KERNEL);
+	if (!states)
+		return -ENOMEM;
+	convert_pss_states(states, px->states, px->state_count);
+	set_xen_guest_handle(perf->states, states);
+
+	/* psd */
+	pdomain = &px->domain_info;
+	convert_psd_pack(&perf->domain_info, pdomain);
+	if (perf->domain_info.num_processors) {
+		if (pdomain->coord_type == DOMAIN_COORD_TYPE_SW_ALL)
+			perf->shared_type = CPUFREQ_SHARED_TYPE_ALL;
+		else if (pdomain->coord_type == DOMAIN_COORD_TYPE_SW_ANY)
+			perf->shared_type = CPUFREQ_SHARED_TYPE_ANY;
+		else if (pdomain->coord_type == DOMAIN_COORD_TYPE_HW_ALL)
+			perf->shared_type = CPUFREQ_SHARED_TYPE_HW;
+	} else
+		perf->shared_type = CPUFREQ_SHARED_TYPE_NONE;
+
+	ret = HYPERVISOR_platform_op(&op);
+	kfree(states);
+	return ret;
+}
+
+static int xen_tx_notifier(struct acpi_processor *pr, int action)
+{
+	return -EINVAL;
+}
+static int xen_hotplug_notifier(struct acpi_processor *pr, int event)
+{
+	return -EINVAL;
+}
+
+static struct processor_extcntl_ops xen_extcntl_ops = {
+	.hotplug		= xen_hotplug_notifier,
+};
+
+void arch_acpi_processor_init_extcntl(const struct processor_extcntl_ops **ops)
+{
+	xen_processor_pmbits = (xen_start_info->flags & SIF_PM_MASK) >> 8;
+
+	if (xen_processor_pmbits & XEN_PROCESSOR_PM_CX)
+		xen_extcntl_ops.pm_ops[PM_TYPE_IDLE] = xen_cx_notifier;
+	if (xen_processor_pmbits & XEN_PROCESSOR_PM_PX)
+		xen_extcntl_ops.pm_ops[PM_TYPE_PERF] = xen_px_notifier;
+	if (xen_processor_pmbits & XEN_PROCESSOR_PM_TX)
+		xen_extcntl_ops.pm_ops[PM_TYPE_THR] = xen_tx_notifier;
+
+	*ops = &xen_extcntl_ops;
+}
+EXPORT_SYMBOL(arch_acpi_processor_init_extcntl);
diff -rpuN linux-2.6.18.8/arch/i386/kernel/acpi/sleep-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/acpi/sleep-xen.c
--- linux-2.6.18.8/arch/i386/kernel/acpi/sleep-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/acpi/sleep-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,113 @@
+/*
+ * sleep.c - x86-specific ACPI sleep support.
+ *
+ *  Copyright (C) 2001-2003 Patrick Mochel
+ *  Copyright (C) 2001-2003 Pavel Machek <pavel@suse.cz>
+ */
+
+#include <linux/acpi.h>
+#include <linux/bootmem.h>
+#include <linux/dmi.h>
+#include <linux/cpumask.h>
+
+#include <asm/smp.h>
+
+#ifndef CONFIG_ACPI_PV_SLEEP
+/* address in low memory of the wakeup routine. */
+unsigned long acpi_wakeup_address = 0;
+unsigned long acpi_video_flags;
+extern char wakeup_start, wakeup_end;
+
+extern unsigned long FASTCALL(acpi_copy_wakeup_routine(unsigned long));
+#endif
+
+/**
+ * acpi_save_state_mem - save kernel state
+ *
+ * Create an identity mapped page table and copy the wakeup routine to
+ * low memory.
+ */
+int acpi_save_state_mem(void)
+{
+#ifndef CONFIG_ACPI_PV_SLEEP
+	if (!acpi_wakeup_address)
+		return 1;
+	memcpy((void *)acpi_wakeup_address, &wakeup_start,
+	       &wakeup_end - &wakeup_start);
+	acpi_copy_wakeup_routine(acpi_wakeup_address);
+#endif
+	return 0;
+}
+
+/*
+ * acpi_restore_state - undo effects of acpi_save_state_mem
+ */
+void acpi_restore_state_mem(void)
+{
+}
+
+/**
+ * acpi_reserve_bootmem - do _very_ early ACPI initialisation
+ *
+ * We allocate a page from the first 1MB of memory for the wakeup
+ * routine for when we come back from a sleep state. The
+ * runtime allocator allows specification of <16MB pages, but not
+ * <1MB pages.
+ */
+void __init acpi_reserve_bootmem(void)
+{
+#ifndef CONFIG_ACPI_PV_SLEEP
+	if ((&wakeup_end - &wakeup_start) > PAGE_SIZE) {
+		printk(KERN_ERR
+		       "ACPI: Wakeup code way too big, S3 disabled.\n");
+		return;
+	}
+
+	acpi_wakeup_address = (unsigned long)alloc_bootmem_low(PAGE_SIZE);
+	if (!acpi_wakeup_address)
+		printk(KERN_ERR "ACPI: Cannot allocate lowmem, S3 disabled.\n");
+#endif
+}
+
+#ifndef CONFIG_ACPI_PV_SLEEP
+static int __init acpi_sleep_setup(char *str)
+{
+	while ((str != NULL) && (*str != '\0')) {
+		if (strncmp(str, "s3_bios", 7) == 0)
+			acpi_video_flags = 1;
+		if (strncmp(str, "s3_mode", 7) == 0)
+			acpi_video_flags |= 2;
+		str = strchr(str, ',');
+		if (str != NULL)
+			str += strspn(str, ", \t");
+	}
+	return 1;
+}
+
+__setup("acpi_sleep=", acpi_sleep_setup);
+
+static __init int reset_videomode_after_s3(struct dmi_system_id *d)
+{
+	acpi_video_flags |= 2;
+	return 0;
+}
+
+static __initdata struct dmi_system_id acpisleep_dmi_table[] = {
+	{			/* Reset video mode after returning from ACPI S3 sleep */
+	 .callback = reset_videomode_after_s3,
+	 .ident = "Toshiba Satellite 4030cdt",
+	 .matches = {
+		     DMI_MATCH(DMI_PRODUCT_NAME, "S4030CDT/4.3"),
+		     },
+	 },
+	{}
+};
+
+static int __init acpisleep_dmi_init(void)
+{
+	dmi_check_system(acpisleep_dmi_table);
+	return 0;
+}
+
+core_initcall(acpisleep_dmi_init);
+#endif /* CONFIG_ACPI_PV_SLEEP */
diff -rpuN linux-2.6.18.8/arch/i386/kernel/apic-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/apic-xen.c
--- linux-2.6.18.8/arch/i386/kernel/apic-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/apic-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,155 @@
+/*
+ *	Local APIC handling, local APIC timers
+ *
+ *	(c) 1999, 2000 Ingo Molnar <mingo@redhat.com>
+ *
+ *	Fixes
+ *	Maciej W. Rozycki	:	Bits for genuine 82489DX APICs;
+ *					thanks to Eric Gilmore
+ *					and Rolf G. Tews
+ *					for testing these extensively.
+ *	Maciej W. Rozycki	:	Various updates and fixes.
+ *	Mikael Pettersson	:	Power Management for UP-APIC.
+ *	Pavel Machek and
+ *	Mikael Pettersson	:	PM converted to driver model.
+ */
+
+#include <linux/init.h>
+
+#include <linux/mm.h>
+#include <linux/delay.h>
+#include <linux/bootmem.h>
+#include <linux/smp_lock.h>
+#include <linux/interrupt.h>
+#include <linux/mc146818rtc.h>
+#include <linux/kernel_stat.h>
+#include <linux/sysdev.h>
+#include <linux/cpu.h>
+#include <linux/module.h>
+
+#include <asm/atomic.h>
+#include <asm/smp.h>
+#include <asm/mtrr.h>
+#include <asm/mpspec.h>
+#include <asm/desc.h>
+#include <asm/arch_hooks.h>
+#include <asm/hpet.h>
+#include <asm/i8253.h>
+#include <asm/nmi.h>
+
+#include <mach_apic.h>
+#include <mach_apicdef.h>
+#include <mach_ipi.h>
+
+#include "io_ports.h"
+
+#ifndef CONFIG_XEN
+/*
+ * cpu_mask that denotes the CPUs that needs timer interrupt coming in as
+ * IPIs in place of local APIC timers
+ */
+static cpumask_t timer_bcast_ipi;
+#endif
+
+/*
+ * Knob to control our willingness to enable the local APIC.
+ */
+int enable_local_apic __initdata = 0; /* -1=force-disable, +1=force-enable */
+
+/*
+ * Debug level
+ */
+int apic_verbosity;
+
+#ifndef CONFIG_XEN
+static int modern_apic(void)
+{
+	unsigned int lvr, version;
+	/* AMD systems use old APIC versions, so check the CPU */
+	if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD &&
+		boot_cpu_data.x86 >= 0xf)
+		return 1;
+	lvr = apic_read(APIC_LVR);
+	version = GET_APIC_VERSION(lvr);
+	return version >= 0x14;
+}
+#endif /* !CONFIG_XEN */
+
+/*
+ * 'what should we do if we get a hw irq event on an illegal vector'.
+ * each architecture has to answer this themselves.
+ */
+void ack_bad_irq(unsigned int irq)
+{
+	printk("unexpected IRQ trap at vector %02x\n", irq);
+	/*
+	 * Currently unexpected vectors happen only on SMP and APIC.
+	 * We _must_ ack these because every local APIC has only N
+	 * irq slots per priority level, and a 'hanging, unacked' IRQ
+	 * holds up an irq slot - in excessive cases (when multiple
+	 * unexpected vectors occur) that might lock up the APIC
+	 * completely.
+	 * But only ack when the APIC is enabled -AK
+	 */
+	if (cpu_has_apic)
+		ack_APIC_irq();
+}
+
+int get_physical_broadcast(void)
+{
+        return 0xff;
+}
+
+#ifndef CONFIG_XEN
+#ifndef CONFIG_SMP
+static void up_apic_timer_interrupt_call(struct pt_regs *regs)
+{
+	int cpu = smp_processor_id();
+
+	/*
+	 * the NMI deadlock-detector uses this.
+	 */
+	per_cpu(irq_stat, cpu).apic_timer_irqs++;
+
+	smp_local_timer_interrupt(regs);
+}
+#endif
+
+void smp_send_timer_broadcast_ipi(struct pt_regs *regs)
+{
+	cpumask_t mask;
+
+	cpus_and(mask, cpu_online_map, timer_bcast_ipi);
+	if (!cpus_empty(mask)) {
+#ifdef CONFIG_SMP
+		send_IPI_mask(mask, LOCAL_TIMER_VECTOR);
+#else
+		/*
+		 * We can directly call the apic timer interrupt handler
+		 * in UP case. Minus all irq related functions
+		 */
+		up_apic_timer_interrupt_call(regs);
+#endif
+	}
+}
+#endif
+
+int setup_profiling_timer(unsigned int multiplier)
+{
+	return -EINVAL;
+}
+
+/*
+ * This initializes the IO-APIC and APIC hardware if this is
+ * a UP kernel.
+ */
+int __init APIC_init_uniprocessor (void)
+{
+#ifdef CONFIG_X86_IO_APIC
+	if (smp_found_config)
+		if (!skip_ioapic_setup && nr_ioapics)
+			setup_IO_APIC();
+#endif
+
+	return 0;
+}
diff -rpuN linux-2.6.18.8/arch/i386/kernel/asm-offsets.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/asm-offsets.c
--- linux-2.6.18.8/arch/i386/kernel/asm-offsets.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/asm-offsets.c	2008-08-21 11:36:07.000000000 +0200
@@ -66,9 +66,14 @@ void foo(void)
 	OFFSET(pbe_orig_address, pbe, orig_address);
 	OFFSET(pbe_next, pbe, next);
 
+#ifndef CONFIG_X86_NO_TSS
 	/* Offset from the sysenter stack to tss.esp0 */
-	DEFINE(TSS_sysenter_esp0, offsetof(struct tss_struct, esp0) -
+	DEFINE(SYSENTER_stack_esp0, offsetof(struct tss_struct, esp0) -
 		 sizeof(struct tss_struct));
+#else
+	/* sysenter stack points directly to esp0 */
+	DEFINE(SYSENTER_stack_esp0, 0);
+#endif
 
 	DEFINE(PAGE_SIZE_asm, PAGE_SIZE);
 	DEFINE(VDSO_PRELINK, VDSO_PRELINK);
diff -rpuN linux-2.6.18.8/arch/i386/kernel/cpu/common-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/cpu/common-xen.c
--- linux-2.6.18.8/arch/i386/kernel/cpu/common-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/cpu/common-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,743 @@
+#include <linux/init.h>
+#include <linux/string.h>
+#include <linux/delay.h>
+#include <linux/smp.h>
+#include <linux/module.h>
+#include <linux/percpu.h>
+#include <linux/bootmem.h>
+#include <asm/semaphore.h>
+#include <asm/processor.h>
+#include <asm/i387.h>
+#include <asm/msr.h>
+#include <asm/io.h>
+#include <asm/mmu_context.h>
+#include <asm/mtrr.h>
+#include <asm/mce.h>
+#ifdef CONFIG_X86_LOCAL_APIC
+#include <asm/mpspec.h>
+#include <asm/apic.h>
+#include <mach_apic.h>
+#else
+#ifdef CONFIG_XEN
+#define phys_pkg_id(a,b) a
+#endif
+#endif
+#include <asm/hypervisor.h>
+
+#include "cpu.h"
+
+DEFINE_PER_CPU(struct Xgt_desc_struct, cpu_gdt_descr);
+EXPORT_PER_CPU_SYMBOL(cpu_gdt_descr);
+
+#ifndef CONFIG_XEN
+DEFINE_PER_CPU(unsigned char, cpu_16bit_stack[CPU_16BIT_STACK_SIZE]);
+EXPORT_PER_CPU_SYMBOL(cpu_16bit_stack);
+#endif
+
+static int cachesize_override __cpuinitdata = -1;
+static int disable_x86_fxsr __cpuinitdata;
+static int disable_x86_serial_nr __cpuinitdata = 1;
+static int disable_x86_sep __cpuinitdata;
+
+struct cpu_dev * cpu_devs[X86_VENDOR_NUM] = {};
+
+extern int disable_pse;
+
+static void default_init(struct cpuinfo_x86 * c)
+{
+	/* Not much we can do here... */
+	/* Check if at least it has cpuid */
+	if (c->cpuid_level == -1) {
+		/* No cpuid. It must be an ancient CPU */
+		if (c->x86 == 4)
+			strcpy(c->x86_model_id, "486");
+		else if (c->x86 == 3)
+			strcpy(c->x86_model_id, "386");
+	}
+}
+
+static struct cpu_dev default_cpu = {
+	.c_init	= default_init,
+	.c_vendor = "Unknown",
+};
+static struct cpu_dev * this_cpu = &default_cpu;
+
+static int __init cachesize_setup(char *str)
+{
+	get_option (&str, &cachesize_override);
+	return 1;
+}
+__setup("cachesize=", cachesize_setup);
+
+int __cpuinit get_model_name(struct cpuinfo_x86 *c)
+{
+	unsigned int *v;
+	char *p, *q;
+
+	if (cpuid_eax(0x80000000) < 0x80000004)
+		return 0;
+
+	v = (unsigned int *) c->x86_model_id;
+	cpuid(0x80000002, &v[0], &v[1], &v[2], &v[3]);
+	cpuid(0x80000003, &v[4], &v[5], &v[6], &v[7]);
+	cpuid(0x80000004, &v[8], &v[9], &v[10], &v[11]);
+	c->x86_model_id[48] = 0;
+
+	/* Intel chips right-justify this string for some dumb reason;
+	   undo that brain damage */
+	p = q = &c->x86_model_id[0];
+	while ( *p == ' ' )
+	     p++;
+	if ( p != q ) {
+	     while ( *p )
+		  *q++ = *p++;
+	     while ( q <= &c->x86_model_id[48] )
+		  *q++ = '\0';	/* Zero-pad the rest */
+	}
+
+	return 1;
+}
+
+
+void __cpuinit display_cacheinfo(struct cpuinfo_x86 *c)
+{
+	unsigned int n, dummy, ecx, edx, l2size;
+
+	n = cpuid_eax(0x80000000);
+
+	if (n >= 0x80000005) {
+		cpuid(0x80000005, &dummy, &dummy, &ecx, &edx);
+		printk(KERN_INFO "CPU: L1 I Cache: %dK (%d bytes/line), D cache %dK (%d bytes/line)\n",
+			edx>>24, edx&0xFF, ecx>>24, ecx&0xFF);
+		c->x86_cache_size=(ecx>>24)+(edx>>24);	
+	}
+
+	if (n < 0x80000006)	/* Some chips just has a large L1. */
+		return;
+
+	ecx = cpuid_ecx(0x80000006);
+	l2size = ecx >> 16;
+	
+	/* do processor-specific cache resizing */
+	if (this_cpu->c_size_cache)
+		l2size = this_cpu->c_size_cache(c,l2size);
+
+	/* Allow user to override all this if necessary. */
+	if (cachesize_override != -1)
+		l2size = cachesize_override;
+
+	if ( l2size == 0 )
+		return;		/* Again, no L2 cache is possible */
+
+	c->x86_cache_size = l2size;
+
+	printk(KERN_INFO "CPU: L2 Cache: %dK (%d bytes/line)\n",
+	       l2size, ecx & 0xFF);
+}
+
+/* Naming convention should be: <Name> [(<Codename>)] */
+/* This table only is used unless init_<vendor>() below doesn't set it; */
+/* in particular, if CPUID levels 0x80000002..4 are supported, this isn't used */
+
+/* Look up CPU names by table lookup. */
+static char __cpuinit *table_lookup_model(struct cpuinfo_x86 *c)
+{
+	struct cpu_model_info *info;
+
+	if ( c->x86_model >= 16 )
+		return NULL;	/* Range check */
+
+	if (!this_cpu)
+		return NULL;
+
+	info = this_cpu->c_models;
+
+	while (info && info->family) {
+		if (info->family == c->x86)
+			return info->model_names[c->x86_model];
+		info++;
+	}
+	return NULL;		/* Not found */
+}
+
+
+static void __cpuinit get_cpu_vendor(struct cpuinfo_x86 *c, int early)
+{
+	char *v = c->x86_vendor_id;
+	int i;
+	static int printed;
+
+	for (i = 0; i < X86_VENDOR_NUM; i++) {
+		if (cpu_devs[i]) {
+			if (!strcmp(v,cpu_devs[i]->c_ident[0]) ||
+			    (cpu_devs[i]->c_ident[1] && 
+			     !strcmp(v,cpu_devs[i]->c_ident[1]))) {
+				c->x86_vendor = i;
+				if (!early)
+					this_cpu = cpu_devs[i];
+				return;
+			}
+		}
+	}
+	if (!printed) {
+		printed++;
+		printk(KERN_ERR "CPU: Vendor unknown, using generic init.\n");
+		printk(KERN_ERR "CPU: Your system may be unstable.\n");
+	}
+	c->x86_vendor = X86_VENDOR_UNKNOWN;
+	this_cpu = &default_cpu;
+}
+
+
+static int __init x86_fxsr_setup(char * s)
+{
+	disable_x86_fxsr = 1;
+	return 1;
+}
+__setup("nofxsr", x86_fxsr_setup);
+
+
+static int __init x86_sep_setup(char * s)
+{
+	disable_x86_sep = 1;
+	return 1;
+}
+__setup("nosep", x86_sep_setup);
+
+
+/* Standard macro to see if a specific flag is changeable */
+static inline int flag_is_changeable_p(u32 flag)
+{
+	u32 f1, f2;
+
+	asm("pushfl\n\t"
+	    "pushfl\n\t"
+	    "popl %0\n\t"
+	    "movl %0,%1\n\t"
+	    "xorl %2,%0\n\t"
+	    "pushl %0\n\t"
+	    "popfl\n\t"
+	    "pushfl\n\t"
+	    "popl %0\n\t"
+	    "popfl\n\t"
+	    : "=&r" (f1), "=&r" (f2)
+	    : "ir" (flag));
+
+	return ((f1^f2) & flag) != 0;
+}
+
+
+/* Probe for the CPUID instruction */
+static int __cpuinit have_cpuid_p(void)
+{
+	return flag_is_changeable_p(X86_EFLAGS_ID);
+}
+
+/* Do minimum CPU detection early.
+   Fields really needed: vendor, cpuid_level, family, model, mask, cache alignment.
+   The others are not touched to avoid unwanted side effects.
+
+   WARNING: this function is only called on the BP.  Don't add code here
+   that is supposed to run on all CPUs. */
+static void __init early_cpu_detect(void)
+{
+	struct cpuinfo_x86 *c = &boot_cpu_data;
+
+	c->x86_cache_alignment = 32;
+
+	if (!have_cpuid_p())
+		return;
+
+	/* Get vendor name */
+	cpuid(0x00000000, &c->cpuid_level,
+	      (int *)&c->x86_vendor_id[0],
+	      (int *)&c->x86_vendor_id[8],
+	      (int *)&c->x86_vendor_id[4]);
+
+	get_cpu_vendor(c, 1);
+
+	c->x86 = 4;
+	if (c->cpuid_level >= 0x00000001) {
+		u32 junk, tfms, cap0, misc;
+		cpuid(0x00000001, &tfms, &misc, &junk, &cap0);
+		c->x86 = (tfms >> 8) & 15;
+		c->x86_model = (tfms >> 4) & 15;
+		if (c->x86 == 0xf)
+			c->x86 += (tfms >> 20) & 0xff;
+		if (c->x86 >= 0x6)
+			c->x86_model += ((tfms >> 16) & 0xF) << 4;
+		c->x86_mask = tfms & 15;
+		if (cap0 & (1<<19))
+			c->x86_cache_alignment = ((misc >> 8) & 0xff) * 8;
+	}
+}
+
+void __cpuinit generic_identify(struct cpuinfo_x86 * c)
+{
+	u32 tfms, xlvl;
+	int ebx;
+
+	if (have_cpuid_p()) {
+		/* Get vendor name */
+		cpuid(0x00000000, &c->cpuid_level,
+		      (int *)&c->x86_vendor_id[0],
+		      (int *)&c->x86_vendor_id[8],
+		      (int *)&c->x86_vendor_id[4]);
+		
+		get_cpu_vendor(c, 0);
+		/* Initialize the standard set of capabilities */
+		/* Note that the vendor-specific code below might override */
+	
+		/* Intel-defined flags: level 0x00000001 */
+		if ( c->cpuid_level >= 0x00000001 ) {
+			u32 capability, excap;
+			cpuid(0x00000001, &tfms, &ebx, &excap, &capability);
+			c->x86_capability[0] = capability;
+			c->x86_capability[4] = excap;
+			c->x86 = (tfms >> 8) & 15;
+			c->x86_model = (tfms >> 4) & 15;
+			if (c->x86 == 0xf)
+				c->x86 += (tfms >> 20) & 0xff;
+			if (c->x86 >= 0x6)
+				c->x86_model += ((tfms >> 16) & 0xF) << 4;
+			c->x86_mask = tfms & 15;
+#ifdef CONFIG_X86_HT
+			c->apicid = phys_pkg_id((ebx >> 24) & 0xFF, 0);
+#else
+			c->apicid = (ebx >> 24) & 0xFF;
+#endif
+		} else {
+			/* Have CPUID level 0 only - unheard of */
+			c->x86 = 4;
+		}
+
+		/* AMD-defined flags: level 0x80000001 */
+		xlvl = cpuid_eax(0x80000000);
+		if ( (xlvl & 0xffff0000) == 0x80000000 ) {
+			if ( xlvl >= 0x80000001 ) {
+				c->x86_capability[1] = cpuid_edx(0x80000001);
+				c->x86_capability[6] = cpuid_ecx(0x80000001);
+			}
+			if ( xlvl >= 0x80000004 )
+				get_model_name(c); /* Default name */
+		}
+	}
+
+	early_intel_workaround(c);
+
+#ifdef CONFIG_X86_HT
+	c->phys_proc_id = (cpuid_ebx(1) >> 24) & 0xff;
+#endif
+}
+
+static void __cpuinit squash_the_stupid_serial_number(struct cpuinfo_x86 *c)
+{
+	if (cpu_has(c, X86_FEATURE_PN) && disable_x86_serial_nr ) {
+		/* Disable processor serial number */
+		unsigned long lo,hi;
+		rdmsr(MSR_IA32_BBL_CR_CTL,lo,hi);
+		lo |= 0x200000;
+		wrmsr(MSR_IA32_BBL_CR_CTL,lo,hi);
+		printk(KERN_NOTICE "CPU serial number disabled.\n");
+		clear_bit(X86_FEATURE_PN, c->x86_capability);
+
+		/* Disabling the serial number may affect the cpuid level */
+		c->cpuid_level = cpuid_eax(0);
+	}
+}
+
+static int __init x86_serial_nr_setup(char *s)
+{
+	disable_x86_serial_nr = 0;
+	return 1;
+}
+__setup("serialnumber", x86_serial_nr_setup);
+
+
+
+/*
+ * This does the hard work of actually picking apart the CPU stuff...
+ */
+void __cpuinit identify_cpu(struct cpuinfo_x86 *c)
+{
+	int i;
+
+	c->loops_per_jiffy = loops_per_jiffy;
+	c->x86_cache_size = -1;
+	c->x86_vendor = X86_VENDOR_UNKNOWN;
+	c->cpuid_level = -1;	/* CPUID not detected */
+	c->x86_model = c->x86_mask = 0;	/* So far unknown... */
+	c->x86_vendor_id[0] = '\0'; /* Unset */
+	c->x86_model_id[0] = '\0';  /* Unset */
+	c->x86_max_cores = 1;
+	memset(&c->x86_capability, 0, sizeof c->x86_capability);
+
+	if (!have_cpuid_p()) {
+		/* First of all, decide if this is a 486 or higher */
+		/* It's a 486 if we can modify the AC flag */
+		if ( flag_is_changeable_p(X86_EFLAGS_AC) )
+			c->x86 = 4;
+		else
+			c->x86 = 3;
+	}
+
+	generic_identify(c);
+
+	printk(KERN_DEBUG "CPU: After generic identify, caps:");
+	for (i = 0; i < NCAPINTS; i++)
+		printk(" %08lx", c->x86_capability[i]);
+	printk("\n");
+
+	if (this_cpu->c_identify) {
+		this_cpu->c_identify(c);
+
+		printk(KERN_DEBUG "CPU: After vendor identify, caps:");
+		for (i = 0; i < NCAPINTS; i++)
+			printk(" %08lx", c->x86_capability[i]);
+		printk("\n");
+	}
+
+	/*
+	 * Vendor-specific initialization.  In this section we
+	 * canonicalize the feature flags, meaning if there are
+	 * features a certain CPU supports which CPUID doesn't
+	 * tell us, CPUID claiming incorrect flags, or other bugs,
+	 * we handle them here.
+	 *
+	 * At the end of this section, c->x86_capability better
+	 * indicate the features this CPU genuinely supports!
+	 */
+	if (this_cpu->c_init)
+		this_cpu->c_init(c);
+
+	/* Disable the PN if appropriate */
+	squash_the_stupid_serial_number(c);
+
+	/*
+	 * The vendor-specific functions might have changed features.  Now
+	 * we do "generic changes."
+	 */
+
+	/* TSC disabled? */
+	if ( tsc_disable )
+		clear_bit(X86_FEATURE_TSC, c->x86_capability);
+
+	/* FXSR disabled? */
+	if (disable_x86_fxsr) {
+		clear_bit(X86_FEATURE_FXSR, c->x86_capability);
+		clear_bit(X86_FEATURE_XMM, c->x86_capability);
+	}
+
+	/* SEP disabled? */
+	if (disable_x86_sep)
+		clear_bit(X86_FEATURE_SEP, c->x86_capability);
+
+	if (disable_pse)
+		clear_bit(X86_FEATURE_PSE, c->x86_capability);
+
+	/* If the model name is still unset, do table lookup. */
+	if ( !c->x86_model_id[0] ) {
+		char *p;
+		p = table_lookup_model(c);
+		if ( p )
+			strcpy(c->x86_model_id, p);
+		else
+			/* Last resort... */
+			sprintf(c->x86_model_id, "%02x/%02x",
+				c->x86, c->x86_model);
+	}
+
+	/* Now the feature flags better reflect actual CPU features! */
+
+	printk(KERN_DEBUG "CPU: After all inits, caps:");
+	for (i = 0; i < NCAPINTS; i++)
+		printk(" %08lx", c->x86_capability[i]);
+	printk("\n");
+
+	/*
+	 * On SMP, boot_cpu_data holds the common feature set between
+	 * all CPUs; so make sure that we indicate which features are
+	 * common between the CPUs.  The first time this routine gets
+	 * executed, c == &boot_cpu_data.
+	 */
+	if ( c != &boot_cpu_data ) {
+		/* AND the already accumulated flags with these */
+		for ( i = 0 ; i < NCAPINTS ; i++ )
+			boot_cpu_data.x86_capability[i] &= c->x86_capability[i];
+	}
+
+	/* Init Machine Check Exception if available. */
+	mcheck_init(c);
+
+	if (c == &boot_cpu_data)
+		sysenter_setup();
+	enable_sep_cpu();
+
+	if (c == &boot_cpu_data)
+		mtrr_bp_init();
+	else
+		mtrr_ap_init();
+}
+
+#ifdef CONFIG_X86_HT
+void __cpuinit detect_ht(struct cpuinfo_x86 *c)
+{
+	u32 	eax, ebx, ecx, edx;
+	int 	index_msb, core_bits;
+
+	cpuid(1, &eax, &ebx, &ecx, &edx);
+
+	if (!cpu_has(c, X86_FEATURE_HT) || cpu_has(c, X86_FEATURE_CMP_LEGACY))
+		return;
+
+	smp_num_siblings = (ebx & 0xff0000) >> 16;
+
+	if (smp_num_siblings == 1) {
+		printk(KERN_INFO  "CPU: Hyper-Threading is disabled\n");
+	} else if (smp_num_siblings > 1 ) {
+
+		if (smp_num_siblings > NR_CPUS) {
+			printk(KERN_WARNING "CPU: Unsupported number of the "
+					"siblings %d", smp_num_siblings);
+			smp_num_siblings = 1;
+			return;
+		}
+
+		index_msb = get_count_order(smp_num_siblings);
+		c->phys_proc_id = phys_pkg_id((ebx >> 24) & 0xFF, index_msb);
+
+		printk(KERN_INFO  "CPU: Physical Processor ID: %d\n",
+		       c->phys_proc_id);
+
+		smp_num_siblings = smp_num_siblings / c->x86_max_cores;
+
+		index_msb = get_count_order(smp_num_siblings) ;
+
+		core_bits = get_count_order(c->x86_max_cores);
+
+		c->cpu_core_id = phys_pkg_id((ebx >> 24) & 0xFF, index_msb) &
+					       ((1 << core_bits) - 1);
+
+		if (c->x86_max_cores > 1)
+			printk(KERN_INFO  "CPU: Processor Core ID: %d\n",
+			       c->cpu_core_id);
+	}
+}
+#endif
+
+void __cpuinit print_cpu_info(struct cpuinfo_x86 *c)
+{
+	char *vendor = NULL;
+
+	if (c->x86_vendor < X86_VENDOR_NUM)
+		vendor = this_cpu->c_vendor;
+	else if (c->cpuid_level >= 0)
+		vendor = c->x86_vendor_id;
+
+	if (vendor && strncmp(c->x86_model_id, vendor, strlen(vendor)))
+		printk("%s ", vendor);
+
+	if (!c->x86_model_id[0])
+		printk("%d86", c->x86);
+	else
+		printk("%s", c->x86_model_id);
+
+	if (c->x86_mask || c->cpuid_level >= 0) 
+		printk(" stepping %02x\n", c->x86_mask);
+	else
+		printk("\n");
+}
+
+cpumask_t cpu_initialized __cpuinitdata = CPU_MASK_NONE;
+
+/* This is hacky. :)
+ * We're emulating future behavior.
+ * In the future, the cpu-specific init functions will be called implicitly
+ * via the magic of initcalls.
+ * They will insert themselves into the cpu_devs structure.
+ * Then, when cpu_init() is called, we can just iterate over that array.
+ */
+
+extern int intel_cpu_init(void);
+extern int cyrix_init_cpu(void);
+extern int nsc_init_cpu(void);
+extern int amd_init_cpu(void);
+extern int centaur_init_cpu(void);
+extern int transmeta_init_cpu(void);
+extern int rise_init_cpu(void);
+extern int nexgen_init_cpu(void);
+extern int umc_init_cpu(void);
+
+void __init early_cpu_init(void)
+{
+	intel_cpu_init();
+	cyrix_init_cpu();
+	nsc_init_cpu();
+	amd_init_cpu();
+	centaur_init_cpu();
+	transmeta_init_cpu();
+	rise_init_cpu();
+	nexgen_init_cpu();
+	umc_init_cpu();
+	early_cpu_detect();
+
+#ifdef CONFIG_DEBUG_PAGEALLOC
+	/* pse is not compatible with on-the-fly unmapping,
+	 * disable it even if the cpus claim to support it.
+	 */
+	clear_bit(X86_FEATURE_PSE, boot_cpu_data.x86_capability);
+	disable_pse = 1;
+#endif
+}
+
+static void __cpuinit cpu_gdt_init(const struct Xgt_desc_struct *gdt_descr)
+{
+	unsigned long frames[16];
+	unsigned long va;
+	int f;
+
+	for (va = gdt_descr->address, f = 0;
+	     va < gdt_descr->address + gdt_descr->size;
+	     va += PAGE_SIZE, f++) {
+		frames[f] = virt_to_mfn(va);
+		make_lowmem_page_readonly(
+			(void *)va, XENFEAT_writable_descriptor_tables);
+	}
+	if (HYPERVISOR_set_gdt(frames, (gdt_descr->size + 1) / 8))
+		BUG();
+}
+
+/*
+ * cpu_init() initializes state that is per-CPU. Some data is already
+ * initialized (naturally) in the bootstrap process, such as the GDT
+ * and IDT. We reload them nevertheless, this function acts as a
+ * 'CPU state barrier', nothing should get across.
+ */
+void __cpuinit cpu_init(void)
+{
+	int cpu = smp_processor_id();
+#ifndef CONFIG_X86_NO_TSS
+	struct tss_struct * t = &per_cpu(init_tss, cpu);
+#endif
+	struct thread_struct *thread = &current->thread;
+	struct desc_struct *gdt;
+	struct Xgt_desc_struct *cpu_gdt_descr = &per_cpu(cpu_gdt_descr, cpu);
+
+	if (cpu_test_and_set(cpu, cpu_initialized)) {
+		printk(KERN_WARNING "CPU#%d already initialized!\n", cpu);
+		for (;;) local_irq_enable();
+	}
+	printk(KERN_INFO "Initializing CPU#%d\n", cpu);
+
+	if (cpu_has_vme || cpu_has_de)
+		clear_in_cr4(X86_CR4_VME|X86_CR4_PVI|X86_CR4_TSD|X86_CR4_DE);
+	if (tsc_disable && cpu_has_tsc) {
+		printk(KERN_NOTICE "Disabling TSC...\n");
+		/**** FIX-HPA: DOES THIS REALLY BELONG HERE? ****/
+		clear_bit(X86_FEATURE_TSC, boot_cpu_data.x86_capability);
+		set_in_cr4(X86_CR4_TSD);
+	}
+
+#ifndef CONFIG_XEN
+	/* The CPU hotplug case */
+	if (cpu_gdt_descr->address) {
+		gdt = (struct desc_struct *)cpu_gdt_descr->address;
+		memset(gdt, 0, PAGE_SIZE);
+		goto old_gdt;
+	}
+	/*
+	 * This is a horrible hack to allocate the GDT.  The problem
+	 * is that cpu_init() is called really early for the boot CPU
+	 * (and hence needs bootmem) but much later for the secondary
+	 * CPUs, when bootmem will have gone away
+	 */
+	if (NODE_DATA(0)->bdata->node_bootmem_map) {
+		gdt = (struct desc_struct *)alloc_bootmem_pages(PAGE_SIZE);
+		/* alloc_bootmem_pages panics on failure, so no check */
+		memset(gdt, 0, PAGE_SIZE);
+	} else {
+		gdt = (struct desc_struct *)get_zeroed_page(GFP_KERNEL);
+		if (unlikely(!gdt)) {
+			printk(KERN_CRIT "CPU%d failed to allocate GDT\n", cpu);
+			for (;;)
+				local_irq_enable();
+		}
+	}
+old_gdt:
+	/*
+	 * Initialize the per-CPU GDT with the boot GDT,
+	 * and set up the GDT descriptor:
+	 */
+ 	memcpy(gdt, cpu_gdt_table, GDT_SIZE);
+
+	/* Set up GDT entry for 16bit stack */
+ 	*(__u64 *)(&gdt[GDT_ENTRY_ESPFIX_SS]) |=
+		((((__u64)stk16_off) << 16) & 0x000000ffffff0000ULL) |
+		((((__u64)stk16_off) << 32) & 0xff00000000000000ULL) |
+		(CPU_16BIT_STACK_SIZE - 1);
+
+	cpu_gdt_descr->size = GDT_SIZE - 1;
+ 	cpu_gdt_descr->address = (unsigned long)gdt;
+#else
+	if (cpu == 0 && cpu_gdt_descr->address == 0) {
+		gdt = (struct desc_struct *)alloc_bootmem_pages(PAGE_SIZE);
+		/* alloc_bootmem_pages panics on failure, so no check */
+		memset(gdt, 0, PAGE_SIZE);
+
+		memcpy(gdt, cpu_gdt_table, GDT_SIZE);
+		
+		cpu_gdt_descr->size = GDT_SIZE;
+		cpu_gdt_descr->address = (unsigned long)gdt;
+	}
+#endif
+
+	cpu_gdt_init(cpu_gdt_descr);
+
+	/*
+	 * Set up and load the per-CPU TSS and LDT
+	 */
+	atomic_inc(&init_mm.mm_count);
+	current->active_mm = &init_mm;
+	if (current->mm)
+		BUG();
+	enter_lazy_tlb(&init_mm, current);
+
+	load_esp0(t, thread);
+
+	load_LDT(&init_mm.context);
+
+#ifdef CONFIG_DOUBLEFAULT
+	/* Set up doublefault TSS pointer in the GDT */
+	__set_tss_desc(cpu, GDT_ENTRY_DOUBLEFAULT_TSS, &doublefault_tss);
+#endif
+
+	/* Clear %fs and %gs. */
+	asm volatile ("xorl %eax, %eax; movl %eax, %fs; movl %eax, %gs");
+
+	/* Clear all 6 debug registers: */
+	set_debugreg(0, 0);
+	set_debugreg(0, 1);
+	set_debugreg(0, 2);
+	set_debugreg(0, 3);
+	set_debugreg(0, 6);
+	set_debugreg(0, 7);
+
+	/*
+	 * Force FPU initialization:
+	 */
+	current_thread_info()->status = 0;
+	clear_used_math();
+	mxcsr_feature_mask_init();
+}
+
+#ifdef CONFIG_HOTPLUG_CPU
+void __cpuinit cpu_uninit(void)
+{
+	int cpu = raw_smp_processor_id();
+	cpu_clear(cpu, cpu_initialized);
+
+	/* lazy TLB state */
+	per_cpu(cpu_tlbstate, cpu).state = 0;
+	per_cpu(cpu_tlbstate, cpu).active_mm = &init_mm;
+}
+#endif
diff -rpuN linux-2.6.18.8/arch/i386/kernel/cpu/cpufreq/powernow-k8.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/cpu/cpufreq/powernow-k8.c
--- linux-2.6.18.8/arch/i386/kernel/cpu/cpufreq/powernow-k8.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/cpu/cpufreq/powernow-k8.c	2008-08-21 11:36:07.000000000 +0200
@@ -46,7 +46,7 @@
 
 #define PFX "powernow-k8: "
 #define BFX PFX "BIOS error: "
-#define VERSION "version 2.00.00"
+#define VERSION "version 2.20.00"
 #include "powernow-k8.h"
 
 /* serialize freq changes  */
@@ -66,36 +66,15 @@ static u32 find_freq_from_fid(u32 fid)
 	return 800 + (fid * 100);
 }
 
-
 /* Return a frequency in KHz, given an input fid */
 static u32 find_khz_freq_from_fid(u32 fid)
 {
 	return 1000 * find_freq_from_fid(fid);
 }
 
-/* Return a frequency in MHz, given an input fid and did */
-static u32 find_freq_from_fiddid(u32 fid, u32 did)
-{
-	return 100 * (fid + 0x10) >> did;
-}
-
-static u32 find_khz_freq_from_fiddid(u32 fid, u32 did)
-{
-	return 1000 * find_freq_from_fiddid(fid, did);
-}
-
-static u32 find_fid_from_pstate(u32 pstate)
-{
-	u32 hi, lo;
-	rdmsr(MSR_PSTATE_DEF_BASE + pstate, lo, hi);
-	return lo & HW_PSTATE_FID_MASK;
-}
-
-static u32 find_did_from_pstate(u32 pstate)
+static u32 find_khz_freq_from_pstate(struct cpufreq_frequency_table *data, u32 pstate)
 {
-	u32 hi, lo;
-	rdmsr(MSR_PSTATE_DEF_BASE + pstate, lo, hi);
-	return (lo & HW_PSTATE_DID_MASK) >> HW_PSTATE_DID_SHIFT;
+	return data[pstate].frequency;
 }
 
 /* Return the vco fid for an input fid
@@ -139,9 +118,7 @@ static int query_current_values_with_pen
 	if (cpu_family == CPU_HW_PSTATE) {
 		rdmsr(MSR_PSTATE_STATUS, lo, hi);
 		i = lo & HW_PSTATE_MASK;
-		rdmsr(MSR_PSTATE_DEF_BASE + i, lo, hi);
-		data->currfid = lo & HW_PSTATE_FID_MASK;
-		data->currdid = (lo & HW_PSTATE_DID_MASK) >> HW_PSTATE_DID_SHIFT;
+		data->currpstate = i;
 		return 0;
 	}
 	do {
@@ -292,7 +269,7 @@ static int decrease_vid_code_by_step(str
 static int transition_pstate(struct powernow_k8_data *data, u32 pstate)
 {
 	wrmsr(MSR_PSTATE_CTRL, pstate, 0);
-	data->currfid = find_fid_from_pstate(pstate);
+	data->currpstate = pstate;
 	return 0;
 }
 
@@ -738,6 +715,7 @@ static int find_psb_table(struct powerno
 
 		data->numps = psb->numps;
 		dprintk("numpstates: 0x%x\n", data->numps);
+		data->starting_core_affinity = cpumask_of_cpu(0);
 		return fill_powernow_table(data, (struct pst_s *)(psb+1), maxvid);
 	}
 	/*
@@ -758,15 +736,43 @@ static int find_psb_table(struct powerno
 #ifdef CONFIG_X86_POWERNOW_K8_ACPI
 static void powernow_k8_acpi_pst_values(struct powernow_k8_data *data, unsigned int index)
 {
-	if (!data->acpi_data.state_count || (cpu_family == CPU_HW_PSTATE))
+	if (!data->acpi_data->state_count || (cpu_family == CPU_HW_PSTATE))
 		return;
 
-	data->irt = (data->acpi_data.states[index].control >> IRT_SHIFT) & IRT_MASK;
-	data->rvo = (data->acpi_data.states[index].control >> RVO_SHIFT) & RVO_MASK;
-	data->exttype = (data->acpi_data.states[index].control >> EXT_TYPE_SHIFT) & EXT_TYPE_MASK;
-	data->plllock = (data->acpi_data.states[index].control >> PLL_L_SHIFT) & PLL_L_MASK;
-	data->vidmvs = 1 << ((data->acpi_data.states[index].control >> MVS_SHIFT) & MVS_MASK);
-	data->vstable = (data->acpi_data.states[index].control >> VST_SHIFT) & VST_MASK;
+	data->irt = (data->acpi_data->states[index].control >> IRT_SHIFT) & IRT_MASK;
+	data->rvo = (data->acpi_data->states[index].control >> RVO_SHIFT) & RVO_MASK;
+	data->exttype = (data->acpi_data->states[index].control >> EXT_TYPE_SHIFT) & EXT_TYPE_MASK;
+	data->plllock = (data->acpi_data->states[index].control >> PLL_L_SHIFT) & PLL_L_MASK;
+	data->vidmvs = 1 << ((data->acpi_data->states[index].control >> MVS_SHIFT) & MVS_MASK);
+	data->vstable = (data->acpi_data->states[index].control >> VST_SHIFT) & VST_MASK;
+}
+
+static struct acpi_processor_performance *acpi_perf_data[NR_CPUS];
+static int preregister_valid = 0;
+
+static int powernow_k8_cpu_preinit_acpi()
+{
+	int i; 
+	struct acpi_processor_performance *data;
+	for_each_possible_cpu(i) {
+		data = kzalloc(sizeof(struct acpi_processor_performance),
+				GFP_KERNEL);
+		if (!data) {
+			int j;
+			for_each_possible_cpu(j) {
+				kfree(acpi_perf_data[j]);
+				acpi_perf_data[j] = NULL;
+			}
+			return -ENODEV;
+		}
+		acpi_perf_data[i] = data;
+	}
+
+	if (acpi_processor_preregister_performance(acpi_perf_data))
+		return -ENODEV;
+	else 
+		preregister_valid = 1;
+	return 0;
 }
 
 static int powernow_k8_cpu_init_acpi(struct powernow_k8_data *data)
@@ -774,28 +780,29 @@ static int powernow_k8_cpu_init_acpi(str
 	struct cpufreq_frequency_table *powernow_table;
 	int ret_val;
 
-	if (acpi_processor_register_performance(&data->acpi_data, data->cpu)) {
+	data->acpi_data = acpi_perf_data[data->cpu];
+	if (acpi_processor_register_performance(data->acpi_data, data->cpu)) {
 		dprintk("register performance failed: bad ACPI data\n");
 		return -EIO;
 	}
 
 	/* verify the data contained in the ACPI structures */
-	if (data->acpi_data.state_count <= 1) {
+	if (data->acpi_data->state_count <= 1) {
 		dprintk("No ACPI P-States\n");
 		goto err_out;
 	}
 
-	if ((data->acpi_data.control_register.space_id != ACPI_ADR_SPACE_FIXED_HARDWARE) ||
-		(data->acpi_data.status_register.space_id != ACPI_ADR_SPACE_FIXED_HARDWARE)) {
+	if ((data->acpi_data->control_register.space_id != ACPI_ADR_SPACE_FIXED_HARDWARE) ||
+		(data->acpi_data->status_register.space_id != ACPI_ADR_SPACE_FIXED_HARDWARE)) {
 		dprintk("Invalid control/status registers (%x - %x)\n",
-			data->acpi_data.control_register.space_id,
-			data->acpi_data.status_register.space_id);
+			data->acpi_data->control_register.space_id,
+			data->acpi_data->status_register.space_id);
 		goto err_out;
 	}
 
 	/* fill in data->powernow_table */
 	powernow_table = kmalloc((sizeof(struct cpufreq_frequency_table)
-		* (data->acpi_data.state_count + 1)), GFP_KERNEL);
+		* (data->acpi_data->state_count + 1)), GFP_KERNEL);
 	if (!powernow_table) {
 		dprintk("powernow_table memory alloc failure\n");
 		goto err_out;
@@ -808,28 +815,43 @@ static int powernow_k8_cpu_init_acpi(str
 	if (ret_val)
 		goto err_out_mem;
 
-	powernow_table[data->acpi_data.state_count].frequency = CPUFREQ_TABLE_END;
-	powernow_table[data->acpi_data.state_count].index = 0;
+	powernow_table[data->acpi_data->state_count].frequency = CPUFREQ_TABLE_END;
+	powernow_table[data->acpi_data->state_count].index = 0;
 	data->powernow_table = powernow_table;
 
 	/* fill in data */
-	data->numps = data->acpi_data.state_count;
+	data->numps = data->acpi_data->state_count;
 	print_basics(data);
 	powernow_k8_acpi_pst_values(data, 0);
 
 	/* notify BIOS that we exist */
 	acpi_processor_notify_smm(THIS_MODULE);
 
+	/* determine affinity, from ACPI if available */
+	if (preregister_valid) {
+		if ((data->acpi_data->shared_type == CPUFREQ_SHARED_TYPE_ALL) ||
+		    (data->acpi_data->shared_type == CPUFREQ_SHARED_TYPE_ANY))
+			data->starting_core_affinity = data->acpi_data->shared_cpu_map;
+		else
+			data->starting_core_affinity = cpumask_of_cpu(data->cpu);
+	} else {
+		/* best guess from family if not */
+		if (cpu_family == CPU_HW_PSTATE)
+			data->starting_core_affinity = cpumask_of_cpu(data->cpu);
+		else
+			data->starting_core_affinity = cpu_core_map[data->cpu];
+	}
+
 	return 0;
 
 err_out_mem:
 	kfree(powernow_table);
 
 err_out:
-	acpi_processor_unregister_performance(&data->acpi_data, data->cpu);
+	acpi_processor_unregister_performance(data->acpi_data, data->cpu);
 
-	/* data->acpi_data.state_count informs us at ->exit() whether ACPI was used */
-	data->acpi_data.state_count = 0;
+	/* data->acpi_data->state_count informs us at ->exit() whether ACPI was used */
+	data->acpi_data->state_count = 0;
 
 	return -ENODEV;
 }
@@ -837,41 +859,23 @@ err_out:
 static int fill_powernow_table_pstate(struct powernow_k8_data *data, struct cpufreq_frequency_table *powernow_table)
 {
 	int i;
+	u32 hi = 0, lo = 0;
+	rdmsr(MSR_PSTATE_CUR_LIMIT, hi, lo);
+	data->max_hw_pstate = (hi & HW_PSTATE_MAX_MASK) >> HW_PSTATE_MAX_SHIFT;
 
-	for (i = 0; i < data->acpi_data.state_count; i++) {
+	for (i = 0; i < data->acpi_data->state_count; i++) {
 		u32 index;
-		u32 hi = 0, lo = 0;
-		u32 fid;
-		u32 did;
 
-		index = data->acpi_data.states[i].control & HW_PSTATE_MASK;
-		if (index > MAX_HW_PSTATE) {
+		index = data->acpi_data->states[i].control & HW_PSTATE_MASK;
+		if (index > data->max_hw_pstate) {
 			printk(KERN_ERR PFX "invalid pstate %d - bad value %d.\n", i, index);
 			printk(KERN_ERR PFX "Please report to BIOS manufacturer\n");
-		}
-		rdmsr(MSR_PSTATE_DEF_BASE + index, lo, hi);
-		if (!(hi & HW_PSTATE_VALID_MASK)) {
-			dprintk("invalid pstate %d, ignoring\n", index);
-			powernow_table[i].frequency = CPUFREQ_ENTRY_INVALID;
 			continue;
 		}
 
-		fid = lo & HW_PSTATE_FID_MASK;
-		did = (lo & HW_PSTATE_DID_MASK) >> HW_PSTATE_DID_SHIFT;
-
-		dprintk("   %d : fid 0x%x, did 0x%x\n", index, fid, did);
+		powernow_table[i].index = index;
+		powernow_table[i].frequency = data->acpi_data->states[i].core_frequency * 1000;
 
-		powernow_table[i].index = index | (fid << HW_FID_INDEX_SHIFT) | (did << HW_DID_INDEX_SHIFT);
-
-		powernow_table[i].frequency = find_khz_freq_from_fiddid(fid, did);
-
-		if (powernow_table[i].frequency != (data->acpi_data.states[i].core_frequency * 1000)) {
-			printk(KERN_INFO PFX "invalid freq entries %u kHz vs. %u kHz\n",
-				powernow_table[i].frequency,
-				(unsigned int) (data->acpi_data.states[i].core_frequency * 1000));
-			powernow_table[i].frequency = CPUFREQ_ENTRY_INVALID;
-			continue;
-		}
 	}
 	return 0;
 }
@@ -880,16 +884,16 @@ static int fill_powernow_table_fidvid(st
 {
 	int i;
 	int cntlofreq = 0;
-	for (i = 0; i < data->acpi_data.state_count; i++) {
+	for (i = 0; i < data->acpi_data->state_count; i++) {
 		u32 fid;
 		u32 vid;
 
 		if (data->exttype) {
-			fid = data->acpi_data.states[i].status & EXT_FID_MASK;
-			vid = (data->acpi_data.states[i].status >> VID_SHIFT) & EXT_VID_MASK;
+			fid = data->acpi_data->states[i].status & EXT_FID_MASK;
+			vid = (data->acpi_data->states[i].status >> VID_SHIFT) & EXT_VID_MASK;
 		} else {
-			fid = data->acpi_data.states[i].control & FID_MASK;
-			vid = (data->acpi_data.states[i].control >> VID_SHIFT) & VID_MASK;
+			fid = data->acpi_data->states[i].control & FID_MASK;
+			vid = (data->acpi_data->states[i].control >> VID_SHIFT) & VID_MASK;
 		}
 
 		dprintk("   %d : fid 0x%x, vid 0x%x\n", i, fid, vid);
@@ -930,10 +934,10 @@ static int fill_powernow_table_fidvid(st
 				cntlofreq = i;
 		}
 
-		if (powernow_table[i].frequency != (data->acpi_data.states[i].core_frequency * 1000)) {
+		if (powernow_table[i].frequency != (data->acpi_data->states[i].core_frequency * 1000)) {
 			printk(KERN_INFO PFX "invalid freq entries %u kHz vs. %u kHz\n",
 				powernow_table[i].frequency,
-				(unsigned int) (data->acpi_data.states[i].core_frequency * 1000));
+				(unsigned int) (data->acpi_data->states[i].core_frequency * 1000));
 			powernow_table[i].frequency = CPUFREQ_ENTRY_INVALID;
 			continue;
 		}
@@ -943,14 +947,15 @@ static int fill_powernow_table_fidvid(st
 
 static void powernow_k8_cpu_exit_acpi(struct powernow_k8_data *data)
 {
-	if (data->acpi_data.state_count)
-		acpi_processor_unregister_performance(&data->acpi_data, data->cpu);
+	if (data->acpi_data->state_count)
+		acpi_processor_unregister_performance(data->acpi_data, data->cpu);
 }
 
 #else
 static int powernow_k8_cpu_init_acpi(struct powernow_k8_data *data) { return -ENODEV; }
 static void powernow_k8_cpu_exit_acpi(struct powernow_k8_data *data) { return; }
 static void powernow_k8_acpi_pst_values(struct powernow_k8_data *data, unsigned int index) { return; }
+static int powernow_k8_cpu_preinit_acpi() { return -ENODEV; }
 #endif /* CONFIG_X86_POWERNOW_K8_ACPI */
 
 /* Take a frequency, and issue the fid/vid transition command */
@@ -1012,22 +1017,18 @@ static int transition_frequency_fidvid(s
 /* Take a frequency, and issue the hardware pstate transition command */
 static int transition_frequency_pstate(struct powernow_k8_data *data, unsigned int index)
 {
-	u32 fid = 0;
-	u32 did = 0;
 	u32 pstate = 0;
 	int res, i;
 	struct cpufreq_freqs freqs;
 
 	dprintk("cpu %d transition to index %u\n", smp_processor_id(), index);
 
-	/* get fid did for hardware pstate transition */
+	/* get MSR index for hardware pstate transition */
 	pstate = index & HW_PSTATE_MASK;
-	if (pstate > MAX_HW_PSTATE)
+	if (pstate > data->max_hw_pstate)
 		return 0;
-	fid = (index & HW_FID_INDEX_MASK) >> HW_FID_INDEX_SHIFT;
-	did = (index & HW_DID_INDEX_MASK) >> HW_DID_INDEX_SHIFT;
-	freqs.old = find_khz_freq_from_fiddid(data->currfid, data->currdid);
-	freqs.new = find_khz_freq_from_fiddid(fid, did);
+	freqs.old = find_khz_freq_from_pstate(data->powernow_table, data->currpstate);
+	freqs.new = find_khz_freq_from_pstate(data->powernow_table, pstate);
 
 	for_each_cpu_mask(i, *(data->available_cores)) {
 		freqs.cpu = i;
@@ -1035,9 +1036,7 @@ static int transition_frequency_pstate(s
 	}
 
 	res = transition_pstate(data, pstate);
-	data->currfid = find_fid_from_pstate(pstate);
-	data->currdid = find_did_from_pstate(pstate);
-	freqs.new = find_khz_freq_from_fiddid(data->currfid, data->currdid);
+	freqs.new = find_khz_freq_from_pstate(data->powernow_table, pstate);
 
 	for_each_cpu_mask(i, *(data->available_cores)) {
 		freqs.cpu = i;
@@ -1082,10 +1081,7 @@ static int powernowk8_target(struct cpuf
 	if (query_current_values_with_pending_wait(data))
 		goto err_out;
 
-	if (cpu_family == CPU_HW_PSTATE)
-		dprintk("targ: curr fid 0x%x, did 0x%x\n",
-			data->currfid, data->currvid);
-	else {
+	if (cpu_family != CPU_HW_PSTATE) {
 		dprintk("targ: curr fid 0x%x, vid 0x%x\n",
 		data->currfid, data->currvid);
 
@@ -1116,7 +1112,7 @@ static int powernowk8_target(struct cpuf
 	mutex_unlock(&fidvid_mutex);
 
 	if (cpu_family == CPU_HW_PSTATE)
-		pol->cur = find_khz_freq_from_fiddid(data->currfid, data->currdid);
+		pol->cur = find_khz_freq_from_pstate(data->powernow_table, newstate);
 	else
 		pol->cur = find_khz_freq_from_fid(data->currfid);
 	ret = 0;
@@ -1164,7 +1160,7 @@ static int __cpuinit powernowk8_cpu_init
 		 * an UP version, and is deprecated by AMD.
 		 */
 		if (num_online_cpus() != 1) {
-			printk(KERN_ERR PFX "MP systems not supported by PSB BIOS structure\n");
+			printk(KERN_ERR PFX "Your BIOS does not provide _PSS objects.  PowerNow! does not work on SMP systems without _PSS objects.  Complain to your BIOS vendor.\n");
 			kfree(data);
 			return -ENODEV;
 		}
@@ -1204,10 +1200,7 @@ static int __cpuinit powernowk8_cpu_init
 	set_cpus_allowed(current, oldmask);
 
 	pol->governor = CPUFREQ_DEFAULT_GOVERNOR;
-	if (cpu_family == CPU_HW_PSTATE)
-		pol->cpus = cpumask_of_cpu(pol->cpu);
-	else
-		pol->cpus = cpu_core_map[pol->cpu];
+	pol->cpus = data->starting_core_affinity;
 	data->available_cores = &(pol->cpus);
 
 	/* Take a crude guess here.
@@ -1216,7 +1209,7 @@ static int __cpuinit powernowk8_cpu_init
 	    + (3 * (1 << data->irt) * 10)) * 1000;
 
 	if (cpu_family == CPU_HW_PSTATE)
-		pol->cur = find_khz_freq_from_fiddid(data->currfid, data->currdid);
+		pol->cur = find_khz_freq_from_pstate(data->powernow_table, data->currpstate);
 	else
 		pol->cur = find_khz_freq_from_fid(data->currfid);
 	dprintk("policy current frequency %d kHz\n", pol->cur);
@@ -1233,8 +1226,7 @@ static int __cpuinit powernowk8_cpu_init
 	cpufreq_frequency_table_get_attr(data->powernow_table, pol->cpu);
 
 	if (cpu_family == CPU_HW_PSTATE)
-		dprintk("cpu_init done, current fid 0x%x, did 0x%x\n",
-			data->currfid, data->currdid);
+		dprintk("cpu_init done, current pstate 0x%x\n", data->currpstate);
 	else
 		dprintk("cpu_init done, current fid 0x%x, vid 0x%x\n",
 			data->currfid, data->currvid);
@@ -1289,7 +1281,10 @@ static unsigned int powernowk8_get (unsi
 	if (query_current_values_with_pending_wait(data))
 		goto out;
 
-	khz = find_khz_freq_from_fid(data->currfid);
+	if (cpu_family == CPU_HW_PSTATE)
+		khz = find_khz_freq_from_pstate(data->powernow_table, data->currpstate);
+	else
+		khz = find_khz_freq_from_fid(data->currfid);
 
 out:
 	set_cpus_allowed(current, oldmask);
@@ -1323,6 +1318,7 @@ static int __cpuinit powernowk8_init(voi
 	}
 
 	if (supported_cpus == num_online_cpus()) {
+		powernow_k8_cpu_preinit_acpi();
 		printk(KERN_INFO PFX "Found %d %s "
 			"processors (" VERSION ")\n", supported_cpus,
 			boot_cpu_data.x86_model_id);
diff -rpuN linux-2.6.18.8/arch/i386/kernel/cpu/cpufreq/powernow-k8.h linux-2.6.18-xen-3.3.0/arch/i386/kernel/cpu/cpufreq/powernow-k8.h
--- linux-2.6.18.8/arch/i386/kernel/cpu/cpufreq/powernow-k8.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/cpu/cpufreq/powernow-k8.h	2008-08-21 11:36:07.000000000 +0200
@@ -1,5 +1,5 @@
 /*
- *  (c) 2003-2006 Advanced Micro Devices, Inc.
+#*  (c) 2003-2006 Advanced Micro Devices, Inc.
  *  Your use of this code is subject to the terms and conditions of the
  *  GNU general public license version 2. See "COPYING" or
  *  http://www.gnu.org/licenses/gpl.html
@@ -10,6 +10,7 @@ struct powernow_k8_data {
 
 	u32 numps;  /* number of p-states */
 	u32 batps;  /* number of p-states supported on battery */
+	u32 max_hw_pstate; /* maximum legal hardware pstate */
 
 	/* these values are constant when the PSB is used to determine
 	 * vid/fid pairings, but are modified during the ->target() call
@@ -21,8 +22,8 @@ struct powernow_k8_data {
 	u32 plllock; /* pll lock time, units 1 us */
         u32 exttype; /* extended interface = 1 */
 
-	/* keep track of the current fid / vid or did */
-	u32 currvid, currfid, currdid;
+	/* keep track of the current fid / vid or pstate */
+	u32 currvid, currfid, currpstate;
 
 	/* the powernow_table includes all frequency and vid/fid pairings:
 	 * fid are the lower 8 bits of the index, vid are the upper 8 bits.
@@ -32,12 +33,13 @@ struct powernow_k8_data {
 #ifdef CONFIG_X86_POWERNOW_K8_ACPI
 	/* the acpi table needs to be kept. it's only available if ACPI was
 	 * used to determine valid frequency/vid/fid states */
-	struct acpi_processor_performance acpi_data;
+	struct acpi_processor_performance *acpi_data;
 #endif
 	/* we need to keep track of associated cores, but let cpufreq
 	 * handle hotplug events - so just point at cpufreq pol->cpus
 	 * structure */
 	cpumask_t *available_cores;
+	cpumask_t starting_core_affinity;
 };
 
 
@@ -87,23 +89,14 @@ struct powernow_k8_data {
 
 /* Hardware Pstate _PSS and MSR definitions */
 #define USE_HW_PSTATE		0x00000080
-#define HW_PSTATE_FID_MASK 	0x0000003f
-#define HW_PSTATE_DID_MASK 	0x000001c0
-#define HW_PSTATE_DID_SHIFT 	6
-#define HW_PSTATE_MASK 		0x00000007
-#define HW_PSTATE_VALID_MASK 	0x80000000
-#define HW_FID_INDEX_SHIFT	8
-#define HW_FID_INDEX_MASK	0x0000ff00
-#define HW_DID_INDEX_SHIFT	16
-#define HW_DID_INDEX_MASK	0x00ff0000
-#define HW_WATTS_MASK		0xff
-#define HW_PWR_DVR_MASK		0x300
-#define HW_PWR_DVR_SHIFT	8
-#define HW_PWR_MAX_MULT		3
-#define MAX_HW_PSTATE		8	/* hw pstate supports up to 8 */
+#define HW_PSTATE_MASK		0x00000007
+#define HW_PSTATE_VALID_MASK	0x80000000
+#define HW_PSTATE_MAX_MASK	0x000000f0
+#define HW_PSTATE_MAX_SHIFT	4
 #define MSR_PSTATE_DEF_BASE 	0xc0010064 /* base of Pstate MSRs */
 #define MSR_PSTATE_STATUS 	0xc0010063 /* Pstate Status MSR */
 #define MSR_PSTATE_CTRL 	0xc0010062 /* Pstate control MSR */
+#define MSR_PSTATE_CUR_LIMIT	0xc0010061 /* pstate current limit MSR */
 
 /* define the two driver architectures */
 #define CPU_OPTERON 0
diff -rpuN linux-2.6.18.8/arch/i386/kernel/cpu/Makefile linux-2.6.18-xen-3.3.0/arch/i386/kernel/cpu/Makefile
--- linux-2.6.18.8/arch/i386/kernel/cpu/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/cpu/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -17,3 +17,4 @@ obj-$(CONFIG_X86_MCE)	+=	mcheck/
 
 obj-$(CONFIG_MTRR)	+= 	mtrr/
 obj-$(CONFIG_CPU_FREQ)	+=	cpufreq/
+
diff -rpuN linux-2.6.18.8/arch/i386/kernel/cpu/mtrr/main-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/cpu/mtrr/main-xen.c
--- linux-2.6.18.8/arch/i386/kernel/cpu/mtrr/main-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/cpu/mtrr/main-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,198 @@
+#include <linux/init.h>
+#include <linux/proc_fs.h>
+#include <linux/ctype.h>
+#include <linux/module.h>
+#include <linux/seq_file.h>
+#include <asm/uaccess.h>
+#include <linux/mutex.h>
+
+#include <asm/mtrr.h>
+#include "mtrr.h"
+
+static DEFINE_MUTEX(mtrr_mutex);
+
+void generic_get_mtrr(unsigned int reg, unsigned long *base,
+		      unsigned int *size, mtrr_type * type)
+{
+	struct xen_platform_op op;
+
+	op.cmd = XENPF_read_memtype;
+	op.u.read_memtype.reg = reg;
+	if (unlikely(HYPERVISOR_platform_op(&op)))
+		memset(&op.u.read_memtype, 0, sizeof(op.u.read_memtype));
+
+	*size = op.u.read_memtype.nr_mfns;
+	*base = op.u.read_memtype.mfn;
+	*type = op.u.read_memtype.type;
+}
+
+struct mtrr_ops generic_mtrr_ops = {
+	.use_intel_if      = 1,
+	.get               = generic_get_mtrr,
+};
+
+struct mtrr_ops *mtrr_if = &generic_mtrr_ops;
+unsigned int num_var_ranges;
+unsigned int *usage_table;
+
+static void __init set_num_var_ranges(void)
+{
+	struct xen_platform_op op;
+
+	for (num_var_ranges = 0; ; num_var_ranges++) {
+		op.cmd = XENPF_read_memtype;
+		op.u.read_memtype.reg = num_var_ranges;
+		if (HYPERVISOR_platform_op(&op) != 0)
+			break;
+	}
+}
+
+static void __init init_table(void)
+{
+	int i, max;
+
+	max = num_var_ranges;
+	if ((usage_table = kmalloc(max * sizeof *usage_table, GFP_KERNEL))
+	    == NULL) {
+		printk(KERN_ERR "mtrr: could not allocate\n");
+		return;
+	}
+	for (i = 0; i < max; i++)
+		usage_table[i] = 0;
+}
+
+int mtrr_add_page(unsigned long base, unsigned long size, 
+		  unsigned int type, char increment)
+{
+	int error;
+	struct xen_platform_op op;
+
+	mutex_lock(&mtrr_mutex);
+
+	op.cmd = XENPF_add_memtype;
+	op.u.add_memtype.mfn     = base;
+	op.u.add_memtype.nr_mfns = size;
+	op.u.add_memtype.type    = type;
+	error = HYPERVISOR_platform_op(&op);
+	if (error) {
+		mutex_unlock(&mtrr_mutex);
+		BUG_ON(error > 0);
+		return error;
+	}
+
+	if (increment)
+		++usage_table[op.u.add_memtype.reg];
+
+	mutex_unlock(&mtrr_mutex);
+
+	return op.u.add_memtype.reg;
+}
+
+static int mtrr_check(unsigned long base, unsigned long size)
+{
+	if ((base & (PAGE_SIZE - 1)) || (size & (PAGE_SIZE - 1))) {
+		printk(KERN_WARNING
+			"mtrr: size and base must be multiples of 4 kiB\n");
+		printk(KERN_DEBUG
+			"mtrr: size: 0x%lx  base: 0x%lx\n", size, base);
+		dump_stack();
+		return -1;
+	}
+	return 0;
+}
+
+int
+mtrr_add(unsigned long base, unsigned long size, unsigned int type,
+	 char increment)
+{
+	if (mtrr_check(base, size))
+		return -EINVAL;
+	return mtrr_add_page(base >> PAGE_SHIFT, size >> PAGE_SHIFT, type,
+			     increment);
+}
+
+int mtrr_del_page(int reg, unsigned long base, unsigned long size)
+{
+	unsigned i;
+	mtrr_type ltype;
+	unsigned long lbase;
+	unsigned int lsize;
+	int error = -EINVAL;
+	struct xen_platform_op op;
+
+	mutex_lock(&mtrr_mutex);
+
+	if (reg < 0) {
+		/*  Search for existing MTRR  */
+		for (i = 0; i < num_var_ranges; ++i) {
+			mtrr_if->get(i, &lbase, &lsize, &ltype);
+			if (lbase == base && lsize == size) {
+				reg = i;
+				break;
+			}
+		}
+		if (reg < 0) {
+			printk(KERN_DEBUG "mtrr: no MTRR for %lx000,%lx000 found\n", base,
+			       size);
+			goto out;
+		}
+	}
+	if (usage_table[reg] < 1) {
+		printk(KERN_WARNING "mtrr: reg: %d has count=0\n", reg);
+		goto out;
+	}
+	if (--usage_table[reg] < 1) {
+		op.cmd = XENPF_del_memtype;
+		op.u.del_memtype.handle = 0;
+		op.u.del_memtype.reg    = reg;
+		error = HYPERVISOR_platform_op(&op);
+		if (error) {
+			BUG_ON(error > 0);
+			goto out;
+		}
+	}
+	error = reg;
+ out:
+	mutex_unlock(&mtrr_mutex);
+	return error;
+}
+
+int
+mtrr_del(int reg, unsigned long base, unsigned long size)
+{
+	if (mtrr_check(base, size))
+		return -EINVAL;
+	return mtrr_del_page(reg, base >> PAGE_SHIFT, size >> PAGE_SHIFT);
+}
+
+EXPORT_SYMBOL(mtrr_add);
+EXPORT_SYMBOL(mtrr_del);
+
+void __init mtrr_bp_init(void)
+{
+}
+
+void mtrr_ap_init(void)
+{
+}
+
+static int __init mtrr_init(void)
+{
+	struct cpuinfo_x86 *c = &boot_cpu_data;
+
+	if (!is_initial_xendomain())
+		return -ENODEV;
+
+	if ((!cpu_has(c, X86_FEATURE_MTRR)) &&
+	    (!cpu_has(c, X86_FEATURE_K6_MTRR)) &&
+	    (!cpu_has(c, X86_FEATURE_CYRIX_ARR)) &&
+	    (!cpu_has(c, X86_FEATURE_CENTAUR_MCR)))
+		return -ENODEV;
+
+	set_num_var_ranges();
+	init_table();
+
+	return 0;
+}
+
+subsys_initcall(mtrr_init);
diff -rpuN linux-2.6.18.8/arch/i386/kernel/cpu/mtrr/Makefile linux-2.6.18-xen-3.3.0/arch/i386/kernel/cpu/mtrr/Makefile
--- linux-2.6.18.8/arch/i386/kernel/cpu/mtrr/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/cpu/mtrr/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -3,3 +3,4 @@ obj-y		+= amd.o
 obj-y		+= cyrix.o
 obj-y		+= centaur.o
 
+obj-$(CONFIG_XEN) := main.o if.o
diff -rpuN linux-2.6.18.8/arch/i386/kernel/crash.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/crash.c
--- linux-2.6.18.8/arch/i386/kernel/crash.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/crash.c	2008-08-21 11:36:07.000000000 +0200
@@ -90,6 +90,7 @@ static void crash_save_self(struct pt_re
 	crash_save_this_cpu(regs, cpu);
 }
 
+#ifndef CONFIG_XEN
 #if defined(CONFIG_SMP) && defined(CONFIG_X86_LOCAL_APIC)
 static atomic_t waiting_for_crash_ipi;
 
@@ -154,6 +155,7 @@ static void nmi_shootdown_cpus(void)
 	/* There are no cpus to shootdown */
 }
 #endif
+#endif /* CONFIG_XEN */
 
 void machine_crash_shutdown(struct pt_regs *regs)
 {
@@ -170,10 +172,12 @@ void machine_crash_shutdown(struct pt_re
 
 	/* Make a note of crashing cpu. Will be used in NMI callback.*/
 	crashing_cpu = smp_processor_id();
+#ifndef CONFIG_XEN
 	nmi_shootdown_cpus();
 	lapic_shutdown();
 #if defined(CONFIG_X86_IO_APIC)
 	disable_IO_APIC();
 #endif
+#endif /* CONFIG_XEN */
 	crash_save_self(regs);
 }
diff -rpuN linux-2.6.18.8/arch/i386/kernel/early_printk-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/early_printk-xen.c
--- linux-2.6.18.8/arch/i386/kernel/early_printk-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/early_printk-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,2 @@
+
+#include "../../x86_64/kernel/early_printk-xen.c"
diff -rpuN linux-2.6.18.8/arch/i386/kernel/entry.S linux-2.6.18-xen-3.3.0/arch/i386/kernel/entry.S
--- linux-2.6.18.8/arch/i386/kernel/entry.S	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/entry.S	2008-08-21 11:36:07.000000000 +0200
@@ -269,7 +269,7 @@ ENTRY(sysenter_entry)
 	CFI_STARTPROC simple
 	CFI_DEF_CFA esp, 0
 	CFI_REGISTER esp, ebp
-	movl TSS_sysenter_esp0(%esp),%esp
+	movl SYSENTER_stack_esp0(%esp),%esp
 sysenter_past_esp:
 	/*
 	 * No need to follow this irqs on/off section: the syscall
@@ -689,7 +689,7 @@ device_not_available_emulate:
  * that sets up the real kernel stack. Check here, since we can't
  * allow the wrong stack to be used.
  *
- * "TSS_sysenter_esp0+12" is because the NMI/debug handler will have
+ * "SYSENTER_stack_esp0+12" is because the NMI/debug handler will have
  * already pushed 3 words if it hits on the sysenter instruction:
  * eflags, cs and eip.
  *
@@ -701,7 +701,7 @@ device_not_available_emulate:
 	cmpw $__KERNEL_CS,4(%esp);		\
 	jne ok;					\
 label:						\
-	movl TSS_sysenter_esp0+offset(%esp),%esp;	\
+	movl SYSENTER_stack_esp0+offset(%esp),%esp;	\
 	pushfl;					\
 	pushl $__KERNEL_CS;			\
 	pushl $sysenter_past_esp
diff -rpuN linux-2.6.18.8/arch/i386/kernel/entry-xen.S linux-2.6.18-xen-3.3.0/arch/i386/kernel/entry-xen.S
--- linux-2.6.18.8/arch/i386/kernel/entry-xen.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/entry-xen.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1238 @@
+/*
+ *  linux/arch/i386/entry.S
+ *
+ *  Copyright (C) 1991, 1992  Linus Torvalds
+ */
+
+/*
+ * entry.S contains the system-call and fault low-level handling routines.
+ * This also contains the timer-interrupt handler, as well as all interrupts
+ * and faults that can result in a task-switch.
+ *
+ * NOTE: This code handles signal-recognition, which happens every time
+ * after a timer-interrupt and after each system call.
+ *
+ * I changed all the .align's to 4 (16 byte alignment), as that's faster
+ * on a 486.
+ *
+ * Stack layout in 'ret_from_system_call':
+ * 	ptrace needs to have all regs on the stack.
+ *	if the order here is changed, it needs to be
+ *	updated in fork.c:copy_process, signal.c:do_signal,
+ *	ptrace.c and ptrace.h
+ *
+ *	 0(%esp) - %ebx
+ *	 4(%esp) - %ecx
+ *	 8(%esp) - %edx
+ *       C(%esp) - %esi
+ *	10(%esp) - %edi
+ *	14(%esp) - %ebp
+ *	18(%esp) - %eax
+ *	1C(%esp) - %ds
+ *	20(%esp) - %es
+ *	24(%esp) - orig_eax
+ *	28(%esp) - %eip
+ *	2C(%esp) - %cs
+ *	30(%esp) - %eflags
+ *	34(%esp) - %oldesp
+ *	38(%esp) - %oldss
+ *
+ * "current" is in register %ebx during any slow entries.
+ */
+
+#include <linux/linkage.h>
+#include <asm/thread_info.h>
+#include <asm/irqflags.h>
+#include <asm/errno.h>
+#include <asm/segment.h>
+#include <asm/smp.h>
+#include <asm/page.h>
+#include <asm/desc.h>
+#include <asm/dwarf2.h>
+#include "irq_vectors.h"
+#include <xen/interface/xen.h>
+
+#define nr_syscalls ((syscall_table_size)/4)
+
+EBX		= 0x00
+ECX		= 0x04
+EDX		= 0x08
+ESI		= 0x0C
+EDI		= 0x10
+EBP		= 0x14
+EAX		= 0x18
+DS		= 0x1C
+ES		= 0x20
+ORIG_EAX	= 0x24
+EIP		= 0x28
+CS		= 0x2C
+EFLAGS		= 0x30
+OLDESP		= 0x34
+OLDSS		= 0x38
+
+CF_MASK		= 0x00000001
+TF_MASK		= 0x00000100
+IF_MASK		= 0x00000200
+DF_MASK		= 0x00000400 
+NT_MASK		= 0x00004000
+VM_MASK		= 0x00020000
+/* Pseudo-eflags. */
+NMI_MASK	= 0x80000000
+
+#ifndef CONFIG_XEN
+#define DISABLE_INTERRUPTS	cli
+#define ENABLE_INTERRUPTS	sti
+#else
+/* Offsets into shared_info_t. */
+#define evtchn_upcall_pending		/* 0 */
+#define evtchn_upcall_mask		1
+
+#define sizeof_vcpu_shift		6
+
+#ifdef CONFIG_SMP
+#define GET_VCPU_INFO		movl TI_cpu(%ebp),%esi			; \
+				shl  $sizeof_vcpu_shift,%esi		; \
+				addl HYPERVISOR_shared_info,%esi
+#else
+#define GET_VCPU_INFO		movl HYPERVISOR_shared_info,%esi
+#endif
+
+#define __DISABLE_INTERRUPTS	movb $1,evtchn_upcall_mask(%esi)
+#define __ENABLE_INTERRUPTS	movb $0,evtchn_upcall_mask(%esi)
+#define DISABLE_INTERRUPTS	GET_VCPU_INFO				; \
+				__DISABLE_INTERRUPTS
+#define ENABLE_INTERRUPTS	GET_VCPU_INFO				; \
+				__ENABLE_INTERRUPTS
+#define __TEST_PENDING		testb $0xFF,evtchn_upcall_pending(%esi)
+#endif
+
+#ifdef CONFIG_PREEMPT
+#define preempt_stop		cli; TRACE_IRQS_OFF
+#else
+#define preempt_stop
+#define resume_kernel		restore_nocheck
+#endif
+
+.macro TRACE_IRQS_IRET
+#ifdef CONFIG_TRACE_IRQFLAGS
+	testl $IF_MASK,EFLAGS(%esp)     # interrupts off?
+	jz 1f
+	TRACE_IRQS_ON
+1:
+#endif
+.endm
+
+#ifdef CONFIG_VM86
+#define resume_userspace_sig	check_userspace
+#else
+#define resume_userspace_sig	resume_userspace
+#endif
+
+#define SAVE_ALL \
+	cld; \
+	pushl %es; \
+	CFI_ADJUST_CFA_OFFSET 4;\
+	/*CFI_REL_OFFSET es, 0;*/\
+	pushl %ds; \
+	CFI_ADJUST_CFA_OFFSET 4;\
+	/*CFI_REL_OFFSET ds, 0;*/\
+	pushl %eax; \
+	CFI_ADJUST_CFA_OFFSET 4;\
+	CFI_REL_OFFSET eax, 0;\
+	pushl %ebp; \
+	CFI_ADJUST_CFA_OFFSET 4;\
+	CFI_REL_OFFSET ebp, 0;\
+	pushl %edi; \
+	CFI_ADJUST_CFA_OFFSET 4;\
+	CFI_REL_OFFSET edi, 0;\
+	pushl %esi; \
+	CFI_ADJUST_CFA_OFFSET 4;\
+	CFI_REL_OFFSET esi, 0;\
+	pushl %edx; \
+	CFI_ADJUST_CFA_OFFSET 4;\
+	CFI_REL_OFFSET edx, 0;\
+	pushl %ecx; \
+	CFI_ADJUST_CFA_OFFSET 4;\
+	CFI_REL_OFFSET ecx, 0;\
+	pushl %ebx; \
+	CFI_ADJUST_CFA_OFFSET 4;\
+	CFI_REL_OFFSET ebx, 0;\
+	movl $(__USER_DS), %edx; \
+	movl %edx, %ds; \
+	movl %edx, %es;
+
+#define RESTORE_INT_REGS \
+	popl %ebx;	\
+	CFI_ADJUST_CFA_OFFSET -4;\
+	CFI_RESTORE ebx;\
+	popl %ecx;	\
+	CFI_ADJUST_CFA_OFFSET -4;\
+	CFI_RESTORE ecx;\
+	popl %edx;	\
+	CFI_ADJUST_CFA_OFFSET -4;\
+	CFI_RESTORE edx;\
+	popl %esi;	\
+	CFI_ADJUST_CFA_OFFSET -4;\
+	CFI_RESTORE esi;\
+	popl %edi;	\
+	CFI_ADJUST_CFA_OFFSET -4;\
+	CFI_RESTORE edi;\
+	popl %ebp;	\
+	CFI_ADJUST_CFA_OFFSET -4;\
+	CFI_RESTORE ebp;\
+	popl %eax;	\
+	CFI_ADJUST_CFA_OFFSET -4;\
+	CFI_RESTORE eax
+
+#define RESTORE_REGS	\
+	RESTORE_INT_REGS; \
+1:	popl %ds;	\
+	CFI_ADJUST_CFA_OFFSET -4;\
+	/*CFI_RESTORE ds;*/\
+2:	popl %es;	\
+	CFI_ADJUST_CFA_OFFSET -4;\
+	/*CFI_RESTORE es;*/\
+.section .fixup,"ax";	\
+3:	movl $0,(%esp);	\
+	jmp 1b;		\
+4:	movl $0,(%esp);	\
+	jmp 2b;		\
+.previous;		\
+.section __ex_table,"a";\
+	.align 4;	\
+	.long 1b,3b;	\
+	.long 2b,4b;	\
+.previous
+
+#define RING0_INT_FRAME \
+	CFI_STARTPROC simple;\
+	CFI_DEF_CFA esp, 3*4;\
+	/*CFI_OFFSET cs, -2*4;*/\
+	CFI_OFFSET eip, -3*4
+
+#define RING0_EC_FRAME \
+	CFI_STARTPROC simple;\
+	CFI_DEF_CFA esp, 4*4;\
+	/*CFI_OFFSET cs, -2*4;*/\
+	CFI_OFFSET eip, -3*4
+
+#define RING0_PTREGS_FRAME \
+	CFI_STARTPROC simple;\
+	CFI_DEF_CFA esp, OLDESP-EBX;\
+	/*CFI_OFFSET cs, CS-OLDESP;*/\
+	CFI_OFFSET eip, EIP-OLDESP;\
+	/*CFI_OFFSET es, ES-OLDESP;*/\
+	/*CFI_OFFSET ds, DS-OLDESP;*/\
+	CFI_OFFSET eax, EAX-OLDESP;\
+	CFI_OFFSET ebp, EBP-OLDESP;\
+	CFI_OFFSET edi, EDI-OLDESP;\
+	CFI_OFFSET esi, ESI-OLDESP;\
+	CFI_OFFSET edx, EDX-OLDESP;\
+	CFI_OFFSET ecx, ECX-OLDESP;\
+	CFI_OFFSET ebx, EBX-OLDESP
+
+ENTRY(ret_from_fork)
+	CFI_STARTPROC
+	pushl %eax
+	CFI_ADJUST_CFA_OFFSET 4
+	call schedule_tail
+	GET_THREAD_INFO(%ebp)
+	popl %eax
+	CFI_ADJUST_CFA_OFFSET -4
+	pushl $0x0202			# Reset kernel eflags
+	CFI_ADJUST_CFA_OFFSET 4
+	popfl
+	CFI_ADJUST_CFA_OFFSET -4
+	jmp syscall_exit
+	CFI_ENDPROC
+
+/*
+ * Return to user mode is not as complex as all this looks,
+ * but we want the default path for a system call return to
+ * go as quickly as possible which is why some of this is
+ * less clear than it otherwise should be.
+ */
+
+	# userspace resumption stub bypassing syscall exit tracing
+	ALIGN
+	RING0_PTREGS_FRAME
+ret_from_exception:
+	preempt_stop
+ret_from_intr:
+	GET_THREAD_INFO(%ebp)
+check_userspace:
+	movl EFLAGS(%esp), %eax		# mix EFLAGS and CS
+	movb CS(%esp), %al
+	testl $(VM_MASK | 2), %eax
+	jz resume_kernel
+ENTRY(resume_userspace)
+	DISABLE_INTERRUPTS		# make sure we don't miss an interrupt
+					# setting need_resched or sigpending
+					# between sampling and the iret
+	movl TI_flags(%ebp), %ecx
+	andl $_TIF_WORK_MASK, %ecx	# is there any work to be done on
+					# int/exception return?
+	jne work_pending
+	jmp restore_all
+
+#ifdef CONFIG_PREEMPT
+ENTRY(resume_kernel)
+	cli
+	cmpl $0,TI_preempt_count(%ebp)	# non-zero preempt_count ?
+	jnz restore_nocheck
+need_resched:
+	movl TI_flags(%ebp), %ecx	# need_resched set ?
+	testb $_TIF_NEED_RESCHED, %cl
+	jz restore_all
+	testl $IF_MASK,EFLAGS(%esp)     # interrupts off (exception path) ?
+	jz restore_all
+	call preempt_schedule_irq
+	jmp need_resched
+#endif
+	CFI_ENDPROC
+
+/* SYSENTER_RETURN points to after the "sysenter" instruction in
+   the vsyscall page.  See vsyscall-sysentry.S, which defines the symbol.  */
+
+	# sysenter call handler stub
+ENTRY(sysenter_entry)
+	CFI_STARTPROC simple
+	CFI_DEF_CFA esp, 0
+	CFI_REGISTER esp, ebp
+	movl SYSENTER_stack_esp0(%esp),%esp
+sysenter_past_esp:
+	/*
+	 * No need to follow this irqs on/off section: the syscall
+	 * disabled irqs and here we enable it straight after entry:
+	 */
+	sti
+	pushl $(__USER_DS)
+	CFI_ADJUST_CFA_OFFSET 4
+	/*CFI_REL_OFFSET ss, 0*/
+	pushl %ebp
+	CFI_ADJUST_CFA_OFFSET 4
+	CFI_REL_OFFSET esp, 0
+	pushfl
+	CFI_ADJUST_CFA_OFFSET 4
+	pushl $(__USER_CS)
+	CFI_ADJUST_CFA_OFFSET 4
+	/*CFI_REL_OFFSET cs, 0*/
+	/*
+	 * Push current_thread_info()->sysenter_return to the stack.
+	 * A tiny bit of offset fixup is necessary - 4*4 means the 4 words
+	 * pushed above; +8 corresponds to copy_thread's esp0 setting.
+	 */
+	pushl (TI_sysenter_return-THREAD_SIZE+8+4*4)(%esp)
+	CFI_ADJUST_CFA_OFFSET 4
+	CFI_REL_OFFSET eip, 0
+
+/*
+ * Load the potential sixth argument from user stack.
+ * Careful about security.
+ */
+	cmpl $__PAGE_OFFSET-3,%ebp
+	jae syscall_fault
+1:	movl (%ebp),%ebp
+.section __ex_table,"a"
+	.align 4
+	.long 1b,syscall_fault
+.previous
+
+	pushl %eax
+	CFI_ADJUST_CFA_OFFSET 4
+	SAVE_ALL
+	GET_THREAD_INFO(%ebp)
+
+	/* Note, _TIF_SECCOMP is bit number 8, and so it needs testw and not testb */
+	testw $(_TIF_SYSCALL_EMU|_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT),TI_flags(%ebp)
+	jnz syscall_trace_entry
+	cmpl $(nr_syscalls), %eax
+	jae syscall_badsys
+	call *sys_call_table(,%eax,4)
+	movl %eax,EAX(%esp)
+	DISABLE_INTERRUPTS
+	TRACE_IRQS_OFF
+	movl TI_flags(%ebp), %ecx
+	testw $_TIF_ALLWORK_MASK, %cx
+	jne syscall_exit_work
+/* if something modifies registers it must also disable sysexit */
+	movl EIP(%esp), %edx
+	movl OLDESP(%esp), %ecx
+	xorl %ebp,%ebp
+#ifdef CONFIG_XEN
+	TRACE_IRQS_ON
+	__ENABLE_INTERRUPTS
+sysexit_scrit:	/**** START OF SYSEXIT CRITICAL REGION ****/
+	__TEST_PENDING
+	jnz  14f			# process more events if necessary...
+	movl ESI(%esp), %esi
+	sysexit
+14:	__DISABLE_INTERRUPTS
+	TRACE_IRQS_OFF
+sysexit_ecrit:	/**** END OF SYSEXIT CRITICAL REGION ****/
+	push %esp
+	call evtchn_do_upcall
+	add  $4,%esp
+	jmp  ret_from_intr
+#else
+	TRACE_IRQS_ON
+	sti
+	sysexit
+#endif /* !CONFIG_XEN */
+	CFI_ENDPROC
+
+	# pv sysenter call handler stub
+ENTRY(sysenter_entry_pv)
+	RING0_INT_FRAME
+	movl $__USER_DS,16(%esp)
+	movl %ebp,12(%esp)
+	movl $__USER_CS,4(%esp)
+	addl $4,%esp
+	/* +5*4 is SS:ESP,EFLAGS,CS:EIP. +8 is esp0 setting. */
+	pushl (TI_sysenter_return-THREAD_SIZE+8+4*4)(%esp)
+/*
+ * Load the potential sixth argument from user stack.
+ * Careful about security.
+ */
+	cmpl $__PAGE_OFFSET-3,%ebp
+	jae syscall_fault
+1:	movl (%ebp),%ebp
+.section __ex_table,"a"
+	.align 4
+	.long 1b,syscall_fault
+.previous
+	/* fall through */
+	CFI_ENDPROC
+ENDPROC(sysenter_entry_pv)
+
+	# system call handler stub
+ENTRY(system_call)
+	RING0_INT_FRAME			# can't unwind into user space anyway
+	pushl %eax			# save orig_eax
+	CFI_ADJUST_CFA_OFFSET 4
+	SAVE_ALL
+	GET_THREAD_INFO(%ebp)
+	testl $TF_MASK,EFLAGS(%esp)
+	jz no_singlestep
+	orl $_TIF_SINGLESTEP,TI_flags(%ebp)
+no_singlestep:
+					# system call tracing in operation / emulation
+	/* Note, _TIF_SECCOMP is bit number 8, and so it needs testw and not testb */
+	testw $(_TIF_SYSCALL_EMU|_TIF_SYSCALL_TRACE|_TIF_SECCOMP|_TIF_SYSCALL_AUDIT),TI_flags(%ebp)
+	jnz syscall_trace_entry
+	cmpl $(nr_syscalls), %eax
+	jae syscall_badsys
+syscall_call:
+	call *sys_call_table(,%eax,4)
+	movl %eax,EAX(%esp)		# store the return value
+syscall_exit:
+	DISABLE_INTERRUPTS		# make sure we don't miss an interrupt
+					# setting need_resched or sigpending
+					# between sampling and the iret
+	TRACE_IRQS_OFF
+	movl TI_flags(%ebp), %ecx
+	testw $_TIF_ALLWORK_MASK, %cx	# current->work
+	jne syscall_exit_work
+
+restore_all:
+#ifndef CONFIG_XEN
+	movl EFLAGS(%esp), %eax		# mix EFLAGS, SS and CS
+	# Warning: OLDSS(%esp) contains the wrong/random values if we
+	# are returning to the kernel.
+	# See comments in process.c:copy_thread() for details.
+	movb OLDSS(%esp), %ah
+	movb CS(%esp), %al
+	andl $(VM_MASK | (4 << 8) | 3), %eax
+	cmpl $((4 << 8) | 3), %eax
+	CFI_REMEMBER_STATE
+	je ldt_ss			# returning to user-space with LDT SS
+restore_nocheck:
+#else
+restore_nocheck:
+	movl EFLAGS(%esp), %eax
+	testl $(VM_MASK|NMI_MASK), %eax
+	CFI_REMEMBER_STATE
+	jnz hypervisor_iret
+	shr $9, %eax			# EAX[0] == IRET_EFLAGS.IF
+	GET_VCPU_INFO
+	andb evtchn_upcall_mask(%esi),%al
+	andb $1,%al			# EAX[0] == IRET_EFLAGS.IF & event_mask
+	CFI_REMEMBER_STATE
+	jnz restore_all_enable_events	#        != 0 => enable event delivery
+#endif
+	TRACE_IRQS_IRET
+restore_nocheck_notrace:
+	RESTORE_REGS
+	addl $4, %esp
+	CFI_ADJUST_CFA_OFFSET -4
+1:	iret
+.section .fixup,"ax"
+iret_exc:
+#ifndef CONFIG_XEN
+	TRACE_IRQS_ON
+	sti
+#endif
+	pushl $0			# no error code
+	pushl $do_iret_error
+	jmp error_code
+.previous
+.section __ex_table,"a"
+	.align 4
+	.long 1b,iret_exc
+.previous
+
+	CFI_RESTORE_STATE
+#ifndef CONFIG_XEN
+ldt_ss:
+	larl OLDSS(%esp), %eax
+	jnz restore_nocheck
+	testl $0x00400000, %eax		# returning to 32bit stack?
+	jnz restore_nocheck		# allright, normal return
+	/* If returning to userspace with 16bit stack,
+	 * try to fix the higher word of ESP, as the CPU
+	 * won't restore it.
+	 * This is an "official" bug of all the x86-compatible
+	 * CPUs, which we can try to work around to make
+	 * dosemu and wine happy. */
+	subl $8, %esp		# reserve space for switch16 pointer
+	CFI_ADJUST_CFA_OFFSET 8
+	cli
+	TRACE_IRQS_OFF
+	movl %esp, %eax
+	/* Set up the 16bit stack frame with switch32 pointer on top,
+	 * and a switch16 pointer on top of the current frame. */
+	call setup_x86_bogus_stack
+	CFI_ADJUST_CFA_OFFSET -8	# frame has moved
+	TRACE_IRQS_IRET
+	RESTORE_REGS
+	lss 20+4(%esp), %esp	# switch to 16bit stack
+1:	iret
+.section __ex_table,"a"
+	.align 4
+	.long 1b,iret_exc
+.previous
+#else
+        ALIGN
+restore_all_enable_events:
+	TRACE_IRQS_ON
+	__ENABLE_INTERRUPTS
+scrit:	/**** START OF CRITICAL REGION ****/
+	__TEST_PENDING
+	jnz  14f			# process more events if necessary...
+	RESTORE_REGS
+	addl $4, %esp
+	CFI_ADJUST_CFA_OFFSET -4
+1:	iret
+.section __ex_table,"a"
+	.align 4
+	.long 1b,iret_exc
+.previous
+14:	__DISABLE_INTERRUPTS
+	TRACE_IRQS_OFF
+	jmp  11f
+ecrit:  /**** END OF CRITICAL REGION ****/
+
+	CFI_RESTORE_STATE
+hypervisor_iret:
+	andl $~NMI_MASK, EFLAGS(%esp)
+	RESTORE_REGS
+	addl $4, %esp
+	CFI_ADJUST_CFA_OFFSET -4
+	jmp  hypercall_page + (__HYPERVISOR_iret * 32)
+#endif
+	CFI_ENDPROC
+
+	# perform work that needs to be done immediately before resumption
+	ALIGN
+	RING0_PTREGS_FRAME		# can't unwind into user space anyway
+work_pending:
+	testb $_TIF_NEED_RESCHED, %cl
+	jz work_notifysig
+work_resched:
+	call schedule
+	DISABLE_INTERRUPTS		# make sure we don't miss an interrupt
+					# setting need_resched or sigpending
+					# between sampling and the iret
+	TRACE_IRQS_OFF
+	movl TI_flags(%ebp), %ecx
+	andl $_TIF_WORK_MASK, %ecx	# is there any work to be done other
+					# than syscall tracing?
+	jz restore_all
+	testb $_TIF_NEED_RESCHED, %cl
+	jnz work_resched
+
+work_notifysig:				# deal with pending signals and
+					# notify-resume requests
+	testl $VM_MASK, EFLAGS(%esp)
+	movl %esp, %eax
+	jne work_notifysig_v86		# returning to kernel-space or
+					# vm86-space
+	xorl %edx, %edx
+	call do_notify_resume
+	jmp resume_userspace_sig
+
+	ALIGN
+work_notifysig_v86:
+#ifdef CONFIG_VM86
+	pushl %ecx			# save ti_flags for do_notify_resume
+	CFI_ADJUST_CFA_OFFSET 4
+	call save_v86_state		# %eax contains pt_regs pointer
+	popl %ecx
+	CFI_ADJUST_CFA_OFFSET -4
+	movl %eax, %esp
+	xorl %edx, %edx
+	call do_notify_resume
+	jmp resume_userspace_sig
+#endif
+
+	# perform syscall exit tracing
+	ALIGN
+syscall_trace_entry:
+	movl $-ENOSYS,EAX(%esp)
+	movl %esp, %eax
+	xorl %edx,%edx
+	call do_syscall_trace
+	cmpl $0, %eax
+	jne resume_userspace		# ret != 0 -> running under PTRACE_SYSEMU,
+					# so must skip actual syscall
+	movl ORIG_EAX(%esp), %eax
+	cmpl $(nr_syscalls), %eax
+	jnae syscall_call
+	jmp syscall_exit
+
+	# perform syscall exit tracing
+	ALIGN
+syscall_exit_work:
+	testb $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SINGLESTEP), %cl
+	jz work_pending
+	TRACE_IRQS_ON
+	ENABLE_INTERRUPTS		# could let do_syscall_trace() call
+					# schedule() instead
+	movl %esp, %eax
+	movl $1, %edx
+	call do_syscall_trace
+	jmp resume_userspace
+	CFI_ENDPROC
+
+	RING0_INT_FRAME			# can't unwind into user space anyway
+syscall_fault:
+	pushl %eax			# save orig_eax
+	CFI_ADJUST_CFA_OFFSET 4
+	SAVE_ALL
+	GET_THREAD_INFO(%ebp)
+	movl $-EFAULT,EAX(%esp)
+	jmp resume_userspace
+
+syscall_badsys:
+	movl $-ENOSYS,EAX(%esp)
+	jmp resume_userspace
+	CFI_ENDPROC
+
+#ifndef CONFIG_XEN
+#define FIXUP_ESPFIX_STACK \
+	movl %esp, %eax; \
+	/* switch to 32bit stack using the pointer on top of 16bit stack */ \
+	lss %ss:CPU_16BIT_STACK_SIZE-8, %esp; \
+	/* copy data from 16bit stack to 32bit stack */ \
+	call fixup_x86_bogus_stack; \
+	/* put ESP to the proper location */ \
+	movl %eax, %esp;
+#define UNWIND_ESPFIX_STACK \
+	pushl %eax; \
+	CFI_ADJUST_CFA_OFFSET 4; \
+	movl %ss, %eax; \
+	/* see if on 16bit stack */ \
+	cmpw $__ESPFIX_SS, %ax; \
+	je 28f; \
+27:	popl %eax; \
+	CFI_ADJUST_CFA_OFFSET -4; \
+.section .fixup,"ax"; \
+28:	movl $__KERNEL_DS, %eax; \
+	movl %eax, %ds; \
+	movl %eax, %es; \
+	/* switch to 32bit stack */ \
+	FIXUP_ESPFIX_STACK; \
+	jmp 27b; \
+.previous
+
+/*
+ * Build the entry stubs and pointer table with
+ * some assembler magic.
+ */
+.data
+ENTRY(interrupt)
+.text
+
+vector=0
+ENTRY(irq_entries_start)
+	RING0_INT_FRAME
+.rept NR_IRQS
+	ALIGN
+ .if vector
+	CFI_ADJUST_CFA_OFFSET -4
+ .endif
+1:	pushl $~(vector)
+	CFI_ADJUST_CFA_OFFSET 4
+	jmp common_interrupt
+.data
+	.long 1b
+.text
+vector=vector+1
+.endr
+
+/*
+ * the CPU automatically disables interrupts when executing an IRQ vector,
+ * so IRQ-flags tracing has to follow that:
+ */
+	ALIGN
+common_interrupt:
+	SAVE_ALL
+	TRACE_IRQS_OFF
+	movl %esp,%eax
+	call do_IRQ
+	jmp ret_from_intr
+	CFI_ENDPROC
+
+#define BUILD_INTERRUPT(name, nr)	\
+ENTRY(name)				\
+	RING0_INT_FRAME;		\
+	pushl $~(nr);			\
+	CFI_ADJUST_CFA_OFFSET 4;	\
+	SAVE_ALL;			\
+	TRACE_IRQS_OFF			\
+	movl %esp,%eax;			\
+	call smp_/**/name;		\
+	jmp ret_from_intr;		\
+	CFI_ENDPROC
+
+/* The include is where all of the SMP etc. interrupts come from */
+#include "entry_arch.h"
+#else
+#define UNWIND_ESPFIX_STACK
+#endif
+
+ENTRY(divide_error)
+	RING0_INT_FRAME
+	pushl $0			# no error code
+	CFI_ADJUST_CFA_OFFSET 4
+	pushl $do_divide_error
+	CFI_ADJUST_CFA_OFFSET 4
+	ALIGN
+error_code:
+	pushl %ds
+	CFI_ADJUST_CFA_OFFSET 4
+	/*CFI_REL_OFFSET ds, 0*/
+	pushl %eax
+	CFI_ADJUST_CFA_OFFSET 4
+	CFI_REL_OFFSET eax, 0
+	xorl %eax, %eax
+	pushl %ebp
+	CFI_ADJUST_CFA_OFFSET 4
+	CFI_REL_OFFSET ebp, 0
+	pushl %edi
+	CFI_ADJUST_CFA_OFFSET 4
+	CFI_REL_OFFSET edi, 0
+	pushl %esi
+	CFI_ADJUST_CFA_OFFSET 4
+	CFI_REL_OFFSET esi, 0
+	pushl %edx
+	CFI_ADJUST_CFA_OFFSET 4
+	CFI_REL_OFFSET edx, 0
+	decl %eax			# eax = -1
+	pushl %ecx
+	CFI_ADJUST_CFA_OFFSET 4
+	CFI_REL_OFFSET ecx, 0
+	pushl %ebx
+	CFI_ADJUST_CFA_OFFSET 4
+	CFI_REL_OFFSET ebx, 0
+	cld
+	pushl %es
+	CFI_ADJUST_CFA_OFFSET 4
+	/*CFI_REL_OFFSET es, 0*/
+	UNWIND_ESPFIX_STACK
+	popl %ecx
+	CFI_ADJUST_CFA_OFFSET -4
+	/*CFI_REGISTER es, ecx*/
+	movl ES(%esp), %edi		# get the function address
+	movl ORIG_EAX(%esp), %edx	# get the error code
+	movl %eax, ORIG_EAX(%esp)
+	movl %ecx, ES(%esp)
+	/*CFI_REL_OFFSET es, ES*/
+	movl $(__USER_DS), %ecx
+	movl %ecx, %ds
+	movl %ecx, %es
+	movl %esp,%eax			# pt_regs pointer
+	call *%edi
+	jmp ret_from_exception
+	CFI_ENDPROC
+
+#ifdef CONFIG_XEN
+# A note on the "critical region" in our callback handler.
+# We want to avoid stacking callback handlers due to events occurring
+# during handling of the last event. To do this, we keep events disabled
+# until we've done all processing. HOWEVER, we must enable events before
+# popping the stack frame (can't be done atomically) and so it would still
+# be possible to get enough handler activations to overflow the stack.
+# Although unlikely, bugs of that kind are hard to track down, so we'd
+# like to avoid the possibility.
+# So, on entry to the handler we detect whether we interrupted an
+# existing activation in its critical region -- if so, we pop the current
+# activation and restart the handler using the previous one.
+#
+# The sysexit critical region is slightly different. sysexit
+# atomically removes the entire stack frame. If we interrupt in the
+# critical region we know that the entire frame is present and correct
+# so we can simply throw away the new one.
+ENTRY(hypervisor_callback)
+	RING0_INT_FRAME
+	pushl %eax
+	CFI_ADJUST_CFA_OFFSET 4
+	SAVE_ALL
+	movl EIP(%esp),%eax
+	cmpl $scrit,%eax
+	jb   11f
+	cmpl $ecrit,%eax
+	jb   critical_region_fixup
+	cmpl $sysexit_scrit,%eax
+	jb   11f
+	cmpl $sysexit_ecrit,%eax
+	ja   11f
+	addl $OLDESP,%esp		# Remove eflags...ebx from stack frame.
+11:	push %esp
+	CFI_ADJUST_CFA_OFFSET 4
+	call evtchn_do_upcall
+	add  $4,%esp
+	CFI_ADJUST_CFA_OFFSET -4
+	jmp  ret_from_intr
+	CFI_ENDPROC
+
+# [How we do the fixup]. We want to merge the current stack frame with the
+# just-interrupted frame. How we do this depends on where in the critical
+# region the interrupted handler was executing, and so how many saved
+# registers are in each frame. We do this quickly using the lookup table
+# 'critical_fixup_table'. For each byte offset in the critical region, it
+# provides the number of bytes which have already been popped from the
+# interrupted stack frame.
+critical_region_fixup:
+	movzbl critical_fixup_table-scrit(%eax),%ecx # %eax contains num bytes popped
+	cmpb $0xff,%cl                  # 0xff => vcpu_info critical region
+	jne  15f
+	xorl %ecx,%ecx
+15:	leal (%esp,%ecx),%esi		# %esi points at end of src region
+	leal OLDESP(%esp),%edi		# %edi points at end of dst region
+	shrl $2,%ecx			# convert words to bytes
+	je   17f			# skip loop if nothing to copy
+16:	subl $4,%esi			# pre-decrementing copy loop
+	subl $4,%edi
+	movl (%esi),%eax
+	movl %eax,(%edi)
+	loop 16b
+17:	movl %edi,%esp			# final %edi is top of merged stack
+	jmp  11b
+
+.section .rodata,"a"
+critical_fixup_table:
+	.byte 0xff,0xff,0xff		# testb $0xff,(%esi) = __TEST_PENDING
+	.byte 0xff,0xff			# jnz  14f
+	.byte 0x00			# pop  %ebx
+	.byte 0x04			# pop  %ecx
+	.byte 0x08			# pop  %edx
+	.byte 0x0c			# pop  %esi
+	.byte 0x10			# pop  %edi
+	.byte 0x14			# pop  %ebp
+	.byte 0x18			# pop  %eax
+	.byte 0x1c			# pop  %ds
+	.byte 0x20			# pop  %es
+	.byte 0x24,0x24,0x24		# add  $4,%esp
+	.byte 0x28			# iret
+	.byte 0xff,0xff,0xff,0xff	# movb $1,1(%esi)
+	.byte 0x00,0x00			# jmp  11b
+.previous
+
+# Hypervisor uses this for application faults while it executes.
+# We get here for two reasons:
+#  1. Fault while reloading DS, ES, FS or GS
+#  2. Fault while executing IRET
+# Category 1 we fix up by reattempting the load, and zeroing the segment
+# register if the load fails.
+# Category 2 we fix up by jumping to do_iret_error. We cannot use the
+# normal Linux return path in this case because if we use the IRET hypercall
+# to pop the stack frame we end up in an infinite loop of failsafe callbacks.
+# We distinguish between categories by maintaining a status value in EAX.
+ENTRY(failsafe_callback)
+	pushl %eax
+	movl $1,%eax
+1:	mov 4(%esp),%ds
+2:	mov 8(%esp),%es
+3:	mov 12(%esp),%fs
+4:	mov 16(%esp),%gs
+	testl %eax,%eax
+	popl %eax
+	jz 5f
+	addl $16,%esp		# EAX != 0 => Category 2 (Bad IRET)
+	jmp iret_exc
+5:	addl $16,%esp		# EAX == 0 => Category 1 (Bad segment)
+	RING0_INT_FRAME
+	pushl $0
+	SAVE_ALL
+	jmp ret_from_exception
+.section .fixup,"ax";		\
+6:	xorl %eax,%eax;		\
+	movl %eax,4(%esp);	\
+	jmp 1b;			\
+7:	xorl %eax,%eax;		\
+	movl %eax,8(%esp);	\
+	jmp 2b;			\
+8:	xorl %eax,%eax;		\
+	movl %eax,12(%esp);	\
+	jmp 3b;			\
+9:	xorl %eax,%eax;		\
+	movl %eax,16(%esp);	\
+	jmp 4b;			\
+.previous;			\
+.section __ex_table,"a";	\
+	.align 4;		\
+	.long 1b,6b;		\
+	.long 2b,7b;		\
+	.long 3b,8b;		\
+	.long 4b,9b;		\
+.previous
+#endif
+	CFI_ENDPROC
+
+ENTRY(coprocessor_error)
+	RING0_INT_FRAME
+	pushl $0
+	CFI_ADJUST_CFA_OFFSET 4
+	pushl $do_coprocessor_error
+	CFI_ADJUST_CFA_OFFSET 4
+	jmp error_code
+	CFI_ENDPROC
+
+ENTRY(simd_coprocessor_error)
+	RING0_INT_FRAME
+	pushl $0
+	CFI_ADJUST_CFA_OFFSET 4
+	pushl $do_simd_coprocessor_error
+	CFI_ADJUST_CFA_OFFSET 4
+	jmp error_code
+	CFI_ENDPROC
+
+ENTRY(device_not_available)
+	RING0_INT_FRAME
+	pushl $-1			# mark this as an int
+	CFI_ADJUST_CFA_OFFSET 4
+	SAVE_ALL
+#ifndef CONFIG_XEN
+	movl %cr0, %eax
+	testl $0x4, %eax		# EM (math emulation bit)
+	je device_available_emulate
+	pushl $0			# temporary storage for ORIG_EIP
+	CFI_ADJUST_CFA_OFFSET 4
+	call math_emulate
+	addl $4, %esp
+	CFI_ADJUST_CFA_OFFSET -4
+	jmp ret_from_exception
+device_available_emulate:
+#endif
+	preempt_stop
+	call math_state_restore
+	jmp ret_from_exception
+	CFI_ENDPROC
+
+#ifndef CONFIG_XEN
+/*
+ * Debug traps and NMI can happen at the one SYSENTER instruction
+ * that sets up the real kernel stack. Check here, since we can't
+ * allow the wrong stack to be used.
+ *
+ * "SYSENTER_stack_esp0+12" is because the NMI/debug handler will have
+ * already pushed 3 words if it hits on the sysenter instruction:
+ * eflags, cs and eip.
+ *
+ * We just load the right stack, and push the three (known) values
+ * by hand onto the new stack - while updating the return eip past
+ * the instruction that would have done it for sysenter.
+ */
+#define FIX_STACK(offset, ok, label)		\
+	cmpw $__KERNEL_CS,4(%esp);		\
+	jne ok;					\
+label:						\
+	movl SYSENTER_stack_esp0+offset(%esp),%esp;	\
+	pushfl;					\
+	pushl $__KERNEL_CS;			\
+	pushl $sysenter_past_esp
+#endif /* CONFIG_XEN */
+
+KPROBE_ENTRY(debug)
+	RING0_INT_FRAME
+#ifndef CONFIG_XEN
+	cmpl $sysenter_entry,(%esp)
+	jne debug_stack_correct
+	FIX_STACK(12, debug_stack_correct, debug_esp_fix_insn)
+debug_stack_correct:
+#endif /* !CONFIG_XEN */
+	pushl $-1			# mark this as an int
+	CFI_ADJUST_CFA_OFFSET 4
+	SAVE_ALL
+	xorl %edx,%edx			# error code 0
+	movl %esp,%eax			# pt_regs pointer
+	call do_debug
+	jmp ret_from_exception
+	CFI_ENDPROC
+	.previous .text
+#ifndef CONFIG_XEN
+/*
+ * NMI is doubly nasty. It can happen _while_ we're handling
+ * a debug fault, and the debug fault hasn't yet been able to
+ * clear up the stack. So we first check whether we got  an
+ * NMI on the sysenter entry path, but after that we need to
+ * check whether we got an NMI on the debug path where the debug
+ * fault happened on the sysenter path.
+ */
+ENTRY(nmi)
+	RING0_INT_FRAME
+	pushl %eax
+	CFI_ADJUST_CFA_OFFSET 4
+	movl %ss, %eax
+	cmpw $__ESPFIX_SS, %ax
+	popl %eax
+	CFI_ADJUST_CFA_OFFSET -4
+	je nmi_16bit_stack
+	cmpl $sysenter_entry,(%esp)
+	je nmi_stack_fixup
+	pushl %eax
+	CFI_ADJUST_CFA_OFFSET 4
+	movl %esp,%eax
+	/* Do not access memory above the end of our stack page,
+	 * it might not exist.
+	 */
+	andl $(THREAD_SIZE-1),%eax
+	cmpl $(THREAD_SIZE-20),%eax
+	popl %eax
+	CFI_ADJUST_CFA_OFFSET -4
+	jae nmi_stack_correct
+	cmpl $sysenter_entry,12(%esp)
+	je nmi_debug_stack_check
+nmi_stack_correct:
+	pushl %eax
+	CFI_ADJUST_CFA_OFFSET 4
+	SAVE_ALL
+	xorl %edx,%edx		# zero error code
+	movl %esp,%eax		# pt_regs pointer
+	call do_nmi
+	jmp restore_nocheck_notrace
+	CFI_ENDPROC
+
+nmi_stack_fixup:
+	FIX_STACK(12,nmi_stack_correct, 1)
+	jmp nmi_stack_correct
+nmi_debug_stack_check:
+	cmpw $__KERNEL_CS,16(%esp)
+	jne nmi_stack_correct
+	cmpl $debug,(%esp)
+	jb nmi_stack_correct
+	cmpl $debug_esp_fix_insn,(%esp)
+	ja nmi_stack_correct
+	FIX_STACK(24,nmi_stack_correct, 1)
+	jmp nmi_stack_correct
+
+nmi_16bit_stack:
+	RING0_INT_FRAME
+	/* create the pointer to lss back */
+	pushl %ss
+	CFI_ADJUST_CFA_OFFSET 4
+	pushl %esp
+	CFI_ADJUST_CFA_OFFSET 4
+	movzwl %sp, %esp
+	addw $4, (%esp)
+	/* copy the iret frame of 12 bytes */
+	.rept 3
+	pushl 16(%esp)
+	CFI_ADJUST_CFA_OFFSET 4
+	.endr
+	pushl %eax
+	CFI_ADJUST_CFA_OFFSET 4
+	SAVE_ALL
+	FIXUP_ESPFIX_STACK		# %eax == %esp
+	CFI_ADJUST_CFA_OFFSET -20	# the frame has now moved
+	xorl %edx,%edx			# zero error code
+	call do_nmi
+	RESTORE_REGS
+	lss 12+4(%esp), %esp		# back to 16bit stack
+1:	iret
+	CFI_ENDPROC
+.section __ex_table,"a"
+	.align 4
+	.long 1b,iret_exc
+.previous
+#else
+ENTRY(nmi)
+	RING0_INT_FRAME
+	pushl %eax
+	CFI_ADJUST_CFA_OFFSET 4
+	SAVE_ALL
+	xorl %edx,%edx		# zero error code
+	movl %esp,%eax		# pt_regs pointer
+	call do_nmi
+	orl  $NMI_MASK, EFLAGS(%esp)
+	jmp restore_all
+	CFI_ENDPROC
+#endif
+
+KPROBE_ENTRY(int3)
+	RING0_INT_FRAME
+	pushl $-1			# mark this as an int
+	CFI_ADJUST_CFA_OFFSET 4
+	SAVE_ALL
+	xorl %edx,%edx		# zero error code
+	movl %esp,%eax		# pt_regs pointer
+	call do_int3
+	jmp ret_from_exception
+	CFI_ENDPROC
+	.previous .text
+
+ENTRY(overflow)
+	RING0_INT_FRAME
+	pushl $0
+	CFI_ADJUST_CFA_OFFSET 4
+	pushl $do_overflow
+	CFI_ADJUST_CFA_OFFSET 4
+	jmp error_code
+	CFI_ENDPROC
+
+ENTRY(bounds)
+	RING0_INT_FRAME
+	pushl $0
+	CFI_ADJUST_CFA_OFFSET 4
+	pushl $do_bounds
+	CFI_ADJUST_CFA_OFFSET 4
+	jmp error_code
+	CFI_ENDPROC
+
+ENTRY(invalid_op)
+	RING0_INT_FRAME
+	pushl $0
+	CFI_ADJUST_CFA_OFFSET 4
+	pushl $do_invalid_op
+	CFI_ADJUST_CFA_OFFSET 4
+	jmp error_code
+	CFI_ENDPROC
+
+ENTRY(coprocessor_segment_overrun)
+	RING0_INT_FRAME
+	pushl $0
+	CFI_ADJUST_CFA_OFFSET 4
+	pushl $do_coprocessor_segment_overrun
+	CFI_ADJUST_CFA_OFFSET 4
+	jmp error_code
+	CFI_ENDPROC
+
+ENTRY(invalid_TSS)
+	RING0_EC_FRAME
+	pushl $do_invalid_TSS
+	CFI_ADJUST_CFA_OFFSET 4
+	jmp error_code
+	CFI_ENDPROC
+
+ENTRY(segment_not_present)
+	RING0_EC_FRAME
+	pushl $do_segment_not_present
+	CFI_ADJUST_CFA_OFFSET 4
+	jmp error_code
+	CFI_ENDPROC
+
+ENTRY(stack_segment)
+	RING0_EC_FRAME
+	pushl $do_stack_segment
+	CFI_ADJUST_CFA_OFFSET 4
+	jmp error_code
+	CFI_ENDPROC
+
+KPROBE_ENTRY(general_protection)
+	RING0_EC_FRAME
+	pushl $do_general_protection
+	CFI_ADJUST_CFA_OFFSET 4
+	jmp error_code
+	CFI_ENDPROC
+	.previous .text
+
+ENTRY(alignment_check)
+	RING0_EC_FRAME
+	pushl $do_alignment_check
+	CFI_ADJUST_CFA_OFFSET 4
+	jmp error_code
+	CFI_ENDPROC
+
+KPROBE_ENTRY(page_fault)
+	RING0_EC_FRAME
+	pushl $do_page_fault
+	CFI_ADJUST_CFA_OFFSET 4
+	jmp error_code
+	CFI_ENDPROC
+	.previous .text
+
+#ifdef CONFIG_X86_MCE
+ENTRY(machine_check)
+	RING0_INT_FRAME
+	pushl $0
+	CFI_ADJUST_CFA_OFFSET 4
+	pushl machine_check_vector
+	CFI_ADJUST_CFA_OFFSET 4
+	jmp error_code
+	CFI_ENDPROC
+#endif
+
+#ifndef CONFIG_XEN
+ENTRY(spurious_interrupt_bug)
+	RING0_INT_FRAME
+	pushl $0
+	CFI_ADJUST_CFA_OFFSET 4
+	pushl $do_spurious_interrupt_bug
+	CFI_ADJUST_CFA_OFFSET 4
+	jmp error_code
+	CFI_ENDPROC
+#endif /* !CONFIG_XEN */
+
+#ifdef CONFIG_STACK_UNWIND
+ENTRY(arch_unwind_init_running)
+	CFI_STARTPROC
+	movl	4(%esp), %edx
+	movl	(%esp), %ecx
+	leal	4(%esp), %eax
+	movl	%ebx, EBX(%edx)
+	xorl	%ebx, %ebx
+	movl	%ebx, ECX(%edx)
+	movl	%ebx, EDX(%edx)
+	movl	%esi, ESI(%edx)
+	movl	%edi, EDI(%edx)
+	movl	%ebp, EBP(%edx)
+	movl	%ebx, EAX(%edx)
+	movl	$__USER_DS, DS(%edx)
+	movl	$__USER_DS, ES(%edx)
+	movl	%ebx, ORIG_EAX(%edx)
+	movl	%ecx, EIP(%edx)
+	movl	12(%esp), %ecx
+	movl	$__KERNEL_CS, CS(%edx)
+	movl	%ebx, EFLAGS(%edx)
+	movl	%eax, OLDESP(%edx)
+	movl	8(%esp), %eax
+	movl	%ecx, 8(%esp)
+	movl	EBX(%edx), %ebx
+	movl	$__KERNEL_DS, OLDSS(%edx)
+	jmpl	*%eax
+	CFI_ENDPROC
+ENDPROC(arch_unwind_init_running)
+#endif
+
+ENTRY(fixup_4gb_segment)
+	RING0_EC_FRAME
+	pushl $do_fixup_4gb_segment
+	CFI_ADJUST_CFA_OFFSET 4
+	jmp error_code
+	CFI_ENDPROC
+
+.section .rodata,"a"
+#include "syscall_table.S"
+
+syscall_table_size=(.-sys_call_table)
diff -rpuN linux-2.6.18.8/arch/i386/kernel/fixup.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/fixup.c
--- linux-2.6.18.8/arch/i386/kernel/fixup.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/fixup.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,88 @@
+/******************************************************************************
+ * fixup.c
+ * 
+ * Binary-rewriting of certain IA32 instructions, on notification by Xen.
+ * Used to avoid repeated slow emulation of common instructions used by the
+ * user-space TLS (Thread-Local Storage) libraries.
+ * 
+ * **** NOTE ****
+ *  Issues with the binary rewriting have caused it to be removed. Instead
+ *  we rely on Xen's emulator to boot the kernel, and then print a banner
+ *  message recommending that the user disables /lib/tls.
+ * 
+ * Copyright (c) 2004, K A Fraser
+ * 
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ * 
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ * 
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#include <linux/init.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/kernel.h>
+#include <linux/delay.h>
+#include <linux/version.h>
+
+#define DP(_f, _args...) printk(KERN_ALERT "  " _f "\n" , ## _args )
+
+fastcall void do_fixup_4gb_segment(struct pt_regs *regs, long error_code)
+{
+	static unsigned long printed = 0;
+	char info[100];
+	int i;
+
+	/* Ignore statically-linked init. */
+	if (current->tgid == 1)
+		return;
+            
+	VOID(HYPERVISOR_vm_assist(VMASST_CMD_disable,
+				  VMASST_TYPE_4gb_segments_notify));
+
+	if (test_and_set_bit(0, &printed))
+		return;
+
+	sprintf(info, "%s (pid=%d)", current->comm, current->tgid);
+
+	DP("");
+	DP("***************************************************************");
+	DP("***************************************************************");
+	DP("** WARNING: Currently emulating unsupported memory accesses  **");
+	DP("**          in /lib/tls glibc libraries. The emulation is    **");
+	DP("**          slow. To ensure full performance you should      **");
+	DP("**          install a 'xen-friendly' (nosegneg) version of   **");
+	DP("**          the library, or disable tls support by executing **");
+	DP("**          the following as root:                           **");
+	DP("**          mv /lib/tls /lib/tls.disabled                    **");
+	DP("** Offending process: %-38.38s **", info);
+	DP("***************************************************************");
+	DP("***************************************************************");
+	DP("");
+
+	for (i = 5; i > 0; i--) {
+		touch_softlockup_watchdog();
+		printk("Pausing... %d", i);
+		mdelay(1000);
+		printk("\b\b\b\b\b\b\b\b\b\b\b\b");
+	}
+
+	printk("Continuing...\n\n");
+}
+
+static int __init fixup_init(void)
+{
+	WARN_ON(HYPERVISOR_vm_assist(VMASST_CMD_enable,
+				     VMASST_TYPE_4gb_segments_notify));
+	return 0;
+}
+__initcall(fixup_init);
diff -rpuN linux-2.6.18.8/arch/i386/kernel/head-xen.S linux-2.6.18-xen-3.3.0/arch/i386/kernel/head-xen.S
--- linux-2.6.18.8/arch/i386/kernel/head-xen.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/head-xen.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,207 @@
+
+
+.text
+#include <linux/elfnote.h>
+#include <linux/threads.h>
+#include <linux/linkage.h>
+#include <asm/segment.h>
+#include <asm/page.h>
+#include <asm/cache.h>
+#include <asm/thread_info.h>
+#include <asm/asm-offsets.h>
+#include <asm/dwarf2.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/elfnote.h>
+
+/*
+ * References to members of the new_cpu_data structure.
+ */
+
+#define X86		new_cpu_data+CPUINFO_x86
+#define X86_VENDOR	new_cpu_data+CPUINFO_x86_vendor
+#define X86_MODEL	new_cpu_data+CPUINFO_x86_model
+#define X86_MASK	new_cpu_data+CPUINFO_x86_mask
+#define X86_HARD_MATH	new_cpu_data+CPUINFO_hard_math
+#define X86_CPUID	new_cpu_data+CPUINFO_cpuid_level
+#define X86_CAPABILITY	new_cpu_data+CPUINFO_x86_capability
+#define X86_VENDOR_ID	new_cpu_data+CPUINFO_x86_vendor_id
+
+#define VIRT_ENTRY_OFFSET 0x0
+.org VIRT_ENTRY_OFFSET
+ENTRY(startup_32)
+	movl %esi,xen_start_info
+	cld
+
+	/* Set up the stack pointer */
+	movl $(init_thread_union+THREAD_SIZE),%esp
+
+	/* get vendor info */
+	xorl %eax,%eax			# call CPUID with 0 -> return vendor ID
+	XEN_CPUID
+	movl %eax,X86_CPUID		# save CPUID level
+	movl %ebx,X86_VENDOR_ID		# lo 4 chars
+	movl %edx,X86_VENDOR_ID+4	# next 4 chars
+	movl %ecx,X86_VENDOR_ID+8	# last 4 chars
+
+	movl $1,%eax		# Use the CPUID instruction to get CPU type
+	XEN_CPUID
+	movb %al,%cl		# save reg for future use
+	andb $0x0f,%ah		# mask processor family
+	movb %ah,X86
+	andb $0xf0,%al		# mask model
+	shrb $4,%al
+	movb %al,X86_MODEL
+	andb $0x0f,%cl		# mask mask revision
+	movb %cl,X86_MASK
+	movl %edx,X86_CAPABILITY
+
+	movb $1,X86_HARD_MATH
+
+	xorl %eax,%eax			# Clear FS/GS and LDT
+	movl %eax,%fs
+	movl %eax,%gs
+	cld			# gcc2 wants the direction flag cleared at all times
+
+	pushl %eax		# fake return address
+	jmp start_kernel
+
+#define HYPERCALL_PAGE_OFFSET 0x1000
+.org HYPERCALL_PAGE_OFFSET
+ENTRY(hypercall_page)
+	CFI_STARTPROC
+.skip 0x1000
+	CFI_ENDPROC
+
+/*
+ * Real beginning of normal "text" segment
+ */
+ENTRY(stext)
+ENTRY(_stext)
+
+/*
+ * BSS section
+ */
+.section ".bss.page_aligned","w"
+ENTRY(empty_zero_page)
+	.fill 4096,1,0
+
+/*
+ * This starts the data section.
+ */
+.data
+
+/*
+ * The Global Descriptor Table contains 28 quadwords, per-CPU.
+ */
+	.align L1_CACHE_BYTES
+ENTRY(cpu_gdt_table)
+	.quad 0x0000000000000000	/* NULL descriptor */
+	.quad 0x0000000000000000	/* 0x0b reserved */
+	.quad 0x0000000000000000	/* 0x13 reserved */
+	.quad 0x0000000000000000	/* 0x1b reserved */
+	.quad 0x0000000000000000	/* 0x20 unused */
+	.quad 0x0000000000000000	/* 0x28 unused */
+	.quad 0x0000000000000000	/* 0x33 TLS entry 1 */
+	.quad 0x0000000000000000	/* 0x3b TLS entry 2 */
+	.quad 0x0000000000000000	/* 0x43 TLS entry 3 */
+	.quad 0x0000000000000000	/* 0x4b reserved */
+	.quad 0x0000000000000000	/* 0x53 reserved */
+	.quad 0x0000000000000000	/* 0x5b reserved */
+
+	.quad 0x00cf9a000000ffff	/* 0x60 kernel 4GB code at 0x00000000 */
+	.quad 0x00cf92000000ffff	/* 0x68 kernel 4GB data at 0x00000000 */
+	.quad 0x00cffa000000ffff	/* 0x73 user 4GB code at 0x00000000 */
+	.quad 0x00cff2000000ffff	/* 0x7b user 4GB data at 0x00000000 */
+
+	.quad 0x0000000000000000	/* 0x80 TSS descriptor */
+	.quad 0x0000000000000000	/* 0x88 LDT descriptor */
+
+	/*
+	 * Segments used for calling PnP BIOS have byte granularity.
+	 * They code segments and data segments have fixed 64k limits,
+	 * the transfer segment sizes are set at run time.
+	 */
+	.quad 0x0000000000000000	/* 0x90 32-bit code */
+	.quad 0x0000000000000000	/* 0x98 16-bit code */
+	.quad 0x0000000000000000	/* 0xa0 16-bit data */
+	.quad 0x0000000000000000	/* 0xa8 16-bit data */
+	.quad 0x0000000000000000	/* 0xb0 16-bit data */
+
+	/*
+	 * The APM segments have byte granularity and their bases
+	 * are set at run time.  All have 64k limits.
+	 */
+	.quad 0x0000000000000000	/* 0xb8 APM CS    code */
+	.quad 0x0000000000000000	/* 0xc0 APM CS 16 code (16 bit) */
+	.quad 0x0000000000000000	/* 0xc8 APM DS    data */
+
+	.quad 0x0000000000000000	/* 0xd0 - ESPFIX 16-bit SS */
+	.quad 0x0000000000000000	/* 0xd8 - unused */
+	.quad 0x0000000000000000	/* 0xe0 - unused */
+	.quad 0x0000000000000000	/* 0xe8 - unused */
+	.quad 0x0000000000000000	/* 0xf0 - unused */
+	.quad 0x0000000000000000	/* 0xf8 - GDT entry 31: double-fault TSS */
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+/*
+ * __xen_guest information
+ */
+.macro utoa value
+ .if (\value) < 0 || (\value) >= 0x10
+	utoa (((\value)>>4)&0x0fffffff)
+ .endif
+ .if ((\value) & 0xf) < 10
+  .byte '0' + ((\value) & 0xf)
+ .else
+  .byte 'A' + ((\value) & 0xf) - 10
+ .endif
+.endm
+
+.section __xen_guest
+	.ascii	"GUEST_OS=linux,GUEST_VER=2.6"
+	.ascii	",XEN_VER=xen-3.0"
+	.ascii	",VIRT_BASE=0x"
+		utoa __PAGE_OFFSET
+	.ascii	",ELF_PADDR_OFFSET=0x"
+		utoa __PAGE_OFFSET
+	.ascii	",VIRT_ENTRY=0x"
+		utoa (__PAGE_OFFSET + __PHYSICAL_START + VIRT_ENTRY_OFFSET)
+	.ascii	",HYPERCALL_PAGE=0x"
+		utoa ((__PHYSICAL_START+HYPERCALL_PAGE_OFFSET)>>PAGE_SHIFT)
+	.ascii  ",FEATURES=writable_page_tables"
+	.ascii	         "|writable_descriptor_tables"
+	.ascii	         "|auto_translated_physmap"
+	.ascii	         "|pae_pgdir_above_4gb"
+	.ascii	         "|supervisor_mode_kernel"
+#ifdef CONFIG_X86_PAE
+	.ascii	",PAE=yes[extended-cr3]"
+#else
+	.ascii	",PAE=no"
+#endif
+	.ascii	",LOADER=generic"
+	.byte	0
+#endif /* CONFIG_XEN_COMPAT <= 0x030002 */
+
+
+	ELFNOTE(Xen, XEN_ELFNOTE_GUEST_OS,       .asciz, "linux")	
+	ELFNOTE(Xen, XEN_ELFNOTE_GUEST_VERSION,  .asciz, "2.6")
+	ELFNOTE(Xen, XEN_ELFNOTE_XEN_VERSION,    .asciz, "xen-3.0")
+	ELFNOTE(Xen, XEN_ELFNOTE_VIRT_BASE,      .long,  __PAGE_OFFSET)
+#if CONFIG_XEN_COMPAT <= 0x030002
+	ELFNOTE(Xen, XEN_ELFNOTE_PADDR_OFFSET,   .long,  __PAGE_OFFSET)
+#else
+	ELFNOTE(Xen, XEN_ELFNOTE_PADDR_OFFSET,   .long,  0)
+#endif
+	ELFNOTE(Xen, XEN_ELFNOTE_ENTRY,          .long,  startup_32)
+	ELFNOTE(Xen, XEN_ELFNOTE_HYPERCALL_PAGE, .long,  hypercall_page)
+	ELFNOTE(Xen, XEN_ELFNOTE_HV_START_LOW,   .long,  HYPERVISOR_VIRT_START)
+	ELFNOTE(Xen, XEN_ELFNOTE_FEATURES,       .asciz, "writable_page_tables|writable_descriptor_tables|auto_translated_physmap|pae_pgdir_above_4gb|supervisor_mode_kernel")
+#ifdef CONFIG_X86_PAE
+	ELFNOTE(Xen, XEN_ELFNOTE_PAE_MODE,       .asciz, "yes")
+	ELFNOTE(Xen, XEN_ELFNOTE_L1_MFN_VALID,   .quad,  _PAGE_PRESENT,_PAGE_PRESENT)
+#else
+	ELFNOTE(Xen, XEN_ELFNOTE_PAE_MODE,       .asciz, "no")
+	ELFNOTE(Xen, XEN_ELFNOTE_L1_MFN_VALID,   .long,  _PAGE_PRESENT,_PAGE_PRESENT)
+#endif
+	ELFNOTE(Xen, XEN_ELFNOTE_LOADER,         .asciz, "generic")
+	ELFNOTE(Xen, XEN_ELFNOTE_SUSPEND_CANCEL, .long,  1)
diff -rpuN linux-2.6.18.8/arch/i386/kernel/init_task-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/init_task-xen.c
--- linux-2.6.18.8/arch/i386/kernel/init_task-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/init_task-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,51 @@
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/sched.h>
+#include <linux/init.h>
+#include <linux/init_task.h>
+#include <linux/fs.h>
+#include <linux/mqueue.h>
+
+#include <asm/uaccess.h>
+#include <asm/pgtable.h>
+#include <asm/desc.h>
+
+static struct fs_struct init_fs = INIT_FS;
+static struct files_struct init_files = INIT_FILES;
+static struct signal_struct init_signals = INIT_SIGNALS(init_signals);
+static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand);
+
+#define swapper_pg_dir ((pgd_t *)NULL)
+struct mm_struct init_mm = INIT_MM(init_mm);
+#undef swapper_pg_dir
+
+EXPORT_SYMBOL(init_mm);
+
+/*
+ * Initial thread structure.
+ *
+ * We need to make sure that this is THREAD_SIZE aligned due to the
+ * way process stacks are handled. This is done by having a special
+ * "init_task" linker map entry..
+ */
+union thread_union init_thread_union 
+	__attribute__((__section__(".data.init_task"))) =
+		{ INIT_THREAD_INFO(init_task) };
+
+/*
+ * Initial task structure.
+ *
+ * All other task structs will be allocated on slabs in fork.c
+ */
+struct task_struct init_task = INIT_TASK(init_task);
+
+EXPORT_SYMBOL(init_task);
+
+#ifndef CONFIG_X86_NO_TSS
+/*
+ * per-CPU TSS segments. Threads are completely 'soft' on Linux,
+ * no more per-task TSS's.
+ */ 
+DEFINE_PER_CPU(struct tss_struct, init_tss) ____cacheline_internodealigned_in_smp = INIT_TSS;
+#endif
+
diff -rpuN linux-2.6.18.8/arch/i386/kernel/io_apic-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/io_apic-xen.c
--- linux-2.6.18.8/arch/i386/kernel/io_apic-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/io_apic-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,2770 @@
+/*
+ *	Intel IO-APIC support for multi-Pentium hosts.
+ *
+ *	Copyright (C) 1997, 1998, 1999, 2000 Ingo Molnar, Hajnalka Szabo
+ *
+ *	Many thanks to Stig Venaas for trying out countless experimental
+ *	patches and reporting/debugging problems patiently!
+ *
+ *	(c) 1999, Multiple IO-APIC support, developed by
+ *	Ken-ichi Yaku <yaku@css1.kbnes.nec.co.jp> and
+ *      Hidemi Kishimoto <kisimoto@css1.kbnes.nec.co.jp>,
+ *	further tested and cleaned up by Zach Brown <zab@redhat.com>
+ *	and Ingo Molnar <mingo@redhat.com>
+ *
+ *	Fixes
+ *	Maciej W. Rozycki	:	Bits for genuine 82489DX APICs;
+ *					thanks to Eric Gilmore
+ *					and Rolf G. Tews
+ *					for testing these extensively
+ *	Paul Diefenbaugh	:	Added full ACPI support
+ */
+
+#include <linux/mm.h>
+#include <linux/interrupt.h>
+#include <linux/init.h>
+#include <linux/delay.h>
+#include <linux/sched.h>
+#include <linux/smp_lock.h>
+#include <linux/mc146818rtc.h>
+#include <linux/compiler.h>
+#include <linux/acpi.h>
+#include <linux/module.h>
+#include <linux/sysdev.h>
+
+#include <asm/io.h>
+#include <asm/smp.h>
+#include <asm/desc.h>
+#include <asm/timer.h>
+#include <asm/i8259.h>
+#include <asm/nmi.h>
+
+#include <mach_apic.h>
+
+#include "io_ports.h"
+
+#ifdef CONFIG_XEN
+
+#include <xen/interface/xen.h>
+#include <xen/interface/physdev.h>
+
+/* Fake i8259 */
+#define make_8259A_irq(_irq)     (io_apic_irqs &= ~(1UL<<(_irq)))
+#define disable_8259A_irq(_irq)  ((void)0)
+#define i8259A_irq_pending(_irq) (0)
+
+unsigned long io_apic_irqs;
+
+static inline unsigned int xen_io_apic_read(unsigned int apic, unsigned int reg)
+{
+	struct physdev_apic apic_op;
+	int ret;
+
+	apic_op.apic_physbase = mp_ioapics[apic].mpc_apicaddr;
+	apic_op.reg = reg;
+	ret = HYPERVISOR_physdev_op(PHYSDEVOP_apic_read, &apic_op);
+	if (ret)
+		return ret;
+	return apic_op.value;
+}
+
+static inline void xen_io_apic_write(unsigned int apic, unsigned int reg, unsigned int value)
+{
+	struct physdev_apic apic_op;
+
+	apic_op.apic_physbase = mp_ioapics[apic].mpc_apicaddr;
+	apic_op.reg = reg;
+	apic_op.value = value;
+	WARN_ON(HYPERVISOR_physdev_op(PHYSDEVOP_apic_write, &apic_op));
+}
+
+#define io_apic_read(a,r)    xen_io_apic_read(a,r)
+#define io_apic_write(a,r,v) xen_io_apic_write(a,r,v)
+
+#endif /* CONFIG_XEN */
+
+int (*ioapic_renumber_irq)(int ioapic, int irq);
+atomic_t irq_mis_count;
+
+/* Where if anywhere is the i8259 connect in external int mode */
+static struct { int pin, apic; } ioapic_i8259 = { -1, -1 };
+
+static DEFINE_SPINLOCK(ioapic_lock);
+static DEFINE_SPINLOCK(vector_lock);
+
+int timer_over_8254 __initdata = 1;
+
+/*
+ *	Is the SiS APIC rmw bug present ?
+ *	-1 = don't know, 0 = no, 1 = yes
+ */
+int sis_apic_bug = -1;
+
+/*
+ * # of IRQ routing registers
+ */
+int nr_ioapic_registers[MAX_IO_APICS];
+
+int disable_timer_pin_1 __initdata;
+
+/*
+ * Rough estimation of how many shared IRQs there are, can
+ * be changed anytime.
+ */
+#define MAX_PLUS_SHARED_IRQS NR_IRQS
+#define PIN_MAP_SIZE (MAX_PLUS_SHARED_IRQS + NR_IRQS)
+
+/*
+ * This is performance-critical, we want to do it O(1)
+ *
+ * the indexing order of this array favors 1:1 mappings
+ * between pins and IRQs.
+ */
+
+static struct irq_pin_list {
+	int apic, pin, next;
+} irq_2_pin[PIN_MAP_SIZE];
+
+int vector_irq[NR_VECTORS] __read_mostly = { [0 ... NR_VECTORS - 1] = -1};
+#ifdef CONFIG_PCI_MSI
+#define vector_to_irq(vector) 	\
+	(platform_legacy_irq(vector) ? vector : vector_irq[vector])
+#else
+#define vector_to_irq(vector)	(vector)
+#endif
+
+/*
+ * The common case is 1:1 IRQ<->pin mappings. Sometimes there are
+ * shared ISA-space IRQs, so we have to support them. We are super
+ * fast in the common case, and fast for shared ISA-space IRQs.
+ */
+static void add_pin_to_irq(unsigned int irq, int apic, int pin)
+{
+	static int first_free_entry = NR_IRQS;
+	struct irq_pin_list *entry = irq_2_pin + irq;
+
+	while (entry->next)
+		entry = irq_2_pin + entry->next;
+
+	if (entry->pin != -1) {
+		entry->next = first_free_entry;
+		entry = irq_2_pin + entry->next;
+		if (++first_free_entry >= PIN_MAP_SIZE)
+			panic("io_apic.c: whoops");
+	}
+	entry->apic = apic;
+	entry->pin = pin;
+}
+
+#ifdef CONFIG_XEN
+#define clear_IO_APIC() ((void)0)
+#else
+/*
+ * Reroute an IRQ to a different pin.
+ */
+static void __init replace_pin_at_irq(unsigned int irq,
+				      int oldapic, int oldpin,
+				      int newapic, int newpin)
+{
+	struct irq_pin_list *entry = irq_2_pin + irq;
+
+	while (1) {
+		if (entry->apic == oldapic && entry->pin == oldpin) {
+			entry->apic = newapic;
+			entry->pin = newpin;
+		}
+		if (!entry->next)
+			break;
+		entry = irq_2_pin + entry->next;
+	}
+}
+
+static void __modify_IO_APIC_irq (unsigned int irq, unsigned long enable, unsigned long disable)
+{
+	struct irq_pin_list *entry = irq_2_pin + irq;
+	unsigned int pin, reg;
+
+	for (;;) {
+		pin = entry->pin;
+		if (pin == -1)
+			break;
+		reg = io_apic_read(entry->apic, 0x10 + pin*2);
+		reg &= ~disable;
+		reg |= enable;
+		io_apic_modify(entry->apic, 0x10 + pin*2, reg);
+		if (!entry->next)
+			break;
+		entry = irq_2_pin + entry->next;
+	}
+}
+
+/* mask = 1 */
+static void __mask_IO_APIC_irq (unsigned int irq)
+{
+	__modify_IO_APIC_irq(irq, 0x00010000, 0);
+}
+
+/* mask = 0 */
+static void __unmask_IO_APIC_irq (unsigned int irq)
+{
+	__modify_IO_APIC_irq(irq, 0, 0x00010000);
+}
+
+/* mask = 1, trigger = 0 */
+static void __mask_and_edge_IO_APIC_irq (unsigned int irq)
+{
+	__modify_IO_APIC_irq(irq, 0x00010000, 0x00008000);
+}
+
+/* mask = 0, trigger = 1 */
+static void __unmask_and_level_IO_APIC_irq (unsigned int irq)
+{
+	__modify_IO_APIC_irq(irq, 0x00008000, 0x00010000);
+}
+
+static void mask_IO_APIC_irq (unsigned int irq)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	__mask_IO_APIC_irq(irq);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+}
+
+static void unmask_IO_APIC_irq (unsigned int irq)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	__unmask_IO_APIC_irq(irq);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+}
+
+static void clear_IO_APIC_pin(unsigned int apic, unsigned int pin)
+{
+	struct IO_APIC_route_entry entry;
+	unsigned long flags;
+	
+	/* Check delivery_mode to be sure we're not clearing an SMI pin */
+	spin_lock_irqsave(&ioapic_lock, flags);
+	*(((int*)&entry) + 0) = io_apic_read(apic, 0x10 + 2 * pin);
+	*(((int*)&entry) + 1) = io_apic_read(apic, 0x11 + 2 * pin);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+	if (entry.delivery_mode == dest_SMI)
+		return;
+
+	/*
+	 * Disable it in the IO-APIC irq-routing table:
+	 */
+	memset(&entry, 0, sizeof(entry));
+	entry.mask = 1;
+	spin_lock_irqsave(&ioapic_lock, flags);
+	io_apic_write(apic, 0x10 + 2 * pin, *(((int *)&entry) + 0));
+	io_apic_write(apic, 0x11 + 2 * pin, *(((int *)&entry) + 1));
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+}
+
+static void clear_IO_APIC (void)
+{
+	int apic, pin;
+
+	for (apic = 0; apic < nr_ioapics; apic++)
+		for (pin = 0; pin < nr_ioapic_registers[apic]; pin++)
+			clear_IO_APIC_pin(apic, pin);
+}
+
+#ifdef CONFIG_SMP
+static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t cpumask)
+{
+	unsigned long flags;
+	int pin;
+	struct irq_pin_list *entry = irq_2_pin + irq;
+	unsigned int apicid_value;
+	cpumask_t tmp;
+	
+	cpus_and(tmp, cpumask, cpu_online_map);
+	if (cpus_empty(tmp))
+		tmp = TARGET_CPUS;
+
+	cpus_and(cpumask, tmp, CPU_MASK_ALL);
+
+	apicid_value = cpu_mask_to_apicid(cpumask);
+	/* Prepare to do the io_apic_write */
+	apicid_value = apicid_value << 24;
+	spin_lock_irqsave(&ioapic_lock, flags);
+	for (;;) {
+		pin = entry->pin;
+		if (pin == -1)
+			break;
+		io_apic_write(entry->apic, 0x10 + 1 + pin*2, apicid_value);
+		if (!entry->next)
+			break;
+		entry = irq_2_pin + entry->next;
+	}
+	set_irq_info(irq, cpumask);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+}
+
+#if defined(CONFIG_IRQBALANCE)
+# include <asm/processor.h>	/* kernel_thread() */
+# include <linux/kernel_stat.h>	/* kstat */
+# include <linux/slab.h>		/* kmalloc() */
+# include <linux/timer.h>	/* time_after() */
+ 
+#ifdef CONFIG_BALANCED_IRQ_DEBUG
+#  define TDprintk(x...) do { printk("<%ld:%s:%d>: ", jiffies, __FILE__, __LINE__); printk(x); } while (0)
+#  define Dprintk(x...) do { TDprintk(x); } while (0)
+# else
+#  define TDprintk(x...) 
+#  define Dprintk(x...) 
+# endif
+
+#define IRQBALANCE_CHECK_ARCH -999
+#define MAX_BALANCED_IRQ_INTERVAL	(5*HZ)
+#define MIN_BALANCED_IRQ_INTERVAL	(HZ/2)
+#define BALANCED_IRQ_MORE_DELTA		(HZ/10)
+#define BALANCED_IRQ_LESS_DELTA		(HZ)
+
+static int irqbalance_disabled __read_mostly = IRQBALANCE_CHECK_ARCH;
+static int physical_balance __read_mostly;
+static long balanced_irq_interval __read_mostly = MAX_BALANCED_IRQ_INTERVAL;
+
+static struct irq_cpu_info {
+	unsigned long * last_irq;
+	unsigned long * irq_delta;
+	unsigned long irq;
+} irq_cpu_data[NR_CPUS];
+
+#define CPU_IRQ(cpu)		(irq_cpu_data[cpu].irq)
+#define LAST_CPU_IRQ(cpu,irq)   (irq_cpu_data[cpu].last_irq[irq])
+#define IRQ_DELTA(cpu,irq) 	(irq_cpu_data[cpu].irq_delta[irq])
+
+#define IDLE_ENOUGH(cpu,now) \
+	(idle_cpu(cpu) && ((now) - per_cpu(irq_stat, (cpu)).idle_timestamp > 1))
+
+#define IRQ_ALLOWED(cpu, allowed_mask)	cpu_isset(cpu, allowed_mask)
+
+#define CPU_TO_PACKAGEINDEX(i) (first_cpu(cpu_sibling_map[i]))
+
+static cpumask_t balance_irq_affinity[NR_IRQS] = {
+	[0 ... NR_IRQS-1] = CPU_MASK_ALL
+};
+
+void set_balance_irq_affinity(unsigned int irq, cpumask_t mask)
+{
+	balance_irq_affinity[irq] = mask;
+}
+
+static unsigned long move(int curr_cpu, cpumask_t allowed_mask,
+			unsigned long now, int direction)
+{
+	int search_idle = 1;
+	int cpu = curr_cpu;
+
+	goto inside;
+
+	do {
+		if (unlikely(cpu == curr_cpu))
+			search_idle = 0;
+inside:
+		if (direction == 1) {
+			cpu++;
+			if (cpu >= NR_CPUS)
+				cpu = 0;
+		} else {
+			cpu--;
+			if (cpu == -1)
+				cpu = NR_CPUS-1;
+		}
+	} while (!cpu_online(cpu) || !IRQ_ALLOWED(cpu,allowed_mask) ||
+			(search_idle && !IDLE_ENOUGH(cpu,now)));
+
+	return cpu;
+}
+
+static inline void balance_irq(int cpu, int irq)
+{
+	unsigned long now = jiffies;
+	cpumask_t allowed_mask;
+	unsigned int new_cpu;
+		
+	if (irqbalance_disabled)
+		return; 
+
+	cpus_and(allowed_mask, cpu_online_map, balance_irq_affinity[irq]);
+	new_cpu = move(cpu, allowed_mask, now, 1);
+	if (cpu != new_cpu) {
+		set_pending_irq(irq, cpumask_of_cpu(new_cpu));
+	}
+}
+
+static inline void rotate_irqs_among_cpus(unsigned long useful_load_threshold)
+{
+	int i, j;
+	Dprintk("Rotating IRQs among CPUs.\n");
+	for_each_online_cpu(i) {
+		for (j = 0; j < NR_IRQS; j++) {
+			if (!irq_desc[j].action)
+				continue;
+			/* Is it a significant load ?  */
+			if (IRQ_DELTA(CPU_TO_PACKAGEINDEX(i),j) <
+						useful_load_threshold)
+				continue;
+			balance_irq(i, j);
+		}
+	}
+	balanced_irq_interval = max((long)MIN_BALANCED_IRQ_INTERVAL,
+		balanced_irq_interval - BALANCED_IRQ_LESS_DELTA);	
+	return;
+}
+
+static void do_irq_balance(void)
+{
+	int i, j;
+	unsigned long max_cpu_irq = 0, min_cpu_irq = (~0);
+	unsigned long move_this_load = 0;
+	int max_loaded = 0, min_loaded = 0;
+	int load;
+	unsigned long useful_load_threshold = balanced_irq_interval + 10;
+	int selected_irq;
+	int tmp_loaded, first_attempt = 1;
+	unsigned long tmp_cpu_irq;
+	unsigned long imbalance = 0;
+	cpumask_t allowed_mask, target_cpu_mask, tmp;
+
+	for_each_possible_cpu(i) {
+		int package_index;
+		CPU_IRQ(i) = 0;
+		if (!cpu_online(i))
+			continue;
+		package_index = CPU_TO_PACKAGEINDEX(i);
+		for (j = 0; j < NR_IRQS; j++) {
+			unsigned long value_now, delta;
+			/* Is this an active IRQ? */
+			if (!irq_desc[j].action)
+				continue;
+			if ( package_index == i )
+				IRQ_DELTA(package_index,j) = 0;
+			/* Determine the total count per processor per IRQ */
+			value_now = (unsigned long) kstat_cpu(i).irqs[j];
+
+			/* Determine the activity per processor per IRQ */
+			delta = value_now - LAST_CPU_IRQ(i,j);
+
+			/* Update last_cpu_irq[][] for the next time */
+			LAST_CPU_IRQ(i,j) = value_now;
+
+			/* Ignore IRQs whose rate is less than the clock */
+			if (delta < useful_load_threshold)
+				continue;
+			/* update the load for the processor or package total */
+			IRQ_DELTA(package_index,j) += delta;
+
+			/* Keep track of the higher numbered sibling as well */
+			if (i != package_index)
+				CPU_IRQ(i) += delta;
+			/*
+			 * We have sibling A and sibling B in the package
+			 *
+			 * cpu_irq[A] = load for cpu A + load for cpu B
+			 * cpu_irq[B] = load for cpu B
+			 */
+			CPU_IRQ(package_index) += delta;
+		}
+	}
+	/* Find the least loaded processor package */
+	for_each_online_cpu(i) {
+		if (i != CPU_TO_PACKAGEINDEX(i))
+			continue;
+		if (min_cpu_irq > CPU_IRQ(i)) {
+			min_cpu_irq = CPU_IRQ(i);
+			min_loaded = i;
+		}
+	}
+	max_cpu_irq = ULONG_MAX;
+
+tryanothercpu:
+	/* Look for heaviest loaded processor.
+	 * We may come back to get the next heaviest loaded processor.
+	 * Skip processors with trivial loads.
+	 */
+	tmp_cpu_irq = 0;
+	tmp_loaded = -1;
+	for_each_online_cpu(i) {
+		if (i != CPU_TO_PACKAGEINDEX(i))
+			continue;
+		if (max_cpu_irq <= CPU_IRQ(i)) 
+			continue;
+		if (tmp_cpu_irq < CPU_IRQ(i)) {
+			tmp_cpu_irq = CPU_IRQ(i);
+			tmp_loaded = i;
+		}
+	}
+
+	if (tmp_loaded == -1) {
+ 	 /* In the case of small number of heavy interrupt sources, 
+	  * loading some of the cpus too much. We use Ingo's original 
+	  * approach to rotate them around.
+	  */
+		if (!first_attempt && imbalance >= useful_load_threshold) {
+			rotate_irqs_among_cpus(useful_load_threshold);
+			return;
+		}
+		goto not_worth_the_effort;
+	}
+	
+	first_attempt = 0;		/* heaviest search */
+	max_cpu_irq = tmp_cpu_irq;	/* load */
+	max_loaded = tmp_loaded;	/* processor */
+	imbalance = (max_cpu_irq - min_cpu_irq) / 2;
+	
+	Dprintk("max_loaded cpu = %d\n", max_loaded);
+	Dprintk("min_loaded cpu = %d\n", min_loaded);
+	Dprintk("max_cpu_irq load = %ld\n", max_cpu_irq);
+	Dprintk("min_cpu_irq load = %ld\n", min_cpu_irq);
+	Dprintk("load imbalance = %lu\n", imbalance);
+
+	/* if imbalance is less than approx 10% of max load, then
+	 * observe diminishing returns action. - quit
+	 */
+	if (imbalance < (max_cpu_irq >> 3)) {
+		Dprintk("Imbalance too trivial\n");
+		goto not_worth_the_effort;
+	}
+
+tryanotherirq:
+	/* if we select an IRQ to move that can't go where we want, then
+	 * see if there is another one to try.
+	 */
+	move_this_load = 0;
+	selected_irq = -1;
+	for (j = 0; j < NR_IRQS; j++) {
+		/* Is this an active IRQ? */
+		if (!irq_desc[j].action)
+			continue;
+		if (imbalance <= IRQ_DELTA(max_loaded,j))
+			continue;
+		/* Try to find the IRQ that is closest to the imbalance
+		 * without going over.
+		 */
+		if (move_this_load < IRQ_DELTA(max_loaded,j)) {
+			move_this_load = IRQ_DELTA(max_loaded,j);
+			selected_irq = j;
+		}
+	}
+	if (selected_irq == -1) {
+		goto tryanothercpu;
+	}
+
+	imbalance = move_this_load;
+	
+	/* For physical_balance case, we accumlated both load
+	 * values in the one of the siblings cpu_irq[],
+	 * to use the same code for physical and logical processors
+	 * as much as possible. 
+	 *
+	 * NOTE: the cpu_irq[] array holds the sum of the load for
+	 * sibling A and sibling B in the slot for the lowest numbered
+	 * sibling (A), _AND_ the load for sibling B in the slot for
+	 * the higher numbered sibling.
+	 *
+	 * We seek the least loaded sibling by making the comparison
+	 * (A+B)/2 vs B
+	 */
+	load = CPU_IRQ(min_loaded) >> 1;
+	for_each_cpu_mask(j, cpu_sibling_map[min_loaded]) {
+		if (load > CPU_IRQ(j)) {
+			/* This won't change cpu_sibling_map[min_loaded] */
+			load = CPU_IRQ(j);
+			min_loaded = j;
+		}
+	}
+
+	cpus_and(allowed_mask,
+		cpu_online_map,
+		balance_irq_affinity[selected_irq]);
+	target_cpu_mask = cpumask_of_cpu(min_loaded);
+	cpus_and(tmp, target_cpu_mask, allowed_mask);
+
+	if (!cpus_empty(tmp)) {
+
+		Dprintk("irq = %d moved to cpu = %d\n",
+				selected_irq, min_loaded);
+		/* mark for change destination */
+		set_pending_irq(selected_irq, cpumask_of_cpu(min_loaded));
+
+		/* Since we made a change, come back sooner to 
+		 * check for more variation.
+		 */
+		balanced_irq_interval = max((long)MIN_BALANCED_IRQ_INTERVAL,
+			balanced_irq_interval - BALANCED_IRQ_LESS_DELTA);	
+		return;
+	}
+	goto tryanotherirq;
+
+not_worth_the_effort:
+	/*
+	 * if we did not find an IRQ to move, then adjust the time interval
+	 * upward
+	 */
+	balanced_irq_interval = min((long)MAX_BALANCED_IRQ_INTERVAL,
+		balanced_irq_interval + BALANCED_IRQ_MORE_DELTA);	
+	Dprintk("IRQ worth rotating not found\n");
+	return;
+}
+
+static int balanced_irq(void *unused)
+{
+	int i;
+	unsigned long prev_balance_time = jiffies;
+	long time_remaining = balanced_irq_interval;
+
+	daemonize("kirqd");
+	
+	/* push everything to CPU 0 to give us a starting point.  */
+	for (i = 0 ; i < NR_IRQS ; i++) {
+		irq_desc[i].pending_mask = cpumask_of_cpu(0);
+		set_pending_irq(i, cpumask_of_cpu(0));
+	}
+
+	for ( ; ; ) {
+		time_remaining = schedule_timeout_interruptible(time_remaining);
+		try_to_freeze();
+		if (time_after(jiffies,
+				prev_balance_time+balanced_irq_interval)) {
+			preempt_disable();
+			do_irq_balance();
+			prev_balance_time = jiffies;
+			time_remaining = balanced_irq_interval;
+			preempt_enable();
+		}
+	}
+	return 0;
+}
+
+static int __init balanced_irq_init(void)
+{
+	int i;
+	struct cpuinfo_x86 *c;
+	cpumask_t tmp;
+
+	cpus_shift_right(tmp, cpu_online_map, 2);
+        c = &boot_cpu_data;
+	/* When not overwritten by the command line ask subarchitecture. */
+	if (irqbalance_disabled == IRQBALANCE_CHECK_ARCH)
+		irqbalance_disabled = NO_BALANCE_IRQ;
+	if (irqbalance_disabled)
+		return 0;
+	
+	 /* disable irqbalance completely if there is only one processor online */
+	if (num_online_cpus() < 2) {
+		irqbalance_disabled = 1;
+		return 0;
+	}
+	/*
+	 * Enable physical balance only if more than 1 physical processor
+	 * is present
+	 */
+	if (smp_num_siblings > 1 && !cpus_empty(tmp))
+		physical_balance = 1;
+
+	for_each_online_cpu(i) {
+		irq_cpu_data[i].irq_delta = kmalloc(sizeof(unsigned long) * NR_IRQS, GFP_KERNEL);
+		irq_cpu_data[i].last_irq = kmalloc(sizeof(unsigned long) * NR_IRQS, GFP_KERNEL);
+		if (irq_cpu_data[i].irq_delta == NULL || irq_cpu_data[i].last_irq == NULL) {
+			printk(KERN_ERR "balanced_irq_init: out of memory");
+			goto failed;
+		}
+		memset(irq_cpu_data[i].irq_delta,0,sizeof(unsigned long) * NR_IRQS);
+		memset(irq_cpu_data[i].last_irq,0,sizeof(unsigned long) * NR_IRQS);
+	}
+	
+	printk(KERN_INFO "Starting balanced_irq\n");
+	if (kernel_thread(balanced_irq, NULL, CLONE_KERNEL) >= 0) 
+		return 0;
+	else 
+		printk(KERN_ERR "balanced_irq_init: failed to spawn balanced_irq");
+failed:
+	for_each_possible_cpu(i) {
+		kfree(irq_cpu_data[i].irq_delta);
+		irq_cpu_data[i].irq_delta = NULL;
+		kfree(irq_cpu_data[i].last_irq);
+		irq_cpu_data[i].last_irq = NULL;
+	}
+	return 0;
+}
+
+int __init irqbalance_disable(char *str)
+{
+	irqbalance_disabled = 1;
+	return 1;
+}
+
+__setup("noirqbalance", irqbalance_disable);
+
+late_initcall(balanced_irq_init);
+#endif /* CONFIG_IRQBALANCE */
+#endif /* CONFIG_SMP */
+#endif
+
+#ifndef CONFIG_SMP
+void fastcall send_IPI_self(int vector)
+{
+#ifndef CONFIG_XEN
+	unsigned int cfg;
+
+	/*
+	 * Wait for idle.
+	 */
+	apic_wait_icr_idle();
+	cfg = APIC_DM_FIXED | APIC_DEST_SELF | vector | APIC_DEST_LOGICAL;
+	/*
+	 * Send the IPI. The write to APIC_ICR fires this off.
+	 */
+	apic_write_around(APIC_ICR, cfg);
+#endif
+}
+#endif /* !CONFIG_SMP */
+
+
+/*
+ * support for broken MP BIOSs, enables hand-redirection of PIRQ0-7 to
+ * specific CPU-side IRQs.
+ */
+
+#define MAX_PIRQS 8
+static int pirq_entries [MAX_PIRQS];
+static int pirqs_enabled;
+int skip_ioapic_setup;
+
+static int __init ioapic_setup(char *str)
+{
+	skip_ioapic_setup = 1;
+	return 1;
+}
+
+__setup("noapic", ioapic_setup);
+
+static int __init ioapic_pirq_setup(char *str)
+{
+	int i, max;
+	int ints[MAX_PIRQS+1];
+
+	get_options(str, ARRAY_SIZE(ints), ints);
+
+	for (i = 0; i < MAX_PIRQS; i++)
+		pirq_entries[i] = -1;
+
+	pirqs_enabled = 1;
+	apic_printk(APIC_VERBOSE, KERN_INFO
+			"PIRQ redirection, working around broken MP-BIOS.\n");
+	max = MAX_PIRQS;
+	if (ints[0] < MAX_PIRQS)
+		max = ints[0];
+
+	for (i = 0; i < max; i++) {
+		apic_printk(APIC_VERBOSE, KERN_DEBUG
+				"... PIRQ%d -> IRQ %d\n", i, ints[i+1]);
+		/*
+		 * PIRQs are mapped upside down, usually.
+		 */
+		pirq_entries[MAX_PIRQS-i-1] = ints[i+1];
+	}
+	return 1;
+}
+
+__setup("pirq=", ioapic_pirq_setup);
+
+/*
+ * Find the IRQ entry number of a certain pin.
+ */
+static int find_irq_entry(int apic, int pin, int type)
+{
+	int i;
+
+	for (i = 0; i < mp_irq_entries; i++)
+		if (mp_irqs[i].mpc_irqtype == type &&
+		    (mp_irqs[i].mpc_dstapic == mp_ioapics[apic].mpc_apicid ||
+		     mp_irqs[i].mpc_dstapic == MP_APIC_ALL) &&
+		    mp_irqs[i].mpc_dstirq == pin)
+			return i;
+
+	return -1;
+}
+
+/*
+ * Find the pin to which IRQ[irq] (ISA) is connected
+ */
+static int __init find_isa_irq_pin(int irq, int type)
+{
+	int i;
+
+	for (i = 0; i < mp_irq_entries; i++) {
+		int lbus = mp_irqs[i].mpc_srcbus;
+
+		if ((mp_bus_id_to_type[lbus] == MP_BUS_ISA ||
+		     mp_bus_id_to_type[lbus] == MP_BUS_EISA ||
+		     mp_bus_id_to_type[lbus] == MP_BUS_MCA ||
+		     mp_bus_id_to_type[lbus] == MP_BUS_NEC98
+		    ) &&
+		    (mp_irqs[i].mpc_irqtype == type) &&
+		    (mp_irqs[i].mpc_srcbusirq == irq))
+
+			return mp_irqs[i].mpc_dstirq;
+	}
+	return -1;
+}
+
+static int __init find_isa_irq_apic(int irq, int type)
+{
+	int i;
+
+	for (i = 0; i < mp_irq_entries; i++) {
+		int lbus = mp_irqs[i].mpc_srcbus;
+
+		if ((mp_bus_id_to_type[lbus] == MP_BUS_ISA ||
+		     mp_bus_id_to_type[lbus] == MP_BUS_EISA ||
+		     mp_bus_id_to_type[lbus] == MP_BUS_MCA ||
+		     mp_bus_id_to_type[lbus] == MP_BUS_NEC98
+		    ) &&
+		    (mp_irqs[i].mpc_irqtype == type) &&
+		    (mp_irqs[i].mpc_srcbusirq == irq))
+			break;
+	}
+	if (i < mp_irq_entries) {
+		int apic;
+		for(apic = 0; apic < nr_ioapics; apic++) {
+			if (mp_ioapics[apic].mpc_apicid == mp_irqs[i].mpc_dstapic)
+				return apic;
+		}
+	}
+
+	return -1;
+}
+
+/*
+ * Find a specific PCI IRQ entry.
+ * Not an __init, possibly needed by modules
+ */
+static int pin_2_irq(int idx, int apic, int pin);
+
+int IO_APIC_get_PCI_irq_vector(int bus, int slot, int pin)
+{
+	int apic, i, best_guess = -1;
+
+	apic_printk(APIC_DEBUG, "querying PCI -> IRQ mapping bus:%d, "
+		"slot:%d, pin:%d.\n", bus, slot, pin);
+	if (mp_bus_id_to_pci_bus[bus] == -1) {
+		printk(KERN_WARNING "PCI BIOS passed nonexistent PCI bus %d!\n", bus);
+		return -1;
+	}
+	for (i = 0; i < mp_irq_entries; i++) {
+		int lbus = mp_irqs[i].mpc_srcbus;
+
+		for (apic = 0; apic < nr_ioapics; apic++)
+			if (mp_ioapics[apic].mpc_apicid == mp_irqs[i].mpc_dstapic ||
+			    mp_irqs[i].mpc_dstapic == MP_APIC_ALL)
+				break;
+
+		if ((mp_bus_id_to_type[lbus] == MP_BUS_PCI) &&
+		    !mp_irqs[i].mpc_irqtype &&
+		    (bus == lbus) &&
+		    (slot == ((mp_irqs[i].mpc_srcbusirq >> 2) & 0x1f))) {
+			int irq = pin_2_irq(i,apic,mp_irqs[i].mpc_dstirq);
+
+			if (!(apic || IO_APIC_IRQ(irq)))
+				continue;
+
+			if (pin == (mp_irqs[i].mpc_srcbusirq & 3))
+				return irq;
+			/*
+			 * Use the first all-but-pin matching entry as a
+			 * best-guess fuzzy result for broken mptables.
+			 */
+			if (best_guess < 0)
+				best_guess = irq;
+		}
+	}
+	return best_guess;
+}
+EXPORT_SYMBOL(IO_APIC_get_PCI_irq_vector);
+
+/*
+ * This function currently is only a helper for the i386 smp boot process where 
+ * we need to reprogram the ioredtbls to cater for the cpus which have come online
+ * so mask in all cases should simply be TARGET_CPUS
+ */
+#ifdef CONFIG_SMP
+#ifndef CONFIG_XEN
+void __init setup_ioapic_dest(void)
+{
+	int pin, ioapic, irq, irq_entry;
+
+	if (skip_ioapic_setup == 1)
+		return;
+
+	for (ioapic = 0; ioapic < nr_ioapics; ioapic++) {
+		for (pin = 0; pin < nr_ioapic_registers[ioapic]; pin++) {
+			irq_entry = find_irq_entry(ioapic, pin, mp_INT);
+			if (irq_entry == -1)
+				continue;
+			irq = pin_2_irq(irq_entry, ioapic, pin);
+			set_ioapic_affinity_irq(irq, TARGET_CPUS);
+		}
+
+	}
+}
+#endif /* !CONFIG_XEN */
+#endif
+
+/*
+ * EISA Edge/Level control register, ELCR
+ */
+static int EISA_ELCR(unsigned int irq)
+{
+	if (irq < 16) {
+		unsigned int port = 0x4d0 + (irq >> 3);
+		return (inb(port) >> (irq & 7)) & 1;
+	}
+	apic_printk(APIC_VERBOSE, KERN_INFO
+			"Broken MPtable reports ISA irq %d\n", irq);
+	return 0;
+}
+
+/* EISA interrupts are always polarity zero and can be edge or level
+ * trigger depending on the ELCR value.  If an interrupt is listed as
+ * EISA conforming in the MP table, that means its trigger type must
+ * be read in from the ELCR */
+
+#define default_EISA_trigger(idx)	(EISA_ELCR(mp_irqs[idx].mpc_srcbusirq))
+#define default_EISA_polarity(idx)	(0)
+
+/* ISA interrupts are always polarity zero edge triggered,
+ * when listed as conforming in the MP table. */
+
+#define default_ISA_trigger(idx)	(0)
+#define default_ISA_polarity(idx)	(0)
+
+/* PCI interrupts are always polarity one level triggered,
+ * when listed as conforming in the MP table. */
+
+#define default_PCI_trigger(idx)	(1)
+#define default_PCI_polarity(idx)	(1)
+
+/* MCA interrupts are always polarity zero level triggered,
+ * when listed as conforming in the MP table. */
+
+#define default_MCA_trigger(idx)	(1)
+#define default_MCA_polarity(idx)	(0)
+
+/* NEC98 interrupts are always polarity zero edge triggered,
+ * when listed as conforming in the MP table. */
+
+#define default_NEC98_trigger(idx)     (0)
+#define default_NEC98_polarity(idx)    (0)
+
+static int __init MPBIOS_polarity(int idx)
+{
+	int bus = mp_irqs[idx].mpc_srcbus;
+	int polarity;
+
+	/*
+	 * Determine IRQ line polarity (high active or low active):
+	 */
+	switch (mp_irqs[idx].mpc_irqflag & 3)
+	{
+		case 0: /* conforms, ie. bus-type dependent polarity */
+		{
+			switch (mp_bus_id_to_type[bus])
+			{
+				case MP_BUS_ISA: /* ISA pin */
+				{
+					polarity = default_ISA_polarity(idx);
+					break;
+				}
+				case MP_BUS_EISA: /* EISA pin */
+				{
+					polarity = default_EISA_polarity(idx);
+					break;
+				}
+				case MP_BUS_PCI: /* PCI pin */
+				{
+					polarity = default_PCI_polarity(idx);
+					break;
+				}
+				case MP_BUS_MCA: /* MCA pin */
+				{
+					polarity = default_MCA_polarity(idx);
+					break;
+				}
+				case MP_BUS_NEC98: /* NEC 98 pin */
+				{
+					polarity = default_NEC98_polarity(idx);
+					break;
+				}
+				default:
+				{
+					printk(KERN_WARNING "broken BIOS!!\n");
+					polarity = 1;
+					break;
+				}
+			}
+			break;
+		}
+		case 1: /* high active */
+		{
+			polarity = 0;
+			break;
+		}
+		case 2: /* reserved */
+		{
+			printk(KERN_WARNING "broken BIOS!!\n");
+			polarity = 1;
+			break;
+		}
+		case 3: /* low active */
+		{
+			polarity = 1;
+			break;
+		}
+		default: /* invalid */
+		{
+			printk(KERN_WARNING "broken BIOS!!\n");
+			polarity = 1;
+			break;
+		}
+	}
+	return polarity;
+}
+
+static int MPBIOS_trigger(int idx)
+{
+	int bus = mp_irqs[idx].mpc_srcbus;
+	int trigger;
+
+	/*
+	 * Determine IRQ trigger mode (edge or level sensitive):
+	 */
+	switch ((mp_irqs[idx].mpc_irqflag>>2) & 3)
+	{
+		case 0: /* conforms, ie. bus-type dependent */
+		{
+			switch (mp_bus_id_to_type[bus])
+			{
+				case MP_BUS_ISA: /* ISA pin */
+				{
+					trigger = default_ISA_trigger(idx);
+					break;
+				}
+				case MP_BUS_EISA: /* EISA pin */
+				{
+					trigger = default_EISA_trigger(idx);
+					break;
+				}
+				case MP_BUS_PCI: /* PCI pin */
+				{
+					trigger = default_PCI_trigger(idx);
+					break;
+				}
+				case MP_BUS_MCA: /* MCA pin */
+				{
+					trigger = default_MCA_trigger(idx);
+					break;
+				}
+				case MP_BUS_NEC98: /* NEC 98 pin */
+				{
+					trigger = default_NEC98_trigger(idx);
+					break;
+				}
+				default:
+				{
+					printk(KERN_WARNING "broken BIOS!!\n");
+					trigger = 1;
+					break;
+				}
+			}
+			break;
+		}
+		case 1: /* edge */
+		{
+			trigger = 0;
+			break;
+		}
+		case 2: /* reserved */
+		{
+			printk(KERN_WARNING "broken BIOS!!\n");
+			trigger = 1;
+			break;
+		}
+		case 3: /* level */
+		{
+			trigger = 1;
+			break;
+		}
+		default: /* invalid */
+		{
+			printk(KERN_WARNING "broken BIOS!!\n");
+			trigger = 0;
+			break;
+		}
+	}
+	return trigger;
+}
+
+static inline int irq_polarity(int idx)
+{
+	return MPBIOS_polarity(idx);
+}
+
+static inline int irq_trigger(int idx)
+{
+	return MPBIOS_trigger(idx);
+}
+
+static int pin_2_irq(int idx, int apic, int pin)
+{
+	int irq, i;
+	int bus = mp_irqs[idx].mpc_srcbus;
+
+	/*
+	 * Debugging check, we are in big trouble if this message pops up!
+	 */
+	if (mp_irqs[idx].mpc_dstirq != pin)
+		printk(KERN_ERR "broken BIOS or MPTABLE parser, ayiee!!\n");
+
+	switch (mp_bus_id_to_type[bus])
+	{
+		case MP_BUS_ISA: /* ISA pin */
+		case MP_BUS_EISA:
+		case MP_BUS_MCA:
+		case MP_BUS_NEC98:
+		{
+			irq = mp_irqs[idx].mpc_srcbusirq;
+			break;
+		}
+		case MP_BUS_PCI: /* PCI pin */
+		{
+			/*
+			 * PCI IRQs are mapped in order
+			 */
+			i = irq = 0;
+			while (i < apic)
+				irq += nr_ioapic_registers[i++];
+			irq += pin;
+
+			/*
+			 * For MPS mode, so far only needed by ES7000 platform
+			 */
+			if (ioapic_renumber_irq)
+				irq = ioapic_renumber_irq(apic, irq);
+
+			break;
+		}
+		default:
+		{
+			printk(KERN_ERR "unknown bus type %d.\n",bus); 
+			irq = 0;
+			break;
+		}
+	}
+
+	/*
+	 * PCI IRQ command line redirection. Yes, limits are hardcoded.
+	 */
+	if ((pin >= 16) && (pin <= 23)) {
+		if (pirq_entries[pin-16] != -1) {
+			if (!pirq_entries[pin-16]) {
+				apic_printk(APIC_VERBOSE, KERN_DEBUG
+						"disabling PIRQ%d\n", pin-16);
+			} else {
+				irq = pirq_entries[pin-16];
+				apic_printk(APIC_VERBOSE, KERN_DEBUG
+						"using PIRQ%d -> IRQ %d\n",
+						pin-16, irq);
+			}
+		}
+	}
+	return irq;
+}
+
+static inline int IO_APIC_irq_trigger(int irq)
+{
+	int apic, idx, pin;
+
+	for (apic = 0; apic < nr_ioapics; apic++) {
+		for (pin = 0; pin < nr_ioapic_registers[apic]; pin++) {
+			idx = find_irq_entry(apic,pin,mp_INT);
+			if ((idx != -1) && (irq == pin_2_irq(idx,apic,pin)))
+				return irq_trigger(idx);
+		}
+	}
+	/*
+	 * nonexistent IRQs are edge default
+	 */
+	return 0;
+}
+
+/* irq_vectors is indexed by the sum of all RTEs in all I/O APICs. */
+u8 irq_vector[NR_IRQ_VECTORS] __read_mostly; /* = { FIRST_DEVICE_VECTOR , 0 }; */
+
+int assign_irq_vector(int irq)
+{
+	unsigned long flags;
+	int vector;
+	struct physdev_irq irq_op;
+
+	BUG_ON(irq != AUTO_ASSIGN && (unsigned)irq >= NR_IRQ_VECTORS);
+
+	spin_lock_irqsave(&vector_lock, flags);
+
+	if (irq != AUTO_ASSIGN && IO_APIC_VECTOR(irq) > 0) {
+		spin_unlock_irqrestore(&vector_lock, flags);
+		return IO_APIC_VECTOR(irq);
+	}
+
+	irq_op.irq = irq;
+	if (HYPERVISOR_physdev_op(PHYSDEVOP_alloc_irq_vector, &irq_op)) {
+		spin_unlock_irqrestore(&vector_lock, flags);
+		return -ENOSPC;
+	}
+
+	vector = irq_op.vector;
+	vector_irq[vector] = irq;
+	if (irq != AUTO_ASSIGN)
+		IO_APIC_VECTOR(irq) = vector;
+
+	spin_unlock_irqrestore(&vector_lock, flags);
+
+	return vector;
+}
+
+#ifndef CONFIG_XEN
+static struct hw_interrupt_type ioapic_level_type;
+static struct hw_interrupt_type ioapic_edge_type;
+
+#define IOAPIC_AUTO	-1
+#define IOAPIC_EDGE	0
+#define IOAPIC_LEVEL	1
+
+static void ioapic_register_intr(int irq, int vector, unsigned long trigger)
+{
+	unsigned idx;
+
+	idx = use_pci_vector() && !platform_legacy_irq(irq) ? vector : irq;
+
+	if ((trigger == IOAPIC_AUTO && IO_APIC_irq_trigger(irq)) ||
+			trigger == IOAPIC_LEVEL)
+		irq_desc[idx].chip = &ioapic_level_type;
+	else
+		irq_desc[idx].chip = &ioapic_edge_type;
+	set_intr_gate(vector, interrupt[idx]);
+}
+#else
+#define ioapic_register_intr(_irq,_vector,_trigger) ((void)0)
+#endif
+
+static void __init setup_IO_APIC_irqs(void)
+{
+	struct IO_APIC_route_entry entry;
+	int apic, pin, idx, irq, first_notcon = 1, vector;
+	unsigned long flags;
+
+	apic_printk(APIC_VERBOSE, KERN_DEBUG "init IO_APIC IRQs\n");
+
+	for (apic = 0; apic < nr_ioapics; apic++) {
+	for (pin = 0; pin < nr_ioapic_registers[apic]; pin++) {
+
+		/*
+		 * add it to the IO-APIC irq-routing table:
+		 */
+		memset(&entry,0,sizeof(entry));
+
+		entry.delivery_mode = INT_DELIVERY_MODE;
+		entry.dest_mode = INT_DEST_MODE;
+		entry.mask = 0;				/* enable IRQ */
+		entry.dest.logical.logical_dest = 
+					cpu_mask_to_apicid(TARGET_CPUS);
+
+		idx = find_irq_entry(apic,pin,mp_INT);
+		if (idx == -1) {
+			if (first_notcon) {
+				apic_printk(APIC_VERBOSE, KERN_DEBUG
+						" IO-APIC (apicid-pin) %d-%d",
+						mp_ioapics[apic].mpc_apicid,
+						pin);
+				first_notcon = 0;
+			} else
+				apic_printk(APIC_VERBOSE, ", %d-%d",
+					mp_ioapics[apic].mpc_apicid, pin);
+			continue;
+		}
+
+		entry.trigger = irq_trigger(idx);
+		entry.polarity = irq_polarity(idx);
+
+		if (irq_trigger(idx)) {
+			entry.trigger = 1;
+			entry.mask = 1;
+		}
+
+		irq = pin_2_irq(idx, apic, pin);
+		/*
+		 * skip adding the timer int on secondary nodes, which causes
+		 * a small but painful rift in the time-space continuum
+		 */
+		if (multi_timer_check(apic, irq))
+			continue;
+		else
+			add_pin_to_irq(irq, apic, pin);
+
+		if (/*!apic &&*/ !IO_APIC_IRQ(irq))
+			continue;
+
+		if (IO_APIC_IRQ(irq)) {
+			vector = assign_irq_vector(irq);
+			entry.vector = vector;
+			ioapic_register_intr(irq, vector, IOAPIC_AUTO);
+		
+			if (!apic && (irq < 16))
+				disable_8259A_irq(irq);
+		}
+		spin_lock_irqsave(&ioapic_lock, flags);
+		io_apic_write(apic, 0x11+2*pin, *(((int *)&entry)+1));
+		io_apic_write(apic, 0x10+2*pin, *(((int *)&entry)+0));
+		set_native_irq_info(irq, TARGET_CPUS);
+		spin_unlock_irqrestore(&ioapic_lock, flags);
+	}
+	}
+
+	if (!first_notcon)
+		apic_printk(APIC_VERBOSE, " not connected.\n");
+}
+
+/*
+ * Set up the 8259A-master output pin:
+ */
+#ifndef CONFIG_XEN
+static void __init setup_ExtINT_IRQ0_pin(unsigned int apic, unsigned int pin, int vector)
+{
+	struct IO_APIC_route_entry entry;
+	unsigned long flags;
+
+	memset(&entry,0,sizeof(entry));
+
+	disable_8259A_irq(0);
+
+	/* mask LVT0 */
+	apic_write_around(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_EXTINT);
+
+	/*
+	 * We use logical delivery to get the timer IRQ
+	 * to the first CPU.
+	 */
+	entry.dest_mode = INT_DEST_MODE;
+	entry.mask = 0;					/* unmask IRQ now */
+	entry.dest.logical.logical_dest = cpu_mask_to_apicid(TARGET_CPUS);
+	entry.delivery_mode = INT_DELIVERY_MODE;
+	entry.polarity = 0;
+	entry.trigger = 0;
+	entry.vector = vector;
+
+	/*
+	 * The timer IRQ doesn't have to know that behind the
+	 * scene we have a 8259A-master in AEOI mode ...
+	 */
+	irq_desc[0].chip = &ioapic_edge_type;
+
+	/*
+	 * Add it to the IO-APIC irq-routing table:
+	 */
+	spin_lock_irqsave(&ioapic_lock, flags);
+	io_apic_write(apic, 0x11+2*pin, *(((int *)&entry)+1));
+	io_apic_write(apic, 0x10+2*pin, *(((int *)&entry)+0));
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	enable_8259A_irq(0);
+}
+
+static inline void UNEXPECTED_IO_APIC(void)
+{
+}
+
+void __init print_IO_APIC(void)
+{
+	int apic, i;
+	union IO_APIC_reg_00 reg_00;
+	union IO_APIC_reg_01 reg_01;
+	union IO_APIC_reg_02 reg_02;
+	union IO_APIC_reg_03 reg_03;
+	unsigned long flags;
+
+	if (apic_verbosity == APIC_QUIET)
+		return;
+
+ 	printk(KERN_DEBUG "number of MP IRQ sources: %d.\n", mp_irq_entries);
+	for (i = 0; i < nr_ioapics; i++)
+		printk(KERN_DEBUG "number of IO-APIC #%d registers: %d.\n",
+		       mp_ioapics[i].mpc_apicid, nr_ioapic_registers[i]);
+
+	/*
+	 * We are a bit conservative about what we expect.  We have to
+	 * know about every hardware change ASAP.
+	 */
+	printk(KERN_INFO "testing the IO APIC.......................\n");
+
+	for (apic = 0; apic < nr_ioapics; apic++) {
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	reg_00.raw = io_apic_read(apic, 0);
+	reg_01.raw = io_apic_read(apic, 1);
+	if (reg_01.bits.version >= 0x10)
+		reg_02.raw = io_apic_read(apic, 2);
+	if (reg_01.bits.version >= 0x20)
+		reg_03.raw = io_apic_read(apic, 3);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	printk(KERN_DEBUG "IO APIC #%d......\n", mp_ioapics[apic].mpc_apicid);
+	printk(KERN_DEBUG ".... register #00: %08X\n", reg_00.raw);
+	printk(KERN_DEBUG ".......    : physical APIC id: %02X\n", reg_00.bits.ID);
+	printk(KERN_DEBUG ".......    : Delivery Type: %X\n", reg_00.bits.delivery_type);
+	printk(KERN_DEBUG ".......    : LTS          : %X\n", reg_00.bits.LTS);
+	if (reg_00.bits.ID >= get_physical_broadcast())
+		UNEXPECTED_IO_APIC();
+	if (reg_00.bits.__reserved_1 || reg_00.bits.__reserved_2)
+		UNEXPECTED_IO_APIC();
+
+	printk(KERN_DEBUG ".... register #01: %08X\n", reg_01.raw);
+	printk(KERN_DEBUG ".......     : max redirection entries: %04X\n", reg_01.bits.entries);
+	if (	(reg_01.bits.entries != 0x0f) && /* older (Neptune) boards */
+		(reg_01.bits.entries != 0x17) && /* typical ISA+PCI boards */
+		(reg_01.bits.entries != 0x1b) && /* Compaq Proliant boards */
+		(reg_01.bits.entries != 0x1f) && /* dual Xeon boards */
+		(reg_01.bits.entries != 0x22) && /* bigger Xeon boards */
+		(reg_01.bits.entries != 0x2E) &&
+		(reg_01.bits.entries != 0x3F)
+	)
+		UNEXPECTED_IO_APIC();
+
+	printk(KERN_DEBUG ".......     : PRQ implemented: %X\n", reg_01.bits.PRQ);
+	printk(KERN_DEBUG ".......     : IO APIC version: %04X\n", reg_01.bits.version);
+	if (	(reg_01.bits.version != 0x01) && /* 82489DX IO-APICs */
+		(reg_01.bits.version != 0x10) && /* oldest IO-APICs */
+		(reg_01.bits.version != 0x11) && /* Pentium/Pro IO-APICs */
+		(reg_01.bits.version != 0x13) && /* Xeon IO-APICs */
+		(reg_01.bits.version != 0x20)    /* Intel P64H (82806 AA) */
+	)
+		UNEXPECTED_IO_APIC();
+	if (reg_01.bits.__reserved_1 || reg_01.bits.__reserved_2)
+		UNEXPECTED_IO_APIC();
+
+	/*
+	 * Some Intel chipsets with IO APIC VERSION of 0x1? don't have reg_02,
+	 * but the value of reg_02 is read as the previous read register
+	 * value, so ignore it if reg_02 == reg_01.
+	 */
+	if (reg_01.bits.version >= 0x10 && reg_02.raw != reg_01.raw) {
+		printk(KERN_DEBUG ".... register #02: %08X\n", reg_02.raw);
+		printk(KERN_DEBUG ".......     : arbitration: %02X\n", reg_02.bits.arbitration);
+		if (reg_02.bits.__reserved_1 || reg_02.bits.__reserved_2)
+			UNEXPECTED_IO_APIC();
+	}
+
+	/*
+	 * Some Intel chipsets with IO APIC VERSION of 0x2? don't have reg_02
+	 * or reg_03, but the value of reg_0[23] is read as the previous read
+	 * register value, so ignore it if reg_03 == reg_0[12].
+	 */
+	if (reg_01.bits.version >= 0x20 && reg_03.raw != reg_02.raw &&
+	    reg_03.raw != reg_01.raw) {
+		printk(KERN_DEBUG ".... register #03: %08X\n", reg_03.raw);
+		printk(KERN_DEBUG ".......     : Boot DT    : %X\n", reg_03.bits.boot_DT);
+		if (reg_03.bits.__reserved_1)
+			UNEXPECTED_IO_APIC();
+	}
+
+	printk(KERN_DEBUG ".... IRQ redirection table:\n");
+
+	printk(KERN_DEBUG " NR Log Phy Mask Trig IRR Pol"
+			  " Stat Dest Deli Vect:   \n");
+
+	for (i = 0; i <= reg_01.bits.entries; i++) {
+		struct IO_APIC_route_entry entry;
+
+		spin_lock_irqsave(&ioapic_lock, flags);
+		*(((int *)&entry)+0) = io_apic_read(apic, 0x10+i*2);
+		*(((int *)&entry)+1) = io_apic_read(apic, 0x11+i*2);
+		spin_unlock_irqrestore(&ioapic_lock, flags);
+
+		printk(KERN_DEBUG " %02x %03X %02X  ",
+			i,
+			entry.dest.logical.logical_dest,
+			entry.dest.physical.physical_dest
+		);
+
+		printk("%1d    %1d    %1d   %1d   %1d    %1d    %1d    %02X\n",
+			entry.mask,
+			entry.trigger,
+			entry.irr,
+			entry.polarity,
+			entry.delivery_status,
+			entry.dest_mode,
+			entry.delivery_mode,
+			entry.vector
+		);
+	}
+	}
+	if (use_pci_vector())
+		printk(KERN_INFO "Using vector-based indexing\n");
+	printk(KERN_DEBUG "IRQ to pin mappings:\n");
+	for (i = 0; i < NR_IRQS; i++) {
+		struct irq_pin_list *entry = irq_2_pin + i;
+		if (entry->pin < 0)
+			continue;
+ 		if (use_pci_vector() && !platform_legacy_irq(i))
+			printk(KERN_DEBUG "IRQ%d ", IO_APIC_VECTOR(i));
+		else
+			printk(KERN_DEBUG "IRQ%d ", i);
+		for (;;) {
+			printk("-> %d:%d", entry->apic, entry->pin);
+			if (!entry->next)
+				break;
+			entry = irq_2_pin + entry->next;
+		}
+		printk("\n");
+	}
+
+	printk(KERN_INFO ".................................... done.\n");
+
+	return;
+}
+
+static void print_APIC_bitfield (int base)
+{
+	unsigned int v;
+	int i, j;
+
+	if (apic_verbosity == APIC_QUIET)
+		return;
+
+	printk(KERN_DEBUG "0123456789abcdef0123456789abcdef\n" KERN_DEBUG);
+	for (i = 0; i < 8; i++) {
+		v = apic_read(base + i*0x10);
+		for (j = 0; j < 32; j++) {
+			if (v & (1<<j))
+				printk("1");
+			else
+				printk("0");
+		}
+		printk("\n");
+	}
+}
+
+void /*__init*/ print_local_APIC(void * dummy)
+{
+	unsigned int v, ver, maxlvt;
+
+	if (apic_verbosity == APIC_QUIET)
+		return;
+
+	printk("\n" KERN_DEBUG "printing local APIC contents on CPU#%d/%d:\n",
+		smp_processor_id(), hard_smp_processor_id());
+	v = apic_read(APIC_ID);
+	printk(KERN_INFO "... APIC ID:      %08x (%01x)\n", v, GET_APIC_ID(v));
+	v = apic_read(APIC_LVR);
+	printk(KERN_INFO "... APIC VERSION: %08x\n", v);
+	ver = GET_APIC_VERSION(v);
+	maxlvt = get_maxlvt();
+
+	v = apic_read(APIC_TASKPRI);
+	printk(KERN_DEBUG "... APIC TASKPRI: %08x (%02x)\n", v, v & APIC_TPRI_MASK);
+
+	if (APIC_INTEGRATED(ver)) {			/* !82489DX */
+		v = apic_read(APIC_ARBPRI);
+		printk(KERN_DEBUG "... APIC ARBPRI: %08x (%02x)\n", v,
+			v & APIC_ARBPRI_MASK);
+		v = apic_read(APIC_PROCPRI);
+		printk(KERN_DEBUG "... APIC PROCPRI: %08x\n", v);
+	}
+
+	v = apic_read(APIC_EOI);
+	printk(KERN_DEBUG "... APIC EOI: %08x\n", v);
+	v = apic_read(APIC_RRR);
+	printk(KERN_DEBUG "... APIC RRR: %08x\n", v);
+	v = apic_read(APIC_LDR);
+	printk(KERN_DEBUG "... APIC LDR: %08x\n", v);
+	v = apic_read(APIC_DFR);
+	printk(KERN_DEBUG "... APIC DFR: %08x\n", v);
+	v = apic_read(APIC_SPIV);
+	printk(KERN_DEBUG "... APIC SPIV: %08x\n", v);
+
+	printk(KERN_DEBUG "... APIC ISR field:\n");
+	print_APIC_bitfield(APIC_ISR);
+	printk(KERN_DEBUG "... APIC TMR field:\n");
+	print_APIC_bitfield(APIC_TMR);
+	printk(KERN_DEBUG "... APIC IRR field:\n");
+	print_APIC_bitfield(APIC_IRR);
+
+	if (APIC_INTEGRATED(ver)) {		/* !82489DX */
+		if (maxlvt > 3)		/* Due to the Pentium erratum 3AP. */
+			apic_write(APIC_ESR, 0);
+		v = apic_read(APIC_ESR);
+		printk(KERN_DEBUG "... APIC ESR: %08x\n", v);
+	}
+
+	v = apic_read(APIC_ICR);
+	printk(KERN_DEBUG "... APIC ICR: %08x\n", v);
+	v = apic_read(APIC_ICR2);
+	printk(KERN_DEBUG "... APIC ICR2: %08x\n", v);
+
+	v = apic_read(APIC_LVTT);
+	printk(KERN_DEBUG "... APIC LVTT: %08x\n", v);
+
+	if (maxlvt > 3) {                       /* PC is LVT#4. */
+		v = apic_read(APIC_LVTPC);
+		printk(KERN_DEBUG "... APIC LVTPC: %08x\n", v);
+	}
+	v = apic_read(APIC_LVT0);
+	printk(KERN_DEBUG "... APIC LVT0: %08x\n", v);
+	v = apic_read(APIC_LVT1);
+	printk(KERN_DEBUG "... APIC LVT1: %08x\n", v);
+
+	if (maxlvt > 2) {			/* ERR is LVT#3. */
+		v = apic_read(APIC_LVTERR);
+		printk(KERN_DEBUG "... APIC LVTERR: %08x\n", v);
+	}
+
+	v = apic_read(APIC_TMICT);
+	printk(KERN_DEBUG "... APIC TMICT: %08x\n", v);
+	v = apic_read(APIC_TMCCT);
+	printk(KERN_DEBUG "... APIC TMCCT: %08x\n", v);
+	v = apic_read(APIC_TDCR);
+	printk(KERN_DEBUG "... APIC TDCR: %08x\n", v);
+	printk("\n");
+}
+
+void print_all_local_APICs (void)
+{
+	on_each_cpu(print_local_APIC, NULL, 1, 1);
+}
+
+void /*__init*/ print_PIC(void)
+{
+	unsigned int v;
+	unsigned long flags;
+
+	if (apic_verbosity == APIC_QUIET)
+		return;
+
+	printk(KERN_DEBUG "\nprinting PIC contents\n");
+
+	spin_lock_irqsave(&i8259A_lock, flags);
+
+	v = inb(0xa1) << 8 | inb(0x21);
+	printk(KERN_DEBUG "... PIC  IMR: %04x\n", v);
+
+	v = inb(0xa0) << 8 | inb(0x20);
+	printk(KERN_DEBUG "... PIC  IRR: %04x\n", v);
+
+	outb(0x0b,0xa0);
+	outb(0x0b,0x20);
+	v = inb(0xa0) << 8 | inb(0x20);
+	outb(0x0a,0xa0);
+	outb(0x0a,0x20);
+
+	spin_unlock_irqrestore(&i8259A_lock, flags);
+
+	printk(KERN_DEBUG "... PIC  ISR: %04x\n", v);
+
+	v = inb(0x4d1) << 8 | inb(0x4d0);
+	printk(KERN_DEBUG "... PIC ELCR: %04x\n", v);
+}
+#endif /* !CONFIG_XEN */
+
+static void __init enable_IO_APIC(void)
+{
+	union IO_APIC_reg_01 reg_01;
+	int i8259_apic, i8259_pin;
+	int i, apic;
+	unsigned long flags;
+
+	for (i = 0; i < PIN_MAP_SIZE; i++) {
+		irq_2_pin[i].pin = -1;
+		irq_2_pin[i].next = 0;
+	}
+	if (!pirqs_enabled)
+		for (i = 0; i < MAX_PIRQS; i++)
+			pirq_entries[i] = -1;
+
+	/*
+	 * The number of IO-APIC IRQ registers (== #pins):
+	 */
+	for (apic = 0; apic < nr_ioapics; apic++) {
+		spin_lock_irqsave(&ioapic_lock, flags);
+		reg_01.raw = io_apic_read(apic, 1);
+		spin_unlock_irqrestore(&ioapic_lock, flags);
+		nr_ioapic_registers[apic] = reg_01.bits.entries+1;
+	}
+	for(apic = 0; apic < nr_ioapics; apic++) {
+		int pin;
+		/* See if any of the pins is in ExtINT mode */
+		for (pin = 0; pin < nr_ioapic_registers[apic]; pin++) {
+			struct IO_APIC_route_entry entry;
+			spin_lock_irqsave(&ioapic_lock, flags);
+			*(((int *)&entry) + 0) = io_apic_read(apic, 0x10 + 2 * pin);
+			*(((int *)&entry) + 1) = io_apic_read(apic, 0x11 + 2 * pin);
+			spin_unlock_irqrestore(&ioapic_lock, flags);
+
+
+			/* If the interrupt line is enabled and in ExtInt mode
+			 * I have found the pin where the i8259 is connected.
+			 */
+			if ((entry.mask == 0) && (entry.delivery_mode == dest_ExtINT)) {
+				ioapic_i8259.apic = apic;
+				ioapic_i8259.pin  = pin;
+				goto found_i8259;
+			}
+		}
+	}
+ found_i8259:
+	/* Look to see what if the MP table has reported the ExtINT */
+	/* If we could not find the appropriate pin by looking at the ioapic
+	 * the i8259 probably is not connected the ioapic but give the
+	 * mptable a chance anyway.
+	 */
+	i8259_pin  = find_isa_irq_pin(0, mp_ExtINT);
+	i8259_apic = find_isa_irq_apic(0, mp_ExtINT);
+	/* Trust the MP table if nothing is setup in the hardware */
+	if ((ioapic_i8259.pin == -1) && (i8259_pin >= 0)) {
+		printk(KERN_WARNING "ExtINT not setup in hardware but reported by MP table\n");
+		ioapic_i8259.pin  = i8259_pin;
+		ioapic_i8259.apic = i8259_apic;
+	}
+	/* Complain if the MP table and the hardware disagree */
+	if (((ioapic_i8259.apic != i8259_apic) || (ioapic_i8259.pin != i8259_pin)) &&
+		(i8259_pin >= 0) && (ioapic_i8259.pin >= 0))
+	{
+		printk(KERN_WARNING "ExtINT in hardware and MP table differ\n");
+	}
+
+	/*
+	 * Do not trust the IO-APIC being empty at bootup
+	 */
+	clear_IO_APIC();
+}
+
+/*
+ * Not an __init, needed by the reboot code
+ */
+void disable_IO_APIC(void)
+{
+	/*
+	 * Clear the IO-APIC before rebooting:
+	 */
+	clear_IO_APIC();
+
+#ifndef CONFIG_XEN
+	/*
+	 * If the i8259 is routed through an IOAPIC
+	 * Put that IOAPIC in virtual wire mode
+	 * so legacy interrupts can be delivered.
+	 */
+	if (ioapic_i8259.pin != -1) {
+		struct IO_APIC_route_entry entry;
+		unsigned long flags;
+
+		memset(&entry, 0, sizeof(entry));
+		entry.mask            = 0; /* Enabled */
+		entry.trigger         = 0; /* Edge */
+		entry.irr             = 0;
+		entry.polarity        = 0; /* High */
+		entry.delivery_status = 0;
+		entry.dest_mode       = 0; /* Physical */
+		entry.delivery_mode   = dest_ExtINT; /* ExtInt */
+		entry.vector          = 0;
+		entry.dest.physical.physical_dest =
+					GET_APIC_ID(apic_read(APIC_ID));
+
+		/*
+		 * Add it to the IO-APIC irq-routing table:
+		 */
+		spin_lock_irqsave(&ioapic_lock, flags);
+		io_apic_write(ioapic_i8259.apic, 0x11+2*ioapic_i8259.pin,
+			*(((int *)&entry)+1));
+		io_apic_write(ioapic_i8259.apic, 0x10+2*ioapic_i8259.pin,
+			*(((int *)&entry)+0));
+		spin_unlock_irqrestore(&ioapic_lock, flags);
+	}
+	disconnect_bsp_APIC(ioapic_i8259.pin != -1);
+#endif
+}
+
+/*
+ * function to set the IO-APIC physical IDs based on the
+ * values stored in the MPC table.
+ *
+ * by Matt Domsch <Matt_Domsch@dell.com>  Tue Dec 21 12:25:05 CST 1999
+ */
+
+#if !defined(CONFIG_XEN) && !defined(CONFIG_X86_NUMAQ)
+static void __init setup_ioapic_ids_from_mpc(void)
+{
+	union IO_APIC_reg_00 reg_00;
+	physid_mask_t phys_id_present_map;
+	int apic;
+	int i;
+	unsigned char old_id;
+	unsigned long flags;
+
+	/*
+	 * Don't check I/O APIC IDs for xAPIC systems.  They have
+	 * no meaning without the serial APIC bus.
+	 */
+	if (!(boot_cpu_data.x86_vendor == X86_VENDOR_INTEL)
+		|| APIC_XAPIC(apic_version[boot_cpu_physical_apicid]))
+		return;
+	/*
+	 * This is broken; anything with a real cpu count has to
+	 * circumvent this idiocy regardless.
+	 */
+	phys_id_present_map = ioapic_phys_id_map(phys_cpu_present_map);
+
+	/*
+	 * Set the IOAPIC ID to the value stored in the MPC table.
+	 */
+	for (apic = 0; apic < nr_ioapics; apic++) {
+
+		/* Read the register 0 value */
+		spin_lock_irqsave(&ioapic_lock, flags);
+		reg_00.raw = io_apic_read(apic, 0);
+		spin_unlock_irqrestore(&ioapic_lock, flags);
+		
+		old_id = mp_ioapics[apic].mpc_apicid;
+
+		if (mp_ioapics[apic].mpc_apicid >= get_physical_broadcast()) {
+			printk(KERN_ERR "BIOS bug, IO-APIC#%d ID is %d in the MPC table!...\n",
+				apic, mp_ioapics[apic].mpc_apicid);
+			printk(KERN_ERR "... fixing up to %d. (tell your hw vendor)\n",
+				reg_00.bits.ID);
+			mp_ioapics[apic].mpc_apicid = reg_00.bits.ID;
+		}
+
+		/*
+		 * Sanity check, is the ID really free? Every APIC in a
+		 * system must have a unique ID or we get lots of nice
+		 * 'stuck on smp_invalidate_needed IPI wait' messages.
+		 */
+		if (check_apicid_used(phys_id_present_map,
+					mp_ioapics[apic].mpc_apicid)) {
+			printk(KERN_ERR "BIOS bug, IO-APIC#%d ID %d is already used!...\n",
+				apic, mp_ioapics[apic].mpc_apicid);
+			for (i = 0; i < get_physical_broadcast(); i++)
+				if (!physid_isset(i, phys_id_present_map))
+					break;
+			if (i >= get_physical_broadcast())
+				panic("Max APIC ID exceeded!\n");
+			printk(KERN_ERR "... fixing up to %d. (tell your hw vendor)\n",
+				i);
+			physid_set(i, phys_id_present_map);
+			mp_ioapics[apic].mpc_apicid = i;
+		} else {
+			physid_mask_t tmp;
+			tmp = apicid_to_cpu_present(mp_ioapics[apic].mpc_apicid);
+			apic_printk(APIC_VERBOSE, "Setting %d in the "
+					"phys_id_present_map\n",
+					mp_ioapics[apic].mpc_apicid);
+			physids_or(phys_id_present_map, phys_id_present_map, tmp);
+		}
+
+
+		/*
+		 * We need to adjust the IRQ routing table
+		 * if the ID changed.
+		 */
+		if (old_id != mp_ioapics[apic].mpc_apicid)
+			for (i = 0; i < mp_irq_entries; i++)
+				if (mp_irqs[i].mpc_dstapic == old_id)
+					mp_irqs[i].mpc_dstapic
+						= mp_ioapics[apic].mpc_apicid;
+
+		/*
+		 * Read the right value from the MPC table and
+		 * write it into the ID register.
+	 	 */
+		apic_printk(APIC_VERBOSE, KERN_INFO
+			"...changing IO-APIC physical APIC ID to %d ...",
+			mp_ioapics[apic].mpc_apicid);
+
+		reg_00.bits.ID = mp_ioapics[apic].mpc_apicid;
+		spin_lock_irqsave(&ioapic_lock, flags);
+		io_apic_write(apic, 0, reg_00.raw);
+		spin_unlock_irqrestore(&ioapic_lock, flags);
+
+		/*
+		 * Sanity check
+		 */
+		spin_lock_irqsave(&ioapic_lock, flags);
+		reg_00.raw = io_apic_read(apic, 0);
+		spin_unlock_irqrestore(&ioapic_lock, flags);
+		if (reg_00.bits.ID != mp_ioapics[apic].mpc_apicid)
+			printk("could not set ID!\n");
+		else
+			apic_printk(APIC_VERBOSE, " ok.\n");
+	}
+}
+#else
+static void __init setup_ioapic_ids_from_mpc(void) { }
+#endif
+
+#ifndef CONFIG_XEN
+/*
+ * There is a nasty bug in some older SMP boards, their mptable lies
+ * about the timer IRQ. We do the following to work around the situation:
+ *
+ *	- timer IRQ defaults to IO-APIC IRQ
+ *	- if this function detects that timer IRQs are defunct, then we fall
+ *	  back to ISA timer IRQs
+ */
+static int __init timer_irq_works(void)
+{
+	unsigned long t1 = jiffies;
+
+	local_irq_enable();
+	/* Let ten ticks pass... */
+	mdelay((10 * 1000) / HZ);
+
+	/*
+	 * Expect a few ticks at least, to be sure some possible
+	 * glue logic does not lock up after one or two first
+	 * ticks in a non-ExtINT mode.  Also the local APIC
+	 * might have cached one ExtINT interrupt.  Finally, at
+	 * least one tick may be lost due to delays.
+	 */
+	if (jiffies - t1 > 4)
+		return 1;
+
+	return 0;
+}
+
+/*
+ * In the SMP+IOAPIC case it might happen that there are an unspecified
+ * number of pending IRQ events unhandled. These cases are very rare,
+ * so we 'resend' these IRQs via IPIs, to the same CPU. It's much
+ * better to do it this way as thus we do not have to be aware of
+ * 'pending' interrupts in the IRQ path, except at this point.
+ */
+/*
+ * Edge triggered needs to resend any interrupt
+ * that was delayed but this is now handled in the device
+ * independent code.
+ */
+
+/*
+ * Starting up a edge-triggered IO-APIC interrupt is
+ * nasty - we need to make sure that we get the edge.
+ * If it is already asserted for some reason, we need
+ * return 1 to indicate that is was pending.
+ *
+ * This is not complete - we should be able to fake
+ * an edge even if it isn't on the 8259A...
+ */
+static unsigned int startup_edge_ioapic_irq(unsigned int irq)
+{
+	int was_pending = 0;
+	unsigned long flags;
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	if (irq < 16) {
+		disable_8259A_irq(irq);
+		if (i8259A_irq_pending(irq))
+			was_pending = 1;
+	}
+	__unmask_IO_APIC_irq(irq);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	return was_pending;
+}
+
+/*
+ * Once we have recorded IRQ_PENDING already, we can mask the
+ * interrupt for real. This prevents IRQ storms from unhandled
+ * devices.
+ */
+static void ack_edge_ioapic_irq(unsigned int irq)
+{
+	move_irq(irq);
+	if ((irq_desc[irq].status & (IRQ_PENDING | IRQ_DISABLED))
+					== (IRQ_PENDING | IRQ_DISABLED))
+		mask_IO_APIC_irq(irq);
+	ack_APIC_irq();
+}
+
+/*
+ * Level triggered interrupts can just be masked,
+ * and shutting down and starting up the interrupt
+ * is the same as enabling and disabling them -- except
+ * with a startup need to return a "was pending" value.
+ *
+ * Level triggered interrupts are special because we
+ * do not touch any IO-APIC register while handling
+ * them. We ack the APIC in the end-IRQ handler, not
+ * in the start-IRQ-handler. Protection against reentrance
+ * from the same interrupt is still provided, both by the
+ * generic IRQ layer and by the fact that an unacked local
+ * APIC does not accept IRQs.
+ */
+static unsigned int startup_level_ioapic_irq (unsigned int irq)
+{
+	unmask_IO_APIC_irq(irq);
+
+	return 0; /* don't check for pending */
+}
+
+static void end_level_ioapic_irq (unsigned int irq)
+{
+	unsigned long v;
+	int i;
+
+	move_irq(irq);
+/*
+ * It appears there is an erratum which affects at least version 0x11
+ * of I/O APIC (that's the 82093AA and cores integrated into various
+ * chipsets).  Under certain conditions a level-triggered interrupt is
+ * erroneously delivered as edge-triggered one but the respective IRR
+ * bit gets set nevertheless.  As a result the I/O unit expects an EOI
+ * message but it will never arrive and further interrupts are blocked
+ * from the source.  The exact reason is so far unknown, but the
+ * phenomenon was observed when two consecutive interrupt requests
+ * from a given source get delivered to the same CPU and the source is
+ * temporarily disabled in between.
+ *
+ * A workaround is to simulate an EOI message manually.  We achieve it
+ * by setting the trigger mode to edge and then to level when the edge
+ * trigger mode gets detected in the TMR of a local APIC for a
+ * level-triggered interrupt.  We mask the source for the time of the
+ * operation to prevent an edge-triggered interrupt escaping meanwhile.
+ * The idea is from Manfred Spraul.  --macro
+ */
+	i = IO_APIC_VECTOR(irq);
+
+	v = apic_read(APIC_TMR + ((i & ~0x1f) >> 1));
+
+	ack_APIC_irq();
+
+	if (!(v & (1 << (i & 0x1f)))) {
+		atomic_inc(&irq_mis_count);
+		spin_lock(&ioapic_lock);
+		__mask_and_edge_IO_APIC_irq(irq);
+		__unmask_and_level_IO_APIC_irq(irq);
+		spin_unlock(&ioapic_lock);
+	}
+}
+
+#ifdef CONFIG_PCI_MSI
+static unsigned int startup_edge_ioapic_vector(unsigned int vector)
+{
+	int irq = vector_to_irq(vector);
+
+	return startup_edge_ioapic_irq(irq);
+}
+
+static void ack_edge_ioapic_vector(unsigned int vector)
+{
+	int irq = vector_to_irq(vector);
+
+	move_native_irq(vector);
+	ack_edge_ioapic_irq(irq);
+}
+
+static unsigned int startup_level_ioapic_vector (unsigned int vector)
+{
+	int irq = vector_to_irq(vector);
+
+	return startup_level_ioapic_irq (irq);
+}
+
+static void end_level_ioapic_vector (unsigned int vector)
+{
+	int irq = vector_to_irq(vector);
+
+	move_native_irq(vector);
+	end_level_ioapic_irq(irq);
+}
+
+static void mask_IO_APIC_vector (unsigned int vector)
+{
+	int irq = vector_to_irq(vector);
+
+	mask_IO_APIC_irq(irq);
+}
+
+static void unmask_IO_APIC_vector (unsigned int vector)
+{
+	int irq = vector_to_irq(vector);
+
+	unmask_IO_APIC_irq(irq);
+}
+
+#ifdef CONFIG_SMP
+static void set_ioapic_affinity_vector (unsigned int vector,
+					cpumask_t cpu_mask)
+{
+	int irq = vector_to_irq(vector);
+
+	set_native_irq_info(vector, cpu_mask);
+	set_ioapic_affinity_irq(irq, cpu_mask);
+}
+#endif
+#endif
+
+static int ioapic_retrigger(unsigned int irq)
+{
+	send_IPI_self(IO_APIC_VECTOR(irq));
+
+	return 1;
+}
+
+/*
+ * Level and edge triggered IO-APIC interrupts need different handling,
+ * so we use two separate IRQ descriptors. Edge triggered IRQs can be
+ * handled with the level-triggered descriptor, but that one has slightly
+ * more overhead. Level-triggered interrupts cannot be handled with the
+ * edge-triggered handler, without risking IRQ storms and other ugly
+ * races.
+ */
+static struct hw_interrupt_type ioapic_edge_type __read_mostly = {
+	.typename 	= "IO-APIC-edge",
+	.startup 	= startup_edge_ioapic,
+	.shutdown 	= shutdown_edge_ioapic,
+	.enable 	= enable_edge_ioapic,
+	.disable 	= disable_edge_ioapic,
+	.ack 		= ack_edge_ioapic,
+	.end 		= end_edge_ioapic,
+#ifdef CONFIG_SMP
+	.set_affinity 	= set_ioapic_affinity,
+#endif
+	.retrigger	= ioapic_retrigger,
+};
+
+static struct hw_interrupt_type ioapic_level_type __read_mostly = {
+	.typename 	= "IO-APIC-level",
+	.startup 	= startup_level_ioapic,
+	.shutdown 	= shutdown_level_ioapic,
+	.enable 	= enable_level_ioapic,
+	.disable 	= disable_level_ioapic,
+	.ack 		= mask_and_ack_level_ioapic,
+	.end 		= end_level_ioapic,
+#ifdef CONFIG_SMP
+	.set_affinity 	= set_ioapic_affinity,
+#endif
+	.retrigger	= ioapic_retrigger,
+};
+#endif /* !CONFIG_XEN */
+
+static inline void init_IO_APIC_traps(void)
+{
+	int irq;
+
+	/*
+	 * NOTE! The local APIC isn't very good at handling
+	 * multiple interrupts at the same interrupt level.
+	 * As the interrupt level is determined by taking the
+	 * vector number and shifting that right by 4, we
+	 * want to spread these out a bit so that they don't
+	 * all fall in the same interrupt level.
+	 *
+	 * Also, we've got to be careful not to trash gate
+	 * 0x80, because int 0x80 is hm, kind of importantish. ;)
+	 */
+	for (irq = 0; irq < NR_IRQS ; irq++) {
+		int tmp = irq;
+		if (use_pci_vector()) {
+			if (!platform_legacy_irq(tmp))
+				if ((tmp = vector_to_irq(tmp)) == -1)
+					continue;
+		}
+		if (IO_APIC_IRQ(tmp) && !IO_APIC_VECTOR(tmp)) {
+			/*
+			 * Hmm.. We don't have an entry for this,
+			 * so default to an old-fashioned 8259
+			 * interrupt if we can..
+			 */
+			if (irq < 16)
+				make_8259A_irq(irq);
+#ifndef CONFIG_XEN
+			else
+				/* Strange. Oh, well.. */
+				irq_desc[irq].chip = &no_irq_type;
+#endif
+		}
+	}
+}
+
+#ifndef CONFIG_XEN
+static void enable_lapic_irq (unsigned int irq)
+{
+	unsigned long v;
+
+	v = apic_read(APIC_LVT0);
+	apic_write_around(APIC_LVT0, v & ~APIC_LVT_MASKED);
+}
+
+static void disable_lapic_irq (unsigned int irq)
+{
+	unsigned long v;
+
+	v = apic_read(APIC_LVT0);
+	apic_write_around(APIC_LVT0, v | APIC_LVT_MASKED);
+}
+
+static void ack_lapic_irq (unsigned int irq)
+{
+	ack_APIC_irq();
+}
+
+static void end_lapic_irq (unsigned int i) { /* nothing */ }
+
+static struct hw_interrupt_type lapic_irq_type __read_mostly = {
+	.typename 	= "local-APIC-edge",
+	.startup 	= NULL, /* startup_irq() not used for IRQ0 */
+	.shutdown 	= NULL, /* shutdown_irq() not used for IRQ0 */
+	.enable 	= enable_lapic_irq,
+	.disable 	= disable_lapic_irq,
+	.ack 		= ack_lapic_irq,
+	.end 		= end_lapic_irq
+};
+
+static void setup_nmi (void)
+{
+	/*
+ 	 * Dirty trick to enable the NMI watchdog ...
+	 * We put the 8259A master into AEOI mode and
+	 * unmask on all local APICs LVT0 as NMI.
+	 *
+	 * The idea to use the 8259A in AEOI mode ('8259A Virtual Wire')
+	 * is from Maciej W. Rozycki - so we do not have to EOI from
+	 * the NMI handler or the timer interrupt.
+	 */ 
+	apic_printk(APIC_VERBOSE, KERN_INFO "activating NMI Watchdog ...");
+
+	on_each_cpu(enable_NMI_through_LVT0, NULL, 1, 1);
+
+	apic_printk(APIC_VERBOSE, " done.\n");
+}
+
+/*
+ * This looks a bit hackish but it's about the only one way of sending
+ * a few INTA cycles to 8259As and any associated glue logic.  ICR does
+ * not support the ExtINT mode, unfortunately.  We need to send these
+ * cycles as some i82489DX-based boards have glue logic that keeps the
+ * 8259A interrupt line asserted until INTA.  --macro
+ */
+static inline void unlock_ExtINT_logic(void)
+{
+	int apic, pin, i;
+	struct IO_APIC_route_entry entry0, entry1;
+	unsigned char save_control, save_freq_select;
+	unsigned long flags;
+
+	pin  = find_isa_irq_pin(8, mp_INT);
+	apic = find_isa_irq_apic(8, mp_INT);
+	if (pin == -1)
+		return;
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	*(((int *)&entry0) + 1) = io_apic_read(apic, 0x11 + 2 * pin);
+	*(((int *)&entry0) + 0) = io_apic_read(apic, 0x10 + 2 * pin);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+	clear_IO_APIC_pin(apic, pin);
+
+	memset(&entry1, 0, sizeof(entry1));
+
+	entry1.dest_mode = 0;			/* physical delivery */
+	entry1.mask = 0;			/* unmask IRQ now */
+	entry1.dest.physical.physical_dest = hard_smp_processor_id();
+	entry1.delivery_mode = dest_ExtINT;
+	entry1.polarity = entry0.polarity;
+	entry1.trigger = 0;
+	entry1.vector = 0;
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	io_apic_write(apic, 0x11 + 2 * pin, *(((int *)&entry1) + 1));
+	io_apic_write(apic, 0x10 + 2 * pin, *(((int *)&entry1) + 0));
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	save_control = CMOS_READ(RTC_CONTROL);
+	save_freq_select = CMOS_READ(RTC_FREQ_SELECT);
+	CMOS_WRITE((save_freq_select & ~RTC_RATE_SELECT) | 0x6,
+		   RTC_FREQ_SELECT);
+	CMOS_WRITE(save_control | RTC_PIE, RTC_CONTROL);
+
+	i = 100;
+	while (i-- > 0) {
+		mdelay(10);
+		if ((CMOS_READ(RTC_INTR_FLAGS) & RTC_PF) == RTC_PF)
+			i -= 10;
+	}
+
+	CMOS_WRITE(save_control, RTC_CONTROL);
+	CMOS_WRITE(save_freq_select, RTC_FREQ_SELECT);
+	clear_IO_APIC_pin(apic, pin);
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	io_apic_write(apic, 0x11 + 2 * pin, *(((int *)&entry0) + 1));
+	io_apic_write(apic, 0x10 + 2 * pin, *(((int *)&entry0) + 0));
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+}
+
+int timer_uses_ioapic_pin_0;
+
+/*
+ * This code may look a bit paranoid, but it's supposed to cooperate with
+ * a wide range of boards and BIOS bugs.  Fortunately only the timer IRQ
+ * is so screwy.  Thanks to Brian Perkins for testing/hacking this beast
+ * fanatically on his truly buggy board.
+ */
+static inline void check_timer(void)
+{
+	int apic1, pin1, apic2, pin2;
+	int vector;
+
+	/*
+	 * get/set the timer IRQ vector:
+	 */
+	disable_8259A_irq(0);
+	vector = assign_irq_vector(0);
+	set_intr_gate(vector, interrupt[0]);
+
+	/*
+	 * Subtle, code in do_timer_interrupt() expects an AEOI
+	 * mode for the 8259A whenever interrupts are routed
+	 * through I/O APICs.  Also IRQ0 has to be enabled in
+	 * the 8259A which implies the virtual wire has to be
+	 * disabled in the local APIC.
+	 */
+	apic_write_around(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_EXTINT);
+	init_8259A(1);
+	timer_ack = 1;
+	if (timer_over_8254 > 0)
+		enable_8259A_irq(0);
+
+	pin1  = find_isa_irq_pin(0, mp_INT);
+	apic1 = find_isa_irq_apic(0, mp_INT);
+	pin2  = ioapic_i8259.pin;
+	apic2 = ioapic_i8259.apic;
+
+	if (pin1 == 0)
+		timer_uses_ioapic_pin_0 = 1;
+
+	printk(KERN_INFO "..TIMER: vector=0x%02X apic1=%d pin1=%d apic2=%d pin2=%d\n",
+		vector, apic1, pin1, apic2, pin2);
+
+	if (pin1 != -1) {
+		/*
+		 * Ok, does IRQ0 through the IOAPIC work?
+		 */
+		unmask_IO_APIC_irq(0);
+		if (timer_irq_works()) {
+			if (nmi_watchdog == NMI_IO_APIC) {
+				disable_8259A_irq(0);
+				setup_nmi();
+				enable_8259A_irq(0);
+			}
+			if (disable_timer_pin_1 > 0)
+				clear_IO_APIC_pin(0, pin1);
+			return;
+		}
+		clear_IO_APIC_pin(apic1, pin1);
+		printk(KERN_ERR "..MP-BIOS bug: 8254 timer not connected to "
+				"IO-APIC\n");
+	}
+
+	printk(KERN_INFO "...trying to set up timer (IRQ0) through the 8259A ... ");
+	if (pin2 != -1) {
+		printk("\n..... (found pin %d) ...", pin2);
+		/*
+		 * legacy devices should be connected to IO APIC #0
+		 */
+		setup_ExtINT_IRQ0_pin(apic2, pin2, vector);
+		if (timer_irq_works()) {
+			printk("works.\n");
+			if (pin1 != -1)
+				replace_pin_at_irq(0, apic1, pin1, apic2, pin2);
+			else
+				add_pin_to_irq(0, apic2, pin2);
+			if (nmi_watchdog == NMI_IO_APIC) {
+				setup_nmi();
+			}
+			return;
+		}
+		/*
+		 * Cleanup, just in case ...
+		 */
+		clear_IO_APIC_pin(apic2, pin2);
+	}
+	printk(" failed.\n");
+
+	if (nmi_watchdog == NMI_IO_APIC) {
+		printk(KERN_WARNING "timer doesn't work through the IO-APIC - disabling NMI Watchdog!\n");
+		nmi_watchdog = 0;
+	}
+
+	printk(KERN_INFO "...trying to set up timer as Virtual Wire IRQ...");
+
+	disable_8259A_irq(0);
+	irq_desc[0].chip = &lapic_irq_type;
+	apic_write_around(APIC_LVT0, APIC_DM_FIXED | vector);	/* Fixed mode */
+	enable_8259A_irq(0);
+
+	if (timer_irq_works()) {
+		printk(" works.\n");
+		return;
+	}
+	apic_write_around(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_FIXED | vector);
+	printk(" failed.\n");
+
+	printk(KERN_INFO "...trying to set up timer as ExtINT IRQ...");
+
+	timer_ack = 0;
+	init_8259A(0);
+	make_8259A_irq(0);
+	apic_write_around(APIC_LVT0, APIC_DM_EXTINT);
+
+	unlock_ExtINT_logic();
+
+	if (timer_irq_works()) {
+		printk(" works.\n");
+		return;
+	}
+	printk(" failed :(.\n");
+	panic("IO-APIC + timer doesn't work!  Boot with apic=debug and send a "
+		"report.  Then try booting with the 'noapic' option");
+}
+#else
+int timer_uses_ioapic_pin_0 = 0;
+#define check_timer() ((void)0)
+#endif
+
+/*
+ *
+ * IRQ's that are handled by the PIC in the MPS IOAPIC case.
+ * - IRQ2 is the cascade IRQ, and cannot be a io-apic IRQ.
+ *   Linux doesn't really care, as it's not actually used
+ *   for any interrupt handling anyway.
+ */
+#define PIC_IRQS	(1 << PIC_CASCADE_IR)
+
+void __init setup_IO_APIC(void)
+{
+	enable_IO_APIC();
+
+	if (acpi_ioapic)
+		io_apic_irqs = ~0;	/* all IRQs go through IOAPIC */
+	else
+		io_apic_irqs = ~PIC_IRQS;
+
+	printk("ENABLING IO-APIC IRQs\n");
+
+	/*
+	 * Set up IO-APIC IRQ routing.
+	 */
+	if (!acpi_ioapic)
+		setup_ioapic_ids_from_mpc();
+#ifndef CONFIG_XEN
+	sync_Arb_IDs();
+#endif
+	setup_IO_APIC_irqs();
+	init_IO_APIC_traps();
+	check_timer();
+	if (!acpi_ioapic)
+		print_IO_APIC();
+}
+
+static int __init setup_disable_8254_timer(char *s)
+{
+	timer_over_8254 = -1;
+	return 1;
+}
+static int __init setup_enable_8254_timer(char *s)
+{
+	timer_over_8254 = 2;
+	return 1;
+}
+
+__setup("disable_8254_timer", setup_disable_8254_timer);
+__setup("enable_8254_timer", setup_enable_8254_timer);
+
+/*
+ *	Called after all the initialization is done. If we didnt find any
+ *	APIC bugs then we can allow the modify fast path
+ */
+ 
+static int __init io_apic_bug_finalize(void)
+{
+	if(sis_apic_bug == -1)
+		sis_apic_bug = 0;
+	if (is_initial_xendomain()) {
+		struct xen_platform_op op = { .cmd = XENPF_platform_quirk };
+		op.u.platform_quirk.quirk_id = sis_apic_bug ?
+			QUIRK_IOAPIC_BAD_REGSEL : QUIRK_IOAPIC_GOOD_REGSEL;
+		VOID(HYPERVISOR_platform_op(&op));
+	}
+	return 0;
+}
+
+late_initcall(io_apic_bug_finalize);
+
+struct sysfs_ioapic_data {
+	struct sys_device dev;
+	struct IO_APIC_route_entry entry[0];
+};
+static struct sysfs_ioapic_data * mp_ioapic_data[MAX_IO_APICS];
+
+static int ioapic_suspend(struct sys_device *dev, pm_message_t state)
+{
+	struct IO_APIC_route_entry *entry;
+	struct sysfs_ioapic_data *data;
+	unsigned long flags;
+	int i;
+	
+	data = container_of(dev, struct sysfs_ioapic_data, dev);
+	entry = data->entry;
+	spin_lock_irqsave(&ioapic_lock, flags);
+	for (i = 0; i < nr_ioapic_registers[dev->id]; i ++, entry ++ ) {
+		*(((int *)entry) + 1) = io_apic_read(dev->id, 0x11 + 2 * i);
+		*(((int *)entry) + 0) = io_apic_read(dev->id, 0x10 + 2 * i);
+	}
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	return 0;
+}
+
+static int ioapic_resume(struct sys_device *dev)
+{
+	struct IO_APIC_route_entry *entry;
+	struct sysfs_ioapic_data *data;
+	unsigned long flags;
+	union IO_APIC_reg_00 reg_00;
+	int i;
+	
+	data = container_of(dev, struct sysfs_ioapic_data, dev);
+	entry = data->entry;
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	reg_00.raw = io_apic_read(dev->id, 0);
+	if (reg_00.bits.ID != mp_ioapics[dev->id].mpc_apicid) {
+		reg_00.bits.ID = mp_ioapics[dev->id].mpc_apicid;
+		io_apic_write(dev->id, 0, reg_00.raw);
+	}
+	for (i = 0; i < nr_ioapic_registers[dev->id]; i ++, entry ++ ) {
+		io_apic_write(dev->id, 0x11+2*i, *(((int *)entry)+1));
+		io_apic_write(dev->id, 0x10+2*i, *(((int *)entry)+0));
+	}
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	return 0;
+}
+
+static struct sysdev_class ioapic_sysdev_class = {
+	set_kset_name("ioapic"),
+	.suspend = ioapic_suspend,
+	.resume = ioapic_resume,
+};
+
+static int __init ioapic_init_sysfs(void)
+{
+	struct sys_device * dev;
+	int i, size, error = 0;
+
+	error = sysdev_class_register(&ioapic_sysdev_class);
+	if (error)
+		return error;
+
+	for (i = 0; i < nr_ioapics; i++ ) {
+		size = sizeof(struct sys_device) + nr_ioapic_registers[i] 
+			* sizeof(struct IO_APIC_route_entry);
+		mp_ioapic_data[i] = kmalloc(size, GFP_KERNEL);
+		if (!mp_ioapic_data[i]) {
+			printk(KERN_ERR "Can't suspend/resume IOAPIC %d\n", i);
+			continue;
+		}
+		memset(mp_ioapic_data[i], 0, size);
+		dev = &mp_ioapic_data[i]->dev;
+		dev->id = i; 
+		dev->cls = &ioapic_sysdev_class;
+		error = sysdev_register(dev);
+		if (error) {
+			kfree(mp_ioapic_data[i]);
+			mp_ioapic_data[i] = NULL;
+			printk(KERN_ERR "Can't suspend/resume IOAPIC %d\n", i);
+			continue;
+		}
+	}
+
+	return 0;
+}
+
+device_initcall(ioapic_init_sysfs);
+
+/* --------------------------------------------------------------------------
+                          ACPI-based IOAPIC Configuration
+   -------------------------------------------------------------------------- */
+
+#ifdef CONFIG_ACPI
+
+int __init io_apic_get_unique_id (int ioapic, int apic_id)
+{
+#ifndef CONFIG_XEN
+	union IO_APIC_reg_00 reg_00;
+	static physid_mask_t apic_id_map = PHYSID_MASK_NONE;
+	physid_mask_t tmp;
+	unsigned long flags;
+	int i = 0;
+
+	/*
+	 * The P4 platform supports up to 256 APIC IDs on two separate APIC 
+	 * buses (one for LAPICs, one for IOAPICs), where predecessors only 
+	 * supports up to 16 on one shared APIC bus.
+	 * 
+	 * TBD: Expand LAPIC/IOAPIC support on P4-class systems to take full
+	 *      advantage of new APIC bus architecture.
+	 */
+
+	if (physids_empty(apic_id_map))
+		apic_id_map = ioapic_phys_id_map(phys_cpu_present_map);
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	reg_00.raw = io_apic_read(ioapic, 0);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	if (apic_id >= get_physical_broadcast()) {
+		printk(KERN_WARNING "IOAPIC[%d]: Invalid apic_id %d, trying "
+			"%d\n", ioapic, apic_id, reg_00.bits.ID);
+		apic_id = reg_00.bits.ID;
+	}
+
+	/*
+	 * Every APIC in a system must have a unique ID or we get lots of nice 
+	 * 'stuck on smp_invalidate_needed IPI wait' messages.
+	 */
+	if (check_apicid_used(apic_id_map, apic_id)) {
+
+		for (i = 0; i < get_physical_broadcast(); i++) {
+			if (!check_apicid_used(apic_id_map, i))
+				break;
+		}
+
+		if (i == get_physical_broadcast())
+			panic("Max apic_id exceeded!\n");
+
+		printk(KERN_WARNING "IOAPIC[%d]: apic_id %d already used, "
+			"trying %d\n", ioapic, apic_id, i);
+
+		apic_id = i;
+	} 
+
+	tmp = apicid_to_cpu_present(apic_id);
+	physids_or(apic_id_map, apic_id_map, tmp);
+
+	if (reg_00.bits.ID != apic_id) {
+		reg_00.bits.ID = apic_id;
+
+		spin_lock_irqsave(&ioapic_lock, flags);
+		io_apic_write(ioapic, 0, reg_00.raw);
+		reg_00.raw = io_apic_read(ioapic, 0);
+		spin_unlock_irqrestore(&ioapic_lock, flags);
+
+		/* Sanity check */
+		if (reg_00.bits.ID != apic_id) {
+			printk("IOAPIC[%d]: Unable to change apic_id!\n", ioapic);
+			return -1;
+		}
+	}
+
+	apic_printk(APIC_VERBOSE, KERN_INFO
+			"IOAPIC[%d]: Assigned apic_id %d\n", ioapic, apic_id);
+#endif /* !CONFIG_XEN */
+
+	return apic_id;
+}
+
+
+int __init io_apic_get_version (int ioapic)
+{
+	union IO_APIC_reg_01	reg_01;
+	unsigned long flags;
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	reg_01.raw = io_apic_read(ioapic, 1);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	return reg_01.bits.version;
+}
+
+
+int __init io_apic_get_redir_entries (int ioapic)
+{
+	union IO_APIC_reg_01	reg_01;
+	unsigned long flags;
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	reg_01.raw = io_apic_read(ioapic, 1);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	return reg_01.bits.entries;
+}
+
+
+int io_apic_set_pci_routing (int ioapic, int pin, int irq, int edge_level, int active_high_low)
+{
+	struct IO_APIC_route_entry entry;
+	unsigned long flags;
+
+	if (!IO_APIC_IRQ(irq)) {
+		printk(KERN_ERR "IOAPIC[%d]: Invalid reference to IRQ 0\n",
+			ioapic);
+		return -EINVAL;
+	}
+
+	/*
+	 * Generate a PCI IRQ routing entry and program the IOAPIC accordingly.
+	 * Note that we mask (disable) IRQs now -- these get enabled when the
+	 * corresponding device driver registers for this IRQ.
+	 */
+
+	memset(&entry,0,sizeof(entry));
+
+	entry.delivery_mode = INT_DELIVERY_MODE;
+	entry.dest_mode = INT_DEST_MODE;
+	entry.dest.logical.logical_dest = cpu_mask_to_apicid(TARGET_CPUS);
+	entry.trigger = edge_level;
+	entry.polarity = active_high_low;
+	entry.mask  = 1;
+
+	/*
+	 * IRQs < 16 are already in the irq_2_pin[] map
+	 */
+	if (irq >= 16)
+		add_pin_to_irq(irq, ioapic, pin);
+
+	entry.vector = assign_irq_vector(irq);
+
+	apic_printk(APIC_DEBUG, KERN_DEBUG "IOAPIC[%d]: Set PCI routing entry "
+		"(%d-%d -> 0x%x -> IRQ %d Mode:%i Active:%i)\n", ioapic,
+		mp_ioapics[ioapic].mpc_apicid, pin, entry.vector, irq,
+		edge_level, active_high_low);
+
+	ioapic_register_intr(irq, entry.vector, edge_level);
+
+	if (!ioapic && (irq < 16))
+		disable_8259A_irq(irq);
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	io_apic_write(ioapic, 0x11+2*pin, *(((int *)&entry)+1));
+	io_apic_write(ioapic, 0x10+2*pin, *(((int *)&entry)+0));
+	set_native_irq_info(use_pci_vector() ? entry.vector : irq, TARGET_CPUS);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	return 0;
+}
+
+#endif /* CONFIG_ACPI */
diff -rpuN linux-2.6.18.8/arch/i386/kernel/ioport-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/ioport-xen.c
--- linux-2.6.18.8/arch/i386/kernel/ioport-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/ioport-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,123 @@
+/*
+ *	linux/arch/i386/kernel/ioport.c
+ *
+ * This contains the io-permission bitmap code - written by obz, with changes
+ * by Linus.
+ */
+
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/capability.h>
+#include <linux/errno.h>
+#include <linux/types.h>
+#include <linux/ioport.h>
+#include <linux/smp.h>
+#include <linux/smp_lock.h>
+#include <linux/stddef.h>
+#include <linux/slab.h>
+#include <linux/thread_info.h>
+#include <xen/interface/physdev.h>
+
+/* Set EXTENT bits starting at BASE in BITMAP to value TURN_ON. */
+static void set_bitmap(unsigned long *bitmap, unsigned int base, unsigned int extent, int new_value)
+{
+	unsigned long mask;
+	unsigned long *bitmap_base = bitmap + (base / BITS_PER_LONG);
+	unsigned int low_index = base & (BITS_PER_LONG-1);
+	int length = low_index + extent;
+
+	if (low_index != 0) {
+		mask = (~0UL << low_index);
+		if (length < BITS_PER_LONG)
+			mask &= ~(~0UL << length);
+		if (new_value)
+			*bitmap_base++ |= mask;
+		else
+			*bitmap_base++ &= ~mask;
+		length -= BITS_PER_LONG;
+	}
+
+	mask = (new_value ? ~0UL : 0UL);
+	while (length >= BITS_PER_LONG) {
+		*bitmap_base++ = mask;
+		length -= BITS_PER_LONG;
+	}
+
+	if (length > 0) {
+		mask = ~(~0UL << length);
+		if (new_value)
+			*bitmap_base++ |= mask;
+		else
+			*bitmap_base++ &= ~mask;
+	}
+}
+
+
+/*
+ * this changes the io permissions bitmap in the current task.
+ */
+asmlinkage long sys_ioperm(unsigned long from, unsigned long num, int turn_on)
+{
+	struct thread_struct * t = &current->thread;
+	unsigned long *bitmap;
+	struct physdev_set_iobitmap set_iobitmap;
+
+	if ((from + num <= from) || (from + num > IO_BITMAP_BITS))
+		return -EINVAL;
+	if (turn_on && !capable(CAP_SYS_RAWIO))
+		return -EPERM;
+
+	/*
+	 * If it's the first ioperm() call in this thread's lifetime, set the
+	 * IO bitmap up. ioperm() is much less timing critical than clone(),
+	 * this is why we delay this operation until now:
+	 */
+	if (!t->io_bitmap_ptr) {
+		bitmap = kmalloc(IO_BITMAP_BYTES, GFP_KERNEL);
+		if (!bitmap)
+			return -ENOMEM;
+
+		memset(bitmap, 0xff, IO_BITMAP_BYTES);
+		t->io_bitmap_ptr = bitmap;
+		set_thread_flag(TIF_IO_BITMAP);
+
+		set_xen_guest_handle(set_iobitmap.bitmap, (char *)bitmap);
+		set_iobitmap.nr_ports = IO_BITMAP_BITS;
+		WARN_ON(HYPERVISOR_physdev_op(PHYSDEVOP_set_iobitmap,
+					      &set_iobitmap));
+	}
+
+	set_bitmap(t->io_bitmap_ptr, from, num, !turn_on);
+
+	return 0;
+}
+
+/*
+ * sys_iopl has to be used when you want to access the IO ports
+ * beyond the 0x3ff range: to get the full 65536 ports bitmapped
+ * you'd need 8kB of bitmaps/process, which is a bit excessive.
+ *
+ * Here we just change the eflags value on the stack: we allow
+ * only the super-user to do it. This depends on the stack-layout
+ * on system-call entry - see also fork() and the signal handling
+ * code.
+ */
+
+asmlinkage long sys_iopl(unsigned long unused)
+{
+	volatile struct pt_regs * regs = (struct pt_regs *) &unused;
+	unsigned int level = regs->ebx;
+	struct thread_struct *t = &current->thread;
+	unsigned int old = (t->iopl >> 12) & 3;
+
+	if (level > 3)
+		return -EINVAL;
+	/* Trying to gain more privileges? */
+	if (level > old) {
+		if (!capable(CAP_SYS_RAWIO))
+			return -EPERM;
+	}
+	t->iopl = level << 12;
+	set_iopl_mask(t->iopl);
+	return 0;
+}
diff -rpuN linux-2.6.18.8/arch/i386/kernel/irq-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/irq-xen.c
--- linux-2.6.18.8/arch/i386/kernel/irq-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/irq-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,324 @@
+/*
+ *	linux/arch/i386/kernel/irq.c
+ *
+ *	Copyright (C) 1992, 1998 Linus Torvalds, Ingo Molnar
+ *
+ * This file contains the lowest level x86-specific interrupt
+ * entry, irq-stacks and irq statistics code. All the remaining
+ * irq logic is done by the generic kernel/irq/ code and
+ * by the x86-specific irq controller code. (e.g. i8259.c and
+ * io_apic.c.)
+ */
+
+#include <asm/uaccess.h>
+#include <linux/module.h>
+#include <linux/seq_file.h>
+#include <linux/interrupt.h>
+#include <linux/kernel_stat.h>
+#include <linux/notifier.h>
+#include <linux/cpu.h>
+#include <linux/delay.h>
+
+DEFINE_PER_CPU(irq_cpustat_t, irq_stat) ____cacheline_internodealigned_in_smp;
+EXPORT_PER_CPU_SYMBOL(irq_stat);
+
+#ifndef CONFIG_X86_LOCAL_APIC
+/*
+ * 'what should we do if we get a hw irq event on an illegal vector'.
+ * each architecture has to answer this themselves.
+ */
+void ack_bad_irq(unsigned int irq)
+{
+	printk("unexpected IRQ trap at vector %02x\n", irq);
+}
+#endif
+
+#ifdef CONFIG_4KSTACKS
+/*
+ * per-CPU IRQ handling contexts (thread information and stack)
+ */
+union irq_ctx {
+	struct thread_info      tinfo;
+	u32                     stack[THREAD_SIZE/sizeof(u32)];
+};
+
+static union irq_ctx *hardirq_ctx[NR_CPUS] __read_mostly;
+static union irq_ctx *softirq_ctx[NR_CPUS] __read_mostly;
+#endif
+
+/*
+ * do_IRQ handles all normal device IRQ's (the special
+ * SMP cross-CPU interrupts have their own specific
+ * handlers).
+ */
+fastcall unsigned int do_IRQ(struct pt_regs *regs)
+{	
+	/* high bit used in ret_from_ code */
+	int irq = ~regs->orig_eax;
+#ifdef CONFIG_4KSTACKS
+	union irq_ctx *curctx, *irqctx;
+	u32 *isp;
+#endif
+
+	if (unlikely((unsigned)irq >= NR_IRQS)) {
+		printk(KERN_EMERG "%s: cannot handle IRQ %d\n",
+					__FUNCTION__, irq);
+		BUG();
+	}
+
+	irq_enter();
+#ifdef CONFIG_DEBUG_STACKOVERFLOW
+	/* Debugging check for stack overflow: is there less than 1KB free? */
+	{
+		long esp;
+
+		__asm__ __volatile__("andl %%esp,%0" :
+					"=r" (esp) : "0" (THREAD_SIZE - 1));
+		if (unlikely(esp < (sizeof(struct thread_info) + STACK_WARN))) {
+			printk("do_IRQ: stack overflow: %ld\n",
+				esp - sizeof(struct thread_info));
+			dump_stack();
+		}
+	}
+#endif
+
+#ifdef CONFIG_4KSTACKS
+
+	curctx = (union irq_ctx *) current_thread_info();
+	irqctx = hardirq_ctx[smp_processor_id()];
+
+	/*
+	 * this is where we switch to the IRQ stack. However, if we are
+	 * already using the IRQ stack (because we interrupted a hardirq
+	 * handler) we can't do that and just have to keep using the
+	 * current stack (which is the irq stack already after all)
+	 */
+	if (curctx != irqctx) {
+		int arg1, arg2, ebx;
+
+		/* build the stack frame on the IRQ stack */
+		isp = (u32*) ((char*)irqctx + sizeof(*irqctx));
+		irqctx->tinfo.task = curctx->tinfo.task;
+		irqctx->tinfo.previous_esp = current_stack_pointer;
+
+		/*
+		 * Copy the softirq bits in preempt_count so that the
+		 * softirq checks work in the hardirq context.
+		 */
+		irqctx->tinfo.preempt_count =
+			(irqctx->tinfo.preempt_count & ~SOFTIRQ_MASK) |
+			(curctx->tinfo.preempt_count & SOFTIRQ_MASK);
+
+		asm volatile(
+			"       xchgl   %%ebx,%%esp      \n"
+			"       call    __do_IRQ         \n"
+			"       movl   %%ebx,%%esp      \n"
+			: "=a" (arg1), "=d" (arg2), "=b" (ebx)
+			:  "0" (irq),   "1" (regs),  "2" (isp)
+			: "memory", "cc", "ecx"
+		);
+	} else
+#endif
+		__do_IRQ(irq, regs);
+
+	irq_exit();
+
+	return 1;
+}
+
+#ifdef CONFIG_4KSTACKS
+
+/*
+ * These should really be __section__(".bss.page_aligned") as well, but
+ * gcc's 3.0 and earlier don't handle that correctly.
+ */
+static char softirq_stack[NR_CPUS * THREAD_SIZE]
+		__attribute__((__aligned__(THREAD_SIZE)));
+
+static char hardirq_stack[NR_CPUS * THREAD_SIZE]
+		__attribute__((__aligned__(THREAD_SIZE)));
+
+/*
+ * allocate per-cpu stacks for hardirq and for softirq processing
+ */
+void irq_ctx_init(int cpu)
+{
+	union irq_ctx *irqctx;
+
+	if (hardirq_ctx[cpu])
+		return;
+
+	irqctx = (union irq_ctx*) &hardirq_stack[cpu*THREAD_SIZE];
+	irqctx->tinfo.task              = NULL;
+	irqctx->tinfo.exec_domain       = NULL;
+	irqctx->tinfo.cpu               = cpu;
+	irqctx->tinfo.preempt_count     = HARDIRQ_OFFSET;
+	irqctx->tinfo.addr_limit        = MAKE_MM_SEG(0);
+
+	hardirq_ctx[cpu] = irqctx;
+
+	irqctx = (union irq_ctx*) &softirq_stack[cpu*THREAD_SIZE];
+	irqctx->tinfo.task              = NULL;
+	irqctx->tinfo.exec_domain       = NULL;
+	irqctx->tinfo.cpu               = cpu;
+	irqctx->tinfo.preempt_count     = 0;
+	irqctx->tinfo.addr_limit        = MAKE_MM_SEG(0);
+
+	softirq_ctx[cpu] = irqctx;
+
+	printk("CPU %u irqstacks, hard=%p soft=%p\n",
+		cpu,hardirq_ctx[cpu],softirq_ctx[cpu]);
+}
+
+void irq_ctx_exit(int cpu)
+{
+	hardirq_ctx[cpu] = NULL;
+}
+
+extern asmlinkage void __do_softirq(void);
+
+asmlinkage void do_softirq(void)
+{
+	unsigned long flags;
+	struct thread_info *curctx;
+	union irq_ctx *irqctx;
+	u32 *isp;
+
+	if (in_interrupt())
+		return;
+
+	local_irq_save(flags);
+
+	if (local_softirq_pending()) {
+		curctx = current_thread_info();
+		irqctx = softirq_ctx[smp_processor_id()];
+		irqctx->tinfo.task = curctx->task;
+		irqctx->tinfo.previous_esp = current_stack_pointer;
+
+		/* build the stack frame on the softirq stack */
+		isp = (u32*) ((char*)irqctx + sizeof(*irqctx));
+
+		asm volatile(
+			"       xchgl   %%ebx,%%esp     \n"
+			"       call    __do_softirq    \n"
+			"       movl    %%ebx,%%esp     \n"
+			: "=b"(isp)
+			: "0"(isp)
+			: "memory", "cc", "edx", "ecx", "eax"
+		);
+		/*
+		 * Shouldnt happen, we returned above if in_interrupt():
+	 	 */
+		WARN_ON_ONCE(softirq_count());
+	}
+
+	local_irq_restore(flags);
+}
+
+EXPORT_SYMBOL(do_softirq);
+#endif
+
+/*
+ * Interrupt statistics:
+ */
+
+atomic_t irq_err_count;
+
+/*
+ * /proc/interrupts printing:
+ */
+
+int show_interrupts(struct seq_file *p, void *v)
+{
+	int i = *(loff_t *) v, j;
+	struct irqaction * action;
+	unsigned long flags;
+
+	if (i == 0) {
+		seq_printf(p, "           ");
+		for_each_online_cpu(j)
+			seq_printf(p, "CPU%-8d",j);
+		seq_putc(p, '\n');
+	}
+
+	if (i < NR_IRQS) {
+		spin_lock_irqsave(&irq_desc[i].lock, flags);
+		action = irq_desc[i].action;
+		if (!action)
+			goto skip;
+		seq_printf(p, "%3d: ",i);
+#ifndef CONFIG_SMP
+		seq_printf(p, "%10u ", kstat_irqs(i));
+#else
+		for_each_online_cpu(j)
+			seq_printf(p, "%10u ", kstat_cpu(j).irqs[i]);
+#endif
+		seq_printf(p, " %14s", irq_desc[i].chip->typename);
+		seq_printf(p, "  %s", action->name);
+
+		for (action=action->next; action; action = action->next)
+			seq_printf(p, ", %s", action->name);
+
+		seq_putc(p, '\n');
+skip:
+		spin_unlock_irqrestore(&irq_desc[i].lock, flags);
+	} else if (i == NR_IRQS) {
+		seq_printf(p, "NMI: ");
+		for_each_online_cpu(j)
+			seq_printf(p, "%10u ", nmi_count(j));
+		seq_putc(p, '\n');
+#ifdef CONFIG_X86_LOCAL_APIC
+		seq_printf(p, "LOC: ");
+		for_each_online_cpu(j)
+			seq_printf(p, "%10u ",
+				per_cpu(irq_stat,j).apic_timer_irqs);
+		seq_putc(p, '\n');
+#endif
+		seq_printf(p, "ERR: %10u\n", atomic_read(&irq_err_count));
+#if defined(CONFIG_X86_IO_APIC)
+		seq_printf(p, "MIS: %10u\n", atomic_read(&irq_mis_count));
+#endif
+	}
+	return 0;
+}
+
+#ifdef CONFIG_HOTPLUG_CPU
+
+void fixup_irqs(cpumask_t map)
+{
+	unsigned int irq;
+	static int warned;
+
+	for (irq = 0; irq < NR_IRQS; irq++) {
+		cpumask_t mask;
+		if (irq == 2)
+			continue;
+
+		cpus_and(mask, irq_desc[irq].affinity, map);
+		if (any_online_cpu(mask) == NR_CPUS) {
+			/*printk("Breaking affinity for irq %i\n", irq);*/
+			mask = map;
+		}
+		if (irq_desc[irq].chip->set_affinity)
+			irq_desc[irq].chip->set_affinity(irq, mask);
+		else if (irq_desc[irq].action && !(warned++))
+			printk("Cannot set affinity for irq %i\n", irq);
+	}
+
+#if 0
+	barrier();
+	/* Ingo Molnar says: "after the IO-APIC masks have been redirected
+	   [note the nop - the interrupt-enable boundary on x86 is two
+	   instructions from sti] - to flush out pending hardirqs and
+	   IPIs. After this point nothing is supposed to reach this CPU." */
+	__asm__ __volatile__("sti; nop; cli");
+	barrier();
+#else
+	/* That doesn't seem sufficient.  Give it 1ms. */
+	local_irq_enable();
+	mdelay(1);
+	local_irq_disable();
+#endif
+}
+#endif
+
diff -rpuN linux-2.6.18.8/arch/i386/kernel/ldt-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/ldt-xen.c
--- linux-2.6.18.8/arch/i386/kernel/ldt-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/ldt-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,270 @@
+/*
+ * linux/kernel/ldt.c
+ *
+ * Copyright (C) 1992 Krishna Balasubramanian and Linus Torvalds
+ * Copyright (C) 1999 Ingo Molnar <mingo@redhat.com>
+ */
+
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/string.h>
+#include <linux/mm.h>
+#include <linux/smp.h>
+#include <linux/smp_lock.h>
+#include <linux/vmalloc.h>
+#include <linux/slab.h>
+
+#include <asm/uaccess.h>
+#include <asm/system.h>
+#include <asm/ldt.h>
+#include <asm/desc.h>
+#include <asm/mmu_context.h>
+
+#ifdef CONFIG_SMP /* avoids "defined but not used" warnig */
+static void flush_ldt(void *null)
+{
+	if (current->active_mm)
+		load_LDT(&current->active_mm->context);
+}
+#endif
+
+static int alloc_ldt(mm_context_t *pc, int mincount, int reload)
+{
+	void *oldldt;
+	void *newldt;
+	int oldsize;
+
+	if (mincount <= pc->size)
+		return 0;
+	oldsize = pc->size;
+	mincount = (mincount+511)&(~511);
+	if (mincount*LDT_ENTRY_SIZE > PAGE_SIZE)
+		newldt = vmalloc(mincount*LDT_ENTRY_SIZE);
+	else
+		newldt = kmalloc(mincount*LDT_ENTRY_SIZE, GFP_KERNEL);
+
+	if (!newldt)
+		return -ENOMEM;
+
+	if (oldsize)
+		memcpy(newldt, pc->ldt, oldsize*LDT_ENTRY_SIZE);
+	oldldt = pc->ldt;
+	memset(newldt+oldsize*LDT_ENTRY_SIZE, 0, (mincount-oldsize)*LDT_ENTRY_SIZE);
+	pc->ldt = newldt;
+	wmb();
+	pc->size = mincount;
+	wmb();
+
+	if (reload) {
+#ifdef CONFIG_SMP
+		cpumask_t mask;
+		preempt_disable();
+#endif
+		make_pages_readonly(
+			pc->ldt,
+			(pc->size * LDT_ENTRY_SIZE) / PAGE_SIZE,
+			XENFEAT_writable_descriptor_tables);
+		load_LDT(pc);
+#ifdef CONFIG_SMP
+		mask = cpumask_of_cpu(smp_processor_id());
+		if (!cpus_equal(current->mm->cpu_vm_mask, mask))
+			smp_call_function(flush_ldt, NULL, 1, 1);
+		preempt_enable();
+#endif
+	}
+	if (oldsize) {
+		make_pages_writable(
+			oldldt,
+			(oldsize * LDT_ENTRY_SIZE) / PAGE_SIZE,
+			XENFEAT_writable_descriptor_tables);
+		if (oldsize*LDT_ENTRY_SIZE > PAGE_SIZE)
+			vfree(oldldt);
+		else
+			kfree(oldldt);
+	}
+	return 0;
+}
+
+static inline int copy_ldt(mm_context_t *new, mm_context_t *old)
+{
+	int err = alloc_ldt(new, old->size, 0);
+	if (err < 0)
+		return err;
+	memcpy(new->ldt, old->ldt, old->size*LDT_ENTRY_SIZE);
+	make_pages_readonly(
+		new->ldt,
+		(new->size * LDT_ENTRY_SIZE) / PAGE_SIZE,
+		XENFEAT_writable_descriptor_tables);
+	return 0;
+}
+
+/*
+ * we do not have to muck with descriptors here, that is
+ * done in switch_mm() as needed.
+ */
+int init_new_context(struct task_struct *tsk, struct mm_struct *mm)
+{
+	struct mm_struct * old_mm;
+	int retval = 0;
+
+	init_MUTEX(&mm->context.sem);
+	mm->context.size = 0;
+	mm->context.has_foreign_mappings = 0;
+	old_mm = current->mm;
+	if (old_mm && old_mm->context.size > 0) {
+		down(&old_mm->context.sem);
+		retval = copy_ldt(&mm->context, &old_mm->context);
+		up(&old_mm->context.sem);
+	}
+	return retval;
+}
+
+/*
+ * No need to lock the MM as we are the last user
+ */
+void destroy_context(struct mm_struct *mm)
+{
+	if (mm->context.size) {
+		if (mm == current->active_mm)
+			clear_LDT();
+		make_pages_writable(
+			mm->context.ldt,
+			(mm->context.size * LDT_ENTRY_SIZE) / PAGE_SIZE,
+			XENFEAT_writable_descriptor_tables);
+		if (mm->context.size*LDT_ENTRY_SIZE > PAGE_SIZE)
+			vfree(mm->context.ldt);
+		else
+			kfree(mm->context.ldt);
+		mm->context.size = 0;
+	}
+}
+
+static int read_ldt(void __user * ptr, unsigned long bytecount)
+{
+	int err;
+	unsigned long size;
+	struct mm_struct * mm = current->mm;
+
+	if (!mm->context.size)
+		return 0;
+	if (bytecount > LDT_ENTRY_SIZE*LDT_ENTRIES)
+		bytecount = LDT_ENTRY_SIZE*LDT_ENTRIES;
+
+	down(&mm->context.sem);
+	size = mm->context.size*LDT_ENTRY_SIZE;
+	if (size > bytecount)
+		size = bytecount;
+
+	err = 0;
+	if (copy_to_user(ptr, mm->context.ldt, size))
+		err = -EFAULT;
+	up(&mm->context.sem);
+	if (err < 0)
+		goto error_return;
+	if (size != bytecount) {
+		/* zero-fill the rest */
+		if (clear_user(ptr+size, bytecount-size) != 0) {
+			err = -EFAULT;
+			goto error_return;
+		}
+	}
+	return bytecount;
+error_return:
+	return err;
+}
+
+static int read_default_ldt(void __user * ptr, unsigned long bytecount)
+{
+	int err;
+	unsigned long size;
+	void *address;
+
+	err = 0;
+	address = &default_ldt[0];
+	size = 5*sizeof(struct desc_struct);
+	if (size > bytecount)
+		size = bytecount;
+
+	err = size;
+	if (copy_to_user(ptr, address, size))
+		err = -EFAULT;
+
+	return err;
+}
+
+static int write_ldt(void __user * ptr, unsigned long bytecount, int oldmode)
+{
+	struct mm_struct * mm = current->mm;
+	__u32 entry_1, entry_2;
+	int error;
+	struct user_desc ldt_info;
+
+	error = -EINVAL;
+	if (bytecount != sizeof(ldt_info))
+		goto out;
+	error = -EFAULT; 	
+	if (copy_from_user(&ldt_info, ptr, sizeof(ldt_info)))
+		goto out;
+
+	error = -EINVAL;
+	if (ldt_info.entry_number >= LDT_ENTRIES)
+		goto out;
+	if (ldt_info.contents == 3) {
+		if (oldmode)
+			goto out;
+		if (ldt_info.seg_not_present == 0)
+			goto out;
+	}
+
+	down(&mm->context.sem);
+	if (ldt_info.entry_number >= mm->context.size) {
+		error = alloc_ldt(&current->mm->context, ldt_info.entry_number+1, 1);
+		if (error < 0)
+			goto out_unlock;
+	}
+
+   	/* Allow LDTs to be cleared by the user. */
+   	if (ldt_info.base_addr == 0 && ldt_info.limit == 0) {
+		if (oldmode || LDT_empty(&ldt_info)) {
+			entry_1 = 0;
+			entry_2 = 0;
+			goto install;
+		}
+	}
+
+	entry_1 = LDT_entry_a(&ldt_info);
+	entry_2 = LDT_entry_b(&ldt_info);
+	if (oldmode)
+		entry_2 &= ~(1 << 20);
+
+	/* Install the new entry ...  */
+install:
+	error = write_ldt_entry(mm->context.ldt, ldt_info.entry_number,
+				entry_1, entry_2);
+
+out_unlock:
+	up(&mm->context.sem);
+out:
+	return error;
+}
+
+asmlinkage int sys_modify_ldt(int func, void __user *ptr, unsigned long bytecount)
+{
+	int ret = -ENOSYS;
+
+	switch (func) {
+	case 0:
+		ret = read_ldt(ptr, bytecount);
+		break;
+	case 1:
+		ret = write_ldt(ptr, bytecount, 1);
+		break;
+	case 2:
+		ret = read_default_ldt(ptr, bytecount);
+		break;
+	case 0x11:
+		ret = write_ldt(ptr, bytecount, 0);
+		break;
+	}
+	return ret;
+}
diff -rpuN linux-2.6.18.8/arch/i386/kernel/machine_kexec.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/machine_kexec.c
--- linux-2.6.18.8/arch/i386/kernel/machine_kexec.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/machine_kexec.c	2008-08-21 11:36:07.000000000 +0200
@@ -19,123 +19,67 @@
 #include <asm/desc.h>
 #include <asm/system.h>
 
-#define PAGE_ALIGNED __attribute__ ((__aligned__(PAGE_SIZE)))
-
-#define L0_ATTR (_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY)
-#define L1_ATTR (_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY)
-#define L2_ATTR (_PAGE_PRESENT)
-
-#define LEVEL0_SIZE (1UL << 12UL)
-
-#ifndef CONFIG_X86_PAE
-#define LEVEL1_SIZE (1UL << 22UL)
-static u32 pgtable_level1[1024] PAGE_ALIGNED;
-
-static void identity_map_page(unsigned long address)
-{
-	unsigned long level1_index, level2_index;
-	u32 *pgtable_level2;
-
-	/* Find the current page table */
-	pgtable_level2 = __va(read_cr3());
+#ifdef CONFIG_XEN
+#include <xen/interface/kexec.h>
+#endif
 
-	/* Find the indexes of the physical address to identity map */
-	level1_index = (address % LEVEL1_SIZE)/LEVEL0_SIZE;
-	level2_index = address / LEVEL1_SIZE;
-
-	/* Identity map the page table entry */
-	pgtable_level1[level1_index] = address | L0_ATTR;
-	pgtable_level2[level2_index] = __pa(pgtable_level1) | L1_ATTR;
-
-	/* Flush the tlb so the new mapping takes effect.
-	 * Global tlb entries are not flushed but that is not an issue.
-	 */
-	load_cr3(pgtable_level2);
-}
+#define PAGE_ALIGNED __attribute__ ((__aligned__(PAGE_SIZE)))
+static u32 kexec_pgd[1024] PAGE_ALIGNED;
+#ifdef CONFIG_X86_PAE
+static u32 kexec_pmd0[1024] PAGE_ALIGNED;
+static u32 kexec_pmd1[1024] PAGE_ALIGNED;
+#endif
+static u32 kexec_pte0[1024] PAGE_ALIGNED;
+static u32 kexec_pte1[1024] PAGE_ALIGNED;
 
-#else
-#define LEVEL1_SIZE (1UL << 21UL)
-#define LEVEL2_SIZE (1UL << 30UL)
-static u64 pgtable_level1[512] PAGE_ALIGNED;
-static u64 pgtable_level2[512] PAGE_ALIGNED;
+#ifdef CONFIG_XEN
 
-static void identity_map_page(unsigned long address)
-{
-	unsigned long level1_index, level2_index, level3_index;
-	u64 *pgtable_level3;
+#define __ma(x) (pfn_to_mfn(__pa((x)) >> PAGE_SHIFT) << PAGE_SHIFT)
 
-	/* Find the current page table */
-	pgtable_level3 = __va(read_cr3());
+#if PAGES_NR > KEXEC_XEN_NO_PAGES
+#error PAGES_NR is greater than KEXEC_XEN_NO_PAGES - Xen support will break
+#endif
 
-	/* Find the indexes of the physical address to identity map */
-	level1_index = (address % LEVEL1_SIZE)/LEVEL0_SIZE;
-	level2_index = (address % LEVEL2_SIZE)/LEVEL1_SIZE;
-	level3_index = address / LEVEL2_SIZE;
-
-	/* Identity map the page table entry */
-	pgtable_level1[level1_index] = address | L0_ATTR;
-	pgtable_level2[level2_index] = __pa(pgtable_level1) | L1_ATTR;
-	set_64bit(&pgtable_level3[level3_index],
-					       __pa(pgtable_level2) | L2_ATTR);
-
-	/* Flush the tlb so the new mapping takes effect.
-	 * Global tlb entries are not flushed but that is not an issue.
-	 */
-	load_cr3(pgtable_level3);
-}
+#if PA_CONTROL_PAGE != 0
+#error PA_CONTROL_PAGE is non zero - Xen support will break
 #endif
 
-static void set_idt(void *newidt, __u16 limit)
+void machine_kexec_setup_load_arg(xen_kexec_image_t *xki, struct kimage *image)
 {
-	struct Xgt_desc_struct curidt;
+	void *control_page;
 
-	/* ia32 supports unaliged loads & stores */
-	curidt.size    = limit;
-	curidt.address = (unsigned long)newidt;
+	memset(xki->page_list, 0, sizeof(xki->page_list));
 
-	load_idt(&curidt);
-};
+	control_page = page_address(image->control_code_page);
+	memcpy(control_page, relocate_kernel, PAGE_SIZE);
 
+	xki->page_list[PA_CONTROL_PAGE] = __ma(control_page);
+	xki->page_list[PA_PGD] = __ma(kexec_pgd);
+#ifdef CONFIG_X86_PAE
+	xki->page_list[PA_PMD_0] = __ma(kexec_pmd0);
+	xki->page_list[PA_PMD_1] = __ma(kexec_pmd1);
+#endif
+	xki->page_list[PA_PTE_0] = __ma(kexec_pte0);
+	xki->page_list[PA_PTE_1] = __ma(kexec_pte1);
 
-static void set_gdt(void *newgdt, __u16 limit)
-{
-	struct Xgt_desc_struct curgdt;
-
-	/* ia32 supports unaligned loads & stores */
-	curgdt.size    = limit;
-	curgdt.address = (unsigned long)newgdt;
-
-	load_gdt(&curgdt);
-};
+}
 
-static void load_segments(void)
+int __init machine_kexec_setup_resources(struct resource *hypervisor,
+					 struct resource *phys_cpus,
+					 int nr_phys_cpus)
 {
-#define __STR(X) #X
-#define STR(X) __STR(X)
+	int k;
+
+	/* The per-cpu crash note resources belong to the hypervisor resource */
+	for (k = 0; k < nr_phys_cpus; k++)
+		request_resource(hypervisor, phys_cpus + k);
 
-	__asm__ __volatile__ (
-		"\tljmp $"STR(__KERNEL_CS)",$1f\n"
-		"\t1:\n"
-		"\tmovl $"STR(__KERNEL_DS)",%%eax\n"
-		"\tmovl %%eax,%%ds\n"
-		"\tmovl %%eax,%%es\n"
-		"\tmovl %%eax,%%fs\n"
-		"\tmovl %%eax,%%gs\n"
-		"\tmovl %%eax,%%ss\n"
-		::: "eax", "memory");
-#undef STR
-#undef __STR
+	return 0;
 }
 
-typedef asmlinkage NORET_TYPE void (*relocate_new_kernel_t)(
-					unsigned long indirection_page,
-					unsigned long reboot_code_buffer,
-					unsigned long start_address,
-					unsigned int has_pae) ATTRIB_NORET;
-
-extern const unsigned char relocate_new_kernel[];
-extern void relocate_new_kernel_end(void);
-extern const unsigned int relocate_new_kernel_size;
+void machine_kexec_register_resources(struct resource *res) { ; }
+
+#endif /* CONFIG_XEN */
 
 /*
  * A architecture hook called to validate the
@@ -163,49 +107,38 @@ void machine_kexec_cleanup(struct kimage
 {
 }
 
+#ifndef CONFIG_XEN
 /*
  * Do not allocate memory (or fail in any way) in machine_kexec().
  * We are past the point of no return, committed to rebooting now.
  */
 NORET_TYPE void machine_kexec(struct kimage *image)
 {
-	unsigned long page_list;
-	unsigned long reboot_code_buffer;
-
-	relocate_new_kernel_t rnk;
+	unsigned long page_list[PAGES_NR];
+	void *control_page;
 
 	/* Interrupts aren't acceptable while we reboot */
 	local_irq_disable();
 
-	/* Compute some offsets */
-	reboot_code_buffer = page_to_pfn(image->control_code_page)
-								<< PAGE_SHIFT;
-	page_list = image->head;
-
-	/* Set up an identity mapping for the reboot_code_buffer */
-	identity_map_page(reboot_code_buffer);
-
-	/* copy it out */
-	memcpy((void *)reboot_code_buffer, relocate_new_kernel,
-						relocate_new_kernel_size);
-
-	/* The segment registers are funny things, they have both a
-	 * visible and an invisible part.  Whenever the visible part is
-	 * set to a specific selector, the invisible part is loaded
-	 * with from a table in memory.  At no other time is the
-	 * descriptor table in memory accessed.
-	 *
-	 * I take advantage of this here by force loading the
-	 * segments, before I zap the gdt with an invalid value.
-	 */
-	load_segments();
-	/* The gdt & idt are now invalid.
-	 * If you want to load them you must set up your own idt & gdt.
-	 */
-	set_gdt(phys_to_virt(0),0);
-	set_idt(phys_to_virt(0),0);
-
-	/* now call it */
-	rnk = (relocate_new_kernel_t) reboot_code_buffer;
-	(*rnk)(page_list, reboot_code_buffer, image->start, cpu_has_pae);
+	control_page = page_address(image->control_code_page);
+	memcpy(control_page, relocate_kernel, PAGE_SIZE);
+
+	page_list[PA_CONTROL_PAGE] = __pa(control_page);
+	page_list[VA_CONTROL_PAGE] = (unsigned long)relocate_kernel;
+	page_list[PA_PGD] = __pa(kexec_pgd);
+	page_list[VA_PGD] = (unsigned long)kexec_pgd;
+#ifdef CONFIG_X86_PAE
+	page_list[PA_PMD_0] = __pa(kexec_pmd0);
+	page_list[VA_PMD_0] = (unsigned long)kexec_pmd0;
+	page_list[PA_PMD_1] = __pa(kexec_pmd1);
+	page_list[VA_PMD_1] = (unsigned long)kexec_pmd1;
+#endif
+	page_list[PA_PTE_0] = __pa(kexec_pte0);
+	page_list[VA_PTE_0] = (unsigned long)kexec_pte0;
+	page_list[PA_PTE_1] = __pa(kexec_pte1);
+	page_list[VA_PTE_1] = (unsigned long)kexec_pte1;
+
+	relocate_kernel((unsigned long)image->head, (unsigned long)page_list,
+			image->start, cpu_has_pae);
 }
+#endif
diff -rpuN linux-2.6.18.8/arch/i386/kernel/Makefile linux-2.6.18-xen-3.3.0/arch/i386/kernel/Makefile
--- linux-2.6.18.8/arch/i386/kernel/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -43,6 +43,7 @@ obj-$(CONFIG_K8_NB)		+= k8.o
 EXTRA_AFLAGS   := -traditional
 
 obj-$(CONFIG_SCx200)		+= scx200.o
+obj-$(CONFIG_XEN)		+= fixup.o
 
 # vsyscall.o contains the vsyscall DSO images as __initdata.
 # We must build both images before we can assemble it.
@@ -80,5 +81,8 @@ $(obj)/vsyscall-syms.o: $(src)/vsyscall.
 			$(obj)/vsyscall-sysenter.o $(obj)/vsyscall-note.o FORCE
 	$(call if_changed,syscall)
 
+early_printk-y            += ../../x86_64/kernel/early_printk.o
 k8-y                      += ../../x86_64/kernel/k8.o
 
+disabled-obj-$(CONFIG_XEN) := i8259.o reboot.o smpboot.o
+%/head.o %/head.s: $(if $(CONFIG_XEN),EXTRA_AFLAGS,dummy) :=
diff -rpuN linux-2.6.18.8/arch/i386/kernel/microcode-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/microcode-xen.c
--- linux-2.6.18.8/arch/i386/kernel/microcode-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/microcode-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,144 @@
+/*
+ *	Intel CPU Microcode Update Driver for Linux
+ *
+ *	Copyright (C) 2000-2004 Tigran Aivazian
+ *
+ *	This driver allows to upgrade microcode on Intel processors
+ *	belonging to IA-32 family - PentiumPro, Pentium II, 
+ *	Pentium III, Xeon, Pentium 4, etc.
+ *
+ *	Reference: Section 8.10 of Volume III, Intel Pentium 4 Manual, 
+ *	Order Number 245472 or free download from:
+ *		
+ *	http://developer.intel.com/design/pentium4/manuals/245472.htm
+ *
+ *	For more information, go to http://www.urbanmyth.org/microcode
+ *
+ *	This program is free software; you can redistribute it and/or
+ *	modify it under the terms of the GNU General Public License
+ *	as published by the Free Software Foundation; either version
+ *	2 of the License, or (at your option) any later version.
+ */
+
+//#define DEBUG /* pr_debug */
+#include <linux/capability.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/sched.h>
+#include <linux/cpumask.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <linux/miscdevice.h>
+#include <linux/spinlock.h>
+#include <linux/mm.h>
+#include <linux/mutex.h>
+#include <linux/syscalls.h>
+
+#include <asm/msr.h>
+#include <asm/uaccess.h>
+#include <asm/processor.h>
+
+MODULE_DESCRIPTION("Intel CPU (IA-32) Microcode Update Driver");
+MODULE_AUTHOR("Tigran Aivazian <tigran@veritas.com>");
+MODULE_LICENSE("GPL");
+
+static int verbose;
+module_param(verbose, int, 0644);
+
+#define MICROCODE_VERSION 	"1.14a-xen"
+
+#define DEFAULT_UCODE_DATASIZE 	(2000) 	  /* 2000 bytes */
+#define MC_HEADER_SIZE		(sizeof (microcode_header_t))  	  /* 48 bytes */
+#define DEFAULT_UCODE_TOTALSIZE (DEFAULT_UCODE_DATASIZE + MC_HEADER_SIZE) /* 2048 bytes */
+
+/* no concurrent ->write()s are allowed on /dev/cpu/microcode */
+static DEFINE_MUTEX(microcode_mutex);
+				
+static int microcode_open (struct inode *unused1, struct file *unused2)
+{
+	return capable(CAP_SYS_RAWIO) ? 0 : -EPERM;
+}
+
+
+static int do_microcode_update (const void __user *ubuf, size_t len)
+{
+	int err;
+	void *kbuf;
+
+	kbuf = vmalloc(len);
+	if (!kbuf)
+		return -ENOMEM;
+
+	if (copy_from_user(kbuf, ubuf, len) == 0) {
+		struct xen_platform_op op;
+
+		op.cmd = XENPF_microcode_update;
+		set_xen_guest_handle(op.u.microcode.data, kbuf);
+		op.u.microcode.length = len;
+		err = HYPERVISOR_platform_op(&op);
+	} else
+		err = -EFAULT;
+
+	vfree(kbuf);
+
+	return err;
+}
+
+static ssize_t microcode_write (struct file *file, const char __user *buf, size_t len, loff_t *ppos)
+{
+	ssize_t ret;
+
+	if (len < MC_HEADER_SIZE) {
+		printk(KERN_ERR "microcode: not enough data\n"); 
+		return -EINVAL;
+	}
+
+	mutex_lock(&microcode_mutex);
+
+	ret = do_microcode_update(buf, len);
+	if (!ret)
+		ret = (ssize_t)len;
+
+	mutex_unlock(&microcode_mutex);
+
+	return ret;
+}
+
+static struct file_operations microcode_fops = {
+	.owner		= THIS_MODULE,
+	.write		= microcode_write,
+	.open		= microcode_open,
+};
+
+static struct miscdevice microcode_dev = {
+	.minor		= MICROCODE_MINOR,
+	.name		= "microcode",
+	.fops		= &microcode_fops,
+};
+
+static int __init microcode_init (void)
+{
+	int error;
+
+	error = misc_register(&microcode_dev);
+	if (error) {
+		printk(KERN_ERR
+			"microcode: can't misc_register on minor=%d\n",
+			MICROCODE_MINOR);
+		return error;
+	}
+
+	printk(KERN_INFO 
+		"IA-32 Microcode Update Driver: v" MICROCODE_VERSION " <tigran@veritas.com>\n");
+	return 0;
+}
+
+static void __exit microcode_exit (void)
+{
+	misc_deregister(&microcode_dev);
+}
+
+module_init(microcode_init)
+module_exit(microcode_exit)
+MODULE_ALIAS_MISCDEV(MICROCODE_MINOR);
diff -rpuN linux-2.6.18.8/arch/i386/kernel/mpparse-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/mpparse-xen.c
--- linux-2.6.18.8/arch/i386/kernel/mpparse-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/mpparse-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1185 @@
+/*
+ *	Intel Multiprocessor Specification 1.1 and 1.4
+ *	compliant MP-table parsing routines.
+ *
+ *	(c) 1995 Alan Cox, Building #3 <alan@redhat.com>
+ *	(c) 1998, 1999, 2000 Ingo Molnar <mingo@redhat.com>
+ *
+ *	Fixes
+ *		Erich Boleyn	:	MP v1.4 and additional changes.
+ *		Alan Cox	:	Added EBDA scanning
+ *		Ingo Molnar	:	various cleanups and rewrites
+ *		Maciej W. Rozycki:	Bits for default MP configurations
+ *		Paul Diefenbaugh:	Added full ACPI support
+ */
+
+#include <linux/mm.h>
+#include <linux/init.h>
+#include <linux/acpi.h>
+#include <linux/delay.h>
+#include <linux/bootmem.h>
+#include <linux/smp_lock.h>
+#include <linux/kernel_stat.h>
+#include <linux/mc146818rtc.h>
+#include <linux/bitops.h>
+
+#include <asm/smp.h>
+#include <asm/acpi.h>
+#include <asm/mtrr.h>
+#include <asm/mpspec.h>
+#include <asm/io_apic.h>
+
+#include <mach_apic.h>
+#include <mach_mpparse.h>
+#include <bios_ebda.h>
+
+/* Have we found an MP table */
+int smp_found_config;
+unsigned int __initdata maxcpus = NR_CPUS;
+
+/*
+ * Various Linux-internal data structures created from the
+ * MP-table.
+ */
+int apic_version [MAX_APICS];
+int mp_bus_id_to_type [MAX_MP_BUSSES];
+int mp_bus_id_to_node [MAX_MP_BUSSES];
+int mp_bus_id_to_local [MAX_MP_BUSSES];
+int quad_local_to_mp_bus_id [NR_CPUS/4][4];
+int mp_bus_id_to_pci_bus [MAX_MP_BUSSES] = { [0 ... MAX_MP_BUSSES-1] = -1 };
+static int mp_current_pci_id;
+
+/* I/O APIC entries */
+struct mpc_config_ioapic mp_ioapics[MAX_IO_APICS];
+
+/* # of MP IRQ source entries */
+struct mpc_config_intsrc mp_irqs[MAX_IRQ_SOURCES];
+
+/* MP IRQ source entries */
+int mp_irq_entries;
+
+int nr_ioapics;
+
+int pic_mode;
+unsigned long mp_lapic_addr;
+
+unsigned int def_to_bigsmp = 0;
+
+/* Processor that is doing the boot up */
+unsigned int boot_cpu_physical_apicid = -1U;
+/* Internal processor count */
+static unsigned int __devinitdata num_processors;
+
+/* Bitmask of physically existing CPUs */
+physid_mask_t phys_cpu_present_map;
+
+u8 bios_cpu_apicid[NR_CPUS] = { [0 ... NR_CPUS-1] = BAD_APICID };
+
+/*
+ * Intel MP BIOS table parsing routines:
+ */
+
+
+/*
+ * Checksum an MP configuration block.
+ */
+
+static int __init mpf_checksum(unsigned char *mp, int len)
+{
+	int sum = 0;
+
+	while (len--)
+		sum += *mp++;
+
+	return sum & 0xFF;
+}
+
+/*
+ * Have to match translation table entries to main table entries by counter
+ * hence the mpc_record variable .... can't see a less disgusting way of
+ * doing this ....
+ */
+
+static int mpc_record; 
+static struct mpc_config_translation *translation_table[MAX_MPC_ENTRY] __initdata;
+
+#ifndef CONFIG_XEN
+static void __devinit MP_processor_info (struct mpc_config_processor *m)
+{
+ 	int ver, apicid;
+	physid_mask_t phys_cpu;
+ 	
+	if (!(m->mpc_cpuflag & CPU_ENABLED))
+		return;
+
+	apicid = mpc_apic_id(m, translation_table[mpc_record]);
+
+	if (m->mpc_featureflag&(1<<0))
+		Dprintk("    Floating point unit present.\n");
+	if (m->mpc_featureflag&(1<<7))
+		Dprintk("    Machine Exception supported.\n");
+	if (m->mpc_featureflag&(1<<8))
+		Dprintk("    64 bit compare & exchange supported.\n");
+	if (m->mpc_featureflag&(1<<9))
+		Dprintk("    Internal APIC present.\n");
+	if (m->mpc_featureflag&(1<<11))
+		Dprintk("    SEP present.\n");
+	if (m->mpc_featureflag&(1<<12))
+		Dprintk("    MTRR  present.\n");
+	if (m->mpc_featureflag&(1<<13))
+		Dprintk("    PGE  present.\n");
+	if (m->mpc_featureflag&(1<<14))
+		Dprintk("    MCA  present.\n");
+	if (m->mpc_featureflag&(1<<15))
+		Dprintk("    CMOV  present.\n");
+	if (m->mpc_featureflag&(1<<16))
+		Dprintk("    PAT  present.\n");
+	if (m->mpc_featureflag&(1<<17))
+		Dprintk("    PSE  present.\n");
+	if (m->mpc_featureflag&(1<<18))
+		Dprintk("    PSN  present.\n");
+	if (m->mpc_featureflag&(1<<19))
+		Dprintk("    Cache Line Flush Instruction present.\n");
+	/* 20 Reserved */
+	if (m->mpc_featureflag&(1<<21))
+		Dprintk("    Debug Trace and EMON Store present.\n");
+	if (m->mpc_featureflag&(1<<22))
+		Dprintk("    ACPI Thermal Throttle Registers  present.\n");
+	if (m->mpc_featureflag&(1<<23))
+		Dprintk("    MMX  present.\n");
+	if (m->mpc_featureflag&(1<<24))
+		Dprintk("    FXSR  present.\n");
+	if (m->mpc_featureflag&(1<<25))
+		Dprintk("    XMM  present.\n");
+	if (m->mpc_featureflag&(1<<26))
+		Dprintk("    Willamette New Instructions  present.\n");
+	if (m->mpc_featureflag&(1<<27))
+		Dprintk("    Self Snoop  present.\n");
+	if (m->mpc_featureflag&(1<<28))
+		Dprintk("    HT  present.\n");
+	if (m->mpc_featureflag&(1<<29))
+		Dprintk("    Thermal Monitor present.\n");
+	/* 30, 31 Reserved */
+
+
+	if (m->mpc_cpuflag & CPU_BOOTPROCESSOR) {
+		Dprintk("    Bootup CPU\n");
+		boot_cpu_physical_apicid = m->mpc_apicid;
+	}
+
+	ver = m->mpc_apicver;
+
+	/*
+	 * Validate version
+	 */
+	if (ver == 0x0) {
+		printk(KERN_WARNING "BIOS bug, APIC version is 0 for CPU#%d! "
+				"fixing up to 0x10. (tell your hw vendor)\n",
+				m->mpc_apicid);
+		ver = 0x10;
+	}
+	apic_version[m->mpc_apicid] = ver;
+
+	phys_cpu = apicid_to_cpu_present(apicid);
+	physids_or(phys_cpu_present_map, phys_cpu_present_map, phys_cpu);
+
+	if (num_processors >= NR_CPUS) {
+		printk(KERN_WARNING "WARNING: NR_CPUS limit of %i reached."
+			"  Processor ignored.\n", NR_CPUS);
+		return;
+	}
+
+	if (num_processors >= maxcpus) {
+		printk(KERN_WARNING "WARNING: maxcpus limit of %i reached."
+			" Processor ignored.\n", maxcpus);
+		return;
+	}
+
+	cpu_set(num_processors, cpu_possible_map);
+	num_processors++;
+
+	/*
+	 * Would be preferable to switch to bigsmp when CONFIG_HOTPLUG_CPU=y
+	 * but we need to work other dependencies like SMP_SUSPEND etc
+	 * before this can be done without some confusion.
+	 * if (CPU_HOTPLUG_ENABLED || num_processors > 8)
+	 *       - Ashok Raj <ashok.raj@intel.com>
+	 */
+	if (num_processors > 8) {
+		switch (boot_cpu_data.x86_vendor) {
+		case X86_VENDOR_INTEL:
+			if (!APIC_XAPIC(ver)) {
+				def_to_bigsmp = 0;
+				break;
+			}
+			/* If P4 and above fall through */
+		case X86_VENDOR_AMD:
+			def_to_bigsmp = 1;
+		}
+	}
+	bios_cpu_apicid[num_processors - 1] = m->mpc_apicid;
+}
+#else
+void __init MP_processor_info (struct mpc_config_processor *m)
+{
+	num_processors++;
+}
+#endif /* CONFIG_XEN */
+
+static void __init MP_bus_info (struct mpc_config_bus *m)
+{
+	char str[7];
+
+	memcpy(str, m->mpc_bustype, 6);
+	str[6] = 0;
+
+	mpc_oem_bus_info(m, str, translation_table[mpc_record]);
+
+	if (m->mpc_busid >= MAX_MP_BUSSES) {
+		printk(KERN_WARNING "MP table busid value (%d) for bustype %s "
+			" is too large, max. supported is %d\n",
+			m->mpc_busid, str, MAX_MP_BUSSES - 1);
+		return;
+	}
+
+	if (strncmp(str, BUSTYPE_ISA, sizeof(BUSTYPE_ISA)-1) == 0) {
+		mp_bus_id_to_type[m->mpc_busid] = MP_BUS_ISA;
+	} else if (strncmp(str, BUSTYPE_EISA, sizeof(BUSTYPE_EISA)-1) == 0) {
+		mp_bus_id_to_type[m->mpc_busid] = MP_BUS_EISA;
+	} else if (strncmp(str, BUSTYPE_PCI, sizeof(BUSTYPE_PCI)-1) == 0) {
+		mpc_oem_pci_bus(m, translation_table[mpc_record]);
+		mp_bus_id_to_type[m->mpc_busid] = MP_BUS_PCI;
+		mp_bus_id_to_pci_bus[m->mpc_busid] = mp_current_pci_id;
+		mp_current_pci_id++;
+	} else if (strncmp(str, BUSTYPE_MCA, sizeof(BUSTYPE_MCA)-1) == 0) {
+		mp_bus_id_to_type[m->mpc_busid] = MP_BUS_MCA;
+	} else if (strncmp(str, BUSTYPE_NEC98, sizeof(BUSTYPE_NEC98)-1) == 0) {
+		mp_bus_id_to_type[m->mpc_busid] = MP_BUS_NEC98;
+	} else {
+		printk(KERN_WARNING "Unknown bustype %s - ignoring\n", str);
+	}
+}
+
+static void __init MP_ioapic_info (struct mpc_config_ioapic *m)
+{
+	if (!(m->mpc_flags & MPC_APIC_USABLE))
+		return;
+
+	printk(KERN_INFO "I/O APIC #%d Version %d at 0x%lX.\n",
+		m->mpc_apicid, m->mpc_apicver, m->mpc_apicaddr);
+	if (nr_ioapics >= MAX_IO_APICS) {
+		printk(KERN_CRIT "Max # of I/O APICs (%d) exceeded (found %d).\n",
+			MAX_IO_APICS, nr_ioapics);
+		panic("Recompile kernel with bigger MAX_IO_APICS!.\n");
+	}
+	if (!m->mpc_apicaddr) {
+		printk(KERN_ERR "WARNING: bogus zero I/O APIC address"
+			" found in MP table, skipping!\n");
+		return;
+	}
+	mp_ioapics[nr_ioapics] = *m;
+	nr_ioapics++;
+}
+
+static void __init MP_intsrc_info (struct mpc_config_intsrc *m)
+{
+	mp_irqs [mp_irq_entries] = *m;
+	Dprintk("Int: type %d, pol %d, trig %d, bus %d,"
+		" IRQ %02x, APIC ID %x, APIC INT %02x\n",
+			m->mpc_irqtype, m->mpc_irqflag & 3,
+			(m->mpc_irqflag >> 2) & 3, m->mpc_srcbus,
+			m->mpc_srcbusirq, m->mpc_dstapic, m->mpc_dstirq);
+	if (++mp_irq_entries == MAX_IRQ_SOURCES)
+		panic("Max # of irq sources exceeded!!\n");
+}
+
+static void __init MP_lintsrc_info (struct mpc_config_lintsrc *m)
+{
+	Dprintk("Lint: type %d, pol %d, trig %d, bus %d,"
+		" IRQ %02x, APIC ID %x, APIC LINT %02x\n",
+			m->mpc_irqtype, m->mpc_irqflag & 3,
+			(m->mpc_irqflag >> 2) &3, m->mpc_srcbusid,
+			m->mpc_srcbusirq, m->mpc_destapic, m->mpc_destapiclint);
+	/*
+	 * Well it seems all SMP boards in existence
+	 * use ExtINT/LVT1 == LINT0 and
+	 * NMI/LVT2 == LINT1 - the following check
+	 * will show us if this assumptions is false.
+	 * Until then we do not have to add baggage.
+	 */
+	if ((m->mpc_irqtype == mp_ExtINT) &&
+		(m->mpc_destapiclint != 0))
+			BUG();
+	if ((m->mpc_irqtype == mp_NMI) &&
+		(m->mpc_destapiclint != 1))
+			BUG();
+}
+
+#ifdef CONFIG_X86_NUMAQ
+static void __init MP_translation_info (struct mpc_config_translation *m)
+{
+	printk(KERN_INFO "Translation: record %d, type %d, quad %d, global %d, local %d\n", mpc_record, m->trans_type, m->trans_quad, m->trans_global, m->trans_local);
+
+	if (mpc_record >= MAX_MPC_ENTRY) 
+		printk(KERN_ERR "MAX_MPC_ENTRY exceeded!\n");
+	else
+		translation_table[mpc_record] = m; /* stash this for later */
+	if (m->trans_quad < MAX_NUMNODES && !node_online(m->trans_quad))
+		node_set_online(m->trans_quad);
+}
+
+/*
+ * Read/parse the MPC oem tables
+ */
+
+static void __init smp_read_mpc_oem(struct mp_config_oemtable *oemtable, \
+	unsigned short oemsize)
+{
+	int count = sizeof (*oemtable); /* the header size */
+	unsigned char *oemptr = ((unsigned char *)oemtable)+count;
+	
+	mpc_record = 0;
+	printk(KERN_INFO "Found an OEM MPC table at %8p - parsing it ... \n", oemtable);
+	if (memcmp(oemtable->oem_signature,MPC_OEM_SIGNATURE,4))
+	{
+		printk(KERN_WARNING "SMP mpc oemtable: bad signature [%c%c%c%c]!\n",
+			oemtable->oem_signature[0],
+			oemtable->oem_signature[1],
+			oemtable->oem_signature[2],
+			oemtable->oem_signature[3]);
+		return;
+	}
+	if (mpf_checksum((unsigned char *)oemtable,oemtable->oem_length))
+	{
+		printk(KERN_WARNING "SMP oem mptable: checksum error!\n");
+		return;
+	}
+	while (count < oemtable->oem_length) {
+		switch (*oemptr) {
+			case MP_TRANSLATION:
+			{
+				struct mpc_config_translation *m=
+					(struct mpc_config_translation *)oemptr;
+				MP_translation_info(m);
+				oemptr += sizeof(*m);
+				count += sizeof(*m);
+				++mpc_record;
+				break;
+			}
+			default:
+			{
+				printk(KERN_WARNING "Unrecognised OEM table entry type! - %d\n", (int) *oemptr);
+				return;
+			}
+		}
+       }
+}
+
+static inline void mps_oem_check(struct mp_config_table *mpc, char *oem,
+		char *productid)
+{
+	if (strncmp(oem, "IBM NUMA", 8))
+		printk("Warning!  May not be a NUMA-Q system!\n");
+	if (mpc->mpc_oemptr)
+		smp_read_mpc_oem((struct mp_config_oemtable *) mpc->mpc_oemptr,
+				mpc->mpc_oemsize);
+}
+#endif	/* CONFIG_X86_NUMAQ */
+
+/*
+ * Read/parse the MPC
+ */
+
+static int __init smp_read_mpc(struct mp_config_table *mpc)
+{
+	char str[16];
+	char oem[10];
+	int count=sizeof(*mpc);
+	unsigned char *mpt=((unsigned char *)mpc)+count;
+
+	if (memcmp(mpc->mpc_signature,MPC_SIGNATURE,4)) {
+		printk(KERN_ERR "SMP mptable: bad signature [0x%x]!\n",
+			*(u32 *)mpc->mpc_signature);
+		return 0;
+	}
+	if (mpf_checksum((unsigned char *)mpc,mpc->mpc_length)) {
+		printk(KERN_ERR "SMP mptable: checksum error!\n");
+		return 0;
+	}
+	if (mpc->mpc_spec!=0x01 && mpc->mpc_spec!=0x04) {
+		printk(KERN_ERR "SMP mptable: bad table version (%d)!!\n",
+			mpc->mpc_spec);
+		return 0;
+	}
+	if (!mpc->mpc_lapic) {
+		printk(KERN_ERR "SMP mptable: null local APIC address!\n");
+		return 0;
+	}
+	memcpy(oem,mpc->mpc_oem,8);
+	oem[8]=0;
+	printk(KERN_INFO "OEM ID: %s ",oem);
+
+	memcpy(str,mpc->mpc_productid,12);
+	str[12]=0;
+	printk("Product ID: %s ",str);
+
+	mps_oem_check(mpc, oem, str);
+
+	printk("APIC at: 0x%lX\n",mpc->mpc_lapic);
+
+	/* 
+	 * Save the local APIC address (it might be non-default) -- but only
+	 * if we're not using ACPI.
+	 */
+	if (!acpi_lapic)
+		mp_lapic_addr = mpc->mpc_lapic;
+
+	/*
+	 *	Now process the configuration blocks.
+	 */
+	mpc_record = 0;
+	while (count < mpc->mpc_length) {
+		switch(*mpt) {
+			case MP_PROCESSOR:
+			{
+				struct mpc_config_processor *m=
+					(struct mpc_config_processor *)mpt;
+				/* ACPI may have already provided this data */
+				if (!acpi_lapic)
+					MP_processor_info(m);
+				mpt += sizeof(*m);
+				count += sizeof(*m);
+				break;
+			}
+			case MP_BUS:
+			{
+				struct mpc_config_bus *m=
+					(struct mpc_config_bus *)mpt;
+				MP_bus_info(m);
+				mpt += sizeof(*m);
+				count += sizeof(*m);
+				break;
+			}
+			case MP_IOAPIC:
+			{
+				struct mpc_config_ioapic *m=
+					(struct mpc_config_ioapic *)mpt;
+				MP_ioapic_info(m);
+				mpt+=sizeof(*m);
+				count+=sizeof(*m);
+				break;
+			}
+			case MP_INTSRC:
+			{
+				struct mpc_config_intsrc *m=
+					(struct mpc_config_intsrc *)mpt;
+
+				MP_intsrc_info(m);
+				mpt+=sizeof(*m);
+				count+=sizeof(*m);
+				break;
+			}
+			case MP_LINTSRC:
+			{
+				struct mpc_config_lintsrc *m=
+					(struct mpc_config_lintsrc *)mpt;
+				MP_lintsrc_info(m);
+				mpt+=sizeof(*m);
+				count+=sizeof(*m);
+				break;
+			}
+			default:
+			{
+				count = mpc->mpc_length;
+				break;
+			}
+		}
+		++mpc_record;
+	}
+	clustered_apic_check();
+	if (!num_processors)
+		printk(KERN_ERR "SMP mptable: no processors registered!\n");
+	return num_processors;
+}
+
+static int __init ELCR_trigger(unsigned int irq)
+{
+	unsigned int port;
+
+	port = 0x4d0 + (irq >> 3);
+	return (inb(port) >> (irq & 7)) & 1;
+}
+
+static void __init construct_default_ioirq_mptable(int mpc_default_type)
+{
+	struct mpc_config_intsrc intsrc;
+	int i;
+	int ELCR_fallback = 0;
+
+	intsrc.mpc_type = MP_INTSRC;
+	intsrc.mpc_irqflag = 0;			/* conforming */
+	intsrc.mpc_srcbus = 0;
+	intsrc.mpc_dstapic = mp_ioapics[0].mpc_apicid;
+
+	intsrc.mpc_irqtype = mp_INT;
+
+	/*
+	 *  If true, we have an ISA/PCI system with no IRQ entries
+	 *  in the MP table. To prevent the PCI interrupts from being set up
+	 *  incorrectly, we try to use the ELCR. The sanity check to see if
+	 *  there is good ELCR data is very simple - IRQ0, 1, 2 and 13 can
+	 *  never be level sensitive, so we simply see if the ELCR agrees.
+	 *  If it does, we assume it's valid.
+	 */
+	if (mpc_default_type == 5) {
+		printk(KERN_INFO "ISA/PCI bus type with no IRQ information... falling back to ELCR\n");
+
+		if (ELCR_trigger(0) || ELCR_trigger(1) || ELCR_trigger(2) || ELCR_trigger(13))
+			printk(KERN_WARNING "ELCR contains invalid data... not using ELCR\n");
+		else {
+			printk(KERN_INFO "Using ELCR to identify PCI interrupts\n");
+			ELCR_fallback = 1;
+		}
+	}
+
+	for (i = 0; i < 16; i++) {
+		switch (mpc_default_type) {
+		case 2:
+			if (i == 0 || i == 13)
+				continue;	/* IRQ0 & IRQ13 not connected */
+			/* fall through */
+		default:
+			if (i == 2)
+				continue;	/* IRQ2 is never connected */
+		}
+
+		if (ELCR_fallback) {
+			/*
+			 *  If the ELCR indicates a level-sensitive interrupt, we
+			 *  copy that information over to the MP table in the
+			 *  irqflag field (level sensitive, active high polarity).
+			 */
+			if (ELCR_trigger(i))
+				intsrc.mpc_irqflag = 13;
+			else
+				intsrc.mpc_irqflag = 0;
+		}
+
+		intsrc.mpc_srcbusirq = i;
+		intsrc.mpc_dstirq = i ? i : 2;		/* IRQ0 to INTIN2 */
+		MP_intsrc_info(&intsrc);
+	}
+
+	intsrc.mpc_irqtype = mp_ExtINT;
+	intsrc.mpc_srcbusirq = 0;
+	intsrc.mpc_dstirq = 0;				/* 8259A to INTIN0 */
+	MP_intsrc_info(&intsrc);
+}
+
+static inline void __init construct_default_ISA_mptable(int mpc_default_type)
+{
+	struct mpc_config_processor processor;
+	struct mpc_config_bus bus;
+	struct mpc_config_ioapic ioapic;
+	struct mpc_config_lintsrc lintsrc;
+	int linttypes[2] = { mp_ExtINT, mp_NMI };
+	int i;
+
+	/*
+	 * local APIC has default address
+	 */
+	mp_lapic_addr = APIC_DEFAULT_PHYS_BASE;
+
+	/*
+	 * 2 CPUs, numbered 0 & 1.
+	 */
+	processor.mpc_type = MP_PROCESSOR;
+	/* Either an integrated APIC or a discrete 82489DX. */
+	processor.mpc_apicver = mpc_default_type > 4 ? 0x10 : 0x01;
+	processor.mpc_cpuflag = CPU_ENABLED;
+	processor.mpc_cpufeature = (boot_cpu_data.x86 << 8) |
+				   (boot_cpu_data.x86_model << 4) |
+				   boot_cpu_data.x86_mask;
+	processor.mpc_featureflag = boot_cpu_data.x86_capability[0];
+	processor.mpc_reserved[0] = 0;
+	processor.mpc_reserved[1] = 0;
+	for (i = 0; i < 2; i++) {
+		processor.mpc_apicid = i;
+		MP_processor_info(&processor);
+	}
+
+	bus.mpc_type = MP_BUS;
+	bus.mpc_busid = 0;
+	switch (mpc_default_type) {
+		default:
+			printk("???\n");
+			printk(KERN_ERR "Unknown standard configuration %d\n",
+				mpc_default_type);
+			/* fall through */
+		case 1:
+		case 5:
+			memcpy(bus.mpc_bustype, "ISA   ", 6);
+			break;
+		case 2:
+		case 6:
+		case 3:
+			memcpy(bus.mpc_bustype, "EISA  ", 6);
+			break;
+		case 4:
+		case 7:
+			memcpy(bus.mpc_bustype, "MCA   ", 6);
+	}
+	MP_bus_info(&bus);
+	if (mpc_default_type > 4) {
+		bus.mpc_busid = 1;
+		memcpy(bus.mpc_bustype, "PCI   ", 6);
+		MP_bus_info(&bus);
+	}
+
+	ioapic.mpc_type = MP_IOAPIC;
+	ioapic.mpc_apicid = 2;
+	ioapic.mpc_apicver = mpc_default_type > 4 ? 0x10 : 0x01;
+	ioapic.mpc_flags = MPC_APIC_USABLE;
+	ioapic.mpc_apicaddr = 0xFEC00000;
+	MP_ioapic_info(&ioapic);
+
+	/*
+	 * We set up most of the low 16 IO-APIC pins according to MPS rules.
+	 */
+	construct_default_ioirq_mptable(mpc_default_type);
+
+	lintsrc.mpc_type = MP_LINTSRC;
+	lintsrc.mpc_irqflag = 0;		/* conforming */
+	lintsrc.mpc_srcbusid = 0;
+	lintsrc.mpc_srcbusirq = 0;
+	lintsrc.mpc_destapic = MP_APIC_ALL;
+	for (i = 0; i < 2; i++) {
+		lintsrc.mpc_irqtype = linttypes[i];
+		lintsrc.mpc_destapiclint = i;
+		MP_lintsrc_info(&lintsrc);
+	}
+}
+
+static struct intel_mp_floating *mpf_found;
+
+/*
+ * Scan the memory blocks for an SMP configuration block.
+ */
+void __init get_smp_config (void)
+{
+	struct intel_mp_floating *mpf = mpf_found;
+
+	/*
+	 * ACPI supports both logical (e.g. Hyper-Threading) and physical 
+	 * processors, where MPS only supports physical.
+	 */
+	if (acpi_lapic && acpi_ioapic) {
+		printk(KERN_INFO "Using ACPI (MADT) for SMP configuration information\n");
+		return;
+	}
+	else if (acpi_lapic)
+		printk(KERN_INFO "Using ACPI for processor (LAPIC) configuration information\n");
+
+	printk(KERN_INFO "Intel MultiProcessor Specification v1.%d\n", mpf->mpf_specification);
+	if (mpf->mpf_feature2 & (1<<7)) {
+		printk(KERN_INFO "    IMCR and PIC compatibility mode.\n");
+		pic_mode = 1;
+	} else {
+		printk(KERN_INFO "    Virtual Wire compatibility mode.\n");
+		pic_mode = 0;
+	}
+
+	/*
+	 * Now see if we need to read further.
+	 */
+	if (mpf->mpf_feature1 != 0) {
+
+		printk(KERN_INFO "Default MP configuration #%d\n", mpf->mpf_feature1);
+		construct_default_ISA_mptable(mpf->mpf_feature1);
+
+	} else if (mpf->mpf_physptr) {
+
+		/*
+		 * Read the physical hardware table.  Anything here will
+		 * override the defaults.
+		 */
+		if (!smp_read_mpc(isa_bus_to_virt(mpf->mpf_physptr))) {
+			smp_found_config = 0;
+			printk(KERN_ERR "BIOS bug, MP table errors detected!...\n");
+			printk(KERN_ERR "... disabling SMP support. (tell your hw vendor)\n");
+			return;
+		}
+		/*
+		 * If there are no explicit MP IRQ entries, then we are
+		 * broken.  We set up most of the low 16 IO-APIC pins to
+		 * ISA defaults and hope it will work.
+		 */
+		if (!mp_irq_entries) {
+			struct mpc_config_bus bus;
+
+			printk(KERN_ERR "BIOS bug, no explicit IRQ entries, using default mptable. (tell your hw vendor)\n");
+
+			bus.mpc_type = MP_BUS;
+			bus.mpc_busid = 0;
+			memcpy(bus.mpc_bustype, "ISA   ", 6);
+			MP_bus_info(&bus);
+
+			construct_default_ioirq_mptable(0);
+		}
+
+	} else
+		BUG();
+
+	printk(KERN_INFO "Processors: %d\n", num_processors);
+	/*
+	 * Only use the first configuration found.
+	 */
+}
+
+static int __init smp_scan_config (unsigned long base, unsigned long length)
+{
+	unsigned long *bp = isa_bus_to_virt(base);
+	struct intel_mp_floating *mpf;
+
+	Dprintk("Scan SMP from %p for %ld bytes.\n", bp,length);
+	if (sizeof(*mpf) != 16)
+		printk("Error: MPF size\n");
+
+	while (length > 0) {
+		mpf = (struct intel_mp_floating *)bp;
+		if ((*bp == SMP_MAGIC_IDENT) &&
+			(mpf->mpf_length == 1) &&
+			!mpf_checksum((unsigned char *)bp, 16) &&
+			((mpf->mpf_specification == 1)
+				|| (mpf->mpf_specification == 4)) ) {
+
+			smp_found_config = 1;
+#ifndef CONFIG_XEN
+			printk(KERN_INFO "found SMP MP-table at %08lx\n",
+						virt_to_phys(mpf));
+			reserve_bootmem(virt_to_phys(mpf), PAGE_SIZE);
+			if (mpf->mpf_physptr) {
+				/*
+				 * We cannot access to MPC table to compute
+				 * table size yet, as only few megabytes from
+				 * the bottom is mapped now.
+				 * PC-9800's MPC table places on the very last
+				 * of physical memory; so that simply reserving
+				 * PAGE_SIZE from mpg->mpf_physptr yields BUG()
+				 * in reserve_bootmem.
+				 */
+				unsigned long size = PAGE_SIZE;
+				unsigned long end = max_low_pfn * PAGE_SIZE;
+				if (mpf->mpf_physptr + size > end)
+					size = end - mpf->mpf_physptr;
+				reserve_bootmem(mpf->mpf_physptr, size);
+			}
+#else
+			printk(KERN_INFO "found SMP MP-table at %08lx\n",
+				((unsigned long)bp - (unsigned long)isa_bus_to_virt(base)) + base);
+#endif
+
+			mpf_found = mpf;
+			return 1;
+		}
+		bp += 4;
+		length -= 16;
+	}
+	return 0;
+}
+
+void __init find_smp_config (void)
+{
+#ifndef CONFIG_XEN
+	unsigned int address;
+#endif
+
+	/*
+	 * FIXME: Linux assumes you have 640K of base ram..
+	 * this continues the error...
+	 *
+	 * 1) Scan the bottom 1K for a signature
+	 * 2) Scan the top 1K of base RAM
+	 * 3) Scan the 64K of bios
+	 */
+	if (smp_scan_config(0x0,0x400) ||
+		smp_scan_config(639*0x400,0x400) ||
+			smp_scan_config(0xF0000,0x10000))
+		return;
+	/*
+	 * If it is an SMP machine we should know now, unless the
+	 * configuration is in an EISA/MCA bus machine with an
+	 * extended bios data area.
+	 *
+	 * there is a real-mode segmented pointer pointing to the
+	 * 4K EBDA area at 0x40E, calculate and scan it here.
+	 *
+	 * NOTE! There are Linux loaders that will corrupt the EBDA
+	 * area, and as such this kind of SMP config may be less
+	 * trustworthy, simply because the SMP table may have been
+	 * stomped on during early boot. These loaders are buggy and
+	 * should be fixed.
+	 *
+	 * MP1.4 SPEC states to only scan first 1K of 4K EBDA.
+	 */
+
+#ifndef CONFIG_XEN
+	address = get_bios_ebda();
+	if (address)
+		smp_scan_config(address, 0x400);
+#endif
+}
+
+int es7000_plat;
+
+/* --------------------------------------------------------------------------
+                            ACPI-based MP Configuration
+   -------------------------------------------------------------------------- */
+
+#ifdef CONFIG_ACPI
+
+void __init mp_register_lapic_address (
+	u64			address)
+{
+#ifndef CONFIG_XEN
+	mp_lapic_addr = (unsigned long) address;
+
+	set_fixmap_nocache(FIX_APIC_BASE, mp_lapic_addr);
+
+	if (boot_cpu_physical_apicid == -1U)
+		boot_cpu_physical_apicid = GET_APIC_ID(apic_read(APIC_ID));
+
+	Dprintk("Boot CPU = %d\n", boot_cpu_physical_apicid);
+#endif
+}
+
+
+void __devinit mp_register_lapic (
+	u8			id, 
+	u8			enabled)
+{
+	struct mpc_config_processor processor;
+	int			boot_cpu = 0;
+	
+	if (MAX_APICS - id <= 0) {
+		printk(KERN_WARNING "Processor #%d invalid (max %d)\n",
+			id, MAX_APICS);
+		return;
+	}
+
+	if (id == boot_cpu_physical_apicid)
+		boot_cpu = 1;
+
+#ifndef CONFIG_XEN
+	processor.mpc_type = MP_PROCESSOR;
+	processor.mpc_apicid = id;
+	processor.mpc_apicver = GET_APIC_VERSION(apic_read(APIC_LVR));
+	processor.mpc_cpuflag = (enabled ? CPU_ENABLED : 0);
+	processor.mpc_cpuflag |= (boot_cpu ? CPU_BOOTPROCESSOR : 0);
+	processor.mpc_cpufeature = (boot_cpu_data.x86 << 8) | 
+		(boot_cpu_data.x86_model << 4) | boot_cpu_data.x86_mask;
+	processor.mpc_featureflag = boot_cpu_data.x86_capability[0];
+	processor.mpc_reserved[0] = 0;
+	processor.mpc_reserved[1] = 0;
+#endif
+
+	MP_processor_info(&processor);
+}
+
+#ifdef	CONFIG_X86_IO_APIC
+
+#define MP_ISA_BUS		0
+#define MP_MAX_IOAPIC_PIN	127
+
+static struct mp_ioapic_routing {
+	int			apic_id;
+	int			gsi_base;
+	int			gsi_end;
+	u32			pin_programmed[4];
+} mp_ioapic_routing[MAX_IO_APICS];
+
+
+static int mp_find_ioapic (
+	int			gsi)
+{
+	int			i = 0;
+
+	/* Find the IOAPIC that manages this GSI. */
+	for (i = 0; i < nr_ioapics; i++) {
+		if ((gsi >= mp_ioapic_routing[i].gsi_base)
+			&& (gsi <= mp_ioapic_routing[i].gsi_end))
+			return i;
+	}
+
+	printk(KERN_ERR "ERROR: Unable to locate IOAPIC for GSI %d\n", gsi);
+
+	return -1;
+}
+	
+
+void __init mp_register_ioapic (
+	u8			id, 
+	u32			address,
+	u32			gsi_base)
+{
+	int			idx = 0;
+	int			tmpid;
+
+	if (nr_ioapics >= MAX_IO_APICS) {
+		printk(KERN_ERR "ERROR: Max # of I/O APICs (%d) exceeded "
+			"(found %d)\n", MAX_IO_APICS, nr_ioapics);
+		panic("Recompile kernel with bigger MAX_IO_APICS!\n");
+	}
+	if (!address) {
+		printk(KERN_ERR "WARNING: Bogus (zero) I/O APIC address"
+			" found in MADT table, skipping!\n");
+		return;
+	}
+
+	idx = nr_ioapics++;
+
+	mp_ioapics[idx].mpc_type = MP_IOAPIC;
+	mp_ioapics[idx].mpc_flags = MPC_APIC_USABLE;
+	mp_ioapics[idx].mpc_apicaddr = address;
+
+#ifndef CONFIG_XEN
+	set_fixmap_nocache(FIX_IO_APIC_BASE_0 + idx, address);
+#endif
+	if ((boot_cpu_data.x86_vendor == X86_VENDOR_INTEL)
+		&& !APIC_XAPIC(apic_version[boot_cpu_physical_apicid]))
+		tmpid = io_apic_get_unique_id(idx, id);
+	else
+		tmpid = id;
+	if (tmpid == -1) {
+		nr_ioapics--;
+		return;
+	}
+	mp_ioapics[idx].mpc_apicid = tmpid;
+	mp_ioapics[idx].mpc_apicver = io_apic_get_version(idx);
+	
+	/* 
+	 * Build basic GSI lookup table to facilitate gsi->io_apic lookups
+	 * and to prevent reprogramming of IOAPIC pins (PCI GSIs).
+	 */
+	mp_ioapic_routing[idx].apic_id = mp_ioapics[idx].mpc_apicid;
+	mp_ioapic_routing[idx].gsi_base = gsi_base;
+	mp_ioapic_routing[idx].gsi_end = gsi_base + 
+		io_apic_get_redir_entries(idx);
+
+	printk("IOAPIC[%d]: apic_id %d, version %d, address 0x%lx, "
+		"GSI %d-%d\n", idx, mp_ioapics[idx].mpc_apicid, 
+		mp_ioapics[idx].mpc_apicver, mp_ioapics[idx].mpc_apicaddr,
+		mp_ioapic_routing[idx].gsi_base,
+		mp_ioapic_routing[idx].gsi_end);
+
+	return;
+}
+
+
+void __init mp_override_legacy_irq (
+	u8			bus_irq,
+	u8			polarity, 
+	u8			trigger, 
+	u32			gsi)
+{
+	struct mpc_config_intsrc intsrc;
+	int			ioapic = -1;
+	int			pin = -1;
+
+	/* 
+	 * Convert 'gsi' to 'ioapic.pin'.
+	 */
+	ioapic = mp_find_ioapic(gsi);
+	if (ioapic < 0)
+		return;
+	pin = gsi - mp_ioapic_routing[ioapic].gsi_base;
+
+	/*
+	 * TBD: This check is for faulty timer entries, where the override
+	 *      erroneously sets the trigger to level, resulting in a HUGE 
+	 *      increase of timer interrupts!
+	 */
+	if ((bus_irq == 0) && (trigger == 3))
+		trigger = 1;
+
+	intsrc.mpc_type = MP_INTSRC;
+	intsrc.mpc_irqtype = mp_INT;
+	intsrc.mpc_irqflag = (trigger << 2) | polarity;
+	intsrc.mpc_srcbus = MP_ISA_BUS;
+	intsrc.mpc_srcbusirq = bus_irq;				       /* IRQ */
+	intsrc.mpc_dstapic = mp_ioapics[ioapic].mpc_apicid;	   /* APIC ID */
+	intsrc.mpc_dstirq = pin;				    /* INTIN# */
+
+	Dprintk("Int: type %d, pol %d, trig %d, bus %d, irq %d, %d-%d\n",
+		intsrc.mpc_irqtype, intsrc.mpc_irqflag & 3, 
+		(intsrc.mpc_irqflag >> 2) & 3, intsrc.mpc_srcbus, 
+		intsrc.mpc_srcbusirq, intsrc.mpc_dstapic, intsrc.mpc_dstirq);
+
+	mp_irqs[mp_irq_entries] = intsrc;
+	if (++mp_irq_entries == MAX_IRQ_SOURCES)
+		panic("Max # of irq sources exceeded!\n");
+
+	return;
+}
+
+void __init mp_config_acpi_legacy_irqs (void)
+{
+	struct mpc_config_intsrc intsrc;
+	int			i = 0;
+	int			ioapic = -1;
+
+	/* 
+	 * Fabricate the legacy ISA bus (bus #31).
+	 */
+	mp_bus_id_to_type[MP_ISA_BUS] = MP_BUS_ISA;
+	Dprintk("Bus #%d is ISA\n", MP_ISA_BUS);
+
+	/*
+	 * Older generations of ES7000 have no legacy identity mappings
+	 */
+	if (es7000_plat == 1)
+		return;
+
+	/* 
+	 * Locate the IOAPIC that manages the ISA IRQs (0-15). 
+	 */
+	ioapic = mp_find_ioapic(0);
+	if (ioapic < 0)
+		return;
+
+	intsrc.mpc_type = MP_INTSRC;
+	intsrc.mpc_irqflag = 0;					/* Conforming */
+	intsrc.mpc_srcbus = MP_ISA_BUS;
+	intsrc.mpc_dstapic = mp_ioapics[ioapic].mpc_apicid;
+
+	/* 
+	 * Use the default configuration for the IRQs 0-15.  Unless
+	 * overriden by (MADT) interrupt source override entries.
+	 */
+	for (i = 0; i < 16; i++) {
+		int idx;
+
+		for (idx = 0; idx < mp_irq_entries; idx++) {
+			struct mpc_config_intsrc *irq = mp_irqs + idx;
+
+			/* Do we already have a mapping for this ISA IRQ? */
+			if (irq->mpc_srcbus == MP_ISA_BUS && irq->mpc_srcbusirq == i)
+				break;
+
+			/* Do we already have a mapping for this IOAPIC pin */
+			if ((irq->mpc_dstapic == intsrc.mpc_dstapic) &&
+				(irq->mpc_dstirq == i))
+				break;
+		}
+
+		if (idx != mp_irq_entries) {
+			printk(KERN_DEBUG "ACPI: IRQ%d used by override.\n", i);
+			continue;			/* IRQ already used */
+		}
+
+		intsrc.mpc_irqtype = mp_INT;
+		intsrc.mpc_srcbusirq = i;		   /* Identity mapped */
+		intsrc.mpc_dstirq = i;
+
+		Dprintk("Int: type %d, pol %d, trig %d, bus %d, irq %d, "
+			"%d-%d\n", intsrc.mpc_irqtype, intsrc.mpc_irqflag & 3, 
+			(intsrc.mpc_irqflag >> 2) & 3, intsrc.mpc_srcbus, 
+			intsrc.mpc_srcbusirq, intsrc.mpc_dstapic, 
+			intsrc.mpc_dstirq);
+
+		mp_irqs[mp_irq_entries] = intsrc;
+		if (++mp_irq_entries == MAX_IRQ_SOURCES)
+			panic("Max # of irq sources exceeded!\n");
+	}
+}
+
+#define MAX_GSI_NUM	4096
+
+int mp_register_gsi (u32 gsi, int triggering, int polarity)
+{
+	int			ioapic = -1;
+	int			ioapic_pin = 0;
+	int			idx, bit = 0;
+	static int		pci_irq = 16;
+	/*
+	 * Mapping between Global System Interrups, which
+	 * represent all possible interrupts, and IRQs
+	 * assigned to actual devices.
+	 */
+	static int		gsi_to_irq[MAX_GSI_NUM];
+
+	/* Don't set up the ACPI SCI because it's already set up */
+	if (acpi_fadt.sci_int == gsi)
+		return gsi;
+
+	ioapic = mp_find_ioapic(gsi);
+	if (ioapic < 0) {
+		printk(KERN_WARNING "No IOAPIC for GSI %u\n", gsi);
+		return gsi;
+	}
+
+	ioapic_pin = gsi - mp_ioapic_routing[ioapic].gsi_base;
+
+	if (ioapic_renumber_irq)
+		gsi = ioapic_renumber_irq(ioapic, gsi);
+
+	/* 
+	 * Avoid pin reprogramming.  PRTs typically include entries  
+	 * with redundant pin->gsi mappings (but unique PCI devices);
+	 * we only program the IOAPIC on the first.
+	 */
+	bit = ioapic_pin % 32;
+	idx = (ioapic_pin < 32) ? 0 : (ioapic_pin / 32);
+	if (idx > 3) {
+		printk(KERN_ERR "Invalid reference to IOAPIC pin "
+			"%d-%d\n", mp_ioapic_routing[ioapic].apic_id, 
+			ioapic_pin);
+		return gsi;
+	}
+	if ((1<<bit) & mp_ioapic_routing[ioapic].pin_programmed[idx]) {
+		Dprintk(KERN_DEBUG "Pin %d-%d already programmed\n",
+			mp_ioapic_routing[ioapic].apic_id, ioapic_pin);
+		return gsi_to_irq[gsi];
+	}
+
+	mp_ioapic_routing[ioapic].pin_programmed[idx] |= (1<<bit);
+
+	if (triggering == ACPI_LEVEL_SENSITIVE) {
+		/*
+		 * For PCI devices assign IRQs in order, avoiding gaps
+		 * due to unused I/O APIC pins.
+		 */
+		int irq = gsi;
+		if (gsi < MAX_GSI_NUM) {
+			/*
+			 * Retain the VIA chipset work-around (gsi > 15), but
+			 * avoid a problem where the 8254 timer (IRQ0) is setup
+			 * via an override (so it's not on pin 0 of the ioapic),
+			 * and at the same time, the pin 0 interrupt is a PCI
+			 * type.  The gsi > 15 test could cause these two pins
+			 * to be shared as IRQ0, and they are not shareable.
+			 * So test for this condition, and if necessary, avoid
+			 * the pin collision.
+			 */
+			if (gsi > 15 || (gsi == 0 && !timer_uses_ioapic_pin_0))
+				gsi = pci_irq++;
+			/*
+			 * Don't assign IRQ used by ACPI SCI
+			 */
+			if (gsi == acpi_fadt.sci_int)
+				gsi = pci_irq++;
+			gsi_to_irq[irq] = gsi;
+		} else {
+			printk(KERN_ERR "GSI %u is too high\n", gsi);
+			return gsi;
+		}
+	}
+
+	io_apic_set_pci_routing(ioapic, ioapic_pin, gsi,
+		    triggering == ACPI_EDGE_SENSITIVE ? 0 : 1,
+		    polarity == ACPI_ACTIVE_HIGH ? 0 : 1);
+	return gsi;
+}
+
+#endif /* CONFIG_X86_IO_APIC */
+#endif /* CONFIG_ACPI */
diff -rpuN linux-2.6.18.8/arch/i386/kernel/pci-dma-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/pci-dma-xen.c
--- linux-2.6.18.8/arch/i386/kernel/pci-dma-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/pci-dma-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,415 @@
+/*
+ * Dynamic DMA mapping support.
+ *
+ * On i386 there is no hardware dynamic DMA address translation,
+ * so consistent alloc/free are merely page allocation/freeing.
+ * The rest of the dynamic DMA mapping interface is implemented
+ * in asm/pci.h.
+ */
+
+#include <linux/types.h>
+#include <linux/mm.h>
+#include <linux/string.h>
+#include <linux/pci.h>
+#include <linux/module.h>
+#include <linux/version.h>
+#include <asm/io.h>
+#include <xen/balloon.h>
+#include <xen/gnttab.h>
+#include <asm/swiotlb.h>
+#include <asm/tlbflush.h>
+#include <asm-i386/mach-xen/asm/swiotlb.h>
+#include <asm-i386/mach-xen/asm/gnttab_dma.h>
+#include <asm/bug.h>
+
+#ifdef __x86_64__
+#include <asm/proto.h>
+
+int iommu_merge __read_mostly = 0;
+EXPORT_SYMBOL(iommu_merge);
+
+dma_addr_t bad_dma_address __read_mostly;
+EXPORT_SYMBOL(bad_dma_address);
+
+/* This tells the BIO block layer to assume merging. Default to off
+   because we cannot guarantee merging later. */
+int iommu_bio_merge __read_mostly = 0;
+EXPORT_SYMBOL(iommu_bio_merge);
+
+int force_iommu __read_mostly= 0;
+
+__init int iommu_setup(char *p)
+{
+    return 1;
+}
+
+void __init pci_iommu_alloc(void)
+{
+#ifdef CONFIG_SWIOTLB
+	pci_swiotlb_init();
+#endif
+}
+
+static int __init pci_iommu_init(void)
+{
+	no_iommu_init();
+	return 0;
+}
+
+/* Must execute after PCI subsystem */
+fs_initcall(pci_iommu_init);
+#endif
+
+struct dma_coherent_mem {
+	void		*virt_base;
+	u32		device_base;
+	int		size;
+	int		flags;
+	unsigned long	*bitmap;
+};
+
+#define IOMMU_BUG_ON(test)				\
+do {							\
+	if (unlikely(test)) {				\
+		printk(KERN_ALERT "Fatal DMA error! "	\
+		       "Please use 'swiotlb=force'\n");	\
+		BUG();					\
+	}						\
+} while (0)
+
+static int check_pages_physically_contiguous(unsigned long pfn, 
+					     unsigned int offset,
+					     size_t length)
+{
+	unsigned long next_mfn;
+	int i;
+	int nr_pages;
+	
+	next_mfn = pfn_to_mfn(pfn);
+	nr_pages = (offset + length + PAGE_SIZE-1) >> PAGE_SHIFT;
+	
+	for (i = 1; i < nr_pages; i++) {
+		if (pfn_to_mfn(++pfn) != ++next_mfn) 
+			return 0;
+	}
+	return 1;
+}
+
+int range_straddles_page_boundary(paddr_t p, size_t size)
+{
+	extern unsigned long *contiguous_bitmap;
+	unsigned long pfn = p >> PAGE_SHIFT;
+	unsigned int offset = p & ~PAGE_MASK;
+
+	if (offset + size <= PAGE_SIZE)
+		return 0;
+	if (test_bit(pfn, contiguous_bitmap))
+		return 0;
+	if (check_pages_physically_contiguous(pfn, offset, size))
+		return 0;
+	return 1;
+}
+
+int
+dma_map_sg(struct device *hwdev, struct scatterlist *sg, int nents,
+	   enum dma_data_direction direction)
+{
+	int i, rc;
+
+	if (direction == DMA_NONE)
+		BUG();
+	WARN_ON(nents == 0 || sg[0].length == 0);
+
+	if (swiotlb) {
+		rc = swiotlb_map_sg(hwdev, sg, nents, direction);
+	} else {
+		for (i = 0; i < nents; i++ ) {
+			BUG_ON(!sg[i].page);
+			sg[i].dma_address =
+				gnttab_dma_map_page(sg[i].page) + sg[i].offset;
+			sg[i].dma_length  = sg[i].length;
+			IOMMU_BUG_ON(address_needs_mapping(
+				hwdev, sg[i].dma_address));
+			IOMMU_BUG_ON(range_straddles_page_boundary(
+				page_to_pseudophys(sg[i].page) + sg[i].offset,
+				sg[i].length));
+		}
+		rc = nents;
+	}
+
+	flush_write_buffers();
+	return rc;
+}
+EXPORT_SYMBOL(dma_map_sg);
+
+void
+dma_unmap_sg(struct device *hwdev, struct scatterlist *sg, int nents,
+	     enum dma_data_direction direction)
+{
+	int i;
+
+	BUG_ON(direction == DMA_NONE);
+	if (swiotlb)
+		swiotlb_unmap_sg(hwdev, sg, nents, direction);
+	else {
+		for (i = 0; i < nents; i++ )
+			gnttab_dma_unmap_page(sg[i].dma_address);
+	}
+}
+EXPORT_SYMBOL(dma_unmap_sg);
+
+#ifdef CONFIG_HIGHMEM
+dma_addr_t
+dma_map_page(struct device *dev, struct page *page, unsigned long offset,
+	     size_t size, enum dma_data_direction direction)
+{
+	dma_addr_t dma_addr;
+
+	BUG_ON(direction == DMA_NONE);
+
+	if (swiotlb) {
+		dma_addr = swiotlb_map_page(
+			dev, page, offset, size, direction);
+	} else {
+		dma_addr = gnttab_dma_map_page(page) + offset;
+		IOMMU_BUG_ON(address_needs_mapping(dev, dma_addr));
+	}
+
+	return dma_addr;
+}
+EXPORT_SYMBOL(dma_map_page);
+
+void
+dma_unmap_page(struct device *dev, dma_addr_t dma_address, size_t size,
+	       enum dma_data_direction direction)
+{
+	BUG_ON(direction == DMA_NONE);
+	if (swiotlb)
+		swiotlb_unmap_page(dev, dma_address, size, direction);
+	else
+		gnttab_dma_unmap_page(dma_address);
+}
+EXPORT_SYMBOL(dma_unmap_page);
+#endif /* CONFIG_HIGHMEM */
+
+int
+dma_mapping_error(dma_addr_t dma_addr)
+{
+	if (swiotlb)
+		return swiotlb_dma_mapping_error(dma_addr);
+	return 0;
+}
+EXPORT_SYMBOL(dma_mapping_error);
+
+int
+dma_supported(struct device *dev, u64 mask)
+{
+	if (swiotlb)
+		return swiotlb_dma_supported(dev, mask);
+	/*
+	 * By default we'll BUG when an infeasible DMA is requested, and
+	 * request swiotlb=force (see IOMMU_BUG_ON).
+	 */
+	return 1;
+}
+EXPORT_SYMBOL(dma_supported);
+
+void *dma_alloc_coherent(struct device *dev, size_t size,
+			   dma_addr_t *dma_handle, gfp_t gfp)
+{
+	void *ret;
+	struct dma_coherent_mem *mem = dev ? dev->dma_mem : NULL;
+	unsigned int order = get_order(size);
+	unsigned long vstart;
+	u64 mask;
+
+	/* ignore region specifiers */
+	gfp &= ~(__GFP_DMA | __GFP_HIGHMEM);
+
+	if (mem) {
+		int page = bitmap_find_free_region(mem->bitmap, mem->size,
+						     order);
+		if (page >= 0) {
+			*dma_handle = mem->device_base + (page << PAGE_SHIFT);
+			ret = mem->virt_base + (page << PAGE_SHIFT);
+			memset(ret, 0, size);
+			return ret;
+		}
+		if (mem->flags & DMA_MEMORY_EXCLUSIVE)
+			return NULL;
+	}
+
+	if (dev == NULL || (dev->coherent_dma_mask < 0xffffffff))
+		gfp |= GFP_DMA;
+
+	vstart = __get_free_pages(gfp, order);
+	ret = (void *)vstart;
+
+	if (dev != NULL && dev->coherent_dma_mask)
+		mask = dev->coherent_dma_mask;
+	else
+		mask = 0xffffffff;
+
+	if (ret != NULL) {
+		if (xen_create_contiguous_region(vstart, order,
+						 fls64(mask)) != 0) {
+			free_pages(vstart, order);
+			return NULL;
+		}
+		memset(ret, 0, size);
+		*dma_handle = virt_to_bus(ret);
+	}
+	return ret;
+}
+EXPORT_SYMBOL(dma_alloc_coherent);
+
+void dma_free_coherent(struct device *dev, size_t size,
+			 void *vaddr, dma_addr_t dma_handle)
+{
+	struct dma_coherent_mem *mem = dev ? dev->dma_mem : NULL;
+	int order = get_order(size);
+	
+	if (mem && vaddr >= mem->virt_base && vaddr < (mem->virt_base + (mem->size << PAGE_SHIFT))) {
+		int page = (vaddr - mem->virt_base) >> PAGE_SHIFT;
+
+		bitmap_release_region(mem->bitmap, page, order);
+	} else {
+		xen_destroy_contiguous_region((unsigned long)vaddr, order);
+		free_pages((unsigned long)vaddr, order);
+	}
+}
+EXPORT_SYMBOL(dma_free_coherent);
+
+#ifdef ARCH_HAS_DMA_DECLARE_COHERENT_MEMORY
+int dma_declare_coherent_memory(struct device *dev, dma_addr_t bus_addr,
+				dma_addr_t device_addr, size_t size, int flags)
+{
+	void __iomem *mem_base;
+	int pages = size >> PAGE_SHIFT;
+	int bitmap_size = (pages + 31)/32;
+
+	if ((flags & (DMA_MEMORY_MAP | DMA_MEMORY_IO)) == 0)
+		goto out;
+	if (!size)
+		goto out;
+	if (dev->dma_mem)
+		goto out;
+
+	/* FIXME: this routine just ignores DMA_MEMORY_INCLUDES_CHILDREN */
+
+	mem_base = ioremap(bus_addr, size);
+	if (!mem_base)
+		goto out;
+
+	dev->dma_mem = kmalloc(sizeof(struct dma_coherent_mem), GFP_KERNEL);
+	if (!dev->dma_mem)
+		goto out;
+	memset(dev->dma_mem, 0, sizeof(struct dma_coherent_mem));
+	dev->dma_mem->bitmap = kmalloc(bitmap_size, GFP_KERNEL);
+	if (!dev->dma_mem->bitmap)
+		goto free1_out;
+	memset(dev->dma_mem->bitmap, 0, bitmap_size);
+
+	dev->dma_mem->virt_base = mem_base;
+	dev->dma_mem->device_base = device_addr;
+	dev->dma_mem->size = pages;
+	dev->dma_mem->flags = flags;
+
+	if (flags & DMA_MEMORY_MAP)
+		return DMA_MEMORY_MAP;
+
+	return DMA_MEMORY_IO;
+
+ free1_out:
+	kfree(dev->dma_mem->bitmap);
+ out:
+	return 0;
+}
+EXPORT_SYMBOL(dma_declare_coherent_memory);
+
+void dma_release_declared_memory(struct device *dev)
+{
+	struct dma_coherent_mem *mem = dev->dma_mem;
+	
+	if(!mem)
+		return;
+	dev->dma_mem = NULL;
+	iounmap(mem->virt_base);
+	kfree(mem->bitmap);
+	kfree(mem);
+}
+EXPORT_SYMBOL(dma_release_declared_memory);
+
+void *dma_mark_declared_memory_occupied(struct device *dev,
+					dma_addr_t device_addr, size_t size)
+{
+	struct dma_coherent_mem *mem = dev->dma_mem;
+	int pages = (size + (device_addr & ~PAGE_MASK) + PAGE_SIZE - 1) >> PAGE_SHIFT;
+	int pos, err;
+
+	if (!mem)
+		return ERR_PTR(-EINVAL);
+
+	pos = (device_addr - mem->device_base) >> PAGE_SHIFT;
+	err = bitmap_allocate_region(mem->bitmap, pos, get_order(pages));
+	if (err != 0)
+		return ERR_PTR(err);
+	return mem->virt_base + (pos << PAGE_SHIFT);
+}
+EXPORT_SYMBOL(dma_mark_declared_memory_occupied);
+#endif /* ARCH_HAS_DMA_DECLARE_COHERENT_MEMORY */
+
+dma_addr_t
+dma_map_single(struct device *dev, void *ptr, size_t size,
+	       enum dma_data_direction direction)
+{
+	dma_addr_t dma;
+
+	if (direction == DMA_NONE)
+		BUG();
+	WARN_ON(size == 0);
+
+	if (swiotlb) {
+		dma = swiotlb_map_single(dev, ptr, size, direction);
+	} else {
+		dma = gnttab_dma_map_page(virt_to_page(ptr)) +
+		      offset_in_page(ptr);
+		IOMMU_BUG_ON(range_straddles_page_boundary(__pa(ptr), size));
+		IOMMU_BUG_ON(address_needs_mapping(dev, dma));
+	}
+
+	flush_write_buffers();
+	return dma;
+}
+EXPORT_SYMBOL(dma_map_single);
+
+void
+dma_unmap_single(struct device *dev, dma_addr_t dma_addr, size_t size,
+		 enum dma_data_direction direction)
+{
+	if (direction == DMA_NONE)
+		BUG();
+	if (swiotlb)
+		swiotlb_unmap_single(dev, dma_addr, size, direction);
+	else
+		gnttab_dma_unmap_page(dma_addr);
+}
+EXPORT_SYMBOL(dma_unmap_single);
+
+void
+dma_sync_single_for_cpu(struct device *dev, dma_addr_t dma_handle, size_t size,
+			enum dma_data_direction direction)
+{
+	if (swiotlb)
+		swiotlb_sync_single_for_cpu(dev, dma_handle, size, direction);
+}
+EXPORT_SYMBOL(dma_sync_single_for_cpu);
+
+void
+dma_sync_single_for_device(struct device *dev, dma_addr_t dma_handle, size_t size,
+                           enum dma_data_direction direction)
+{
+	if (swiotlb)
+		swiotlb_sync_single_for_device(dev, dma_handle, size, direction);
+}
+EXPORT_SYMBOL(dma_sync_single_for_device);
diff -rpuN linux-2.6.18.8/arch/i386/kernel/process-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/process-xen.c
--- linux-2.6.18.8/arch/i386/kernel/process-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/process-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,877 @@
+/*
+ *  linux/arch/i386/kernel/process.c
+ *
+ *  Copyright (C) 1995  Linus Torvalds
+ *
+ *  Pentium III FXSR, SSE support
+ *	Gareth Hughes <gareth@valinux.com>, May 2000
+ */
+
+/*
+ * This file handles the architecture-dependent parts of process handling..
+ */
+
+#include <stdarg.h>
+
+#include <linux/cpu.h>
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/fs.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/elfcore.h>
+#include <linux/smp.h>
+#include <linux/smp_lock.h>
+#include <linux/stddef.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <linux/user.h>
+#include <linux/a.out.h>
+#include <linux/interrupt.h>
+#include <linux/utsname.h>
+#include <linux/delay.h>
+#include <linux/reboot.h>
+#include <linux/init.h>
+#include <linux/mc146818rtc.h>
+#include <linux/module.h>
+#include <linux/kallsyms.h>
+#include <linux/ptrace.h>
+#include <linux/random.h>
+
+#include <asm/uaccess.h>
+#include <asm/pgtable.h>
+#include <asm/system.h>
+#include <asm/io.h>
+#include <asm/ldt.h>
+#include <asm/processor.h>
+#include <asm/i387.h>
+#include <asm/desc.h>
+#include <asm/vm86.h>
+#ifdef CONFIG_MATH_EMULATION
+#include <asm/math_emu.h>
+#endif
+
+#include <xen/interface/physdev.h>
+#include <xen/interface/vcpu.h>
+#include <xen/cpu_hotplug.h>
+
+#include <linux/err.h>
+
+#include <asm/tlbflush.h>
+#include <asm/cpu.h>
+
+asmlinkage void ret_from_fork(void) __asm__("ret_from_fork");
+
+static int hlt_counter;
+
+unsigned long boot_option_idle_override = 0;
+EXPORT_SYMBOL(boot_option_idle_override);
+
+/*
+ * Return saved PC of a blocked thread.
+ */
+unsigned long thread_saved_pc(struct task_struct *tsk)
+{
+	return ((unsigned long *)tsk->thread.esp)[3];
+}
+
+/*
+ * Powermanagement idle function, if any..
+ */
+void (*pm_idle)(void);
+EXPORT_SYMBOL(pm_idle);
+static DEFINE_PER_CPU(unsigned int, cpu_idle_state);
+
+void disable_hlt(void)
+{
+	hlt_counter++;
+}
+
+EXPORT_SYMBOL(disable_hlt);
+
+void enable_hlt(void)
+{
+	hlt_counter--;
+}
+
+EXPORT_SYMBOL(enable_hlt);
+
+/*
+ * On SMP it's slightly faster (but much more power-consuming!)
+ * to poll the ->work.need_resched flag instead of waiting for the
+ * cross-CPU IPI to arrive. Use this option with caution.
+ */
+static void poll_idle (void)
+{
+	local_irq_enable();
+
+	asm volatile(
+		"2:"
+		"testl %0, %1;"
+		"rep; nop;"
+		"je 2b;"
+		: : "i"(_TIF_NEED_RESCHED), "m" (current_thread_info()->flags));
+}
+
+static void xen_idle(void)
+{
+	local_irq_disable();
+
+	if (need_resched())
+		local_irq_enable();
+	else {
+		current_thread_info()->status &= ~TS_POLLING;
+		smp_mb__after_clear_bit();
+		safe_halt();
+		current_thread_info()->status |= TS_POLLING;
+	}
+}
+#ifdef CONFIG_APM_MODULE
+EXPORT_SYMBOL(default_idle);
+#endif
+
+#ifdef CONFIG_HOTPLUG_CPU
+extern cpumask_t cpu_initialized;
+static inline void play_dead(void)
+{
+	idle_task_exit();
+	local_irq_disable();
+	cpu_clear(smp_processor_id(), cpu_initialized);
+	preempt_enable_no_resched();
+	VOID(HYPERVISOR_vcpu_op(VCPUOP_down, smp_processor_id(), NULL));
+	cpu_bringup();
+}
+#else
+static inline void play_dead(void)
+{
+	BUG();
+}
+#endif /* CONFIG_HOTPLUG_CPU */
+
+/*
+ * The idle thread. There's no useful work to be
+ * done, so just try to conserve power and have a
+ * low exit latency (ie sit in a loop waiting for
+ * somebody to say that they'd like to reschedule)
+ */
+void cpu_idle(void)
+{
+	int cpu = smp_processor_id();
+
+	current_thread_info()->status |= TS_POLLING;
+
+	/* endless idle loop with no priority at all */
+	while (1) {
+		while (!need_resched()) {
+			void (*idle)(void);
+
+			if (__get_cpu_var(cpu_idle_state))
+				__get_cpu_var(cpu_idle_state) = 0;
+
+			rmb();
+			idle = xen_idle; /* no alternatives */
+
+			if (cpu_is_offline(cpu))
+				play_dead();
+
+			__get_cpu_var(irq_stat).idle_timestamp = jiffies;
+			idle();
+		}
+		preempt_enable_no_resched();
+		schedule();
+		preempt_disable();
+	}
+}
+
+void cpu_idle_wait(void)
+{
+	unsigned int cpu, this_cpu = get_cpu();
+	cpumask_t map;
+
+	set_cpus_allowed(current, cpumask_of_cpu(this_cpu));
+	put_cpu();
+
+	cpus_clear(map);
+	for_each_online_cpu(cpu) {
+		per_cpu(cpu_idle_state, cpu) = 1;
+		cpu_set(cpu, map);
+	}
+
+	__get_cpu_var(cpu_idle_state) = 0;
+
+	wmb();
+	do {
+		ssleep(1);
+		for_each_online_cpu(cpu) {
+			if (cpu_isset(cpu, map) && !per_cpu(cpu_idle_state, cpu))
+				cpu_clear(cpu, map);
+		}
+		cpus_and(map, map, cpu_online_map);
+	} while (!cpus_empty(map));
+}
+EXPORT_SYMBOL_GPL(cpu_idle_wait);
+
+void __devinit select_idle_routine(const struct cpuinfo_x86 *c)
+{
+}
+
+static int __init idle_setup (char *str)
+{
+	if (!strncmp(str, "poll", 4)) {
+		printk("using polling idle threads.\n");
+		pm_idle = poll_idle;
+	}
+
+	boot_option_idle_override = 1;
+	return 1;
+}
+
+__setup("idle=", idle_setup);
+
+void show_regs(struct pt_regs * regs)
+{
+	unsigned long cr0 = 0L, cr2 = 0L, cr3 = 0L, cr4 = 0L;
+
+	printk("\n");
+	printk("Pid: %d, comm: %20s\n", current->pid, current->comm);
+	printk("EIP: %04x:[<%08lx>] CPU: %d\n",0xffff & regs->xcs,regs->eip, smp_processor_id());
+	print_symbol("EIP is at %s\n", regs->eip);
+
+	if (user_mode_vm(regs))
+		printk(" ESP: %04x:%08lx",0xffff & regs->xss,regs->esp);
+	printk(" EFLAGS: %08lx    %s  (%s %.*s)\n",
+	       regs->eflags, print_tainted(), system_utsname.release,
+	       (int)strcspn(system_utsname.version, " "),
+	       system_utsname.version);
+	printk("EAX: %08lx EBX: %08lx ECX: %08lx EDX: %08lx\n",
+		regs->eax,regs->ebx,regs->ecx,regs->edx);
+	printk("ESI: %08lx EDI: %08lx EBP: %08lx",
+		regs->esi, regs->edi, regs->ebp);
+	printk(" DS: %04x ES: %04x\n",
+		0xffff & regs->xds,0xffff & regs->xes);
+
+	cr0 = read_cr0();
+	cr2 = read_cr2();
+	cr3 = read_cr3();
+	cr4 = read_cr4_safe();
+	printk("CR0: %08lx CR2: %08lx CR3: %08lx CR4: %08lx\n", cr0, cr2, cr3, cr4);
+	show_trace(NULL, regs, &regs->esp);
+}
+
+/*
+ * This gets run with %ebx containing the
+ * function to call, and %edx containing
+ * the "args".
+ */
+extern void kernel_thread_helper(void);
+__asm__(".section .text\n"
+	".align 4\n"
+	"kernel_thread_helper:\n\t"
+	"movl %edx,%eax\n\t"
+	"pushl %edx\n\t"
+	"call *%ebx\n\t"
+	"pushl %eax\n\t"
+	"call do_exit\n"
+	".previous");
+
+/*
+ * Create a kernel thread
+ */
+int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
+{
+	struct pt_regs regs;
+
+	memset(&regs, 0, sizeof(regs));
+
+	regs.ebx = (unsigned long) fn;
+	regs.edx = (unsigned long) arg;
+
+	regs.xds = __USER_DS;
+	regs.xes = __USER_DS;
+	regs.orig_eax = -1;
+	regs.eip = (unsigned long) kernel_thread_helper;
+	regs.xcs = GET_KERNEL_CS();
+	regs.eflags = X86_EFLAGS_IF | X86_EFLAGS_SF | X86_EFLAGS_PF | 0x2;
+
+	/* Ok, create the new process.. */
+	return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, &regs, 0, NULL, NULL);
+}
+EXPORT_SYMBOL(kernel_thread);
+
+/*
+ * Free current thread data structures etc..
+ */
+void exit_thread(void)
+{
+	/* The process may have allocated an io port bitmap... nuke it. */
+	if (unlikely(test_thread_flag(TIF_IO_BITMAP))) {
+		struct task_struct *tsk = current;
+		struct thread_struct *t = &tsk->thread;
+		struct physdev_set_iobitmap set_iobitmap;
+		memset(&set_iobitmap, 0, sizeof(set_iobitmap));
+		WARN_ON(HYPERVISOR_physdev_op(PHYSDEVOP_set_iobitmap,
+					      &set_iobitmap));
+		kfree(t->io_bitmap_ptr);
+		t->io_bitmap_ptr = NULL;
+		clear_thread_flag(TIF_IO_BITMAP);
+	}
+}
+
+void flush_thread(void)
+{
+	struct task_struct *tsk = current;
+
+	memset(tsk->thread.debugreg, 0, sizeof(unsigned long)*8);
+	memset(tsk->thread.tls_array, 0, sizeof(tsk->thread.tls_array));	
+	clear_tsk_thread_flag(tsk, TIF_DEBUG);
+	/*
+	 * Forget coprocessor state..
+	 */
+	clear_fpu(tsk);
+	clear_used_math();
+}
+
+void release_thread(struct task_struct *dead_task)
+{
+	BUG_ON(dead_task->mm);
+	release_vm86_irqs(dead_task);
+}
+
+/*
+ * This gets called before we allocate a new thread and copy
+ * the current task into it.
+ */
+void prepare_to_copy(struct task_struct *tsk)
+{
+	unlazy_fpu(tsk);
+}
+
+int copy_thread(int nr, unsigned long clone_flags, unsigned long esp,
+	unsigned long unused,
+	struct task_struct * p, struct pt_regs * regs)
+{
+	struct pt_regs * childregs;
+	struct task_struct *tsk;
+	int err;
+
+	childregs = task_pt_regs(p);
+	*childregs = *regs;
+	childregs->eax = 0;
+	childregs->esp = esp;
+
+	p->thread.esp = (unsigned long) childregs;
+	p->thread.esp0 = (unsigned long) (childregs+1);
+
+	p->thread.eip = (unsigned long) ret_from_fork;
+
+	savesegment(fs,p->thread.fs);
+	savesegment(gs,p->thread.gs);
+
+	tsk = current;
+	if (unlikely(test_tsk_thread_flag(tsk, TIF_IO_BITMAP))) {
+		p->thread.io_bitmap_ptr = kmalloc(IO_BITMAP_BYTES, GFP_KERNEL);
+		if (!p->thread.io_bitmap_ptr) {
+			p->thread.io_bitmap_max = 0;
+			return -ENOMEM;
+		}
+		memcpy(p->thread.io_bitmap_ptr, tsk->thread.io_bitmap_ptr,
+			IO_BITMAP_BYTES);
+		set_tsk_thread_flag(p, TIF_IO_BITMAP);
+	}
+
+	/*
+	 * Set a new TLS for the child thread?
+	 */
+	if (clone_flags & CLONE_SETTLS) {
+		struct desc_struct *desc;
+		struct user_desc info;
+		int idx;
+
+		err = -EFAULT;
+		if (copy_from_user(&info, (void __user *)childregs->esi, sizeof(info)))
+			goto out;
+		err = -EINVAL;
+		if (LDT_empty(&info))
+			goto out;
+
+		idx = info.entry_number;
+		if (idx < GDT_ENTRY_TLS_MIN || idx > GDT_ENTRY_TLS_MAX)
+			goto out;
+
+		desc = p->thread.tls_array + idx - GDT_ENTRY_TLS_MIN;
+		desc->a = LDT_entry_a(&info);
+		desc->b = LDT_entry_b(&info);
+	}
+
+	p->thread.iopl = current->thread.iopl;
+
+	err = 0;
+ out:
+	if (err && p->thread.io_bitmap_ptr) {
+		kfree(p->thread.io_bitmap_ptr);
+		p->thread.io_bitmap_max = 0;
+	}
+	return err;
+}
+
+/*
+ * fill in the user structure for a core dump..
+ */
+void dump_thread(struct pt_regs * regs, struct user * dump)
+{
+	int i;
+
+/* changed the size calculations - should hopefully work better. lbt */
+	dump->magic = CMAGIC;
+	dump->start_code = 0;
+	dump->start_stack = regs->esp & ~(PAGE_SIZE - 1);
+	dump->u_tsize = ((unsigned long) current->mm->end_code) >> PAGE_SHIFT;
+	dump->u_dsize = ((unsigned long) (current->mm->brk + (PAGE_SIZE-1))) >> PAGE_SHIFT;
+	dump->u_dsize -= dump->u_tsize;
+	dump->u_ssize = 0;
+	for (i = 0; i < 8; i++)
+		dump->u_debugreg[i] = current->thread.debugreg[i];  
+
+	if (dump->start_stack < TASK_SIZE)
+		dump->u_ssize = ((unsigned long) (TASK_SIZE - dump->start_stack)) >> PAGE_SHIFT;
+
+	dump->regs.ebx = regs->ebx;
+	dump->regs.ecx = regs->ecx;
+	dump->regs.edx = regs->edx;
+	dump->regs.esi = regs->esi;
+	dump->regs.edi = regs->edi;
+	dump->regs.ebp = regs->ebp;
+	dump->regs.eax = regs->eax;
+	dump->regs.ds = regs->xds;
+	dump->regs.es = regs->xes;
+	savesegment(fs,dump->regs.fs);
+	savesegment(gs,dump->regs.gs);
+	dump->regs.orig_eax = regs->orig_eax;
+	dump->regs.eip = regs->eip;
+	dump->regs.cs = regs->xcs;
+	dump->regs.eflags = regs->eflags;
+	dump->regs.esp = regs->esp;
+	dump->regs.ss = regs->xss;
+
+	dump->u_fpvalid = dump_fpu (regs, &dump->i387);
+}
+EXPORT_SYMBOL(dump_thread);
+
+/* 
+ * Capture the user space registers if the task is not running (in user space)
+ */
+int dump_task_regs(struct task_struct *tsk, elf_gregset_t *regs)
+{
+	struct pt_regs ptregs = *task_pt_regs(tsk);
+	ptregs.xcs &= 0xffff;
+	ptregs.xds &= 0xffff;
+	ptregs.xes &= 0xffff;
+	ptregs.xss &= 0xffff;
+
+	elf_core_copy_regs(regs, &ptregs);
+
+	return 1;
+}
+
+static noinline void __switch_to_xtra(struct task_struct *next_p)
+{
+	struct thread_struct *next;
+
+	next = &next_p->thread;
+
+	if (test_tsk_thread_flag(next_p, TIF_DEBUG)) {
+		set_debugreg(next->debugreg[0], 0);
+		set_debugreg(next->debugreg[1], 1);
+		set_debugreg(next->debugreg[2], 2);
+		set_debugreg(next->debugreg[3], 3);
+		/* no 4 and 5 */
+		set_debugreg(next->debugreg[6], 6);
+		set_debugreg(next->debugreg[7], 7);
+	}
+}
+
+/*
+ * This function selects if the context switch from prev to next
+ * has to tweak the TSC disable bit in the cr4.
+ */
+static inline void disable_tsc(struct task_struct *prev_p,
+			       struct task_struct *next_p)
+{
+	struct thread_info *prev, *next;
+
+	/*
+	 * gcc should eliminate the ->thread_info dereference if
+	 * has_secure_computing returns 0 at compile time (SECCOMP=n).
+	 */
+	prev = task_thread_info(prev_p);
+	next = task_thread_info(next_p);
+
+	if (has_secure_computing(prev) || has_secure_computing(next)) {
+		/* slow path here */
+		if (has_secure_computing(prev) &&
+		    !has_secure_computing(next)) {
+			write_cr4(read_cr4() & ~X86_CR4_TSD);
+		} else if (!has_secure_computing(prev) &&
+			   has_secure_computing(next))
+			write_cr4(read_cr4() | X86_CR4_TSD);
+	}
+}
+
+/*
+ *	switch_to(x,yn) should switch tasks from x to y.
+ *
+ * We fsave/fwait so that an exception goes off at the right time
+ * (as a call from the fsave or fwait in effect) rather than to
+ * the wrong process. Lazy FP saving no longer makes any sense
+ * with modern CPU's, and this simplifies a lot of things (SMP
+ * and UP become the same).
+ *
+ * NOTE! We used to use the x86 hardware context switching. The
+ * reason for not using it any more becomes apparent when you
+ * try to recover gracefully from saved state that is no longer
+ * valid (stale segment register values in particular). With the
+ * hardware task-switch, there is no way to fix up bad state in
+ * a reasonable manner.
+ *
+ * The fact that Intel documents the hardware task-switching to
+ * be slow is a fairly red herring - this code is not noticeably
+ * faster. However, there _is_ some room for improvement here,
+ * so the performance issues may eventually be a valid point.
+ * More important, however, is the fact that this allows us much
+ * more flexibility.
+ *
+ * The return value (in %eax) will be the "prev" task after
+ * the task-switch, and shows up in ret_from_fork in entry.S,
+ * for example.
+ */
+struct task_struct fastcall * __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
+{
+	struct thread_struct *prev = &prev_p->thread,
+				 *next = &next_p->thread;
+	int cpu = smp_processor_id();
+#ifndef CONFIG_X86_NO_TSS
+	struct tss_struct *tss = &per_cpu(init_tss, cpu);
+#endif
+#if CONFIG_XEN_COMPAT > 0x030002
+	struct physdev_set_iopl iopl_op;
+	struct physdev_set_iobitmap iobmp_op;
+#else
+	struct physdev_op _pdo[2], *pdo = _pdo;
+#define iopl_op pdo->u.set_iopl
+#define iobmp_op pdo->u.set_iobitmap
+#endif
+	multicall_entry_t _mcl[8], *mcl = _mcl;
+
+	/* XEN NOTE: FS/GS saved in switch_mm(), not here. */
+
+	/*
+	 * This is basically '__unlazy_fpu', except that we queue a
+	 * multicall to indicate FPU task switch, rather than
+	 * synchronously trapping to Xen.
+	 */
+	if (prev_p->thread_info->status & TS_USEDFPU) {
+		__save_init_fpu(prev_p); /* _not_ save_init_fpu() */
+		mcl->op      = __HYPERVISOR_fpu_taskswitch;
+		mcl->args[0] = 1;
+		mcl++;
+	}
+#if 0 /* lazy fpu sanity check */
+	else BUG_ON(!(read_cr0() & 8));
+#endif
+
+	/*
+	 * Reload esp0.
+	 * This is load_esp0(tss, next) with a multicall.
+	 */
+	mcl->op      = __HYPERVISOR_stack_switch;
+	mcl->args[0] = __KERNEL_DS;
+	mcl->args[1] = next->esp0;
+	mcl++;
+
+	/*
+	 * Load the per-thread Thread-Local Storage descriptor.
+	 * This is load_TLS(next, cpu) with multicalls.
+	 */
+#define C(i) do {							\
+	if (unlikely(next->tls_array[i].a != prev->tls_array[i].a ||	\
+		     next->tls_array[i].b != prev->tls_array[i].b)) {	\
+		mcl->op = __HYPERVISOR_update_descriptor;		\
+		*(u64 *)&mcl->args[0] =	virt_to_machine(		\
+			&get_cpu_gdt_table(cpu)[GDT_ENTRY_TLS_MIN + i]);\
+		*(u64 *)&mcl->args[2] = *(u64 *)&next->tls_array[i];	\
+		mcl++;							\
+	}								\
+} while (0)
+	C(0); C(1); C(2);
+#undef C
+
+	if (unlikely(prev->iopl != next->iopl)) {
+		iopl_op.iopl = (next->iopl == 0) ? 1 : (next->iopl >> 12) & 3;
+#if CONFIG_XEN_COMPAT > 0x030002
+		mcl->op      = __HYPERVISOR_physdev_op;
+		mcl->args[0] = PHYSDEVOP_set_iopl;
+		mcl->args[1] = (unsigned long)&iopl_op;
+#else
+		mcl->op      = __HYPERVISOR_physdev_op_compat;
+		pdo->cmd     = PHYSDEVOP_set_iopl;
+		mcl->args[0] = (unsigned long)pdo++;
+#endif
+		mcl++;
+	}
+
+	if (unlikely(prev->io_bitmap_ptr || next->io_bitmap_ptr)) {
+		set_xen_guest_handle(iobmp_op.bitmap,
+				     (char *)next->io_bitmap_ptr);
+		iobmp_op.nr_ports = next->io_bitmap_ptr ? IO_BITMAP_BITS : 0;
+#if CONFIG_XEN_COMPAT > 0x030002
+		mcl->op      = __HYPERVISOR_physdev_op;
+		mcl->args[0] = PHYSDEVOP_set_iobitmap;
+		mcl->args[1] = (unsigned long)&iobmp_op;
+#else
+		mcl->op      = __HYPERVISOR_physdev_op_compat;
+		pdo->cmd     = PHYSDEVOP_set_iobitmap;
+		mcl->args[0] = (unsigned long)pdo++;
+#endif
+		mcl++;
+	}
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+	BUG_ON(pdo > _pdo + ARRAY_SIZE(_pdo));
+#endif
+	BUG_ON(mcl > _mcl + ARRAY_SIZE(_mcl));
+	if (unlikely(HYPERVISOR_multicall_check(_mcl, mcl - _mcl, NULL)))
+		BUG();
+
+	/*
+	 * Restore %fs and %gs if needed.
+	 *
+	 * Glibc normally makes %fs be zero, and %gs is one of
+	 * the TLS segments.
+	 */
+	if (unlikely(next->fs))
+		loadsegment(fs, next->fs);
+
+	if (next->gs)
+		loadsegment(gs, next->gs);
+
+	/*
+	 * Now maybe handle debug registers
+	 */
+	if (unlikely(task_thread_info(next_p)->flags & _TIF_WORK_CTXSW))
+		__switch_to_xtra(next_p);
+
+	disable_tsc(prev_p, next_p);
+
+	return prev_p;
+}
+
+asmlinkage int sys_fork(struct pt_regs regs)
+{
+	return do_fork(SIGCHLD, regs.esp, &regs, 0, NULL, NULL);
+}
+
+asmlinkage int sys_clone(struct pt_regs regs)
+{
+	unsigned long clone_flags;
+	unsigned long newsp;
+	int __user *parent_tidptr, *child_tidptr;
+
+	clone_flags = regs.ebx;
+	newsp = regs.ecx;
+	parent_tidptr = (int __user *)regs.edx;
+	child_tidptr = (int __user *)regs.edi;
+	if (!newsp)
+		newsp = regs.esp;
+	return do_fork(clone_flags, newsp, &regs, 0, parent_tidptr, child_tidptr);
+}
+
+/*
+ * This is trivial, and on the face of it looks like it
+ * could equally well be done in user mode.
+ *
+ * Not so, for quite unobvious reasons - register pressure.
+ * In user mode vfork() cannot have a stack frame, and if
+ * done by calling the "clone()" system call directly, you
+ * do not have enough call-clobbered registers to hold all
+ * the information you need.
+ */
+asmlinkage int sys_vfork(struct pt_regs regs)
+{
+	return do_fork(CLONE_VFORK | CLONE_VM | SIGCHLD, regs.esp, &regs, 0, NULL, NULL);
+}
+
+/*
+ * sys_execve() executes a new program.
+ */
+asmlinkage int sys_execve(struct pt_regs regs)
+{
+	int error;
+	char * filename;
+
+	filename = getname((char __user *) regs.ebx);
+	error = PTR_ERR(filename);
+	if (IS_ERR(filename))
+		goto out;
+	error = do_execve(filename,
+			(char __user * __user *) regs.ecx,
+			(char __user * __user *) regs.edx,
+			&regs);
+	if (error == 0) {
+		task_lock(current);
+		current->ptrace &= ~PT_DTRACE;
+		task_unlock(current);
+		/* Make sure we don't return using sysenter.. */
+		set_thread_flag(TIF_IRET);
+	}
+	putname(filename);
+out:
+	return error;
+}
+
+#define top_esp                (THREAD_SIZE - sizeof(unsigned long))
+#define top_ebp                (THREAD_SIZE - 2*sizeof(unsigned long))
+
+unsigned long get_wchan(struct task_struct *p)
+{
+	unsigned long ebp, esp, eip;
+	unsigned long stack_page;
+	int count = 0;
+	if (!p || p == current || p->state == TASK_RUNNING)
+		return 0;
+	stack_page = (unsigned long)task_stack_page(p);
+	esp = p->thread.esp;
+	if (!stack_page || esp < stack_page || esp > top_esp+stack_page)
+		return 0;
+	/* include/asm-i386/system.h:switch_to() pushes ebp last. */
+	ebp = *(unsigned long *) esp;
+	do {
+		if (ebp < stack_page || ebp > top_ebp+stack_page)
+			return 0;
+		eip = *(unsigned long *) (ebp+4);
+		if (!in_sched_functions(eip))
+			return eip;
+		ebp = *(unsigned long *) ebp;
+	} while (count++ < 16);
+	return 0;
+}
+
+/*
+ * sys_alloc_thread_area: get a yet unused TLS descriptor index.
+ */
+static int get_free_idx(void)
+{
+	struct thread_struct *t = &current->thread;
+	int idx;
+
+	for (idx = 0; idx < GDT_ENTRY_TLS_ENTRIES; idx++)
+		if (desc_empty(t->tls_array + idx))
+			return idx + GDT_ENTRY_TLS_MIN;
+	return -ESRCH;
+}
+
+/*
+ * Set a given TLS descriptor:
+ */
+asmlinkage int sys_set_thread_area(struct user_desc __user *u_info)
+{
+	struct thread_struct *t = &current->thread;
+	struct user_desc info;
+	struct desc_struct *desc;
+	int cpu, idx;
+
+	if (copy_from_user(&info, u_info, sizeof(info)))
+		return -EFAULT;
+	idx = info.entry_number;
+
+	/*
+	 * index -1 means the kernel should try to find and
+	 * allocate an empty descriptor:
+	 */
+	if (idx == -1) {
+		idx = get_free_idx();
+		if (idx < 0)
+			return idx;
+		if (put_user(idx, &u_info->entry_number))
+			return -EFAULT;
+	}
+
+	if (idx < GDT_ENTRY_TLS_MIN || idx > GDT_ENTRY_TLS_MAX)
+		return -EINVAL;
+
+	desc = t->tls_array + idx - GDT_ENTRY_TLS_MIN;
+
+	/*
+	 * We must not get preempted while modifying the TLS.
+	 */
+	cpu = get_cpu();
+
+	if (LDT_empty(&info)) {
+		desc->a = 0;
+		desc->b = 0;
+	} else {
+		desc->a = LDT_entry_a(&info);
+		desc->b = LDT_entry_b(&info);
+	}
+	load_TLS(t, cpu);
+
+	put_cpu();
+
+	return 0;
+}
+
+/*
+ * Get the current Thread-Local Storage area:
+ */
+
+#define GET_BASE(desc) ( \
+	(((desc)->a >> 16) & 0x0000ffff) | \
+	(((desc)->b << 16) & 0x00ff0000) | \
+	( (desc)->b        & 0xff000000)   )
+
+#define GET_LIMIT(desc) ( \
+	((desc)->a & 0x0ffff) | \
+	 ((desc)->b & 0xf0000) )
+	
+#define GET_32BIT(desc)		(((desc)->b >> 22) & 1)
+#define GET_CONTENTS(desc)	(((desc)->b >> 10) & 3)
+#define GET_WRITABLE(desc)	(((desc)->b >>  9) & 1)
+#define GET_LIMIT_PAGES(desc)	(((desc)->b >> 23) & 1)
+#define GET_PRESENT(desc)	(((desc)->b >> 15) & 1)
+#define GET_USEABLE(desc)	(((desc)->b >> 20) & 1)
+
+asmlinkage int sys_get_thread_area(struct user_desc __user *u_info)
+{
+	struct user_desc info;
+	struct desc_struct *desc;
+	int idx;
+
+	if (get_user(idx, &u_info->entry_number))
+		return -EFAULT;
+	if (idx < GDT_ENTRY_TLS_MIN || idx > GDT_ENTRY_TLS_MAX)
+		return -EINVAL;
+
+	memset(&info, 0, sizeof(info));
+
+	desc = current->thread.tls_array + idx - GDT_ENTRY_TLS_MIN;
+
+	info.entry_number = idx;
+	info.base_addr = GET_BASE(desc);
+	info.limit = GET_LIMIT(desc);
+	info.seg_32bit = GET_32BIT(desc);
+	info.contents = GET_CONTENTS(desc);
+	info.read_exec_only = !GET_WRITABLE(desc);
+	info.limit_in_pages = GET_LIMIT_PAGES(desc);
+	info.seg_not_present = !GET_PRESENT(desc);
+	info.useable = GET_USEABLE(desc);
+
+	if (copy_to_user(u_info, &info, sizeof(info)))
+		return -EFAULT;
+	return 0;
+}
+
+unsigned long arch_align_stack(unsigned long sp)
+{
+	if (randomize_va_space)
+		sp -= get_random_int() % 8192;
+	return sp & ~0xf;
+}
diff -rpuN linux-2.6.18.8/arch/i386/kernel/quirks-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/quirks-xen.c
--- linux-2.6.18.8/arch/i386/kernel/quirks-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/quirks-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,47 @@
+/*
+ * This file contains work-arounds for x86 and x86_64 platform bugs.
+ */
+#include <linux/pci.h>
+#include <linux/irq.h>
+
+#if defined(CONFIG_X86_IO_APIC) && (defined(CONFIG_SMP) || defined(CONFIG_XEN)) && defined(CONFIG_PCI)
+
+static void __devinit quirk_intel_irqbalance(struct pci_dev *dev)
+{
+	u8 config, rev;
+	u32 word;
+
+	/* BIOS may enable hardware IRQ balancing for
+	 * E7520/E7320/E7525(revision ID 0x9 and below)
+	 * based platforms.
+	 * Disable SW irqbalance/affinity on those platforms.
+	 */
+	pci_read_config_byte(dev, PCI_CLASS_REVISION, &rev);
+	if (rev > 0x9)
+		return;
+
+	printk(KERN_INFO "Intel E7520/7320/7525 detected.");
+
+	/* enable access to config space*/
+	pci_read_config_byte(dev, 0xf4, &config);
+	pci_write_config_byte(dev, 0xf4, config|0x2);
+
+	/* read xTPR register */
+	raw_pci_ops->read(0, 0, 0x40, 0x4c, 2, &word);
+
+	if (!(word & (1 << 13))) {
+		struct xen_platform_op op;
+		printk(KERN_INFO "Disabling irq balancing and affinity\n");
+		op.cmd = XENPF_platform_quirk;
+		op.u.platform_quirk.quirk_id = QUIRK_NOIRQBALANCING;
+		WARN_ON(HYPERVISOR_platform_op(&op));
+	}
+
+	/* put back the original value for config space*/
+	if (!(config & 0x2))
+		pci_write_config_byte(dev, 0xf4, config);
+}
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_INTEL,	PCI_DEVICE_ID_INTEL_E7320_MCH,	quirk_intel_irqbalance);
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_INTEL,	PCI_DEVICE_ID_INTEL_E7525_MCH,	quirk_intel_irqbalance);
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_INTEL,	PCI_DEVICE_ID_INTEL_E7520_MCH,	quirk_intel_irqbalance);
+#endif
diff -rpuN linux-2.6.18.8/arch/i386/kernel/relocate_kernel.S linux-2.6.18-xen-3.3.0/arch/i386/kernel/relocate_kernel.S
--- linux-2.6.18.8/arch/i386/kernel/relocate_kernel.S	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/relocate_kernel.S	2008-08-21 11:36:07.000000000 +0200
@@ -7,16 +7,138 @@
  */
 
 #include <linux/linkage.h>
+#include <asm/page.h>
+#include <asm/kexec.h>
+
+/*
+ * Must be relocatable PIC code callable as a C function
+ */
+
+#define PTR(x) (x << 2)
+#define PAGE_ALIGNED (1 << PAGE_SHIFT)
+#define PAGE_ATTR 0x63 /* _PAGE_PRESENT|_PAGE_RW|_PAGE_ACCESSED|_PAGE_DIRTY */
+#define PAE_PGD_ATTR 0x01 /* _PAGE_PRESENT */
+
+	.text
+	.align PAGE_ALIGNED
+	.globl relocate_kernel
+relocate_kernel:
+	movl	8(%esp), %ebp /* list of pages */
+
+#ifdef CONFIG_X86_PAE
+	/* map the control page at its virtual address */
+
+	movl	PTR(VA_PGD)(%ebp), %edi
+	movl	PTR(VA_CONTROL_PAGE)(%ebp), %eax
+	andl	$0xc0000000, %eax
+	shrl	$27, %eax
+	addl	%edi, %eax
+
+	movl	PTR(PA_PMD_0)(%ebp), %edx
+	orl	$PAE_PGD_ATTR, %edx
+	movl	%edx, (%eax)
+
+	movl	PTR(VA_PMD_0)(%ebp), %edi
+	movl	PTR(VA_CONTROL_PAGE)(%ebp), %eax
+	andl	$0x3fe00000, %eax
+	shrl	$18, %eax
+	addl	%edi, %eax
+
+	movl	PTR(PA_PTE_0)(%ebp), %edx
+	orl	$PAGE_ATTR, %edx
+	movl	%edx, (%eax)
+
+	movl	PTR(VA_PTE_0)(%ebp), %edi
+	movl	PTR(VA_CONTROL_PAGE)(%ebp), %eax
+	andl	$0x001ff000, %eax
+	shrl	$9, %eax
+	addl	%edi, %eax
+
+	movl	PTR(PA_CONTROL_PAGE)(%ebp), %edx
+	orl	$PAGE_ATTR, %edx
+	movl	%edx, (%eax)
+
+	/* identity map the control page at its physical address */
+
+	movl	PTR(VA_PGD)(%ebp), %edi
+	movl	PTR(PA_CONTROL_PAGE)(%ebp), %eax
+	andl	$0xc0000000, %eax
+	shrl	$27, %eax
+	addl	%edi, %eax
+
+	movl	PTR(PA_PMD_1)(%ebp), %edx
+	orl	$PAE_PGD_ATTR, %edx
+	movl	%edx, (%eax)
+
+	movl	PTR(VA_PMD_1)(%ebp), %edi
+	movl	PTR(PA_CONTROL_PAGE)(%ebp), %eax
+	andl	$0x3fe00000, %eax
+	shrl	$18, %eax
+	addl	%edi, %eax
+
+	movl	PTR(PA_PTE_1)(%ebp), %edx
+	orl	$PAGE_ATTR, %edx
+	movl	%edx, (%eax)
+
+	movl	PTR(VA_PTE_1)(%ebp), %edi
+	movl	PTR(PA_CONTROL_PAGE)(%ebp), %eax
+	andl	$0x001ff000, %eax
+	shrl	$9, %eax
+	addl	%edi, %eax
+
+	movl	PTR(PA_CONTROL_PAGE)(%ebp), %edx
+	orl	$PAGE_ATTR, %edx
+	movl	%edx, (%eax)
+#else
+	/* map the control page at its virtual address */
+
+	movl	PTR(VA_PGD)(%ebp), %edi
+	movl	PTR(VA_CONTROL_PAGE)(%ebp), %eax
+	andl	$0xffc00000, %eax
+	shrl	$20, %eax
+	addl	%edi, %eax
+
+	movl	PTR(PA_PTE_0)(%ebp), %edx
+	orl	$PAGE_ATTR, %edx
+	movl	%edx, (%eax)
+
+	movl	PTR(VA_PTE_0)(%ebp), %edi
+	movl	PTR(VA_CONTROL_PAGE)(%ebp), %eax
+	andl	$0x003ff000, %eax
+	shrl	$10, %eax
+	addl	%edi, %eax
+
+	movl	PTR(PA_CONTROL_PAGE)(%ebp), %edx
+	orl	$PAGE_ATTR, %edx
+	movl	%edx, (%eax)
+
+	/* identity map the control page at its physical address */
+
+	movl	PTR(VA_PGD)(%ebp), %edi
+	movl	PTR(PA_CONTROL_PAGE)(%ebp), %eax
+	andl	$0xffc00000, %eax
+	shrl	$20, %eax
+	addl	%edi, %eax
+
+	movl	PTR(PA_PTE_1)(%ebp), %edx
+	orl	$PAGE_ATTR, %edx
+	movl	%edx, (%eax)
+
+	movl	PTR(VA_PTE_1)(%ebp), %edi
+	movl	PTR(PA_CONTROL_PAGE)(%ebp), %eax
+	andl	$0x003ff000, %eax
+	shrl	$10, %eax
+	addl	%edi, %eax
+
+	movl	PTR(PA_CONTROL_PAGE)(%ebp), %edx
+	orl	$PAGE_ATTR, %edx
+	movl	%edx, (%eax)
+#endif
 
-	/*
-	 * Must be relocatable PIC code callable as a C function, that once
-	 * it starts can not use the previous processes stack.
-	 */
-	.globl relocate_new_kernel
 relocate_new_kernel:
 	/* read the arguments and say goodbye to the stack */
 	movl  4(%esp), %ebx /* page_list */
-	movl  8(%esp), %ebp /* reboot_code_buffer */
+	movl  8(%esp), %ebp /* list of pages */
 	movl  12(%esp), %edx /* start address */
 	movl  16(%esp), %ecx /* cpu_has_pae */
 
@@ -24,11 +146,57 @@ relocate_new_kernel:
 	pushl $0
 	popfl
 
-	/* set a new stack at the bottom of our page... */
-	lea   4096(%ebp), %esp
+	/* get physical address of control page now */
+	/* this is impossible after page table switch */
+	movl	PTR(PA_CONTROL_PAGE)(%ebp), %edi
 
-	/* store the parameters back on the stack */
-	pushl   %edx /* store the start address */
+	/* switch to new set of page tables */
+	movl	PTR(PA_PGD)(%ebp), %eax
+	movl	%eax, %cr3
+
+	/* setup idt */
+	movl	%edi, %eax
+	addl	$(idt_48 - relocate_kernel), %eax
+	lidtl	(%eax)
+
+	/* setup gdt */
+	movl	%edi, %eax
+	addl	$(gdt - relocate_kernel), %eax
+	movl	%edi, %esi
+	addl	$((gdt_48 - relocate_kernel) + 2), %esi
+	movl	%eax, (%esi)
+	
+	movl	%edi, %eax
+	addl	$(gdt_48 - relocate_kernel), %eax
+	lgdtl	(%eax)
+
+	/* setup data segment registers */
+	mov	$(gdt_ds - gdt), %eax
+	mov	%eax, %ds
+	mov	%eax, %es
+	mov	%eax, %fs
+	mov	%eax, %gs
+	mov	%eax, %ss
+	
+	/* setup a new stack at the end of the physical control page */
+	lea	4096(%edi), %esp
+
+	/* load new code segment and jump to identity mapped page */
+	movl	%edi, %esi
+	xorl	%eax, %eax
+	pushl	%eax
+	pushl	%esi
+	pushl	%eax
+	movl	$(gdt_cs - gdt), %eax
+	pushl	%eax	
+	movl	%edi, %eax
+	addl	$(identity_mapped - relocate_kernel),%eax
+	pushl	%eax
+	iretl
+
+identity_mapped:
+	/* store the start address on the stack */
+	pushl   %edx
 
 	/* Set cr0 to a known state:
 	 * 31 0 == Paging disabled
@@ -113,8 +281,20 @@ relocate_new_kernel:
 	xorl    %edi, %edi
 	xorl    %ebp, %ebp
 	ret
-relocate_new_kernel_end:
 
-	.globl relocate_new_kernel_size
-relocate_new_kernel_size:
-	.long relocate_new_kernel_end - relocate_new_kernel
+	.align	16
+gdt:
+	.quad	0x0000000000000000	/* NULL descriptor */
+gdt_cs:	
+	.quad	0x00cf9a000000ffff	/* kernel 4GB code at 0x00000000 */
+gdt_ds:
+	.quad	0x00cf92000000ffff	/* kernel 4GB data at 0x00000000 */
+gdt_end:
+	
+gdt_48:
+	.word	gdt_end - gdt - 1	/* limit */
+	.long	0			/* base - filled in by code above */
+
+idt_48:
+	.word	0			/* limit */
+	.long	0			/* base */
diff -rpuN linux-2.6.18.8/arch/i386/kernel/setup.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/setup.c
--- linux-2.6.18.8/arch/i386/kernel/setup.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/setup.c	2008-08-21 11:36:07.000000000 +0200
@@ -261,7 +261,12 @@ static struct resource standard_io_resou
 }, {
 	.name	= "keyboard",
 	.start	= 0x0060,
-	.end	= 0x006f,
+	.end	= 0x0060,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_IO
+}, {
+	.name	= "keyboard",
+	.start	= 0x0064,
+	.end	= 0x0064,
 	.flags	= IORESOURCE_BUSY | IORESOURCE_IO
 }, {
 	.name	= "dma page reg",
@@ -956,6 +961,28 @@ efi_memory_present_wrapper(unsigned long
 	return 0;
 }
 
+/*
+ * This function checks if any part of the range <start,end> is mapped
+ * with type.
+ */
+int
+e820_any_mapped(u64 start, u64 end, unsigned type)
+{
+	int i;
+
+	for (i = 0; i < e820.nr_map; i++) {
+		const struct e820entry *ei = &e820.map[i];
+
+		if (type && ei->type != type)
+			continue;
+		if (ei->addr >= end || ei->addr + ei->size <= start)
+			continue;
+		return 1;
+	}
+	return 0;
+}
+EXPORT_SYMBOL_GPL(e820_any_mapped);
+
  /*
   * This function checks if the entire range <start,end> is mapped with type.
   *
diff -rpuN linux-2.6.18.8/arch/i386/kernel/setup-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/setup-xen.c
--- linux-2.6.18.8/arch/i386/kernel/setup-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/setup-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1924 @@
+/*
+ *  linux/arch/i386/kernel/setup.c
+ *
+ *  Copyright (C) 1995  Linus Torvalds
+ *
+ *  Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999
+ *
+ *  Memory region support
+ *	David Parsons <orc@pell.chi.il.us>, July-August 1999
+ *
+ *  Added E820 sanitization routine (removes overlapping memory regions);
+ *  Brian Moyle <bmoyle@mvista.com>, February 2001
+ *
+ * Moved CPU detection code to cpu/${cpu}.c
+ *    Patrick Mochel <mochel@osdl.org>, March 2002
+ *
+ *  Provisions for empty E820 memory regions (reported by certain BIOSes).
+ *  Alex Achenbach <xela@slit.de>, December 2002.
+ *
+ */
+
+/*
+ * This file handles the architecture-dependent parts of initialization
+ */
+
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/mmzone.h>
+#include <linux/screen_info.h>
+#include <linux/ioport.h>
+#include <linux/acpi.h>
+#include <linux/apm_bios.h>
+#include <linux/initrd.h>
+#include <linux/bootmem.h>
+#include <linux/seq_file.h>
+#include <linux/platform_device.h>
+#include <linux/console.h>
+#include <linux/mca.h>
+#include <linux/root_dev.h>
+#include <linux/highmem.h>
+#include <linux/module.h>
+#include <linux/efi.h>
+#include <linux/init.h>
+#include <linux/edd.h>
+#include <linux/nodemask.h>
+#include <linux/kernel.h>
+#include <linux/percpu.h>
+#include <linux/notifier.h>
+#include <linux/kexec.h>
+#include <linux/crash_dump.h>
+#include <linux/dmi.h>
+#include <linux/pfn.h>
+
+#include <video/edid.h>
+
+#include <asm/apic.h>
+#include <asm/e820.h>
+#include <asm/mpspec.h>
+#include <asm/setup.h>
+#include <asm/arch_hooks.h>
+#include <asm/sections.h>
+#include <asm/io_apic.h>
+#include <asm/ist.h>
+#include <asm/io.h>
+#include <asm/hypervisor.h>
+#include <xen/interface/physdev.h>
+#include <xen/interface/memory.h>
+#include <xen/features.h>
+#include <xen/firmware.h>
+#include <xen/xencons.h>
+#include <setup_arch.h>
+#include <bios_ebda.h>
+
+#ifdef CONFIG_XEN
+#include <xen/interface/kexec.h>
+#endif
+
+/* Forward Declaration. */
+void __init find_max_pfn(void);
+
+static int xen_panic_event(struct notifier_block *, unsigned long, void *);
+static struct notifier_block xen_panic_block = {
+	xen_panic_event, NULL, 0 /* try to go last */
+};
+
+extern char hypercall_page[PAGE_SIZE];
+EXPORT_SYMBOL(hypercall_page);
+
+int disable_pse __devinitdata = 0;
+
+/*
+ * Machine setup..
+ */
+
+#ifdef CONFIG_EFI
+int efi_enabled = 0;
+EXPORT_SYMBOL(efi_enabled);
+#endif
+
+/* cpu data as detected by the assembly code in head.S */
+struct cpuinfo_x86 new_cpu_data __initdata = { 0, 0, 0, 0, -1, 1, 0, 0, -1 };
+/* common cpu data for all cpus */
+struct cpuinfo_x86 boot_cpu_data __read_mostly = { 0, 0, 0, 0, -1, 1, 0, 0, -1 };
+EXPORT_SYMBOL(boot_cpu_data);
+
+unsigned long mmu_cr4_features;
+
+#ifdef	CONFIG_ACPI
+	int acpi_disabled = 0;
+#else
+	int acpi_disabled = 1;
+#endif
+EXPORT_SYMBOL(acpi_disabled);
+
+#ifdef	CONFIG_ACPI
+int __initdata acpi_force = 0;
+extern acpi_interrupt_flags	acpi_sci_flags;
+#endif
+
+/* for MCA, but anyone else can use it if they want */
+unsigned int machine_id;
+#ifdef CONFIG_MCA
+EXPORT_SYMBOL(machine_id);
+#endif
+unsigned int machine_submodel_id;
+unsigned int BIOS_revision;
+unsigned int mca_pentium_flag;
+
+/* For PCI or other memory-mapped resources */
+unsigned long pci_mem_start = 0x10000000;
+#ifdef CONFIG_PCI
+EXPORT_SYMBOL(pci_mem_start);
+#endif
+
+/* Boot loader ID as an integer, for the benefit of proc_dointvec */
+int bootloader_type;
+
+/* user-defined highmem size */
+static unsigned int highmem_pages = -1;
+
+/*
+ * Setup options
+ */
+struct drive_info_struct { char dummy[32]; } drive_info;
+#if defined(CONFIG_BLK_DEV_IDE) || defined(CONFIG_BLK_DEV_HD) || \
+    defined(CONFIG_BLK_DEV_IDE_MODULE) || defined(CONFIG_BLK_DEV_HD_MODULE)
+EXPORT_SYMBOL(drive_info);
+#endif
+struct screen_info screen_info;
+EXPORT_SYMBOL(screen_info);
+struct apm_info apm_info;
+EXPORT_SYMBOL(apm_info);
+struct sys_desc_table_struct {
+	unsigned short length;
+	unsigned char table[0];
+};
+struct edid_info edid_info;
+EXPORT_SYMBOL_GPL(edid_info);
+#ifndef CONFIG_XEN
+#define copy_edid() (edid_info = EDID_INFO)
+#endif
+struct ist_info ist_info;
+#if defined(CONFIG_X86_SPEEDSTEP_SMI) || \
+	defined(CONFIG_X86_SPEEDSTEP_SMI_MODULE)
+EXPORT_SYMBOL(ist_info);
+#endif
+struct e820map e820;
+#ifdef CONFIG_XEN
+struct e820map machine_e820;
+#endif
+
+extern void early_cpu_init(void);
+extern void generic_apic_probe(char *);
+extern int root_mountflags;
+
+unsigned long saved_videomode;
+
+#define RAMDISK_IMAGE_START_MASK  	0x07FF
+#define RAMDISK_PROMPT_FLAG		0x8000
+#define RAMDISK_LOAD_FLAG		0x4000	
+
+static char command_line[COMMAND_LINE_SIZE];
+
+unsigned char __initdata boot_params[PARAM_SIZE];
+
+static struct resource data_resource = {
+	.name	= "Kernel data",
+	.start	= 0,
+	.end	= 0,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_MEM
+};
+
+static struct resource code_resource = {
+	.name	= "Kernel code",
+	.start	= 0,
+	.end	= 0,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_MEM
+};
+
+static struct resource system_rom_resource = {
+	.name	= "System ROM",
+	.start	= 0xf0000,
+	.end	= 0xfffff,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM
+};
+
+static struct resource extension_rom_resource = {
+	.name	= "Extension ROM",
+	.start	= 0xe0000,
+	.end	= 0xeffff,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM
+};
+
+static struct resource adapter_rom_resources[] = { {
+	.name 	= "Adapter ROM",
+	.start	= 0xc8000,
+	.end	= 0,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM
+}, {
+	.name 	= "Adapter ROM",
+	.start	= 0,
+	.end	= 0,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM
+}, {
+	.name 	= "Adapter ROM",
+	.start	= 0,
+	.end	= 0,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM
+}, {
+	.name 	= "Adapter ROM",
+	.start	= 0,
+	.end	= 0,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM
+}, {
+	.name 	= "Adapter ROM",
+	.start	= 0,
+	.end	= 0,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM
+}, {
+	.name 	= "Adapter ROM",
+	.start	= 0,
+	.end	= 0,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM
+} };
+
+#define ADAPTER_ROM_RESOURCES \
+	(sizeof adapter_rom_resources / sizeof adapter_rom_resources[0])
+
+static struct resource video_rom_resource = {
+	.name 	= "Video ROM",
+	.start	= 0xc0000,
+	.end	= 0xc7fff,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM
+};
+
+static struct resource video_ram_resource = {
+	.name	= "Video RAM area",
+	.start	= 0xa0000,
+	.end	= 0xbffff,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_MEM
+};
+
+static struct resource standard_io_resources[] = { {
+	.name	= "dma1",
+	.start	= 0x0000,
+	.end	= 0x001f,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_IO
+}, {
+	.name	= "pic1",
+	.start	= 0x0020,
+	.end	= 0x0021,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_IO
+}, {
+	.name   = "timer0",
+	.start	= 0x0040,
+	.end    = 0x0043,
+	.flags  = IORESOURCE_BUSY | IORESOURCE_IO
+}, {
+	.name   = "timer1",
+	.start  = 0x0050,
+	.end    = 0x0053,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_IO
+}, {
+	.name	= "keyboard",
+	.start	= 0x0060,
+	.end	= 0x0060,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_IO
+}, {
+	.name	= "keyboard",
+	.start	= 0x0064,
+	.end	= 0x0064,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_IO
+}, {
+	.name	= "dma page reg",
+	.start	= 0x0080,
+	.end	= 0x008f,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_IO
+}, {
+	.name	= "pic2",
+	.start	= 0x00a0,
+	.end	= 0x00a1,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_IO
+}, {
+	.name	= "dma2",
+	.start	= 0x00c0,
+	.end	= 0x00df,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_IO
+}, {
+	.name	= "fpu",
+	.start	= 0x00f0,
+	.end	= 0x00ff,
+	.flags	= IORESOURCE_BUSY | IORESOURCE_IO
+} };
+
+#define STANDARD_IO_RESOURCES \
+	(sizeof standard_io_resources / sizeof standard_io_resources[0])
+
+#define romsignature(x) (*(unsigned short *)(x) == 0xaa55)
+
+static int __init romchecksum(unsigned char *rom, unsigned long length)
+{
+	unsigned char *p, sum = 0;
+
+	for (p = rom; p < rom + length; p++)
+		sum += *p;
+	return sum == 0;
+}
+
+static void __init probe_roms(void)
+{
+	unsigned long start, length, upper;
+	unsigned char *rom;
+	int	      i;
+
+#ifdef CONFIG_XEN
+	/* Nothing to do if not running in dom0. */
+	if (!is_initial_xendomain())
+		return;
+#endif
+
+	/* video rom */
+	upper = adapter_rom_resources[0].start;
+	for (start = video_rom_resource.start; start < upper; start += 2048) {
+		rom = isa_bus_to_virt(start);
+		if (!romsignature(rom))
+			continue;
+
+		video_rom_resource.start = start;
+
+		/* 0 < length <= 0x7f * 512, historically */
+		length = rom[2] * 512;
+
+		/* if checksum okay, trust length byte */
+		if (length && romchecksum(rom, length))
+			video_rom_resource.end = start + length - 1;
+
+		request_resource(&iomem_resource, &video_rom_resource);
+		break;
+	}
+
+	start = (video_rom_resource.end + 1 + 2047) & ~2047UL;
+	if (start < upper)
+		start = upper;
+
+	/* system rom */
+	request_resource(&iomem_resource, &system_rom_resource);
+	upper = system_rom_resource.start;
+
+	/* check for extension rom (ignore length byte!) */
+	rom = isa_bus_to_virt(extension_rom_resource.start);
+	if (romsignature(rom)) {
+		length = extension_rom_resource.end - extension_rom_resource.start + 1;
+		if (romchecksum(rom, length)) {
+			request_resource(&iomem_resource, &extension_rom_resource);
+			upper = extension_rom_resource.start;
+		}
+	}
+
+	/* check for adapter roms on 2k boundaries */
+	for (i = 0; i < ADAPTER_ROM_RESOURCES && start < upper; start += 2048) {
+		rom = isa_bus_to_virt(start);
+		if (!romsignature(rom))
+			continue;
+
+		/* 0 < length <= 0x7f * 512, historically */
+		length = rom[2] * 512;
+
+		/* but accept any length that fits if checksum okay */
+		if (!length || start + length > upper || !romchecksum(rom, length))
+			continue;
+
+		adapter_rom_resources[i].start = start;
+		adapter_rom_resources[i].end = start + length - 1;
+		request_resource(&iomem_resource, &adapter_rom_resources[i]);
+
+		start = adapter_rom_resources[i++].end & ~2047UL;
+	}
+}
+
+/*
+ * Point at the empty zero page to start with. We map the real shared_info
+ * page as soon as fixmap is up and running.
+ */
+shared_info_t *HYPERVISOR_shared_info = (shared_info_t *)empty_zero_page;
+EXPORT_SYMBOL(HYPERVISOR_shared_info);
+
+unsigned long *phys_to_machine_mapping;
+unsigned long *pfn_to_mfn_frame_list_list, *pfn_to_mfn_frame_list[16];
+EXPORT_SYMBOL(phys_to_machine_mapping);
+
+/* Raw start-of-day parameters from the hypervisor. */
+start_info_t *xen_start_info;
+EXPORT_SYMBOL(xen_start_info);
+
+void __init add_memory_region(unsigned long long start,
+			      unsigned long long size, int type)
+{
+	int x;
+
+	if (!efi_enabled) {
+       		x = e820.nr_map;
+
+		if (x == E820MAX) {
+		    printk(KERN_ERR "Ooops! Too many entries in the memory map!\n");
+		    return;
+		}
+
+		e820.map[x].addr = start;
+		e820.map[x].size = size;
+		e820.map[x].type = type;
+		e820.nr_map++;
+	}
+} /* add_memory_region */
+
+static void __init limit_regions(unsigned long long size)
+{
+	unsigned long long current_addr = 0;
+	int i;
+
+	if (efi_enabled) {
+		efi_memory_desc_t *md;
+		void *p;
+
+		for (p = memmap.map, i = 0; p < memmap.map_end;
+			p += memmap.desc_size, i++) {
+			md = p;
+			current_addr = md->phys_addr + (md->num_pages << 12);
+			if (md->type == EFI_CONVENTIONAL_MEMORY) {
+				if (current_addr >= size) {
+					md->num_pages -=
+						(((current_addr-size) + PAGE_SIZE-1) >> PAGE_SHIFT);
+					memmap.nr_map = i + 1;
+					return;
+				}
+			}
+		}
+	}
+	for (i = 0; i < e820.nr_map; i++) {
+		current_addr = e820.map[i].addr + e820.map[i].size;
+		if (current_addr < size)
+			continue;
+
+		if (e820.map[i].type != E820_RAM)
+			continue;
+
+		if (e820.map[i].addr >= size) {
+			/*
+			 * This region starts past the end of the
+			 * requested size, skip it completely.
+			 */
+			e820.nr_map = i;
+		} else {
+			e820.nr_map = i + 1;
+			e820.map[i].size -= current_addr - size;
+		}
+		return;
+	}
+#ifdef CONFIG_XEN
+	if (i==e820.nr_map && current_addr < size) {
+		/*
+                 * The e820 map finished before our requested size so
+                 * extend the final entry to the requested address.
+                 */
+		--i;
+		if (e820.map[i].type == E820_RAM)
+			e820.map[i].size -= current_addr - size;
+		else
+			add_memory_region(current_addr, size - current_addr, E820_RAM);
+	}
+#endif
+}
+
+#define E820_DEBUG	1
+
+static void __init print_memory_map(char *who)
+{
+	int i;
+
+	for (i = 0; i < e820.nr_map; i++) {
+		printk(" %s: %016Lx - %016Lx ", who,
+			e820.map[i].addr,
+			e820.map[i].addr + e820.map[i].size);
+		switch (e820.map[i].type) {
+		case E820_RAM:	printk("(usable)\n");
+				break;
+		case E820_RESERVED:
+				printk("(reserved)\n");
+				break;
+		case E820_ACPI:
+				printk("(ACPI data)\n");
+				break;
+		case E820_NVS:
+				printk("(ACPI NVS)\n");
+				break;
+		default:	printk("type %lu\n", e820.map[i].type);
+				break;
+		}
+	}
+}
+
+/*
+ * Sanitize the BIOS e820 map.
+ *
+ * Some e820 responses include overlapping entries.  The following 
+ * replaces the original e820 map with a new one, removing overlaps.
+ *
+ */
+struct change_member {
+	struct e820entry *pbios; /* pointer to original bios entry */
+	unsigned long long addr; /* address for this change point */
+};
+static struct change_member change_point_list[2*E820MAX] __initdata;
+static struct change_member *change_point[2*E820MAX] __initdata;
+static struct e820entry *overlap_list[E820MAX] __initdata;
+static struct e820entry new_bios[E820MAX] __initdata;
+
+int __init sanitize_e820_map(struct e820entry * biosmap, char * pnr_map)
+{
+	struct change_member *change_tmp;
+	unsigned long current_type, last_type;
+	unsigned long long last_addr;
+	int chgidx, still_changing;
+	int overlap_entries;
+	int new_bios_entry;
+	int old_nr, new_nr, chg_nr;
+	int i;
+
+	/*
+		Visually we're performing the following (1,2,3,4 = memory types)...
+
+		Sample memory map (w/overlaps):
+		   ____22__________________
+		   ______________________4_
+		   ____1111________________
+		   _44_____________________
+		   11111111________________
+		   ____________________33__
+		   ___________44___________
+		   __________33333_________
+		   ______________22________
+		   ___________________2222_
+		   _________111111111______
+		   _____________________11_
+		   _________________4______
+
+		Sanitized equivalent (no overlap):
+		   1_______________________
+		   _44_____________________
+		   ___1____________________
+		   ____22__________________
+		   ______11________________
+		   _________1______________
+		   __________3_____________
+		   ___________44___________
+		   _____________33_________
+		   _______________2________
+		   ________________1_______
+		   _________________4______
+		   ___________________2____
+		   ____________________33__
+		   ______________________4_
+	*/
+
+	/* if there's only one memory region, don't bother */
+	if (*pnr_map < 2)
+		return -1;
+
+	old_nr = *pnr_map;
+
+	/* bail out if we find any unreasonable addresses in bios map */
+	for (i=0; i<old_nr; i++)
+		if (biosmap[i].addr + biosmap[i].size < biosmap[i].addr)
+			return -1;
+
+	/* create pointers for initial change-point information (for sorting) */
+	for (i=0; i < 2*old_nr; i++)
+		change_point[i] = &change_point_list[i];
+
+	/* record all known change-points (starting and ending addresses),
+	   omitting those that are for empty memory regions */
+	chgidx = 0;
+	for (i=0; i < old_nr; i++)	{
+		if (biosmap[i].size != 0) {
+			change_point[chgidx]->addr = biosmap[i].addr;
+			change_point[chgidx++]->pbios = &biosmap[i];
+			change_point[chgidx]->addr = biosmap[i].addr + biosmap[i].size;
+			change_point[chgidx++]->pbios = &biosmap[i];
+		}
+	}
+	chg_nr = chgidx;    	/* true number of change-points */
+
+	/* sort change-point list by memory addresses (low -> high) */
+	still_changing = 1;
+	while (still_changing)	{
+		still_changing = 0;
+		for (i=1; i < chg_nr; i++)  {
+			/* if <current_addr> > <last_addr>, swap */
+			/* or, if current=<start_addr> & last=<end_addr>, swap */
+			if ((change_point[i]->addr < change_point[i-1]->addr) ||
+				((change_point[i]->addr == change_point[i-1]->addr) &&
+				 (change_point[i]->addr == change_point[i]->pbios->addr) &&
+				 (change_point[i-1]->addr != change_point[i-1]->pbios->addr))
+			   )
+			{
+				change_tmp = change_point[i];
+				change_point[i] = change_point[i-1];
+				change_point[i-1] = change_tmp;
+				still_changing=1;
+			}
+		}
+	}
+
+	/* create a new bios memory map, removing overlaps */
+	overlap_entries=0;	 /* number of entries in the overlap table */
+	new_bios_entry=0;	 /* index for creating new bios map entries */
+	last_type = 0;		 /* start with undefined memory type */
+	last_addr = 0;		 /* start with 0 as last starting address */
+	/* loop through change-points, determining affect on the new bios map */
+	for (chgidx=0; chgidx < chg_nr; chgidx++)
+	{
+		/* keep track of all overlapping bios entries */
+		if (change_point[chgidx]->addr == change_point[chgidx]->pbios->addr)
+		{
+			/* add map entry to overlap list (> 1 entry implies an overlap) */
+			overlap_list[overlap_entries++]=change_point[chgidx]->pbios;
+		}
+		else
+		{
+			/* remove entry from list (order independent, so swap with last) */
+			for (i=0; i<overlap_entries; i++)
+			{
+				if (overlap_list[i] == change_point[chgidx]->pbios)
+					overlap_list[i] = overlap_list[overlap_entries-1];
+			}
+			overlap_entries--;
+		}
+		/* if there are overlapping entries, decide which "type" to use */
+		/* (larger value takes precedence -- 1=usable, 2,3,4,4+=unusable) */
+		current_type = 0;
+		for (i=0; i<overlap_entries; i++)
+			if (overlap_list[i]->type > current_type)
+				current_type = overlap_list[i]->type;
+		/* continue building up new bios map based on this information */
+		if (current_type != last_type)	{
+			if (last_type != 0)	 {
+				new_bios[new_bios_entry].size =
+					change_point[chgidx]->addr - last_addr;
+				/* move forward only if the new size was non-zero */
+				if (new_bios[new_bios_entry].size != 0)
+					if (++new_bios_entry >= E820MAX)
+						break; 	/* no more space left for new bios entries */
+			}
+			if (current_type != 0)	{
+				new_bios[new_bios_entry].addr = change_point[chgidx]->addr;
+				new_bios[new_bios_entry].type = current_type;
+				last_addr=change_point[chgidx]->addr;
+			}
+			last_type = current_type;
+		}
+	}
+	new_nr = new_bios_entry;   /* retain count for new bios entries */
+
+	/* copy new bios mapping into original location */
+	memcpy(biosmap, new_bios, new_nr*sizeof(struct e820entry));
+	*pnr_map = new_nr;
+
+	return 0;
+}
+
+/*
+ * Copy the BIOS e820 map into a safe place.
+ *
+ * Sanity-check it while we're at it..
+ *
+ * If we're lucky and live on a modern system, the setup code
+ * will have given us a memory map that we can use to properly
+ * set up memory.  If we aren't, we'll fake a memory map.
+ *
+ * We check to see that the memory map contains at least 2 elements
+ * before we'll use it, because the detection code in setup.S may
+ * not be perfect and most every PC known to man has two memory
+ * regions: one from 0 to 640k, and one from 1mb up.  (The IBM
+ * thinkpad 560x, for example, does not cooperate with the memory
+ * detection code.)
+ */
+int __init copy_e820_map(struct e820entry * biosmap, int nr_map)
+{
+#ifndef CONFIG_XEN
+	/* Only one memory region (or negative)? Ignore it */
+	if (nr_map < 2)
+		return -1;
+#else
+	BUG_ON(nr_map < 1);
+#endif
+
+	do {
+		unsigned long long start = biosmap->addr;
+		unsigned long long size = biosmap->size;
+		unsigned long long end = start + size;
+		unsigned long type = biosmap->type;
+
+		/* Overflow in 64 bits? Ignore the memory map. */
+		if (start > end)
+			return -1;
+
+#ifndef CONFIG_XEN
+		/*
+		 * Some BIOSes claim RAM in the 640k - 1M region.
+		 * Not right. Fix it up.
+		 */
+		if (type == E820_RAM) {
+			if (start < 0x100000ULL && end > 0xA0000ULL) {
+				if (start < 0xA0000ULL)
+					add_memory_region(start, 0xA0000ULL-start, type);
+				if (end <= 0x100000ULL)
+					continue;
+				start = 0x100000ULL;
+				size = end - start;
+			}
+		}
+#endif
+		add_memory_region(start, size, type);
+	} while (biosmap++,--nr_map);
+
+#ifdef CONFIG_XEN
+	if (is_initial_xendomain()) {
+		struct xen_memory_map memmap;
+
+		memmap.nr_entries = E820MAX;
+		set_xen_guest_handle(memmap.buffer, machine_e820.map);
+
+		if (HYPERVISOR_memory_op(XENMEM_machine_memory_map, &memmap))
+			BUG();
+		machine_e820.nr_map = memmap.nr_entries;
+	} else
+		machine_e820 = e820;
+#endif
+
+	return 0;
+}
+
+#if defined(CONFIG_EDD) || defined(CONFIG_EDD_MODULE)
+struct edd edd;
+#ifdef CONFIG_EDD_MODULE
+EXPORT_SYMBOL(edd);
+#endif
+#ifndef CONFIG_XEN
+/**
+ * copy_edd() - Copy the BIOS EDD information
+ *              from boot_params into a safe place.
+ *
+ */
+static inline void copy_edd(void)
+{
+     memcpy(edd.mbr_signature, EDD_MBR_SIGNATURE, sizeof(edd.mbr_signature));
+     memcpy(edd.edd_info, EDD_BUF, sizeof(edd.edd_info));
+     edd.mbr_signature_nr = EDD_MBR_SIG_NR;
+     edd.edd_info_nr = EDD_NR;
+}
+#endif
+#else
+static inline void copy_edd(void)
+{
+}
+#endif
+
+static void __init parse_cmdline_early (char ** cmdline_p)
+{
+	char c = ' ', *to = command_line, *from = saved_command_line;
+	int len = 0, max_cmdline;
+	int userdef = 0;
+
+	if ((max_cmdline = MAX_GUEST_CMDLINE) > COMMAND_LINE_SIZE)
+		max_cmdline = COMMAND_LINE_SIZE;
+	memcpy(saved_command_line, xen_start_info->cmd_line, max_cmdline);
+	/* Save unparsed command line copy for /proc/cmdline */
+	saved_command_line[max_cmdline-1] = '\0';
+
+	for (;;) {
+		if (c != ' ')
+			goto next_char;
+		/*
+		 * "mem=nopentium" disables the 4MB page tables.
+		 * "mem=XXX[kKmM]" defines a memory region from HIGH_MEM
+		 * to <mem>, overriding the bios size.
+		 * "memmap=XXX[KkmM]@XXX[KkmM]" defines a memory region from
+		 * <start> to <start>+<mem>, overriding the bios size.
+		 *
+		 * HPA tells me bootloaders need to parse mem=, so no new
+		 * option should be mem=  [also see Documentation/i386/boot.txt]
+		 */
+		if (!memcmp(from, "mem=", 4)) {
+			if (to != command_line)
+				to--;
+			if (!memcmp(from+4, "nopentium", 9)) {
+				from += 9+4;
+				clear_bit(X86_FEATURE_PSE, boot_cpu_data.x86_capability);
+				disable_pse = 1;
+			} else {
+				/* If the user specifies memory size, we
+				 * limit the BIOS-provided memory map to
+				 * that size. exactmap can be used to specify
+				 * the exact map. mem=number can be used to
+				 * trim the existing memory map.
+				 */
+				unsigned long long mem_size;
+ 
+				mem_size = memparse(from+4, &from);
+				limit_regions(mem_size);
+				userdef=1;
+			}
+		}
+
+		else if (!memcmp(from, "memmap=", 7)) {
+			if (to != command_line)
+				to--;
+			if (!memcmp(from+7, "exactmap", 8)) {
+#ifdef CONFIG_CRASH_DUMP
+				/* If we are doing a crash dump, we
+				 * still need to know the real mem
+				 * size before original memory map is
+				 * reset.
+				 */
+				find_max_pfn();
+				saved_max_pfn = max_pfn;
+#endif
+				from += 8+7;
+				e820.nr_map = 0;
+				userdef = 1;
+			} else {
+				/* If the user specifies memory size, we
+				 * limit the BIOS-provided memory map to
+				 * that size. exactmap can be used to specify
+				 * the exact map. mem=number can be used to
+				 * trim the existing memory map.
+				 */
+				unsigned long long start_at, mem_size;
+ 
+				mem_size = memparse(from+7, &from);
+				if (*from == '@') {
+					start_at = memparse(from+1, &from);
+					add_memory_region(start_at, mem_size, E820_RAM);
+				} else if (*from == '#') {
+					start_at = memparse(from+1, &from);
+					add_memory_region(start_at, mem_size, E820_ACPI);
+				} else if (*from == '$') {
+					start_at = memparse(from+1, &from);
+					add_memory_region(start_at, mem_size, E820_RESERVED);
+				} else {
+					limit_regions(mem_size);
+					userdef=1;
+				}
+			}
+		}
+
+		else if (!memcmp(from, "noexec=", 7))
+			noexec_setup(from + 7);
+
+
+#ifdef  CONFIG_X86_MPPARSE
+		/*
+		 * If the BIOS enumerates physical processors before logical,
+		 * maxcpus=N at enumeration-time can be used to disable HT.
+		 */
+		else if (!memcmp(from, "maxcpus=", 8)) {
+			extern unsigned int maxcpus;
+
+			maxcpus = simple_strtoul(from + 8, NULL, 0);
+		}
+#endif
+
+#ifdef CONFIG_ACPI
+		/* "acpi=off" disables both ACPI table parsing and interpreter */
+		else if (!memcmp(from, "acpi=off", 8)) {
+			disable_acpi();
+		}
+
+		/* acpi=force to over-ride black-list */
+		else if (!memcmp(from, "acpi=force", 10)) {
+			acpi_force = 1;
+			acpi_ht = 1;
+			acpi_disabled = 0;
+		}
+
+		/* acpi=strict disables out-of-spec workarounds */
+		else if (!memcmp(from, "acpi=strict", 11)) {
+			acpi_strict = 1;
+		}
+
+		/* Limit ACPI just to boot-time to enable HT */
+		else if (!memcmp(from, "acpi=ht", 7)) {
+			if (!acpi_force)
+				disable_acpi();
+			acpi_ht = 1;
+		}
+		
+		/* "pci=noacpi" disable ACPI IRQ routing and PCI scan */
+		else if (!memcmp(from, "pci=noacpi", 10)) {
+			acpi_disable_pci();
+		}
+		/* "acpi=noirq" disables ACPI interrupt routing */
+		else if (!memcmp(from, "acpi=noirq", 10)) {
+			acpi_noirq_set();
+		}
+
+		else if (!memcmp(from, "acpi_sci=edge", 13))
+			acpi_sci_flags.trigger =  1;
+
+		else if (!memcmp(from, "acpi_sci=level", 14))
+			acpi_sci_flags.trigger = 3;
+
+		else if (!memcmp(from, "acpi_sci=high", 13))
+			acpi_sci_flags.polarity = 1;
+
+		else if (!memcmp(from, "acpi_sci=low", 12))
+			acpi_sci_flags.polarity = 3;
+
+#ifdef CONFIG_X86_IO_APIC
+		else if (!memcmp(from, "acpi_skip_timer_override", 24))
+			acpi_skip_timer_override = 1;
+
+		if (!memcmp(from, "disable_timer_pin_1", 19))
+			disable_timer_pin_1 = 1;
+		if (!memcmp(from, "enable_timer_pin_1", 18))
+			disable_timer_pin_1 = -1;
+
+		/* disable IO-APIC */
+		else if (!memcmp(from, "noapic", 6))
+			disable_ioapic_setup();
+#endif /* CONFIG_X86_IO_APIC */
+#endif /* CONFIG_ACPI */
+
+#ifdef CONFIG_X86_LOCAL_APIC
+		/* enable local APIC */
+		else if (!memcmp(from, "lapic", 5))
+			lapic_enable();
+
+		/* disable local APIC */
+		else if (!memcmp(from, "nolapic", 6))
+			lapic_disable();
+#endif /* CONFIG_X86_LOCAL_APIC */
+
+#ifdef CONFIG_KEXEC
+		/* crashkernel=size@addr specifies the location to reserve for
+		 * a crash kernel.  By reserving this memory we guarantee
+		 * that linux never set's it up as a DMA target.
+		 * Useful for holding code to do something appropriate
+		 * after a kernel panic.
+		 */
+		else if (!memcmp(from, "crashkernel=", 12)) {
+#ifndef CONFIG_XEN
+			unsigned long size, base;
+			size = memparse(from+12, &from);
+			if (*from == '@') {
+				base = memparse(from+1, &from);
+				/* FIXME: Do I want a sanity check
+				 * to validate the memory range?
+				 */
+				crashk_res.start = base;
+				crashk_res.end   = base + size - 1;
+			}
+#else
+			printk("Ignoring crashkernel command line, "
+			       "parameter will be supplied by xen\n");
+#endif
+		}
+#endif
+#ifdef CONFIG_PROC_VMCORE
+		/* elfcorehdr= specifies the location of elf core header
+		 * stored by the crashed kernel.
+		 */
+		else if (!memcmp(from, "elfcorehdr=", 11))
+			elfcorehdr_addr = memparse(from+11, &from);
+#endif
+
+		/*
+		 * highmem=size forces highmem to be exactly 'size' bytes.
+		 * This works even on boxes that have no highmem otherwise.
+		 * This also works to reduce highmem size on bigger boxes.
+		 */
+		else if (!memcmp(from, "highmem=", 8))
+			highmem_pages = memparse(from+8, &from) >> PAGE_SHIFT;
+	
+		/*
+		 * vmalloc=size forces the vmalloc area to be exactly 'size'
+		 * bytes. This can be used to increase (or decrease) the
+		 * vmalloc area - the default is 128m.
+		 */
+		else if (!memcmp(from, "vmalloc=", 8))
+			__VMALLOC_RESERVE = memparse(from+8, &from);
+
+	next_char:
+		c = *(from++);
+		if (!c)
+			break;
+		if (COMMAND_LINE_SIZE <= ++len)
+			break;
+		*(to++) = c;
+	}
+	*to = '\0';
+	*cmdline_p = command_line;
+	if (userdef) {
+		printk(KERN_INFO "user-defined physical RAM map:\n");
+		print_memory_map("user");
+	}
+}
+
+/*
+ * Callback for efi_memory_walk.
+ */
+static int __init
+efi_find_max_pfn(unsigned long start, unsigned long end, void *arg)
+{
+	unsigned long *max_pfn = arg, pfn;
+
+	if (start < end) {
+		pfn = PFN_UP(end -1);
+		if (pfn > *max_pfn)
+			*max_pfn = pfn;
+	}
+	return 0;
+}
+
+static int __init
+efi_memory_present_wrapper(unsigned long start, unsigned long end, void *arg)
+{
+	memory_present(0, start, end);
+	return 0;
+}
+
+/*
+ * This function checks if any part of the range <start,end> is mapped
+ * with type.
+ */
+int
+e820_any_mapped(u64 start, u64 end, unsigned type)
+{
+	int i;
+
+#ifndef CONFIG_XEN
+	for (i = 0; i < e820.nr_map; i++) {
+		const struct e820entry *ei = &e820.map[i];
+#else
+	if (!is_initial_xendomain())
+		return 0;
+	for (i = 0; i < machine_e820.nr_map; ++i) {
+		const struct e820entry *ei = &machine_e820.map[i];
+#endif
+
+		if (type && ei->type != type)
+			continue;
+		if (ei->addr >= end || ei->addr + ei->size <= start)
+			continue;
+		return 1;
+	}
+	return 0;
+}
+EXPORT_SYMBOL_GPL(e820_any_mapped);
+
+ /*
+  * This function checks if the entire range <start,end> is mapped with type.
+  *
+  * Note: this function only works correct if the e820 table is sorted and
+  * not-overlapping, which is the case
+  */
+int __init
+e820_all_mapped(unsigned long s, unsigned long e, unsigned type)
+{
+	u64 start = s;
+	u64 end = e;
+	int i;
+
+#ifndef CONFIG_XEN
+	for (i = 0; i < e820.nr_map; i++) {
+		struct e820entry *ei = &e820.map[i];
+#else
+	if (!is_initial_xendomain())
+		return 0;
+	for (i = 0; i < machine_e820.nr_map; ++i) {
+		const struct e820entry *ei = &machine_e820.map[i];
+#endif
+		if (type && ei->type != type)
+			continue;
+		/* is the region (part) in overlap with the current region ?*/
+		if (ei->addr >= end || ei->addr + ei->size <= start)
+			continue;
+		/* if the region is at the beginning of <start,end> we move
+		 * start to the end of the region since it's ok until there
+		 */
+		if (ei->addr <= start)
+			start = ei->addr + ei->size;
+		/* if start is now at or beyond end, we're done, full
+		 * coverage */
+		if (start >= end)
+			return 1; /* we're done */
+	}
+	return 0;
+}
+
+/*
+ * Find the highest page frame number we have available
+ */
+void __init find_max_pfn(void)
+{
+	int i;
+
+	max_pfn = 0;
+	if (efi_enabled) {
+		efi_memmap_walk(efi_find_max_pfn, &max_pfn);
+		efi_memmap_walk(efi_memory_present_wrapper, NULL);
+		return;
+	}
+
+	for (i = 0; i < e820.nr_map; i++) {
+		unsigned long start, end;
+		/* RAM? */
+		if (e820.map[i].type != E820_RAM)
+			continue;
+		start = PFN_UP(e820.map[i].addr);
+		end = PFN_DOWN(e820.map[i].addr + e820.map[i].size);
+		if (start >= end)
+			continue;
+		if (end > max_pfn)
+			max_pfn = end;
+		memory_present(0, start, end);
+	}
+}
+
+/*
+ * Determine low and high memory ranges:
+ */
+unsigned long __init find_max_low_pfn(void)
+{
+	unsigned long max_low_pfn;
+
+	max_low_pfn = max_pfn;
+	if (max_low_pfn > MAXMEM_PFN) {
+		if (highmem_pages == -1)
+			highmem_pages = max_pfn - MAXMEM_PFN;
+		if (highmem_pages + MAXMEM_PFN < max_pfn)
+			max_pfn = MAXMEM_PFN + highmem_pages;
+		if (highmem_pages + MAXMEM_PFN > max_pfn) {
+			printk("only %luMB highmem pages available, ignoring highmem size of %uMB.\n", pages_to_mb(max_pfn - MAXMEM_PFN), pages_to_mb(highmem_pages));
+			highmem_pages = 0;
+		}
+		max_low_pfn = MAXMEM_PFN;
+#ifndef CONFIG_HIGHMEM
+		/* Maximum memory usable is what is directly addressable */
+		printk(KERN_WARNING "Warning only %ldMB will be used.\n",
+					MAXMEM>>20);
+		if (max_pfn > MAX_NONPAE_PFN)
+			printk(KERN_WARNING "Use a PAE enabled kernel.\n");
+		else
+			printk(KERN_WARNING "Use a HIGHMEM enabled kernel.\n");
+		max_pfn = MAXMEM_PFN;
+#else /* !CONFIG_HIGHMEM */
+#ifndef CONFIG_X86_PAE
+		if (max_pfn > MAX_NONPAE_PFN) {
+			max_pfn = MAX_NONPAE_PFN;
+			printk(KERN_WARNING "Warning only 4GB will be used.\n");
+			printk(KERN_WARNING "Use a PAE enabled kernel.\n");
+		}
+#endif /* !CONFIG_X86_PAE */
+#endif /* !CONFIG_HIGHMEM */
+	} else {
+		if (highmem_pages == -1)
+			highmem_pages = 0;
+#ifdef CONFIG_HIGHMEM
+		if (highmem_pages >= max_pfn) {
+			printk(KERN_ERR "highmem size specified (%uMB) is bigger than pages available (%luMB)!.\n", pages_to_mb(highmem_pages), pages_to_mb(max_pfn));
+			highmem_pages = 0;
+		}
+		if (highmem_pages) {
+			if (max_low_pfn-highmem_pages < 64*1024*1024/PAGE_SIZE){
+				printk(KERN_ERR "highmem size %uMB results in smaller than 64MB lowmem, ignoring it.\n", pages_to_mb(highmem_pages));
+				highmem_pages = 0;
+			}
+			max_low_pfn -= highmem_pages;
+		}
+#else
+		if (highmem_pages)
+			printk(KERN_ERR "ignoring highmem size on non-highmem kernel!\n");
+#endif
+	}
+	return max_low_pfn;
+}
+
+/*
+ * Free all available memory for boot time allocation.  Used
+ * as a callback function by efi_memory_walk()
+ */
+
+static int __init
+free_available_memory(unsigned long start, unsigned long end, void *arg)
+{
+	/* check max_low_pfn */
+	if (start >= (max_low_pfn << PAGE_SHIFT))
+		return 0;
+	if (end >= (max_low_pfn << PAGE_SHIFT))
+		end = max_low_pfn << PAGE_SHIFT;
+	if (start < end)
+		free_bootmem(start, end - start);
+
+	return 0;
+}
+/*
+ * Register fully available low RAM pages with the bootmem allocator.
+ */
+static void __init register_bootmem_low_pages(unsigned long max_low_pfn)
+{
+	int i;
+
+	if (efi_enabled) {
+		efi_memmap_walk(free_available_memory, NULL);
+		return;
+	}
+	for (i = 0; i < e820.nr_map; i++) {
+		unsigned long curr_pfn, last_pfn, size;
+		/*
+		 * Reserve usable low memory
+		 */
+		if (e820.map[i].type != E820_RAM)
+			continue;
+		/*
+		 * We are rounding up the start address of usable memory:
+		 */
+		curr_pfn = PFN_UP(e820.map[i].addr);
+		if (curr_pfn >= max_low_pfn)
+			continue;
+		/*
+		 * ... and at the end of the usable range downwards:
+		 */
+		last_pfn = PFN_DOWN(e820.map[i].addr + e820.map[i].size);
+
+#ifdef CONFIG_XEN
+		/*
+                 * Truncate to the number of actual pages currently
+                 * present.
+                 */
+		if (last_pfn > xen_start_info->nr_pages)
+			last_pfn = xen_start_info->nr_pages;
+#endif
+
+		if (last_pfn > max_low_pfn)
+			last_pfn = max_low_pfn;
+
+		/*
+		 * .. finally, did all the rounding and playing
+		 * around just make the area go away?
+		 */
+		if (last_pfn <= curr_pfn)
+			continue;
+
+		size = last_pfn - curr_pfn;
+		free_bootmem(PFN_PHYS(curr_pfn), PFN_PHYS(size));
+	}
+}
+
+#ifndef CONFIG_XEN
+/*
+ * workaround for Dell systems that neglect to reserve EBDA
+ */
+static void __init reserve_ebda_region(void)
+{
+	unsigned int addr;
+	addr = get_bios_ebda();
+	if (addr)
+		reserve_bootmem(addr, PAGE_SIZE);	
+}
+#endif
+
+#ifndef CONFIG_NEED_MULTIPLE_NODES
+void __init setup_bootmem_allocator(void);
+static unsigned long __init setup_memory(void)
+{
+	/*
+	 * partially used pages are not usable - thus
+	 * we are rounding upwards:
+	 */
+ 	min_low_pfn = PFN_UP(__pa(xen_start_info->pt_base)) +
+		xen_start_info->nr_pt_frames;
+
+	find_max_pfn();
+
+	max_low_pfn = find_max_low_pfn();
+
+#ifdef CONFIG_HIGHMEM
+	highstart_pfn = highend_pfn = max_pfn;
+	if (max_pfn > max_low_pfn) {
+		highstart_pfn = max_low_pfn;
+	}
+	printk(KERN_NOTICE "%ldMB HIGHMEM available.\n",
+		pages_to_mb(highend_pfn - highstart_pfn));
+#endif
+	printk(KERN_NOTICE "%ldMB LOWMEM available.\n",
+			pages_to_mb(max_low_pfn));
+
+	setup_bootmem_allocator();
+
+	return max_low_pfn;
+}
+
+void __init zone_sizes_init(void)
+{
+	unsigned long zones_size[MAX_NR_ZONES] = {0, 0, 0};
+	unsigned int max_dma, low;
+
+	max_dma = virt_to_phys((char *)MAX_DMA_ADDRESS) >> PAGE_SHIFT;
+	low = max_low_pfn;
+
+	if (low < max_dma)
+		zones_size[ZONE_DMA] = low;
+	else {
+		zones_size[ZONE_DMA] = max_dma;
+		zones_size[ZONE_NORMAL] = low - max_dma;
+#ifdef CONFIG_HIGHMEM
+		zones_size[ZONE_HIGHMEM] = highend_pfn - low;
+#endif
+	}
+	free_area_init(zones_size);
+}
+#else
+extern unsigned long __init setup_memory(void);
+extern void zone_sizes_init(void);
+#endif /* !CONFIG_NEED_MULTIPLE_NODES */
+
+void __init setup_bootmem_allocator(void)
+{
+	unsigned long bootmap_size;
+	/*
+	 * Initialize the boot-time allocator (with low memory only):
+	 */
+	bootmap_size = init_bootmem(min_low_pfn, max_low_pfn);
+
+	register_bootmem_low_pages(max_low_pfn);
+
+	/*
+	 * Reserve the bootmem bitmap itself as well. We do this in two
+	 * steps (first step was init_bootmem()) because this catches
+	 * the (very unlikely) case of us accidentally initializing the
+	 * bootmem allocator with an invalid RAM area.
+	 */
+	reserve_bootmem(__PHYSICAL_START, (PFN_PHYS(min_low_pfn) +
+			 bootmap_size + PAGE_SIZE-1) - (__PHYSICAL_START));
+
+#ifndef CONFIG_XEN
+	/*
+	 * reserve physical page 0 - it's a special BIOS page on many boxes,
+	 * enabling clean reboots, SMP operation, laptop functions.
+	 */
+	reserve_bootmem(0, PAGE_SIZE);
+
+	/* reserve EBDA region, it's a 4K region */
+	reserve_ebda_region();
+
+    /* could be an AMD 768MPX chipset. Reserve a page  before VGA to prevent
+       PCI prefetch into it (errata #56). Usually the page is reserved anyways,
+       unless you have no PS/2 mouse plugged in. */
+	if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD &&
+	    boot_cpu_data.x86 == 6)
+	     reserve_bootmem(0xa0000 - 4096, 4096);
+
+#ifdef CONFIG_SMP
+	/*
+	 * But first pinch a few for the stack/trampoline stuff
+	 * FIXME: Don't need the extra page at 4K, but need to fix
+	 * trampoline before removing it. (see the GDT stuff)
+	 */
+	reserve_bootmem(PAGE_SIZE, PAGE_SIZE);
+#endif
+#ifdef CONFIG_ACPI_SLEEP
+	/*
+	 * Reserve low memory region for sleep support.
+	 */
+	acpi_reserve_bootmem();
+#endif
+#endif /* !CONFIG_XEN */
+
+#ifdef CONFIG_BLK_DEV_INITRD
+	if (xen_start_info->mod_start) {
+		if (INITRD_START + INITRD_SIZE <= (max_low_pfn << PAGE_SHIFT)) {
+			/*reserve_bootmem(INITRD_START, INITRD_SIZE);*/
+			initrd_start = INITRD_START + PAGE_OFFSET;
+			initrd_end = initrd_start+INITRD_SIZE;
+			initrd_below_start_ok = 1;
+		}
+		else {
+			printk(KERN_ERR "initrd extends beyond end of memory "
+			    "(0x%08lx > 0x%08lx)\ndisabling initrd\n",
+			    INITRD_START + INITRD_SIZE,
+			    max_low_pfn << PAGE_SHIFT);
+			initrd_start = 0;
+		}
+	}
+#endif
+#ifdef CONFIG_KEXEC
+#ifdef CONFIG_XEN
+	xen_machine_kexec_setup_resources();
+#else
+	if (crashk_res.start != crashk_res.end)
+		reserve_bootmem(crashk_res.start,
+			crashk_res.end - crashk_res.start + 1);
+#endif
+#endif
+}
+
+/*
+ * The node 0 pgdat is initialized before all of these because
+ * it's needed for bootmem.  node>0 pgdats have their virtual
+ * space allocated before the pagetables are in place to access
+ * them, so they can't be cleared then.
+ *
+ * This should all compile down to nothing when NUMA is off.
+ */
+void __init remapped_pgdat_init(void)
+{
+	int nid;
+
+	for_each_online_node(nid) {
+		if (nid != 0)
+			memset(NODE_DATA(nid), 0, sizeof(struct pglist_data));
+	}
+}
+
+/*
+ * Request address space for all standard RAM and ROM resources
+ * and also for regions reported as reserved by the e820.
+ */
+static void __init
+legacy_init_iomem_resources(struct e820entry *e820, int nr_map,
+			    struct resource *code_resource,
+			    struct resource *data_resource)
+{
+	int i;
+
+	probe_roms();
+
+	for (i = 0; i < nr_map; i++) {
+		struct resource *res;
+#ifndef CONFIG_RESOURCES_64BIT
+		if (e820[i].addr + e820[i].size > 0x100000000ULL)
+			continue;
+#endif
+		res = kzalloc(sizeof(struct resource), GFP_ATOMIC);
+		switch (e820[i].type) {
+		case E820_RAM:	res->name = "System RAM"; break;
+		case E820_ACPI:	res->name = "ACPI Tables"; break;
+		case E820_NVS:	res->name = "ACPI Non-volatile Storage"; break;
+		default:	res->name = "reserved";
+		}
+		res->start = e820[i].addr;
+		res->end = res->start + e820[i].size - 1;
+		res->flags = IORESOURCE_MEM | IORESOURCE_BUSY;
+		if (request_resource(&iomem_resource, res)) {
+			kfree(res);
+			continue;
+		}
+		if (e820[i].type == E820_RAM) {
+			/*
+			 *  We don't know which RAM region contains kernel data,
+			 *  so we try it repeatedly and let the resource manager
+			 *  test it.
+			 */
+#ifndef CONFIG_XEN
+			request_resource(res, code_resource);
+			request_resource(res, data_resource);
+#endif
+#ifdef CONFIG_KEXEC
+			if (crashk_res.start != crashk_res.end)
+			     request_resource(res, &crashk_res);
+#ifdef CONFIG_XEN
+			xen_machine_kexec_register_resources(res);
+#endif
+#endif
+		}
+	}
+}
+
+/*
+ * Locate a unused range of the physical address space below 4G which
+ * can be used for PCI mappings.
+ */
+static void __init
+e820_setup_gap(struct e820entry *e820, int nr_map)
+{
+	unsigned long gapstart, gapsize, round;
+	unsigned long long last;
+	int i;
+
+	/*
+	 * Search for the bigest gap in the low 32 bits of the e820
+	 * memory space.
+	 */
+	last = 0x100000000ull;
+	gapstart = 0x10000000;
+	gapsize = 0x400000;
+	i = nr_map;
+	while (--i >= 0) {
+		unsigned long long start = e820[i].addr;
+		unsigned long long end = start + e820[i].size;
+
+		/*
+		 * Since "last" is at most 4GB, we know we'll
+		 * fit in 32 bits if this condition is true
+		 */
+		if (last > end) {
+			unsigned long gap = last - end;
+
+			if (gap > gapsize) {
+				gapsize = gap;
+				gapstart = end;
+			}
+		}
+		if (start < last)
+			last = start;
+	}
+
+	/*
+	 * See how much we want to round up: start off with
+	 * rounding to the next 1MB area.
+	 */
+	round = 0x100000;
+	while ((gapsize >> 4) > round)
+		round += round;
+	/* Fun with two's complement */
+	pci_mem_start = (gapstart + round) & -round;
+
+	printk("Allocating PCI resources starting at %08lx (gap: %08lx:%08lx)\n",
+		pci_mem_start, gapstart, gapsize);
+}
+
+/*
+ * Request address space for all standard resources
+ *
+ * This is called just before pcibios_init(), which is also a
+ * subsys_initcall, but is linked in later (in arch/i386/pci/common.c).
+ */
+static int __init request_standard_resources(void)
+{
+	int	      i;
+
+	/* Nothing to do if not running in dom0. */
+	if (!is_initial_xendomain())
+		return 0;
+
+	printk("Setting up standard PCI resources\n");
+#ifdef CONFIG_XEN
+	legacy_init_iomem_resources(machine_e820.map, machine_e820.nr_map,
+				    &code_resource, &data_resource);
+#else
+	if (efi_enabled)
+		efi_initialize_iomem_resources(&code_resource, &data_resource);
+	else
+		legacy_init_iomem_resources(e820.map, e820.nr_map,
+					    &code_resource, &data_resource);
+#endif
+
+	/* EFI systems may still have VGA */
+	request_resource(&iomem_resource, &video_ram_resource);
+
+	/* request I/O space for devices used on all i[345]86 PCs */
+	for (i = 0; i < STANDARD_IO_RESOURCES; i++)
+		request_resource(&ioport_resource, &standard_io_resources[i]);
+	return 0;
+}
+
+subsys_initcall(request_standard_resources);
+
+static void __init register_memory(void)
+{
+#ifdef CONFIG_XEN
+	if (is_initial_xendomain())
+		e820_setup_gap(machine_e820.map, machine_e820.nr_map);
+	else
+#endif
+		e820_setup_gap(e820.map, e820.nr_map);
+}
+
+#ifdef CONFIG_MCA
+static void set_mca_bus(int x)
+{
+	MCA_bus = x;
+}
+#else
+static void set_mca_bus(int x) { }
+#endif
+
+/*
+ * Determine if we were loaded by an EFI loader.  If so, then we have also been
+ * passed the efi memmap, systab, etc., so we should use these data structures
+ * for initialization.  Note, the efi init code path is determined by the
+ * global efi_enabled. This allows the same kernel image to be used on existing
+ * systems (with a traditional BIOS) as well as on EFI systems.
+ */
+void __init setup_arch(char **cmdline_p)
+{
+	int i, j, k, fpp;
+	struct physdev_set_iopl set_iopl;
+	unsigned long max_low_pfn;
+	unsigned long p2m_pages;
+
+	/* Force a quick death if the kernel panics (not domain 0). */
+	extern int panic_timeout;
+	if (!panic_timeout && !is_initial_xendomain())
+		panic_timeout = 1;
+
+	/* Register a call for panic conditions. */
+	atomic_notifier_chain_register(&panic_notifier_list, &xen_panic_block);
+
+	WARN_ON(HYPERVISOR_vm_assist(VMASST_CMD_enable,
+				     VMASST_TYPE_4gb_segments));
+	WARN_ON(HYPERVISOR_vm_assist(VMASST_CMD_enable,
+				     VMASST_TYPE_writable_pagetables));
+
+	memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data));
+	pre_setup_arch_hook();
+	early_cpu_init();
+#ifdef CONFIG_SMP
+	prefill_possible_map();
+#endif
+
+	/*
+	 * FIXME: This isn't an official loader_type right
+	 * now but does currently work with elilo.
+	 * If we were configured as an EFI kernel, check to make
+	 * sure that we were loaded correctly from elilo and that
+	 * the system table is valid.  If not, then initialize normally.
+	 */
+#ifdef CONFIG_EFI
+	if ((LOADER_TYPE == 0x50) && EFI_SYSTAB)
+		efi_enabled = 1;
+#endif
+
+	/* This must be initialized to UNNAMED_MAJOR for ipconfig to work
+	   properly.  Setting ROOT_DEV to default to /dev/ram0 breaks initrd.
+	*/
+	ROOT_DEV = MKDEV(UNNAMED_MAJOR,0);
+ 	drive_info = DRIVE_INFO;
+ 	screen_info = SCREEN_INFO;
+	copy_edid();
+	apm_info.bios = APM_BIOS_INFO;
+	ist_info = IST_INFO;
+	saved_videomode = VIDEO_MODE;
+	if( SYS_DESC_TABLE.length != 0 ) {
+		set_mca_bus(SYS_DESC_TABLE.table[3] & 0x2);
+		machine_id = SYS_DESC_TABLE.table[0];
+		machine_submodel_id = SYS_DESC_TABLE.table[1];
+		BIOS_revision = SYS_DESC_TABLE.table[2];
+	}
+	bootloader_type = LOADER_TYPE;
+
+	if (is_initial_xendomain()) {
+		const struct dom0_vga_console_info *info =
+			(void *)((char *)xen_start_info +
+			         xen_start_info->console.dom0.info_off);
+
+		dom0_init_screen_info(info,
+		                      xen_start_info->console.dom0.info_size);
+		xen_start_info->console.domU.mfn = 0;
+		xen_start_info->console.domU.evtchn = 0;
+	} else
+		screen_info.orig_video_isVGA = 0;
+
+#ifdef CONFIG_BLK_DEV_RAM
+	rd_image_start = RAMDISK_FLAGS & RAMDISK_IMAGE_START_MASK;
+	rd_prompt = ((RAMDISK_FLAGS & RAMDISK_PROMPT_FLAG) != 0);
+	rd_doload = ((RAMDISK_FLAGS & RAMDISK_LOAD_FLAG) != 0);
+#endif
+
+	ARCH_SETUP
+	if (efi_enabled)
+		efi_init();
+	else {
+		printk(KERN_INFO "BIOS-provided physical RAM map:\n");
+		print_memory_map(machine_specific_memory_setup());
+	}
+
+	copy_edd();
+
+	if (!MOUNT_ROOT_RDONLY)
+		root_mountflags &= ~MS_RDONLY;
+	init_mm.start_code = (unsigned long) _text;
+	init_mm.end_code = (unsigned long) _etext;
+	init_mm.end_data = (unsigned long) _edata;
+	init_mm.brk = (PFN_UP(__pa(xen_start_info->pt_base)) +
+		       xen_start_info->nr_pt_frames) << PAGE_SHIFT;
+
+	code_resource.start = virt_to_phys(_text);
+	code_resource.end = virt_to_phys(_etext)-1;
+	data_resource.start = virt_to_phys(_etext);
+	data_resource.end = virt_to_phys(_edata)-1;
+
+	parse_cmdline_early(cmdline_p);
+
+#ifdef CONFIG_EARLY_PRINTK
+	{
+		char *s = strstr(*cmdline_p, "earlyprintk=");
+		if (s) {
+			setup_early_printk(strchr(s, '=') + 1);
+			printk("early console enabled\n");
+		}
+	}
+#endif
+
+	max_low_pfn = setup_memory();
+
+	/*
+	 * NOTE: before this point _nobody_ is allowed to allocate
+	 * any memory using the bootmem allocator.  Although the
+	 * alloctor is now initialised only the first 8Mb of the kernel
+	 * virtual address space has been mapped.  All allocations before
+	 * paging_init() has completed must use the alloc_bootmem_low_pages()
+	 * variant (which allocates DMA'able memory) and care must be taken
+	 * not to exceed the 8Mb limit.
+	 */
+
+#ifdef CONFIG_SMP
+	smp_alloc_memory(); /* AP processor realmode stacks in low memory*/
+#endif
+	paging_init();
+	remapped_pgdat_init();
+	sparse_init();
+	zone_sizes_init();
+
+#ifdef CONFIG_X86_FIND_SMP_CONFIG
+	/*
+	 * Find and reserve possible boot-time SMP configuration:
+	 */
+	find_smp_config();
+#endif
+
+	p2m_pages = max_pfn;
+	if (xen_start_info->nr_pages > max_pfn) {
+		/*
+		 * the max_pfn was shrunk (probably by mem= or highmem=
+		 * kernel parameter); shrink reservation with the HV
+		 */
+		struct xen_memory_reservation reservation = {
+			.address_bits = 0,
+			.extent_order = 0,
+			.domid = DOMID_SELF
+		};
+		unsigned int difference;
+		int ret;
+
+		difference = xen_start_info->nr_pages - max_pfn;
+
+		set_xen_guest_handle(reservation.extent_start,
+				     ((unsigned long *)xen_start_info->mfn_list) + max_pfn);
+		reservation.nr_extents = difference;
+		ret = HYPERVISOR_memory_op(XENMEM_decrease_reservation,
+					   &reservation);
+		BUG_ON (ret != difference);
+	}
+	else if (max_pfn > xen_start_info->nr_pages)
+		p2m_pages = xen_start_info->nr_pages;
+
+	/* Make sure we have a correctly sized P->M table. */
+	if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+		phys_to_machine_mapping = alloc_bootmem_low_pages(
+		     max_pfn * sizeof(unsigned long));
+		memset(phys_to_machine_mapping, ~0,
+		       max_pfn * sizeof(unsigned long));
+		memcpy(phys_to_machine_mapping,
+		       (unsigned long *)xen_start_info->mfn_list,
+		       p2m_pages * sizeof(unsigned long));
+		free_bootmem(
+		     __pa(xen_start_info->mfn_list),
+		     PFN_PHYS(PFN_UP(xen_start_info->nr_pages *
+				     sizeof(unsigned long))));
+
+		/*
+		 * Initialise the list of the frames that specify the list of
+		 * frames that make up the p2m table. Used by save/restore
+		 */
+		pfn_to_mfn_frame_list_list = alloc_bootmem_low_pages(PAGE_SIZE);
+
+		fpp = PAGE_SIZE/sizeof(unsigned long);
+		for (i=0, j=0, k=-1; i< max_pfn; i+=fpp, j++) {
+			if ((j % fpp) == 0) {
+				k++;
+				BUG_ON(k>=16);
+				pfn_to_mfn_frame_list[k] =
+					alloc_bootmem_low_pages(PAGE_SIZE);
+				pfn_to_mfn_frame_list_list[k] =
+					virt_to_mfn(pfn_to_mfn_frame_list[k]);
+				j=0;
+			}
+			pfn_to_mfn_frame_list[k][j] =
+				virt_to_mfn(&phys_to_machine_mapping[i]);
+		}
+		HYPERVISOR_shared_info->arch.max_pfn = max_pfn;
+		HYPERVISOR_shared_info->arch.pfn_to_mfn_frame_list_list =
+		     virt_to_mfn(pfn_to_mfn_frame_list_list);
+	}
+
+	/* Mark all ISA DMA channels in-use - using them wouldn't work. */
+	for (i = 0; i < MAX_DMA_CHANNELS; ++i)
+		if (i != 4 && request_dma(i, "xen") != 0)
+			BUG();
+
+	/*
+	 * NOTE: at this point the bootmem allocator is fully available.
+	 */
+
+	if (is_initial_xendomain())
+		dmi_scan_machine();
+
+#ifdef CONFIG_X86_GENERICARCH
+	generic_apic_probe(*cmdline_p);
+#endif	
+	if (efi_enabled)
+		efi_map_memmap();
+
+	set_iopl.iopl = 1;
+	WARN_ON(HYPERVISOR_physdev_op(PHYSDEVOP_set_iopl, &set_iopl));
+
+#ifdef CONFIG_ACPI
+	if (!is_initial_xendomain()) {
+		printk(KERN_INFO "ACPI in unprivileged domain disabled\n");
+		acpi_disabled = 1;
+		acpi_ht = 0;
+	}
+
+	/*
+	 * Parse the ACPI tables for possible boot-time SMP configuration.
+	 */
+	acpi_boot_table_init();
+#endif
+
+#ifdef CONFIG_X86_IO_APIC
+	check_acpi_pci();	/* Checks more than just ACPI actually */
+#endif
+
+#ifdef CONFIG_ACPI
+	acpi_boot_init();
+
+#if defined(CONFIG_SMP) && defined(CONFIG_X86_PC)
+	if (def_to_bigsmp)
+		printk(KERN_WARNING "More than 8 CPUs detected and "
+			"CONFIG_X86_PC cannot handle it.\nUse "
+			"CONFIG_X86_GENERICARCH or CONFIG_X86_BIGSMP.\n");
+#endif
+#endif
+#ifdef CONFIG_X86_LOCAL_APIC
+	if (smp_found_config)
+		get_smp_config();
+#endif
+
+	register_memory();
+
+	if (is_initial_xendomain()) {
+#ifdef CONFIG_VT
+#if defined(CONFIG_VGA_CONSOLE)
+		if (!efi_enabled ||
+		    (efi_mem_type(0xa0000) != EFI_CONVENTIONAL_MEMORY))
+			conswitchp = &vga_con;
+#elif defined(CONFIG_DUMMY_CONSOLE)
+		conswitchp = &dummy_con;
+#endif
+#endif
+	} else {
+#if defined(CONFIG_VT) && defined(CONFIG_DUMMY_CONSOLE)
+		conswitchp = &dummy_con;
+#endif
+	}
+	tsc_init();
+}
+
+static int
+xen_panic_event(struct notifier_block *this, unsigned long event, void *ptr)
+{
+	HYPERVISOR_shutdown(SHUTDOWN_crash);
+	/* we're never actually going to get here... */
+	return NOTIFY_DONE;
+}
+
+static __init int add_pcspkr(void)
+{
+	struct platform_device *pd;
+	int ret;
+
+	if (!is_initial_xendomain())
+		return 0;
+
+	pd = platform_device_alloc("pcspkr", -1);
+	if (!pd)
+		return -ENOMEM;
+
+	ret = platform_device_add(pd);
+	if (ret)
+		platform_device_put(pd);
+
+	return ret;
+}
+device_initcall(add_pcspkr);
+
+/*
+ * Local Variables:
+ * mode:c
+ * c-file-style:"k&r"
+ * c-basic-offset:8
+ * End:
+ */
diff -rpuN linux-2.6.18.8/arch/i386/kernel/smp-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/smp-xen.c
--- linux-2.6.18.8/arch/i386/kernel/smp-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/smp-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,605 @@
+/*
+ *	Intel SMP support routines.
+ *
+ *	(c) 1995 Alan Cox, Building #3 <alan@redhat.com>
+ *	(c) 1998-99, 2000 Ingo Molnar <mingo@redhat.com>
+ *
+ *	This code is released under the GNU General Public License version 2 or
+ *	later.
+ */
+
+#include <linux/init.h>
+
+#include <linux/mm.h>
+#include <linux/delay.h>
+#include <linux/spinlock.h>
+#include <linux/smp_lock.h>
+#include <linux/kernel_stat.h>
+#include <linux/mc146818rtc.h>
+#include <linux/cache.h>
+#include <linux/interrupt.h>
+#include <linux/cpu.h>
+#include <linux/module.h>
+
+#include <asm/mtrr.h>
+#include <asm/tlbflush.h>
+#if 0
+#include <mach_apic.h>
+#endif
+#include <xen/evtchn.h>
+
+/*
+ *	Some notes on x86 processor bugs affecting SMP operation:
+ *
+ *	Pentium, Pentium Pro, II, III (and all CPUs) have bugs.
+ *	The Linux implications for SMP are handled as follows:
+ *
+ *	Pentium III / [Xeon]
+ *		None of the E1AP-E3AP errata are visible to the user.
+ *
+ *	E1AP.	see PII A1AP
+ *	E2AP.	see PII A2AP
+ *	E3AP.	see PII A3AP
+ *
+ *	Pentium II / [Xeon]
+ *		None of the A1AP-A3AP errata are visible to the user.
+ *
+ *	A1AP.	see PPro 1AP
+ *	A2AP.	see PPro 2AP
+ *	A3AP.	see PPro 7AP
+ *
+ *	Pentium Pro
+ *		None of 1AP-9AP errata are visible to the normal user,
+ *	except occasional delivery of 'spurious interrupt' as trap #15.
+ *	This is very rare and a non-problem.
+ *
+ *	1AP.	Linux maps APIC as non-cacheable
+ *	2AP.	worked around in hardware
+ *	3AP.	fixed in C0 and above steppings microcode update.
+ *		Linux does not use excessive STARTUP_IPIs.
+ *	4AP.	worked around in hardware
+ *	5AP.	symmetric IO mode (normal Linux operation) not affected.
+ *		'noapic' mode has vector 0xf filled out properly.
+ *	6AP.	'noapic' mode might be affected - fixed in later steppings
+ *	7AP.	We do not assume writes to the LVT deassering IRQs
+ *	8AP.	We do not enable low power mode (deep sleep) during MP bootup
+ *	9AP.	We do not use mixed mode
+ *
+ *	Pentium
+ *		There is a marginal case where REP MOVS on 100MHz SMP
+ *	machines with B stepping processors can fail. XXX should provide
+ *	an L1cache=Writethrough or L1cache=off option.
+ *
+ *		B stepping CPUs may hang. There are hardware work arounds
+ *	for this. We warn about it in case your board doesn't have the work
+ *	arounds. Basically thats so I can tell anyone with a B stepping
+ *	CPU and SMP problems "tough".
+ *
+ *	Specific items [From Pentium Processor Specification Update]
+ *
+ *	1AP.	Linux doesn't use remote read
+ *	2AP.	Linux doesn't trust APIC errors
+ *	3AP.	We work around this
+ *	4AP.	Linux never generated 3 interrupts of the same priority
+ *		to cause a lost local interrupt.
+ *	5AP.	Remote read is never used
+ *	6AP.	not affected - worked around in hardware
+ *	7AP.	not affected - worked around in hardware
+ *	8AP.	worked around in hardware - we get explicit CS errors if not
+ *	9AP.	only 'noapic' mode affected. Might generate spurious
+ *		interrupts, we log only the first one and count the
+ *		rest silently.
+ *	10AP.	not affected - worked around in hardware
+ *	11AP.	Linux reads the APIC between writes to avoid this, as per
+ *		the documentation. Make sure you preserve this as it affects
+ *		the C stepping chips too.
+ *	12AP.	not affected - worked around in hardware
+ *	13AP.	not affected - worked around in hardware
+ *	14AP.	we always deassert INIT during bootup
+ *	15AP.	not affected - worked around in hardware
+ *	16AP.	not affected - worked around in hardware
+ *	17AP.	not affected - worked around in hardware
+ *	18AP.	not affected - worked around in hardware
+ *	19AP.	not affected - worked around in BIOS
+ *
+ *	If this sounds worrying believe me these bugs are either ___RARE___,
+ *	or are signal timing bugs worked around in hardware and there's
+ *	about nothing of note with C stepping upwards.
+ */
+
+DEFINE_PER_CPU(struct tlb_state, cpu_tlbstate) ____cacheline_aligned = { &init_mm, 0, };
+
+/*
+ * the following functions deal with sending IPIs between CPUs.
+ *
+ * We use 'broadcast', CPU->CPU IPIs and self-IPIs too.
+ */
+
+static inline int __prepare_ICR (unsigned int shortcut, int vector)
+{
+	unsigned int icr = shortcut | APIC_DEST_LOGICAL;
+
+	switch (vector) {
+	default:
+		icr |= APIC_DM_FIXED | vector;
+		break;
+	case NMI_VECTOR:
+		icr |= APIC_DM_NMI;
+		break;
+	}
+	return icr;
+}
+
+static inline int __prepare_ICR2 (unsigned int mask)
+{
+	return SET_APIC_DEST_FIELD(mask);
+}
+
+DECLARE_PER_CPU(int, ipi_to_irq[NR_IPIS]);
+
+static inline void __send_IPI_one(unsigned int cpu, int vector)
+{
+	int irq = per_cpu(ipi_to_irq, cpu)[vector];
+	BUG_ON(irq < 0);
+	notify_remote_via_irq(irq);
+}
+
+void __send_IPI_shortcut(unsigned int shortcut, int vector)
+{
+	int cpu;
+
+	switch (shortcut) {
+	case APIC_DEST_SELF:
+		__send_IPI_one(smp_processor_id(), vector);
+		break;
+	case APIC_DEST_ALLBUT:
+		for (cpu = 0; cpu < NR_CPUS; ++cpu) {
+			if (cpu == smp_processor_id())
+				continue;
+			if (cpu_isset(cpu, cpu_online_map)) {
+				__send_IPI_one(cpu, vector);
+			}
+		}
+		break;
+	default:
+		printk("XXXXXX __send_IPI_shortcut %08x vector %d\n", shortcut,
+		       vector);
+		break;
+	}
+}
+
+void fastcall send_IPI_self(int vector)
+{
+	__send_IPI_shortcut(APIC_DEST_SELF, vector);
+}
+
+/*
+ * This is only used on smaller machines.
+ */
+void send_IPI_mask_bitmask(cpumask_t mask, int vector)
+{
+	unsigned long flags;
+	unsigned int cpu;
+
+	local_irq_save(flags);
+	WARN_ON(cpus_addr(mask)[0] & ~cpus_addr(cpu_online_map)[0]);
+
+	for (cpu = 0; cpu < NR_CPUS; ++cpu) {
+		if (cpu_isset(cpu, mask)) {
+			__send_IPI_one(cpu, vector);
+		}
+	}
+
+	local_irq_restore(flags);
+}
+
+void send_IPI_mask_sequence(cpumask_t mask, int vector)
+{
+
+	send_IPI_mask_bitmask(mask, vector);
+}
+
+#include <mach_ipi.h> /* must come after the send_IPI functions above for inlining */
+
+#if 0 /* XEN */
+/*
+ *	Smarter SMP flushing macros. 
+ *		c/o Linus Torvalds.
+ *
+ *	These mean you can really definitely utterly forget about
+ *	writing to user space from interrupts. (Its not allowed anyway).
+ *
+ *	Optimizations Manfred Spraul <manfred@colorfullife.com>
+ */
+
+static cpumask_t flush_cpumask;
+static struct mm_struct * flush_mm;
+static unsigned long flush_va;
+static DEFINE_SPINLOCK(tlbstate_lock);
+#define FLUSH_ALL	0xffffffff
+
+/*
+ * We cannot call mmdrop() because we are in interrupt context, 
+ * instead update mm->cpu_vm_mask.
+ *
+ * We need to reload %cr3 since the page tables may be going
+ * away from under us..
+ */
+static inline void leave_mm (unsigned long cpu)
+{
+	if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_OK)
+		BUG();
+	cpu_clear(cpu, per_cpu(cpu_tlbstate, cpu).active_mm->cpu_vm_mask);
+	load_cr3(swapper_pg_dir);
+}
+
+/*
+ *
+ * The flush IPI assumes that a thread switch happens in this order:
+ * [cpu0: the cpu that switches]
+ * 1) switch_mm() either 1a) or 1b)
+ * 1a) thread switch to a different mm
+ * 1a1) cpu_clear(cpu, old_mm->cpu_vm_mask);
+ * 	Stop ipi delivery for the old mm. This is not synchronized with
+ * 	the other cpus, but smp_invalidate_interrupt ignore flush ipis
+ * 	for the wrong mm, and in the worst case we perform a superflous
+ * 	tlb flush.
+ * 1a2) set cpu_tlbstate to TLBSTATE_OK
+ * 	Now the smp_invalidate_interrupt won't call leave_mm if cpu0
+ *	was in lazy tlb mode.
+ * 1a3) update cpu_tlbstate[].active_mm
+ * 	Now cpu0 accepts tlb flushes for the new mm.
+ * 1a4) cpu_set(cpu, new_mm->cpu_vm_mask);
+ * 	Now the other cpus will send tlb flush ipis.
+ * 1a4) change cr3.
+ * 1b) thread switch without mm change
+ *	cpu_tlbstate[].active_mm is correct, cpu0 already handles
+ *	flush ipis.
+ * 1b1) set cpu_tlbstate to TLBSTATE_OK
+ * 1b2) test_and_set the cpu bit in cpu_vm_mask.
+ * 	Atomically set the bit [other cpus will start sending flush ipis],
+ * 	and test the bit.
+ * 1b3) if the bit was 0: leave_mm was called, flush the tlb.
+ * 2) switch %%esp, ie current
+ *
+ * The interrupt must handle 2 special cases:
+ * - cr3 is changed before %%esp, ie. it cannot use current->{active_,}mm.
+ * - the cpu performs speculative tlb reads, i.e. even if the cpu only
+ *   runs in kernel space, the cpu could load tlb entries for user space
+ *   pages.
+ *
+ * The good news is that cpu_tlbstate is local to each cpu, no
+ * write/read ordering problems.
+ */
+
+/*
+ * TLB flush IPI:
+ *
+ * 1) Flush the tlb entries if the cpu uses the mm that's being flushed.
+ * 2) Leave the mm if we are in the lazy tlb mode.
+ */
+
+irqreturn_t smp_invalidate_interrupt(int irq, void *dev_id,
+				     struct pt_regs *regs)
+{
+	unsigned long cpu;
+
+	cpu = get_cpu();
+
+	if (!cpu_isset(cpu, flush_cpumask))
+		goto out;
+		/* 
+		 * This was a BUG() but until someone can quote me the
+		 * line from the intel manual that guarantees an IPI to
+		 * multiple CPUs is retried _only_ on the erroring CPUs
+		 * its staying as a return
+		 *
+		 * BUG();
+		 */
+		 
+	if (flush_mm == per_cpu(cpu_tlbstate, cpu).active_mm) {
+		if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_OK) {
+			if (flush_va == FLUSH_ALL)
+				local_flush_tlb();
+			else
+				__flush_tlb_one(flush_va);
+		} else
+			leave_mm(cpu);
+	}
+	smp_mb__before_clear_bit();
+	cpu_clear(cpu, flush_cpumask);
+	smp_mb__after_clear_bit();
+out:
+	put_cpu_no_resched();
+
+	return IRQ_HANDLED;
+}
+
+static void flush_tlb_others(cpumask_t cpumask, struct mm_struct *mm,
+						unsigned long va)
+{
+	/*
+	 * A couple of (to be removed) sanity checks:
+	 *
+	 * - current CPU must not be in mask
+	 * - mask must exist :)
+	 */
+	BUG_ON(cpus_empty(cpumask));
+	BUG_ON(cpu_isset(smp_processor_id(), cpumask));
+	BUG_ON(!mm);
+
+	/* If a CPU which we ran on has gone down, OK. */
+	cpus_and(cpumask, cpumask, cpu_online_map);
+	if (cpus_empty(cpumask))
+		return;
+
+	/*
+	 * i'm not happy about this global shared spinlock in the
+	 * MM hot path, but we'll see how contended it is.
+	 * Temporarily this turns IRQs off, so that lockups are
+	 * detected by the NMI watchdog.
+	 */
+	spin_lock(&tlbstate_lock);
+	
+	flush_mm = mm;
+	flush_va = va;
+#if NR_CPUS <= BITS_PER_LONG
+	atomic_set_mask(cpumask, &flush_cpumask);
+#else
+	{
+		int k;
+		unsigned long *flush_mask = (unsigned long *)&flush_cpumask;
+		unsigned long *cpu_mask = (unsigned long *)&cpumask;
+		for (k = 0; k < BITS_TO_LONGS(NR_CPUS); ++k)
+			atomic_set_mask(cpu_mask[k], &flush_mask[k]);
+	}
+#endif
+	/*
+	 * We have to send the IPI only to
+	 * CPUs affected.
+	 */
+	send_IPI_mask(cpumask, INVALIDATE_TLB_VECTOR);
+
+	while (!cpus_empty(flush_cpumask))
+		/* nothing. lockup detection does not belong here */
+		mb();
+
+	flush_mm = NULL;
+	flush_va = 0;
+	spin_unlock(&tlbstate_lock);
+}
+	
+void flush_tlb_current_task(void)
+{
+	struct mm_struct *mm = current->mm;
+	cpumask_t cpu_mask;
+
+	preempt_disable();
+	cpu_mask = mm->cpu_vm_mask;
+	cpu_clear(smp_processor_id(), cpu_mask);
+
+	local_flush_tlb();
+	if (!cpus_empty(cpu_mask))
+		flush_tlb_others(cpu_mask, mm, FLUSH_ALL);
+	preempt_enable();
+}
+
+void flush_tlb_mm (struct mm_struct * mm)
+{
+	cpumask_t cpu_mask;
+
+	preempt_disable();
+	cpu_mask = mm->cpu_vm_mask;
+	cpu_clear(smp_processor_id(), cpu_mask);
+
+	if (current->active_mm == mm) {
+		if (current->mm)
+			local_flush_tlb();
+		else
+			leave_mm(smp_processor_id());
+	}
+	if (!cpus_empty(cpu_mask))
+		flush_tlb_others(cpu_mask, mm, FLUSH_ALL);
+
+	preempt_enable();
+}
+
+void flush_tlb_page(struct vm_area_struct * vma, unsigned long va)
+{
+	struct mm_struct *mm = vma->vm_mm;
+	cpumask_t cpu_mask;
+
+	preempt_disable();
+	cpu_mask = mm->cpu_vm_mask;
+	cpu_clear(smp_processor_id(), cpu_mask);
+
+	if (current->active_mm == mm) {
+		if(current->mm)
+			__flush_tlb_one(va);
+		else
+		 	leave_mm(smp_processor_id());
+	}
+
+	if (!cpus_empty(cpu_mask))
+		flush_tlb_others(cpu_mask, mm, va);
+
+	preempt_enable();
+}
+EXPORT_SYMBOL(flush_tlb_page);
+
+static void do_flush_tlb_all(void* info)
+{
+	unsigned long cpu = smp_processor_id();
+
+	__flush_tlb_all();
+	if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_LAZY)
+		leave_mm(cpu);
+}
+
+void flush_tlb_all(void)
+{
+	on_each_cpu(do_flush_tlb_all, NULL, 1, 1);
+}
+
+#endif /* XEN */
+
+/*
+ * this function sends a 'reschedule' IPI to another CPU.
+ * it goes straight through and wastes no time serializing
+ * anything. Worst case is that we lose a reschedule ...
+ */
+void smp_send_reschedule(int cpu)
+{
+	WARN_ON(cpu_is_offline(cpu));
+	send_IPI_mask(cpumask_of_cpu(cpu), RESCHEDULE_VECTOR);
+}
+
+/*
+ * Structure and data for smp_call_function(). This is designed to minimise
+ * static memory requirements. It also looks cleaner.
+ */
+static DEFINE_SPINLOCK(call_lock);
+
+struct call_data_struct {
+	void (*func) (void *info);
+	void *info;
+	atomic_t started;
+	atomic_t finished;
+	int wait;
+};
+
+void lock_ipi_call_lock(void)
+{
+	spin_lock_irq(&call_lock);
+}
+
+void unlock_ipi_call_lock(void)
+{
+	spin_unlock_irq(&call_lock);
+}
+
+static struct call_data_struct *call_data;
+
+/**
+ * smp_call_function(): Run a function on all other CPUs.
+ * @func: The function to run. This must be fast and non-blocking.
+ * @info: An arbitrary pointer to pass to the function.
+ * @nonatomic: currently unused.
+ * @wait: If true, wait (atomically) until function has completed on other CPUs.
+ *
+ * Returns 0 on success, else a negative status code. Does not return until
+ * remote CPUs are nearly ready to execute <<func>> or are or have executed.
+ *
+ * You must not call this function with disabled interrupts or from a
+ * hardware interrupt handler or from a bottom half handler.
+ */
+int smp_call_function (void (*func) (void *info), void *info, int nonatomic,
+			int wait)
+{
+	struct call_data_struct data;
+	int cpus;
+
+	/* Holding any lock stops cpus from going down. */
+	spin_lock(&call_lock);
+	cpus = num_online_cpus() - 1;
+	if (!cpus) {
+		spin_unlock(&call_lock);
+		return 0;
+	}
+
+	/* Can deadlock when called with interrupts disabled */
+	WARN_ON(irqs_disabled());
+
+	data.func = func;
+	data.info = info;
+	atomic_set(&data.started, 0);
+	data.wait = wait;
+	if (wait)
+		atomic_set(&data.finished, 0);
+
+	call_data = &data;
+	mb();
+	
+	/* Send a message to all other CPUs and wait for them to respond */
+	send_IPI_allbutself(CALL_FUNCTION_VECTOR);
+
+	/* Wait for response */
+	while (atomic_read(&data.started) != cpus)
+		cpu_relax();
+
+	if (wait)
+		while (atomic_read(&data.finished) != cpus)
+			cpu_relax();
+	spin_unlock(&call_lock);
+
+	return 0;
+}
+EXPORT_SYMBOL(smp_call_function);
+
+static void stop_this_cpu (void * dummy)
+{
+	/*
+	 * Remove this CPU:
+	 */
+	cpu_clear(smp_processor_id(), cpu_online_map);
+	local_irq_disable();
+	disable_all_local_evtchn();
+	if (cpu_data[smp_processor_id()].hlt_works_ok)
+		for(;;) halt();
+	for (;;);
+}
+
+/*
+ * this function calls the 'stop' function on all other CPUs in the system.
+ */
+
+void smp_send_stop(void)
+{
+	smp_call_function(stop_this_cpu, NULL, 1, 0);
+
+	local_irq_disable();
+	disable_all_local_evtchn();
+	local_irq_enable();
+}
+
+/*
+ * Reschedule call back. Nothing to do,
+ * all the work is done automatically when
+ * we return from the interrupt.
+ */
+irqreturn_t smp_reschedule_interrupt(int irq, void *dev_id,
+				     struct pt_regs *regs)
+{
+
+	return IRQ_HANDLED;
+}
+
+#include <linux/kallsyms.h>
+irqreturn_t smp_call_function_interrupt(int irq, void *dev_id,
+					struct pt_regs *regs)
+{
+	void (*func) (void *info) = call_data->func;
+	void *info = call_data->info;
+	int wait = call_data->wait;
+
+	/*
+	 * Notify initiating CPU that I've grabbed the data and am
+	 * about to execute the function
+	 */
+	mb();
+	atomic_inc(&call_data->started);
+	/*
+	 * At this point the info structure may be out of scope unless wait==1
+	 */
+	irq_enter();
+	(*func)(info);
+	irq_exit();
+
+	if (wait) {
+		mb();
+		atomic_inc(&call_data->finished);
+	}
+
+	return IRQ_HANDLED;
+}
+
diff -rpuN linux-2.6.18.8/arch/i386/kernel/sysenter.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/sysenter.c
--- linux-2.6.18.8/arch/i386/kernel/sysenter.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/sysenter.c	2008-08-21 11:36:07.000000000 +0200
@@ -23,6 +23,10 @@
 #include <asm/pgtable.h>
 #include <asm/unistd.h>
 
+#ifdef CONFIG_XEN
+#include <xen/interface/callback.h>
+#endif
+
 /*
  * Should the kernel map a VDSO page into processes and pass its
  * address down to glibc upon exec()?
@@ -44,6 +48,7 @@ extern asmlinkage void sysenter_entry(vo
 
 void enable_sep_cpu(void)
 {
+#ifndef CONFIG_XEN
 	int cpu = get_cpu();
 	struct tss_struct *tss = &per_cpu(init_tss, cpu);
 
@@ -57,7 +62,36 @@ void enable_sep_cpu(void)
 	wrmsr(MSR_IA32_SYSENTER_CS, __KERNEL_CS, 0);
 	wrmsr(MSR_IA32_SYSENTER_ESP, tss->esp1, 0);
 	wrmsr(MSR_IA32_SYSENTER_EIP, (unsigned long) sysenter_entry, 0);
-	put_cpu();	
+#else
+	extern asmlinkage void sysenter_entry_pv(void);
+	static struct callback_register sysenter = {
+		.type = CALLBACKTYPE_sysenter,
+		.address = { __KERNEL_CS, (unsigned long)sysenter_entry_pv },
+	};
+
+	if (!boot_cpu_has(X86_FEATURE_SEP))
+		return;
+
+	get_cpu();
+
+	if (xen_feature(XENFEAT_supervisor_mode_kernel))
+		sysenter.address.eip = (unsigned long)sysenter_entry;
+
+	switch (HYPERVISOR_callback_op(CALLBACKOP_register, &sysenter)) {
+	case 0:
+		break;
+#if CONFIG_XEN_COMPAT < 0x030200
+	case -ENOSYS:
+		sysenter.type = CALLBACKTYPE_sysenter_deprecated;
+		if (HYPERVISOR_callback_op(CALLBACKOP_register, &sysenter) == 0)
+			break;
+#endif
+	default:
+		clear_bit(X86_FEATURE_SEP, boot_cpu_data.x86_capability);
+		break;
+	}
+#endif
+	put_cpu();
 }
 
 /*
@@ -75,11 +109,6 @@ int __init sysenter_setup(void)
 #ifdef CONFIG_COMPAT_VDSO
 	__set_fixmap(FIX_VDSO, __pa(syscall_page), PAGE_READONLY);
 	printk("Compat vDSO mapped to %08lx.\n", __fix_to_virt(FIX_VDSO));
-#else
-	/*
-	 * In the non-compat case the ELF coredumping code needs the fixmap:
-	 */
-	__set_fixmap(FIX_VDSO, __pa(syscall_page), PAGE_KERNEL_RO);
 #endif
 
 	if (!boot_cpu_has(X86_FEATURE_SEP)) {
@@ -142,6 +171,13 @@ int arch_setup_additional_pages(struct l
 	vma->vm_end = addr + PAGE_SIZE;
 	/* MAYWRITE to allow gdb to COW and set breakpoints */
 	vma->vm_flags = VM_READ|VM_EXEC|VM_MAYREAD|VM_MAYEXEC|VM_MAYWRITE;
+	/*
+	 * Make sure the vDSO gets into every core dump.
+	 * Dumping its contents makes post-mortem fully interpretable later
+	 * without matching up the same kernel and hardware config to see
+	 * what PC values meant.
+	 */
+	vma->vm_flags |= VM_ALWAYSDUMP;
 	vma->vm_flags |= mm->def_flags;
 	vma->vm_page_prot = protection_map[vma->vm_flags & 7];
 	vma->vm_ops = &syscall_vm_ops;
diff -rpuN linux-2.6.18.8/arch/i386/kernel/time-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/time-xen.c
--- linux-2.6.18.8/arch/i386/kernel/time-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/time-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1209 @@
+/*
+ *  linux/arch/i386/kernel/time.c
+ *
+ *  Copyright (C) 1991, 1992, 1995  Linus Torvalds
+ *
+ * This file contains the PC-specific time handling details:
+ * reading the RTC at bootup, etc..
+ * 1994-07-02    Alan Modra
+ *	fixed set_rtc_mmss, fixed time.year for >= 2000, new mktime
+ * 1995-03-26    Markus Kuhn
+ *      fixed 500 ms bug at call to set_rtc_mmss, fixed DS12887
+ *      precision CMOS clock update
+ * 1996-05-03    Ingo Molnar
+ *      fixed time warps in do_[slow|fast]_gettimeoffset()
+ * 1997-09-10	Updated NTP code according to technical memorandum Jan '96
+ *		"A Kernel Model for Precision Timekeeping" by Dave Mills
+ * 1998-09-05    (Various)
+ *	More robust do_fast_gettimeoffset() algorithm implemented
+ *	(works with APM, Cyrix 6x86MX and Centaur C6),
+ *	monotonic gettimeofday() with fast_get_timeoffset(),
+ *	drift-proof precision TSC calibration on boot
+ *	(C. Scott Ananian <cananian@alumni.princeton.edu>, Andrew D.
+ *	Balsa <andrebalsa@altern.org>, Philip Gladstone <philip@raptor.com>;
+ *	ported from 2.0.35 Jumbo-9 by Michael Krause <m.krause@tu-harburg.de>).
+ * 1998-12-16    Andrea Arcangeli
+ *	Fixed Jumbo-9 code in 2.1.131: do_gettimeofday was missing 1 jiffy
+ *	because was not accounting lost_ticks.
+ * 1998-12-24 Copyright (C) 1998  Andrea Arcangeli
+ *	Fixed a xtime SMP race (we need the xtime_lock rw spinlock to
+ *	serialize accesses to xtime/lost_ticks).
+ */
+
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/param.h>
+#include <linux/string.h>
+#include <linux/mm.h>
+#include <linux/interrupt.h>
+#include <linux/time.h>
+#include <linux/delay.h>
+#include <linux/init.h>
+#include <linux/smp.h>
+#include <linux/module.h>
+#include <linux/sysdev.h>
+#include <linux/bcd.h>
+#include <linux/efi.h>
+#include <linux/mca.h>
+#include <linux/sysctl.h>
+#include <linux/percpu.h>
+#include <linux/kernel_stat.h>
+#include <linux/posix-timers.h>
+#include <linux/cpufreq.h>
+
+#include <asm/io.h>
+#include <asm/smp.h>
+#include <asm/irq.h>
+#include <asm/msr.h>
+#include <asm/delay.h>
+#include <asm/mpspec.h>
+#include <asm/uaccess.h>
+#include <asm/processor.h>
+#include <asm/timer.h>
+#include <asm/sections.h>
+
+#include "mach_time.h"
+
+#include <linux/timex.h>
+
+#include <asm/hpet.h>
+
+#include <asm/arch_hooks.h>
+
+#include <xen/evtchn.h>
+#include <xen/interface/vcpu.h>
+
+#if defined (__i386__)
+#include <asm/i8259.h>
+#endif
+
+int pit_latch_buggy;              /* extern */
+
+#if defined(__x86_64__)
+unsigned long vxtime_hz = PIT_TICK_RATE;
+struct vxtime_data __vxtime __section_vxtime;   /* for vsyscalls */
+volatile unsigned long __jiffies __section_jiffies = INITIAL_JIFFIES;
+unsigned long __wall_jiffies __section_wall_jiffies = INITIAL_JIFFIES;
+struct timespec __xtime __section_xtime;
+struct timezone __sys_tz __section_sys_tz;
+#endif
+
+unsigned int cpu_khz;	/* Detected as we calibrate the TSC */
+EXPORT_SYMBOL(cpu_khz);
+
+extern unsigned long wall_jiffies;
+
+DEFINE_SPINLOCK(rtc_lock);
+EXPORT_SYMBOL(rtc_lock);
+
+extern struct init_timer_opts timer_tsc_init;
+extern struct timer_opts timer_tsc;
+#define timer_none timer_tsc
+
+/* These are peridically updated in shared_info, and then copied here. */
+struct shadow_time_info {
+	u64 tsc_timestamp;     /* TSC at last update of time vals.  */
+	u64 system_timestamp;  /* Time, in nanosecs, since boot.    */
+	u32 tsc_to_nsec_mul;
+	u32 tsc_to_usec_mul;
+	int tsc_shift;
+	u32 version;
+};
+static DEFINE_PER_CPU(struct shadow_time_info, shadow_time);
+static struct timespec shadow_tv;
+static u32 shadow_tv_version;
+
+static struct timeval monotonic_tv;
+static spinlock_t monotonic_lock = SPIN_LOCK_UNLOCKED;
+
+/* Keep track of last time we did processing/updating of jiffies and xtime. */
+static u64 processed_system_time;   /* System time (ns) at last processing. */
+static DEFINE_PER_CPU(u64, processed_system_time);
+
+/* How much CPU time was spent blocked and how much was 'stolen'? */
+static DEFINE_PER_CPU(u64, processed_stolen_time);
+static DEFINE_PER_CPU(u64, processed_blocked_time);
+
+/* Current runstate of each CPU (updated automatically by the hypervisor). */
+static DEFINE_PER_CPU(struct vcpu_runstate_info, runstate);
+
+/* Must be signed, as it's compared with s64 quantities which can be -ve. */
+#define NS_PER_TICK (1000000000LL/HZ)
+
+static void __clock_was_set(void *unused)
+{
+	clock_was_set();
+}
+static DECLARE_WORK(clock_was_set_work, __clock_was_set, NULL);
+
+/*
+ * GCC 4.3 can turn loops over an induction variable into division. We do
+ * not support arbitrary 64-bit division, and so must break the induction.
+ */
+#define clobber_induction_variable(v) asm ( "" : "+r" (v) )
+
+static inline void __normalize_time(time_t *sec, s64 *nsec)
+{
+	while (*nsec >= NSEC_PER_SEC) {
+		clobber_induction_variable(*nsec);
+		(*nsec) -= NSEC_PER_SEC;
+		(*sec)++;
+	}
+	while (*nsec < 0) {
+		clobber_induction_variable(*nsec);
+		(*nsec) += NSEC_PER_SEC;
+		(*sec)--;
+	}
+}
+
+/* Does this guest OS track Xen time, or set its wall clock independently? */
+static int independent_wallclock = 0;
+static int __init __independent_wallclock(char *str)
+{
+	independent_wallclock = 1;
+	return 1;
+}
+__setup("independent_wallclock", __independent_wallclock);
+
+/* Permitted clock jitter, in nsecs, beyond which a warning will be printed. */
+static unsigned long permitted_clock_jitter = 10000000UL; /* 10ms */
+static int __init __permitted_clock_jitter(char *str)
+{
+	permitted_clock_jitter = simple_strtoul(str, NULL, 0);
+	return 1;
+}
+__setup("permitted_clock_jitter=", __permitted_clock_jitter);
+
+#if 0
+static void delay_tsc(unsigned long loops)
+{
+	unsigned long bclock, now;
+
+	rdtscl(bclock);
+	do {
+		rep_nop();
+		rdtscl(now);
+	} while ((now - bclock) < loops);
+}
+
+struct timer_opts timer_tsc = {
+	.name = "tsc",
+	.delay = delay_tsc,
+};
+#endif
+
+/*
+ * Scale a 64-bit delta by scaling and multiplying by a 32-bit fraction,
+ * yielding a 64-bit result.
+ */
+static inline u64 scale_delta(u64 delta, u32 mul_frac, int shift)
+{
+	u64 product;
+#ifdef __i386__
+	u32 tmp1, tmp2;
+#endif
+
+	if (shift < 0)
+		delta >>= -shift;
+	else
+		delta <<= shift;
+
+#ifdef __i386__
+	__asm__ (
+		"mul  %5       ; "
+		"mov  %4,%%eax ; "
+		"mov  %%edx,%4 ; "
+		"mul  %5       ; "
+		"xor  %5,%5    ; "
+		"add  %4,%%eax ; "
+		"adc  %5,%%edx ; "
+		: "=A" (product), "=r" (tmp1), "=r" (tmp2)
+		: "a" ((u32)delta), "1" ((u32)(delta >> 32)), "2" (mul_frac) );
+#else
+	__asm__ (
+		"mul %%rdx ; shrd $32,%%rdx,%%rax"
+		: "=a" (product) : "0" (delta), "d" ((u64)mul_frac) );
+#endif
+
+	return product;
+}
+
+#if 0 /* defined (__i386__) */
+int read_current_timer(unsigned long *timer_val)
+{
+	rdtscl(*timer_val);
+	return 0;
+}
+#endif
+
+void init_cpu_khz(void)
+{
+	u64 __cpu_khz = 1000000ULL << 32;
+	struct vcpu_time_info *info = &vcpu_info(0)->time;
+	do_div(__cpu_khz, info->tsc_to_system_mul);
+	if (info->tsc_shift < 0)
+		cpu_khz = __cpu_khz << -info->tsc_shift;
+	else
+		cpu_khz = __cpu_khz >> info->tsc_shift;
+}
+
+static u64 get_nsec_offset(struct shadow_time_info *shadow)
+{
+	u64 now, delta;
+	rdtscll(now);
+	delta = now - shadow->tsc_timestamp;
+	return scale_delta(delta, shadow->tsc_to_nsec_mul, shadow->tsc_shift);
+}
+
+static unsigned long get_usec_offset(struct shadow_time_info *shadow)
+{
+	u64 now, delta;
+	rdtscll(now);
+	delta = now - shadow->tsc_timestamp;
+	return scale_delta(delta, shadow->tsc_to_usec_mul, shadow->tsc_shift);
+}
+
+static void __update_wallclock(time_t sec, long nsec)
+{
+	long wtm_nsec, xtime_nsec;
+	time_t wtm_sec, xtime_sec;
+	u64 tmp, wc_nsec;
+
+	/* Adjust wall-clock time base based on wall_jiffies ticks. */
+	wc_nsec = processed_system_time;
+	wc_nsec += sec * (u64)NSEC_PER_SEC;
+	wc_nsec += nsec;
+	wc_nsec -= (jiffies - wall_jiffies) * (u64)NS_PER_TICK;
+
+	/* Split wallclock base into seconds and nanoseconds. */
+	tmp = wc_nsec;
+	xtime_nsec = do_div(tmp, 1000000000);
+	xtime_sec  = (time_t)tmp;
+
+	wtm_sec  = wall_to_monotonic.tv_sec + (xtime.tv_sec - xtime_sec);
+	wtm_nsec = wall_to_monotonic.tv_nsec + (xtime.tv_nsec - xtime_nsec);
+
+	set_normalized_timespec(&xtime, xtime_sec, xtime_nsec);
+	set_normalized_timespec(&wall_to_monotonic, wtm_sec, wtm_nsec);
+
+	ntp_clear();
+}
+
+static void update_wallclock(void)
+{
+	shared_info_t *s = HYPERVISOR_shared_info;
+
+	do {
+		shadow_tv_version = s->wc_version;
+		rmb();
+		shadow_tv.tv_sec  = s->wc_sec;
+		shadow_tv.tv_nsec = s->wc_nsec;
+		rmb();
+	} while ((s->wc_version & 1) | (shadow_tv_version ^ s->wc_version));
+
+	if (!independent_wallclock)
+		__update_wallclock(shadow_tv.tv_sec, shadow_tv.tv_nsec);
+}
+
+/*
+ * Reads a consistent set of time-base values from Xen, into a shadow data
+ * area.
+ */
+static void get_time_values_from_xen(unsigned int cpu)
+{
+	struct vcpu_time_info   *src;
+	struct shadow_time_info *dst;
+	unsigned long flags;
+	u32 pre_version, post_version;
+
+	src = &vcpu_info(cpu)->time;
+	dst = &per_cpu(shadow_time, cpu);
+
+	local_irq_save(flags);
+
+	do {
+		pre_version = dst->version = src->version;
+		rmb();
+		dst->tsc_timestamp     = src->tsc_timestamp;
+		dst->system_timestamp  = src->system_time;
+		dst->tsc_to_nsec_mul   = src->tsc_to_system_mul;
+		dst->tsc_shift         = src->tsc_shift;
+		rmb();
+		post_version = src->version;
+	} while ((pre_version & 1) | (pre_version ^ post_version));
+
+	dst->tsc_to_usec_mul = dst->tsc_to_nsec_mul / 1000;
+
+	local_irq_restore(flags);
+}
+
+static inline int time_values_up_to_date(unsigned int cpu)
+{
+	struct vcpu_time_info   *src;
+	struct shadow_time_info *dst;
+
+	src = &vcpu_info(cpu)->time;
+	dst = &per_cpu(shadow_time, cpu);
+
+	rmb();
+	return (dst->version == src->version);
+}
+
+/*
+ * This is a special lock that is owned by the CPU and holds the index
+ * register we are working with.  It is required for NMI access to the
+ * CMOS/RTC registers.  See include/asm-i386/mc146818rtc.h for details.
+ */
+volatile unsigned long cmos_lock = 0;
+EXPORT_SYMBOL(cmos_lock);
+
+/* Routines for accessing the CMOS RAM/RTC. */
+unsigned char rtc_cmos_read(unsigned char addr)
+{
+	unsigned char val;
+	lock_cmos_prefix(addr);
+	outb_p(addr, RTC_PORT(0));
+	val = inb_p(RTC_PORT(1));
+	lock_cmos_suffix(addr);
+	return val;
+}
+EXPORT_SYMBOL(rtc_cmos_read);
+
+void rtc_cmos_write(unsigned char val, unsigned char addr)
+{
+	lock_cmos_prefix(addr);
+	outb_p(addr, RTC_PORT(0));
+	outb_p(val, RTC_PORT(1));
+	lock_cmos_suffix(addr);
+}
+EXPORT_SYMBOL(rtc_cmos_write);
+
+/*
+ * This version of gettimeofday has microsecond resolution
+ * and better than microsecond precision on fast x86 machines with TSC.
+ */
+void do_gettimeofday(struct timeval *tv)
+{
+	unsigned long seq;
+	unsigned long usec, sec;
+	unsigned long flags;
+	s64 nsec;
+	unsigned int cpu;
+	struct shadow_time_info *shadow;
+	u32 local_time_version;
+
+	cpu = get_cpu();
+	shadow = &per_cpu(shadow_time, cpu);
+
+	do {
+		unsigned long lost;
+
+		local_time_version = shadow->version;
+		seq = read_seqbegin(&xtime_lock);
+
+		usec = get_usec_offset(shadow);
+		lost = jiffies - wall_jiffies;
+
+		if (unlikely(lost))
+			usec += lost * (USEC_PER_SEC / HZ);
+
+		sec = xtime.tv_sec;
+		usec += (xtime.tv_nsec / NSEC_PER_USEC);
+
+		nsec = shadow->system_timestamp - processed_system_time;
+		__normalize_time(&sec, &nsec);
+		usec += (long)nsec / NSEC_PER_USEC;
+
+		if (unlikely(!time_values_up_to_date(cpu))) {
+			/*
+			 * We may have blocked for a long time,
+			 * rendering our calculations invalid
+			 * (e.g. the time delta may have
+			 * overflowed). Detect that and recalculate
+			 * with fresh values.
+			 */
+			get_time_values_from_xen(cpu);
+			continue;
+		}
+	} while (read_seqretry(&xtime_lock, seq) ||
+		 (local_time_version != shadow->version));
+
+	put_cpu();
+
+	while (usec >= USEC_PER_SEC) {
+		usec -= USEC_PER_SEC;
+		sec++;
+	}
+
+	spin_lock_irqsave(&monotonic_lock, flags);
+	if ((sec > monotonic_tv.tv_sec) ||
+	    ((sec == monotonic_tv.tv_sec) && (usec > monotonic_tv.tv_usec)))
+	{
+		monotonic_tv.tv_sec = sec;
+		monotonic_tv.tv_usec = usec;
+	} else {
+		sec = monotonic_tv.tv_sec;
+		usec = monotonic_tv.tv_usec;
+	}
+	spin_unlock_irqrestore(&monotonic_lock, flags);
+
+	tv->tv_sec = sec;
+	tv->tv_usec = usec;
+}
+
+EXPORT_SYMBOL(do_gettimeofday);
+
+int do_settimeofday(struct timespec *tv)
+{
+	time_t sec;
+	s64 nsec;
+	unsigned int cpu;
+	struct shadow_time_info *shadow;
+	struct xen_platform_op op;
+
+	if ((unsigned long)tv->tv_nsec >= NSEC_PER_SEC)
+		return -EINVAL;
+
+	cpu = get_cpu();
+	shadow = &per_cpu(shadow_time, cpu);
+
+	write_seqlock_irq(&xtime_lock);
+
+	/*
+	 * Ensure we don't get blocked for a long time so that our time delta
+	 * overflows. If that were to happen then our shadow time values would
+	 * be stale, so we can retry with fresh ones.
+	 */
+	for (;;) {
+		nsec = tv->tv_nsec - get_nsec_offset(shadow);
+		if (time_values_up_to_date(cpu))
+			break;
+		get_time_values_from_xen(cpu);
+	}
+	sec = tv->tv_sec;
+	__normalize_time(&sec, &nsec);
+
+	if (is_initial_xendomain() && !independent_wallclock) {
+		op.cmd = XENPF_settime;
+		op.u.settime.secs        = sec;
+		op.u.settime.nsecs       = nsec;
+		op.u.settime.system_time = shadow->system_timestamp;
+		WARN_ON(HYPERVISOR_platform_op(&op));
+		update_wallclock();
+	} else if (independent_wallclock) {
+		nsec -= shadow->system_timestamp;
+		__normalize_time(&sec, &nsec);
+		__update_wallclock(sec, nsec);
+	}
+
+	/* Reset monotonic gettimeofday() timeval. */
+	spin_lock(&monotonic_lock);
+	monotonic_tv.tv_sec = 0;
+	monotonic_tv.tv_usec = 0;
+	spin_unlock(&monotonic_lock);
+
+	write_sequnlock_irq(&xtime_lock);
+
+	put_cpu();
+
+	clock_was_set();
+	return 0;
+}
+
+EXPORT_SYMBOL(do_settimeofday);
+
+static void sync_xen_wallclock(unsigned long dummy);
+static DEFINE_TIMER(sync_xen_wallclock_timer, sync_xen_wallclock, 0, 0);
+static void sync_xen_wallclock(unsigned long dummy)
+{
+	time_t sec;
+	s64 nsec;
+	struct xen_platform_op op;
+
+	if (!ntp_synced() || independent_wallclock || !is_initial_xendomain())
+		return;
+
+	write_seqlock_irq(&xtime_lock);
+
+	sec  = xtime.tv_sec;
+	nsec = xtime.tv_nsec + ((jiffies - wall_jiffies) * (u64)NS_PER_TICK);
+	__normalize_time(&sec, &nsec);
+
+	op.cmd = XENPF_settime;
+	op.u.settime.secs        = sec;
+	op.u.settime.nsecs       = nsec;
+	op.u.settime.system_time = processed_system_time;
+	WARN_ON(HYPERVISOR_platform_op(&op));
+
+	update_wallclock();
+
+	write_sequnlock_irq(&xtime_lock);
+
+	/* Once per minute. */
+	mod_timer(&sync_xen_wallclock_timer, jiffies + 60*HZ);
+}
+
+static int set_rtc_mmss(unsigned long nowtime)
+{
+	int retval;
+	unsigned long flags;
+
+	if (independent_wallclock || !is_initial_xendomain())
+		return 0;
+
+	/* gets recalled with irq locally disabled */
+	/* XXX - does irqsave resolve this? -johnstul */
+	spin_lock_irqsave(&rtc_lock, flags);
+	if (efi_enabled)
+		retval = efi_set_rtc_mmss(nowtime);
+	else
+		retval = mach_set_rtc_mmss(nowtime);
+	spin_unlock_irqrestore(&rtc_lock, flags);
+
+	return retval;
+}
+
+/* monotonic_clock(): returns # of nanoseconds passed since time_init()
+ *		Note: This function is required to return accurate
+ *		time even in the absence of multiple timer ticks.
+ */
+unsigned long long monotonic_clock(void)
+{
+	unsigned int cpu = get_cpu();
+	struct shadow_time_info *shadow = &per_cpu(shadow_time, cpu);
+	u64 time;
+	u32 local_time_version;
+
+	do {
+		local_time_version = shadow->version;
+		barrier();
+		time = shadow->system_timestamp + get_nsec_offset(shadow);
+		if (!time_values_up_to_date(cpu))
+			get_time_values_from_xen(cpu);
+		barrier();
+	} while (local_time_version != shadow->version);
+
+	put_cpu();
+
+	return time;
+}
+EXPORT_SYMBOL(monotonic_clock);
+
+#ifdef __x86_64__
+unsigned long long sched_clock(void)
+{
+	return monotonic_clock();
+}
+#endif
+
+#if defined(CONFIG_SMP) && defined(CONFIG_FRAME_POINTER)
+unsigned long profile_pc(struct pt_regs *regs)
+{
+	unsigned long pc = instruction_pointer(regs);
+
+#ifdef __x86_64__
+	/* Assume the lock function has either no stack frame or only a single word.
+	   This checks if the address on the stack looks like a kernel text address.
+	   There is a small window for false hits, but in that case the tick
+	   is just accounted to the spinlock function.
+	   Better would be to write these functions in assembler again
+	   and check exactly. */
+	if (!user_mode_vm(regs) && in_lock_functions(pc)) {
+		char *v = *(char **)regs->rsp;
+		if ((v >= _stext && v <= _etext) ||
+			(v >= _sinittext && v <= _einittext) ||
+			(v >= (char *)MODULES_VADDR  && v <= (char *)MODULES_END))
+			return (unsigned long)v;
+		return ((unsigned long *)regs->rsp)[1];
+	}
+#else
+	if (!user_mode_vm(regs) && in_lock_functions(pc))
+		return *(unsigned long *)(regs->ebp + 4);
+#endif
+
+	return pc;
+}
+EXPORT_SYMBOL(profile_pc);
+#endif
+
+/*
+ * This is the same as the above, except we _also_ save the current
+ * Time Stamp Counter value at the time of the timer interrupt, so that
+ * we later on can estimate the time of day more exactly.
+ */
+irqreturn_t timer_interrupt(int irq, void *dev_id, struct pt_regs *regs)
+{
+	s64 delta, delta_cpu, stolen, blocked;
+	u64 sched_time;
+	unsigned int i, cpu = smp_processor_id();
+	struct shadow_time_info *shadow = &per_cpu(shadow_time, cpu);
+	struct vcpu_runstate_info *runstate = &per_cpu(runstate, cpu);
+
+	/*
+	 * Here we are in the timer irq handler. We just have irqs locally
+	 * disabled but we don't know if the timer_bh is running on the other
+	 * CPU. We need to avoid to SMP race with it. NOTE: we don' t need
+	 * the irq version of write_lock because as just said we have irq
+	 * locally disabled. -arca
+	 */
+	write_seqlock(&xtime_lock);
+
+	do {
+		get_time_values_from_xen(cpu);
+
+		/* Obtain a consistent snapshot of elapsed wallclock cycles. */
+		delta = delta_cpu =
+			shadow->system_timestamp + get_nsec_offset(shadow);
+		delta     -= processed_system_time;
+		delta_cpu -= per_cpu(processed_system_time, cpu);
+
+		/*
+		 * Obtain a consistent snapshot of stolen/blocked cycles. We
+		 * can use state_entry_time to detect if we get preempted here.
+		 */
+		do {
+			sched_time = runstate->state_entry_time;
+			barrier();
+			stolen = runstate->time[RUNSTATE_runnable] +
+				runstate->time[RUNSTATE_offline] -
+				per_cpu(processed_stolen_time, cpu);
+			blocked = runstate->time[RUNSTATE_blocked] -
+				per_cpu(processed_blocked_time, cpu);
+			barrier();
+		} while (sched_time != runstate->state_entry_time);
+	} while (!time_values_up_to_date(cpu));
+
+	if ((unlikely(delta < -(s64)permitted_clock_jitter) ||
+	     unlikely(delta_cpu < -(s64)permitted_clock_jitter))
+	    && printk_ratelimit()) {
+		printk("Timer ISR/%u: Time went backwards: "
+		       "delta=%lld delta_cpu=%lld shadow=%lld "
+		       "off=%lld processed=%lld cpu_processed=%lld\n",
+		       cpu, delta, delta_cpu, shadow->system_timestamp,
+		       (s64)get_nsec_offset(shadow),
+		       processed_system_time,
+		       per_cpu(processed_system_time, cpu));
+		for (i = 0; i < num_online_cpus(); i++)
+			printk(" %d: %lld\n", i,
+			       per_cpu(processed_system_time, i));
+	}
+
+	/* System-wide jiffy work. */
+	while (delta >= NS_PER_TICK) {
+		delta -= NS_PER_TICK;
+		processed_system_time += NS_PER_TICK;
+		do_timer(regs);
+	}
+
+	if (shadow_tv_version != HYPERVISOR_shared_info->wc_version) {
+		update_wallclock();
+		if (keventd_up())
+			schedule_work(&clock_was_set_work);
+	}
+
+	write_sequnlock(&xtime_lock);
+
+	/*
+	 * Account stolen ticks.
+	 * HACK: Passing NULL to account_steal_time()
+	 * ensures that the ticks are accounted as stolen.
+	 */
+	if ((stolen > 0) && (delta_cpu > 0)) {
+		delta_cpu -= stolen;
+		if (unlikely(delta_cpu < 0))
+			stolen += delta_cpu; /* clamp local-time progress */
+		do_div(stolen, NS_PER_TICK);
+		per_cpu(processed_stolen_time, cpu) += stolen * NS_PER_TICK;
+		per_cpu(processed_system_time, cpu) += stolen * NS_PER_TICK;
+		account_steal_time(NULL, (cputime_t)stolen);
+	}
+
+	/*
+	 * Account blocked ticks.
+	 * HACK: Passing idle_task to account_steal_time()
+	 * ensures that the ticks are accounted as idle/wait.
+	 */
+	if ((blocked > 0) && (delta_cpu > 0)) {
+		delta_cpu -= blocked;
+		if (unlikely(delta_cpu < 0))
+			blocked += delta_cpu; /* clamp local-time progress */
+		do_div(blocked, NS_PER_TICK);
+		per_cpu(processed_blocked_time, cpu) += blocked * NS_PER_TICK;
+		per_cpu(processed_system_time, cpu)  += blocked * NS_PER_TICK;
+		account_steal_time(idle_task(cpu), (cputime_t)blocked);
+	}
+
+	/* Account user/system ticks. */
+	if (delta_cpu > 0) {
+		do_div(delta_cpu, NS_PER_TICK);
+		per_cpu(processed_system_time, cpu) += delta_cpu * NS_PER_TICK;
+		if (user_mode_vm(regs))
+			account_user_time(current, (cputime_t)delta_cpu);
+		else
+			account_system_time(current, HARDIRQ_OFFSET,
+					    (cputime_t)delta_cpu);
+	}
+
+	/* Offlined for more than a few seconds? Avoid lockup warnings. */
+	if (stolen > 5*HZ)
+		touch_softlockup_watchdog();
+
+	/* Local timer processing (see update_process_times()). */
+	run_local_timers();
+	if (rcu_pending(cpu))
+		rcu_check_callbacks(cpu, user_mode_vm(regs));
+	scheduler_tick();
+	run_posix_cpu_timers(current);
+	profile_tick(CPU_PROFILING, regs);
+
+	return IRQ_HANDLED;
+}
+
+static void init_missing_ticks_accounting(unsigned int cpu)
+{
+	struct vcpu_register_runstate_memory_area area;
+	struct vcpu_runstate_info *runstate = &per_cpu(runstate, cpu);
+	int rc;
+
+	memset(runstate, 0, sizeof(*runstate));
+
+	area.addr.v = runstate;
+	rc = HYPERVISOR_vcpu_op(VCPUOP_register_runstate_memory_area, cpu, &area);
+	WARN_ON(rc && rc != -ENOSYS);
+
+	per_cpu(processed_blocked_time, cpu) =
+		runstate->time[RUNSTATE_blocked];
+	per_cpu(processed_stolen_time, cpu) =
+		runstate->time[RUNSTATE_runnable] +
+		runstate->time[RUNSTATE_offline];
+}
+
+/* not static: needed by APM */
+unsigned long get_cmos_time(void)
+{
+	unsigned long retval;
+	unsigned long flags;
+
+	spin_lock_irqsave(&rtc_lock, flags);
+
+	if (efi_enabled)
+		retval = efi_get_time();
+	else
+		retval = mach_get_cmos_time();
+
+	spin_unlock_irqrestore(&rtc_lock, flags);
+
+	return retval;
+}
+EXPORT_SYMBOL(get_cmos_time);
+
+static void sync_cmos_clock(unsigned long dummy);
+
+static DEFINE_TIMER(sync_cmos_timer, sync_cmos_clock, 0, 0);
+
+static void sync_cmos_clock(unsigned long dummy)
+{
+	struct timeval now, next;
+	int fail = 1;
+
+	/*
+	 * If we have an externally synchronized Linux clock, then update
+	 * CMOS clock accordingly every ~11 minutes. Set_rtc_mmss() has to be
+	 * called as close as possible to 500 ms before the new second starts.
+	 * This code is run on a timer.  If the clock is set, that timer
+	 * may not expire at the correct time.  Thus, we adjust...
+	 */
+	if (!ntp_synced())
+		/*
+		 * Not synced, exit, do not restart a timer (if one is
+		 * running, let it run out).
+		 */
+		return;
+
+	do_gettimeofday(&now);
+	if (now.tv_usec >= USEC_AFTER - ((unsigned) TICK_SIZE) / 2 &&
+	    now.tv_usec <= USEC_BEFORE + ((unsigned) TICK_SIZE) / 2)
+		fail = set_rtc_mmss(now.tv_sec);
+
+	next.tv_usec = USEC_AFTER - now.tv_usec;
+	if (next.tv_usec <= 0)
+		next.tv_usec += USEC_PER_SEC;
+
+	if (!fail)
+		next.tv_sec = 659;
+	else
+		next.tv_sec = 0;
+
+	if (next.tv_usec >= USEC_PER_SEC) {
+		next.tv_sec++;
+		next.tv_usec -= USEC_PER_SEC;
+	}
+	mod_timer(&sync_cmos_timer, jiffies + timeval_to_jiffies(&next));
+}
+
+void notify_arch_cmos_timer(void)
+{
+	mod_timer(&sync_cmos_timer, jiffies + 1);
+	mod_timer(&sync_xen_wallclock_timer, jiffies + 1);
+}
+
+static int timer_resume(struct sys_device *dev)
+{
+	extern void time_resume(void);
+	time_resume();
+	return 0;
+}
+
+static struct sysdev_class timer_sysclass = {
+	.resume = timer_resume,
+	set_kset_name("timer"),
+};
+
+
+/* XXX this driverfs stuff should probably go elsewhere later -john */
+static struct sys_device device_timer = {
+	.id	= 0,
+	.cls	= &timer_sysclass,
+};
+
+static int time_init_device(void)
+{
+	int error = sysdev_class_register(&timer_sysclass);
+	if (!error)
+		error = sysdev_register(&device_timer);
+	return error;
+}
+
+device_initcall(time_init_device);
+
+#ifdef CONFIG_HPET_TIMER
+extern void (*late_time_init)(void);
+/* Duplicate of time_init() below, with hpet_enable part added */
+static void __init hpet_time_init(void)
+{
+	xtime.tv_sec = get_cmos_time();
+	xtime.tv_nsec = (INITIAL_JIFFIES % HZ) * (NSEC_PER_SEC / HZ);
+	set_normalized_timespec(&wall_to_monotonic,
+		-xtime.tv_sec, -xtime.tv_nsec);
+
+	if ((hpet_enable() >= 0) && hpet_use_timer) {
+		printk("Using HPET for base-timer\n");
+	}
+
+	time_init_hook();
+}
+#endif
+
+/* Dynamically-mapped IRQ. */
+DEFINE_PER_CPU(int, timer_irq);
+
+extern void (*late_time_init)(void);
+static void setup_cpu0_timer_irq(void)
+{
+	per_cpu(timer_irq, 0) =
+		bind_virq_to_irqhandler(
+			VIRQ_TIMER,
+			0,
+			timer_interrupt,
+			SA_INTERRUPT,
+			"timer0",
+			NULL);
+	BUG_ON(per_cpu(timer_irq, 0) < 0);
+}
+
+static struct vcpu_set_periodic_timer xen_set_periodic_tick = {
+	.period_ns = NS_PER_TICK
+};
+
+void __init time_init(void)
+{
+#ifdef CONFIG_HPET_TIMER
+	if (is_hpet_capable()) {
+		/*
+		 * HPET initialization needs to do memory-mapped io. So, let
+		 * us do a late initialization after mem_init().
+		 */
+		late_time_init = hpet_time_init;
+		return;
+	}
+#endif
+
+	switch (HYPERVISOR_vcpu_op(VCPUOP_set_periodic_timer, 0,
+				   &xen_set_periodic_tick)) {
+	case 0:
+#if CONFIG_XEN_COMPAT <= 0x030004
+	case -ENOSYS:
+#endif
+		break;
+	default:
+		BUG();
+	}
+
+	get_time_values_from_xen(0);
+
+	processed_system_time = per_cpu(shadow_time, 0).system_timestamp;
+	per_cpu(processed_system_time, 0) = processed_system_time;
+	init_missing_ticks_accounting(0);
+
+	update_wallclock();
+
+	init_cpu_khz();
+	printk(KERN_INFO "Xen reported: %u.%03u MHz processor.\n",
+	       cpu_khz / 1000, cpu_khz % 1000);
+
+#if defined(__x86_64__)
+	vxtime.mode = VXTIME_TSC;
+	vxtime.quot = (1000000L << 32) / vxtime_hz;
+	vxtime.tsc_quot = (1000L << 32) / cpu_khz;
+	sync_core();
+	rdtscll(vxtime.last_tsc);
+#endif
+
+	/* Cannot request_irq() until kmem is initialised. */
+	late_time_init = setup_cpu0_timer_irq;
+}
+
+/* Convert jiffies to system time. */
+u64 jiffies_to_st(unsigned long j)
+{
+	unsigned long seq;
+	long delta;
+	u64 st;
+
+	do {
+		seq = read_seqbegin(&xtime_lock);
+		delta = j - jiffies;
+		if (delta < 1) {
+			/* Triggers in some wrap-around cases, but that's okay:
+			 * we just end up with a shorter timeout. */
+			st = processed_system_time + NS_PER_TICK;
+		} else if (((unsigned long)delta >> (BITS_PER_LONG-3)) != 0) {
+			/* Very long timeout means there is no pending timer.
+			 * We indicate this to Xen by passing zero timeout. */
+			st = 0;
+		} else {
+			st = processed_system_time + delta * (u64)NS_PER_TICK;
+		}
+	} while (read_seqretry(&xtime_lock, seq));
+
+	return st;
+}
+EXPORT_SYMBOL(jiffies_to_st);
+
+/*
+ * stop_hz_timer / start_hz_timer - enter/exit 'tickless mode' on an idle cpu
+ * These functions are based on implementations from arch/s390/kernel/time.c
+ */
+static void stop_hz_timer(void)
+{
+	struct vcpu_set_singleshot_timer singleshot;
+	unsigned int cpu = smp_processor_id();
+	unsigned long j;
+	int rc;
+
+	cpu_set(cpu, nohz_cpu_mask);
+
+	/* See matching smp_mb in rcu_start_batch in rcupdate.c.  These mbs  */
+	/* ensure that if __rcu_pending (nested in rcu_needs_cpu) fetches a  */
+	/* value of rcp->cur that matches rdp->quiescbatch and allows us to  */
+	/* stop the hz timer then the cpumasks created for subsequent values */
+	/* of cur in rcu_start_batch are guaranteed to pick up the updated   */
+	/* nohz_cpu_mask and so will not depend on this cpu.                 */
+
+	smp_mb();
+
+	/* Leave ourselves in tick mode if rcu or softirq or timer pending. */
+	if (rcu_needs_cpu(cpu) || local_softirq_pending() ||
+	    (j = next_timer_interrupt(), time_before_eq(j, jiffies))) {
+		cpu_clear(cpu, nohz_cpu_mask);
+		j = jiffies + 1;
+	}
+
+	singleshot.timeout_abs_ns = jiffies_to_st(j) + NS_PER_TICK/2;
+	singleshot.flags = 0;
+	rc = HYPERVISOR_vcpu_op(VCPUOP_set_singleshot_timer, cpu, &singleshot);
+#if CONFIG_XEN_COMPAT <= 0x030004
+	if (rc) {
+		BUG_ON(rc != -ENOSYS);
+		rc = HYPERVISOR_set_timer_op(singleshot.timeout_abs_ns);
+	}
+#endif
+	BUG_ON(rc);
+}
+
+static void start_hz_timer(void)
+{
+	cpu_clear(smp_processor_id(), nohz_cpu_mask);
+}
+
+void raw_safe_halt(void)
+{
+	stop_hz_timer();
+	/* Blocking includes an implicit local_irq_enable(). */
+	HYPERVISOR_block();
+	start_hz_timer();
+}
+EXPORT_SYMBOL(raw_safe_halt);
+
+void halt(void)
+{
+	if (irqs_disabled())
+		VOID(HYPERVISOR_vcpu_op(VCPUOP_down, smp_processor_id(), NULL));
+}
+EXPORT_SYMBOL(halt);
+
+/* No locking required. Interrupts are disabled on all CPUs. */
+void time_resume(void)
+{
+	unsigned int cpu;
+
+	init_cpu_khz();
+
+	for_each_online_cpu(cpu) {
+		switch (HYPERVISOR_vcpu_op(VCPUOP_set_periodic_timer, cpu,
+					   &xen_set_periodic_tick)) {
+		case 0:
+#if CONFIG_XEN_COMPAT <= 0x030004
+		case -ENOSYS:
+#endif
+			break;
+		default:
+			BUG();
+		}
+		get_time_values_from_xen(cpu);
+		per_cpu(processed_system_time, cpu) =
+			per_cpu(shadow_time, 0).system_timestamp;
+		init_missing_ticks_accounting(cpu);
+	}
+
+	processed_system_time = per_cpu(shadow_time, 0).system_timestamp;
+
+	update_wallclock();
+}
+
+#ifdef CONFIG_SMP
+static char timer_name[NR_CPUS][15];
+
+int __cpuinit local_setup_timer(unsigned int cpu)
+{
+	int seq, irq;
+
+	BUG_ON(cpu == 0);
+
+	switch (HYPERVISOR_vcpu_op(VCPUOP_set_periodic_timer, cpu,
+			   &xen_set_periodic_tick)) {
+	case 0:
+#if CONFIG_XEN_COMPAT <= 0x030004
+	case -ENOSYS:
+#endif
+		break;
+	default:
+		BUG();
+	}
+
+	do {
+		seq = read_seqbegin(&xtime_lock);
+		/* Use cpu0 timestamp: cpu's shadow is not initialised yet. */
+		per_cpu(processed_system_time, cpu) =
+			per_cpu(shadow_time, 0).system_timestamp;
+		init_missing_ticks_accounting(cpu);
+	} while (read_seqretry(&xtime_lock, seq));
+
+	sprintf(timer_name[cpu], "timer%u", cpu);
+	irq = bind_virq_to_irqhandler(VIRQ_TIMER,
+				      cpu,
+				      timer_interrupt,
+				      SA_INTERRUPT,
+				      timer_name[cpu],
+				      NULL);
+	if (irq < 0)
+		return irq;
+	per_cpu(timer_irq, cpu) = irq;
+
+	return 0;
+}
+
+void __cpuexit local_teardown_timer(unsigned int cpu)
+{
+	BUG_ON(cpu == 0);
+	unbind_from_irqhandler(per_cpu(timer_irq, cpu), NULL);
+}
+#endif
+
+#ifdef CONFIG_CPU_FREQ
+static int time_cpufreq_notifier(struct notifier_block *nb, unsigned long val, 
+				void *data)
+{
+	struct cpufreq_freqs *freq = data;
+	struct xen_platform_op op;
+
+	if (cpu_has(&cpu_data[freq->cpu], X86_FEATURE_CONSTANT_TSC))
+		return 0;
+
+	if (val == CPUFREQ_PRECHANGE)
+		return 0;
+
+	op.cmd = XENPF_change_freq;
+	op.u.change_freq.flags = 0;
+	op.u.change_freq.cpu = freq->cpu;
+	op.u.change_freq.freq = (u64)freq->new * 1000;
+	WARN_ON(HYPERVISOR_platform_op(&op));
+
+	return 0;
+}
+
+static struct notifier_block time_cpufreq_notifier_block = {
+	.notifier_call = time_cpufreq_notifier
+};
+
+static int __init cpufreq_time_setup(void)
+{
+	if (!cpufreq_register_notifier(&time_cpufreq_notifier_block,
+			CPUFREQ_TRANSITION_NOTIFIER)) {
+		printk(KERN_ERR "failed to set up cpufreq notifier\n");
+		return -ENODEV;
+	}
+	return 0;
+}
+
+core_initcall(cpufreq_time_setup);
+#endif
+
+/*
+ * /proc/sys/xen: This really belongs in another file. It can stay here for
+ * now however.
+ */
+static ctl_table xen_subtable[] = {
+	{
+		.ctl_name	= 1,
+		.procname	= "independent_wallclock",
+		.data		= &independent_wallclock,
+		.maxlen		= sizeof(independent_wallclock),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec
+	},
+	{
+		.ctl_name	= 2,
+		.procname	= "permitted_clock_jitter",
+		.data		= &permitted_clock_jitter,
+		.maxlen		= sizeof(permitted_clock_jitter),
+		.mode		= 0644,
+		.proc_handler	= proc_doulongvec_minmax
+	},
+	{ 0 }
+};
+static ctl_table xen_table[] = {
+	{
+		.ctl_name	= 123,
+		.procname	= "xen",
+		.mode		= 0555,
+		.child		= xen_subtable},
+	{ 0 }
+};
+static int __init xen_sysctl_init(void)
+{
+	(void)register_sysctl_table(xen_table, 0);
+	return 0;
+}
+__initcall(xen_sysctl_init);
diff -rpuN linux-2.6.18.8/arch/i386/kernel/traps.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/traps.c
--- linux-2.6.18.8/arch/i386/kernel/traps.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/traps.c	2008-08-21 11:36:07.000000000 +0200
@@ -642,18 +642,11 @@ static void mem_parity_error(unsigned ch
 
 static void io_check_error(unsigned char reason, struct pt_regs * regs)
 {
-	unsigned long i;
-
 	printk(KERN_EMERG "NMI: IOCK error (debug interrupt?)\n");
 	show_registers(regs);
 
 	/* Re-enable the IOCK line, wait for a few seconds */
-	reason = (reason & 0xf) | 8;
-	outb(reason, 0x61);
-	i = 2000;
-	while (--i) udelay(1000);
-	reason &= ~8;
-	outb(reason, 0x61);
+	clear_io_check_error(reason);
 }
 
 static void unknown_nmi_error(unsigned char reason, struct pt_regs * regs)
diff -rpuN linux-2.6.18.8/arch/i386/kernel/traps-xen.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/traps-xen.c
--- linux-2.6.18.8/arch/i386/kernel/traps-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/traps-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1190 @@
+/*
+ *  linux/arch/i386/traps.c
+ *
+ *  Copyright (C) 1991, 1992  Linus Torvalds
+ *
+ *  Pentium III FXSR, SSE support
+ *	Gareth Hughes <gareth@valinux.com>, May 2000
+ */
+
+/*
+ * 'Traps.c' handles hardware traps and faults after we have saved some
+ * state in 'asm.s'.
+ */
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/string.h>
+#include <linux/errno.h>
+#include <linux/timer.h>
+#include <linux/mm.h>
+#include <linux/init.h>
+#include <linux/delay.h>
+#include <linux/spinlock.h>
+#include <linux/interrupt.h>
+#include <linux/highmem.h>
+#include <linux/kallsyms.h>
+#include <linux/ptrace.h>
+#include <linux/utsname.h>
+#include <linux/kprobes.h>
+#include <linux/kexec.h>
+#include <linux/unwind.h>
+
+#ifdef CONFIG_EISA
+#include <linux/ioport.h>
+#include <linux/eisa.h>
+#endif
+
+#ifdef CONFIG_MCA
+#include <linux/mca.h>
+#endif
+
+#include <asm/processor.h>
+#include <asm/system.h>
+#include <asm/uaccess.h>
+#include <asm/io.h>
+#include <asm/atomic.h>
+#include <asm/debugreg.h>
+#include <asm/desc.h>
+#include <asm/i387.h>
+#include <asm/nmi.h>
+#include <asm/unwind.h>
+#include <asm/smp.h>
+#include <asm/arch_hooks.h>
+#include <asm/kdebug.h>
+
+#include <linux/module.h>
+
+#include "mach_traps.h"
+
+asmlinkage int system_call(void);
+
+struct desc_struct default_ldt[] = { { 0, 0 }, { 0, 0 }, { 0, 0 },
+		{ 0, 0 }, { 0, 0 } };
+
+/* Do we ignore FPU interrupts ? */
+char ignore_fpu_irq = 0;
+
+#ifndef CONFIG_X86_NO_IDT
+/*
+ * The IDT has to be page-aligned to simplify the Pentium
+ * F0 0F bug workaround.. We have a special link segment
+ * for this.
+ */
+struct desc_struct idt_table[256] __attribute__((__section__(".data.idt"))) = { {0, 0}, };
+#endif
+
+asmlinkage void divide_error(void);
+asmlinkage void debug(void);
+asmlinkage void nmi(void);
+asmlinkage void int3(void);
+asmlinkage void overflow(void);
+asmlinkage void bounds(void);
+asmlinkage void invalid_op(void);
+asmlinkage void device_not_available(void);
+asmlinkage void coprocessor_segment_overrun(void);
+asmlinkage void invalid_TSS(void);
+asmlinkage void segment_not_present(void);
+asmlinkage void stack_segment(void);
+asmlinkage void general_protection(void);
+asmlinkage void page_fault(void);
+asmlinkage void coprocessor_error(void);
+asmlinkage void simd_coprocessor_error(void);
+asmlinkage void alignment_check(void);
+#ifndef CONFIG_XEN
+asmlinkage void spurious_interrupt_bug(void);
+#else
+asmlinkage void fixup_4gb_segment(void);
+#endif
+asmlinkage void machine_check(void);
+
+static int kstack_depth_to_print = 24;
+#ifdef CONFIG_STACK_UNWIND
+static int call_trace = 1;
+#else
+#define call_trace (-1)
+#endif
+ATOMIC_NOTIFIER_HEAD(i386die_chain);
+
+int register_die_notifier(struct notifier_block *nb)
+{
+	vmalloc_sync_all();
+	return atomic_notifier_chain_register(&i386die_chain, nb);
+}
+EXPORT_SYMBOL(register_die_notifier); /* used modular by kdb */
+
+int unregister_die_notifier(struct notifier_block *nb)
+{
+	return atomic_notifier_chain_unregister(&i386die_chain, nb);
+}
+EXPORT_SYMBOL(unregister_die_notifier); /* used modular by kdb */
+
+static inline int valid_stack_ptr(struct thread_info *tinfo, void *p)
+{
+	return	p > (void *)tinfo &&
+		p < (void *)tinfo + THREAD_SIZE - 3;
+}
+
+/*
+ * Print one address/symbol entries per line.
+ */
+static inline void print_addr_and_symbol(unsigned long addr, char *log_lvl)
+{
+	printk(" [<%08lx>] ", addr);
+
+	print_symbol("%s\n", addr);
+}
+
+static inline unsigned long print_context_stack(struct thread_info *tinfo,
+				unsigned long *stack, unsigned long ebp,
+				char *log_lvl)
+{
+	unsigned long addr;
+
+#ifdef	CONFIG_FRAME_POINTER
+	while (valid_stack_ptr(tinfo, (void *)ebp)) {
+		addr = *(unsigned long *)(ebp + 4);
+		print_addr_and_symbol(addr, log_lvl);
+		/*
+		 * break out of recursive entries (such as
+		 * end_of_stack_stop_unwind_function):
+	 	 */
+		if (ebp == *(unsigned long *)ebp)
+			break;
+		ebp = *(unsigned long *)ebp;
+	}
+#else
+	while (valid_stack_ptr(tinfo, stack)) {
+		addr = *stack++;
+		if (__kernel_text_address(addr))
+			print_addr_and_symbol(addr, log_lvl);
+	}
+#endif
+	return ebp;
+}
+
+static asmlinkage int
+show_trace_unwind(struct unwind_frame_info *info, void *log_lvl)
+{
+	int n = 0;
+
+	while (unwind(info) == 0 && UNW_PC(info)) {
+		n++;
+		print_addr_and_symbol(UNW_PC(info), log_lvl);
+		if (arch_unw_user_mode(info))
+			break;
+	}
+	return n;
+}
+
+static void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,
+			       unsigned long *stack, char *log_lvl)
+{
+	unsigned long ebp;
+
+	if (!task)
+		task = current;
+
+	if (call_trace >= 0) {
+		int unw_ret = 0;
+		struct unwind_frame_info info;
+
+		if (regs) {
+			if (unwind_init_frame_info(&info, task, regs) == 0)
+				unw_ret = show_trace_unwind(&info, log_lvl);
+		} else if (task == current)
+			unw_ret = unwind_init_running(&info, show_trace_unwind, log_lvl);
+		else {
+			if (unwind_init_blocked(&info, task) == 0)
+				unw_ret = show_trace_unwind(&info, log_lvl);
+		}
+		if (unw_ret > 0) {
+			if (call_trace == 1 && !arch_unw_user_mode(&info)) {
+				print_symbol("DWARF2 unwinder stuck at %s\n",
+					     UNW_PC(&info));
+				if (UNW_SP(&info) >= PAGE_OFFSET) {
+					printk("Leftover inexact backtrace:\n");
+					stack = (void *)UNW_SP(&info);
+				} else
+					printk("Full inexact backtrace again:\n");
+			} else if (call_trace >= 1)
+				return;
+			else
+				printk("Full inexact backtrace again:\n");
+		} else
+			printk("Inexact backtrace:\n");
+	}
+
+	if (task == current) {
+		/* Grab ebp right from our regs */
+		asm ("movl %%ebp, %0" : "=r" (ebp) : );
+	} else {
+		/* ebp is the last reg pushed by switch_to */
+		ebp = *(unsigned long *) task->thread.esp;
+	}
+
+	while (1) {
+		struct thread_info *context;
+		context = (struct thread_info *)
+			((unsigned long)stack & (~(THREAD_SIZE - 1)));
+		ebp = print_context_stack(context, stack, ebp, log_lvl);
+		stack = (unsigned long*)context->previous_esp;
+		if (!stack)
+			break;
+		printk("%s =======================\n", log_lvl);
+	}
+}
+
+void show_trace(struct task_struct *task, struct pt_regs *regs, unsigned long * stack)
+{
+	show_trace_log_lvl(task, regs, stack, "");
+}
+
+static void show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs,
+			       unsigned long *esp, char *log_lvl)
+{
+	unsigned long *stack;
+	int i;
+
+	if (esp == NULL) {
+		if (task)
+			esp = (unsigned long*)task->thread.esp;
+		else
+			esp = (unsigned long *)&esp;
+	}
+
+	stack = esp;
+	for(i = 0; i < kstack_depth_to_print; i++) {
+		if (kstack_end(stack))
+			break;
+		if (i && ((i % 8) == 0))
+			printk("\n%s       ", log_lvl);
+		printk("%08lx ", *stack++);
+	}
+	printk("\n%sCall Trace:\n", log_lvl);
+	show_trace_log_lvl(task, regs, esp, log_lvl);
+}
+
+void show_stack(struct task_struct *task, unsigned long *esp)
+{
+	printk("       ");
+	show_stack_log_lvl(task, NULL, esp, "");
+}
+
+/*
+ * The architecture-independent dump_stack generator
+ */
+void dump_stack(void)
+{
+	unsigned long stack;
+
+	show_trace(current, NULL, &stack);
+}
+
+EXPORT_SYMBOL(dump_stack);
+
+void show_registers(struct pt_regs *regs)
+{
+	int i;
+	int in_kernel = 1;
+	unsigned long esp;
+	unsigned short ss;
+
+	esp = (unsigned long) (&regs->esp);
+	savesegment(ss, ss);
+	if (user_mode_vm(regs)) {
+		in_kernel = 0;
+		esp = regs->esp;
+		ss = regs->xss & 0xffff;
+	}
+	print_modules();
+	printk(KERN_EMERG "CPU:    %d\nEIP:    %04x:[<%08lx>]    %s VLI\n"
+			"EFLAGS: %08lx   (%s %.*s) \n",
+		smp_processor_id(), 0xffff & regs->xcs, regs->eip,
+		print_tainted(), regs->eflags, system_utsname.release,
+		(int)strcspn(system_utsname.version, " "),
+		system_utsname.version);
+	print_symbol(KERN_EMERG "EIP is at %s\n", regs->eip);
+	printk(KERN_EMERG "eax: %08lx   ebx: %08lx   ecx: %08lx   edx: %08lx\n",
+		regs->eax, regs->ebx, regs->ecx, regs->edx);
+	printk(KERN_EMERG "esi: %08lx   edi: %08lx   ebp: %08lx   esp: %08lx\n",
+		regs->esi, regs->edi, regs->ebp, esp);
+	printk(KERN_EMERG "ds: %04x   es: %04x   ss: %04x\n",
+		regs->xds & 0xffff, regs->xes & 0xffff, ss);
+	printk(KERN_EMERG "Process %.*s (pid: %d, ti=%p task=%p task.ti=%p)",
+		TASK_COMM_LEN, current->comm, current->pid,
+		current_thread_info(), current, current->thread_info);
+	/*
+	 * When in-kernel, we also print out the stack and code at the
+	 * time of the fault..
+	 */
+	if (in_kernel) {
+		u8 __user *eip;
+
+		printk("\n" KERN_EMERG "Stack: ");
+		show_stack_log_lvl(NULL, regs, (unsigned long *)esp, KERN_EMERG);
+
+		printk(KERN_EMERG "Code: ");
+
+		eip = (u8 __user *)regs->eip - 43;
+		for (i = 0; i < 64; i++, eip++) {
+			unsigned char c;
+
+			if (eip < (u8 __user *)PAGE_OFFSET || __get_user(c, eip)) {
+				printk(" Bad EIP value.");
+				break;
+			}
+			if (eip == (u8 __user *)regs->eip)
+				printk("<%02x> ", c);
+			else
+				printk("%02x ", c);
+		}
+	}
+	printk("\n");
+}	
+
+static void handle_BUG(struct pt_regs *regs)
+{
+	unsigned long eip = regs->eip;
+	unsigned short ud2;
+
+	if (eip < PAGE_OFFSET)
+		return;
+	if (__get_user(ud2, (unsigned short __user *)eip))
+		return;
+	if (ud2 != 0x0b0f)
+		return;
+
+	printk(KERN_EMERG "------------[ cut here ]------------\n");
+
+#ifdef CONFIG_DEBUG_BUGVERBOSE
+	do {
+		unsigned short line;
+		char *file;
+		char c;
+
+		if (__get_user(line, (unsigned short __user *)(eip + 2)))
+			break;
+		if (__get_user(file, (char * __user *)(eip + 4)) ||
+		    (unsigned long)file < PAGE_OFFSET || __get_user(c, file))
+			file = "<bad filename>";
+
+		printk(KERN_EMERG "kernel BUG at %s:%d!\n", file, line);
+		return;
+	} while (0);
+#endif
+	printk(KERN_EMERG "Kernel BUG at [verbose debug info unavailable]\n");
+}
+
+/* This is gone through when something in the kernel
+ * has done something bad and is about to be terminated.
+*/
+void die(const char * str, struct pt_regs * regs, long err)
+{
+	static struct {
+		spinlock_t lock;
+		u32 lock_owner;
+		int lock_owner_depth;
+	} die = {
+		.lock =			SPIN_LOCK_UNLOCKED,
+		.lock_owner =		-1,
+		.lock_owner_depth =	0
+	};
+	static int die_counter;
+	unsigned long flags;
+
+	oops_enter();
+
+	if (die.lock_owner != raw_smp_processor_id()) {
+		console_verbose();
+		spin_lock_irqsave(&die.lock, flags);
+		die.lock_owner = smp_processor_id();
+		die.lock_owner_depth = 0;
+		bust_spinlocks(1);
+	}
+	else
+		local_save_flags(flags);
+
+	if (++die.lock_owner_depth < 3) {
+		int nl = 0;
+		unsigned long esp;
+		unsigned short ss;
+
+		handle_BUG(regs);
+		printk(KERN_EMERG "%s: %04lx [#%d]\n", str, err & 0xffff, ++die_counter);
+#ifdef CONFIG_PREEMPT
+		printk(KERN_EMERG "PREEMPT ");
+		nl = 1;
+#endif
+#ifdef CONFIG_SMP
+		if (!nl)
+			printk(KERN_EMERG);
+		printk("SMP ");
+		nl = 1;
+#endif
+#ifdef CONFIG_DEBUG_PAGEALLOC
+		if (!nl)
+			printk(KERN_EMERG);
+		printk("DEBUG_PAGEALLOC");
+		nl = 1;
+#endif
+		if (nl)
+			printk("\n");
+		if (notify_die(DIE_OOPS, str, regs, err,
+					current->thread.trap_no, SIGSEGV) !=
+				NOTIFY_STOP) {
+			show_registers(regs);
+			/* Executive summary in case the oops scrolled away */
+			esp = (unsigned long) (&regs->esp);
+			savesegment(ss, ss);
+			if (user_mode(regs)) {
+				esp = regs->esp;
+				ss = regs->xss & 0xffff;
+			}
+			printk(KERN_EMERG "EIP: [<%08lx>] ", regs->eip);
+			print_symbol("%s", regs->eip);
+			printk(" SS:ESP %04x:%08lx\n", ss, esp);
+		}
+		else
+			regs = NULL;
+  	} else
+		printk(KERN_EMERG "Recursive die() failure, output suppressed\n");
+
+	bust_spinlocks(0);
+	die.lock_owner = -1;
+	spin_unlock_irqrestore(&die.lock, flags);
+
+	if (!regs)
+		return;
+
+	if (kexec_should_crash(current))
+		crash_kexec(regs);
+
+	if (in_interrupt())
+		panic("Fatal exception in interrupt");
+
+	if (panic_on_oops)
+		panic("Fatal exception");
+
+	oops_exit();
+	do_exit(SIGSEGV);
+}
+
+static inline void die_if_kernel(const char * str, struct pt_regs * regs, long err)
+{
+	if (!user_mode_vm(regs))
+		die(str, regs, err);
+}
+
+static void __kprobes do_trap(int trapnr, int signr, char *str, int vm86,
+			      struct pt_regs * regs, long error_code,
+			      siginfo_t *info)
+{
+	struct task_struct *tsk = current;
+	tsk->thread.error_code = error_code;
+	tsk->thread.trap_no = trapnr;
+
+	if (regs->eflags & VM_MASK) {
+		if (vm86)
+			goto vm86_trap;
+		goto trap_signal;
+	}
+
+	if (!user_mode(regs))
+		goto kernel_trap;
+
+	trap_signal: {
+		if (info)
+			force_sig_info(signr, info, tsk);
+		else
+			force_sig(signr, tsk);
+		return;
+	}
+
+	kernel_trap: {
+		if (!fixup_exception(regs))
+			die(str, regs, error_code);
+		return;
+	}
+
+	vm86_trap: {
+		int ret = handle_vm86_trap((struct kernel_vm86_regs *) regs, error_code, trapnr);
+		if (ret) goto trap_signal;
+		return;
+	}
+}
+
+#define DO_ERROR(trapnr, signr, str, name) \
+fastcall void do_##name(struct pt_regs * regs, long error_code) \
+{ \
+	if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) \
+						== NOTIFY_STOP) \
+		return; \
+	do_trap(trapnr, signr, str, 0, regs, error_code, NULL); \
+}
+
+#define DO_ERROR_INFO(trapnr, signr, str, name, sicode, siaddr) \
+fastcall void do_##name(struct pt_regs * regs, long error_code) \
+{ \
+	siginfo_t info; \
+	info.si_signo = signr; \
+	info.si_errno = 0; \
+	info.si_code = sicode; \
+	info.si_addr = (void __user *)siaddr; \
+	if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) \
+						== NOTIFY_STOP) \
+		return; \
+	do_trap(trapnr, signr, str, 0, regs, error_code, &info); \
+}
+
+#define DO_VM86_ERROR(trapnr, signr, str, name) \
+fastcall void do_##name(struct pt_regs * regs, long error_code) \
+{ \
+	if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) \
+						== NOTIFY_STOP) \
+		return; \
+	do_trap(trapnr, signr, str, 1, regs, error_code, NULL); \
+}
+
+#define DO_VM86_ERROR_INFO(trapnr, signr, str, name, sicode, siaddr) \
+fastcall void do_##name(struct pt_regs * regs, long error_code) \
+{ \
+	siginfo_t info; \
+	info.si_signo = signr; \
+	info.si_errno = 0; \
+	info.si_code = sicode; \
+	info.si_addr = (void __user *)siaddr; \
+	if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) \
+						== NOTIFY_STOP) \
+		return; \
+	do_trap(trapnr, signr, str, 1, regs, error_code, &info); \
+}
+
+DO_VM86_ERROR_INFO( 0, SIGFPE,  "divide error", divide_error, FPE_INTDIV, regs->eip)
+#ifndef CONFIG_KPROBES
+DO_VM86_ERROR( 3, SIGTRAP, "int3", int3)
+#endif
+DO_VM86_ERROR( 4, SIGSEGV, "overflow", overflow)
+DO_VM86_ERROR( 5, SIGSEGV, "bounds", bounds)
+DO_ERROR_INFO( 6, SIGILL,  "invalid opcode", invalid_op, ILL_ILLOPN, regs->eip)
+DO_ERROR( 9, SIGFPE,  "coprocessor segment overrun", coprocessor_segment_overrun)
+DO_ERROR(10, SIGSEGV, "invalid TSS", invalid_TSS)
+DO_ERROR(11, SIGBUS,  "segment not present", segment_not_present)
+DO_ERROR(12, SIGBUS,  "stack segment", stack_segment)
+DO_ERROR_INFO(17, SIGBUS, "alignment check", alignment_check, BUS_ADRALN, 0)
+DO_ERROR_INFO(32, SIGSEGV, "iret exception", iret_error, ILL_BADSTK, 0)
+
+fastcall void __kprobes do_general_protection(struct pt_regs * regs,
+					      long error_code)
+{
+	current->thread.error_code = error_code;
+	current->thread.trap_no = 13;
+
+	if (regs->eflags & VM_MASK)
+		goto gp_in_vm86;
+
+	if (!user_mode(regs))
+		goto gp_in_kernel;
+
+	current->thread.error_code = error_code;
+	current->thread.trap_no = 13;
+	force_sig(SIGSEGV, current);
+	return;
+
+gp_in_vm86:
+	local_irq_enable();
+	handle_vm86_fault((struct kernel_vm86_regs *) regs, error_code);
+	return;
+
+gp_in_kernel:
+	if (!fixup_exception(regs)) {
+		if (notify_die(DIE_GPF, "general protection fault", regs,
+				error_code, 13, SIGSEGV) == NOTIFY_STOP)
+			return;
+		die("general protection fault", regs, error_code);
+	}
+}
+
+static void mem_parity_error(unsigned char reason, struct pt_regs * regs)
+{
+	printk(KERN_EMERG "Uhhuh. NMI received. Dazed and confused, but trying "
+			"to continue\n");
+	printk(KERN_EMERG "You probably have a hardware problem with your RAM "
+			"chips\n");
+
+	/* Clear and disable the memory parity error line. */
+	clear_mem_error(reason);
+}
+
+static void io_check_error(unsigned char reason, struct pt_regs * regs)
+{
+	printk(KERN_EMERG "NMI: IOCK error (debug interrupt?)\n");
+	show_registers(regs);
+
+	/* Re-enable the IOCK line, wait for a few seconds */
+	clear_io_check_error(reason);
+}
+
+static void unknown_nmi_error(unsigned char reason, struct pt_regs * regs)
+{
+#ifdef CONFIG_MCA
+	/* Might actually be able to figure out what the guilty party
+	* is. */
+	if( MCA_bus ) {
+		mca_handle_nmi();
+		return;
+	}
+#endif
+	printk("Uhhuh. NMI received for unknown reason %02x on CPU %d.\n",
+		reason, smp_processor_id());
+	printk("Dazed and confused, but trying to continue\n");
+	printk("Do you have a strange power saving mode enabled?\n");
+}
+
+static DEFINE_SPINLOCK(nmi_print_lock);
+
+void die_nmi (struct pt_regs *regs, const char *msg)
+{
+	if (notify_die(DIE_NMIWATCHDOG, msg, regs, 0, 2, SIGINT) ==
+	    NOTIFY_STOP)
+		return;
+
+	spin_lock(&nmi_print_lock);
+	/*
+	* We are in trouble anyway, lets at least try
+	* to get a message out.
+	*/
+	bust_spinlocks(1);
+	printk(KERN_EMERG "%s", msg);
+	printk(" on CPU%d, eip %08lx, registers:\n",
+		smp_processor_id(), regs->eip);
+	show_registers(regs);
+	printk(KERN_EMERG "console shuts up ...\n");
+	console_silent();
+	spin_unlock(&nmi_print_lock);
+	bust_spinlocks(0);
+
+	/* If we are in kernel we are probably nested up pretty bad
+	 * and might aswell get out now while we still can.
+	*/
+	if (!user_mode_vm(regs)) {
+		current->thread.trap_no = 2;
+		crash_kexec(regs);
+	}
+
+	do_exit(SIGSEGV);
+}
+
+static void default_do_nmi(struct pt_regs * regs)
+{
+	unsigned char reason = 0;
+
+	/* Only the BSP gets external NMIs from the system.  */
+	if (!smp_processor_id())
+		reason = get_nmi_reason();
+ 
+	if (!(reason & 0xc0)) {
+		if (notify_die(DIE_NMI_IPI, "nmi_ipi", regs, reason, 2, SIGINT)
+							== NOTIFY_STOP)
+			return;
+#ifdef CONFIG_X86_LOCAL_APIC
+		/*
+		 * Ok, so this is none of the documented NMI sources,
+		 * so it must be the NMI watchdog.
+		 */
+		if (nmi_watchdog) {
+			nmi_watchdog_tick(regs);
+			return;
+		}
+#endif
+		unknown_nmi_error(reason, regs);
+		return;
+	}
+	if (notify_die(DIE_NMI, "nmi", regs, reason, 2, SIGINT) == NOTIFY_STOP)
+		return;
+	if (reason & 0x80)
+		mem_parity_error(reason, regs);
+	if (reason & 0x40)
+		io_check_error(reason, regs);
+	/*
+	 * Reassert NMI in case it became active meanwhile
+	 * as it's edge-triggered.
+	 */
+	reassert_nmi();
+}
+
+static int dummy_nmi_callback(struct pt_regs * regs, int cpu)
+{
+	return 0;
+}
+ 
+static nmi_callback_t nmi_callback = dummy_nmi_callback;
+ 
+fastcall void do_nmi(struct pt_regs * regs, long error_code)
+{
+	int cpu;
+
+	nmi_enter();
+
+	cpu = smp_processor_id();
+
+	++nmi_count(cpu);
+
+	if (!rcu_dereference(nmi_callback)(regs, cpu))
+		default_do_nmi(regs);
+
+	nmi_exit();
+}
+
+void set_nmi_callback(nmi_callback_t callback)
+{
+	vmalloc_sync_all();
+	rcu_assign_pointer(nmi_callback, callback);
+}
+EXPORT_SYMBOL_GPL(set_nmi_callback);
+
+void unset_nmi_callback(void)
+{
+	nmi_callback = dummy_nmi_callback;
+}
+EXPORT_SYMBOL_GPL(unset_nmi_callback);
+
+#ifdef CONFIG_KPROBES
+fastcall void __kprobes do_int3(struct pt_regs *regs, long error_code)
+{
+	if (notify_die(DIE_INT3, "int3", regs, error_code, 3, SIGTRAP)
+			== NOTIFY_STOP)
+		return;
+	/* This is an interrupt gate, because kprobes wants interrupts
+	disabled.  Normal trap handlers don't. */
+	restore_interrupts(regs);
+	do_trap(3, SIGTRAP, "int3", 1, regs, error_code, NULL);
+}
+#endif
+
+/*
+ * Our handling of the processor debug registers is non-trivial.
+ * We do not clear them on entry and exit from the kernel. Therefore
+ * it is possible to get a watchpoint trap here from inside the kernel.
+ * However, the code in ./ptrace.c has ensured that the user can
+ * only set watchpoints on userspace addresses. Therefore the in-kernel
+ * watchpoint trap can only occur in code which is reading/writing
+ * from user space. Such code must not hold kernel locks (since it
+ * can equally take a page fault), therefore it is safe to call
+ * force_sig_info even though that claims and releases locks.
+ * 
+ * Code in ./signal.c ensures that the debug control register
+ * is restored before we deliver any signal, and therefore that
+ * user code runs with the correct debug control register even though
+ * we clear it here.
+ *
+ * Being careful here means that we don't have to be as careful in a
+ * lot of more complicated places (task switching can be a bit lazy
+ * about restoring all the debug state, and ptrace doesn't have to
+ * find every occurrence of the TF bit that could be saved away even
+ * by user code)
+ */
+fastcall void __kprobes do_debug(struct pt_regs * regs, long error_code)
+{
+	unsigned int condition;
+	struct task_struct *tsk = current;
+
+	get_debugreg(condition, 6);
+
+	if (notify_die(DIE_DEBUG, "debug", regs, condition, error_code,
+					SIGTRAP) == NOTIFY_STOP)
+		return;
+	/* It's safe to allow irq's after DR6 has been saved */
+	if (regs->eflags & X86_EFLAGS_IF)
+		local_irq_enable();
+
+	/* Mask out spurious debug traps due to lazy DR7 setting */
+	if (condition & (DR_TRAP0|DR_TRAP1|DR_TRAP2|DR_TRAP3)) {
+		if (!tsk->thread.debugreg[7])
+			goto clear_dr7;
+	}
+
+	if (regs->eflags & VM_MASK)
+		goto debug_vm86;
+
+	/* Save debug status register where ptrace can see it */
+	tsk->thread.debugreg[6] = condition;
+
+	/*
+	 * Single-stepping through TF: make sure we ignore any events in
+	 * kernel space (but re-enable TF when returning to user mode).
+	 */
+	if (condition & DR_STEP) {
+		/*
+		 * We already checked v86 mode above, so we can
+		 * check for kernel mode by just checking the CPL
+		 * of CS.
+		 */
+		if (!user_mode(regs))
+			goto clear_TF_reenable;
+	}
+
+	/* Ok, finally something we can handle */
+	send_sigtrap(tsk, regs, error_code);
+
+	/* Disable additional traps. They'll be re-enabled when
+	 * the signal is delivered.
+	 */
+clear_dr7:
+	set_debugreg(0, 7);
+	return;
+
+debug_vm86:
+	handle_vm86_trap((struct kernel_vm86_regs *) regs, error_code, 1);
+	return;
+
+clear_TF_reenable:
+	set_tsk_thread_flag(tsk, TIF_SINGLESTEP);
+	regs->eflags &= ~TF_MASK;
+	return;
+}
+
+/*
+ * Note that we play around with the 'TS' bit in an attempt to get
+ * the correct behaviour even in the presence of the asynchronous
+ * IRQ13 behaviour
+ */
+void math_error(void __user *eip)
+{
+	struct task_struct * task;
+	siginfo_t info;
+	unsigned short cwd, swd;
+
+	/*
+	 * Save the info for the exception handler and clear the error.
+	 */
+	task = current;
+	save_init_fpu(task);
+	task->thread.trap_no = 16;
+	task->thread.error_code = 0;
+	info.si_signo = SIGFPE;
+	info.si_errno = 0;
+	info.si_code = __SI_FAULT;
+	info.si_addr = eip;
+	/*
+	 * (~cwd & swd) will mask out exceptions that are not set to unmasked
+	 * status.  0x3f is the exception bits in these regs, 0x200 is the
+	 * C1 reg you need in case of a stack fault, 0x040 is the stack
+	 * fault bit.  We should only be taking one exception at a time,
+	 * so if this combination doesn't produce any single exception,
+	 * then we have a bad program that isn't syncronizing its FPU usage
+	 * and it will suffer the consequences since we won't be able to
+	 * fully reproduce the context of the exception
+	 */
+	cwd = get_fpu_cwd(task);
+	swd = get_fpu_swd(task);
+	switch (swd & ~cwd & 0x3f) {
+		case 0x000: /* No unmasked exception */
+			return;
+		default:    /* Multiple exceptions */
+			break;
+		case 0x001: /* Invalid Op */
+			/*
+			 * swd & 0x240 == 0x040: Stack Underflow
+			 * swd & 0x240 == 0x240: Stack Overflow
+			 * User must clear the SF bit (0x40) if set
+			 */
+			info.si_code = FPE_FLTINV;
+			break;
+		case 0x002: /* Denormalize */
+		case 0x010: /* Underflow */
+			info.si_code = FPE_FLTUND;
+			break;
+		case 0x004: /* Zero Divide */
+			info.si_code = FPE_FLTDIV;
+			break;
+		case 0x008: /* Overflow */
+			info.si_code = FPE_FLTOVF;
+			break;
+		case 0x020: /* Precision */
+			info.si_code = FPE_FLTRES;
+			break;
+	}
+	force_sig_info(SIGFPE, &info, task);
+}
+
+fastcall void do_coprocessor_error(struct pt_regs * regs, long error_code)
+{
+	ignore_fpu_irq = 1;
+	math_error((void __user *)regs->eip);
+}
+
+static void simd_math_error(void __user *eip)
+{
+	struct task_struct * task;
+	siginfo_t info;
+	unsigned short mxcsr;
+
+	/*
+	 * Save the info for the exception handler and clear the error.
+	 */
+	task = current;
+	save_init_fpu(task);
+	task->thread.trap_no = 19;
+	task->thread.error_code = 0;
+	info.si_signo = SIGFPE;
+	info.si_errno = 0;
+	info.si_code = __SI_FAULT;
+	info.si_addr = eip;
+	/*
+	 * The SIMD FPU exceptions are handled a little differently, as there
+	 * is only a single status/control register.  Thus, to determine which
+	 * unmasked exception was caught we must mask the exception mask bits
+	 * at 0x1f80, and then use these to mask the exception bits at 0x3f.
+	 */
+	mxcsr = get_fpu_mxcsr(task);
+	switch (~((mxcsr & 0x1f80) >> 7) & (mxcsr & 0x3f)) {
+		case 0x000:
+		default:
+			break;
+		case 0x001: /* Invalid Op */
+			info.si_code = FPE_FLTINV;
+			break;
+		case 0x002: /* Denormalize */
+		case 0x010: /* Underflow */
+			info.si_code = FPE_FLTUND;
+			break;
+		case 0x004: /* Zero Divide */
+			info.si_code = FPE_FLTDIV;
+			break;
+		case 0x008: /* Overflow */
+			info.si_code = FPE_FLTOVF;
+			break;
+		case 0x020: /* Precision */
+			info.si_code = FPE_FLTRES;
+			break;
+	}
+	force_sig_info(SIGFPE, &info, task);
+}
+
+fastcall void do_simd_coprocessor_error(struct pt_regs * regs,
+					  long error_code)
+{
+	if (cpu_has_xmm) {
+		/* Handle SIMD FPU exceptions on PIII+ processors. */
+		ignore_fpu_irq = 1;
+		simd_math_error((void __user *)regs->eip);
+	} else {
+		/*
+		 * Handle strange cache flush from user space exception
+		 * in all other cases.  This is undocumented behaviour.
+		 */
+		if (regs->eflags & VM_MASK) {
+			handle_vm86_fault((struct kernel_vm86_regs *)regs,
+					  error_code);
+			return;
+		}
+		current->thread.trap_no = 19;
+		current->thread.error_code = error_code;
+		die_if_kernel("cache flush denied", regs, error_code);
+		force_sig(SIGSEGV, current);
+	}
+}
+
+#ifndef CONFIG_XEN
+fastcall void do_spurious_interrupt_bug(struct pt_regs * regs,
+					  long error_code)
+{
+#if 0
+	/* No need to warn about this any longer. */
+	printk("Ignoring P6 Local APIC Spurious Interrupt Bug...\n");
+#endif
+}
+
+fastcall void setup_x86_bogus_stack(unsigned char * stk)
+{
+	unsigned long *switch16_ptr, *switch32_ptr;
+	struct pt_regs *regs;
+	unsigned long stack_top, stack_bot;
+	unsigned short iret_frame16_off;
+	int cpu = smp_processor_id();
+	/* reserve the space on 32bit stack for the magic switch16 pointer */
+	memmove(stk, stk + 8, sizeof(struct pt_regs));
+	switch16_ptr = (unsigned long *)(stk + sizeof(struct pt_regs));
+	regs = (struct pt_regs *)stk;
+	/* now the switch32 on 16bit stack */
+	stack_bot = (unsigned long)&per_cpu(cpu_16bit_stack, cpu);
+	stack_top = stack_bot +	CPU_16BIT_STACK_SIZE;
+	switch32_ptr = (unsigned long *)(stack_top - 8);
+	iret_frame16_off = CPU_16BIT_STACK_SIZE - 8 - 20;
+	/* copy iret frame on 16bit stack */
+	memcpy((void *)(stack_bot + iret_frame16_off), &regs->eip, 20);
+	/* fill in the switch pointers */
+	switch16_ptr[0] = (regs->esp & 0xffff0000) | iret_frame16_off;
+	switch16_ptr[1] = __ESPFIX_SS;
+	switch32_ptr[0] = (unsigned long)stk + sizeof(struct pt_regs) +
+		8 - CPU_16BIT_STACK_SIZE;
+	switch32_ptr[1] = __KERNEL_DS;
+}
+
+fastcall unsigned char * fixup_x86_bogus_stack(unsigned short sp)
+{
+	unsigned long *switch32_ptr;
+	unsigned char *stack16, *stack32;
+	unsigned long stack_top, stack_bot;
+	int len;
+	int cpu = smp_processor_id();
+	stack_bot = (unsigned long)&per_cpu(cpu_16bit_stack, cpu);
+	stack_top = stack_bot +	CPU_16BIT_STACK_SIZE;
+	switch32_ptr = (unsigned long *)(stack_top - 8);
+	/* copy the data from 16bit stack to 32bit stack */
+	len = CPU_16BIT_STACK_SIZE - 8 - sp;
+	stack16 = (unsigned char *)(stack_bot + sp);
+	stack32 = (unsigned char *)
+		(switch32_ptr[0] + CPU_16BIT_STACK_SIZE - 8 - len);
+	memcpy(stack32, stack16, len);
+	return stack32;
+}
+#endif
+
+/*
+ *  'math_state_restore()' saves the current math information in the
+ * old math state array, and gets the new ones from the current task
+ *
+ * Careful.. There are problems with IBM-designed IRQ13 behaviour.
+ * Don't touch unless you *really* know how it works.
+ *
+ * Must be called with kernel preemption disabled (in this case,
+ * local interrupts are disabled at the call-site in entry.S).
+ */
+asmlinkage void math_state_restore(struct pt_regs regs)
+{
+	struct thread_info *thread = current_thread_info();
+	struct task_struct *tsk = thread->task;
+
+	/* NB. 'clts' is done for us by Xen during virtual trap. */
+	if (!tsk_used_math(tsk))
+		init_fpu(tsk);
+	restore_fpu(tsk);
+	thread->status |= TS_USEDFPU;	/* So we fnsave on switch_to() */
+}
+
+#ifndef CONFIG_MATH_EMULATION
+
+asmlinkage void math_emulate(long arg)
+{
+	printk(KERN_EMERG "math-emulation not enabled and no coprocessor found.\n");
+	printk(KERN_EMERG "killing %s.\n",current->comm);
+	force_sig(SIGFPE,current);
+	schedule();
+}
+
+#endif /* CONFIG_MATH_EMULATION */
+
+#ifdef CONFIG_X86_F00F_BUG
+void __init trap_init_f00f_bug(void)
+{
+	__set_fixmap(FIX_F00F_IDT, __pa(&idt_table), PAGE_KERNEL_RO);
+
+	/*
+	 * Update the IDT descriptor and reload the IDT so that
+	 * it uses the read-only mapped virtual address.
+	 */
+	idt_descr.address = fix_to_virt(FIX_F00F_IDT);
+	load_idt(&idt_descr);
+}
+#endif
+
+
+/*
+ * NB. All these are "trap gates" (i.e. events_mask isn't set) except
+ * for those that specify <dpl>|4 in the second field.
+ */
+static trap_info_t __cpuinitdata trap_table[] = {
+	{  0, 0, __KERNEL_CS, (unsigned long)divide_error		},
+	{  1, 0|4, __KERNEL_CS, (unsigned long)debug			},
+	{  3, 3|4, __KERNEL_CS, (unsigned long)int3			},
+	{  4, 3, __KERNEL_CS, (unsigned long)overflow			},
+	{  5, 0, __KERNEL_CS, (unsigned long)bounds			},
+	{  6, 0, __KERNEL_CS, (unsigned long)invalid_op			},
+	{  7, 0|4, __KERNEL_CS, (unsigned long)device_not_available	},
+	{  9, 0, __KERNEL_CS, (unsigned long)coprocessor_segment_overrun },
+	{ 10, 0, __KERNEL_CS, (unsigned long)invalid_TSS		},
+	{ 11, 0, __KERNEL_CS, (unsigned long)segment_not_present	},
+	{ 12, 0, __KERNEL_CS, (unsigned long)stack_segment		},
+	{ 13, 0, __KERNEL_CS, (unsigned long)general_protection		},
+	{ 14, 0|4, __KERNEL_CS, (unsigned long)page_fault		},
+	{ 15, 0, __KERNEL_CS, (unsigned long)fixup_4gb_segment		},
+	{ 16, 0, __KERNEL_CS, (unsigned long)coprocessor_error		},
+	{ 17, 0, __KERNEL_CS, (unsigned long)alignment_check		},
+#ifdef CONFIG_X86_MCE
+	{ 18, 0, __KERNEL_CS, (unsigned long)machine_check		},
+#endif
+	{ 19, 0, __KERNEL_CS, (unsigned long)simd_coprocessor_error	},
+	{ SYSCALL_VECTOR,  3, __KERNEL_CS, (unsigned long)system_call	},
+	{  0, 0,	   0, 0						}
+};
+
+void __init trap_init(void)
+{
+	int ret;
+
+	ret = HYPERVISOR_set_trap_table(trap_table);
+	if (ret)
+		printk("HYPERVISOR_set_trap_table failed: error %d\n", ret);
+
+	if (cpu_has_fxsr) {
+		/*
+		 * Verify that the FXSAVE/FXRSTOR data will be 16-byte aligned.
+		 * Generates a compile-time "error: zero width for bit-field" if
+		 * the alignment is wrong.
+		 */
+		struct fxsrAlignAssert {
+			int _:!(offsetof(struct task_struct,
+					thread.i387.fxsave) & 15);
+		};
+
+		printk(KERN_INFO "Enabling fast FPU save and restore... ");
+		set_in_cr4(X86_CR4_OSFXSR);
+		printk("done.\n");
+	}
+	if (cpu_has_xmm) {
+		printk(KERN_INFO "Enabling unmasked SIMD FPU exception "
+				"support... ");
+		set_in_cr4(X86_CR4_OSXMMEXCPT);
+		printk("done.\n");
+	}
+
+	/*
+	 * Should be a barrier for any external CPU state.
+	 */
+	cpu_init();
+}
+
+void __cpuinit smp_trap_init(trap_info_t *trap_ctxt)
+{
+	const trap_info_t *t = trap_table;
+
+	for (t = trap_table; t->address; t++) {
+		trap_ctxt[t->vector].flags = t->flags;
+		trap_ctxt[t->vector].cs = t->cs;
+		trap_ctxt[t->vector].address = t->address;
+	}
+}
+
+static int __init kstack_setup(char *s)
+{
+	kstack_depth_to_print = simple_strtoul(s, NULL, 0);
+	return 1;
+}
+__setup("kstack=", kstack_setup);
+
+#ifdef CONFIG_STACK_UNWIND
+static int __init call_trace_setup(char *s)
+{
+	if (strcmp(s, "old") == 0)
+		call_trace = -1;
+	else if (strcmp(s, "both") == 0)
+		call_trace = 0;
+	else if (strcmp(s, "newfallback") == 0)
+		call_trace = 1;
+	else if (strcmp(s, "new") == 2)
+		call_trace = 2;
+	return 1;
+}
+__setup("call_trace=", call_trace_setup);
+#endif
diff -rpuN linux-2.6.18.8/arch/i386/kernel/vm86.c linux-2.6.18-xen-3.3.0/arch/i386/kernel/vm86.c
--- linux-2.6.18.8/arch/i386/kernel/vm86.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/vm86.c	2008-08-21 11:36:07.000000000 +0200
@@ -97,7 +97,9 @@
 struct pt_regs * FASTCALL(save_v86_state(struct kernel_vm86_regs * regs));
 struct pt_regs * fastcall save_v86_state(struct kernel_vm86_regs * regs)
 {
+#ifndef CONFIG_X86_NO_TSS
 	struct tss_struct *tss;
+#endif
 	struct pt_regs *ret;
 	unsigned long tmp;
 
@@ -122,12 +124,16 @@ struct pt_regs * fastcall save_v86_state
 		do_exit(SIGSEGV);
 	}
 
+#ifndef CONFIG_X86_NO_TSS
 	tss = &per_cpu(init_tss, get_cpu());
+#endif
 	current->thread.esp0 = current->thread.saved_esp0;
 	current->thread.sysenter_cs = __KERNEL_CS;
 	load_esp0(tss, &current->thread);
 	current->thread.saved_esp0 = 0;
+#ifndef CONFIG_X86_NO_TSS
 	put_cpu();
+#endif
 
 	loadsegment(fs, current->thread.saved_fs);
 	loadsegment(gs, current->thread.saved_gs);
@@ -251,7 +257,9 @@ out:
 
 static void do_sys_vm86(struct kernel_vm86_struct *info, struct task_struct *tsk)
 {
+#ifndef CONFIG_X86_NO_TSS
 	struct tss_struct *tss;
+#endif
 	long eax;
 /*
  * make sure the vm86() system call doesn't try to do anything silly
@@ -296,12 +304,16 @@ static void do_sys_vm86(struct kernel_vm
 	savesegment(fs, tsk->thread.saved_fs);
 	savesegment(gs, tsk->thread.saved_gs);
 
+#ifndef CONFIG_X86_NO_TSS
 	tss = &per_cpu(init_tss, get_cpu());
+#endif
 	tsk->thread.esp0 = (unsigned long) &info->VM86_TSS_ESP0;
 	if (cpu_has_sep)
 		tsk->thread.sysenter_cs = 0;
 	load_esp0(tss, &tsk->thread);
+#ifndef CONFIG_X86_NO_TSS
 	put_cpu();
+#endif
 
 	tsk->thread.screen_bitmap = info->screen_bitmap;
 	if (info->flags & VM86_SCREEN_BITMAP)
diff -rpuN linux-2.6.18.8/arch/i386/kernel/vmlinux.lds.S linux-2.6.18-xen-3.3.0/arch/i386/kernel/vmlinux.lds.S
--- linux-2.6.18.8/arch/i386/kernel/vmlinux.lds.S	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/vmlinux.lds.S	2008-08-21 11:36:07.000000000 +0200
@@ -13,6 +13,12 @@ OUTPUT_FORMAT("elf32-i386", "elf32-i386"
 OUTPUT_ARCH(i386)
 ENTRY(phys_startup_32)
 jiffies = jiffies_64;
+
+PHDRS {
+	text PT_LOAD FLAGS(5);	/* R_E */
+	data PT_LOAD FLAGS(7);	/* RWE */
+	note PT_NOTE FLAGS(4);	/* R__ */
+}
 SECTIONS
 {
   . = __KERNEL_START;
@@ -26,7 +32,7 @@ SECTIONS
 	KPROBES_TEXT
 	*(.fixup)
 	*(.gnu.warning)
-	} = 0x9090
+	} :text = 0x9090
 
   _etext = .;			/* End of text section */
 
@@ -45,10 +51,11 @@ SECTIONS
   __tracedata_end = .;
 
   /* writeable */
+  . = ALIGN(4096);
   .data : AT(ADDR(.data) - LOAD_OFFSET) {	/* Data */
 	*(.data)
 	CONSTRUCTORS
-	}
+	} :data
 
   . = ALIGN(4096);
   __nosave_begin = .;
@@ -184,4 +191,6 @@ SECTIONS
   STABS_DEBUG
 
   DWARF_DEBUG
+
+  NOTES
 }
diff -rpuN linux-2.6.18.8/arch/i386/kernel/vsyscall-note-xen.S linux-2.6.18-xen-3.3.0/arch/i386/kernel/vsyscall-note-xen.S
--- linux-2.6.18.8/arch/i386/kernel/vsyscall-note-xen.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/kernel/vsyscall-note-xen.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,32 @@
+/*
+ * This supplies .note.* sections to go into the PT_NOTE inside the vDSO text.
+ * Here we can supply some information useful to userland.
+ * First we get the vanilla i386 note that supplies the kernel version info.
+ */
+
+#include "vsyscall-note.S"
+
+/*
+ * Now we add a special note telling glibc's dynamic linker a fake hardware
+ * flavor that it will use to choose the search path for libraries in the
+ * same way it uses real hardware capabilities like "mmx".
+ * We supply "nosegneg" as the fake capability, to indicate that we
+ * do not like negative offsets in instructions using segment overrides,
+ * since we implement those inefficiently.  This makes it possible to
+ * install libraries optimized to avoid those access patterns in someplace
+ * like /lib/i686/tls/nosegneg.  Note that an /etc/ld.so.conf.d/file
+ * corresponding to the bits here is needed to make ldconfig work right.
+ * It should contain:
+ *	hwcap 0 nosegneg
+ * to match the mapping of bit to name that we give here.
+ */
+#define NOTE_KERNELCAP_BEGIN(ncaps, mask) \
+	ASM_ELF_NOTE_BEGIN(".note.kernelcap", "a", "GNU", 2) \
+	.long ncaps, mask
+#define NOTE_KERNELCAP(bit, name) \
+	.byte bit; .asciz name
+#define NOTE_KERNELCAP_END ASM_ELF_NOTE_END
+
+NOTE_KERNELCAP_BEGIN(1, 1)
+NOTE_KERNELCAP(0, "nosegneg")
+NOTE_KERNELCAP_END
diff -rpuN linux-2.6.18.8/arch/i386/lib/Makefile linux-2.6.18-xen-3.3.0/arch/i386/lib/Makefile
--- linux-2.6.18.8/arch/i386/lib/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/lib/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -7,3 +7,4 @@ lib-y = checksum.o delay.o usercopy.o ge
 	bitops.o
 
 lib-$(CONFIG_X86_USE_3DNOW) += mmx.o
+lib-$(CONFIG_XEN_SCRUB_PAGES) += scrub.o
diff -rpuN linux-2.6.18.8/arch/i386/lib/scrub.c linux-2.6.18-xen-3.3.0/arch/i386/lib/scrub.c
--- linux-2.6.18.8/arch/i386/lib/scrub.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/lib/scrub.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,21 @@
+#include <asm/cpufeature.h>
+#include <asm/page.h>
+#include <asm/processor.h>
+
+void scrub_pages(void *v, unsigned int count)
+{
+	if (likely(cpu_has_xmm2)) {
+		unsigned long n = count * (PAGE_SIZE / sizeof(long) / 4);
+
+		for (; n--; v += sizeof(long) * 4)
+			asm("movnti %1,(%0)\n\t"
+			    "movnti %1,%c2(%0)\n\t"
+			    "movnti %1,2*%c2(%0)\n\t"
+			    "movnti %1,3*%c2(%0)\n\t"
+			    : : "r" (v), "r" (0L), "i" (sizeof(long))
+			    : "memory");
+		asm volatile("sfence" : : : "memory");
+	} else
+		for (; count--; v += PAGE_SIZE)
+			clear_page(v);
+}
diff -rpuN linux-2.6.18.8/arch/i386/mach-xen/Makefile linux-2.6.18-xen-3.3.0/arch/i386/mach-xen/Makefile
--- linux-2.6.18.8/arch/i386/mach-xen/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/mach-xen/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,5 @@
+#
+# Makefile for the linux kernel.
+#
+
+obj-y				:= setup.o
diff -rpuN linux-2.6.18.8/arch/i386/mach-xen/setup.c linux-2.6.18-xen-3.3.0/arch/i386/mach-xen/setup.c
--- linux-2.6.18.8/arch/i386/mach-xen/setup.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/mach-xen/setup.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,158 @@
+/*
+ *	Machine specific setup for generic
+ */
+
+#include <linux/mm.h>
+#include <linux/smp.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/module.h>
+#include <asm/acpi.h>
+#include <asm/arch_hooks.h>
+#include <asm/e820.h>
+#include <asm/setup.h>
+#include <asm/fixmap.h>
+
+#include <xen/interface/callback.h>
+#include <xen/interface/memory.h>
+
+#ifdef CONFIG_HOTPLUG_CPU
+#define DEFAULT_SEND_IPI	(1)
+#else
+#define DEFAULT_SEND_IPI	(0)
+#endif
+
+int no_broadcast=DEFAULT_SEND_IPI;
+
+static __init int no_ipi_broadcast(char *str)
+{
+	get_option(&str, &no_broadcast);
+	printk ("Using %s mode\n", no_broadcast ? "No IPI Broadcast" :
+											"IPI Broadcast");
+	return 1;
+}
+
+__setup("no_ipi_broadcast", no_ipi_broadcast);
+
+static int __init print_ipi_mode(void)
+{
+	printk ("Using IPI %s mode\n", no_broadcast ? "No-Shortcut" :
+											"Shortcut");
+	return 0;
+}
+
+late_initcall(print_ipi_mode);
+
+/**
+ * machine_specific_memory_setup - Hook for machine specific memory setup.
+ *
+ * Description:
+ *	This is included late in kernel/setup.c so that it can make
+ *	use of all of the static functions.
+ **/
+
+char * __init machine_specific_memory_setup(void)
+{
+	int rc;
+	struct xen_memory_map memmap;
+	/*
+	 * This is rather large for a stack variable but this early in
+	 * the boot process we know we have plenty slack space.
+	 */
+	struct e820entry map[E820MAX];
+
+	memmap.nr_entries = E820MAX;
+	set_xen_guest_handle(memmap.buffer, map);
+
+	rc = HYPERVISOR_memory_op(XENMEM_memory_map, &memmap);
+	if ( rc == -ENOSYS ) {
+		memmap.nr_entries = 1;
+		map[0].addr = 0ULL;
+		map[0].size = PFN_PHYS((unsigned long long)xen_start_info->nr_pages);
+		/* 8MB slack (to balance backend allocations). */
+		map[0].size += 8ULL << 20;
+		map[0].type = E820_RAM;
+		rc = 0;
+	}
+	BUG_ON(rc);
+
+	sanitize_e820_map(map, (char *)&memmap.nr_entries);
+
+	BUG_ON(copy_e820_map(map, (char)memmap.nr_entries) < 0);
+
+	return "Xen";
+}
+
+
+extern void hypervisor_callback(void);
+extern void failsafe_callback(void);
+extern void nmi(void);
+
+unsigned long *machine_to_phys_mapping = (void *)MACH2PHYS_VIRT_START;
+EXPORT_SYMBOL(machine_to_phys_mapping);
+unsigned int machine_to_phys_order;
+EXPORT_SYMBOL(machine_to_phys_order);
+
+void __init pre_setup_arch_hook(void)
+{
+	struct xen_machphys_mapping mapping;
+	unsigned long machine_to_phys_nr_ents;
+	struct xen_platform_parameters pp;
+
+	init_mm.pgd = swapper_pg_dir = (pgd_t *)xen_start_info->pt_base;
+
+	setup_xen_features();
+
+	if (HYPERVISOR_xen_version(XENVER_platform_parameters, &pp) == 0)
+		set_fixaddr_top(pp.virt_start);
+
+	if (HYPERVISOR_memory_op(XENMEM_machphys_mapping, &mapping) == 0) {
+		machine_to_phys_mapping = (unsigned long *)mapping.v_start;
+		machine_to_phys_nr_ents = mapping.max_mfn + 1;
+	} else
+		machine_to_phys_nr_ents = MACH2PHYS_NR_ENTRIES;
+	machine_to_phys_order = fls(machine_to_phys_nr_ents - 1);
+
+	if (!xen_feature(XENFEAT_auto_translated_physmap))
+		phys_to_machine_mapping =
+			(unsigned long *)xen_start_info->mfn_list;
+}
+
+void __init machine_specific_arch_setup(void)
+{
+	int ret;
+	static struct callback_register __initdata event = {
+		.type = CALLBACKTYPE_event,
+		.address = { __KERNEL_CS, (unsigned long)hypervisor_callback },
+	};
+	static struct callback_register __initdata failsafe = {
+		.type = CALLBACKTYPE_failsafe,
+		.address = { __KERNEL_CS, (unsigned long)failsafe_callback },
+	};
+	static struct callback_register __initdata nmi_cb = {
+		.type = CALLBACKTYPE_nmi,
+		.address = { __KERNEL_CS, (unsigned long)nmi },
+	};
+
+	ret = HYPERVISOR_callback_op(CALLBACKOP_register, &event);
+	if (ret == 0)
+		ret = HYPERVISOR_callback_op(CALLBACKOP_register, &failsafe);
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (ret == -ENOSYS)
+		ret = HYPERVISOR_set_callbacks(
+			event.address.cs, event.address.eip,
+			failsafe.address.cs, failsafe.address.eip);
+#endif
+	BUG_ON(ret);
+
+	ret = HYPERVISOR_callback_op(CALLBACKOP_register, &nmi_cb);
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (ret == -ENOSYS) {
+		static struct xennmi_callback __initdata cb = {
+			.handler_address = (unsigned long)nmi
+		};
+
+		HYPERVISOR_nmi_op(XENNMI_register_callback, &cb);
+	}
+#endif
+}
diff -rpuN linux-2.6.18.8/arch/i386/Makefile linux-2.6.18-xen-3.3.0/arch/i386/Makefile
--- linux-2.6.18.8/arch/i386/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -71,6 +71,10 @@ mcore-$(CONFIG_X86_BIGSMP)	:= mach-defau
 mflags-$(CONFIG_X86_SUMMIT) := -Iinclude/asm-i386/mach-summit
 mcore-$(CONFIG_X86_SUMMIT)  := mach-default
 
+# Xen subarch support
+mflags-$(CONFIG_X86_XEN)	:= -Iinclude/asm-i386/mach-xen
+mcore-$(CONFIG_X86_XEN)		:= mach-xen
+
 # generic subarchitecture
 mflags-$(CONFIG_X86_GENERICARCH) := -Iinclude/asm-i386/mach-generic
 mcore-$(CONFIG_X86_GENERICARCH) := mach-default
@@ -102,9 +106,20 @@ AFLAGS += $(mflags-y)
 
 boot := arch/i386/boot
 
-PHONY += zImage bzImage compressed zlilo bzlilo \
+PHONY += zImage bzImage vmlinuz compressed zlilo bzlilo \
          zdisk bzdisk fdimage fdimage144 fdimage288 isoimage install
 
+ifdef CONFIG_XEN
+CPPFLAGS := -D__XEN_INTERFACE_VERSION__=$(CONFIG_XEN_INTERFACE_VERSION) \
+	-Iinclude$(if $(KBUILD_SRC),2)/asm/mach-xen $(CPPFLAGS)
+all: vmlinuz
+
+# KBUILD_IMAGE specifies the target image being built
+KBUILD_IMAGE := $(boot)/vmlinuz
+
+vmlinuz: vmlinux
+	$(Q)$(MAKE) $(build)=$(boot) $(KBUILD_IMAGE)
+else
 all: bzImage
 
 # KBUILD_IMAGE specify target image being built
@@ -124,6 +139,7 @@ zdisk bzdisk: vmlinux
 
 fdimage fdimage144 fdimage288 isoimage: vmlinux
 	$(Q)$(MAKE) $(build)=$(boot) BOOTIMAGE=$(KBUILD_IMAGE) $@
+endif
 
 install:
 	$(Q)$(MAKE) $(build)=$(boot) BOOTIMAGE=$(KBUILD_IMAGE) install
diff -rpuN linux-2.6.18.8/arch/i386/mm/fault-xen.c linux-2.6.18-xen-3.3.0/arch/i386/mm/fault-xen.c
--- linux-2.6.18.8/arch/i386/mm/fault-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/mm/fault-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,779 @@
+/*
+ *  linux/arch/i386/mm/fault.c
+ *
+ *  Copyright (C) 1995  Linus Torvalds
+ */
+
+#include <linux/signal.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/ptrace.h>
+#include <linux/mman.h>
+#include <linux/mm.h>
+#include <linux/smp.h>
+#include <linux/smp_lock.h>
+#include <linux/interrupt.h>
+#include <linux/init.h>
+#include <linux/tty.h>
+#include <linux/vt_kern.h>		/* For unblank_screen() */
+#include <linux/highmem.h>
+#include <linux/module.h>
+#include <linux/kprobes.h>
+
+#include <asm/system.h>
+#include <asm/uaccess.h>
+#include <asm/desc.h>
+#include <asm/kdebug.h>
+
+extern void die(const char *,struct pt_regs *,long);
+
+#ifdef CONFIG_KPROBES
+ATOMIC_NOTIFIER_HEAD(notify_page_fault_chain);
+int register_page_fault_notifier(struct notifier_block *nb)
+{
+	vmalloc_sync_all();
+	return atomic_notifier_chain_register(&notify_page_fault_chain, nb);
+}
+
+int unregister_page_fault_notifier(struct notifier_block *nb)
+{
+	return atomic_notifier_chain_unregister(&notify_page_fault_chain, nb);
+}
+
+static inline int notify_page_fault(enum die_val val, const char *str,
+			struct pt_regs *regs, long err, int trap, int sig)
+{
+	struct die_args args = {
+		.regs = regs,
+		.str = str,
+		.err = err,
+		.trapnr = trap,
+		.signr = sig
+	};
+	return atomic_notifier_call_chain(&notify_page_fault_chain, val, &args);
+}
+#else
+static inline int notify_page_fault(enum die_val val, const char *str,
+			struct pt_regs *regs, long err, int trap, int sig)
+{
+	return NOTIFY_DONE;
+}
+#endif
+
+
+/*
+ * Unlock any spinlocks which will prevent us from getting the
+ * message out 
+ */
+void bust_spinlocks(int yes)
+{
+	int loglevel_save = console_loglevel;
+
+	if (yes) {
+		oops_in_progress = 1;
+		return;
+	}
+#ifdef CONFIG_VT
+	unblank_screen();
+#endif
+	oops_in_progress = 0;
+	/*
+	 * OK, the message is on the console.  Now we call printk()
+	 * without oops_in_progress set so that printk will give klogd
+	 * a poke.  Hold onto your hats...
+	 */
+	console_loglevel = 15;		/* NMI oopser may have shut the console up */
+	printk(" ");
+	console_loglevel = loglevel_save;
+}
+
+/*
+ * Return EIP plus the CS segment base.  The segment limit is also
+ * adjusted, clamped to the kernel/user address space (whichever is
+ * appropriate), and returned in *eip_limit.
+ *
+ * The segment is checked, because it might have been changed by another
+ * task between the original faulting instruction and here.
+ *
+ * If CS is no longer a valid code segment, or if EIP is beyond the
+ * limit, or if it is a kernel address when CS is not a kernel segment,
+ * then the returned value will be greater than *eip_limit.
+ * 
+ * This is slow, but is very rarely executed.
+ */
+static inline unsigned long get_segment_eip(struct pt_regs *regs,
+					    unsigned long *eip_limit)
+{
+	unsigned long eip = regs->eip;
+	unsigned seg = regs->xcs & 0xffff;
+	u32 seg_ar, seg_limit, base, *desc;
+
+	/* Unlikely, but must come before segment checks. */
+	if (unlikely(regs->eflags & VM_MASK)) {
+		base = seg << 4;
+		*eip_limit = base + 0xffff;
+		return base + (eip & 0xffff);
+	}
+
+	/* The standard kernel/user address space limit. */
+	*eip_limit = (seg & 2) ? USER_DS.seg : KERNEL_DS.seg;
+	
+	/* By far the most common cases. */
+	if (likely(seg == __USER_CS || seg == GET_KERNEL_CS()))
+		return eip;
+
+	/* Check the segment exists, is within the current LDT/GDT size,
+	   that kernel/user (ring 0..3) has the appropriate privilege,
+	   that it's a code segment, and get the limit. */
+	__asm__ ("larl %3,%0; lsll %3,%1"
+		 : "=&r" (seg_ar), "=r" (seg_limit) : "0" (0), "rm" (seg));
+	if ((~seg_ar & 0x9800) || eip > seg_limit) {
+		*eip_limit = 0;
+		return 1;	 /* So that returned eip > *eip_limit. */
+	}
+
+	/* Get the GDT/LDT descriptor base. 
+	   When you look for races in this code remember that
+	   LDT and other horrors are only used in user space. */
+	if (seg & (1<<2)) {
+		/* Must lock the LDT while reading it. */
+		down(&current->mm->context.sem);
+		desc = current->mm->context.ldt;
+		desc = (void *)desc + (seg & ~7);
+	} else {
+		/* Must disable preemption while reading the GDT. */
+ 		desc = (u32 *)get_cpu_gdt_table(get_cpu());
+		desc = (void *)desc + (seg & ~7);
+	}
+
+	/* Decode the code segment base from the descriptor */
+	base = get_desc_base((unsigned long *)desc);
+
+	if (seg & (1<<2)) { 
+		up(&current->mm->context.sem);
+	} else
+		put_cpu();
+
+	/* Adjust EIP and segment limit, and clamp at the kernel limit.
+	   It's legitimate for segments to wrap at 0xffffffff. */
+	seg_limit += base;
+	if (seg_limit < *eip_limit && seg_limit >= base)
+		*eip_limit = seg_limit;
+	return eip + base;
+}
+
+/* 
+ * Sometimes AMD Athlon/Opteron CPUs report invalid exceptions on prefetch.
+ * Check that here and ignore it.
+ */
+static int __is_prefetch(struct pt_regs *regs, unsigned long addr)
+{ 
+	unsigned long limit;
+	unsigned long instr = get_segment_eip (regs, &limit);
+	int scan_more = 1;
+	int prefetch = 0; 
+	int i;
+
+	for (i = 0; scan_more && i < 15; i++) { 
+		unsigned char opcode;
+		unsigned char instr_hi;
+		unsigned char instr_lo;
+
+		if (instr > limit)
+			break;
+		if (__get_user(opcode, (unsigned char __user *) instr))
+			break; 
+
+		instr_hi = opcode & 0xf0; 
+		instr_lo = opcode & 0x0f; 
+		instr++;
+
+		switch (instr_hi) { 
+		case 0x20:
+		case 0x30:
+			/* Values 0x26,0x2E,0x36,0x3E are valid x86 prefixes. */
+			scan_more = ((instr_lo & 7) == 0x6);
+			break;
+			
+		case 0x60:
+			/* 0x64 thru 0x67 are valid prefixes in all modes. */
+			scan_more = (instr_lo & 0xC) == 0x4;
+			break;		
+		case 0xF0:
+			/* 0xF0, 0xF2, and 0xF3 are valid prefixes */
+			scan_more = !instr_lo || (instr_lo>>1) == 1;
+			break;			
+		case 0x00:
+			/* Prefetch instruction is 0x0F0D or 0x0F18 */
+			scan_more = 0;
+			if (instr > limit)
+				break;
+			if (__get_user(opcode, (unsigned char __user *) instr))
+				break;
+			prefetch = (instr_lo == 0xF) &&
+				(opcode == 0x0D || opcode == 0x18);
+			break;			
+		default:
+			scan_more = 0;
+			break;
+		} 
+	}
+	return prefetch;
+}
+
+static inline int is_prefetch(struct pt_regs *regs, unsigned long addr,
+			      unsigned long error_code)
+{
+	if (unlikely(boot_cpu_data.x86_vendor == X86_VENDOR_AMD &&
+		     boot_cpu_data.x86 >= 6)) {
+		/* Catch an obscure case of prefetch inside an NX page. */
+		if (nx_enabled && (error_code & 16))
+			return 0;
+		return __is_prefetch(regs, addr);
+	}
+	return 0;
+} 
+
+static noinline void force_sig_info_fault(int si_signo, int si_code,
+	unsigned long address, struct task_struct *tsk)
+{
+	siginfo_t info;
+
+	info.si_signo = si_signo;
+	info.si_errno = 0;
+	info.si_code = si_code;
+	info.si_addr = (void __user *)address;
+	force_sig_info(si_signo, &info, tsk);
+}
+
+fastcall void do_invalid_op(struct pt_regs *, unsigned long);
+
+#ifdef CONFIG_X86_PAE
+static void dump_fault_path(unsigned long address)
+{
+	unsigned long *p, page;
+	unsigned long mfn; 
+
+	page = read_cr3();
+	p  = (unsigned long *)__va(page);
+	p += (address >> 30) * 2;
+	printk(KERN_ALERT "%08lx -> *pde = %08lx:%08lx\n", page, p[1], p[0]);
+	if (p[0] & _PAGE_PRESENT) {
+		mfn  = (p[0] >> PAGE_SHIFT) | (p[1] << 20);
+		page = mfn_to_pfn(mfn) << PAGE_SHIFT; 
+		p  = (unsigned long *)__va(page);
+		address &= 0x3fffffff;
+		p += (address >> 21) * 2;
+		printk(KERN_ALERT "%08lx -> *pme = %08lx:%08lx\n", 
+		       page, p[1], p[0]);
+		mfn  = (p[0] >> PAGE_SHIFT) | (p[1] << 20);
+#ifdef CONFIG_HIGHPTE
+		if (mfn_to_pfn(mfn) >= highstart_pfn)
+			return;
+#endif
+		if (p[0] & _PAGE_PRESENT) {
+			page = mfn_to_pfn(mfn) << PAGE_SHIFT; 
+			p  = (unsigned long *) __va(page);
+			address &= 0x001fffff;
+			p += (address >> 12) * 2;
+			printk(KERN_ALERT "%08lx -> *pte = %08lx:%08lx\n",
+			       page, p[1], p[0]);
+		}
+	}
+}
+#else
+static void dump_fault_path(unsigned long address)
+{
+	unsigned long page;
+
+	page = read_cr3();
+	page = ((unsigned long *) __va(page))[address >> 22];
+	if (oops_may_print())
+		printk(KERN_ALERT "*pde = ma %08lx pa %08lx\n", page,
+		       machine_to_phys(page));
+	/*
+	 * We must not directly access the pte in the highpte
+	 * case if the page table is located in highmem.
+	 * And lets rather not kmap-atomic the pte, just in case
+	 * it's allocated already.
+	 */
+#ifdef CONFIG_HIGHPTE
+	if ((page >> PAGE_SHIFT) >= highstart_pfn)
+		return;
+#endif
+	if ((page & 1) && oops_may_print()) {
+		page &= PAGE_MASK;
+		address &= 0x003ff000;
+		page = machine_to_phys(page);
+		page = ((unsigned long *) __va(page))[address >> PAGE_SHIFT];
+		printk(KERN_ALERT "*pte = ma %08lx pa %08lx\n", page,
+		       machine_to_phys(page));
+	}
+}
+#endif
+
+static int spurious_fault(struct pt_regs *regs,
+			  unsigned long address,
+			  unsigned long error_code)
+{
+	pgd_t *pgd;
+	pud_t *pud;
+	pmd_t *pmd;
+	pte_t *pte;
+
+	/* Reserved-bit violation or user access to kernel space? */
+	if (error_code & 0x0c)
+		return 0;
+
+	pgd = init_mm.pgd + pgd_index(address);
+	if (!pgd_present(*pgd))
+		return 0;
+
+	pud = pud_offset(pgd, address);
+	if (!pud_present(*pud))
+		return 0;
+
+	pmd = pmd_offset(pud, address);
+	if (!pmd_present(*pmd))
+		return 0;
+
+	pte = pte_offset_kernel(pmd, address);
+	if (!pte_present(*pte))
+		return 0;
+	if ((error_code & 0x02) && !pte_write(*pte))
+		return 0;
+#ifdef CONFIG_X86_PAE
+	if ((error_code & 0x10) && (__pte_val(*pte) & _PAGE_NX))
+		return 0;
+#endif
+
+	return 1;
+}
+
+static inline pmd_t *vmalloc_sync_one(pgd_t *pgd, unsigned long address)
+{
+	unsigned index = pgd_index(address);
+	pgd_t *pgd_k;
+	pud_t *pud, *pud_k;
+	pmd_t *pmd, *pmd_k;
+
+	pgd += index;
+	pgd_k = init_mm.pgd + index;
+
+	if (!pgd_present(*pgd_k))
+		return NULL;
+
+	/*
+	 * set_pgd(pgd, *pgd_k); here would be useless on PAE
+	 * and redundant with the set_pmd() on non-PAE. As would
+	 * set_pud.
+	 */
+
+	pud = pud_offset(pgd, address);
+	pud_k = pud_offset(pgd_k, address);
+	if (!pud_present(*pud_k))
+		return NULL;
+
+	pmd = pmd_offset(pud, address);
+	pmd_k = pmd_offset(pud_k, address);
+	if (!pmd_present(*pmd_k))
+		return NULL;
+	if (!pmd_present(*pmd))
+#if CONFIG_XEN_COMPAT > 0x030002
+		set_pmd(pmd, *pmd_k);
+#else
+		/*
+		 * When running on older Xen we must launder *pmd_k through
+		 * pmd_val() to ensure that _PAGE_PRESENT is correctly set.
+		 */
+		set_pmd(pmd, __pmd(pmd_val(*pmd_k)));
+#endif
+	else
+		BUG_ON(pmd_page(*pmd) != pmd_page(*pmd_k));
+	return pmd_k;
+}
+
+/*
+ * Handle a fault on the vmalloc or module mapping area
+ *
+ * This assumes no large pages in there.
+ */
+static inline int vmalloc_fault(unsigned long address)
+{
+	unsigned long pgd_paddr;
+	pmd_t *pmd_k;
+	pte_t *pte_k;
+	/*
+	 * Synchronize this task's top level page-table
+	 * with the 'reference' page table.
+	 *
+	 * Do _not_ use "current" here. We might be inside
+	 * an interrupt in the middle of a task switch..
+	 */
+	pgd_paddr = read_cr3();
+	pmd_k = vmalloc_sync_one(__va(pgd_paddr), address);
+	if (!pmd_k)
+		return -1;
+	pte_k = pte_offset_kernel(pmd_k, address);
+	if (!pte_present(*pte_k))
+		return -1;
+	return 0;
+}
+
+/*
+ * This routine handles page faults.  It determines the address,
+ * and the problem, and then passes it off to one of the appropriate
+ * routines.
+ *
+ * error_code:
+ *	bit 0 == 0 means no page found, 1 means protection fault
+ *	bit 1 == 0 means read, 1 means write
+ *	bit 2 == 0 means kernel, 1 means user-mode
+ *	bit 3 == 1 means use of reserved bit detected
+ *	bit 4 == 1 means fault was an instruction fetch
+ */
+fastcall void __kprobes do_page_fault(struct pt_regs *regs,
+				      unsigned long error_code)
+{
+	struct task_struct *tsk;
+	struct mm_struct *mm;
+	struct vm_area_struct * vma;
+	unsigned long address;
+	int write, si_code;
+
+	/* get the address */
+        address = read_cr2();
+
+	/* Set the "privileged fault" bit to something sane. */
+	error_code &= ~4;
+	error_code |= (regs->xcs & 2) << 1;
+	if (regs->eflags & X86_EFLAGS_VM)
+		error_code |= 4;
+
+	tsk = current;
+
+	si_code = SEGV_MAPERR;
+
+	/*
+	 * We fault-in kernel-space virtual memory on-demand. The
+	 * 'reference' page table is init_mm.pgd.
+	 *
+	 * NOTE! We MUST NOT take any locks for this case. We may
+	 * be in an interrupt or a critical region, and should
+	 * only copy the information from the master page table,
+	 * nothing more.
+	 *
+	 * This verifies that the fault happens in kernel space
+	 * (error_code & 4) == 0, and that the fault was not a
+	 * protection error (error_code & 9) == 0.
+	 */
+	if (unlikely(address >= TASK_SIZE)) {
+#ifdef CONFIG_XEN
+		/* Faults in hypervisor area can never be patched up. */
+		if (address >= hypervisor_virt_start)
+			goto bad_area_nosemaphore;
+#endif
+		if (!(error_code & 0x0000000d) && vmalloc_fault(address) >= 0)
+			return;
+		/* Can take a spurious fault if mapping changes R/O -> R/W. */
+		if (spurious_fault(regs, address, error_code))
+			return;
+		if (notify_page_fault(DIE_PAGE_FAULT, "page fault", regs, error_code, 14,
+						SIGSEGV) == NOTIFY_STOP)
+			return;
+		/* 
+		 * Don't take the mm semaphore here. If we fixup a prefetch
+		 * fault we could otherwise deadlock.
+		 */
+		goto bad_area_nosemaphore;
+	}
+
+	if (notify_page_fault(DIE_PAGE_FAULT, "page fault", regs, error_code, 14,
+					SIGSEGV) == NOTIFY_STOP)
+		return;
+
+	/* It's safe to allow irq's after cr2 has been saved and the vmalloc
+	   fault has been handled. */
+	if (regs->eflags & (X86_EFLAGS_IF|VM_MASK))
+		local_irq_enable();
+
+	mm = tsk->mm;
+
+	/*
+	 * If we're in an interrupt, have no user context or are running in an
+	 * atomic region then we must not take the fault..
+	 */
+	if (in_atomic() || !mm)
+		goto bad_area_nosemaphore;
+
+	/* When running in the kernel we expect faults to occur only to
+	 * addresses in user space.  All other faults represent errors in the
+	 * kernel and should generate an OOPS.  Unfortunatly, in the case of an
+	 * erroneous fault occurring in a code path which already holds mmap_sem
+	 * we will deadlock attempting to validate the fault against the
+	 * address space.  Luckily the kernel only validly references user
+	 * space from well defined areas of code, which are listed in the
+	 * exceptions table.
+	 *
+	 * As the vast majority of faults will be valid we will only perform
+	 * the source reference check when there is a possibilty of a deadlock.
+	 * Attempt to lock the address space, if we cannot we then validate the
+	 * source.  If this is invalid we can skip the address space check,
+	 * thus avoiding the deadlock.
+	 */
+	if (!down_read_trylock(&mm->mmap_sem)) {
+		if ((error_code & 4) == 0 &&
+		    !search_exception_tables(regs->eip))
+			goto bad_area_nosemaphore;
+		down_read(&mm->mmap_sem);
+	}
+
+	vma = find_vma(mm, address);
+	if (!vma)
+		goto bad_area;
+	if (vma->vm_start <= address)
+		goto good_area;
+	if (!(vma->vm_flags & VM_GROWSDOWN))
+		goto bad_area;
+	if (error_code & 4) {
+		/*
+		 * Accessing the stack below %esp is always a bug.
+		 * The large cushion allows instructions like enter
+		 * and pusha to work.  ("enter $65535,$31" pushes
+		 * 32 pointers and then decrements %esp by 65535.)
+		 */
+		if (address + 65536 + 32 * sizeof(unsigned long) < regs->esp)
+			goto bad_area;
+	}
+	if (expand_stack(vma, address))
+		goto bad_area;
+/*
+ * Ok, we have a good vm_area for this memory access, so
+ * we can handle it..
+ */
+good_area:
+	si_code = SEGV_ACCERR;
+	write = 0;
+	switch (error_code & 3) {
+		default:	/* 3: write, present */
+#ifdef TEST_VERIFY_AREA
+			if (regs->cs == GET_KERNEL_CS())
+				printk("WP fault at %08lx\n", regs->eip);
+#endif
+			/* fall through */
+		case 2:		/* write, not present */
+			if (!(vma->vm_flags & VM_WRITE))
+				goto bad_area;
+			write++;
+			break;
+		case 1:		/* read, present */
+			goto bad_area;
+		case 0:		/* read, not present */
+			if (!(vma->vm_flags & (VM_READ | VM_EXEC)))
+				goto bad_area;
+	}
+
+ survive:
+	/*
+	 * If for any reason at all we couldn't handle the fault,
+	 * make sure we exit gracefully rather than endlessly redo
+	 * the fault.
+	 */
+	switch (handle_mm_fault(mm, vma, address, write)) {
+		case VM_FAULT_MINOR:
+			tsk->min_flt++;
+			break;
+		case VM_FAULT_MAJOR:
+			tsk->maj_flt++;
+			break;
+		case VM_FAULT_SIGBUS:
+			goto do_sigbus;
+		case VM_FAULT_OOM:
+			goto out_of_memory;
+		default:
+			BUG();
+	}
+
+	/*
+	 * Did it hit the DOS screen memory VA from vm86 mode?
+	 */
+	if (regs->eflags & VM_MASK) {
+		unsigned long bit = (address - 0xA0000) >> PAGE_SHIFT;
+		if (bit < 32)
+			tsk->thread.screen_bitmap |= 1 << bit;
+	}
+	up_read(&mm->mmap_sem);
+	return;
+
+/*
+ * Something tried to access memory that isn't in our memory map..
+ * Fix it, but check if it's kernel or user first..
+ */
+bad_area:
+	up_read(&mm->mmap_sem);
+
+bad_area_nosemaphore:
+	/* User mode accesses just cause a SIGSEGV */
+	if (error_code & 4) {
+		/* 
+		 * Valid to do another page fault here because this one came 
+		 * from user space.
+		 */
+		if (is_prefetch(regs, address, error_code))
+			return;
+
+		tsk->thread.cr2 = address;
+		/* Kernel addresses are always protection faults */
+		tsk->thread.error_code = error_code | (address >= TASK_SIZE);
+		tsk->thread.trap_no = 14;
+		force_sig_info_fault(SIGSEGV, si_code, address, tsk);
+		return;
+	}
+
+#ifdef CONFIG_X86_F00F_BUG
+	/*
+	 * Pentium F0 0F C7 C8 bug workaround.
+	 */
+	if (boot_cpu_data.f00f_bug) {
+		unsigned long nr;
+		
+		nr = (address - idt_descr.address) >> 3;
+
+		if (nr == 6) {
+			do_invalid_op(regs, 0);
+			return;
+		}
+	}
+#endif
+
+no_context:
+	/* Are we prepared to handle this kernel fault?  */
+	if (fixup_exception(regs))
+		return;
+
+	/* 
+	 * Valid to do another page fault here, because if this fault
+	 * had been triggered by is_prefetch fixup_exception would have 
+	 * handled it.
+	 */
+ 	if (is_prefetch(regs, address, error_code))
+ 		return;
+
+/*
+ * Oops. The kernel tried to access some bad page. We'll have to
+ * terminate things with extreme prejudice.
+ */
+
+	bust_spinlocks(1);
+
+	if (oops_may_print()) {
+	#ifdef CONFIG_X86_PAE
+		if (error_code & 16) {
+			pte_t *pte = lookup_address(address);
+
+			if (pte && pte_present(*pte) && !pte_exec_kernel(*pte))
+				printk(KERN_CRIT "kernel tried to execute "
+					"NX-protected page - exploit attempt? "
+					"(uid: %d)\n", current->uid);
+		}
+	#endif
+		if (address < PAGE_SIZE)
+			printk(KERN_ALERT "BUG: unable to handle kernel NULL "
+					"pointer dereference");
+		else
+			printk(KERN_ALERT "BUG: unable to handle kernel paging"
+					" request");
+		printk(" at virtual address %08lx\n",address);
+		printk(KERN_ALERT " printing eip:\n");
+		printk("%08lx\n", regs->eip);
+	}
+	dump_fault_path(address);
+	tsk->thread.cr2 = address;
+	tsk->thread.trap_no = 14;
+	tsk->thread.error_code = error_code;
+	die("Oops", regs, error_code);
+	bust_spinlocks(0);
+	do_exit(SIGKILL);
+
+/*
+ * We ran out of memory, or some other thing happened to us that made
+ * us unable to handle the page fault gracefully.
+ */
+out_of_memory:
+	up_read(&mm->mmap_sem);
+	if (tsk->pid == 1) {
+		yield();
+		down_read(&mm->mmap_sem);
+		goto survive;
+	}
+	printk("VM: killing process %s\n", tsk->comm);
+	if (error_code & 4)
+		do_exit(SIGKILL);
+	goto no_context;
+
+do_sigbus:
+	up_read(&mm->mmap_sem);
+
+	/* Kernel mode? Handle exceptions or die */
+	if (!(error_code & 4))
+		goto no_context;
+
+	/* User space => ok to do another page fault */
+	if (is_prefetch(regs, address, error_code))
+		return;
+
+	tsk->thread.cr2 = address;
+	tsk->thread.error_code = error_code;
+	tsk->thread.trap_no = 14;
+	force_sig_info_fault(SIGBUS, BUS_ADRERR, address, tsk);
+}
+
+#if !HAVE_SHARED_KERNEL_PMD
+void vmalloc_sync_all(void)
+{
+	/*
+	 * Note that races in the updates of insync and start aren't
+	 * problematic: insync can only get set bits added, and updates to
+	 * start are only improving performance (without affecting correctness
+	 * if undone).
+	 * XEN: To work on PAE, we need to iterate over PMDs rather than PGDs.
+	 *      This change works just fine with 2-level paging too.
+	 */
+#define sync_index(a) ((a) >> PMD_SHIFT)
+	static DECLARE_BITMAP(insync, PTRS_PER_PGD*PTRS_PER_PMD);
+	static unsigned long start = TASK_SIZE;
+	unsigned long address;
+
+	BUILD_BUG_ON(TASK_SIZE & ~PGDIR_MASK);
+	for (address = start;
+	     address >= TASK_SIZE && address < hypervisor_virt_start;
+	     address += 1UL << PMD_SHIFT) {
+		if (!test_bit(sync_index(address), insync)) {
+			unsigned long flags;
+			struct page *page;
+
+			spin_lock_irqsave(&pgd_lock, flags);
+			/* XEN: failure path assumes non-empty pgd_list. */
+			if (unlikely(!pgd_list)) {
+				spin_unlock_irqrestore(&pgd_lock, flags);
+				return;
+			}
+			for (page = pgd_list; page; page =
+					(struct page *)page->index)
+				if (!vmalloc_sync_one(page_address(page),
+								address)) {
+					BUG_ON(page != pgd_list);
+					break;
+				}
+			spin_unlock_irqrestore(&pgd_lock, flags);
+			if (!page)
+				set_bit(sync_index(address), insync);
+		}
+		if (address == start && test_bit(sync_index(address), insync))
+			start = address + (1UL << PMD_SHIFT);
+	}
+}
+#endif
diff -rpuN linux-2.6.18.8/arch/i386/mm/highmem-xen.c linux-2.6.18-xen-3.3.0/arch/i386/mm/highmem-xen.c
--- linux-2.6.18.8/arch/i386/mm/highmem-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/mm/highmem-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,136 @@
+#include <linux/highmem.h>
+#include <linux/module.h>
+
+void *kmap(struct page *page)
+{
+	might_sleep();
+	if (!PageHighMem(page))
+		return page_address(page);
+	return kmap_high(page);
+}
+
+void kunmap(struct page *page)
+{
+	if (in_interrupt())
+		BUG();
+	if (!PageHighMem(page))
+		return;
+	kunmap_high(page);
+}
+
+/*
+ * kmap_atomic/kunmap_atomic is significantly faster than kmap/kunmap because
+ * no global lock is needed and because the kmap code must perform a global TLB
+ * invalidation when the kmap pool wraps.
+ *
+ * However when holding an atomic kmap is is not legal to sleep, so atomic
+ * kmaps are appropriate for short, tight code paths only.
+ */
+static void *__kmap_atomic(struct page *page, enum km_type type, pgprot_t prot)
+{
+	enum fixed_addresses idx;
+	unsigned long vaddr;
+
+	/* even !CONFIG_PREEMPT needs this, for in_atomic in do_page_fault */
+	inc_preempt_count();
+	if (!PageHighMem(page))
+		return page_address(page);
+
+	idx = type + KM_TYPE_NR*smp_processor_id();
+	vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx);
+#ifdef CONFIG_DEBUG_HIGHMEM
+	if (!pte_none(*(kmap_pte-idx)))
+		BUG();
+#endif
+	set_pte_at_sync(&init_mm, vaddr, kmap_pte-idx, mk_pte(page, prot));
+
+	return (void*) vaddr;
+}
+
+void *kmap_atomic(struct page *page, enum km_type type)
+{
+	return __kmap_atomic(page, type, kmap_prot);
+}
+
+/* Same as kmap_atomic but with PAGE_KERNEL_RO page protection. */
+void *kmap_atomic_pte(struct page *page, enum km_type type)
+{
+	return __kmap_atomic(page, type,
+	                     test_bit(PG_pinned, &page->flags)
+	                     ? PAGE_KERNEL_RO : kmap_prot);
+}
+
+void kunmap_atomic(void *kvaddr, enum km_type type)
+{
+#if defined(CONFIG_DEBUG_HIGHMEM) || defined(CONFIG_XEN)
+	unsigned long vaddr = (unsigned long) kvaddr & PAGE_MASK;
+	enum fixed_addresses idx = type + KM_TYPE_NR*smp_processor_id();
+
+	if (vaddr < FIXADDR_START) { // FIXME
+		dec_preempt_count();
+		preempt_check_resched();
+		return;
+	}
+#endif
+
+#if defined(CONFIG_DEBUG_HIGHMEM)
+	if (vaddr != __fix_to_virt(FIX_KMAP_BEGIN+idx))
+		BUG();
+
+	/*
+	 * force other mappings to Oops if they'll try to access
+	 * this pte without first remap it
+	 */
+	pte_clear(&init_mm, vaddr, kmap_pte-idx);
+	__flush_tlb_one(vaddr);
+#elif defined(CONFIG_XEN)
+	/*
+	 * We must ensure there are no dangling pagetable references when
+	 * returning memory to Xen (decrease_reservation).
+	 * XXX TODO: We could make this faster by only zapping when
+	 * kmap_flush_unused is called but that is trickier and more invasive.
+	 */
+	pte_clear(&init_mm, vaddr, kmap_pte-idx);
+#endif
+
+	dec_preempt_count();
+	preempt_check_resched();
+}
+
+/* This is the same as kmap_atomic() but can map memory that doesn't
+ * have a struct page associated with it.
+ */
+void *kmap_atomic_pfn(unsigned long pfn, enum km_type type)
+{
+	enum fixed_addresses idx;
+	unsigned long vaddr;
+
+	inc_preempt_count();
+
+	idx = type + KM_TYPE_NR*smp_processor_id();
+	vaddr = __fix_to_virt(FIX_KMAP_BEGIN + idx);
+	set_pte(kmap_pte-idx, pfn_pte(pfn, kmap_prot));
+	__flush_tlb_one(vaddr);
+
+	return (void*) vaddr;
+}
+
+struct page *kmap_atomic_to_page(void *ptr)
+{
+	unsigned long idx, vaddr = (unsigned long)ptr;
+	pte_t *pte;
+
+	if (vaddr < FIXADDR_START)
+		return virt_to_page(ptr);
+
+	idx = virt_to_fix(vaddr);
+	pte = kmap_pte - (idx - FIX_KMAP_BEGIN);
+	return pte_page(*pte);
+}
+
+EXPORT_SYMBOL(kmap);
+EXPORT_SYMBOL(kunmap);
+EXPORT_SYMBOL(kmap_atomic);
+EXPORT_SYMBOL(kmap_atomic_pte);
+EXPORT_SYMBOL(kunmap_atomic);
+EXPORT_SYMBOL(kmap_atomic_to_page);
diff -rpuN linux-2.6.18.8/arch/i386/mm/hypervisor.c linux-2.6.18-xen-3.3.0/arch/i386/mm/hypervisor.c
--- linux-2.6.18.8/arch/i386/mm/hypervisor.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/mm/hypervisor.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,602 @@
+/******************************************************************************
+ * mm/hypervisor.c
+ * 
+ * Update page tables via the hypervisor.
+ * 
+ * Copyright (c) 2002-2004, K A Fraser
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/vmalloc.h>
+#include <asm/page.h>
+#include <asm/pgtable.h>
+#include <asm/hypervisor.h>
+#include <xen/balloon.h>
+#include <xen/features.h>
+#include <xen/interface/memory.h>
+#include <linux/module.h>
+#include <linux/percpu.h>
+#include <asm/tlbflush.h>
+#include <linux/highmem.h>
+
+void xen_l1_entry_update(pte_t *ptr, pte_t val)
+{
+	mmu_update_t u;
+#ifdef CONFIG_HIGHPTE
+	u.ptr = ((unsigned long)ptr >= (unsigned long)high_memory) ?
+		arbitrary_virt_to_machine(ptr) : virt_to_machine(ptr);
+#else
+	u.ptr = virt_to_machine(ptr);
+#endif
+	u.val = __pte_val(val);
+	BUG_ON(HYPERVISOR_mmu_update(&u, 1, NULL, DOMID_SELF) < 0);
+}
+EXPORT_SYMBOL_GPL(xen_l1_entry_update);
+
+void xen_l2_entry_update(pmd_t *ptr, pmd_t val)
+{
+	mmu_update_t u;
+	u.ptr = virt_to_machine(ptr);
+	u.val = __pmd_val(val);
+	BUG_ON(HYPERVISOR_mmu_update(&u, 1, NULL, DOMID_SELF) < 0);
+}
+
+#if defined(CONFIG_X86_PAE) || defined(CONFIG_X86_64)
+void xen_l3_entry_update(pud_t *ptr, pud_t val)
+{
+	mmu_update_t u;
+	u.ptr = virt_to_machine(ptr);
+	u.val = __pud_val(val);
+	BUG_ON(HYPERVISOR_mmu_update(&u, 1, NULL, DOMID_SELF) < 0);
+}
+#endif
+
+#ifdef CONFIG_X86_64
+void xen_l4_entry_update(pgd_t *ptr, pgd_t val)
+{
+	mmu_update_t u;
+	u.ptr = virt_to_machine(ptr);
+	u.val = __pgd_val(val);
+	BUG_ON(HYPERVISOR_mmu_update(&u, 1, NULL, DOMID_SELF) < 0);
+}
+#endif /* CONFIG_X86_64 */
+
+void xen_pt_switch(unsigned long ptr)
+{
+	struct mmuext_op op;
+	op.cmd = MMUEXT_NEW_BASEPTR;
+	op.arg1.mfn = pfn_to_mfn(ptr >> PAGE_SHIFT);
+	BUG_ON(HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF) < 0);
+}
+
+void xen_new_user_pt(unsigned long ptr)
+{
+	struct mmuext_op op;
+	op.cmd = MMUEXT_NEW_USER_BASEPTR;
+	op.arg1.mfn = pfn_to_mfn(ptr >> PAGE_SHIFT);
+	BUG_ON(HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF) < 0);
+}
+
+void xen_tlb_flush(void)
+{
+	struct mmuext_op op;
+	op.cmd = MMUEXT_TLB_FLUSH_LOCAL;
+	BUG_ON(HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF) < 0);
+}
+EXPORT_SYMBOL(xen_tlb_flush);
+
+void xen_invlpg(unsigned long ptr)
+{
+	struct mmuext_op op;
+	op.cmd = MMUEXT_INVLPG_LOCAL;
+	op.arg1.linear_addr = ptr & PAGE_MASK;
+	BUG_ON(HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF) < 0);
+}
+EXPORT_SYMBOL(xen_invlpg);
+
+#ifdef CONFIG_SMP
+
+void xen_tlb_flush_all(void)
+{
+	struct mmuext_op op;
+	op.cmd = MMUEXT_TLB_FLUSH_ALL;
+	BUG_ON(HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF) < 0);
+}
+
+void xen_tlb_flush_mask(cpumask_t *mask)
+{
+	struct mmuext_op op;
+	if ( cpus_empty(*mask) )
+		return;
+	op.cmd = MMUEXT_TLB_FLUSH_MULTI;
+	set_xen_guest_handle(op.arg2.vcpumask, mask->bits);
+	BUG_ON(HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF) < 0);
+}
+
+void xen_invlpg_all(unsigned long ptr)
+{
+	struct mmuext_op op;
+	op.cmd = MMUEXT_INVLPG_ALL;
+	op.arg1.linear_addr = ptr & PAGE_MASK;
+	BUG_ON(HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF) < 0);
+}
+
+void xen_invlpg_mask(cpumask_t *mask, unsigned long ptr)
+{
+	struct mmuext_op op;
+	if ( cpus_empty(*mask) )
+		return;
+	op.cmd = MMUEXT_INVLPG_MULTI;
+	op.arg1.linear_addr = ptr & PAGE_MASK;
+	set_xen_guest_handle(op.arg2.vcpumask, mask->bits);
+	BUG_ON(HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF) < 0);
+}
+
+#endif /* CONFIG_SMP */
+
+void xen_pgd_pin(unsigned long ptr)
+{
+	struct mmuext_op op;
+#ifdef CONFIG_X86_64
+	op.cmd = MMUEXT_PIN_L4_TABLE;
+#elif defined(CONFIG_X86_PAE)
+	op.cmd = MMUEXT_PIN_L3_TABLE;
+#else
+	op.cmd = MMUEXT_PIN_L2_TABLE;
+#endif
+	op.arg1.mfn = pfn_to_mfn(ptr >> PAGE_SHIFT);
+	BUG_ON(HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF) < 0);
+}
+
+void xen_pgd_unpin(unsigned long ptr)
+{
+	struct mmuext_op op;
+	op.cmd = MMUEXT_UNPIN_TABLE;
+	op.arg1.mfn = pfn_to_mfn(ptr >> PAGE_SHIFT);
+	BUG_ON(HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF) < 0);
+}
+
+void xen_set_ldt(const void *ptr, unsigned int ents)
+{
+	struct mmuext_op op;
+	op.cmd = MMUEXT_SET_LDT;
+	op.arg1.linear_addr = (unsigned long)ptr;
+	op.arg2.nr_ents     = ents;
+	BUG_ON(HYPERVISOR_mmuext_op(&op, 1, NULL, DOMID_SELF) < 0);
+}
+
+/*
+ * Bitmap is indexed by page number. If bit is set, the page is part of a
+ * xen_create_contiguous_region() area of memory.
+ */
+unsigned long *contiguous_bitmap;
+
+static void contiguous_bitmap_set(
+	unsigned long first_page, unsigned long nr_pages)
+{
+	unsigned long start_off, end_off, curr_idx, end_idx;
+
+	curr_idx  = first_page / BITS_PER_LONG;
+	start_off = first_page & (BITS_PER_LONG-1);
+	end_idx   = (first_page + nr_pages) / BITS_PER_LONG;
+	end_off   = (first_page + nr_pages) & (BITS_PER_LONG-1);
+
+	if (curr_idx == end_idx) {
+		contiguous_bitmap[curr_idx] |=
+			((1UL<<end_off)-1) & -(1UL<<start_off);
+	} else {
+		contiguous_bitmap[curr_idx] |= -(1UL<<start_off);
+		while ( ++curr_idx < end_idx )
+			contiguous_bitmap[curr_idx] = ~0UL;
+		contiguous_bitmap[curr_idx] |= (1UL<<end_off)-1;
+	}
+}
+
+static void contiguous_bitmap_clear(
+	unsigned long first_page, unsigned long nr_pages)
+{
+	unsigned long start_off, end_off, curr_idx, end_idx;
+
+	curr_idx  = first_page / BITS_PER_LONG;
+	start_off = first_page & (BITS_PER_LONG-1);
+	end_idx   = (first_page + nr_pages) / BITS_PER_LONG;
+	end_off   = (first_page + nr_pages) & (BITS_PER_LONG-1);
+
+	if (curr_idx == end_idx) {
+		contiguous_bitmap[curr_idx] &=
+			-(1UL<<end_off) | ((1UL<<start_off)-1);
+	} else {
+		contiguous_bitmap[curr_idx] &= (1UL<<start_off)-1;
+		while ( ++curr_idx != end_idx )
+			contiguous_bitmap[curr_idx] = 0;
+		contiguous_bitmap[curr_idx] &= -(1UL<<end_off);
+	}
+}
+
+/* Protected by balloon_lock. */
+#define MAX_CONTIG_ORDER 9 /* 2MB */
+static unsigned long discontig_frames[1<<MAX_CONTIG_ORDER];
+static unsigned long limited_frames[1<<MAX_CONTIG_ORDER];
+static multicall_entry_t cr_mcl[1<<MAX_CONTIG_ORDER];
+
+/* Ensure multi-page extents are contiguous in machine memory. */
+int xen_create_contiguous_region(
+	unsigned long vstart, unsigned int order, unsigned int address_bits)
+{
+	unsigned long *in_frames = discontig_frames, out_frame;
+	unsigned long  frame, flags;
+	unsigned int   i;
+	int            rc, success;
+	struct xen_memory_exchange exchange = {
+		.in = {
+			.nr_extents   = 1UL << order,
+			.extent_order = 0,
+			.domid        = DOMID_SELF
+		},
+		.out = {
+			.nr_extents   = 1,
+			.extent_order = order,
+			.address_bits = address_bits,
+			.domid        = DOMID_SELF
+		}
+	};
+
+	/*
+	 * Currently an auto-translated guest will not perform I/O, nor will
+	 * it require PAE page directories below 4GB. Therefore any calls to
+	 * this function are redundant and can be ignored.
+	 */
+	if (xen_feature(XENFEAT_auto_translated_physmap))
+		return 0;
+
+	if (unlikely(order > MAX_CONTIG_ORDER))
+		return -ENOMEM;
+
+	set_xen_guest_handle(exchange.in.extent_start, in_frames);
+	set_xen_guest_handle(exchange.out.extent_start, &out_frame);
+
+	scrub_pages((void *)vstart, 1 << order);
+
+	balloon_lock(flags);
+
+	/* 1. Zap current PTEs, remembering MFNs. */
+	for (i = 0; i < (1U<<order); i++) {
+		in_frames[i] = pfn_to_mfn((__pa(vstart) >> PAGE_SHIFT) + i);
+		MULTI_update_va_mapping(cr_mcl + i, vstart + (i*PAGE_SIZE),
+					__pte_ma(0), 0);
+		set_phys_to_machine((__pa(vstart)>>PAGE_SHIFT)+i,
+			INVALID_P2M_ENTRY);
+	}
+	if (HYPERVISOR_multicall_check(cr_mcl, i, NULL))
+		BUG();
+
+	/* 2. Get a new contiguous memory extent. */
+	out_frame = __pa(vstart) >> PAGE_SHIFT;
+	rc = HYPERVISOR_memory_op(XENMEM_exchange, &exchange);
+	success = (exchange.nr_exchanged == (1UL << order));
+	BUG_ON(!success && ((exchange.nr_exchanged != 0) || (rc == 0)));
+	BUG_ON(success && (rc != 0));
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (unlikely(rc == -ENOSYS)) {
+		/* Compatibility when XENMEM_exchange is unsupported. */
+		if (HYPERVISOR_memory_op(XENMEM_decrease_reservation,
+					 &exchange.in) != (1UL << order))
+			BUG();
+		success = (HYPERVISOR_memory_op(XENMEM_populate_physmap,
+						&exchange.out) == 1);
+		if (!success) {
+			/* Couldn't get special memory: fall back to normal. */
+			for (i = 0; i < (1U<<order); i++)
+				in_frames[i] = (__pa(vstart)>>PAGE_SHIFT) + i;
+			if (HYPERVISOR_memory_op(XENMEM_populate_physmap,
+						 &exchange.in) != (1UL<<order))
+				BUG();
+		}
+	}
+#endif
+
+	/* 3. Map the new extent in place of old pages. */
+	for (i = 0; i < (1U<<order); i++) {
+		frame = success ? (out_frame + i) : in_frames[i];
+		MULTI_update_va_mapping(cr_mcl + i, vstart + (i*PAGE_SIZE),
+					pfn_pte_ma(frame, PAGE_KERNEL), 0);
+		set_phys_to_machine((__pa(vstart)>>PAGE_SHIFT)+i, frame);
+	}
+
+	cr_mcl[i - 1].args[MULTI_UVMFLAGS_INDEX] = order
+						   ? UVMF_TLB_FLUSH|UVMF_ALL
+						   : UVMF_INVLPG|UVMF_ALL;
+	if (HYPERVISOR_multicall_check(cr_mcl, i, NULL))
+		BUG();
+
+	if (success)
+		contiguous_bitmap_set(__pa(vstart) >> PAGE_SHIFT,
+				      1UL << order);
+
+	balloon_unlock(flags);
+
+	return success ? 0 : -ENOMEM;
+}
+EXPORT_SYMBOL_GPL(xen_create_contiguous_region);
+
+void xen_destroy_contiguous_region(unsigned long vstart, unsigned int order)
+{
+	unsigned long *out_frames = discontig_frames, in_frame;
+	unsigned long  frame, flags;
+	unsigned int   i;
+	int            rc, success;
+	struct xen_memory_exchange exchange = {
+		.in = {
+			.nr_extents   = 1,
+			.extent_order = order,
+			.domid        = DOMID_SELF
+		},
+		.out = {
+			.nr_extents   = 1UL << order,
+			.extent_order = 0,
+			.domid        = DOMID_SELF
+		}
+	};
+
+	if (xen_feature(XENFEAT_auto_translated_physmap) ||
+	    !test_bit(__pa(vstart) >> PAGE_SHIFT, contiguous_bitmap))
+		return;
+
+	if (unlikely(order > MAX_CONTIG_ORDER))
+		return;
+
+	set_xen_guest_handle(exchange.in.extent_start, &in_frame);
+	set_xen_guest_handle(exchange.out.extent_start, out_frames);
+
+	scrub_pages((void *)vstart, 1 << order);
+
+	balloon_lock(flags);
+
+	contiguous_bitmap_clear(__pa(vstart) >> PAGE_SHIFT, 1UL << order);
+
+	/* 1. Find start MFN of contiguous extent. */
+	in_frame = pfn_to_mfn(__pa(vstart) >> PAGE_SHIFT);
+
+	/* 2. Zap current PTEs. */
+	for (i = 0; i < (1U<<order); i++) {
+		MULTI_update_va_mapping(cr_mcl + i, vstart + (i*PAGE_SIZE),
+					__pte_ma(0), 0);
+		set_phys_to_machine((__pa(vstart)>>PAGE_SHIFT)+i,
+			INVALID_P2M_ENTRY);
+		out_frames[i] = (__pa(vstart) >> PAGE_SHIFT) + i;
+	}
+	if (HYPERVISOR_multicall_check(cr_mcl, i, NULL))
+		BUG();
+
+	/* 3. Do the exchange for non-contiguous MFNs. */
+	rc = HYPERVISOR_memory_op(XENMEM_exchange, &exchange);
+	success = (exchange.nr_exchanged == 1);
+	BUG_ON(!success && ((exchange.nr_exchanged != 0) || (rc == 0)));
+	BUG_ON(success && (rc != 0));
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (unlikely(rc == -ENOSYS)) {
+		/* Compatibility when XENMEM_exchange is unsupported. */
+		if (HYPERVISOR_memory_op(XENMEM_decrease_reservation,
+					 &exchange.in) != 1)
+			BUG();
+		if (HYPERVISOR_memory_op(XENMEM_populate_physmap,
+					 &exchange.out) != (1UL << order))
+			BUG();
+		success = 1;
+	}
+#endif
+
+	/* 4. Map new pages in place of old pages. */
+	for (i = 0; i < (1U<<order); i++) {
+		frame = success ? out_frames[i] : (in_frame + i);
+		MULTI_update_va_mapping(cr_mcl + i, vstart + (i*PAGE_SIZE),
+					pfn_pte_ma(frame, PAGE_KERNEL), 0);
+		set_phys_to_machine((__pa(vstart)>>PAGE_SHIFT)+i, frame);
+	}
+
+	cr_mcl[i - 1].args[MULTI_UVMFLAGS_INDEX] = order
+						   ? UVMF_TLB_FLUSH|UVMF_ALL
+						   : UVMF_INVLPG|UVMF_ALL;
+	if (HYPERVISOR_multicall_check(cr_mcl, i, NULL))
+		BUG();
+
+	balloon_unlock(flags);
+}
+EXPORT_SYMBOL_GPL(xen_destroy_contiguous_region);
+
+int xen_limit_pages_to_max_mfn(
+	struct page *pages, unsigned int order, unsigned int address_bits)
+{
+	unsigned long flags, frame;
+	unsigned long *in_frames = discontig_frames, *out_frames = limited_frames;
+	struct page *page;
+	unsigned int i, n, nr_mcl;
+	int rc, success;
+	DECLARE_BITMAP(limit_map, 1 << MAX_CONTIG_ORDER);
+
+	struct xen_memory_exchange exchange = {
+		.in = {
+			.extent_order = 0,
+			.domid        = DOMID_SELF
+		},
+		.out = {
+			.extent_order = 0,
+			.address_bits = address_bits,
+			.domid        = DOMID_SELF
+		}
+	};
+
+	if (xen_feature(XENFEAT_auto_translated_physmap))
+		return 0;
+
+	if (unlikely(order > MAX_CONTIG_ORDER))
+		return -ENOMEM;
+
+	bitmap_zero(limit_map, 1U << order);
+	set_xen_guest_handle(exchange.in.extent_start, in_frames);
+	set_xen_guest_handle(exchange.out.extent_start, out_frames);
+
+	/* 0. Scrub the pages. */
+	for (i = 0, n = 0; i < 1U<<order ; i++) {
+		page = &pages[i];
+		if (!(pfn_to_mfn(page_to_pfn(page)) >> (address_bits - PAGE_SHIFT)))
+			continue;
+		__set_bit(i, limit_map);
+
+		if (!PageHighMem(page))
+			scrub_pages(page_address(page), 1);
+#ifdef CONFIG_XEN_SCRUB_PAGES
+		else {
+			scrub_pages(kmap(page), 1);
+			kunmap(page);
+			++n;
+		}
+#endif
+	}
+	if (bitmap_empty(limit_map, 1U << order))
+		return 0;
+
+	if (n)
+		kmap_flush_unused();
+
+	balloon_lock(flags);
+
+	/* 1. Zap current PTEs (if any), remembering MFNs. */
+	for (i = 0, n = 0, nr_mcl = 0; i < (1U<<order); i++) {
+		if(!test_bit(i, limit_map))
+			continue;
+		page = &pages[i];
+
+		out_frames[n] = page_to_pfn(page);
+		in_frames[n] = pfn_to_mfn(out_frames[n]);
+
+		if (!PageHighMem(page))
+			MULTI_update_va_mapping(cr_mcl + nr_mcl++,
+						(unsigned long)page_address(page),
+						__pte_ma(0), 0);
+
+		set_phys_to_machine(out_frames[n], INVALID_P2M_ENTRY);
+		++n;
+	}
+	if (nr_mcl && HYPERVISOR_multicall_check(cr_mcl, nr_mcl, NULL))
+		BUG();
+
+	/* 2. Get new memory below the required limit. */
+	exchange.in.nr_extents = n;
+	exchange.out.nr_extents = n;
+	rc = HYPERVISOR_memory_op(XENMEM_exchange, &exchange);
+	success = (exchange.nr_exchanged == n);
+	BUG_ON(!success && ((exchange.nr_exchanged != 0) || (rc == 0)));
+	BUG_ON(success && (rc != 0));
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (unlikely(rc == -ENOSYS)) {
+		/* Compatibility when XENMEM_exchange is unsupported. */
+		if (HYPERVISOR_memory_op(XENMEM_decrease_reservation,
+					 &exchange.in) != n)
+			BUG();
+		if (HYPERVISOR_memory_op(XENMEM_populate_physmap,
+					 &exchange.out) != n)
+			BUG();
+		success = 1;
+	}
+#endif
+
+	/* 3. Map the new pages in place of old pages. */
+	for (i = 0, n = 0, nr_mcl = 0; i < (1U<<order); i++) {
+		if(!test_bit(i, limit_map))
+			continue;
+		page = &pages[i];
+
+		frame = success ? out_frames[n] : in_frames[n];
+
+		if (!PageHighMem(page))
+			MULTI_update_va_mapping(cr_mcl + nr_mcl++,
+						(unsigned long)page_address(page),
+						pfn_pte_ma(frame, PAGE_KERNEL), 0);
+
+		set_phys_to_machine(page_to_pfn(page), frame);
+		++n;
+	}
+	if (nr_mcl) {
+		cr_mcl[nr_mcl - 1].args[MULTI_UVMFLAGS_INDEX] = order
+							        ? UVMF_TLB_FLUSH|UVMF_ALL
+							        : UVMF_INVLPG|UVMF_ALL;
+		if (HYPERVISOR_multicall_check(cr_mcl, nr_mcl, NULL))
+			BUG();
+	}
+
+	balloon_unlock(flags);
+
+	return success ? 0 : -ENOMEM;
+}
+EXPORT_SYMBOL_GPL(xen_limit_pages_to_max_mfn);
+
+#ifdef __i386__
+int write_ldt_entry(void *ldt, int entry, __u32 entry_a, __u32 entry_b)
+{
+	__u32 *lp = (__u32 *)((char *)ldt + entry * 8);
+	maddr_t mach_lp = arbitrary_virt_to_machine(lp);
+	return HYPERVISOR_update_descriptor(
+		mach_lp, (u64)entry_a | ((u64)entry_b<<32));
+}
+#endif
+
+#define MAX_BATCHED_FULL_PTES 32
+
+int xen_change_pte_range(struct mm_struct *mm, pmd_t *pmd,
+			 unsigned long addr, unsigned long end, pgprot_t newprot)
+{
+	int rc = 0, i = 0;
+	mmu_update_t u[MAX_BATCHED_FULL_PTES];
+	pte_t *pte;
+	spinlock_t *ptl;
+
+	if (!xen_feature(XENFEAT_mmu_pt_update_preserve_ad))
+		return 0;
+
+	pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
+	do {
+		if (pte_present(*pte)) {
+			u[i].ptr = (__pmd_val(*pmd) & PHYSICAL_PAGE_MASK)
+				   | ((unsigned long)pte & ~PAGE_MASK)
+				   | MMU_PT_UPDATE_PRESERVE_AD;
+			u[i].val = __pte_val(pte_modify(*pte, newprot));
+			if (++i == MAX_BATCHED_FULL_PTES) {
+				if ((rc = HYPERVISOR_mmu_update(
+					&u[0], i, NULL, DOMID_SELF)) != 0)
+					break;
+				i = 0;
+			}
+		}
+	} while (pte++, addr += PAGE_SIZE, addr != end);
+	if (i)
+		rc = HYPERVISOR_mmu_update( &u[0], i, NULL, DOMID_SELF);
+	pte_unmap_unlock(pte - 1, ptl);
+	BUG_ON(rc && rc != -ENOSYS);
+	return !rc;
+}
diff -rpuN linux-2.6.18.8/arch/i386/mm/init-xen.c linux-2.6.18-xen-3.3.0/arch/i386/mm/init-xen.c
--- linux-2.6.18.8/arch/i386/mm/init-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/mm/init-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,847 @@
+/*
+ *  linux/arch/i386/mm/init.c
+ *
+ *  Copyright (C) 1995  Linus Torvalds
+ *
+ *  Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999
+ */
+
+#include <linux/module.h>
+#include <linux/signal.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/ptrace.h>
+#include <linux/mman.h>
+#include <linux/mm.h>
+#include <linux/hugetlb.h>
+#include <linux/swap.h>
+#include <linux/smp.h>
+#include <linux/init.h>
+#include <linux/highmem.h>
+#include <linux/pagemap.h>
+#include <linux/poison.h>
+#include <linux/bootmem.h>
+#include <linux/slab.h>
+#include <linux/proc_fs.h>
+#include <linux/efi.h>
+#include <linux/memory_hotplug.h>
+#include <linux/initrd.h>
+#include <linux/cpumask.h>
+#include <linux/dma-mapping.h>
+#include <linux/scatterlist.h>
+
+#include <asm/processor.h>
+#include <asm/system.h>
+#include <asm/uaccess.h>
+#include <asm/pgtable.h>
+#include <asm/dma.h>
+#include <asm/fixmap.h>
+#include <asm/e820.h>
+#include <asm/apic.h>
+#include <asm/tlb.h>
+#include <asm/tlbflush.h>
+#include <asm/sections.h>
+#include <asm/hypervisor.h>
+#include <asm/swiotlb.h>
+
+extern unsigned long *contiguous_bitmap;
+
+unsigned int __VMALLOC_RESERVE = 128 << 20;
+
+DEFINE_PER_CPU(struct mmu_gather, mmu_gathers);
+unsigned long highstart_pfn, highend_pfn;
+
+static int noinline do_test_wp_bit(void);
+
+/*
+ * Creates a middle page table and puts a pointer to it in the
+ * given global directory entry. This only returns the gd entry
+ * in non-PAE compilation mode, since the middle layer is folded.
+ */
+static pmd_t * __init one_md_table_init(pgd_t *pgd)
+{
+	pud_t *pud;
+	pmd_t *pmd_table;
+		
+#ifdef CONFIG_X86_PAE
+	pmd_table = (pmd_t *) alloc_bootmem_low_pages(PAGE_SIZE);
+	make_lowmem_page_readonly(pmd_table, XENFEAT_writable_page_tables);
+	set_pgd(pgd, __pgd(__pa(pmd_table) | _PAGE_PRESENT));
+	pud = pud_offset(pgd, 0);
+	if (pmd_table != pmd_offset(pud, 0)) 
+		BUG();
+#else
+	pud = pud_offset(pgd, 0);
+	pmd_table = pmd_offset(pud, 0);
+#endif
+
+	return pmd_table;
+}
+
+/*
+ * Create a page table and place a pointer to it in a middle page
+ * directory entry.
+ */
+static pte_t * __init one_page_table_init(pmd_t *pmd)
+{
+	if (pmd_none(*pmd)) {
+		pte_t *page_table = (pte_t *) alloc_bootmem_low_pages(PAGE_SIZE);
+		make_lowmem_page_readonly(page_table,
+					  XENFEAT_writable_page_tables);
+		set_pmd(pmd, __pmd(__pa(page_table) | _PAGE_TABLE));
+		if (page_table != pte_offset_kernel(pmd, 0))
+			BUG();	
+
+		return page_table;
+	}
+	
+	return pte_offset_kernel(pmd, 0);
+}
+
+/*
+ * This function initializes a certain range of kernel virtual memory 
+ * with new bootmem page tables, everywhere page tables are missing in
+ * the given range.
+ */
+
+/*
+ * NOTE: The pagetables are allocated contiguous on the physical space 
+ * so we can cache the place of the first one and move around without 
+ * checking the pgd every time.
+ */
+static void __init page_table_range_init (unsigned long start, unsigned long end, pgd_t *pgd_base)
+{
+	pgd_t *pgd;
+	pud_t *pud;
+	pmd_t *pmd;
+	int pgd_idx, pmd_idx;
+	unsigned long vaddr;
+
+	vaddr = start;
+	pgd_idx = pgd_index(vaddr);
+	pmd_idx = pmd_index(vaddr);
+	pgd = pgd_base + pgd_idx;
+
+	for ( ; (pgd_idx < PTRS_PER_PGD) && (vaddr != end); pgd++, pgd_idx++) {
+		if (pgd_none(*pgd)) 
+			one_md_table_init(pgd);
+		pud = pud_offset(pgd, vaddr);
+		pmd = pmd_offset(pud, vaddr);
+		for (; (pmd_idx < PTRS_PER_PMD) && (vaddr != end); pmd++, pmd_idx++) {
+			if (vaddr < hypervisor_virt_start && pmd_none(*pmd))
+				one_page_table_init(pmd);
+
+			vaddr += PMD_SIZE;
+		}
+		pmd_idx = 0;
+	}
+}
+
+static inline int is_kernel_text(unsigned long addr)
+{
+	if (addr >= PAGE_OFFSET && addr <= (unsigned long)__init_end)
+		return 1;
+	return 0;
+}
+
+/*
+ * This maps the physical memory to kernel virtual address space, a total 
+ * of max_low_pfn pages, by creating page tables starting from address 
+ * PAGE_OFFSET.
+ */
+static void __init kernel_physical_mapping_init(pgd_t *pgd_base)
+{
+	unsigned long pfn;
+	pgd_t *pgd;
+	pmd_t *pmd;
+	pte_t *pte;
+	int pgd_idx, pmd_idx, pte_ofs;
+
+	unsigned long max_ram_pfn = xen_start_info->nr_pages;
+	if (max_ram_pfn > max_low_pfn)
+		max_ram_pfn = max_low_pfn;
+
+	pgd_idx = pgd_index(PAGE_OFFSET);
+	pgd = pgd_base + pgd_idx;
+	pfn = 0;
+	pmd_idx = pmd_index(PAGE_OFFSET);
+	pte_ofs = pte_index(PAGE_OFFSET);
+
+	for (; pgd_idx < PTRS_PER_PGD; pgd++, pgd_idx++) {
+#ifdef CONFIG_XEN
+		/*
+		 * Native linux hasn't PAE-paging enabled yet at this
+		 * point.  When running as xen domain we are in PAE
+		 * mode already, thus we can't simply hook a empty
+		 * pmd.  That would kill the mappings we are currently
+		 * using ...
+		 */
+		pmd = pmd_offset(pud_offset(pgd, PAGE_OFFSET), PAGE_OFFSET);
+#else
+		pmd = one_md_table_init(pgd);
+#endif
+		if (pfn >= max_low_pfn)
+			continue;
+		pmd += pmd_idx;
+		for (; pmd_idx < PTRS_PER_PMD && pfn < max_low_pfn; pmd++, pmd_idx++) {
+			unsigned int address = pfn * PAGE_SIZE + PAGE_OFFSET;
+			if (address >= hypervisor_virt_start)
+				continue;
+
+			/* Map with big pages if possible, otherwise create normal page tables. */
+			if (cpu_has_pse) {
+				unsigned int address2 = (pfn + PTRS_PER_PTE - 1) * PAGE_SIZE + PAGE_OFFSET + PAGE_SIZE-1;
+
+				if (is_kernel_text(address) || is_kernel_text(address2))
+					set_pmd(pmd, pfn_pmd(pfn, PAGE_KERNEL_LARGE_EXEC));
+				else
+					set_pmd(pmd, pfn_pmd(pfn, PAGE_KERNEL_LARGE));
+				pfn += PTRS_PER_PTE;
+			} else {
+				pte = one_page_table_init(pmd);
+
+				pte += pte_ofs;
+				for (; pte_ofs < PTRS_PER_PTE && pfn < max_low_pfn; pte++, pfn++, pte_ofs++) {
+						/* XEN: Only map initial RAM allocation. */
+						if ((pfn >= max_ram_pfn) || pte_present(*pte))
+							continue;
+						if (is_kernel_text(address))
+							set_pte(pte, pfn_pte(pfn, PAGE_KERNEL_EXEC));
+						else
+							set_pte(pte, pfn_pte(pfn, PAGE_KERNEL));
+				}
+				pte_ofs = 0;
+			}
+		}
+		pmd_idx = 0;
+	}
+}
+
+#ifndef CONFIG_XEN
+
+static inline int page_kills_ppro(unsigned long pagenr)
+{
+	if (pagenr >= 0x70000 && pagenr <= 0x7003F)
+		return 1;
+	return 0;
+}
+
+#else
+
+#define page_kills_ppro(p)	0
+
+#endif
+
+extern int is_available_memory(efi_memory_desc_t *);
+
+int page_is_ram(unsigned long pagenr)
+{
+	int i;
+	unsigned long addr, end;
+
+	if (efi_enabled) {
+		efi_memory_desc_t *md;
+		void *p;
+
+		for (p = memmap.map; p < memmap.map_end; p += memmap.desc_size) {
+			md = p;
+			if (!is_available_memory(md))
+				continue;
+			addr = (md->phys_addr+PAGE_SIZE-1) >> PAGE_SHIFT;
+			end = (md->phys_addr + (md->num_pages << EFI_PAGE_SHIFT)) >> PAGE_SHIFT;
+
+			if ((pagenr >= addr) && (pagenr < end))
+				return 1;
+		}
+		return 0;
+	}
+
+	for (i = 0; i < e820.nr_map; i++) {
+
+		if (e820.map[i].type != E820_RAM)	/* not usable memory */
+			continue;
+		/*
+		 *	!!!FIXME!!! Some BIOSen report areas as RAM that
+		 *	are not. Notably the 640->1Mb area. We need a sanity
+		 *	check here.
+		 */
+		addr = (e820.map[i].addr+PAGE_SIZE-1) >> PAGE_SHIFT;
+		end = (e820.map[i].addr+e820.map[i].size) >> PAGE_SHIFT;
+		if  ((pagenr >= addr) && (pagenr < end))
+			return 1;
+	}
+	return 0;
+}
+
+#ifdef CONFIG_HIGHMEM
+pte_t *kmap_pte;
+pgprot_t kmap_prot;
+
+#define kmap_get_fixmap_pte(vaddr)					\
+	pte_offset_kernel(pmd_offset(pud_offset(pgd_offset_k(vaddr), vaddr), (vaddr)), (vaddr))
+
+static void __init kmap_init(void)
+{
+	unsigned long kmap_vstart;
+
+	/* cache the first kmap pte */
+	kmap_vstart = __fix_to_virt(FIX_KMAP_BEGIN);
+	kmap_pte = kmap_get_fixmap_pte(kmap_vstart);
+
+	kmap_prot = PAGE_KERNEL;
+}
+
+static void __init permanent_kmaps_init(pgd_t *pgd_base)
+{
+	pgd_t *pgd;
+	pud_t *pud;
+	pmd_t *pmd;
+	pte_t *pte;
+	unsigned long vaddr;
+
+	vaddr = PKMAP_BASE;
+	page_table_range_init(vaddr, vaddr + PAGE_SIZE*LAST_PKMAP, pgd_base);
+
+	pgd = swapper_pg_dir + pgd_index(vaddr);
+	pud = pud_offset(pgd, vaddr);
+	pmd = pmd_offset(pud, vaddr);
+	pte = pte_offset_kernel(pmd, vaddr);
+	pkmap_page_table = pte;	
+}
+
+static void __meminit free_new_highpage(struct page *page, int pfn)
+{
+	init_page_count(page);
+	if (pfn < xen_start_info->nr_pages)
+		__free_page(page);
+	totalhigh_pages++;
+}
+
+void __init add_one_highpage_init(struct page *page, int pfn, int bad_ppro)
+{
+	if (page_is_ram(pfn) && !(bad_ppro && page_kills_ppro(pfn))) {
+		ClearPageReserved(page);
+		free_new_highpage(page, pfn);
+	} else
+		SetPageReserved(page);
+}
+
+static int add_one_highpage_hotplug(struct page *page, unsigned long pfn)
+{
+	free_new_highpage(page, pfn);
+	totalram_pages++;
+#ifdef CONFIG_FLATMEM
+	max_mapnr = max(pfn, max_mapnr);
+#endif
+	num_physpages++;
+	return 0;
+}
+
+/*
+ * Not currently handling the NUMA case.
+ * Assuming single node and all memory that
+ * has been added dynamically that would be
+ * onlined here is in HIGHMEM
+ */
+void online_page(struct page *page)
+{
+	ClearPageReserved(page);
+	add_one_highpage_hotplug(page, page_to_pfn(page));
+}
+
+
+#ifdef CONFIG_NUMA
+extern void set_highmem_pages_init(int);
+#else
+static void __init set_highmem_pages_init(int bad_ppro)
+{
+	int pfn;
+	for (pfn = highstart_pfn; pfn < highend_pfn; pfn++)
+		add_one_highpage_init(pfn_to_page(pfn), pfn, bad_ppro);
+	totalram_pages += totalhigh_pages;
+}
+#endif /* CONFIG_FLATMEM */
+
+#else
+#define kmap_init() do { } while (0)
+#define permanent_kmaps_init(pgd_base) do { } while (0)
+#define set_highmem_pages_init(bad_ppro) do { } while (0)
+#endif /* CONFIG_HIGHMEM */
+
+unsigned long long __PAGE_KERNEL = _PAGE_KERNEL;
+EXPORT_SYMBOL(__PAGE_KERNEL);
+unsigned long long __PAGE_KERNEL_EXEC = _PAGE_KERNEL_EXEC;
+
+#ifdef CONFIG_NUMA
+extern void __init remap_numa_kva(void);
+#else
+#define remap_numa_kva() do {} while (0)
+#endif
+
+pgd_t *swapper_pg_dir;
+
+static void __init pagetable_init (void)
+{
+	unsigned long vaddr;
+	pgd_t *pgd_base = (pgd_t *)xen_start_info->pt_base;
+
+	/* Enable PSE if available */
+	if (cpu_has_pse) {
+		set_in_cr4(X86_CR4_PSE);
+	}
+
+	/* Enable PGE if available */
+	if (cpu_has_pge) {
+		set_in_cr4(X86_CR4_PGE);
+		__PAGE_KERNEL |= _PAGE_GLOBAL;
+		__PAGE_KERNEL_EXEC |= _PAGE_GLOBAL;
+	}
+
+	kernel_physical_mapping_init(pgd_base);
+	remap_numa_kva();
+
+	/*
+	 * Fixed mappings, only the page table structure has to be
+	 * created - mappings will be set by set_fixmap():
+	 */
+	vaddr = __fix_to_virt(__end_of_fixed_addresses - 1) & PMD_MASK;
+	page_table_range_init(vaddr, hypervisor_virt_start, pgd_base);
+
+	permanent_kmaps_init(pgd_base);
+}
+
+#if defined(CONFIG_SOFTWARE_SUSPEND) || defined(CONFIG_ACPI_SLEEP)
+/*
+ * Swap suspend & friends need this for resume because things like the intel-agp
+ * driver might have split up a kernel 4MB mapping.
+ */
+char __nosavedata swsusp_pg_dir[PAGE_SIZE]
+	__attribute__ ((aligned (PAGE_SIZE)));
+
+static inline void save_pg_dir(void)
+{
+	memcpy(swsusp_pg_dir, swapper_pg_dir, PAGE_SIZE);
+}
+#else
+static inline void save_pg_dir(void)
+{
+}
+#endif
+
+void zap_low_mappings (void)
+{
+	int i;
+
+	save_pg_dir();
+
+	/*
+	 * Zap initial low-memory mappings.
+	 *
+	 * Note that "pgd_clear()" doesn't do it for
+	 * us, because pgd_clear() is a no-op on i386.
+	 */
+	for (i = 0; i < USER_PTRS_PER_PGD; i++)
+#if defined(CONFIG_X86_PAE) && !defined(CONFIG_XEN)
+		set_pgd(swapper_pg_dir+i, __pgd(1 + __pa(empty_zero_page)));
+#else
+		set_pgd(swapper_pg_dir+i, __pgd(0));
+#endif
+	flush_tlb_all();
+}
+
+static int disable_nx __initdata = 0;
+u64 __supported_pte_mask __read_mostly = ~_PAGE_NX;
+EXPORT_SYMBOL(__supported_pte_mask);
+
+/*
+ * noexec = on|off
+ *
+ * Control non executable mappings.
+ *
+ * on      Enable
+ * off     Disable
+ */
+void __init noexec_setup(const char *str)
+{
+	if (!strncmp(str, "on",2) && cpu_has_nx) {
+		__supported_pte_mask |= _PAGE_NX;
+		disable_nx = 0;
+	} else if (!strncmp(str,"off",3)) {
+		disable_nx = 1;
+		__supported_pte_mask &= ~_PAGE_NX;
+	}
+}
+
+int nx_enabled = 0;
+#ifdef CONFIG_X86_PAE
+
+static void __init set_nx(void)
+{
+	unsigned int v[4], l, h;
+
+	if (cpu_has_pae && (cpuid_eax(0x80000000) > 0x80000001)) {
+		cpuid(0x80000001, &v[0], &v[1], &v[2], &v[3]);
+		if ((v[3] & (1 << 20)) && !disable_nx) {
+			rdmsr(MSR_EFER, l, h);
+			l |= EFER_NX;
+			wrmsr(MSR_EFER, l, h);
+			nx_enabled = 1;
+			__supported_pte_mask |= _PAGE_NX;
+		}
+	}
+}
+
+/*
+ * Enables/disables executability of a given kernel page and
+ * returns the previous setting.
+ */
+int __init set_kernel_exec(unsigned long vaddr, int enable)
+{
+	pte_t *pte;
+	int ret = 1;
+
+	if (!nx_enabled)
+		goto out;
+
+	pte = lookup_address(vaddr);
+	BUG_ON(!pte);
+
+	if (!pte_exec_kernel(*pte))
+		ret = 0;
+
+	if (enable)
+		pte->pte_high &= ~(1 << (_PAGE_BIT_NX - 32));
+	else
+		pte->pte_high |= 1 << (_PAGE_BIT_NX - 32);
+	__flush_tlb_all();
+out:
+	return ret;
+}
+
+#endif
+
+/*
+ * paging_init() sets up the page tables - note that the first 8MB are
+ * already mapped by head.S.
+ *
+ * This routines also unmaps the page at virtual kernel address 0, so
+ * that we can trap those pesky NULL-reference errors in the kernel.
+ */
+void __init paging_init(void)
+{
+	int i;
+
+#ifdef CONFIG_X86_PAE
+	set_nx();
+	if (nx_enabled)
+		printk("NX (Execute Disable) protection: active\n");
+#endif
+
+	pagetable_init();
+
+#if defined(CONFIG_X86_PAE) && !defined(CONFIG_XEN)
+	/*
+	 * We will bail out later - printk doesn't work right now so
+	 * the user would just see a hanging kernel.
+	 * when running as xen domain we are already in PAE mode at
+	 * this point.
+	 */
+	if (cpu_has_pae)
+		set_in_cr4(X86_CR4_PAE);
+#endif
+	__flush_tlb_all();
+
+	kmap_init();
+
+	/* Switch to the real shared_info page, and clear the
+	 * dummy page. */
+	set_fixmap(FIX_SHARED_INFO, xen_start_info->shared_info);
+	HYPERVISOR_shared_info = (shared_info_t *)fix_to_virt(FIX_SHARED_INFO);
+	memset(empty_zero_page, 0, sizeof(empty_zero_page));
+
+	/* Setup mapping of lower 1st MB */
+	for (i = 0; i < NR_FIX_ISAMAPS; i++)
+		if (is_initial_xendomain())
+			set_fixmap(FIX_ISAMAP_BEGIN - i, i * PAGE_SIZE);
+		else
+			__set_fixmap(FIX_ISAMAP_BEGIN - i,
+				     virt_to_machine(empty_zero_page),
+				     PAGE_KERNEL_RO);
+}
+
+/*
+ * Test if the WP bit works in supervisor mode. It isn't supported on 386's
+ * and also on some strange 486's (NexGen etc.). All 586+'s are OK. This
+ * used to involve black magic jumps to work around some nasty CPU bugs,
+ * but fortunately the switch to using exceptions got rid of all that.
+ */
+
+static void __init test_wp_bit(void)
+{
+	printk("Checking if this processor honours the WP bit even in supervisor mode... ");
+
+	/* Any page-aligned address will do, the test is non-destructive */
+	__set_fixmap(FIX_WP_TEST, __pa(&swapper_pg_dir), PAGE_READONLY);
+	boot_cpu_data.wp_works_ok = do_test_wp_bit();
+	clear_fixmap(FIX_WP_TEST);
+
+	if (!boot_cpu_data.wp_works_ok) {
+		printk("No.\n");
+#ifdef CONFIG_X86_WP_WORKS_OK
+		panic("This kernel doesn't support CPU's with broken WP. Recompile it for a 386!");
+#endif
+	} else {
+		printk("Ok.\n");
+	}
+}
+
+static void __init set_max_mapnr_init(void)
+{
+#ifdef CONFIG_HIGHMEM
+	num_physpages = highend_pfn;
+#else
+	num_physpages = max_low_pfn;
+#endif
+#ifdef CONFIG_FLATMEM
+	max_mapnr = num_physpages;
+#endif
+}
+
+static struct kcore_list kcore_mem, kcore_vmalloc; 
+
+void __init mem_init(void)
+{
+	extern int ppro_with_ram_bug(void);
+	int codesize, reservedpages, datasize, initsize;
+	int tmp;
+	int bad_ppro;
+	unsigned long pfn;
+
+	contiguous_bitmap = alloc_bootmem_low_pages(
+		(max_low_pfn + 2*BITS_PER_LONG) >> 3);
+	BUG_ON(!contiguous_bitmap);
+	memset(contiguous_bitmap, 0, (max_low_pfn + 2*BITS_PER_LONG) >> 3);
+
+#if defined(CONFIG_SWIOTLB)
+	swiotlb_init();	
+#endif
+
+#ifdef CONFIG_FLATMEM
+	if (!mem_map)
+		BUG();
+#endif
+	
+	bad_ppro = ppro_with_ram_bug();
+
+#ifdef CONFIG_HIGHMEM
+	/* check that fixmap and pkmap do not overlap */
+	if (PKMAP_BASE+LAST_PKMAP*PAGE_SIZE >= FIXADDR_START) {
+		printk(KERN_ERR "fixmap and kmap areas overlap - this will crash\n");
+		printk(KERN_ERR "pkstart: %lxh pkend: %lxh fixstart %lxh\n",
+				PKMAP_BASE, PKMAP_BASE+LAST_PKMAP*PAGE_SIZE, FIXADDR_START);
+		BUG();
+	}
+#endif
+ 
+	set_max_mapnr_init();
+
+#ifdef CONFIG_HIGHMEM
+	high_memory = (void *) __va(highstart_pfn * PAGE_SIZE - 1) + 1;
+#else
+	high_memory = (void *) __va(max_low_pfn * PAGE_SIZE - 1) + 1;
+#endif
+	printk("vmalloc area: %lx-%lx, maxmem %lx\n",
+	       VMALLOC_START,VMALLOC_END,MAXMEM);
+	BUG_ON(VMALLOC_START > VMALLOC_END);
+	
+	/* this will put all low memory onto the freelists */
+	totalram_pages += free_all_bootmem();
+	/* XEN: init and count low-mem pages outside initial allocation. */
+	for (pfn = xen_start_info->nr_pages; pfn < max_low_pfn; pfn++) {
+		ClearPageReserved(pfn_to_page(pfn));
+		init_page_count(pfn_to_page(pfn));
+		totalram_pages++;
+	}
+
+	reservedpages = 0;
+	for (tmp = 0; tmp < max_low_pfn; tmp++)
+		/*
+		 * Only count reserved RAM pages
+		 */
+		if (page_is_ram(tmp) && PageReserved(pfn_to_page(tmp)))
+			reservedpages++;
+
+	set_highmem_pages_init(bad_ppro);
+
+	codesize =  (unsigned long) &_etext - (unsigned long) &_text;
+	datasize =  (unsigned long) &_edata - (unsigned long) &_etext;
+	initsize =  (unsigned long) &__init_end - (unsigned long) &__init_begin;
+
+	kclist_add(&kcore_mem, __va(0), max_low_pfn << PAGE_SHIFT); 
+	kclist_add(&kcore_vmalloc, (void *)VMALLOC_START, 
+		   VMALLOC_END-VMALLOC_START);
+
+	printk(KERN_INFO "Memory: %luk/%luk available (%dk kernel code, %dk reserved, %dk data, %dk init, %ldk highmem)\n",
+		(unsigned long) nr_free_pages() << (PAGE_SHIFT-10),
+		num_physpages << (PAGE_SHIFT-10),
+		codesize >> 10,
+		reservedpages << (PAGE_SHIFT-10),
+		datasize >> 10,
+		initsize >> 10,
+		(unsigned long) (totalhigh_pages << (PAGE_SHIFT-10))
+	       );
+
+#ifdef CONFIG_X86_PAE
+	if (!cpu_has_pae)
+		panic("cannot execute a PAE-enabled kernel on a PAE-less CPU!");
+#endif
+	if (boot_cpu_data.wp_works_ok < 0)
+		test_wp_bit();
+
+	/*
+	 * Subtle. SMP is doing it's boot stuff late (because it has to
+	 * fork idle threads) - but it also needs low mappings for the
+	 * protected-mode entry to work. We zap these entries only after
+	 * the WP-bit has been tested.
+	 */
+#ifndef CONFIG_SMP
+	zap_low_mappings();
+#endif
+
+	set_bit(PG_pinned, &virt_to_page(init_mm.pgd)->flags);
+}
+
+/*
+ * this is for the non-NUMA, single node SMP system case.
+ * Specifically, in the case of x86, we will always add
+ * memory to the highmem for now.
+ */
+#ifdef CONFIG_MEMORY_HOTPLUG
+#ifndef CONFIG_NEED_MULTIPLE_NODES
+int arch_add_memory(int nid, u64 start, u64 size)
+{
+	struct pglist_data *pgdata = &contig_page_data;
+	struct zone *zone = pgdata->node_zones + MAX_NR_ZONES-1;
+	unsigned long start_pfn = start >> PAGE_SHIFT;
+	unsigned long nr_pages = size >> PAGE_SHIFT;
+
+	return __add_pages(zone, start_pfn, nr_pages);
+}
+
+int remove_memory(u64 start, u64 size)
+{
+	return -EINVAL;
+}
+#endif
+#endif
+
+kmem_cache_t *pgd_cache;
+kmem_cache_t *pmd_cache;
+
+void __init pgtable_cache_init(void)
+{
+	if (PTRS_PER_PMD > 1) {
+		pmd_cache = kmem_cache_create("pmd",
+					PTRS_PER_PMD*sizeof(pmd_t),
+					PTRS_PER_PMD*sizeof(pmd_t),
+					0,
+					pmd_ctor,
+					NULL);
+		if (!pmd_cache)
+			panic("pgtable_cache_init(): cannot create pmd cache");
+	}
+	pgd_cache = kmem_cache_create("pgd",
+#ifndef CONFIG_XEN
+				PTRS_PER_PGD*sizeof(pgd_t),
+				PTRS_PER_PGD*sizeof(pgd_t),
+#else
+				PAGE_SIZE,
+				PAGE_SIZE,
+#endif
+				0,
+				pgd_ctor,
+				PTRS_PER_PMD == 1 ? pgd_dtor : NULL);
+	if (!pgd_cache)
+		panic("pgtable_cache_init(): Cannot create pgd cache");
+}
+
+/*
+ * This function cannot be __init, since exceptions don't work in that
+ * section.  Put this after the callers, so that it cannot be inlined.
+ */
+static int noinline do_test_wp_bit(void)
+{
+	char tmp_reg;
+	int flag;
+
+	__asm__ __volatile__(
+		"	movb %0,%1	\n"
+		"1:	movb %1,%0	\n"
+		"	xorl %2,%2	\n"
+		"2:			\n"
+		".section __ex_table,\"a\"\n"
+		"	.align 4	\n"
+		"	.long 1b,2b	\n"
+		".previous		\n"
+		:"=m" (*(char *)fix_to_virt(FIX_WP_TEST)),
+		 "=q" (tmp_reg),
+		 "=r" (flag)
+		:"2" (1)
+		:"memory");
+	
+	return flag;
+}
+
+#ifdef CONFIG_DEBUG_RODATA
+
+void mark_rodata_ro(void)
+{
+	unsigned long addr = (unsigned long)__start_rodata;
+
+	for (; addr < (unsigned long)__end_rodata; addr += PAGE_SIZE)
+		change_page_attr(virt_to_page(addr), 1, PAGE_KERNEL_RO);
+
+	printk("Write protecting the kernel read-only data: %uk\n",
+			(__end_rodata - __start_rodata) >> 10);
+
+	/*
+	 * change_page_attr() requires a global_flush_tlb() call after it.
+	 * We do this after the printk so that if something went wrong in the
+	 * change, the printk gets out at least to give a better debug hint
+	 * of who is the culprit.
+	 */
+	global_flush_tlb();
+}
+#endif
+
+void free_init_pages(char *what, unsigned long begin, unsigned long end)
+{
+	unsigned long addr;
+
+	for (addr = begin; addr < end; addr += PAGE_SIZE) {
+		ClearPageReserved(virt_to_page(addr));
+		init_page_count(virt_to_page(addr));
+		memset((void *)addr, POISON_FREE_INITMEM, PAGE_SIZE);
+		free_page(addr);
+		totalram_pages++;
+	}
+	printk(KERN_INFO "Freeing %s: %ldk freed\n", what, (end - begin) >> 10);
+}
+
+void free_initmem(void)
+{
+	free_init_pages("unused kernel memory",
+			(unsigned long)(&__init_begin),
+			(unsigned long)(&__init_end));
+}
+
+#ifdef CONFIG_BLK_DEV_INITRD
+void free_initrd_mem(unsigned long start, unsigned long end)
+{
+	free_init_pages("initrd memory", start, end);
+}
+#endif
+
diff -rpuN linux-2.6.18.8/arch/i386/mm/ioremap-xen.c linux-2.6.18-xen-3.3.0/arch/i386/mm/ioremap-xen.c
--- linux-2.6.18.8/arch/i386/mm/ioremap-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/mm/ioremap-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,443 @@
+/*
+ * arch/i386/mm/ioremap.c
+ *
+ * Re-map IO memory to kernel address space so that we can access it.
+ * This is needed for high PCI addresses that aren't mapped in the
+ * 640k-1MB IO memory area on PC's
+ *
+ * (C) Copyright 1995 1996 Linus Torvalds
+ */
+
+#include <linux/vmalloc.h>
+#include <linux/init.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <asm/io.h>
+#include <asm/fixmap.h>
+#include <asm/cacheflush.h>
+#include <asm/tlbflush.h>
+#include <asm/pgtable.h>
+#include <asm/pgalloc.h>
+
+#define ISA_START_ADDRESS	0x0
+#define ISA_END_ADDRESS		0x100000
+
+static int direct_remap_area_pte_fn(pte_t *pte, 
+				    struct page *pmd_page,
+				    unsigned long address, 
+				    void *data)
+{
+	mmu_update_t **v = (mmu_update_t **)data;
+
+	BUG_ON(!pte_none(*pte));
+
+	(*v)->ptr = ((u64)pfn_to_mfn(page_to_pfn(pmd_page)) <<
+		     PAGE_SHIFT) | ((unsigned long)pte & ~PAGE_MASK);
+	(*v)++;
+
+	return 0;
+}
+
+static int __direct_remap_pfn_range(struct mm_struct *mm,
+				    unsigned long address, 
+				    unsigned long mfn,
+				    unsigned long size, 
+				    pgprot_t prot,
+				    domid_t  domid)
+{
+	int rc;
+	unsigned long i, start_address;
+	mmu_update_t *u, *v, *w;
+
+	u = v = w = (mmu_update_t *)__get_free_page(GFP_KERNEL|__GFP_REPEAT);
+	if (u == NULL)
+		return -ENOMEM;
+
+	start_address = address;
+
+	flush_cache_all();
+
+	for (i = 0; i < size; i += PAGE_SIZE) {
+		if ((v - u) == (PAGE_SIZE / sizeof(mmu_update_t))) {
+			/* Flush a full batch after filling in the PTE ptrs. */
+			rc = apply_to_page_range(mm, start_address, 
+						 address - start_address,
+						 direct_remap_area_pte_fn, &w);
+			if (rc)
+				goto out;
+			rc = -EFAULT;
+			if (HYPERVISOR_mmu_update(u, v - u, NULL, domid) < 0)
+				goto out;
+			v = w = u;
+			start_address = address;
+		}
+
+		/*
+		 * Fill in the machine address: PTE ptr is done later by
+		 * apply_to_page_range(). 
+		 */
+		v->val = __pte_val(pfn_pte_ma(mfn, prot)) | _PAGE_IO;
+
+		mfn++;
+		address += PAGE_SIZE; 
+		v++;
+	}
+
+	if (v != u) {
+		/* Final batch. */
+		rc = apply_to_page_range(mm, start_address,
+					 address - start_address,
+					 direct_remap_area_pte_fn, &w);
+		if (rc)
+			goto out;
+		rc = -EFAULT;
+		if (unlikely(HYPERVISOR_mmu_update(u, v - u, NULL, domid) < 0))
+			goto out;
+	}
+
+	rc = 0;
+
+ out:
+	flush_tlb_all();
+
+	free_page((unsigned long)u);
+
+	return rc;
+}
+
+int direct_remap_pfn_range(struct vm_area_struct *vma,
+			   unsigned long address, 
+			   unsigned long mfn,
+			   unsigned long size, 
+			   pgprot_t prot,
+			   domid_t  domid)
+{
+	if (xen_feature(XENFEAT_auto_translated_physmap))
+		return remap_pfn_range(vma, address, mfn, size, prot);
+
+	if (domid == DOMID_SELF)
+		return -EINVAL;
+
+	vma->vm_flags |= VM_IO | VM_RESERVED;
+
+	vma->vm_mm->context.has_foreign_mappings = 1;
+
+	return __direct_remap_pfn_range(
+		vma->vm_mm, address, mfn, size, prot, domid);
+}
+EXPORT_SYMBOL(direct_remap_pfn_range);
+
+int direct_kernel_remap_pfn_range(unsigned long address, 
+				  unsigned long mfn,
+				  unsigned long size, 
+				  pgprot_t prot,
+				  domid_t  domid)
+{
+	return __direct_remap_pfn_range(
+		&init_mm, address, mfn, size, prot, domid);
+}
+EXPORT_SYMBOL(direct_kernel_remap_pfn_range);
+
+static int lookup_pte_fn(
+	pte_t *pte, struct page *pmd_page, unsigned long addr, void *data)
+{
+	uint64_t *ptep = (uint64_t *)data;
+	if (ptep)
+		*ptep = ((uint64_t)pfn_to_mfn(page_to_pfn(pmd_page)) <<
+			 PAGE_SHIFT) | ((unsigned long)pte & ~PAGE_MASK);
+	return 0;
+}
+
+int create_lookup_pte_addr(struct mm_struct *mm, 
+			   unsigned long address,
+			   uint64_t *ptep)
+{
+	return apply_to_page_range(mm, address, PAGE_SIZE,
+				   lookup_pte_fn, ptep);
+}
+
+EXPORT_SYMBOL(create_lookup_pte_addr);
+
+static int noop_fn(
+	pte_t *pte, struct page *pmd_page, unsigned long addr, void *data)
+{
+	return 0;
+}
+
+int touch_pte_range(struct mm_struct *mm,
+		    unsigned long address,
+		    unsigned long size)
+{
+	return apply_to_page_range(mm, address, size, noop_fn, NULL);
+} 
+
+EXPORT_SYMBOL(touch_pte_range);
+
+/*
+ * Does @address reside within a non-highmem page that is local to this virtual
+ * machine (i.e., not an I/O page, nor a memory page belonging to another VM).
+ * See the comment that accompanies mfn_to_local_pfn() in page.h to understand
+ * why this works.
+ */
+static inline int is_local_lowmem(unsigned long address)
+{
+	extern unsigned long max_low_pfn;
+	return (mfn_to_local_pfn(address >> PAGE_SHIFT) < max_low_pfn);
+}
+
+/*
+ * Generic mapping function (not visible outside):
+ */
+
+/*
+ * Remap an arbitrary physical address space into the kernel virtual
+ * address space. Needed when the kernel wants to access high addresses
+ * directly.
+ *
+ * NOTE! We need to allow non-page-aligned mappings too: we will obviously
+ * have to convert them into an offset in a page-aligned mapping, but the
+ * caller shouldn't need to know that small detail.
+ */
+void __iomem * __ioremap(unsigned long phys_addr, unsigned long size, unsigned long flags)
+{
+	void __iomem * addr;
+	struct vm_struct * area;
+	unsigned long offset, last_addr;
+	domid_t domid = DOMID_IO;
+
+	/* Don't allow wraparound or zero size */
+	last_addr = phys_addr + size - 1;
+	if (!size || last_addr < phys_addr)
+		return NULL;
+
+	/*
+	 * Don't remap the low PCI/ISA area, it's always mapped..
+	 */
+	if (is_initial_xendomain() &&
+	    phys_addr >= ISA_START_ADDRESS && last_addr < ISA_END_ADDRESS)
+		return (void __iomem *) isa_bus_to_virt(phys_addr);
+
+	/*
+	 * Don't allow anybody to remap normal RAM that we're using..
+	 */
+	if (is_local_lowmem(phys_addr)) {
+		char *t_addr, *t_end;
+		struct page *page;
+
+		t_addr = bus_to_virt(phys_addr);
+		t_end = t_addr + (size - 1);
+	   
+		for(page = virt_to_page(t_addr); page <= virt_to_page(t_end); page++)
+			if(!PageReserved(page))
+				return NULL;
+
+		domid = DOMID_SELF;
+	}
+
+	/*
+	 * Mappings have to be page-aligned
+	 */
+	offset = phys_addr & ~PAGE_MASK;
+	phys_addr &= PAGE_MASK;
+	size = PAGE_ALIGN(last_addr+1) - phys_addr;
+
+	/*
+	 * Ok, go for it..
+	 */
+	area = get_vm_area(size, VM_IOREMAP | (flags << 20));
+	if (!area)
+		return NULL;
+	area->phys_addr = phys_addr;
+	addr = (void __iomem *) area->addr;
+	flags |= _KERNPG_TABLE;
+	if (__direct_remap_pfn_range(&init_mm, (unsigned long)addr,
+				     phys_addr>>PAGE_SHIFT,
+				     size, __pgprot(flags), domid)) {
+		vunmap((void __force *) addr);
+		return NULL;
+	}
+	return (void __iomem *) (offset + (char __iomem *)addr);
+}
+EXPORT_SYMBOL(__ioremap);
+
+/**
+ * ioremap_nocache     -   map bus memory into CPU space
+ * @offset:    bus address of the memory
+ * @size:      size of the resource to map
+ *
+ * ioremap_nocache performs a platform specific sequence of operations to
+ * make bus memory CPU accessible via the readb/readw/readl/writeb/
+ * writew/writel functions and the other mmio helpers. The returned
+ * address is not guaranteed to be usable directly as a virtual
+ * address. 
+ *
+ * This version of ioremap ensures that the memory is marked uncachable
+ * on the CPU as well as honouring existing caching rules from things like
+ * the PCI bus. Note that there are other caches and buffers on many 
+ * busses. In particular driver authors should read up on PCI writes
+ *
+ * It's useful if some control registers are in such an area and
+ * write combining or read caching is not desirable:
+ * 
+ * Must be freed with iounmap.
+ */
+
+void __iomem *ioremap_nocache (unsigned long phys_addr, unsigned long size)
+{
+	unsigned long last_addr;
+	void __iomem *p = __ioremap(phys_addr, size, _PAGE_PCD);
+	if (!p) 
+		return p; 
+
+	/* Guaranteed to be > phys_addr, as per __ioremap() */
+	last_addr = phys_addr + size - 1;
+
+	if (is_local_lowmem(last_addr)) { 
+		struct page *ppage = virt_to_page(bus_to_virt(phys_addr));
+		unsigned long npages;
+
+		phys_addr &= PAGE_MASK;
+
+		/* This might overflow and become zero.. */
+		last_addr = PAGE_ALIGN(last_addr);
+
+		/* .. but that's ok, because modulo-2**n arithmetic will make
+	 	* the page-aligned "last - first" come out right.
+	 	*/
+		npages = (last_addr - phys_addr) >> PAGE_SHIFT;
+
+		if (change_page_attr(ppage, npages, PAGE_KERNEL_NOCACHE) < 0) { 
+			iounmap(p); 
+			p = NULL;
+		}
+		global_flush_tlb();
+	}
+
+	return p;					
+}
+EXPORT_SYMBOL(ioremap_nocache);
+
+/**
+ * iounmap - Free a IO remapping
+ * @addr: virtual address from ioremap_*
+ *
+ * Caller must ensure there is only one unmapping for the same pointer.
+ */
+void iounmap(volatile void __iomem *addr)
+{
+	struct vm_struct *p, *o;
+
+	if ((void __force *)addr <= high_memory)
+		return;
+
+	/*
+	 * __ioremap special-cases the PCI/ISA range by not instantiating a
+	 * vm_area and by simply returning an address into the kernel mapping
+	 * of ISA space.   So handle that here.
+	 */
+	if ((unsigned long) addr >= fix_to_virt(FIX_ISAMAP_BEGIN))
+		return;
+
+	addr = (volatile void __iomem *)(PAGE_MASK & (unsigned long __force)addr);
+
+	/* Use the vm area unlocked, assuming the caller
+	   ensures there isn't another iounmap for the same address
+	   in parallel. Reuse of the virtual address is prevented by
+	   leaving it in the global lists until we're done with it.
+	   cpa takes care of the direct mappings. */
+	read_lock(&vmlist_lock);
+	for (p = vmlist; p; p = p->next) {
+		if (p->addr == addr)
+			break;
+	}
+	read_unlock(&vmlist_lock);
+
+	if (!p) {
+		printk("iounmap: bad address %p\n", addr);
+		dump_stack();
+		return;
+	}
+
+	/* Reset the direct mapping. Can block */
+	if ((p->flags >> 20) && is_local_lowmem(p->phys_addr)) {
+		/* p->size includes the guard page, but cpa doesn't like that */
+		change_page_attr(virt_to_page(bus_to_virt(p->phys_addr)),
+				 (p->size - PAGE_SIZE) >> PAGE_SHIFT,
+				 PAGE_KERNEL);
+		global_flush_tlb();
+	} 
+
+	/* Finally remove it */
+	o = remove_vm_area((void *)addr);
+	BUG_ON(p != o || o == NULL);
+	kfree(p); 
+}
+EXPORT_SYMBOL(iounmap);
+
+void __init *bt_ioremap(unsigned long phys_addr, unsigned long size)
+{
+	unsigned long offset, last_addr;
+	unsigned int nrpages;
+	enum fixed_addresses idx;
+
+	/* Don't allow wraparound or zero size */
+	last_addr = phys_addr + size - 1;
+	if (!size || last_addr < phys_addr)
+		return NULL;
+
+	/*
+	 * Don't remap the low PCI/ISA area, it's always mapped..
+	 */
+	if (is_initial_xendomain() &&
+	    phys_addr >= ISA_START_ADDRESS && last_addr < ISA_END_ADDRESS)
+		return isa_bus_to_virt(phys_addr);
+
+	/*
+	 * Mappings have to be page-aligned
+	 */
+	offset = phys_addr & ~PAGE_MASK;
+	phys_addr &= PAGE_MASK;
+	size = PAGE_ALIGN(last_addr) - phys_addr;
+
+	/*
+	 * Mappings have to fit in the FIX_BTMAP area.
+	 */
+	nrpages = size >> PAGE_SHIFT;
+	if (nrpages > NR_FIX_BTMAPS)
+		return NULL;
+
+	/*
+	 * Ok, go for it..
+	 */
+	idx = FIX_BTMAP_BEGIN;
+	while (nrpages > 0) {
+		set_fixmap(idx, phys_addr);
+		phys_addr += PAGE_SIZE;
+		--idx;
+		--nrpages;
+	}
+	return (void*) (offset + fix_to_virt(FIX_BTMAP_BEGIN));
+}
+
+void __init bt_iounmap(void *addr, unsigned long size)
+{
+	unsigned long virt_addr;
+	unsigned long offset;
+	unsigned int nrpages;
+	enum fixed_addresses idx;
+
+	virt_addr = (unsigned long)addr;
+	if (virt_addr < fix_to_virt(FIX_BTMAP_BEGIN))
+		return;
+	if (virt_addr >= fix_to_virt(FIX_ISAMAP_BEGIN))
+		return;
+	offset = virt_addr & ~PAGE_MASK;
+	nrpages = PAGE_ALIGN(offset + size - 1) >> PAGE_SHIFT;
+
+	idx = FIX_BTMAP_BEGIN;
+	while (nrpages > 0) {
+		clear_fixmap(idx);
+		--idx;
+		--nrpages;
+	}
+}
diff -rpuN linux-2.6.18.8/arch/i386/mm/Makefile linux-2.6.18-xen-3.3.0/arch/i386/mm/Makefile
--- linux-2.6.18.8/arch/i386/mm/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/mm/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -8,3 +8,4 @@ obj-$(CONFIG_NUMA) += discontig.o
 obj-$(CONFIG_HUGETLB_PAGE) += hugetlbpage.o
 obj-$(CONFIG_HIGHMEM) += highmem.o
 obj-$(CONFIG_BOOT_IOREMAP) += boot_ioremap.o
+obj-$(CONFIG_XEN) += hypervisor.o
diff -rpuN linux-2.6.18.8/arch/i386/mm/pageattr.c linux-2.6.18-xen-3.3.0/arch/i386/mm/pageattr.c
--- linux-2.6.18.8/arch/i386/mm/pageattr.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/mm/pageattr.c	2008-08-21 11:36:07.000000000 +0200
@@ -84,7 +84,7 @@ static void set_pmd_pte(pte_t *kpte, uns
 	unsigned long flags;
 
 	set_pte_atomic(kpte, pte); 	/* change init_mm */
-	if (PTRS_PER_PMD > 1)
+	if (HAVE_SHARED_KERNEL_PMD)
 		return;
 
 	spin_lock_irqsave(&pgd_lock, flags);
diff -rpuN linux-2.6.18.8/arch/i386/mm/pgtable.c linux-2.6.18-xen-3.3.0/arch/i386/mm/pgtable.c
--- linux-2.6.18.8/arch/i386/mm/pgtable.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/mm/pgtable.c	2008-08-21 11:36:07.000000000 +0200
@@ -12,6 +12,7 @@
 #include <linux/slab.h>
 #include <linux/pagemap.h>
 #include <linux/spinlock.h>
+#include <linux/module.h>
 
 #include <asm/system.h>
 #include <asm/pgtable.h>
@@ -137,6 +138,10 @@ void set_pmd_pfn(unsigned long vaddr, un
 	__flush_tlb_one(vaddr);
 }
 
+static int nr_fixmaps = 0;
+unsigned long __FIXADDR_TOP = 0xfffff000;
+EXPORT_SYMBOL(__FIXADDR_TOP);
+
 void __set_fixmap (enum fixed_addresses idx, unsigned long phys, pgprot_t flags)
 {
 	unsigned long address = __fix_to_virt(idx);
@@ -146,6 +151,13 @@ void __set_fixmap (enum fixed_addresses 
 		return;
 	}
 	set_pte_pfn(address, phys >> PAGE_SHIFT, flags);
+	nr_fixmaps++;
+}
+
+void set_fixaddr_top(unsigned long top)
+{
+	BUG_ON(nr_fixmaps > 0);
+	__FIXADDR_TOP = top - PAGE_SIZE;
 }
 
 pte_t *pte_alloc_one_kernel(struct mm_struct *mm, unsigned long address)
@@ -214,9 +226,10 @@ void pgd_ctor(void *pgd, kmem_cache_t *c
 		spin_lock_irqsave(&pgd_lock, flags);
 	}
 
-	clone_pgd_range((pgd_t *)pgd + USER_PTRS_PER_PGD,
-			swapper_pg_dir + USER_PTRS_PER_PGD,
-			KERNEL_PGD_PTRS);
+	if (PTRS_PER_PMD == 1 || HAVE_SHARED_KERNEL_PMD)
+		clone_pgd_range((pgd_t *)pgd + USER_PTRS_PER_PGD,
+				swapper_pg_dir + USER_PTRS_PER_PGD,
+				KERNEL_PGD_PTRS);
 	if (PTRS_PER_PMD > 1)
 		return;
 
@@ -248,6 +261,30 @@ pgd_t *pgd_alloc(struct mm_struct *mm)
 			goto out_oom;
 		set_pgd(&pgd[i], __pgd(1 + __pa(pmd)));
 	}
+
+	if (!HAVE_SHARED_KERNEL_PMD) {
+		unsigned long flags;
+
+		for (i = USER_PTRS_PER_PGD; i < PTRS_PER_PGD; i++) {
+			pmd_t *pmd = kmem_cache_alloc(pmd_cache, GFP_KERNEL);
+			if (!pmd)
+				goto out_oom;
+			set_pgd(&pgd[USER_PTRS_PER_PGD], __pgd(1 + __pa(pmd)));
+		}
+
+		spin_lock_irqsave(&pgd_lock, flags);
+		for (i = USER_PTRS_PER_PGD; i < PTRS_PER_PGD; i++) {
+			unsigned long v = (unsigned long)i << PGDIR_SHIFT;
+			pgd_t *kpgd = pgd_offset_k(v);
+			pud_t *kpud = pud_offset(kpgd, v);
+			pmd_t *kpmd = pmd_offset(kpud, v);
+			pmd_t *pmd = (void *)__va(pgd_val(pgd[i])-1);
+			memcpy(pmd, kpmd, PAGE_SIZE);
+		}
+		pgd_list_add(pgd);
+		spin_unlock_irqrestore(&pgd_lock, flags);
+	}
+
 	return pgd;
 
 out_oom:
@@ -262,9 +299,23 @@ void pgd_free(pgd_t *pgd)
 	int i;
 
 	/* in the PAE case user pgd entries are overwritten before usage */
-	if (PTRS_PER_PMD > 1)
-		for (i = 0; i < USER_PTRS_PER_PGD; ++i)
-			kmem_cache_free(pmd_cache, (void *)__va(pgd_val(pgd[i])-1));
+	if (PTRS_PER_PMD > 1) {
+		for (i = 0; i < USER_PTRS_PER_PGD; ++i) {
+			pmd_t *pmd = (void *)__va(pgd_val(pgd[i])-1);
+			kmem_cache_free(pmd_cache, pmd);
+		}
+		if (!HAVE_SHARED_KERNEL_PMD) {
+			unsigned long flags;
+			spin_lock_irqsave(&pgd_lock, flags);
+			pgd_list_del(pgd);
+			spin_unlock_irqrestore(&pgd_lock, flags);
+			for (i = USER_PTRS_PER_PGD; i < PTRS_PER_PGD; i++) {
+				pmd_t *pmd = (void *)__va(pgd_val(pgd[i])-1);
+				memset(pmd, 0, PTRS_PER_PMD*sizeof(pmd_t));
+				kmem_cache_free(pmd_cache, pmd);
+			}
+		}
+	}
 	/* in the non-PAE case, free_pgtables() clears user pgd entries */
 	kmem_cache_free(pgd_cache, pgd);
 }
diff -rpuN linux-2.6.18.8/arch/i386/mm/pgtable-xen.c linux-2.6.18-xen-3.3.0/arch/i386/mm/pgtable-xen.c
--- linux-2.6.18.8/arch/i386/mm/pgtable-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/mm/pgtable-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,725 @@
+/*
+ *  linux/arch/i386/mm/pgtable.c
+ */
+
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/mm.h>
+#include <linux/swap.h>
+#include <linux/smp.h>
+#include <linux/highmem.h>
+#include <linux/slab.h>
+#include <linux/pagemap.h>
+#include <linux/spinlock.h>
+#include <linux/module.h>
+
+#include <asm/system.h>
+#include <asm/pgtable.h>
+#include <asm/pgalloc.h>
+#include <asm/fixmap.h>
+#include <asm/e820.h>
+#include <asm/tlb.h>
+#include <asm/tlbflush.h>
+#include <asm/io.h>
+#include <asm/mmu_context.h>
+
+#include <xen/features.h>
+#include <asm/hypervisor.h>
+
+static void pgd_test_and_unpin(pgd_t *pgd);
+
+void show_mem(void)
+{
+	int total = 0, reserved = 0;
+	int shared = 0, cached = 0;
+	int highmem = 0;
+	struct page *page;
+	pg_data_t *pgdat;
+	unsigned long i;
+	unsigned long flags;
+
+	printk(KERN_INFO "Mem-info:\n");
+	show_free_areas();
+	printk(KERN_INFO "Free swap:       %6ldkB\n", nr_swap_pages<<(PAGE_SHIFT-10));
+	for_each_online_pgdat(pgdat) {
+		pgdat_resize_lock(pgdat, &flags);
+		for (i = 0; i < pgdat->node_spanned_pages; ++i) {
+			page = pgdat_page_nr(pgdat, i);
+			total++;
+			if (PageHighMem(page))
+				highmem++;
+			if (PageReserved(page))
+				reserved++;
+			else if (PageSwapCache(page))
+				cached++;
+			else if (page_count(page))
+				shared += page_count(page) - 1;
+		}
+		pgdat_resize_unlock(pgdat, &flags);
+	}
+	printk(KERN_INFO "%d pages of RAM\n", total);
+	printk(KERN_INFO "%d pages of HIGHMEM\n", highmem);
+	printk(KERN_INFO "%d reserved pages\n", reserved);
+	printk(KERN_INFO "%d pages shared\n", shared);
+	printk(KERN_INFO "%d pages swap cached\n", cached);
+
+	printk(KERN_INFO "%lu pages dirty\n", global_page_state(NR_FILE_DIRTY));
+	printk(KERN_INFO "%lu pages writeback\n",
+					global_page_state(NR_WRITEBACK));
+	printk(KERN_INFO "%lu pages mapped\n", global_page_state(NR_FILE_MAPPED));
+	printk(KERN_INFO "%lu pages slab\n", global_page_state(NR_SLAB));
+	printk(KERN_INFO "%lu pages pagetables\n",
+					global_page_state(NR_PAGETABLE));
+}
+
+/*
+ * Associate a large virtual page frame with a given physical page frame 
+ * and protection flags for that frame. pfn is for the base of the page,
+ * vaddr is what the page gets mapped to - both must be properly aligned. 
+ * The pmd must already be instantiated. Assumes PAE mode.
+ */ 
+void set_pmd_pfn(unsigned long vaddr, unsigned long pfn, pgprot_t flags)
+{
+	pgd_t *pgd;
+	pud_t *pud;
+	pmd_t *pmd;
+
+	if (vaddr & (PMD_SIZE-1)) {		/* vaddr is misaligned */
+		printk(KERN_WARNING "set_pmd_pfn: vaddr misaligned\n");
+		return; /* BUG(); */
+	}
+	if (pfn & (PTRS_PER_PTE-1)) {		/* pfn is misaligned */
+		printk(KERN_WARNING "set_pmd_pfn: pfn misaligned\n");
+		return; /* BUG(); */
+	}
+	pgd = swapper_pg_dir + pgd_index(vaddr);
+	if (pgd_none(*pgd)) {
+		printk(KERN_WARNING "set_pmd_pfn: pgd_none\n");
+		return; /* BUG(); */
+	}
+	pud = pud_offset(pgd, vaddr);
+	pmd = pmd_offset(pud, vaddr);
+	set_pmd(pmd, pfn_pmd(pfn, flags));
+	/*
+	 * It's enough to flush this one mapping.
+	 * (PGE mappings get flushed as well)
+	 */
+	__flush_tlb_one(vaddr);
+}
+
+static int nr_fixmaps = 0;
+unsigned long hypervisor_virt_start = HYPERVISOR_VIRT_START;
+unsigned long __FIXADDR_TOP = (HYPERVISOR_VIRT_START - 2 * PAGE_SIZE);
+EXPORT_SYMBOL(__FIXADDR_TOP);
+
+void __init set_fixaddr_top(unsigned long top)
+{
+	BUG_ON(nr_fixmaps > 0);
+	hypervisor_virt_start = top;
+	__FIXADDR_TOP = hypervisor_virt_start - 2 * PAGE_SIZE;
+}
+
+void __set_fixmap (enum fixed_addresses idx, maddr_t phys, pgprot_t flags)
+{
+	unsigned long address = __fix_to_virt(idx);
+	pte_t pte;
+
+	if (idx >= __end_of_fixed_addresses) {
+		BUG();
+		return;
+	}
+	switch (idx) {
+	case FIX_WP_TEST:
+	case FIX_VDSO:
+		pte = pfn_pte(phys >> PAGE_SHIFT, flags);
+		break;
+	default:
+		pte = pfn_pte_ma(phys >> PAGE_SHIFT, flags);
+		break;
+	}
+	if (HYPERVISOR_update_va_mapping(address, pte,
+					 UVMF_INVLPG|UVMF_ALL))
+		BUG();
+	nr_fixmaps++;
+}
+
+pte_t *pte_alloc_one_kernel(struct mm_struct *mm, unsigned long address)
+{
+	pte_t *pte = (pte_t *)__get_free_page(GFP_KERNEL|__GFP_REPEAT|__GFP_ZERO);
+	if (pte)
+		make_lowmem_page_readonly(pte, XENFEAT_writable_page_tables);
+	return pte;
+}
+
+struct page *pte_alloc_one(struct mm_struct *mm, unsigned long address)
+{
+	struct page *pte;
+
+#ifdef CONFIG_HIGHPTE
+	pte = alloc_pages(GFP_KERNEL|__GFP_HIGHMEM|__GFP_REPEAT|__GFP_ZERO, 0);
+#else
+	pte = alloc_pages(GFP_KERNEL|__GFP_REPEAT|__GFP_ZERO, 0);
+#endif
+	if (pte) {
+		SetPageForeign(pte, pte_free);
+		init_page_count(pte);
+	}
+	return pte;
+}
+
+void pte_free(struct page *pte)
+{
+	unsigned long pfn = page_to_pfn(pte);
+
+	if (!PageHighMem(pte)) {
+		unsigned long va = (unsigned long)__va(pfn << PAGE_SHIFT);
+
+		if (!pte_write(*virt_to_ptep(va)))
+			if (HYPERVISOR_update_va_mapping(
+				va, pfn_pte(pfn, PAGE_KERNEL), 0))
+				BUG();
+	} else
+		clear_bit(PG_pinned, &pte->flags);
+
+	ClearPageForeign(pte);
+	init_page_count(pte);
+
+	__free_page(pte);
+}
+
+void pmd_ctor(void *pmd, kmem_cache_t *cache, unsigned long flags)
+{
+	memset(pmd, 0, PTRS_PER_PMD*sizeof(pmd_t));
+}
+
+/*
+ * List of all pgd's needed for non-PAE so it can invalidate entries
+ * in both cached and uncached pgd's; not needed for PAE since the
+ * kernel pmd is shared. If PAE were not to share the pmd a similar
+ * tactic would be needed. This is essentially codepath-based locking
+ * against pageattr.c; it is the unique case in which a valid change
+ * of kernel pagetables can't be lazily synchronized by vmalloc faults.
+ * vmalloc faults work because attached pagetables are never freed.
+ * The locking scheme was chosen on the basis of manfred's
+ * recommendations and having no core impact whatsoever.
+ * -- wli
+ */
+DEFINE_SPINLOCK(pgd_lock);
+struct page *pgd_list;
+
+static inline void pgd_list_add(pgd_t *pgd)
+{
+	struct page *page = virt_to_page(pgd);
+	page->index = (unsigned long)pgd_list;
+	if (pgd_list)
+		set_page_private(pgd_list, (unsigned long)&page->index);
+	pgd_list = page;
+	set_page_private(page, (unsigned long)&pgd_list);
+}
+
+static inline void pgd_list_del(pgd_t *pgd)
+{
+	struct page *next, **pprev, *page = virt_to_page(pgd);
+	next = (struct page *)page->index;
+	pprev = (struct page **)page_private(page);
+	*pprev = next;
+	if (next)
+		set_page_private(next, (unsigned long)pprev);
+}
+
+void pgd_ctor(void *pgd, kmem_cache_t *cache, unsigned long unused)
+{
+	unsigned long flags;
+
+	if (PTRS_PER_PMD > 1) {
+		if (HAVE_SHARED_KERNEL_PMD)
+			clone_pgd_range((pgd_t *)pgd + USER_PTRS_PER_PGD,
+					swapper_pg_dir + USER_PTRS_PER_PGD,
+					KERNEL_PGD_PTRS);
+	} else {
+		spin_lock_irqsave(&pgd_lock, flags);
+		clone_pgd_range((pgd_t *)pgd + USER_PTRS_PER_PGD,
+				swapper_pg_dir + USER_PTRS_PER_PGD,
+				KERNEL_PGD_PTRS);
+		memset(pgd, 0, USER_PTRS_PER_PGD*sizeof(pgd_t));
+		pgd_list_add(pgd);
+		spin_unlock_irqrestore(&pgd_lock, flags);
+	}
+}
+
+/* never called when PTRS_PER_PMD > 1 */
+void pgd_dtor(void *pgd, kmem_cache_t *cache, unsigned long unused)
+{
+	unsigned long flags; /* can be called from interrupt context */
+
+	spin_lock_irqsave(&pgd_lock, flags);
+	pgd_list_del(pgd);
+	spin_unlock_irqrestore(&pgd_lock, flags);
+
+	pgd_test_and_unpin(pgd);
+}
+
+pgd_t *pgd_alloc(struct mm_struct *mm)
+{
+	int i;
+	pgd_t *pgd = kmem_cache_alloc(pgd_cache, GFP_KERNEL);
+	pmd_t **pmd;
+	unsigned long flags;
+
+	pgd_test_and_unpin(pgd);
+
+	if (PTRS_PER_PMD == 1 || !pgd)
+		return pgd;
+
+	if (HAVE_SHARED_KERNEL_PMD) {
+		for (i = 0; i < USER_PTRS_PER_PGD; ++i) {
+			pmd_t *pmd = kmem_cache_alloc(pmd_cache, GFP_KERNEL);
+			if (!pmd)
+				goto out_oom;
+			set_pgd(&pgd[i], __pgd(1 + __pa(pmd)));
+		}
+		return pgd;
+	}
+
+	/*
+	 * We can race save/restore (if we sleep during a GFP_KERNEL memory
+	 * allocation). We therefore store virtual addresses of pmds as they
+	 * do not change across save/restore, and poke the machine addresses
+	 * into the pgdir under the pgd_lock.
+	 */
+	pmd = kmalloc(PTRS_PER_PGD * sizeof(pmd_t *), GFP_KERNEL);
+	if (!pmd) {
+		kmem_cache_free(pgd_cache, pgd);
+		return NULL;
+	}
+
+	/* Allocate pmds, remember virtual addresses. */
+	for (i = 0; i < PTRS_PER_PGD; ++i) {
+		pmd[i] = kmem_cache_alloc(pmd_cache, GFP_KERNEL);
+		if (!pmd[i])
+			goto out_oom;
+	}
+
+	spin_lock_irqsave(&pgd_lock, flags);
+
+	/* Protect against save/restore: move below 4GB under pgd_lock. */
+	if (!xen_feature(XENFEAT_pae_pgdir_above_4gb)) {
+		int rc = xen_create_contiguous_region(
+			(unsigned long)pgd, 0, 32);
+		if (rc) {
+			spin_unlock_irqrestore(&pgd_lock, flags);
+			goto out_oom;
+		}
+	}
+
+	/* Copy kernel pmd contents and write-protect the new pmds. */
+	for (i = USER_PTRS_PER_PGD; i < PTRS_PER_PGD; i++) {
+		unsigned long v = (unsigned long)i << PGDIR_SHIFT;
+		pgd_t *kpgd = pgd_offset_k(v);
+		pud_t *kpud = pud_offset(kpgd, v);
+		pmd_t *kpmd = pmd_offset(kpud, v);
+		memcpy(pmd[i], kpmd, PAGE_SIZE);
+		make_lowmem_page_readonly(
+			pmd[i], XENFEAT_writable_page_tables);
+	}
+
+	/* It is safe to poke machine addresses of pmds under the pmd_lock. */
+	for (i = 0; i < PTRS_PER_PGD; i++)
+		set_pgd(&pgd[i], __pgd(1 + __pa(pmd[i])));
+
+	/* Ensure this pgd gets picked up and pinned on save/restore. */
+	pgd_list_add(pgd);
+
+	spin_unlock_irqrestore(&pgd_lock, flags);
+
+	kfree(pmd);
+
+	return pgd;
+
+out_oom:
+	if (HAVE_SHARED_KERNEL_PMD) {
+		for (i--; i >= 0; i--)
+			kmem_cache_free(pmd_cache,
+					(void *)__va(pgd_val(pgd[i])-1));
+	} else {
+		for (i--; i >= 0; i--)
+			kmem_cache_free(pmd_cache, pmd[i]);
+		kfree(pmd);
+	}
+	kmem_cache_free(pgd_cache, pgd);
+	return NULL;
+}
+
+void pgd_free(pgd_t *pgd)
+{
+	int i;
+
+	/*
+	 * After this the pgd should not be pinned for the duration of this
+	 * function's execution. We should never sleep and thus never race:
+	 *  1. User pmds will not become write-protected under our feet due
+	 *     to a concurrent mm_pin_all().
+	 *  2. The machine addresses in PGD entries will not become invalid
+	 *     due to a concurrent save/restore.
+	 */
+	pgd_test_and_unpin(pgd);
+
+	/* in the PAE case user pgd entries are overwritten before usage */
+	if (PTRS_PER_PMD > 1) {
+		for (i = 0; i < USER_PTRS_PER_PGD; ++i) {
+			pmd_t *pmd = (void *)__va(pgd_val(pgd[i])-1);
+			kmem_cache_free(pmd_cache, pmd);
+		}
+
+		if (!HAVE_SHARED_KERNEL_PMD) {
+			unsigned long flags;
+			spin_lock_irqsave(&pgd_lock, flags);
+			pgd_list_del(pgd);
+			spin_unlock_irqrestore(&pgd_lock, flags);
+
+			for (i = USER_PTRS_PER_PGD; i < PTRS_PER_PGD; i++) {
+				pmd_t *pmd = (void *)__va(pgd_val(pgd[i])-1);
+				make_lowmem_page_writable(
+					pmd, XENFEAT_writable_page_tables);
+				memset(pmd, 0, PTRS_PER_PMD*sizeof(pmd_t));
+				kmem_cache_free(pmd_cache, pmd);
+			}
+
+			if (!xen_feature(XENFEAT_pae_pgdir_above_4gb))
+				xen_destroy_contiguous_region(
+					(unsigned long)pgd, 0);
+		}
+	}
+
+	/* in the non-PAE case, free_pgtables() clears user pgd entries */
+	kmem_cache_free(pgd_cache, pgd);
+}
+
+void make_lowmem_page_readonly(void *va, unsigned int feature)
+{
+	pte_t *pte;
+	int rc;
+
+	if (xen_feature(feature))
+		return;
+
+	pte = virt_to_ptep(va);
+	rc = HYPERVISOR_update_va_mapping(
+		(unsigned long)va, pte_wrprotect(*pte), 0);
+	BUG_ON(rc);
+}
+
+void make_lowmem_page_writable(void *va, unsigned int feature)
+{
+	pte_t *pte;
+	int rc;
+
+	if (xen_feature(feature))
+		return;
+
+	pte = virt_to_ptep(va);
+	rc = HYPERVISOR_update_va_mapping(
+		(unsigned long)va, pte_mkwrite(*pte), 0);
+	BUG_ON(rc);
+}
+
+void make_page_readonly(void *va, unsigned int feature)
+{
+	pte_t *pte;
+	int rc;
+
+	if (xen_feature(feature))
+		return;
+
+	pte = virt_to_ptep(va);
+	rc = HYPERVISOR_update_va_mapping(
+		(unsigned long)va, pte_wrprotect(*pte), 0);
+	if (rc) /* fallback? */
+		xen_l1_entry_update(pte, pte_wrprotect(*pte));
+	if ((unsigned long)va >= (unsigned long)high_memory) {
+		unsigned long pfn = pte_pfn(*pte);
+#ifdef CONFIG_HIGHMEM
+		if (pfn >= highstart_pfn)
+			kmap_flush_unused(); /* flush stale writable kmaps */
+		else
+#endif
+			make_lowmem_page_readonly(
+				phys_to_virt(pfn << PAGE_SHIFT), feature); 
+	}
+}
+
+void make_page_writable(void *va, unsigned int feature)
+{
+	pte_t *pte;
+	int rc;
+
+	if (xen_feature(feature))
+		return;
+
+	pte = virt_to_ptep(va);
+	rc = HYPERVISOR_update_va_mapping(
+		(unsigned long)va, pte_mkwrite(*pte), 0);
+	if (rc) /* fallback? */
+		xen_l1_entry_update(pte, pte_mkwrite(*pte));
+	if ((unsigned long)va >= (unsigned long)high_memory) {
+		unsigned long pfn = pte_pfn(*pte); 
+#ifdef CONFIG_HIGHMEM
+		if (pfn < highstart_pfn)
+#endif
+			make_lowmem_page_writable(
+				phys_to_virt(pfn << PAGE_SHIFT), feature);
+	}
+}
+
+void make_pages_readonly(void *va, unsigned int nr, unsigned int feature)
+{
+	if (xen_feature(feature))
+		return;
+
+	while (nr-- != 0) {
+		make_page_readonly(va, feature);
+		va = (void *)((unsigned long)va + PAGE_SIZE);
+	}
+}
+
+void make_pages_writable(void *va, unsigned int nr, unsigned int feature)
+{
+	if (xen_feature(feature))
+		return;
+
+	while (nr-- != 0) {
+		make_page_writable(va, feature);
+		va = (void *)((unsigned long)va + PAGE_SIZE);
+	}
+}
+
+static void _pin_lock(struct mm_struct *mm, int lock) {
+	if (lock)
+		spin_lock(&mm->page_table_lock);
+#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
+	/* While mm->page_table_lock protects us against insertions and
+	 * removals of higher level page table pages, it doesn't protect
+	 * against updates of pte-s. Such updates, however, require the
+	 * pte pages to be in consistent state (unpinned+writable or
+	 * pinned+readonly). The pinning and attribute changes, however
+	 * cannot be done atomically, which is why such updates must be
+	 * prevented from happening concurrently.
+	 * Note that no pte lock can ever elsewhere be acquired nesting
+	 * with an already acquired one in the same mm, or with the mm's
+	 * page_table_lock already acquired, as that would break in the
+	 * non-split case (where all these are actually resolving to the
+	 * one page_table_lock). Thus acquiring all of them here is not
+	 * going to result in dead locks, and the order of acquires
+	 * doesn't matter.
+	 */
+	{
+		pgd_t *pgd = mm->pgd;
+		unsigned g;
+
+		for (g = 0; g < USER_PTRS_PER_PGD; g++, pgd++) {
+			pud_t *pud;
+			unsigned u;
+
+			if (pgd_none(*pgd))
+				continue;
+			pud = pud_offset(pgd, 0);
+			for (u = 0; u < PTRS_PER_PUD; u++, pud++) {
+				pmd_t *pmd;
+				unsigned m;
+
+				if (pud_none(*pud))
+					continue;
+				pmd = pmd_offset(pud, 0);
+				for (m = 0; m < PTRS_PER_PMD; m++, pmd++) {
+					spinlock_t *ptl;
+
+					if (pmd_none(*pmd))
+						continue;
+					ptl = pte_lockptr(0, pmd);
+					if (lock)
+						spin_lock(ptl);
+					else
+						spin_unlock(ptl);
+				}
+			}
+		}
+	}
+#endif
+	if (!lock)
+		spin_unlock(&mm->page_table_lock);
+}
+#define pin_lock(mm) _pin_lock(mm, 1)
+#define pin_unlock(mm) _pin_lock(mm, 0)
+
+#define PIN_BATCH 4
+static DEFINE_PER_CPU(multicall_entry_t[PIN_BATCH], pb_mcl);
+
+static inline unsigned int pgd_walk_set_prot(struct page *page, pgprot_t flags,
+                                             unsigned int cpu, unsigned seq)
+{
+	unsigned long pfn = page_to_pfn(page);
+
+	if (PageHighMem(page)) {
+		if (pgprot_val(flags) & _PAGE_RW)
+			clear_bit(PG_pinned, &page->flags);
+		else
+			set_bit(PG_pinned, &page->flags);
+	} else {
+		MULTI_update_va_mapping(per_cpu(pb_mcl, cpu) + seq,
+				(unsigned long)__va(pfn << PAGE_SHIFT),
+				pfn_pte(pfn, flags), 0);
+		if (unlikely(++seq == PIN_BATCH)) {
+			if (unlikely(HYPERVISOR_multicall_check(per_cpu(pb_mcl, cpu),
+		                                                PIN_BATCH, NULL)))
+				BUG();
+			seq = 0;
+		}
+	}
+
+	return seq;
+}
+
+static void pgd_walk(pgd_t *pgd_base, pgprot_t flags)
+{
+	pgd_t *pgd = pgd_base;
+	pud_t *pud;
+	pmd_t *pmd;
+	int    g, u, m;
+	unsigned int cpu, seq;
+
+	if (xen_feature(XENFEAT_auto_translated_physmap))
+		return;
+
+	cpu = get_cpu();
+
+	for (g = 0, seq = 0; g < USER_PTRS_PER_PGD; g++, pgd++) {
+		if (pgd_none(*pgd))
+			continue;
+		pud = pud_offset(pgd, 0);
+		if (PTRS_PER_PUD > 1) /* not folded */
+			seq = pgd_walk_set_prot(virt_to_page(pud),flags,cpu,seq);
+		for (u = 0; u < PTRS_PER_PUD; u++, pud++) {
+			if (pud_none(*pud))
+				continue;
+			pmd = pmd_offset(pud, 0);
+			if (PTRS_PER_PMD > 1) /* not folded */
+				seq = pgd_walk_set_prot(virt_to_page(pmd),flags,cpu,seq);
+			for (m = 0; m < PTRS_PER_PMD; m++, pmd++) {
+				if (pmd_none(*pmd))
+					continue;
+				seq = pgd_walk_set_prot(pmd_page(*pmd),flags,cpu,seq);
+			}
+		}
+	}
+
+	if (likely(seq != 0)) {
+		MULTI_update_va_mapping(per_cpu(pb_mcl, cpu) + seq,
+			(unsigned long)pgd_base,
+			pfn_pte(virt_to_phys(pgd_base)>>PAGE_SHIFT, flags),
+			UVMF_TLB_FLUSH);
+		if (unlikely(HYPERVISOR_multicall_check(per_cpu(pb_mcl, cpu),
+		                                        seq + 1, NULL)))
+			BUG();
+	} else if(HYPERVISOR_update_va_mapping((unsigned long)pgd_base,
+			pfn_pte(virt_to_phys(pgd_base)>>PAGE_SHIFT, flags),
+			UVMF_TLB_FLUSH))
+		BUG();
+
+	put_cpu();
+}
+
+static void __pgd_pin(pgd_t *pgd)
+{
+	pgd_walk(pgd, PAGE_KERNEL_RO);
+	kmap_flush_unused();
+	xen_pgd_pin(__pa(pgd));
+	set_bit(PG_pinned, &virt_to_page(pgd)->flags);
+}
+
+static void __pgd_unpin(pgd_t *pgd)
+{
+	xen_pgd_unpin(__pa(pgd));
+	pgd_walk(pgd, PAGE_KERNEL);
+	clear_bit(PG_pinned, &virt_to_page(pgd)->flags);
+}
+
+static void pgd_test_and_unpin(pgd_t *pgd)
+{
+	if (test_bit(PG_pinned, &virt_to_page(pgd)->flags))
+		__pgd_unpin(pgd);
+}
+
+void mm_pin(struct mm_struct *mm)
+{
+	if (xen_feature(XENFEAT_writable_page_tables))
+		return;
+	pin_lock(mm);
+	__pgd_pin(mm->pgd);
+	pin_unlock(mm);
+}
+
+void mm_unpin(struct mm_struct *mm)
+{
+	if (xen_feature(XENFEAT_writable_page_tables))
+		return;
+	pin_lock(mm);
+	__pgd_unpin(mm->pgd);
+	pin_unlock(mm);
+}
+
+void mm_pin_all(void)
+{
+	struct page *page;
+	unsigned long flags;
+
+	if (xen_feature(XENFEAT_writable_page_tables))
+		return;
+
+	/*
+	 * Allow uninterrupted access to the pgd_list. Also protects
+	 * __pgd_pin() by disabling preemption.
+	 * All other CPUs must be at a safe point (e.g., in stop_machine
+	 * or offlined entirely).
+	 */
+	spin_lock_irqsave(&pgd_lock, flags);
+	for (page = pgd_list; page; page = (struct page *)page->index) {
+		if (!test_bit(PG_pinned, &page->flags))
+			__pgd_pin((pgd_t *)page_address(page));
+	}
+	spin_unlock_irqrestore(&pgd_lock, flags);
+}
+
+void _arch_dup_mmap(struct mm_struct *mm)
+{
+	if (!test_bit(PG_pinned, &virt_to_page(mm->pgd)->flags))
+		mm_pin(mm);
+}
+
+void _arch_exit_mmap(struct mm_struct *mm)
+{
+	struct task_struct *tsk = current;
+
+	task_lock(tsk);
+
+	/*
+	 * We aggressively remove defunct pgd from cr3. We execute unmap_vmas()
+	 * *much* faster this way, as no tlb flushes means bigger wrpt batches.
+	 */
+	if (tsk->active_mm == mm) {
+		tsk->active_mm = &init_mm;
+		atomic_inc(&init_mm.mm_count);
+
+		switch_mm(mm, &init_mm, tsk);
+
+		atomic_dec(&mm->mm_count);
+		BUG_ON(atomic_read(&mm->mm_count) == 0);
+	}
+
+	task_unlock(tsk);
+
+	if (test_bit(PG_pinned, &virt_to_page(mm->pgd)->flags) &&
+	    (atomic_read(&mm->mm_count) == 1) &&
+	    !mm->context.has_foreign_mappings)
+		mm_unpin(mm);
+}
diff -rpuN linux-2.6.18.8/arch/i386/oprofile/Makefile linux-2.6.18-xen-3.3.0/arch/i386/oprofile/Makefile
--- linux-2.6.18.8/arch/i386/oprofile/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/oprofile/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -6,7 +6,14 @@ DRIVER_OBJS = $(addprefix ../../../drive
 		oprofilefs.o oprofile_stats.o  \
 		timer_int.o )
 
+ifdef CONFIG_XEN
+XENOPROF_COMMON_OBJS = $(addprefix ../../../drivers/xen/xenoprof/, \
+			 xenoprofile.o)
+oprofile-y				:= $(DRIVER_OBJS) \
+					   $(XENOPROF_COMMON_OBJS) xenoprof.o
+else 
 oprofile-y				:= $(DRIVER_OBJS) init.o backtrace.o
 oprofile-$(CONFIG_X86_LOCAL_APIC) 	+= nmi_int.o op_model_athlon.o \
 					   op_model_ppro.o op_model_p4.o
 oprofile-$(CONFIG_X86_IO_APIC)		+= nmi_timer_int.o
+endif
diff -rpuN linux-2.6.18.8/arch/i386/oprofile/xenoprof.c linux-2.6.18-xen-3.3.0/arch/i386/oprofile/xenoprof.c
--- linux-2.6.18.8/arch/i386/oprofile/xenoprof.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/oprofile/xenoprof.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,179 @@
+/**
+ * @file xenoprof.c
+ *
+ * @remark Copyright 2002 OProfile authors
+ * @remark Read the file COPYING
+ *
+ * @author John Levon <levon@movementarian.org>
+ *
+ * Modified by Aravind Menon and Jose Renato Santos for Xen
+ * These modifications are:
+ * Copyright (C) 2005 Hewlett-Packard Co.
+ *
+ * x86-specific part
+ * Copyright (c) 2006 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ */
+
+#include <linux/init.h>
+#include <linux/oprofile.h>
+#include <linux/sched.h>
+#include <asm/pgtable.h>
+
+#include <xen/driver_util.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/xenoprof.h>
+#include <xen/xenoprof.h>
+#include "op_counter.h"
+
+static unsigned int num_events = 0;
+
+void __init xenoprof_arch_init_counter(struct xenoprof_init *init)
+{
+	num_events = init->num_events;
+	/* just in case - make sure we do not overflow event list 
+	   (i.e. counter_config list) */
+	if (num_events > OP_MAX_COUNTER) {
+		num_events = OP_MAX_COUNTER;
+		init->num_events = num_events;
+	}
+}
+
+void xenoprof_arch_counter(void)
+{
+	int i;
+	struct xenoprof_counter counter;
+
+	for (i=0; i<num_events; i++) {
+		counter.ind       = i;
+		counter.count     = (uint64_t)counter_config[i].count;
+		counter.enabled   = (uint32_t)counter_config[i].enabled;
+		counter.event     = (uint32_t)counter_config[i].event;
+		counter.kernel    = (uint32_t)counter_config[i].kernel;
+		counter.user      = (uint32_t)counter_config[i].user;
+		counter.unit_mask = (uint64_t)counter_config[i].unit_mask;
+		WARN_ON(HYPERVISOR_xenoprof_op(XENOPROF_counter,
+					       &counter));
+	}
+}
+
+void xenoprof_arch_start(void) 
+{
+	/* nothing */
+}
+
+void xenoprof_arch_stop(void)
+{
+	/* nothing */
+}
+
+void xenoprof_arch_unmap_shared_buffer(struct xenoprof_shared_buffer * sbuf)
+{
+	if (sbuf->buffer) {
+		vunmap(sbuf->buffer);
+		sbuf->buffer = NULL;
+	}
+}
+
+int xenoprof_arch_map_shared_buffer(struct xenoprof_get_buffer * get_buffer,
+				    struct xenoprof_shared_buffer * sbuf)
+{
+	int npages, ret;
+	struct vm_struct *area;
+
+	sbuf->buffer = NULL;
+	if ( (ret = HYPERVISOR_xenoprof_op(XENOPROF_get_buffer, get_buffer)) )
+		return ret;
+
+	npages = (get_buffer->bufsize * get_buffer->nbuf - 1) / PAGE_SIZE + 1;
+
+	area = alloc_vm_area(npages * PAGE_SIZE);
+	if (area == NULL)
+		return -ENOMEM;
+
+	if ( (ret = direct_kernel_remap_pfn_range(
+		      (unsigned long)area->addr,
+		      get_buffer->buf_gmaddr >> PAGE_SHIFT,
+		      npages * PAGE_SIZE, __pgprot(_KERNPG_TABLE),
+		      DOMID_SELF)) ) {
+		vunmap(area->addr);
+		return ret;
+	}
+
+	sbuf->buffer = area->addr;
+	return ret;
+}
+
+int xenoprof_arch_set_passive(struct xenoprof_passive * pdomain,
+			      struct xenoprof_shared_buffer * sbuf)
+{
+	int ret;
+	int npages;
+	struct vm_struct *area;
+	pgprot_t prot = __pgprot(_KERNPG_TABLE);
+
+	sbuf->buffer = NULL;
+	ret = HYPERVISOR_xenoprof_op(XENOPROF_set_passive, pdomain);
+	if (ret)
+		goto out;
+
+	npages = (pdomain->bufsize * pdomain->nbuf - 1) / PAGE_SIZE + 1;
+
+	area = alloc_vm_area(npages * PAGE_SIZE);
+	if (area == NULL) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	ret = direct_kernel_remap_pfn_range(
+		(unsigned long)area->addr,
+		pdomain->buf_gmaddr >> PAGE_SHIFT,
+		npages * PAGE_SIZE, prot, DOMID_SELF);
+	if (ret) {
+		vunmap(area->addr);
+		goto out;
+	}
+	sbuf->buffer = area->addr;
+
+out:
+	return ret;
+}
+
+struct op_counter_config counter_config[OP_MAX_COUNTER];
+
+int xenoprof_create_files(struct super_block * sb, struct dentry * root)
+{
+	unsigned int i;
+
+	for (i = 0; i < num_events; ++i) {
+		struct dentry * dir;
+		char buf[2];
+ 
+		snprintf(buf, 2, "%d", i);
+		dir = oprofilefs_mkdir(sb, root, buf);
+		oprofilefs_create_ulong(sb, dir, "enabled",
+					&counter_config[i].enabled);
+		oprofilefs_create_ulong(sb, dir, "event",
+					&counter_config[i].event);
+		oprofilefs_create_ulong(sb, dir, "count",
+					&counter_config[i].count);
+		oprofilefs_create_ulong(sb, dir, "unit_mask",
+					&counter_config[i].unit_mask);
+		oprofilefs_create_ulong(sb, dir, "kernel",
+					&counter_config[i].kernel);
+		oprofilefs_create_ulong(sb, dir, "user",
+					&counter_config[i].user);
+	}
+
+	return 0;
+}
+
+int __init oprofile_arch_init(struct oprofile_operations * ops)
+{
+	return xenoprofile_init(ops);
+}
+
+void oprofile_arch_exit(void)
+{
+	xenoprofile_exit();
+}
diff -rpuN linux-2.6.18.8/arch/i386/pci/irq.c linux-2.6.18-xen-3.3.0/arch/i386/pci/irq.c
--- linux-2.6.18.8/arch/i386/pci/irq.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/pci/irq.c	2008-08-21 11:36:07.000000000 +0200
@@ -543,6 +543,16 @@ static __init int intel_router_probe(str
 		case PCI_DEVICE_ID_INTEL_ICH8_2:
 		case PCI_DEVICE_ID_INTEL_ICH8_3:
 		case PCI_DEVICE_ID_INTEL_ICH8_4:
+		case PCI_DEVICE_ID_INTEL_ICH9_0:
+		case PCI_DEVICE_ID_INTEL_ICH9_1:
+		case PCI_DEVICE_ID_INTEL_ICH9_2:
+		case PCI_DEVICE_ID_INTEL_ICH9_3:
+		case PCI_DEVICE_ID_INTEL_ICH9_4:
+		case PCI_DEVICE_ID_INTEL_ICH9_5:
+		case PCI_DEVICE_ID_INTEL_ICH10_0:
+		case PCI_DEVICE_ID_INTEL_ICH10_1:
+		case PCI_DEVICE_ID_INTEL_ICH10_2:
+		case PCI_DEVICE_ID_INTEL_ICH10_3:
 			r->name = "PIIX/ICH";
 			r->get = pirq_piix_get;
 			r->set = pirq_piix_set;
diff -rpuN linux-2.6.18.8/arch/i386/pci/irq-xen.c linux-2.6.18-xen-3.3.0/arch/i386/pci/irq-xen.c
--- linux-2.6.18.8/arch/i386/pci/irq-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/pci/irq-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1215 @@
+/*
+ *	Low-Level PCI Support for PC -- Routing of Interrupts
+ *
+ *	(c) 1999--2000 Martin Mares <mj@ucw.cz>
+ */
+
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/pci.h>
+#include <linux/init.h>
+#include <linux/slab.h>
+#include <linux/interrupt.h>
+#include <linux/dmi.h>
+#include <asm/io.h>
+#include <asm/smp.h>
+#include <asm/io_apic.h>
+#include <linux/irq.h>
+#include <linux/acpi.h>
+
+#include "pci.h"
+
+#define PIRQ_SIGNATURE	(('$' << 0) + ('P' << 8) + ('I' << 16) + ('R' << 24))
+#define PIRQ_VERSION 0x0100
+
+static int broken_hp_bios_irq9;
+static int acer_tm360_irqrouting;
+
+static struct irq_routing_table *pirq_table;
+
+static int pirq_enable_irq(struct pci_dev *dev);
+
+/*
+ * Never use: 0, 1, 2 (timer, keyboard, and cascade)
+ * Avoid using: 13, 14 and 15 (FP error and IDE).
+ * Penalize: 3, 4, 6, 7, 12 (known ISA uses: serial, floppy, parallel and mouse)
+ */
+unsigned int pcibios_irq_mask = 0xfff8;
+
+static int pirq_penalty[16] = {
+	1000000, 1000000, 1000000, 1000, 1000, 0, 1000, 1000,
+	0, 0, 0, 0, 1000, 100000, 100000, 100000
+};
+
+struct irq_router {
+	char *name;
+	u16 vendor, device;
+	int (*get)(struct pci_dev *router, struct pci_dev *dev, int pirq);
+	int (*set)(struct pci_dev *router, struct pci_dev *dev, int pirq, int new);
+};
+
+struct irq_router_handler {
+	u16 vendor;
+	int (*probe)(struct irq_router *r, struct pci_dev *router, u16 device);
+};
+
+int (*pcibios_enable_irq)(struct pci_dev *dev) = NULL;
+void (*pcibios_disable_irq)(struct pci_dev *dev) = NULL;
+
+/*
+ *  Check passed address for the PCI IRQ Routing Table signature
+ *  and perform checksum verification.
+ */
+
+static inline struct irq_routing_table * pirq_check_routing_table(u8 *addr)
+{
+	struct irq_routing_table *rt;
+	int i;
+	u8 sum;
+
+	rt = (struct irq_routing_table *) addr;
+	if (rt->signature != PIRQ_SIGNATURE ||
+	    rt->version != PIRQ_VERSION ||
+	    rt->size % 16 ||
+	    rt->size < sizeof(struct irq_routing_table))
+		return NULL;
+	sum = 0;
+	for (i=0; i < rt->size; i++)
+		sum += addr[i];
+	if (!sum) {
+		DBG(KERN_DEBUG "PCI: Interrupt Routing Table found at 0x%p\n", rt);
+		return rt;
+	}
+	return NULL;
+}
+
+
+
+/*
+ *  Search 0xf0000 -- 0xfffff for the PCI IRQ Routing Table.
+ */
+
+static struct irq_routing_table * __init pirq_find_routing_table(void)
+{
+	u8 *addr;
+	struct irq_routing_table *rt;
+
+#ifdef CONFIG_XEN
+	if (!is_initial_xendomain())
+		return NULL;
+#endif
+	if (pirq_table_addr) {
+		rt = pirq_check_routing_table((u8 *) isa_bus_to_virt(pirq_table_addr));
+		if (rt)
+			return rt;
+		printk(KERN_WARNING "PCI: PIRQ table NOT found at pirqaddr\n");
+	}
+	for(addr = (u8 *) isa_bus_to_virt(0xf0000); addr < (u8 *) isa_bus_to_virt(0x100000); addr += 16) {
+		rt = pirq_check_routing_table(addr);
+		if (rt)
+			return rt;
+	}
+	return NULL;
+}
+
+/*
+ *  If we have a IRQ routing table, use it to search for peer host
+ *  bridges.  It's a gross hack, but since there are no other known
+ *  ways how to get a list of buses, we have to go this way.
+ */
+
+static void __init pirq_peer_trick(void)
+{
+	struct irq_routing_table *rt = pirq_table;
+	u8 busmap[256];
+	int i;
+	struct irq_info *e;
+
+	memset(busmap, 0, sizeof(busmap));
+	for(i=0; i < (rt->size - sizeof(struct irq_routing_table)) / sizeof(struct irq_info); i++) {
+		e = &rt->slots[i];
+#ifdef DEBUG
+		{
+			int j;
+			DBG(KERN_DEBUG "%02x:%02x slot=%02x", e->bus, e->devfn/8, e->slot);
+			for(j=0; j<4; j++)
+				DBG(" %d:%02x/%04x", j, e->irq[j].link, e->irq[j].bitmap);
+			DBG("\n");
+		}
+#endif
+		busmap[e->bus] = 1;
+	}
+	for(i = 1; i < 256; i++) {
+		if (!busmap[i] || pci_find_bus(0, i))
+			continue;
+		if (pci_scan_bus(i, &pci_root_ops, NULL))
+			printk(KERN_INFO "PCI: Discovered primary peer bus %02x [IRQ]\n", i);
+	}
+	pcibios_last_bus = -1;
+}
+
+/*
+ *  Code for querying and setting of IRQ routes on various interrupt routers.
+ */
+
+void eisa_set_level_irq(unsigned int irq)
+{
+	unsigned char mask = 1 << (irq & 7);
+	unsigned int port = 0x4d0 + (irq >> 3);
+	unsigned char val;
+	static u16 eisa_irq_mask;
+
+	if (irq >= 16 || (1 << irq) & eisa_irq_mask)
+		return;
+
+	eisa_irq_mask |= (1 << irq);
+	printk(KERN_DEBUG "PCI: setting IRQ %u as level-triggered\n", irq);
+	val = inb(port);
+	if (!(val & mask)) {
+		DBG(KERN_DEBUG " -> edge");
+		outb(val | mask, port);
+	}
+}
+
+/*
+ * Common IRQ routing practice: nybbles in config space,
+ * offset by some magic constant.
+ */
+static unsigned int read_config_nybble(struct pci_dev *router, unsigned offset, unsigned nr)
+{
+	u8 x;
+	unsigned reg = offset + (nr >> 1);
+
+	pci_read_config_byte(router, reg, &x);
+	return (nr & 1) ? (x >> 4) : (x & 0xf);
+}
+
+static void write_config_nybble(struct pci_dev *router, unsigned offset, unsigned nr, unsigned int val)
+{
+	u8 x;
+	unsigned reg = offset + (nr >> 1);
+
+	pci_read_config_byte(router, reg, &x);
+	x = (nr & 1) ? ((x & 0x0f) | (val << 4)) : ((x & 0xf0) | val);
+	pci_write_config_byte(router, reg, x);
+}
+
+/*
+ * ALI pirq entries are damn ugly, and completely undocumented.
+ * This has been figured out from pirq tables, and it's not a pretty
+ * picture.
+ */
+static int pirq_ali_get(struct pci_dev *router, struct pci_dev *dev, int pirq)
+{
+	static const unsigned char irqmap[16] = { 0, 9, 3, 10, 4, 5, 7, 6, 1, 11, 0, 12, 0, 14, 0, 15 };
+
+	return irqmap[read_config_nybble(router, 0x48, pirq-1)];
+}
+
+static int pirq_ali_set(struct pci_dev *router, struct pci_dev *dev, int pirq, int irq)
+{
+	static const unsigned char irqmap[16] = { 0, 8, 0, 2, 4, 5, 7, 6, 0, 1, 3, 9, 11, 0, 13, 15 };
+	unsigned int val = irqmap[irq];
+		
+	if (val) {
+		write_config_nybble(router, 0x48, pirq-1, val);
+		return 1;
+	}
+	return 0;
+}
+
+/*
+ * The Intel PIIX4 pirq rules are fairly simple: "pirq" is
+ * just a pointer to the config space.
+ */
+static int pirq_piix_get(struct pci_dev *router, struct pci_dev *dev, int pirq)
+{
+	u8 x;
+
+	pci_read_config_byte(router, pirq, &x);
+	return (x < 16) ? x : 0;
+}
+
+static int pirq_piix_set(struct pci_dev *router, struct pci_dev *dev, int pirq, int irq)
+{
+	pci_write_config_byte(router, pirq, irq);
+	return 1;
+}
+
+/*
+ * The VIA pirq rules are nibble-based, like ALI,
+ * but without the ugly irq number munging.
+ * However, PIRQD is in the upper instead of lower 4 bits.
+ */
+static int pirq_via_get(struct pci_dev *router, struct pci_dev *dev, int pirq)
+{
+	return read_config_nybble(router, 0x55, pirq == 4 ? 5 : pirq);
+}
+
+static int pirq_via_set(struct pci_dev *router, struct pci_dev *dev, int pirq, int irq)
+{
+	write_config_nybble(router, 0x55, pirq == 4 ? 5 : pirq, irq);
+	return 1;
+}
+
+/*
+ * The VIA pirq rules are nibble-based, like ALI,
+ * but without the ugly irq number munging.
+ * However, for 82C586, nibble map is different .
+ */
+static int pirq_via586_get(struct pci_dev *router, struct pci_dev *dev, int pirq)
+{
+	static const unsigned int pirqmap[5] = { 3, 2, 5, 1, 1 };
+	return read_config_nybble(router, 0x55, pirqmap[pirq-1]);
+}
+
+static int pirq_via586_set(struct pci_dev *router, struct pci_dev *dev, int pirq, int irq)
+{
+	static const unsigned int pirqmap[5] = { 3, 2, 5, 1, 1 };
+	write_config_nybble(router, 0x55, pirqmap[pirq-1], irq);
+	return 1;
+}
+
+/*
+ * ITE 8330G pirq rules are nibble-based
+ * FIXME: pirqmap may be { 1, 0, 3, 2 },
+ * 	  2+3 are both mapped to irq 9 on my system
+ */
+static int pirq_ite_get(struct pci_dev *router, struct pci_dev *dev, int pirq)
+{
+	static const unsigned char pirqmap[4] = { 1, 0, 2, 3 };
+	return read_config_nybble(router,0x43, pirqmap[pirq-1]);
+}
+
+static int pirq_ite_set(struct pci_dev *router, struct pci_dev *dev, int pirq, int irq)
+{
+	static const unsigned char pirqmap[4] = { 1, 0, 2, 3 };
+	write_config_nybble(router, 0x43, pirqmap[pirq-1], irq);
+	return 1;
+}
+
+/*
+ * OPTI: high four bits are nibble pointer..
+ * I wonder what the low bits do?
+ */
+static int pirq_opti_get(struct pci_dev *router, struct pci_dev *dev, int pirq)
+{
+	return read_config_nybble(router, 0xb8, pirq >> 4);
+}
+
+static int pirq_opti_set(struct pci_dev *router, struct pci_dev *dev, int pirq, int irq)
+{
+	write_config_nybble(router, 0xb8, pirq >> 4, irq);
+	return 1;
+}
+
+/*
+ * Cyrix: nibble offset 0x5C
+ * 0x5C bits 7:4 is INTB bits 3:0 is INTA 
+ * 0x5D bits 7:4 is INTD bits 3:0 is INTC
+ */
+static int pirq_cyrix_get(struct pci_dev *router, struct pci_dev *dev, int pirq)
+{
+	return read_config_nybble(router, 0x5C, (pirq-1)^1);
+}
+
+static int pirq_cyrix_set(struct pci_dev *router, struct pci_dev *dev, int pirq, int irq)
+{
+	write_config_nybble(router, 0x5C, (pirq-1)^1, irq);
+	return 1;
+}
+
+/*
+ *	PIRQ routing for SiS 85C503 router used in several SiS chipsets.
+ *	We have to deal with the following issues here:
+ *	- vendors have different ideas about the meaning of link values
+ *	- some onboard devices (integrated in the chipset) have special
+ *	  links and are thus routed differently (i.e. not via PCI INTA-INTD)
+ *	- different revision of the router have a different layout for
+ *	  the routing registers, particularly for the onchip devices
+ *
+ *	For all routing registers the common thing is we have one byte
+ *	per routeable link which is defined as:
+ *		 bit 7      IRQ mapping enabled (0) or disabled (1)
+ *		 bits [6:4] reserved (sometimes used for onchip devices)
+ *		 bits [3:0] IRQ to map to
+ *		     allowed: 3-7, 9-12, 14-15
+ *		     reserved: 0, 1, 2, 8, 13
+ *
+ *	The config-space registers located at 0x41/0x42/0x43/0x44 are
+ *	always used to route the normal PCI INT A/B/C/D respectively.
+ *	Apparently there are systems implementing PCI routing table using
+ *	link values 0x01-0x04 and others using 0x41-0x44 for PCI INTA..D.
+ *	We try our best to handle both link mappings.
+ *	
+ *	Currently (2003-05-21) it appears most SiS chipsets follow the
+ *	definition of routing registers from the SiS-5595 southbridge.
+ *	According to the SiS 5595 datasheets the revision id's of the
+ *	router (ISA-bridge) should be 0x01 or 0xb0.
+ *
+ *	Furthermore we've also seen lspci dumps with revision 0x00 and 0xb1.
+ *	Looks like these are used in a number of SiS 5xx/6xx/7xx chipsets.
+ *	They seem to work with the current routing code. However there is
+ *	some concern because of the two USB-OHCI HCs (original SiS 5595
+ *	had only one). YMMV.
+ *
+ *	Onchip routing for router rev-id 0x01/0xb0 and probably 0x00/0xb1:
+ *
+ *	0x61:	IDEIRQ:
+ *		bits [6:5] must be written 01
+ *		bit 4 channel-select primary (0), secondary (1)
+ *
+ *	0x62:	USBIRQ:
+ *		bit 6 OHCI function disabled (0), enabled (1)
+ *	
+ *	0x6a:	ACPI/SCI IRQ: bits 4-6 reserved
+ *
+ *	0x7e:	Data Acq. Module IRQ - bits 4-6 reserved
+ *
+ *	We support USBIRQ (in addition to INTA-INTD) and keep the
+ *	IDE, ACPI and DAQ routing untouched as set by the BIOS.
+ *
+ *	Currently the only reported exception is the new SiS 65x chipset
+ *	which includes the SiS 69x southbridge. Here we have the 85C503
+ *	router revision 0x04 and there are changes in the register layout
+ *	mostly related to the different USB HCs with USB 2.0 support.
+ *
+ *	Onchip routing for router rev-id 0x04 (try-and-error observation)
+ *
+ *	0x60/0x61/0x62/0x63:	1xEHCI and 3xOHCI (companion) USB-HCs
+ *				bit 6-4 are probably unused, not like 5595
+ */
+
+#define PIRQ_SIS_IRQ_MASK	0x0f
+#define PIRQ_SIS_IRQ_DISABLE	0x80
+#define PIRQ_SIS_USB_ENABLE	0x40
+
+static int pirq_sis_get(struct pci_dev *router, struct pci_dev *dev, int pirq)
+{
+	u8 x;
+	int reg;
+
+	reg = pirq;
+	if (reg >= 0x01 && reg <= 0x04)
+		reg += 0x40;
+	pci_read_config_byte(router, reg, &x);
+	return (x & PIRQ_SIS_IRQ_DISABLE) ? 0 : (x & PIRQ_SIS_IRQ_MASK);
+}
+
+static int pirq_sis_set(struct pci_dev *router, struct pci_dev *dev, int pirq, int irq)
+{
+	u8 x;
+	int reg;
+
+	reg = pirq;
+	if (reg >= 0x01 && reg <= 0x04)
+		reg += 0x40;
+	pci_read_config_byte(router, reg, &x);
+	x &= ~(PIRQ_SIS_IRQ_MASK | PIRQ_SIS_IRQ_DISABLE);
+	x |= irq ? irq: PIRQ_SIS_IRQ_DISABLE;
+	pci_write_config_byte(router, reg, x);
+	return 1;
+}
+
+
+/*
+ * VLSI: nibble offset 0x74 - educated guess due to routing table and
+ *       config space of VLSI 82C534 PCI-bridge/router (1004:0102)
+ *       Tested on HP OmniBook 800 covering PIRQ 1, 2, 4, 8 for onboard
+ *       devices, PIRQ 3 for non-pci(!) soundchip and (untested) PIRQ 6
+ *       for the busbridge to the docking station.
+ */
+
+static int pirq_vlsi_get(struct pci_dev *router, struct pci_dev *dev, int pirq)
+{
+	if (pirq > 8) {
+		printk(KERN_INFO "VLSI router pirq escape (%d)\n", pirq);
+		return 0;
+	}
+	return read_config_nybble(router, 0x74, pirq-1);
+}
+
+static int pirq_vlsi_set(struct pci_dev *router, struct pci_dev *dev, int pirq, int irq)
+{
+	if (pirq > 8) {
+		printk(KERN_INFO "VLSI router pirq escape (%d)\n", pirq);
+		return 0;
+	}
+	write_config_nybble(router, 0x74, pirq-1, irq);
+	return 1;
+}
+
+/*
+ * ServerWorks: PCI interrupts mapped to system IRQ lines through Index
+ * and Redirect I/O registers (0x0c00 and 0x0c01).  The Index register
+ * format is (PCIIRQ## | 0x10), e.g.: PCIIRQ10=0x1a.  The Redirect
+ * register is a straight binary coding of desired PIC IRQ (low nibble).
+ *
+ * The 'link' value in the PIRQ table is already in the correct format
+ * for the Index register.  There are some special index values:
+ * 0x00 for ACPI (SCI), 0x01 for USB, 0x02 for IDE0, 0x04 for IDE1,
+ * and 0x03 for SMBus.
+ */
+static int pirq_serverworks_get(struct pci_dev *router, struct pci_dev *dev, int pirq)
+{
+	outb_p(pirq, 0xc00);
+	return inb(0xc01) & 0xf;
+}
+
+static int pirq_serverworks_set(struct pci_dev *router, struct pci_dev *dev, int pirq, int irq)
+{
+	outb_p(pirq, 0xc00);
+	outb_p(irq, 0xc01);
+	return 1;
+}
+
+/* Support for AMD756 PCI IRQ Routing
+ * Jhon H. Caicedo <jhcaiced@osso.org.co>
+ * Jun/21/2001 0.2.0 Release, fixed to use "nybble" functions... (jhcaiced)
+ * Jun/19/2001 Alpha Release 0.1.0 (jhcaiced)
+ * The AMD756 pirq rules are nibble-based
+ * offset 0x56 0-3 PIRQA  4-7  PIRQB
+ * offset 0x57 0-3 PIRQC  4-7  PIRQD
+ */
+static int pirq_amd756_get(struct pci_dev *router, struct pci_dev *dev, int pirq)
+{
+	u8 irq;
+	irq = 0;
+	if (pirq <= 4)
+	{
+		irq = read_config_nybble(router, 0x56, pirq - 1);
+	}
+	printk(KERN_INFO "AMD756: dev %04x:%04x, router pirq : %d get irq : %2d\n",
+		dev->vendor, dev->device, pirq, irq);
+	return irq;
+}
+
+static int pirq_amd756_set(struct pci_dev *router, struct pci_dev *dev, int pirq, int irq)
+{
+	printk(KERN_INFO "AMD756: dev %04x:%04x, router pirq : %d SET irq : %2d\n", 
+		dev->vendor, dev->device, pirq, irq);
+	if (pirq <= 4)
+	{
+		write_config_nybble(router, 0x56, pirq - 1, irq);
+	}
+	return 1;
+}
+
+#ifdef CONFIG_PCI_BIOS
+
+static int pirq_bios_set(struct pci_dev *router, struct pci_dev *dev, int pirq, int irq)
+{
+	struct pci_dev *bridge;
+	int pin = pci_get_interrupt_pin(dev, &bridge);
+	return pcibios_set_irq_routing(bridge, pin, irq);
+}
+
+#endif
+
+static __init int intel_router_probe(struct irq_router *r, struct pci_dev *router, u16 device)
+{
+	static struct pci_device_id __initdata pirq_440gx[] = {
+		{ PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_82443GX_0) },
+		{ PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_82443GX_2) },
+		{ },
+	};
+
+	/* 440GX has a proprietary PIRQ router -- don't use it */
+	if (pci_dev_present(pirq_440gx))
+		return 0;
+
+	switch(device)
+	{
+		case PCI_DEVICE_ID_INTEL_82371FB_0:
+		case PCI_DEVICE_ID_INTEL_82371SB_0:
+		case PCI_DEVICE_ID_INTEL_82371AB_0:
+		case PCI_DEVICE_ID_INTEL_82371MX:
+		case PCI_DEVICE_ID_INTEL_82443MX_0:
+		case PCI_DEVICE_ID_INTEL_82801AA_0:
+		case PCI_DEVICE_ID_INTEL_82801AB_0:
+		case PCI_DEVICE_ID_INTEL_82801BA_0:
+		case PCI_DEVICE_ID_INTEL_82801BA_10:
+		case PCI_DEVICE_ID_INTEL_82801CA_0:
+		case PCI_DEVICE_ID_INTEL_82801CA_12:
+		case PCI_DEVICE_ID_INTEL_82801DB_0:
+		case PCI_DEVICE_ID_INTEL_82801E_0:
+		case PCI_DEVICE_ID_INTEL_82801EB_0:
+		case PCI_DEVICE_ID_INTEL_ESB_1:
+		case PCI_DEVICE_ID_INTEL_ICH6_0:
+		case PCI_DEVICE_ID_INTEL_ICH6_1:
+		case PCI_DEVICE_ID_INTEL_ICH7_0:
+		case PCI_DEVICE_ID_INTEL_ICH7_1:
+		case PCI_DEVICE_ID_INTEL_ICH7_30:
+		case PCI_DEVICE_ID_INTEL_ICH7_31:
+		case PCI_DEVICE_ID_INTEL_ESB2_0:
+		case PCI_DEVICE_ID_INTEL_ICH8_0:
+		case PCI_DEVICE_ID_INTEL_ICH8_1:
+		case PCI_DEVICE_ID_INTEL_ICH8_2:
+		case PCI_DEVICE_ID_INTEL_ICH8_3:
+		case PCI_DEVICE_ID_INTEL_ICH8_4:
+		case PCI_DEVICE_ID_INTEL_ICH9_0:
+		case PCI_DEVICE_ID_INTEL_ICH9_1:
+		case PCI_DEVICE_ID_INTEL_ICH9_2:
+		case PCI_DEVICE_ID_INTEL_ICH9_3:
+		case PCI_DEVICE_ID_INTEL_ICH9_4:
+		case PCI_DEVICE_ID_INTEL_ICH9_5:
+		case PCI_DEVICE_ID_INTEL_ICH10_0:
+		case PCI_DEVICE_ID_INTEL_ICH10_1:
+		case PCI_DEVICE_ID_INTEL_ICH10_2:
+		case PCI_DEVICE_ID_INTEL_ICH10_3:
+			r->name = "PIIX/ICH";
+			r->get = pirq_piix_get;
+			r->set = pirq_piix_set;
+			return 1;
+	}
+	return 0;
+}
+
+static __init int via_router_probe(struct irq_router *r,
+				struct pci_dev *router, u16 device)
+{
+	/* FIXME: We should move some of the quirk fixup stuff here */
+
+	/*
+	 * work arounds for some buggy BIOSes
+	 */
+	if (device == PCI_DEVICE_ID_VIA_82C586_0) {
+		switch(router->device) {
+		case PCI_DEVICE_ID_VIA_82C686:
+			/*
+			 * Asus k7m bios wrongly reports 82C686A
+			 * as 586-compatible
+			 */
+			device = PCI_DEVICE_ID_VIA_82C686;
+			break;
+		case PCI_DEVICE_ID_VIA_8235:
+			/**
+			 * Asus a7v-x bios wrongly reports 8235
+			 * as 586-compatible
+			 */
+			device = PCI_DEVICE_ID_VIA_8235;
+			break;
+		}
+	}
+
+	switch(device) {
+	case PCI_DEVICE_ID_VIA_82C586_0:
+		r->name = "VIA";
+		r->get = pirq_via586_get;
+		r->set = pirq_via586_set;
+		return 1;
+	case PCI_DEVICE_ID_VIA_82C596:
+	case PCI_DEVICE_ID_VIA_82C686:
+	case PCI_DEVICE_ID_VIA_8231:
+	case PCI_DEVICE_ID_VIA_8233A:
+	case PCI_DEVICE_ID_VIA_8235:
+	case PCI_DEVICE_ID_VIA_8237:
+		/* FIXME: add new ones for 8233/5 */
+		r->name = "VIA";
+		r->get = pirq_via_get;
+		r->set = pirq_via_set;
+		return 1;
+	}
+	return 0;
+}
+
+static __init int vlsi_router_probe(struct irq_router *r, struct pci_dev *router, u16 device)
+{
+	switch(device)
+	{
+		case PCI_DEVICE_ID_VLSI_82C534:
+			r->name = "VLSI 82C534";
+			r->get = pirq_vlsi_get;
+			r->set = pirq_vlsi_set;
+			return 1;
+	}
+	return 0;
+}
+
+
+static __init int serverworks_router_probe(struct irq_router *r, struct pci_dev *router, u16 device)
+{
+	switch(device)
+	{
+		case PCI_DEVICE_ID_SERVERWORKS_OSB4:
+		case PCI_DEVICE_ID_SERVERWORKS_CSB5:
+			r->name = "ServerWorks";
+			r->get = pirq_serverworks_get;
+			r->set = pirq_serverworks_set;
+			return 1;
+	}
+	return 0;
+}
+
+static __init int sis_router_probe(struct irq_router *r, struct pci_dev *router, u16 device)
+{
+	if (device != PCI_DEVICE_ID_SI_503)
+		return 0;
+		
+	r->name = "SIS";
+	r->get = pirq_sis_get;
+	r->set = pirq_sis_set;
+	return 1;
+}
+
+static __init int cyrix_router_probe(struct irq_router *r, struct pci_dev *router, u16 device)
+{
+	switch(device)
+	{
+		case PCI_DEVICE_ID_CYRIX_5520:
+			r->name = "NatSemi";
+			r->get = pirq_cyrix_get;
+			r->set = pirq_cyrix_set;
+			return 1;
+	}
+	return 0;
+}
+
+static __init int opti_router_probe(struct irq_router *r, struct pci_dev *router, u16 device)
+{
+	switch(device)
+	{
+		case PCI_DEVICE_ID_OPTI_82C700:
+			r->name = "OPTI";
+			r->get = pirq_opti_get;
+			r->set = pirq_opti_set;
+			return 1;
+	}
+	return 0;
+}
+
+static __init int ite_router_probe(struct irq_router *r, struct pci_dev *router, u16 device)
+{
+	switch(device)
+	{
+		case PCI_DEVICE_ID_ITE_IT8330G_0:
+			r->name = "ITE";
+			r->get = pirq_ite_get;
+			r->set = pirq_ite_set;
+			return 1;
+	}
+	return 0;
+}
+
+static __init int ali_router_probe(struct irq_router *r, struct pci_dev *router, u16 device)
+{
+	switch(device)
+	{
+	case PCI_DEVICE_ID_AL_M1533:
+	case PCI_DEVICE_ID_AL_M1563:
+		printk(KERN_DEBUG "PCI: Using ALI IRQ Router\n");
+		r->name = "ALI";
+		r->get = pirq_ali_get;
+		r->set = pirq_ali_set;
+		return 1;
+	}
+	return 0;
+}
+
+static __init int amd_router_probe(struct irq_router *r, struct pci_dev *router, u16 device)
+{
+	switch(device)
+	{
+		case PCI_DEVICE_ID_AMD_VIPER_740B:
+			r->name = "AMD756";
+			break;
+		case PCI_DEVICE_ID_AMD_VIPER_7413:
+			r->name = "AMD766";
+			break;
+		case PCI_DEVICE_ID_AMD_VIPER_7443:
+			r->name = "AMD768";
+			break;
+		default:
+			return 0;
+	}
+	r->get = pirq_amd756_get;
+	r->set = pirq_amd756_set;
+	return 1;
+}
+		
+static __initdata struct irq_router_handler pirq_routers[] = {
+	{ PCI_VENDOR_ID_INTEL, intel_router_probe },
+	{ PCI_VENDOR_ID_AL, ali_router_probe },
+	{ PCI_VENDOR_ID_ITE, ite_router_probe },
+	{ PCI_VENDOR_ID_VIA, via_router_probe },
+	{ PCI_VENDOR_ID_OPTI, opti_router_probe },
+	{ PCI_VENDOR_ID_SI, sis_router_probe },
+	{ PCI_VENDOR_ID_CYRIX, cyrix_router_probe },
+	{ PCI_VENDOR_ID_VLSI, vlsi_router_probe },
+	{ PCI_VENDOR_ID_SERVERWORKS, serverworks_router_probe },
+	{ PCI_VENDOR_ID_AMD, amd_router_probe },
+	/* Someone with docs needs to add the ATI Radeon IGP */
+	{ 0, NULL }
+};
+static struct irq_router pirq_router;
+static struct pci_dev *pirq_router_dev;
+
+
+/*
+ *	FIXME: should we have an option to say "generic for
+ *	chipset" ?
+ */
+ 
+static void __init pirq_find_router(struct irq_router *r)
+{
+	struct irq_routing_table *rt = pirq_table;
+	struct irq_router_handler *h;
+
+#ifdef CONFIG_PCI_BIOS
+	if (!rt->signature) {
+		printk(KERN_INFO "PCI: Using BIOS for IRQ routing\n");
+		r->set = pirq_bios_set;
+		r->name = "BIOS";
+		return;
+	}
+#endif
+
+	/* Default unless a driver reloads it */
+	r->name = "default";
+	r->get = NULL;
+	r->set = NULL;
+	
+	DBG(KERN_DEBUG "PCI: Attempting to find IRQ router for %04x:%04x\n",
+	    rt->rtr_vendor, rt->rtr_device);
+
+	pirq_router_dev = pci_find_slot(rt->rtr_bus, rt->rtr_devfn);
+	if (!pirq_router_dev) {
+		DBG(KERN_DEBUG "PCI: Interrupt router not found at "
+			"%02x:%02x\n", rt->rtr_bus, rt->rtr_devfn);
+		return;
+	}
+
+	for( h = pirq_routers; h->vendor; h++) {
+		/* First look for a router match */
+		if (rt->rtr_vendor == h->vendor && h->probe(r, pirq_router_dev, rt->rtr_device))
+			break;
+		/* Fall back to a device match */
+		if (pirq_router_dev->vendor == h->vendor && h->probe(r, pirq_router_dev, pirq_router_dev->device))
+			break;
+	}
+	printk(KERN_INFO "PCI: Using IRQ router %s [%04x/%04x] at %s\n",
+		pirq_router.name,
+		pirq_router_dev->vendor,
+		pirq_router_dev->device,
+		pci_name(pirq_router_dev));
+}
+
+static struct irq_info *pirq_get_info(struct pci_dev *dev)
+{
+	struct irq_routing_table *rt = pirq_table;
+	int entries = (rt->size - sizeof(struct irq_routing_table)) / sizeof(struct irq_info);
+	struct irq_info *info;
+
+	for (info = rt->slots; entries--; info++)
+		if (info->bus == dev->bus->number && PCI_SLOT(info->devfn) == PCI_SLOT(dev->devfn))
+			return info;
+	return NULL;
+}
+
+static int pcibios_lookup_irq(struct pci_dev *dev, int assign)
+{
+	u8 pin;
+	struct irq_info *info;
+	int i, pirq, newirq;
+	int irq = 0;
+	u32 mask;
+	struct irq_router *r = &pirq_router;
+	struct pci_dev *dev2 = NULL;
+	char *msg = NULL;
+
+	/* Find IRQ pin */
+	pci_read_config_byte(dev, PCI_INTERRUPT_PIN, &pin);
+	if (!pin) {
+		DBG(KERN_DEBUG " -> no interrupt pin\n");
+		return 0;
+	}
+	pin = pin - 1;
+
+	/* Find IRQ routing entry */
+
+	if (!pirq_table)
+		return 0;
+	
+	DBG(KERN_DEBUG "IRQ for %s[%c]", pci_name(dev), 'A' + pin);
+	info = pirq_get_info(dev);
+	if (!info) {
+		DBG(" -> not found in routing table\n" KERN_DEBUG);
+		return 0;
+	}
+	pirq = info->irq[pin].link;
+	mask = info->irq[pin].bitmap;
+	if (!pirq) {
+		DBG(" -> not routed\n" KERN_DEBUG);
+		return 0;
+	}
+	DBG(" -> PIRQ %02x, mask %04x, excl %04x", pirq, mask, pirq_table->exclusive_irqs);
+	mask &= pcibios_irq_mask;
+
+	/* Work around broken HP Pavilion Notebooks which assign USB to
+	   IRQ 9 even though it is actually wired to IRQ 11 */
+
+	if (broken_hp_bios_irq9 && pirq == 0x59 && dev->irq == 9) {
+		dev->irq = 11;
+		pci_write_config_byte(dev, PCI_INTERRUPT_LINE, 11);
+		r->set(pirq_router_dev, dev, pirq, 11);
+	}
+
+	/* same for Acer Travelmate 360, but with CB and irq 11 -> 10 */
+	if (acer_tm360_irqrouting && dev->irq == 11 && dev->vendor == PCI_VENDOR_ID_O2) {
+		pirq = 0x68;
+		mask = 0x400;
+		dev->irq = r->get(pirq_router_dev, dev, pirq);
+		pci_write_config_byte(dev, PCI_INTERRUPT_LINE, dev->irq);
+	}
+
+	/*
+	 * Find the best IRQ to assign: use the one
+	 * reported by the device if possible.
+	 */
+	newirq = dev->irq;
+	if (newirq && !((1 << newirq) & mask)) {
+		if ( pci_probe & PCI_USE_PIRQ_MASK) newirq = 0;
+		else printk("\n" KERN_WARNING
+			"PCI: IRQ %i for device %s doesn't match PIRQ mask "
+			"- try pci=usepirqmask\n" KERN_DEBUG, newirq,
+			pci_name(dev));
+	}
+	if (!newirq && assign) {
+		for (i = 0; i < 16; i++) {
+			if (!(mask & (1 << i)))
+				continue;
+			if (pirq_penalty[i] < pirq_penalty[newirq] && can_request_irq(i, IRQF_SHARED))
+				newirq = i;
+		}
+	}
+	DBG(" -> newirq=%d", newirq);
+
+	/* Check if it is hardcoded */
+	if ((pirq & 0xf0) == 0xf0) {
+		irq = pirq & 0xf;
+		DBG(" -> hardcoded IRQ %d\n", irq);
+		msg = "Hardcoded";
+	} else if ( r->get && (irq = r->get(pirq_router_dev, dev, pirq)) && \
+	((!(pci_probe & PCI_USE_PIRQ_MASK)) || ((1 << irq) & mask)) ) {
+		DBG(" -> got IRQ %d\n", irq);
+		msg = "Found";
+		eisa_set_level_irq(irq);
+	} else if (newirq && r->set && (dev->class >> 8) != PCI_CLASS_DISPLAY_VGA) {
+		DBG(" -> assigning IRQ %d", newirq);
+		if (r->set(pirq_router_dev, dev, pirq, newirq)) {
+			eisa_set_level_irq(newirq);
+			DBG(" ... OK\n");
+			msg = "Assigned";
+			irq = newirq;
+		}
+	}
+
+	if (!irq) {
+		DBG(" ... failed\n");
+		if (newirq && mask == (1 << newirq)) {
+			msg = "Guessed";
+			irq = newirq;
+		} else
+			return 0;
+	}
+	printk(KERN_INFO "PCI: %s IRQ %d for device %s\n", msg, irq, pci_name(dev));
+
+	/* Update IRQ for all devices with the same pirq value */
+	while ((dev2 = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev2)) != NULL) {
+		pci_read_config_byte(dev2, PCI_INTERRUPT_PIN, &pin);
+		if (!pin)
+			continue;
+		pin--;
+		info = pirq_get_info(dev2);
+		if (!info)
+			continue;
+		if (info->irq[pin].link == pirq) {
+			/* We refuse to override the dev->irq information. Give a warning! */
+		    	if ( dev2->irq && dev2->irq != irq && \
+			(!(pci_probe & PCI_USE_PIRQ_MASK) || \
+			((1 << dev2->irq) & mask)) ) {
+#ifndef CONFIG_PCI_MSI
+		    		printk(KERN_INFO "IRQ routing conflict for %s, have irq %d, want irq %d\n",
+				       pci_name(dev2), dev2->irq, irq);
+#endif
+		    		continue;
+		    	}
+			dev2->irq = irq;
+			pirq_penalty[irq]++;
+			if (dev != dev2)
+				printk(KERN_INFO "PCI: Sharing IRQ %d with %s\n", irq, pci_name(dev2));
+		}
+	}
+	return 1;
+}
+
+static void __init pcibios_fixup_irqs(void)
+{
+	struct pci_dev *dev = NULL;
+	u8 pin;
+
+	DBG(KERN_DEBUG "PCI: IRQ fixup\n");
+	while ((dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev)) != NULL) {
+		/*
+		 * If the BIOS has set an out of range IRQ number, just ignore it.
+		 * Also keep track of which IRQ's are already in use.
+		 */
+		if (dev->irq >= 16) {
+			DBG(KERN_DEBUG "%s: ignoring bogus IRQ %d\n", pci_name(dev), dev->irq);
+			dev->irq = 0;
+		}
+		/* If the IRQ is already assigned to a PCI device, ignore its ISA use penalty */
+		if (pirq_penalty[dev->irq] >= 100 && pirq_penalty[dev->irq] < 100000)
+			pirq_penalty[dev->irq] = 0;
+		pirq_penalty[dev->irq]++;
+	}
+
+	dev = NULL;
+	while ((dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev)) != NULL) {
+		pci_read_config_byte(dev, PCI_INTERRUPT_PIN, &pin);
+#ifdef CONFIG_X86_IO_APIC
+		/*
+		 * Recalculate IRQ numbers if we use the I/O APIC.
+		 */
+		if (io_apic_assign_pci_irqs)
+		{
+			int irq;
+
+			if (pin) {
+				pin--;		/* interrupt pins are numbered starting from 1 */
+				irq = IO_APIC_get_PCI_irq_vector(dev->bus->number, PCI_SLOT(dev->devfn), pin);
+	/*
+	 * Busses behind bridges are typically not listed in the MP-table.
+	 * In this case we have to look up the IRQ based on the parent bus,
+	 * parent slot, and pin number. The SMP code detects such bridged
+	 * busses itself so we should get into this branch reliably.
+	 */
+				if (irq < 0 && dev->bus->parent) { /* go back to the bridge */
+					struct pci_dev * bridge = dev->bus->self;
+
+					pin = (pin + PCI_SLOT(dev->devfn)) % 4;
+					irq = IO_APIC_get_PCI_irq_vector(bridge->bus->number, 
+							PCI_SLOT(bridge->devfn), pin);
+					if (irq >= 0)
+						printk(KERN_WARNING "PCI: using PPB %s[%c] to get irq %d\n",
+							pci_name(bridge), 'A' + pin, irq);
+				}
+				if (irq >= 0) {
+					if (use_pci_vector() &&
+						!platform_legacy_irq(irq))
+						irq = IO_APIC_VECTOR(irq);
+
+					printk(KERN_INFO "PCI->APIC IRQ transform: %s[%c] -> IRQ %d\n",
+						pci_name(dev), 'A' + pin, irq);
+					dev->irq = irq;
+				}
+			}
+		}
+#endif
+		/*
+		 * Still no IRQ? Try to lookup one...
+		 */
+		if (pin && !dev->irq)
+			pcibios_lookup_irq(dev, 0);
+	}
+}
+
+/*
+ * Work around broken HP Pavilion Notebooks which assign USB to
+ * IRQ 9 even though it is actually wired to IRQ 11
+ */
+static int __init fix_broken_hp_bios_irq9(struct dmi_system_id *d)
+{
+	if (!broken_hp_bios_irq9) {
+		broken_hp_bios_irq9 = 1;
+		printk(KERN_INFO "%s detected - fixing broken IRQ routing\n", d->ident);
+	}
+	return 0;
+}
+
+/*
+ * Work around broken Acer TravelMate 360 Notebooks which assign
+ * Cardbus to IRQ 11 even though it is actually wired to IRQ 10
+ */
+static int __init fix_acer_tm360_irqrouting(struct dmi_system_id *d)
+{
+	if (!acer_tm360_irqrouting) {
+		acer_tm360_irqrouting = 1;
+		printk(KERN_INFO "%s detected - fixing broken IRQ routing\n", d->ident);
+	}
+	return 0;
+}
+
+static struct dmi_system_id __initdata pciirq_dmi_table[] = {
+	{
+		.callback = fix_broken_hp_bios_irq9,
+		.ident = "HP Pavilion N5400 Series Laptop",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Hewlett-Packard"),
+			DMI_MATCH(DMI_BIOS_VERSION, "GE.M1.03"),
+			DMI_MATCH(DMI_PRODUCT_VERSION, "HP Pavilion Notebook Model GE"),
+			DMI_MATCH(DMI_BOARD_VERSION, "OmniBook N32N-736"),
+		},
+	},
+	{
+		.callback = fix_acer_tm360_irqrouting,
+		.ident = "Acer TravelMate 36x Laptop",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Acer"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "TravelMate 360"),
+		},
+	},
+	{ }
+};
+
+static int __init pcibios_irq_init(void)
+{
+	DBG(KERN_DEBUG "PCI: IRQ init\n");
+
+	if (pcibios_enable_irq || raw_pci_ops == NULL)
+		return 0;
+
+	dmi_check_system(pciirq_dmi_table);
+
+	pirq_table = pirq_find_routing_table();
+
+#ifdef CONFIG_PCI_BIOS
+	if (!pirq_table && (pci_probe & PCI_BIOS_IRQ_SCAN))
+		pirq_table = pcibios_get_irq_routing_table();
+#endif
+	if (pirq_table) {
+		pirq_peer_trick();
+		pirq_find_router(&pirq_router);
+		if (pirq_table->exclusive_irqs) {
+			int i;
+			for (i=0; i<16; i++)
+				if (!(pirq_table->exclusive_irqs & (1 << i)))
+					pirq_penalty[i] += 100;
+		}
+		/* If we're using the I/O APIC, avoid using the PCI IRQ routing table */
+		if (io_apic_assign_pci_irqs)
+			pirq_table = NULL;
+	}
+
+	pcibios_enable_irq = pirq_enable_irq;
+
+	pcibios_fixup_irqs();
+	return 0;
+}
+
+subsys_initcall(pcibios_irq_init);
+
+
+static void pirq_penalize_isa_irq(int irq, int active)
+{
+	/*
+	 *  If any ISAPnP device reports an IRQ in its list of possible
+	 *  IRQ's, we try to avoid assigning it to PCI devices.
+	 */
+	if (irq < 16) {
+		if (active)
+			pirq_penalty[irq] += 1000;
+		else
+			pirq_penalty[irq] += 100;
+	}
+}
+
+void pcibios_penalize_isa_irq(int irq, int active)
+{
+#ifdef CONFIG_ACPI
+	if (!acpi_noirq)
+		acpi_penalize_isa_irq(irq, active);
+	else
+#endif
+		pirq_penalize_isa_irq(irq, active);
+}
+
+static int pirq_enable_irq(struct pci_dev *dev)
+{
+	u8 pin;
+	struct pci_dev *temp_dev;
+
+	pci_read_config_byte(dev, PCI_INTERRUPT_PIN, &pin);
+	if (pin && !pcibios_lookup_irq(dev, 1) && !dev->irq) {
+		char *msg = "";
+
+		pin--;		/* interrupt pins are numbered starting from 1 */
+
+		if (io_apic_assign_pci_irqs) {
+			int irq;
+
+			irq = IO_APIC_get_PCI_irq_vector(dev->bus->number, PCI_SLOT(dev->devfn), pin);
+			/*
+			 * Busses behind bridges are typically not listed in the MP-table.
+			 * In this case we have to look up the IRQ based on the parent bus,
+			 * parent slot, and pin number. The SMP code detects such bridged
+			 * busses itself so we should get into this branch reliably.
+			 */
+			temp_dev = dev;
+			while (irq < 0 && dev->bus->parent) { /* go back to the bridge */
+				struct pci_dev * bridge = dev->bus->self;
+
+				pin = (pin + PCI_SLOT(dev->devfn)) % 4;
+				irq = IO_APIC_get_PCI_irq_vector(bridge->bus->number, 
+						PCI_SLOT(bridge->devfn), pin);
+				if (irq >= 0)
+					printk(KERN_WARNING "PCI: using PPB %s[%c] to get irq %d\n",
+						pci_name(bridge), 'A' + pin, irq);
+				dev = bridge;
+			}
+			dev = temp_dev;
+			if (irq >= 0) {
+#ifdef CONFIG_PCI_MSI
+				if (!platform_legacy_irq(irq))
+					irq = IO_APIC_VECTOR(irq);
+#endif
+				printk(KERN_INFO "PCI->APIC IRQ transform: %s[%c] -> IRQ %d\n",
+					pci_name(dev), 'A' + pin, irq);
+				dev->irq = irq;
+				return 0;
+			} else
+				msg = " Probably buggy MP table.";
+		} else if (pci_probe & PCI_BIOS_IRQ_SCAN)
+			msg = "";
+		else
+			msg = " Please try using pci=biosirq.";
+
+		/* With IDE legacy devices the IRQ lookup failure is not a problem.. */
+		if (dev->class >> 8 == PCI_CLASS_STORAGE_IDE && !(dev->class & 0x5))
+			return 0;
+
+		printk(KERN_WARNING "PCI: No IRQ known for interrupt pin %c of device %s.%s\n",
+		       'A' + pin, pci_name(dev), msg);
+	}
+	return 0;
+}
+
+int pci_vector_resources(int last, int nr_released)
+{
+	int count = nr_released;
+
+	int next = last;
+	int offset = (last % 8);
+
+	while (next < FIRST_SYSTEM_VECTOR) {
+		next += 8;
+#ifdef CONFIG_X86_64
+		if (next == IA32_SYSCALL_VECTOR)
+			continue;
+#else
+		if (next == SYSCALL_VECTOR)
+			continue;
+#endif
+		count++;
+		if (next >= FIRST_SYSTEM_VECTOR) {
+			if (offset%8) {
+				next = FIRST_DEVICE_VECTOR + offset;
+				offset++;
+				continue;
+			}
+			count--;
+		}
+	}
+
+	return count;
+}
diff -rpuN linux-2.6.18.8/arch/i386/pci/Makefile linux-2.6.18-xen-3.3.0/arch/i386/pci/Makefile
--- linux-2.6.18.8/arch/i386/pci/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/pci/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -4,6 +4,10 @@ obj-$(CONFIG_PCI_BIOS)		+= pcbios.o
 obj-$(CONFIG_PCI_MMCONFIG)	+= mmconfig.o direct.o
 obj-$(CONFIG_PCI_DIRECT)	+= direct.o
 
+# pcifront should be after pcbios.o, mmconfig.o, and direct.o as it should only
+# take over if direct access to the PCI bus is unavailable
+obj-$(CONFIG_XEN_PCIDEV_FRONTEND)	+= pcifront.o
+
 pci-y				:= fixup.o
 pci-$(CONFIG_ACPI)		+= acpi.o
 pci-y				+= legacy.o irq.o
diff -rpuN linux-2.6.18.8/arch/i386/pci/pcifront.c linux-2.6.18-xen-3.3.0/arch/i386/pci/pcifront.c
--- linux-2.6.18.8/arch/i386/pci/pcifront.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/pci/pcifront.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,55 @@
+/*
+ * PCI Frontend Stub - puts some "dummy" functions in to the Linux x86 PCI core
+ *                     to support the Xen PCI Frontend's operation
+ *
+ *   Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/pci.h>
+#include <asm/acpi.h>
+#include "pci.h"
+
+static int pcifront_enable_irq(struct pci_dev *dev)
+{
+	u8 irq;
+	pci_read_config_byte(dev, PCI_INTERRUPT_LINE, &irq);
+	dev->irq = irq;
+
+	return 0;
+}
+
+extern u8 pci_cache_line_size;
+
+static int __init pcifront_x86_stub_init(void)
+{
+	struct cpuinfo_x86 *c = &boot_cpu_data;
+
+	/* Only install our method if we haven't found real hardware already */
+	if (raw_pci_ops)
+		return 0;
+
+	printk(KERN_INFO "PCI: setting up Xen PCI frontend stub\n");
+
+	/* Copied from arch/i386/pci/common.c */
+	pci_cache_line_size = 32 >> 2;
+	if (c->x86 >= 6 && c->x86_vendor == X86_VENDOR_AMD)
+		pci_cache_line_size = 64 >> 2;	/* K7 & K8 */
+	else if (c->x86 > 6 && c->x86_vendor == X86_VENDOR_INTEL)
+		pci_cache_line_size = 128 >> 2;	/* P4 */
+
+	/* On x86, we need to disable the normal IRQ routing table and
+	 * just ask the backend
+	 */
+	pcibios_enable_irq = pcifront_enable_irq;
+	pcibios_disable_irq = NULL;
+
+#ifdef CONFIG_ACPI
+	/* Keep ACPI out of the picture */
+	acpi_noirq = 1;
+#endif
+
+	return 0;
+}
+
+arch_initcall(pcifront_x86_stub_init);
diff -rpuN linux-2.6.18.8/arch/i386/power/cpu.c linux-2.6.18-xen-3.3.0/arch/i386/power/cpu.c
--- linux-2.6.18.8/arch/i386/power/cpu.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/power/cpu.c	2008-08-21 11:36:07.000000000 +0200
@@ -62,11 +62,12 @@ static void do_fpu_end(void)
 
 static void fix_processor_context(void)
 {
+#ifndef CONFIG_X86_NO_TSS
 	int cpu = smp_processor_id();
 	struct tss_struct * t = &per_cpu(init_tss, cpu);
 
 	set_tss_desc(cpu,t);	/* This just modifies memory; should not be necessary. But... This is necessary, because 386 hardware has concept of busy TSS or some similar stupidity. */
-
+#endif
 	load_TR_desc();				/* This does ltr */
 	load_LDT(&current->active_mm->context);	/* This does lldt */
 
diff -rpuN linux-2.6.18.8/arch/i386/power/Makefile linux-2.6.18-xen-3.3.0/arch/i386/power/Makefile
--- linux-2.6.18.8/arch/i386/power/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/i386/power/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -1,2 +1,4 @@
-obj-$(CONFIG_PM)		+= cpu.o
+obj-$(subst m,y,$(CONFIG_APM))	+= cpu.o
+obj-$(CONFIG_SOFTWARE_SUSPEND)	+= cpu.o
+obj-$(CONFIG_ACPI_SLEEP)	+= cpu.o
 obj-$(CONFIG_SOFTWARE_SUSPEND)	+= swsusp.o
diff -rpuN linux-2.6.18.8/arch/ia64/hp/common/sba_iommu.c linux-2.6.18-xen-3.3.0/arch/ia64/hp/common/sba_iommu.c
--- linux-2.6.18.8/arch/ia64/hp/common/sba_iommu.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/hp/common/sba_iommu.c	2008-08-21 11:36:07.000000000 +0200
@@ -42,6 +42,11 @@
 #include <asm/system.h>		/* wmb() */
 
 #include <asm/acpi-ext.h>
+#include <asm/maddr.h>		/* range_straddles_page_boundary() */
+#ifdef CONFIG_XEN
+#include <xen/gnttab.h>
+#include <asm/gnttab_dma.h>
+#endif
 
 #define PFX "IOC: "
 
@@ -198,6 +203,9 @@ struct ioc {
 	void __iomem	*ioc_hpa;	/* I/O MMU base address */
 	char		*res_map;	/* resource map, bit == pdir entry */
 	u64		*pdir_base;	/* physical base address */
+#ifdef CONFIG_XEN
+	u64		*xen_virt_cache;
+#endif
 	unsigned long	ibase;		/* pdir IOV Space base */
 	unsigned long	imask;		/* pdir IOV Space mask */
 
@@ -762,14 +770,21 @@ sba_free_range(struct ioc *ioc, dma_addr
  * on the vba.
  */
 
-#if 1
-#define sba_io_pdir_entry(pdir_ptr, vba) *pdir_ptr = ((vba & ~0xE000000000000FFFULL)	\
-						      | 0x8000000000000000ULL)
+#ifndef CONFIG_XEN
+#define sba_io_pdir_entry(ioc, pdir_ptr, vba) *pdir_ptr =	\
+	((virt_to_bus((void *)vba) & ~0xFFFULL) | 0x8000000000000000ULL)
 #else
 void SBA_INLINE
-sba_io_pdir_entry(u64 *pdir_ptr, unsigned long vba)
+sba_io_pdir_entry(struct ioc *ioc, u64 *pdir_ptr, unsigned long vba)
 {
-	*pdir_ptr = ((vba & ~0xE000000000000FFFULL) | 0x80000000000000FFULL);
+	*pdir_ptr = ((virt_to_bus((void *)vba) & ~0xFFFULL) |
+		    0x80000000000000FFULL);
+#ifdef CONFIG_XEN
+	if (is_running_on_xen()) {
+		int pide = ((u64)pdir_ptr - (u64)ioc->pdir_base) >> 3;
+		ioc->xen_virt_cache[pide] = vba;
+	}
+#endif
 }
 #endif
 
@@ -784,6 +799,12 @@ mark_clean (void *addr, size_t size)
 {
 	unsigned long pg_addr, end;
 
+#ifdef CONFIG_XEN
+	/* XXX: Bad things happen starting domUs when this is enabled. */
+	if (is_running_on_xen())
+		return;
+#endif
+
 	pg_addr = PAGE_ALIGN((unsigned long) addr);
 	end = (unsigned long) addr + size;
 	while (pg_addr + PAGE_SIZE <= end) {
@@ -850,6 +871,10 @@ sba_mark_invalid(struct ioc *ioc, dma_ad
 		*/
 		ioc->pdir_base[off] = (0x80000000000000FFULL | prefetch_spill_page);
 #endif
+#ifdef CONFIG_XEN
+		if (is_running_on_xen())
+			ioc->xen_virt_cache[off] = 0UL;
+#endif
 	} else {
 		u32 t = get_iovp_order(byte_cnt) + iovp_shift;
 
@@ -865,6 +890,10 @@ sba_mark_invalid(struct ioc *ioc, dma_ad
 #else
 			ioc->pdir_base[off] = (0x80000000000000FFULL | prefetch_spill_page);
 #endif
+#ifdef CONFIG_XEN
+			if (is_running_on_xen())
+				ioc->xen_virt_cache[off] = 0UL;
+#endif
 			off++;
 			byte_cnt -= iovp_size;
 		} while (byte_cnt > 0);
@@ -894,15 +923,29 @@ sba_map_single(struct device *dev, void 
 	unsigned long flags;
 #endif
 #ifdef ALLOW_IOV_BYPASS
-	unsigned long pci_addr = virt_to_phys(addr);
+	unsigned long pci_addr;
+#endif
+
+#ifdef CONFIG_XEN
+	if (is_running_on_xen()) {
+		void* tmp_addr = addr;
+		size_t tmp_size = size;
+		do {
+			gnttab_dma_use_page(virt_to_page(tmp_addr));
+			tmp_addr += PAGE_SIZE;
+			tmp_size -= min(tmp_size, PAGE_SIZE);
+		} while (tmp_size);
+	}
 #endif
 
 #ifdef ALLOW_IOV_BYPASS
+	pci_addr = virt_to_bus(addr);
 	ASSERT(to_pci_dev(dev)->dma_mask);
 	/*
  	** Check if the PCI device can DMA to ptr... if so, just return ptr
  	*/
-	if (likely((pci_addr & ~to_pci_dev(dev)->dma_mask) == 0)) {
+	if (likely((pci_addr & ~to_pci_dev(dev)->dma_mask) == 0 &&
+                   !range_straddles_page_boundary(__pa(addr), size))) {
 		/*
  		** Device is bit capable of DMA'ing to the buffer...
 		** just return the PCI address of ptr
@@ -944,7 +987,7 @@ sba_map_single(struct device *dev, void 
 
 	while (size > 0) {
 		ASSERT(((u8 *)pdir_start)[7] == 0); /* verify availability */
-		sba_io_pdir_entry(pdir_start, (unsigned long) addr);
+		sba_io_pdir_entry(ioc, pdir_start, (unsigned long) addr);
 
 		DBG_RUN("     pdir 0x%p %lx\n", pdir_start, *pdir_start);
 
@@ -973,13 +1016,29 @@ sba_mark_clean(struct ioc *ioc, dma_addr
 	void	*addr;
 
 	if (size <= iovp_size) {
-		addr = phys_to_virt(ioc->pdir_base[off] &
-		                    ~0xE000000000000FFFULL);
+#ifdef CONFIG_XEN
+		if (is_running_on_xen())
+			addr = (void *)ioc->xen_virt_cache[off];
+		else
+			addr = bus_to_virt(ioc->pdir_base[off] &
+					   ~0xE000000000000FFFULL);
+#else
+		addr = bus_to_virt(ioc->pdir_base[off] &
+				   ~0xE000000000000FFFULL);
+#endif
 		mark_clean(addr, size);
 	} else {
 		do {
-			addr = phys_to_virt(ioc->pdir_base[off] &
-			                    ~0xE000000000000FFFULL);
+#ifdef CONFIG_XEN
+			if (is_running_on_xen())
+				addr = (void *)ioc->xen_virt_cache[off];
+			else
+				addr = bus_to_virt(ioc->pdir_base[off] &
+						   ~0xE000000000000FFFULL);
+#else
+			addr = bus_to_virt(ioc->pdir_base[off] &
+					   ~0xE000000000000FFFULL);
+#endif
 			mark_clean(addr, min(size, iovp_size));
 			off++;
 			size -= iovp_size;
@@ -988,6 +1047,34 @@ sba_mark_clean(struct ioc *ioc, dma_addr
 }
 #endif
 
+#ifdef CONFIG_XEN
+static void
+sba_gnttab_dma_unmap_page(struct ioc *ioc, dma_addr_t iova, size_t size)
+{
+	u32 iovp = (u32) SBA_IOVP(ioc,iova);
+	int off = PDIR_INDEX(iovp);
+	struct page *page;
+
+	if (size <= iovp_size) {
+		BUG_ON(!ioc->xen_virt_cache[off]);
+		page = virt_to_page(ioc->xen_virt_cache[off]);
+		__gnttab_dma_unmap_page(page);
+	} else {
+		struct page *last_page = (struct page *)~0UL;
+		do {
+			BUG_ON(!ioc->xen_virt_cache[off]);
+			page = virt_to_page(ioc->xen_virt_cache[off]);
+			if (page != last_page) {
+				__gnttab_dma_unmap_page(page);
+				last_page = page;
+			}
+			off++;
+			size -= iovp_size;
+		} while (size > 0);
+	}
+}
+#endif
+
 /**
  * sba_unmap_single - unmap one IOVA and free resources
  * @dev: instance of PCI owned by the driver that's asking.
@@ -1018,7 +1105,16 @@ void sba_unmap_single(struct device *dev
 
 #ifdef ENABLE_MARK_CLEAN
 		if (dir == DMA_FROM_DEVICE) {
-			mark_clean(phys_to_virt(iova), size);
+			mark_clean(bus_to_virt(iova), size);
+		}
+#endif
+#ifdef CONFIG_XEN
+		if (is_running_on_xen()) {
+			do {
+				gnttab_dma_unmap_page(iova);
+				iova += PAGE_SIZE;
+				size -= min(size,PAGE_SIZE);
+			} while (size);
 		}
 #endif
 		return;
@@ -1037,6 +1133,10 @@ void sba_unmap_single(struct device *dev
 	if (dir == DMA_FROM_DEVICE)
 		sba_mark_clean(ioc, iova, size);
 #endif
+#ifdef CONFIG_XEN
+	if (is_running_on_xen())
+		sba_gnttab_dma_unmap_page(ioc, iova, size);
+#endif
 
 #if DELAYED_RESOURCE_CNT > 0
 	spin_lock_irqsave(&ioc->saved_lock, flags);
@@ -1102,9 +1202,14 @@ sba_alloc_coherent (struct device *dev, 
 		return NULL;
 
 	memset(addr, 0, size);
-	*dma_handle = virt_to_phys(addr);
 
 #ifdef ALLOW_IOV_BYPASS
+#ifdef CONFIG_XEN
+	if (xen_create_contiguous_region((unsigned long)addr, get_order(size),
+					 fls64(dev->coherent_dma_mask)))
+		goto iommu_map;
+#endif
+	*dma_handle = virt_to_bus(addr);
 	ASSERT(dev->coherent_dma_mask);
 	/*
  	** Check if the PCI device can DMA to ptr... if so, just return ptr
@@ -1115,6 +1220,9 @@ sba_alloc_coherent (struct device *dev, 
 
 		return addr;
 	}
+#ifdef CONFIG_XEN
+iommu_map:
+#endif
 #endif
 
 	/*
@@ -1138,6 +1246,13 @@ sba_alloc_coherent (struct device *dev, 
  */
 void sba_free_coherent (struct device *dev, size_t size, void *vaddr, dma_addr_t dma_handle)
 {
+#if defined(ALLOW_IOV_BYPASS) && defined(CONFIG_XEN)
+	struct ioc *ioc = GET_IOC(dev);
+
+	if (likely((dma_handle & ioc->imask) != ioc->ibase))
+		xen_destroy_contiguous_region((unsigned long)vaddr,
+					      get_order(size));
+#endif
 	sba_unmap_single(dev, dma_handle, size, 0);
 	free_pages((unsigned long) vaddr, get_order(size));
 }
@@ -1219,7 +1334,7 @@ sba_fill_pdir(
 			dma_offset=0;	/* only want offset on first chunk */
 			cnt = ROUNDUP(cnt, iovp_size);
 			do {
-				sba_io_pdir_entry(pdirp, vaddr);
+				sba_io_pdir_entry(ioc, pdirp, vaddr);
 				vaddr += iovp_size;
 				cnt -= iovp_size;
 				pdirp++;
@@ -1406,7 +1521,11 @@ int sba_map_sg(struct device *dev, struc
 	if (likely((ioc->dma_mask & ~to_pci_dev(dev)->dma_mask) == 0)) {
 		for (sg = sglist ; filled < nents ; filled++, sg++){
 			sg->dma_length = sg->length;
-			sg->dma_address = virt_to_phys(sba_sg_address(sg));
+#ifdef CONFIG_XEN
+			sg->dma_address = gnttab_dma_map_page(sg->page) + sg->offset;
+#else
+			sg->dma_address = virt_to_bus(sba_sg_address(sg));
+#endif
 		}
 		return filled;
 	}
@@ -1430,6 +1549,15 @@ int sba_map_sg(struct device *dev, struc
 
 	prefetch(ioc->res_hint);
 
+#ifdef CONFIG_XEN
+	if (is_running_on_xen()) {
+		int i;
+
+		for (i = 0; i < nents; i++)
+			gnttab_dma_use_page(sglist[i].page);
+	}
+#endif
+
 	/*
 	** First coalesce the chunks and allocate I/O pdir space
 	**
@@ -1562,11 +1690,25 @@ ioc_iova_init(struct ioc *ioc)
 
 	memset(ioc->pdir_base, 0, ioc->pdir_size);
 
+#ifdef CONFIG_XEN
+	/* The page table needs to be pinned in Xen memory */
+	if (xen_create_contiguous_region((unsigned long)ioc->pdir_base,
+					 get_order(ioc->pdir_size), 0))
+		panic(PFX "Couldn't contiguously map I/O Page Table\n");
+
+	ioc->xen_virt_cache = (void *) __get_free_pages(
+					GFP_KERNEL, get_order(ioc->pdir_size));
+	if (!ioc->xen_virt_cache)
+		panic(PFX "Couldn't allocate Xen virtual address cache\n");
+
+	memset(ioc->xen_virt_cache, 0, ioc->pdir_size);
+#endif
+
 	DBG_INIT("%s() IOV page size %ldK pdir %p size %x\n", __FUNCTION__,
 		iovp_size >> 10, ioc->pdir_base, ioc->pdir_size);
 
 	ASSERT(ALIGN((unsigned long) ioc->pdir_base, 4*1024) == (unsigned long) ioc->pdir_base);
-	WRITE_REG(virt_to_phys(ioc->pdir_base), ioc->ioc_hpa + IOC_PDIR_BASE);
+	WRITE_REG(virt_to_bus(ioc->pdir_base), ioc->ioc_hpa + IOC_PDIR_BASE);
 
 	/*
 	** If an AGP device is present, only use half of the IOV space
@@ -1603,7 +1745,7 @@ ioc_iova_init(struct ioc *ioc)
 		for ( ; (u64) poison_addr < addr + iovp_size; poison_addr += poison_size)
 			memcpy(poison_addr, spill_poison, poison_size);
 
-		prefetch_spill_page = virt_to_phys(addr);
+		prefetch_spill_page = virt_to_bus(addr);
 
 		DBG_INIT("%s() prefetch spill addr: 0x%lx\n", __FUNCTION__, prefetch_spill_page);
 	}
diff -rpuN linux-2.6.18.8/arch/ia64/Kconfig linux-2.6.18-xen-3.3.0/arch/ia64/Kconfig
--- linux-2.6.18.8/arch/ia64/Kconfig	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/Kconfig	2008-08-21 11:36:07.000000000 +0200
@@ -58,6 +58,28 @@ config GENERIC_IOMAP
 	bool
 	default y
 
+config XEN
+	bool "Xen hypervisor support"
+	default y
+	select XEN_XENCOMM
+	help
+	  Enable Xen hypervisor support.  Resulting kernel runs
+	  both as a guest OS on Xen and natively on hardware.
+
+config XEN_IA64_EXPOSE_P2M
+	bool "Xen/IA64 exposure p2m table"
+	depends on XEN
+	default y
+	help
+	  expose p2m from xen
+
+config XEN_IA64_EXPOSE_P2M_USE_DTR
+	bool "Xen/IA64 map p2m table with dtr"
+	depends on XEN_IA64_EXPOSE_P2M
+	default y
+	help
+	  use dtr to map the exposed p2m table
+
 config SCHED_NO_NO_OMIT_FRAME_POINTER
 	bool
 	default y
@@ -133,6 +155,10 @@ config IA64_SGI_SN2
 config IA64_HP_SIM
 	bool "Ski-simulator"
 
+config IA64_XEN
+	bool "Xen guest"
+	depends on XEN
+
 endchoice
 
 choice
@@ -431,6 +457,29 @@ config SGI_SN
 
 source "drivers/sn/Kconfig"
 
+config KEXEC
+	bool "kexec system call (EXPERIMENTAL)"
+	depends on EXPERIMENTAL && !IA64_HP_SIM && (!SMP || HOTPLUG_CPU) && !XEN_UNPRIVILEGED_GUEST
+	help
+	  kexec is a system call that implements the ability to shutdown your
+	  current kernel, and to start another kernel.  It is like a reboot
+	  but it is indepedent of the system firmware.   And like a reboot
+	  you can start any kernel with it, not just Linux.
+
+	  The name comes from the similiarity to the exec system call.
+
+	  It is an ongoing process to be certain the hardware in a machine
+	  is properly shutdown, so do not be surprised if this code does not
+	  initially work for you.  It may help to enable device hotplugging
+	  support.  As of this writing the exact hardware interface is
+	  strongly in flux, so no good recommendation can be made.
+
+config CRASH_DUMP
+	  bool "kernel crash dumps (EXPERIMENTAL)"
+	  depends on EXPERIMENTAL && IA64_MCA_RECOVERY && !IA64_HP_SIM && (!SMP || HOTPLUG_CPU)
+	  help
+	    Generate crash dump after being started by kexec.
+
 source "drivers/firmware/Kconfig"
 
 source "fs/Kconfig.binfmt"
@@ -465,6 +514,21 @@ config PCI_DOMAINS
 	bool
 	default PCI
 
+config XEN_PCIDEV_FRONTEND
+	bool "Xen PCI Frontend"
+	depends on PCI && XEN
+	default y
+	help
+	  The PCI device frontend driver allows the kernel to import arbitrary
+	  PCI devices from a PCI backend to support PCI driver domains.
+
+config XEN_PCIDEV_FE_DEBUG
+	bool "Xen PCI Frontend Debugging"
+	depends on XEN_PCIDEV_FRONTEND
+	default n
+	help
+	  Enables some debug statements within the PCI Frontend.
+
 source "drivers/pci/pcie/Kconfig"
 
 source "drivers/pci/Kconfig"
@@ -528,3 +592,16 @@ source "arch/ia64/Kconfig.debug"
 source "security/Kconfig"
 
 source "crypto/Kconfig"
+
+#
+# override default values of drivers/xen/Kconfig
+#
+if XEN
+config XEN_SMPBOOT
+	default n
+
+config XEN_DEVMEM
+	default n
+endif
+
+source "drivers/xen/Kconfig"
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/acpi.c linux-2.6.18-xen-3.3.0/arch/ia64/kernel/acpi.c
--- linux-2.6.18.8/arch/ia64/kernel/acpi.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/acpi.c	2008-08-21 11:36:07.000000000 +0200
@@ -109,6 +109,10 @@ const char *acpi_get_sysname(void)
 		return "hpzx1";
 	} else if (!strcmp(hdr->oem_id, "SGI")) {
 		return "sn2";
+#ifdef CONFIG_XEN
+	} else if (is_running_on_xen() && !strcmp(hdr->oem_id, "XEN")) {
+		return "xen";
+#endif
 	}
 
 	return "dig";
@@ -123,6 +127,8 @@ const char *acpi_get_sysname(void)
 	return "sn2";
 # elif defined (CONFIG_IA64_DIG)
 	return "dig";
+# elif defined (CONFIG_IA64_XEN)
+	return "xen";
 # else
 #	error Unknown platform.  Fix acpi.c.
 # endif
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/asm-offsets.c linux-2.6.18-xen-3.3.0/arch/ia64/kernel/asm-offsets.c
--- linux-2.6.18.8/arch/ia64/kernel/asm-offsets.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/asm-offsets.c	2008-08-21 11:36:07.000000000 +0200
@@ -268,4 +268,29 @@ void foo(void)
 	DEFINE(IA64_TIME_SOURCE_MMIO64, TIME_SOURCE_MMIO64);
 	DEFINE(IA64_TIME_SOURCE_MMIO32, TIME_SOURCE_MMIO32);
 	DEFINE(IA64_TIMESPEC_TV_NSEC_OFFSET, offsetof (struct timespec, tv_nsec));
+
+#ifdef CONFIG_XEN
+	BLANK();
+
+#define DEFINE_MAPPED_REG_OFS(sym, field) \
+	DEFINE(sym, (XMAPPEDREGS_OFS + offsetof(mapped_regs_t, field)))
+
+	DEFINE_MAPPED_REG_OFS(XSI_PSR_I_ADDR_OFS, interrupt_mask_addr);
+	DEFINE_MAPPED_REG_OFS(XSI_IPSR_OFS, ipsr);
+	DEFINE_MAPPED_REG_OFS(XSI_IIP_OFS, iip);
+	DEFINE_MAPPED_REG_OFS(XSI_IFS_OFS, ifs);
+	DEFINE_MAPPED_REG_OFS(XSI_PRECOVER_IFS_OFS, precover_ifs);
+	DEFINE_MAPPED_REG_OFS(XSI_ISR_OFS, isr);
+	DEFINE_MAPPED_REG_OFS(XSI_IFA_OFS, ifa);
+	DEFINE_MAPPED_REG_OFS(XSI_IIPA_OFS, iipa);
+	DEFINE_MAPPED_REG_OFS(XSI_IIM_OFS, iim);
+	DEFINE_MAPPED_REG_OFS(XSI_IHA_OFS, iha);
+	DEFINE_MAPPED_REG_OFS(XSI_ITIR_OFS, itir);
+	DEFINE_MAPPED_REG_OFS(XSI_PSR_IC_OFS, interrupt_collection_enabled);
+	DEFINE_MAPPED_REG_OFS(XSI_BANKNUM_OFS, banknum);
+	DEFINE_MAPPED_REG_OFS(XSI_BANK0_R16_OFS, bank0_regs[0]);
+	DEFINE_MAPPED_REG_OFS(XSI_BANK1_R16_OFS, bank1_regs[0]);
+	DEFINE_MAPPED_REG_OFS(XSI_B0NATS_OFS, vbnat);
+	DEFINE_MAPPED_REG_OFS(XSI_B1NATS_OFS, vnat);    
+#endif /* CONFIG_XEN */
 }
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/crash.c linux-2.6.18-xen-3.3.0/arch/ia64/kernel/crash.c
--- linux-2.6.18.8/arch/ia64/kernel/crash.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/crash.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,222 @@
+/*
+ * arch/ia64/kernel/crash.c
+ *
+ * Architecture specific (ia64) functions for kexec based crash dumps.
+ *
+ * Created by: Khalid Aziz <khalid.aziz@hp.com>
+ * Copyright (C) 2005 Hewlett-Packard Development Company, L.P.
+ * Copyright (C) 2005 Intel Corp	Zou Nan hai <nanhai.zou@intel.com>
+ *
+ */
+#include <linux/smp.h>
+#include <linux/delay.h>
+#include <linux/crash_dump.h>
+#include <linux/bootmem.h>
+#include <linux/kexec.h>
+#include <linux/elfcore.h>
+#include <linux/sysctl.h>
+#include <linux/init.h>
+
+#include <asm/kdebug.h>
+#include <asm/mca.h>
+
+int kdump_status[NR_CPUS];
+atomic_t kdump_cpu_freezed;
+atomic_t kdump_in_progress;
+int kdump_on_init = 1;
+
+static inline Elf64_Word
+*append_elf_note(Elf64_Word *buf, char *name, unsigned type, void *data,
+		size_t data_len)
+{
+	struct elf_note *note = (struct elf_note *)buf;
+	note->n_namesz = strlen(name) + 1;
+	note->n_descsz = data_len;
+	note->n_type   = type;
+	buf += (sizeof(*note) + 3)/4;
+	memcpy(buf, name, note->n_namesz);
+	buf += (note->n_namesz + 3)/4;
+	memcpy(buf, data, data_len);
+	buf += (data_len + 3)/4;
+	return buf;
+}
+
+static void
+final_note(void *buf)
+{
+	memset(buf, 0, sizeof(struct elf_note));
+}
+
+extern void ia64_dump_cpu_regs(void *);
+
+static DEFINE_PER_CPU(struct elf_prstatus, elf_prstatus);
+
+void
+crash_save_this_cpu()
+{
+	void *buf;
+	unsigned long cfm, sof, sol;
+
+	int cpu = smp_processor_id();
+	struct elf_prstatus *prstatus = &per_cpu(elf_prstatus, cpu);
+
+	elf_greg_t *dst = (elf_greg_t *)&(prstatus->pr_reg);
+	memset(prstatus, 0, sizeof(*prstatus));
+	prstatus->pr_pid = current->pid;
+
+	ia64_dump_cpu_regs(dst);
+	cfm = dst[43];
+	sol = (cfm >> 7) & 0x7f;
+	sof = cfm & 0x7f;
+	dst[46] = (unsigned long)ia64_rse_skip_regs((unsigned long *)dst[46],
+			sof - sol);
+
+	buf = (u64 *) per_cpu_ptr(crash_notes, cpu);
+	if (!buf)
+		return;
+	buf = append_elf_note(buf, "CORE", NT_PRSTATUS, prstatus,
+			sizeof(*prstatus));
+	final_note(buf);
+}
+
+static int
+kdump_wait_cpu_freeze(void)
+{
+	int cpu_num = num_online_cpus() - 1;
+	int timeout = 1000;
+	while(timeout-- > 0) {
+		if (atomic_read(&kdump_cpu_freezed) == cpu_num)
+			return 0;
+		udelay(1000);
+	}
+	return 1;
+}
+
+void
+machine_crash_shutdown(struct pt_regs *pt)
+{
+	/* This function is only called after the system
+	 * has paniced or is otherwise in a critical state.
+	 * The minimum amount of code to allow a kexec'd kernel
+	 * to run successfully needs to happen here.
+	 *
+	 * In practice this means shooting down the other cpus in
+	 * an SMP system.
+	 */
+	kexec_disable_iosapic();
+#ifdef CONFIG_SMP
+	kdump_smp_send_stop();
+	if (kdump_wait_cpu_freeze() && kdump_on_init) 	{
+		//not all cpu response to IPI, send INIT to freeze them
+		kdump_smp_send_init();
+	}
+#endif
+}
+
+static void
+machine_kdump_on_init(void)
+{
+	local_irq_disable();
+	kexec_disable_iosapic();
+	machine_kexec(ia64_kimage);
+}
+
+void
+kdump_cpu_freeze(struct unw_frame_info *info, void *arg)
+{
+	int cpuid;
+	local_irq_disable();
+	cpuid = smp_processor_id();
+	crash_save_this_cpu();
+	current->thread.ksp = (__u64)info->sw - 16;
+	atomic_inc(&kdump_cpu_freezed);
+	kdump_status[cpuid] = 1;
+	mb();
+	if (cpuid == 0) {
+		for (;;)
+			cpu_relax();
+	} else
+		ia64_jump_to_sal(&sal_boot_rendez_state[cpuid]);
+}
+
+static int
+kdump_init_notifier(struct notifier_block *self, unsigned long val, void *data)
+{
+	struct die_args *args = data;
+
+	if (!kdump_on_init)
+		return NOTIFY_DONE;
+
+	if (val != DIE_INIT_MONARCH_ENTER &&
+	    val != DIE_INIT_SLAVE_ENTER &&
+	    val != DIE_MCA_RENDZVOUS_LEAVE &&
+	    val != DIE_MCA_MONARCH_LEAVE)
+		return NOTIFY_DONE;
+
+	/* There really ought to be a check here to see if this
+	 * is a machine check rendevous. The kexec code that
+	 * was merged around 2.6.20-rc1 includes such a check.
+	 * But the check relies on infastructure that is not
+	 * available in 2.6.16. */
+
+	switch (val) {
+		case DIE_INIT_MONARCH_ENTER:
+			machine_kdump_on_init();
+			break;
+		case DIE_INIT_SLAVE_ENTER:
+			unw_init_running(kdump_cpu_freeze, NULL);
+			break;
+		case DIE_MCA_RENDZVOUS_LEAVE:
+			if (atomic_read(&kdump_in_progress))
+				unw_init_running(kdump_cpu_freeze, NULL);
+			break;
+		case DIE_MCA_MONARCH_LEAVE:
+		     /* die_register->signr indicate if MCA is recoverable */
+			if (!args->signr)
+				machine_kdump_on_init();
+			break;
+	}
+	return NOTIFY_DONE;
+}
+
+#ifdef CONFIG_SYSCTL
+static ctl_table kdump_on_init_table[] = {
+	{
+		.ctl_name = CTL_UNNUMBERED,
+		.procname = "kdump_on_init",
+		.data = &kdump_on_init,
+		.maxlen = sizeof(int),
+		.mode = 0644,
+		.proc_handler = &proc_dointvec,
+	},
+	{ .ctl_name = 0 }
+};
+
+static ctl_table sys_table[] = {
+	{
+	  .ctl_name = CTL_KERN,
+	  .procname = "kernel",
+	  .mode = 0555,
+	  .child = kdump_on_init_table,
+	},
+	{ .ctl_name = 0 }
+};
+#endif
+
+static int
+machine_crash_setup(void)
+{
+	static struct notifier_block kdump_init_notifier_nb = {
+		.notifier_call = kdump_init_notifier,
+	};
+	int ret;
+	if((ret = register_die_notifier(&kdump_init_notifier_nb)) != 0)
+		return ret;
+#ifdef CONFIG_SYSCTL
+	register_sysctl_table(sys_table, 0);
+#endif
+	return 0;
+}
+
+__initcall(machine_crash_setup);
+
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/crash_dump.c linux-2.6.18-xen-3.3.0/arch/ia64/kernel/crash_dump.c
--- linux-2.6.18.8/arch/ia64/kernel/crash_dump.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/crash_dump.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,48 @@
+/*
+ *	kernel/crash_dump.c - Memory preserving reboot related code.
+ *
+ *	Created by: Simon Horman <horms@verge.net.au>
+ *	Original code moved from kernel/crash.c
+ *	Original code comment copied from the i386 version of this file
+ */
+
+#include <linux/errno.h>
+#include <linux/types.h>
+
+#include <linux/uaccess.h>
+
+/**
+ * copy_oldmem_page - copy one page from "oldmem"
+ * @pfn: page frame number to be copied
+ * @buf: target memory address for the copy; this can be in kernel address
+ *	space or user address space (see @userbuf)
+ * @csize: number of bytes to copy
+ * @offset: offset in bytes into the page (based on pfn) to begin the copy
+ * @userbuf: if set, @buf is in user address space, use copy_to_user(),
+ *	otherwise @buf is in kernel address space, use memcpy().
+ *
+ * Copy a page from "oldmem". For this page, there is no pte mapped
+ * in the current kernel. We stitch up a pte, similar to kmap_atomic.
+ *
+ * Calling copy_to_user() in atomic context is not desirable. Hence first
+ * copying the data to a pre-allocated kernel page and then copying to user
+ * space in non-atomic context.
+ */
+ssize_t
+copy_oldmem_page(unsigned long pfn, char *buf,
+		size_t csize, unsigned long offset, int userbuf)
+{
+	void  *vaddr;
+
+	if (!csize)
+		return 0;
+	vaddr = __va(pfn<<PAGE_SHIFT);
+	if (userbuf) {
+		if (copy_to_user(buf, (vaddr + offset), csize)) {
+			return -EFAULT;
+		}
+	} else
+		memcpy(buf, (vaddr + offset), csize);
+	return csize;
+}
+
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/efi.c linux-2.6.18-xen-3.3.0/arch/ia64/kernel/efi.c
--- linux-2.6.18.8/arch/ia64/kernel/efi.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/efi.c	2008-08-21 11:36:07.000000000 +0200
@@ -21,11 +21,13 @@
  *	Skip non-WB memory and ignore empty memory ranges.
  */
 #include <linux/module.h>
+#include <linux/bootmem.h>
 #include <linux/kernel.h>
 #include <linux/init.h>
 #include <linux/types.h>
 #include <linux/time.h>
 #include <linux/efi.h>
+#include <linux/kexec.h>
 
 #include <asm/io.h>
 #include <asm/kregs.h>
@@ -34,6 +36,11 @@
 #include <asm/processor.h>
 #include <asm/mca.h>
 
+#ifdef CONFIG_PROC_IOMEM_MACHINE
+#include <xen/interface/memory.h>
+#include <asm/hypercall.h>
+#endif
+
 #define EFI_DEBUG	0
 
 extern efi_status_t efi_call_phys (void *, ...);
@@ -41,7 +48,7 @@ extern efi_status_t efi_call_phys (void 
 struct efi efi;
 EXPORT_SYMBOL(efi);
 static efi_runtime_services_t *runtime;
-static unsigned long mem_limit = ~0UL, max_addr = ~0UL;
+static unsigned long mem_limit = ~0UL, max_addr = ~0UL, min_addr = 0UL;
 
 #define efi_call_virt(f, args...)	(*(f))(args)
 
@@ -421,6 +428,8 @@ efi_init (void)
 			mem_limit = memparse(cp + 4, &cp);
 		} else if (memcmp(cp, "max_addr=", 9) == 0) {
 			max_addr = GRANULEROUNDDOWN(memparse(cp + 9, &cp));
+		} else if (memcmp(cp, "min_addr=", 9) == 0) {
+			min_addr = GRANULEROUNDDOWN(memparse(cp + 9, &cp));
 		} else {
 			while (*cp != ' ' && *cp)
 				++cp;
@@ -428,6 +437,8 @@ efi_init (void)
 				++cp;
 		}
 	}
+	if (min_addr != 0UL)
+		printk(KERN_INFO "Ignoring memory below %luMB\n", min_addr >> 20);
 	if (max_addr != ~0UL)
 		printk(KERN_INFO "Ignoring memory above %luMB\n", max_addr >> 20);
 
@@ -894,7 +905,8 @@ find_memmap_space (void)
 		as = max(contig_low, md->phys_addr);
 		ae = min(contig_high, efi_md_end(md));
 
-		/* keep within max_addr= command line arg */
+		/* keep within max_addr= and min_addr= command line arg */
+		as = max(as, min_addr);
 		ae = min(ae, max_addr);
 		if (ae <= as)
 			continue;
@@ -965,6 +977,11 @@ efi_memmap_init(unsigned long *s, unsign
 		if (!is_available_memory(md))
 			continue;
 
+#ifdef CONFIG_CRASH_DUMP
+		/* saved_max_pfn should ignore max_addr= command line arg */
+		if (saved_max_pfn < (efi_md_end(md) >> PAGE_SHIFT))
+			saved_max_pfn = (efi_md_end(md) >> PAGE_SHIFT);
+#endif
 		/*
 		 * Round ends inward to granule boundaries
 		 * Give trimmings to uncached allocator
@@ -1004,7 +1021,8 @@ efi_memmap_init(unsigned long *s, unsign
 		} else
 			ae = efi_md_end(md);
 
-		/* keep within max_addr= command line arg */
+		/* keep within max_addr= and min_addr= command line arg */
+		as = max(as, min_addr);
 		ae = min(ae, max_addr);
 		if (ae <= as)
 			continue;
@@ -1033,21 +1051,22 @@ efi_memmap_init(unsigned long *s, unsign
 	*e = (u64)++k;
 }
 
-void
-efi_initialize_iomem_resources(struct resource *code_resource,
-			       struct resource *data_resource)
+#define EFI_INITIALISE_PHYS 0x1
+#define EFI_INITIALISE_MACH 0x2
+#define EFI_INITIALISE_ALL  (EFI_INITIALISE_PHYS|EFI_INITIALISE_MACH)
+
+static void
+efi_initialize_resources(void *efi_map_start, void *efi_map_end,
+			 u64 efi_desc_size, struct resource *root_resource,
+			 struct resource *code_resource,
+			 struct resource *data_resource, unsigned flag)
 {
 	struct resource *res;
-	void *efi_map_start, *efi_map_end, *p;
+	void *p;
 	efi_memory_desc_t *md;
-	u64 efi_desc_size;
 	char *name;
 	unsigned long flags;
 
-	efi_map_start = __va(ia64_boot_param->efi_memmap);
-	efi_map_end   = efi_map_start + ia64_boot_param->efi_memmap_size;
-	efi_desc_size = ia64_boot_param->efi_memdesc_size;
-
 	res = NULL;
 
 	for (p = efi_map_start; p < efi_map_end; p += efi_desc_size) {
@@ -1106,7 +1125,7 @@ efi_initialize_iomem_resources(struct re
 		res->end = md->phys_addr + (md->num_pages << EFI_PAGE_SHIFT) - 1;
 		res->flags = flags;
 
-		if (insert_resource(&iomem_resource, res) < 0)
+		if (insert_resource(root_resource, res) < 0)
 			kfree(res);
 		else {
 			/*
@@ -1114,8 +1133,135 @@ efi_initialize_iomem_resources(struct re
 			 * kernel data so we try it repeatedly and
 			 * let the resource manager test it.
 			 */
-			insert_resource(res, code_resource);
-			insert_resource(res, data_resource);
+			if (flag & EFI_INITIALISE_PHYS) {
+				insert_resource(res, code_resource);
+				insert_resource(res, data_resource);
+			}
+#ifdef CONFIG_KEXEC
+			if (flag & EFI_INITIALISE_MACH) {
+				insert_resource(res, &efi_memmap_res);
+				insert_resource(res, &boot_param_res);
+				if (crashk_res.end > crashk_res.start)
+					insert_resource(res, &crashk_res);
+#ifdef CONFIG_XEN
+				if (is_initial_xendomain())
+					xen_machine_kexec_register_resources(
+								res);
+#endif
+			}
+#endif
 		}
 	}
 }
+
+#ifdef CONFIG_PROC_IOMEM_MACHINE
+static void
+efi_initialize_iomem_machine_resources(void)
+{
+	unsigned long size;
+	xen_memory_map_t memmap;
+	xen_ia64_memmap_info_t *memmap_info = NULL;
+	void *efi_map_start, *efi_map_end;
+	u64 efi_desc_size;
+	int ret;
+
+	/* It would be nice if it wasn't neccessary to loop like this */
+	for (size = 1024; 1; size += 1024) {
+		memmap_info = kmalloc(size, GFP_KERNEL);
+		if (memmap_info == NULL)
+			return; /* -ENOMEM, but no way to return error */
+
+		memmap.nr_entries = size;
+		set_xen_guest_handle(memmap.buffer, memmap_info);
+		ret = HYPERVISOR_memory_op(XENMEM_machine_memory_map, &memmap);
+		if (!ret)
+			break;
+
+		kfree(memmap_info);
+	}
+
+	efi_map_start = &memmap_info->memdesc;
+	efi_map_end = efi_map_start + memmap_info->efi_memmap_size;
+	efi_desc_size = memmap_info->efi_memdesc_size;
+	efi_initialize_resources(efi_map_start, efi_map_end, efi_desc_size,
+				 &iomem_machine_resource, NULL, NULL,
+				 EFI_INITIALISE_MACH);
+
+	kfree(memmap_info);
+}
+#endif
+
+void
+efi_initialize_iomem_resources(struct resource *code_resource,
+			       struct resource *data_resource)
+{
+	void *efi_map_start, *efi_map_end;
+	u64 efi_desc_size;
+
+	efi_map_start = __va(ia64_boot_param->efi_memmap);
+	efi_map_end   = efi_map_start + ia64_boot_param->efi_memmap_size;
+	efi_desc_size = ia64_boot_param->efi_memdesc_size;
+
+#ifdef CONFIG_PROC_IOMEM_MACHINE
+	if (is_initial_xendomain()) {
+		efi_initialize_resources(efi_map_start, efi_map_end,
+					 efi_desc_size, &iomem_resource,
+					 code_resource, data_resource,
+					 EFI_INITIALISE_PHYS);
+		efi_initialize_iomem_machine_resources();
+	}
+	else
+#endif
+		efi_initialize_resources(efi_map_start, efi_map_end,
+					 efi_desc_size, &iomem_resource,
+					 code_resource, data_resource,
+					 EFI_INITIALISE_ALL);
+}
+
+
+
+#ifdef CONFIG_KEXEC
+/* find a block of memory aligned to 64M exclude reserved regions
+   rsvd_regions are sorted
+ */
+unsigned long
+kdump_find_rsvd_region (unsigned long size,
+		struct rsvd_region *r, int n)
+{
+  int i;
+  u64 start, end;
+  u64 alignment = 1UL << _PAGE_SIZE_64M;
+  void *efi_map_start, *efi_map_end, *p;
+  efi_memory_desc_t *md;
+  u64 efi_desc_size;
+
+  efi_map_start = __va(ia64_boot_param->efi_memmap);
+  efi_map_end   = efi_map_start + ia64_boot_param->efi_memmap_size;
+  efi_desc_size = ia64_boot_param->efi_memdesc_size;
+
+  for (p = efi_map_start; p < efi_map_end; p += efi_desc_size) {
+	  md = p;
+	  if (!efi_wb(md))
+		  continue;
+	  start = ALIGN(md->phys_addr, alignment);
+	  end = efi_md_end(md);
+	  for (i = 0; i < n; i++) {
+		if (__pa(r[i].start) >= start && __pa(r[i].end) < end) {
+			if (__pa(r[i].start) > start + size)
+				return start;
+			start = ALIGN(__pa(r[i].end), alignment);
+			if (i < n-1 && __pa(r[i+1].start) < start + size)
+				continue;
+			else
+				break;
+		}
+	  }
+	  if (end > start + size)
+		return start;
+  }
+
+  printk(KERN_WARNING "Cannot reserve 0x%lx byte of memory for crashdump\n",
+	size);
+  return ~0UL;
+}
+#endif
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/entry.S linux-2.6.18-xen-3.3.0/arch/ia64/kernel/entry.S
--- linux-2.6.18.8/arch/ia64/kernel/entry.S	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/entry.S	2008-08-21 11:36:07.000000000 +0200
@@ -180,7 +180,7 @@ END(sys_clone)
  *	called.  The code starting at .map relies on this.  The rest of the code
  *	doesn't care about the interrupt masking status.
  */
-GLOBAL_ENTRY(ia64_switch_to)
+GLOBAL_ENTRY(__ia64_switch_to)
 	.prologue
 	alloc r16=ar.pfs,1,0,0,0
 	DO_SAVE_SWITCH_STACK
@@ -234,7 +234,7 @@ GLOBAL_ENTRY(ia64_switch_to)
 	;;
 	srlz.d
 	br.cond.sptk .done
-END(ia64_switch_to)
+END(__ia64_switch_to)
 
 /*
  * Note that interrupts are enabled during save_switch_stack and load_switch_stack.  This
@@ -375,7 +375,7 @@ END(save_switch_stack)
  *	- b7 holds address to return to
  *	- must not touch r8-r11
  */
-ENTRY(load_switch_stack)
+GLOBAL_ENTRY(load_switch_stack)
 	.prologue
 	.altrp b7
 
@@ -510,7 +510,7 @@ END(clone)
 	 * because some system calls (such as ia64_execve) directly
 	 * manipulate ar.pfs.
 	 */
-GLOBAL_ENTRY(ia64_trace_syscall)
+GLOBAL_ENTRY(__ia64_trace_syscall)
 	PT_REGS_UNWIND_INFO(0)
 	/*
 	 * We need to preserve the scratch registers f6-f11 in case the system
@@ -582,7 +582,7 @@ strace_error:
 (p6)	mov r10=-1
 (p6)	mov r8=r9
 	br.cond.sptk .strace_save_retval
-END(ia64_trace_syscall)
+END(__ia64_trace_syscall)
 
 	/*
 	 * When traced and returning from sigreturn, we invoke syscall_trace but then
@@ -601,7 +601,7 @@ GLOBAL_ENTRY(ia64_strace_leave_kernel)
 .ret4:	br.cond.sptk ia64_leave_kernel
 END(ia64_strace_leave_kernel)
 
-GLOBAL_ENTRY(ia64_ret_from_clone)
+GLOBAL_ENTRY(__ia64_ret_from_clone)
 	PT_REGS_UNWIND_INFO(0)
 {	/*
 	 * Some versions of gas generate bad unwind info if the first instruction of a
@@ -627,7 +627,7 @@ GLOBAL_ENTRY(ia64_ret_from_clone)
 	cmp.ne p6,p0=r2,r0
 (p6)	br.cond.spnt .strace_check_retval
 	;;					// added stop bits to prevent r8 dependency
-END(ia64_ret_from_clone)
+END(__ia64_ret_from_clone)
 	// fall through
 GLOBAL_ENTRY(ia64_ret_from_syscall)
 	PT_REGS_UNWIND_INFO(0)
@@ -635,8 +635,11 @@ GLOBAL_ENTRY(ia64_ret_from_syscall)
 	adds r2=PT(R8)+16,sp			// r2 = &pt_regs.r8
 	mov r10=r0				// clear error indication in r10
 (p7)	br.cond.spnt handle_syscall_error	// handle potential syscall failure
+	;;
+	// don't fall through, ia64_leave_syscall may be #define'd
+	br.cond.sptk.few ia64_leave_syscall
+	;;
 END(ia64_ret_from_syscall)
-	// fall through
 /*
  * ia64_leave_syscall(): Same as ia64_leave_kernel, except that it doesn't
  *	need to switch to bank 0 and doesn't restore the scratch registers.
@@ -681,7 +684,7 @@ END(ia64_ret_from_syscall)
  *	      ar.csd: cleared
  *	      ar.ssd: cleared
  */
-ENTRY(ia64_leave_syscall)
+GLOBAL_ENTRY(__ia64_leave_syscall)
 	PT_REGS_UNWIND_INFO(0)
 	/*
 	 * work.need_resched etc. mustn't get changed by this CPU before it returns to
@@ -789,7 +792,7 @@ ENTRY(ia64_leave_syscall)
 	mov.m ar.ssd=r0			// M2   clear ar.ssd
 	mov f11=f0			// F    clear f11
 	br.cond.sptk.many rbs_switch	// B
-END(ia64_leave_syscall)
+END(__ia64_leave_syscall)
 
 #ifdef CONFIG_IA32_SUPPORT
 GLOBAL_ENTRY(ia64_ret_from_ia32_execve)
@@ -801,10 +804,13 @@ GLOBAL_ENTRY(ia64_ret_from_ia32_execve)
 	st8.spill [r2]=r8	// store return value in slot for r8 and set unat bit
 	.mem.offset 8,0
 	st8.spill [r3]=r0	// clear error indication in slot for r10 and set unat bit
+	;;
+	// don't fall through, ia64_leave_kernel may be #define'd
+	br.cond.sptk.few ia64_leave_kernel
+	;;
 END(ia64_ret_from_ia32_execve)
-	// fall through
 #endif /* CONFIG_IA32_SUPPORT */
-GLOBAL_ENTRY(ia64_leave_kernel)
+GLOBAL_ENTRY(__ia64_leave_kernel)
 	PT_REGS_UNWIND_INFO(0)
 	/*
 	 * work.need_resched etc. mustn't get changed by this CPU before it returns to
@@ -1135,7 +1141,7 @@ skip_rbs_switch:
 	ld8 r10=[r3]
 	br.cond.sptk.many .work_processed_syscall	// re-check
 
-END(ia64_leave_kernel)
+END(__ia64_leave_kernel)
 
 ENTRY(handle_syscall_error)
 	/*
@@ -1175,7 +1181,7 @@ END(ia64_invoke_schedule_tail)
 	 * be set up by the caller.  We declare 8 input registers so the system call
 	 * args get preserved, in case we need to restart a system call.
 	 */
-ENTRY(notify_resume_user)
+GLOBAL_ENTRY(notify_resume_user)
 	.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(8)
 	alloc loc1=ar.pfs,8,2,3,0 // preserve all eight input regs in case of syscall restart!
 	mov r9=ar.unat
@@ -1263,7 +1269,7 @@ ENTRY(sys_rt_sigreturn)
 	adds sp=16,sp
 	;;
 	ld8 r9=[sp]				// load new ar.unat
-	mov.sptk b7=r8,ia64_leave_kernel
+	mov.sptk b7=r8,__ia64_leave_kernel
 	;;
 	mov ar.unat=r9
 	br.many b7
@@ -1575,7 +1581,7 @@ sys_call_table:
 	data8 sys_mq_timedreceive		// 1265
 	data8 sys_mq_notify
 	data8 sys_mq_getsetattr
-	data8 sys_ni_syscall			// reserved for kexec_load
+	data8 sys_kexec_load
 	data8 sys_ni_syscall			// reserved for vserver
 	data8 sys_waitid			// 1270
 	data8 sys_add_key
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/fsys.S linux-2.6.18-xen-3.3.0/arch/ia64/kernel/fsys.S
--- linux-2.6.18.8/arch/ia64/kernel/fsys.S	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/fsys.S	2008-08-21 11:36:07.000000000 +0200
@@ -516,11 +516,34 @@ ENTRY(fsys_fallback_syscall)
 	adds r17=-1024,r15
 	movl r14=sys_call_table
 	;;
+#ifdef CONFIG_XEN
+	movl r18=running_on_xen;;
+	ld4 r18=[r18];;
+	// p14 = running_on_xen
+	// p15 = !running_on_xen
+	cmp.ne p14,p15=r0,r18
+	;;    
+(p14)	movl r18=XSI_PSR_I_ADDR;;
+(p14)	ld8 r18=[r18]
+(p14)	mov r29=1;;
+(p14)	st1 [r18]=r29
+(p15)	rsm psr.i
+#else    
 	rsm psr.i
+#endif    
 	shladd r18=r17,3,r14
 	;;
 	ld8 r18=[r18]				// load normal (heavy-weight) syscall entry-point
+#ifdef CONFIG_XEN
+(p14)	mov r27=r8
+(p14)	XEN_HYPER_GET_PSR
+	;;
+(p14)	mov r29=r8
+(p14)	mov r8=r27
+(p15)	mov r29=psr				// read psr (12 cyc load latency)
+#else    
 	mov r29=psr				// read psr (12 cyc load latency)
+#endif    
 	mov r27=ar.rsc
 	mov r21=ar.fpsr
 	mov r26=ar.pfs
@@ -632,7 +655,25 @@ GLOBAL_ENTRY(fsys_bubble_down)
 	mov rp=r14				// I0   set the real return addr
 	and r3=_TIF_SYSCALL_TRACEAUDIT,r3	// A
 	;;
+#ifdef CONFIG_XEN
+	movl r14=running_on_xen;;
+	ld4 r14=[r14];;
+	// p14 = running_on_xen
+	// p15 = !running_on_xen
+	cmp.ne p14,p15=r0,r14
+	;; 
+(p14)	movl r28=XSI_PSR_I_ADDR;;
+(p14)	ld8 r28=[r28];;
+(p14)	adds r28=-1,r28;;			// event_pending
+(p14)	ld1 r14=[r28];;
+(p14)	cmp.ne.unc p13,p14=r14,r0;;
+(p13)	XEN_HYPER_SSM_I
+(p14)	adds r28=1,r28;;			// event_mask
+(p14)	st1 [r28]=r0;;
+(p15)	ssm psr.i
+#else
 	ssm psr.i				// M2   we're on kernel stacks now, reenable irqs
+#endif    
 	cmp.eq p8,p0=r3,r0			// A
 (p10)	br.cond.spnt.many ia64_ret_from_syscall	// B    return if bad call-frame or r15 is a NaT
 
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/gate.lds.S linux-2.6.18-xen-3.3.0/arch/ia64/kernel/gate.lds.S
--- linux-2.6.18.8/arch/ia64/kernel/gate.lds.S	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/gate.lds.S	2008-08-21 11:36:07.000000000 +0200
@@ -28,6 +28,24 @@ SECTIONS
   . = GATE_ADDR + 0x500;
 
   .data.patch			: {
+#ifdef __XEN_IA64_VDSO_PARAVIRT
+#define __start_gate_mckinley_e9_patchlist \
+	__start_gate_mckinley_e9_patchlist_xen
+#define __end_gate_mckinley_e9_patchlist \
+	__end_gate_mckinley_e9_patchlist_xen
+#define __start_gate_vtop_patchlist \
+	__start_gate_vtop_patchlist_xen
+#define __end_gate_vtop_patchlist \
+	__end_gate_vtop_patchlist_xen
+#define __start_gate_fsyscall_patchlist \
+	__start_gate_fsyscall_patchlist_xen
+#define __end_gate_fsyscall_patchlist \
+	__end_gate_fsyscall_patchlist_xen
+#define __start_gate_brl_fsys_bubble_down_patchlist \
+	__start_gate_brl_fsys_bubble_down_patchlist_xen
+#define __end_gate_brl_fsys_bubble_down_patchlist \
+	__end_gate_brl_fsys_bubble_down_patchlist_xen
+#endif
 				    __start_gate_mckinley_e9_patchlist = .;
 				    *(.data.patch.mckinley_e9)
 				    __end_gate_mckinley_e9_patchlist = .;
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/gate.S linux-2.6.18-xen-3.3.0/arch/ia64/kernel/gate.S
--- linux-2.6.18.8/arch/ia64/kernel/gate.S	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/gate.S	2008-08-21 11:36:07.000000000 +0200
@@ -32,102 +32,6 @@
 [1:](pr)brl.cond.sptk 0;				\
 	.xdata4 ".data.patch.brl_fsys_bubble_down", 1b-.
 
-GLOBAL_ENTRY(__kernel_syscall_via_break)
-	.prologue
-	.altrp b6
-	.body
-	/*
-	 * Note: for (fast) syscall restart to work, the break instruction must be
-	 *	 the first one in the bundle addressed by syscall_via_break.
-	 */
-{ .mib
-	break 0x100000
-	nop.i 0
-	br.ret.sptk.many b6
-}
-END(__kernel_syscall_via_break)
-
-/*
- * On entry:
- *	r11 = saved ar.pfs
- *	r15 = system call #
- *	b0  = saved return address
- *	b6  = return address
- * On exit:
- *	r11 = saved ar.pfs
- *	r15 = system call #
- *	b0  = saved return address
- *	all other "scratch" registers:	undefined
- *	all "preserved" registers:	same as on entry
- */
-
-GLOBAL_ENTRY(__kernel_syscall_via_epc)
-	.prologue
-	.altrp b6
-	.body
-{
-	/*
-	 * Note: the kernel cannot assume that the first two instructions in this
-	 * bundle get executed.  The remaining code must be safe even if
-	 * they do not get executed.
-	 */
-	adds r17=-1024,r15			// A
-	mov r10=0				// A    default to successful syscall execution
-	epc					// B	causes split-issue
-}
-	;;
-	rsm psr.be | psr.i			// M2 (5 cyc to srlz.d)
-	LOAD_FSYSCALL_TABLE(r14)		// X
-	;;
-	mov r16=IA64_KR(CURRENT)		// M2 (12 cyc)
-	shladd r18=r17,3,r14			// A
-	mov r19=NR_syscalls-1			// A
-	;;
-	lfetch [r18]				// M0|1
-	mov r29=psr				// M2 (12 cyc)
-	// If r17 is a NaT, p6 will be zero
-	cmp.geu p6,p7=r19,r17			// A    (sysnr > 0 && sysnr < 1024+NR_syscalls)?
-	;;
-	mov r21=ar.fpsr				// M2 (12 cyc)
-	tnat.nz p10,p9=r15			// I0
-	mov.i r26=ar.pfs			// I0 (would stall anyhow due to srlz.d...)
-	;;
-	srlz.d					// M0 (forces split-issue) ensure PSR.BE==0
-(p6)	ld8 r18=[r18]				// M0|1
-	nop.i 0
-	;;
-	nop.m 0
-(p6)	tbit.z.unc p8,p0=r18,0			// I0 (dual-issues with "mov b7=r18"!)
-	nop.i 0
-	;;
-(p8)	ssm psr.i
-(p6)	mov b7=r18				// I0
-(p8)	br.dptk.many b7				// B
-
-	mov r27=ar.rsc				// M2 (12 cyc)
-/*
- * brl.cond doesn't work as intended because the linker would convert this branch
- * into a branch to a PLT.  Perhaps there will be a way to avoid this with some
- * future version of the linker.  In the meantime, we just use an indirect branch
- * instead.
- */
-#ifdef CONFIG_ITANIUM
-(p6)	add r14=-8,r14				// r14 <- addr of fsys_bubble_down entry
-	;;
-(p6)	ld8 r14=[r14]				// r14 <- fsys_bubble_down
-	;;
-(p6)	mov b7=r14
-(p6)	br.sptk.many b7
-#else
-	BRL_COND_FSYS_BUBBLE_DOWN(p6)
-#endif
-	ssm psr.i
-	mov r10=-1
-(p10)	mov r8=EINVAL
-(p9)	mov r8=ENOSYS
-	FSYS_RETURN
-END(__kernel_syscall_via_epc)
-
 #	define ARG0_OFF		(16 + IA64_SIGFRAME_ARG0_OFFSET)
 #	define ARG1_OFF		(16 + IA64_SIGFRAME_ARG1_OFFSET)
 #	define ARG2_OFF		(16 + IA64_SIGFRAME_ARG2_OFFSET)
@@ -373,3 +277,154 @@ restore_rbs:
 	// invala not necessary as that will happen when returning to user-mode
 	br.cond.sptk back_from_restore_rbs
 END(__kernel_sigtramp)
+
+GLOBAL_ENTRY(__kernel_syscall_via_break)
+	.prologue
+	.altrp b6
+	.body
+	/*
+	 * Note: for (fast) syscall restart to work, the break instruction must be
+	 *	 the first one in the bundle addressed by syscall_via_break.
+	 */
+{ .mib
+	break 0x100000
+	nop.i 0
+	br.ret.sptk.many b6
+}
+END(__kernel_syscall_via_break)
+
+/*
+ * On entry:
+ *	r11 = saved ar.pfs
+ *	r15 = system call #
+ *	b0  = saved return address
+ *	b6  = return address
+ * On exit:
+ *	r11 = saved ar.pfs
+ *	r15 = system call #
+ *	b0  = saved return address
+ *	all other "scratch" registers:	undefined
+ *	all "preserved" registers:	same as on entry
+ */
+
+GLOBAL_ENTRY(__kernel_syscall_via_epc)
+	.prologue
+	.altrp b6
+	.body
+{
+	/*
+	 * Note: the kernel cannot assume that the first two instructions in this
+	 * bundle get executed.  The remaining code must be safe even if
+	 * they do not get executed.
+	 */
+	adds r17=-1024,r15			// A
+	mov r10=0				// A    default to successful syscall execution
+	epc					// B	causes split-issue
+}
+	;;
+#ifdef __XEN_IA64_VDSO_PARAVIRT
+	// r20 = 1
+	// r22 = &vcpu->vcpu_info->evtchn_upcall_mask
+	// r24 = &vcpu->vcpu_info->evtchn_upcall_pending
+	// r25 = tmp
+	// r31 = tmp
+	// p11 = tmp
+	// p14 = tmp
+	mov r20=1
+	movl r22=XSI_PSR_I_ADDR
+	;;
+	ld8 r22=[r22]
+	;;
+	st1 [r22]=r20
+	rum psr.be
+	adds r24=-1,r22
+#else
+	rsm psr.be | psr.i			// M2 (5 cyc to srlz.d)
+#endif
+	LOAD_FSYSCALL_TABLE(r14)		// X
+	;;
+	mov r16=IA64_KR(CURRENT)		// M2 (12 cyc)
+	shladd r18=r17,3,r14			// A
+	mov r19=NR_syscalls-1			// A
+#ifdef __XEN_IA64_VDSO_PARAVIRT
+	XEN_HYPER_GET_PSR
+	;;
+	lfetch [r18]				// M0|1
+	mov r29=r8
+#else
+	;;
+	lfetch [r18]				// M0|1
+	mov r29=psr				// M2 (12 cyc)
+#endif
+	// If r17 is a NaT, p6 will be zero
+	cmp.geu p6,p7=r19,r17			// A    (sysnr > 0 && sysnr < 1024+NR_syscalls)?
+	;;
+	mov r21=ar.fpsr				// M2 (12 cyc)
+	tnat.nz p10,p9=r15			// I0
+	mov.i r26=ar.pfs			// I0 (would stall anyhow due to srlz.d...)
+	;;
+	srlz.d					// M0 (forces split-issue) ensure PSR.BE==0
+(p6)	ld8 r18=[r18]				// M0|1
+	nop.i 0
+	;;
+	nop.m 0
+(p6)	tbit.z.unc p8,p0=r18,0			// I0 (dual-issues with "mov b7=r18"!)
+#ifdef __XEN_IA64_VDSO_PARAVIRT
+	
+#define XEN_SET_PSR_I(pred)		\
+(pred)	ld1 r31=[r22];			\
+	;; ;				\
+(pred)	st1 [r22]=r0;			\
+(pred)	cmp.ne.unc p14,p0=r0,r31;	\
+	;; ;				\
+(p14)	ld1 r25=[r24];			\
+	;; ;				\
+(p14)	cmp.ne.unc p11,p0=r0,r25;	\
+	;; ;				\
+(p11)	XEN_HYPER_SSM_I;
+
+	;; 
+	XEN_SET_PSR_I(p8)
+#else
+	nop.i 0
+	;;
+(p8)	ssm psr.i
+#endif
+(p6)	mov b7=r18				// I0
+(p8)	br.dptk.many b7				// B
+
+	mov r27=ar.rsc				// M2 (12 cyc)
+/*
+ * brl.cond doesn't work as intended because the linker would convert this branch
+ * into a branch to a PLT.  Perhaps there will be a way to avoid this with some
+ * future version of the linker.  In the meantime, we just use an indirect branch
+ * instead.
+ */
+#ifdef CONFIG_ITANIUM
+(p6)	add r14=-8,r14				// r14 <- addr of fsys_bubble_down entry
+	;;
+(p6)	ld8 r14=[r14]				// r14 <- fsys_bubble_down
+	;;
+(p6)	mov b7=r14
+(p6)	br.sptk.many b7
+#else
+	BRL_COND_FSYS_BUBBLE_DOWN(p6)
+#endif
+#ifdef __XEN_IA64_VDSO_PARAVIRT
+	XEN_SET_PSR_I(p0)
+#else
+	ssm psr.i
+#endif
+	mov r10=-1
+(p10)	mov r8=EINVAL
+(p9)	mov r8=ENOSYS
+	FSYS_RETURN
+#ifdef __KERNEL_SYSCALL_VIA_EPC_PADDING
+	/*
+	 * All values/sizes of __kernel_xxx symbol in gate.so and xengate.so
+	 * must be same to each other.
+	 * Adjust symbol size in gate.so to be same to the one in xengate.so.
+	 */
+.include "arch/ia64/kernel/gate-skip.s"
+#endif
+END(__kernel_syscall_via_epc)
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/head.S linux-2.6.18-xen-3.3.0/arch/ia64/kernel/head.S
--- linux-2.6.18.8/arch/ia64/kernel/head.S	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/head.S	2008-08-21 11:36:07.000000000 +0200
@@ -367,6 +367,12 @@ start_ap:
 	;;
 (isBP)	st8 [r2]=r28		// save the address of the boot param area passed by the bootloader
 
+#ifdef CONFIG_XEN
+	//  Note: isBP is used by the subprogram.
+	br.call.sptk.many rp=early_xen_setup
+	;;
+#endif
+
 #ifdef CONFIG_SMP
 (isAP)	br.call.sptk.many rp=start_secondary
 .ret0:
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/iosapic.c linux-2.6.18-xen-3.3.0/arch/ia64/kernel/iosapic.c
--- linux-2.6.18.8/arch/ia64/kernel/iosapic.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/iosapic.c	2008-08-21 11:36:07.000000000 +0200
@@ -159,6 +159,75 @@ static unsigned char pcat_compat __devin
 static int iosapic_kmalloc_ok;
 static LIST_HEAD(free_rte_list);
 
+#ifdef CONFIG_XEN
+#include <xen/interface/xen.h>
+#include <xen/interface/physdev.h>
+#include <asm/hypervisor.h>
+static inline unsigned int xen_iosapic_read(char __iomem *iosapic, unsigned int reg)
+{
+	struct physdev_apic apic_op;
+	int ret;
+
+	apic_op.apic_physbase = (unsigned long)iosapic -
+					__IA64_UNCACHED_OFFSET;
+	apic_op.reg = reg;
+	ret = HYPERVISOR_physdev_op(PHYSDEVOP_apic_read, &apic_op);
+	if (ret)
+		return ret;
+	return apic_op.value;
+}
+
+static inline void xen_iosapic_write(char __iomem *iosapic, unsigned int reg, u32 val)
+{
+	struct physdev_apic apic_op;
+
+	apic_op.apic_physbase = (unsigned long)iosapic - 
+					__IA64_UNCACHED_OFFSET;
+	apic_op.reg = reg;
+	apic_op.value = val;
+	HYPERVISOR_physdev_op(PHYSDEVOP_apic_write, &apic_op);
+}
+
+static inline unsigned int iosapic_read(char __iomem *iosapic, unsigned int reg)
+{
+	if (!is_running_on_xen()) {
+		writel(reg, iosapic + IOSAPIC_REG_SELECT);
+		return readl(iosapic + IOSAPIC_WINDOW);
+	} else
+		return xen_iosapic_read(iosapic, reg);
+}
+
+static inline void iosapic_write(char __iomem *iosapic, unsigned int reg, u32 val)
+{
+	if (!is_running_on_xen()) {
+		writel(reg, iosapic + IOSAPIC_REG_SELECT);
+		writel(val, iosapic + IOSAPIC_WINDOW);
+	} else
+		xen_iosapic_write(iosapic, reg, val);
+}
+
+int xen_assign_irq_vector(int irq)
+{
+	struct physdev_irq irq_op;
+
+	irq_op.irq = irq;
+	if (HYPERVISOR_physdev_op(PHYSDEVOP_alloc_irq_vector, &irq_op))
+		return -ENOSPC;
+
+	return irq_op.vector;
+}
+
+void xen_free_irq_vector(int vector)
+{
+	struct physdev_irq irq_op;
+
+	irq_op.vector = vector;
+	if (HYPERVISOR_physdev_op(PHYSDEVOP_free_irq_vector, &irq_op))
+		printk(KERN_WARNING "%s: xen_free_irq_vecotr fail vector=%d\n",
+		       __FUNCTION__, vector);
+}
+#endif /* XEN */
+
 /*
  * Find an IOSAPIC associated with a GSI
  */
@@ -288,6 +357,27 @@ nop (unsigned int irq)
 	/* do nothing... */
 }
 
+
+#ifdef CONFIG_KEXEC
+void
+kexec_disable_iosapic(void)
+{
+	struct iosapic_intr_info *info;
+	struct iosapic_rte_info *rte;
+	u8 vec = 0;
+	for (info = iosapic_intr_info; info <
+			iosapic_intr_info + IA64_NUM_VECTORS; ++info, ++vec) {
+		list_for_each_entry(rte, &info->rtes,
+				rte_list) {
+			iosapic_write(rte->addr,
+					IOSAPIC_RTE_LOW(rte->rte_index),
+					IOSAPIC_MASK|vec);
+			iosapic_eoi(rte->addr, vec);
+		}
+	}
+}
+#endif
+
 static void
 mask_irq (unsigned int irq)
 {
@@ -653,6 +743,9 @@ register_intr (unsigned int gsi, int vec
 	iosapic_intr_info[vector].dmode    = delivery;
 	iosapic_intr_info[vector].trigger  = trigger;
 
+	if (is_running_on_xen())
+		return 0;
+
 	if (trigger == IOSAPIC_EDGE)
 		irq_type = &irq_type_iosapic_edge;
 	else
@@ -1015,6 +1108,9 @@ iosapic_system_init (int system_pcat_com
 	}
 
 	pcat_compat = system_pcat_compat;
+	if (is_running_on_xen())
+		return;
+
 	if (pcat_compat) {
 		/*
 		 * Disable the compatibility mode interrupts (8259 style),
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/irq_ia64.c linux-2.6.18-xen-3.3.0/arch/ia64/kernel/irq_ia64.c
--- linux-2.6.18.8/arch/ia64/kernel/irq_ia64.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/irq_ia64.c	2008-08-21 11:36:07.000000000 +0200
@@ -30,6 +30,9 @@
 #include <linux/smp_lock.h>
 #include <linux/threads.h>
 #include <linux/bitops.h>
+#ifdef CONFIG_XEN
+#include <linux/cpu.h>
+#endif
 
 #include <asm/delay.h>
 #include <asm/intrinsics.h>
@@ -69,6 +72,13 @@ int
 assign_irq_vector (int irq)
 {
 	int pos, vector;
+
+#ifdef CONFIG_XEN
+	if (is_running_on_xen()) {
+		extern int xen_assign_irq_vector(int);
+		return xen_assign_irq_vector(irq);
+	}
+#endif
  again:
 	pos = find_first_zero_bit(ia64_vector_mask, IA64_NUM_DEVICE_VECTORS);
 	vector = IA64_FIRST_DEVICE_VECTOR + pos;
@@ -87,6 +97,13 @@ free_irq_vector (int vector)
 	if (vector < IA64_FIRST_DEVICE_VECTOR || vector > IA64_LAST_DEVICE_VECTOR)
 		return;
 
+#ifdef CONFIG_XEN
+	if (is_running_on_xen()) {
+		extern void xen_free_irq_vector(int);
+		xen_free_irq_vector(vector);
+		return;
+	}
+#endif
 	pos = vector - IA64_FIRST_DEVICE_VECTOR;
 	if (!test_and_clear_bit(pos, ia64_vector_mask))
 		printk(KERN_WARNING "%s: double free!\n", __FUNCTION__);
@@ -240,12 +257,342 @@ static struct irqaction ipi_irqaction = 
 };
 #endif
 
+#ifdef CONFIG_XEN
+#include <xen/evtchn.h>
+#include <xen/interface/callback.h>
+
+static DEFINE_PER_CPU(int, timer_irq) = -1;
+static DEFINE_PER_CPU(int, ipi_irq) = -1;
+static DEFINE_PER_CPU(int, resched_irq) = -1;
+static DEFINE_PER_CPU(int, cmc_irq) = -1;
+static DEFINE_PER_CPU(int, cmcp_irq) = -1;
+static DEFINE_PER_CPU(int, cpep_irq) = -1;
+static char timer_name[NR_CPUS][15];
+static char ipi_name[NR_CPUS][15];
+static char resched_name[NR_CPUS][15];
+static char cmc_name[NR_CPUS][15];
+static char cmcp_name[NR_CPUS][15];
+static char cpep_name[NR_CPUS][15];
+
+struct saved_irq {
+	unsigned int irq;
+	struct irqaction *action;
+};
+/* 16 should be far optimistic value, since only several percpu irqs
+ * are registered early.
+ */
+#define MAX_LATE_IRQ	16
+static struct saved_irq saved_percpu_irqs[MAX_LATE_IRQ];
+static unsigned short late_irq_cnt = 0;
+static unsigned short saved_irq_cnt = 0;
+static int xen_slab_ready = 0;
+
+#ifdef CONFIG_SMP
+/* Dummy stub. Though we may check RESCHEDULE_VECTOR before __do_IRQ,
+ * it ends up to issue several memory accesses upon percpu data and
+ * thus adds unnecessary traffic to other paths.
+ */
+static irqreturn_t
+handle_reschedule(int irq, void *dev_id, struct pt_regs *regs)
+{
+
+	return IRQ_HANDLED;
+}
+
+static struct irqaction resched_irqaction = {
+	.handler =	handle_reschedule,
+	.flags =	SA_INTERRUPT,
+	.name =		"RESCHED"
+};
+#endif
+
+/*
+ * This is xen version percpu irq registration, which needs bind
+ * to xen specific evtchn sub-system. One trick here is that xen
+ * evtchn binding interface depends on kmalloc because related
+ * port needs to be freed at device/cpu down. So we cache the
+ * registration on BSP before slab is ready and then deal them
+ * at later point. For rest instances happening after slab ready,
+ * we hook them to xen evtchn immediately.
+ *
+ * FIXME: MCA is not supported by far, and thus "nomca" boot param is
+ * required.
+ */
+static void
+xen_register_percpu_irq(unsigned int cpu, unsigned int vec,
+			 struct irqaction *action, int save)
+{
+	irq_desc_t *desc;
+	int irq = 0;
+
+	if (xen_slab_ready) {
+		switch (vec) {
+		case IA64_TIMER_VECTOR:
+			sprintf(timer_name[cpu], "%s%d", action->name, cpu);
+			irq = bind_virq_to_irqhandler(VIRQ_ITC, cpu,
+				action->handler, action->flags,
+				timer_name[cpu], action->dev_id);
+			per_cpu(timer_irq,cpu) = irq;
+			break;
+		case IA64_IPI_RESCHEDULE:
+			sprintf(resched_name[cpu], "%s%d", action->name, cpu);
+			irq = bind_ipi_to_irqhandler(RESCHEDULE_VECTOR, cpu,
+				action->handler, action->flags,
+				resched_name[cpu], action->dev_id);
+			per_cpu(resched_irq,cpu) = irq;
+			break;
+		case IA64_IPI_VECTOR:
+			sprintf(ipi_name[cpu], "%s%d", action->name, cpu);
+			irq = bind_ipi_to_irqhandler(IPI_VECTOR, cpu,
+				action->handler, action->flags,
+				ipi_name[cpu], action->dev_id);
+			per_cpu(ipi_irq,cpu) = irq;
+			break;
+		case IA64_CMC_VECTOR:
+			sprintf(cmc_name[cpu], "%s%d", action->name, cpu);
+			irq = bind_virq_to_irqhandler(VIRQ_MCA_CMC, cpu,
+			                              action->handler,
+			                              action->flags,
+			                              cmc_name[cpu],
+			                              action->dev_id);
+			per_cpu(cmc_irq,cpu) = irq;
+			break;
+		case IA64_CMCP_VECTOR:
+			sprintf(cmcp_name[cpu], "%s%d", action->name, cpu);
+			irq = bind_ipi_to_irqhandler(CMCP_VECTOR, cpu,
+			                             action->handler,
+			                             action->flags,
+			                             cmcp_name[cpu],
+			                             action->dev_id);
+			per_cpu(cmcp_irq,cpu) = irq;
+			break;
+		case IA64_CPEP_VECTOR:
+			sprintf(cpep_name[cpu], "%s%d", action->name, cpu);
+			irq = bind_ipi_to_irqhandler(CPEP_VECTOR, cpu,
+			                             action->handler,
+			                             action->flags,
+			                             cpep_name[cpu],
+			                             action->dev_id);
+			per_cpu(cpep_irq,cpu) = irq;
+			break;
+		case IA64_CPE_VECTOR:
+		case IA64_MCA_RENDEZ_VECTOR:
+		case IA64_PERFMON_VECTOR:
+		case IA64_MCA_WAKEUP_VECTOR:
+		case IA64_SPURIOUS_INT_VECTOR:
+			/* No need to complain, these aren't supported. */
+			break;
+		default:
+			printk(KERN_WARNING "Percpu irq %d is unsupported "
+			       "by xen!\n", vec);
+			break;
+		}
+		BUG_ON(irq < 0);
+
+		if (irq > 0) {
+			/*
+			 * Mark percpu.  Without this, migrate_irqs() will
+			 * mark the interrupt for migrations and trigger it
+			 * on cpu hotplug.
+			 */
+			desc = irq_desc + irq;
+			desc->status |= IRQ_PER_CPU;
+		}
+	} 
+
+	/* For BSP, we cache registered percpu irqs, and then re-walk
+	 * them when initializing APs
+	 */
+	if (!cpu && save) {
+		BUG_ON(saved_irq_cnt == MAX_LATE_IRQ);
+		saved_percpu_irqs[saved_irq_cnt].irq = vec;
+		saved_percpu_irqs[saved_irq_cnt].action = action;
+		saved_irq_cnt++;
+		if (!xen_slab_ready)
+			late_irq_cnt++;
+	}
+}
+
+static void
+xen_bind_early_percpu_irq (void)
+{
+	int i;
+
+	xen_slab_ready = 1;
+	/* There's no race when accessing this cached array, since only
+	 * BSP will face with such step shortly
+	 */
+	for (i = 0; i < late_irq_cnt; i++)
+		xen_register_percpu_irq(smp_processor_id(),
+					saved_percpu_irqs[i].irq,
+		                        saved_percpu_irqs[i].action, 0);
+}
+
+/* FIXME: There's no obvious point to check whether slab is ready. So
+ * a hack is used here by utilizing a late time hook.
+ */
+extern void (*late_time_init)(void);
+extern char xen_event_callback;
+extern void xen_init_IRQ(void);
+
+#ifdef CONFIG_HOTPLUG_CPU
+static int __devinit
+unbind_evtchn_callback(struct notifier_block *nfb,
+                       unsigned long action, void *hcpu)
+{
+	unsigned int cpu = (unsigned long)hcpu;
+
+	if (action == CPU_DEAD) {
+		/* Unregister evtchn.  */
+		if (per_cpu(cpep_irq,cpu) >= 0) {
+			unbind_from_irqhandler(per_cpu(cpep_irq, cpu), NULL);
+			per_cpu(cpep_irq, cpu) = -1;
+		}
+		if (per_cpu(cmcp_irq,cpu) >= 0) {
+			unbind_from_irqhandler(per_cpu(cmcp_irq, cpu), NULL);
+			per_cpu(cmcp_irq, cpu) = -1;
+		}
+		if (per_cpu(cmc_irq,cpu) >= 0) {
+			unbind_from_irqhandler(per_cpu(cmc_irq, cpu), NULL);
+			per_cpu(cmc_irq, cpu) = -1;
+		}
+		if (per_cpu(ipi_irq,cpu) >= 0) {
+			unbind_from_irqhandler (per_cpu(ipi_irq, cpu), NULL);
+			per_cpu(ipi_irq, cpu) = -1;
+		}
+		if (per_cpu(resched_irq,cpu) >= 0) {
+			unbind_from_irqhandler (per_cpu(resched_irq, cpu),
+						NULL);
+			per_cpu(resched_irq, cpu) = -1;
+		}
+		if (per_cpu(timer_irq,cpu) >= 0) {
+			unbind_from_irqhandler (per_cpu(timer_irq, cpu), NULL);
+			per_cpu(timer_irq, cpu) = -1;
+		}
+	}
+	return NOTIFY_OK;
+}
+
+static struct notifier_block unbind_evtchn_notifier = {
+	.notifier_call = unbind_evtchn_callback,
+	.priority = 0
+};
+#endif
+
+DECLARE_PER_CPU(int, ipi_to_irq[NR_IPIS]);
+void xen_smp_intr_init_early(unsigned int cpu)
+{
+#ifdef CONFIG_SMP
+	unsigned int i;
+
+	for (i = 0; i < saved_irq_cnt; i++)
+		xen_register_percpu_irq(cpu, saved_percpu_irqs[i].irq,
+		                        saved_percpu_irqs[i].action, 0);
+#endif
+}
+
+void xen_smp_intr_init(void)
+{
+#ifdef CONFIG_SMP
+	unsigned int cpu = smp_processor_id();
+	struct callback_register event = {
+		.type = CALLBACKTYPE_event,
+		.address = (unsigned long)&xen_event_callback,
+	};
+
+	if (cpu == 0) {
+		/* Initialization was already done for boot cpu.  */
+#ifdef CONFIG_HOTPLUG_CPU
+		/* Register the notifier only once.  */
+		register_cpu_notifier(&unbind_evtchn_notifier);
+#endif
+		return;
+	}
+
+	/* This should be piggyback when setup vcpu guest context */
+	BUG_ON(HYPERVISOR_callback_op(CALLBACKOP_register, &event));
+#endif /* CONFIG_SMP */
+}
+
+void
+xen_irq_init(void)
+{
+	struct callback_register event = {
+		.type = CALLBACKTYPE_event,
+		.address = (unsigned long)&xen_event_callback,
+	};
+
+	xen_init_IRQ();
+	BUG_ON(HYPERVISOR_callback_op(CALLBACKOP_register, &event));
+	late_time_init = xen_bind_early_percpu_irq;
+#ifdef CONFIG_SMP
+	register_percpu_irq(IA64_IPI_RESCHEDULE, &resched_irqaction);
+#endif
+}
+
+void
+xen_platform_send_ipi(int cpu, int vector, int delivery_mode, int redirect)
+{
+	int irq = -1;
+	extern void xen_send_ipi(int cpu, int vec);
+
+#ifdef CONFIG_SMP
+	/* TODO: we need to call vcpu_up here */
+	if (unlikely(vector == ap_wakeup_vector)) {
+		/* XXX
+		 * This should be in __cpu_up(cpu) in ia64 smpboot.c
+		 * like x86. But don't want to modify it,
+		 * keep it untouched.
+		 */
+		xen_smp_intr_init_early(cpu);
+
+		xen_send_ipi (cpu, vector);
+		//vcpu_prepare_and_up(cpu);
+		return;
+	}
+#endif
+
+	switch (vector) {
+		case IA64_IPI_VECTOR:
+			irq = per_cpu(ipi_to_irq, cpu)[IPI_VECTOR];
+			break;
+		case IA64_IPI_RESCHEDULE:
+			irq = per_cpu(ipi_to_irq, cpu)[RESCHEDULE_VECTOR];
+			break;
+		case IA64_CMCP_VECTOR:
+			irq = per_cpu(ipi_to_irq, cpu)[CMCP_VECTOR];
+			break;
+		case IA64_CPEP_VECTOR:
+			irq = per_cpu(ipi_to_irq, cpu)[CPEP_VECTOR];
+			break;
+		case IA64_TIMER_VECTOR:
+			xen_send_ipi(cpu, vector);
+			return;
+		default:
+			printk(KERN_WARNING "Unsupported IPI type 0x%x\n",
+			       vector);
+			irq = 0;
+			break;
+	}		
+	
+	BUG_ON(irq < 0);
+	notify_remote_via_irq(irq);
+	return;
+}
+#endif /* CONFIG_XEN */
+
 void
 register_percpu_irq (ia64_vector vec, struct irqaction *action)
 {
 	irq_desc_t *desc;
 	unsigned int irq;
 
+#ifdef CONFIG_XEN
+	if (is_running_on_xen())
+		return xen_register_percpu_irq(smp_processor_id(), 
+					       vec, action, 1);
+#endif
+
 	for (irq = 0; irq < NR_IRQS; ++irq)
 		if (irq_to_vector(irq) == vec) {
 			desc = irq_desc + irq;
@@ -267,6 +614,10 @@ init_IRQ (void)
 	pfm_init_percpu();
 #endif
 	platform_irq_init();
+#ifdef CONFIG_XEN
+	if (is_running_on_xen() && !ia64_platform_is("xen"))
+		xen_irq_init();
+#endif
 }
 
 void
@@ -276,6 +627,13 @@ ia64_send_ipi (int cpu, int vector, int 
 	unsigned long ipi_data;
 	unsigned long phys_cpu_id;
 
+#ifdef CONFIG_XEN
+	if (is_running_on_xen()) {
+		xen_platform_send_ipi(cpu, vector, delivery_mode, redirect);
+		return;
+	}
+#endif
+
 #ifdef CONFIG_SMP
 	phys_cpu_id = cpu_physical_id(cpu);
 #else
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/machine_kexec.c linux-2.6.18-xen-3.3.0/arch/ia64/kernel/machine_kexec.c
--- linux-2.6.18.8/arch/ia64/kernel/machine_kexec.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/machine_kexec.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,204 @@
+/*
+ * arch/ia64/kernel/machine_kexec.c
+ *
+ * Handle transition of Linux booting another kernel
+ * Copyright (C) 2005 Hewlett-Packard Development Comapny, L.P.
+ * Copyright (C) 2005 Khalid Aziz <khalid.aziz@hp.com>
+ * Copyright (C) 2006 Intel Corp, Zou Nan hai <nanhai.zou@intel.com>
+ *
+ * This source code is licensed under the GNU General Public License,
+ * Version 2.  See the file COPYING for more details.
+ */
+
+#include <linux/mm.h>
+#include <linux/kexec.h>
+#include <linux/cpu.h>
+#include <linux/irq.h>
+#include <asm/mmu_context.h>
+#include <asm/setup.h>
+#include <asm/delay.h>
+#include <asm/meminit.h>
+#ifdef CONFIG_XEN
+#include <xen/interface/kexec.h>
+#include <asm/kexec.h>
+#endif
+
+typedef void (*relocate_new_kernel_t)(unsigned long, unsigned long,
+		struct ia64_boot_param *, unsigned long);
+
+struct kimage *ia64_kimage;
+
+struct resource efi_memmap_res = {
+        .name  = "EFI Memory Map",
+        .start = 0,
+        .end   = 0,
+        .flags = IORESOURCE_BUSY | IORESOURCE_MEM
+};
+
+struct resource boot_param_res = {
+        .name  = "Boot parameter",
+        .start = 0,
+        .end   = 0,
+        .flags = IORESOURCE_BUSY | IORESOURCE_MEM
+};
+
+
+/*
+ * Do what every setup is needed on image and the
+ * reboot code buffer to allow us to avoid allocations
+ * later.
+ */
+int machine_kexec_prepare(struct kimage *image)
+{
+	void *control_code_buffer;
+	const unsigned long *func;
+
+	func = (unsigned long *)&relocate_new_kernel;
+	/* Pre-load control code buffer to minimize work in kexec path */
+	control_code_buffer = page_address(image->control_code_page);
+	memcpy((void *)control_code_buffer, (const void *)func[0],
+			relocate_new_kernel_size);
+	flush_icache_range((unsigned long)control_code_buffer,
+			(unsigned long)control_code_buffer + relocate_new_kernel_size);
+	ia64_kimage = image;
+
+	return 0;
+}
+
+void machine_kexec_cleanup(struct kimage *image)
+{
+}
+
+#ifndef CONFIG_XEN
+void machine_shutdown(void)
+{
+	int cpu;
+
+	for_each_online_cpu(cpu) {
+		if (cpu != smp_processor_id())
+			cpu_down(cpu);
+	}
+	kexec_disable_iosapic();
+}
+
+/*
+ * Do not allocate memory (or fail in any way) in machine_kexec().
+ * We are past the point of no return, committed to rebooting now.
+ */
+extern void *efi_get_pal_addr(void);
+static void ia64_machine_kexec(struct unw_frame_info *info, void *arg)
+{
+	struct kimage *image = arg;
+	relocate_new_kernel_t rnk;
+	void *pal_addr = efi_get_pal_addr();
+	unsigned long code_addr = (unsigned long)page_address(image->control_code_page);
+	unsigned long vector;
+	int ii;
+
+	if (image->type == KEXEC_TYPE_CRASH) {
+		crash_save_this_cpu();
+		current->thread.ksp = (__u64)info->sw - 16;
+	}
+
+	/* Interrupts aren't acceptable while we reboot */
+	local_irq_disable();
+
+	/* Mask CMC and Performance Monitor interrupts */
+	ia64_setreg(_IA64_REG_CR_PMV, 1 << 16);
+	ia64_setreg(_IA64_REG_CR_CMCV, 1 << 16);
+
+	/* Mask ITV and Local Redirect Registers */
+	ia64_set_itv(1 << 16);
+	ia64_set_lrr0(1 << 16);
+	ia64_set_lrr1(1 << 16);
+
+	/* terminate possible nested in-service interrupts */
+	for (ii = 0; ii < 16; ii++)
+		ia64_eoi();
+
+	/* unmask TPR and clear any pending interrupts */
+	ia64_setreg(_IA64_REG_CR_TPR, 0);
+	ia64_srlz_d();
+	vector = ia64_get_ivr();
+	while (vector != IA64_SPURIOUS_INT_VECTOR) {
+		ia64_eoi();
+		vector = ia64_get_ivr();
+	}
+	platform_kernel_launch_event();
+	rnk = (relocate_new_kernel_t)&code_addr;
+	(*rnk)(image->head, image->start, ia64_boot_param,
+		     GRANULEROUNDDOWN((unsigned long) pal_addr));
+	BUG();
+}
+
+void machine_kexec(struct kimage *image)
+{
+	unw_init_running(ia64_machine_kexec, image);
+	for(;;);
+}
+#else /* CONFIG_XEN */
+void machine_kexec_setup_load_arg(xen_kexec_image_t *xki,struct kimage *image)
+{
+	xki->reboot_code_buffer =
+		kexec_page_to_pfn(image->control_code_page) << PAGE_SHIFT;
+}
+
+static struct resource xen_hypervisor_heap_res;
+
+int __init machine_kexec_setup_resources(struct resource *hypervisor,
+					 struct resource *phys_cpus,
+					 int nr_phys_cpus)
+{
+	xen_kexec_range_t range;
+	int k;
+
+	/* fill in xen_hypervisor_heap_res with hypervisor heap
+	 * machine address range
+	 */
+
+	memset(&range, 0, sizeof(range));
+	range.range = KEXEC_RANGE_MA_XENHEAP;
+
+	if (HYPERVISOR_kexec_op(KEXEC_CMD_kexec_get_range, &range))
+		return -1;
+
+	xen_hypervisor_heap_res.name = "Hypervisor heap";
+	xen_hypervisor_heap_res.start = range.start;
+	xen_hypervisor_heap_res.end = range.start + range.size - 1;
+	xen_hypervisor_heap_res.flags = IORESOURCE_BUSY | IORESOURCE_MEM;
+
+	/* The per-cpu crash note  resources belong inside the
+	 * hypervisor heap resource */
+	for (k = 0; k < nr_phys_cpus; k++)
+		request_resource(&xen_hypervisor_heap_res, phys_cpus + k);
+
+	/* fill in efi_memmap_res with EFI memmap machine address range */
+
+	memset(&range, 0, sizeof(range));
+	range.range = KEXEC_RANGE_MA_EFI_MEMMAP;
+
+	if (HYPERVISOR_kexec_op(KEXEC_CMD_kexec_get_range, &range))
+		return -1;
+
+	efi_memmap_res.start = range.start;
+	efi_memmap_res.end = range.start + range.size - 1;
+
+	/* fill in boot_param_res with boot parameter machine address range */
+
+	memset(&range, 0, sizeof(range));
+	range.range = KEXEC_RANGE_MA_BOOT_PARAM;
+
+	if (HYPERVISOR_kexec_op(KEXEC_CMD_kexec_get_range, &range))
+		return -1;
+
+	boot_param_res.start = range.start;
+	boot_param_res.end = range.start + range.size - 1;
+
+	return 0;
+}
+
+void machine_kexec_register_resources(struct resource *res)
+{
+	request_resource(res, &xen_hypervisor_heap_res);
+}
+#endif /* CONFIG_XEN */
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/Makefile linux-2.6.18-xen-3.3.0/arch/ia64/kernel/Makefile
--- linux-2.6.18.8/arch/ia64/kernel/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -28,6 +28,8 @@ obj-$(CONFIG_IA64_CYCLONE)	+= cyclone.o
 obj-$(CONFIG_CPU_FREQ)		+= cpufreq/
 obj-$(CONFIG_IA64_MCA_RECOVERY)	+= mca_recovery.o
 obj-$(CONFIG_KPROBES)		+= kprobes.o jprobes.o
+obj-$(CONFIG_KEXEC)		+= machine_kexec.o relocate_kernel.o crash.o
+obj-$(CONFIG_CRASH_DUMP)	+= crash_dump.o
 obj-$(CONFIG_IA64_UNCACHED_ALLOCATOR)	+= uncached.o
 obj-$(CONFIG_AUDIT)		+= audit.o
 mca_recovery-y			+= mca_drv.o mca_drv_asm.o
@@ -61,3 +63,61 @@ $(obj)/gate-syms.o: $(obj)/gate.lds $(ob
 # We must build gate.so before we can assemble it.
 # Note: kbuild does not track this dependency due to usage of .incbin
 $(obj)/gate-data.o: $(obj)/gate.so
+
+#
+# gate page paravirtualization for xen
+#
+obj-$(CONFIG_XEN) += xengate-data.o
+
+ifeq ($(CONFIG_XEN), y)
+# The gate DSO image is built using a special linker script.
+targets += xengate.so xengate-syms.o
+endif
+
+extra-$(CONFIG_XEN) += xengate.so xengate.lds xengate.o
+
+AFLAGS_xengate.o += -D__XEN_IA64_VDSO_PARAVIRT
+$(obj)/xengate.o: $(src)/gate.S FORCE
+	$(call if_changed_dep,as_o_S)
+
+CPPFLAGS_xengate.lds := -P -C -U$(ARCH) -D__XEN_IA64_VDSO_PARAVIRT
+$(obj)/xengate.lds: $(src)/gate.lds.S
+	$(call if_changed_dep,cpp_lds_S)
+
+GATECFLAGS_xengate.so = -shared -s -Wl,-soname=linux-gate.so.1 \
+		     $(call ld-option, -Wl$(comma)--hash-style=sysv)
+$(obj)/xengate.so: $(obj)/xengate.lds $(obj)/xengate.o FORCE
+	$(call if_changed,gate)
+
+ifeq ($(CONFIG_XEN), y)
+$(obj)/built-in.o: $(obj)/xengate-syms.o
+$(obj)/built-in.o: ld_flags += -R $(obj)/xengate-syms.o
+$(obj)/mca_recovery.o: $(obj)/gate-syms.o $(obj)/xengate-syms.o
+endif
+
+GATECFLAGS_xengate-syms.o = -r
+$(obj)/xengate-syms.o: $(obj)/xengate.lds $(obj)/xengate.o FORCE
+	$(call if_changed,gate)
+$(obj)/xengate-data.o: $(obj)/xengate.so
+
+#
+# .tmp_gate.o to calculate padding size for __kernel_syscall_via_epc
+#
+extra-$(CONFIG_XEN) += gate-skip.s .tmp_gate.o
+
+ifeq ($(CONFIG_XEN), y)
+AFLAGS_gate.o += -D__KERNEL_SYSCALL_VIA_EPC_PADDING
+$(obj)/gate.o: $(obj)/gate-skip.s FORCE
+endif
+
+$(obj)/.tmp_gate.o: $(src)/gate.S FORCE
+	$(call if_changed_dep,as_o_S)
+
+quiet_cmd_gate_size = GATE_SIZE	$@
+      cmd_gate_size = $(NM) --extern-only --print-size $(obj)/xengate.o | \
+	$(AWK) '/__kernel_syscall_via_epc/{printf "\t.skip 0x"$$2" - "}' > $@; \
+	$(NM) --extern-only --print-size $(obj)/.tmp_gate.o | \
+	$(AWK) '/__kernel_syscall_via_epc/{printf "0x"$$2"\n"}' >> $@
+
+$(obj)/gate-skip.s: $(obj)/xengate.o $(obj)/.tmp_gate.o FORCE
+	$(call if_changed,gate_size)
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/mca.c linux-2.6.18-xen-3.3.0/arch/ia64/kernel/mca.c
--- linux-2.6.18.8/arch/ia64/kernel/mca.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/mca.c	2008-08-21 11:36:07.000000000 +0200
@@ -79,6 +79,7 @@
 #include <asm/system.h>
 #include <asm/sal.h>
 #include <asm/mca.h>
+#include <asm/kexec.h>
 
 #include <asm/irq.h>
 #include <asm/hw_irq.h>
@@ -160,11 +161,33 @@ typedef struct ia64_state_log_s
 
 static ia64_state_log_t ia64_state_log[IA64_MAX_LOG_TYPES];
 
+#ifdef CONFIG_XEN
+DEFINE_SPINLOCK(ia64_mca_xencomm_lock);
+LIST_HEAD(ia64_mca_xencomm_list);
+
+#define IA64_MCA_XENCOMM_ALLOCATE(rec, desc) \
+	if (is_running_on_xen()) { \
+		ia64_mca_xencomm_t *entry; \
+		entry = alloc_bootmem(sizeof(ia64_mca_xencomm_t)); \
+		entry->record = rec; \
+		entry->handle = desc; \
+		list_add(&entry->list, &ia64_mca_xencomm_list); \
+	}
+#define IA64_LOG_ALLOCATE(it, size) \
+	{ia64_err_rec_t *rec; \
+	ia64_state_log[it].isl_log[IA64_LOG_CURR_INDEX(it)] = rec = \
+		(ia64_err_rec_t *)alloc_bootmem(size); \
+	IA64_MCA_XENCOMM_ALLOCATE(rec, xencomm_map(rec, size)); \
+	ia64_state_log[it].isl_log[IA64_LOG_NEXT_INDEX(it)] = rec = \
+		(ia64_err_rec_t *)alloc_bootmem(size); \
+	IA64_MCA_XENCOMM_ALLOCATE(rec, xencomm_map(rec, size));}
+#else
 #define IA64_LOG_ALLOCATE(it, size) \
 	{ia64_state_log[it].isl_log[IA64_LOG_CURR_INDEX(it)] = \
 		(ia64_err_rec_t *)alloc_bootmem(size); \
 	ia64_state_log[it].isl_log[IA64_LOG_NEXT_INDEX(it)] = \
 		(ia64_err_rec_t *)alloc_bootmem(size);}
+#endif
 #define IA64_LOG_LOCK_INIT(it) spin_lock_init(&ia64_state_log[it].isl_lock)
 #define IA64_LOG_LOCK(it)      spin_lock_irqsave(&ia64_state_log[it].isl_lock, s)
 #define IA64_LOG_UNLOCK(it)    spin_unlock_irqrestore(&ia64_state_log[it].isl_lock,s)
@@ -1066,7 +1089,12 @@ ia64_mca_handler(struct pt_regs *regs, s
 		rh->severity = sal_log_severity_corrected;
 		ia64_sal_clear_state_info(SAL_INFO_TYPE_MCA);
 		sos->os_status = IA64_MCA_CORRECTED;
-	}
+ 	} else {
+#ifdef CONFIG_KEXEC
+		atomic_set(&kdump_in_progress, 1);
+		monarch_cpu = -1;
+#endif
+ 	}
 	if (notify_die(DIE_MCA_MONARCH_LEAVE, "MCA", regs, (long)&nd, 0, recover)
 			== NOTIFY_STOP)
 		ia64_mca_spin(__FUNCTION__);
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/pal.S linux-2.6.18-xen-3.3.0/arch/ia64/kernel/pal.S
--- linux-2.6.18.8/arch/ia64/kernel/pal.S	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/pal.S	2008-08-21 11:36:07.000000000 +0200
@@ -16,6 +16,7 @@
 #include <asm/processor.h>
 
 	.data
+	.globl pal_entry_point
 pal_entry_point:
 	data8 ia64_pal_default_handler
 	.text
@@ -53,7 +54,7 @@ END(ia64_pal_default_handler)
  * in4	       1 ==> clear psr.ic,  0 ==> don't clear psr.ic
  *
  */
-GLOBAL_ENTRY(ia64_pal_call_static)
+GLOBAL_ENTRY(__ia64_pal_call_static)
 	.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(5)
 	alloc loc1 = ar.pfs,5,5,0,0
 	movl loc2 = pal_entry_point
@@ -90,7 +91,7 @@ GLOBAL_ENTRY(ia64_pal_call_static)
 	;;
 	srlz.d				// seralize restoration of psr.l
 	br.ret.sptk.many b0
-END(ia64_pal_call_static)
+END(__ia64_pal_call_static)
 
 /*
  * Make a PAL call using the stacked registers calling convention.
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/patch.c linux-2.6.18-xen-3.3.0/arch/ia64/kernel/patch.c
--- linux-2.6.18.8/arch/ia64/kernel/patch.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/patch.c	2008-08-21 11:36:07.000000000 +0200
@@ -184,9 +184,37 @@ patch_brl_fsys_bubble_down (unsigned lon
 	ia64_srlz_i();
 }
 
+#ifdef CONFIG_XEN
+void __init
+ia64_patch_gate_xen (void)
+{
+	extern char __start_gate_mckinley_e9_patchlist_xen[], __end_gate_mckinley_e9_patchlist_xen[];
+	extern char __start_gate_vtop_patchlist_xen[], __end_gate_vtop_patchlist_xen[];
+	extern char __start_gate_fsyscall_patchlist_xen[], __end_gate_fsyscall_patchlist_xen[];
+	extern char __start_gate_brl_fsys_bubble_down_patchlist_xen[], __end_gate_brl_fsys_bubble_down_patchlist_xen[];
+#	define START(name)	((unsigned long) __start_gate_##name##_patchlist_xen)
+#	define END(name)	((unsigned long)__end_gate_##name##_patchlist_xen)
+
+	patch_fsyscall_table(START(fsyscall), END(fsyscall));
+	patch_brl_fsys_bubble_down(START(brl_fsys_bubble_down), END(brl_fsys_bubble_down));
+	ia64_patch_vtop(START(vtop), END(vtop));
+	ia64_patch_mckinley_e9(START(mckinley_e9), END(mckinley_e9));
+
+#	undef START
+#	undef END
+}
+#else
+#define ia64_patch_gate_xen()	do { } while (0)
+#endif
+
 void __init
 ia64_patch_gate (void)
 {
+	if (is_running_on_xen()) {
+		ia64_patch_gate_xen();
+		return;
+	}
+	
 #	define START(name)	((unsigned long) __start_gate_##name##_patchlist)
 #	define END(name)	((unsigned long)__end_gate_##name##_patchlist)
 
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/perfmon.c linux-2.6.18-xen-3.3.0/arch/ia64/kernel/perfmon.c
--- linux-2.6.18.8/arch/ia64/kernel/perfmon.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/perfmon.c	2008-08-21 11:36:07.000000000 +0200
@@ -52,6 +52,31 @@
 #include <asm/delay.h>
 
 #ifdef CONFIG_PERFMON
+#include <asm/hypervisor.h>
+#ifdef CONFIG_XEN
+//#include <xen/xenoprof.h>
+#include <xen/interface/xenoprof.h>
+
+static int xenoprof_is_primary = 0;
+#define init_xenoprof_primary(is_primary)  (xenoprof_is_primary = (is_primary))
+#define is_xenoprof_primary()	(xenoprof_is_primary)
+#define XEN_NOT_SUPPORTED_YET						\
+	do {								\
+		if (is_running_on_xen()) {				\
+			printk("%s is not supported yet under xen.\n",	\
+			       __func__);				\
+			return -ENOSYS;					\
+		}							\
+	} while (0)
+#else
+#define init_xenoprof_primary(is_primary)	do { } while (0)
+#define is_xenoprof_primary()			(0)
+#define XEN_NOT_SUPPORTED_YET			do { } while (0)
+#define HYPERVISOR_perfmon_op(cmd, arg, count)	(0)
+#define HYPERVISOR_xenoprof_op(op, arg)		({(void)arg;0;})
+struct xenoprof_init { /* dummy */ };
+#endif
+
 /*
  * perfmon context state
  */
@@ -1514,6 +1539,7 @@ pfm_read(struct file *filp, char __user 
 	ssize_t ret;
 	unsigned long flags;
   	DECLARE_WAITQUEUE(wait, current);
+	XEN_NOT_SUPPORTED_YET;
 	if (PFM_IS_FILE(filp) == 0) {
 		printk(KERN_ERR "perfmon: pfm_poll: bad magic [%d]\n", current->pid);
 		return -EINVAL;
@@ -2112,6 +2138,15 @@ doit:
 	 */
 	if (free_possible) pfm_context_free(ctx);
 
+	if (is_running_on_xen()) {
+		if (is_xenoprof_primary()) {
+			int ret = HYPERVISOR_perfmon_op(PFM_DESTROY_CONTEXT,
+			                                NULL, 0);
+			if (ret)
+				printk("%s:%d PFM_DESTROY_CONTEXT hypercall "
+				       "failed\n", __func__, __LINE__);
+		}
+	}
 	return 0;
 }
 
@@ -2735,6 +2770,23 @@ pfm_context_create(pfm_context_t *ctx, v
 	 */
 	pfm_reset_pmu_state(ctx);
 
+	if (is_running_on_xen()) {
+		/*
+		 * kludge to get xenoprof.is_primary.
+		 * XENOPROF_init/ia64 is nop. so it is safe to call it here.
+		 */
+		struct xenoprof_init init;
+		ret = HYPERVISOR_xenoprof_op(XENOPROF_init, &init);
+		if (ret)
+			goto buffer_error;
+		init_xenoprof_primary(init.is_primary);
+
+		if (is_xenoprof_primary()) {
+			ret = HYPERVISOR_perfmon_op(PFM_CREATE_CONTEXT, arg, 0);
+			if (ret)
+				goto buffer_error;
+		}
+	}
 	return 0;
 
 buffer_error:
@@ -2871,6 +2923,12 @@ pfm_write_pmcs(pfm_context_t *ctx, void 
 	pfm_reg_check_t	wr_func;
 #define PFM_CHECK_PMC_PM(x, y, z) ((x)->ctx_fl_system ^ PMC_PM(y, z))
 
+  	if (is_running_on_xen()) {
+		if (is_xenoprof_primary())
+			return HYPERVISOR_perfmon_op(PFM_WRITE_PMCS,
+			                             arg, count);
+		return 0;
+  	}
 	state     = ctx->ctx_state;
 	is_loaded = state == PFM_CTX_LOADED ? 1 : 0;
 	is_system = ctx->ctx_fl_system;
@@ -3111,6 +3169,12 @@ pfm_write_pmds(pfm_context_t *ctx, void 
 	int ret = -EINVAL;
 	pfm_reg_check_t wr_func;
 
+  	if (is_running_on_xen()) {
+		if (is_xenoprof_primary())
+			return HYPERVISOR_perfmon_op(PFM_WRITE_PMDS,
+			                             arg, count);
+		return 0;
+  	}
 
 	state     = ctx->ctx_state;
 	is_loaded = state == PFM_CTX_LOADED ? 1 : 0;
@@ -3308,6 +3372,7 @@ pfm_read_pmds(pfm_context_t *ctx, void *
 	int is_loaded, is_system, is_counting, expert_mode;
 	int ret = -EINVAL;
 	pfm_reg_check_t rd_func;
+	XEN_NOT_SUPPORTED_YET;
 
 	/*
 	 * access is possible when loaded only for
@@ -3559,6 +3624,7 @@ pfm_restart(pfm_context_t *ctx, void *ar
 	pfm_ovfl_ctrl_t rst_ctrl;
 	int state, is_system;
 	int ret = 0;
+	XEN_NOT_SUPPORTED_YET;
 
 	state     = ctx->ctx_state;
 	fmt       = ctx->ctx_buf_fmt;
@@ -3708,6 +3774,7 @@ static int
 pfm_debug(pfm_context_t *ctx, void *arg, int count, struct pt_regs *regs)
 {
 	unsigned int m = *(unsigned int *)arg;
+	XEN_NOT_SUPPORTED_YET;
 
 	pfm_sysctl.debug = m == 0 ? 0 : 1;
 
@@ -3978,6 +4045,8 @@ pfm_get_features(pfm_context_t *ctx, voi
 {
 	pfarg_features_t *req = (pfarg_features_t *)arg;
 
+	if (is_running_on_xen())
+		return HYPERVISOR_perfmon_op(PFM_GET_FEATURES, &arg, 0);
 	req->ft_version = PFM_VERSION;
 	return 0;
 }
@@ -3989,6 +4058,12 @@ pfm_stop(pfm_context_t *ctx, void *arg, 
 	struct task_struct *task = PFM_CTX_TASK(ctx);
 	int state, is_system;
 
+  	if (is_running_on_xen()) {
+		if (is_xenoprof_primary())
+			return HYPERVISOR_perfmon_op(PFM_STOP, NULL, 0);
+		return 0;
+  	}
+
 	state     = ctx->ctx_state;
 	is_system = ctx->ctx_fl_system;
 
@@ -4077,6 +4152,11 @@ pfm_start(pfm_context_t *ctx, void *arg,
 	struct pt_regs *tregs;
 	int state, is_system;
 
+  	if (is_running_on_xen()) {
+		if (is_xenoprof_primary())
+			return HYPERVISOR_perfmon_op(PFM_START, NULL, 0);
+		return 0;
+  	}
 	state     = ctx->ctx_state;
 	is_system = ctx->ctx_fl_system;
 
@@ -4159,6 +4239,7 @@ pfm_get_pmc_reset(pfm_context_t *ctx, vo
 	unsigned int cnum;
 	int i;
 	int ret = -EINVAL;
+	XEN_NOT_SUPPORTED_YET;
 
 	for (i = 0; i < count; i++, req++) {
 
@@ -4217,6 +4298,11 @@ pfm_context_load(pfm_context_t *ctx, voi
 	int ret = 0;
 	int state, is_system, set_dbregs = 0;
 
+  	if (is_running_on_xen()) {
+		if (is_xenoprof_primary())
+			return HYPERVISOR_perfmon_op(PFM_LOAD_CONTEXT, arg, 0);
+		return 0;
+  	}
 	state     = ctx->ctx_state;
 	is_system = ctx->ctx_fl_system;
 	/*
@@ -4465,6 +4551,12 @@ pfm_context_unload(pfm_context_t *ctx, v
 	int prev_state, is_system;
 	int ret;
 
+  	if (is_running_on_xen()) {
+		if (is_xenoprof_primary())
+			return HYPERVISOR_perfmon_op(PFM_UNLOAD_CONTEXT,
+			                             NULL, 0);
+		return 0;
+  	}
 	DPRINT(("ctx_state=%d task [%d]\n", ctx->ctx_state, task ? task->pid : -1));
 
 	prev_state = ctx->ctx_state;
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/relocate_kernel.S linux-2.6.18-xen-3.3.0/arch/ia64/kernel/relocate_kernel.S
--- linux-2.6.18.8/arch/ia64/kernel/relocate_kernel.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/relocate_kernel.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,338 @@
+/*
+ * arch/ia64/kernel/relocate_kernel.S
+ *
+ * Relocate kexec'able kernel and start it
+ *
+ * Copyright (C) 2005 Hewlett-Packard Development Company, L.P.
+ * Copyright (C) 2005 Khalid Aziz  <khalid.aziz@hp.com>
+ * Copyright (C) 2005 Intel Corp,  Zou Nan hai <nanhai.zou@intel.com>
+ *
+ * This source code is licensed under the GNU General Public License,
+ * Version 2.  See the file COPYING for more details.
+ */
+#include <asm/asmmacro.h>
+#include <asm/kregs.h>
+#include <asm/page.h>
+#include <asm/pgtable.h>
+#include <asm/mca_asm.h>
+
+       /* Must be relocatable PIC code callable as a C function
+        */
+GLOBAL_ENTRY(relocate_new_kernel)
+	.prologue
+	alloc r31=ar.pfs,4,0,0,0
+        .body
+.reloc_entry:
+{
+	rsm psr.i| psr.ic
+	mov r2=ip
+}
+	;;
+{
+        flushrs                         // must be first insn in group
+        srlz.i
+}
+	;;
+	dep r2=0,r2,61,3		//to physical address
+	;;
+	//first switch to physical mode
+	add r3=1f-.reloc_entry, r2
+	movl r16 = IA64_PSR_AC|IA64_PSR_BN|IA64_PSR_IC
+	mov ar.rsc=0	          	// put RSE in enforced lazy mode
+	;;
+	add sp=(memory_stack_end - 16 - .reloc_entry),r2
+	add r8=(register_stack - .reloc_entry),r2
+	;;
+	mov r18=ar.rnat
+	mov ar.bspstore=r8
+	;;
+        mov cr.ipsr=r16
+        mov cr.iip=r3
+        mov cr.ifs=r0
+	srlz.i
+	;;
+	mov ar.rnat=r18
+	rfi
+	;;
+1:
+	//physical mode code begin
+	mov b6=in1
+#ifdef CONFIG_XEN
+	mov r28=in2		//already a physical address
+#else
+	dep r28=0,in2,61,3	//to physical address
+
+	// purge all TC entries
+#define O(member)       IA64_CPUINFO_##member##_OFFSET
+        GET_THIS_PADDR(r2, cpu_info)    // load phys addr of cpu_info into r2
+        ;;
+        addl r17=O(PTCE_STRIDE),r2
+        addl r2=O(PTCE_BASE),r2
+        ;;
+        ld8 r18=[r2],(O(PTCE_COUNT)-O(PTCE_BASE));;    	// r18=ptce_base
+        ld4 r19=[r2],4                                  // r19=ptce_count[0]
+        ld4 r21=[r17],4                                 // r21=ptce_stride[0]
+        ;;
+        ld4 r20=[r2]                                    // r20=ptce_count[1]
+        ld4 r22=[r17]                                   // r22=ptce_stride[1]
+        mov r24=r0
+        ;;
+        adds r20=-1,r20
+        ;;
+#undef O
+2:
+        cmp.ltu p6,p7=r24,r19
+(p7)    br.cond.dpnt.few 4f
+        mov ar.lc=r20
+3:
+        ptc.e r18
+        ;;
+        add r18=r22,r18
+        br.cloop.sptk.few 3b
+        ;;
+        add r18=r21,r18
+        add r24=1,r24
+        ;;
+        br.sptk.few 2b
+4:
+        srlz.i
+        ;;
+	//purge TR entry for kernel text and data
+        movl r16=KERNEL_START
+        mov r18=KERNEL_TR_PAGE_SHIFT<<2
+        ;;
+        ptr.i r16, r18
+        ptr.d r16, r18
+        ;;
+        srlz.i
+        ;;
+
+	// purge TR entry for percpu data
+        movl r16=PERCPU_ADDR
+        mov r18=PERCPU_PAGE_SHIFT<<2
+        ;;
+        ptr.d r16,r18
+        ;;
+        srlz.d
+	;;
+
+        // purge TR entry for pal code
+        mov r16=in3
+        mov r18=IA64_GRANULE_SHIFT<<2
+        ;;
+        ptr.i r16,r18
+        ;;
+        srlz.i
+	;;
+
+        // purge TR entry for stack
+        mov r16=IA64_KR(CURRENT_STACK)
+        ;;
+        shl r16=r16,IA64_GRANULE_SHIFT
+        movl r19=PAGE_OFFSET
+        ;;
+        add r16=r19,r16
+        mov r18=IA64_GRANULE_SHIFT<<2
+        ;;
+        ptr.d r16,r18
+        ;;
+        srlz.i
+	;;
+#endif /* ! CONFIG_XEN */
+
+	//copy segments
+	movl r16=PAGE_MASK
+        mov  r30=in0                    // in0 is page_list
+        br.sptk.few .dest_page
+	;;
+.loop:
+	ld8  r30=[in0], 8;;
+.dest_page:
+	tbit.z p0, p6=r30, 0;;    	// 0x1 dest page
+(p6)	and r17=r30, r16
+(p6)	br.cond.sptk.few .loop;;
+
+	tbit.z p0, p6=r30, 1;;		// 0x2 indirect page
+(p6)	and in0=r30, r16
+(p6)	br.cond.sptk.few .loop;;
+
+	tbit.z p0, p6=r30, 2;;		// 0x4 end flag
+(p6)	br.cond.sptk.few .end_loop;;
+
+	tbit.z p6, p0=r30, 3;;		// 0x8 source page
+(p6)	br.cond.sptk.few .loop
+
+	and r18=r30, r16
+
+	// simple copy page, may optimize later
+	movl r14=PAGE_SIZE/8 - 1;;
+	mov ar.lc=r14;;
+1:
+	ld8 r14=[r18], 8;;
+	st8 [r17]=r14;;
+	fc.i r17
+	add r17=8, r17
+	br.ctop.sptk.few 1b
+	br.sptk.few .loop
+	;;
+
+.end_loop:
+	sync.i			// for fc.i
+	;;
+	srlz.i
+	;;
+	srlz.d
+	;;
+	br.call.sptk.many b0=b6;;
+
+.align  32
+memory_stack:
+	.fill           8192, 1, 0
+memory_stack_end:
+register_stack:
+	.fill           8192, 1, 0
+register_stack_end:
+relocate_new_kernel_end:
+END(relocate_new_kernel)
+
+.global relocate_new_kernel_size
+relocate_new_kernel_size:
+	data8	relocate_new_kernel_end - relocate_new_kernel
+
+GLOBAL_ENTRY(ia64_dump_cpu_regs)
+        .prologue
+        alloc loc0=ar.pfs,1,2,0,0
+        .body
+        mov     ar.rsc=0                // put RSE in enforced lazy mode
+        add     loc1=4*8, in0           // save r4 and r5 first
+        ;;
+{
+        flushrs                         // flush dirty regs to backing store
+        srlz.i
+}
+        st8 [loc1]=r4, 8
+        ;;
+        st8 [loc1]=r5, 8
+        ;;
+        add loc1=32*8, in0
+        mov r4=ar.rnat
+        ;;
+        st8 [in0]=r0, 8			// r0
+        st8 [loc1]=r4, 8		// rnat
+        mov r5=pr
+        ;;
+        st8 [in0]=r1, 8			// r1
+        st8 [loc1]=r5, 8		// pr
+        mov r4=b0
+        ;;
+        st8 [in0]=r2, 8			// r2
+        st8 [loc1]=r4, 8		// b0
+        mov r5=b1;
+        ;;
+        st8 [in0]=r3, 24		// r3
+        st8 [loc1]=r5, 8		// b1
+        mov r4=b2
+        ;;
+        st8 [in0]=r6, 8			// r6
+        st8 [loc1]=r4, 8		// b2
+	mov r5=b3
+        ;;
+        st8 [in0]=r7, 8			// r7
+        st8 [loc1]=r5, 8		// b3
+        mov r4=b4
+        ;;
+        st8 [in0]=r8, 8			// r8
+        st8 [loc1]=r4, 8		// b4
+        mov r5=b5
+        ;;
+        st8 [in0]=r9, 8			// r9
+        st8 [loc1]=r5, 8		// b5
+        mov r4=b6
+        ;;
+        st8 [in0]=r10, 8		// r10
+        st8 [loc1]=r5, 8		// b6
+        mov r5=b7
+        ;;
+        st8 [in0]=r11, 8		// r11
+        st8 [loc1]=r5, 8		// b7
+        mov r4=b0
+        ;;
+        st8 [in0]=r12, 8		// r12
+        st8 [loc1]=r4, 8		// ip
+        mov r5=loc0
+	;;
+        st8 [in0]=r13, 8		// r13
+        extr.u r5=r5, 0, 38		// ar.pfs.pfm
+	mov r4=r0			// user mask
+        ;;
+        st8 [in0]=r14, 8		// r14
+        st8 [loc1]=r5, 8		// cfm
+        ;;
+        st8 [in0]=r15, 8		// r15
+        st8 [loc1]=r4, 8        	// user mask
+	mov r5=ar.rsc
+        ;;
+        st8 [in0]=r16, 8		// r16
+        st8 [loc1]=r5, 8        	// ar.rsc
+        mov r4=ar.bsp
+        ;;
+        st8 [in0]=r17, 8		// r17
+        st8 [loc1]=r4, 8        	// ar.bsp
+        mov r5=ar.bspstore
+        ;;
+        st8 [in0]=r18, 8		// r18
+        st8 [loc1]=r5, 8        	// ar.bspstore
+        mov r4=ar.rnat
+        ;;
+        st8 [in0]=r19, 8		// r19
+        st8 [loc1]=r4, 8        	// ar.rnat
+        mov r5=ar.ccv
+        ;;
+        st8 [in0]=r20, 8		// r20
+	st8 [loc1]=r5, 8        	// ar.ccv
+        mov r4=ar.unat
+        ;;
+        st8 [in0]=r21, 8		// r21
+        st8 [loc1]=r4, 8        	// ar.unat
+        mov r5 = ar.fpsr
+        ;;
+        st8 [in0]=r22, 8		// r22
+        st8 [loc1]=r5, 8        	// ar.fpsr
+        mov r4 = ar.unat
+        ;;
+        st8 [in0]=r23, 8		// r23
+        st8 [loc1]=r4, 8        	// unat
+        mov r5 = ar.fpsr
+        ;;
+        st8 [in0]=r24, 8		// r24
+        st8 [loc1]=r5, 8        	// fpsr
+        mov r4 = ar.pfs
+        ;;
+        st8 [in0]=r25, 8		// r25
+        st8 [loc1]=r4, 8        	// ar.pfs
+        mov r5 = ar.lc
+        ;;
+        st8 [in0]=r26, 8		// r26
+        st8 [loc1]=r5, 8        	// ar.lc
+        mov r4 = ar.ec
+        ;;
+        st8 [in0]=r27, 8		// r27
+        st8 [loc1]=r4, 8        	// ar.ec
+        mov r5 = ar.csd
+        ;;
+        st8 [in0]=r28, 8		// r28
+        st8 [loc1]=r5, 8        	// ar.csd
+        mov r4 = ar.ssd
+        ;;
+        st8 [in0]=r29, 8		// r29
+        st8 [loc1]=r4, 8        	// ar.ssd
+        ;;
+        st8 [in0]=r30, 8		// r30
+        ;;
+	st8 [in0]=r31, 8		// r31
+        mov ar.pfs=loc0
+        ;;
+        br.ret.sptk.many rp
+END(ia64_dump_cpu_regs)
+
+
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/salinfo.c linux-2.6.18-xen-3.3.0/arch/ia64/kernel/salinfo.c
--- linux-2.6.18.8/arch/ia64/kernel/salinfo.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/salinfo.c	2008-08-21 11:36:07.000000000 +0200
@@ -375,6 +375,25 @@ salinfo_log_open(struct inode *inode, st
 		data->open = 0;
 		return -ENOMEM;
 	}
+#ifdef CONFIG_XEN
+	if (is_running_on_xen()) {
+		ia64_mca_xencomm_t *entry;
+		unsigned long flags;
+
+		entry = vmalloc(sizeof(ia64_mca_xencomm_t));
+		if (!entry) {
+			data->open = 0;
+			vfree(data->log_buffer);
+			return -ENOMEM;
+		}
+		entry->record = data->log_buffer;
+		entry->handle = xencomm_map(data->log_buffer, 
+					ia64_sal_get_state_info_size(data->type));
+		spin_lock_irqsave(&ia64_mca_xencomm_lock, flags);
+		list_add(&entry->list, &ia64_mca_xencomm_list);
+		spin_unlock_irqrestore(&ia64_mca_xencomm_lock, flags);
+	}
+#endif
 
 	return 0;
 }
@@ -386,6 +405,30 @@ salinfo_log_release(struct inode *inode,
 	struct salinfo_data *data = entry->data;
 
 	if (data->state == STATE_NO_DATA) {
+#ifdef CONFIG_XEN
+		if (is_running_on_xen()) {
+			struct list_head *pos, *n;
+			ia64_mca_xencomm_t *found_entry = NULL;
+			unsigned long flags;
+
+			spin_lock_irqsave(&ia64_mca_xencomm_lock, flags);
+			list_for_each_safe(pos, n, &ia64_mca_xencomm_list) {
+				ia64_mca_xencomm_t *entry;
+
+				entry = list_entry(pos, ia64_mca_xencomm_t, list);
+				if (entry->record == data->log_buffer) {
+					list_del(&entry->list);
+					found_entry = entry;
+					break;
+				}
+			}
+			spin_unlock_irqrestore(&ia64_mca_xencomm_lock, flags);
+			if (found_entry) {
+				xencomm_free(found_entry->handle);
+				vfree(found_entry);
+			}
+		}
+#endif
 		vfree(data->log_buffer);
 		vfree(data->oemdata);
 		data->log_buffer = NULL;
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/setup.c linux-2.6.18-xen-3.3.0/arch/ia64/kernel/setup.c
--- linux-2.6.18.8/arch/ia64/kernel/setup.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/setup.c	2008-08-21 11:36:07.000000000 +0200
@@ -43,6 +43,8 @@
 #include <linux/initrd.h>
 #include <linux/pm.h>
 #include <linux/cpufreq.h>
+#include <linux/kexec.h>
+#include <linux/crash_dump.h>
 
 #include <asm/ia32.h>
 #include <asm/machvec.h>
@@ -60,6 +62,12 @@
 #include <asm/system.h>
 #include <asm/unistd.h>
 #include <asm/system.h>
+#ifdef CONFIG_XEN
+#include <asm/hypervisor.h>
+#include <asm/xen/xencomm.h>
+#include <xen/xencons.h>
+#endif
+#include <linux/dma-mapping.h>
 
 #if defined(CONFIG_SMP) && (IA64_CPU_SIZE > PAGE_SIZE)
 # error "struct cpuinfo_ia64 too big!"
@@ -70,6 +78,34 @@ unsigned long __per_cpu_offset[NR_CPUS];
 EXPORT_SYMBOL(__per_cpu_offset);
 #endif
 
+#ifdef CONFIG_XEN
+static void
+xen_panic_hypercall(struct unw_frame_info *info, void *arg)
+{
+	current->thread.ksp = (__u64)info->sw - 16;
+	HYPERVISOR_shutdown(SHUTDOWN_crash);
+	/* we're never actually going to get here... */
+}
+
+static int
+xen_panic_event(struct notifier_block *this, unsigned long event, void *ptr)
+{
+	unw_init_running(xen_panic_hypercall, NULL);
+	/* we're never actually going to get here... */
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block xen_panic_block = {
+	xen_panic_event, NULL, 0 /* try to go last */
+};
+
+void xen_pm_power_off(void)
+{
+	local_irq_disable();
+	HYPERVISOR_shutdown(SHUTDOWN_poweroff);
+}
+#endif
+
 extern void ia64_setup_printk_clock(void);
 
 DEFINE_PER_CPU(struct cpuinfo_ia64, cpu_info);
@@ -242,6 +278,14 @@ reserve_memory (void)
 	rsvd_region[n].end   = (unsigned long) ia64_imva(_end);
 	n++;
 
+#ifdef CONFIG_XEN
+	if (is_running_on_xen()) {
+		rsvd_region[n].start = (unsigned long)__va((HYPERVISOR_shared_info->arch.start_info_pfn << PAGE_SHIFT));
+		rsvd_region[n].end   = rsvd_region[n].start + PAGE_SIZE;
+		n++;
+ 	}
+#endif
+
 #ifdef CONFIG_BLK_DEV_INITRD
 	if (ia64_boot_param->initrd_start) {
 		rsvd_region[n].start = (unsigned long)__va(ia64_boot_param->initrd_start);
@@ -253,6 +297,56 @@ reserve_memory (void)
 	efi_memmap_init(&rsvd_region[n].start, &rsvd_region[n].end);
 	n++;
 
+#ifdef CONFIG_KEXEC
+	/* crashkernel=size@offset specifies the size to reserve for a crash
+	 * kernel. If offset is 0, then it is determined automatically.
+	 * By reserving this memory we guarantee that linux never set's it
+	 * up as a DMA target.Useful for holding code to do something
+	 * appropriate after a kernel panic.
+	 */
+	{
+		char *from = strstr(saved_command_line, "crashkernel=");
+		unsigned long base, size;
+#ifdef CONFIG_XEN
+		if (is_initial_xendomain() && from)
+				printk("Ignoring crashkernel command line, "
+				       "parameter will be supplied by xen\n");
+		else {
+#endif
+		if (from) {
+			size = memparse(from + 12, &from);
+			if (*from == '@')
+				base = memparse(from+1, &from);
+			else
+				base = 0;
+			if (size) {
+				if (!base) {
+					sort_regions(rsvd_region, n);
+					base = kdump_find_rsvd_region(size,
+							      	rsvd_region, n);
+					}
+				if (base != ~0UL) {
+					rsvd_region[n].start =
+						(unsigned long)__va(base);
+					rsvd_region[n].end =
+						(unsigned long)__va(base + size);
+					n++;
+					crashk_res.start = base;
+					crashk_res.end = base + size - 1;
+				}
+			}
+		}
+		efi_memmap_res.start = ia64_boot_param->efi_memmap;
+                efi_memmap_res.end = efi_memmap_res.start +
+                        ia64_boot_param->efi_memmap_size;
+                boot_param_res.start = kexec_virt_to_phys(ia64_boot_param);
+                boot_param_res.end = boot_param_res.start +
+                        sizeof(*ia64_boot_param);
+#ifdef CONFIG_XEN
+		}
+#endif
+	}
+#endif
 	/* end of memory marker */
 	rsvd_region[n].start = ~0UL;
 	rsvd_region[n].end   = ~0UL;
@@ -264,6 +358,7 @@ reserve_memory (void)
 	sort_regions(rsvd_region, num_rsvd_regions);
 }
 
+
 /**
  * find_initrd - get initrd parameters from the boot parameter structure
  *
@@ -397,11 +492,49 @@ static __init int setup_nomca(char *s)
 }
 early_param("nomca", setup_nomca);
 
+#ifdef CONFIG_PROC_VMCORE
+/* elfcorehdr= specifies the location of elf core header
+ * stored by the crashed kernel.
+ */
+static int __init parse_elfcorehdr(char *arg)
+{
+	if (!arg)
+		return -EINVAL;
+
+        elfcorehdr_addr = memparse(arg, &arg);
+	return 0;
+}
+early_param("elfcorehdr", parse_elfcorehdr);
+#endif /* CONFIG_PROC_VMCORE */
+
 void __init
 setup_arch (char **cmdline_p)
 {
+#ifdef CONFIG_XEN
+	shared_info_t *s = NULL;
+	if (is_running_on_xen()) {
+		s = HYPERVISOR_shared_info;
+		xen_start_info = __va(s->arch.start_info_pfn << PAGE_SHIFT);
+	}
+#endif
+
 	unw_init();
 
+#ifdef CONFIG_XEN
+	if (is_running_on_xen()) {
+		/* Must be done before any hypercall.  */
+		xencomm_initialize();
+
+		setup_xen_features();
+		/* Register a call for panic conditions. */
+		atomic_notifier_chain_register(&panic_notifier_list,
+		                               &xen_panic_block);
+		pm_power_off = xen_pm_power_off;
+
+		xen_ia64_enable_opt_feature();
+	}
+#endif
+
 	ia64_patch_vtop((u64) __start___vtop_patchlist, (u64) __end___vtop_patchlist);
 
 	*cmdline_p = __va(ia64_boot_param->command_line);
@@ -462,6 +595,57 @@ setup_arch (char **cmdline_p)
 	acpi_boot_init();
 #endif
 
+#ifdef CONFIG_XEN
+	if (is_running_on_xen()) {
+		printk("Running on Xen! start_info_pfn=0x%lx nr_pages=%ld "
+		       "flags=0x%x\n", s->arch.start_info_pfn,
+		       xen_start_info->nr_pages, xen_start_info->flags);
+
+		/*
+		 * If a console= is NOT specified, we assume using the
+		 * xencons console is desired.  By default, this is xvc0
+		 * for both dom0 and domU.
+		 */
+		if (!strstr(*cmdline_p, "console=")) {
+			char *p, *q, name[5] = "xvc";
+			int offset = 0;
+
+#if defined(CONFIG_VGA_CONSOLE)
+			/*
+			 * conswitchp might be set intelligently from the
+			 * PCDP code.  If set to VGA console, use it.
+			 */
+			if (is_initial_xendomain() && conswitchp == &vga_con)
+				strncpy(name, "tty", 3);
+#endif
+
+			p = strstr(*cmdline_p, "xencons=");
+
+			if (p) {
+				p += 8;
+				if (!strncmp(p, "ttyS", 4)) {
+					strncpy(name, p, 4);
+					p += 4;
+					offset = simple_strtol(p, &q, 10);
+					if (p == q)
+						offset = 0;
+				} else if (!strncmp(p, "tty", 3) ||
+				           !strncmp(p, "xvc", 3)) {
+					strncpy(name, p, 3);
+					p += 3;
+					offset = simple_strtol(p, &q, 10);
+					if (p == q)
+						offset = 0;
+				} else if (!strncmp(p, "off", 3))
+					offset = -1;
+			}
+
+			if (offset >= 0)
+				add_preferred_console(name, offset, NULL);
+		}
+	}
+#endif
+
 #ifdef CONFIG_VT
 	if (!conswitchp) {
 # if defined(CONFIG_DUMMY_CONSOLE)
@@ -481,11 +665,28 @@ setup_arch (char **cmdline_p)
 #endif
 
 	/* enable IA-64 Machine Check Abort Handling unless disabled */
+#ifdef CONFIG_XEN
+	if (is_running_on_xen() && !is_initial_xendomain()) {
+		nomca = 1;
+#if !defined(CONFIG_VT) || !defined(CONFIG_DUMMY_CONSOLE)
+		conswitchp = NULL;
+#endif
+	}
+#endif
 	if (!nomca)
 		ia64_mca_init();
 
 	platform_setup(cmdline_p);
+#ifdef CONFIG_XEN
+	if (is_running_on_xen() && !ia64_platform_is("xen")) {
+		extern ia64_mv_setup_t xen_setup;
+		xen_setup(cmdline_p);
+	}
+#endif
 	paging_init();
+#ifdef CONFIG_XEN
+	xen_contiguous_bitmap_init(max_pfn);
+#endif
 }
 
 /*
@@ -870,6 +1071,13 @@ cpu_init (void)
 	/* size of physical stacked register partition plus 8 bytes: */
 	__get_cpu_var(ia64_phys_stacked_size_p8) = num_phys_stacked*8 + 8;
 	platform_cpu_init();
+#ifdef CONFIG_XEN
+	if (is_running_on_xen() && !ia64_platform_is("xen")) {
+		extern ia64_mv_cpu_init_t xen_cpu_init;
+		xen_cpu_init();
+	}
+#endif
+
 	pm_idle = default_idle;
 }
 
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/smp.c linux-2.6.18-xen-3.3.0/arch/ia64/kernel/smp.c
--- linux-2.6.18.8/arch/ia64/kernel/smp.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/smp.c	2008-08-21 11:36:07.000000000 +0200
@@ -30,6 +30,7 @@
 #include <linux/delay.h>
 #include <linux/efi.h>
 #include <linux/bitops.h>
+#include <linux/kexec.h>
 
 #include <asm/atomic.h>
 #include <asm/current.h>
@@ -66,6 +67,7 @@ static volatile struct call_data_struct 
 
 #define IPI_CALL_FUNC		0
 #define IPI_CPU_STOP		1
+#define IPI_KDUMP_CPU_STOP	3
 
 /* This needs to be cacheline aligned because it is written to by *other* CPUs.  */
 static DEFINE_PER_CPU(u64, ipi_operation) ____cacheline_aligned;
@@ -155,7 +157,11 @@ handle_IPI (int irq, void *dev_id, struc
 			      case IPI_CPU_STOP:
 				stop_this_cpu();
 				break;
-
+#ifdef CONFIG_KEXEC
+			      case IPI_KDUMP_CPU_STOP:
+				unw_init_running(kdump_cpu_freeze, NULL);
+				break;
+#endif
 			      default:
 				printk(KERN_CRIT "Unknown IPI on CPU %d: %lu\n", this_cpu, which);
 				break;
@@ -213,6 +219,26 @@ send_IPI_self (int op)
 	send_IPI_single(smp_processor_id(), op);
 }
 
+#ifdef CONFIG_KEXEC
+void
+kdump_smp_send_stop()
+{
+ 	send_IPI_allbutself(IPI_KDUMP_CPU_STOP);
+}
+
+void
+kdump_smp_send_init()
+{
+	unsigned int cpu, self_cpu;
+	self_cpu = smp_processor_id();
+	for_each_online_cpu(cpu) {
+		if (cpu != self_cpu) {
+			if(kdump_status[cpu] == 0)
+				platform_send_ipi(cpu, 0, IA64_IPI_DM_INIT, 0);
+		}
+	}
+}
+#endif
 /*
  * Called with preeemption disabled.
  */
@@ -328,10 +354,14 @@ int
 smp_call_function (void (*func) (void *info), void *info, int nonatomic, int wait)
 {
 	struct call_data_struct data;
-	int cpus = num_online_cpus()-1;
+	int cpus;
 
-	if (!cpus)
+	spin_lock(&call_lock);
+	cpus = num_online_cpus()-1;
+	if (!cpus) {
+		spin_unlock(&call_lock);
 		return 0;
+	}
 
 	/* Can deadlock when called with interrupts disabled */
 	WARN_ON(irqs_disabled());
@@ -343,8 +373,6 @@ smp_call_function (void (*func) (void *i
 	if (wait)
 		atomic_set(&data.finished, 0);
 
-	spin_lock(&call_lock);
-
 	call_data = &data;
 	mb();	/* ensure store to call_data precedes setting of IPI_CALL_FUNC */
 	send_IPI_allbutself(IPI_CALL_FUNC);
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/time.c linux-2.6.18-xen-3.3.0/arch/ia64/kernel/time.c
--- linux-2.6.18.8/arch/ia64/kernel/time.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/time.c	2008-08-21 11:36:07.000000000 +0200
@@ -29,6 +29,14 @@
 #include <asm/sections.h>
 #include <asm/system.h>
 
+#include <asm/hypervisor.h>
+#ifdef CONFIG_XEN
+#include <linux/kernel_stat.h>
+#include <linux/posix-timers.h>
+#include <xen/interface/vcpu.h>
+#include <asm/percpu.h>
+#endif
+
 extern unsigned long wall_jiffies;
 
 volatile int time_keeper_id = 0; /* smp_processor_id() of time-keeper */
@@ -40,16 +48,109 @@ EXPORT_SYMBOL(last_cli_ip);
 
 #endif
 
+#ifdef CONFIG_XEN
+DEFINE_PER_CPU(struct vcpu_runstate_info, runstate);
+DEFINE_PER_CPU(unsigned long, processed_stolen_time);
+DEFINE_PER_CPU(unsigned long, processed_blocked_time);
+#define NS_PER_TICK (1000000000LL/HZ)
+#endif
+
 static struct time_interpolator itc_interpolator = {
 	.shift = 16,
 	.mask = 0xffffffffffffffffLL,
 	.source = TIME_SOURCE_CPU
 };
 
+#ifdef CONFIG_XEN
+static unsigned long 
+consider_steal_time(unsigned long new_itm, struct pt_regs *regs)
+{
+	unsigned long stolen, blocked, sched_time;
+	unsigned long delta_itm = 0, stolentick = 0;
+	int i, cpu = smp_processor_id();
+	struct vcpu_runstate_info *runstate;
+	struct task_struct *p = current;
+
+	runstate = &per_cpu(runstate, smp_processor_id());
+
+	do {
+		sched_time = runstate->state_entry_time;
+		mb();
+		stolen = runstate->time[RUNSTATE_runnable] + 
+			 runstate->time[RUNSTATE_offline] -
+			 per_cpu(processed_stolen_time, cpu);
+		blocked = runstate->time[RUNSTATE_blocked] -
+			  per_cpu(processed_blocked_time, cpu);
+		mb();
+	} while (sched_time != runstate->state_entry_time);
+
+	/*
+	 * Check for vcpu migration effect
+	 * In this case, itc value is reversed.
+	 * This causes huge stolen value.  
+	 * This function just checks and reject this effect.
+	 */
+	if (!time_after_eq(runstate->time[RUNSTATE_blocked],
+			   per_cpu(processed_blocked_time, cpu)))
+		blocked = 0;
+
+	if (!time_after_eq(runstate->time[RUNSTATE_runnable] +
+			   runstate->time[RUNSTATE_offline],
+			   per_cpu(processed_stolen_time, cpu)))
+		stolen = 0;
+
+	if (!time_after(delta_itm + new_itm, ia64_get_itc()))
+		stolentick = ia64_get_itc() - delta_itm - new_itm;
+
+	do_div(stolentick, NS_PER_TICK);
+	stolentick++;
+
+	do_div(stolen, NS_PER_TICK);
+
+	if (stolen > stolentick)
+		stolen = stolentick;
+
+	stolentick -= stolen;
+	do_div(blocked, NS_PER_TICK);
+
+	if (blocked > stolentick)
+		blocked = stolentick;
+
+	if (stolen > 0 || blocked > 0) {
+		account_steal_time(NULL, jiffies_to_cputime(stolen)); 
+		account_steal_time(idle_task(cpu), jiffies_to_cputime(blocked)); 
+		run_local_timers();
+
+		if (rcu_pending(cpu))
+			rcu_check_callbacks(cpu, user_mode(regs));
+
+		scheduler_tick();
+		run_posix_cpu_timers(p);
+		delta_itm += local_cpu_data->itm_delta * (stolen + blocked);
+
+		if (cpu == time_keeper_id) {
+			write_seqlock(&xtime_lock);
+			for(i = 0; i < stolen + blocked; i++)
+				do_timer(regs);
+			local_cpu_data->itm_next = delta_itm + new_itm;
+			write_sequnlock(&xtime_lock);
+		} else {
+			local_cpu_data->itm_next = delta_itm + new_itm;
+		}
+		per_cpu(processed_stolen_time,cpu) += NS_PER_TICK * stolen;
+		per_cpu(processed_blocked_time,cpu) += NS_PER_TICK * blocked;
+	}
+	return delta_itm; 
+}
+#else
+#define consider_steal_time(new_itm, regs) (0)
+#endif
+
 static irqreturn_t
 timer_interrupt (int irq, void *dev_id, struct pt_regs *regs)
 {
 	unsigned long new_itm;
+	unsigned long delta_itm; /* XEN */
 
 	if (unlikely(cpu_is_offline(smp_processor_id()))) {
 		return IRQ_HANDLED;
@@ -65,6 +166,13 @@ timer_interrupt (int irq, void *dev_id, 
 
 	profile_tick(CPU_PROFILING, regs);
 
+	if (is_running_on_xen()) {
+		delta_itm = consider_steal_time(new_itm, regs);
+		new_itm += delta_itm;
+		if (time_after(new_itm, ia64_get_itc()) && delta_itm)
+			goto skip_process_time_accounting;
+	}
+
 	while (1) {
 		update_process_times(user_mode(regs));
 
@@ -88,6 +196,8 @@ timer_interrupt (int irq, void *dev_id, 
 			break;
 	}
 
+skip_process_time_accounting:	/* XEN */
+
 	do {
 		/*
 		 * If we're too close to the next clock tick for
@@ -142,6 +252,85 @@ static int __init nojitter_setup(char *s
 
 __setup("nojitter", nojitter_setup);
 
+#ifdef CONFIG_XEN
+/* taken from i386/kernel/time-xen.c */
+static void init_missing_ticks_accounting(int cpu)
+{
+	struct vcpu_register_runstate_memory_area area;
+	struct vcpu_runstate_info *runstate = &per_cpu(runstate, cpu);
+	int rc;
+
+	memset(runstate, 0, sizeof(*runstate));
+
+	area.addr.v = runstate;
+	rc = HYPERVISOR_vcpu_op(VCPUOP_register_runstate_memory_area, cpu, &area);
+	WARN_ON(rc && rc != -ENOSYS);
+
+	per_cpu(processed_blocked_time, cpu) = runstate->time[RUNSTATE_blocked];
+	per_cpu(processed_stolen_time, cpu) = runstate->time[RUNSTATE_runnable]
+					    + runstate->time[RUNSTATE_offline];
+}
+
+static int xen_ia64_settimefoday_after_resume;
+
+static int __init __xen_ia64_settimeofday_after_resume(char *str)
+{
+	xen_ia64_settimefoday_after_resume = 1;
+	return 1;
+}
+
+__setup("xen_ia64_settimefoday_after_resume",
+	 __xen_ia64_settimeofday_after_resume);
+
+/* Called after suspend, to resume time.  */
+void
+time_resume(void)
+{
+	unsigned int cpu;
+	
+	/* Just trigger a tick.  */
+	ia64_cpu_local_tick();
+
+	if (xen_ia64_settimefoday_after_resume) {
+		/* do_settimeofday() resets timer interplator */
+		struct timespec xen_time;
+		int ret;
+		efi_gettimeofday(&xen_time);
+
+		ret = do_settimeofday(&xen_time);
+		WARN_ON(ret);
+	} else {
+#if 0
+		/* adjust EFI time */
+		struct timespec my_time = CURRENT_TIME;
+		struct timespec xen_time;
+		static timespec diff;
+		struct xen_domctl domctl;
+		int ret;
+
+		efi_gettimeofday(&xen_time);
+		diff = timespec_sub(&xen_time, &my_time);
+		domctl.cmd = XEN_DOMCTL_settimeoffset;
+		domctl.domain = DOMID_SELF;
+		domctl.u.settimeoffset.timeoffset_seconds = diff.tv_sec;
+		ret = HYPERVISOR_domctl_op(&domctl);
+		WARN_ON(ret);
+#endif
+		/* Time interpolator remembers the last timer status.
+		   Forget it */
+		write_seqlock_irq(&xtime_lock);
+		time_interpolator_reset();
+		write_sequnlock_irq(&xtime_lock);
+	}
+
+	for_each_online_cpu(cpu)
+		init_missing_ticks_accounting(cpu);
+
+	touch_softlockup_watchdog();
+}
+#else
+#define init_missing_ticks_accounting(cpu) do {} while (0)
+#endif
 
 void __devinit
 ia64_init_itm (void)
@@ -225,6 +414,12 @@ ia64_init_itm (void)
 		register_time_interpolator(&itc_interpolator);
 	}
 
+	if (is_running_on_xen())
+		init_missing_ticks_accounting(smp_processor_id());
+
+	/* avoid softlock up message when cpu is unplug and plugged again. */
+	touch_softlockup_watchdog();
+
 	/* Setup the CPU local timer tick */
 	ia64_cpu_local_tick();
 }
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/vmlinux.lds.S linux-2.6.18-xen-3.3.0/arch/ia64/kernel/vmlinux.lds.S
--- linux-2.6.18.8/arch/ia64/kernel/vmlinux.lds.S	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/vmlinux.lds.S	2008-08-21 11:36:07.000000000 +0200
@@ -183,6 +183,12 @@ SECTIONS
 	  __start_gate_section = .;
 	  *(.data.gate)
 	  __stop_gate_section = .;
+#if defined(CONFIG_XEN)
+	  . = ALIGN(PAGE_SIZE);
+	  __start_xen_gate_section = .;
+	  *(.data.gate.xen)
+	  __stop_xen_gate_section = .;
+#endif
 	}
   . = ALIGN(PAGE_SIZE);		/* make sure the gate page doesn't expose kernel data */
 
diff -rpuN linux-2.6.18.8/arch/ia64/kernel/xengate-data.S linux-2.6.18-xen-3.3.0/arch/ia64/kernel/xengate-data.S
--- linux-2.6.18.8/arch/ia64/kernel/xengate-data.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/kernel/xengate-data.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,3 @@
+	.section .data.gate.xen, "aw"
+
+	.incbin "arch/ia64/kernel/xengate.so"
diff -rpuN linux-2.6.18.8/arch/ia64/Makefile linux-2.6.18-xen-3.3.0/arch/ia64/Makefile
--- linux-2.6.18.8/arch/ia64/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -45,6 +45,12 @@ ifeq ($(call cc-version),0304)
 endif
 
 CFLAGS += $(cflags-y)
+
+cppflags-$(CONFIG_XEN) += \
+	-D__XEN_INTERFACE_VERSION__=$(CONFIG_XEN_INTERFACE_VERSION)
+
+CPPFLAGS += $(cppflags-y)
+
 head-y := arch/ia64/kernel/head.o arch/ia64/kernel/init_task.o
 
 libs-y				+= arch/ia64/lib/
@@ -54,7 +60,9 @@ core-$(CONFIG_IA64_DIG) 	+= arch/ia64/di
 core-$(CONFIG_IA64_GENERIC) 	+= arch/ia64/dig/
 core-$(CONFIG_IA64_HP_ZX1)	+= arch/ia64/dig/
 core-$(CONFIG_IA64_HP_ZX1_SWIOTLB) += arch/ia64/dig/
+core-$(CONFIG_IA64_XEN)		+= arch/ia64/dig/
 core-$(CONFIG_IA64_SGI_SN2)	+= arch/ia64/sn/
+core-$(CONFIG_XEN)		+= arch/ia64/xen/
 
 drivers-$(CONFIG_PCI)		+= arch/ia64/pci/
 drivers-$(CONFIG_IA64_HP_SIM)	+= arch/ia64/hp/sim/
@@ -87,8 +95,8 @@ CLEAN_FILES += vmlinux.gz bootloader
 boot:	lib/lib.a vmlinux
 	$(Q)$(MAKE) $(build)=$(boot) $@
 
-install: vmlinux.gz
-	sh $(srctree)/arch/ia64/install.sh $(KERNELRELEASE) $< System.map "$(INSTALL_PATH)"
+install:
+	-yes | sh $(srctree)/arch/ia64/install.sh $(KERNELRELEASE) vmlinux.gz System.map "$(INSTALL_PATH)"
 
 define archhelp
   echo '* compressed	- Build compressed kernel image'
diff -rpuN linux-2.6.18.8/arch/ia64/mm/contig.c linux-2.6.18-xen-3.3.0/arch/ia64/mm/contig.c
--- linux-2.6.18.8/arch/ia64/mm/contig.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/mm/contig.c	2008-08-21 11:36:07.000000000 +0200
@@ -18,6 +18,9 @@
 #include <linux/efi.h>
 #include <linux/mm.h>
 #include <linux/swap.h>
+#if defined(CONFIG_XEN) && defined(CONFIG_KEXEC)
+#include <linux/kexec.h>
+#endif
 
 #include <asm/meminit.h>
 #include <asm/pgalloc.h>
@@ -172,8 +175,12 @@ find_memory (void)
 	/* Free all available memory, then mark bootmem-map as being in use. */
 	efi_memmap_walk(filter_rsvd_memory, free_bootmem);
 	reserve_bootmem(bootmap_start, bootmap_size);
+#if defined(CONFIG_XEN) && defined(CONFIG_KEXEC)
+	xen_machine_kexec_setup_resources();
+#endif
 
 	find_initrd();
+
 }
 
 #ifdef CONFIG_SMP
diff -rpuN linux-2.6.18.8/arch/ia64/mm/discontig.c linux-2.6.18-xen-3.3.0/arch/ia64/mm/discontig.c
--- linux-2.6.18.8/arch/ia64/mm/discontig.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/mm/discontig.c	2008-08-21 11:36:07.000000000 +0200
@@ -21,6 +21,9 @@
 #include <linux/acpi.h>
 #include <linux/efi.h>
 #include <linux/nodemask.h>
+#if defined(CONFIG_XEN) && defined(CONFIG_KEXEC)
+#include <linux/kexec.h>
+#endif
 #include <asm/pgalloc.h>
 #include <asm/tlb.h>
 #include <asm/meminit.h>
@@ -502,6 +505,9 @@ void __init find_memory(void)
 	reserve_pernode_space();
 	memory_less_nodes();
 	initialize_pernode_data();
+#if defined(CONFIG_XEN) && defined(CONFIG_KEXEC)
+	xen_machine_kexec_setup_resources();
+#endif
 
 	max_pfn = max_low_pfn;
 
diff -rpuN linux-2.6.18.8/arch/ia64/mm/init.c linux-2.6.18-xen-3.3.0/arch/ia64/mm/init.c
--- linux-2.6.18.8/arch/ia64/mm/init.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/mm/init.c	2008-08-21 11:36:07.000000000 +0200
@@ -303,16 +303,34 @@ static void __init
 setup_gate (void)
 {
 	struct page *page;
+	void *gate_page_addr = __start_gate_section;
+	
+#ifdef CONFIG_XEN
+	unsigned long unused_gate;
+	extern char __start_xen_gate_section[];
+	if (is_running_on_xen()) {
+		gate_page_addr = __start_xen_gate_section;
+		unused_gate = (unsigned long)ia64_imva(__start_gate_section);
+	} else
+		unused_gate =
+			(unsigned long)ia64_imva(__start_xen_gate_section);
+#ifndef HAVE_BUGGY_SEGREL
+	ClearPageReserved(virt_to_page(unused_gate));
+	init_page_count(virt_to_page(unused_gate));
+	free_page(unused_gate);
+	++totalram_pages;
+#endif
+#endif
 
 	/*
 	 * Map the gate page twice: once read-only to export the ELF
 	 * headers etc. and once execute-only page to enable
 	 * privilege-promotion via "epc":
 	 */
-	page = virt_to_page(ia64_imva(__start_gate_section));
+	page = virt_to_page(ia64_imva(gate_page_addr));
 	put_kernel_page(page, GATE_ADDR, PAGE_READONLY);
 #ifdef HAVE_BUGGY_SEGREL
-	page = virt_to_page(ia64_imva(__start_gate_section + PAGE_SIZE));
+	page = virt_to_page(ia64_imva(gate_page_addr + PAGE_SIZE));
 	put_kernel_page(page, GATE_ADDR + PAGE_SIZE, PAGE_GATE);
 #else
 	put_kernel_page(page, GATE_ADDR + PERCPU_PAGE_SIZE, PAGE_GATE);
diff -rpuN linux-2.6.18.8/arch/ia64/mm/ioremap.c linux-2.6.18-xen-3.3.0/arch/ia64/mm/ioremap.c
--- linux-2.6.18.8/arch/ia64/mm/ioremap.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/mm/ioremap.c	2008-08-21 11:36:07.000000000 +0200
@@ -16,6 +16,9 @@
 static inline void __iomem *
 __ioremap (unsigned long offset, unsigned long size)
 {
+	offset = HYPERVISOR_ioremap(offset, size);
+	if (IS_ERR_VALUE(offset))
+		return NULL;
 	return (void __iomem *) (__IA64_UNCACHED_OFFSET | offset);
 }
 
diff -rpuN linux-2.6.18.8/arch/ia64/oprofile/init.c linux-2.6.18-xen-3.3.0/arch/ia64/oprofile/init.c
--- linux-2.6.18.8/arch/ia64/oprofile/init.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/oprofile/init.c	2008-08-21 11:36:07.000000000 +0200
@@ -11,6 +11,8 @@
 #include <linux/oprofile.h>
 #include <linux/init.h>
 #include <linux/errno.h>
+#include <asm/hypervisor.h>
+#include "oprofile_perfmon.h"
  
 extern int perfmon_init(struct oprofile_operations * ops);
 extern void perfmon_exit(void);
@@ -20,6 +22,13 @@ int __init oprofile_arch_init(struct opr
 {
 	int ret = -ENODEV;
 
+	if (is_running_on_xen()) {
+		ret = xen_perfmon_init();
+		if (ret)
+			return ret;
+		return xenoprofile_init(ops);
+	}
+
 #ifdef CONFIG_PERFMON
 	/* perfmon_init() can fail, but we have no way to report it */
 	ret = perfmon_init(ops);
@@ -32,6 +41,12 @@ int __init oprofile_arch_init(struct opr
 
 void oprofile_arch_exit(void)
 {
+	if (is_running_on_xen()) {
+		xenoprofile_exit();
+		xen_perfmon_exit();
+		return;
+	}
+
 #ifdef CONFIG_PERFMON
 	perfmon_exit();
 #endif
diff -rpuN linux-2.6.18.8/arch/ia64/oprofile/Makefile linux-2.6.18-xen-3.3.0/arch/ia64/oprofile/Makefile
--- linux-2.6.18.8/arch/ia64/oprofile/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/oprofile/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -8,3 +8,7 @@ DRIVER_OBJS := $(addprefix ../../../driv
 
 oprofile-y := $(DRIVER_OBJS) init.o backtrace.o
 oprofile-$(CONFIG_PERFMON) += perfmon.o
+ifeq ($(CONFIG_XEN), y)
+oprofile-$(CONFIG_PERFMON) += xenoprof.o \
+	../../../drivers/xen/xenoprof/xenoprofile.o
+endif
diff -rpuN linux-2.6.18.8/arch/ia64/oprofile/oprofile_perfmon.h linux-2.6.18-xen-3.3.0/arch/ia64/oprofile/oprofile_perfmon.h
--- linux-2.6.18.8/arch/ia64/oprofile/oprofile_perfmon.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/oprofile/oprofile_perfmon.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,30 @@
+#ifndef OPROFILE_PERFMON_H
+#define OPROFILE_PERFMON_H
+
+#ifdef CONFIG_PERFMON
+#ifdef CONFIG_XEN
+int __perfmon_init(void);
+void __perfmon_exit(void);
+int perfmon_start(void);
+void perfmon_stop(void);
+#endif
+#else
+#define __perfmon_init()	(-ENOSYS)
+#define __perfmon_exit()	do {} while (0)
+#endif /* CONFIG_PERFMON */
+
+#ifdef CONFIG_XEN
+#define STATIC_IF_NO_XEN	/* nothing */
+#define xen_perfmon_init()	__perfmon_init()
+#define xen_perfmon_exit()	__perfmon_exit()
+extern int xenoprofile_init(struct oprofile_operations * ops);
+extern void xenoprofile_exit(void);
+#else
+#define STATIC_IF_NO_XEN	static
+#define xen_perfmon_init()	(-ENOSYS)
+#define xen_perfmon_exit()	do {} while (0)
+#define xenoprofile_init(ops)	(-ENOSYS)
+#define xenoprofile_exit()	do {} while (0)
+#endif /* CONFIG_XEN */
+
+#endif /* OPROFILE_PERFMON_H */
diff -rpuN linux-2.6.18.8/arch/ia64/oprofile/perfmon.c linux-2.6.18-xen-3.3.0/arch/ia64/oprofile/perfmon.c
--- linux-2.6.18.8/arch/ia64/oprofile/perfmon.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/oprofile/perfmon.c	2008-08-21 11:36:07.000000000 +0200
@@ -13,6 +13,7 @@
 #include <asm/perfmon.h>
 #include <asm/ptrace.h>
 #include <asm/errno.h>
+#include "oprofile_perfmon.h"
 
 static int allow_ints;
 
@@ -33,14 +34,16 @@ perfmon_handler(struct task_struct *task
 }
 
 
-static int perfmon_start(void)
+STATIC_IF_NO_XEN
+int perfmon_start(void)
 {
 	allow_ints = 1;
 	return 0;
 }
 
 
-static void perfmon_stop(void)
+STATIC_IF_NO_XEN
+void perfmon_stop(void)
 {
 	allow_ints = 0;
 }
@@ -75,16 +78,35 @@ static char * get_cpu_type(void)
 
 static int using_perfmon;
 
-int perfmon_init(struct oprofile_operations * ops)
+STATIC_IF_NO_XEN
+int __perfmon_init(void)
 {
 	int ret = pfm_register_buffer_fmt(&oprofile_fmt);
 	if (ret)
 		return -ENODEV;
 
+	using_perfmon = 1;
+	return 0;
+}
+
+STATIC_IF_NO_XEN
+void __perfmon_exit(void)
+{
+	if (!using_perfmon)
+		return;
+
+	pfm_unregister_buffer_fmt(oprofile_fmt.fmt_uuid);
+}
+
+int perfmon_init(struct oprofile_operations * ops)
+{
+	int ret = __perfmon_init();
+	if (ret)
+		return -ENODEV;
+
 	ops->cpu_type = get_cpu_type();
 	ops->start = perfmon_start;
 	ops->stop = perfmon_stop;
-	using_perfmon = 1;
 	printk(KERN_INFO "oprofile: using perfmon.\n");
 	return 0;
 }
@@ -92,8 +114,5 @@ int perfmon_init(struct oprofile_operati
 
 void perfmon_exit(void)
 {
-	if (!using_perfmon)
-		return;
-
-	pfm_unregister_buffer_fmt(oprofile_fmt.fmt_uuid);
+	__perfmon_exit();
 }
diff -rpuN linux-2.6.18.8/arch/ia64/oprofile/xenoprof.c linux-2.6.18-xen-3.3.0/arch/ia64/oprofile/xenoprof.c
--- linux-2.6.18.8/arch/ia64/oprofile/xenoprof.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/oprofile/xenoprof.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,142 @@
+/******************************************************************************
+ * xenoprof ia64 specific part
+ *
+ * Copyright (c) 2006 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+#include <linux/init.h>
+#include <linux/oprofile.h>
+#include <linux/ioport.h>
+
+#include <xen/driver_util.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/xenoprof.h>
+#include <xen/xenoprof.h>
+
+#include "oprofile_perfmon.h"
+
+void __init xenoprof_arch_init_counter(struct xenoprof_init *init)
+{
+	init->num_events = 0; /* perfmon manages. */
+}
+
+void xenoprof_arch_counter(void)
+{
+	/* nothing. perfmon does. */
+}
+
+void xenoprof_arch_start(void) 
+{
+	perfmon_start();
+}
+
+void xenoprof_arch_stop(void)
+{
+	perfmon_stop();
+}
+
+/* XXX move them to an appropriate header file. */
+struct resource* xen_ia64_allocate_resource(unsigned long size);
+void xen_ia64_release_resource(struct resource *res);
+void xen_ia64_unmap_resource(struct resource *res);
+
+struct resource*
+xenoprof_ia64_allocate_resource(int32_t max_samples)
+{
+	unsigned long bufsize;
+
+	/* XXX add hypercall to get bufsize? */
+	/*     this value is taken from alloc_xenoprof_struct(). */
+#if 0
+	bufsize = NR_CPUS * (sizeof(struct xenoprof_buf) +
+			     (max_samples - 1) * sizeof(struct event_log));
+	bufsize = PAGE_ALIGN(bufsize) + PAGE_SIZE;
+#else
+#define MAX_OPROF_SHARED_PAGES 32
+	bufsize = (MAX_OPROF_SHARED_PAGES + 1) * PAGE_SIZE;
+#endif
+	return xen_ia64_allocate_resource(bufsize);
+}
+
+void xenoprof_arch_unmap_shared_buffer(struct xenoprof_shared_buffer *sbuf)
+{
+	if (sbuf->buffer) {
+		xen_ia64_unmap_resource(sbuf->arch.res);
+		sbuf->buffer = NULL;
+		sbuf->arch.res = NULL;
+	}
+}
+
+int xenoprof_arch_map_shared_buffer(struct xenoprof_get_buffer *get_buffer,
+                                    struct xenoprof_shared_buffer *sbuf)
+{
+	int ret;
+	struct resource *res;
+
+	sbuf->buffer = NULL;
+	sbuf->arch.res = NULL;
+
+	res = xenoprof_ia64_allocate_resource(get_buffer->max_samples);
+	if (IS_ERR(res))
+		return PTR_ERR(res);
+
+	get_buffer->buf_gmaddr = res->start;
+
+	ret = HYPERVISOR_xenoprof_op(XENOPROF_get_buffer, get_buffer);
+	if (ret) {
+		xen_ia64_release_resource(res);
+		return ret;
+	}
+
+	BUG_ON((res->end - res->start + 1) <
+	       get_buffer->bufsize * get_buffer->nbuf);
+
+	sbuf->buffer = __va(res->start);
+	sbuf->arch.res = res;
+
+	return ret;
+}
+
+int xenoprof_arch_set_passive(struct xenoprof_passive *pdomain,
+                              struct xenoprof_shared_buffer *sbuf)
+{
+	int ret;
+	struct resource *res;
+
+	sbuf->buffer = NULL;
+	sbuf->arch.res = NULL;
+
+	res = xenoprof_ia64_allocate_resource(pdomain->max_samples);
+	if (IS_ERR(res))
+		return PTR_ERR(res);
+
+	pdomain->buf_gmaddr = res->start;
+
+	ret = HYPERVISOR_xenoprof_op(XENOPROF_set_passive, pdomain);
+	if (ret) {
+		xen_ia64_release_resource(res);
+		return ret;
+	}
+
+	BUG_ON((res->end - res->start + 1) < pdomain->bufsize * pdomain->nbuf);
+
+	sbuf->buffer = __va(res->start);
+	sbuf->arch.res = res;
+
+	return ret;
+}
diff -rpuN linux-2.6.18.8/arch/ia64/pci/pci.c linux-2.6.18-xen-3.3.0/arch/ia64/pci/pci.c
--- linux-2.6.18.8/arch/ia64/pci/pci.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/pci/pci.c	2008-08-21 11:36:07.000000000 +0200
@@ -30,6 +30,15 @@
 #include <asm/irq.h>
 #include <asm/hw_irq.h>
 
+#ifdef CONFIG_XEN
+struct ioremap_issue_list {
+	struct list_head		listp;
+	unsigned long			start;
+	unsigned long			end;
+};
+typedef struct ioremap_issue_list	ioremap_issue_list_t;
+#endif /* CONFIG_XEN */
+
 /*
  * Low-level SAL-based PCI configuration access functions. Note that SAL
  * calls are already serialized (via sal_lock), so we don't need another
@@ -165,6 +174,11 @@ new_space (u64 phys_base, int sparse)
 	io_space[i].mmio_base = mmio_base;
 	io_space[i].sparse = sparse;
 
+#ifdef CONFIG_XEN
+	if (is_initial_xendomain())
+		HYPERVISOR_add_io_space(phys_base, sparse, i);
+#endif
+
 	return i;
 }
 
@@ -332,6 +346,169 @@ pcibios_setup_root_windows(struct pci_bu
 	}
 }
 
+#ifdef CONFIG_XEN
+static void __devinit
+__cleanup_issue_list(struct list_head *top)
+{
+	ioremap_issue_list_t *ptr, *tmp_ptr;
+
+	list_for_each_entry_safe(ptr, tmp_ptr, top, listp) {
+		list_del(&(ptr->listp));
+		kfree(ptr);
+	}
+}
+
+static int __devinit
+__add_issue_list(unsigned long start, unsigned long end, struct list_head *top)
+{
+	ioremap_issue_list_t *ptr, *new;
+
+	if (start > end) {
+		printk(KERN_ERR "%s: Internal error (start addr > end addr)\n",
+		       __FUNCTION__);
+		return 0;
+	}
+
+	/*
+	 * Head of the resource structure list contains
+	 * dummy val.(start=0, end=~0), so skip it
+	 */
+	if ((start == 0) && (end == ~0))
+		return 0;
+
+	start &= PAGE_MASK;
+	end |= ~PAGE_MASK;
+
+	/* We can merge specified address range into existing entry */
+	list_for_each_entry(ptr, top, listp) {
+		if ((ptr->start > end + 1) || (ptr->end + 1 < start))
+			continue;
+		ptr->start = min(start, ptr->start);
+		ptr->end = max(end, ptr->end);
+		return 0;
+	}
+
+	/* We could not merge, so create new entry */
+	new = kmalloc(sizeof(ioremap_issue_list_t), GFP_KERNEL);
+	if (new == NULL) {
+		printk(KERN_ERR "%s: Could not allocate memory. "
+		       "HYPERVISOR_ioremap will not be issued\n",
+		       __FUNCTION__);
+		return -ENOMEM;
+	}
+
+	new->start = start;
+	new->end = end;
+
+	/* Insert the new entry to the list by ascending order */
+	if (list_empty(top)) {
+		list_add_tail(&(new->listp), top);
+		return 0;
+	}
+	list_for_each_entry(ptr, top, listp) {
+		if (new->start > ptr->start)
+			continue;
+		list_add(&(new->listp), ((struct list_head *)ptr)->prev);
+		return 0;
+	}
+	list_add_tail(&(new->listp), top);
+
+	return 0;
+}
+
+static int __devinit
+__make_issue_list(struct resource *ptr, struct list_head *top)
+{
+	int ret;
+
+	if (ptr->child) {
+		ret = __make_issue_list(ptr->child, top);
+		if (ret)
+			return ret;
+	}
+	if (ptr->sibling) {
+		ret = __make_issue_list(ptr->sibling, top);
+		if (ret)
+			return ret;
+	}
+
+	if (ptr->flags & IORESOURCE_MEM) {
+		ret = __add_issue_list(ptr->start, ptr->end, top);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static void __devinit
+__compress_issue_list(struct list_head *top)
+{
+	ioremap_issue_list_t *ptr, *tmp_ptr, *next;
+	int compressed;
+
+	/*
+	 * Merge adjacent entries, if overlapped
+	 * (entries are sorted by ascending order)
+	 */
+	list_for_each_entry_safe(ptr, tmp_ptr, top, listp) {
+		if (list_is_last((struct list_head *)ptr, top))
+			continue;
+
+		next = (ioremap_issue_list_t *)
+			(((struct list_head *)ptr)->next);
+		if (next->start <= (ptr->end) + 1) {
+			next->start = min(ptr->start, next->start);
+			next->end   = max(ptr->end, next->end);
+
+			list_del(&(ptr->listp));
+			kfree(ptr);
+		}
+	}
+}
+
+static int __devinit
+__issue_ioremap(struct list_head *top)
+{
+	ioremap_issue_list_t *ptr, *tmp_ptr;
+	unsigned int offset;
+
+	list_for_each_entry_safe(ptr, tmp_ptr, top, listp) {
+		offset = HYPERVISOR_ioremap(ptr->start,
+					    ptr->end - ptr->start + 1);
+		if (offset == ~0) {
+			printk(KERN_ERR "%s: HYPERVISOR_ioremap() failed. "
+			       "Address Range: 0x%016lx-0x%016lx\n",
+			       __FUNCTION__, ptr->start, ptr->end);
+		}
+
+		list_del(&(ptr->listp));
+		kfree(ptr);
+	}
+	
+	return 0;
+}
+
+static int __devinit
+do_ioremap_on_resource_list(struct resource *top)
+{
+	LIST_HEAD(ioremap_issue_list_top);
+	int ret;
+
+	ret = __make_issue_list(top, &ioremap_issue_list_top);
+	if (ret) {
+		__cleanup_issue_list(&ioremap_issue_list_top);
+		return ret;
+	}
+
+	__compress_issue_list(&ioremap_issue_list_top);
+
+	(void)__issue_ioremap(&ioremap_issue_list_top);
+
+	return 0;
+}
+#endif /* CONFIG_XEN */
+
 struct pci_bus * __devinit
 pci_acpi_scan_root(struct acpi_device *device, int domain, int bus)
 {
@@ -375,6 +552,18 @@ pci_acpi_scan_root(struct acpi_device *d
 	if (pbus)
 		pcibios_setup_root_windows(pbus, controller);
 
+#ifdef CONFIG_XEN
+	if (is_initial_xendomain()) {
+		if (do_ioremap_on_resource_list(&iomem_resource) != 0) {
+			printk(KERN_ERR
+			       "%s: Counld not issue HYPERVISOR_ioremap "
+			       "due to lack of memory or hypercall failure\n",
+			       __FUNCTION__);
+			goto out3;
+		}
+	}
+#endif /* CONFIG_XEN */
+
 	return pbus;
 
 out3:
@@ -818,3 +1007,31 @@ int pci_vector_resources(int last, int n
 
 	return count;
 }
+
+#ifdef CONFIG_XEN
+void __devinit xen_add_resource(struct pci_controller *controller,
+				unsigned int domain, unsigned int bus,
+				struct acpi_resource *resource)
+{
+	struct pci_root_info info;
+	char *name;
+
+	name = kmalloc(16, GFP_KERNEL);
+	if (!name)
+		return;
+
+	sprintf(name, "PCI Bus %04x:%02x", domain, bus);
+	info.controller = controller;
+	info.name = name;
+
+	add_window(resource, &info);
+}
+EXPORT_SYMBOL(xen_add_resource);
+
+void __devinit xen_pcibios_setup_root_windows(struct pci_bus *bus,
+					      struct pci_controller *controller)
+{
+	pcibios_setup_root_windows(bus, controller);
+}
+EXPORT_SYMBOL(xen_pcibios_setup_root_windows);
+#endif
diff -rpuN linux-2.6.18.8/arch/ia64/sn/kernel/setup.c linux-2.6.18-xen-3.3.0/arch/ia64/sn/kernel/setup.c
--- linux-2.6.18.8/arch/ia64/sn/kernel/setup.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/sn/kernel/setup.c	2008-08-21 11:36:07.000000000 +0200
@@ -763,5 +763,13 @@ int sn_prom_feature_available(int id)
 		return 0;
 	return test_bit(id, sn_prom_features);
 }
+
+void
+sn_kernel_launch_event(void)
+{
+	/* ignore status until we understand possible failure, if any*/
+	if (ia64_sn_kernel_launch_event())
+		printk(KERN_ERR "KEXEC is not supported in this PROM, Please update the PROM.\n");
+}
 EXPORT_SYMBOL(sn_prom_feature_available);
 
diff -rpuN linux-2.6.18.8/arch/ia64/sn/pci/pcibr/pcibr_provider.c linux-2.6.18-xen-3.3.0/arch/ia64/sn/pci/pcibr/pcibr_provider.c
--- linux-2.6.18.8/arch/ia64/sn/pci/pcibr/pcibr_provider.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/sn/pci/pcibr/pcibr_provider.c	2008-08-21 11:36:07.000000000 +0200
@@ -15,6 +15,7 @@
 #include <asm/sn/pcibus_provider_defs.h>
 #include <asm/sn/pcidev.h>
 #include <asm/sn/sn_sal.h>
+#include <asm/sn/pic.h>
 #include <asm/sn/sn2/sn_hwperf.h>
 #include "xtalk/xwidgetdev.h"
 #include "xtalk/hubdev.h"
@@ -129,9 +130,9 @@ pcibr_bus_fixup(struct pcibus_bussoft *p
 	}
 
 	memcpy(soft, prom_bussoft, sizeof(struct pcibus_info));
-	soft->pbi_buscommon.bs_base =
-	    (((u64) soft->pbi_buscommon.
-	      bs_base << 4) >> 4) | __IA64_UNCACHED_OFFSET;
+	soft->pbi_buscommon.bs_base = (unsigned long)
+		ioremap(REGION_OFFSET(soft->pbi_buscommon.bs_base),
+			sizeof(struct pic));
 
 	spin_lock_init(&soft->pbi_lock);
 
diff -rpuN linux-2.6.18.8/arch/ia64/sn/pci/tioca_provider.c linux-2.6.18-xen-3.3.0/arch/ia64/sn/pci/tioca_provider.c
--- linux-2.6.18.8/arch/ia64/sn/pci/tioca_provider.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/sn/pci/tioca_provider.c	2008-08-21 11:36:07.000000000 +0200
@@ -611,7 +611,9 @@ tioca_bus_fixup(struct pcibus_bussoft *p
 		return NULL;
 
 	memcpy(tioca_common, prom_bussoft, sizeof(struct tioca_common));
-	tioca_common->ca_common.bs_base |= __IA64_UNCACHED_OFFSET;
+	tioca_common->ca_common.bs_base = (unsigned long)
+		ioremap(REGION_OFFSET(tioca_common->ca_common.bs_base),
+			sizeof(struct tioca_common));
 
 	/* init kernel-private area */
 
diff -rpuN linux-2.6.18.8/arch/ia64/sn/pci/tioce_provider.c linux-2.6.18-xen-3.3.0/arch/ia64/sn/pci/tioce_provider.c
--- linux-2.6.18.8/arch/ia64/sn/pci/tioce_provider.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/sn/pci/tioce_provider.c	2008-08-21 11:36:07.000000000 +0200
@@ -1006,7 +1006,9 @@ tioce_bus_fixup(struct pcibus_bussoft *p
 		return NULL;
 
 	memcpy(tioce_common, prom_bussoft, sizeof(struct tioce_common));
-	tioce_common->ce_pcibus.bs_base |= __IA64_UNCACHED_OFFSET;
+	tioce_common->ce_pcibus.bs_base = (unsigned long)
+		ioremap(REGION_OFFSET(tioce_common->ce_pcibus.bs_base),
+			sizeof(struct tioce_common));
 
 	tioce_kern = tioce_kern_init(tioce_common);
 	if (tioce_kern == NULL) {
diff -rpuN linux-2.6.18.8/arch/ia64/xen/hypercall.S linux-2.6.18-xen-3.3.0/arch/ia64/xen/hypercall.S
--- linux-2.6.18.8/arch/ia64/xen/hypercall.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/xen/hypercall.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,141 @@
+/*
+ * Support routines for Xen hypercalls
+ *
+ * Copyright (C) 2005 Dan Magenheimer <dan.magenheimer@hp.com>
+ */
+
+#include <asm/asmmacro.h>
+#include <asm/intrinsics.h>
+
+#ifdef __INTEL_COMPILER
+# undef ASM_SUPPORTED
+#else
+# define ASM_SUPPORTED
+#endif
+
+#ifndef ASM_SUPPORTED
+GLOBAL_ENTRY(xen_get_psr)
+	XEN_HYPER_GET_PSR
+	br.ret.sptk.many rp
+	;;
+END(xen_get_psr)
+
+GLOBAL_ENTRY(xen_get_ivr)
+	XEN_HYPER_GET_IVR
+	br.ret.sptk.many rp
+	;;
+END(xen_get_ivr)
+
+GLOBAL_ENTRY(xen_get_tpr)
+	XEN_HYPER_GET_TPR
+	br.ret.sptk.many rp
+	;;
+END(xen_get_tpr)
+
+GLOBAL_ENTRY(xen_set_tpr)
+	mov r8=r32
+	XEN_HYPER_SET_TPR
+	br.ret.sptk.many rp
+	;;
+END(xen_set_tpr)
+
+GLOBAL_ENTRY(xen_eoi)
+	mov r8=r32
+	XEN_HYPER_EOI
+	br.ret.sptk.many rp
+	;;
+END(xen_eoi)
+
+GLOBAL_ENTRY(xen_thash)
+	mov r8=r32
+	XEN_HYPER_THASH
+	br.ret.sptk.many rp
+	;;
+END(xen_thash)
+
+GLOBAL_ENTRY(xen_set_itm)
+	mov r8=r32
+	XEN_HYPER_SET_ITM
+	br.ret.sptk.many rp
+	;;
+END(xen_set_itm)
+
+GLOBAL_ENTRY(xen_ptcga)
+	mov r8=r32
+	mov r9=r33
+	XEN_HYPER_PTC_GA
+	br.ret.sptk.many rp
+	;;
+END(xen_ptcga)
+
+GLOBAL_ENTRY(xen_get_rr)
+	mov r8=r32
+	XEN_HYPER_GET_RR
+	br.ret.sptk.many rp
+	;;
+END(xen_get_rr)
+
+GLOBAL_ENTRY(xen_set_rr)
+	mov r8=r32
+	mov r9=r33
+	XEN_HYPER_SET_RR
+	br.ret.sptk.many rp
+	;;
+END(xen_set_rr)
+
+GLOBAL_ENTRY(xen_set_kr)
+	mov r8=r32
+	mov r9=r33
+	XEN_HYPER_SET_KR
+	br.ret.sptk.many rp
+END(xen_set_kr)
+
+GLOBAL_ENTRY(xen_fc)
+	mov r8=r32
+	XEN_HYPER_FC
+	br.ret.sptk.many rp
+END(xen_fc)
+
+GLOBAL_ENTRY(xen_get_cpuid)
+	mov r8=r32
+	XEN_HYPER_GET_CPUID
+	br.ret.sptk.many rp
+END(xen_get_cpuid)
+
+GLOBAL_ENTRY(xen_get_pmd)
+	mov r8=r32
+	XEN_HYPER_GET_PMD
+	br.ret.sptk.many rp
+END(xen_get_pmd)
+
+#ifdef CONFIG_IA32_SUPPORT
+GLOBAL_ENTRY(xen_get_eflag)
+	XEN_HYPER_GET_EFLAG
+	br.ret.sptk.many rp
+END(xen_get_eflag)
+	
+// some bits aren't set if pl!=0, see SDM vol1 3.1.8
+GLOBAL_ENTRY(xen_set_eflag)
+	mov r8=r32
+	XEN_HYPER_SET_EFLAG
+	br.ret.sptk.many rp
+END(xen_set_eflag)
+#endif /* CONFIG_IA32_SUPPORT */
+#endif /* ASM_SUPPORTED */
+
+GLOBAL_ENTRY(xen_send_ipi)
+	mov r14=r32
+	mov r15=r33
+	mov r2=0x400
+	break 0x1000
+	;;
+	br.ret.sptk.many rp
+	;;
+END(xen_send_ipi)
+
+GLOBAL_ENTRY(__hypercall)
+	mov r2=r37
+	break 0x1000
+	br.ret.sptk.many b0
+	;; 
+END(__hypercall)
diff -rpuN linux-2.6.18.8/arch/ia64/xen/hypervisor.c linux-2.6.18-xen-3.3.0/arch/ia64/xen/hypervisor.c
--- linux-2.6.18.8/arch/ia64/xen/hypervisor.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/xen/hypervisor.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1526 @@
+/******************************************************************************
+ * include/asm-ia64/shadow.h
+ *
+ * Copyright (c) 2006 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+
+#include <linux/spinlock.h>
+#include <linux/bootmem.h>
+#include <linux/module.h>
+#include <linux/vmalloc.h>
+#include <linux/efi.h>
+#include <asm/page.h>
+#include <asm/pgalloc.h>
+#include <asm/meminit.h>
+#include <asm/hypervisor.h>
+#include <asm/hypercall.h>
+#include <xen/interface/memory.h>
+#include <xen/xencons.h>
+#include <xen/balloon.h>
+
+shared_info_t *HYPERVISOR_shared_info __read_mostly =
+	(shared_info_t *)XSI_BASE;
+EXPORT_SYMBOL(HYPERVISOR_shared_info);
+
+start_info_t *xen_start_info;
+EXPORT_SYMBOL(xen_start_info);
+
+EXPORT_SYMBOL(running_on_xen);
+
+#ifdef CONFIG_XEN_IA64_EXPOSE_P2M
+static int p2m_expose_init(void);
+#else
+#define p2m_expose_init() (-ENOSYS)
+#define p2m_expose_resume() ((void)0)
+#endif
+
+EXPORT_SYMBOL(__hypercall);
+
+void __init
+xen_setup(char **cmdline_p)
+{
+	struct dom0_vga_console_info *info;
+	extern void dig_setup(char **cmdline_p);
+
+	if (ia64_platform_is("xen"))
+		dig_setup(cmdline_p);
+
+	if (!is_running_on_xen() || !is_initial_xendomain())
+		return;
+
+	info = (void *)((char *)xen_start_info +
+			xen_start_info->console.dom0.info_off);
+	dom0_init_screen_info(info, xen_start_info->console.dom0.info_size);
+
+	xen_start_info->console.domU.mfn = 0;
+	xen_start_info->console.domU.evtchn = 0;
+}
+
+void __cpuinit
+xen_cpu_init(void)
+{
+	extern void xen_smp_intr_init(void);
+	xen_smp_intr_init();
+}
+
+/*
+ *XXX same as i386, x86_64 contiguous_bitmap_set(), contiguous_bitmap_clear()
+ * move those to lib/contiguous_bitmap?
+ *XXX discontigmem/sparsemem
+ */
+
+/*
+ * Bitmap is indexed by page number. If bit is set, the page is part of a
+ * xen_create_contiguous_region() area of memory.
+ */
+unsigned long *contiguous_bitmap __read_mostly;
+
+#ifdef CONFIG_VIRTUAL_MEM_MAP
+/* Following logic is stolen from create_mem_map_table() for virtual memmap */
+static int
+create_contiguous_bitmap(u64 start, u64 end, void *arg)
+{
+	unsigned long address, start_page, end_page;
+	unsigned long bitmap_start, bitmap_end;
+	unsigned char *bitmap;
+	int node;
+	pgd_t *pgd;
+	pud_t *pud;
+	pmd_t *pmd;
+	pte_t *pte;
+
+	bitmap_start = (unsigned long)contiguous_bitmap +
+		       ((__pa(start) >> PAGE_SHIFT) >> 3);
+	bitmap_end = (unsigned long)contiguous_bitmap +
+		     (((__pa(end) >> PAGE_SHIFT) + 2 * BITS_PER_LONG) >> 3);
+
+	start_page = bitmap_start & PAGE_MASK;
+	end_page = PAGE_ALIGN(bitmap_end);
+	node = paddr_to_nid(__pa(start));
+
+	bitmap = alloc_bootmem_pages_node(NODE_DATA(node),
+					  end_page - start_page);
+	BUG_ON(!bitmap);
+	memset(bitmap, 0, end_page - start_page);
+
+	for (address = start_page; address < end_page; address += PAGE_SIZE) {
+		pgd = pgd_offset_k(address);
+		if (pgd_none(*pgd))
+			pgd_populate(&init_mm, pgd,
+				     alloc_bootmem_pages_node(NODE_DATA(node),
+							      PAGE_SIZE));
+		pud = pud_offset(pgd, address);
+
+		if (pud_none(*pud))
+			pud_populate(&init_mm, pud,
+				     alloc_bootmem_pages_node(NODE_DATA(node),
+							      PAGE_SIZE));
+		pmd = pmd_offset(pud, address);
+
+		if (pmd_none(*pmd))
+			pmd_populate_kernel(&init_mm, pmd,
+					    alloc_bootmem_pages_node
+					    (NODE_DATA(node), PAGE_SIZE));
+		pte = pte_offset_kernel(pmd, address);
+
+		if (pte_none(*pte))
+			set_pte(pte,
+				pfn_pte(__pa(bitmap + (address - start_page))
+					>> PAGE_SHIFT, PAGE_KERNEL));
+	}
+	return 0;
+}
+#endif
+
+static void
+__contiguous_bitmap_init(unsigned long size)
+{
+	contiguous_bitmap = alloc_bootmem_pages(size);
+	BUG_ON(!contiguous_bitmap);
+	memset(contiguous_bitmap, 0, size);
+}
+
+void
+xen_contiguous_bitmap_init(unsigned long end_pfn)
+{
+	unsigned long size = (end_pfn + 2 * BITS_PER_LONG) >> 3;
+#ifndef CONFIG_VIRTUAL_MEM_MAP
+	__contiguous_bitmap_init(size);
+#else
+	unsigned long max_gap = 0;
+
+	efi_memmap_walk(find_largest_hole, (u64*)&max_gap);
+	if (max_gap < LARGE_GAP) {
+		__contiguous_bitmap_init(size);
+	} else {
+		unsigned long map_size = PAGE_ALIGN(size);
+		vmalloc_end -= map_size;
+		contiguous_bitmap = (unsigned long*)vmalloc_end;
+		efi_memmap_walk(create_contiguous_bitmap, NULL);
+	}
+#endif
+}
+
+#if 0
+int
+contiguous_bitmap_test(void* p)
+{
+	return test_bit(__pa(p) >> PAGE_SHIFT, contiguous_bitmap);
+}
+#endif
+
+static void contiguous_bitmap_set(
+	unsigned long first_page, unsigned long nr_pages)
+{
+	unsigned long start_off, end_off, curr_idx, end_idx;
+
+	curr_idx  = first_page / BITS_PER_LONG;
+	start_off = first_page & (BITS_PER_LONG-1);
+	end_idx   = (first_page + nr_pages) / BITS_PER_LONG;
+	end_off   = (first_page + nr_pages) & (BITS_PER_LONG-1);
+
+	if (curr_idx == end_idx) {
+		contiguous_bitmap[curr_idx] |=
+			((1UL<<end_off)-1) & -(1UL<<start_off);
+	} else {
+		contiguous_bitmap[curr_idx] |= -(1UL<<start_off);
+		while ( ++curr_idx < end_idx )
+			contiguous_bitmap[curr_idx] = ~0UL;
+		contiguous_bitmap[curr_idx] |= (1UL<<end_off)-1;
+	}
+}
+
+static void contiguous_bitmap_clear(
+	unsigned long first_page, unsigned long nr_pages)
+{
+	unsigned long start_off, end_off, curr_idx, end_idx;
+
+	curr_idx  = first_page / BITS_PER_LONG;
+	start_off = first_page & (BITS_PER_LONG-1);
+	end_idx   = (first_page + nr_pages) / BITS_PER_LONG;
+	end_off   = (first_page + nr_pages) & (BITS_PER_LONG-1);
+
+	if (curr_idx == end_idx) {
+		contiguous_bitmap[curr_idx] &=
+			-(1UL<<end_off) | ((1UL<<start_off)-1);
+	} else {
+		contiguous_bitmap[curr_idx] &= (1UL<<start_off)-1;
+		while ( ++curr_idx != end_idx )
+			contiguous_bitmap[curr_idx] = 0;
+		contiguous_bitmap[curr_idx] &= -(1UL<<end_off);
+	}
+}
+
+/*
+ * __xen_create_contiguous_region(), __xen_destroy_contiguous_region()
+ * are based on i386 xen_create_contiguous_region(),
+ * xen_destroy_contiguous_region()
+ */
+
+/* Protected by balloon_lock. */
+#define MAX_CONTIG_ORDER 7
+static unsigned long discontig_frames[1<<MAX_CONTIG_ORDER];
+
+/* Ensure multi-page extents are contiguous in machine memory. */
+int
+__xen_create_contiguous_region(unsigned long vstart,
+			       unsigned int order, unsigned int address_bits)
+{
+	unsigned long error = 0;
+	unsigned long gphys = __pa(vstart);
+	unsigned long start_gpfn = gphys >> PAGE_SHIFT;
+	unsigned long num_gpfn = 1 << order;
+	unsigned long i;
+	unsigned long flags;
+
+	unsigned long *in_frames = discontig_frames, out_frame;
+	int success;
+	struct xen_memory_exchange exchange = {
+		.in = {
+			.nr_extents   = num_gpfn,
+			.extent_order = 0,
+			.domid        = DOMID_SELF
+		},
+		.out = {
+			 .nr_extents   = 1,
+			 .extent_order = order,
+			 .address_bits = address_bits,
+			 .domid        = DOMID_SELF
+		 },
+		.nr_exchanged = 0
+	};
+
+	if (unlikely(order > MAX_CONTIG_ORDER))
+		return -ENOMEM;
+	
+	set_xen_guest_handle(exchange.in.extent_start, in_frames);
+	set_xen_guest_handle(exchange.out.extent_start, &out_frame);
+
+	scrub_pages(vstart, num_gpfn);
+
+	balloon_lock(flags);
+
+	/* Get a new contiguous memory extent. */
+	for (i = 0; i < num_gpfn; i++)
+		in_frames[i] = start_gpfn + i;
+	out_frame = start_gpfn;
+	error = HYPERVISOR_memory_op(XENMEM_exchange, &exchange);
+	success = (exchange.nr_exchanged == num_gpfn);
+	BUG_ON(!success && ((exchange.nr_exchanged != 0) || (error == 0)));
+	BUG_ON(success && (error != 0));
+	if (unlikely(error == -ENOSYS)) {
+		/* Compatibility when XENMEM_exchange is unsupported. */
+		error = HYPERVISOR_memory_op(XENMEM_decrease_reservation,
+					     &exchange.in);
+		BUG_ON(error != num_gpfn);
+		error = HYPERVISOR_memory_op(XENMEM_populate_physmap,
+					     &exchange.out);
+		if (error != 1) {
+			/* Couldn't get special memory: fall back to normal. */
+			for (i = 0; i < num_gpfn; i++) {
+				in_frames[i] = start_gpfn + i;
+			}
+			error = HYPERVISOR_memory_op(XENMEM_populate_physmap,
+						     &exchange.in);
+			BUG_ON(error != num_gpfn);
+			success = 0;
+		} else
+			success = 1;
+	}
+	if (success)
+		contiguous_bitmap_set(start_gpfn, num_gpfn);
+#if 0
+	if (success) {
+		unsigned long mfn;
+		unsigned long mfn_prev = ~0UL;
+		for (i = 0; i < num_gpfn; i++) {
+			mfn = pfn_to_mfn_for_dma(start_gpfn + i);
+			if (mfn_prev != ~0UL && mfn != mfn_prev + 1) {
+				xprintk("\n");
+				xprintk("%s:%d order %d "
+					"start 0x%lx bus 0x%lx "
+					"machine 0x%lx\n",
+					__func__, __LINE__, order,
+					vstart, virt_to_bus((void*)vstart),
+					phys_to_machine_for_dma(gphys));
+				xprintk("mfn: ");
+				for (i = 0; i < num_gpfn; i++) {
+					mfn = pfn_to_mfn_for_dma(
+						start_gpfn + i);
+					xprintk("0x%lx ", mfn);
+				}
+				xprintk("\n");
+				break;
+			}
+			mfn_prev = mfn;
+		}
+	}
+#endif
+	balloon_unlock(flags);
+	return success? 0: -ENOMEM;
+}
+
+void
+__xen_destroy_contiguous_region(unsigned long vstart, unsigned int order)
+{
+	unsigned long flags;
+	unsigned long error = 0;
+	unsigned long start_gpfn = __pa(vstart) >> PAGE_SHIFT;
+	unsigned long num_gpfn = 1UL << order;
+	unsigned long i;
+
+	unsigned long *out_frames = discontig_frames, in_frame;
+	int            success;
+	struct xen_memory_exchange exchange = {
+		.in = {
+			.nr_extents   = 1,
+			.extent_order = order,
+			.domid        = DOMID_SELF
+		},
+		.out = {
+			 .nr_extents   = num_gpfn,
+			 .extent_order = 0,
+			 .address_bits = 0,
+			 .domid        = DOMID_SELF
+		 },
+		.nr_exchanged = 0
+	};
+	
+
+	if (!test_bit(start_gpfn, contiguous_bitmap))
+		return;
+
+	if (unlikely(order > MAX_CONTIG_ORDER))
+		return;
+
+	set_xen_guest_handle(exchange.in.extent_start, &in_frame);
+	set_xen_guest_handle(exchange.out.extent_start, out_frames);
+
+	scrub_pages(vstart, num_gpfn);
+
+	balloon_lock(flags);
+
+	contiguous_bitmap_clear(start_gpfn, num_gpfn);
+
+	/* Do the exchange for non-contiguous MFNs. */
+	in_frame = start_gpfn;
+	for (i = 0; i < num_gpfn; i++)
+		out_frames[i] = start_gpfn + i;
+	error = HYPERVISOR_memory_op(XENMEM_exchange, &exchange);
+	success = (exchange.nr_exchanged == 1);
+	BUG_ON(!success && ((exchange.nr_exchanged != 0) || (error == 0)));
+	BUG_ON(success && (error != 0));
+	if (unlikely(error == -ENOSYS)) {
+		/* Compatibility when XENMEM_exchange is unsupported. */
+		error = HYPERVISOR_memory_op(XENMEM_decrease_reservation,
+					     &exchange.in);
+		BUG_ON(error != 1);
+
+		error = HYPERVISOR_memory_op(XENMEM_populate_physmap,
+					     &exchange.out);
+		BUG_ON(error != num_gpfn);
+	}
+	balloon_unlock(flags);
+}
+
+int
+xen_limit_pages_to_max_mfn(struct page *pages, unsigned int order,
+			   unsigned int address_bits)
+{
+	return xen_create_contiguous_region((unsigned long)page_address(pages),
+					    order, address_bits);
+}
+
+/****************************************************************************
+ * grant table hack
+ * cmd: GNTTABOP_xxx
+ */
+#include <linux/mm.h>
+#include <xen/interface/xen.h>
+#include <xen/gnttab.h>
+
+void *arch_gnttab_alloc_shared(unsigned long *frames)
+{
+	return __va(frames[0] << PAGE_SHIFT);
+}
+
+static void
+gnttab_map_grant_ref_pre(struct gnttab_map_grant_ref *uop)
+{
+	uint32_t flags;
+
+	flags = uop->flags;
+
+	if (flags & GNTMAP_host_map) {
+		if (flags & GNTMAP_application_map) {
+			xprintd("GNTMAP_application_map is not supported yet:"
+				" flags 0x%x\n", flags);
+			BUG();
+		}
+		if (flags & GNTMAP_contains_pte) {
+			xprintd("GNTMAP_contains_pte is not supported yet"
+				" flags 0x%x\n", flags);
+			BUG();
+		}
+	} else if (flags & GNTMAP_device_map) {
+		xprintd("GNTMAP_device_map is not supported yet 0x%x\n",
+			flags);
+		BUG(); /* XXX not yet. actually this flag is not used. */
+	} else {
+		BUG();
+	}
+}
+
+int
+HYPERVISOR_grant_table_op(unsigned int cmd, void *uop, unsigned int count)
+{
+	if (cmd == GNTTABOP_map_grant_ref) {
+		unsigned int i;
+		for (i = 0; i < count; i++) {
+			gnttab_map_grant_ref_pre(
+				(struct gnttab_map_grant_ref*)uop + i);
+		}
+	}
+	return xencomm_hypercall_grant_table_op(cmd, uop, count);
+}
+EXPORT_SYMBOL(HYPERVISOR_grant_table_op);
+
+/**************************************************************************
+ * foreign mapping
+ */
+#include <linux/efi.h>
+#include <asm/meminit.h> /* for IA64_GRANULE_SIZE, GRANULEROUND{UP,DOWN}() */
+
+static unsigned long privcmd_resource_min = 0;
+/* Xen/ia64 currently can handle pseudo physical address bits up to
+ * (PAGE_SHIFT * 3) */
+static unsigned long privcmd_resource_max =
+	GRANULEROUNDDOWN((1UL << (PAGE_SHIFT * 3)) - 1);
+static unsigned long privcmd_resource_align = IA64_GRANULE_SIZE;
+
+static unsigned long
+md_end_addr(const efi_memory_desc_t *md)
+{
+	return md->phys_addr + (md->num_pages << EFI_PAGE_SHIFT);
+}
+
+#define XEN_IA64_PRIVCMD_LEAST_GAP_SIZE	(1024 * 1024 * 1024UL)
+static int
+xen_ia64_privcmd_check_size(unsigned long start, unsigned long end)
+{
+	return (start < end &&
+		(end - start) > XEN_IA64_PRIVCMD_LEAST_GAP_SIZE);
+}
+
+static int __init
+xen_ia64_privcmd_init(void)
+{
+	void *efi_map_start, *efi_map_end, *p;
+	u64 efi_desc_size;
+	efi_memory_desc_t *md;
+	unsigned long tmp_min;
+	unsigned long tmp_max;
+	unsigned long gap_size;
+	unsigned long prev_end;
+
+	if (!is_running_on_xen())
+		return -1;
+
+	efi_map_start = __va(ia64_boot_param->efi_memmap);
+	efi_map_end   = efi_map_start + ia64_boot_param->efi_memmap_size;
+	efi_desc_size = ia64_boot_param->efi_memdesc_size;
+
+	/* at first check the used highest address */
+	for (p = efi_map_start; p < efi_map_end; p += efi_desc_size) {
+		/* nothing */;
+	}
+	md = p - efi_desc_size;
+	privcmd_resource_min = GRANULEROUNDUP(md_end_addr(md));
+	if (xen_ia64_privcmd_check_size(privcmd_resource_min,
+					privcmd_resource_max))
+		goto out;
+
+	/* the used highest address is too large.
+	 * try to find the largest gap. */
+	tmp_min = privcmd_resource_max;
+	tmp_max = 0;
+	gap_size = 0;
+	prev_end = 0;
+	for (p = efi_map_start;
+	     p < efi_map_end - efi_desc_size;
+	     p += efi_desc_size) {
+		unsigned long end;
+		efi_memory_desc_t* next;
+		unsigned long next_start;
+
+		md = p;
+		end = md_end_addr(md);
+		if (end > privcmd_resource_max)
+			break;
+		if (end < prev_end) {
+			/* work around. 
+			 * Xen may pass incompletely sorted memory
+			 * descriptors like
+			 * [x, x + length]
+			 * [x, x]
+			 * this order should be reversed. */
+			continue;
+		}
+		next = p + efi_desc_size;
+		next_start = next->phys_addr;
+		if (next_start > privcmd_resource_max)
+			next_start = privcmd_resource_max;
+		if (end < next_start && gap_size < (next_start - end)) {
+			tmp_min = end;
+			tmp_max = next_start;
+			gap_size = tmp_max - tmp_min;
+		}
+		prev_end = end;
+	}
+
+	privcmd_resource_min = GRANULEROUNDUP(tmp_min);
+	if (xen_ia64_privcmd_check_size(privcmd_resource_min, tmp_max)) {
+		privcmd_resource_max = tmp_max;
+		goto out;
+	}
+
+	privcmd_resource_min = tmp_min;
+	privcmd_resource_max = tmp_max;
+	if (!xen_ia64_privcmd_check_size(privcmd_resource_min,
+					 privcmd_resource_max)) {
+		/* Any large enough gap isn't found.
+		 * go ahead anyway with the warning hoping that large region
+		 * won't be requested. */
+		printk(KERN_WARNING "xen privcmd: "
+		       "large enough region for privcmd mmap is not found.\n");
+	}
+
+out:
+	printk(KERN_INFO "xen privcmd uses pseudo physical addr range "
+	       "[0x%lx, 0x%lx] (%ldMB)\n",
+	       privcmd_resource_min, privcmd_resource_max, 
+	       (privcmd_resource_max - privcmd_resource_min) >> 20);
+	BUG_ON(privcmd_resource_min >= privcmd_resource_max);
+
+	/* XXX this should be somewhere appropriate */
+	(void)p2m_expose_init();
+
+	return 0;
+}
+late_initcall(xen_ia64_privcmd_init);
+
+struct xen_ia64_privcmd_entry {
+	atomic_t	map_count;
+#define INVALID_GPFN	(~0UL)
+	unsigned long	gpfn;
+};
+
+struct xen_ia64_privcmd_range {
+	atomic_t			ref_count;
+	unsigned long			pgoff; /* in PAGE_SIZE */
+	struct resource			*res;
+
+	/* for foreign domain p2m mapping */
+	void				*private;
+	void (*callback)(struct xen_ia64_privcmd_range *range, void *arg);
+
+	unsigned long			num_entries;
+	struct xen_ia64_privcmd_entry	entries[0];
+};
+
+struct xen_ia64_privcmd_vma {
+	int				is_privcmd_mmapped;
+	struct xen_ia64_privcmd_range	*range;
+
+	unsigned long			num_entries;
+	struct xen_ia64_privcmd_entry	*entries;
+};
+
+static void
+xen_ia64_privcmd_init_entry(struct xen_ia64_privcmd_entry *entry)
+{
+	atomic_set(&entry->map_count, 0);
+	entry->gpfn = INVALID_GPFN;
+}
+
+static int
+xen_ia64_privcmd_entry_mmap(struct vm_area_struct *vma,
+			    unsigned long addr,
+			    struct xen_ia64_privcmd_range *privcmd_range,
+			    int i,
+			    unsigned long gmfn,
+			    pgprot_t prot,
+			    domid_t domid)
+{
+	int error = 0;
+	struct xen_ia64_privcmd_entry *entry = &privcmd_range->entries[i];
+	unsigned long gpfn;
+	unsigned long flags;
+
+	if ((addr & ~PAGE_MASK) != 0 || gmfn == INVALID_MFN) {
+		error = -EINVAL;
+		goto out;
+	}
+
+	if (entry->gpfn != INVALID_GPFN) {
+		error = -EBUSY;
+		goto out;
+	}
+	gpfn = (privcmd_range->res->start >> PAGE_SHIFT) + i;
+
+	flags = ASSIGN_writable;
+	if (pgprot_val(prot) == PROT_READ)
+		flags = ASSIGN_readonly;
+	error = HYPERVISOR_add_physmap_with_gmfn(gpfn, gmfn, flags, domid);
+	if (error != 0)
+		goto out;
+
+	prot = vma->vm_page_prot;
+	error = remap_pfn_range(vma, addr, gpfn, 1 << PAGE_SHIFT, prot);
+	/*
+	 * VM_PFNMAP is set in remap_pfn_range().
+	 * Reset the flag to avoid BUG_ON() in do_no_page().
+	 */
+	vma->vm_flags &= ~VM_PFNMAP;
+
+	if (error != 0) {
+		error = HYPERVISOR_zap_physmap(gpfn, 0);
+		if (error)
+			BUG(); /* XXX */
+	} else {
+		atomic_inc(&entry->map_count);
+		entry->gpfn = gpfn;
+	}
+
+out:
+	return error;
+}
+
+static void
+xen_ia64_privcmd_entry_munmap(struct xen_ia64_privcmd_range *privcmd_range,
+			      int i)
+{
+	struct xen_ia64_privcmd_entry *entry = &privcmd_range->entries[i];
+	unsigned long gpfn = entry->gpfn;
+	/* gpfn = (privcmd_range->res->start >> PAGE_SHIFT) +
+  		(vma->vm_pgoff - privcmd_range->pgoff); */
+	int error;
+
+	error = HYPERVISOR_zap_physmap(gpfn, 0);
+	if (error)
+		BUG(); /* XXX */
+	entry->gpfn = INVALID_GPFN;
+}
+
+static void
+xen_ia64_privcmd_entry_open(struct xen_ia64_privcmd_range *privcmd_range,
+			    int i)
+{
+	struct xen_ia64_privcmd_entry *entry = &privcmd_range->entries[i];
+	if (entry->gpfn != INVALID_GPFN)
+		atomic_inc(&entry->map_count);
+	else
+		BUG_ON(atomic_read(&entry->map_count) != 0);
+}
+
+static void
+xen_ia64_privcmd_entry_close(struct xen_ia64_privcmd_range *privcmd_range,
+			     int i)
+{
+	struct xen_ia64_privcmd_entry *entry = &privcmd_range->entries[i];
+	if (entry->gpfn != INVALID_GPFN &&
+	    atomic_dec_and_test(&entry->map_count))
+		xen_ia64_privcmd_entry_munmap(privcmd_range, i);
+}
+
+static void xen_ia64_privcmd_vma_open(struct vm_area_struct *vma);
+static void xen_ia64_privcmd_vma_close(struct vm_area_struct *vma);
+
+static struct page *
+xen_ia64_privcmd_vma_nopage(struct vm_area_struct *vma,
+			    unsigned long address,
+			    int *type)
+{
+	return NOPAGE_SIGBUS;
+}
+
+struct vm_operations_struct xen_ia64_privcmd_vm_ops = {
+	.open = xen_ia64_privcmd_vma_open,
+	.close = xen_ia64_privcmd_vma_close,
+	.nopage = xen_ia64_privcmd_vma_nopage
+};
+
+static void
+__xen_ia64_privcmd_vma_open(struct vm_area_struct *vma,
+			    struct xen_ia64_privcmd_vma *privcmd_vma,
+			    struct xen_ia64_privcmd_range *privcmd_range)
+{
+	unsigned long entry_offset = vma->vm_pgoff - privcmd_range->pgoff;
+	unsigned long num_entries =
+		(vma->vm_end - vma->vm_start) >> PAGE_SHIFT;
+	unsigned long i;
+
+	BUG_ON(entry_offset < 0);
+	BUG_ON(entry_offset + num_entries > privcmd_range->num_entries);
+
+	privcmd_vma->range = privcmd_range;
+	privcmd_vma->num_entries = num_entries;
+	privcmd_vma->entries = &privcmd_range->entries[entry_offset];
+	vma->vm_private_data = privcmd_vma;
+	for (i = 0; i < privcmd_vma->num_entries; i++)
+		xen_ia64_privcmd_entry_open(privcmd_range, entry_offset + i);
+
+	vma->vm_private_data = privcmd_vma;
+	vma->vm_ops = &xen_ia64_privcmd_vm_ops;
+}
+
+static void
+xen_ia64_privcmd_vma_open(struct vm_area_struct *vma)
+{
+	struct xen_ia64_privcmd_vma *old_privcmd_vma =
+		(struct xen_ia64_privcmd_vma*)vma->vm_private_data;
+	struct xen_ia64_privcmd_vma *privcmd_vma =
+		(struct xen_ia64_privcmd_vma*)vma->vm_private_data;
+	struct xen_ia64_privcmd_range *privcmd_range = privcmd_vma->range;
+
+	atomic_inc(&privcmd_range->ref_count);
+	/* vm_op->open() can't fail. */
+	privcmd_vma = kmalloc(sizeof(*privcmd_vma), GFP_KERNEL | __GFP_NOFAIL);
+	/* copy original value if necessary */
+	privcmd_vma->is_privcmd_mmapped = old_privcmd_vma->is_privcmd_mmapped;
+
+	__xen_ia64_privcmd_vma_open(vma, privcmd_vma, privcmd_range);
+}
+
+static void
+xen_ia64_privcmd_vma_close(struct vm_area_struct *vma)
+{
+	struct xen_ia64_privcmd_vma *privcmd_vma =
+		(struct xen_ia64_privcmd_vma*)vma->vm_private_data;
+	struct xen_ia64_privcmd_range *privcmd_range = privcmd_vma->range;
+	unsigned long entry_offset = vma->vm_pgoff - privcmd_range->pgoff;
+	unsigned long i;
+
+	for (i = 0; i < privcmd_vma->num_entries; i++) {
+		xen_ia64_privcmd_entry_close(privcmd_range, entry_offset + i);
+		cond_resched();
+	}
+	vma->vm_private_data = NULL;
+	kfree(privcmd_vma);
+
+	if (atomic_dec_and_test(&privcmd_range->ref_count)) {
+#if 1
+		for (i = 0; i < privcmd_range->num_entries; i++) {
+			struct xen_ia64_privcmd_entry *entry =
+				&privcmd_range->entries[i];
+			BUG_ON(atomic_read(&entry->map_count) != 0);
+			BUG_ON(entry->gpfn != INVALID_GPFN);
+		}
+#endif
+		if (privcmd_range->callback)
+			(*privcmd_range->callback)(privcmd_range,
+						   privcmd_range->private);
+		release_resource(privcmd_range->res);
+		kfree(privcmd_range->res);
+		vfree(privcmd_range);
+	}
+}
+
+int
+privcmd_enforce_singleshot_mapping(struct vm_area_struct *vma)
+{
+	struct xen_ia64_privcmd_vma *privcmd_vma =
+		(struct xen_ia64_privcmd_vma *)vma->vm_private_data;
+	return (xchg(&privcmd_vma->is_privcmd_mmapped, 1) == 0);
+}
+
+int
+privcmd_mmap(struct file * file, struct vm_area_struct * vma)
+{
+	int error;
+	unsigned long size = vma->vm_end - vma->vm_start;
+	unsigned long num_entries = size >> PAGE_SHIFT;
+	struct xen_ia64_privcmd_range *privcmd_range = NULL;
+	struct xen_ia64_privcmd_vma *privcmd_vma = NULL;
+	struct resource *res = NULL;
+	unsigned long i;
+	BUG_ON(!is_running_on_xen());
+
+	BUG_ON(file->private_data != NULL);
+
+	error = -ENOMEM;
+	privcmd_range =
+		vmalloc(sizeof(*privcmd_range) +
+			sizeof(privcmd_range->entries[0]) * num_entries);
+	if (privcmd_range == NULL)
+		goto out_enomem0;
+	privcmd_vma = kmalloc(sizeof(*privcmd_vma), GFP_KERNEL);
+	if (privcmd_vma == NULL)
+		goto out_enomem1;
+	privcmd_vma->is_privcmd_mmapped = 0;
+
+	res = kzalloc(sizeof(*res), GFP_KERNEL);
+	if (res == NULL)
+		goto out_enomem1;
+	res->name = "Xen privcmd mmap";
+	error = allocate_resource(&iomem_resource, res, size,
+				  privcmd_resource_min, privcmd_resource_max,
+				  privcmd_resource_align, NULL, NULL);
+	if (error)
+		goto out_enomem1;
+	privcmd_range->res = res;
+
+	/* DONTCOPY is essential for Xen as copy_page_range is broken. */
+	vma->vm_flags |= VM_RESERVED | VM_IO | VM_DONTCOPY;
+
+	atomic_set(&privcmd_range->ref_count, 1);
+	privcmd_range->pgoff = vma->vm_pgoff;
+	privcmd_range->num_entries = num_entries;
+	privcmd_range->private = NULL;
+	privcmd_range->callback = NULL;
+	for (i = 0; i < privcmd_range->num_entries; i++)
+		xen_ia64_privcmd_init_entry(&privcmd_range->entries[i]);
+
+	__xen_ia64_privcmd_vma_open(vma, privcmd_vma, privcmd_range);
+	return 0;
+
+out_enomem1:
+	kfree(res);
+	kfree(privcmd_vma);
+out_enomem0:
+	vfree(privcmd_range);
+	return error;
+}
+
+int
+direct_remap_pfn_range(struct vm_area_struct *vma,
+		       unsigned long address,	/* process virtual address */
+		       unsigned long gmfn,	/* gmfn, gmfn + 1, ... gmfn + size/PAGE_SIZE */
+		       unsigned long size,
+		       pgprot_t prot,
+		       domid_t  domid)		/* target domain */
+{
+	struct xen_ia64_privcmd_vma *privcmd_vma =
+		(struct xen_ia64_privcmd_vma*)vma->vm_private_data;
+	struct xen_ia64_privcmd_range *privcmd_range = privcmd_vma->range;
+	unsigned long entry_offset = vma->vm_pgoff - privcmd_range->pgoff;
+
+	unsigned long i;
+	unsigned long offset;
+	int error = 0;
+	BUG_ON(!is_running_on_xen());
+
+#if 0
+	if (prot != vm->vm_page_prot)
+		return -EINVAL;
+#endif
+
+	i = (address - vma->vm_start) >> PAGE_SHIFT;
+	for (offset = 0; offset < size; offset += PAGE_SIZE) {
+		error = xen_ia64_privcmd_entry_mmap(vma, (address + offset) & PAGE_MASK, privcmd_range, entry_offset + i, gmfn, prot, domid);
+		if (error != 0)
+			break;
+
+		i++;
+		gmfn++;
+	}
+
+	return error;
+}
+
+
+/**************************************************************************
+ * expose p2m table
+ */
+#ifdef CONFIG_XEN_IA64_EXPOSE_P2M
+#include <linux/cpu.h>
+#include <asm/uaccess.h>
+
+int p2m_initialized __read_mostly = 0;
+
+unsigned long p2m_min_low_pfn __read_mostly;
+unsigned long p2m_max_low_pfn __read_mostly;
+unsigned long p2m_convert_min_pfn __read_mostly;
+unsigned long p2m_convert_max_pfn __read_mostly;
+
+static struct resource p2m_resource = {
+	.name    = "Xen p2m table",
+	.flags   = IORESOURCE_MEM,
+};
+static unsigned long p2m_assign_start_pfn __read_mostly;
+static unsigned long p2m_assign_end_pfn __read_mostly;
+static unsigned long p2m_expose_size;	/* this is referenced only when resume.
+					 * so __read_mostly doesn't make sense.
+					 */
+volatile const pte_t *p2m_pte __read_mostly;
+
+#define GRANULE_PFN	PTRS_PER_PTE
+static unsigned long p2m_granule_pfn __read_mostly = GRANULE_PFN;
+
+#define ROUNDDOWN(x, y)  ((x) & ~((y) - 1))
+#define ROUNDUP(x, y)    (((x) + (y) - 1) & ~((y) - 1))
+
+#define P2M_PREFIX	"Xen p2m: "
+
+static int xen_ia64_p2m_expose __read_mostly = 1;
+module_param(xen_ia64_p2m_expose, int, 0);
+MODULE_PARM_DESC(xen_ia64_p2m_expose,
+		 "enable/disable xen/ia64 p2m exposure optimization\n");
+
+#ifdef CONFIG_XEN_IA64_EXPOSE_P2M_USE_DTR
+static int xen_ia64_p2m_expose_use_dtr __read_mostly = 1;
+module_param(xen_ia64_p2m_expose_use_dtr, int, 0);
+MODULE_PARM_DESC(xen_ia64_p2m_expose_use_dtr,
+		 "use/unuse dtr to map exposed p2m table\n");
+
+static const int p2m_page_shifts[] = {
+	_PAGE_SIZE_4K,
+	_PAGE_SIZE_8K,
+	_PAGE_SIZE_16K,
+	_PAGE_SIZE_64K,
+	_PAGE_SIZE_256K,
+	_PAGE_SIZE_1M,
+	_PAGE_SIZE_4M,
+	_PAGE_SIZE_16M,
+	_PAGE_SIZE_64M,
+	_PAGE_SIZE_256M,
+};
+
+struct p2m_itr_arg {
+	unsigned long vaddr;
+	unsigned long pteval;
+	unsigned long log_page_size;
+};
+static struct p2m_itr_arg p2m_itr_arg __read_mostly;
+
+/* This should be in asm-ia64/kregs.h */
+#define IA64_TR_P2M_TABLE	3
+
+static void
+p2m_itr(void *info)
+{
+	struct p2m_itr_arg *arg = (struct p2m_itr_arg*)info;
+	ia64_itr(0x2, IA64_TR_P2M_TABLE,
+		 arg->vaddr, arg->pteval, arg->log_page_size);
+	ia64_srlz_d();
+}
+
+static int
+p2m_expose_dtr_call(struct notifier_block *self,
+		    unsigned long event, void *ptr)
+{
+	unsigned int cpu = (unsigned int)(long)ptr;
+	if (event != CPU_ONLINE)
+		return 0;
+	if (p2m_initialized && xen_ia64_p2m_expose_use_dtr) {
+		unsigned int me = get_cpu();
+		if (cpu == me)
+			p2m_itr(&p2m_itr_arg);
+		else
+			smp_call_function_single(cpu, &p2m_itr, &p2m_itr_arg,
+						 1, 1);
+		put_cpu();
+	}
+	return 0;
+}
+
+static struct notifier_block p2m_expose_dtr_hotplug_notifier = {
+	.notifier_call = p2m_expose_dtr_call,
+	.next          = NULL,
+	.priority      = 0
+};
+#endif
+
+static inline unsigned long
+p2m_table_size(unsigned long num_pfn)
+{
+	return ((num_pfn + PTRS_PER_PTE - 1) / PTRS_PER_PTE) << PAGE_SHIFT;
+}
+
+static int
+p2m_expose_init(void)
+{
+	unsigned long num_pfn;
+	unsigned long p2m_size = 0;
+	unsigned long align = ~0UL;
+	int error = 0;
+#ifdef CONFIG_XEN_IA64_EXPOSE_P2M_USE_DTR
+	int i;
+	unsigned long log_page_size = 0;
+#endif
+
+	if (!xen_ia64_p2m_expose)
+		return -ENOSYS;
+	if (p2m_initialized)
+		return 0;
+
+#ifdef CONFIG_XEN_IA64_EXPOSE_P2M_USE_DTR
+	error = register_cpu_notifier(&p2m_expose_dtr_hotplug_notifier);
+	if (error < 0)
+		return error;
+#endif
+
+	lock_cpu_hotplug();
+	if (p2m_initialized)
+		goto out;
+
+#ifdef CONFIG_DISCONTIGMEM
+	p2m_min_low_pfn = min_low_pfn;
+	p2m_max_low_pfn = max_low_pfn;
+#else
+	p2m_min_low_pfn = 0;
+	p2m_max_low_pfn = max_pfn;
+#endif
+
+#ifdef CONFIG_XEN_IA64_EXPOSE_P2M_USE_DTR
+	if (xen_ia64_p2m_expose_use_dtr) {
+		unsigned long page_size = 0;
+		unsigned long granule_pfn = 0;
+		p2m_size = p2m_table_size(p2m_max_low_pfn - p2m_min_low_pfn);
+		for (i = 0;
+		     i < sizeof(p2m_page_shifts)/sizeof(p2m_page_shifts[0]);
+		     i++) {
+			log_page_size = p2m_page_shifts[i];
+			page_size = 1UL << log_page_size;
+			if (page_size < p2m_size)
+				continue;
+
+			granule_pfn = max(page_size >> PAGE_SHIFT,
+					  p2m_granule_pfn);
+			p2m_convert_min_pfn = ROUNDDOWN(p2m_min_low_pfn,
+							granule_pfn);
+			p2m_convert_max_pfn = ROUNDUP(p2m_max_low_pfn,
+						      granule_pfn);
+			num_pfn = p2m_convert_max_pfn - p2m_convert_min_pfn;
+			p2m_expose_size = num_pfn << PAGE_SHIFT;
+			p2m_size = p2m_table_size(num_pfn);
+			p2m_size = ROUNDUP(p2m_size,
+					   granule_pfn << PAGE_SHIFT);
+			if (p2m_size == page_size)
+				break;
+		}
+		if (p2m_size != page_size) {
+			printk(KERN_ERR "p2m_size != page_size\n");
+			error = -EINVAL;
+			goto out;
+		}
+		align = max(privcmd_resource_align, granule_pfn << PAGE_SHIFT);
+	} else
+#endif
+	{
+		BUG_ON(p2m_granule_pfn & (p2m_granule_pfn - 1));
+		p2m_convert_min_pfn = ROUNDDOWN(p2m_min_low_pfn,
+						p2m_granule_pfn);
+		p2m_convert_max_pfn = ROUNDUP(p2m_max_low_pfn,
+					      p2m_granule_pfn);
+		num_pfn = p2m_convert_max_pfn - p2m_convert_min_pfn;
+		p2m_expose_size = num_pfn << PAGE_SHIFT;
+		p2m_size = p2m_table_size(num_pfn);
+		p2m_size = ROUNDUP(p2m_size, p2m_granule_pfn << PAGE_SHIFT);
+		align = max(privcmd_resource_align,
+			    p2m_granule_pfn << PAGE_SHIFT);
+	}
+	
+	/* use privcmd region */
+	error = allocate_resource(&iomem_resource, &p2m_resource, p2m_size,
+				  privcmd_resource_min, privcmd_resource_max,
+				  align, NULL, NULL);
+	if (error) {
+		printk(KERN_ERR P2M_PREFIX
+		       "can't allocate region for p2m exposure "
+		       "[0x%016lx, 0x%016lx] 0x%016lx\n",
+		       p2m_convert_min_pfn, p2m_convert_max_pfn, p2m_size);
+		goto out;
+	}
+
+	p2m_assign_start_pfn = p2m_resource.start >> PAGE_SHIFT;
+	p2m_assign_end_pfn = p2m_resource.end >> PAGE_SHIFT;
+	
+	error = HYPERVISOR_expose_p2m(p2m_convert_min_pfn,
+				      p2m_assign_start_pfn,
+				      p2m_expose_size, p2m_granule_pfn);
+	if (error) {
+		printk(KERN_ERR P2M_PREFIX "failed expose p2m hypercall %d\n",
+		       error);
+		printk(KERN_ERR P2M_PREFIX "conv 0x%016lx assign 0x%016lx "
+		       "expose_size 0x%016lx granule 0x%016lx\n",
+		       p2m_convert_min_pfn, p2m_assign_start_pfn,
+		       p2m_expose_size, p2m_granule_pfn);;
+		release_resource(&p2m_resource);
+		goto out;
+	}
+	p2m_pte = (volatile const pte_t*)pfn_to_kaddr(p2m_assign_start_pfn);
+#ifdef CONFIG_XEN_IA64_EXPOSE_P2M_USE_DTR
+	if (xen_ia64_p2m_expose_use_dtr) {
+		p2m_itr_arg.vaddr = (unsigned long)__va(p2m_assign_start_pfn
+							<< PAGE_SHIFT);
+		p2m_itr_arg.pteval = pte_val(pfn_pte(p2m_assign_start_pfn,
+						     PAGE_KERNEL));
+		p2m_itr_arg.log_page_size = log_page_size;
+		smp_mb();
+		smp_call_function(&p2m_itr, &p2m_itr_arg, 1, 1);
+		p2m_itr(&p2m_itr_arg);
+	}
+#endif	
+	smp_mb();
+	p2m_initialized = 1;
+	printk(P2M_PREFIX "assign p2m table of [0x%016lx, 0x%016lx)\n",
+	       p2m_convert_min_pfn << PAGE_SHIFT,
+	       (p2m_convert_max_pfn << PAGE_SHIFT) + PAGE_SIZE);
+	printk(P2M_PREFIX "to [0x%016lx, 0x%016lx) (%ld KBytes)\n",
+	       p2m_assign_start_pfn << PAGE_SHIFT,
+	       (p2m_assign_end_pfn << PAGE_SHIFT) + PAGE_SIZE,
+	       p2m_size / 1024);
+out:
+	unlock_cpu_hotplug();
+	return error;
+}
+
+#ifdef notyet
+void
+p2m_expose_cleanup(void)
+{
+	BUG_ON(!p2m_initialized);
+#ifdef CONFIG_XEN_IA64_EXPOSE_P2M_USE_DTR
+	unregister_cpu_notifier(&p2m_expose_dtr_hotplug_notifier);
+#endif
+	release_resource(&p2m_resource);
+}
+#endif
+
+static void
+p2m_expose_resume(void)
+{
+	int error;
+
+	if (!xen_ia64_p2m_expose || !p2m_initialized)
+		return;
+
+	/*
+	 * We can't call {lock, unlock}_cpu_hotplug() because
+	 * they require process context.
+	 * We don't need them because we're the only one cpu and
+	 * interrupts are masked when resume.
+	 */
+	error = HYPERVISOR_expose_p2m(p2m_convert_min_pfn,
+				      p2m_assign_start_pfn,
+				      p2m_expose_size, p2m_granule_pfn);
+	if (error) {
+		printk(KERN_ERR P2M_PREFIX "failed expose p2m hypercall %d\n",
+		       error);
+		printk(KERN_ERR P2M_PREFIX "conv 0x%016lx assign 0x%016lx "
+		       "expose_size 0x%016lx granule 0x%016lx\n",
+		       p2m_convert_min_pfn, p2m_assign_start_pfn,
+		       p2m_expose_size, p2m_granule_pfn);;
+		p2m_initialized = 0;
+		smp_mb();
+		ia64_ptr(0x2, p2m_itr_arg.vaddr, p2m_itr_arg.log_page_size);
+		
+		/*
+		 * We can't call those clean up functions because they
+		 * require process context.
+		 */
+#if 0
+#ifdef CONFIG_XEN_IA64_EXPOSE_P2M_USE_DTR
+		if (xen_ia64_p2m_expose_use_dtr)
+			unregister_cpu_notifier(
+				&p2m_expose_dtr_hotplug_notifier);
+#endif
+		release_resource(&p2m_resource);
+#endif
+	}
+}
+
+/* XXX inlinize? */
+unsigned long
+p2m_phystomach(unsigned long gpfn)
+{
+	volatile const pte_t *pte;
+	unsigned long mfn;
+	unsigned long pteval;
+	
+	if (!p2m_initialized ||
+	    gpfn < p2m_min_low_pfn || gpfn > p2m_max_low_pfn
+	    /* || !pfn_valid(gpfn) */)
+		return INVALID_MFN;
+	pte = p2m_pte + (gpfn - p2m_convert_min_pfn);
+
+	mfn = INVALID_MFN;
+	if (likely(__get_user(pteval, (unsigned long __user *)pte) == 0 &&
+		   pte_present(__pte(pteval)) &&
+		   pte_pfn(__pte(pteval)) != (INVALID_MFN >> PAGE_SHIFT)))
+		mfn = (pteval & _PFN_MASK) >> PAGE_SHIFT;
+
+	return mfn;
+}
+
+EXPORT_SYMBOL_GPL(p2m_initialized);
+EXPORT_SYMBOL_GPL(p2m_min_low_pfn);
+EXPORT_SYMBOL_GPL(p2m_max_low_pfn);
+EXPORT_SYMBOL_GPL(p2m_convert_min_pfn);
+EXPORT_SYMBOL_GPL(p2m_convert_max_pfn);
+EXPORT_SYMBOL_GPL(p2m_pte);
+EXPORT_SYMBOL_GPL(p2m_phystomach);
+
+/**************************************************************************
+ * foreign domain p2m mapping
+ */
+#include <asm/xen/xencomm.h>
+#include <xen/public/privcmd.h>
+
+struct foreign_p2m_private {
+	unsigned long	gpfn;
+	domid_t		domid;
+};
+
+static void
+xen_foreign_p2m_unexpose(struct xen_ia64_privcmd_range *privcmd_range,
+			 void *arg)
+{
+	struct foreign_p2m_private *private = (struct foreign_p2m_private*)arg;
+	int ret;
+
+	privcmd_range->private = NULL;
+	privcmd_range->callback = NULL;
+
+	ret = HYPERVISOR_unexpose_foreign_p2m(private->gpfn, private->domid);
+	if (ret)
+		printk(KERN_WARNING
+		       "unexpose_foreign_p2m hypercall failed.\n");
+	kfree(private);
+}
+
+int
+xen_foreign_p2m_expose(privcmd_hypercall_t *hypercall)
+{
+	/*
+	 * hypercall->
+	 * arg0: cmd = IA64_DOM0VP_expose_foreign_p2m
+	 * arg1: va
+	 * arg2: domid
+	 * arg3: __user* memmap_info
+	 * arg4: flags
+	 */
+
+	int ret = 0;
+	struct mm_struct *mm = current->mm;
+
+	unsigned long vaddr = hypercall->arg[1];
+	domid_t domid = hypercall->arg[2];
+	struct xen_ia64_memmap_info __user *u_memmap_info =
+		(struct xen_ia64_memmap_info __user *)hypercall->arg[3];
+
+	struct xen_ia64_memmap_info memmap_info;
+	size_t memmap_size;
+	struct xen_ia64_memmap_info *k_memmap_info = NULL;
+	unsigned long max_gpfn;
+	unsigned long p2m_size;
+	struct resource *res;
+	unsigned long gpfn;
+
+	struct vm_area_struct *vma;
+	void *p;
+	unsigned long prev_src_gpfn_end;
+
+	struct xen_ia64_privcmd_vma *privcmd_vma;
+	struct xen_ia64_privcmd_range *privcmd_range;
+	struct foreign_p2m_private *private = NULL;
+
+	BUG_ON(hypercall->arg[0] != IA64_DOM0VP_expose_foreign_p2m);
+
+	private = kmalloc(sizeof(*private), GFP_KERNEL);
+	if (private == NULL)
+		goto kfree_out;
+
+	if (copy_from_user(&memmap_info, u_memmap_info, sizeof(memmap_info)))
+		return -EFAULT;
+	/* memmap_info integrity check */
+	if (memmap_info.efi_memdesc_size < sizeof(efi_memory_desc_t) ||
+	    memmap_info.efi_memmap_size < memmap_info.efi_memdesc_size ||
+	    (memmap_info.efi_memmap_size % memmap_info.efi_memdesc_size)
+	    != 0) {
+		ret = -EINVAL;
+		goto kfree_out;
+	}
+
+	memmap_size = sizeof(*k_memmap_info) + memmap_info.efi_memmap_size;
+	k_memmap_info = kmalloc(memmap_size, GFP_KERNEL);
+	if (k_memmap_info == NULL)
+		return -ENOMEM;
+	if (copy_from_user(k_memmap_info, u_memmap_info, memmap_size)) {
+		ret = -EFAULT;
+		goto kfree_out;
+	}
+	/* k_memmap_info integrity check is done by the expose foreng p2m 
+	   hypercall */
+
+	max_gpfn = HYPERVISOR_memory_op(XENMEM_maximum_gpfn, &domid);
+	if (max_gpfn < 0) {
+		ret = max_gpfn;
+		goto kfree_out;
+	}
+	p2m_size = p2m_table_size(max_gpfn + 1);
+
+	down_write(&mm->mmap_sem);
+
+	vma = find_vma(mm, vaddr);
+	if (vma == NULL || vma->vm_ops != &xen_ia64_privcmd_vm_ops ||
+	    vaddr != vma->vm_start ||
+	    (vma->vm_flags & VM_WRITE) || (vma->vm_flags & VM_EXEC) ||
+	    !privcmd_enforce_singleshot_mapping(vma))
+		goto mmap_out;
+
+	privcmd_vma = (struct xen_ia64_privcmd_vma*)vma->vm_private_data;
+	res = privcmd_vma->range->res;
+	if (p2m_size > (res->end - res->start + 1) ||
+	    p2m_size > vma->vm_end - vma->vm_start) {
+		ret = -EINVAL;
+		goto mmap_out;
+	}
+	
+	gpfn = res->start >> PAGE_SHIFT;
+	/*
+	 * arg0: dest_gpfn
+	 * arg1: domid
+	 * arg2: XEN_GUEST_HANDLE(char) buffer: memmap_info
+	 * arg3: flags
+	 * The hypercall checks its intergirty/simplfies it and 
+	 * copy it back for us.
+	 */
+	ret = xencomm_arch_expose_foreign_p2m(gpfn, domid,
+	      xencomm_map_no_alloc(k_memmap_info, memmap_size),
+	      hypercall->arg[4]);
+	if (ret)
+		goto mmap_out;
+
+	privcmd_range = (struct xen_ia64_privcmd_range*)privcmd_vma->range;
+	prev_src_gpfn_end = 0;
+	for (p = k_memmap_info->memdesc;
+	     p < (void*)&k_memmap_info->memdesc[0] +
+		     k_memmap_info->efi_memmap_size;
+	     p += k_memmap_info->efi_memdesc_size) {
+		efi_memory_desc_t* md = p;
+		unsigned long src_gpfn = md->phys_addr >> PAGE_SHIFT;
+		unsigned long src_gpfn_end =
+			(md->phys_addr + (md->num_pages << EFI_PAGE_SHIFT)) >>
+			PAGE_SHIFT;
+		unsigned long num_src_gpfn;
+		unsigned long gpfn_offset;
+		unsigned long size;
+		unsigned int i;
+
+		if (src_gpfn <= prev_src_gpfn_end)
+			src_gpfn = prev_src_gpfn_end + 1;
+		if (src_gpfn_end <= prev_src_gpfn_end)
+			continue;
+
+		src_gpfn &= ~(PTRS_PER_PTE - 1);
+		src_gpfn_end = (src_gpfn_end + PTRS_PER_PTE - 1) &
+			~(PTRS_PER_PTE - 1);
+		num_src_gpfn = src_gpfn_end - src_gpfn;
+		gpfn_offset = src_gpfn / PTRS_PER_PTE;
+		size = p2m_table_size(num_src_gpfn);
+
+		prev_src_gpfn_end = src_gpfn_end;
+		ret = remap_pfn_range(vma,
+				      vaddr + (gpfn_offset << PAGE_SHIFT), 
+				      gpfn + gpfn_offset, size,
+				      vma->vm_page_prot);
+		if (ret) {
+			for (i = 0; i < gpfn + gpfn_offset; i++) {
+				struct xen_ia64_privcmd_entry *entry =
+					&privcmd_range->entries[i];
+				BUG_ON(atomic_read(&entry->map_count) != 1 &&
+				       atomic_read(&entry->map_count) != 0);
+				atomic_set(&entry->map_count, 0);
+				entry->gpfn = INVALID_GPFN;
+			}
+			(void)HYPERVISOR_unexpose_foreign_p2m(gpfn, domid);
+			goto mmap_out;
+		}
+
+		for (i = gpfn_offset;
+		     i < gpfn_offset + (size >> PAGE_SHIFT);
+		     i++) {
+			struct xen_ia64_privcmd_entry *entry =
+				&privcmd_range->entries[i];
+			BUG_ON(atomic_read(&entry->map_count) != 0);
+			BUG_ON(entry->gpfn != INVALID_GPFN);
+			atomic_inc(&entry->map_count);
+			entry->gpfn = gpfn + i;
+		}
+	}
+
+	private->gpfn = gpfn;
+	private->domid = domid;
+
+	privcmd_range->callback = &xen_foreign_p2m_unexpose;
+	privcmd_range->private = private;
+
+mmap_out:
+	up_write(&mm->mmap_sem);
+kfree_out:
+	kfree(k_memmap_info);
+	if (ret != 0)
+		kfree(private);
+	return ret;
+}
+#endif
+
+/**************************************************************************
+ * for xenoprof
+ */
+struct resource*
+xen_ia64_allocate_resource(unsigned long size)
+{
+	struct resource *res;
+	int error;
+	
+	res = kzalloc(sizeof(*res), GFP_KERNEL);
+	if (res == NULL)
+		return ERR_PTR(-ENOMEM);
+
+	res->name = "Xen";
+	res->flags = IORESOURCE_MEM;
+	error = allocate_resource(&iomem_resource, res, PAGE_ALIGN(size),
+				  privcmd_resource_min, privcmd_resource_max,
+				  IA64_GRANULE_SIZE, NULL, NULL);
+	if (error) {
+		kfree(res);
+		return ERR_PTR(error);
+	}
+	return res;
+}
+EXPORT_SYMBOL_GPL(xen_ia64_allocate_resource);
+
+void
+xen_ia64_release_resource(struct resource *res)
+{
+	release_resource(res);
+	kfree(res);
+}
+EXPORT_SYMBOL_GPL(xen_ia64_release_resource);
+
+void
+xen_ia64_unmap_resource(struct resource *res)
+{
+	unsigned long gpfn = res->start >> PAGE_SHIFT;
+	unsigned long nr_pages = (res->end - res->start) >> PAGE_SHIFT;
+	unsigned long i;
+	
+	for (i = 0; i < nr_pages; i++) {
+		int error = HYPERVISOR_zap_physmap(gpfn + i, 0);
+		if (error)
+			printk(KERN_ERR
+			       "%s:%d zap_phsymap failed %d gpfn %lx\n",
+			       __func__, __LINE__, error, gpfn + i);
+	}
+	xen_ia64_release_resource(res);
+}
+EXPORT_SYMBOL_GPL(xen_ia64_unmap_resource);
+
+/**************************************************************************
+ * opt feature
+ */
+void
+xen_ia64_enable_opt_feature(void)
+{
+	/* Enable region 7 identity map optimizations in Xen */
+	struct xen_ia64_opt_feature optf;
+
+	optf.cmd = XEN_IA64_OPTF_IDENT_MAP_REG7;
+	optf.on = XEN_IA64_OPTF_ON;
+	optf.pgprot = pgprot_val(PAGE_KERNEL);
+	optf.key = 0;	/* No key on linux. */
+	HYPERVISOR_opt_feature(&optf);
+}
+
+/**************************************************************************
+ * suspend/resume
+ */
+void
+xen_post_suspend(int suspend_cancelled)
+{
+	if (suspend_cancelled)
+		return;
+	
+	p2m_expose_resume();
+	xen_ia64_enable_opt_feature();
+	/* add more if necessary */
+}
diff -rpuN linux-2.6.18.8/arch/ia64/xen/machvec.c linux-2.6.18-xen-3.3.0/arch/ia64/xen/machvec.c
--- linux-2.6.18.8/arch/ia64/xen/machvec.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/xen/machvec.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,4 @@
+#define MACHVEC_PLATFORM_NAME           xen
+#define MACHVEC_PLATFORM_HEADER         <asm/machvec_xen.h>
+#include <asm/machvec_init.h>
+
diff -rpuN linux-2.6.18.8/arch/ia64/xen/Makefile linux-2.6.18-xen-3.3.0/arch/ia64/xen/Makefile
--- linux-2.6.18.8/arch/ia64/xen/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/xen/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,9 @@
+#
+# Makefile for Xen components
+#
+
+obj-y := hypercall.o xenivt.o xenentry.o xensetup.o xenpal.o \
+	 hypervisor.o util.o xencomm.o xcom_hcall.o \
+	 xcom_privcmd.o xcom_asm.o xen_dma.o
+
+obj-$(CONFIG_IA64_GENERIC) += machvec.o
diff -rpuN linux-2.6.18.8/arch/ia64/xen/swiotlb.c linux-2.6.18-xen-3.3.0/arch/ia64/xen/swiotlb.c
--- linux-2.6.18.8/arch/ia64/xen/swiotlb.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/xen/swiotlb.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,906 @@
+/*
+ * Dynamic DMA mapping support.
+ *
+ * This implementation is for IA-64 and EM64T platforms that do not support
+ * I/O TLBs (aka DMA address translation hardware).
+ * Copyright (C) 2000 Asit Mallick <Asit.K.Mallick@intel.com>
+ * Copyright (C) 2000 Goutham Rao <goutham.rao@intel.com>
+ * Copyright (C) 2000, 2003 Hewlett-Packard Co
+ *	David Mosberger-Tang <davidm@hpl.hp.com>
+ *
+ * 03/05/07 davidm	Switch from PCI-DMA to generic device DMA API.
+ * 00/12/13 davidm	Rename to swiotlb.c and add mark_clean() to avoid
+ *			unnecessary i-cache flushing.
+ * 04/07/.. ak		Better overflow handling. Assorted fixes.
+ * 05/09/10 linville	Add support for syncing ranges, support syncing for
+ *			DMA_BIDIRECTIONAL mappings, miscellaneous cleanup.
+ */
+
+#include <linux/cache.h>
+#include <linux/dma-mapping.h>
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/spinlock.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/ctype.h>
+
+#include <asm/io.h>
+#include <asm/dma.h>
+#include <asm/scatterlist.h>
+
+#include <linux/init.h>
+#include <linux/bootmem.h>
+
+#ifdef CONFIG_XEN
+#include <xen/gnttab.h>
+#include <asm/gnttab_dma.h>
+/*
+ * What DMA mask should Xen use to remap the bounce buffer pool?  Most
+ * reports seem to indicate 30 bits is sufficient, except maybe for old
+ * sound cards that we probably don't care about anyway.  If we need to,
+ * we could put in some smarts to try to lower, but hopefully it's not
+ * necessary.
+ */
+#define DMA_BITS	(30)
+#endif
+
+#define OFFSET(val,align) ((unsigned long)	\
+	                   ( (val) & ( (align) - 1)))
+
+#define SG_ENT_VIRT_ADDRESS(sg)	(page_address((sg)->page) + (sg)->offset)
+#define SG_ENT_PHYS_ADDRESS(SG)	virt_to_bus(SG_ENT_VIRT_ADDRESS(SG))
+
+/*
+ * Maximum allowable number of contiguous slabs to map,
+ * must be a power of 2.  What is the appropriate value ?
+ * The complexity of {map,unmap}_single is linearly dependent on this value.
+ */
+#define IO_TLB_SEGSIZE	128
+
+/*
+ * log of the size of each IO TLB slab.  The number of slabs is command line
+ * controllable.
+ */
+#define IO_TLB_SHIFT 11
+
+#define SLABS_PER_PAGE (1 << (PAGE_SHIFT - IO_TLB_SHIFT))
+
+/*
+ * Minimum IO TLB size to bother booting with.  Systems with mainly
+ * 64bit capable cards will only lightly use the swiotlb.  If we can't
+ * allocate a contiguous 1MB, we're probably in trouble anyway.
+ */
+#define IO_TLB_MIN_SLABS ((1<<20) >> IO_TLB_SHIFT)
+
+/*
+ * Enumeration for sync targets
+ */
+enum dma_sync_target {
+	SYNC_FOR_CPU = 0,
+	SYNC_FOR_DEVICE = 1,
+};
+
+int swiotlb_force;
+
+/*
+ * Used to do a quick range check in swiotlb_unmap_single and
+ * swiotlb_sync_single_*, to see if the memory was in fact allocated by this
+ * API.
+ */
+static char *io_tlb_start, *io_tlb_end;
+
+/*
+ * The number of IO TLB blocks (in groups of 64) betweeen io_tlb_start and
+ * io_tlb_end.  This is command line adjustable via setup_io_tlb_npages.
+ */
+static unsigned long io_tlb_nslabs;
+
+/*
+ * When the IOMMU overflows we return a fallback buffer. This sets the size.
+ */
+static unsigned long io_tlb_overflow = 32*1024;
+
+void *io_tlb_overflow_buffer;
+
+/*
+ * This is a free list describing the number of free entries available from
+ * each index
+ */
+static unsigned int *io_tlb_list;
+static unsigned int io_tlb_index;
+
+/*
+ * We need to save away the original address corresponding to a mapped entry
+ * for the sync operations.
+ */
+static unsigned char **io_tlb_orig_addr;
+
+/*
+ * Protect the above data structures in the map and unmap calls
+ */
+static DEFINE_SPINLOCK(io_tlb_lock);
+
+static int __init
+setup_io_tlb_npages(char *str)
+{
+	if (isdigit(*str)) {
+		io_tlb_nslabs = simple_strtoul(str, &str, 0);
+		/* avoid tail segment of size < IO_TLB_SEGSIZE */
+		io_tlb_nslabs = ALIGN(io_tlb_nslabs, IO_TLB_SEGSIZE);
+	}
+	if (*str == ',')
+		++str;
+	if (!strcmp(str, "force"))
+		swiotlb_force = 1;
+	return 1;
+}
+__setup("swiotlb=", setup_io_tlb_npages);
+/* make io_tlb_overflow tunable too? */
+
+/*
+ * Statically reserve bounce buffer space and initialize bounce buffer data
+ * structures for the software IO TLB used to implement the DMA API.
+ */
+void
+swiotlb_init_with_default_size (size_t default_size)
+{
+	unsigned long i;
+
+	if (!io_tlb_nslabs) {
+		io_tlb_nslabs = (default_size >> IO_TLB_SHIFT);
+		io_tlb_nslabs = ALIGN(io_tlb_nslabs, IO_TLB_SEGSIZE);
+	}
+
+#ifdef CONFIG_XEN
+	if (is_running_on_xen())
+		io_tlb_nslabs = roundup_pow_of_two(io_tlb_nslabs);
+#endif
+	/*
+	 * Get IO TLB memory from the low pages
+	 */
+	io_tlb_start = alloc_bootmem_low_pages(io_tlb_nslabs * (1 << IO_TLB_SHIFT));
+	if (!io_tlb_start)
+		panic("Cannot allocate SWIOTLB buffer");
+	io_tlb_end = io_tlb_start + io_tlb_nslabs * (1 << IO_TLB_SHIFT);
+
+#ifdef CONFIG_XEN
+	for (i = 0 ; i < io_tlb_nslabs ; i += IO_TLB_SEGSIZE) {
+		if (xen_create_contiguous_region(
+				(unsigned long)io_tlb_start +
+				(i << IO_TLB_SHIFT),
+				get_order(IO_TLB_SEGSIZE << IO_TLB_SHIFT),
+				DMA_BITS))
+			panic("Failed to setup Xen contiguous region");
+	}
+#endif
+
+	/*
+	 * Allocate and initialize the free list array.  This array is used
+	 * to find contiguous free memory regions of size up to IO_TLB_SEGSIZE
+	 * between io_tlb_start and io_tlb_end.
+	 */
+	io_tlb_list = alloc_bootmem(io_tlb_nslabs * sizeof(int));
+	for (i = 0; i < io_tlb_nslabs; i++)
+ 		io_tlb_list[i] = IO_TLB_SEGSIZE - OFFSET(i, IO_TLB_SEGSIZE);
+	io_tlb_index = 0;
+	io_tlb_orig_addr = alloc_bootmem(io_tlb_nslabs * sizeof(char *));
+
+	/*
+	 * Get the overflow emergency buffer
+	 */
+	io_tlb_overflow_buffer = alloc_bootmem_low(io_tlb_overflow);
+#ifdef CONFIG_XEN
+	if (xen_create_contiguous_region((unsigned long)io_tlb_overflow_buffer,
+					 get_order(io_tlb_overflow), DMA_BITS))
+		panic("Failed to setup Xen contiguous region for overflow");
+#endif
+	printk(KERN_INFO "Placing software IO TLB between 0x%lx - 0x%lx\n",
+	       virt_to_phys(io_tlb_start), virt_to_phys(io_tlb_end));
+}
+
+void
+swiotlb_init (void)
+{
+	swiotlb_init_with_default_size(64 * (1<<20));	/* default to 64MB */
+}
+
+/*
+ * Systems with larger DMA zones (those that don't support ISA) can
+ * initialize the swiotlb later using the slab allocator if needed.
+ * This should be just like above, but with some error catching.
+ */
+int
+swiotlb_late_init_with_default_size (size_t default_size)
+{
+	unsigned long i, req_nslabs = io_tlb_nslabs;
+	unsigned int order;
+
+	if (!io_tlb_nslabs) {
+		io_tlb_nslabs = (default_size >> IO_TLB_SHIFT);
+		io_tlb_nslabs = ALIGN(io_tlb_nslabs, IO_TLB_SEGSIZE);
+	}
+
+#ifdef CONFIG_XEN
+	if (is_running_on_xen())
+		io_tlb_nslabs = roundup_pow_of_two(io_tlb_nslabs);
+#endif
+	/*
+	 * Get IO TLB memory from the low pages
+	 */
+	order = get_order(io_tlb_nslabs * (1 << IO_TLB_SHIFT));
+	io_tlb_nslabs = SLABS_PER_PAGE << order;
+
+	while ((SLABS_PER_PAGE << order) > IO_TLB_MIN_SLABS) {
+		io_tlb_start = (char *)__get_free_pages(GFP_DMA | __GFP_NOWARN,
+		                                        order);
+		if (io_tlb_start)
+			break;
+		order--;
+	}
+
+	if (!io_tlb_start)
+		goto cleanup1;
+
+	if (order != get_order(io_tlb_nslabs * (1 << IO_TLB_SHIFT))) {
+		printk(KERN_WARNING "Warning: only able to allocate %ld MB "
+		       "for software IO TLB\n", (PAGE_SIZE << order) >> 20);
+		io_tlb_nslabs = SLABS_PER_PAGE << order;
+	}
+	io_tlb_end = io_tlb_start + io_tlb_nslabs * (1 << IO_TLB_SHIFT);
+	memset(io_tlb_start, 0, io_tlb_nslabs * (1 << IO_TLB_SHIFT));
+
+#ifdef CONFIG_XEN
+	for (i = 0 ; i < io_tlb_nslabs ; i += IO_TLB_SEGSIZE) {
+		if (xen_create_contiguous_region(
+				(unsigned long)io_tlb_start +
+				(i << IO_TLB_SHIFT),
+				get_order(IO_TLB_SEGSIZE << IO_TLB_SHIFT),
+				DMA_BITS))
+			panic("Failed to setup Xen contiguous region");
+	}
+#endif
+	/*
+	 * Allocate and initialize the free list array.  This array is used
+	 * to find contiguous free memory regions of size up to IO_TLB_SEGSIZE
+	 * between io_tlb_start and io_tlb_end.
+	 */
+	io_tlb_list = (unsigned int *)__get_free_pages(GFP_KERNEL,
+	                              get_order(io_tlb_nslabs * sizeof(int)));
+	if (!io_tlb_list)
+		goto cleanup2;
+
+	for (i = 0; i < io_tlb_nslabs; i++)
+ 		io_tlb_list[i] = IO_TLB_SEGSIZE - OFFSET(i, IO_TLB_SEGSIZE);
+	io_tlb_index = 0;
+
+	io_tlb_orig_addr = (unsigned char **)__get_free_pages(GFP_KERNEL,
+	                           get_order(io_tlb_nslabs * sizeof(char *)));
+	if (!io_tlb_orig_addr)
+		goto cleanup3;
+
+	memset(io_tlb_orig_addr, 0, io_tlb_nslabs * sizeof(char *));
+
+	/*
+	 * Get the overflow emergency buffer
+	 */
+	io_tlb_overflow_buffer = (void *)__get_free_pages(GFP_DMA,
+	                                          get_order(io_tlb_overflow));
+	if (!io_tlb_overflow_buffer)
+		goto cleanup4;
+
+#ifdef CONFIG_XEN
+	if (xen_create_contiguous_region((unsigned long)io_tlb_overflow_buffer,
+					 get_order(io_tlb_overflow), DMA_BITS))
+		panic("Failed to setup Xen contiguous region for overflow");
+#endif
+	printk(KERN_INFO "Placing %ldMB software IO TLB between 0x%lx - "
+	       "0x%lx\n", (io_tlb_nslabs * (1 << IO_TLB_SHIFT)) >> 20,
+	       virt_to_phys(io_tlb_start), virt_to_phys(io_tlb_end));
+
+	return 0;
+
+cleanup4:
+	free_pages((unsigned long)io_tlb_orig_addr, get_order(io_tlb_nslabs *
+	                                                      sizeof(char *)));
+	io_tlb_orig_addr = NULL;
+cleanup3:
+	free_pages((unsigned long)io_tlb_list, get_order(io_tlb_nslabs *
+	                                                 sizeof(int)));
+	io_tlb_list = NULL;
+	io_tlb_end = NULL;
+cleanup2:
+	free_pages((unsigned long)io_tlb_start, order);
+	io_tlb_start = NULL;
+cleanup1:
+	io_tlb_nslabs = req_nslabs;
+	return -ENOMEM;
+}
+
+static inline int
+address_needs_mapping(struct device *hwdev, dma_addr_t addr)
+{
+	dma_addr_t mask = 0xffffffff;
+	/* If the device has a mask, use it, otherwise default to 32 bits */
+	if (hwdev && hwdev->dma_mask)
+		mask = *hwdev->dma_mask;
+	return (addr & ~mask) != 0;
+}
+
+/*
+ * Allocates bounce buffer and returns its kernel virtual address.
+ */
+static void *
+map_single(struct device *hwdev, char *buffer, size_t size, int dir)
+{
+	unsigned long flags;
+	char *dma_addr;
+	unsigned int nslots, stride, index, wrap;
+	char *slot_buf;
+	int i;
+
+	/*
+	 * For mappings greater than a page, we limit the stride (and
+	 * hence alignment) to a page size.
+	 */
+	nslots = ALIGN(size, 1 << IO_TLB_SHIFT) >> IO_TLB_SHIFT;
+	if (size > PAGE_SIZE)
+		stride = (1 << (PAGE_SHIFT - IO_TLB_SHIFT));
+	else
+		stride = 1;
+
+	BUG_ON(!nslots);
+
+	/*
+	 * Find suitable number of IO TLB entries size that will fit this
+	 * request and allocate a buffer from that IO TLB pool.
+	 */
+	spin_lock_irqsave(&io_tlb_lock, flags);
+	{
+		wrap = index = ALIGN(io_tlb_index, stride);
+
+		if (index >= io_tlb_nslabs)
+			wrap = index = 0;
+
+		do {
+			/*
+			 * If we find a slot that indicates we have 'nslots'
+			 * number of contiguous buffers, we allocate the
+			 * buffers from that slot and mark the entries as '0'
+			 * indicating unavailable.
+			 */
+			if (io_tlb_list[index] >= nslots) {
+				int count = 0;
+
+				for (i = index; i < (int) (index + nslots); i++)
+					io_tlb_list[i] = 0;
+				for (i = index - 1; (OFFSET(i, IO_TLB_SEGSIZE) != IO_TLB_SEGSIZE -1) && io_tlb_list[i]; i--)
+					io_tlb_list[i] = ++count;
+				dma_addr = io_tlb_start + (index << IO_TLB_SHIFT);
+
+				/*
+				 * Update the indices to avoid searching in
+				 * the next round.
+				 */
+				io_tlb_index = ((index + nslots) < io_tlb_nslabs
+						? (index + nslots) : 0);
+
+				goto found;
+			}
+			index += stride;
+			if (index >= io_tlb_nslabs)
+				index = 0;
+		} while (index != wrap);
+
+		spin_unlock_irqrestore(&io_tlb_lock, flags);
+		return NULL;
+	}
+  found:
+	spin_unlock_irqrestore(&io_tlb_lock, flags);
+
+	/*
+	 * Save away the mapping from the original address to the DMA address.
+	 * This is needed when we sync the memory.  Then we sync the buffer if
+	 * needed.
+	 */
+	slot_buf = buffer;
+	for (i = 0; i < nslots; i++) {
+		io_tlb_orig_addr[index + i] = slot_buf;
+		slot_buf += 1 << IO_TLB_SHIFT;
+	}
+	if (dir == DMA_TO_DEVICE || dir == DMA_BIDIRECTIONAL)
+		memcpy(dma_addr, buffer, size);
+
+	return dma_addr;
+}
+
+/*
+ * dma_addr is the kernel virtual address of the bounce buffer to unmap.
+ */
+static void
+unmap_single(struct device *hwdev, char *dma_addr, size_t size, int dir)
+{
+	unsigned long flags;
+	int i, count, nslots = ALIGN(size, 1 << IO_TLB_SHIFT) >> IO_TLB_SHIFT;
+	int index = (dma_addr - io_tlb_start) >> IO_TLB_SHIFT;
+	char *buffer = io_tlb_orig_addr[index];
+
+	/*
+	 * First, sync the memory before unmapping the entry
+	 */
+	if (buffer && ((dir == DMA_FROM_DEVICE) || (dir == DMA_BIDIRECTIONAL)))
+		/*
+		 * bounce... copy the data back into the original buffer * and
+		 * delete the bounce buffer.
+		 */
+		memcpy(buffer, dma_addr, size);
+
+	/*
+	 * Return the buffer to the free list by setting the corresponding
+	 * entries to indicate the number of contigous entries available.
+	 * While returning the entries to the free list, we merge the entries
+	 * with slots below and above the pool being returned.
+	 */
+	spin_lock_irqsave(&io_tlb_lock, flags);
+	{
+		count = ((index + nslots) < ALIGN(index + 1, IO_TLB_SEGSIZE) ?
+			 io_tlb_list[index + nslots] : 0);
+		/*
+		 * Step 1: return the slots to the free list, merging the
+		 * slots with superceeding slots
+		 */
+		for (i = index + nslots - 1; i >= index; i--)
+			io_tlb_list[i] = ++count;
+		/*
+		 * Step 2: merge the returned slots with the preceding slots,
+		 * if available (non zero)
+		 */
+		for (i = index - 1; (OFFSET(i, IO_TLB_SEGSIZE) != IO_TLB_SEGSIZE -1) && io_tlb_list[i]; i--)
+			io_tlb_list[i] = ++count;
+	}
+	spin_unlock_irqrestore(&io_tlb_lock, flags);
+}
+
+static void
+sync_single(struct device *hwdev, char *dma_addr, size_t size,
+	    int dir, int target)
+{
+	int index = (dma_addr - io_tlb_start) >> IO_TLB_SHIFT;
+	char *buffer = io_tlb_orig_addr[index];
+
+	switch (target) {
+	case SYNC_FOR_CPU:
+		if (likely(dir == DMA_FROM_DEVICE || dir == DMA_BIDIRECTIONAL))
+			memcpy(buffer, dma_addr, size);
+		else
+			BUG_ON(dir != DMA_TO_DEVICE);
+		break;
+	case SYNC_FOR_DEVICE:
+		if (likely(dir == DMA_TO_DEVICE || dir == DMA_BIDIRECTIONAL))
+			memcpy(dma_addr, buffer, size);
+		else
+			BUG_ON(dir != DMA_FROM_DEVICE);
+		break;
+	default:
+		BUG();
+	}
+}
+
+void *
+swiotlb_alloc_coherent(struct device *hwdev, size_t size,
+		       dma_addr_t *dma_handle, gfp_t flags)
+{
+	unsigned long dev_addr;
+	void *ret;
+	int order = get_order(size);
+
+	/*
+	 * XXX fix me: the DMA API should pass us an explicit DMA mask
+	 * instead, or use ZONE_DMA32 (ia64 overloads ZONE_DMA to be a ~32
+	 * bit range instead of a 16MB one).
+	 */
+	flags |= GFP_DMA;
+
+	ret = (void *)__get_free_pages(flags, order);
+#ifdef CONFIG_XEN
+	if (ret && is_running_on_xen()) {
+		if (xen_create_contiguous_region((unsigned long)ret, order,
+					fls64(hwdev->coherent_dma_mask))) {
+			free_pages((unsigned long)ret, order);
+			ret = NULL;
+		} else {
+			/*
+			 * Short circuit the rest, xen_create_contiguous_region
+			 * should fail if it didn't give us an address within
+			 * the mask requested.  
+			 */
+			memset(ret, 0, size);
+			*dma_handle = virt_to_bus(ret);
+			return ret;
+		}
+	}
+#endif
+	if (ret && address_needs_mapping(hwdev, virt_to_bus(ret))) {
+		/*
+		 * The allocated memory isn't reachable by the device.
+		 * Fall back on swiotlb_map_single().
+		 */
+		free_pages((unsigned long) ret, order);
+		ret = NULL;
+	}
+	if (!ret) {
+		/*
+		 * We are either out of memory or the device can't DMA
+		 * to GFP_DMA memory; fall back on
+		 * swiotlb_map_single(), which will grab memory from
+		 * the lowest available address range.
+		 */
+		dma_addr_t handle;
+		handle = swiotlb_map_single(NULL, NULL, size, DMA_FROM_DEVICE);
+		if (swiotlb_dma_mapping_error(handle))
+			return NULL;
+
+		ret = bus_to_virt(handle);
+	}
+
+	memset(ret, 0, size);
+	dev_addr = virt_to_bus(ret);
+
+	/* Confirm address can be DMA'd by device */
+	if (address_needs_mapping(hwdev, dev_addr)) {
+		printk("hwdev DMA mask = 0x%016Lx, dev_addr = 0x%016lx\n",
+		       (unsigned long long)*hwdev->dma_mask, dev_addr);
+		panic("swiotlb_alloc_coherent: allocated memory is out of "
+		      "range for device");
+	}
+	*dma_handle = dev_addr;
+	return ret;
+}
+
+void
+swiotlb_free_coherent(struct device *hwdev, size_t size, void *vaddr,
+		      dma_addr_t dma_handle)
+{
+	if (!(vaddr >= (void *)io_tlb_start
+                    && vaddr < (void *)io_tlb_end)) {
+#ifdef CONFIG_XEN
+		xen_destroy_contiguous_region((unsigned long)vaddr,
+					      get_order(size));
+#endif
+		free_pages((unsigned long) vaddr, get_order(size));
+	} else
+		/* DMA_TO_DEVICE to avoid memcpy in unmap_single */
+		swiotlb_unmap_single (hwdev, dma_handle, size, DMA_TO_DEVICE);
+}
+
+static void
+swiotlb_full(struct device *dev, size_t size, int dir, int do_panic)
+{
+	/*
+	 * Ran out of IOMMU space for this operation. This is very bad.
+	 * Unfortunately the drivers cannot handle this operation properly.
+	 * unless they check for dma_mapping_error (most don't)
+	 * When the mapping is small enough return a static buffer to limit
+	 * the damage, or panic when the transfer is too big.
+	 */
+	printk(KERN_ERR "DMA: Out of SW-IOMMU space for %lu bytes at "
+	       "device %s\n", size, dev ? dev->bus_id : "?");
+
+	if (size > io_tlb_overflow && do_panic) {
+		if (dir == DMA_FROM_DEVICE || dir == DMA_BIDIRECTIONAL)
+			panic("DMA: Memory would be corrupted\n");
+		if (dir == DMA_TO_DEVICE || dir == DMA_BIDIRECTIONAL)
+			panic("DMA: Random memory would be DMAed\n");
+	}
+}
+
+/*
+ * Map a single buffer of the indicated size for DMA in streaming mode.  The
+ * physical address to use is returned.
+ *
+ * Once the device is given the dma address, the device owns this memory until
+ * either swiotlb_unmap_single or swiotlb_dma_sync_single is performed.
+ */
+dma_addr_t
+swiotlb_map_single(struct device *hwdev, void *ptr, size_t size, int dir)
+{
+	unsigned long dev_addr = gnttab_dma_map_virt(ptr);
+	void *map;
+
+	BUG_ON(dir == DMA_NONE);
+	/*
+	 * If the pointer passed in happens to be in the device's DMA window,
+	 * we can safely return the device addr and not worry about bounce
+	 * buffering it.
+	 */
+	if (!range_straddles_page_boundary(__pa(ptr), size) &&
+	    !address_needs_mapping(hwdev, dev_addr) && !swiotlb_force)
+		return dev_addr;
+
+	__gnttab_dma_unmap_page(virt_to_page(ptr));
+	/*
+	 * Oh well, have to allocate and map a bounce buffer.
+	 */
+	map = map_single(hwdev, ptr, size, dir);
+	if (!map) {
+		swiotlb_full(hwdev, size, dir, 1);
+		map = io_tlb_overflow_buffer;
+	}
+
+	dev_addr = virt_to_bus(map);
+
+	/*
+	 * Ensure that the address returned is DMA'ble
+	 */
+	if (address_needs_mapping(hwdev, dev_addr))
+		panic("map_single: bounce buffer is not DMA'ble");
+
+	return dev_addr;
+}
+
+/*
+ * Since DMA is i-cache coherent, any (complete) pages that were written via
+ * DMA can be marked as "clean" so that lazy_mmu_prot_update() doesn't have to
+ * flush them when they get mapped into an executable vm-area.
+ */
+static void
+mark_clean(void *addr, size_t size)
+{
+	unsigned long pg_addr, end;
+
+#ifdef CONFIG_XEN
+	/* XXX: Bad things happen when starting domUs if this is enabled. */
+	if (is_running_on_xen())
+		return;
+#endif
+
+	pg_addr = PAGE_ALIGN((unsigned long) addr);
+	end = (unsigned long) addr + size;
+	while (pg_addr + PAGE_SIZE <= end) {
+		struct page *page = virt_to_page(pg_addr);
+		set_bit(PG_arch_1, &page->flags);
+		pg_addr += PAGE_SIZE;
+	}
+}
+
+/*
+ * Unmap a single streaming mode DMA translation.  The dma_addr and size must
+ * match what was provided for in a previous swiotlb_map_single call.  All
+ * other usages are undefined.
+ *
+ * After this call, reads by the cpu to the buffer are guaranteed to see
+ * whatever the device wrote there.
+ */
+void
+swiotlb_unmap_single(struct device *hwdev, dma_addr_t dev_addr, size_t size,
+		     int dir)
+{
+	char *dma_addr = bus_to_virt(dev_addr);
+
+	BUG_ON(dir == DMA_NONE);
+	if (dma_addr >= io_tlb_start && dma_addr < io_tlb_end)
+		unmap_single(hwdev, dma_addr, size, dir);
+	else {
+		__gnttab_dma_unmap_page(virt_to_page(dma_addr));
+		if (dir == DMA_FROM_DEVICE)
+			mark_clean(dma_addr, size);
+	}
+}
+
+/*
+ * Make physical memory consistent for a single streaming mode DMA translation
+ * after a transfer.
+ *
+ * If you perform a swiotlb_map_single() but wish to interrogate the buffer
+ * using the cpu, yet do not wish to teardown the dma mapping, you must
+ * call this function before doing so.  At the next point you give the dma
+ * address back to the card, you must first perform a
+ * swiotlb_dma_sync_for_device, and then the device again owns the buffer
+ */
+static inline void
+swiotlb_sync_single(struct device *hwdev, dma_addr_t dev_addr,
+		    size_t size, int dir, int target)
+{
+	char *dma_addr = bus_to_virt(dev_addr);
+
+	BUG_ON(dir == DMA_NONE);
+	if (dma_addr >= io_tlb_start && dma_addr < io_tlb_end)
+		sync_single(hwdev, dma_addr, size, dir, target);
+	else if (dir == DMA_FROM_DEVICE)
+		mark_clean(dma_addr, size);
+}
+
+void
+swiotlb_sync_single_for_cpu(struct device *hwdev, dma_addr_t dev_addr,
+			    size_t size, int dir)
+{
+	swiotlb_sync_single(hwdev, dev_addr, size, dir, SYNC_FOR_CPU);
+}
+
+void
+swiotlb_sync_single_for_device(struct device *hwdev, dma_addr_t dev_addr,
+			       size_t size, int dir)
+{
+	swiotlb_sync_single(hwdev, dev_addr, size, dir, SYNC_FOR_DEVICE);
+}
+
+/*
+ * Same as above, but for a sub-range of the mapping.
+ */
+static inline void
+swiotlb_sync_single_range(struct device *hwdev, dma_addr_t dev_addr,
+			  unsigned long offset, size_t size,
+			  int dir, int target)
+{
+	char *dma_addr = bus_to_virt(dev_addr) + offset;
+
+	BUG_ON(dir == DMA_NONE);
+	if (dma_addr >= io_tlb_start && dma_addr < io_tlb_end)
+		sync_single(hwdev, dma_addr, size, dir, target);
+	else if (dir == DMA_FROM_DEVICE)
+		mark_clean(dma_addr, size);
+}
+
+void
+swiotlb_sync_single_range_for_cpu(struct device *hwdev, dma_addr_t dev_addr,
+				  unsigned long offset, size_t size, int dir)
+{
+	swiotlb_sync_single_range(hwdev, dev_addr, offset, size, dir,
+				  SYNC_FOR_CPU);
+}
+
+void
+swiotlb_sync_single_range_for_device(struct device *hwdev, dma_addr_t dev_addr,
+				     unsigned long offset, size_t size, int dir)
+{
+	swiotlb_sync_single_range(hwdev, dev_addr, offset, size, dir,
+				  SYNC_FOR_DEVICE);
+}
+
+/*
+ * Map a set of buffers described by scatterlist in streaming mode for DMA.
+ * This is the scatter-gather version of the above swiotlb_map_single
+ * interface.  Here the scatter gather list elements are each tagged with the
+ * appropriate dma address and length.  They are obtained via
+ * sg_dma_{address,length}(SG).
+ *
+ * NOTE: An implementation may be able to use a smaller number of
+ *       DMA address/length pairs than there are SG table elements.
+ *       (for example via virtual mapping capabilities)
+ *       The routine returns the number of addr/length pairs actually
+ *       used, at most nents.
+ *
+ * Device ownership issues as mentioned above for swiotlb_map_single are the
+ * same here.
+ */
+int
+swiotlb_map_sg(struct device *hwdev, struct scatterlist *sg, int nelems,
+	       int dir)
+{
+	void *addr;
+	unsigned long dev_addr;
+	int i;
+
+	BUG_ON(dir == DMA_NONE);
+
+	for (i = 0; i < nelems; i++, sg++) {
+		addr = SG_ENT_VIRT_ADDRESS(sg);
+		dev_addr = gnttab_dma_map_virt(addr);
+		if (swiotlb_force ||
+		    range_straddles_page_boundary(page_to_pseudophys(sg->page)
+						  + sg->offset, sg->length) ||
+		    address_needs_mapping(hwdev, dev_addr)) {
+			void *map;
+			__gnttab_dma_unmap_page(sg->page);
+			map = map_single(hwdev, addr, sg->length, dir);
+			sg->dma_address = virt_to_bus(map);
+			if (!map) {
+				/* Don't panic here, we expect map_sg users
+				   to do proper error handling. */
+				swiotlb_full(hwdev, sg->length, dir, 0);
+				swiotlb_unmap_sg(hwdev, sg - i, i, dir);
+				sg[0].dma_length = 0;
+				return 0;
+			}
+		} else
+			sg->dma_address = dev_addr;
+		sg->dma_length = sg->length;
+	}
+	return nelems;
+}
+
+/*
+ * Unmap a set of streaming mode DMA translations.  Again, cpu read rules
+ * concerning calls here are the same as for swiotlb_unmap_single() above.
+ */
+void
+swiotlb_unmap_sg(struct device *hwdev, struct scatterlist *sg, int nelems,
+		 int dir)
+{
+	int i;
+
+	BUG_ON(dir == DMA_NONE);
+
+	for (i = 0; i < nelems; i++, sg++)
+		if (sg->dma_address != SG_ENT_PHYS_ADDRESS(sg))
+			unmap_single(hwdev, (void *) bus_to_virt(sg->dma_address), sg->dma_length, dir);
+		else {
+			__gnttab_dma_unmap_page(sg->page);
+			if (dir == DMA_FROM_DEVICE)
+				mark_clean(SG_ENT_VIRT_ADDRESS(sg),
+					   sg->dma_length);
+		}
+}
+
+/*
+ * Make physical memory consistent for a set of streaming mode DMA translations
+ * after a transfer.
+ *
+ * The same as swiotlb_sync_single_* but for a scatter-gather list, same rules
+ * and usage.
+ */
+static inline void
+swiotlb_sync_sg(struct device *hwdev, struct scatterlist *sg,
+		int nelems, int dir, int target)
+{
+	int i;
+
+	BUG_ON(dir == DMA_NONE);
+
+	for (i = 0; i < nelems; i++, sg++)
+		if (sg->dma_address != SG_ENT_PHYS_ADDRESS(sg))
+			sync_single(hwdev, (void *) sg->dma_address,
+				    sg->dma_length, dir, target);
+}
+
+void
+swiotlb_sync_sg_for_cpu(struct device *hwdev, struct scatterlist *sg,
+			int nelems, int dir)
+{
+	swiotlb_sync_sg(hwdev, sg, nelems, dir, SYNC_FOR_CPU);
+}
+
+void
+swiotlb_sync_sg_for_device(struct device *hwdev, struct scatterlist *sg,
+			   int nelems, int dir)
+{
+	swiotlb_sync_sg(hwdev, sg, nelems, dir, SYNC_FOR_DEVICE);
+}
+
+int
+swiotlb_dma_mapping_error(dma_addr_t dma_addr)
+{
+	return (dma_addr == virt_to_bus(io_tlb_overflow_buffer));
+}
+
+/*
+ * Return whether the given device DMA address mask can be supported
+ * properly.  For example, if your device can only drive the low 24-bits
+ * during bus mastering, then you would pass 0x00ffffff as the mask to
+ * this function.
+ */
+int
+swiotlb_dma_supported (struct device *hwdev, u64 mask)
+{
+#ifdef CONFIG_XEN
+	return (virt_to_bus(io_tlb_end - 1)) <= mask;
+#else
+	return (virt_to_bus(io_tlb_end) - 1) <= mask;
+#endif
+}
+
+EXPORT_SYMBOL(swiotlb_init);
+EXPORT_SYMBOL(swiotlb_map_single);
+EXPORT_SYMBOL(swiotlb_unmap_single);
+EXPORT_SYMBOL(swiotlb_map_sg);
+EXPORT_SYMBOL(swiotlb_unmap_sg);
+EXPORT_SYMBOL(swiotlb_sync_single_for_cpu);
+EXPORT_SYMBOL(swiotlb_sync_single_for_device);
+EXPORT_SYMBOL_GPL(swiotlb_sync_single_range_for_cpu);
+EXPORT_SYMBOL_GPL(swiotlb_sync_single_range_for_device);
+EXPORT_SYMBOL(swiotlb_sync_sg_for_cpu);
+EXPORT_SYMBOL(swiotlb_sync_sg_for_device);
+EXPORT_SYMBOL(swiotlb_dma_mapping_error);
+EXPORT_SYMBOL(swiotlb_alloc_coherent);
+EXPORT_SYMBOL(swiotlb_free_coherent);
+EXPORT_SYMBOL(swiotlb_dma_supported);
diff -rpuN linux-2.6.18.8/arch/ia64/xen/util.c linux-2.6.18-xen-3.3.0/arch/ia64/xen/util.c
--- linux-2.6.18.8/arch/ia64/xen/util.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/xen/util.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,102 @@
+/******************************************************************************
+ * arch/ia64/xen/util.c
+ * This file is the ia64 counterpart of drivers/xen/util.c
+ *
+ * Copyright (c) 2006 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <asm/uaccess.h>
+#include <xen/driver_util.h>
+#include <xen/interface/memory.h>
+#include <asm/hypercall.h>
+
+struct vm_struct *alloc_vm_area(unsigned long size)
+{
+	int order;
+	unsigned long virt;
+	unsigned long nr_pages;
+	struct vm_struct *area;
+
+	order = get_order(size);
+	virt = __get_free_pages(GFP_KERNEL, order);
+	if (virt == 0)
+		goto err0;
+	nr_pages = 1 << order;
+	scrub_pages(virt, nr_pages);
+
+	area = kmalloc(sizeof(*area), GFP_KERNEL);
+	if (area == NULL)
+		goto err1;
+
+        area->flags = VM_IOREMAP; /* XXX */
+        area->addr = (void*)virt;
+        area->size = size;
+        area->pages = NULL; /* XXX */
+        area->nr_pages = nr_pages;
+        area->phys_addr = 0; 	/* xenbus_map_ring_valloc uses this field!  */
+
+	return area;
+
+err1:
+	free_pages(virt, order);
+err0:
+	return NULL;
+}
+EXPORT_SYMBOL_GPL(alloc_vm_area);
+
+void free_vm_area(struct vm_struct *area)
+{
+	unsigned int order = get_order(area->size);
+	unsigned long i;
+	unsigned long phys_addr = __pa(area->addr);
+
+	/* This area is used for foreign page mappping.
+	 * So underlying machine page may not be assigned. */
+	for (i = 0; i < (1 << order); i++) {
+		unsigned long ret;
+		unsigned long gpfn = (phys_addr >> PAGE_SHIFT) + i;
+		struct xen_memory_reservation reservation = {
+			.nr_extents   = 1,
+			.address_bits = 0,
+			.extent_order = 0,
+			.domid        = DOMID_SELF
+		};
+		set_xen_guest_handle(reservation.extent_start, &gpfn);
+		ret = HYPERVISOR_memory_op(XENMEM_populate_physmap,
+					   &reservation);
+		BUG_ON(ret != 1);
+	}
+	free_pages((unsigned long)area->addr, order);
+	kfree(area);
+}
+EXPORT_SYMBOL_GPL(free_vm_area);
+
+/*
+ * Local variables:
+ *  c-file-style: "linux"
+ *  indent-tabs-mode: t
+ *  c-indent-level: 8
+ *  c-basic-offset: 8
+ *  tab-width: 8
+ * End:
+ */
diff -rpuN linux-2.6.18.8/arch/ia64/xen/xcom_asm.S linux-2.6.18-xen-3.3.0/arch/ia64/xen/xcom_asm.S
--- linux-2.6.18.8/arch/ia64/xen/xcom_asm.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/xen/xcom_asm.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,27 @@
+/*
+ * xencomm suspend support
+ * Support routines for Xen
+ *
+ * Copyright (C) 2005 Dan Magenheimer <dan.magenheimer@hp.com>
+ */
+#include <asm/asmmacro.h>
+#include <xen/interface/xen.h>
+
+/*
+ * Stub for suspend.
+ * Just force the stacked registers to be written in memory.
+ */
+GLOBAL_ENTRY(xencomm_arch_hypercall_suspend)
+	;; 
+	alloc r20=ar.pfs,0,0,6,0
+	mov r2=__HYPERVISOR_sched_op
+	;; 
+	/* We don't want to deal with RSE.  */
+	flushrs
+	mov r33=r32
+	mov r32=2 // SCHEDOP_shutdown
+	;;
+	break 0x1000
+	;; 
+	br.ret.sptk.many b0
+END(xencomm_arch_hypercall_suspend)
diff -rpuN linux-2.6.18.8/arch/ia64/xen/xcom_hcall.c linux-2.6.18-xen-3.3.0/arch/ia64/xen/xcom_hcall.c
--- linux-2.6.18.8/arch/ia64/xen/xcom_hcall.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/xen/xcom_hcall.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,671 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
+ *
+ *          Tristan Gingold <tristan.gingold@bull.net>
+ *
+ *          Copyright (c) 2007
+ *          Isaku Yamahata <yamahata at valinux co jp>
+ *                          VA Linux Systems Japan K.K.
+ *          consolidate mini and inline version.
+ */
+#include <linux/types.h>
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/gfp.h>
+#include <linux/module.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/platform.h>
+#include <xen/interface/memory.h>
+#include <xen/interface/xencomm.h>
+#include <xen/interface/version.h>
+#include <xen/interface/sched.h>
+#include <xen/interface/event_channel.h>
+#include <xen/interface/physdev.h>
+#include <xen/interface/grant_table.h>
+#include <xen/interface/callback.h>
+#include <xen/interface/xsm/acm_ops.h>
+#include <xen/interface/hvm/params.h>
+#include <xen/interface/xenoprof.h>
+#include <xen/interface/vcpu.h>
+#include <xen/interface/kexec.h>
+#include <asm/hypervisor.h>
+#include <asm/page.h>
+#include <asm/uaccess.h>
+#include <asm/xen/xencomm.h>
+#include <asm/perfmon.h>
+
+/* Xencomm notes:
+ * This file defines hypercalls to be used by xencomm.  The hypercalls simply
+ * create inlines or mini descriptors for pointers and then call the raw arch
+ * hypercall xencomm_arch_hypercall_XXX
+ *
+ * If the arch wants to directly use these hypercalls, simply define macros
+ * in asm/hypercall.h, eg:
+ *  #define HYPERVISOR_sched_op xencomm_hypercall_sched_op
+ * 
+ * The arch may also define HYPERVISOR_xxx as a function and do more operations
+ * before/after doing the hypercall.
+ *
+ * Note: because only inline or mini descriptors are created these functions
+ * must only be called with in kernel memory parameters.
+ */
+
+int
+xencomm_hypercall_console_io(int cmd, int count, char *str)
+{
+	return xencomm_arch_hypercall_console_io
+		(cmd, count, xencomm_map_no_alloc(str, count));
+}
+EXPORT_SYMBOL_GPL(xencomm_hypercall_console_io);
+
+int
+xencomm_hypercall_event_channel_op(int cmd, void *op)
+{
+	struct xencomm_handle *desc;
+	desc = xencomm_map_no_alloc(op, sizeof(evtchn_op_t));
+	if (desc == NULL)
+		return -EINVAL;
+
+	return xencomm_arch_hypercall_event_channel_op(cmd, desc);
+}
+EXPORT_SYMBOL_GPL(xencomm_hypercall_event_channel_op);
+
+int
+xencomm_hypercall_xen_version(int cmd, void *arg)
+{
+	struct xencomm_handle *desc;
+	unsigned int argsize;
+
+	switch (cmd) {
+	case XENVER_version:
+		/* do not actually pass an argument */
+		return xencomm_arch_hypercall_xen_version(cmd, 0);
+	case XENVER_extraversion:
+		argsize = sizeof(xen_extraversion_t);
+		break;
+	case XENVER_compile_info:
+		argsize = sizeof(xen_compile_info_t);
+		break;
+	case XENVER_capabilities:
+		argsize = sizeof(xen_capabilities_info_t);
+		break;
+	case XENVER_changeset:
+		argsize = sizeof(xen_changeset_info_t);
+		break;
+	case XENVER_platform_parameters:
+		argsize = sizeof(xen_platform_parameters_t);
+		break;
+	case XENVER_pagesize:
+		argsize = (arg == NULL) ? 0 : sizeof(void *);
+		break;
+	case XENVER_get_features:
+		argsize = (arg == NULL) ? 0 : sizeof(xen_feature_info_t);
+		break;
+
+	default:
+		printk("%s: unknown version op %d\n", __func__, cmd);
+		return -ENOSYS;
+	}
+
+	desc = xencomm_map_no_alloc(arg, argsize);
+	if (desc == NULL)
+		return -EINVAL;
+
+	return xencomm_arch_hypercall_xen_version(cmd, desc);
+}
+EXPORT_SYMBOL_GPL(xencomm_hypercall_xen_version);
+
+int
+xencomm_hypercall_physdev_op(int cmd, void *op)
+{
+	unsigned int argsize;
+	
+	switch (cmd) {
+	case PHYSDEVOP_apic_read:
+	case PHYSDEVOP_apic_write:
+		argsize = sizeof(physdev_apic_t);
+		break;
+	case PHYSDEVOP_alloc_irq_vector:
+	case PHYSDEVOP_free_irq_vector:
+		argsize = sizeof(physdev_irq_t);
+		break;
+	case PHYSDEVOP_irq_status_query:
+		argsize = sizeof(physdev_irq_status_query_t);
+		break;
+	case PHYSDEVOP_manage_pci_add:
+	case PHYSDEVOP_manage_pci_remove:
+		argsize = sizeof(physdev_manage_pci_t);
+		break;
+	case PHYSDEVOP_map_pirq:
+		argsize = sizeof(physdev_map_pirq_t);
+		break;
+	case PHYSDEVOP_unmap_pirq:
+		argsize = sizeof(physdev_unmap_pirq_t);
+		break;
+
+	default:
+		printk("%s: unknown physdev op %d\n", __func__, cmd);
+		return -ENOSYS;
+	}
+	
+	return xencomm_arch_hypercall_physdev_op
+		(cmd, xencomm_map_no_alloc(op, argsize));
+}
+
+static int
+xencommize_grant_table_op(struct xencomm_mini **xc_area,
+			  unsigned int cmd, void *op, unsigned int count,
+			  struct xencomm_handle **desc)
+{
+	struct xencomm_handle *desc1;
+	unsigned int argsize;
+
+	switch (cmd) {
+	case GNTTABOP_map_grant_ref:
+		argsize = sizeof(struct gnttab_map_grant_ref);
+		break;
+	case GNTTABOP_unmap_grant_ref:
+		argsize = sizeof(struct gnttab_unmap_grant_ref);
+		break;
+	case GNTTABOP_unmap_and_replace:
+		argsize = sizeof(struct gnttab_unmap_and_replace);
+		break;
+	case GNTTABOP_setup_table:
+	{
+		struct gnttab_setup_table *setup = op;
+
+		argsize = sizeof(*setup);
+
+		if (count != 1)
+			return -EINVAL;
+		desc1 = __xencomm_map_no_alloc
+			(xen_guest_handle(setup->frame_list),
+			 setup->nr_frames *
+			 sizeof(*xen_guest_handle(setup->frame_list)),
+			 *xc_area);
+		if (desc1 == NULL)
+			return -EINVAL;
+		(*xc_area)++;
+		set_xen_guest_handle(setup->frame_list, (void *)desc1);
+		break;
+	}
+	case GNTTABOP_dump_table:
+		argsize = sizeof(struct gnttab_dump_table);
+		break;
+	case GNTTABOP_transfer:
+		argsize = sizeof(struct gnttab_transfer);
+		break;
+	case GNTTABOP_copy:
+		argsize = sizeof(struct gnttab_copy);
+		break;
+	case GNTTABOP_query_size:
+		argsize = sizeof(struct gnttab_query_size);
+		break;
+	default:
+		printk("%s: unknown hypercall grant table op %d\n",
+		       __func__, cmd);
+		BUG();
+	}
+
+	*desc = __xencomm_map_no_alloc(op, count * argsize, *xc_area);
+	if (*desc == NULL)
+		return -EINVAL;
+	(*xc_area)++;
+
+	return 0;
+}
+
+int
+xencomm_hypercall_grant_table_op(unsigned int cmd, void *op,
+                                      unsigned int count)
+{
+	int rc;
+	struct xencomm_handle *desc;
+	XENCOMM_MINI_ALIGNED(xc_area, 2);
+
+	rc = xencommize_grant_table_op(&xc_area, cmd, op, count, &desc);
+	if (rc)
+		return rc;
+
+	return xencomm_arch_hypercall_grant_table_op(cmd, desc, count);
+}
+EXPORT_SYMBOL_GPL(xencomm_hypercall_grant_table_op);
+
+int
+xencomm_hypercall_sched_op(int cmd, void *arg)
+{
+	struct xencomm_handle *desc;
+	unsigned int argsize;
+
+	switch (cmd) {
+	case SCHEDOP_yield:
+	case SCHEDOP_block:
+		argsize = 0;
+		break;
+	case SCHEDOP_shutdown:
+		argsize = sizeof(sched_shutdown_t);
+		break;
+	case SCHEDOP_remote_shutdown:
+		argsize = sizeof(sched_remote_shutdown_t);
+		break;
+	case SCHEDOP_poll:
+	{
+		sched_poll_t *poll = arg;
+		struct xencomm_handle *ports;
+
+		argsize = sizeof(sched_poll_t);
+		ports = xencomm_map_no_alloc(xen_guest_handle(poll->ports),
+				     sizeof(*xen_guest_handle(poll->ports)));
+
+		set_xen_guest_handle(poll->ports, (void *)ports);
+		break;
+	}
+	default:
+		printk("%s: unknown sched op %d\n", __func__, cmd);
+		return -ENOSYS;
+	}
+	
+	desc = xencomm_map_no_alloc(arg, argsize);
+	if (desc == NULL)
+		return -EINVAL;
+
+	return xencomm_arch_hypercall_sched_op(cmd, desc);
+}
+EXPORT_SYMBOL_GPL(xencomm_hypercall_sched_op);
+
+int
+xencomm_hypercall_multicall(void *call_list, int nr_calls)
+{
+	int rc;
+	int i;
+	multicall_entry_t *mce;
+	struct xencomm_handle *desc;
+	XENCOMM_MINI_ALIGNED(xc_area, nr_calls * 2);
+
+	for (i = 0; i < nr_calls; i++) {
+		mce = (multicall_entry_t *)call_list + i;
+
+		switch (mce->op) {
+		case __HYPERVISOR_update_va_mapping:
+		case __HYPERVISOR_mmu_update:
+			/* No-op on ia64.  */
+			break;
+		case __HYPERVISOR_grant_table_op:
+			rc = xencommize_grant_table_op
+				(&xc_area,
+				 mce->args[0], (void *)mce->args[1],
+				 mce->args[2], &desc);
+			if (rc)
+				return rc;
+			mce->args[1] = (unsigned long)desc;
+			break;
+		case __HYPERVISOR_memory_op:
+		default:
+			printk("%s: unhandled multicall op entry op %lu\n",
+			       __func__, mce->op);
+			return -ENOSYS;
+		}
+	}
+
+	desc = xencomm_map_no_alloc(call_list,
+				    nr_calls * sizeof(multicall_entry_t));
+	if (desc == NULL)
+		return -EINVAL;
+
+	return xencomm_arch_hypercall_multicall(desc, nr_calls);
+}
+EXPORT_SYMBOL_GPL(xencomm_hypercall_multicall);
+
+int
+xencomm_hypercall_callback_op(int cmd, void *arg)
+{
+	unsigned int argsize;
+	switch (cmd)
+	{
+	case CALLBACKOP_register:
+		argsize = sizeof(struct callback_register);
+		break;
+	case CALLBACKOP_unregister:
+		argsize = sizeof(struct callback_unregister);
+		break;
+	default:
+		printk("%s: unknown callback op %d\n", __func__, cmd);
+		return -ENOSYS;
+	}
+
+	return xencomm_arch_hypercall_callback_op
+		(cmd, xencomm_map_no_alloc(arg, argsize));
+}
+
+static int
+xencommize_memory_reservation(struct xencomm_mini *xc_area,
+			      xen_memory_reservation_t *mop)
+{
+	struct xencomm_handle *desc;
+
+	desc = __xencomm_map_no_alloc(xen_guest_handle(mop->extent_start),
+			mop->nr_extents *
+			sizeof(*xen_guest_handle(mop->extent_start)),
+			xc_area);
+	if (desc == NULL)
+		return -EINVAL;
+
+	set_xen_guest_handle(mop->extent_start, (void *)desc);
+	return 0;
+}
+
+int
+xencomm_hypercall_memory_op(unsigned int cmd, void *arg)
+{
+	XEN_GUEST_HANDLE(xen_pfn_t) extent_start_va[2];
+	xen_memory_reservation_t *xmr = NULL, *xme_in = NULL, *xme_out = NULL;
+	xen_memory_map_t *memmap = NULL;
+	XEN_GUEST_HANDLE(void) buffer;
+	int rc;
+	struct xencomm_handle *desc;
+	unsigned int argsize;
+	XENCOMM_MINI_ALIGNED(xc_area, 2);
+
+	switch (cmd) {
+	case XENMEM_increase_reservation:
+	case XENMEM_decrease_reservation:
+	case XENMEM_populate_physmap:
+		xmr = (xen_memory_reservation_t *)arg;
+		set_xen_guest_handle(extent_start_va[0],
+				     xen_guest_handle(xmr->extent_start));
+
+		argsize = sizeof(*xmr);
+		rc = xencommize_memory_reservation(xc_area, xmr);
+		if (rc)
+			return rc;
+		xc_area++;
+		break;
+
+	case XENMEM_maximum_gpfn:
+		argsize = 0;
+		break;
+
+	case XENMEM_maximum_ram_page:
+		argsize = 0;
+		break;
+
+	case XENMEM_exchange:
+		xme_in  = &((xen_memory_exchange_t *)arg)->in;
+		xme_out = &((xen_memory_exchange_t *)arg)->out;
+		set_xen_guest_handle(extent_start_va[0],
+				     xen_guest_handle(xme_in->extent_start));
+		set_xen_guest_handle(extent_start_va[1],
+				     xen_guest_handle(xme_out->extent_start));
+
+		argsize = sizeof(xen_memory_exchange_t);
+		rc = xencommize_memory_reservation(xc_area, xme_in);
+		if (rc)
+			return rc;
+		xc_area++;
+		rc = xencommize_memory_reservation(xc_area, xme_out);
+		if (rc)
+			return rc;
+		xc_area++;
+		break;
+
+	case XENMEM_add_to_physmap:
+		argsize = sizeof(xen_add_to_physmap_t);
+		break;
+
+	case XENMEM_machine_memory_map:
+		argsize = sizeof(*memmap);
+		memmap = (xen_memory_map_t *)arg;
+		set_xen_guest_handle(buffer, xen_guest_handle(memmap->buffer));
+		desc = xencomm_map_no_alloc(xen_guest_handle(memmap->buffer),
+					      memmap->nr_entries);
+		if (desc == NULL)
+			return -EINVAL;
+		set_xen_guest_handle(memmap->buffer, (void *)desc);
+		break;
+
+	default:
+		printk("%s: unknown memory op %d\n", __func__, cmd);
+		return -ENOSYS;
+	}
+
+	desc = xencomm_map_no_alloc(arg, argsize);
+	if (desc == NULL)
+		return -EINVAL;
+
+	rc = xencomm_arch_hypercall_memory_op(cmd, desc);
+
+	switch (cmd) {
+	case XENMEM_increase_reservation:
+	case XENMEM_decrease_reservation:
+	case XENMEM_populate_physmap:
+		set_xen_guest_handle(xmr->extent_start,
+				     xen_guest_handle(extent_start_va[0]));
+		break;
+
+	case XENMEM_exchange:
+		set_xen_guest_handle(xme_in->extent_start,
+				     xen_guest_handle(extent_start_va[0]));
+		set_xen_guest_handle(xme_out->extent_start,
+				     xen_guest_handle(extent_start_va[1]));
+		break;
+
+	case XENMEM_machine_memory_map:
+		set_xen_guest_handle(memmap->buffer, xen_guest_handle(buffer));
+		break;
+	}
+
+	return rc;
+}
+EXPORT_SYMBOL_GPL(xencomm_hypercall_memory_op);
+
+unsigned long
+xencomm_hypercall_hvm_op(int cmd, void *arg)
+{
+	struct xencomm_handle *desc;
+	unsigned int argsize;
+
+	switch (cmd) {
+	case HVMOP_get_param:
+	case HVMOP_set_param:
+		argsize = sizeof(xen_hvm_param_t);
+		break;
+	default:
+		printk("%s: unknown HVMOP %d\n", __func__, cmd);
+		return -EINVAL;
+	}
+
+	desc = xencomm_map_no_alloc(arg, argsize);
+	if (desc == NULL)
+		return -EINVAL;
+
+	return xencomm_arch_hypercall_hvm_op(cmd, desc);
+}
+EXPORT_SYMBOL_GPL(xencomm_hypercall_hvm_op);
+
+int
+xencomm_hypercall_suspend(unsigned long srec)
+{
+	struct sched_shutdown arg;
+
+	arg.reason = SHUTDOWN_suspend;
+
+	return xencomm_arch_hypercall_suspend(
+		xencomm_map_no_alloc(&arg, sizeof(arg)));
+}
+
+int
+xencomm_hypercall_xenoprof_op(int op, void *arg)
+{
+	unsigned int argsize;
+	struct xencomm_handle *desc;
+
+	switch (op) {
+	case XENOPROF_init:
+		argsize = sizeof(xenoprof_init_t);
+		break;
+	case XENOPROF_set_active:
+		argsize = sizeof(domid_t);
+		break;
+	case XENOPROF_set_passive:
+		argsize = sizeof(xenoprof_passive_t);
+		break;
+	case XENOPROF_counter:
+		argsize = sizeof(xenoprof_counter_t);
+		break;
+	case XENOPROF_get_buffer:
+		argsize = sizeof(xenoprof_get_buffer_t);
+		break;
+
+	case XENOPROF_reset_active_list:
+	case XENOPROF_reset_passive_list:
+	case XENOPROF_reserve_counters:
+	case XENOPROF_setup_events:
+	case XENOPROF_enable_virq:
+	case XENOPROF_start:
+	case XENOPROF_stop:
+	case XENOPROF_disable_virq:
+	case XENOPROF_release_counters:
+	case XENOPROF_shutdown:
+		return xencomm_arch_hypercall_xenoprof_op(op, arg);
+
+	default:
+		printk("%s: op %d isn't supported\n", __func__, op);
+		return -ENOSYS;
+	}
+
+	desc = xencomm_map_no_alloc(arg, argsize);
+	if (desc == NULL)
+		return -EINVAL;
+
+	return xencomm_arch_hypercall_xenoprof_op(op, desc);
+}
+EXPORT_SYMBOL_GPL(xencomm_hypercall_xenoprof_op);
+
+int
+xencomm_hypercall_perfmon_op(unsigned long cmd, void* arg,
+                                  unsigned long count)
+{
+	unsigned int argsize;
+	struct xencomm_handle *desc;
+
+	switch (cmd) {
+	case PFM_GET_FEATURES:
+		argsize = sizeof(pfarg_features_t);
+		break;
+	case PFM_CREATE_CONTEXT:
+		argsize = sizeof(pfarg_context_t);
+		break;
+	case PFM_LOAD_CONTEXT:
+		argsize = sizeof(pfarg_load_t);
+		break;
+	case PFM_WRITE_PMCS:
+	case PFM_WRITE_PMDS:
+		argsize = sizeof(pfarg_reg_t) * count;
+		break;
+
+	case PFM_DESTROY_CONTEXT:
+	case PFM_UNLOAD_CONTEXT:
+	case PFM_START:
+	case PFM_STOP:
+		return xencomm_arch_hypercall_perfmon_op(cmd, arg, count);
+
+	default:
+		printk("%s:%d cmd %ld isn't supported\n",
+		       __func__, __LINE__, cmd);
+		BUG();
+	}
+
+	desc = xencomm_map_no_alloc(arg, argsize);
+	if (desc == NULL)
+		return -EINVAL;
+
+	return xencomm_arch_hypercall_perfmon_op(cmd, desc, count);
+}
+EXPORT_SYMBOL_GPL(xencomm_hypercall_perfmon_op);
+
+long
+xencomm_hypercall_vcpu_op(int cmd, int cpu, void *arg)
+{
+	unsigned int argsize;
+	switch (cmd) {
+	case VCPUOP_register_runstate_memory_area: {
+		vcpu_register_runstate_memory_area_t *area =
+			(vcpu_register_runstate_memory_area_t *)arg;
+		argsize = sizeof(*arg);
+		set_xen_guest_handle(area->addr.h,
+		     (void *)xencomm_map_no_alloc(area->addr.v,
+						  sizeof(area->addr.v)));
+		break;
+	}
+
+	default:
+		printk("%s: unknown vcpu op %d\n", __func__, cmd);
+		return -ENOSYS;
+	}
+
+	return xencomm_arch_hypercall_vcpu_op(cmd, cpu,
+					xencomm_map_no_alloc(arg, argsize));
+}
+
+long
+xencomm_hypercall_opt_feature(void *arg)
+{
+	return xencomm_arch_hypercall_opt_feature(
+		xencomm_map_no_alloc(arg,
+				     sizeof(struct xen_ia64_opt_feature)));
+}
+
+int
+xencomm_hypercall_fpswa_revision(unsigned int *revision)
+{
+	struct xencomm_handle *desc;
+
+	desc = xencomm_map_no_alloc(revision, sizeof(*revision));
+	if (desc == NULL)
+		return -EINVAL;
+
+	return xencomm_arch_hypercall_fpswa_revision(desc);
+}
+EXPORT_SYMBOL_GPL(xencomm_hypercall_fpswa_revision);
+
+int
+xencomm_hypercall_kexec_op(int cmd, void *arg)
+{
+	unsigned int argsize;
+	struct xencomm_handle *desc;
+
+	switch (cmd) {
+	case KEXEC_CMD_kexec_get_range:
+		argsize = sizeof(xen_kexec_range_t);
+		break;
+	case KEXEC_CMD_kexec_load:
+	case KEXEC_CMD_kexec_unload:
+		argsize = sizeof(xen_kexec_load_t);
+		break;
+	case KEXEC_CMD_kexec:
+		argsize = sizeof(xen_kexec_exec_t);
+		break;
+	default:
+		printk("%s:%d cmd %d isn't supported\n",
+		       __func__, __LINE__, cmd);
+		BUG();
+	}
+
+	desc = xencomm_map_no_alloc(arg, argsize);
+	if (desc == NULL)
+		return -EINVAL;
+
+	return xencomm_arch_hypercall_kexec_op(cmd, desc);
+}
diff -rpuN linux-2.6.18.8/arch/ia64/xen/xcom_privcmd.c linux-2.6.18-xen-3.3.0/arch/ia64/xen/xcom_privcmd.c
--- linux-2.6.18.8/arch/ia64/xen/xcom_privcmd.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/xen/xcom_privcmd.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,902 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
+ *
+ * Authors: Hollis Blanchard <hollisb@us.ibm.com>
+ *          Tristan Gingold <tristan.gingold@bull.net>
+ */
+#include <linux/types.h>
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/gfp.h>
+#include <linux/module.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/platform.h>
+#define __XEN__
+#include <xen/interface/domctl.h>
+#include <xen/interface/sysctl.h>
+#include <xen/interface/memory.h>
+#include <xen/interface/version.h>
+#include <xen/interface/event_channel.h>
+#include <xen/interface/xsm/acm_ops.h>
+#include <xen/interface/hvm/params.h>
+#include <xen/interface/arch-ia64/debug_op.h>
+#include <xen/public/privcmd.h>
+#include <asm/hypercall.h>
+#include <asm/page.h>
+#include <asm/uaccess.h>
+#include <asm/xen/xencomm.h>
+
+#define ROUND_DIV(v,s) (((v) + (s) - 1) / (s))
+
+static int
+xencomm_privcmd_platform_op(privcmd_hypercall_t *hypercall)
+{
+	struct xen_platform_op kern_op;
+	struct xen_platform_op __user *user_op = (struct xen_platform_op __user *)hypercall->arg[0];
+	struct xencomm_handle *op_desc;
+	struct xencomm_handle *desc = NULL;
+	int ret = 0;
+
+	if (copy_from_user(&kern_op, user_op, sizeof(struct xen_platform_op)))
+		return -EFAULT;
+
+	if (kern_op.interface_version != XENPF_INTERFACE_VERSION)
+		return -EACCES;
+
+	op_desc = xencomm_map_no_alloc(&kern_op, sizeof(kern_op));
+
+	switch (kern_op.cmd) {
+	default:
+		printk("%s: unknown platform cmd %d\n", __func__, kern_op.cmd);
+		return -ENOSYS;
+	}
+
+	if (ret) {
+		/* error mapping the nested pointer */
+		return ret;
+	}
+
+	ret = xencomm_arch_hypercall_platform_op(op_desc);
+
+	/* FIXME: should we restore the handle?  */
+	if (copy_to_user(user_op, &kern_op, sizeof(struct xen_platform_op)))
+		ret = -EFAULT;
+
+	xencomm_free(desc);
+	return ret;
+}
+
+static int
+xencomm_privcmd_sysctl(privcmd_hypercall_t *hypercall)
+{
+	xen_sysctl_t kern_op;
+	xen_sysctl_t __user *user_op;
+	struct xencomm_handle *op_desc;
+	struct xencomm_handle *desc = NULL;
+	struct xencomm_handle *desc1 = NULL;
+	int ret = 0;
+
+	user_op = (xen_sysctl_t __user *)hypercall->arg[0];
+
+	if (copy_from_user(&kern_op, user_op, sizeof(xen_sysctl_t)))
+		return -EFAULT;
+
+	if (kern_op.interface_version != XEN_SYSCTL_INTERFACE_VERSION)
+		return -EACCES;
+
+	op_desc = xencomm_map_no_alloc(&kern_op, sizeof(kern_op));
+
+	switch (kern_op.cmd) {
+	case XEN_SYSCTL_readconsole:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.readconsole.buffer),
+			kern_op.u.readconsole.count);
+		if (xen_guest_handle(kern_op.u.readconsole.buffer) != NULL &&
+		    kern_op.u.readconsole.count > 0 && desc == NULL)
+			return -ENOMEM;
+		set_xen_guest_handle(kern_op.u.readconsole.buffer,
+		                     (void *)desc);
+		break;
+	case XEN_SYSCTL_tbuf_op:
+	case XEN_SYSCTL_sched_id:
+	case XEN_SYSCTL_availheap:
+		break;
+	case XEN_SYSCTL_perfc_op:
+	{
+		struct xencomm_handle *tmp_desc;
+		xen_sysctl_t tmp_op = {
+			.cmd = XEN_SYSCTL_perfc_op,
+			.interface_version = XEN_SYSCTL_INTERFACE_VERSION,
+			.u.perfc_op = {
+				.cmd = XEN_SYSCTL_PERFCOP_query,
+				/* .desc.p = NULL, */
+				/* .val.p = NULL, */
+			},
+		};
+
+		if (xen_guest_handle(kern_op.u.perfc_op.desc) == NULL) {
+			if (xen_guest_handle(kern_op.u.perfc_op.val) != NULL)
+				return -EINVAL;
+			break;
+		}
+
+		/* query the buffer size for xencomm */
+		tmp_desc = xencomm_map_no_alloc(&tmp_op, sizeof(tmp_op));
+		ret = xencomm_arch_hypercall_sysctl(tmp_desc);
+		if (ret)
+			return ret;
+
+		desc = xencomm_map(xen_guest_handle(kern_op.u.perfc_op.desc),
+				   tmp_op.u.perfc_op.nr_counters *
+				   sizeof(xen_sysctl_perfc_desc_t));
+		if (xen_guest_handle(kern_op.u.perfc_op.desc) != NULL &&
+		    tmp_op.u.perfc_op.nr_counters > 0 && desc == NULL)
+			return -ENOMEM;
+
+		set_xen_guest_handle(kern_op.u.perfc_op.desc, (void *)desc);
+
+		desc1 = xencomm_map(xen_guest_handle(kern_op.u.perfc_op.val),
+				    tmp_op.u.perfc_op.nr_vals *
+				    sizeof(xen_sysctl_perfc_val_t));
+		if (xen_guest_handle(kern_op.u.perfc_op.val) != NULL &&
+		    tmp_op.u.perfc_op.nr_vals > 0 && desc1 == NULL) {
+			xencomm_free(desc);
+			return -ENOMEM;
+		}
+
+		set_xen_guest_handle(kern_op.u.perfc_op.val, (void *)desc1);
+		break;
+	}
+	case XEN_SYSCTL_getdomaininfolist:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.getdomaininfolist.buffer),
+			kern_op.u.getdomaininfolist.max_domains *
+			sizeof(xen_domctl_getdomaininfo_t));
+		if (xen_guest_handle(kern_op.u.getdomaininfolist.buffer) !=
+		    NULL && kern_op.u.getdomaininfolist.max_domains > 0 &&
+		    desc == NULL)
+			return -ENOMEM;
+		set_xen_guest_handle(kern_op.u.getdomaininfolist.buffer,
+				     (void *)desc);
+		break;
+	case XEN_SYSCTL_debug_keys:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.debug_keys.keys),
+			kern_op.u.debug_keys.nr_keys);
+		if (xen_guest_handle(kern_op.u.debug_keys.keys) != NULL &&
+		    kern_op.u.debug_keys.nr_keys > 0 && desc == NULL)
+			return -ENOMEM;
+		set_xen_guest_handle(kern_op.u.debug_keys.keys,
+				     (void *)desc);
+		break;
+
+	case XEN_SYSCTL_physinfo:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.physinfo.cpu_to_node),
+			kern_op.u.physinfo.max_cpu_id * sizeof(uint32_t));
+		if (xen_guest_handle(kern_op.u.physinfo.cpu_to_node) != NULL &&
+		    kern_op.u.physinfo.max_cpu_id > 0 && desc == NULL)
+			return -ENOMEM;
+
+		set_xen_guest_handle(kern_op.u.physinfo.cpu_to_node,
+		                     (void *)desc);
+		break;
+	default:
+		printk("%s: unknown sysctl cmd %d\n", __func__, kern_op.cmd);
+		return -ENOSYS;
+	}
+
+	if (ret) {
+		/* error mapping the nested pointer */
+		return ret;
+	}
+
+	ret = xencomm_arch_hypercall_sysctl(op_desc);
+
+	/* FIXME: should we restore the handles?  */
+	if (copy_to_user(user_op, &kern_op, sizeof(xen_sysctl_t)))
+		ret = -EFAULT;
+
+	xencomm_free(desc);
+	xencomm_free(desc1);
+	return ret;
+}
+
+static int
+xencomm_privcmd_domctl(privcmd_hypercall_t *hypercall)
+{
+	xen_domctl_t kern_op;
+	xen_domctl_t __user *user_op;
+	struct xencomm_handle *op_desc;
+	struct xencomm_handle *desc = NULL;
+	int ret = 0;
+
+	user_op = (xen_domctl_t __user *)hypercall->arg[0];
+
+	if (copy_from_user(&kern_op, user_op, sizeof(xen_domctl_t)))
+		return -EFAULT;
+
+	if (kern_op.interface_version != XEN_DOMCTL_INTERFACE_VERSION)
+		return -EACCES;
+
+	op_desc = xencomm_map_no_alloc(&kern_op, sizeof(kern_op));
+
+	switch (kern_op.cmd) {
+	case XEN_DOMCTL_createdomain:
+	case XEN_DOMCTL_destroydomain:
+	case XEN_DOMCTL_pausedomain:
+	case XEN_DOMCTL_unpausedomain:
+	case XEN_DOMCTL_resumedomain:
+	case XEN_DOMCTL_getdomaininfo:
+		break;
+	case XEN_DOMCTL_getmemlist:
+	{
+		unsigned long nr_pages = kern_op.u.getmemlist.max_pfns;
+
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.getmemlist.buffer),
+			nr_pages * sizeof(unsigned long));
+		if (xen_guest_handle(kern_op.u.getmemlist.buffer) != NULL &&
+		    nr_pages > 0 && desc == NULL)
+			return -ENOMEM;
+		set_xen_guest_handle(kern_op.u.getmemlist.buffer,
+		                     (void *)desc);
+		break;
+	}
+	case XEN_DOMCTL_getpageframeinfo:
+		break;
+	case XEN_DOMCTL_getpageframeinfo2:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.getpageframeinfo2.array),
+			kern_op.u.getpageframeinfo2.num);
+		if (xen_guest_handle(kern_op.u.getpageframeinfo2.array) !=
+		    NULL && kern_op.u.getpageframeinfo2.num > 0 &&
+		    desc == NULL)
+			return -ENOMEM;
+		set_xen_guest_handle(kern_op.u.getpageframeinfo2.array,
+		                     (void *)desc);
+		break;
+	case XEN_DOMCTL_shadow_op:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.shadow_op.dirty_bitmap),
+			ROUND_DIV(kern_op.u.shadow_op.pages, 8));
+		if (xen_guest_handle(kern_op.u.shadow_op.dirty_bitmap) != NULL
+		    && kern_op.u.shadow_op.pages > 0 && desc == NULL)
+			return -ENOMEM;
+		set_xen_guest_handle(kern_op.u.shadow_op.dirty_bitmap,
+		                     (void *)desc);
+		break;
+	case XEN_DOMCTL_max_mem:
+		break;
+	case XEN_DOMCTL_setvcpucontext:
+	case XEN_DOMCTL_getvcpucontext:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.vcpucontext.ctxt),
+			sizeof(vcpu_guest_context_t));
+		if (xen_guest_handle(kern_op.u.vcpucontext.ctxt) != NULL &&
+		    desc == NULL)
+			return -ENOMEM;
+		set_xen_guest_handle(kern_op.u.vcpucontext.ctxt, (void *)desc);
+		break;
+	case XEN_DOMCTL_getvcpuinfo:
+		break;
+	case XEN_DOMCTL_setvcpuaffinity:
+	case XEN_DOMCTL_getvcpuaffinity:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.vcpuaffinity.cpumap.bitmap),
+			ROUND_DIV(kern_op.u.vcpuaffinity.cpumap.nr_cpus, 8));
+		if (xen_guest_handle(kern_op.u.vcpuaffinity.cpumap.bitmap) !=
+		    NULL && kern_op.u.vcpuaffinity.cpumap.nr_cpus > 0 &&
+		    desc == NULL)
+			return -ENOMEM;
+		set_xen_guest_handle(kern_op.u.vcpuaffinity.cpumap.bitmap,
+		                     (void *)desc);
+		break;
+	case XEN_DOMCTL_gethvmcontext:
+	case XEN_DOMCTL_sethvmcontext:
+		if (kern_op.u.hvmcontext.size > 0)
+			desc = xencomm_map(
+				xen_guest_handle(kern_op.u.hvmcontext.buffer),
+				kern_op.u.hvmcontext.size);
+		if (xen_guest_handle(kern_op.u.hvmcontext.buffer) != NULL &&
+		    kern_op.u.hvmcontext.size > 0 && desc == NULL)
+			return -ENOMEM;
+		set_xen_guest_handle(kern_op.u.hvmcontext.buffer, (void*)desc);
+		break;
+	case XEN_DOMCTL_max_vcpus:
+	case XEN_DOMCTL_scheduler_op:
+	case XEN_DOMCTL_setdomainhandle:
+	case XEN_DOMCTL_setdebugging:
+	case XEN_DOMCTL_irq_permission:
+	case XEN_DOMCTL_iomem_permission:
+	case XEN_DOMCTL_ioport_permission:
+	case XEN_DOMCTL_hypercall_init:
+	case XEN_DOMCTL_arch_setup:
+	case XEN_DOMCTL_settimeoffset:
+	case XEN_DOMCTL_sendtrigger:
+	case XEN_DOMCTL_set_opt_feature:
+	case XEN_DOMCTL_assign_device:
+	case XEN_DOMCTL_subscribe:
+		break;
+	case XEN_DOMCTL_pin_mem_cacheattr:
+		return -ENOSYS;
+	default:
+		printk("%s: unknown domctl cmd %d\n", __func__, kern_op.cmd);
+		return -ENOSYS;
+	}
+
+	if (ret) {
+		/* error mapping the nested pointer */
+		return ret;
+	}
+
+	ret = xencomm_arch_hypercall_domctl (op_desc);
+
+	/* FIXME: should we restore the handle?  */
+	if (copy_to_user(user_op, &kern_op, sizeof(xen_domctl_t)))
+		ret = -EFAULT;
+
+	xencomm_free(desc);
+	return ret;
+}
+
+static int
+xencomm_privcmd_xsm_op(privcmd_hypercall_t *hypercall)
+{
+	void __user *arg = (void __user *)hypercall->arg[0];
+	xen_acmctl_t kern_arg;
+	struct xencomm_handle *op_desc;
+	struct xencomm_handle *desc = NULL;
+	int ret;
+
+	if (copy_from_user(&kern_arg, arg, sizeof(kern_arg)))
+		return -EFAULT;
+	if (kern_arg.interface_version != ACM_INTERFACE_VERSION)
+		return -ENOSYS;
+	
+	switch (kern_arg.cmd) {
+	case ACMOP_getssid: {
+		op_desc = xencomm_map_no_alloc(&kern_arg, sizeof(kern_arg));
+
+		desc = xencomm_map(
+			xen_guest_handle(kern_arg.u.getssid.ssidbuf),
+			kern_arg.u.getssid.ssidbuf_size);
+		if (xen_guest_handle(kern_arg.u.getssid.ssidbuf) != NULL &&
+		    kern_arg.u.getssid.ssidbuf_size > 0 && desc == NULL)
+			return -ENOMEM;
+
+		set_xen_guest_handle(kern_arg.u.getssid.ssidbuf, (void *)desc);
+
+		ret = xencomm_arch_hypercall_xsm_op(op_desc);
+
+		xencomm_free(desc);
+
+		if (copy_to_user(arg, &kern_arg, sizeof(kern_arg)))
+			return -EFAULT;
+		return ret;
+	}
+	default:
+		printk("%s: unknown acm_op cmd %d\n", __func__, kern_arg.cmd);
+		return -ENOSYS;
+	}
+
+	return ret;
+}
+
+static int
+xencomm_privcmd_memory_reservation_op(privcmd_hypercall_t *hypercall)
+{
+	const unsigned long cmd = hypercall->arg[0];
+	int ret = 0;
+	xen_memory_reservation_t kern_op;
+	xen_memory_reservation_t __user *user_op;
+	struct xencomm_handle *desc = NULL;
+	struct xencomm_handle *desc_op;
+
+	user_op = (xen_memory_reservation_t __user *)hypercall->arg[1];
+	if (copy_from_user(&kern_op, user_op,
+			   sizeof(xen_memory_reservation_t)))
+		return -EFAULT;
+	desc_op = xencomm_map_no_alloc(&kern_op, sizeof(kern_op));
+
+	if (!xen_guest_handle(kern_op.extent_start)) {
+		ret = xencomm_arch_hypercall_memory_op(cmd, desc_op);
+		if (ret < 0)
+			return ret;
+	} else {
+		xen_ulong_t nr_done = 0;
+		xen_ulong_t nr_extents = kern_op.nr_extents;
+		void *addr = xen_guest_handle(kern_op.extent_start);
+			
+		/*
+		 * Work around.
+		 *   Xencomm has single page size limit caused
+		 *   by xencomm_alloc()/xencomm_free() so that
+		 *   we have to repeat the hypercall.
+		 *   This limitation can be removed.
+		 */
+#define MEMORYOP_XENCOMM_LIMIT						\
+		(((((PAGE_SIZE - sizeof(struct xencomm_desc)) /		\
+		    sizeof(uint64_t)) - 2) * PAGE_SIZE) /		\
+		 sizeof(*xen_guest_handle(kern_op.extent_start)))
+
+		/*
+		 * Work around.
+		 *   Even if the above limitation is removed,
+		 *   the hypercall with large number of extents 
+		 *   may cause the soft lockup warning.
+		 *   In order to avoid the warning, we limit
+		 *   the number of extents and repeat the hypercall.
+		 *   The following value is determined by evaluation.
+		 *   Time of one hypercall should be smaller than
+		 *   a vcpu time slice. The time with current
+		 *   MEMORYOP_MAX_EXTENTS is around 5 msec.
+		 *   If the following limit causes some issues,
+		 *   we should decrease this value.
+		 *
+		 *   Another way would be that start with small value and
+		 *   increase adoptively measuring hypercall time.
+		 *   It might be over-kill.
+		 */
+#define MEMORYOP_MAX_EXTENTS	(MEMORYOP_XENCOMM_LIMIT / 512)
+
+		while (nr_extents > 0) {
+			xen_ulong_t nr_tmp = nr_extents;
+			if (nr_tmp > MEMORYOP_MAX_EXTENTS)
+				nr_tmp = MEMORYOP_MAX_EXTENTS;
+
+			kern_op.nr_extents = nr_tmp;
+			desc = xencomm_map
+				(addr + nr_done * sizeof(*xen_guest_handle(kern_op.extent_start)),
+				 nr_tmp * sizeof(*xen_guest_handle(kern_op.extent_start)));
+			if (addr != NULL && nr_tmp > 0 && desc == NULL)
+				return nr_done > 0 ? nr_done : -ENOMEM;
+
+			set_xen_guest_handle(kern_op.extent_start,
+					     (void *)desc);
+
+			ret = xencomm_arch_hypercall_memory_op(cmd, desc_op);
+			xencomm_free(desc);
+			if (ret < 0)
+				return nr_done > 0 ? nr_done : ret;
+
+			nr_done += ret;
+			nr_extents -= ret;
+			if (ret < nr_tmp)
+				break;
+
+			/*
+			 * prevent softlock up message.
+			 * give cpu to soft lockup kernel thread.
+			 */
+			if (nr_extents > 0)
+				schedule();
+		}
+		ret = nr_done;
+		set_xen_guest_handle(kern_op.extent_start, addr);
+	}
+
+	if (copy_to_user(user_op, &kern_op, sizeof(xen_memory_reservation_t)))
+		return -EFAULT;
+
+	return ret;
+}
+
+static int
+xencomm_privcmd_memory_op(privcmd_hypercall_t *hypercall)
+{
+	const unsigned long cmd = hypercall->arg[0];
+	int ret = 0;
+
+	switch (cmd) {
+	case XENMEM_increase_reservation:
+	case XENMEM_decrease_reservation:
+	case XENMEM_populate_physmap:
+		return xencomm_privcmd_memory_reservation_op(hypercall);
+	case XENMEM_maximum_gpfn:
+	{
+		domid_t kern_domid;
+		domid_t __user *user_domid;
+		struct xencomm_handle *desc;
+
+		user_domid = (domid_t __user *)hypercall->arg[1];
+		if (copy_from_user(&kern_domid, user_domid, sizeof(domid_t)))
+			return -EFAULT;
+		desc = xencomm_map_no_alloc(&kern_domid, sizeof(kern_domid));
+
+		ret = xencomm_arch_hypercall_memory_op(cmd, desc);
+
+		return ret;
+	}
+	case XENMEM_translate_gpfn_list:
+	{
+		xen_translate_gpfn_list_t kern_op;
+		xen_translate_gpfn_list_t __user *user_op;
+		struct xencomm_handle *desc_gpfn = NULL;
+		struct xencomm_handle *desc_mfn = NULL;
+		struct xencomm_handle *desc_op;
+		void *addr;
+
+		user_op = (xen_translate_gpfn_list_t __user *)
+			hypercall->arg[1];
+		if (copy_from_user(&kern_op, user_op,
+		                   sizeof(xen_translate_gpfn_list_t)))
+			return -EFAULT;
+		desc_op = xencomm_map_no_alloc(&kern_op, sizeof(kern_op));
+
+		if (kern_op.nr_gpfns) {
+			/* gpfn_list.  */
+			addr = xen_guest_handle(kern_op.gpfn_list);
+
+			desc_gpfn = xencomm_map(addr, kern_op.nr_gpfns *
+			                     sizeof(*xen_guest_handle
+			                            (kern_op.gpfn_list)));
+			if (addr != NULL && kern_op.nr_gpfns > 0 &&
+			    desc_gpfn == NULL)
+				return -ENOMEM;
+			set_xen_guest_handle(kern_op.gpfn_list,
+			                     (void *)desc_gpfn);
+
+			/* mfn_list.  */
+			addr = xen_guest_handle(kern_op.mfn_list);
+
+			desc_mfn = xencomm_map(addr, kern_op.nr_gpfns *
+			                     sizeof(*xen_guest_handle
+			                            (kern_op.mfn_list)));
+			if (addr != NULL && kern_op.nr_gpfns > 0 &&
+			    desc_mfn == NULL) {
+				xencomm_free(desc_gpfn);
+				return -ENOMEM;
+			}
+
+			set_xen_guest_handle(kern_op.mfn_list,
+			                     (void *)desc_mfn);
+		}
+
+		ret = xencomm_arch_hypercall_memory_op(cmd, desc_op);
+
+		xencomm_free(desc_gpfn);
+		xencomm_free(desc_mfn);
+
+		if (ret != 0)
+			return ret;
+
+		return ret;
+	}
+	default:
+		printk("%s: unknown memory op %lu\n", __func__, cmd);
+		ret = -ENOSYS;
+	}
+	return ret;
+}
+
+static int
+xencomm_privcmd_xen_version(privcmd_hypercall_t *hypercall)
+{
+	int cmd = hypercall->arg[0];
+	void __user *arg = (void __user *)hypercall->arg[1];
+	struct xencomm_handle *desc;
+	size_t argsize;
+	int rc;
+
+	switch (cmd) {
+	case XENVER_version:
+		/* do not actually pass an argument */
+		return xencomm_arch_hypercall_xen_version(cmd, 0);
+	case XENVER_extraversion:
+		argsize = sizeof(xen_extraversion_t);
+		break;
+	case XENVER_compile_info:
+		argsize = sizeof(xen_compile_info_t);
+		break;
+	case XENVER_capabilities:
+		argsize = sizeof(xen_capabilities_info_t);
+		break;
+	case XENVER_changeset:
+		argsize = sizeof(xen_changeset_info_t);
+		break;
+	case XENVER_platform_parameters:
+		argsize = sizeof(xen_platform_parameters_t);
+		break;
+	case XENVER_pagesize:
+		argsize = (arg == NULL) ? 0 : sizeof(void *);
+		break;
+	case XENVER_get_features:
+		argsize = (arg == NULL) ? 0 : sizeof(xen_feature_info_t);
+		break;
+
+	default:
+		printk("%s: unknown version op %d\n", __func__, cmd);
+		return -ENOSYS;
+	}
+
+	desc = xencomm_map(arg, argsize);
+	if (arg != NULL && argsize > 0 && desc == NULL)
+		return -ENOMEM;
+
+	rc = xencomm_arch_hypercall_xen_version(cmd, desc);
+
+	xencomm_free(desc);
+
+	return rc;
+}
+
+static int
+xencomm_privcmd_event_channel_op(privcmd_hypercall_t *hypercall)
+{
+	int cmd = hypercall->arg[0];
+	struct xencomm_handle *desc;
+	unsigned int argsize;
+	int ret;
+
+	switch (cmd) {
+	case EVTCHNOP_alloc_unbound:
+		argsize = sizeof(evtchn_alloc_unbound_t);
+		break;
+
+	case EVTCHNOP_status:
+		argsize = sizeof(evtchn_status_t);
+		break;
+
+	default:
+		printk("%s: unknown EVTCHNOP %d\n", __func__, cmd);
+		return -EINVAL;
+	}
+
+	desc = xencomm_map((void *)hypercall->arg[1], argsize);
+	if ((void *)hypercall->arg[1] != NULL && argsize > 0 && desc == NULL)
+		return -ENOMEM;
+
+	ret = xencomm_arch_hypercall_event_channel_op(cmd, desc);
+
+	xencomm_free(desc);
+	return ret;
+}
+
+static int
+xencomm_privcmd_hvm_op_track_dirty_vram(privcmd_hypercall_t *hypercall)
+{
+#if 1
+	/*
+	 * At this moment HVMOP_track_dirty_vram isn't implemented
+	 * on xen/ia64 so that it just returns -ENOSYS.
+	 * Don't issue hypercall to get -ENOSYS.
+	 * When the hypercall is implemented, enable the following codes.
+	 */
+	return -ENOSYS;
+#else
+	int cmd = hypercall->arg[0];
+	struct xen_hvm_track_dirty_vram *user_op = (void*)hypercall->arg[1];
+	struct xen_hvm_track_dirty_vram kern_op;
+	struct xencomm_handle *desc;
+	struct xencomm_handle *bitmap_desc;
+	int ret;
+
+	BUG_ON(cmd != HVMOP_track_dirty_vram);
+	if (copy_from_user(&kern_op, user_op, sizeof(kern_op)))
+		return -EFAULT;
+	desc = xencomm_map_no_alloc(&kern_op, sizeof(kern_op));
+	bitmap_desc = xencomm_map(xen_guest_handle(kern_op.dirty_bitmap),
+				  kern_op.nr * sizeof(uint8_t));
+	if (bitmap_desc == NULL)
+		return -ENOMEM;
+	set_xen_guest_handle(kern_op.dirty_bitmap, (void*)bitmap_desc);
+	ret = xencomm_arch_hypercall_hvm_op(cmd, desc);
+	xencomm_free(bitmap_desc);
+
+	return ret;
+#endif
+}
+
+static int
+xencomm_privcmd_hvm_op(privcmd_hypercall_t *hypercall)
+{
+	int cmd = hypercall->arg[0];
+	struct xencomm_handle *desc;
+	unsigned int argsize;
+	int ret;
+
+	switch (cmd) {
+	case HVMOP_get_param:
+	case HVMOP_set_param:
+		argsize = sizeof(xen_hvm_param_t);
+		break;
+	case HVMOP_set_pci_intx_level:
+		argsize = sizeof(xen_hvm_set_pci_intx_level_t);
+		break;
+	case HVMOP_set_isa_irq_level:
+		argsize = sizeof(xen_hvm_set_isa_irq_level_t);
+		break;
+	case HVMOP_set_pci_link_route:
+		argsize = sizeof(xen_hvm_set_pci_link_route_t);
+		break;
+	case HVMOP_set_mem_type:
+		argsize = sizeof(xen_hvm_set_mem_type_t);
+		break;
+
+	case HVMOP_track_dirty_vram:
+		return xencomm_privcmd_hvm_op_track_dirty_vram(hypercall);
+
+	default:
+		printk("%s: unknown HVMOP %d\n", __func__, cmd);
+		return -EINVAL;
+	}
+
+	desc = xencomm_map((void *)hypercall->arg[1], argsize);
+	if ((void *)hypercall->arg[1] != NULL && argsize > 0 && desc == NULL)
+		return -ENOMEM;
+
+	ret = xencomm_arch_hypercall_hvm_op(cmd, desc);
+
+	xencomm_free(desc);
+	return ret;
+}
+
+static int
+xencomm_privcmd_sched_op(privcmd_hypercall_t *hypercall)
+{
+	int cmd = hypercall->arg[0];
+	struct xencomm_handle *desc;
+	unsigned int argsize;
+	int ret;
+
+	switch (cmd) {
+	case SCHEDOP_remote_shutdown:
+		argsize = sizeof(sched_remote_shutdown_t);
+		break;
+	default:
+		printk("%s: unknown SCHEDOP %d\n", __func__, cmd);
+		return -EINVAL;
+	}
+
+	desc = xencomm_map((void *)hypercall->arg[1], argsize);
+	if ((void *)hypercall->arg[1] != NULL && argsize > 0 && desc == NULL)
+		return -ENOMEM;
+
+	ret = xencomm_arch_hypercall_sched_op(cmd, desc);
+
+	xencomm_free(desc);
+	return ret;
+}
+
+static int
+xencomm_privcmd_ia64_dom0vp_op(privcmd_hypercall_t *hypercall)
+{
+	int cmd = hypercall->arg[0];
+	int ret;
+
+	switch (cmd) {
+	case IA64_DOM0VP_fpswa_revision: {
+		unsigned int revision;
+		unsigned int __user *revision_user =
+			(unsigned int* __user)hypercall->arg[1];
+		struct xencomm_handle *desc;
+		desc = xencomm_map(&revision, sizeof(revision));
+		if (desc == NULL)
+			return -ENOMEM;
+
+		ret = xencomm_arch_hypercall_fpswa_revision(desc);
+		xencomm_free(desc);
+		if (ret)
+			break;
+		if (copy_to_user(revision_user, &revision, sizeof(revision)))
+			ret = -EFAULT;
+		break;
+	}
+#ifdef CONFIG_XEN_IA64_EXPOSE_P2M
+	case IA64_DOM0VP_expose_foreign_p2m:
+		ret = xen_foreign_p2m_expose(hypercall);
+		break;
+#endif
+	default:
+		printk("%s: unknown IA64 DOM0VP op %d\n", __func__, cmd);
+		ret = -EINVAL;
+		break;
+	}
+	return ret;
+}
+
+static int
+xencomm_privcmd_ia64_debug_op(privcmd_hypercall_t *hypercall)
+{
+	int cmd = hypercall->arg[0];
+	unsigned long domain = hypercall->arg[1];
+	struct xencomm_handle *desc;
+	int ret;
+
+	switch (cmd) {
+	case XEN_IA64_DEBUG_OP_SET_FLAGS:
+	case XEN_IA64_DEBUG_OP_GET_FLAGS:
+		break;
+	default:
+		printk("%s: unknown IA64 DEBUGOP %d\n", __func__, cmd);
+		return -EINVAL;
+	}
+
+	desc = xencomm_map((void *)hypercall->arg[2],
+			   sizeof(xen_ia64_debug_op_t));
+	if (desc == NULL)
+		return -ENOMEM;
+
+	ret = xencomm_arch_hypercall_ia64_debug_op(cmd, domain, desc);
+
+	xencomm_free(desc);
+	return ret;	
+}
+
+static int
+xencomm_privcmd_ia64_physdev_op(privcmd_hypercall_t *hypercall)
+{
+	int cmd = hypercall->arg[0];
+	struct xencomm_handle *desc;
+	unsigned int argsize;
+	int ret;
+
+	switch (cmd) {
+	case PHYSDEVOP_map_pirq:
+		argsize = sizeof(physdev_map_pirq_t);
+		break;
+	case PHYSDEVOP_unmap_pirq:
+		argsize = sizeof(physdev_unmap_pirq_t);
+		break;
+	default:
+		printk("%s: unknown PHYSDEVOP %d\n", __func__, cmd);
+		return -EINVAL;
+	}
+
+	desc = xencomm_map((void *)hypercall->arg[1], argsize);
+	if ((void *)hypercall->arg[1] != NULL && argsize > 0 && desc == NULL)
+		return -ENOMEM;
+
+	ret = xencomm_arch_hypercall_physdev_op(cmd, desc);
+
+	xencomm_free(desc);
+	return ret;
+}
+
+int
+privcmd_hypercall(privcmd_hypercall_t *hypercall)
+{
+	switch (hypercall->op) {
+	case __HYPERVISOR_platform_op:
+		return xencomm_privcmd_platform_op(hypercall);
+	case __HYPERVISOR_domctl:
+		return xencomm_privcmd_domctl(hypercall);
+	case __HYPERVISOR_sysctl:
+		return xencomm_privcmd_sysctl(hypercall);
+	case __HYPERVISOR_xsm_op:
+		return xencomm_privcmd_xsm_op(hypercall);
+	case __HYPERVISOR_xen_version:
+		return xencomm_privcmd_xen_version(hypercall);
+	case __HYPERVISOR_memory_op:
+		return xencomm_privcmd_memory_op(hypercall);
+	case __HYPERVISOR_event_channel_op:
+		return xencomm_privcmd_event_channel_op(hypercall);
+	case __HYPERVISOR_hvm_op:
+		return xencomm_privcmd_hvm_op(hypercall);
+	case __HYPERVISOR_sched_op:
+		return xencomm_privcmd_sched_op(hypercall);
+	case __HYPERVISOR_ia64_dom0vp_op:
+		return xencomm_privcmd_ia64_dom0vp_op(hypercall);
+	case __HYPERVISOR_ia64_debug_op:
+		return xencomm_privcmd_ia64_debug_op(hypercall);
+	case __HYPERVISOR_physdev_op:
+		return xencomm_privcmd_ia64_physdev_op(hypercall);
+	default:
+		printk("%s: unknown hcall (%ld)\n", __func__, hypercall->op);
+		return -ENOSYS;
+	}
+}
+
diff -rpuN linux-2.6.18.8/arch/ia64/xen/xencomm.c linux-2.6.18-xen-3.3.0/arch/ia64/xen/xencomm.c
--- linux-2.6.18.8/arch/ia64/xen/xencomm.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/xen/xencomm.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,109 @@
+/*
+ * Copyright (C) 2006 Hollis Blanchard <hollisb@us.ibm.com>, IBM Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ * 
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ * 
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ */
+
+#include <linux/gfp.h>
+#include <linux/mm.h>
+#include <xen/interface/xen.h>
+#include <asm/page.h>
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+#include <asm/xen/xencomm.h>
+
+static unsigned long kernel_start_pa;
+
+void
+xencomm_initialize (void)
+{
+	kernel_start_pa = KERNEL_START - ia64_tpa(KERNEL_START);
+}
+
+/* Translate virtual address to physical address.  */
+unsigned long
+xencomm_vtop(unsigned long vaddr)
+{
+#ifndef CONFIG_VMX_GUEST
+	struct page *page;
+	struct vm_area_struct *vma;
+#endif
+
+	if (vaddr == 0)
+		return 0;
+
+#ifdef __ia64__
+	if (REGION_NUMBER(vaddr) == 5) {
+		pgd_t *pgd;
+		pud_t *pud;
+		pmd_t *pmd;
+		pte_t *ptep;
+
+		/* On ia64, TASK_SIZE refers to current.  It is not initialized
+		   during boot.
+		   Furthermore the kernel is relocatable and __pa() doesn't
+		   work on  addresses.  */
+		if (vaddr >= KERNEL_START
+		    && vaddr < (KERNEL_START + KERNEL_TR_PAGE_SIZE)) {
+			return vaddr - kernel_start_pa;
+		}
+
+		/* In kernel area -- virtually mapped.  */
+		pgd = pgd_offset_k(vaddr);
+		if (pgd_none(*pgd) || pgd_bad(*pgd))
+			return ~0UL;
+
+		pud = pud_offset(pgd, vaddr);
+		if (pud_none(*pud) || pud_bad(*pud))
+			return ~0UL;
+
+		pmd = pmd_offset(pud, vaddr);
+		if (pmd_none(*pmd) || pmd_bad(*pmd))
+			return ~0UL;
+
+		ptep = pte_offset_kernel(pmd, vaddr);
+		if (!ptep)
+			return ~0UL;
+
+		return (pte_val(*ptep) & _PFN_MASK) | (vaddr & ~PAGE_MASK);
+	}
+#endif
+
+	if (vaddr > TASK_SIZE) {
+		/* kernel address */
+		return __pa(vaddr);
+	}
+
+
+#ifdef CONFIG_VMX_GUEST
+	/* No privcmd within vmx guest.  */
+	return ~0UL;
+#else
+	/* XXX double-check (lack of) locking */
+	vma = find_extend_vma(current->mm, vaddr);
+	if (!vma)
+		return ~0UL;
+
+	/* We assume the page is modified.  */
+	page = follow_page(vma, vaddr, FOLL_WRITE | FOLL_TOUCH);
+	if (!page)
+		return ~0UL;
+
+	return (page_to_pfn(page) << PAGE_SHIFT) | (vaddr & ~PAGE_MASK);
+#endif
+}
diff -rpuN linux-2.6.18.8/arch/ia64/xen/xen_dma.c linux-2.6.18-xen-3.3.0/arch/ia64/xen/xen_dma.c
--- linux-2.6.18.8/arch/ia64/xen/xen_dma.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/xen/xen_dma.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,190 @@
+/*
+ * Copyright (C) 2007 Hewlett-Packard Development Company, L.P.
+ * 	Alex Williamson <alex.williamson@hp.com>
+ *
+ * Basic DMA mapping services for Xen guests.
+ * Based on arch/i386/kernel/pci-dma-xen.c.
+ *
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ */
+
+#include <linux/bitops.h>
+#include <linux/dma-mapping.h>
+#include <linux/mm.h>
+#include <asm/scatterlist.h>
+#include <xen/gnttab.h>
+#include <asm/gnttab_dma.h>
+
+#define IOMMU_BUG_ON(test)					\
+do {								\
+	if (unlikely(test)) {					\
+		printk(KERN_ALERT "Fatal DMA error!\n");	\
+		BUG();						\
+	}							\
+} while (0)
+
+static int check_pages_physically_contiguous(unsigned long pfn, 
+					     unsigned int offset,
+					     size_t length)
+{
+	unsigned long next_bus;
+	int i;
+	int nr_pages;
+
+	next_bus = pfn_to_mfn_for_dma(pfn);
+	nr_pages = (offset + length + PAGE_SIZE-1) >> PAGE_SHIFT;
+
+	for (i = 1; i < nr_pages; i++) {
+		if (pfn_to_mfn_for_dma(++pfn) != ++next_bus) 
+			return 0;
+	}
+	return 1;
+}
+
+int range_straddles_page_boundary(paddr_t p, size_t size)
+{
+	extern unsigned long *contiguous_bitmap;
+	unsigned long pfn = p >> PAGE_SHIFT;
+	unsigned int offset = p & ~PAGE_MASK;
+
+	if (!is_running_on_xen())
+		return 0;
+
+	if (offset + size <= PAGE_SIZE)
+		return 0;
+	if (test_bit(pfn, contiguous_bitmap))
+		return 0;
+	if (check_pages_physically_contiguous(pfn, offset, size))
+		return 0;
+	return 1;
+}
+
+/*
+ * This should be broken out of swiotlb and put in a common place
+ * when merged with upstream Linux.
+ */
+static inline int
+address_needs_mapping(struct device *dev, dma_addr_t addr)
+{
+	dma_addr_t mask = 0xffffffff;
+
+	/* If the device has a mask, use it, otherwise default to 32 bits */
+	if (dev && dev->dma_mask)
+		mask = *dev->dma_mask;
+	return (addr & ~mask) != 0;
+}
+
+int
+xen_map_sg(struct device *dev, struct scatterlist *sg, int nents,
+	   int direction)
+{
+	int i;
+
+	for (i = 0 ; i < nents ; i++) {
+		sg[i].dma_address = gnttab_dma_map_page(sg[i].page) + sg[i].offset;
+		sg[i].dma_length  = sg[i].length;
+
+		IOMMU_BUG_ON(address_needs_mapping(dev, sg[i].dma_address));
+		IOMMU_BUG_ON(range_straddles_page_boundary( 
+			page_to_pseudophys(sg[i].page) + sg[i].offset,
+			sg[i].length));
+	}
+
+	return nents;
+}
+EXPORT_SYMBOL(xen_map_sg);
+
+void
+xen_unmap_sg(struct device *dev, struct scatterlist *sg, int nents,
+	     int direction)
+{
+	int i;
+	for (i = 0; i < nents; i++)
+		__gnttab_dma_unmap_page(sg[i].page);
+}
+EXPORT_SYMBOL(xen_unmap_sg);
+
+int
+xen_dma_mapping_error(dma_addr_t dma_addr)
+{
+	return 0;
+}
+EXPORT_SYMBOL(xen_dma_mapping_error);
+
+int
+xen_dma_supported(struct device *dev, u64 mask)
+{
+	return 1;
+}
+EXPORT_SYMBOL(xen_dma_supported);
+
+void *
+xen_alloc_coherent(struct device *dev, size_t size,
+		   dma_addr_t *dma_handle, gfp_t gfp)
+{
+	unsigned long vaddr;
+	unsigned int order = get_order(size);
+
+	vaddr = __get_free_pages(gfp, order);
+
+	if (!vaddr)
+		return NULL;
+
+	if (xen_create_contiguous_region(vaddr, order,
+					 fls64(dev->coherent_dma_mask))) {
+		free_pages(vaddr, order);
+		return NULL;
+	}
+
+	memset((void *)vaddr, 0, size);
+	*dma_handle = virt_to_bus((void *)vaddr);
+
+	return (void *)vaddr;
+}
+EXPORT_SYMBOL(xen_alloc_coherent);
+
+void
+xen_free_coherent(struct device *dev, size_t size,
+		      void *vaddr, dma_addr_t dma_handle)
+{
+	unsigned int order =  get_order(size);
+
+	xen_destroy_contiguous_region((unsigned long)vaddr, order);
+	free_pages((unsigned long)vaddr, order);
+}
+EXPORT_SYMBOL(xen_free_coherent);
+
+dma_addr_t
+xen_map_single(struct device *dev, void *ptr, size_t size,
+	       int direction)
+{
+	dma_addr_t dma_addr = gnttab_dma_map_virt(ptr);
+
+	IOMMU_BUG_ON(range_straddles_page_boundary(__pa(ptr), size));
+	IOMMU_BUG_ON(address_needs_mapping(dev, dma_addr));
+
+	return dma_addr;
+}
+EXPORT_SYMBOL(xen_map_single);
+
+void
+xen_unmap_single(struct device *dev, dma_addr_t dma_addr, size_t size,
+		 int direction)
+{
+	gnttab_dma_unmap_page(dma_addr);
+}
+EXPORT_SYMBOL(xen_unmap_single);
diff -rpuN linux-2.6.18.8/arch/ia64/xen/xenentry.S linux-2.6.18-xen-3.3.0/arch/ia64/xen/xenentry.S
--- linux-2.6.18.8/arch/ia64/xen/xenentry.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/xen/xenentry.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,931 @@
+/*
+ * ia64/xen/entry.S
+ *
+ * Alternate kernel routines for Xen.  Heavily leveraged from
+ *   ia64/kernel/entry.S
+ *
+ * Copyright (C) 2005 Hewlett-Packard Co
+ *	Dan Magenheimer <dan.magenheimer@.hp.com>
+ */
+
+#include <asm/asmmacro.h>
+#include <asm/cache.h>
+#include <asm/errno.h>
+#include <asm/kregs.h>
+#include <asm/asm-offsets.h>
+#include <asm/pgtable.h>
+#include <asm/percpu.h>
+#include <asm/processor.h>
+#include <asm/thread_info.h>
+#include <asm/unistd.h>
+
+#ifdef CONFIG_XEN
+#include "xenminstate.h"
+#else
+#include "minstate.h"
+#endif
+
+/*
+ * prev_task <- ia64_switch_to(struct task_struct *next)
+ *	With Ingo's new scheduler, interrupts are disabled when this routine gets
+ *	called.  The code starting at .map relies on this.  The rest of the code
+ *	doesn't care about the interrupt masking status.
+ */
+#ifdef CONFIG_XEN
+GLOBAL_ENTRY(xen_switch_to)
+	.prologue
+	alloc r16=ar.pfs,1,0,0,0
+	movl r22=running_on_xen;;
+	ld4 r22=[r22];;
+	cmp.eq p7,p0=r22,r0
+(p7)	br.cond.sptk.many __ia64_switch_to;;
+#else
+GLOBAL_ENTRY(ia64_switch_to)
+	.prologue
+	alloc r16=ar.pfs,1,0,0,0
+#endif
+	DO_SAVE_SWITCH_STACK
+	.body
+
+	adds r22=IA64_TASK_THREAD_KSP_OFFSET,r13
+	movl r25=init_task
+	mov r27=IA64_KR(CURRENT_STACK)
+	adds r21=IA64_TASK_THREAD_KSP_OFFSET,in0
+	dep r20=0,in0,61,3		// physical address of "next"
+	;;
+	st8 [r22]=sp			// save kernel stack pointer of old task
+	shr.u r26=r20,IA64_GRANULE_SHIFT
+	cmp.eq p7,p6=r25,in0
+	;;
+	/*
+	 * If we've already mapped this task's page, we can skip doing it again.
+	 */
+(p6)	cmp.eq p7,p6=r26,r27
+(p6)	br.cond.dpnt .map
+	;;
+.done:
+	ld8 sp=[r21]			// load kernel stack pointer of new task
+#ifdef CONFIG_XEN
+	// update "current" application register
+	mov r8=IA64_KR_CURRENT
+	mov r9=in0;;
+	XEN_HYPER_SET_KR
+#else
+	mov IA64_KR(CURRENT)=in0	// update "current" application register
+#endif
+	mov r8=r13			// return pointer to previously running task
+	mov r13=in0			// set "current" pointer
+	;;
+	DO_LOAD_SWITCH_STACK
+
+#ifdef CONFIG_SMP
+	sync.i				// ensure "fc"s done by this CPU are visible on other CPUs
+#endif
+	br.ret.sptk.many rp		// boogie on out in new context
+
+.map:
+#ifdef CONFIG_XEN
+    movl r25=XSI_PSR_IC			// clear psr.ic
+    ;;
+    st4 [r25]=r0
+    ;;
+#else
+	rsm psr.ic			// interrupts (psr.i) are already disabled here
+#endif
+	movl r25=PAGE_KERNEL
+	;;
+	srlz.d
+	or r23=r25,r20			// construct PA | page properties
+	mov r25=IA64_GRANULE_SHIFT<<2
+	;;
+#ifdef CONFIG_XEN
+	movl r8=XSI_ITIR
+	;;
+	st8 [r8]=r25
+	;;
+	movl r8=XSI_IFA
+	;;
+	st8 [r8]=in0			 // VA of next task...
+	;;
+	mov r25=IA64_TR_CURRENT_STACK
+	// remember last page we mapped...
+	mov r8=IA64_KR_CURRENT_STACK
+	mov r9=r26;;
+	XEN_HYPER_SET_KR;;
+#else
+	mov cr.itir=r25
+	mov cr.ifa=in0			// VA of next task...
+	;;
+	mov r25=IA64_TR_CURRENT_STACK
+	mov IA64_KR(CURRENT_STACK)=r26	// remember last page we mapped...
+#endif
+	;;
+	itr.d dtr[r25]=r23		// wire in new mapping...
+#ifdef CONFIG_XEN
+	;;
+	srlz.d
+	mov r9=1
+	movl r8=XSI_PSR_IC
+	;;
+	st4 [r8]=r9
+	;;
+#else    
+	ssm psr.ic			// reenable the psr.ic bit
+	;;
+	srlz.d
+#endif
+	br.cond.sptk .done
+#ifdef CONFIG_XEN
+END(xen_switch_to)
+#else
+END(ia64_switch_to)
+#endif
+
+	/*
+	 * Invoke a system call, but do some tracing before and after the call.
+	 * We MUST preserve the current register frame throughout this routine
+	 * because some system calls (such as ia64_execve) directly
+	 * manipulate ar.pfs.
+	 */
+#ifdef CONFIG_XEN
+GLOBAL_ENTRY(xen_trace_syscall)
+	PT_REGS_UNWIND_INFO(0)
+	movl r16=running_on_xen;;
+	ld4 r16=[r16];;
+	cmp.eq p7,p0=r16,r0
+(p7)	br.cond.sptk.many __ia64_trace_syscall;;
+#else
+GLOBAL_ENTRY(ia64_trace_syscall)
+	PT_REGS_UNWIND_INFO(0)
+#endif
+	/*
+	 * We need to preserve the scratch registers f6-f11 in case the system
+	 * call is sigreturn.
+	 */
+	adds r16=PT(F6)+16,sp
+	adds r17=PT(F7)+16,sp
+	;;
+ 	stf.spill [r16]=f6,32
+ 	stf.spill [r17]=f7,32
+	;;
+ 	stf.spill [r16]=f8,32
+ 	stf.spill [r17]=f9,32
+	;;
+ 	stf.spill [r16]=f10
+ 	stf.spill [r17]=f11
+	br.call.sptk.many rp=syscall_trace_enter // give parent a chance to catch syscall args
+	adds r16=PT(F6)+16,sp
+	adds r17=PT(F7)+16,sp
+	;;
+	ldf.fill f6=[r16],32
+	ldf.fill f7=[r17],32
+	;;
+	ldf.fill f8=[r16],32
+	ldf.fill f9=[r17],32
+	;;
+	ldf.fill f10=[r16]
+	ldf.fill f11=[r17]
+	// the syscall number may have changed, so re-load it and re-calculate the
+	// syscall entry-point:
+	adds r15=PT(R15)+16,sp			// r15 = &pt_regs.r15 (syscall #)
+	;;
+	ld8 r15=[r15]
+	mov r3=NR_syscalls - 1
+	;;
+	adds r15=-1024,r15
+	movl r16=sys_call_table
+	;;
+	shladd r20=r15,3,r16			// r20 = sys_call_table + 8*(syscall-1024)
+	cmp.leu p6,p7=r15,r3
+	;;
+(p6)	ld8 r20=[r20]				// load address of syscall entry point
+(p7)	movl r20=sys_ni_syscall
+	;;
+	mov b6=r20
+	br.call.sptk.many rp=b6			// do the syscall
+.strace_check_retval:
+	cmp.lt p6,p0=r8,r0			// syscall failed?
+	adds r2=PT(R8)+16,sp			// r2 = &pt_regs.r8
+	adds r3=PT(R10)+16,sp			// r3 = &pt_regs.r10
+	mov r10=0
+(p6)	br.cond.sptk strace_error		// syscall failed ->
+	;;					// avoid RAW on r10
+.strace_save_retval:
+.mem.offset 0,0; st8.spill [r2]=r8		// store return value in slot for r8
+.mem.offset 8,0; st8.spill [r3]=r10		// clear error indication in slot for r10
+	br.call.sptk.many rp=syscall_trace_leave // give parent a chance to catch return value
+.ret3:
+(pUStk)	cmp.eq.unc p6,p0=r0,r0			// p6 <- pUStk
+	br.cond.sptk .work_pending_syscall_end
+
+strace_error:
+	ld8 r3=[r2]				// load pt_regs.r8
+	sub r9=0,r8				// negate return value to get errno value
+	;;
+	cmp.ne p6,p0=r3,r0			// is pt_regs.r8!=0?
+	adds r3=16,r2				// r3=&pt_regs.r10
+	;;
+(p6)	mov r10=-1
+(p6)	mov r8=r9
+	br.cond.sptk .strace_save_retval
+#ifdef CONFIG_XEN
+END(xen_trace_syscall)
+#else
+END(ia64_trace_syscall)
+#endif
+
+#ifdef CONFIG_XEN
+GLOBAL_ENTRY(xen_ret_from_clone)
+	PT_REGS_UNWIND_INFO(0)
+	movl r16=running_on_xen;;
+	ld4 r16=[r16];;
+	cmp.eq p7,p0=r16,r0
+(p7)	br.cond.sptk.many __ia64_ret_from_clone;;
+#else	
+GLOBAL_ENTRY(ia64_ret_from_clone)
+	PT_REGS_UNWIND_INFO(0)
+#endif	
+{	/*
+	 * Some versions of gas generate bad unwind info if the first instruction of a
+	 * procedure doesn't go into the first slot of a bundle.  This is a workaround.
+	 */
+	nop.m 0
+	nop.i 0
+	/*
+	 * We need to call schedule_tail() to complete the scheduling process.
+	 * Called by ia64_switch_to() after do_fork()->copy_thread().  r8 contains the
+	 * address of the previously executing task.
+	 */
+	br.call.sptk.many rp=ia64_invoke_schedule_tail
+}
+.ret8:
+	adds r2=TI_FLAGS+IA64_TASK_SIZE,r13
+	;;
+	ld4 r2=[r2]
+	;;
+	mov r8=0
+	and r2=_TIF_SYSCALL_TRACEAUDIT,r2
+	;;
+	cmp.ne p6,p0=r2,r0
+(p6)	br.cond.spnt .strace_check_retval
+	;;					// added stop bits to prevent r8 dependency
+#ifdef CONFIG_XEN
+	br.cond.sptk ia64_ret_from_syscall
+END(xen_ret_from_clone)
+#else
+END(ia64_ret_from_clone)
+#endif		
+/*
+ * ia64_leave_syscall(): Same as ia64_leave_kernel, except that it doesn't
+ *	need to switch to bank 0 and doesn't restore the scratch registers.
+ *	To avoid leaking kernel bits, the scratch registers are set to
+ *	the following known-to-be-safe values:
+ *
+ *		  r1: restored (global pointer)
+ *		  r2: cleared
+ *		  r3: 1 (when returning to user-level)
+ *	      r8-r11: restored (syscall return value(s))
+ *		 r12: restored (user-level stack pointer)
+ *		 r13: restored (user-level thread pointer)
+ *		 r14: set to __kernel_syscall_via_epc
+ *		 r15: restored (syscall #)
+ *	     r16-r17: cleared
+ *		 r18: user-level b6
+ *		 r19: cleared
+ *		 r20: user-level ar.fpsr
+ *		 r21: user-level b0
+ *		 r22: cleared
+ *		 r23: user-level ar.bspstore
+ *		 r24: user-level ar.rnat
+ *		 r25: user-level ar.unat
+ *		 r26: user-level ar.pfs
+ *		 r27: user-level ar.rsc
+ *		 r28: user-level ip
+ *		 r29: user-level psr
+ *		 r30: user-level cfm
+ *		 r31: user-level pr
+ *	      f6-f11: cleared
+ *		  pr: restored (user-level pr)
+ *		  b0: restored (user-level rp)
+ *	          b6: restored
+ *		  b7: set to __kernel_syscall_via_epc
+ *	     ar.unat: restored (user-level ar.unat)
+ *	      ar.pfs: restored (user-level ar.pfs)
+ *	      ar.rsc: restored (user-level ar.rsc)
+ *	     ar.rnat: restored (user-level ar.rnat)
+ *	 ar.bspstore: restored (user-level ar.bspstore)
+ *	     ar.fpsr: restored (user-level ar.fpsr)
+ *	      ar.ccv: cleared
+ *	      ar.csd: cleared
+ *	      ar.ssd: cleared
+ */
+#ifdef CONFIG_XEN
+GLOBAL_ENTRY(xen_leave_syscall)
+	PT_REGS_UNWIND_INFO(0)
+	movl r22=running_on_xen;;
+	ld4 r22=[r22];;
+	cmp.eq p7,p0=r22,r0
+(p7)	br.cond.sptk.many __ia64_leave_syscall;;
+#else
+ENTRY(ia64_leave_syscall)
+	PT_REGS_UNWIND_INFO(0)
+#endif
+	/*
+	 * work.need_resched etc. mustn't get changed by this CPU before it returns to
+	 * user- or fsys-mode, hence we disable interrupts early on.
+	 *
+	 * p6 controls whether current_thread_info()->flags needs to be check for
+	 * extra work.  We always check for extra work when returning to user-level.
+	 * With CONFIG_PREEMPT, we also check for extra work when the preempt_count
+	 * is 0.  After extra work processing has been completed, execution
+	 * resumes at .work_processed_syscall with p6 set to 1 if the extra-work-check
+	 * needs to be redone.
+	 */
+#ifdef CONFIG_PREEMPT
+	rsm psr.i				// disable interrupts
+	cmp.eq pLvSys,p0=r0,r0			// pLvSys=1: leave from syscall
+(pKStk) adds r20=TI_PRE_COUNT+IA64_TASK_SIZE,r13
+	;;
+	.pred.rel.mutex pUStk,pKStk
+(pKStk) ld4 r21=[r20]			// r21 <- preempt_count
+(pUStk)	mov r21=0			// r21 <- 0
+	;;
+	cmp.eq p6,p0=r21,r0		// p6 <- pUStk || (preempt_count == 0)
+#else /* !CONFIG_PREEMPT */
+#ifdef CONFIG_XEN
+	movl r2=XSI_PSR_I_ADDR
+	mov r18=1
+	;;
+	ld8 r2=[r2]
+	;;
+(pUStk)	st1 [r2]=r18
+#else
+(pUStk)	rsm psr.i
+#endif
+	cmp.eq pLvSys,p0=r0,r0		// pLvSys=1: leave from syscall
+(pUStk)	cmp.eq.unc p6,p0=r0,r0		// p6 <- pUStk
+#endif
+.work_processed_syscall:
+	adds r2=PT(LOADRS)+16,r12
+	adds r3=PT(AR_BSPSTORE)+16,r12
+	adds r18=TI_FLAGS+IA64_TASK_SIZE,r13
+	;;
+(p6)	ld4 r31=[r18]				// load current_thread_info()->flags
+	ld8 r19=[r2],PT(B6)-PT(LOADRS)		// load ar.rsc value for "loadrs"
+	nop.i 0
+	;;
+	mov r16=ar.bsp				// M2  get existing backing store pointer
+	ld8 r18=[r2],PT(R9)-PT(B6)		// load b6
+(p6)	and r15=TIF_WORK_MASK,r31		// any work other than TIF_SYSCALL_TRACE?
+	;;
+	ld8 r23=[r3],PT(R11)-PT(AR_BSPSTORE)	// load ar.bspstore (may be garbage)
+(p6)	cmp4.ne.unc p6,p0=r15, r0		// any special work pending?
+(p6)	br.cond.spnt .work_pending_syscall
+	;;
+	// start restoring the state saved on the kernel stack (struct pt_regs):
+	ld8 r9=[r2],PT(CR_IPSR)-PT(R9)
+	ld8 r11=[r3],PT(CR_IIP)-PT(R11)
+(pNonSys) break 0		//      bug check: we shouldn't be here if pNonSys is TRUE!
+	;;
+	invala			// M0|1 invalidate ALAT
+#ifdef CONFIG_XEN
+	movl r28=XSI_PSR_I_ADDR
+	movl r29=XSI_PSR_IC
+	;;
+	ld8 r28=[r28]
+	mov r30=1
+	;;
+	st1	[r28]=r30
+	st4	[r29]=r0	// note: clears both vpsr.i and vpsr.ic!
+	;;
+#else
+	rsm psr.i | psr.ic	// M2   turn off interrupts and interruption collection
+#endif
+	cmp.eq p9,p0=r0,r0	// A    set p9 to indicate that we should restore cr.ifs
+
+	ld8 r29=[r2],16		// M0|1 load cr.ipsr
+	ld8 r28=[r3],16		// M0|1 load cr.iip
+	mov r22=r0		// A    clear r22
+	;;
+	ld8 r30=[r2],16		// M0|1 load cr.ifs
+	ld8 r25=[r3],16		// M0|1 load ar.unat
+(pUStk) add r14=IA64_TASK_THREAD_ON_USTACK_OFFSET,r13
+	;;
+	ld8 r26=[r2],PT(B0)-PT(AR_PFS)	// M0|1 load ar.pfs
+#ifdef CONFIG_XEN
+(pKStk)	mov r21=r8
+(pKStk)	XEN_HYPER_GET_PSR
+	;;
+(pKStk)	mov r22=r8
+(pKStk)	mov r8=r21
+	;;
+#else    
+(pKStk)	mov r22=psr			// M2   read PSR now that interrupts are disabled
+#endif
+	nop 0
+	;;
+	ld8 r21=[r2],PT(AR_RNAT)-PT(B0) // M0|1 load b0
+	ld8 r27=[r3],PT(PR)-PT(AR_RSC)	// M0|1 load ar.rsc
+	mov f6=f0			// F    clear f6
+	;;
+	ld8 r24=[r2],PT(AR_FPSR)-PT(AR_RNAT)	// M0|1 load ar.rnat (may be garbage)
+	ld8 r31=[r3],PT(R1)-PT(PR)		// M0|1 load predicates
+	mov f7=f0				// F    clear f7
+	;;
+	ld8 r20=[r2],PT(R12)-PT(AR_FPSR)	// M0|1 load ar.fpsr
+	ld8.fill r1=[r3],16			// M0|1 load r1
+(pUStk) mov r17=1				// A
+	;;
+(pUStk) st1 [r14]=r17				// M2|3
+	ld8.fill r13=[r3],16			// M0|1
+	mov f8=f0				// F    clear f8
+	;;
+	ld8.fill r12=[r2]			// M0|1 restore r12 (sp)
+	ld8.fill r15=[r3]			// M0|1 restore r15
+	mov b6=r18				// I0   restore b6
+
+	addl r17=THIS_CPU(ia64_phys_stacked_size_p8),r0 // A
+	mov f9=f0					// F    clear f9
+(pKStk) br.cond.dpnt.many skip_rbs_switch		// B
+
+	srlz.d				// M0   ensure interruption collection is off (for cover)
+	shr.u r18=r19,16		// I0|1 get byte size of existing "dirty" partition
+#ifdef CONFIG_XEN
+	XEN_HYPER_COVER;
+#else
+	cover				// B    add current frame into dirty partition & set cr.ifs
+#endif
+	;;
+(pUStk) ld4 r17=[r17]			// M0|1 r17 = cpu_data->phys_stacked_size_p8
+	mov r19=ar.bsp			// M2   get new backing store pointer
+	mov f10=f0			// F    clear f10
+
+	nop.m 0
+	movl r14=__kernel_syscall_via_epc // X
+	;;
+	mov.m ar.csd=r0			// M2   clear ar.csd
+	mov.m ar.ccv=r0			// M2   clear ar.ccv
+	mov b7=r14			// I0   clear b7 (hint with __kernel_syscall_via_epc)
+
+	mov.m ar.ssd=r0			// M2   clear ar.ssd
+	mov f11=f0			// F    clear f11
+	br.cond.sptk.many rbs_switch	// B
+#ifdef CONFIG_XEN
+END(xen_leave_syscall)
+#else
+END(ia64_leave_syscall)
+#endif
+
+#ifdef CONFIG_XEN
+GLOBAL_ENTRY(xen_leave_kernel)
+	PT_REGS_UNWIND_INFO(0)
+	movl r22=running_on_xen;;
+	ld4 r22=[r22];;
+	cmp.eq p7,p0=r22,r0
+(p7)	br.cond.sptk.many __ia64_leave_kernel;;
+#else
+GLOBAL_ENTRY(ia64_leave_kernel)
+	PT_REGS_UNWIND_INFO(0)
+#endif
+	/*
+	 * work.need_resched etc. mustn't get changed by this CPU before it returns to
+	 * user- or fsys-mode, hence we disable interrupts early on.
+	 *
+	 * p6 controls whether current_thread_info()->flags needs to be check for
+	 * extra work.  We always check for extra work when returning to user-level.
+	 * With CONFIG_PREEMPT, we also check for extra work when the preempt_count
+	 * is 0.  After extra work processing has been completed, execution
+	 * resumes at .work_processed_syscall with p6 set to 1 if the extra-work-check
+	 * needs to be redone.
+	 */
+#ifdef CONFIG_PREEMPT
+	rsm psr.i				// disable interrupts
+	cmp.eq p0,pLvSys=r0,r0			// pLvSys=0: leave from kernel
+(pKStk)	adds r20=TI_PRE_COUNT+IA64_TASK_SIZE,r13
+	;;
+	.pred.rel.mutex pUStk,pKStk
+(pKStk)	ld4 r21=[r20]			// r21 <- preempt_count
+(pUStk)	mov r21=0			// r21 <- 0
+	;;
+	cmp.eq p6,p0=r21,r0		// p6 <- pUStk || (preempt_count == 0)
+#else
+#ifdef CONFIG_XEN
+(pUStk)	movl r17=XSI_PSR_I_ADDR
+(pUStk)	mov r31=1
+		;;
+(pUStk) 	ld8 r17=[r17]
+		;;
+(pUStk)	st1 [r17]=r31
+	;;
+#else
+(pUStk)	rsm psr.i
+#endif
+	cmp.eq p0,pLvSys=r0,r0		// pLvSys=0: leave from kernel
+(pUStk)	cmp.eq.unc p6,p0=r0,r0		// p6 <- pUStk
+#endif
+.work_processed_kernel:
+	adds r17=TI_FLAGS+IA64_TASK_SIZE,r13
+	;;
+(p6)	ld4 r31=[r17]				// load current_thread_info()->flags
+	adds r21=PT(PR)+16,r12
+	;;
+
+	lfetch [r21],PT(CR_IPSR)-PT(PR)
+	adds r2=PT(B6)+16,r12
+	adds r3=PT(R16)+16,r12
+	;;
+	lfetch [r21]
+	ld8 r28=[r2],8		// load b6
+	adds r29=PT(R24)+16,r12
+
+	ld8.fill r16=[r3],PT(AR_CSD)-PT(R16)
+	adds r30=PT(AR_CCV)+16,r12
+(p6)	and r19=TIF_WORK_MASK,r31		// any work other than TIF_SYSCALL_TRACE?
+	;;
+	ld8.fill r24=[r29]
+	ld8 r15=[r30]		// load ar.ccv
+(p6)	cmp4.ne.unc p6,p0=r19, r0		// any special work pending?
+	;;
+	ld8 r29=[r2],16		// load b7
+	ld8 r30=[r3],16		// load ar.csd
+(p6)	br.cond.spnt .work_pending
+	;;
+	ld8 r31=[r2],16		// load ar.ssd
+	ld8.fill r8=[r3],16
+	;;
+	ld8.fill r9=[r2],16
+	ld8.fill r10=[r3],PT(R17)-PT(R10)
+	;;
+	ld8.fill r11=[r2],PT(R18)-PT(R11)
+	ld8.fill r17=[r3],16
+	;;
+	ld8.fill r18=[r2],16
+	ld8.fill r19=[r3],16
+	;;
+	ld8.fill r20=[r2],16
+	ld8.fill r21=[r3],16
+	mov ar.csd=r30
+	mov ar.ssd=r31
+	;;
+#ifdef CONFIG_XEN
+	movl r23=XSI_PSR_I_ADDR
+	movl r22=XSI_PSR_IC
+	;;
+	ld8 r23=[r23]
+	mov r25=1
+	;;
+	st1 [r23]=r25
+	st4 [r22]=r0		// note: clears both vpsr.i and vpsr.ic!
+	;;
+#else
+	rsm psr.i | psr.ic	// initiate turning off of interrupt and interruption collection
+#endif
+	invala			// invalidate ALAT
+	;;
+	ld8.fill r22=[r2],24
+	ld8.fill r23=[r3],24
+	mov b6=r28
+	;;
+	ld8.fill r25=[r2],16
+	ld8.fill r26=[r3],16
+	mov b7=r29
+	;;
+	ld8.fill r27=[r2],16
+	ld8.fill r28=[r3],16
+	;;
+	ld8.fill r29=[r2],16
+	ld8.fill r30=[r3],24
+	;;
+	ld8.fill r31=[r2],PT(F9)-PT(R31)
+	adds r3=PT(F10)-PT(F6),r3
+	;;
+	ldf.fill f9=[r2],PT(F6)-PT(F9)
+	ldf.fill f10=[r3],PT(F8)-PT(F10)
+	;;
+	ldf.fill f6=[r2],PT(F7)-PT(F6)
+	;;
+	ldf.fill f7=[r2],PT(F11)-PT(F7)
+	ldf.fill f8=[r3],32
+	;;
+	srlz.d	// ensure that inter. collection is off (VHPT is don't care, since text is pinned)
+	mov ar.ccv=r15
+	;;
+	ldf.fill f11=[r2]
+#ifdef CONFIG_XEN
+	;;
+	// r16-r31 all now hold bank1 values
+	mov r15=ar.unat
+	movl r2=XSI_BANK1_R16
+	movl r3=XSI_BANK1_R16+8
+	;;
+.mem.offset 0,0; st8.spill [r2]=r16,16
+.mem.offset 8,0; st8.spill [r3]=r17,16
+	;;
+.mem.offset 0,0; st8.spill [r2]=r18,16
+.mem.offset 8,0; st8.spill [r3]=r19,16
+	;;
+.mem.offset 0,0; st8.spill [r2]=r20,16
+.mem.offset 8,0; st8.spill [r3]=r21,16
+	;;
+.mem.offset 0,0; st8.spill [r2]=r22,16
+.mem.offset 8,0; st8.spill [r3]=r23,16
+	;;
+.mem.offset 0,0; st8.spill [r2]=r24,16
+.mem.offset 8,0; st8.spill [r3]=r25,16
+	;;
+.mem.offset 0,0; st8.spill [r2]=r26,16
+.mem.offset 8,0; st8.spill [r3]=r27,16
+	;;
+.mem.offset 0,0; st8.spill [r2]=r28,16
+.mem.offset 8,0; st8.spill [r3]=r29,16
+	;;
+.mem.offset 0,0; st8.spill [r2]=r30,16
+.mem.offset 8,0; st8.spill [r3]=r31,16
+	;;
+	mov r3=ar.unat
+	movl r2=XSI_B1NAT
+	;;
+	st8 [r2]=r3
+	mov ar.unat=r15
+	movl r2=XSI_BANKNUM;;
+	st4 [r2]=r0;
+#else
+	bsw.0			// switch back to bank 0 (no stop bit required beforehand...)
+#endif
+	;;
+(pUStk)	mov r18=IA64_KR(CURRENT)// M2 (12 cycle read latency)
+	adds r16=PT(CR_IPSR)+16,r12
+	adds r17=PT(CR_IIP)+16,r12
+
+#ifdef CONFIG_XEN    
+(pKStk)	mov r29=r8
+(pKStk)	XEN_HYPER_GET_PSR
+	;;
+(pKStk)	mov r22=r8
+(pKStk)	mov r8=r29
+	;;
+#else
+(pKStk)	mov r22=psr		// M2 read PSR now that interrupts are disabled
+#endif
+	nop.i 0
+	nop.i 0
+	;;
+	ld8 r29=[r16],16	// load cr.ipsr
+	ld8 r28=[r17],16	// load cr.iip
+	;;
+	ld8 r30=[r16],16	// load cr.ifs
+	ld8 r25=[r17],16	// load ar.unat
+	;;
+	ld8 r26=[r16],16	// load ar.pfs
+	ld8 r27=[r17],16	// load ar.rsc
+	cmp.eq p9,p0=r0,r0	// set p9 to indicate that we should restore cr.ifs
+	;;
+	ld8 r24=[r16],16	// load ar.rnat (may be garbage)
+	ld8 r23=[r17],16	// load ar.bspstore (may be garbage)
+	;;
+	ld8 r31=[r16],16	// load predicates
+	ld8 r21=[r17],16	// load b0
+	;;
+	ld8 r19=[r16],16	// load ar.rsc value for "loadrs"
+	ld8.fill r1=[r17],16	// load r1
+	;;
+	ld8.fill r12=[r16],16
+	ld8.fill r13=[r17],16
+(pUStk)	adds r18=IA64_TASK_THREAD_ON_USTACK_OFFSET,r18
+	;;
+	ld8 r20=[r16],16	// ar.fpsr
+	ld8.fill r15=[r17],16
+	;;
+	ld8.fill r14=[r16],16
+	ld8.fill r2=[r17]
+(pUStk)	mov r17=1
+	;;
+	ld8.fill r3=[r16]
+(pUStk)	st1 [r18]=r17		// restore current->thread.on_ustack
+	shr.u r18=r19,16	// get byte size of existing "dirty" partition
+	;;
+	mov r16=ar.bsp		// get existing backing store pointer
+	addl r17=THIS_CPU(ia64_phys_stacked_size_p8),r0
+	;;
+	ld4 r17=[r17]		// r17 = cpu_data->phys_stacked_size_p8
+(pKStk)	br.cond.dpnt skip_rbs_switch
+
+	/*
+	 * Restore user backing store.
+	 *
+	 * NOTE: alloc, loadrs, and cover can't be predicated.
+	 */
+(pNonSys) br.cond.dpnt dont_preserve_current_frame
+
+#ifdef CONFIG_XEN
+	XEN_HYPER_COVER;
+#else
+	cover				// add current frame into dirty partition and set cr.ifs
+#endif
+	;;
+	mov r19=ar.bsp			// get new backing store pointer
+rbs_switch:
+	sub r16=r16,r18			// krbs = old bsp - size of dirty partition
+	cmp.ne p9,p0=r0,r0		// clear p9 to skip restore of cr.ifs
+	;;
+	sub r19=r19,r16			// calculate total byte size of dirty partition
+	add r18=64,r18			// don't force in0-in7 into memory...
+	;;
+	shl r19=r19,16			// shift size of dirty partition into loadrs position
+	;;
+dont_preserve_current_frame:
+	/*
+	 * To prevent leaking bits between the kernel and user-space,
+	 * we must clear the stacked registers in the "invalid" partition here.
+	 * Not pretty, but at least it's fast (3.34 registers/cycle on Itanium,
+	 * 5 registers/cycle on McKinley).
+	 */
+#	define pRecurse	p6
+#	define pReturn	p7
+#ifdef CONFIG_ITANIUM
+#	define Nregs	10
+#else
+#	define Nregs	14
+#endif
+	alloc loc0=ar.pfs,2,Nregs-2,2,0
+	shr.u loc1=r18,9		// RNaTslots <= floor(dirtySize / (64*8))
+	sub r17=r17,r18			// r17 = (physStackedSize + 8) - dirtySize
+	;;
+	mov ar.rsc=r19			// load ar.rsc to be used for "loadrs"
+	shladd in0=loc1,3,r17
+	mov in1=0
+	;;
+	TEXT_ALIGN(32)
+rse_clear_invalid:
+#ifdef CONFIG_ITANIUM
+	// cycle 0
+ { .mii
+	alloc loc0=ar.pfs,2,Nregs-2,2,0
+	cmp.lt pRecurse,p0=Nregs*8,in0	// if more than Nregs regs left to clear, (re)curse
+	add out0=-Nregs*8,in0
+}{ .mfb
+	add out1=1,in1			// increment recursion count
+	nop.f 0
+	nop.b 0				// can't do br.call here because of alloc (WAW on CFM)
+	;;
+}{ .mfi	// cycle 1
+	mov loc1=0
+	nop.f 0
+	mov loc2=0
+}{ .mib
+	mov loc3=0
+	mov loc4=0
+(pRecurse) br.call.sptk.many b0=rse_clear_invalid
+
+}{ .mfi	// cycle 2
+	mov loc5=0
+	nop.f 0
+	cmp.ne pReturn,p0=r0,in1	// if recursion count != 0, we need to do a br.ret
+}{ .mib
+	mov loc6=0
+	mov loc7=0
+(pReturn) br.ret.sptk.many b0
+}
+#else /* !CONFIG_ITANIUM */
+	alloc loc0=ar.pfs,2,Nregs-2,2,0
+	cmp.lt pRecurse,p0=Nregs*8,in0	// if more than Nregs regs left to clear, (re)curse
+	add out0=-Nregs*8,in0
+	add out1=1,in1			// increment recursion count
+	mov loc1=0
+	mov loc2=0
+	;;
+	mov loc3=0
+	mov loc4=0
+	mov loc5=0
+	mov loc6=0
+	mov loc7=0
+(pRecurse) br.call.dptk.few b0=rse_clear_invalid
+	;;
+	mov loc8=0
+	mov loc9=0
+	cmp.ne pReturn,p0=r0,in1	// if recursion count != 0, we need to do a br.ret
+	mov loc10=0
+	mov loc11=0
+(pReturn) br.ret.dptk.many b0
+#endif /* !CONFIG_ITANIUM */
+#	undef pRecurse
+#	undef pReturn
+	;;
+	alloc r17=ar.pfs,0,0,0,0	// drop current register frame
+	;;
+	loadrs
+	;;
+skip_rbs_switch:
+	mov ar.unat=r25		// M2
+(pKStk)	extr.u r22=r22,21,1	// I0 extract current value of psr.pp from r22
+(pLvSys)mov r19=r0		// A  clear r19 for leave_syscall, no-op otherwise
+	;;
+(pUStk)	mov ar.bspstore=r23	// M2
+(pKStk)	dep r29=r22,r29,21,1	// I0 update ipsr.pp with psr.pp
+(pLvSys)mov r16=r0		// A  clear r16 for leave_syscall, no-op otherwise
+	;;
+#ifdef CONFIG_XEN
+	movl r25=XSI_IPSR
+	;;
+	st8[r25]=r29,XSI_IFS_OFS-XSI_IPSR_OFS
+	;;
+#else
+	mov cr.ipsr=r29		// M2
+#endif
+	mov ar.pfs=r26		// I0
+(pLvSys)mov r17=r0		// A  clear r17 for leave_syscall, no-op otherwise
+
+#ifdef CONFIG_XEN
+(p9)	st8 [r25]=r30
+	;;
+	adds r25=XSI_IIP_OFS-XSI_IFS_OFS,r25
+	;;
+#else
+(p9)	mov cr.ifs=r30		// M2
+#endif
+	mov b0=r21		// I0
+(pLvSys)mov r18=r0		// A  clear r18 for leave_syscall, no-op otherwise
+
+	mov ar.fpsr=r20		// M2
+#ifdef CONFIG_XEN
+	st8	[r25]=r28
+#else
+	mov cr.iip=r28		// M2
+#endif
+	nop 0
+	;;
+(pUStk)	mov ar.rnat=r24		// M2 must happen with RSE in lazy mode
+	nop 0
+(pLvSys)mov r2=r0
+
+	mov ar.rsc=r27		// M2
+	mov pr=r31,-1		// I0
+#ifdef CONFIG_XEN
+	;;
+	XEN_HYPER_RFI;
+#else
+	rfi			// B
+#endif
+
+	/*
+	 * On entry:
+	 *	r20 = &current->thread_info->pre_count (if CONFIG_PREEMPT)
+	 *	r31 = current->thread_info->flags
+	 * On exit:
+	 *	p6 = TRUE if work-pending-check needs to be redone
+	 */
+.work_pending_syscall:
+	add r2=-8,r2
+	add r3=-8,r3
+	;;
+	st8 [r2]=r8
+	st8 [r3]=r10
+.work_pending:
+	tbit.z p6,p0=r31,TIF_NEED_RESCHED		// current_thread_info()->need_resched==0?
+(p6)	br.cond.sptk.few .notify
+#ifdef CONFIG_PREEMPT
+(pKStk) dep r21=-1,r0,PREEMPT_ACTIVE_BIT,1
+	;;
+(pKStk) st4 [r20]=r21
+	ssm psr.i		// enable interrupts
+#endif
+	br.call.spnt.many rp=schedule
+.ret9:	cmp.eq p6,p0=r0,r0				// p6 <- 1
+#ifdef CONFIG_XEN
+	movl r2=XSI_PSR_I_ADDR
+	mov r20=1
+	;;
+	ld8 r2=[r2]
+	;;
+	st1 [r2]=r20
+#else
+	rsm psr.i		// disable interrupts
+#endif
+	;;
+#ifdef CONFIG_PREEMPT
+(pKStk)	adds r20=TI_PRE_COUNT+IA64_TASK_SIZE,r13
+	;;
+(pKStk)	st4 [r20]=r0		// preempt_count() <- 0
+#endif
+(pLvSys)br.cond.sptk.few  .work_pending_syscall_end
+	br.cond.sptk.many .work_processed_kernel	// re-check
+
+.notify:
+(pUStk)	br.call.spnt.many rp=notify_resume_user
+.ret10:	cmp.ne p6,p0=r0,r0				// p6 <- 0
+(pLvSys)br.cond.sptk.few  .work_pending_syscall_end
+	br.cond.sptk.many .work_processed_kernel	// don't re-check
+
+.work_pending_syscall_end:
+	adds r2=PT(R8)+16,r12
+	adds r3=PT(R10)+16,r12
+	;;
+	ld8 r8=[r2]
+	ld8 r10=[r3]
+	br.cond.sptk.many .work_processed_syscall	// re-check
+
+#ifdef CONFIG_XEN
+END(xen_leave_kernel)
+#else
+END(ia64_leave_kernel)
+#endif
diff -rpuN linux-2.6.18.8/arch/ia64/xen/xenivt.S linux-2.6.18-xen-3.3.0/arch/ia64/xen/xenivt.S
--- linux-2.6.18.8/arch/ia64/xen/xenivt.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/xen/xenivt.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,2178 @@
+/*
+ * arch/ia64/xen/ivt.S
+ *
+ * Copyright (C) 2005 Hewlett-Packard Co
+ *	Dan Magenheimer <dan.magenheimer@hp.com>
+ */
+/*
+ * This file defines the interruption vector table used by the CPU.
+ * It does not include one entry per possible cause of interruption.
+ *
+ * The first 20 entries of the table contain 64 bundles each while the
+ * remaining 48 entries contain only 16 bundles each.
+ *
+ * The 64 bundles are used to allow inlining the whole handler for critical
+ * interruptions like TLB misses.
+ *
+ *  For each entry, the comment is as follows:
+ *
+ *		// 0x1c00 Entry 7 (size 64 bundles) Data Key Miss (12,51)
+ *  entry offset ----/     /         /                  /          /
+ *  entry number ---------/         /                  /          /
+ *  size of the entry -------------/                  /          /
+ *  vector name -------------------------------------/          /
+ *  interruptions triggering this vector ----------------------/
+ *
+ * The table is 32KB in size and must be aligned on 32KB boundary.
+ * (The CPU ignores the 15 lower bits of the address)
+ *
+ * Table is based upon EAS2.6 (Oct 1999)
+ */
+
+#include <asm/asmmacro.h>
+#include <asm/break.h>
+#include <asm/ia32.h>
+#include <asm/kregs.h>
+#include <asm/asm-offsets.h>
+#include <asm/pgtable.h>
+#include <asm/processor.h>
+#include <asm/ptrace.h>
+#include <asm/system.h>
+#include <asm/thread_info.h>
+#include <asm/unistd.h>
+#include <asm/errno.h>
+
+#ifdef CONFIG_XEN
+#define ia64_ivt xen_ivt
+#endif
+
+#if 1
+# define PSR_DEFAULT_BITS	psr.ac
+#else
+# define PSR_DEFAULT_BITS	0
+#endif
+
+#if 0
+  /*
+   * This lets you track the last eight faults that occurred on the CPU.  Make sure ar.k2 isn't
+   * needed for something else before enabling this...
+   */
+# define DBG_FAULT(i)	mov r16=ar.k2;;	shl r16=r16,8;;	add r16=(i),r16;;mov ar.k2=r16
+#else
+# define DBG_FAULT(i)
+#endif
+
+#define MINSTATE_VIRT	/* needed by minstate.h */
+#include "xenminstate.h"
+
+#define FAULT(n)									\
+	mov r31=pr;									\
+	mov r19=n;;			/* prepare to save predicates */		\
+	br.sptk.many dispatch_to_fault_handler
+
+	.section .text.ivt,"ax"
+
+	.align 32768	// align on 32KB boundary
+	.global ia64_ivt
+ia64_ivt:
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x0000 Entry 0 (size 64 bundles) VHPT Translation (8,20,47)
+ENTRY(vhpt_miss)
+	DBG_FAULT(0)
+	/*
+	 * The VHPT vector is invoked when the TLB entry for the virtual page table
+	 * is missing.  This happens only as a result of a previous
+	 * (the "original") TLB miss, which may either be caused by an instruction
+	 * fetch or a data access (or non-access).
+	 *
+	 * What we do here is normal TLB miss handing for the _original_ miss,
+	 * followed by inserting the TLB entry for the virtual page table page
+	 * that the VHPT walker was attempting to access.  The latter gets
+	 * inserted as long as page table entry above pte level have valid
+	 * mappings for the faulting address.  The TLB entry for the original
+	 * miss gets inserted only if the pte entry indicates that the page is
+	 * present.
+	 *
+	 * do_page_fault gets invoked in the following cases:
+	 *	- the faulting virtual address uses unimplemented address bits
+	 *	- the faulting virtual address has no valid page table mapping
+	 */
+#ifdef CONFIG_XEN
+	movl r16=XSI_IFA
+	;;
+	ld8 r16=[r16]
+#ifdef CONFIG_HUGETLB_PAGE
+	movl r18=PAGE_SHIFT
+	movl r25=XSI_ITIR
+	;;
+	ld8 r25=[r25]
+#endif
+	;;
+#else
+	mov r16=cr.ifa				// get address that caused the TLB miss
+#ifdef CONFIG_HUGETLB_PAGE
+	movl r18=PAGE_SHIFT
+	mov r25=cr.itir
+#endif
+#endif
+	;;
+#ifdef CONFIG_XEN
+	XEN_HYPER_RSM_PSR_DT;
+#else
+	rsm psr.dt				// use physical addressing for data
+#endif
+	mov r31=pr				// save the predicate registers
+	mov r19=IA64_KR(PT_BASE)		// get page table base address
+	shl r21=r16,3				// shift bit 60 into sign bit
+	shr.u r17=r16,61			// get the region number into r17
+	;;
+	shr.u r22=r21,3
+#ifdef CONFIG_HUGETLB_PAGE
+	extr.u r26=r25,2,6
+	;;
+	cmp.ne p8,p0=r18,r26
+	sub r27=r26,r18
+	;;
+(p8)	dep r25=r18,r25,2,6
+(p8)	shr r22=r22,r27
+#endif
+	;;
+	cmp.eq p6,p7=5,r17			// is IFA pointing into to region 5?
+	shr.u r18=r22,PGDIR_SHIFT		// get bottom portion of pgd index bit
+	;;
+(p7)	dep r17=r17,r19,(PAGE_SHIFT-3),3	// put region number bits in place
+
+	srlz.d
+	LOAD_PHYSICAL(p6, r19, swapper_pg_dir)	// region 5 is rooted at swapper_pg_dir
+
+	.pred.rel "mutex", p6, p7
+(p6)	shr.u r21=r21,PGDIR_SHIFT+PAGE_SHIFT
+(p7)	shr.u r21=r21,PGDIR_SHIFT+PAGE_SHIFT-3
+	;;
+(p6)	dep r17=r18,r19,3,(PAGE_SHIFT-3)	// r17=pgd_offset for region 5
+(p7)	dep r17=r18,r17,3,(PAGE_SHIFT-6)	// r17=pgd_offset for region[0-4]
+	cmp.eq p7,p6=0,r21			// unused address bits all zeroes?
+#ifdef CONFIG_PGTABLE_4
+	shr.u r28=r22,PUD_SHIFT			// shift pud index into position
+#else
+	shr.u r18=r22,PMD_SHIFT			// shift pmd index into position
+#endif
+	;;
+	ld8 r17=[r17]				// get *pgd (may be 0)
+	;;
+(p7)	cmp.eq p6,p7=r17,r0			// was pgd_present(*pgd) == NULL?
+#ifdef CONFIG_PGTABLE_4
+	dep r28=r28,r17,3,(PAGE_SHIFT-3)	// r28=pud_offset(pgd,addr)
+	;;
+	shr.u r18=r22,PMD_SHIFT			// shift pmd index into position
+(p7)	ld8 r29=[r28]				// get *pud (may be 0)
+	;;
+(p7)	cmp.eq.or.andcm p6,p7=r29,r0		// was pud_present(*pud) == NULL?
+	dep r17=r18,r29,3,(PAGE_SHIFT-3)	// r17=pmd_offset(pud,addr)
+#else
+	dep r17=r18,r17,3,(PAGE_SHIFT-3)	// r17=pmd_offset(pgd,addr)
+#endif
+	;;
+(p7)	ld8 r20=[r17]				// get *pmd (may be 0)
+	shr.u r19=r22,PAGE_SHIFT		// shift pte index into position
+	;;
+(p7)	cmp.eq.or.andcm p6,p7=r20,r0		// was pmd_present(*pmd) == NULL?
+	dep r21=r19,r20,3,(PAGE_SHIFT-3)	// r21=pte_offset(pmd,addr)
+	;;
+(p7)	ld8 r18=[r21]				// read *pte
+#ifdef CONFIG_XEN
+	movl r19=XSI_ISR
+	;;
+	ld8 r19=[r19]
+#else
+	mov r19=cr.isr				// cr.isr bit 32 tells us if this is an insn miss
+#endif
+	;;
+(p7)	tbit.z p6,p7=r18,_PAGE_P_BIT		// page present bit cleared?
+#ifdef CONFIG_XEN
+	movl r22=XSI_IHA
+	;;
+	ld8 r22=[r22]
+#else
+	mov r22=cr.iha				// get the VHPT address that caused the TLB miss
+#endif
+	;;					// avoid RAW on p7
+(p7)	tbit.nz.unc p10,p11=r19,32		// is it an instruction TLB miss?
+	dep r23=0,r20,0,PAGE_SHIFT		// clear low bits to get page address
+	;;
+#ifdef CONFIG_XEN
+	mov r24=r8
+	mov r8=r18
+	;;
+(p10)	XEN_HYPER_ITC_I
+	;;
+(p11)	XEN_HYPER_ITC_D
+	;;
+	mov r8=r24
+	;;
+#else
+(p10)	itc.i r18				// insert the instruction TLB entry
+(p11)	itc.d r18				// insert the data TLB entry
+#endif
+(p6)	br.cond.spnt.many page_fault		// handle bad address/page not present (page fault)
+#ifdef CONFIG_XEN
+	movl r24=XSI_IFA
+	;;
+	st8 [r24]=r22
+	;;
+#else
+	mov cr.ifa=r22
+#endif
+
+#ifdef CONFIG_HUGETLB_PAGE
+(p8)	mov cr.itir=r25				// change to default page-size for VHPT
+#endif
+
+	/*
+	 * Now compute and insert the TLB entry for the virtual page table.  We never
+	 * execute in a page table page so there is no need to set the exception deferral
+	 * bit.
+	 */
+	adds r24=__DIRTY_BITS_NO_ED|_PAGE_PL_0|_PAGE_AR_RW,r23
+	;;
+#ifdef CONFIG_XEN
+(p7)	mov r25=r8
+(p7)	mov r8=r24
+	;;
+(p7)	XEN_HYPER_ITC_D
+	;;
+(p7)	mov r8=r25
+	;;
+#else
+(p7)	itc.d r24
+#endif
+	;;
+#ifdef CONFIG_SMP
+	/*
+	 * Tell the assemblers dependency-violation checker that the above "itc" instructions
+	 * cannot possibly affect the following loads:
+	 */
+	dv_serialize_data
+
+	/*
+	 * Re-check pagetable entry.  If they changed, we may have received a ptc.g
+	 * between reading the pagetable and the "itc".  If so, flush the entry we
+	 * inserted and retry.  At this point, we have:
+	 *
+	 * r28 = equivalent of pud_offset(pgd, ifa)
+	 * r17 = equivalent of pmd_offset(pud, ifa)
+	 * r21 = equivalent of pte_offset(pmd, ifa)
+	 *
+	 * r29 = *pud
+	 * r20 = *pmd
+	 * r18 = *pte
+	 */
+	ld8 r25=[r21]				// read *pte again
+	ld8 r26=[r17]				// read *pmd again
+#ifdef CONFIG_PGTABLE_4
+	ld8 r19=[r28]				// read *pud again
+#endif
+	cmp.ne p6,p7=r0,r0
+	;;
+	cmp.ne.or.andcm p6,p7=r26,r20		// did *pmd change
+#ifdef CONFIG_PGTABLE_4
+	cmp.ne.or.andcm p6,p7=r19,r29		// did *pud change
+#endif
+	mov r27=PAGE_SHIFT<<2
+	;;
+(p6)	ptc.l r22,r27				// purge PTE page translation
+(p7)	cmp.ne.or.andcm p6,p7=r25,r18		// did *pte change
+	;;
+(p6)	ptc.l r16,r27				// purge translation
+#endif
+
+	mov pr=r31,-1				// restore predicate registers
+#ifdef CONFIG_XEN
+	XEN_HYPER_RFI
+	dv_serialize_data
+#else
+	rfi
+#endif
+END(vhpt_miss)
+
+	.org ia64_ivt+0x400
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x0400 Entry 1 (size 64 bundles) ITLB (21)
+ENTRY(itlb_miss)
+	DBG_FAULT(1)
+	/*
+	 * The ITLB handler accesses the PTE via the virtually mapped linear
+	 * page table.  If a nested TLB miss occurs, we switch into physical
+	 * mode, walk the page table, and then re-execute the PTE read and
+	 * go on normally after that.
+	 */
+#ifdef CONFIG_XEN
+	movl r16=XSI_IFA
+	;;
+	ld8 r16=[r16]
+#else
+	mov r16=cr.ifa				// get virtual address
+#endif
+	mov r29=b0				// save b0
+	mov r31=pr				// save predicates
+.itlb_fault:
+#ifdef CONFIG_XEN
+	movl r17=XSI_IHA
+	;;
+	ld8 r17=[r17]				// get virtual address of L3 PTE
+#else
+	mov r17=cr.iha				// get virtual address of PTE
+#endif
+	movl r30=1f				// load nested fault continuation point
+	;;
+1:	ld8 r18=[r17]				// read *pte
+	;;
+	mov b0=r29
+	tbit.z p6,p0=r18,_PAGE_P_BIT		// page present bit cleared?
+(p6)	br.cond.spnt page_fault
+	;;
+#ifdef CONFIG_XEN
+	mov r19=r8
+	mov r8=r18
+	;;
+	XEN_HYPER_ITC_I
+	;;
+	mov r8=r19
+#else
+	itc.i r18
+#endif
+	;;
+#ifdef CONFIG_SMP
+	/*
+	 * Tell the assemblers dependency-violation checker that the above "itc" instructions
+	 * cannot possibly affect the following loads:
+	 */
+	dv_serialize_data
+
+	ld8 r19=[r17]				// read *pte again and see if same
+	mov r20=PAGE_SHIFT<<2			// setup page size for purge
+	;;
+	cmp.ne p7,p0=r18,r19
+	;;
+(p7)	ptc.l r16,r20
+#endif
+	mov pr=r31,-1
+#ifdef CONFIG_XEN
+	XEN_HYPER_RFI
+	dv_serialize_data
+#else
+	rfi
+#endif
+END(itlb_miss)
+
+	.org ia64_ivt+0x0800
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x0800 Entry 2 (size 64 bundles) DTLB (9,48)
+ENTRY(dtlb_miss)
+	DBG_FAULT(2)
+	/*
+	 * The DTLB handler accesses the PTE via the virtually mapped linear
+	 * page table.  If a nested TLB miss occurs, we switch into physical
+	 * mode, walk the page table, and then re-execute the PTE read and
+	 * go on normally after that.
+	 */
+#ifdef CONFIG_XEN
+	movl r16=XSI_IFA
+	;;
+	ld8 r16=[r16]
+#else
+	mov r16=cr.ifa				// get virtual address
+#endif
+	mov r29=b0				// save b0
+	mov r31=pr				// save predicates
+dtlb_fault:
+#ifdef CONFIG_XEN
+	movl r17=XSI_IHA
+	;;
+	ld8 r17=[r17]				// get virtual address of L3 PTE
+#else
+	mov r17=cr.iha				// get virtual address of PTE
+#endif
+	movl r30=1f				// load nested fault continuation point
+	;;
+1:	ld8 r18=[r17]				// read *pte
+	;;
+	mov b0=r29
+	tbit.z p6,p0=r18,_PAGE_P_BIT		// page present bit cleared?
+(p6)	br.cond.spnt page_fault
+	;;
+#ifdef CONFIG_XEN
+	mov r19=r8
+	mov r8=r18
+	;;
+	XEN_HYPER_ITC_D
+	;;
+	mov r8=r19
+	;;
+#else
+	itc.d r18
+#endif
+	;;
+#ifdef CONFIG_SMP
+	/*
+	 * Tell the assemblers dependency-violation checker that the above "itc" instructions
+	 * cannot possibly affect the following loads:
+	 */
+	dv_serialize_data
+
+	ld8 r19=[r17]				// read *pte again and see if same
+	mov r20=PAGE_SHIFT<<2			// setup page size for purge
+	;;
+	cmp.ne p7,p0=r18,r19
+	;;
+(p7)	ptc.l r16,r20
+#endif
+	mov pr=r31,-1
+#ifdef CONFIG_XEN
+	XEN_HYPER_RFI
+	dv_serialize_data
+#else
+	rfi
+#endif
+END(dtlb_miss)
+
+	.org ia64_ivt+0x0c00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x0c00 Entry 3 (size 64 bundles) Alt ITLB (19)
+ENTRY(alt_itlb_miss)
+	DBG_FAULT(3)
+#ifdef CONFIG_XEN
+	movl r31=XSI_IPSR
+	;;
+	ld8 r21=[r31],XSI_IFA_OFS-XSI_IPSR_OFS	// get ipsr, point to ifa
+	movl r17=PAGE_KERNEL
+	;;
+	ld8 r16=[r31]		// get ifa
+#else
+	mov r16=cr.ifa		// get address that caused the TLB miss
+	movl r17=PAGE_KERNEL
+	mov r21=cr.ipsr
+#endif
+	movl r19=(((1 << IA64_MAX_PHYS_BITS) - 1) & ~0xfff)
+	mov r31=pr
+	;;
+#ifdef CONFIG_DISABLE_VHPT
+	shr.u r22=r16,61			// get the region number into r21
+	;;
+	cmp.gt p8,p0=6,r22			// user mode
+	;;
+#ifndef CONFIG_XEN
+(p8)	thash r17=r16
+	;;
+(p8)	mov cr.iha=r17
+#endif
+(p8)	mov r29=b0				// save b0
+(p8)	br.cond.dptk .itlb_fault
+#endif
+	extr.u r23=r21,IA64_PSR_CPL0_BIT,2	// extract psr.cpl
+	and r19=r19,r16		// clear ed, reserved bits, and PTE control bits
+	shr.u r18=r16,57	// move address bit 61 to bit 4
+	;;
+	andcm r18=0x10,r18	// bit 4=~address-bit(61)
+	cmp.ne p8,p0=r0,r23	// psr.cpl != 0?
+	or r19=r17,r19		// insert PTE control bits into r19
+	;;
+	or r19=r19,r18		// set bit 4 (uncached) if the access was to region 6
+(p8)	br.cond.spnt page_fault
+	;;
+#ifdef CONFIG_XEN
+	mov r18=r8
+	mov r8=r19
+	;;
+	XEN_HYPER_ITC_I
+	;;
+	mov r8=r18
+	;;
+	mov pr=r31,-1
+	;;
+	XEN_HYPER_RFI;
+#else
+	itc.i r19		// insert the TLB entry
+	mov pr=r31,-1
+	rfi
+#endif
+END(alt_itlb_miss)
+
+	.org ia64_ivt+0x1000
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x1000 Entry 4 (size 64 bundles) Alt DTLB (7,46)
+ENTRY(alt_dtlb_miss)
+	DBG_FAULT(4)
+#ifdef CONFIG_XEN
+	movl r31=XSI_IPSR
+	;;
+	ld8 r21=[r31],XSI_ISR_OFS-XSI_IPSR_OFS	// get ipsr, point to isr
+	movl r17=PAGE_KERNEL
+	;;
+	ld8 r20=[r31],XSI_IFA_OFS-XSI_ISR_OFS	// get isr, point to ifa
+	movl r19=(((1 << IA64_MAX_PHYS_BITS) - 1) & ~0xfff)
+	;;
+	ld8 r16=[r31]		// get ifa
+#else
+	mov r16=cr.ifa		// get address that caused the TLB miss
+	movl r17=PAGE_KERNEL
+	mov r20=cr.isr
+	movl r19=(((1 << IA64_MAX_PHYS_BITS) - 1) & ~0xfff)
+	mov r21=cr.ipsr
+#endif
+	mov r31=pr
+	;;
+#ifdef CONFIG_DISABLE_VHPT
+	shr.u r22=r16,61			// get the region number into r21
+	;;
+	cmp.gt p8,p0=6,r22			// access to region 0-5
+	;;
+#ifndef CONFIG_XEN
+(p8)	thash r17=r16
+	;;
+(p8)	mov cr.iha=r17
+#endif
+(p8)	mov r29=b0				// save b0
+(p8)	br.cond.dptk dtlb_fault
+#endif
+	extr.u r23=r21,IA64_PSR_CPL0_BIT,2	// extract psr.cpl
+	and r22=IA64_ISR_CODE_MASK,r20		// get the isr.code field
+	tbit.nz p6,p7=r20,IA64_ISR_SP_BIT	// is speculation bit on?
+	shr.u r18=r16,57			// move address bit 61 to bit 4
+	and r19=r19,r16				// clear ed, reserved bits, and PTE control bits
+	tbit.nz p9,p0=r20,IA64_ISR_NA_BIT	// is non-access bit on?
+	;;
+	andcm r18=0x10,r18	// bit 4=~address-bit(61)
+	cmp.ne p8,p0=r0,r23
+(p9)	cmp.eq.or.andcm p6,p7=IA64_ISR_CODE_LFETCH,r22	// check isr.code field
+(p8)	br.cond.spnt page_fault
+
+	dep r21=-1,r21,IA64_PSR_ED_BIT,1
+	or r19=r19,r17		// insert PTE control bits into r19
+	;;
+	or r19=r19,r18		// set bit 4 (uncached) if the access was to region 6
+(p6)	mov cr.ipsr=r21
+	;;
+#ifdef CONFIG_XEN
+(p7)	mov r18=r8
+(p7)	mov r8=r19
+	;;
+(p7)	XEN_HYPER_ITC_D
+	;;
+(p7)	mov r8=r18
+	;;
+	mov pr=r31,-1
+	;;
+	XEN_HYPER_RFI;
+#else
+(p7)	itc.d r19		// insert the TLB entry
+	mov pr=r31,-1
+	rfi
+#endif
+END(alt_dtlb_miss)
+
+	.org ia64_ivt+0x1400
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x1400 Entry 5 (size 64 bundles) Data nested TLB (6,45)
+ENTRY(nested_dtlb_miss)
+	/*
+	 * In the absence of kernel bugs, we get here when the virtually mapped linear
+	 * page table is accessed non-speculatively (e.g., in the Dirty-bit, Instruction
+	 * Access-bit, or Data Access-bit faults).  If the DTLB entry for the virtual page
+	 * table is missing, a nested TLB miss fault is triggered and control is
+	 * transferred to this point.  When this happens, we lookup the pte for the
+	 * faulting address by walking the page table in physical mode and return to the
+	 * continuation point passed in register r30 (or call page_fault if the address is
+	 * not mapped).
+	 *
+	 * Input:	r16:	faulting address
+	 *		r29:	saved b0
+	 *		r30:	continuation address
+	 *		r31:	saved pr
+	 *
+	 * Output:	r17:	physical address of PTE of faulting address
+	 *		r29:	saved b0
+	 *		r30:	continuation address
+	 *		r31:	saved pr
+	 *
+	 * Clobbered:	b0, r18, r19, r21, r22, psr.dt (cleared)
+	 */
+#ifdef CONFIG_XEN
+	XEN_HYPER_RSM_PSR_DT;
+#else
+	rsm psr.dt				// switch to using physical data addressing
+#endif
+	mov r19=IA64_KR(PT_BASE)		// get the page table base address
+	shl r21=r16,3				// shift bit 60 into sign bit
+#ifdef CONFIG_XEN
+	movl r18=XSI_ITIR
+	;;
+	ld8 r18=[r18]
+#else
+	mov r18=cr.itir
+#endif
+	;;
+	shr.u r17=r16,61			// get the region number into r17
+	extr.u r18=r18,2,6			// get the faulting page size
+	;;
+	cmp.eq p6,p7=5,r17			// is faulting address in region 5?
+	add r22=-PAGE_SHIFT,r18			// adjustment for hugetlb address
+	add r18=PGDIR_SHIFT-PAGE_SHIFT,r18
+	;;
+	shr.u r22=r16,r22
+	shr.u r18=r16,r18
+(p7)	dep r17=r17,r19,(PAGE_SHIFT-3),3	// put region number bits in place
+
+	srlz.d
+	LOAD_PHYSICAL(p6, r19, swapper_pg_dir)	// region 5 is rooted at swapper_pg_dir
+
+	.pred.rel "mutex", p6, p7
+(p6)	shr.u r21=r21,PGDIR_SHIFT+PAGE_SHIFT
+(p7)	shr.u r21=r21,PGDIR_SHIFT+PAGE_SHIFT-3
+	;;
+(p6)	dep r17=r18,r19,3,(PAGE_SHIFT-3)	// r17=pgd_offset for region 5
+(p7)	dep r17=r18,r17,3,(PAGE_SHIFT-6)	// r17=pgd_offset for region[0-4]
+	cmp.eq p7,p6=0,r21			// unused address bits all zeroes?
+#ifdef CONFIG_PGTABLE_4
+	shr.u r18=r22,PUD_SHIFT			// shift pud index into position
+#else
+	shr.u r18=r22,PMD_SHIFT			// shift pmd index into position
+#endif
+	;;
+	ld8 r17=[r17]				// get *pgd (may be 0)
+	;;
+(p7)	cmp.eq p6,p7=r17,r0			// was pgd_present(*pgd) == NULL?
+	dep r17=r18,r17,3,(PAGE_SHIFT-3)	// r17=p[u|m]d_offset(pgd,addr)
+	;;
+#ifdef CONFIG_PGTABLE_4
+(p7)	ld8 r17=[r17]				// get *pud (may be 0)
+	shr.u r18=r22,PMD_SHIFT			// shift pmd index into position
+	;;
+(p7)	cmp.eq.or.andcm p6,p7=r17,r0		// was pud_present(*pud) == NULL?
+	dep r17=r18,r17,3,(PAGE_SHIFT-3)	// r17=pmd_offset(pud,addr)
+	;;
+#endif
+(p7)	ld8 r17=[r17]				// get *pmd (may be 0)
+	shr.u r19=r22,PAGE_SHIFT		// shift pte index into position
+	;;
+(p7)	cmp.eq.or.andcm p6,p7=r17,r0		// was pmd_present(*pmd) == NULL?
+	dep r17=r19,r17,3,(PAGE_SHIFT-3)	// r17=pte_offset(pmd,addr);
+(p6)	br.cond.spnt page_fault
+	mov b0=r30
+	br.sptk.many b0				// return to continuation point
+END(nested_dtlb_miss)
+
+	.org ia64_ivt+0x1800
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x1800 Entry 6 (size 64 bundles) Instruction Key Miss (24)
+ENTRY(ikey_miss)
+	DBG_FAULT(6)
+	FAULT(6)
+END(ikey_miss)
+
+	//-----------------------------------------------------------------------------------
+	// call do_page_fault (predicates are in r31, psr.dt may be off, r16 is faulting address)
+ENTRY(page_fault)
+#ifdef CONFIG_XEN
+	XEN_HYPER_SSM_PSR_DT
+#else
+	ssm psr.dt
+	;;
+	srlz.i
+#endif
+	;;
+	SAVE_MIN_WITH_COVER
+	alloc r15=ar.pfs,0,0,3,0
+#ifdef CONFIG_XEN
+	movl r3=XSI_ISR
+	;;
+	ld8 out1=[r3],XSI_IFA_OFS-XSI_ISR_OFS	// get vcr.isr, point to ifa
+	;;
+	ld8 out0=[r3]				// get vcr.ifa
+	mov r14=1
+	;;
+	add r3=XSI_PSR_IC_OFS-XSI_IFA_OFS, r3	// point to vpsr.ic
+	;;
+	st4 [r3]=r14				// vpsr.ic = 1
+	adds r3=8,r2				// set up second base pointer
+	;;
+	sum PSR_DEFAULT_BITS
+#else
+	mov out0=cr.ifa
+	mov out1=cr.isr
+	adds r3=8,r2				// set up second base pointer
+	;;
+	ssm psr.ic | PSR_DEFAULT_BITS
+#endif
+	;;
+	srlz.i					// guarantee that interruption collectin is on
+	;;
+#ifdef CONFIG_XEN
+    
+#define MASK_TO_PEND_OFS    (-1)
+    
+(p15)	movl r14=XSI_PSR_I_ADDR
+	;;
+(p15)	ld8 r14=[r14]
+	;;
+(p15)	st1 [r14]=r0,MASK_TO_PEND_OFS		// if (p15) vpsr.i = 1
+	;;		// if (p15) (vcpu->vcpu_info->evtchn_upcall_mask)=0
+(p15)	ld1 r14=[r14]	// if (vcpu->vcpu_info->evtchn_upcall_pending)
+	;;
+(p15)	cmp.ne	p15,p0=r14,r0
+	;;
+(p15)	XEN_HYPER_SSM_I
+#else
+(p15)	ssm psr.i				// restore psr.i
+#endif
+	movl r14=ia64_leave_kernel
+	;;
+	SAVE_REST
+	mov rp=r14
+	;;
+	adds out2=16,r12			// out2 = pointer to pt_regs
+	br.call.sptk.many b6=ia64_do_page_fault	// ignore return address
+END(page_fault)
+
+	.org ia64_ivt+0x1c00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x1c00 Entry 7 (size 64 bundles) Data Key Miss (12,51)
+ENTRY(dkey_miss)
+	DBG_FAULT(7)
+	FAULT(7)
+END(dkey_miss)
+
+	.org ia64_ivt+0x2000
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x2000 Entry 8 (size 64 bundles) Dirty-bit (54)
+ENTRY(dirty_bit)
+	DBG_FAULT(8)
+	/*
+	 * What we do here is to simply turn on the dirty bit in the PTE.  We need to
+	 * update both the page-table and the TLB entry.  To efficiently access the PTE,
+	 * we address it through the virtual page table.  Most likely, the TLB entry for
+	 * the relevant virtual page table page is still present in the TLB so we can
+	 * normally do this without additional TLB misses.  In case the necessary virtual
+	 * page table TLB entry isn't present, we take a nested TLB miss hit where we look
+	 * up the physical address of the L3 PTE and then continue at label 1 below.
+	 */
+#ifdef CONFIG_XEN
+	movl r16=XSI_IFA
+	;;
+	ld8 r16=[r16]
+	;;
+#else
+	mov r16=cr.ifa				// get the address that caused the fault
+#endif
+	movl r30=1f				// load continuation point in case of nested fault
+	;;
+#ifdef CONFIG_XEN
+	mov r18=r8;
+	mov r8=r16;
+	XEN_HYPER_THASH;;
+	mov r17=r8;
+	mov r8=r18;;
+#else
+	thash r17=r16				// compute virtual address of L3 PTE
+#endif
+	mov r29=b0				// save b0 in case of nested fault
+	mov r31=pr				// save pr
+#ifdef CONFIG_SMP
+	mov r28=ar.ccv				// save ar.ccv
+	;;
+1:	ld8 r18=[r17]
+	;;					// avoid RAW on r18
+	mov ar.ccv=r18				// set compare value for cmpxchg
+	or r25=_PAGE_D|_PAGE_A,r18		// set the dirty and accessed bits
+	tbit.z p7,p6 = r18,_PAGE_P_BIT		// Check present bit
+	;;
+(p6)	cmpxchg8.acq r26=[r17],r25,ar.ccv	// Only update if page is present
+	mov r24=PAGE_SHIFT<<2
+	;;
+(p6)	cmp.eq p6,p7=r26,r18			// Only compare if page is present
+	;;
+#ifdef CONFIG_XEN
+(p6)	mov r18=r8
+(p6)	mov r8=r25
+	;;
+(p6)	XEN_HYPER_ITC_D
+	;;
+(p6)	mov r8=r18
+#else
+(p6)	itc.d r25				// install updated PTE
+#endif	
+	;;
+	/*
+	 * Tell the assemblers dependency-violation checker that the above "itc" instructions
+	 * cannot possibly affect the following loads:
+	 */
+	dv_serialize_data
+
+	ld8 r18=[r17]				// read PTE again
+	;;
+	cmp.eq p6,p7=r18,r25			// is it same as the newly installed
+	;;
+(p7)	ptc.l r16,r24
+	mov b0=r29				// restore b0
+	mov ar.ccv=r28
+#else
+	;;
+1:	ld8 r18=[r17]
+	;;					// avoid RAW on r18
+	or r18=_PAGE_D|_PAGE_A,r18		// set the dirty and accessed bits
+	mov b0=r29				// restore b0
+	;;
+	st8 [r17]=r18				// store back updated PTE
+	itc.d r18				// install updated PTE
+#endif
+	mov pr=r31,-1				// restore pr
+#ifdef CONFIG_XEN
+	XEN_HYPER_RFI
+	dv_serialize_data
+#else
+	rfi
+#endif
+END(dirty_bit)
+
+	.org ia64_ivt+0x2400
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x2400 Entry 9 (size 64 bundles) Instruction Access-bit (27)
+ENTRY(iaccess_bit)
+	DBG_FAULT(9)
+	// Like Entry 8, except for instruction access
+#ifdef CONFIG_XEN
+	movl r16=XSI_IFA
+	;;
+	ld8 r16=[r16]
+	;;
+#else
+	mov r16=cr.ifa				// get the address that caused the fault
+#endif
+	movl r30=1f				// load continuation point in case of nested fault
+	mov r31=pr				// save predicates
+#ifdef CONFIG_ITANIUM
+	/*
+	 * Erratum 10 (IFA may contain incorrect address) has "NoFix" status.
+	 */
+	mov r17=cr.ipsr
+	;;
+	mov r18=cr.iip
+	tbit.z p6,p0=r17,IA64_PSR_IS_BIT	// IA64 instruction set?
+	;;
+(p6)	mov r16=r18				// if so, use cr.iip instead of cr.ifa
+#endif /* CONFIG_ITANIUM */
+	;;
+#ifdef CONFIG_XEN
+	mov r18=r8;
+	mov r8=r16;
+	XEN_HYPER_THASH;;
+	mov r17=r8;
+	mov r8=r18;;
+#else
+	thash r17=r16				// compute virtual address of L3 PTE
+#endif
+	mov r29=b0				// save b0 in case of nested fault)
+#ifdef CONFIG_SMP
+	mov r28=ar.ccv				// save ar.ccv
+	;;
+1:	ld8 r18=[r17]
+	;;
+	mov ar.ccv=r18				// set compare value for cmpxchg
+	or r25=_PAGE_A,r18			// set the accessed bit
+	tbit.z p7,p6 = r18,_PAGE_P_BIT	 	// Check present bit
+	;;
+(p6)	cmpxchg8.acq r26=[r17],r25,ar.ccv	// Only if page present
+	mov r24=PAGE_SHIFT<<2
+	;;
+(p6)	cmp.eq p6,p7=r26,r18			// Only if page present
+	;;
+#ifdef CONFIG_XEN
+	mov r26=r8
+	mov r8=r25
+	;;
+(p6)	XEN_HYPER_ITC_I
+	;;
+	mov r8=r26
+	;;
+#else
+(p6)	itc.i r25				// install updated PTE
+#endif
+	;;
+	/*
+	 * Tell the assemblers dependency-violation checker that the above "itc" instructions
+	 * cannot possibly affect the following loads:
+	 */
+	dv_serialize_data
+
+	ld8 r18=[r17]				// read PTE again
+	;;
+	cmp.eq p6,p7=r18,r25			// is it same as the newly installed
+	;;
+(p7)	ptc.l r16,r24
+	mov b0=r29				// restore b0
+	mov ar.ccv=r28
+#else /* !CONFIG_SMP */
+	;;
+1:	ld8 r18=[r17]
+	;;
+	or r18=_PAGE_A,r18			// set the accessed bit
+	mov b0=r29				// restore b0
+	;;
+	st8 [r17]=r18				// store back updated PTE
+	itc.i r18				// install updated PTE
+#endif /* !CONFIG_SMP */
+	mov pr=r31,-1
+#ifdef CONFIG_XEN
+	XEN_HYPER_RFI
+	dv_serialize_data
+#else
+	rfi
+#endif
+END(iaccess_bit)
+
+	.org ia64_ivt+0x2800
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x2800 Entry 10 (size 64 bundles) Data Access-bit (15,55)
+ENTRY(daccess_bit)
+	DBG_FAULT(10)
+	// Like Entry 8, except for data access
+#ifdef CONFIG_XEN
+	movl r16=XSI_IFA
+	;;
+	ld8 r16=[r16]
+	;;
+#else
+	mov r16=cr.ifa				// get the address that caused the fault
+#endif
+	movl r30=1f				// load continuation point in case of nested fault
+	;;
+#ifdef CONFIG_XEN
+	mov r18=r8
+	mov r8=r16
+	XEN_HYPER_THASH
+	;;
+	mov r17=r8
+	mov r8=r18
+	;;
+#else
+	thash r17=r16				// compute virtual address of L3 PTE
+#endif
+	mov r31=pr
+	mov r29=b0				// save b0 in case of nested fault)
+#ifdef CONFIG_SMP
+	mov r28=ar.ccv				// save ar.ccv
+	;;
+1:	ld8 r18=[r17]
+	;;					// avoid RAW on r18
+	mov ar.ccv=r18				// set compare value for cmpxchg
+	or r25=_PAGE_A,r18			// set the dirty bit
+	tbit.z p7,p6 = r18,_PAGE_P_BIT		// Check present bit
+	;;
+(p6)	cmpxchg8.acq r26=[r17],r25,ar.ccv	// Only if page is present
+	mov r24=PAGE_SHIFT<<2
+	;;
+(p6)	cmp.eq p6,p7=r26,r18			// Only if page is present
+	;;
+#ifdef CONFIG_XEN
+	mov r26=r8
+	mov r8=r25
+	;;
+(p6)	XEN_HYPER_ITC_D
+	;;
+	mov r8=r26
+	;;
+#else
+(p6)	itc.d r25				// install updated PTE
+#endif
+	/*
+	 * Tell the assemblers dependency-violation checker that the above "itc" instructions
+	 * cannot possibly affect the following loads:
+	 */
+	dv_serialize_data
+	;;
+	ld8 r18=[r17]				// read PTE again
+	;;
+	cmp.eq p6,p7=r18,r25			// is it same as the newly installed
+	;;
+(p7)	ptc.l r16,r24
+	mov ar.ccv=r28
+#else
+	;;
+1:	ld8 r18=[r17]
+	;;					// avoid RAW on r18
+	or r18=_PAGE_A,r18			// set the accessed bit
+	;;
+	st8 [r17]=r18				// store back updated PTE
+	itc.d r18				// install updated PTE
+#endif
+	mov b0=r29				// restore b0
+	mov pr=r31,-1
+#ifdef CONFIG_XEN
+	XEN_HYPER_RFI
+	dv_serialize_data
+#else
+	rfi
+#endif
+END(daccess_bit)
+
+	.org ia64_ivt+0x2c00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x2c00 Entry 11 (size 64 bundles) Break instruction (33)
+ENTRY(break_fault)
+	/*
+	 * The streamlined system call entry/exit paths only save/restore the initial part
+	 * of pt_regs.  This implies that the callers of system-calls must adhere to the
+	 * normal procedure calling conventions.
+	 *
+	 *   Registers to be saved & restored:
+	 *	CR registers: cr.ipsr, cr.iip, cr.ifs
+	 *	AR registers: ar.unat, ar.pfs, ar.rsc, ar.rnat, ar.bspstore, ar.fpsr
+	 * 	others: pr, b0, b6, loadrs, r1, r11, r12, r13, r15
+	 *   Registers to be restored only:
+	 * 	r8-r11: output value from the system call.
+	 *
+	 * During system call exit, scratch registers (including r15) are modified/cleared
+	 * to prevent leaking bits from kernel to user level.
+	 */
+	DBG_FAULT(11)
+	mov.m r16=IA64_KR(CURRENT)		// M2 r16 <- current task (12 cyc)
+#ifdef CONFIG_XEN
+	movl r22=XSI_IPSR
+	;;
+	ld8 r29=[r22],XSI_IIM_OFS-XSI_IPSR_OFS	// get ipsr, point to iip
+#else
+	mov r29=cr.ipsr				// M2 (12 cyc)
+#endif
+	mov r31=pr				// I0 (2 cyc)
+
+#ifdef CONFIG_XEN
+	;;
+	ld8 r17=[r22],XSI_IIP_OFS-XSI_IIM_OFS
+#else
+	mov r17=cr.iim				// M2 (2 cyc)
+#endif
+	mov.m r27=ar.rsc			// M2 (12 cyc)
+	mov r18=__IA64_BREAK_SYSCALL		// A
+
+	mov.m ar.rsc=0				// M2
+	mov.m r21=ar.fpsr			// M2 (12 cyc)
+	mov r19=b6				// I0 (2 cyc)
+	;;
+	mov.m r23=ar.bspstore			// M2 (12 cyc)
+	mov.m r24=ar.rnat			// M2 (5 cyc)
+	mov.i r26=ar.pfs			// I0 (2 cyc)
+
+	invala					// M0|1
+	nop.m 0					// M
+	mov r20=r1				// A			save r1
+
+	nop.m 0
+	movl r30=sys_call_table			// X
+
+#ifdef CONFIG_XEN
+	ld8 r28=[r22]
+#else
+	mov r28=cr.iip				// M2 (2 cyc)
+#endif
+	cmp.eq p0,p7=r18,r17			// I0 is this a system call?
+(p7)	br.cond.spnt non_syscall		// B  no ->
+	//
+	// From this point on, we are definitely on the syscall-path
+	// and we can use (non-banked) scratch registers.
+	//
+///////////////////////////////////////////////////////////////////////
+	mov r1=r16				// A    move task-pointer to "addl"-addressable reg
+	mov r2=r16				// A    setup r2 for ia64_syscall_setup
+	add r9=TI_FLAGS+IA64_TASK_SIZE,r16	// A	r9 = &current_thread_info()->flags
+
+	adds r16=IA64_TASK_THREAD_ON_USTACK_OFFSET,r16
+	adds r15=-1024,r15			// A    subtract 1024 from syscall number
+	mov r3=NR_syscalls - 1
+	;;
+	ld1.bias r17=[r16]			// M0|1 r17 = current->thread.on_ustack flag
+	ld4 r9=[r9]				// M0|1 r9 = current_thread_info()->flags
+	extr.u r8=r29,41,2			// I0   extract ei field from cr.ipsr
+
+	shladd r30=r15,3,r30			// A    r30 = sys_call_table + 8*(syscall-1024)
+	addl r22=IA64_RBS_OFFSET,r1		// A    compute base of RBS
+	cmp.leu p6,p7=r15,r3			// A    syscall number in range?
+	;;
+
+	lfetch.fault.excl.nt1 [r22]		// M0|1 prefetch RBS
+(p6)	ld8 r30=[r30]				// M0|1 load address of syscall entry point
+	tnat.nz.or p7,p0=r15			// I0	is syscall nr a NaT?
+
+	mov.m ar.bspstore=r22			// M2   switch to kernel RBS
+	cmp.eq p8,p9=2,r8			// A    isr.ei==2?
+	;;
+
+(p8)	mov r8=0				// A    clear ei to 0
+(p7)	movl r30=sys_ni_syscall			// X
+
+(p8)	adds r28=16,r28				// A    switch cr.iip to next bundle
+(p9)	adds r8=1,r8				// A    increment ei to next slot
+	nop.i 0
+	;;
+
+	mov.m r25=ar.unat			// M2 (5 cyc)
+	dep r29=r8,r29,41,2			// I0   insert new ei into cr.ipsr
+	adds r15=1024,r15			// A    restore original syscall number
+	//
+	// If any of the above loads miss in L1D, we'll stall here until
+	// the data arrives.
+	//
+///////////////////////////////////////////////////////////////////////
+	st1 [r16]=r0				// M2|3 clear current->thread.on_ustack flag
+	mov b6=r30				// I0   setup syscall handler branch reg early
+	cmp.eq pKStk,pUStk=r0,r17		// A    were we on kernel stacks already?
+
+	and r9=_TIF_SYSCALL_TRACEAUDIT,r9	// A    mask trace or audit
+	mov r18=ar.bsp				// M2 (12 cyc)
+(pKStk)	br.cond.spnt .break_fixup		// B	we're already in kernel-mode -- fix up RBS
+	;;
+.back_from_break_fixup:
+(pUStk)	addl r1=IA64_STK_OFFSET-IA64_PT_REGS_SIZE,r1 // A    compute base of memory stack
+	cmp.eq p14,p0=r9,r0			// A    are syscalls being traced/audited?
+	br.call.sptk.many b7=ia64_syscall_setup	// B
+1:
+	mov ar.rsc=0x3				// M2   set eager mode, pl 0, LE, loadrs=0
+	nop 0
+#ifdef CONFIG_XEN
+	mov r2=b0; br.call.sptk b0=xen_bsw1;; mov b0=r2;;
+#else
+	bsw.1					// B (6 cyc) regs are saved, switch to bank 1
+#endif
+	;;
+
+#ifdef CONFIG_XEN
+	movl r16=XSI_PSR_IC
+	mov r3=1
+	;;
+	st4 [r16]=r3,XSI_PSR_I_ADDR_OFS-XSI_PSR_IC_OFS	// vpsr.ic = 1
+#else
+	ssm psr.ic | PSR_DEFAULT_BITS		// M2	now it's safe to re-enable intr.-collection
+#endif
+	movl r3=ia64_ret_from_syscall		// X
+	;;
+
+	srlz.i					// M0   ensure interruption collection is on
+	mov rp=r3				// I0   set the real return addr
+(p10)	br.cond.spnt.many ia64_ret_from_syscall	// B    return if bad call-frame or r15 is a NaT
+
+#ifdef CONFIG_XEN
+(p15)	ld8 r16=[r16]				// vpsr.i
+	;;
+(p15)	st1 [r16]=r0,MASK_TO_PEND_OFS		// if (p15) vpsr.i = 1
+	;;		// if (p15) (vcpu->vcpu_info->evtchn_upcall_mask)=0
+(p15)	ld1 r2=[r16]	// if (vcpu->vcpu_info->evtchn_upcall_pending)
+	;;
+(p15)	cmp.ne.unc p6,p0=r2,r0
+	;;
+(p6)	XEN_HYPER_SSM_I				//   do a real ssm psr.i
+#else
+(p15)	ssm psr.i				// M2   restore psr.i
+#endif
+(p14)	br.call.sptk.many b6=b6			// B    invoke syscall-handker (ignore return addr)
+	br.cond.spnt.many ia64_trace_syscall	// B	do syscall-tracing thingamagic
+	// NOT REACHED
+///////////////////////////////////////////////////////////////////////
+	// On entry, we optimistically assumed that we're coming from user-space.
+	// For the rare cases where a system-call is done from within the kernel,
+	// we fix things up at this point:
+.break_fixup:
+	add r1=-IA64_PT_REGS_SIZE,sp		// A    allocate space for pt_regs structure
+	mov ar.rnat=r24				// M2	restore kernel's AR.RNAT
+	;;
+	mov ar.bspstore=r23			// M2	restore kernel's AR.BSPSTORE
+	br.cond.sptk .back_from_break_fixup
+END(break_fault)
+
+	.org ia64_ivt+0x3000
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x3000 Entry 12 (size 64 bundles) External Interrupt (4)
+ENTRY(interrupt)
+	DBG_FAULT(12)
+	mov r31=pr		// prepare to save predicates
+	;;
+	SAVE_MIN_WITH_COVER	// uses r31; defines r2 and r3
+#ifdef CONFIG_XEN
+	movl r3=XSI_PSR_IC
+	mov r14=1
+	;;
+	st4 [r3]=r14
+#else
+	ssm psr.ic | PSR_DEFAULT_BITS
+#endif
+	;;
+	adds r3=8,r2		// set up second base pointer for SAVE_REST
+	srlz.i			// ensure everybody knows psr.ic is back on
+	;;
+	SAVE_REST
+	;;
+	alloc r14=ar.pfs,0,0,2,0 // must be first in an insn group
+#ifdef CONFIG_XEN
+	;;
+	XEN_HYPER_GET_IVR
+	;;
+	mov out0=r8		// pass cr.ivr as first arg
+#else
+	mov out0=cr.ivr		// pass cr.ivr as first arg
+#endif
+	add out1=16,sp		// pass pointer to pt_regs as second arg
+	;;
+	srlz.d			// make sure we see the effect of cr.ivr
+	movl r14=ia64_leave_kernel
+	;;
+	mov rp=r14
+	br.call.sptk.many b6=ia64_handle_irq
+END(interrupt)
+
+	.org ia64_ivt+0x3400
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x3400 Entry 13 (size 64 bundles) Reserved
+	DBG_FAULT(13)
+	FAULT(13)
+
+	.org ia64_ivt+0x3800
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x3800 Entry 14 (size 64 bundles) Reserved
+	DBG_FAULT(14)
+	FAULT(14)
+
+	/*
+	 * There is no particular reason for this code to be here, other than that
+	 * there happens to be space here that would go unused otherwise.  If this
+	 * fault ever gets "unreserved", simply moved the following code to a more
+	 * suitable spot...
+	 *
+	 * ia64_syscall_setup() is a separate subroutine so that it can
+	 *	allocate stacked registers so it can safely demine any
+	 *	potential NaT values from the input registers.
+	 *
+	 * On entry:
+	 *	- executing on bank 0 or bank 1 register set (doesn't matter)
+	 *	-  r1: stack pointer
+	 *	-  r2: current task pointer
+	 *	-  r3: preserved
+	 *	- r11: original contents (saved ar.pfs to be saved)
+	 *	- r12: original contents (sp to be saved)
+	 *	- r13: original contents (tp to be saved)
+	 *	- r15: original contents (syscall # to be saved)
+	 *	- r18: saved bsp (after switching to kernel stack)
+	 *	- r19: saved b6
+	 *	- r20: saved r1 (gp)
+	 *	- r21: saved ar.fpsr
+	 *	- r22: kernel's register backing store base (krbs_base)
+	 *	- r23: saved ar.bspstore
+	 *	- r24: saved ar.rnat
+	 *	- r25: saved ar.unat
+	 *	- r26: saved ar.pfs
+	 *	- r27: saved ar.rsc
+	 *	- r28: saved cr.iip
+	 *	- r29: saved cr.ipsr
+	 *	- r31: saved pr
+	 *	-  b0: original contents (to be saved)
+	 * On exit:
+	 *	-  p10: TRUE if syscall is invoked with more than 8 out
+	 *		registers or r15's Nat is true
+	 *	-  r1: kernel's gp
+	 *	-  r3: preserved (same as on entry)
+	 *	-  r8: -EINVAL if p10 is true
+	 *	- r12: points to kernel stack
+	 *	- r13: points to current task
+	 *	- r14: preserved (same as on entry)
+	 *	- p13: preserved
+	 *	- p15: TRUE if interrupts need to be re-enabled
+	 *	- ar.fpsr: set to kernel settings
+	 *	-  b6: preserved (same as on entry)
+	 */
+#ifndef CONFIG_XEN
+GLOBAL_ENTRY(ia64_syscall_setup)
+#if PT(B6) != 0
+# error This code assumes that b6 is the first field in pt_regs.
+#endif
+	st8 [r1]=r19				// save b6
+	add r16=PT(CR_IPSR),r1			// initialize first base pointer
+	add r17=PT(R11),r1			// initialize second base pointer
+	;;
+	alloc r19=ar.pfs,8,0,0,0		// ensure in0-in7 are writable
+	st8 [r16]=r29,PT(AR_PFS)-PT(CR_IPSR)	// save cr.ipsr
+	tnat.nz p8,p0=in0
+
+	st8.spill [r17]=r11,PT(CR_IIP)-PT(R11)	// save r11
+	tnat.nz p9,p0=in1
+(pKStk)	mov r18=r0				// make sure r18 isn't NaT
+	;;
+
+	st8 [r16]=r26,PT(CR_IFS)-PT(AR_PFS)	// save ar.pfs
+	st8 [r17]=r28,PT(AR_UNAT)-PT(CR_IIP)	// save cr.iip
+	mov r28=b0				// save b0 (2 cyc)
+	;;
+
+	st8 [r17]=r25,PT(AR_RSC)-PT(AR_UNAT)	// save ar.unat
+	dep r19=0,r19,38,26			// clear all bits but 0..37 [I0]
+(p8)	mov in0=-1
+	;;
+
+	st8 [r16]=r19,PT(AR_RNAT)-PT(CR_IFS)	// store ar.pfs.pfm in cr.ifs
+	extr.u r11=r19,7,7	// I0		// get sol of ar.pfs
+	and r8=0x7f,r19		// A		// get sof of ar.pfs
+
+	st8 [r17]=r27,PT(AR_BSPSTORE)-PT(AR_RSC)// save ar.rsc
+	tbit.nz p15,p0=r29,IA64_PSR_I_BIT // I0
+(p9)	mov in1=-1
+	;;
+
+(pUStk) sub r18=r18,r22				// r18=RSE.ndirty*8
+	tnat.nz p10,p0=in2
+	add r11=8,r11
+	;;
+(pKStk) adds r16=PT(PR)-PT(AR_RNAT),r16		// skip over ar_rnat field
+(pKStk) adds r17=PT(B0)-PT(AR_BSPSTORE),r17	// skip over ar_bspstore field
+	tnat.nz p11,p0=in3
+	;;
+(p10)	mov in2=-1
+	tnat.nz p12,p0=in4				// [I0]
+(p11)	mov in3=-1
+	;;
+(pUStk) st8 [r16]=r24,PT(PR)-PT(AR_RNAT)	// save ar.rnat
+(pUStk) st8 [r17]=r23,PT(B0)-PT(AR_BSPSTORE)	// save ar.bspstore
+	shl r18=r18,16				// compute ar.rsc to be used for "loadrs"
+	;;
+	st8 [r16]=r31,PT(LOADRS)-PT(PR)		// save predicates
+	st8 [r17]=r28,PT(R1)-PT(B0)		// save b0
+	tnat.nz p13,p0=in5				// [I0]
+	;;
+	st8 [r16]=r18,PT(R12)-PT(LOADRS)	// save ar.rsc value for "loadrs"
+	st8.spill [r17]=r20,PT(R13)-PT(R1)	// save original r1
+(p12)	mov in4=-1
+	;;
+
+.mem.offset 0,0; st8.spill [r16]=r12,PT(AR_FPSR)-PT(R12)	// save r12
+.mem.offset 8,0; st8.spill [r17]=r13,PT(R15)-PT(R13)		// save r13
+(p13)	mov in5=-1
+	;;
+	st8 [r16]=r21,PT(R8)-PT(AR_FPSR)	// save ar.fpsr
+	tnat.nz p13,p0=in6
+	cmp.lt p10,p9=r11,r8	// frame size can't be more than local+8
+	;;
+	mov r8=1
+(p9)	tnat.nz p10,p0=r15
+	adds r12=-16,r1		// switch to kernel memory stack (with 16 bytes of scratch)
+
+	st8.spill [r17]=r15			// save r15
+	tnat.nz p8,p0=in7
+	nop.i 0
+
+	mov r13=r2				// establish `current'
+	movl r1=__gp				// establish kernel global pointer
+	;;
+	st8 [r16]=r8		// ensure pt_regs.r8 != 0 (see handle_syscall_error)
+(p13)	mov in6=-1
+(p8)	mov in7=-1
+
+	cmp.eq pSys,pNonSys=r0,r0		// set pSys=1, pNonSys=0
+	movl r17=FPSR_DEFAULT
+	;;
+	mov.m ar.fpsr=r17			// set ar.fpsr to kernel default value
+(p10)	mov r8=-EINVAL
+	br.ret.sptk.many b7
+END(ia64_syscall_setup)
+#endif
+
+	.org ia64_ivt+0x3c00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x3c00 Entry 15 (size 64 bundles) Reserved
+	DBG_FAULT(15)
+	FAULT(15)
+
+	/*
+	 * Squatting in this space ...
+	 *
+	 * This special case dispatcher for illegal operation faults allows preserved
+	 * registers to be modified through a callback function (asm only) that is handed
+	 * back from the fault handler in r8. Up to three arguments can be passed to the
+	 * callback function by returning an aggregate with the callback as its first
+	 * element, followed by the arguments.
+	 */
+ENTRY(dispatch_illegal_op_fault)
+	.prologue
+	.body
+	SAVE_MIN_WITH_COVER
+	ssm psr.ic | PSR_DEFAULT_BITS
+	;;
+	srlz.i		// guarantee that interruption collection is on
+	;;
+(p15)	ssm psr.i	// restore psr.i
+	adds r3=8,r2	// set up second base pointer for SAVE_REST
+	;;
+	alloc r14=ar.pfs,0,0,1,0	// must be first in insn group
+	mov out0=ar.ec
+	;;
+	SAVE_REST
+	PT_REGS_UNWIND_INFO(0)
+	;;
+	br.call.sptk.many rp=ia64_illegal_op_fault
+.ret0:	;;
+	alloc r14=ar.pfs,0,0,3,0	// must be first in insn group
+	mov out0=r9
+	mov out1=r10
+	mov out2=r11
+	movl r15=ia64_leave_kernel
+	;;
+	mov rp=r15
+	mov b6=r8
+	;;
+	cmp.ne p6,p0=0,r8
+(p6)	br.call.dpnt.many b6=b6		// call returns to ia64_leave_kernel
+	br.sptk.many ia64_leave_kernel
+END(dispatch_illegal_op_fault)
+
+	.org ia64_ivt+0x4000
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x4000 Entry 16 (size 64 bundles) Reserved
+	DBG_FAULT(16)
+	FAULT(16)
+
+	.org ia64_ivt+0x4400
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x4400 Entry 17 (size 64 bundles) Reserved
+	DBG_FAULT(17)
+	FAULT(17)
+
+ENTRY(non_syscall)
+	mov ar.rsc=r27			// restore ar.rsc before SAVE_MIN_WITH_COVER
+	;;
+	SAVE_MIN_WITH_COVER
+
+	// There is no particular reason for this code to be here, other than that
+	// there happens to be space here that would go unused otherwise.  If this
+	// fault ever gets "unreserved", simply moved the following code to a more
+	// suitable spot...
+
+	alloc r14=ar.pfs,0,0,2,0
+	mov out0=cr.iim
+	add out1=16,sp
+	adds r3=8,r2			// set up second base pointer for SAVE_REST
+
+	ssm psr.ic | PSR_DEFAULT_BITS
+	;;
+	srlz.i				// guarantee that interruption collection is on
+	;;
+(p15)	ssm psr.i			// restore psr.i
+	movl r15=ia64_leave_kernel
+	;;
+	SAVE_REST
+	mov rp=r15
+	;;
+	br.call.sptk.many b6=ia64_bad_break	// avoid WAW on CFM and ignore return addr
+END(non_syscall)
+
+	.org ia64_ivt+0x4800
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x4800 Entry 18 (size 64 bundles) Reserved
+	DBG_FAULT(18)
+	FAULT(18)
+
+	/*
+	 * There is no particular reason for this code to be here, other than that
+	 * there happens to be space here that would go unused otherwise.  If this
+	 * fault ever gets "unreserved", simply moved the following code to a more
+	 * suitable spot...
+	 */
+
+ENTRY(dispatch_unaligned_handler)
+	SAVE_MIN_WITH_COVER
+	;;
+	alloc r14=ar.pfs,0,0,2,0		// now it's safe (must be first in insn group!)
+	mov out0=cr.ifa
+	adds out1=16,sp
+
+	ssm psr.ic | PSR_DEFAULT_BITS
+	;;
+	srlz.i					// guarantee that interruption collection is on
+	;;
+(p15)	ssm psr.i				// restore psr.i
+	adds r3=8,r2				// set up second base pointer
+	;;
+	SAVE_REST
+	movl r14=ia64_leave_kernel
+	;;
+	mov rp=r14
+	br.sptk.many ia64_prepare_handle_unaligned
+END(dispatch_unaligned_handler)
+
+	.org ia64_ivt+0x4c00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x4c00 Entry 19 (size 64 bundles) Reserved
+	DBG_FAULT(19)
+	FAULT(19)
+
+	/*
+	 * There is no particular reason for this code to be here, other than that
+	 * there happens to be space here that would go unused otherwise.  If this
+	 * fault ever gets "unreserved", simply moved the following code to a more
+	 * suitable spot...
+	 */
+
+ENTRY(dispatch_to_fault_handler)
+	/*
+	 * Input:
+	 *	psr.ic:	off
+	 *	r19:	fault vector number (e.g., 24 for General Exception)
+	 *	r31:	contains saved predicates (pr)
+	 */
+	SAVE_MIN_WITH_COVER_R19
+	alloc r14=ar.pfs,0,0,5,0
+	mov out0=r15
+#ifdef CONFIG_XEN
+	movl out1=XSI_ISR
+	;;
+	adds out2=XSI_IFA-XSI_ISR,out1
+	adds out3=XSI_IIM-XSI_ISR,out1
+	adds out4=XSI_ITIR-XSI_ISR,out1
+	;;
+	ld8 out1=[out1]
+	ld8 out2=[out2]
+	ld8 out3=[out4]
+	ld8 out4=[out4]
+	;;
+#else
+	mov out1=cr.isr
+	mov out2=cr.ifa
+	mov out3=cr.iim
+	mov out4=cr.itir
+	;;
+#endif
+	ssm psr.ic | PSR_DEFAULT_BITS
+	;;
+	srlz.i					// guarantee that interruption collection is on
+	;;
+(p15)	ssm psr.i				// restore psr.i
+	adds r3=8,r2				// set up second base pointer for SAVE_REST
+	;;
+	SAVE_REST
+	movl r14=ia64_leave_kernel
+	;;
+	mov rp=r14
+	br.call.sptk.many b6=ia64_fault
+END(dispatch_to_fault_handler)
+
+//
+// --- End of long entries, Beginning of short entries
+//
+
+	.org ia64_ivt+0x5000
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x5000 Entry 20 (size 16 bundles) Page Not Present (10,22,49)
+ENTRY(page_not_present)
+	DBG_FAULT(20)
+	mov r16=cr.ifa
+	rsm psr.dt
+	/*
+	 * The Linux page fault handler doesn't expect non-present pages to be in
+	 * the TLB.  Flush the existing entry now, so we meet that expectation.
+	 */
+	mov r17=PAGE_SHIFT<<2
+	;;
+	ptc.l r16,r17
+	;;
+	mov r31=pr
+	srlz.d
+	br.sptk.many page_fault
+END(page_not_present)
+
+	.org ia64_ivt+0x5100
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x5100 Entry 21 (size 16 bundles) Key Permission (13,25,52)
+ENTRY(key_permission)
+	DBG_FAULT(21)
+	mov r16=cr.ifa
+	rsm psr.dt
+	mov r31=pr
+	;;
+	srlz.d
+	br.sptk.many page_fault
+END(key_permission)
+
+	.org ia64_ivt+0x5200
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x5200 Entry 22 (size 16 bundles) Instruction Access Rights (26)
+ENTRY(iaccess_rights)
+	DBG_FAULT(22)
+	mov r16=cr.ifa
+	rsm psr.dt
+	mov r31=pr
+	;;
+	srlz.d
+	br.sptk.many page_fault
+END(iaccess_rights)
+
+	.org ia64_ivt+0x5300
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x5300 Entry 23 (size 16 bundles) Data Access Rights (14,53)
+ENTRY(daccess_rights)
+	DBG_FAULT(23)
+#ifdef CONFIG_XEN
+	movl r16=XSI_IFA
+	;;
+	ld8 r16=[r16]
+	;;
+	XEN_HYPER_RSM_PSR_DT
+#else
+	mov r16=cr.ifa
+	rsm psr.dt
+#endif
+	mov r31=pr
+	;;
+	srlz.d
+	br.sptk.many page_fault
+END(daccess_rights)
+
+	.org ia64_ivt+0x5400
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x5400 Entry 24 (size 16 bundles) General Exception (5,32,34,36,38,39)
+ENTRY(general_exception)
+	DBG_FAULT(24)
+	mov r16=cr.isr
+	mov r31=pr
+	;;
+	cmp4.eq p6,p0=0,r16
+(p6)	br.sptk.many dispatch_illegal_op_fault
+	;;
+	mov r19=24		// fault number
+	br.sptk.many dispatch_to_fault_handler
+END(general_exception)
+
+	.org ia64_ivt+0x5500
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x5500 Entry 25 (size 16 bundles) Disabled FP-Register (35)
+ENTRY(disabled_fp_reg)
+	DBG_FAULT(25)
+	rsm psr.dfh		// ensure we can access fph
+	;;
+	srlz.d
+	mov r31=pr
+	mov r19=25
+	br.sptk.many dispatch_to_fault_handler
+END(disabled_fp_reg)
+
+	.org ia64_ivt+0x5600
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x5600 Entry 26 (size 16 bundles) Nat Consumption (11,23,37,50)
+ENTRY(nat_consumption)
+	DBG_FAULT(26)
+
+	mov r16=cr.ipsr
+	mov r17=cr.isr
+	mov r31=pr				// save PR
+	;;
+	and r18=0xf,r17				// r18 = cr.ipsr.code{3:0}
+	tbit.z p6,p0=r17,IA64_ISR_NA_BIT
+	;;
+	cmp.ne.or p6,p0=IA64_ISR_CODE_LFETCH,r18
+	dep r16=-1,r16,IA64_PSR_ED_BIT,1
+(p6)	br.cond.spnt 1f		// branch if (cr.ispr.na == 0 || cr.ipsr.code{3:0} != LFETCH)
+	;;
+	mov cr.ipsr=r16		// set cr.ipsr.na
+	mov pr=r31,-1
+	;;
+	rfi
+
+1:	mov pr=r31,-1
+	;;
+	FAULT(26)
+END(nat_consumption)
+
+	.org ia64_ivt+0x5700
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x5700 Entry 27 (size 16 bundles) Speculation (40)
+ENTRY(speculation_vector)
+	DBG_FAULT(27)
+	/*
+	 * A [f]chk.[as] instruction needs to take the branch to the recovery code but
+	 * this part of the architecture is not implemented in hardware on some CPUs, such
+	 * as Itanium.  Thus, in general we need to emulate the behavior.  IIM contains
+	 * the relative target (not yet sign extended).  So after sign extending it we
+	 * simply add it to IIP.  We also need to reset the EI field of the IPSR to zero,
+	 * i.e., the slot to restart into.
+	 *
+	 * cr.imm contains zero_ext(imm21)
+	 */
+	mov r18=cr.iim
+	;;
+	mov r17=cr.iip
+	shl r18=r18,43			// put sign bit in position (43=64-21)
+	;;
+
+	mov r16=cr.ipsr
+	shr r18=r18,39			// sign extend (39=43-4)
+	;;
+
+	add r17=r17,r18			// now add the offset
+	;;
+	mov cr.iip=r17
+	dep r16=0,r16,41,2		// clear EI
+	;;
+
+	mov cr.ipsr=r16
+	;;
+
+#ifdef CONFIG_XEN
+	XEN_HYPER_RFI;
+#else
+	rfi				// and go back
+#endif
+END(speculation_vector)
+
+	.org ia64_ivt+0x5800
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x5800 Entry 28 (size 16 bundles) Reserved
+	DBG_FAULT(28)
+	FAULT(28)
+
+	.org ia64_ivt+0x5900
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x5900 Entry 29 (size 16 bundles) Debug (16,28,56)
+ENTRY(debug_vector)
+	DBG_FAULT(29)
+	FAULT(29)
+END(debug_vector)
+
+	.org ia64_ivt+0x5a00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x5a00 Entry 30 (size 16 bundles) Unaligned Reference (57)
+ENTRY(unaligned_access)
+	DBG_FAULT(30)
+	mov r31=pr		// prepare to save predicates
+	;;
+	br.sptk.many dispatch_unaligned_handler
+END(unaligned_access)
+
+	.org ia64_ivt+0x5b00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x5b00 Entry 31 (size 16 bundles) Unsupported Data Reference (57)
+ENTRY(unsupported_data_reference)
+	DBG_FAULT(31)
+	FAULT(31)
+END(unsupported_data_reference)
+
+	.org ia64_ivt+0x5c00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x5c00 Entry 32 (size 16 bundles) Floating-Point Fault (64)
+ENTRY(floating_point_fault)
+	DBG_FAULT(32)
+	FAULT(32)
+END(floating_point_fault)
+
+	.org ia64_ivt+0x5d00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x5d00 Entry 33 (size 16 bundles) Floating Point Trap (66)
+ENTRY(floating_point_trap)
+	DBG_FAULT(33)
+	FAULT(33)
+END(floating_point_trap)
+
+	.org ia64_ivt+0x5e00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x5e00 Entry 34 (size 16 bundles) Lower Privilege Transfer Trap (66)
+ENTRY(lower_privilege_trap)
+	DBG_FAULT(34)
+	FAULT(34)
+END(lower_privilege_trap)
+
+	.org ia64_ivt+0x5f00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x5f00 Entry 35 (size 16 bundles) Taken Branch Trap (68)
+ENTRY(taken_branch_trap)
+	DBG_FAULT(35)
+	FAULT(35)
+END(taken_branch_trap)
+
+	.org ia64_ivt+0x6000
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x6000 Entry 36 (size 16 bundles) Single Step Trap (69)
+ENTRY(single_step_trap)
+	DBG_FAULT(36)
+	FAULT(36)
+END(single_step_trap)
+
+	.org ia64_ivt+0x6100
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x6100 Entry 37 (size 16 bundles) Reserved
+	DBG_FAULT(37)
+	FAULT(37)
+
+	.org ia64_ivt+0x6200
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x6200 Entry 38 (size 16 bundles) Reserved
+	DBG_FAULT(38)
+	FAULT(38)
+
+	.org ia64_ivt+0x6300
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x6300 Entry 39 (size 16 bundles) Reserved
+	DBG_FAULT(39)
+	FAULT(39)
+
+	.org ia64_ivt+0x6400
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x6400 Entry 40 (size 16 bundles) Reserved
+	DBG_FAULT(40)
+	FAULT(40)
+
+	.org ia64_ivt+0x6500
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x6500 Entry 41 (size 16 bundles) Reserved
+	DBG_FAULT(41)
+	FAULT(41)
+
+	.org ia64_ivt+0x6600
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x6600 Entry 42 (size 16 bundles) Reserved
+	DBG_FAULT(42)
+	FAULT(42)
+
+	.org ia64_ivt+0x6700
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x6700 Entry 43 (size 16 bundles) Reserved
+	DBG_FAULT(43)
+	FAULT(43)
+
+	.org ia64_ivt+0x6800
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x6800 Entry 44 (size 16 bundles) Reserved
+	DBG_FAULT(44)
+	FAULT(44)
+
+	.org ia64_ivt+0x6900
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x6900 Entry 45 (size 16 bundles) IA-32 Exeception (17,18,29,41,42,43,44,58,60,61,62,72,73,75,76,77)
+ENTRY(ia32_exception)
+	DBG_FAULT(45)
+	FAULT(45)
+END(ia32_exception)
+
+	.org ia64_ivt+0x6a00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x6a00 Entry 46 (size 16 bundles) IA-32 Intercept  (30,31,59,70,71)
+ENTRY(ia32_intercept)
+	DBG_FAULT(46)
+#ifdef	CONFIG_IA32_SUPPORT
+	mov r31=pr
+	mov r16=cr.isr
+	;;
+	extr.u r17=r16,16,8	// get ISR.code
+	mov r18=ar.eflag
+	mov r19=cr.iim		// old eflag value
+	;;
+	cmp.ne p6,p0=2,r17
+(p6)	br.cond.spnt 1f		// not a system flag fault
+	xor r16=r18,r19
+	;;
+	extr.u r17=r16,18,1	// get the eflags.ac bit
+	;;
+	cmp.eq p6,p0=0,r17
+(p6)	br.cond.spnt 1f		// eflags.ac bit didn't change
+	;;
+	mov pr=r31,-1		// restore predicate registers
+#ifdef CONFIG_XEN
+	XEN_HYPER_RFI;
+#else
+	rfi
+#endif
+
+1:
+#endif	// CONFIG_IA32_SUPPORT
+	FAULT(46)
+END(ia32_intercept)
+
+	.org ia64_ivt+0x6b00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x6b00 Entry 47 (size 16 bundles) IA-32 Interrupt  (74)
+ENTRY(ia32_interrupt)
+	DBG_FAULT(47)
+#ifdef CONFIG_IA32_SUPPORT
+	mov r31=pr
+	br.sptk.many dispatch_to_ia32_handler
+#else
+	FAULT(47)
+#endif
+END(ia32_interrupt)
+
+	.org ia64_ivt+0x6c00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x6c00 Entry 48 (size 16 bundles) Reserved
+	DBG_FAULT(48)
+	FAULT(48)
+
+	.org ia64_ivt+0x6d00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x6d00 Entry 49 (size 16 bundles) Reserved
+	DBG_FAULT(49)
+	FAULT(49)
+
+	.org ia64_ivt+0x6e00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x6e00 Entry 50 (size 16 bundles) Reserved
+	DBG_FAULT(50)
+	FAULT(50)
+
+	.org ia64_ivt+0x6f00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x6f00 Entry 51 (size 16 bundles) Reserved
+	DBG_FAULT(51)
+	FAULT(51)
+
+	.org ia64_ivt+0x7000
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x7000 Entry 52 (size 16 bundles) Reserved
+	DBG_FAULT(52)
+	FAULT(52)
+
+	.org ia64_ivt+0x7100
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x7100 Entry 53 (size 16 bundles) Reserved
+	DBG_FAULT(53)
+	FAULT(53)
+
+	.org ia64_ivt+0x7200
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x7200 Entry 54 (size 16 bundles) Reserved
+	DBG_FAULT(54)
+	FAULT(54)
+
+	.org ia64_ivt+0x7300
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x7300 Entry 55 (size 16 bundles) Reserved
+	DBG_FAULT(55)
+	FAULT(55)
+
+	.org ia64_ivt+0x7400
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x7400 Entry 56 (size 16 bundles) Reserved
+	DBG_FAULT(56)
+	FAULT(56)
+
+	.org ia64_ivt+0x7500
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x7500 Entry 57 (size 16 bundles) Reserved
+	DBG_FAULT(57)
+	FAULT(57)
+
+	.org ia64_ivt+0x7600
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x7600 Entry 58 (size 16 bundles) Reserved
+	DBG_FAULT(58)
+	FAULT(58)
+
+	.org ia64_ivt+0x7700
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x7700 Entry 59 (size 16 bundles) Reserved
+	DBG_FAULT(59)
+	FAULT(59)
+
+	.org ia64_ivt+0x7800
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x7800 Entry 60 (size 16 bundles) Reserved
+	DBG_FAULT(60)
+	FAULT(60)
+
+	.org ia64_ivt+0x7900
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x7900 Entry 61 (size 16 bundles) Reserved
+	DBG_FAULT(61)
+	FAULT(61)
+
+	.org ia64_ivt+0x7a00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x7a00 Entry 62 (size 16 bundles) Reserved
+	DBG_FAULT(62)
+	FAULT(62)
+
+	.org ia64_ivt+0x7b00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x7b00 Entry 63 (size 16 bundles) Reserved
+	DBG_FAULT(63)
+	FAULT(63)
+
+	.org ia64_ivt+0x7c00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x7c00 Entry 64 (size 16 bundles) Reserved
+	DBG_FAULT(64)
+	FAULT(64)
+
+	.org ia64_ivt+0x7d00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x7d00 Entry 65 (size 16 bundles) Reserved
+	DBG_FAULT(65)
+	FAULT(65)
+
+	.org ia64_ivt+0x7e00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x7e00 Entry 66 (size 16 bundles) Reserved
+	DBG_FAULT(66)
+	FAULT(66)
+
+	.org ia64_ivt+0x7f00
+/////////////////////////////////////////////////////////////////////////////////////////
+// 0x7f00 Entry 67 (size 16 bundles) Reserved
+	DBG_FAULT(67)
+	FAULT(67)
+
+#ifdef CONFIG_IA32_SUPPORT
+
+	/*
+	 * There is no particular reason for this code to be here, other than that
+	 * there happens to be space here that would go unused otherwise.  If this
+	 * fault ever gets "unreserved", simply moved the following code to a more
+	 * suitable spot...
+	 */
+
+	// IA32 interrupt entry point
+
+ENTRY(dispatch_to_ia32_handler)
+	SAVE_MIN
+	;;
+	mov r14=cr.isr
+	ssm psr.ic | PSR_DEFAULT_BITS
+	;;
+	srlz.i					// guarantee that interruption collection is on
+	;;
+(p15)	ssm psr.i
+	adds r3=8,r2		// Base pointer for SAVE_REST
+	;;
+	SAVE_REST
+	;;
+	mov r15=0x80
+	shr r14=r14,16		// Get interrupt number
+	;;
+	cmp.ne p6,p0=r14,r15
+(p6)	br.call.dpnt.many b6=non_ia32_syscall
+
+	adds r14=IA64_PT_REGS_R8_OFFSET + 16,sp	// 16 byte hole per SW conventions
+	adds r15=IA64_PT_REGS_R1_OFFSET + 16,sp
+	;;
+	cmp.eq pSys,pNonSys=r0,r0 // set pSys=1, pNonSys=0
+	ld8 r8=[r14]		// get r8
+	;;
+	st8 [r15]=r8		// save original EAX in r1 (IA32 procs don't use the GP)
+	;;
+	alloc r15=ar.pfs,0,0,6,0	// must first in an insn group
+	;;
+	ld4 r8=[r14],8		// r8 == eax (syscall number)
+	mov r15=IA32_NR_syscalls
+	;;
+	cmp.ltu.unc p6,p7=r8,r15
+	ld4 out1=[r14],8	// r9 == ecx
+	;;
+	ld4 out2=[r14],8	// r10 == edx
+	;;
+	ld4 out0=[r14]		// r11 == ebx
+	adds r14=(IA64_PT_REGS_R13_OFFSET) + 16,sp
+	;;
+	ld4 out5=[r14],PT(R14)-PT(R13)	// r13 == ebp
+	;;
+	ld4 out3=[r14],PT(R15)-PT(R14)	// r14 == esi
+	adds r2=TI_FLAGS+IA64_TASK_SIZE,r13
+	;;
+	ld4 out4=[r14]		// r15 == edi
+	movl r16=ia32_syscall_table
+	;;
+(p6)	shladd r16=r8,3,r16	// force ni_syscall if not valid syscall number
+	ld4 r2=[r2]		// r2 = current_thread_info()->flags
+	;;
+	ld8 r16=[r16]
+	and r2=_TIF_SYSCALL_TRACEAUDIT,r2	// mask trace or audit
+	;;
+	mov b6=r16
+	movl r15=ia32_ret_from_syscall
+	cmp.eq p8,p0=r2,r0
+	;;
+	mov rp=r15
+(p8)	br.call.sptk.many b6=b6
+	br.cond.sptk ia32_trace_syscall
+
+non_ia32_syscall:
+	alloc r15=ar.pfs,0,0,2,0
+	mov out0=r14				// interrupt #
+	add out1=16,sp				// pointer to pt_regs
+	;;			// avoid WAW on CFM
+	br.call.sptk.many rp=ia32_bad_interrupt
+.ret1:	movl r15=ia64_leave_kernel
+	;;
+	mov rp=r15
+	br.ret.sptk.many rp
+END(dispatch_to_ia32_handler)
+#endif /* CONFIG_IA32_SUPPORT */
+
+#ifdef CONFIG_XEN
+	.section .text,"ax"
+GLOBAL_ENTRY(xen_event_callback)
+	mov r31=pr		// prepare to save predicates
+	;;
+	SAVE_MIN_WITH_COVER	// uses r31; defines r2 and r3
+	;;
+	movl r3=XSI_PSR_IC
+	mov r14=1
+	;;
+	st4 [r3]=r14
+	;;
+	adds r3=8,r2		// set up second base pointer for SAVE_REST
+	srlz.i			// ensure everybody knows psr.ic is back on
+	;;
+	SAVE_REST
+	;;
+1:
+	alloc r14=ar.pfs,0,0,1,0 // must be first in an insn group
+	add out0=16,sp		// pass pointer to pt_regs as first arg
+	;;
+	br.call.sptk.many b0=evtchn_do_upcall
+	;;
+	movl r20=XSI_PSR_I_ADDR
+	;;
+	ld8 r20=[r20]
+	;;
+	adds r20=-1,r20		// vcpu_info->evtchn_upcall_pending
+	;;
+	ld1 r20=[r20]
+	;;
+	cmp.ne p6,p0=r20,r0	// if there are pending events, 
+	(p6) br.spnt.few 1b	// call evtchn_do_upcall again.
+	br.sptk.many ia64_leave_kernel   
+END(xen_event_callback)
+
+
+	/*
+	 * There is no particular reason for this code to be here, other than that
+	 * there happens to be space here that would go unused otherwise.  If this
+	 * fault ever gets "unreserved", simply moved the following code to a more
+	 * suitable spot...
+	 */
+
+GLOBAL_ENTRY(xen_bsw1)
+	/* FIXME: THIS CODE IS NOT NaT SAFE! */
+	mov r14=ar.unat
+	movl r30=XSI_B1NAT
+	;;
+	ld8 r30=[r30];;
+	mov ar.unat=r30
+	movl r30=XSI_BANKNUM;
+	mov r31=1;;
+	st4 [r30]=r31;
+	movl r30=XSI_BANK1_R16;
+	movl r31=XSI_BANK1_R16+8;;
+	ld8.fill r16=[r30],16; ld8.fill r17=[r31],16;;
+	ld8.fill r18=[r30],16; ld8.fill r19=[r31],16;;
+	ld8.fill r20=[r30],16; ld8.fill r21=[r31],16;;
+	ld8.fill r22=[r30],16; ld8.fill r23=[r31],16;;
+	ld8.fill r24=[r30],16; ld8.fill r25=[r31],16;;
+	ld8.fill r26=[r30],16; ld8.fill r27=[r31],16;;
+	ld8.fill r28=[r30],16; ld8.fill r29=[r31],16;;
+	ld8.fill r30=[r30]; ld8.fill r31=[r31];;
+	mov ar.unat=r14
+	br.ret.sptk.many b0
+END(xen_bsw1)
+
+   
+#endif
diff -rpuN linux-2.6.18.8/arch/ia64/xen/xenminstate.h linux-2.6.18-xen-3.3.0/arch/ia64/xen/xenminstate.h
--- linux-2.6.18.8/arch/ia64/xen/xenminstate.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/xen/xenminstate.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,358 @@
+#include <asm/cache.h>
+
+#ifdef CONFIG_XEN
+#include "../kernel/entry.h"
+#else
+#include "entry.h"
+#endif
+
+/*
+ * For ivt.s we want to access the stack virtually so we don't have to disable translation
+ * on interrupts.
+ *
+ *  On entry:
+ *	r1:	pointer to current task (ar.k6)
+ */
+#define MINSTATE_START_SAVE_MIN_VIRT								\
+(pUStk)	mov ar.rsc=0;		/* set enforced lazy mode, pl 0, little-endian, loadrs=0 */	\
+	;;											\
+(pUStk)	mov.m r24=ar.rnat;									\
+(pUStk)	addl r22=IA64_RBS_OFFSET,r1;			/* compute base of RBS */		\
+(pKStk) mov r1=sp;					/* get sp  */				\
+	;;											\
+(pUStk) lfetch.fault.excl.nt1 [r22];								\
+(pUStk)	addl r1=IA64_STK_OFFSET-IA64_PT_REGS_SIZE,r1;	/* compute base of memory stack */	\
+(pUStk)	mov r23=ar.bspstore;				/* save ar.bspstore */			\
+	;;											\
+(pUStk)	mov ar.bspstore=r22;				/* switch to kernel RBS */		\
+(pKStk) addl r1=-IA64_PT_REGS_SIZE,r1;			/* if in kernel mode, use sp (r12) */	\
+	;;											\
+(pUStk)	mov r18=ar.bsp;										\
+(pUStk)	mov ar.rsc=0x3;		/* set eager mode, pl 0, little-endian, loadrs=0 */		\
+
+#define MINSTATE_END_SAVE_MIN_VIRT								\
+	bsw.1;			/* switch back to bank 1 (must be last in insn group) */	\
+	;;
+
+/*
+ * For mca_asm.S we want to access the stack physically since the state is saved before we
+ * go virtual and don't want to destroy the iip or ipsr.
+ */
+#define MINSTATE_START_SAVE_MIN_PHYS								\
+(pKStk) mov r3=IA64_KR(PER_CPU_DATA);;								\
+(pKStk) addl r3=THIS_CPU(ia64_mca_data),r3;;							\
+(pKStk) ld8 r3 = [r3];;										\
+(pKStk) addl r3=IA64_MCA_CPU_INIT_STACK_OFFSET,r3;;						\
+(pKStk) addl sp=IA64_STK_OFFSET-IA64_PT_REGS_SIZE,r3;						\
+(pUStk)	mov ar.rsc=0;		/* set enforced lazy mode, pl 0, little-endian, loadrs=0 */	\
+(pUStk)	addl r22=IA64_RBS_OFFSET,r1;		/* compute base of register backing store */	\
+	;;											\
+(pUStk)	mov r24=ar.rnat;									\
+(pUStk)	addl r1=IA64_STK_OFFSET-IA64_PT_REGS_SIZE,r1;	/* compute base of memory stack */	\
+(pUStk)	mov r23=ar.bspstore;				/* save ar.bspstore */			\
+(pUStk)	dep r22=-1,r22,61,3;			/* compute kernel virtual addr of RBS */	\
+	;;											\
+(pKStk) addl r1=-IA64_PT_REGS_SIZE,r1;		/* if in kernel mode, use sp (r12) */		\
+(pUStk)	mov ar.bspstore=r22;			/* switch to kernel RBS */			\
+	;;											\
+(pUStk)	mov r18=ar.bsp;										\
+(pUStk)	mov ar.rsc=0x3;		/* set eager mode, pl 0, little-endian, loadrs=0 */		\
+
+#define MINSTATE_END_SAVE_MIN_PHYS								\
+	dep r12=-1,r12,61,3;		/* make sp a kernel virtual address */			\
+	;;
+
+#ifdef MINSTATE_VIRT
+# define MINSTATE_GET_CURRENT(reg)	mov reg=IA64_KR(CURRENT)
+# define MINSTATE_START_SAVE_MIN	MINSTATE_START_SAVE_MIN_VIRT
+# define MINSTATE_END_SAVE_MIN		MINSTATE_END_SAVE_MIN_VIRT
+#endif
+
+#ifdef MINSTATE_PHYS
+# define MINSTATE_GET_CURRENT(reg)	mov reg=IA64_KR(CURRENT);; tpa reg=reg
+# define MINSTATE_START_SAVE_MIN	MINSTATE_START_SAVE_MIN_PHYS
+# define MINSTATE_END_SAVE_MIN		MINSTATE_END_SAVE_MIN_PHYS
+#endif
+
+/*
+ * DO_SAVE_MIN switches to the kernel stacks (if necessary) and saves
+ * the minimum state necessary that allows us to turn psr.ic back
+ * on.
+ *
+ * Assumed state upon entry:
+ *	psr.ic: off
+ *	r31:	contains saved predicates (pr)
+ *
+ * Upon exit, the state is as follows:
+ *	psr.ic: off
+ *	 r2 = points to &pt_regs.r16
+ *	 r8 = contents of ar.ccv
+ *	 r9 = contents of ar.csd
+ *	r10 = contents of ar.ssd
+ *	r11 = FPSR_DEFAULT
+ *	r12 = kernel sp (kernel virtual address)
+ *	r13 = points to current task_struct (kernel virtual address)
+ *	p15 = TRUE if psr.i is set in cr.ipsr
+ *	predicate registers (other than p2, p3, and p15), b6, r3, r14, r15:
+ *		preserved
+ * CONFIG_XEN note: p6/p7 are not preserved
+ *
+ * Note that psr.ic is NOT turned on by this macro.  This is so that
+ * we can pass interruption state as arguments to a handler.
+ */
+#ifdef CONFIG_XEN
+#define DO_SAVE_MIN(COVER,SAVE_IFS,EXTRA)							\
+	MINSTATE_GET_CURRENT(r16);	/* M (or M;;I) */					\
+	mov r27=ar.rsc;			/* M */							\
+	mov r20=r1;			/* A */							\
+	mov r25=ar.unat;		/* M */							\
+	/* mov r29=cr.ipsr;		/* M */							\
+	movl r29=XSI_IPSR;;									\
+	ld8 r29=[r29];;										\
+	mov r26=ar.pfs;			/* I */							\
+	/* mov r28=cr.iip;		/* M */							\
+	movl r28=XSI_IIP;;									\
+	ld8 r28=[r28];;										\
+	mov r21=ar.fpsr;		/* M */							\
+	COVER;			/* B;; (or nothing) */					\
+	;;											\
+	adds r16=IA64_TASK_THREAD_ON_USTACK_OFFSET,r16;						\
+	;;											\
+	ld1 r17=[r16];				/* load current->thread.on_ustack flag */	\
+	st1 [r16]=r0;				/* clear current->thread.on_ustack flag */	\
+	adds r1=-IA64_TASK_THREAD_ON_USTACK_OFFSET,r16						\
+	/* switch from user to kernel RBS: */							\
+	;;											\
+	invala;				/* M */							\
+	/* SAVE_IFS; /* see xen special handling below */						\
+	cmp.eq pKStk,pUStk=r0,r17;		/* are we in kernel mode already? */		\
+	;;											\
+	MINSTATE_START_SAVE_MIN									\
+	adds r17=2*L1_CACHE_BYTES,r1;		/* really: biggest cache-line size */		\
+	adds r16=PT(CR_IPSR),r1;								\
+	;;											\
+	lfetch.fault.excl.nt1 [r17],L1_CACHE_BYTES;						\
+	st8 [r16]=r29;		/* save cr.ipsr */						\
+	;;											\
+	lfetch.fault.excl.nt1 [r17];								\
+	tbit.nz p15,p0=r29,IA64_PSR_I_BIT;							\
+	mov r29=b0										\
+	;;											\
+	adds r16=PT(R8),r1;	/* initialize first base pointer */				\
+	adds r17=PT(R9),r1;	/* initialize second base pointer */				\
+(pKStk)	mov r18=r0;		/* make sure r18 isn't NaT */					\
+	;;											\
+.mem.offset 0,0; st8.spill [r16]=r8,16;								\
+.mem.offset 8,0; st8.spill [r17]=r9,16;								\
+        ;;											\
+.mem.offset 0,0; st8.spill [r16]=r10,24;							\
+.mem.offset 8,0; st8.spill [r17]=r11,24;							\
+        ;;											\
+	/* xen special handling for possibly lazy cover */					\
+	movl r8=XSI_PRECOVER_IFS;								\
+	;;											\
+	ld8 r30=[r8];										\
+	;;											\
+	st8 [r16]=r28,16;	/* save cr.iip */						\
+	st8 [r17]=r30,16;	/* save cr.ifs */						\
+(pUStk)	sub r18=r18,r22;	/* r18=RSE.ndirty*8 */						\
+	mov r8=ar.ccv;										\
+	mov r9=ar.csd;										\
+	mov r10=ar.ssd;										\
+	movl r11=FPSR_DEFAULT;   /* L-unit */							\
+	;;											\
+	st8 [r16]=r25,16;	/* save ar.unat */						\
+	st8 [r17]=r26,16;	/* save ar.pfs */						\
+	shl r18=r18,16;		/* compute ar.rsc to be used for "loadrs" */			\
+	;;											\
+	st8 [r16]=r27,16;	/* save ar.rsc */						\
+(pUStk)	st8 [r17]=r24,16;	/* save ar.rnat */						\
+(pKStk)	adds r17=16,r17;	/* skip over ar_rnat field */					\
+	;;			/* avoid RAW on r16 & r17 */					\
+(pUStk)	st8 [r16]=r23,16;	/* save ar.bspstore */						\
+	st8 [r17]=r31,16;	/* save predicates */						\
+(pKStk)	adds r16=16,r16;	/* skip over ar_bspstore field */				\
+	;;											\
+	st8 [r16]=r29,16;	/* save b0 */							\
+	st8 [r17]=r18,16;	/* save ar.rsc value for "loadrs" */				\
+	cmp.eq pNonSys,pSys=r0,r0	/* initialize pSys=0, pNonSys=1 */			\
+	;;											\
+.mem.offset 0,0; st8.spill [r16]=r20,16;	/* save original r1 */				\
+.mem.offset 8,0; st8.spill [r17]=r12,16;							\
+	adds r12=-16,r1;	/* switch to kernel memory stack (with 16 bytes of scratch) */	\
+	;;											\
+.mem.offset 0,0; st8.spill [r16]=r13,16;							\
+.mem.offset 8,0; st8.spill [r17]=r21,16;	/* save ar.fpsr */				\
+	mov r13=IA64_KR(CURRENT);	/* establish `current' */				\
+	;;											\
+.mem.offset 0,0; st8.spill [r16]=r15,16;							\
+.mem.offset 8,0; st8.spill [r17]=r14,16;							\
+	;;											\
+.mem.offset 0,0; st8.spill [r16]=r2,16;								\
+.mem.offset 8,0; st8.spill [r17]=r3,16;								\
+	;;											\
+	EXTRA;											\
+	mov r2=b0; br.call.sptk b0=xen_bsw1;; mov b0=r2;					\
+	adds r2=IA64_PT_REGS_R16_OFFSET,r1;							\
+	;;											\
+	movl r1=__gp;		/* establish kernel global pointer */				\
+	;;											\
+	/* MINSTATE_END_SAVE_MIN */
+#else
+#define DO_SAVE_MIN(COVER,SAVE_IFS,EXTRA)							\
+	MINSTATE_GET_CURRENT(r16);	/* M (or M;;I) */					\
+	mov r27=ar.rsc;			/* M */							\
+	mov r20=r1;			/* A */							\
+	mov r25=ar.unat;		/* M */							\
+	mov r29=cr.ipsr;		/* M */							\
+	mov r26=ar.pfs;			/* I */							\
+	mov r28=cr.iip;			/* M */							\
+	mov r21=ar.fpsr;		/* M */							\
+	COVER;				/* B;; (or nothing) */					\
+	;;											\
+	adds r16=IA64_TASK_THREAD_ON_USTACK_OFFSET,r16;						\
+	;;											\
+	ld1 r17=[r16];				/* load current->thread.on_ustack flag */	\
+	st1 [r16]=r0;				/* clear current->thread.on_ustack flag */	\
+	adds r1=-IA64_TASK_THREAD_ON_USTACK_OFFSET,r16						\
+	/* switch from user to kernel RBS: */							\
+	;;											\
+	invala;				/* M */							\
+	SAVE_IFS;										\
+	cmp.eq pKStk,pUStk=r0,r17;		/* are we in kernel mode already? */		\
+	;;											\
+	MINSTATE_START_SAVE_MIN									\
+	adds r17=2*L1_CACHE_BYTES,r1;		/* really: biggest cache-line size */		\
+	adds r16=PT(CR_IPSR),r1;								\
+	;;											\
+	lfetch.fault.excl.nt1 [r17],L1_CACHE_BYTES;						\
+	st8 [r16]=r29;		/* save cr.ipsr */						\
+	;;											\
+	lfetch.fault.excl.nt1 [r17];								\
+	tbit.nz p15,p0=r29,IA64_PSR_I_BIT;							\
+	mov r29=b0										\
+	;;											\
+	adds r16=PT(R8),r1;	/* initialize first base pointer */				\
+	adds r17=PT(R9),r1;	/* initialize second base pointer */				\
+(pKStk)	mov r18=r0;		/* make sure r18 isn't NaT */					\
+	;;											\
+.mem.offset 0,0; st8.spill [r16]=r8,16;								\
+.mem.offset 8,0; st8.spill [r17]=r9,16;								\
+        ;;											\
+.mem.offset 0,0; st8.spill [r16]=r10,24;							\
+.mem.offset 8,0; st8.spill [r17]=r11,24;							\
+        ;;											\
+	st8 [r16]=r28,16;	/* save cr.iip */						\
+	st8 [r17]=r30,16;	/* save cr.ifs */						\
+(pUStk)	sub r18=r18,r22;	/* r18=RSE.ndirty*8 */						\
+	mov r8=ar.ccv;										\
+	mov r9=ar.csd;										\
+	mov r10=ar.ssd;										\
+	movl r11=FPSR_DEFAULT;   /* L-unit */							\
+	;;											\
+	st8 [r16]=r25,16;	/* save ar.unat */						\
+	st8 [r17]=r26,16;	/* save ar.pfs */						\
+	shl r18=r18,16;		/* compute ar.rsc to be used for "loadrs" */			\
+	;;											\
+	st8 [r16]=r27,16;	/* save ar.rsc */						\
+(pUStk)	st8 [r17]=r24,16;	/* save ar.rnat */						\
+(pKStk)	adds r17=16,r17;	/* skip over ar_rnat field */					\
+	;;			/* avoid RAW on r16 & r17 */					\
+(pUStk)	st8 [r16]=r23,16;	/* save ar.bspstore */						\
+	st8 [r17]=r31,16;	/* save predicates */						\
+(pKStk)	adds r16=16,r16;	/* skip over ar_bspstore field */				\
+	;;											\
+	st8 [r16]=r29,16;	/* save b0 */							\
+	st8 [r17]=r18,16;	/* save ar.rsc value for "loadrs" */				\
+	cmp.eq pNonSys,pSys=r0,r0	/* initialize pSys=0, pNonSys=1 */			\
+	;;											\
+.mem.offset 0,0; st8.spill [r16]=r20,16;	/* save original r1 */				\
+.mem.offset 8,0; st8.spill [r17]=r12,16;							\
+	adds r12=-16,r1;	/* switch to kernel memory stack (with 16 bytes of scratch) */	\
+	;;											\
+.mem.offset 0,0; st8.spill [r16]=r13,16;							\
+.mem.offset 8,0; st8.spill [r17]=r21,16;	/* save ar.fpsr */				\
+	mov r13=IA64_KR(CURRENT);	/* establish `current' */				\
+	;;											\
+.mem.offset 0,0; st8.spill [r16]=r15,16;							\
+.mem.offset 8,0; st8.spill [r17]=r14,16;							\
+	;;											\
+.mem.offset 0,0; st8.spill [r16]=r2,16;								\
+.mem.offset 8,0; st8.spill [r17]=r3,16;								\
+	adds r2=IA64_PT_REGS_R16_OFFSET,r1;							\
+	;;											\
+	EXTRA;											\
+	movl r1=__gp;		/* establish kernel global pointer */				\
+	;;											\
+	MINSTATE_END_SAVE_MIN
+#endif
+
+/*
+ * SAVE_REST saves the remainder of pt_regs (with psr.ic on).
+ *
+ * Assumed state upon entry:
+ *	psr.ic: on
+ *	r2:	points to &pt_regs.r16
+ *	r3:	points to &pt_regs.r17
+ *	r8:	contents of ar.ccv
+ *	r9:	contents of ar.csd
+ *	r10:	contents of ar.ssd
+ *	r11:	FPSR_DEFAULT
+ *
+ * Registers r14 and r15 are guaranteed not to be touched by SAVE_REST.
+ */
+#define SAVE_REST				\
+.mem.offset 0,0; st8.spill [r2]=r16,16;		\
+.mem.offset 8,0; st8.spill [r3]=r17,16;		\
+	;;					\
+.mem.offset 0,0; st8.spill [r2]=r18,16;		\
+.mem.offset 8,0; st8.spill [r3]=r19,16;		\
+	;;					\
+.mem.offset 0,0; st8.spill [r2]=r20,16;		\
+.mem.offset 8,0; st8.spill [r3]=r21,16;		\
+	mov r18=b6;				\
+	;;					\
+.mem.offset 0,0; st8.spill [r2]=r22,16;		\
+.mem.offset 8,0; st8.spill [r3]=r23,16;		\
+	mov r19=b7;				\
+	;;					\
+.mem.offset 0,0; st8.spill [r2]=r24,16;		\
+.mem.offset 8,0; st8.spill [r3]=r25,16;		\
+	;;					\
+.mem.offset 0,0; st8.spill [r2]=r26,16;		\
+.mem.offset 8,0; st8.spill [r3]=r27,16;		\
+	;;					\
+.mem.offset 0,0; st8.spill [r2]=r28,16;		\
+.mem.offset 8,0; st8.spill [r3]=r29,16;		\
+	;;					\
+.mem.offset 0,0; st8.spill [r2]=r30,16;		\
+.mem.offset 8,0; st8.spill [r3]=r31,32;		\
+	;;					\
+	mov ar.fpsr=r11;	/* M-unit */	\
+	st8 [r2]=r8,8;		/* ar.ccv */	\
+	adds r24=PT(B6)-PT(F7),r3;		\
+	;;					\
+	stf.spill [r2]=f6,32;			\
+	stf.spill [r3]=f7,32;			\
+	;;					\
+	stf.spill [r2]=f8,32;			\
+	stf.spill [r3]=f9,32;			\
+	;;					\
+	stf.spill [r2]=f10;			\
+	stf.spill [r3]=f11;			\
+	adds r25=PT(B7)-PT(F11),r3;		\
+	;;					\
+	st8 [r24]=r18,16;       /* b6 */	\
+	st8 [r25]=r19,16;       /* b7 */	\
+	;;					\
+	st8 [r24]=r9;        	/* ar.csd */	\
+	st8 [r25]=r10;      	/* ar.ssd */	\
+	;;
+
+#define SAVE_MIN_WITH_COVER	DO_SAVE_MIN(cover, mov r30=cr.ifs,)
+#define SAVE_MIN_WITH_COVER_R19	DO_SAVE_MIN(cover, mov r30=cr.ifs, mov r15=r19)
+#ifdef CONFIG_XEN
+#define SAVE_MIN		break 0;; /* FIXME: non-cover version only for ia32 support? */
+#else
+#define SAVE_MIN		DO_SAVE_MIN(     , mov r30=r0, )
+#endif
diff -rpuN linux-2.6.18.8/arch/ia64/xen/xenpal.S linux-2.6.18-xen-3.3.0/arch/ia64/xen/xenpal.S
--- linux-2.6.18.8/arch/ia64/xen/xenpal.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/xen/xenpal.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,85 @@
+/*
+ * ia64/xen/xenpal.S
+ *
+ * Alternate PAL  routines for Xen.  Heavily leveraged from
+ *   ia64/kernel/pal.S
+ *
+ * Copyright (C) 2005 Hewlett-Packard Co
+ *	Dan Magenheimer <dan.magenheimer@.hp.com>
+ */
+
+#include <asm/asmmacro.h>
+#include <asm/processor.h>
+
+GLOBAL_ENTRY(xen_pal_call_static)
+	.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(5)
+	alloc loc1 = ar.pfs,5,5,0,0
+#ifdef CONFIG_XEN
+	movl r22=running_on_xen;;
+	ld4 r22=[r22];;
+	cmp.eq p7,p0=r22,r0
+(p7)	br.cond.spnt.many __ia64_pal_call_static;;
+#endif
+	movl loc2 = pal_entry_point
+1:	{
+	  mov r28 = in0
+	  mov r29 = in1
+	  mov r8 = ip
+	}
+	;;
+	ld8 loc2 = [loc2]		// loc2 <- entry point
+	tbit.nz p6,p7 = in4, 0
+	adds r8 = 1f-1b,r8
+	mov loc4=ar.rsc			// save RSE configuration
+	;;
+	mov ar.rsc=0			// put RSE in enforced lazy, LE mode
+#ifdef CONFIG_XEN
+	mov r9 = r8
+	XEN_HYPER_GET_PSR
+	;;
+	mov loc3 = r8
+	mov r8 = r9
+	;;
+#else    
+	mov loc3 = psr
+#endif    
+	mov loc0 = rp
+	.body
+	mov r30 = in2
+
+#ifdef CONFIG_XEN
+	// this is low priority for paravirtualization, but is called
+	// from the idle loop so confuses privop counting
+	movl r31=XSI_PSR_I_ADDR
+	;;
+	ld8 r31=[r31]
+	mov r22=1
+	;;
+	st1 [r31]=r22
+	;;  
+(p6)	movl r31=XSI_PSR_IC
+	;;
+(p6)	st4.rel [r31]=r0
+	;;
+	mov r31 = in3
+	mov b7 = loc2
+	;;
+#else
+(p6)	rsm psr.i | psr.ic
+	mov r31 = in3
+	mov b7 = loc2
+
+(p7)	rsm psr.i
+	;;
+(p6)	srlz.i
+#endif
+	mov rp = r8
+	br.cond.sptk.many b7
+1:	mov psr.l = loc3
+	mov ar.rsc = loc4		// restore RSE configuration
+	mov ar.pfs = loc1
+	mov rp = loc0
+	;;
+	srlz.d				// seralize restoration of psr.l
+	br.ret.sptk.many b0
+END(xen_pal_call_static)
diff -rpuN linux-2.6.18.8/arch/ia64/xen/xensetup.S linux-2.6.18-xen-3.3.0/arch/ia64/xen/xensetup.S
--- linux-2.6.18.8/arch/ia64/xen/xensetup.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/ia64/xen/xensetup.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,40 @@
+/*
+ * Support routines for Xen
+ *
+ * Copyright (C) 2005 Dan Magenheimer <dan.magenheimer@hp.com>
+ */
+
+#include <asm/processor.h>
+#include <asm/asmmacro.h>
+
+	.section .data.read_mostly
+	.align 8
+	.global running_on_xen
+running_on_xen:
+	data4 0
+	.previous
+
+#define isBP	p3	// are we the Bootstrap Processor?
+
+	.text
+GLOBAL_ENTRY(early_xen_setup)
+	mov r8=ar.rsc		// Initialized in head.S
+(isBP)	movl r9=running_on_xen;;
+	extr.u r8=r8,2,2;;	// Extract pl fields
+	cmp.eq p7,p0=r8,r0	// p7: !running on xen
+	mov r8=1		// booleanize.
+(p7)	br.ret.sptk.many rp;;
+(isBP)	st4 [r9]=r8
+	movl r10=xen_ivt;;
+	
+	mov cr.iva=r10
+
+	/* Set xsi base.  */
+#define FW_HYPERCALL_SET_SHARED_INFO_VA			0x600
+(isBP)	mov r2=FW_HYPERCALL_SET_SHARED_INFO_VA
+(isBP)	movl r28=XSI_BASE;;
+(isBP)	break 0x1000;;
+
+	br.ret.sptk.many rp
+	;;
+END(early_xen_setup)
diff -rpuN linux-2.6.18.8/arch/powerpc/boot/Makefile linux-2.6.18-xen-3.3.0/arch/powerpc/boot/Makefile
--- linux-2.6.18.8/arch/powerpc/boot/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/boot/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -36,8 +36,11 @@ zliblinuxheader := zlib.h zconf.h zutil.
 $(addprefix $(obj)/,$(zlib) main.o): $(addprefix $(obj)/,$(zliblinuxheader)) $(addprefix $(obj)/,$(zlibheader))
 #$(addprefix $(obj)/,main.o): $(addprefix $(obj)/,zlib.h)
 
+xen_guest-y = xen_guest.S
+
 src-boot := crt0.S string.S prom.c stdio.c main.c div64.S
 src-boot += $(zlib)
+src-boot += $(xen_guest-$(CONFIG_XEN))
 src-boot := $(addprefix $(obj)/, $(src-boot))
 obj-boot := $(addsuffix .o, $(basename $(src-boot)))
 
diff -rpuN linux-2.6.18.8/arch/powerpc/boot/xen_guest.S linux-2.6.18-xen-3.3.0/arch/powerpc/boot/xen_guest.S
--- linux-2.6.18.8/arch/powerpc/boot/xen_guest.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/boot/xen_guest.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,7 @@
+	.section __xen_guest
+	.ascii	"GUEST_OS=linux"
+	.ascii  ",GUEST_VER=xen-3.0"
+	.ascii	",XEN_VER=xen-3.0"
+	.ascii	",VIRT_BASE=0x0"
+	.ascii	",LOADER=generic"
+	.byte	0
diff -rpuN linux-2.6.18.8/arch/powerpc/configs/xen_maple_defconfig linux-2.6.18-xen-3.3.0/arch/powerpc/configs/xen_maple_defconfig
--- linux-2.6.18.8/arch/powerpc/configs/xen_maple_defconfig	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/configs/xen_maple_defconfig	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1342 @@
+#
+# Automatically generated make config: don't edit
+# Linux kernel version: 2.6.17
+# Mon Jan 15 23:48:47 2007
+#
+CONFIG_PPC64=y
+CONFIG_64BIT=y
+CONFIG_PPC_MERGE=y
+CONFIG_MMU=y
+CONFIG_GENERIC_HARDIRQS=y
+CONFIG_IRQ_PER_CPU=y
+CONFIG_RWSEM_XCHGADD_ALGORITHM=y
+CONFIG_GENERIC_HWEIGHT=y
+CONFIG_GENERIC_CALIBRATE_DELAY=y
+CONFIG_GENERIC_FIND_NEXT_BIT=y
+CONFIG_PPC=y
+CONFIG_EARLY_PRINTK=y
+CONFIG_COMPAT=y
+CONFIG_SYSVIPC_COMPAT=y
+CONFIG_SCHED_NO_NO_OMIT_FRAME_POINTER=y
+CONFIG_ARCH_MAY_HAVE_PC_FDC=y
+CONFIG_PPC_OF=y
+CONFIG_PPC_UDBG_16550=y
+CONFIG_GENERIC_TBSYNC=y
+# CONFIG_DEFAULT_UIMAGE is not set
+
+#
+# Processor support
+#
+CONFIG_POWER4_ONLY=y
+CONFIG_POWER4=y
+CONFIG_PPC_FPU=y
+CONFIG_ALTIVEC=y
+CONFIG_PPC_STD_MMU=y
+CONFIG_VIRT_CPU_ACCOUNTING=y
+CONFIG_SMP=y
+CONFIG_NR_CPUS=32
+CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
+
+#
+# Code maturity level options
+#
+CONFIG_EXPERIMENTAL=y
+CONFIG_LOCK_KERNEL=y
+CONFIG_INIT_ENV_ARG_LIMIT=32
+
+#
+# General setup
+#
+CONFIG_LOCALVERSION="-Xen"
+CONFIG_LOCALVERSION_AUTO=y
+CONFIG_SWAP=y
+CONFIG_SYSVIPC=y
+# CONFIG_POSIX_MQUEUE is not set
+# CONFIG_BSD_PROCESS_ACCT is not set
+CONFIG_SYSCTL=y
+# CONFIG_AUDIT is not set
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+# CONFIG_CPUSETS is not set
+# CONFIG_RELAY is not set
+CONFIG_INITRAMFS_SOURCE=""
+# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set
+# CONFIG_EMBEDDED is not set
+CONFIG_KALLSYMS=y
+# CONFIG_KALLSYMS_ALL is not set
+# CONFIG_KALLSYMS_EXTRA_PASS is not set
+CONFIG_HOTPLUG=y
+CONFIG_PRINTK=y
+CONFIG_BUG=y
+CONFIG_ELF_CORE=y
+CONFIG_BASE_FULL=y
+CONFIG_RT_MUTEXES=y
+CONFIG_FUTEX=y
+CONFIG_EPOLL=y
+CONFIG_SHMEM=y
+CONFIG_SLAB=y
+# CONFIG_TINY_SHMEM is not set
+CONFIG_BASE_SMALL=0
+# CONFIG_SLOB is not set
+
+#
+# Loadable module support
+#
+CONFIG_MODULES=y
+CONFIG_MODULE_UNLOAD=y
+# CONFIG_MODULE_FORCE_UNLOAD is not set
+CONFIG_MODVERSIONS=y
+CONFIG_MODULE_SRCVERSION_ALL=y
+CONFIG_KMOD=y
+CONFIG_STOP_MACHINE=y
+
+#
+# Block layer
+#
+# CONFIG_BLK_DEV_IO_TRACE is not set
+
+#
+# IO Schedulers
+#
+CONFIG_IOSCHED_NOOP=y
+# CONFIG_IOSCHED_AS is not set
+CONFIG_IOSCHED_DEADLINE=y
+# CONFIG_IOSCHED_CFQ is not set
+# CONFIG_DEFAULT_AS is not set
+CONFIG_DEFAULT_DEADLINE=y
+# CONFIG_DEFAULT_CFQ is not set
+# CONFIG_DEFAULT_NOOP is not set
+CONFIG_DEFAULT_IOSCHED="deadline"
+
+#
+# Platform support
+#
+CONFIG_PPC_MULTIPLATFORM=y
+# CONFIG_PPC_ISERIES is not set
+# CONFIG_EMBEDDED6xx is not set
+# CONFIG_APUS is not set
+CONFIG_PPC_PSERIES=y
+# CONFIG_PPC_PMAC is not set
+CONFIG_PPC_MAPLE=y
+# CONFIG_PPC_CELL is not set
+# CONFIG_PPC_CELL_NATIVE is not set
+# CONFIG_PPC_IBM_CELL_BLADE is not set
+# CONFIG_UDBG_RTAS_CONSOLE is not set
+CONFIG_PPC_XEN=y
+CONFIG_XICS=y
+CONFIG_U3_DART=y
+CONFIG_MPIC=y
+CONFIG_PPC_RTAS=y
+CONFIG_RTAS_ERROR_LOGGING=y
+CONFIG_RTAS_PROC=y
+CONFIG_RTAS_FLASH=y
+# CONFIG_MMIO_NVRAM is not set
+CONFIG_MPIC_BROKEN_U3=y
+CONFIG_IBMVIO=y
+# CONFIG_IBMEBUS is not set
+# CONFIG_PPC_MPC106 is not set
+CONFIG_PPC_970_NAP=y
+# CONFIG_CPU_FREQ is not set
+# CONFIG_WANT_EARLY_SERIAL is not set
+
+#
+# Kernel options
+#
+CONFIG_HZ_100=y
+# CONFIG_HZ_250 is not set
+# CONFIG_HZ_1000 is not set
+CONFIG_HZ=100
+CONFIG_PREEMPT_NONE=y
+# CONFIG_PREEMPT_VOLUNTARY is not set
+# CONFIG_PREEMPT is not set
+# CONFIG_PREEMPT_BKL is not set
+CONFIG_BINFMT_ELF=y
+# CONFIG_BINFMT_MISC is not set
+CONFIG_FORCE_MAX_ZONEORDER=13
+CONFIG_IOMMU_VMERGE=y
+# CONFIG_HOTPLUG_CPU is not set
+CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y
+# CONFIG_KEXEC is not set
+# CONFIG_CRASH_DUMP is not set
+CONFIG_IRQ_ALL_CPUS=y
+# CONFIG_PPC_SPLPAR is not set
+CONFIG_EEH=y
+CONFIG_SCANLOG=y
+CONFIG_LPARCFG=y
+CONFIG_NUMA=y
+CONFIG_NODES_SHIFT=4
+CONFIG_ARCH_SELECT_MEMORY_MODEL=y
+CONFIG_ARCH_SPARSEMEM_ENABLE=y
+CONFIG_ARCH_SPARSEMEM_DEFAULT=y
+CONFIG_SELECT_MEMORY_MODEL=y
+# CONFIG_FLATMEM_MANUAL is not set
+# CONFIG_DISCONTIGMEM_MANUAL is not set
+CONFIG_SPARSEMEM_MANUAL=y
+CONFIG_SPARSEMEM=y
+CONFIG_NEED_MULTIPLE_NODES=y
+CONFIG_HAVE_MEMORY_PRESENT=y
+# CONFIG_SPARSEMEM_STATIC is not set
+CONFIG_SPARSEMEM_EXTREME=y
+CONFIG_MEMORY_HOTPLUG=y
+CONFIG_SPLIT_PTLOCK_CPUS=4
+CONFIG_MIGRATION=y
+CONFIG_RESOURCES_64BIT=y
+CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID=y
+CONFIG_ARCH_MEMORY_PROBE=y
+# CONFIG_PPC_64K_PAGES is not set
+# CONFIG_SCHED_SMT is not set
+CONFIG_PROC_DEVICETREE=y
+# CONFIG_CMDLINE_BOOL is not set
+# CONFIG_PM is not set
+CONFIG_SECCOMP=y
+CONFIG_ISA_DMA_API=y
+
+#
+# Bus options
+#
+CONFIG_GENERIC_ISA_DMA=y
+CONFIG_PPC_I8259=y
+# CONFIG_PPC_INDIRECT_PCI is not set
+CONFIG_PCI=y
+CONFIG_PCI_DOMAINS=y
+# CONFIG_PCIEPORTBUS is not set
+# CONFIG_PCI_DEBUG is not set
+
+#
+# PCCARD (PCMCIA/CardBus) support
+#
+# CONFIG_PCCARD is not set
+
+#
+# PCI Hotplug Support
+#
+# CONFIG_HOTPLUG_PCI is not set
+CONFIG_KERNEL_START=0xc000000000000000
+
+#
+# Networking
+#
+CONFIG_NET=y
+
+#
+# Networking options
+#
+# CONFIG_NETDEBUG is not set
+CONFIG_PACKET=y
+CONFIG_PACKET_MMAP=y
+CONFIG_UNIX=y
+CONFIG_XFRM=y
+# CONFIG_XFRM_USER is not set
+# CONFIG_NET_KEY is not set
+CONFIG_INET=y
+CONFIG_IP_MULTICAST=y
+CONFIG_IP_ADVANCED_ROUTER=y
+CONFIG_ASK_IP_FIB_HASH=y
+# CONFIG_IP_FIB_TRIE is not set
+CONFIG_IP_FIB_HASH=y
+CONFIG_IP_MULTIPLE_TABLES=y
+CONFIG_IP_ROUTE_FWMARK=y
+CONFIG_IP_ROUTE_MULTIPATH=y
+# CONFIG_IP_ROUTE_MULTIPATH_CACHED is not set
+# CONFIG_IP_ROUTE_VERBOSE is not set
+CONFIG_IP_PNP=y
+CONFIG_IP_PNP_DHCP=y
+CONFIG_IP_PNP_BOOTP=y
+CONFIG_IP_PNP_RARP=y
+CONFIG_NET_IPIP=y
+# CONFIG_NET_IPGRE is not set
+CONFIG_IP_MROUTE=y
+CONFIG_IP_PIMSM_V1=y
+CONFIG_IP_PIMSM_V2=y
+# CONFIG_ARPD is not set
+CONFIG_SYN_COOKIES=y
+# CONFIG_INET_AH is not set
+# CONFIG_INET_ESP is not set
+# CONFIG_INET_IPCOMP is not set
+# CONFIG_INET_XFRM_TUNNEL is not set
+CONFIG_INET_TUNNEL=y
+CONFIG_INET_XFRM_MODE_TRANSPORT=y
+CONFIG_INET_XFRM_MODE_TUNNEL=y
+CONFIG_INET_DIAG=y
+CONFIG_INET_TCP_DIAG=y
+# CONFIG_TCP_CONG_ADVANCED is not set
+CONFIG_TCP_CONG_BIC=y
+
+#
+# IP: Virtual Server Configuration
+#
+# CONFIG_IP_VS is not set
+CONFIG_IPV6=y
+CONFIG_IPV6_PRIVACY=y
+CONFIG_IPV6_ROUTER_PREF=y
+CONFIG_IPV6_ROUTE_INFO=y
+CONFIG_INET6_AH=y
+CONFIG_INET6_ESP=y
+CONFIG_INET6_IPCOMP=y
+CONFIG_INET6_XFRM_TUNNEL=y
+CONFIG_INET6_TUNNEL=y
+CONFIG_INET6_XFRM_MODE_TRANSPORT=y
+CONFIG_INET6_XFRM_MODE_TUNNEL=y
+CONFIG_IPV6_TUNNEL=y
+# CONFIG_NETWORK_SECMARK is not set
+CONFIG_NETFILTER=y
+# CONFIG_NETFILTER_DEBUG is not set
+CONFIG_BRIDGE_NETFILTER=y
+
+#
+# Core Netfilter Configuration
+#
+CONFIG_NETFILTER_NETLINK=y
+CONFIG_NETFILTER_NETLINK_QUEUE=y
+CONFIG_NETFILTER_NETLINK_LOG=y
+CONFIG_NETFILTER_XTABLES=y
+CONFIG_NETFILTER_XT_TARGET_CLASSIFY=y
+CONFIG_NETFILTER_XT_TARGET_CONNMARK=y
+CONFIG_NETFILTER_XT_TARGET_MARK=y
+CONFIG_NETFILTER_XT_TARGET_NFQUEUE=y
+CONFIG_NETFILTER_XT_TARGET_NOTRACK=y
+CONFIG_NETFILTER_XT_MATCH_COMMENT=y
+CONFIG_NETFILTER_XT_MATCH_CONNBYTES=y
+CONFIG_NETFILTER_XT_MATCH_CONNMARK=y
+CONFIG_NETFILTER_XT_MATCH_CONNTRACK=y
+CONFIG_NETFILTER_XT_MATCH_DCCP=y
+CONFIG_NETFILTER_XT_MATCH_ESP=y
+CONFIG_NETFILTER_XT_MATCH_HELPER=y
+CONFIG_NETFILTER_XT_MATCH_LENGTH=y
+CONFIG_NETFILTER_XT_MATCH_LIMIT=y
+CONFIG_NETFILTER_XT_MATCH_MAC=y
+CONFIG_NETFILTER_XT_MATCH_MARK=y
+CONFIG_NETFILTER_XT_MATCH_POLICY=y
+CONFIG_NETFILTER_XT_MATCH_MULTIPORT=y
+CONFIG_NETFILTER_XT_MATCH_PHYSDEV=y
+CONFIG_NETFILTER_XT_MATCH_PKTTYPE=y
+CONFIG_NETFILTER_XT_MATCH_QUOTA=y
+CONFIG_NETFILTER_XT_MATCH_REALM=y
+CONFIG_NETFILTER_XT_MATCH_SCTP=y
+CONFIG_NETFILTER_XT_MATCH_STATE=y
+CONFIG_NETFILTER_XT_MATCH_STATISTIC=y
+CONFIG_NETFILTER_XT_MATCH_STRING=y
+CONFIG_NETFILTER_XT_MATCH_TCPMSS=y
+
+#
+# IP: Netfilter Configuration
+#
+CONFIG_IP_NF_CONNTRACK=y
+CONFIG_IP_NF_CT_ACCT=y
+CONFIG_IP_NF_CONNTRACK_MARK=y
+CONFIG_IP_NF_CONNTRACK_EVENTS=y
+CONFIG_IP_NF_CONNTRACK_NETLINK=y
+CONFIG_IP_NF_CT_PROTO_SCTP=y
+CONFIG_IP_NF_FTP=y
+CONFIG_IP_NF_IRC=y
+# CONFIG_IP_NF_NETBIOS_NS is not set
+CONFIG_IP_NF_TFTP=y
+CONFIG_IP_NF_AMANDA=y
+CONFIG_IP_NF_PPTP=y
+# CONFIG_IP_NF_H323 is not set
+# CONFIG_IP_NF_SIP is not set
+# CONFIG_IP_NF_QUEUE is not set
+CONFIG_IP_NF_IPTABLES=y
+CONFIG_IP_NF_MATCH_IPRANGE=y
+CONFIG_IP_NF_MATCH_TOS=y
+CONFIG_IP_NF_MATCH_RECENT=y
+CONFIG_IP_NF_MATCH_ECN=y
+CONFIG_IP_NF_MATCH_DSCP=y
+CONFIG_IP_NF_MATCH_AH=y
+CONFIG_IP_NF_MATCH_TTL=y
+CONFIG_IP_NF_MATCH_OWNER=y
+CONFIG_IP_NF_MATCH_ADDRTYPE=y
+CONFIG_IP_NF_MATCH_HASHLIMIT=y
+CONFIG_IP_NF_FILTER=y
+CONFIG_IP_NF_TARGET_REJECT=y
+CONFIG_IP_NF_TARGET_LOG=y
+CONFIG_IP_NF_TARGET_ULOG=y
+CONFIG_IP_NF_TARGET_TCPMSS=y
+CONFIG_IP_NF_NAT=y
+CONFIG_IP_NF_NAT_NEEDED=y
+CONFIG_IP_NF_TARGET_MASQUERADE=y
+CONFIG_IP_NF_TARGET_REDIRECT=y
+CONFIG_IP_NF_TARGET_NETMAP=y
+CONFIG_IP_NF_TARGET_SAME=y
+CONFIG_IP_NF_NAT_SNMP_BASIC=y
+CONFIG_IP_NF_NAT_IRC=y
+CONFIG_IP_NF_NAT_FTP=y
+CONFIG_IP_NF_NAT_TFTP=y
+CONFIG_IP_NF_NAT_AMANDA=y
+CONFIG_IP_NF_NAT_PPTP=y
+CONFIG_IP_NF_MANGLE=y
+CONFIG_IP_NF_TARGET_TOS=y
+CONFIG_IP_NF_TARGET_ECN=y
+CONFIG_IP_NF_TARGET_DSCP=y
+CONFIG_IP_NF_TARGET_TTL=y
+CONFIG_IP_NF_TARGET_CLUSTERIP=y
+CONFIG_IP_NF_RAW=y
+CONFIG_IP_NF_ARPTABLES=y
+CONFIG_IP_NF_ARPFILTER=y
+CONFIG_IP_NF_ARP_MANGLE=y
+
+#
+# IPv6: Netfilter Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP6_NF_QUEUE is not set
+CONFIG_IP6_NF_IPTABLES=y
+CONFIG_IP6_NF_MATCH_RT=y
+CONFIG_IP6_NF_MATCH_OPTS=y
+CONFIG_IP6_NF_MATCH_FRAG=y
+CONFIG_IP6_NF_MATCH_HL=y
+CONFIG_IP6_NF_MATCH_OWNER=y
+CONFIG_IP6_NF_MATCH_IPV6HEADER=y
+CONFIG_IP6_NF_MATCH_AH=y
+CONFIG_IP6_NF_MATCH_EUI64=y
+CONFIG_IP6_NF_FILTER=y
+CONFIG_IP6_NF_TARGET_LOG=y
+CONFIG_IP6_NF_TARGET_REJECT=y
+CONFIG_IP6_NF_MANGLE=y
+CONFIG_IP6_NF_TARGET_HL=y
+CONFIG_IP6_NF_RAW=y
+
+#
+# Bridge: Netfilter Configuration
+#
+CONFIG_BRIDGE_NF_EBTABLES=y
+CONFIG_BRIDGE_EBT_BROUTE=y
+CONFIG_BRIDGE_EBT_T_FILTER=y
+CONFIG_BRIDGE_EBT_T_NAT=y
+CONFIG_BRIDGE_EBT_802_3=y
+CONFIG_BRIDGE_EBT_AMONG=y
+CONFIG_BRIDGE_EBT_ARP=y
+CONFIG_BRIDGE_EBT_IP=y
+CONFIG_BRIDGE_EBT_LIMIT=y
+CONFIG_BRIDGE_EBT_MARK=y
+CONFIG_BRIDGE_EBT_PKTTYPE=y
+CONFIG_BRIDGE_EBT_STP=y
+CONFIG_BRIDGE_EBT_VLAN=y
+CONFIG_BRIDGE_EBT_ARPREPLY=y
+CONFIG_BRIDGE_EBT_DNAT=y
+CONFIG_BRIDGE_EBT_MARK_T=y
+CONFIG_BRIDGE_EBT_REDIRECT=y
+CONFIG_BRIDGE_EBT_SNAT=y
+CONFIG_BRIDGE_EBT_LOG=y
+CONFIG_BRIDGE_EBT_ULOG=y
+
+#
+# DCCP Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP_DCCP is not set
+
+#
+# SCTP Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP_SCTP is not set
+
+#
+# TIPC Configuration (EXPERIMENTAL)
+#
+# CONFIG_TIPC is not set
+# CONFIG_ATM is not set
+CONFIG_BRIDGE=y
+CONFIG_VLAN_8021Q=y
+# CONFIG_DECNET is not set
+CONFIG_LLC=y
+# CONFIG_LLC2 is not set
+# CONFIG_IPX is not set
+# CONFIG_ATALK is not set
+# CONFIG_X25 is not set
+# CONFIG_LAPB is not set
+# CONFIG_NET_DIVERT is not set
+# CONFIG_ECONET is not set
+# CONFIG_WAN_ROUTER is not set
+
+#
+# QoS and/or fair queueing
+#
+# CONFIG_NET_SCHED is not set
+CONFIG_NET_CLS_ROUTE=y
+
+#
+# Network testing
+#
+# CONFIG_NET_PKTGEN is not set
+# CONFIG_HAMRADIO is not set
+# CONFIG_IRDA is not set
+# CONFIG_BT is not set
+# CONFIG_IEEE80211 is not set
+
+#
+# Device Drivers
+#
+
+#
+# Generic Driver Options
+#
+CONFIG_STANDALONE=y
+CONFIG_PREVENT_FIRMWARE_BUILD=y
+CONFIG_FW_LOADER=y
+# CONFIG_DEBUG_DRIVER is not set
+# CONFIG_SYS_HYPERVISOR is not set
+
+#
+# Connector - unified userspace <-> kernelspace linker
+#
+# CONFIG_CONNECTOR is not set
+
+#
+# Memory Technology Devices (MTD)
+#
+# CONFIG_MTD is not set
+
+#
+# Parallel port support
+#
+# CONFIG_PARPORT is not set
+
+#
+# Plug and Play support
+#
+
+#
+# Block devices
+#
+# CONFIG_BLK_DEV_FD is not set
+# CONFIG_BLK_CPQ_DA is not set
+# CONFIG_BLK_CPQ_CISS_DA is not set
+# CONFIG_BLK_DEV_DAC960 is not set
+# CONFIG_BLK_DEV_UMEM is not set
+# CONFIG_BLK_DEV_COW_COMMON is not set
+CONFIG_BLK_DEV_LOOP=y
+# CONFIG_BLK_DEV_CRYPTOLOOP is not set
+# CONFIG_BLK_DEV_NBD is not set
+# CONFIG_BLK_DEV_SX8 is not set
+CONFIG_BLK_DEV_RAM=y
+CONFIG_BLK_DEV_RAM_COUNT=16
+CONFIG_BLK_DEV_RAM_SIZE=10240
+CONFIG_BLK_DEV_INITRD=y
+# CONFIG_CDROM_PKTCDVD is not set
+# CONFIG_ATA_OVER_ETH is not set
+
+#
+# ATA/ATAPI/MFM/RLL support
+#
+CONFIG_IDE=y
+CONFIG_BLK_DEV_IDE=y
+
+#
+# Please see Documentation/ide.txt for help/info on IDE drives
+#
+# CONFIG_BLK_DEV_IDE_SATA is not set
+CONFIG_BLK_DEV_IDEDISK=y
+# CONFIG_IDEDISK_MULTI_MODE is not set
+CONFIG_BLK_DEV_IDECD=y
+# CONFIG_BLK_DEV_IDETAPE is not set
+# CONFIG_BLK_DEV_IDEFLOPPY is not set
+# CONFIG_BLK_DEV_IDESCSI is not set
+CONFIG_IDE_TASK_IOCTL=y
+
+#
+# IDE chipset support/bugfixes
+#
+CONFIG_IDE_GENERIC=y
+CONFIG_BLK_DEV_IDEPCI=y
+CONFIG_IDEPCI_SHARE_IRQ=y
+# CONFIG_BLK_DEV_OFFBOARD is not set
+CONFIG_BLK_DEV_GENERIC=y
+# CONFIG_BLK_DEV_OPTI621 is not set
+# CONFIG_BLK_DEV_SL82C105 is not set
+CONFIG_BLK_DEV_IDEDMA_PCI=y
+# CONFIG_BLK_DEV_IDEDMA_FORCED is not set
+CONFIG_IDEDMA_PCI_AUTO=y
+# CONFIG_IDEDMA_ONLYDISK is not set
+# CONFIG_BLK_DEV_AEC62XX is not set
+# CONFIG_BLK_DEV_ALI15X3 is not set
+CONFIG_BLK_DEV_AMD74XX=y
+# CONFIG_BLK_DEV_CMD64X is not set
+# CONFIG_BLK_DEV_TRIFLEX is not set
+# CONFIG_BLK_DEV_CY82C693 is not set
+# CONFIG_BLK_DEV_CS5520 is not set
+# CONFIG_BLK_DEV_CS5530 is not set
+# CONFIG_BLK_DEV_HPT34X is not set
+# CONFIG_BLK_DEV_HPT366 is not set
+# CONFIG_BLK_DEV_SC1200 is not set
+# CONFIG_BLK_DEV_PIIX is not set
+# CONFIG_BLK_DEV_IT821X is not set
+# CONFIG_BLK_DEV_NS87415 is not set
+# CONFIG_BLK_DEV_PDC202XX_OLD is not set
+# CONFIG_BLK_DEV_PDC202XX_NEW is not set
+# CONFIG_BLK_DEV_SVWKS is not set
+# CONFIG_BLK_DEV_SIIMAGE is not set
+# CONFIG_BLK_DEV_SLC90E66 is not set
+# CONFIG_BLK_DEV_TRM290 is not set
+# CONFIG_BLK_DEV_VIA82CXXX is not set
+# CONFIG_IDE_ARM is not set
+CONFIG_BLK_DEV_IDEDMA=y
+# CONFIG_IDEDMA_IVB is not set
+CONFIG_IDEDMA_AUTO=y
+# CONFIG_BLK_DEV_HD is not set
+
+#
+# SCSI device support
+#
+# CONFIG_RAID_ATTRS is not set
+CONFIG_SCSI=y
+# CONFIG_SCSI_PROC_FS is not set
+
+#
+# SCSI support type (disk, tape, CD-ROM)
+#
+CONFIG_BLK_DEV_SD=y
+# CONFIG_CHR_DEV_ST is not set
+# CONFIG_CHR_DEV_OSST is not set
+CONFIG_BLK_DEV_SR=y
+# CONFIG_BLK_DEV_SR_VENDOR is not set
+CONFIG_CHR_DEV_SG=y
+# CONFIG_CHR_DEV_SCH is not set
+
+#
+# Some SCSI devices (e.g. CD jukebox) support multiple LUNs
+#
+# CONFIG_SCSI_MULTI_LUN is not set
+# CONFIG_SCSI_CONSTANTS is not set
+# CONFIG_SCSI_LOGGING is not set
+
+#
+# SCSI Transport Attributes
+#
+CONFIG_SCSI_SPI_ATTRS=y
+CONFIG_SCSI_FC_ATTRS=y
+CONFIG_SCSI_ISCSI_ATTRS=y
+CONFIG_SCSI_SAS_ATTRS=y
+
+#
+# SCSI low-level drivers
+#
+# CONFIG_ISCSI_TCP is not set
+# CONFIG_BLK_DEV_3W_XXXX_RAID is not set
+# CONFIG_SCSI_3W_9XXX is not set
+# CONFIG_SCSI_ACARD is not set
+# CONFIG_SCSI_AACRAID is not set
+# CONFIG_SCSI_AIC7XXX is not set
+# CONFIG_SCSI_AIC7XXX_OLD is not set
+# CONFIG_SCSI_AIC79XX is not set
+# CONFIG_MEGARAID_NEWGEN is not set
+# CONFIG_MEGARAID_LEGACY is not set
+# CONFIG_MEGARAID_SAS is not set
+# CONFIG_SCSI_SATA is not set
+# CONFIG_SCSI_HPTIOP is not set
+# CONFIG_SCSI_BUSLOGIC is not set
+# CONFIG_SCSI_DMX3191D is not set
+# CONFIG_SCSI_EATA is not set
+# CONFIG_SCSI_FUTURE_DOMAIN is not set
+# CONFIG_SCSI_GDTH is not set
+# CONFIG_SCSI_IPS is not set
+# CONFIG_SCSI_IBMVSCSI is not set
+# CONFIG_SCSI_INITIO is not set
+# CONFIG_SCSI_INIA100 is not set
+# CONFIG_SCSI_SYM53C8XX_2 is not set
+CONFIG_SCSI_IPR=y
+# CONFIG_SCSI_IPR_TRACE is not set
+# CONFIG_SCSI_IPR_DUMP is not set
+# CONFIG_SCSI_QLOGIC_1280 is not set
+# CONFIG_SCSI_QLA_FC is not set
+# CONFIG_SCSI_LPFC is not set
+# CONFIG_SCSI_DC395x is not set
+# CONFIG_SCSI_DC390T is not set
+# CONFIG_SCSI_DEBUG is not set
+
+#
+# Multi-device support (RAID and LVM)
+#
+# CONFIG_MD is not set
+
+#
+# Fusion MPT device support
+#
+# CONFIG_FUSION is not set
+# CONFIG_FUSION_SPI is not set
+# CONFIG_FUSION_FC is not set
+# CONFIG_FUSION_SAS is not set
+
+#
+# IEEE 1394 (FireWire) support
+#
+# CONFIG_IEEE1394 is not set
+
+#
+# I2O device support
+#
+# CONFIG_I2O is not set
+
+#
+# Macintosh device drivers
+#
+# CONFIG_WINDFARM is not set
+
+#
+# Network device support
+#
+CONFIG_NETDEVICES=y
+# CONFIG_DUMMY is not set
+# CONFIG_BONDING is not set
+# CONFIG_EQUALIZER is not set
+# CONFIG_TUN is not set
+
+#
+# ARCnet devices
+#
+# CONFIG_ARCNET is not set
+
+#
+# PHY device support
+#
+# CONFIG_PHYLIB is not set
+
+#
+# Ethernet (10 or 100Mbit)
+#
+CONFIG_NET_ETHERNET=y
+CONFIG_MII=y
+# CONFIG_HAPPYMEAL is not set
+# CONFIG_SUNGEM is not set
+# CONFIG_CASSINI is not set
+# CONFIG_NET_VENDOR_3COM is not set
+
+#
+# Tulip family network device support
+#
+# CONFIG_NET_TULIP is not set
+# CONFIG_HP100 is not set
+CONFIG_IBMVETH=y
+CONFIG_NET_PCI=y
+# CONFIG_PCNET32 is not set
+CONFIG_AMD8111_ETH=y
+# CONFIG_AMD8111E_NAPI is not set
+# CONFIG_ADAPTEC_STARFIRE is not set
+# CONFIG_B44 is not set
+# CONFIG_FORCEDETH is not set
+# CONFIG_DGRS is not set
+# CONFIG_EEPRO100 is not set
+# CONFIG_E100 is not set
+# CONFIG_FEALNX is not set
+# CONFIG_NATSEMI is not set
+# CONFIG_NE2K_PCI is not set
+# CONFIG_8139CP is not set
+# CONFIG_8139TOO is not set
+# CONFIG_SIS900 is not set
+# CONFIG_EPIC100 is not set
+# CONFIG_SUNDANCE is not set
+# CONFIG_VIA_RHINE is not set
+
+#
+# Ethernet (1000 Mbit)
+#
+# CONFIG_ACENIC is not set
+# CONFIG_DL2K is not set
+CONFIG_E1000=y
+# CONFIG_E1000_NAPI is not set
+# CONFIG_E1000_DISABLE_PACKET_SPLIT is not set
+# CONFIG_NS83820 is not set
+# CONFIG_HAMACHI is not set
+# CONFIG_YELLOWFIN is not set
+# CONFIG_R8169 is not set
+# CONFIG_SIS190 is not set
+# CONFIG_SKGE is not set
+CONFIG_SKY2=y
+CONFIG_SK98LIN=y
+# CONFIG_VIA_VELOCITY is not set
+CONFIG_TIGON3=y
+CONFIG_BNX2=y
+# CONFIG_MV643XX_ETH is not set
+
+#
+# Ethernet (10000 Mbit)
+#
+# CONFIG_CHELSIO_T1 is not set
+# CONFIG_IXGB is not set
+# CONFIG_S2IO is not set
+# CONFIG_MYRI10GE is not set
+
+#
+# Token Ring devices
+#
+# CONFIG_TR is not set
+
+#
+# Wireless LAN (non-hamradio)
+#
+# CONFIG_NET_RADIO is not set
+
+#
+# Wan interfaces
+#
+# CONFIG_WAN is not set
+# CONFIG_FDDI is not set
+# CONFIG_HIPPI is not set
+# CONFIG_PPP is not set
+# CONFIG_SLIP is not set
+# CONFIG_NET_FC is not set
+# CONFIG_SHAPER is not set
+# CONFIG_NETCONSOLE is not set
+# CONFIG_NETPOLL is not set
+# CONFIG_NET_POLL_CONTROLLER is not set
+
+#
+# ISDN subsystem
+#
+# CONFIG_ISDN is not set
+
+#
+# Telephony Support
+#
+# CONFIG_PHONE is not set
+
+#
+# Input device support
+#
+CONFIG_INPUT=y
+
+#
+# Userland interfaces
+#
+CONFIG_INPUT_MOUSEDEV=y
+# CONFIG_INPUT_MOUSEDEV_PSAUX is not set
+CONFIG_INPUT_MOUSEDEV_SCREEN_X=1600
+CONFIG_INPUT_MOUSEDEV_SCREEN_Y=1200
+# CONFIG_INPUT_JOYDEV is not set
+# CONFIG_INPUT_TSDEV is not set
+# CONFIG_INPUT_EVDEV is not set
+# CONFIG_INPUT_EVBUG is not set
+
+#
+# Input Device Drivers
+#
+# CONFIG_INPUT_KEYBOARD is not set
+# CONFIG_INPUT_MOUSE is not set
+# CONFIG_INPUT_JOYSTICK is not set
+# CONFIG_INPUT_TOUCHSCREEN is not set
+# CONFIG_INPUT_MISC is not set
+
+#
+# Hardware I/O ports
+#
+# CONFIG_SERIO is not set
+# CONFIG_GAMEPORT is not set
+
+#
+# Character devices
+#
+CONFIG_VT=y
+CONFIG_VT_CONSOLE=y
+CONFIG_HW_CONSOLE=y
+# CONFIG_VT_HW_CONSOLE_BINDING is not set
+# CONFIG_SERIAL_NONSTANDARD is not set
+
+#
+# Serial drivers
+#
+
+#
+# Non-8250 serial port support
+#
+# CONFIG_SERIAL_ICOM is not set
+# CONFIG_SERIAL_JSM is not set
+CONFIG_UNIX98_PTYS=y
+CONFIG_LEGACY_PTYS=y
+CONFIG_LEGACY_PTY_COUNT=256
+CONFIG_HVC_DRIVER=y
+CONFIG_HVC_CONSOLE=y
+# CONFIG_HVC_RTAS is not set
+# CONFIG_HVCS is not set
+
+#
+# IPMI
+#
+# CONFIG_IPMI_HANDLER is not set
+
+#
+# Watchdog Cards
+#
+# CONFIG_WATCHDOG is not set
+CONFIG_HW_RANDOM=y
+CONFIG_GEN_RTC=y
+# CONFIG_GEN_RTC_X is not set
+# CONFIG_DTLK is not set
+# CONFIG_R3964 is not set
+# CONFIG_APPLICOM is not set
+
+#
+# Ftape, the floppy tape device driver
+#
+# CONFIG_AGP is not set
+# CONFIG_DRM is not set
+# CONFIG_RAW_DRIVER is not set
+# CONFIG_HANGCHECK_TIMER is not set
+
+#
+# TPM devices
+#
+# CONFIG_TCG_TPM is not set
+# CONFIG_TELCLOCK is not set
+
+#
+# I2C support
+#
+CONFIG_I2C=y
+CONFIG_I2C_CHARDEV=y
+
+#
+# I2C Algorithms
+#
+CONFIG_I2C_ALGOBIT=y
+# CONFIG_I2C_ALGOPCF is not set
+# CONFIG_I2C_ALGOPCA is not set
+
+#
+# I2C Hardware Bus support
+#
+# CONFIG_I2C_ALI1535 is not set
+# CONFIG_I2C_ALI1563 is not set
+# CONFIG_I2C_ALI15X3 is not set
+# CONFIG_I2C_AMD756 is not set
+CONFIG_I2C_AMD8111=y
+# CONFIG_I2C_I801 is not set
+# CONFIG_I2C_I810 is not set
+# CONFIG_I2C_PIIX4 is not set
+# CONFIG_I2C_NFORCE2 is not set
+# CONFIG_I2C_OCORES is not set
+# CONFIG_I2C_PARPORT_LIGHT is not set
+# CONFIG_I2C_PROSAVAGE is not set
+# CONFIG_I2C_SAVAGE4 is not set
+# CONFIG_I2C_SIS5595 is not set
+# CONFIG_I2C_SIS630 is not set
+# CONFIG_I2C_SIS96X is not set
+# CONFIG_I2C_STUB is not set
+# CONFIG_I2C_VIA is not set
+# CONFIG_I2C_VIAPRO is not set
+# CONFIG_I2C_VOODOO3 is not set
+# CONFIG_I2C_PCA_ISA is not set
+
+#
+# Miscellaneous I2C Chip support
+#
+# CONFIG_SENSORS_DS1337 is not set
+# CONFIG_SENSORS_DS1374 is not set
+# CONFIG_SENSORS_EEPROM is not set
+# CONFIG_SENSORS_PCF8574 is not set
+# CONFIG_SENSORS_PCA9539 is not set
+# CONFIG_SENSORS_PCF8591 is not set
+# CONFIG_SENSORS_MAX6875 is not set
+# CONFIG_I2C_DEBUG_CORE is not set
+# CONFIG_I2C_DEBUG_ALGO is not set
+# CONFIG_I2C_DEBUG_BUS is not set
+# CONFIG_I2C_DEBUG_CHIP is not set
+
+#
+# SPI support
+#
+# CONFIG_SPI is not set
+# CONFIG_SPI_MASTER is not set
+
+#
+# Dallas's 1-wire bus
+#
+
+#
+# Hardware Monitoring support
+#
+# CONFIG_HWMON is not set
+# CONFIG_HWMON_VID is not set
+
+#
+# Misc devices
+#
+
+#
+# Multimedia devices
+#
+# CONFIG_VIDEO_DEV is not set
+CONFIG_VIDEO_V4L2=y
+
+#
+# Digital Video Broadcasting Devices
+#
+# CONFIG_DVB is not set
+
+#
+# Graphics support
+#
+CONFIG_FIRMWARE_EDID=y
+# CONFIG_FB is not set
+
+#
+# Console display driver support
+#
+# CONFIG_VGA_CONSOLE is not set
+CONFIG_DUMMY_CONSOLE=y
+
+#
+# Sound
+#
+# CONFIG_SOUND is not set
+
+#
+# USB support
+#
+CONFIG_USB_ARCH_HAS_HCD=y
+CONFIG_USB_ARCH_HAS_OHCI=y
+CONFIG_USB_ARCH_HAS_EHCI=y
+# CONFIG_USB is not set
+
+#
+# NOTE: USB_STORAGE enables SCSI, and 'SCSI disk support'
+#
+
+#
+# USB Gadget Support
+#
+# CONFIG_USB_GADGET is not set
+
+#
+# MMC/SD Card support
+#
+# CONFIG_MMC is not set
+
+#
+# LED devices
+#
+# CONFIG_NEW_LEDS is not set
+
+#
+# LED drivers
+#
+
+#
+# LED Triggers
+#
+
+#
+# InfiniBand support
+#
+CONFIG_INFINIBAND=y
+CONFIG_INFINIBAND_USER_MAD=y
+CONFIG_INFINIBAND_USER_ACCESS=y
+CONFIG_INFINIBAND_ADDR_TRANS=y
+CONFIG_INFINIBAND_MTHCA=y
+CONFIG_INFINIBAND_MTHCA_DEBUG=y
+CONFIG_INFINIBAND_IPOIB=y
+CONFIG_INFINIBAND_IPOIB_DEBUG=y
+CONFIG_INFINIBAND_IPOIB_DEBUG_DATA=y
+CONFIG_INFINIBAND_SRP=y
+# CONFIG_INFINIBAND_ISER is not set
+
+#
+# EDAC - error detection and reporting (RAS) (EXPERIMENTAL)
+#
+
+#
+# Real Time Clock
+#
+# CONFIG_RTC_CLASS is not set
+
+#
+# DMA Engine support
+#
+# CONFIG_DMA_ENGINE is not set
+
+#
+# DMA Clients
+#
+
+#
+# DMA Devices
+#
+
+#
+# File systems
+#
+CONFIG_EXT2_FS=y
+# CONFIG_EXT2_FS_XATTR is not set
+CONFIG_EXT2_FS_XIP=y
+CONFIG_FS_XIP=y
+CONFIG_EXT3_FS=y
+# CONFIG_EXT3_FS_XATTR is not set
+CONFIG_JBD=y
+# CONFIG_JBD_DEBUG is not set
+CONFIG_REISERFS_FS=y
+# CONFIG_REISERFS_CHECK is not set
+CONFIG_REISERFS_PROC_INFO=y
+CONFIG_REISERFS_FS_XATTR=y
+CONFIG_REISERFS_FS_POSIX_ACL=y
+CONFIG_REISERFS_FS_SECURITY=y
+# CONFIG_JFS_FS is not set
+CONFIG_FS_POSIX_ACL=y
+# CONFIG_XFS_FS is not set
+# CONFIG_OCFS2_FS is not set
+# CONFIG_MINIX_FS is not set
+# CONFIG_ROMFS_FS is not set
+CONFIG_INOTIFY=y
+CONFIG_INOTIFY_USER=y
+# CONFIG_QUOTA is not set
+CONFIG_DNOTIFY=y
+CONFIG_AUTOFS_FS=y
+# CONFIG_AUTOFS4_FS is not set
+# CONFIG_FUSE_FS is not set
+
+#
+# CD-ROM/DVD Filesystems
+#
+CONFIG_ISO9660_FS=y
+# CONFIG_JOLIET is not set
+# CONFIG_ZISOFS is not set
+# CONFIG_UDF_FS is not set
+
+#
+# DOS/FAT/NT Filesystems
+#
+CONFIG_FAT_FS=y
+CONFIG_MSDOS_FS=y
+CONFIG_VFAT_FS=y
+CONFIG_FAT_DEFAULT_CODEPAGE=437
+CONFIG_FAT_DEFAULT_IOCHARSET="iso8859-1"
+# CONFIG_NTFS_FS is not set
+
+#
+# Pseudo filesystems
+#
+CONFIG_PROC_FS=y
+CONFIG_PROC_KCORE=y
+CONFIG_SYSFS=y
+CONFIG_TMPFS=y
+# CONFIG_HUGETLB_PAGE is not set
+CONFIG_RAMFS=y
+# CONFIG_CONFIGFS_FS is not set
+
+#
+# Miscellaneous filesystems
+#
+# CONFIG_ADFS_FS is not set
+# CONFIG_AFFS_FS is not set
+# CONFIG_HFS_FS is not set
+# CONFIG_HFSPLUS_FS is not set
+# CONFIG_BEFS_FS is not set
+# CONFIG_BFS_FS is not set
+# CONFIG_EFS_FS is not set
+CONFIG_CRAMFS=y
+# CONFIG_VXFS_FS is not set
+# CONFIG_HPFS_FS is not set
+# CONFIG_QNX4FS_FS is not set
+# CONFIG_SYSV_FS is not set
+# CONFIG_UFS_FS is not set
+
+#
+# Network File Systems
+#
+CONFIG_NFS_FS=y
+CONFIG_NFS_V3=y
+CONFIG_NFS_V3_ACL=y
+CONFIG_NFS_V4=y
+# CONFIG_NFS_DIRECTIO is not set
+# CONFIG_NFSD is not set
+CONFIG_ROOT_NFS=y
+CONFIG_LOCKD=y
+CONFIG_LOCKD_V4=y
+CONFIG_NFS_ACL_SUPPORT=y
+CONFIG_NFS_COMMON=y
+CONFIG_SUNRPC=y
+CONFIG_SUNRPC_GSS=y
+CONFIG_RPCSEC_GSS_KRB5=y
+# CONFIG_RPCSEC_GSS_SPKM3 is not set
+# CONFIG_SMB_FS is not set
+# CONFIG_CIFS is not set
+# CONFIG_CIFS_DEBUG2 is not set
+# CONFIG_NCP_FS is not set
+# CONFIG_CODA_FS is not set
+# CONFIG_AFS_FS is not set
+# CONFIG_9P_FS is not set
+
+#
+# Partition Types
+#
+CONFIG_PARTITION_ADVANCED=y
+# CONFIG_ACORN_PARTITION is not set
+# CONFIG_OSF_PARTITION is not set
+# CONFIG_AMIGA_PARTITION is not set
+# CONFIG_ATARI_PARTITION is not set
+CONFIG_MAC_PARTITION=y
+CONFIG_MSDOS_PARTITION=y
+# CONFIG_BSD_DISKLABEL is not set
+# CONFIG_MINIX_SUBPARTITION is not set
+# CONFIG_SOLARIS_X86_PARTITION is not set
+# CONFIG_UNIXWARE_DISKLABEL is not set
+# CONFIG_LDM_PARTITION is not set
+# CONFIG_SGI_PARTITION is not set
+# CONFIG_ULTRIX_PARTITION is not set
+# CONFIG_SUN_PARTITION is not set
+# CONFIG_KARMA_PARTITION is not set
+# CONFIG_EFI_PARTITION is not set
+
+#
+# Native Language Support
+#
+CONFIG_NLS=y
+CONFIG_NLS_DEFAULT="utf-8"
+# CONFIG_NLS_CODEPAGE_437 is not set
+# CONFIG_NLS_CODEPAGE_737 is not set
+# CONFIG_NLS_CODEPAGE_775 is not set
+# CONFIG_NLS_CODEPAGE_850 is not set
+# CONFIG_NLS_CODEPAGE_852 is not set
+# CONFIG_NLS_CODEPAGE_855 is not set
+# CONFIG_NLS_CODEPAGE_857 is not set
+# CONFIG_NLS_CODEPAGE_860 is not set
+# CONFIG_NLS_CODEPAGE_861 is not set
+# CONFIG_NLS_CODEPAGE_862 is not set
+# CONFIG_NLS_CODEPAGE_863 is not set
+# CONFIG_NLS_CODEPAGE_864 is not set
+# CONFIG_NLS_CODEPAGE_865 is not set
+# CONFIG_NLS_CODEPAGE_866 is not set
+# CONFIG_NLS_CODEPAGE_869 is not set
+# CONFIG_NLS_CODEPAGE_936 is not set
+# CONFIG_NLS_CODEPAGE_950 is not set
+# CONFIG_NLS_CODEPAGE_932 is not set
+# CONFIG_NLS_CODEPAGE_949 is not set
+# CONFIG_NLS_CODEPAGE_874 is not set
+# CONFIG_NLS_ISO8859_8 is not set
+# CONFIG_NLS_CODEPAGE_1250 is not set
+# CONFIG_NLS_CODEPAGE_1251 is not set
+# CONFIG_NLS_ASCII is not set
+# CONFIG_NLS_ISO8859_1 is not set
+# CONFIG_NLS_ISO8859_2 is not set
+# CONFIG_NLS_ISO8859_3 is not set
+# CONFIG_NLS_ISO8859_4 is not set
+# CONFIG_NLS_ISO8859_5 is not set
+# CONFIG_NLS_ISO8859_6 is not set
+# CONFIG_NLS_ISO8859_7 is not set
+# CONFIG_NLS_ISO8859_9 is not set
+# CONFIG_NLS_ISO8859_13 is not set
+# CONFIG_NLS_ISO8859_14 is not set
+# CONFIG_NLS_ISO8859_15 is not set
+# CONFIG_NLS_KOI8_R is not set
+# CONFIG_NLS_KOI8_U is not set
+CONFIG_NLS_UTF8=y
+
+#
+# Library routines
+#
+CONFIG_CRC_CCITT=y
+# CONFIG_CRC16 is not set
+CONFIG_CRC32=y
+# CONFIG_LIBCRC32C is not set
+CONFIG_ZLIB_INFLATE=y
+CONFIG_ZLIB_DEFLATE=y
+CONFIG_TEXTSEARCH=y
+CONFIG_TEXTSEARCH_KMP=y
+CONFIG_TEXTSEARCH_BM=y
+CONFIG_TEXTSEARCH_FSM=y
+CONFIG_PLIST=y
+
+#
+# Instrumentation Support
+#
+# CONFIG_PROFILING is not set
+# CONFIG_KPROBES is not set
+
+#
+# Kernel hacking
+#
+# CONFIG_PRINTK_TIME is not set
+CONFIG_MAGIC_SYSRQ=y
+# CONFIG_UNUSED_SYMBOLS is not set
+CONFIG_DEBUG_KERNEL=y
+CONFIG_LOG_BUF_SHIFT=17
+CONFIG_DETECT_SOFTLOCKUP=y
+# CONFIG_SCHEDSTATS is not set
+CONFIG_DEBUG_SLAB=y
+CONFIG_DEBUG_MUTEXES=y
+# CONFIG_DEBUG_RT_MUTEXES is not set
+# CONFIG_RT_MUTEX_TESTER is not set
+# CONFIG_DEBUG_SPINLOCK is not set
+CONFIG_DEBUG_SPINLOCK_SLEEP=y
+# CONFIG_DEBUG_KOBJECT is not set
+CONFIG_DEBUG_INFO=y
+CONFIG_DEBUG_FS=y
+# CONFIG_DEBUG_VM is not set
+CONFIG_FORCED_INLINING=y
+# CONFIG_RCU_TORTURE_TEST is not set
+CONFIG_DEBUG_STACKOVERFLOW=y
+CONFIG_DEBUG_STACK_USAGE=y
+CONFIG_DEBUGGER=y
+CONFIG_XMON=y
+CONFIG_XMON_DEFAULT=y
+# CONFIG_IRQSTACKS is not set
+CONFIG_BOOTX_TEXT=y
+# CONFIG_PPC_EARLY_DEBUG is not set
+
+#
+# Security options
+#
+# CONFIG_KEYS is not set
+# CONFIG_SECURITY is not set
+CONFIG_XEN=y
+CONFIG_XEN_INTERFACE_VERSION=0x00030202
+
+#
+# XEN
+#
+CONFIG_XEN_PRIVILEGED_GUEST=y
+CONFIG_XEN_UNPRIVILEGED_GUEST=y
+CONFIG_XEN_PRIVCMD=y
+CONFIG_XEN_BACKEND=y
+# CONFIG_XEN_PCIDEV_BACKEND is not set
+CONFIG_XEN_BLKDEV_BACKEND=y
+CONFIG_XEN_XENBUS_DEV=y
+CONFIG_XEN_NETDEV_BACKEND=y
+# CONFIG_XEN_NETDEV_PIPELINED_TRANSMITTER is not set
+CONFIG_XEN_NETDEV_LOOPBACK=y
+# CONFIG_XEN_TPMDEV_BACKEND is not set
+CONFIG_XEN_BLKDEV_FRONTEND=y
+CONFIG_XEN_NETDEV_FRONTEND=y
+CONFIG_XEN_SCRUB_PAGES=y
+CONFIG_XEN_DISABLE_SERIAL=y
+CONFIG_XEN_SYSFS=y
+# CONFIG_XEN_COMPAT_030002_AND_LATER is not set
+CONFIG_XEN_COMPAT_LATEST_ONLY=y
+# CONFIG_XEN_COMPAT_030002 is not set
+CONFIG_HAVE_ARCH_ALLOC_SKB=y
+CONFIG_HAVE_ARCH_DEV_ALLOC_SKB=y
+CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
+CONFIG_NO_IDLE_HZ=y
+CONFIG_XEN_DEVMEM=y
+CONFIG_XEN_SKBUFF=y
+CONFIG_XEN_REBOOT=y
+CONFIG_XEN_XENCOMM=y
+
+#
+# Cryptographic options
+#
+CONFIG_CRYPTO=y
+CONFIG_CRYPTO_HMAC=y
+# CONFIG_CRYPTO_NULL is not set
+# CONFIG_CRYPTO_MD4 is not set
+CONFIG_CRYPTO_MD5=y
+CONFIG_CRYPTO_SHA1=y
+# CONFIG_CRYPTO_SHA256 is not set
+# CONFIG_CRYPTO_SHA512 is not set
+# CONFIG_CRYPTO_WP512 is not set
+# CONFIG_CRYPTO_TGR192 is not set
+CONFIG_CRYPTO_DES=y
+# CONFIG_CRYPTO_BLOWFISH is not set
+# CONFIG_CRYPTO_TWOFISH is not set
+# CONFIG_CRYPTO_SERPENT is not set
+# CONFIG_CRYPTO_AES is not set
+# CONFIG_CRYPTO_CAST5 is not set
+# CONFIG_CRYPTO_CAST6 is not set
+# CONFIG_CRYPTO_TEA is not set
+# CONFIG_CRYPTO_ARC4 is not set
+# CONFIG_CRYPTO_KHAZAD is not set
+# CONFIG_CRYPTO_ANUBIS is not set
+CONFIG_CRYPTO_DEFLATE=y
+# CONFIG_CRYPTO_MICHAEL_MIC is not set
+# CONFIG_CRYPTO_CRC32C is not set
+# CONFIG_CRYPTO_TEST is not set
+
+#
+# Hardware crypto devices
+#
diff -rpuN linux-2.6.18.8/arch/powerpc/Kconfig linux-2.6.18-xen-3.3.0/arch/powerpc/Kconfig
--- linux-2.6.18.8/arch/powerpc/Kconfig	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/Kconfig	2008-08-21 11:36:07.000000000 +0200
@@ -439,6 +439,17 @@ config UDBG_RTAS_CONSOLE
 	bool
 	default n
 
+config PPC_XEN
+       bool "Enable Xen compatible kernel"
+       depends on PPC_MULTIPLATFORM && PPC64 && PPC_MAPLE && PPC_PSERIES && SMP
+       select XEN
+       select XEN_PRIVILEGED_GUEST
+       select XEN_UNPRIVILEGED_GUEST
+       select XEN_XENCOMM
+       
+       help
+	  This option will compile a kernel compatible with Xen hypervisor
+
 config XICS
 	depends on PPC_PSERIES
 	bool
@@ -1080,6 +1091,8 @@ source "arch/powerpc/Kconfig.debug"
 
 source "security/Kconfig"
 
+source "drivers/xen/Kconfig"
+
 config KEYS_COMPAT
 	bool
 	depends on COMPAT && KEYS
diff -rpuN linux-2.6.18.8/arch/powerpc/Kconfig.debug linux-2.6.18-xen-3.3.0/arch/powerpc/Kconfig.debug
--- linux-2.6.18.8/arch/powerpc/Kconfig.debug	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/Kconfig.debug	2008-08-21 11:36:07.000000000 +0200
@@ -160,6 +160,20 @@ config PPC_EARLY_DEBUG_ISERIES
 	  Select this to enable early debugging for legacy iSeries. You need
 	  to hit "Ctrl-x Ctrl-x" to see the messages on the console.
 
+config PPC_EARLY_DEBUG_XEN_DOM0
+	bool "Xen Dom0 Console"
+	depends on PPC_XEN
+	help
+	  Select this to enable early debugging for Xen Dom0. Setting
+	  this will result in a kernel that may not work as a DomU.
+
+config PPC_EARLY_DEBUG_XEN_DOMU
+	bool "Xen DomU Console"
+	depends on PPC_XEN && XEN_UNPRIVILEGED_GUEST
+	help
+	  Select this to enable early debugging for Xen DomU. Setting
+	  this will result in a kernel that may not work as a Dom0.
+
 endchoice
 
 endmenu
diff -rpuN linux-2.6.18.8/arch/powerpc/kernel/cpu_setup_power4.S linux-2.6.18-xen-3.3.0/arch/powerpc/kernel/cpu_setup_power4.S
--- linux-2.6.18.8/arch/powerpc/kernel/cpu_setup_power4.S	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/kernel/cpu_setup_power4.S	2008-08-21 11:36:07.000000000 +0200
@@ -73,6 +73,13 @@ _GLOBAL(__970_cpu_preinit)
 	blr
 
 _GLOBAL(__setup_cpu_ppc970)
+	/*
+	 * Do nothing if not running in HV mode
+	 */
+	mfmsr	r0
+	rldicl.	r0,r0,4,63
+	beqlr
+
 	mfspr	r0,SPRN_HID0
 	li	r11,5			/* clear DOZE and SLEEP */
 	rldimi	r0,r11,52,8		/* set NAP and DPM */
diff -rpuN linux-2.6.18.8/arch/powerpc/kernel/prom_init.c linux-2.6.18-xen-3.3.0/arch/powerpc/kernel/prom_init.c
--- linux-2.6.18.8/arch/powerpc/kernel/prom_init.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/kernel/prom_init.c	2008-08-21 11:36:07.000000000 +0200
@@ -188,6 +188,7 @@ static unsigned long __initdata prom_tce
 #define PLATFORM_LPAR		0x0001
 #define PLATFORM_POWERMAC	0x0400
 #define PLATFORM_GENERIC	0x0500
+#define PLATFORM_GENERIC_XEN	(PLATFORM_GENERIC | PLATFORM_LPAR)
 
 static int __initdata of_platform;
 
@@ -1529,6 +1530,14 @@ static int __init prom_find_machine_type
 	phandle rtas;
 	int x;
 #endif
+#ifdef CONFIG_PPC_XEN
+	phandle xen;
+
+	xen = call_prom("finddevice", 1, 1, ADDR("/xen"));
+	if (PHANDLE_VALID(xen)) {
+		return PLATFORM_GENERIC_XEN;
+	}
+#endif
 
 	/* Look for a PowerMac */
 	len = prom_getprop(_prom->root, "compatible",
@@ -2262,6 +2271,31 @@ unsigned long __init prom_init(unsigned 
 	if (RELOC(of_platform) == PLATFORM_PSERIES)
 		prom_initialize_tce_table();
 #endif
+#ifdef CONFIG_PPC_XEN
+	if (RELOC(of_platform) & PLATFORM_LPAR) {
+		phandle xen;
+
+		prom_debug("XXX:checking for Xen OF package\n");
+
+		xen = call_prom("finddevice", 1, 1, ADDR("/xen"));
+		if (PHANDLE_VALID(xen)) {
+			u64 res[2];
+			int l;
+			ulong base;
+
+			l = prom_getprop(xen, "reserved", res, sizeof (res));
+			if (l != sizeof(res)) {
+				prom_panic("Xen reserved prop not exist\n");
+			}
+			
+			base = alloc_down(res[1], PAGE_SIZE, 0);
+			if (base != res[0]) {
+				prom_panic("XSI != alloc_down()\n");
+			}
+			reserve_mem(res[0], res[1]);
+		}
+	}	
+#endif
 
 	/*
 	 * On non-powermacs, try to instantiate RTAS and puts all CPUs
diff -rpuN linux-2.6.18.8/arch/powerpc/kernel/setup-common.c linux-2.6.18-xen-3.3.0/arch/powerpc/kernel/setup-common.c
--- linux-2.6.18.8/arch/powerpc/kernel/setup-common.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/kernel/setup-common.c	2008-08-21 11:36:07.000000000 +0200
@@ -387,6 +387,12 @@ void __init smp_setup_cpu_maps(void)
 		}
 	}
 
+	if (machine_is(xen)) {
+		/* something more inteligent perhaps? */
+		for (cpu = 0; cpu < NR_CPUS; cpu++)
+			cpu_set(cpu, cpu_possible_map);
+	}
+
 #ifdef CONFIG_PPC64
 	/*
 	 * On pSeries LPAR, we need to know how many cpus
diff -rpuN linux-2.6.18.8/arch/powerpc/kernel/udbg.c linux-2.6.18-xen-3.3.0/arch/powerpc/kernel/udbg.c
--- linux-2.6.18.8/arch/powerpc/kernel/udbg.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/kernel/udbg.c	2008-08-21 11:36:07.000000000 +0200
@@ -45,6 +45,9 @@ void __init udbg_early_init(void)
 #elif defined(CONFIG_PPC_EARLY_DEBUG_ISERIES)
 	/* For iSeries - hit Ctrl-x Ctrl-x to see the output */
 	udbg_init_iseries();
+#elif defined(CONFIG_PPC_EARLY_DEBUG_XEN_DOM0) || \
+	defined(CONFIG_PPC_EARLY_DEBUG_XEN_DOMU)
+	udbg_init_xen();
 #endif
 }
 
diff -rpuN linux-2.6.18.8/arch/powerpc/kernel/vdso.c linux-2.6.18-xen-3.3.0/arch/powerpc/kernel/vdso.c
--- linux-2.6.18.8/arch/powerpc/kernel/vdso.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/kernel/vdso.c	2008-08-21 11:36:07.000000000 +0200
@@ -282,6 +282,13 @@ int arch_setup_additional_pages(struct l
 	 * pages though
 	 */
 	vma->vm_flags = VM_READ|VM_EXEC|VM_MAYREAD|VM_MAYWRITE|VM_MAYEXEC;
+	/*
+	 * Make sure the vDSO gets into every core dump.
+	 * Dumping its contents makes post-mortem fully interpretable later
+	 * without matching up the same kernel and hardware config to see
+	 * what PC values meant.
+	 */
+	vma->vm_flags |= VM_ALWAYSDUMP;
 	vma->vm_flags |= mm->def_flags;
 	vma->vm_page_prot = protection_map[vma->vm_flags & 0x7];
 	vma->vm_ops = &vdso_vmops;
diff -rpuN linux-2.6.18.8/arch/powerpc/Makefile linux-2.6.18-xen-3.3.0/arch/powerpc/Makefile
--- linux-2.6.18.8/arch/powerpc/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -65,6 +65,7 @@ CPPFLAGS-$(CONFIG_PPC32) := -Iarch/$(ARC
 AFLAGS-$(CONFIG_PPC32)	:= -Iarch/$(ARCH)
 CFLAGS-$(CONFIG_PPC64)	:= -mminimal-toc -mtraceback=none  -mcall-aixdesc
 CFLAGS-$(CONFIG_PPC32)	:= -Iarch/$(ARCH) -ffixed-r2 -mmultiple
+CFLAGS-$(CONFIG_PPC_XEN) += -Iinclude/asm-$(ARCH)/xen
 CPPFLAGS	+= $(CPPFLAGS-y)
 AFLAGS		+= $(AFLAGS-y)
 CFLAGS		+= -msoft-float -pipe $(CFLAGS-y)
diff -rpuN linux-2.6.18.8/arch/powerpc/mm/slb_low.S linux-2.6.18-xen-3.3.0/arch/powerpc/mm/slb_low.S
--- linux-2.6.18.8/arch/powerpc/mm/slb_low.S	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/mm/slb_low.S	2008-08-21 11:36:07.000000000 +0200
@@ -51,6 +51,23 @@ _GLOBAL(slb_allocate_realmode)
 	*/
 	bne	cr7,1f
 
+#ifdef CONFIG_PPC_XEN
+_GLOBAL(slb_miss_kernel_load_xen_nop)
+	b 3f
+	/* Need to check if it is in the part of our XEN Foreign Map */
+	rldicl	r9,r3,30,63		/* get Xen region */
+	cmpldi	cr7,r9,1		/* cmp this bit set to 1 */
+	bne	cr7,3f
+	/* Xen Linear mapping encoding bits, the "li" instruction below
+	 * could be patched below (like the other pages of the linear map)
+	 * if we ever wish to map anything other that 4K pages in 
+	 * this region, right now it is fine as zero.
+	 */
+_GLOBAL(slb_miss_kernel_load_xen_linear)
+	li	r11,0
+	b	slb_finish_load
+3:
+#endif
 	/* Linear mapping encoding bits, the "li" instruction below will
 	 * be patched by the kernel at boot
 	 */
diff -rpuN linux-2.6.18.8/arch/powerpc/platforms/Makefile linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/Makefile
--- linux-2.6.18.8/arch/powerpc/platforms/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -12,6 +12,8 @@ obj-$(CONFIG_PPC_85xx)		+= 85xx/
 obj-$(CONFIG_PPC_86xx)		+= 86xx/
 obj-$(CONFIG_PPC_PSERIES)	+= pseries/
 obj-$(CONFIG_PPC_ISERIES)	+= iseries/
+# must occur before xen hosting platforms
+obj-$(CONFIG_PPC_XEN)		+= xen/
 obj-$(CONFIG_PPC_MAPLE)		+= maple/
 obj-$(CONFIG_PPC_CELL)		+= cell/
 obj-$(CONFIG_EMBEDDED6xx)	+= embedded6xx/
diff -rpuN linux-2.6.18.8/arch/powerpc/platforms/pseries/iommu.c linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/pseries/iommu.c
--- linux-2.6.18.8/arch/powerpc/platforms/pseries/iommu.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/pseries/iommu.c	2008-08-21 11:36:07.000000000 +0200
@@ -531,6 +531,17 @@ static void iommu_dev_setup_pSeriesLP(st
 	 * already allocated.
 	 */
 	dn = pci_device_to_OF_node(dev);
+	if (dn == NULL) {
+#ifdef CONFIG_PPC_XEN
+		/* this becomes possible for Xen Dom0 */
+		DBG("%s, dev %p (%s) has no OF devtree entree\n", __func__,
+		    dev, pci_name(dev));
+		return;
+#else
+		panic("%s, dev %p (%s) has no OF devtree entree\n", __func__,
+		      dev, pci_name(dev));
+#endif
+	}
 
 	for (pdn = dn; pdn && PCI_DN(pdn) && !PCI_DN(pdn)->iommu_table;
 	     pdn = pdn->parent) {
diff -rpuN linux-2.6.18.8/arch/powerpc/platforms/xen/balloon.c linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/balloon.c
--- linux-2.6.18.8/arch/powerpc/platforms/xen/balloon.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/balloon.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,82 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ *
+ * Copyright (C) IBM Corp. 2006
+ *
+ * Authors: Jimi Xenidis <jimix@watson.ibm.com>
+ */
+
+#include <linux/module.h>
+#include <linux/mm.h>
+#include <asm/hypervisor.h>
+#include "setup.h"
+
+/*
+ * FIXME: Port balloon driver, if ever
+ */
+
+struct page **alloc_empty_pages_and_pagevec(int nr_pages)
+{
+	struct page *page, **pagevec;
+	int i;
+
+	pagevec = kmalloc(sizeof(*pagevec) * nr_pages, GFP_KERNEL);
+	if (pagevec == NULL)
+		return  NULL;
+
+	for (i = 0; i < nr_pages; i++) {
+		page = alloc_foreign_page();
+		BUG_ON(page == NULL);
+		pagevec[i] = page;
+		/* There is no real page backing us yet so it cannot
+		 * be scrubbed */
+	}
+
+	return pagevec;
+}
+
+void free_empty_pages_and_pagevec(struct page **pagevec, int nr_pages)
+{
+	int i;
+
+	if (pagevec == NULL)
+		return;
+
+	for (i = 0; i < nr_pages; i++) {
+		free_foreign_page(pagevec[i]);
+	}
+	
+	kfree(pagevec);
+}
+
+void balloon_dealloc_empty_page_range(
+	struct page *page, unsigned long nr_pages)
+{
+	__free_pages(page, get_order(nr_pages * PAGE_SIZE));
+}
+
+void balloon_update_driver_allowance(long delta)
+{
+}
+
+void balloon_release_driver_page(struct page *page)
+{
+	BUG();
+}
+
+EXPORT_SYMBOL_GPL(balloon_update_driver_allowance);
+EXPORT_SYMBOL_GPL(alloc_empty_pages_and_pagevec);
+EXPORT_SYMBOL_GPL(free_empty_pages_and_pagevec);
+EXPORT_SYMBOL_GPL(balloon_release_driver_page);
diff -rpuN linux-2.6.18.8/arch/powerpc/platforms/xen/gnttab.c linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/gnttab.c
--- linux-2.6.18.8/arch/powerpc/platforms/xen/gnttab.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/gnttab.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,468 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ *
+ * Copyright (C) IBM Corp. 2006
+ *
+ * Authors: Jimi Xenidis <jimix@watson.ibm.com>
+ */
+
+#include <linux/config.h>
+#include <linux/vmalloc.h>
+#include <linux/memory_hotplug.h>
+#include <xen/gnttab.h>
+#include <asm/hypervisor.h>
+#include <xen/interface/grant_table.h>
+#include <asm/pgtable.h>
+#include <asm/sections.h>
+#include <asm/io.h>
+#include <asm/machdep.h>
+#include <asm/prom.h>
+#include <asm/cacheflush.h>
+#include "setup.h"
+#include "../pseries/plpar_wrappers.h"
+
+#undef DEBUG
+
+#ifdef DEBUG
+#define DBG(fmt...) printk(KERN_EMERG fmt)
+#else
+#define DBG(fmt...)
+#endif
+
+#define NR_GRANT_FRAMES 4
+
+struct address_space xen_foreign_dummy_mapping;
+
+static ulong foreign_map_pfn;
+static ulong foreign_map_pgs;
+static unsigned long *foreign_map_bitmap;
+
+
+/* hijack _mapcount */
+static inline int gnt_mapcount(struct page *page)
+{
+	return atomic_read(&(page)->_mapcount) + 1;
+}
+
+static inline int gnt_map(struct page *page)
+{
+	/* return true is transition from -1 to 0 */
+	return atomic_inc_and_test(&page->_mapcount);
+}
+
+static inline int gnt_unmap(struct page *page)
+{
+	int val;
+
+	val = atomic_dec_return(&page->_mapcount);
+	if (val < -1) {
+		atomic_inc(&page->_mapcount);
+		printk(KERN_EMERG "%s: %d\n", __func__, val);
+	}
+
+	return (val == -1);
+}
+
+
+static long map_to_linear(ulong paddr)
+{
+	unsigned long vaddr;
+	int psize;
+	unsigned long mode;
+	int slot;
+	uint shift;
+	unsigned long tmp_mode;
+
+	psize = MMU_PAGE_4K;
+	shift = mmu_psize_defs[psize].shift;
+	mode = _PAGE_ACCESSED | _PAGE_DIRTY | _PAGE_COHERENT | PP_RWXX;
+	vaddr = (ulong)__va(paddr);
+
+	{
+		unsigned long vpn, hash, hpteg;
+		unsigned long vsid = get_kernel_vsid(vaddr);
+		unsigned long va = (vsid << 28) | (vaddr & 0x0fffffff);
+
+		vpn = va >> shift;
+		tmp_mode = mode;
+		
+		/* Make non-kernel text non-executable */
+		if (!in_kernel_text(vaddr))
+			tmp_mode = mode | HPTE_R_N;
+
+		hash = hpt_hash(va, shift);
+		hpteg = ((hash & htab_hash_mask) * HPTES_PER_GROUP);
+
+		BUG_ON(!ppc_md.hpte_insert);
+		slot = ppc_md.hpte_insert(hpteg, va, paddr,
+					  tmp_mode, HPTE_V_BOLTED, psize);
+		if (slot < 0)
+			printk(KERN_EMERG
+			       "%s: no more bolted entries "
+			       "HTAB[0x%lx]: 0x%lx\n",
+			       __func__, hpteg, paddr);
+	}
+	return slot;
+}
+
+static unsigned long get_hpte_vsid(ulong slot)
+{
+ 	unsigned long dword0;
+	unsigned long lpar_rc;
+	unsigned long dummy_word1;
+	unsigned long flags;
+
+	/* Read 1 pte at a time                        */
+	/* Do not need RPN to logical page translation */
+	/* No cross CEC PFT access                     */
+	flags = 0;
+
+	lpar_rc = plpar_pte_read(flags, slot, &dword0, &dummy_word1);
+
+	BUG_ON(lpar_rc != H_SUCCESS);
+
+	return dword0;
+}
+
+static long find_hpte_slot(unsigned long va, int psize)
+{
+	unsigned long hash;
+	unsigned long i, j;
+	long slot;
+	unsigned long want_v, hpte_v;
+
+	hash = hpt_hash(va, mmu_psize_defs[psize].shift);
+	want_v = hpte_encode_v(va, psize);
+
+	for (j = 0; j < 2; j++) {
+		slot = (hash & htab_hash_mask) * HPTES_PER_GROUP;
+		for (i = 0; i < HPTES_PER_GROUP; i++) {
+			hpte_v = get_hpte_vsid(slot);
+
+			if (HPTE_V_COMPARE(hpte_v, want_v)
+			    && (hpte_v & HPTE_V_VALID)
+			    && (!!(hpte_v & HPTE_V_SECONDARY) == j)) {
+				/* HPTE matches */
+				if (j)
+					slot = -slot;
+				return slot;
+			}
+			++slot;
+		}
+		hash = ~hash;
+	}
+
+	return -1;
+} 
+
+static long find_map_slot(ulong ea)
+{
+	int psize = MMU_PAGE_4K;
+	ulong vsid;
+	ulong va;
+
+	vsid = get_kernel_vsid(ea);
+	va = (vsid << 28) | (ea & 0x0fffffff);
+	
+	return find_hpte_slot(va, psize);
+}
+
+
+static void gnttab_pre_unmap_grant_ref(
+	struct gnttab_unmap_grant_ref *unmap, int count)
+{
+	long slot;
+	int i;
+	ulong ea;
+	unsigned long dummy1, dummy2;
+	ulong flags;
+
+	/* paranoia */
+	local_irq_save(flags);
+
+	for (i = 0 ; i < count; i++) {
+		struct page *page;
+
+		ea = (ulong)__va(unmap[i].host_addr);
+		page = virt_to_page(ea);
+		
+		if (!gnt_unmap(page)) {
+			DBG("%s[0x%x]: skip: 0x%lx, mapcount 0x%x\n",
+			    __func__, i, ea, gnt_mapcount(page));
+			continue;
+		}
+		slot = find_map_slot(ea);
+		if (slot < 0) {
+			printk(KERN_EMERG "%s: PTE not found: 0x%lx\n",
+			       __func__, ea);
+			continue;
+		}
+
+		DBG("%s[0x%x]: 0x%lx: mapcount: 0x%x\n",
+		    __func__, i, ea, gnt_mapcount(page));
+		plpar_pte_remove(0, slot, 0, &dummy1, &dummy2);
+	}
+	local_irq_restore(flags);
+}
+
+static void gnttab_post_map_grant_ref(
+	struct gnttab_map_grant_ref *map, int count)
+{
+	int i;
+	long slot;
+	ulong flags;
+
+	/* paranoia */
+	local_irq_save(flags);
+
+	for (i = 0 ; i < count; i++) {
+		ulong pa = map[i].host_addr;
+		struct page *page;
+
+		if (map[i].status != GNTST_okay) {
+			printk(KERN_EMERG "%s: status, skip\n", __func__);
+			continue;
+		}
+
+		BUG_ON(pa < (foreign_map_pfn << PAGE_SHIFT));
+		BUG_ON(pa >= (foreign_map_pfn << PAGE_SHIFT) + 
+		       (foreign_map_pgs << PAGE_SHIFT));
+
+		page = virt_to_page(__va(pa));
+
+		if (gnt_map(page)) {
+#ifdef DEBUG			
+			/* we need to get smarted than this */
+			slot = find_map_slot((ulong)__va(pa));
+			if (slot >= 0) {
+				DBG("%s: redundant 0x%lx\n", __func__, pa);
+				continue;
+			}
+#endif
+			slot = map_to_linear(pa);
+			DBG("%s[0x%x]: 0x%lx, mapcount:0x%x\n",
+			    __func__, i, pa, gnt_mapcount(page));
+
+		} else {
+			DBG("%s[0x%x] skip 0x%lx, mapcount:0x%x\n",
+			    __func__, i, pa, gnt_mapcount(page));
+		}
+	}
+	local_irq_restore(flags);
+}
+
+int HYPERVISOR_grant_table_op(unsigned int cmd, void *op, unsigned int count)
+{
+	void *desc;
+	void *frame_list = NULL;
+	int argsize;
+	int ret = -ENOMEM;
+
+	switch (cmd) {
+	case GNTTABOP_map_grant_ref:
+		argsize = sizeof(struct gnttab_map_grant_ref);
+		break;
+	case GNTTABOP_unmap_grant_ref:
+		gnttab_pre_unmap_grant_ref(op, count);
+		argsize = sizeof(struct gnttab_unmap_grant_ref);
+		break;
+	case GNTTABOP_setup_table: {
+		struct gnttab_setup_table setup;
+
+		memcpy(&setup, op, sizeof(setup));
+		argsize = sizeof(setup);
+
+		frame_list = xencomm_map(
+			xen_guest_handle(setup.frame_list),
+			(sizeof(*xen_guest_handle(setup.frame_list)) 
+			* setup.nr_frames));
+
+		if (frame_list == NULL)
+			return -ENOMEM;
+
+		set_xen_guest_handle(setup.frame_list, frame_list);
+		memcpy(op, &setup, sizeof(setup));
+		}
+		break;
+	case GNTTABOP_dump_table:
+		argsize = sizeof(struct gnttab_dump_table);
+		break;
+	case GNTTABOP_transfer:
+		BUG();
+		argsize = sizeof(struct gnttab_transfer);
+		break;
+	case GNTTABOP_copy:
+		argsize = sizeof(struct gnttab_transfer);
+		break;
+	case GNTTABOP_query_size:
+		argsize = sizeof(struct gnttab_query_size);
+		break;
+	default:
+		printk(KERN_EMERG "%s: unknown grant table op %d\n",
+		       __func__, cmd);
+		return -ENOSYS;
+	}
+
+	desc = xencomm_map_no_alloc(op, argsize);
+	if (desc) {
+		ret = plpar_hcall_norets(XEN_MARK(__HYPERVISOR_grant_table_op),
+					 cmd, desc, count);
+		if (!ret && cmd == GNTTABOP_map_grant_ref)
+			gnttab_post_map_grant_ref(op, count);
+		xencomm_free(desc);
+	}
+	xencomm_free(frame_list);
+
+	return ret;
+}
+EXPORT_SYMBOL(HYPERVISOR_grant_table_op);
+
+static ulong find_grant_maps(void)
+{
+	struct device_node *xen;
+	u64 *gm;
+	u64 _gm[2];
+	u64 expect;
+
+	/* This value is currently hardcoded into the SLB logic that
+	 * it written in assempler, See
+	 * slb_miss_kernel_load_xen_linear for more information.
+	 * Anything else and we can not run. */
+	expect = 34 - PAGE_SHIFT;
+
+	xen = of_find_node_by_path("/xen");
+
+	/* 
+	 * The foreign is 2x2 Cells.
+	 * The first entry is log2 of the base page frame.
+	 * The second is the number of pages
+	 */
+	gm = (u64 *)get_property(xen, "foreign-map", NULL);
+	if (gm == NULL) {
+		if (!is_initial_xendomain()) {
+			printk("OF: /xen/foreign-map not present\n");
+			_gm[0] = expect;
+			_gm[1] = 2048;
+			gm = _gm;
+		} else
+			panic("OF: /xen/foreign-map must be present\n");
+	}
+
+	if (gm[0] != expect)
+		panic("foreign-map is 0x%lx, expect 0x%lx\n",
+		      gm[0], expect);
+
+	foreign_map_pfn = 1UL << gm[0];
+	return gm[1];
+}
+
+static void setup_foreign_segment(void)
+{
+	extern int *slb_miss_kernel_load_xen_nop;
+	ulong iaddr = (ulong)slb_miss_kernel_load_xen_nop;
+
+	/* By default Linux will branch around this logic we replace
+	 * the branch with a NOP to turn the logic on */
+	*slb_miss_kernel_load_xen_nop = 0x60000000;
+	flush_icache_range(iaddr, iaddr + 4);
+}
+
+struct page *alloc_foreign_page(void)
+{
+	ulong bit;
+	do {
+		bit = find_first_zero_bit(foreign_map_bitmap,
+					  foreign_map_pgs);
+		if (bit >= foreign_map_pgs)
+			return NULL;
+	} while (test_and_set_bit(bit, foreign_map_bitmap) == 1);
+
+	return pfn_to_page(foreign_map_pfn + bit);
+}
+
+void free_foreign_page(struct page *page)
+{
+	ulong bit = page_to_pfn(page) - foreign_map_pfn;
+
+	BUG_ON(bit >= foreign_map_pgs);
+	BUG_ON(!test_bit(bit, foreign_map_bitmap));
+
+	clear_bit(bit, foreign_map_bitmap);
+}
+
+static void setup_grant_area(void)
+{
+	ulong pgs;
+	int err;
+	struct zone *zone;
+	struct pglist_data *pgdata;
+	int nid;
+
+	pgs = find_grant_maps();
+	setup_foreign_segment();
+
+	printk("%s: Xen VIO will use a foreign address space of 0x%lx pages\n",
+	       __func__, pgs);
+
+	/* add pages to the zone */
+	nid = 0;
+	pgdata = NODE_DATA(nid);
+	zone = pgdata->node_zones;
+
+	err = __add_pages(zone, foreign_map_pfn, pgs);
+
+	if (err < 0) {
+		printk(KERN_EMERG "%s: add_pages(0x%lx, 0x%lx) = %d\n",
+		       __func__, foreign_map_pfn, pgs, err);
+		BUG();
+	}
+
+	/* create a bitmap to manage these pages */
+	foreign_map_bitmap = kmalloc(BITS_TO_LONGS(pgs) * sizeof(long),
+				     GFP_KERNEL);
+	if (foreign_map_bitmap == NULL) {
+		printk(KERN_EMERG 
+		       "%s: could not allocate foreign_map_bitmap to "
+		       "manage 0x%lx foreign pages\n", __func__, pgs);
+		BUG();
+	}
+	/* I'm paranoid so make sure we assign the top bits so we
+	 * don't give them away */
+	bitmap_fill(&foreign_map_bitmap[BITS_TO_LONGS(pgs) - 1],
+		    BITS_PER_LONG);
+	/* now clear all the real bits */
+	bitmap_zero(foreign_map_bitmap, pgs);
+
+	foreign_map_pgs = pgs;
+}
+
+void *arch_gnttab_alloc_shared(unsigned long *frames)
+{
+	void *shared;
+	ulong pa = frames[0] << PAGE_SHIFT;
+	static int resume;
+
+	shared = ioremap(pa, PAGE_SIZE * NR_GRANT_FRAMES);
+	BUG_ON(shared == NULL);
+	printk("%s: grant table at %p\n", __func__, shared);
+
+	/* no need to do the rest of this if we are resuming */
+	if (!resume)
+		setup_grant_area();
+
+	resume = 1;
+
+	return shared;
+}
diff -rpuN linux-2.6.18.8/arch/powerpc/platforms/xen/hcall.c linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/hcall.c
--- linux-2.6.18.8/arch/powerpc/platforms/xen/hcall.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/hcall.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,891 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ *
+ * Copyright (C) IBM Corp. 2006, 2007
+ *
+ * Authors: Hollis Blanchard <hollisb@us.ibm.com>
+ */
+
+#include <linux/types.h>
+#include <linux/errno.h>
+#include <linux/kernel.h>
+#include <linux/gfp.h>
+#include <linux/module.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/domctl.h>
+#include <xen/interface/sysctl.h>
+#include <xen/interface/platform.h>
+#include <xen/interface/memory.h>
+#include <xen/interface/xencomm.h>
+#include <xen/interface/version.h>
+#include <xen/interface/sched.h>
+#include <xen/interface/event_channel.h>
+#include <xen/interface/physdev.h>
+#include <xen/interface/vcpu.h>
+#include <xen/interface/xsm/acm_ops.h>
+#include <xen/interface/kexec.h>
+#include <xen/public/privcmd.h>
+#include <asm/hypercall.h>
+#include <asm/page.h>
+#include <asm/uaccess.h>
+#include <asm/hvcall.h>
+#include "setup.h"
+
+/* Xencomm notes:
+ *
+ * For kernel memory, we assume that virtually contiguous pages are also
+ * physically contiguous. This allows us to avoid creating descriptors for
+ * kernel hypercalls, such as console and event channel operations.
+ *
+ * In general, we need a xencomm descriptor to cover the top-level data
+ * structure (e.g. the domctl op), plus another for every embedded pointer to
+ * another data structure (i.e. for every GUEST_HANDLE).
+ */
+
+int HYPERVISOR_console_io(int cmd, int count, char *str)
+{
+	struct xencomm_handle *desc;
+	int rc;
+
+	desc = xencomm_map_no_alloc(str, count);
+	if (desc == NULL)
+		return -EINVAL;
+
+	rc = plpar_hcall_norets(XEN_MARK(__HYPERVISOR_console_io),
+				  cmd, count, desc);
+
+	xencomm_free(desc);
+
+	return rc;
+}
+EXPORT_SYMBOL(HYPERVISOR_console_io);
+
+int HYPERVISOR_event_channel_op(int cmd, void *op)
+{
+	int rc;
+
+	struct xencomm_handle *desc =
+		xencomm_map_no_alloc(op, sizeof(evtchn_op_t));
+	if (desc == NULL)
+		return -EINVAL;
+
+	rc = plpar_hcall_norets(XEN_MARK(__HYPERVISOR_event_channel_op),
+				cmd, desc);
+
+	xencomm_free(desc);
+
+	return rc;
+
+}
+EXPORT_SYMBOL(HYPERVISOR_event_channel_op);
+
+int HYPERVISOR_xen_version(int cmd, void *arg)
+{
+	struct xencomm_handle *desc;
+	const unsigned long hcall = __HYPERVISOR_xen_version;
+	int argsize;
+	int rc;
+
+	switch (cmd) {
+	case XENVER_version:
+		/* do not actually pass an argument */
+		return plpar_hcall_norets(XEN_MARK(hcall), cmd, 0);
+	case XENVER_extraversion:
+		argsize = sizeof(xen_extraversion_t);
+		break;
+	case XENVER_compile_info:
+		argsize = sizeof(xen_compile_info_t);
+		break;
+	case XENVER_capabilities:
+		argsize = sizeof(xen_capabilities_info_t);
+		break;
+	case XENVER_changeset:
+		argsize = sizeof(xen_changeset_info_t);
+		break;
+	case XENVER_platform_parameters:
+		argsize = sizeof(xen_platform_parameters_t);
+		break;
+	case XENVER_pagesize:
+		if (arg == NULL)
+			argsize = 0;
+		else
+			argsize = sizeof(void *);
+		break;
+	case XENVER_get_features:
+		argsize = sizeof(xen_feature_info_t);
+		break;
+	default:
+		printk(KERN_ERR "%s: unknown version cmd %d\n", __func__, cmd);
+		return -ENOSYS;
+	}
+
+	/* desc could be NULL in the case of XENVER_pagesize with NULL arg */
+	desc = xencomm_map(arg, argsize);
+
+	rc = plpar_hcall_norets(XEN_MARK(hcall), cmd, desc);
+
+	xencomm_free(desc);
+
+	return rc;
+}
+EXPORT_SYMBOL(HYPERVISOR_xen_version);
+
+
+int HYPERVISOR_physdev_op(int cmd, void *op)
+{
+	struct xencomm_handle *desc =
+		xencomm_map_no_alloc(op, sizeof(physdev_op_t));
+	int rc;
+
+	if (desc == NULL)
+		return -EINVAL;
+
+	rc = plpar_hcall_norets(XEN_MARK(__HYPERVISOR_physdev_op),
+				cmd, desc);
+
+	xencomm_free(desc);
+
+	return rc;
+}
+EXPORT_SYMBOL(HYPERVISOR_physdev_op);
+
+int HYPERVISOR_sched_op(int cmd, void *arg)
+{
+	int argsize = 0;
+	int rc = -EINVAL;
+	struct xencomm_handle *desc;
+	struct xencomm_handle *ports = NULL;
+
+	switch (cmd) {
+	case SCHEDOP_yield:
+	case SCHEDOP_block:
+		return plpar_hcall_norets(XEN_MARK(__HYPERVISOR_sched_op),
+					  cmd, 0);
+		break;
+
+	case SCHEDOP_poll: {
+		struct sched_poll sched_poll;
+
+		argsize = sizeof(struct sched_poll);
+
+		memcpy(&sched_poll, arg, sizeof(sched_poll));
+
+		ports = xencomm_map(
+				xen_guest_handle(sched_poll.ports),
+				(sizeof(evtchn_port_t) * sched_poll.nr_ports));
+
+		if (ports == NULL)
+			return -ENOMEM;
+
+		set_xen_guest_handle(sched_poll.ports, (evtchn_port_t *)ports);
+		memcpy(arg, &sched_poll, sizeof(sched_poll));
+
+		}
+		break;
+	case SCHEDOP_shutdown:
+		argsize = sizeof(struct sched_shutdown);
+		break;
+	case SCHEDOP_remote_shutdown:
+		argsize = sizeof(struct sched_remote_shutdown);
+		break;
+	default:
+		printk(KERN_ERR "%s: unknown sched op %d\n", __func__, cmd);
+		return -ENOSYS;
+	}
+
+	desc = xencomm_map_no_alloc(arg, argsize);
+	if (desc) {
+		rc = plpar_hcall_norets(XEN_MARK(__HYPERVISOR_sched_op),
+					cmd, desc);
+		xencomm_free(desc);
+	}
+
+	xencomm_free(ports);
+
+	return rc;
+}
+EXPORT_SYMBOL(HYPERVISOR_sched_op);
+
+int HYPERVISOR_suspend(unsigned long srec)
+{
+	int cmd = SCHEDOP_shutdown;
+	struct sched_shutdown sched_shutdown = {
+		.reason = SHUTDOWN_suspend,
+	};
+	struct xencomm_handle *desc;
+
+	desc = xencomm_map_no_alloc(&sched_shutdown, sizeof(struct sched_shutdown));
+
+	return plpar_hcall_norets(XEN_MARK(__HYPERVISOR_sched_op),
+				  cmd, desc, srec);
+}
+EXPORT_SYMBOL(HYPERVISOR_suspend);
+
+int HYPERVISOR_kexec_op(unsigned long op, void *args)
+{
+	unsigned long argsize;
+	struct xencomm_handle *desc;
+
+	switch (op) {
+		case KEXEC_CMD_kexec_get_range:
+			argsize = sizeof(struct xen_kexec_range);
+			break;
+		case KEXEC_CMD_kexec_load:
+			argsize = sizeof(struct xen_kexec_load);
+			break;
+		case KEXEC_CMD_kexec_unload:
+			argsize = sizeof(struct xen_kexec_load);
+			break;
+		case KEXEC_CMD_kexec:
+			argsize = sizeof(struct xen_kexec_exec);
+			break;
+		default:
+			return -ENOSYS;
+	}
+	desc = xencomm_map_no_alloc(args, argsize);
+
+	return plpar_hcall_norets(XEN_MARK(__HYPERVISOR_kexec_op),
+				  op, desc);
+}
+EXPORT_SYMBOL(HYPERVISOR_kexec_op);
+
+int HYPERVISOR_poll(
+	evtchn_port_t *ports, unsigned int nr_ports, u64 timeout)
+{
+	struct sched_poll sched_poll = {
+		.nr_ports = nr_ports,
+		.timeout = jiffies_to_ns(timeout)
+	};
+	set_xen_guest_handle(sched_poll.ports, ports);
+
+	return HYPERVISOR_sched_op(SCHEDOP_poll, &sched_poll);
+}
+EXPORT_SYMBOL(HYPERVISOR_poll);
+
+typedef ulong (mf_t)(ulong arg0, ...);
+
+static mf_t *multicall_funcs[] = {
+	[__HYPERVISOR_grant_table_op] = (mf_t *)HYPERVISOR_grant_table_op,
+};
+
+int HYPERVISOR_multicall(void *call_list, int nr_calls)
+{
+	/* we blow out the multicall because the xencomm stuff is jsut
+	 * too tricky */
+	multicall_entry_t *mcl = (multicall_entry_t *)call_list;
+	multicall_entry_t *c;
+	int i;
+	mf_t *mf;
+	int res;
+	ulong flags;
+
+	/* let make sure all the calls are supported */
+	for (i = 0; i < nr_calls; i++) {
+		mf = multicall_funcs[mcl[i].op];
+		BUG_ON(mf == NULL);
+	}
+	/* disable interrupts until we are done all calls */
+	local_irq_save(flags);
+	for (i = 0; i < nr_calls; i++) {
+		/* lookup supported multicalls */
+		c = &mcl[i];
+		mf = multicall_funcs[c->op];
+		res = mf(c->args[0], c->args[1], c->args[2],
+			 c->args[3], c->args[4], c->args[5]);
+		c->result = res;
+	}
+	local_irq_restore(flags);
+	return 0;
+}
+EXPORT_SYMBOL(HYPERVISOR_multicall);
+
+
+/* privcmd operations: */
+
+static int xenppc_privcmd_domctl(privcmd_hypercall_t *hypercall)
+{
+	xen_domctl_t kern_op;
+	xen_domctl_t __user *user_op = (xen_domctl_t __user *)hypercall->arg[0];
+	struct xencomm_handle *op_desc;
+	struct xencomm_handle *desc = NULL;
+	int ret = 0;
+
+	if (copy_from_user(&kern_op, user_op, sizeof(xen_domctl_t)))
+		return -EFAULT;
+
+	if (kern_op.interface_version != XEN_DOMCTL_INTERFACE_VERSION) {
+		printk(KERN_WARNING "%s: %s %x != %x\n", __func__, current->comm,
+				kern_op.interface_version, XEN_DOMCTL_INTERFACE_VERSION);
+		return -EACCES;
+	}
+
+	op_desc = xencomm_map(&kern_op, sizeof(xen_domctl_t));
+	if (op_desc == NULL)
+		return -ENOMEM;
+
+	switch (kern_op.cmd) {
+	case XEN_DOMCTL_createdomain:
+	case XEN_DOMCTL_destroydomain:
+	case XEN_DOMCTL_pausedomain:
+	case XEN_DOMCTL_unpausedomain:
+	case XEN_DOMCTL_getdomaininfo:
+		break;
+	case XEN_DOMCTL_getmemlist:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.getmemlist.buffer),
+			kern_op.u.getmemlist.max_pfns * sizeof(unsigned long));
+
+		if (desc == NULL)
+			ret = -ENOMEM;
+
+		set_xen_guest_handle(kern_op.u.getmemlist.buffer,
+				     (void *)desc);
+		break;
+	case XEN_DOMCTL_getpageframeinfo:
+		break;
+	case XEN_DOMCTL_getpageframeinfo2:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.getpageframeinfo2.array),
+			kern_op.u.getpageframeinfo2.num);
+
+		if (desc == NULL)
+			ret = -ENOMEM;
+
+		set_xen_guest_handle(kern_op.u.getpageframeinfo2.array,
+				     (void *)desc);
+		break;
+	case XEN_DOMCTL_shadow_op:
+
+		if (xen_guest_handle(kern_op.u.shadow_op.dirty_bitmap))
+		{
+			desc = xencomm_map(
+				xen_guest_handle(kern_op.u.shadow_op.dirty_bitmap),
+				kern_op.u.shadow_op.pages * sizeof(unsigned long));
+
+			if (desc == NULL)
+				ret = -ENOMEM;
+
+			set_xen_guest_handle(kern_op.u.shadow_op.dirty_bitmap,
+				    	 (void *)desc);
+		}
+		break;
+	case XEN_DOMCTL_max_mem:
+		break;
+	case XEN_DOMCTL_setvcpucontext:
+	case XEN_DOMCTL_getvcpucontext:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.vcpucontext.ctxt),
+			sizeof(vcpu_guest_context_t));
+
+		if (desc == NULL)
+			ret = -ENOMEM;
+
+		set_xen_guest_handle(kern_op.u.vcpucontext.ctxt,
+				     (void *)desc);
+		break;
+	case XEN_DOMCTL_getvcpuinfo:
+		break;
+	case XEN_DOMCTL_setvcpuaffinity:
+	case XEN_DOMCTL_getvcpuaffinity:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.vcpuaffinity.cpumap.bitmap),
+			(kern_op.u.vcpuaffinity.cpumap.nr_cpus + 7) / 8);
+
+		if (desc == NULL)
+			ret = -ENOMEM;
+
+		set_xen_guest_handle(kern_op.u.vcpuaffinity.cpumap.bitmap,
+				     (void *)desc);
+		break;
+	case XEN_DOMCTL_max_vcpus:
+	case XEN_DOMCTL_scheduler_op:
+	case XEN_DOMCTL_setdomainhandle:
+	case XEN_DOMCTL_setdebugging:
+	case XEN_DOMCTL_irq_permission:
+	case XEN_DOMCTL_iomem_permission:
+	case XEN_DOMCTL_ioport_permission:
+	case XEN_DOMCTL_hypercall_init:
+	case XEN_DOMCTL_arch_setup:
+	case XEN_DOMCTL_settimeoffset:
+	case XEN_DOMCTL_real_mode_area:
+		break;
+	default:
+		printk(KERN_ERR "%s: unknown domctl cmd %d\n", __func__, kern_op.cmd);
+		return -ENOSYS;
+	}
+
+	if (ret)
+		goto out; /* error mapping the nested pointer */
+
+	ret = plpar_hcall_norets(XEN_MARK(hypercall->op),op_desc);
+
+	if (copy_to_user(user_op, &kern_op, sizeof(xen_domctl_t)))
+		ret = -EFAULT;
+
+out:
+	xencomm_free(desc);
+	xencomm_free(op_desc);
+	return ret;
+}
+
+static int xenppc_privcmd_sysctl(privcmd_hypercall_t *hypercall)
+{
+	xen_sysctl_t kern_op;
+	xen_sysctl_t __user *user_op = (xen_sysctl_t __user *)hypercall->arg[0];
+	struct xencomm_handle *op_desc;
+	struct xencomm_handle *desc = NULL;
+	int ret = 0;
+
+	if (copy_from_user(&kern_op, user_op, sizeof(xen_sysctl_t)))
+		return -EFAULT;
+
+	if (kern_op.interface_version != XEN_SYSCTL_INTERFACE_VERSION) {
+		printk(KERN_WARNING "%s: %s %x != %x\n", __func__, current->comm,
+				kern_op.interface_version, XEN_SYSCTL_INTERFACE_VERSION);
+		return -EACCES;
+	}
+
+	op_desc = xencomm_map(&kern_op, sizeof(xen_sysctl_t));
+
+	if (op_desc == NULL)
+		return -ENOMEM;
+
+	switch (kern_op.cmd) {
+	case XEN_SYSCTL_readconsole:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.readconsole.buffer),
+			kern_op.u.readconsole.count);
+
+		if (desc == NULL)
+			ret = -ENOMEM;
+
+		set_xen_guest_handle(kern_op.u.readconsole.buffer,
+				     (void *)desc);
+		break;
+	case XEN_SYSCTL_tbuf_op:
+	case XEN_SYSCTL_physinfo:
+	case XEN_SYSCTL_sched_id:
+		break;
+	case XEN_SYSCTL_perfc_op:
+		/* XXX this requires *two* embedded xencomm mappings (desc and val),
+		 * and I don't feel like it right now. */
+		printk(KERN_ERR "%s: unknown sysctl cmd %d\n", __func__, kern_op.cmd);
+		return -ENOSYS;
+	case XEN_SYSCTL_getdomaininfolist:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.getdomaininfolist.buffer),
+			kern_op.u.getdomaininfolist.max_domains *
+					sizeof(xen_domctl_getdomaininfo_t));
+
+		if (desc == NULL)
+			ret = -ENOMEM;
+
+		set_xen_guest_handle(kern_op.u.getdomaininfolist.buffer,
+				     (void *)desc);
+		break;
+	default:
+		printk(KERN_ERR "%s: unknown sysctl cmd %d\n", __func__, kern_op.cmd);
+		return -ENOSYS;
+	}
+
+	if (ret)
+		goto out; /* error mapping the nested pointer */
+
+	ret = plpar_hcall_norets(XEN_MARK(hypercall->op), op_desc);
+
+	if (copy_to_user(user_op, &kern_op, sizeof(xen_sysctl_t)))
+		ret = -EFAULT;
+
+out:
+	xencomm_free(desc);
+	xencomm_free(op_desc);
+	return ret;
+}
+
+static int xenppc_privcmd_platform_op(privcmd_hypercall_t *hypercall)
+{
+	xen_platform_op_t kern_op;
+	xen_platform_op_t __user *user_op =
+			(xen_platform_op_t __user *)hypercall->arg[0];
+	struct xencomm_handle *op_desc;
+	struct xencomm_handle *desc = NULL;
+	int ret = 0;
+
+	if (copy_from_user(&kern_op, user_op, sizeof(xen_platform_op_t)))
+		return -EFAULT;
+
+	if (kern_op.interface_version != XENPF_INTERFACE_VERSION) {
+		printk(KERN_WARNING "%s: %s %x != %x\n", __func__, current->comm,
+				kern_op.interface_version, XENPF_INTERFACE_VERSION);
+		return -EACCES;
+	}
+
+	op_desc = xencomm_map(&kern_op, sizeof(xen_platform_op_t));
+
+	if (op_desc == NULL)
+		return -ENOMEM;
+
+	switch (kern_op.cmd) {
+	case XENPF_settime:
+	case XENPF_add_memtype:
+	case XENPF_del_memtype:
+	case XENPF_read_memtype:
+	case XENPF_microcode_update:
+	case XENPF_platform_quirk:
+		break;
+	default:
+		printk(KERN_ERR "%s: unknown platform_op cmd %d\n", __func__,
+				kern_op.cmd);
+		return -ENOSYS;
+	}
+
+	if (ret)
+		goto out; /* error mapping the nested pointer */
+
+	ret = plpar_hcall_norets(XEN_MARK(hypercall->op), op_desc);
+
+	if (copy_to_user(user_op, &kern_op, sizeof(xen_platform_op_t)))
+		ret = -EFAULT;
+
+out:
+	xencomm_free(desc);
+	xencomm_free(op_desc);
+	return ret;
+}
+
+int HYPERVISOR_memory_op(unsigned int cmd, void *arg)
+{
+	int ret;
+	struct xencomm_handle *op_desc;
+	xen_memory_reservation_t *mop;
+
+
+	mop = (xen_memory_reservation_t *)arg;
+
+	op_desc = xencomm_map(mop, sizeof(xen_memory_reservation_t));
+
+	if (op_desc == NULL)
+		return -ENOMEM;
+
+	switch (cmd) {
+	case XENMEM_increase_reservation:
+	case XENMEM_decrease_reservation:
+	case XENMEM_populate_physmap: {
+		struct xencomm_handle *desc = NULL;
+
+		if (xen_guest_handle(mop->extent_start)) {
+			desc = xencomm_map(
+				xen_guest_handle(mop->extent_start),
+				mop->nr_extents *
+				sizeof(*xen_guest_handle(mop->extent_start)));
+
+			if (desc == NULL) {
+				ret = -ENOMEM;
+				goto out;
+			}
+
+			set_xen_guest_handle(mop->extent_start,
+					     (void *)desc);
+		}
+
+		ret = plpar_hcall_norets(XEN_MARK(__HYPERVISOR_memory_op),
+					cmd, op_desc);
+
+		xencomm_free(desc);
+		}
+		break;
+
+	case XENMEM_maximum_ram_page:
+		/* arg is NULL so we can call thru here */
+		ret = plpar_hcall_norets(XEN_MARK(__HYPERVISOR_memory_op),
+					cmd, NULL);
+		break;
+	default:
+		printk(KERN_ERR "%s: unknown memory op %d\n", __func__, cmd);
+		ret = -ENOSYS;
+	}
+
+out:
+	xencomm_free(op_desc);
+	return ret;
+}
+EXPORT_SYMBOL(HYPERVISOR_memory_op);
+
+static int xenppc_privcmd_memory_op(privcmd_hypercall_t *hypercall)
+{
+	xen_memory_reservation_t kern_op;
+	xen_memory_reservation_t __user *user_op;
+	const unsigned long cmd = hypercall->arg[0];
+	int ret = 0;
+
+	user_op = (xen_memory_reservation_t __user *)hypercall->arg[1];
+	if (copy_from_user(&kern_op, user_op,
+			   sizeof(xen_memory_reservation_t)))
+		return -EFAULT;
+
+	ret = HYPERVISOR_memory_op(cmd, &kern_op);
+	if (ret >= 0) {
+		if (copy_to_user(user_op, &kern_op,
+				 sizeof(xen_memory_reservation_t)))
+			return -EFAULT;
+	}
+	return ret;
+}
+
+static int xenppc_privcmd_version(privcmd_hypercall_t *hypercall)
+{
+	return HYPERVISOR_xen_version(hypercall->arg[0],
+			(void *)hypercall->arg[1]);
+}
+
+static int xenppc_privcmd_event_channel_op(privcmd_hypercall_t *hypercall)
+{
+	struct xencomm_handle *desc;
+	unsigned int argsize;
+	int ret;
+
+	switch (hypercall->arg[0]) {
+	case EVTCHNOP_alloc_unbound:
+		argsize = sizeof(evtchn_alloc_unbound_t);
+		break;
+
+	case EVTCHNOP_status:
+		argsize = sizeof(evtchn_status_t);
+		break;
+
+	default:
+		printk(KERN_ERR "%s: unknown EVTCHNOP (%ld)\n",
+		       __func__, hypercall->arg[0]);
+		return -EINVAL;
+	}
+
+	desc = xencomm_map((void *)hypercall->arg[1], argsize);
+
+	if (desc == NULL)
+		return -ENOMEM;
+
+	ret = plpar_hcall_norets(XEN_MARK(hypercall->op), hypercall->arg[0],
+				desc);
+
+	xencomm_free(desc);
+	return ret;
+}
+
+static int xenppc_acmcmd_op(privcmd_hypercall_t *hypercall)
+{
+	xen_acmctl_t kern_op;
+	xen_acmctl_t __user *user_op = (xen_acmctl_t __user *)hypercall->arg[0];
+	void *op_desc;
+	void *desc = NULL, *desc2 = NULL, *desc3 = NULL, *desc4 = NULL;
+	int ret = 0;
+
+	if (copy_from_user(&kern_op, user_op, sizeof(xen_acmctl_t)))
+		return -EFAULT;
+
+	if (kern_op.interface_version != ACM_INTERFACE_VERSION) {
+		printk(KERN_WARNING "%s: %s %x != %x\n", __func__, current->comm,
+				kern_op.interface_version, ACM_INTERFACE_VERSION);
+		return -EACCES;
+	}
+
+	op_desc = xencomm_map(&kern_op, sizeof(xen_acmctl_t));
+	if (op_desc == NULL)
+		return -ENOMEM;
+
+	switch (kern_op.cmd) {
+	case ACMOP_setpolicy:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.setpolicy.pushcache),
+			kern_op.u.setpolicy.pushcache_size);
+
+		if (desc == NULL)
+			ret = -ENOMEM;
+
+		set_xen_guest_handle(kern_op.u.setpolicy.pushcache,
+		                     desc);
+		break;
+	case ACMOP_getpolicy:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.getpolicy.pullcache),
+			kern_op.u.getpolicy.pullcache_size);
+
+		if (desc == NULL)
+			ret = -ENOMEM;
+
+		set_xen_guest_handle(kern_op.u.getpolicy.pullcache,
+		                     desc);
+		break;
+	case ACMOP_dumpstats:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.dumpstats.pullcache),
+			kern_op.u.dumpstats.pullcache_size);
+
+		if (desc == NULL)
+			ret = -ENOMEM;
+
+		set_xen_guest_handle(kern_op.u.dumpstats.pullcache,
+		                     desc);
+		break;
+	case ACMOP_getssid:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.getssid.ssidbuf),
+			kern_op.u.getssid.ssidbuf_size);
+
+		if (desc == NULL)
+			ret = -ENOMEM;
+
+		set_xen_guest_handle(kern_op.u.getssid.ssidbuf,
+		                     desc);
+		break;
+	case ACMOP_getdecision:
+		break;
+	case ACMOP_chgpolicy:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.change_policy.policy_pushcache),
+			kern_op.u.change_policy.policy_pushcache_size);
+		desc2 = xencomm_map(
+		 	 xen_guest_handle(kern_op.u.change_policy.del_array),
+		 	 kern_op.u.change_policy.delarray_size);
+		desc3 = xencomm_map(
+		 	 xen_guest_handle(kern_op.u.change_policy.chg_array),
+		 	 kern_op.u.change_policy.chgarray_size);
+		desc4 = xencomm_map(
+		 	 xen_guest_handle(kern_op.u.change_policy.err_array),
+		 	 kern_op.u.change_policy.errarray_size);
+
+		if (desc  == NULL || desc2 == NULL ||
+			desc3 == NULL || desc4 == NULL) {
+			ret = -ENOMEM;
+			goto out;
+		}
+
+		set_xen_guest_handle(kern_op.u.change_policy.policy_pushcache,
+		                     desc);
+		set_xen_guest_handle(kern_op.u.change_policy.del_array,
+		                     desc2);
+		set_xen_guest_handle(kern_op.u.change_policy.chg_array,
+		                     desc3);
+		set_xen_guest_handle(kern_op.u.change_policy.err_array,
+		                     desc4);
+		break;
+	case ACMOP_relabeldoms:
+		desc = xencomm_map(
+			xen_guest_handle(kern_op.u.relabel_doms.relabel_map),
+			kern_op.u.relabel_doms.relabel_map_size);
+		desc2 = xencomm_map(
+			xen_guest_handle(kern_op.u.relabel_doms.err_array),
+			kern_op.u.relabel_doms.errarray_size);
+
+		if (desc  == NULL || desc2 == NULL) {
+			ret = -ENOMEM;
+			goto out;
+		}
+
+		set_xen_guest_handle(kern_op.u.relabel_doms.relabel_map,
+		                     desc);
+		set_xen_guest_handle(kern_op.u.relabel_doms.err_array,
+		                     desc2);
+		break;
+	default:
+		printk(KERN_ERR "%s: unknown/unsupported acmctl cmd %d\n",
+		       __func__, kern_op.cmd);
+		return -ENOSYS;
+	}
+
+	if (ret)
+		goto out; /* error mapping the nested pointer */
+
+	ret = plpar_hcall_norets(XEN_MARK(hypercall->op),op_desc);
+
+	if (copy_to_user(user_op, &kern_op, sizeof(xen_acmctl_t)))
+		ret = -EFAULT;
+
+out:
+	xencomm_free(desc);
+	xencomm_free(desc2);
+	xencomm_free(desc3);
+	xencomm_free(desc4);
+	xencomm_free(op_desc);
+	return ret;
+}
+
+
+/* The PowerPC hypervisor runs in a separate address space from Linux
+ * kernel/userspace, i.e. real mode. We must therefore translate userspace
+ * pointers to something the hypervisor can make sense of. */
+int privcmd_hypercall(privcmd_hypercall_t *hypercall)
+{
+	switch (hypercall->op) {
+	case __HYPERVISOR_domctl:
+		return xenppc_privcmd_domctl(hypercall);
+	case __HYPERVISOR_sysctl:
+		return xenppc_privcmd_sysctl(hypercall);
+	case __HYPERVISOR_platform_op:
+		return xenppc_privcmd_platform_op(hypercall);
+	case __HYPERVISOR_memory_op:
+		return xenppc_privcmd_memory_op(hypercall);
+	case __HYPERVISOR_xen_version:
+		return xenppc_privcmd_version(hypercall);
+	case __HYPERVISOR_event_channel_op:
+		return xenppc_privcmd_event_channel_op(hypercall);
+	case __HYPERVISOR_acm_op:
+		return xenppc_acmcmd_op(hypercall);
+	default:
+		printk(KERN_ERR "%s: unknown hcall (%ld)\n", __func__, hypercall->op);
+		/* maybe we'll get lucky and the hcall needs no translation. */
+		return plpar_hcall_norets(XEN_MARK(hypercall->op),
+				hypercall->arg[0],
+				hypercall->arg[1],
+				hypercall->arg[2],
+				hypercall->arg[3],
+				hypercall->arg[4]);
+	}
+}
+
+int HYPERVISOR_vcpu_op(int cmd, int vcpuid, void *extra_args)
+{
+	int argsize;
+	const unsigned long hcall = __HYPERVISOR_vcpu_op;
+	struct xencomm_handle *desc;
+	int rc;
+
+	switch (cmd) {
+	case  VCPUOP_initialise:
+		argsize = sizeof(vcpu_guest_context_t);
+		break;
+	case VCPUOP_up:
+	case VCPUOP_down:
+	case VCPUOP_is_up:
+		return plpar_hcall_norets(XEN_MARK(hcall), cmd, vcpuid, 0);
+
+	case VCPUOP_get_runstate_info:
+		argsize = sizeof (vcpu_runstate_info_t);
+		break;
+	default:
+		printk(KERN_ERR "%s: unknown version cmd %d\n", __func__, cmd);
+		return -ENOSYS;
+	}
+
+	desc = xencomm_map_no_alloc(extra_args, argsize);
+
+	if (desc == NULL)
+		return -EINVAL;
+
+	rc = plpar_hcall_norets(XEN_MARK(hcall), cmd, vcpuid, desc);
+
+	xencomm_free(desc);
+
+	return rc;
+}
diff -rpuN linux-2.6.18.8/arch/powerpc/platforms/xen/Makefile linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/Makefile
--- linux-2.6.18.8/arch/powerpc/platforms/xen/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,20 @@
+obj-y	+= gnttab.o
+obj-y	+= hcall.o
+obj-y	+= reboot.o
+obj-y	+= setup.o
+obj-y	+= smp.o
+obj-y	+= time.o
+obj-y	+= udbg_xen.o
+obj-y	+= xen_guest.o
+obj-y	+= xencomm.o
+
+# we need the latest __XEN_INTERFACE_VERSION__ (see xen-compat.h)
+CFLAGS_hcall.o += -D__XEN_TOOLS__
+
+ifndef CONFIG_XEN_BALLOON
+obj-y += balloon.o
+endif
+
+ifndef CONFIG_XEN_UTIL
+obj-y	+= util.o
+endif
diff -rpuN linux-2.6.18.8/arch/powerpc/platforms/xen/reboot.c linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/reboot.c
--- linux-2.6.18.8/arch/powerpc/platforms/xen/reboot.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/reboot.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,53 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ *
+ * Copyright (C) IBM Corp. 2006
+ *
+ * Authors: Jimi Xenidis <jimix@watson.ibm.com>
+ */
+
+#include <linux/module.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/io/console.h>
+#include <xen/xencons.h>
+#include <asm/hypervisor.h>
+#include <asm/machdep.h>
+
+static void domain_machine_restart(char * __unused)
+{
+	/* We really want to get pending console data out before we die. */
+	xencons_force_flush();
+	HYPERVISOR_shutdown(SHUTDOWN_reboot);
+}
+
+static void domain_machine_power_off(void)
+{
+	/* We really want to get pending console data out before we die. */
+	xencons_force_flush();
+	HYPERVISOR_shutdown(SHUTDOWN_poweroff);
+}
+
+void xen_reboot_init(struct machdep_calls *md)
+{
+	if (md != NULL) {
+		ppc_md.restart	 = md->restart;
+		ppc_md.power_off = md->power_off;
+		ppc_md.halt	 = md->halt;
+	} else {
+		ppc_md.restart	 = domain_machine_restart;
+		ppc_md.power_off = domain_machine_power_off;
+		ppc_md.halt	 = domain_machine_power_off;
+	}
+}
diff -rpuN linux-2.6.18.8/arch/powerpc/platforms/xen/setup.c linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/setup.c
--- linux-2.6.18.8/arch/powerpc/platforms/xen/setup.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/setup.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,336 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ *
+ * Copyright (C) IBM Corp. 2006
+ *
+ * Authors: Jimi Xenidis <jimix@watson.ibm.com>
+ */
+
+#define DEBUG
+#define CONFIG_SHARE_MPIC
+
+#include <linux/module.h>
+#include <linux/rwsem.h>
+#include <linux/delay.h>
+#include <linux/console.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/sched.h>
+#include <xen/evtchn.h>
+#include <xen/features.h>
+#include <xen/xencons.h>
+#include <asm/udbg.h>
+#include <asm/pgtable.h>
+#include <asm/prom.h>
+#include <asm/iommu.h>
+#include <asm/mmu.h>
+#include <asm/abs_addr.h>
+#include <asm/machdep.h>
+#include <asm/hypervisor.h>
+#include <asm/time.h>
+#include <asm/pmc.h>
+#include "setup.h"
+
+#ifdef DEBUG
+#define DBG(fmt...) udbg_printf(fmt)
+#else
+#define DBG(fmt...)
+#endif
+
+/* Apperently on other arches this could be used before its defined,
+ * this should not be the case in PPC */
+shared_info_t *HYPERVISOR_shared_info = (shared_info_t *)NULL;
+EXPORT_SYMBOL(HYPERVISOR_shared_info);
+
+/* Raw start-of-day parameters from the hypervisor. */
+static start_info_t xsi;
+start_info_t *xen_start_info;
+EXPORT_SYMBOL(xen_start_info);
+
+extern struct machdep_calls mach_maple_md;
+extern void maple_pci_init(void);
+
+static unsigned long foreign_mfn_flag;
+
+/* Must be called with &vma->vm_mm->mmap_sem locked for write */
+int direct_remap_pfn_range(struct vm_area_struct *vma,
+		unsigned long address, 
+		unsigned long mfn,
+		unsigned long size, 
+		pgprot_t prot,
+		domid_t  domid)
+{
+	int rc;
+
+	/* Set the MFN flag to tell Xen that this is not a PFN. */
+	printk("%s: mapping mfn 0x%lx (size 0x%lx) -> 0x%lx\n", __func__,
+			mfn, size, mfn | foreign_mfn_flag);
+	mfn = mfn | foreign_mfn_flag;
+
+	WARN_ON(!rwsem_is_locked(&vma->vm_mm->mmap_sem));
+	rc = remap_pfn_range(vma, address, mfn, size, prot);
+
+	return rc;
+}
+
+static void __init xen_fw_feature_init(void)
+{
+	DBG(" -> %s\n", __func__);
+
+	powerpc_firmware_features = 0;
+
+	powerpc_firmware_features |= FW_FEATURE_LPAR;
+	powerpc_firmware_features |= FW_FEATURE_TCE | FW_FEATURE_DABR;
+		
+	printk(KERN_INFO "firmware_features = 0x%lx\n", 
+			powerpc_firmware_features);
+
+	DBG(" <- %s\n", __func__);
+}
+
+/* if these were global then I could get them from the pseries/setup.c */
+static int pseries_set_dabr(unsigned long dabr)
+{
+	return plpar_hcall_norets(H_SET_DABR, dabr);
+}
+
+static int pseries_set_xdabr(unsigned long dabr)
+{
+	/* We want to catch accesses from kernel and userspace */
+	return plpar_hcall_norets(H_SET_XDABR, dabr,
+			H_DABRX_KERNEL | H_DABRX_USER);
+}
+
+/* 
+ * Early initialization.
+ */
+static void __init xenppc_init_early(void)
+{
+	struct device_node *xen;
+
+	DBG(" -> %s\n", __func__);
+
+	xen = of_find_node_by_path("/xen");
+
+	xen_start_info = &xsi;
+
+	/* fill out start_info_t from devtree */
+	if ((char *)get_property(xen, "privileged", NULL))
+		xen_start_info->flags |= SIF_PRIVILEGED;
+	if ((char *)get_property(xen, "initdomain", NULL))
+		xen_start_info->flags |= SIF_INITDOMAIN;
+	xen_start_info->shared_info = *((u64 *)get_property(xen, 
+	   "shared-info", NULL));
+
+	/* only look for store and console for guest domains */
+	if (xen_start_info->flags == 0) {
+		struct device_node *console = of_find_node_by_path("/xen/console");
+		struct device_node *store = of_find_node_by_path("/xen/store");
+
+		xen_start_info->store_mfn = (*((u64 *)get_property(store,
+		   "reg", NULL))) >> PAGE_SHIFT;
+		xen_start_info->store_evtchn = *((u32 *)get_property(store,
+		   "interrupts", NULL));
+		xen_start_info->console.domU.mfn = (*((u64 *)get_property(console,
+		   "reg", NULL))) >> PAGE_SHIFT;
+		xen_start_info->console.domU.evtchn = *((u32 *)get_property(console,
+		   "interrupts", NULL));
+	}
+
+	HYPERVISOR_shared_info = __va(xen_start_info->shared_info);
+
+	udbg_init_xen();
+
+	DBG("xen_start_info at %p\n", xen_start_info);
+	DBG("    magic          %s\n", xen_start_info->magic);
+	DBG("    flags          %x\n", xen_start_info->flags);
+	DBG("    shared_info    %lx, %p\n",
+	    xen_start_info->shared_info, HYPERVISOR_shared_info);
+	DBG("    store_mfn      %llx\n", xen_start_info->store_mfn);
+	DBG("    store_evtchn   %x\n", xen_start_info->store_evtchn);
+	DBG("    console_mfn    %llx\n", xen_start_info->console.domU.mfn);
+	DBG("    console_evtchn %x\n", xen_start_info->console.domU.evtchn);
+
+	xen_setup_time(&mach_maple_md);
+
+	add_preferred_console("xvc", 0, NULL);
+
+	if (get_property(xen, "power-control", NULL))
+		xen_reboot_init(&mach_maple_md);
+	else
+		xen_reboot_init(NULL);
+
+	if (is_initial_xendomain()) {
+		u64 *mfnflag = (u64 *)get_property(xen, "mfn-flag", NULL);
+		if (mfnflag) {
+			foreign_mfn_flag = (1UL << mfnflag[0]);
+			printk("OF: using 0x%lx as foreign mfn flag\n", foreign_mfn_flag);
+		} else
+			printk("OF: /xen/mfn-base must be present it build guests\n");
+	}
+
+	/* get the domain features */
+	setup_xen_features();
+
+	DBG("Hello World I'm Maple Xen-LPAR!\n");
+
+	if (firmware_has_feature(FW_FEATURE_DABR))
+		ppc_md.set_dabr = pseries_set_dabr;
+	else if (firmware_has_feature(FW_FEATURE_XDABR))
+		ppc_md.set_dabr = pseries_set_xdabr;
+
+	iommu_init_early_pSeries();
+
+	DBG(" <- %s\n", __func__);
+}
+
+/*
+ * this interface is limiting
+ */
+static int running_on_xen;
+int is_running_on_xen(void)
+{
+	return running_on_xen;
+}
+EXPORT_SYMBOL(is_running_on_xen);
+
+static void xenppc_power_save(void)
+{
+	/* SCHEDOP_yield could immediately return. Instead, we
+	 * want to idle in the Xen idle domain, so use
+	 * SCHEDOP_block with a one-shot timer. */
+	/* XXX do tickless stuff here. See
+	 * linux-2.6-xen-sparse/arch/xen/i386/kernel/time.c */
+	u64 now_ns = tb_to_ns(get_tb());
+	u64 offset_ns = jiffies_to_ns(1);
+	int rc;
+
+	rc = HYPERVISOR_set_timer_op(now_ns + offset_ns);
+	BUG_ON(rc != 0);
+
+	HYPERVISOR_sched_op(SCHEDOP_block, NULL);
+}
+
+void __init xenppc_setup_arch(void)
+{
+	/* init to some ~sane value until calibrate_delay() runs */
+	loops_per_jiffy = 50000000;
+
+	/* Lookup PCI hosts */
+	if (is_initial_xendomain())
+		maple_pci_init();
+
+#ifdef CONFIG_DUMMY_CONSOLE
+	conswitchp = &dummy_con;
+#endif
+#ifdef CONFIG_SMP
+	/* let them fly */
+	xen_setup_smp();
+#endif
+
+	printk(KERN_INFO "Using Xen idle loop\n");
+}
+
+static int __init xen_probe_flat_dt(unsigned long node,
+				    const char *uname, int depth,
+				    void *data)
+{
+	if (depth != 1)
+		return 0;
+	if (strcmp(uname, "xen") != 0)
+ 		return 0;
+
+	running_on_xen = 1;
+
+	return 1;
+}
+
+/*
+ * Called very early, MMU is off, device-tree isn't unflattened
+ */
+/* forward ref */
+struct machdep_calls __initdata xen_md;
+static int __init xenppc_probe(void)
+{
+	of_scan_flat_dt(xen_probe_flat_dt, NULL);
+
+	if (!running_on_xen)
+		return 0;
+
+	xen_fw_feature_init();
+
+	hpte_init_lpar();
+
+	return 1;
+}
+
+static void __init xenppc_progress(char *s, unsigned short hex)
+{
+	printk("*** %04x : %s\n", hex, s ? s : "");
+}
+
+unsigned int xenppc_get_irq(struct pt_regs *regs)
+{
+	evtchn_do_upcall(regs);
+	/* evtchn_do_upcall() handles all pending event channels directly, so there
+	 * is nothing for do_IRQ() to do.
+	 * XXX This means we aren't using IRQ stacks. */
+	return NO_IRQ;
+}
+
+static void xenppc_enable_pmcs(void)
+{
+	unsigned long set, reset;
+
+	power4_enable_pmcs();
+
+	set = 1UL << 63;
+	reset = 0;
+	plpar_hcall_norets(H_PERFMON, set, reset);
+}
+
+#ifdef CONFIG_KEXEC
+void xen_machine_kexec(struct kimage *image)
+{
+	panic("%s(%p): called\n", __func__, image);
+}
+
+int xen_machine_kexec_prepare(struct kimage *image)
+{
+	panic("%s(%p): called\n", __func__, image);
+}
+
+void xen_machine_crash_shutdown(struct pt_regs *regs)
+{
+	panic("%s(%p): called\n", __func__, regs);
+}       
+#endif
+
+define_machine(xen) {
+	.name			= "Xen-Maple",
+	.probe			= xenppc_probe,
+	.setup_arch		= xenppc_setup_arch,
+	.init_early		= xenppc_init_early,
+	.init_IRQ		= xen_init_IRQ,
+	.get_irq		= xenppc_get_irq,
+	.calibrate_decr		= generic_calibrate_decr,
+	.progress		= xenppc_progress,
+	.power_save		= xenppc_power_save,
+	.enable_pmcs	= xenppc_enable_pmcs,
+#ifdef CONFIG_KEXEC
+	.machine_kexec		= xen_machine_kexec,
+	.machine_kexec_prepare	= xen_machine_kexec_prepare,
+	.machine_crash_shutdown	= xen_machine_crash_shutdown,
+#endif
+};
diff -rpuN linux-2.6.18.8/arch/powerpc/platforms/xen/setup.h linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/setup.h
--- linux-2.6.18.8/arch/powerpc/platforms/xen/setup.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/setup.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,47 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ *
+ * Copyright (C) IBM Corp. 2006
+ *
+ * Authors: Jimi Xenidis <jimix@watson.ibm.com>
+ */
+
+#include <asm/machdep.h>
+#include <asm/time.h>
+
+extern void evtchn_init_IRQ(void);
+extern void xen_init_IRQ(void);
+extern void xen_reboot_init(struct machdep_calls *);
+extern void xen_maple_init_IRQ(void);
+extern unsigned int xen_get_irq(struct pt_regs *regs);
+
+static inline u64 tb_to_ns(u64 tb)
+{
+	if (likely(tb_ticks_per_sec)) {
+		return tb * (1000000000UL / tb_ticks_per_sec);
+	}
+	return 0;
+}
+
+static inline u64 jiffies_to_ns(unsigned long j) 
+{
+	return j * (1000000000UL / HZ);
+}
+
+extern struct page *alloc_foreign_page(void);
+extern void free_foreign_page(struct page *page);
+
+extern void __init xen_setup_time(struct machdep_calls *host_md);
+extern void xen_setup_smp(void);
diff -rpuN linux-2.6.18.8/arch/powerpc/platforms/xen/smp.c linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/smp.c
--- linux-2.6.18.8/arch/powerpc/platforms/xen/smp.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/smp.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,444 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ *
+ * Copyright (C) IBM Corp. 2006
+ *
+ * Authors: Jimi Xenidis <jimix@watson.ibm.com>
+ */
+
+#include <linux/kernel.h>
+#include <linux/config.h>
+#include <linux/bootmem.h>
+#include <linux/irq.h>
+#include <linux/smp.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/vcpu.h>
+#include <xen/evtchn.h>
+#include <asm/prom.h>
+#include <asm/udbg.h>
+#include <asm/hypervisor.h>
+#include "setup.h"
+
+#undef DEBUG
+
+#ifdef DEBUG
+#define DBG(fmt...) printk(KERN_EMERG fmt)
+#else
+#define DBG(fmt...)
+#endif
+
+static inline void *xen_of_alloc(ulong size)
+{
+	if (mem_init_done)
+		return kmalloc(size, GFP_KERNEL);
+	return alloc_bootmem(size);
+}
+static inline void xen_of_free(void *ptr)
+{
+	/* if this happens with the boot allocator then we are screwed */
+	BUG_ON(!mem_init_done);
+	kfree(ptr);
+}
+
+static struct property *dup_prop(struct property *op)
+{
+	struct property *np;
+	void *p;
+	ulong sz;
+
+
+	/* allocate everything in one go in case it fails */
+	sz = sizeof (*np); /* prop node */
+	sz += strlen(op->name) + 1; /* prop name */
+	sz += op->length; /* prop value */
+		
+	p = xen_of_alloc(sz);
+	if (!p)
+		return NULL;
+	memset(p, 0, sz);
+
+	/* prop node first */
+	np = p;
+	p += sizeof (*np);
+
+	/* value next becuase we want it aligned */
+	np->value = p;
+	p += op->length;
+
+	/* name */
+	np->name = p;
+
+	/* copy it all */
+	strcpy(np->name, op->name);
+	np->length = op->length;
+	memcpy(np->value, op->value, np->length);
+
+	return np;
+}
+
+static int dup_properties(struct device_node *dst, struct device_node *src)
+{
+	struct property *op;
+	struct property *np;
+	struct property *lp;
+	int rc = 0;
+
+	DBG("%s: duping to new cpu node: %s\n", __func__, dst->full_name);
+
+	np = lp = NULL;
+	for (op = src->properties; op != 0; op = op->next) {
+		lp = np;
+		np = dup_prop(op);
+		if (!np)
+			break;
+
+		prom_add_property(dst, np);
+	}
+
+	if (!np) {
+		DBG("%s: FAILED duping: %s\n", __func__, dst->full_name);
+		/* we could not allocate enuff so free what we have
+		 * allocated */
+		rc = -ENOMEM;
+		for (op = dst->properties; lp && op != lp; op = op->next)
+			xen_of_free(op);
+	}
+
+	return rc;
+}
+
+/* returns added device node so it can be added to procfs in the case
+ * of hotpluging */
+static struct device_node *xen_add_vcpu_node(struct device_node *boot_cpu,
+					     uint cpu)
+{
+	struct device_node *new_cpu;
+	struct property *pp;
+	void *p;
+	int sz;
+	int type_sz;
+	int name_sz;
+
+	DBG("%s: boot cpu: %s\n", __func__, boot_cpu->full_name);
+
+	/* allocate in one shot in case we fail */
+	name_sz = strlen(boot_cpu->name) + 1;
+	type_sz = strlen(boot_cpu->type) + 1;
+
+	sz = sizeof (*new_cpu);	/* the node */
+	sz += strlen(boot_cpu->full_name) + 3; /* full_name */
+	sz += name_sz; /* name */
+	sz += type_sz; /* type */
+
+	p = xen_of_alloc(sz);
+	if (!p)
+		return NULL;
+	memset(p, 0, sz);
+
+	/* the node */
+	new_cpu = p;
+	p += sizeof (*new_cpu);
+	
+	/* name */
+	new_cpu->name = p;
+	strcpy(new_cpu->name, boot_cpu->name);
+	p += name_sz;
+	
+	/* type */
+	new_cpu->type = p;
+	strcpy(new_cpu->type, boot_cpu->type);
+	p += type_sz;
+
+	/* full_name */
+	new_cpu->full_name = p;
+
+	/* assemble new full_name */
+	pp = of_find_property(boot_cpu, "name", NULL);
+	if (!pp)
+		panic("%s: no name prop\n", __func__);
+
+	DBG("%s: name is: %s = %s\n", __func__, pp->name, pp->value);
+	sprintf(new_cpu->full_name, "/cpus/%s@%u", pp->value, cpu);
+
+	if (dup_properties(new_cpu, boot_cpu)) {
+		xen_of_free(new_cpu);
+		return NULL;
+	}
+
+	/* fixup reg property */
+	DBG("%s: updating reg: %d\n", __func__, cpu);
+	pp = of_find_property(new_cpu, "reg", NULL);
+	if (!pp)
+		panic("%s: no reg prop\n", __func__);
+	*(int *)pp->value = cpu;
+
+	if (mem_init_done)
+		OF_MARK_DYNAMIC(new_cpu);
+
+	kref_init(&new_cpu->kref);
+
+	/* insert the node */
+	new_cpu->parent = of_get_parent(boot_cpu);
+	of_attach_node(new_cpu);
+	of_node_put(new_cpu->parent);
+
+	return new_cpu;
+}
+
+static void cpu_initialize_context(unsigned int vcpu, ulong entry)
+{
+	vcpu_guest_context_t ctxt;
+
+	memset(&ctxt.user_regs, 0x55, sizeof(ctxt.user_regs));
+
+	ctxt.user_regs.pc = entry;
+	ctxt.user_regs.msr = 0;
+	ctxt.user_regs.gprs[1] = 0; /* Linux uses its own stack */
+	ctxt.user_regs.gprs[3] = vcpu;
+
+	/* XXX verify this *** */
+	/* There is a buggy kernel that does not zero the "local_paca", so
+	 * we must make sure this register is 0 */
+	ctxt.user_regs.gprs[13] = 0;
+
+	DBG("%s: initializing vcpu: %d\n", __func__, vcpu);
+
+	if (HYPERVISOR_vcpu_op(VCPUOP_initialise, vcpu, &ctxt))
+		panic("%s: VCPUOP_initialise failed, vcpu: %d\n",
+		       __func__, vcpu);
+
+}
+
+static int xen_start_vcpu(uint vcpu, ulong entry)
+{
+	DBG("%s: starting vcpu: %d\n", __func__, vcpu);
+
+	cpu_initialize_context(vcpu, entry);
+
+	DBG("%s: Spinning up vcpu: %d\n", __func__, vcpu);
+	return HYPERVISOR_vcpu_op(VCPUOP_up, vcpu, NULL);
+}
+
+extern void __secondary_hold(void);
+extern unsigned long __secondary_hold_spinloop;
+extern unsigned long __secondary_hold_acknowledge;
+
+static void xen_boot_secondary_vcpus(void)
+{
+	int vcpu;
+	int rc;
+	const unsigned long mark = (unsigned long)-1;
+	unsigned long *spinloop = &__secondary_hold_spinloop;
+	unsigned long *acknowledge = &__secondary_hold_acknowledge;
+#ifdef CONFIG_PPC64
+	/* __secondary_hold is actually a descriptor, not the text address */
+	unsigned long secondary_hold = __pa(*(unsigned long *)__secondary_hold);
+#else
+	unsigned long secondary_hold = __pa(__secondary_hold);
+#endif
+	struct device_node *boot_cpu;
+
+	DBG("%s: finding CPU node\n", __func__);
+	boot_cpu = of_find_node_by_type(NULL, "cpu");
+	if (!boot_cpu)
+		panic("%s: Cannot find Booting CPU node\n", __func__);
+
+	/* Set the common spinloop variable, so all of the secondary cpus
+	 * will block when they are awakened from their OF spinloop.
+	 * This must occur for both SMP and non SMP kernels, since OF will
+	 * be trashed when we move the kernel.
+	 */
+	*spinloop = 0;
+
+	DBG("%s: Searching for all vcpu numbers > 0\n", __func__);
+	/* try and start as many as we can */
+	for (vcpu = 1; vcpu < NR_CPUS; vcpu++) {
+		int i;
+
+		rc = HYPERVISOR_vcpu_op(VCPUOP_is_up, vcpu, NULL);
+		if (rc < 0)
+			continue;
+
+		DBG("%s: Found vcpu: %d\n", __func__, vcpu);
+		/* Init the acknowledge var which will be reset by
+		 * the secondary cpu when it awakens from its OF
+		 * spinloop.
+		 */
+		*acknowledge = mark;
+
+		DBG("%s: Starting vcpu: %d at pc: 0x%lx\n", __func__,
+		    vcpu, secondary_hold);
+		rc = xen_start_vcpu(vcpu, secondary_hold);
+		if (rc)
+			panic("%s: xen_start_vpcu() failed\n", __func__);
+
+
+		DBG("%s: Waiting for ACK on vcpu: %d\n", __func__, vcpu);
+		for (i = 0; (i < 100000000) && (*acknowledge == mark); i++)
+			mb();
+
+		if (*acknowledge == vcpu)
+			DBG("%s: Recieved for ACK on vcpu: %d\n",
+			    __func__, vcpu);
+
+		xen_add_vcpu_node(boot_cpu, vcpu);
+
+		cpu_set(vcpu, cpu_present_map);
+		set_hard_smp_processor_id(vcpu, vcpu);
+	}
+	of_node_put(boot_cpu);
+	DBG("%s: end...\n", __func__);
+}
+
+static int __init smp_xen_probe(void)
+{
+	return cpus_weight(cpu_present_map);
+}
+
+static irqreturn_t xen_ppc_msg_reschedule(int irq, void *dev_id,
+					  struct pt_regs *regs)
+{
+	smp_message_recv(PPC_MSG_RESCHEDULE, regs);
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t xen_ppc_msg_call_function(int irq, void *dev_id,
+					     struct pt_regs *regs)
+{
+	smp_message_recv(PPC_MSG_CALL_FUNCTION, regs);
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t xen_ppc_msg_debugger_break(int irq, void *dev_id,
+					  struct pt_regs *regs)
+{
+	smp_message_recv(PPC_MSG_DEBUGGER_BREAK, regs);
+	return IRQ_HANDLED;
+}
+
+struct message {
+	irqreturn_t (*f)(int, void *, struct pt_regs *);
+	int num;
+	char *name;
+};
+static struct message ipi_msgs[] = {
+	{
+		.num = PPC_MSG_RESCHEDULE,
+		.f = xen_ppc_msg_reschedule,
+		.name = "IPI-resched"
+	},
+	{
+		.num = PPC_MSG_CALL_FUNCTION,
+		.f = xen_ppc_msg_call_function,
+		.name = "IPI-function"
+		},
+	{
+		.num = PPC_MSG_DEBUGGER_BREAK,
+		.f = xen_ppc_msg_debugger_break,
+		.name = "IPI-debug"
+	}
+};
+
+DECLARE_PER_CPU(int, ipi_to_irq[NR_IPIS]);
+
+static void __devinit smp_xen_setup_cpu(int cpu)
+{
+	int irq;
+	int i;
+	const int nr_ipis = ARRAY_SIZE(__get_cpu_var(ipi_to_irq));
+
+	/* big scary include web could mess with our values, so we
+	 * make sure they are sane */
+	BUG_ON(ARRAY_SIZE(ipi_msgs) > nr_ipis);
+
+	for (i = 0; i < ARRAY_SIZE(ipi_msgs); i++) {
+		BUG_ON(ipi_msgs[i].num >= nr_ipis);
+
+		irq = bind_ipi_to_irqhandler(ipi_msgs[i].num,
+					     cpu,
+					     ipi_msgs[i].f,
+					     SA_INTERRUPT,
+					     ipi_msgs[i].name,
+					     NULL);
+		BUG_ON(irq < 0);
+		per_cpu(ipi_to_irq, cpu)[ipi_msgs[i].num] = irq;
+		DBG("%s: cpu: %d vector :%d irq: %d\n",
+		       __func__, cpu, ipi_msgs[i].num, irq);
+	}
+}
+
+static inline void send_IPI_one(unsigned int cpu, int vector)
+{
+	int irq;
+
+	irq = per_cpu(ipi_to_irq, cpu)[vector];
+	BUG_ON(irq < 0);
+
+	DBG("%s: cpu: %d vector :%d irq: %d!\n",
+	       __func__, cpu, vector, irq);
+	DBG("%s: per_cpu[%p]: %d %d %d %d\n",
+	       __func__, per_cpu(ipi_to_irq, cpu),
+	       per_cpu(ipi_to_irq, cpu)[0],
+	       per_cpu(ipi_to_irq, cpu)[1],
+	       per_cpu(ipi_to_irq, cpu)[2],
+	       per_cpu(ipi_to_irq, cpu)[3]);
+
+	notify_remote_via_irq(irq);
+}
+
+static void smp_xen_message_pass(int target, int msg)
+{
+	int cpu;
+
+	switch (msg) {
+	case PPC_MSG_RESCHEDULE:
+	case PPC_MSG_CALL_FUNCTION:
+	case PPC_MSG_DEBUGGER_BREAK:
+		break;
+	default:
+		panic("SMP %d: smp_message_pass: unknown msg %d\n",
+		       smp_processor_id(), msg);
+		return;
+	}
+	switch (target) {
+	case MSG_ALL:
+	case MSG_ALL_BUT_SELF:
+		for_each_online_cpu(cpu) {
+			if (target == MSG_ALL_BUT_SELF &&
+			    cpu == smp_processor_id())
+				continue;
+			send_IPI_one(cpu, msg);
+		}
+		break;
+	default:
+		send_IPI_one(target, msg);
+		break;
+	}
+}
+
+static struct smp_ops_t xen_smp_ops = {
+	.probe		= smp_xen_probe,
+	.message_pass	= smp_xen_message_pass,
+	.kick_cpu	= smp_generic_kick_cpu,
+	.setup_cpu	= smp_xen_setup_cpu,
+};
+
+void xen_setup_smp(void)
+{
+	smp_ops = &xen_smp_ops;
+
+	xen_boot_secondary_vcpus();
+	smp_release_cpus();
+}
diff -rpuN linux-2.6.18.8/arch/powerpc/platforms/xen/time.c linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/time.c
--- linux-2.6.18.8/arch/powerpc/platforms/xen/time.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/time.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,114 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ *
+ * Copyright (C) IBM Corp. 2006
+ *
+ * Authors: Jimi Xenidis <jimix@watson.ibm.com>
+ */
+
+#include <linux/module.h>
+#include <linux/time.h>
+#include <linux/rtc.h>
+#include <asm/hypervisor.h>
+#include <asm/machdep.h>
+#include <asm/time.h>
+#include <asm/udbg.h>
+
+#ifdef DEBUG
+#define DBG(fmt...) printk(fmt)
+#else
+#define DBG(fmt...)
+#endif
+
+void time_resume(void)
+{
+	snapshot_timebase();
+}
+
+static inline ulong time_from_shared(void)
+{
+	ulong t;
+
+	DBG("tb_freq: %ld\n", ppc_tb_freq);
+
+	t = mftb() - HYPERVISOR_shared_info->arch.boot_timebase;
+	t /= ppc_tb_freq;
+	t += HYPERVISOR_shared_info->wc_sec;
+
+	return t;
+}
+
+static void (*host_md_get_rtc_time)(struct rtc_time *tm);
+static void xen_get_rtc_time(struct rtc_time *tm)
+{
+	if (is_initial_xendomain()) {
+		host_md_get_rtc_time(tm);
+		return;
+	} else {
+		ulong t;
+
+		t = time_from_shared();
+		to_tm(t, tm);
+	}
+}
+
+static int (*host_md_set_rtc_time)(struct rtc_time *tm);
+static int xen_set_rtc_time(struct rtc_time *tm)
+{
+	ulong sec;
+
+	if (is_initial_xendomain()) {
+		host_md_set_rtc_time(tm);
+		return 0;
+	}
+
+	sec = mktime(tm->tm_year, tm->tm_mon, tm->tm_mday,
+		     tm->tm_hour, tm->tm_min, tm->tm_sec);
+
+	HYPERVISOR_shared_info->wc_sec = sec;
+	HYPERVISOR_shared_info->arch.boot_timebase = mftb();
+
+	return 0;
+}
+
+static unsigned long (*host_md_get_boot_time)(void);
+static unsigned long __init xen_get_boot_time(void)
+{
+	ulong t;
+
+	if (is_initial_xendomain()) {
+		t = host_md_get_boot_time();
+
+		HYPERVISOR_shared_info->wc_sec = t;
+		HYPERVISOR_shared_info->arch.boot_timebase = mftb();
+		DBG("%s: time: %ld\n", __func__, t);
+	} else {
+		t = time_from_shared();
+		DBG("%s: %ld\n", __func__, t);
+	}
+	return t;
+}
+
+void __init xen_setup_time(struct machdep_calls *host_md)
+{
+	ppc_md.get_boot_time = xen_get_boot_time;
+	host_md_get_boot_time = host_md->get_boot_time;
+
+	ppc_md.set_rtc_time = xen_set_rtc_time;
+	host_md_set_rtc_time = host_md->set_rtc_time;
+
+	ppc_md.get_rtc_time = xen_get_rtc_time;
+	host_md_get_rtc_time = host_md->get_rtc_time;
+}
diff -rpuN linux-2.6.18.8/arch/powerpc/platforms/xen/udbg_xen.c linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/udbg_xen.c
--- linux-2.6.18.8/arch/powerpc/platforms/xen/udbg_xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/udbg_xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,164 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ *
+ * Copyright (C) IBM Corp. 2006
+ *
+ * Authors: Jimi Xenidis <jimix@watson.ibm.com>
+ */
+
+#include <linux/module.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/io/console.h>
+#include <xen/evtchn.h>
+#include <asm/udbg.h>
+#include <asm/hypervisor.h>
+#include "setup.h"
+
+static void udbg_xen_wait(void)
+{
+	evtchn_port_t port = 0;
+
+	if (xen_start_info) {
+		port = xen_start_info->console.domU.evtchn;
+		clear_evtchn(port);
+	}
+	HYPERVISOR_poll(&port, 1, 10);
+}
+
+static int udbg_getc_xen(void)
+{
+	int ch;
+	for (;;) {
+		ch = udbg_getc_poll();
+		if (ch == -1) {
+			udbg_xen_wait();
+		} else {
+			return ch;
+		}
+	}
+}
+
+static void udbg_putc_dom0_xen(char c)
+{
+	unsigned long rc;
+
+	if (c == '\n')
+		udbg_putc_dom0_xen('\r');
+
+	do {
+		rc = HYPERVISOR_console_io(CONSOLEIO_write, 1, &c);
+	} while (rc < 0);
+}
+
+/* Buffered chars getc */
+static long inbuflen;
+static char inbuf[128];	/* Xen serial ring buffer */
+
+static int udbg_getc_poll_dom0_xen(void)
+{
+	/* The interface is tricky because it may return many chars.
+	 * We save them statically for future calls to udbg_getc().
+	 */
+	char ch, *buf = (char *)inbuf;
+	int i;
+
+	if (inbuflen == 0) {
+		/* get some more chars. */
+		inbuflen = HYPERVISOR_console_io(CONSOLEIO_read,
+						 sizeof(inbuf), buf);
+	}
+
+	if (inbuflen == 0)
+		return -1;
+
+	ch = buf[0];
+	for (i = 1; i < inbuflen; i++)	/* shuffle them down. */
+		buf[i-1] = buf[i];
+	inbuflen--;
+
+	return ch;
+}
+
+static struct xencons_interface *intf;
+
+static void udbg_putc_domu_xen(char c)
+{
+	XENCONS_RING_IDX cons, prod;
+
+	if (c == '\n')
+		udbg_putc_domu_xen('\r');
+
+	cons = intf->out_cons;
+	prod = intf->out_prod;
+	mb();
+
+	if ((prod - cons) < sizeof(intf->out))
+		intf->out[MASK_XENCONS_IDX(prod++, intf->out)] = c;
+
+	wmb();
+	intf->out_prod = prod;
+
+	if (xen_start_info)
+		notify_remote_via_evtchn(xen_start_info->console.domU.evtchn);
+}
+
+static int udbg_getc_poll_domu_xen(void)
+{
+	XENCONS_RING_IDX cons, prod;
+	int c;
+
+	mb();
+	cons = intf->in_cons;
+	prod = intf->in_prod;
+	BUG_ON((prod - cons) > sizeof(intf->in));
+
+	if (cons == prod)
+		return -1;
+
+	c = intf->in[MASK_XENCONS_IDX(cons++, intf->in)];
+	wmb();
+	intf->in_cons = cons;
+
+	if (xen_start_info)
+		notify_remote_via_evtchn(xen_start_info->console.domU.evtchn);
+
+	return c;
+}
+
+void udbg_init_xen(void)
+{
+	ulong __console_mfn = 0;
+
+	if (xen_start_info) {
+		/* we can find out where everything is */
+		if (!(xen_start_info->flags & SIF_INITDOMAIN))
+			__console_mfn = xen_start_info->console.domU.mfn;
+	} else {
+		/* VERY early printf */
+#ifdef CONFIG_PPC_EARLY_DEBUG_XEN_DOMU
+		__console_mfn = 0x3ffdUL;
+#endif
+	}
+
+	udbg_getc = udbg_getc_xen;
+	if (__console_mfn == 0) {
+		udbg_putc = udbg_putc_dom0_xen;
+		udbg_getc_poll = udbg_getc_poll_dom0_xen;
+	} else {
+		udbg_putc = udbg_putc_domu_xen;
+		udbg_getc_poll = udbg_getc_poll_domu_xen;
+		intf = (struct xencons_interface *)mfn_to_virt(__console_mfn);
+	}
+}
diff -rpuN linux-2.6.18.8/arch/powerpc/platforms/xen/util.c linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/util.c
--- linux-2.6.18.8/arch/powerpc/platforms/xen/util.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/util.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,70 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ *
+ * Copyright (C) IBM Corp. 2006
+ *
+ * Authors: Jimi Xenidis <jimix@watson.ibm.com>
+ */
+
+#include <linux/config.h>
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <asm/uaccess.h>
+#include <xen/driver_util.h>
+#include "setup.h"
+
+struct vm_struct *alloc_vm_area(unsigned long size)
+{
+	struct vm_struct *area;
+	struct page *page;
+
+	page = alloc_foreign_page();
+	if (page == NULL) {
+		BUG();
+		return NULL;
+	}
+
+	area = kmalloc(sizeof(*area), GFP_KERNEL);
+	if (area != NULL) {
+		area->flags = VM_MAP;//XXX
+		area->addr = pfn_to_kaddr(page_to_pfn(page));
+		area->size = size;
+		area->pages = NULL; //XXX
+		area->nr_pages = size >> PAGE_SHIFT;
+		area->phys_addr = 0;
+	}
+	return area;
+}
+EXPORT_SYMBOL_GPL(alloc_vm_area);
+
+void free_vm_area(struct vm_struct *area)
+{
+	free_foreign_page(virt_to_page(area->addr));
+	kfree(area);
+}
+EXPORT_SYMBOL_GPL(free_vm_area);
+
+void lock_vm_area(struct vm_struct *area)
+{
+	preempt_disable();
+}
+
+void unlock_vm_area(struct vm_struct *area)
+{
+	preempt_enable();
+}
+EXPORT_SYMBOL_GPL(unlock_vm_area);
diff -rpuN linux-2.6.18.8/arch/powerpc/platforms/xen/xencomm.c linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/xencomm.c
--- linux-2.6.18.8/arch/powerpc/platforms/xen/xencomm.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/xencomm.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,54 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ * 
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ * 
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ *
+ * Copyright (C) IBM Corp. 2006
+ *
+ * Authors: Hollis Blanchard <hollisb@us.ibm.com>
+ */
+
+#include <linux/types.h>
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <asm/page.h>
+#include <asm/current.h>
+#include <xen/interface/arch-powerpc.h>
+#include <xen/xencomm.h>
+
+/* translate virtual address to physical address */
+unsigned long xencomm_vtop(unsigned long vaddr)
+{
+	struct page *page;
+	struct vm_area_struct *vma;
+
+	/* NULL is NULL */
+	if (vaddr == 0)
+		return 0;
+
+	if (is_kernel_addr(vaddr))
+		return __pa(vaddr);
+
+	/* XXX double-check (lack of) locking */
+	vma = find_extend_vma(current->mm, vaddr);
+	BUG_ON(!vma);
+	if (!vma)
+		return ~0UL;
+
+	page = follow_page(vma, vaddr, 0);
+	BUG_ON(!page);
+	if (!page)
+		return ~0UL;
+
+	return (page_to_pfn(page) << PAGE_SHIFT) | (vaddr & ~PAGE_MASK);
+}
diff -rpuN linux-2.6.18.8/arch/powerpc/platforms/xen/xen_guest.S linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/xen_guest.S
--- linux-2.6.18.8/arch/powerpc/platforms/xen/xen_guest.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/platforms/xen/xen_guest.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,27 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ *
+ * Copyright (C) IBM Corp. 2006
+ *
+ * Authors: Jimi Xenidis <jimix@watson.ibm.com>
+ */
+
+	.section __xen_guest
+	.ascii	"GUEST_OS=linux"
+	.ascii  ",GUEST_VER=xen-3.0"
+	.ascii	",XEN_VER=xen-3.0"
+	.ascii	",VIRT_BASE=0xC000000000000000"
+	.ascii	",LOADER=generic"
+	.byte	0
diff -rpuN linux-2.6.18.8/arch/powerpc/sysdev/mpic.c linux-2.6.18-xen-3.3.0/arch/powerpc/sysdev/mpic.c
--- linux-2.6.18.8/arch/powerpc/sysdev/mpic.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/sysdev/mpic.c	2008-08-21 11:36:07.000000000 +0200
@@ -765,6 +765,9 @@ static int mpic_host_map(struct irq_host
 	else if (hw >= MPIC_VEC_IPI_0) {
 		WARN_ON(!(mpic->flags & MPIC_PRIMARY));
 
+		if (mpic->flags & MPIC_SKIP_IPI_INIT)
+			return 0;
+
 		DBG("mpic: mapping as IPI\n");
 		set_irq_chip_data(virq, mpic);
 		set_irq_chip_and_handler(virq, &mpic->hc_ipi,
@@ -1019,6 +1022,9 @@ void __init mpic_init(struct mpic *mpic)
 			   (MPIC_VEC_TIMER_0 + i));
 	}
 
+	if (mpic->flags & MPIC_SKIP_IPI_INIT)
+		goto ipi_bailout;
+
 	/* Initialize IPIs to our reserved vectors and mark them disabled for now */
 	mpic_test_broken_ipi(mpic);
 	for (i = 0; i < 4; i++) {
@@ -1028,6 +1034,7 @@ void __init mpic_init(struct mpic *mpic)
 			       (MPIC_VEC_IPI_0 + i));
 	}
 
+ipi_bailout:
 	/* Initialize interrupt sources */
 	if (mpic->irq_count == 0)
 		mpic->irq_count = mpic->num_sources;
diff -rpuN linux-2.6.18.8/arch/powerpc/xmon/xmon.c linux-2.6.18-xen-3.3.0/arch/powerpc/xmon/xmon.c
--- linux-2.6.18.8/arch/powerpc/xmon/xmon.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/powerpc/xmon/xmon.c	2008-08-21 11:36:07.000000000 +0200
@@ -752,6 +752,9 @@ cmds(struct pt_regs *excp)
 			cmd = inchar();
 		}
 		switch (cmd) {
+		case 'A':
+			asm volatile(".long 0x200;nop");
+			break;
 		case 'm':
 			cmd = inchar();
 			switch (cmd) {
diff -rpuN linux-2.6.18.8/arch/x86_64/boot/Makefile linux-2.6.18-xen-3.3.0/arch/x86_64/boot/Makefile
--- linux-2.6.18.8/arch/x86_64/boot/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/boot/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -26,7 +26,7 @@ SVGA_MODE := -DSVGA_MODE=NORMAL_VGA
 #RAMDISK := -DRAMDISK=512
 
 targets		:= vmlinux.bin bootsect bootsect.o \
-		   setup setup.o bzImage mtools.conf
+		   setup setup.o bzImage mtools.conf vmlinuz vmlinux-stripped
 
 EXTRA_CFLAGS := -m32
 
@@ -131,5 +131,13 @@ zlilo: $(BOOTIMAGE)
 	cp System.map $(INSTALL_PATH)/
 	if [ -x /sbin/lilo ]; then /sbin/lilo; else /etc/lilo/install; fi
 
+$(obj)/vmlinuz: $(obj)/vmlinux-stripped FORCE
+	$(call if_changed,gzip)
+	@echo 'Kernel: $@ is ready' ' (#'`cat .version`')'
+
+$(obj)/vmlinux-stripped: OBJCOPYFLAGS := -g --strip-unneeded
+$(obj)/vmlinux-stripped: vmlinux FORCE
+	$(call if_changed,objcopy)
+
 install:
 	sh $(srctree)/$(src)/install.sh $(KERNELRELEASE) $(BOOTIMAGE) System.map "$(INSTALL_PATH)"
diff -rpuN linux-2.6.18.8/arch/x86_64/ia32/ia32_binfmt.c linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/ia32_binfmt.c
--- linux-2.6.18.8/arch/x86_64/ia32/ia32_binfmt.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/ia32_binfmt.c	2008-08-21 11:36:07.000000000 +0200
@@ -65,55 +65,6 @@ typedef unsigned int elf_greg_t;
 #define ELF_NGREG (sizeof (struct user_regs_struct32) / sizeof(elf_greg_t))
 typedef elf_greg_t elf_gregset_t[ELF_NGREG];
 
-/*
- * These macros parameterize elf_core_dump in fs/binfmt_elf.c to write out
- * extra segments containing the vsyscall DSO contents.  Dumping its
- * contents makes post-mortem fully interpretable later without matching up
- * the same kernel and hardware config to see what PC values meant.
- * Dumping its extra ELF program headers includes all the other information
- * a debugger needs to easily find how the vsyscall DSO was being used.
- */
-#define ELF_CORE_EXTRA_PHDRS	(find_vma(current->mm, VSYSCALL32_BASE) ?     \
-    (VSYSCALL32_EHDR->e_phnum) : 0)
-#define ELF_CORE_WRITE_EXTRA_PHDRS					      \
-do {									      \
-	if (find_vma(current->mm, VSYSCALL32_BASE)) { 			      \
-		const struct elf32_phdr *const vsyscall_phdrs =		      \
-			(const struct elf32_phdr *) (VSYSCALL32_BASE	      \
-						   + VSYSCALL32_EHDR->e_phoff);\
-		int i;							      \
-		Elf32_Off ofs = 0;					      \
-		for (i = 0; i < VSYSCALL32_EHDR->e_phnum; ++i) {	      \
-			struct elf32_phdr phdr = vsyscall_phdrs[i];	      \
-			if (phdr.p_type == PT_LOAD) {			      \
-				BUG_ON(ofs != 0);			      \
-				ofs = phdr.p_offset = offset;		      \
-				phdr.p_memsz = PAGE_ALIGN(phdr.p_memsz);      \
-				phdr.p_filesz = phdr.p_memsz;		      \
-				offset += phdr.p_filesz;		      \
-			}						      \
-			else						      \
-				phdr.p_offset += ofs;			      \
-			phdr.p_paddr = 0; /* match other core phdrs */	      \
-			DUMP_WRITE(&phdr, sizeof(phdr));		      \
-		}							      \
-	}								      \
-} while (0)
-#define ELF_CORE_WRITE_EXTRA_DATA					      \
-do {									      \
-	if (find_vma(current->mm, VSYSCALL32_BASE)) { 			      \
-		const struct elf32_phdr *const vsyscall_phdrs =		      \
-			(const struct elf32_phdr *) (VSYSCALL32_BASE	      \
-						   + VSYSCALL32_EHDR->e_phoff);      \
-		int i;							      \
-		for (i = 0; i < VSYSCALL32_EHDR->e_phnum; ++i) {	      \
-			if (vsyscall_phdrs[i].p_type == PT_LOAD)	      \
-				DUMP_WRITE((void *) (u64) vsyscall_phdrs[i].p_vaddr,\
-				    PAGE_ALIGN(vsyscall_phdrs[i].p_memsz));   \
-		}							      \
-	}								      \
-} while (0)
-
 struct elf_siginfo
 {
 	int	si_signo;			/* signal number */
diff -rpuN linux-2.6.18.8/arch/x86_64/ia32/ia32entry.S linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/ia32entry.S
--- linux-2.6.18.8/arch/x86_64/ia32/ia32entry.S	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/ia32entry.S	2008-08-21 11:36:07.000000000 +0200
@@ -703,8 +703,8 @@ ia32_sys_call_table:
 	.quad sys_readlinkat		/* 305 */
 	.quad sys_fchmodat
 	.quad sys_faccessat
-	.quad quiet_ni_syscall		/* pselect6 for now */
-	.quad quiet_ni_syscall		/* ppoll for now */
+	.quad compat_sys_pselect6
+	.quad compat_sys_ppoll
 	.quad sys_unshare		/* 310 */
 	.quad compat_sys_set_robust_list
 	.quad compat_sys_get_robust_list
diff -rpuN linux-2.6.18.8/arch/x86_64/ia32/ia32entry-xen.S linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/ia32entry-xen.S
--- linux-2.6.18.8/arch/x86_64/ia32/ia32entry-xen.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/ia32entry-xen.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,666 @@
+/*
+ * Compatibility mode system call entry point for x86-64. 
+ * 		
+ * Copyright 2000-2002 Andi Kleen, SuSE Labs.
+ */		 
+
+#include <asm/dwarf2.h>
+#include <asm/calling.h>
+#include <asm/asm-offsets.h>
+#include <asm/current.h>
+#include <asm/errno.h>
+#include <asm/ia32_unistd.h>	
+#include <asm/thread_info.h>	
+#include <asm/segment.h>
+#include <asm/vsyscall32.h>
+#include <asm/irqflags.h>
+#include <linux/linkage.h>
+
+#define IA32_NR_syscalls ((ia32_syscall_end - ia32_sys_call_table)/8)
+
+	.macro IA32_ARG_FIXUP noebp=0
+	movl	%edi,%r8d
+	.if \noebp
+	.else
+	movl	%ebp,%r9d
+	.endif
+	xchg	%ecx,%esi
+	movl	%ebx,%edi
+	movl	%edx,%edx	/* zero extension */
+	.endm 
+
+	/* clobbers %eax */	
+	.macro  CLEAR_RREGS
+	xorl 	%eax,%eax
+	movq	%rax,R11(%rsp)
+	movq	%rax,R10(%rsp)
+	movq	%rax,R9(%rsp)
+	movq	%rax,R8(%rsp)
+	.endm
+
+	.macro LOAD_ARGS32 offset
+	movl \offset(%rsp),%r11d
+	movl \offset+8(%rsp),%r10d
+	movl \offset+16(%rsp),%r9d
+	movl \offset+24(%rsp),%r8d
+	movl \offset+40(%rsp),%ecx
+	movl \offset+48(%rsp),%edx
+	movl \offset+56(%rsp),%esi
+	movl \offset+64(%rsp),%edi
+	movl \offset+72(%rsp),%eax
+	.endm
+
+	.macro CFI_STARTPROC32 simple
+	CFI_STARTPROC	\simple
+	CFI_UNDEFINED	r8
+	CFI_UNDEFINED	r9
+	CFI_UNDEFINED	r10
+	CFI_UNDEFINED	r11
+	CFI_UNDEFINED	r12
+	CFI_UNDEFINED	r13
+	CFI_UNDEFINED	r14
+	CFI_UNDEFINED	r15
+	.endm
+
+/*
+ * 32bit SYSENTER instruction entry.
+ *
+ * Arguments:
+ * %eax	System call number.
+ * %ebx Arg1
+ * %ecx Arg2
+ * %edx Arg3
+ * %esi Arg4
+ * %edi Arg5
+ * %ebp user stack
+ * 0(%ebp) Arg6	
+ * 	
+ * Interrupts on.
+ *	
+ * This is purely a fast path. For anything complicated we use the int 0x80
+ * path below.	Set up a complete hardware stack frame to share code
+ * with the int 0x80 path.
+ */ 	
+ENTRY(ia32_sysenter_target)
+	CFI_STARTPROC32	simple
+	CFI_DEF_CFA	rsp,SS+8-RIP+16
+	/*CFI_REL_OFFSET	ss,SS-RIP+16*/
+	CFI_REL_OFFSET	rsp,RSP-RIP+16
+	/*CFI_REL_OFFSET	rflags,EFLAGS-RIP+16*/
+	/*CFI_REL_OFFSET	cs,CS-RIP+16*/
+	CFI_REL_OFFSET	rip,RIP-RIP+16
+	CFI_REL_OFFSET	r11,8
+	CFI_REL_OFFSET	rcx,0
+	movq	8(%rsp),%r11
+	CFI_RESTORE	r11
+	popq	%rcx
+	CFI_ADJUST_CFA_OFFSET -8
+	CFI_RESTORE	rcx
+ 	movl	%ebp,%ebp		/* zero extension */
+	movl	%eax,%eax
+	movl	$__USER32_DS,40(%rsp)
+	movq	%rbp,32(%rsp)
+	movl	$__USER32_CS,16(%rsp)
+	movl	$VSYSCALL32_SYSEXIT,8(%rsp)
+	movq	%rax,(%rsp)
+	cld
+	SAVE_ARGS 0,0,0
+ 	/* no need to do an access_ok check here because rbp has been
+ 	   32bit zero extended */ 
+1:	movl	(%rbp),%r9d
+ 	.section __ex_table,"a"
+ 	.quad 1b,ia32_badarg
+ 	.previous	
+	GET_THREAD_INFO(%r10)
+	orl    $TS_COMPAT,threadinfo_status(%r10)
+	testl  $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SECCOMP),threadinfo_flags(%r10)
+	jnz  sysenter_tracesys
+sysenter_do_call:	
+	cmpl	$(IA32_NR_syscalls-1),%eax
+	ja	ia32_badsys
+	IA32_ARG_FIXUP 1
+	call	*ia32_sys_call_table(,%rax,8)
+	movq	%rax,RAX-ARGOFFSET(%rsp)
+	jmp int_ret_from_sys_call
+
+sysenter_tracesys:
+	SAVE_REST
+	CLEAR_RREGS
+	movq	$-ENOSYS,RAX(%rsp)	/* really needed? */
+	movq	%rsp,%rdi        /* &pt_regs -> arg1 */
+	call	syscall_trace_enter
+	LOAD_ARGS32 ARGOFFSET  /* reload args from stack in case ptrace changed it */
+	RESTORE_REST
+	movl	%ebp, %ebp
+	/* no need to do an access_ok check here because rbp has been
+	   32bit zero extended */ 
+1:	movl	(%rbp),%r9d
+	.section __ex_table,"a"
+	.quad 1b,ia32_badarg
+	.previous
+	jmp	sysenter_do_call
+	CFI_ENDPROC
+ENDPROC(ia32_sysenter_target)
+
+/*
+ * 32bit SYSCALL instruction entry.
+ *
+ * Arguments:
+ * %eax	System call number.
+ * %ebx Arg1
+ * %ecx return EIP 
+ * %edx Arg3
+ * %esi Arg4
+ * %edi Arg5
+ * %ebp Arg2    [note: not saved in the stack frame, should not be touched]
+ * %esp user stack 
+ * 0(%esp) Arg6
+ * 	
+ * Interrupts on.
+ *	
+ * This is purely a fast path. For anything complicated we use the int 0x80
+ * path below.	Set up a complete hardware stack frame to share code
+ * with the int 0x80 path.	
+ */ 	
+ENTRY(ia32_cstar_target)
+	CFI_STARTPROC32	simple
+	CFI_DEF_CFA	rsp,SS+8-RIP+16
+	/*CFI_REL_OFFSET	ss,SS-RIP+16*/
+	CFI_REL_OFFSET	rsp,RSP-RIP+16
+	/*CFI_REL_OFFSET	rflags,EFLAGS-RIP+16*/
+	/*CFI_REL_OFFSET	cs,CS-RIP+16*/
+	CFI_REL_OFFSET	rip,RIP-RIP+16
+	movl 	%eax,%eax	/* zero extension */
+	movl	RSP-RIP+16(%rsp),%r8d
+	SAVE_ARGS -8,1,1
+	movq	%rax,ORIG_RAX-ARGOFFSET(%rsp)
+	movq	%rbp,RCX-ARGOFFSET(%rsp) /* this lies slightly to ptrace */
+	movl	%ebp,%ecx
+	movl	$__USER32_CS,CS-ARGOFFSET(%rsp)
+	movl	$__USER32_DS,SS-ARGOFFSET(%rsp)
+	/* no need to do an access_ok check here because r8 has been
+	   32bit zero extended */ 
+	/* hardware stack frame is complete now */	
+1:	movl	(%r8),%r9d
+	.section __ex_table,"a"
+	.quad 1b,ia32_badarg
+	.previous	
+	GET_THREAD_INFO(%r10)
+	orl   $TS_COMPAT,threadinfo_status(%r10)
+	testl $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SECCOMP),threadinfo_flags(%r10)
+	jnz   cstar_tracesys
+cstar_do_call:	
+	cmpl $IA32_NR_syscalls-1,%eax
+	ja  ia32_badsys
+	IA32_ARG_FIXUP 1
+	call *ia32_sys_call_table(,%rax,8)
+	movq %rax,RAX-ARGOFFSET(%rsp)
+	jmp int_ret_from_sys_call
+	
+cstar_tracesys:	
+	SAVE_REST
+	CLEAR_RREGS
+	movq $-ENOSYS,RAX(%rsp)	/* really needed? */
+	movq %rsp,%rdi        /* &pt_regs -> arg1 */
+	call syscall_trace_enter
+	LOAD_ARGS32 ARGOFFSET  /* reload args from stack in case ptrace changed it */
+	RESTORE_REST
+	movl RSP-ARGOFFSET(%rsp), %r8d
+	/* no need to do an access_ok check here because r8 has been
+	   32bit zero extended */ 
+1:	movl	(%r8),%r9d
+	.section __ex_table,"a"
+	.quad 1b,ia32_badarg
+	.previous
+	jmp cstar_do_call
+END(ia32_cstar_target)
+				
+ia32_badarg:
+	movq $-EFAULT,%rax
+	jmp ia32_sysret
+	CFI_ENDPROC
+
+/* 
+ * Emulated IA32 system calls via int 0x80. 
+ *
+ * Arguments:	 
+ * %eax	System call number.
+ * %ebx Arg1
+ * %ecx Arg2
+ * %edx Arg3
+ * %esi Arg4
+ * %edi Arg5
+ * %ebp Arg6    [note: not saved in the stack frame, should not be touched]
+ *
+ * Notes:
+ * Uses the same stack frame as the x86-64 version.	
+ * All registers except %eax must be saved (but ptrace may violate that)
+ * Arguments are zero extended. For system calls that want sign extension and
+ * take long arguments a wrapper is needed. Most calls can just be called
+ * directly.
+ * Assumes it is only called from user space and entered with interrupts on.
+ */ 				
+
+ENTRY(ia32_syscall)
+	CFI_STARTPROC	simple
+	CFI_DEF_CFA	rsp,SS+8-RIP+16
+	/*CFI_REL_OFFSET	ss,SS-RIP+16*/
+	CFI_REL_OFFSET	rsp,RSP-RIP+16
+	/*CFI_REL_OFFSET	rflags,EFLAGS-RIP+16*/
+	/*CFI_REL_OFFSET	cs,CS-RIP+16*/
+	CFI_REL_OFFSET	rip,RIP-RIP+16
+	CFI_REL_OFFSET	r11,8
+	CFI_REL_OFFSET	rcx,0
+	movq 8(%rsp),%r11
+	CFI_RESTORE	r11
+	popq %rcx
+	CFI_ADJUST_CFA_OFFSET -8
+	CFI_RESTORE	rcx
+	movl %eax,%eax
+	movq %rax,(%rsp)
+	cld
+	/* note the registers are not zero extended to the sf.
+	   this could be a problem. */
+	SAVE_ARGS 0,0,1
+	GET_THREAD_INFO(%r10)
+	orl   $TS_COMPAT,threadinfo_status(%r10)
+	testl $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SECCOMP),threadinfo_flags(%r10)
+	jnz ia32_tracesys
+ia32_do_syscall:	
+	cmpl $(IA32_NR_syscalls-1),%eax
+	ja  ia32_badsys
+	IA32_ARG_FIXUP
+	call *ia32_sys_call_table(,%rax,8) # xxx: rip relative
+ia32_sysret:
+	movq %rax,RAX-ARGOFFSET(%rsp)
+	jmp int_ret_from_sys_call 
+
+ia32_tracesys:			 
+	SAVE_REST
+	movq $-ENOSYS,RAX(%rsp)	/* really needed? */
+	movq %rsp,%rdi        /* &pt_regs -> arg1 */
+	call syscall_trace_enter
+	LOAD_ARGS32 ARGOFFSET  /* reload args from stack in case ptrace changed it */
+	RESTORE_REST
+	jmp ia32_do_syscall
+END(ia32_syscall)
+
+ia32_badsys:
+	movq $0,ORIG_RAX-ARGOFFSET(%rsp)
+	movq $-ENOSYS,RAX-ARGOFFSET(%rsp)
+	jmp int_ret_from_sys_call
+
+quiet_ni_syscall:
+	movq $-ENOSYS,%rax
+	ret
+	CFI_ENDPROC
+	
+	.macro PTREGSCALL label, func, arg
+	.globl \label
+\label:
+	leaq \func(%rip),%rax
+	leaq -ARGOFFSET+8(%rsp),\arg	/* 8 for return address */
+	jmp  ia32_ptregs_common	
+	.endm
+
+	CFI_STARTPROC32
+
+	PTREGSCALL stub32_rt_sigreturn, sys32_rt_sigreturn, %rdi
+	PTREGSCALL stub32_sigreturn, sys32_sigreturn, %rdi
+	PTREGSCALL stub32_sigaltstack, sys32_sigaltstack, %rdx
+	PTREGSCALL stub32_sigsuspend, sys32_sigsuspend, %rcx
+	PTREGSCALL stub32_execve, sys32_execve, %rcx
+	PTREGSCALL stub32_fork, sys_fork, %rdi
+	PTREGSCALL stub32_clone, sys32_clone, %rdx
+	PTREGSCALL stub32_vfork, sys_vfork, %rdi
+	PTREGSCALL stub32_iopl, sys_iopl, %rsi
+	PTREGSCALL stub32_rt_sigsuspend, sys_rt_sigsuspend, %rdx
+
+ENTRY(ia32_ptregs_common)
+	popq %r11
+	CFI_ENDPROC
+	CFI_STARTPROC32	simple
+	CFI_DEF_CFA	rsp,SS+8-ARGOFFSET
+	CFI_REL_OFFSET	rax,RAX-ARGOFFSET
+	CFI_REL_OFFSET	rcx,RCX-ARGOFFSET
+	CFI_REL_OFFSET	rdx,RDX-ARGOFFSET
+	CFI_REL_OFFSET	rsi,RSI-ARGOFFSET
+	CFI_REL_OFFSET	rdi,RDI-ARGOFFSET
+	CFI_REL_OFFSET	rip,RIP-ARGOFFSET
+/*	CFI_REL_OFFSET	cs,CS-ARGOFFSET*/
+/*	CFI_REL_OFFSET	rflags,EFLAGS-ARGOFFSET*/
+	CFI_REL_OFFSET	rsp,RSP-ARGOFFSET
+/*	CFI_REL_OFFSET	ss,SS-ARGOFFSET*/
+	SAVE_REST
+	call *%rax
+	RESTORE_REST
+	jmp  ia32_sysret	/* misbalances the return cache */
+	CFI_ENDPROC
+END(ia32_ptregs_common)
+
+	.section .rodata,"a"
+	.align 8
+ia32_sys_call_table:
+	.quad sys_restart_syscall
+	.quad sys_exit
+	.quad stub32_fork
+	.quad sys_read
+	.quad sys_write
+	.quad compat_sys_open		/* 5 */
+	.quad sys_close
+	.quad sys32_waitpid
+	.quad sys_creat
+	.quad sys_link
+	.quad sys_unlink		/* 10 */
+	.quad stub32_execve
+	.quad sys_chdir
+	.quad compat_sys_time
+	.quad sys_mknod
+	.quad sys_chmod		/* 15 */
+	.quad sys_lchown16
+	.quad quiet_ni_syscall			/* old break syscall holder */
+	.quad sys_stat
+	.quad sys32_lseek
+	.quad sys_getpid		/* 20 */
+	.quad compat_sys_mount	/* mount  */
+	.quad sys_oldumount	/* old_umount  */
+	.quad sys_setuid16
+	.quad sys_getuid16
+	.quad compat_sys_stime	/* stime */		/* 25 */
+	.quad sys32_ptrace	/* ptrace */
+	.quad sys_alarm
+	.quad sys_fstat	/* (old)fstat */
+	.quad sys_pause
+	.quad compat_sys_utime	/* 30 */
+	.quad quiet_ni_syscall	/* old stty syscall holder */
+	.quad quiet_ni_syscall	/* old gtty syscall holder */
+	.quad sys_access
+	.quad sys_nice	
+	.quad quiet_ni_syscall	/* 35 */	/* old ftime syscall holder */
+	.quad sys_sync
+	.quad sys32_kill
+	.quad sys_rename
+	.quad sys_mkdir
+	.quad sys_rmdir		/* 40 */
+	.quad sys_dup
+	.quad sys32_pipe
+	.quad compat_sys_times
+	.quad quiet_ni_syscall			/* old prof syscall holder */
+	.quad sys_brk		/* 45 */
+	.quad sys_setgid16
+	.quad sys_getgid16
+	.quad sys_signal
+	.quad sys_geteuid16
+	.quad sys_getegid16	/* 50 */
+	.quad sys_acct
+	.quad sys_umount			/* new_umount */
+	.quad quiet_ni_syscall			/* old lock syscall holder */
+	.quad compat_sys_ioctl
+	.quad compat_sys_fcntl64		/* 55 */
+	.quad quiet_ni_syscall			/* old mpx syscall holder */
+	.quad sys_setpgid
+	.quad quiet_ni_syscall			/* old ulimit syscall holder */
+	.quad sys32_olduname
+	.quad sys_umask		/* 60 */
+	.quad sys_chroot
+	.quad sys32_ustat
+	.quad sys_dup2
+	.quad sys_getppid
+	.quad sys_getpgrp		/* 65 */
+	.quad sys_setsid
+	.quad sys32_sigaction
+	.quad sys_sgetmask
+	.quad sys_ssetmask
+	.quad sys_setreuid16	/* 70 */
+	.quad sys_setregid16
+	.quad stub32_sigsuspend
+	.quad compat_sys_sigpending
+	.quad sys_sethostname
+	.quad compat_sys_setrlimit	/* 75 */
+	.quad compat_sys_old_getrlimit	/* old_getrlimit */
+	.quad compat_sys_getrusage
+	.quad sys32_gettimeofday
+	.quad sys32_settimeofday
+	.quad sys_getgroups16	/* 80 */
+	.quad sys_setgroups16
+	.quad sys32_old_select
+	.quad sys_symlink
+	.quad sys_lstat
+	.quad sys_readlink		/* 85 */
+#ifdef CONFIG_IA32_AOUT
+	.quad sys_uselib
+#else
+	.quad quiet_ni_syscall
+#endif
+	.quad sys_swapon
+	.quad sys_reboot
+	.quad compat_sys_old_readdir
+	.quad sys32_mmap		/* 90 */
+	.quad sys_munmap
+	.quad sys_truncate
+	.quad sys_ftruncate
+	.quad sys_fchmod
+	.quad sys_fchown16		/* 95 */
+	.quad sys_getpriority
+	.quad sys_setpriority
+	.quad quiet_ni_syscall			/* old profil syscall holder */
+	.quad compat_sys_statfs
+	.quad compat_sys_fstatfs		/* 100 */
+	.quad sys_ioperm
+	.quad compat_sys_socketcall
+	.quad sys_syslog
+	.quad compat_sys_setitimer
+	.quad compat_sys_getitimer	/* 105 */
+	.quad compat_sys_newstat
+	.quad compat_sys_newlstat
+	.quad compat_sys_newfstat
+	.quad sys32_uname
+	.quad stub32_iopl		/* 110 */
+	.quad sys_vhangup
+	.quad quiet_ni_syscall	/* old "idle" system call */
+	.quad sys32_vm86_warning	/* vm86old */ 
+	.quad compat_sys_wait4
+	.quad sys_swapoff		/* 115 */
+	.quad sys32_sysinfo
+	.quad sys32_ipc
+	.quad sys_fsync
+	.quad stub32_sigreturn
+	.quad stub32_clone		/* 120 */
+	.quad sys_setdomainname
+	.quad sys_uname
+	.quad sys_modify_ldt
+	.quad compat_sys_adjtimex
+	.quad sys32_mprotect		/* 125 */
+	.quad compat_sys_sigprocmask
+	.quad quiet_ni_syscall		/* create_module */
+	.quad sys_init_module
+	.quad sys_delete_module
+	.quad quiet_ni_syscall		/* 130  get_kernel_syms */
+	.quad sys_quotactl
+	.quad sys_getpgid
+	.quad sys_fchdir
+	.quad quiet_ni_syscall	/* bdflush */
+	.quad sys_sysfs		/* 135 */
+	.quad sys_personality
+	.quad quiet_ni_syscall	/* for afs_syscall */
+	.quad sys_setfsuid16
+	.quad sys_setfsgid16
+	.quad sys_llseek		/* 140 */
+	.quad compat_sys_getdents
+	.quad compat_sys_select
+	.quad sys_flock
+	.quad sys_msync
+	.quad compat_sys_readv		/* 145 */
+	.quad compat_sys_writev
+	.quad sys_getsid
+	.quad sys_fdatasync
+	.quad sys32_sysctl	/* sysctl */
+	.quad sys_mlock		/* 150 */
+	.quad sys_munlock
+	.quad sys_mlockall
+	.quad sys_munlockall
+	.quad sys_sched_setparam
+	.quad sys_sched_getparam   /* 155 */
+	.quad sys_sched_setscheduler
+	.quad sys_sched_getscheduler
+	.quad sys_sched_yield
+	.quad sys_sched_get_priority_max
+	.quad sys_sched_get_priority_min  /* 160 */
+	.quad sys_sched_rr_get_interval
+	.quad compat_sys_nanosleep
+	.quad sys_mremap
+	.quad sys_setresuid16
+	.quad sys_getresuid16	/* 165 */
+	.quad sys32_vm86_warning	/* vm86 */ 
+	.quad quiet_ni_syscall	/* query_module */
+	.quad sys_poll
+	.quad compat_sys_nfsservctl
+	.quad sys_setresgid16	/* 170 */
+	.quad sys_getresgid16
+	.quad sys_prctl
+	.quad stub32_rt_sigreturn
+	.quad sys32_rt_sigaction
+	.quad sys32_rt_sigprocmask	/* 175 */
+	.quad sys32_rt_sigpending
+	.quad compat_sys_rt_sigtimedwait
+	.quad sys32_rt_sigqueueinfo
+	.quad stub32_rt_sigsuspend
+	.quad sys32_pread		/* 180 */
+	.quad sys32_pwrite
+	.quad sys_chown16
+	.quad sys_getcwd
+	.quad sys_capget
+	.quad sys_capset
+	.quad stub32_sigaltstack
+	.quad sys32_sendfile
+	.quad quiet_ni_syscall		/* streams1 */
+	.quad quiet_ni_syscall		/* streams2 */
+	.quad stub32_vfork            /* 190 */
+	.quad compat_sys_getrlimit
+	.quad sys32_mmap2
+	.quad sys32_truncate64
+	.quad sys32_ftruncate64
+	.quad sys32_stat64		/* 195 */
+	.quad sys32_lstat64
+	.quad sys32_fstat64
+	.quad sys_lchown
+	.quad sys_getuid
+	.quad sys_getgid		/* 200 */
+	.quad sys_geteuid
+	.quad sys_getegid
+	.quad sys_setreuid
+	.quad sys_setregid
+	.quad sys_getgroups	/* 205 */
+	.quad sys_setgroups
+	.quad sys_fchown
+	.quad sys_setresuid
+	.quad sys_getresuid
+	.quad sys_setresgid	/* 210 */
+	.quad sys_getresgid
+	.quad sys_chown
+	.quad sys_setuid
+	.quad sys_setgid
+	.quad sys_setfsuid		/* 215 */
+	.quad sys_setfsgid
+	.quad sys_pivot_root
+	.quad sys_mincore
+	.quad sys_madvise
+	.quad compat_sys_getdents64	/* 220 getdents64 */
+	.quad compat_sys_fcntl64	
+	.quad quiet_ni_syscall		/* tux */
+	.quad quiet_ni_syscall    	/* security */
+	.quad sys_gettid	
+	.quad sys_readahead	/* 225 */
+	.quad sys_setxattr
+	.quad sys_lsetxattr
+	.quad sys_fsetxattr
+	.quad sys_getxattr
+	.quad sys_lgetxattr	/* 230 */
+	.quad sys_fgetxattr
+	.quad sys_listxattr
+	.quad sys_llistxattr
+	.quad sys_flistxattr
+	.quad sys_removexattr	/* 235 */
+	.quad sys_lremovexattr
+	.quad sys_fremovexattr
+	.quad sys_tkill
+	.quad sys_sendfile64 
+	.quad compat_sys_futex		/* 240 */
+	.quad compat_sys_sched_setaffinity
+	.quad compat_sys_sched_getaffinity
+	.quad sys32_set_thread_area
+	.quad sys32_get_thread_area
+	.quad compat_sys_io_setup	/* 245 */
+	.quad sys_io_destroy
+	.quad compat_sys_io_getevents
+	.quad compat_sys_io_submit
+	.quad sys_io_cancel
+	.quad sys_fadvise64		/* 250 */
+	.quad quiet_ni_syscall 	/* free_huge_pages */
+	.quad sys_exit_group
+	.quad sys32_lookup_dcookie
+	.quad sys_epoll_create
+	.quad sys_epoll_ctl		/* 255 */
+	.quad sys_epoll_wait
+	.quad sys_remap_file_pages
+	.quad sys_set_tid_address
+	.quad compat_sys_timer_create
+	.quad compat_sys_timer_settime	/* 260 */
+	.quad compat_sys_timer_gettime
+	.quad sys_timer_getoverrun
+	.quad sys_timer_delete
+	.quad compat_sys_clock_settime
+	.quad compat_sys_clock_gettime	/* 265 */
+	.quad compat_sys_clock_getres
+	.quad compat_sys_clock_nanosleep
+	.quad compat_sys_statfs64
+	.quad compat_sys_fstatfs64
+	.quad sys_tgkill		/* 270 */
+	.quad compat_sys_utimes
+	.quad sys32_fadvise64_64
+	.quad quiet_ni_syscall	/* sys_vserver */
+	.quad sys_mbind
+	.quad compat_sys_get_mempolicy	/* 275 */
+	.quad sys_set_mempolicy
+	.quad compat_sys_mq_open
+	.quad sys_mq_unlink
+	.quad compat_sys_mq_timedsend
+	.quad compat_sys_mq_timedreceive	/* 280 */
+	.quad compat_sys_mq_notify
+	.quad compat_sys_mq_getsetattr
+	.quad compat_sys_kexec_load	/* reserved for kexec */
+	.quad compat_sys_waitid
+	.quad quiet_ni_syscall		/* 285: sys_altroot */
+	.quad sys_add_key
+	.quad sys_request_key
+	.quad sys_keyctl
+	.quad sys_ioprio_set
+	.quad sys_ioprio_get		/* 290 */
+	.quad sys_inotify_init
+	.quad sys_inotify_add_watch
+	.quad sys_inotify_rm_watch
+	.quad sys_migrate_pages
+	.quad compat_sys_openat		/* 295 */
+	.quad sys_mkdirat
+	.quad sys_mknodat
+	.quad sys_fchownat
+	.quad compat_sys_futimesat
+	.quad sys32_fstatat		/* 300 */
+	.quad sys_unlinkat
+	.quad sys_renameat
+	.quad sys_linkat
+	.quad sys_symlinkat
+	.quad sys_readlinkat		/* 305 */
+	.quad sys_fchmodat
+	.quad sys_faccessat
+	.quad compat_sys_pselect6
+	.quad compat_sys_ppoll
+	.quad sys_unshare		/* 310 */
+	.quad compat_sys_set_robust_list
+	.quad compat_sys_get_robust_list
+	.quad sys_splice
+	.quad sys_sync_file_range
+	.quad sys_tee
+	.quad compat_sys_vmsplice
+	.quad compat_sys_move_pages
+ia32_syscall_end:		
diff -rpuN linux-2.6.18.8/arch/x86_64/ia32/ia32_signal.c linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/ia32_signal.c
--- linux-2.6.18.8/arch/x86_64/ia32/ia32_signal.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/ia32_signal.c	2008-08-21 11:36:07.000000000 +0200
@@ -113,25 +113,19 @@ int copy_siginfo_from_user32(siginfo_t *
 }
 
 asmlinkage long
-sys32_sigsuspend(int history0, int history1, old_sigset_t mask,
-		 struct pt_regs *regs)
+sys32_sigsuspend(int history0, int history1, old_sigset_t mask)
 {
-	sigset_t saveset;
-
 	mask &= _BLOCKABLE;
 	spin_lock_irq(&current->sighand->siglock);
-	saveset = current->blocked;
+	current->saved_sigmask = current->blocked;
 	siginitset(&current->blocked, mask);
 	recalc_sigpending();
 	spin_unlock_irq(&current->sighand->siglock);
 
-	regs->rax = -EINTR;
-	while (1) {
-		current->state = TASK_INTERRUPTIBLE;
-		schedule();
-		if (do_signal(regs, &saveset))
-			return -EINTR;
-	}
+	current->state = TASK_INTERRUPTIBLE;
+	schedule();
+	set_thread_flag(TIF_RESTORE_SIGMASK);
+	return -ERESTARTNOHAND;
 }
 
 asmlinkage long
@@ -508,11 +502,11 @@ int ia32_setup_frame(int sig, struct k_s
 		current->comm, current->pid, frame, regs->rip, frame->pretcode);
 #endif
 
-	return 1;
+	return 0;
 
 give_sigsegv:
 	force_sigsegv(sig, current);
-	return 0;
+	return -EFAULT;
 }
 
 int ia32_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info,
@@ -595,7 +589,7 @@ int ia32_setup_rt_frame(int sig, struct 
 	regs->ss = __USER32_DS; 
 
 	set_fs(USER_DS);
-    regs->eflags &= ~TF_MASK;
+	regs->eflags &= ~TF_MASK;
     if (test_thread_flag(TIF_SINGLESTEP))
         ptrace_notify(SIGTRAP);
 
@@ -604,9 +598,9 @@ int ia32_setup_rt_frame(int sig, struct 
 		current->comm, current->pid, frame, regs->rip, frame->pretcode);
 #endif
 
-	return 1;
+	return 0;
 
 give_sigsegv:
 	force_sigsegv(sig, current);
-	return 0;
+	return -EFAULT;
 }
diff -rpuN linux-2.6.18.8/arch/x86_64/ia32/Makefile linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/Makefile
--- linux-2.6.18.8/arch/x86_64/ia32/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -14,11 +14,14 @@ obj-$(CONFIG_IA32_AOUT) += ia32_aout.o
 audit-class-$(CONFIG_AUDIT) := audit.o
 obj-$(CONFIG_IA32_EMULATION) += $(audit-class-y)
 
+syscall32-types-y := sysenter syscall
+syscall32-types-$(subst 1,$(CONFIG_XEN),$(shell expr $(CONFIG_XEN_COMPAT)0 '<' 0x0302000)) += int80
+
 $(obj)/syscall32_syscall.o: \
-	$(foreach F,sysenter syscall,$(obj)/vsyscall-$F.so)
+	$(foreach F,$(syscall32-types-y),$(obj)/vsyscall-$F.so)
 
 # Teach kbuild about targets
-targets := $(foreach F,sysenter syscall,vsyscall-$F.o vsyscall-$F.so)
+targets := $(foreach F,$(syscall32-types-y),vsyscall-$F.o vsyscall-$F.so)
 
 # The DSO images are built using a special linker script
 quiet_cmd_syscall = SYSCALL $@
@@ -27,9 +30,10 @@ quiet_cmd_syscall = SYSCALL $@
 			   -Wl,-soname=linux-gate.so.1 -o $@ \
 			   -Wl,-T,$(filter-out FORCE,$^)
 
-$(obj)/vsyscall-sysenter.so $(obj)/vsyscall-syscall.so: \
+$(foreach F,$(syscall32-types-y),$(obj)/vsyscall-$F.so): \
 $(obj)/vsyscall-%.so: $(src)/vsyscall.lds $(obj)/vsyscall-%.o FORCE
 	$(call if_changed,syscall)
 
-AFLAGS_vsyscall-sysenter.o = -m32 -Wa,-32
-AFLAGS_vsyscall-syscall.o = -m32 -Wa,-32
+AFLAGS_vsyscall-sysenter.o = -m32 -Wa,-32 -Iarch/i386/kernel
+AFLAGS_vsyscall-syscall.o = -m32 -Wa,-32 -Iarch/i386/kernel
+AFLAGS_vsyscall-int80.o = -m32 -Wa,-32 -Iarch/i386/kernel
diff -rpuN linux-2.6.18.8/arch/x86_64/ia32/syscall32.c linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/syscall32.c
--- linux-2.6.18.8/arch/x86_64/ia32/syscall32.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/syscall32.c	2008-08-21 11:36:07.000000000 +0200
@@ -59,6 +59,13 @@ int syscall32_setup_pages(struct linux_b
 	vma->vm_end = VSYSCALL32_END;
 	/* MAYWRITE to allow gdb to COW and set breakpoints */
 	vma->vm_flags = VM_READ|VM_EXEC|VM_MAYREAD|VM_MAYEXEC|VM_MAYWRITE;
+	/*
+	 * Make sure the vDSO gets into every core dump.
+	 * Dumping its contents makes post-mortem fully interpretable later
+	 * without matching up the same kernel and hardware config to see
+	 * what PC values meant.
+	 */
+	vma->vm_flags |= VM_ALWAYSDUMP;
 	vma->vm_flags |= mm->def_flags;
 	vma->vm_page_prot = protection_map[vma->vm_flags & 7];
 	vma->vm_ops = &syscall32_vm_ops;
diff -rpuN linux-2.6.18.8/arch/x86_64/ia32/syscall32_syscall-xen.S linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/syscall32_syscall-xen.S
--- linux-2.6.18.8/arch/x86_64/ia32/syscall32_syscall-xen.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/syscall32_syscall-xen.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,28 @@
+/* 32bit VDSOs mapped into user space. */
+
+	.section ".init.data","aw"
+
+#if CONFIG_XEN_COMPAT < 0x030200
+
+	.globl syscall32_int80
+	.globl syscall32_int80_end
+
+syscall32_int80:
+	.incbin "arch/x86_64/ia32/vsyscall-int80.so"
+syscall32_int80_end:
+
+#endif
+
+	.globl syscall32_syscall
+	.globl syscall32_syscall_end
+
+syscall32_syscall:
+	.incbin "arch/x86_64/ia32/vsyscall-syscall.so"
+syscall32_syscall_end:
+
+	.globl syscall32_sysenter
+	.globl syscall32_sysenter_end
+
+syscall32_sysenter:
+	.incbin "arch/x86_64/ia32/vsyscall-sysenter.so"
+syscall32_sysenter_end:
diff -rpuN linux-2.6.18.8/arch/x86_64/ia32/syscall32-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/syscall32-xen.c
--- linux-2.6.18.8/arch/x86_64/ia32/syscall32-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/syscall32-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,137 @@
+/* Copyright 2002,2003 Andi Kleen, SuSE Labs */
+
+/* vsyscall handling for 32bit processes. Map a stub page into it 
+   on demand because 32bit cannot reach the kernel's fixmaps */
+
+#include <linux/mm.h>
+#include <linux/string.h>
+#include <linux/kernel.h>
+#include <linux/gfp.h>
+#include <linux/init.h>
+#include <linux/stringify.h>
+#include <linux/security.h>
+#include <asm/proto.h>
+#include <asm/tlbflush.h>
+#include <asm/ia32_unistd.h>
+#include <xen/interface/callback.h>
+
+extern unsigned char syscall32_syscall[], syscall32_syscall_end[];
+extern unsigned char syscall32_sysenter[], syscall32_sysenter_end[];
+extern int sysctl_vsyscall32;
+
+char *syscall32_page; 
+static int use_sysenter = -1;
+
+#if CONFIG_XEN_COMPAT < 0x030200
+extern unsigned char syscall32_int80[], syscall32_int80_end[];
+static int use_int80 = 1;
+#endif
+
+static struct page *
+syscall32_nopage(struct vm_area_struct *vma, unsigned long adr, int *type)
+{
+	struct page *p = virt_to_page(adr - vma->vm_start + syscall32_page);
+	get_page(p);
+	return p;
+}
+
+/* Prevent VMA merging */
+static void syscall32_vma_close(struct vm_area_struct *vma)
+{
+}
+
+static struct vm_operations_struct syscall32_vm_ops = {
+	.close = syscall32_vma_close,
+	.nopage = syscall32_nopage,
+};
+
+struct linux_binprm;
+
+/* Setup a VMA at program startup for the vsyscall page */
+int syscall32_setup_pages(struct linux_binprm *bprm, int exstack)
+{
+	int npages = (VSYSCALL32_END - VSYSCALL32_BASE) >> PAGE_SHIFT;
+	struct vm_area_struct *vma;
+	struct mm_struct *mm = current->mm;
+	int ret;
+
+	vma = kmem_cache_alloc(vm_area_cachep, SLAB_KERNEL);
+	if (!vma)
+		return -ENOMEM;
+
+	memset(vma, 0, sizeof(struct vm_area_struct));
+	/* Could randomize here */
+	vma->vm_start = VSYSCALL32_BASE;
+	vma->vm_end = VSYSCALL32_END;
+	/* MAYWRITE to allow gdb to COW and set breakpoints */
+	vma->vm_flags = VM_READ|VM_EXEC|VM_MAYREAD|VM_MAYEXEC|VM_MAYWRITE;
+	vma->vm_flags |= mm->def_flags;
+	vma->vm_page_prot = protection_map[vma->vm_flags & 7];
+	vma->vm_ops = &syscall32_vm_ops;
+	vma->vm_mm = mm;
+
+	down_write(&mm->mmap_sem);
+	if ((ret = insert_vm_struct(mm, vma))) {
+		up_write(&mm->mmap_sem);
+		kmem_cache_free(vm_area_cachep, vma);
+		return ret;
+	}
+	mm->total_vm += npages;
+	up_write(&mm->mmap_sem);
+	return 0;
+}
+
+static int __init init_syscall32(void)
+{ 
+	syscall32_page = (void *)get_zeroed_page(GFP_KERNEL); 
+	if (!syscall32_page) 
+		panic("Cannot allocate syscall32 page"); 
+
+#if CONFIG_XEN_COMPAT < 0x030200
+	if (use_int80) {
+		memcpy(syscall32_page, syscall32_int80,
+		       syscall32_int80_end - syscall32_int80);
+	} else
+#endif
+ 	if (use_sysenter > 0) {
+ 		memcpy(syscall32_page, syscall32_sysenter,
+ 		       syscall32_sysenter_end - syscall32_sysenter);
+ 	} else {
+  		memcpy(syscall32_page, syscall32_syscall,
+  		       syscall32_syscall_end - syscall32_syscall);
+  	}	
+	return 0;
+} 
+
+/*
+ * This must be done early in case we have an initrd containing 32-bit
+ * binaries (e.g., hotplug). This could be pushed upstream to arch/x86_64.
+ */	
+core_initcall(init_syscall32); 
+
+/* May not be __init: called during resume */
+void syscall32_cpu_init(void)
+{
+	static const struct callback_register cstar = {
+		.type = CALLBACKTYPE_syscall32,
+		.address = (unsigned long)ia32_cstar_target
+	};
+	static const struct callback_register sysenter = {
+		.type = CALLBACKTYPE_sysenter,
+		.address = (unsigned long)ia32_sysenter_target
+	};
+
+	/* Load these always in case some future AMD CPU supports
+	   SYSENTER from compat mode too. */
+	if ((HYPERVISOR_callback_op(CALLBACKOP_register, &sysenter) < 0) ||
+	    (HYPERVISOR_callback_op(CALLBACKOP_register, &cstar) < 0))
+#if CONFIG_XEN_COMPAT < 0x030200
+		return;
+	use_int80 = 0;
+#else
+		BUG();
+#endif
+
+	if (use_sysenter < 0)
+		use_sysenter = (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL);
+}
diff -rpuN linux-2.6.18.8/arch/x86_64/ia32/vsyscall-int80.S linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/vsyscall-int80.S
--- linux-2.6.18.8/arch/x86_64/ia32/vsyscall-int80.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/vsyscall-int80.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,58 @@
+/*
+ * Code for the vsyscall page.  This version uses the old int $0x80 method.
+ *
+ * NOTE:
+ * 1) __kernel_vsyscall _must_ be first in this page.
+ * 2) there are alignment constraints on this stub, see vsyscall-sigreturn.S
+ *    for details.
+ */
+#include <asm/ia32_unistd.h>
+#include <asm/asm-offsets.h>
+
+	.code32
+	.text
+	.section .text.vsyscall,"ax"
+	.globl __kernel_vsyscall
+	.type __kernel_vsyscall,@function
+__kernel_vsyscall:
+.LSTART_vsyscall:
+	int $0x80
+	ret
+.LEND_vsyscall:
+	.size __kernel_vsyscall,.-.LSTART_vsyscall
+	.previous
+
+	.section .eh_frame,"a",@progbits
+.LSTARTFRAME:
+	.long .LENDCIE-.LSTARTCIE
+.LSTARTCIE:
+	.long 0			/* CIE ID */
+	.byte 1			/* Version number */
+	.string "zR"		/* NUL-terminated augmentation string */
+	.uleb128 1		/* Code alignment factor */
+	.sleb128 -4		/* Data alignment factor */
+	.byte 8			/* Return address register column */
+	.uleb128 1		/* Augmentation value length */
+	.byte 0x1b		/* DW_EH_PE_pcrel|DW_EH_PE_sdata4. */
+	.byte 0x0c		/* DW_CFA_def_cfa */
+	.uleb128 4
+	.uleb128 4
+	.byte 0x88		/* DW_CFA_offset, column 0x8 */
+	.uleb128 1
+	.align 4
+.LENDCIE:
+
+	.long .LENDFDE1-.LSTARTFDE1	/* Length FDE */
+.LSTARTFDE1:
+	.long .LSTARTFDE1-.LSTARTFRAME	/* CIE pointer */
+	.long .LSTART_vsyscall-.	/* PC-relative start address */
+	.long .LEND_vsyscall-.LSTART_vsyscall
+	.uleb128 0			/* Augmentation length */
+	.align 4
+.LENDFDE1:
+		
+/*
+ * Get the common code for the sigreturn entry points.
+ */
+#define SYSCALL_ENTER_KERNEL    int $0x80
+#include "vsyscall-sigreturn.S"
diff -rpuN linux-2.6.18.8/arch/x86_64/ia32/vsyscall-sigreturn.S linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/vsyscall-sigreturn.S
--- linux-2.6.18.8/arch/x86_64/ia32/vsyscall-sigreturn.S	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/ia32/vsyscall-sigreturn.S	2008-08-21 11:36:07.000000000 +0200
@@ -139,5 +139,5 @@ __kernel_rt_sigreturn:
 	.align 4
 .LENDFDE3:
 
-#include "../../i386/kernel/vsyscall-note.S"
+#include <vsyscall-note.S>
 
diff -rpuN linux-2.6.18.8/arch/x86_64/Kconfig linux-2.6.18-xen-3.3.0/arch/x86_64/Kconfig
--- linux-2.6.18.8/arch/x86_64/Kconfig	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/Kconfig	2008-08-21 11:36:07.000000000 +0200
@@ -135,6 +135,23 @@ config GENERIC_CPU
 
 endchoice
 
+config X86_64_XEN
+	bool "Enable Xen compatible kernel"
+	select XEN
+	select SWIOTLB
+	help
+	  This option will compile a kernel compatible with Xen hypervisor
+
+config X86_NO_TSS
+	bool
+	depends on X86_64_XEN
+	default y
+
+config X86_NO_IDT
+	bool
+	depends on X86_64_XEN
+	default y
+
 #
 # Define implied options from the CPU selection here
 #
@@ -155,6 +172,7 @@ config X86_INTERNODE_CACHE_BYTES
 
 config X86_TSC
 	bool
+	depends on !X86_64_XEN
 	default y
 
 config X86_GOOD_APIC
@@ -197,7 +215,7 @@ config X86_CPUID
 
 config X86_HT
 	bool
-	depends on SMP && !MK8
+	depends on SMP && !MK8 && !X86_64_XEN
 	default y
 
 config MATH_EMULATION
@@ -211,14 +229,22 @@ config EISA
 
 config X86_IO_APIC
 	bool
+	depends !XEN_UNPRIVILEGED_GUEST
+	default y
+
+config X86_XEN_GENAPIC
+	bool
+	depends on X86_64_XEN
 	default y
 
 config X86_LOCAL_APIC
 	bool
+	depends !XEN_UNPRIVILEGED_GUEST
 	default y
 
 config MTRR
 	bool "MTRR (Memory Type Range Register) support"
+	depends on !XEN_UNPRIVILEGED_GUEST
 	---help---
 	  On Intel P6 family processors (Pentium Pro, Pentium II and later)
 	  the Memory Type Range Registers (MTRRs) may be used to control
@@ -259,7 +285,7 @@ config SMP
 
 config SCHED_SMT
 	bool "SMT (Hyperthreading) scheduler support"
-	depends on SMP
+	depends on SMP && !X86_64_XEN
 	default n
 	help
 	  SMT scheduler support improves the CPU scheduler's decision making
@@ -269,7 +295,7 @@ config SCHED_SMT
 
 config SCHED_MC
 	bool "Multi-core scheduler support"
-	depends on SMP
+	depends on SMP && !X86_64_XEN
 	default y
 	help
 	  Multi-core scheduler support improves the CPU scheduler's decision
@@ -280,7 +306,7 @@ source "kernel/Kconfig.preempt"
 
 config NUMA
        bool "Non Uniform Memory Access (NUMA) Support"
-       depends on SMP
+       depends on SMP && !X86_64_XEN
        help
 	 Enable NUMA (Non Uniform Memory Access) support. The kernel 
 	 will try to allocate memory used by a CPU on the local memory 
@@ -341,7 +367,7 @@ config ARCH_DISCONTIGMEM_DEFAULT
 
 config ARCH_SPARSEMEM_ENABLE
 	def_bool y
-	depends on (NUMA || EXPERIMENTAL)
+	depends on (NUMA || EXPERIMENTAL) && !X86_64_XEN
 
 config ARCH_MEMORY_PROBE
 	def_bool y
@@ -365,6 +391,7 @@ config NR_CPUS
 	int "Maximum number of CPUs (2-256)"
 	range 2 255
 	depends on SMP
+	default "16" if X86_64_XEN
 	default "8"
 	help
 	  This allows you to specify the maximum number of CPUs which this
@@ -387,6 +414,7 @@ config ARCH_ENABLE_MEMORY_HOTPLUG
 
 config HPET_TIMER
 	bool
+	depends on !X86_64_XEN
 	default y
 	help
 	  Use the IA-PC HPET (High Precision Event Timer) to manage
@@ -407,7 +435,7 @@ config IOMMU
 	default y
 	select SWIOTLB
 	select AGP
-	depends on PCI
+	depends on PCI && !X86_64_XEN
 	help
 	  Support for full DMA access of devices with 32bit memory access only
 	  on systems with more than 3GB. This is usually needed for USB,
@@ -423,7 +451,7 @@ config CALGARY_IOMMU
 	bool "IBM Calgary IOMMU support"
 	default y
 	select SWIOTLB
-	depends on PCI && EXPERIMENTAL
+	depends on PCI && !X86_64_XEN && EXPERIMENTAL
 	help
 	  Support for hardware IOMMUs in IBM's xSeries x366 and x460
 	  systems. Needed to run systems with more than 3GB of memory
@@ -444,6 +472,7 @@ config SWIOTLB
 
 config X86_MCE
 	bool "Machine check support" if EMBEDDED
+	depends on !X86_64_XEN
 	default y
 	help
 	   Include a machine check error handler to report hardware errors.
@@ -469,7 +498,7 @@ config X86_MCE_AMD
 
 config KEXEC
 	bool "kexec system call (EXPERIMENTAL)"
-	depends on EXPERIMENTAL
+	depends on EXPERIMENTAL && !XEN_UNPRIVILEGED_GUEST
 	help
 	  kexec is a system call that implements the ability to shutdown your
 	  current kernel, and to start another kernel.  It is like a reboot
@@ -564,8 +593,9 @@ config GENERIC_PENDING_IRQ
 	default y
 
 menu "Power management options"
+	depends on !XEN_UNPRIVILEGED_GUEST
 
-source kernel/power/Kconfig
+source "kernel/power/Kconfig"
 
 source "drivers/acpi/Kconfig"
 
@@ -588,6 +618,22 @@ config PCI_MMCONFIG
 	bool "Support mmconfig PCI config space access"
 	depends on PCI && ACPI
 
+config XEN_PCIDEV_FRONTEND
+	bool "Xen PCI Frontend"
+	depends on PCI && X86_64_XEN
+	select HOTPLUG
+	default y
+	help
+	  The PCI device frontend driver allows the kernel to import arbitrary
+	  PCI devices from a PCI backend to support PCI driver domains.
+
+config XEN_PCIDEV_FE_DEBUG
+	bool "Xen PCI Frontend Debugging"
+	depends on XEN_PCIDEV_FRONTEND
+	default n
+	help
+	  Enables some debug statements within the PCI Frontend.
+
 source "drivers/pci/pcie/Kconfig"
 
 source "drivers/pci/Kconfig"
@@ -658,4 +704,6 @@ source "security/Kconfig"
 
 source "crypto/Kconfig"
 
+source "drivers/xen/Kconfig"
+
 source "lib/Kconfig"
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/acpi/Makefile linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/acpi/Makefile
--- linux-2.6.18.8/arch/x86_64/kernel/acpi/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/acpi/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -5,5 +5,10 @@ obj-$(CONFIG_ACPI_SLEEP)	+= sleep.o wake
 ifneq ($(CONFIG_ACPI_PROCESSOR),)
 obj-y			+= processor.o
 processor-y		:= ../../../i386/kernel/acpi/processor.o ../../../i386/kernel/acpi/cstate.o
+processor-$(CONFIG_XEN)	:= ../../../i386/kernel/acpi/processor.o
+ifneq ($(CONFIG_PROCESSOR_EXTERNAL_CONTROL),)
+processor-$(CONFIG_XEN)	+= ../../../i386/kernel/acpi/processor_extcntl_xen.o
+endif
 endif
 
+disabled-obj-$(CONFIG_XEN) := wakeup.o
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/acpi/sleep-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/acpi/sleep-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/acpi/sleep-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/acpi/sleep-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,146 @@
+/*
+ *  acpi.c - Architecture-Specific Low-Level ACPI Support
+ *
+ *  Copyright (C) 2001, 2002 Paul Diefenbaugh <paul.s.diefenbaugh@intel.com>
+ *  Copyright (C) 2001 Jun Nakajima <jun.nakajima@intel.com>
+ *  Copyright (C) 2001 Patrick Mochel <mochel@osdl.org>
+ *  Copyright (C) 2002 Andi Kleen, SuSE Labs (x86-64 port)
+ *  Copyright (C) 2003 Pavel Machek, SuSE Labs
+ *
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ */
+
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/types.h>
+#include <linux/stddef.h>
+#include <linux/slab.h>
+#include <linux/pci.h>
+#include <linux/bootmem.h>
+#include <linux/acpi.h>
+#include <linux/cpumask.h>
+
+#include <asm/mpspec.h>
+#include <asm/io.h>
+#include <asm/apic.h>
+#include <asm/apicdef.h>
+#include <asm/page.h>
+#include <asm/pgtable.h>
+#include <asm/pgalloc.h>
+#include <asm/io_apic.h>
+#include <asm/proto.h>
+#include <asm/tlbflush.h>
+
+/* --------------------------------------------------------------------------
+                              Low-Level Sleep Support
+   -------------------------------------------------------------------------- */
+
+#ifdef CONFIG_ACPI_SLEEP
+
+#ifndef CONFIG_ACPI_PV_SLEEP
+/* address in low memory of the wakeup routine. */
+unsigned long acpi_wakeup_address = 0;
+unsigned long acpi_video_flags;
+extern char wakeup_start, wakeup_end;
+
+extern unsigned long FASTCALL(acpi_copy_wakeup_routine(unsigned long));
+
+static pgd_t low_ptr;
+
+static void init_low_mapping(void)
+{
+	pgd_t *slot0 = pgd_offset(current->mm, 0UL);
+	low_ptr = *slot0;
+	set_pgd(slot0, *pgd_offset(current->mm, PAGE_OFFSET));
+	WARN_ON(num_online_cpus() != 1);
+	local_flush_tlb();
+}
+#endif
+
+/**
+ * acpi_save_state_mem - save kernel state
+ *
+ * Create an identity mapped page table and copy the wakeup routine to
+ * low memory.
+ */
+int acpi_save_state_mem(void)
+{
+#ifndef CONFIG_ACPI_PV_SLEEP
+	init_low_mapping();
+
+	memcpy((void *)acpi_wakeup_address, &wakeup_start,
+	       &wakeup_end - &wakeup_start);
+	acpi_copy_wakeup_routine(acpi_wakeup_address);
+#endif
+	return 0;
+}
+
+/*
+ * acpi_restore_state
+ */
+void acpi_restore_state_mem(void)
+{
+#ifndef CONFIG_ACPI_PV_SLEEP
+	set_pgd(pgd_offset(current->mm, 0UL), low_ptr);
+	local_flush_tlb();
+#endif
+}
+
+/**
+ * acpi_reserve_bootmem - do _very_ early ACPI initialisation
+ *
+ * We allocate a page in low memory for the wakeup
+ * routine for when we come back from a sleep state. The
+ * runtime allocator allows specification of <16M pages, but not
+ * <1M pages.
+ */
+void __init acpi_reserve_bootmem(void)
+{
+#ifndef CONFIG_ACPI_PV_SLEEP
+	acpi_wakeup_address = (unsigned long)alloc_bootmem_low(PAGE_SIZE);
+	if ((&wakeup_end - &wakeup_start) > PAGE_SIZE)
+		printk(KERN_CRIT
+		       "ACPI: Wakeup code way too big, will crash on attempt to suspend\n");
+#endif
+}
+
+#ifndef CONFIG_ACPI_PV_SLEEP
+static int __init acpi_sleep_setup(char *str)
+{
+	while ((str != NULL) && (*str != '\0')) {
+		if (strncmp(str, "s3_bios", 7) == 0)
+			acpi_video_flags = 1;
+		if (strncmp(str, "s3_mode", 7) == 0)
+			acpi_video_flags |= 2;
+		str = strchr(str, ',');
+		if (str != NULL)
+			str += strspn(str, ", \t");
+	}
+
+	return 1;
+}
+
+__setup("acpi_sleep=", acpi_sleep_setup);
+#endif				/* CONFIG_ACPI_PV_SLEEP */
+
+#endif				/*CONFIG_ACPI_SLEEP */
+
+void acpi_pci_link_exit(void)
+{
+}
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/apic-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/apic-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/apic-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/apic-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,197 @@
+/*
+ *	Local APIC handling, local APIC timers
+ *
+ *	(c) 1999, 2000 Ingo Molnar <mingo@redhat.com>
+ *
+ *	Fixes
+ *	Maciej W. Rozycki	:	Bits for genuine 82489DX APICs;
+ *					thanks to Eric Gilmore
+ *					and Rolf G. Tews
+ *					for testing these extensively.
+ *	Maciej W. Rozycki	:	Various updates and fixes.
+ *	Mikael Pettersson	:	Power Management for UP-APIC.
+ *	Pavel Machek and
+ *	Mikael Pettersson	:	PM converted to driver model.
+ */
+
+#include <linux/init.h>
+
+#include <linux/mm.h>
+#include <linux/delay.h>
+#include <linux/bootmem.h>
+#include <linux/smp_lock.h>
+#include <linux/interrupt.h>
+#include <linux/mc146818rtc.h>
+#include <linux/kernel_stat.h>
+#include <linux/sysdev.h>
+#include <linux/module.h>
+
+#include <asm/atomic.h>
+#include <asm/smp.h>
+#include <asm/mtrr.h>
+#include <asm/mpspec.h>
+#include <asm/desc.h>
+#include <asm/arch_hooks.h>
+#include <asm/hpet.h>
+#include <asm/idle.h>
+
+int apic_verbosity;
+
+/*
+ * 'what should we do if we get a hw irq event on an illegal vector'.
+ * each architecture has to answer this themselves.
+ */
+void ack_bad_irq(unsigned int irq)
+{
+	printk("unexpected IRQ trap at vector %02x\n", irq);
+	/*
+	 * Currently unexpected vectors happen only on SMP and APIC.
+	 * We _must_ ack these because every local APIC has only N
+	 * irq slots per priority level, and a 'hanging, unacked' IRQ
+	 * holds up an irq slot - in excessive cases (when multiple
+	 * unexpected vectors occur) that might lock up the APIC
+	 * completely.
+  	 * But don't ack when the APIC is disabled. -AK
+	 */
+	if (!disable_apic)
+		ack_APIC_irq();
+}
+
+int setup_profiling_timer(unsigned int multiplier)
+{
+	return -EINVAL;
+}
+
+void smp_local_timer_interrupt(struct pt_regs *regs)
+{
+	profile_tick(CPU_PROFILING, regs);
+#ifndef CONFIG_XEN
+#ifdef CONFIG_SMP
+		update_process_times(user_mode(regs));
+#endif
+#endif
+	/*
+	 * We take the 'long' return path, and there every subsystem
+	 * grabs the appropriate locks (kernel lock/ irq lock).
+	 *
+	 * we might want to decouple profiling from the 'long path',
+	 * and do the profiling totally in assembly.
+	 *
+	 * Currently this isn't too much of an issue (performance wise),
+	 * we can take more than 100K local irqs per second on a 100 MHz P5.
+	 */
+}
+
+/*
+ * Local APIC timer interrupt. This is the most natural way for doing
+ * local interrupts, but local timer interrupts can be emulated by
+ * broadcast interrupts too. [in case the hw doesn't support APIC timers]
+ *
+ * [ if a single-CPU system runs an SMP kernel then we call the local
+ *   interrupt as well. Thus we cannot inline the local irq ... ]
+ */
+void smp_apic_timer_interrupt(struct pt_regs *regs)
+{
+	/*
+	 * the NMI deadlock-detector uses this.
+	 */
+	add_pda(apic_timer_irqs, 1);
+
+	/*
+	 * NOTE! We'd better ACK the irq immediately,
+	 * because timer handling can be slow.
+	 */
+	ack_APIC_irq();
+	/*
+	 * update_process_times() expects us to have done irq_enter().
+	 * Besides, if we don't timer interrupts ignore the global
+	 * interrupt lock, which is the WrongThing (tm) to do.
+	 */
+	exit_idle();
+	irq_enter();
+	smp_local_timer_interrupt(regs);
+	irq_exit();
+}
+
+/*
+ * This interrupt should _never_ happen with our APIC/SMP architecture
+ */
+asmlinkage void smp_spurious_interrupt(void)
+{
+	unsigned int v;
+	exit_idle();
+	irq_enter();
+	/*
+	 * Check if this really is a spurious interrupt and ACK it
+	 * if it is a vectored one.  Just in case...
+	 * Spurious interrupts should not be ACKed.
+	 */
+	v = apic_read(APIC_ISR + ((SPURIOUS_APIC_VECTOR & ~0x1f) >> 1));
+	if (v & (1 << (SPURIOUS_APIC_VECTOR & 0x1f)))
+		ack_APIC_irq();
+
+#if 0
+	static unsigned long last_warning; 
+	static unsigned long skipped; 
+
+	/* see sw-dev-man vol 3, chapter 7.4.13.5 */
+	if (time_before(last_warning+30*HZ,jiffies)) { 
+		printk(KERN_INFO "spurious APIC interrupt on CPU#%d, %ld skipped.\n",
+		       smp_processor_id(), skipped);
+		last_warning = jiffies; 
+		skipped = 0;
+	} else { 
+		skipped++; 
+	} 
+#endif 
+	irq_exit();
+}
+
+/*
+ * This interrupt should never happen with our APIC/SMP architecture
+ */
+
+asmlinkage void smp_error_interrupt(void)
+{
+	unsigned int v, v1;
+
+	exit_idle();
+	irq_enter();
+	/* First tickle the hardware, only then report what went on. -- REW */
+	v = apic_read(APIC_ESR);
+	apic_write(APIC_ESR, 0);
+	v1 = apic_read(APIC_ESR);
+	ack_APIC_irq();
+	atomic_inc(&irq_err_count);
+
+	/* Here is what the APIC error bits mean:
+	   0: Send CS error
+	   1: Receive CS error
+	   2: Send accept error
+	   3: Receive accept error
+	   4: Reserved
+	   5: Send illegal vector
+	   6: Received illegal vector
+	   7: Illegal register address
+	*/
+	printk (KERN_DEBUG "APIC error on CPU%d: %02x(%02x)\n",
+	        smp_processor_id(), v , v1);
+	irq_exit();
+}
+
+int disable_apic;
+
+/*
+ * This initializes the IO-APIC and APIC hardware if this is
+ * a UP kernel.
+ */
+int __init APIC_init_uniprocessor (void)
+{
+#ifdef CONFIG_X86_IO_APIC
+	if (smp_found_config)
+		if (!skip_ioapic_setup && nr_ioapics)
+			setup_IO_APIC();
+#endif
+
+	return 1;
+}
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/asm-offsets.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/asm-offsets.c
--- linux-2.6.18.8/arch/x86_64/kernel/asm-offsets.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/asm-offsets.c	2008-08-21 11:36:07.000000000 +0200
@@ -67,8 +67,10 @@ int main(void)
 	DEFINE(pbe_address, offsetof(struct pbe, address));
 	DEFINE(pbe_orig_address, offsetof(struct pbe, orig_address));
 	DEFINE(pbe_next, offsetof(struct pbe, next));
+#ifndef CONFIG_X86_NO_TSS
 	BLANK();
 	DEFINE(TSS_ist, offsetof(struct tss_struct, ist));
+#endif
 	BLANK();
 	DEFINE(crypto_tfm_ctx_offset, offsetof(struct crypto_tfm, __crt_ctx));
 	return 0;
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/crash.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/crash.c
--- linux-2.6.18.8/arch/x86_64/kernel/crash.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/crash.c	2008-08-21 11:36:07.000000000 +0200
@@ -92,6 +92,7 @@ static void crash_save_self(struct pt_re
 	crash_save_this_cpu(regs, cpu);
 }
 
+#ifndef CONFIG_XEN
 #ifdef CONFIG_SMP
 static atomic_t waiting_for_crash_ipi;
 
@@ -156,6 +157,7 @@ static void nmi_shootdown_cpus(void)
 	/* There are no cpus to shootdown */
 }
 #endif
+#endif /* CONFIG_XEN */
 
 void machine_crash_shutdown(struct pt_regs *regs)
 {
@@ -173,6 +175,8 @@ void machine_crash_shutdown(struct pt_re
 
 	/* Make a note of crashing cpu. Will be used in NMI callback.*/
 	crashing_cpu = smp_processor_id();
+
+#ifndef CONFIG_XEN
 	nmi_shootdown_cpus();
 
 	if(cpu_has_apic)
@@ -181,6 +185,6 @@ void machine_crash_shutdown(struct pt_re
 #if defined(CONFIG_X86_IO_APIC)
 	disable_IO_APIC();
 #endif
-
+#endif /* CONFIG_XEN */
 	crash_save_self(regs);
 }
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/e820.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/e820.c
--- linux-2.6.18.8/arch/x86_64/kernel/e820.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/e820.c	2008-08-21 11:36:07.000000000 +0200
@@ -93,7 +93,7 @@ static inline int bad_addr(unsigned long
  * This function checks if any part of the range <start,end> is mapped
  * with type.
  */
-int __meminit
+int
 e820_any_mapped(unsigned long start, unsigned long end, unsigned type)
 { 
 	int i;
@@ -107,6 +107,7 @@ e820_any_mapped(unsigned long start, uns
 	} 
 	return 0;
 }
+EXPORT_SYMBOL_GPL(e820_any_mapped);
 
 /*
  * This function checks if the entire range <start,end> is mapped with type.
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/e820-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/e820-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/e820-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/e820-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,798 @@
+/* 
+ * Handle the memory map.
+ * The functions here do the job until bootmem takes over.
+ *
+ *  Getting sanitize_e820_map() in sync with i386 version by applying change:
+ *  -  Provisions for empty E820 memory regions (reported by certain BIOSes).
+ *     Alex Achenbach <xela@slit.de>, December 2002.
+ *  Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
+ *
+ */
+#include <linux/kernel.h>
+#include <linux/types.h>
+#include <linux/init.h>
+#include <linux/bootmem.h>
+#include <linux/ioport.h>
+#include <linux/string.h>
+#include <linux/kexec.h>
+#include <linux/module.h>
+
+#include <asm/pgtable.h>
+#include <asm/page.h>
+#include <asm/e820.h>
+#include <asm/proto.h>
+#include <asm/bootsetup.h>
+#include <asm/sections.h>
+#include <xen/interface/memory.h>
+
+/* 
+ * PFN of last memory page.
+ */
+unsigned long end_pfn; 
+EXPORT_SYMBOL(end_pfn);
+
+/* 
+ * end_pfn only includes RAM, while end_pfn_map includes all e820 entries.
+ * The direct mapping extends to end_pfn_map, so that we can directly access
+ * apertures, ACPI and other tables without having to play with fixmaps.
+ */ 
+unsigned long end_pfn_map; 
+
+/* 
+ * Last pfn which the user wants to use.
+ */
+unsigned long end_user_pfn = MAXMEM>>PAGE_SHIFT;  
+
+extern struct resource code_resource, data_resource;
+
+#ifdef CONFIG_XEN
+extern struct e820map machine_e820;
+#endif
+
+/* Check for some hardcoded bad areas that early boot is not allowed to touch */ 
+static inline int bad_addr(unsigned long *addrp, unsigned long size)
+{ 
+	unsigned long addr = *addrp, last = addr + size; 
+
+#ifndef CONFIG_XEN
+	/* various gunk below that needed for SMP startup */
+	if (addr < 0x8000) { 
+		*addrp = 0x8000;
+		return 1; 
+	}
+
+	/* direct mapping tables of the kernel */
+	if (last >= table_start<<PAGE_SHIFT && addr < table_end<<PAGE_SHIFT) { 
+		*addrp = table_end << PAGE_SHIFT; 
+		return 1;
+	} 
+
+	/* initrd */ 
+#ifdef CONFIG_BLK_DEV_INITRD
+	if (LOADER_TYPE && INITRD_START && last >= INITRD_START && 
+	    addr < INITRD_START+INITRD_SIZE) { 
+		*addrp = INITRD_START + INITRD_SIZE; 
+		return 1;
+	} 
+#endif
+	/* kernel code + 640k memory hole (later should not be needed, but 
+	   be paranoid for now) */
+	if (last >= 640*1024 && addr < 1024*1024) {
+		*addrp = 1024*1024;
+		return 1;
+	}
+	if (last >= __pa_symbol(&_text) && last < __pa_symbol(&_end)) {
+		*addrp = __pa_symbol(&_end);
+		return 1;
+	}
+
+	if (last >= ebda_addr && addr < ebda_addr + ebda_size) {
+		*addrp = ebda_addr + ebda_size;
+		return 1;
+	}
+
+	/* XXX ramdisk image here? */ 
+#else
+	if (last < (table_end<<PAGE_SHIFT)) {
+		*addrp = table_end << PAGE_SHIFT;
+		return 1;
+	}
+#endif
+	return 0;
+} 
+
+/*
+ * This function checks if any part of the range <start,end> is mapped
+ * with type.
+ */
+int e820_any_mapped(unsigned long start, unsigned long end, unsigned type)
+{ 
+	int i;
+
+#ifndef CONFIG_XEN
+	for (i = 0; i < e820.nr_map; i++) { 
+		struct e820entry *ei = &e820.map[i]; 
+#else
+	if (!is_initial_xendomain())
+		return 0;
+	for (i = 0; i < machine_e820.nr_map; i++) {
+		const struct e820entry *ei = &machine_e820.map[i];
+#endif
+
+		if (type && ei->type != type) 
+			continue;
+		if (ei->addr >= end || ei->addr + ei->size <= start)
+			continue; 
+		return 1; 
+	} 
+	return 0;
+}
+EXPORT_SYMBOL_GPL(e820_any_mapped);
+
+/*
+ * This function checks if the entire range <start,end> is mapped with type.
+ *
+ * Note: this function only works correct if the e820 table is sorted and
+ * not-overlapping, which is the case
+ */
+int __init e820_all_mapped(unsigned long start, unsigned long end, unsigned type)
+{
+	int i;
+
+#ifndef CONFIG_XEN
+	for (i = 0; i < e820.nr_map; i++) {
+		struct e820entry *ei = &e820.map[i];
+#else
+	if (!is_initial_xendomain())
+		return 0;
+	for (i = 0; i < machine_e820.nr_map; i++) {
+		const struct e820entry *ei = &machine_e820.map[i];
+#endif
+
+		if (type && ei->type != type)
+			continue;
+		/* is the region (part) in overlap with the current region ?*/
+		if (ei->addr >= end || ei->addr + ei->size <= start)
+			continue;
+
+		/* if the region is at the beginning of <start,end> we move
+		 * start to the end of the region since it's ok until there
+		 */
+		if (ei->addr <= start)
+			start = ei->addr + ei->size;
+		/* if start is now at or beyond end, we're done, full coverage */
+		if (start >= end)
+			return 1; /* we're done */
+	}
+	return 0;
+}
+
+/* 
+ * Find a free area in a specific range. 
+ */ 
+unsigned long __init find_e820_area(unsigned long start, unsigned long end, unsigned size) 
+{ 
+	int i; 
+	for (i = 0; i < e820.nr_map; i++) { 
+		struct e820entry *ei = &e820.map[i]; 
+		unsigned long addr = ei->addr, last; 
+		if (ei->type != E820_RAM) 
+			continue; 
+		if (addr < start) 
+			addr = start;
+		if (addr > ei->addr + ei->size) 
+			continue; 
+		while (bad_addr(&addr, size) && addr+size <= ei->addr+ei->size)
+			;
+		last = addr + size;
+		if (last > ei->addr + ei->size)
+			continue;
+		if (last > end) 
+			continue;
+		return addr; 
+	} 
+	return -1UL;		
+} 
+
+/* 
+ * Free bootmem based on the e820 table for a node.
+ */
+void __init e820_bootmem_free(pg_data_t *pgdat, unsigned long start,unsigned long end)
+{
+	int i;
+	for (i = 0; i < e820.nr_map; i++) {
+		struct e820entry *ei = &e820.map[i]; 
+		unsigned long last, addr;
+
+		if (ei->type != E820_RAM || 
+		    ei->addr+ei->size <= start || 
+		    ei->addr >= end)
+			continue;
+
+		addr = round_up(ei->addr, PAGE_SIZE);
+		if (addr < start) 
+			addr = start;
+
+		last = round_down(ei->addr + ei->size, PAGE_SIZE); 
+		if (last >= end)
+			last = end; 
+
+		if (last > addr && last-addr >= PAGE_SIZE)
+			free_bootmem_node(pgdat, addr, last-addr);
+	}
+}
+
+/*
+ * Find the highest page frame number we have available
+ */
+unsigned long __init e820_end_of_ram(void)
+{
+	int i;
+	unsigned long end_pfn = 0;
+	
+	for (i = 0; i < e820.nr_map; i++) {
+		struct e820entry *ei = &e820.map[i]; 
+		unsigned long start, end;
+
+		start = round_up(ei->addr, PAGE_SIZE); 
+		end = round_down(ei->addr + ei->size, PAGE_SIZE); 
+		if (start >= end)
+			continue;
+		if (ei->type == E820_RAM) { 
+		if (end > end_pfn<<PAGE_SHIFT)
+			end_pfn = end>>PAGE_SHIFT;
+		} else { 
+			if (end > end_pfn_map<<PAGE_SHIFT) 
+				end_pfn_map = end>>PAGE_SHIFT;
+		} 
+	}
+
+	if (end_pfn > end_pfn_map) 
+		end_pfn_map = end_pfn;
+	if (end_pfn_map > MAXMEM>>PAGE_SHIFT)
+		end_pfn_map = MAXMEM>>PAGE_SHIFT;
+	if (end_pfn > end_user_pfn)
+		end_pfn = end_user_pfn;
+	if (end_pfn > end_pfn_map) 
+		end_pfn = end_pfn_map; 
+
+	return end_pfn;	
+}
+
+/* 
+ * Compute how much memory is missing in a range.
+ * Unlike the other functions in this file the arguments are in page numbers.
+ */
+unsigned long __init
+e820_hole_size(unsigned long start_pfn, unsigned long end_pfn)
+{
+	unsigned long ram = 0;
+	unsigned long start = start_pfn << PAGE_SHIFT;
+	unsigned long end = end_pfn << PAGE_SHIFT;
+	int i;
+	for (i = 0; i < e820.nr_map; i++) {
+		struct e820entry *ei = &e820.map[i];
+		unsigned long last, addr;
+
+		if (ei->type != E820_RAM ||
+		    ei->addr+ei->size <= start ||
+		    ei->addr >= end)
+			continue;
+
+		addr = round_up(ei->addr, PAGE_SIZE);
+		if (addr < start)
+			addr = start;
+
+		last = round_down(ei->addr + ei->size, PAGE_SIZE);
+		if (last >= end)
+			last = end;
+
+		if (last > addr)
+			ram += last - addr;
+	}
+	return ((end - start) - ram) >> PAGE_SHIFT;
+}
+
+/*
+ * Mark e820 reserved areas as busy for the resource manager.
+ */
+void __init e820_reserve_resources(struct e820entry *e820, int nr_map)
+{
+	int i;
+	for (i = 0; i < nr_map; i++) {
+		struct resource *res;
+		res = alloc_bootmem_low(sizeof(struct resource));
+		switch (e820[i].type) {
+		case E820_RAM:	res->name = "System RAM"; break;
+		case E820_ACPI:	res->name = "ACPI Tables"; break;
+		case E820_NVS:	res->name = "ACPI Non-volatile Storage"; break;
+		default:	res->name = "reserved";
+		}
+		res->start = e820[i].addr;
+		res->end = res->start + e820[i].size - 1;
+		res->flags = IORESOURCE_MEM | IORESOURCE_BUSY;
+		request_resource(&iomem_resource, res);
+		if (e820[i].type == E820_RAM) {
+			/*
+			 *  We don't know which RAM region contains kernel data,
+			 *  so we try it repeatedly and let the resource manager
+			 *  test it.
+			 */
+#ifndef CONFIG_XEN
+			request_resource(res, &code_resource);
+			request_resource(res, &data_resource);
+#endif
+#ifdef CONFIG_KEXEC
+			if (crashk_res.start != crashk_res.end)
+				request_resource(res, &crashk_res);
+#ifdef CONFIG_XEN
+			xen_machine_kexec_register_resources(res);
+#endif
+#endif
+		}
+	}
+}
+
+/* 
+ * Add a memory region to the kernel e820 map.
+ */ 
+void __init add_memory_region(unsigned long start, unsigned long size, int type)
+{
+	int x = e820.nr_map;
+
+	if (x == E820MAX) {
+		printk(KERN_ERR "Ooops! Too many entries in the memory map!\n");
+		return;
+	}
+
+	e820.map[x].addr = start;
+	e820.map[x].size = size;
+	e820.map[x].type = type;
+	e820.nr_map++;
+}
+
+void __init e820_print_map(char *who)
+{
+	int i;
+
+	for (i = 0; i < e820.nr_map; i++) {
+		printk(" %s: %016Lx - %016Lx ", who,
+			(unsigned long long) e820.map[i].addr,
+			(unsigned long long) (e820.map[i].addr + e820.map[i].size));
+		switch (e820.map[i].type) {
+		case E820_RAM:	printk("(usable)\n");
+				break;
+		case E820_RESERVED:
+				printk("(reserved)\n");
+				break;
+		case E820_ACPI:
+				printk("(ACPI data)\n");
+				break;
+		case E820_NVS:
+				printk("(ACPI NVS)\n");
+				break;
+		default:	printk("type %u\n", e820.map[i].type);
+				break;
+		}
+	}
+}
+
+/*
+ * Sanitize the BIOS e820 map.
+ *
+ * Some e820 responses include overlapping entries.  The following 
+ * replaces the original e820 map with a new one, removing overlaps.
+ *
+ */
+static int __init sanitize_e820_map(struct e820entry * biosmap, char * pnr_map)
+{
+	struct change_member {
+		struct e820entry *pbios; /* pointer to original bios entry */
+		unsigned long long addr; /* address for this change point */
+	};
+	static struct change_member change_point_list[2*E820MAX] __initdata;
+	static struct change_member *change_point[2*E820MAX] __initdata;
+	static struct e820entry *overlap_list[E820MAX] __initdata;
+	static struct e820entry new_bios[E820MAX] __initdata;
+	struct change_member *change_tmp;
+	unsigned long current_type, last_type;
+	unsigned long long last_addr;
+	int chgidx, still_changing;
+	int overlap_entries;
+	int new_bios_entry;
+	int old_nr, new_nr, chg_nr;
+	int i;
+
+	/*
+		Visually we're performing the following (1,2,3,4 = memory types)...
+
+		Sample memory map (w/overlaps):
+		   ____22__________________
+		   ______________________4_
+		   ____1111________________
+		   _44_____________________
+		   11111111________________
+		   ____________________33__
+		   ___________44___________
+		   __________33333_________
+		   ______________22________
+		   ___________________2222_
+		   _________111111111______
+		   _____________________11_
+		   _________________4______
+
+		Sanitized equivalent (no overlap):
+		   1_______________________
+		   _44_____________________
+		   ___1____________________
+		   ____22__________________
+		   ______11________________
+		   _________1______________
+		   __________3_____________
+		   ___________44___________
+		   _____________33_________
+		   _______________2________
+		   ________________1_______
+		   _________________4______
+		   ___________________2____
+		   ____________________33__
+		   ______________________4_
+	*/
+
+	/* if there's only one memory region, don't bother */
+	if (*pnr_map < 2)
+		return -1;
+
+	old_nr = *pnr_map;
+
+	/* bail out if we find any unreasonable addresses in bios map */
+	for (i=0; i<old_nr; i++)
+		if (biosmap[i].addr + biosmap[i].size < biosmap[i].addr)
+			return -1;
+
+	/* create pointers for initial change-point information (for sorting) */
+	for (i=0; i < 2*old_nr; i++)
+		change_point[i] = &change_point_list[i];
+
+	/* record all known change-points (starting and ending addresses),
+	   omitting those that are for empty memory regions */
+	chgidx = 0;
+	for (i=0; i < old_nr; i++)	{
+		if (biosmap[i].size != 0) {
+			change_point[chgidx]->addr = biosmap[i].addr;
+			change_point[chgidx++]->pbios = &biosmap[i];
+			change_point[chgidx]->addr = biosmap[i].addr + biosmap[i].size;
+			change_point[chgidx++]->pbios = &biosmap[i];
+		}
+	}
+	chg_nr = chgidx;
+
+	/* sort change-point list by memory addresses (low -> high) */
+	still_changing = 1;
+	while (still_changing)	{
+		still_changing = 0;
+		for (i=1; i < chg_nr; i++)  {
+			/* if <current_addr> > <last_addr>, swap */
+			/* or, if current=<start_addr> & last=<end_addr>, swap */
+			if ((change_point[i]->addr < change_point[i-1]->addr) ||
+				((change_point[i]->addr == change_point[i-1]->addr) &&
+				 (change_point[i]->addr == change_point[i]->pbios->addr) &&
+				 (change_point[i-1]->addr != change_point[i-1]->pbios->addr))
+			   )
+			{
+				change_tmp = change_point[i];
+				change_point[i] = change_point[i-1];
+				change_point[i-1] = change_tmp;
+				still_changing=1;
+			}
+		}
+	}
+
+	/* create a new bios memory map, removing overlaps */
+	overlap_entries=0;	 /* number of entries in the overlap table */
+	new_bios_entry=0;	 /* index for creating new bios map entries */
+	last_type = 0;		 /* start with undefined memory type */
+	last_addr = 0;		 /* start with 0 as last starting address */
+	/* loop through change-points, determining affect on the new bios map */
+	for (chgidx=0; chgidx < chg_nr; chgidx++)
+	{
+		/* keep track of all overlapping bios entries */
+		if (change_point[chgidx]->addr == change_point[chgidx]->pbios->addr)
+		{
+			/* add map entry to overlap list (> 1 entry implies an overlap) */
+			overlap_list[overlap_entries++]=change_point[chgidx]->pbios;
+		}
+		else
+		{
+			/* remove entry from list (order independent, so swap with last) */
+			for (i=0; i<overlap_entries; i++)
+			{
+				if (overlap_list[i] == change_point[chgidx]->pbios)
+					overlap_list[i] = overlap_list[overlap_entries-1];
+			}
+			overlap_entries--;
+		}
+		/* if there are overlapping entries, decide which "type" to use */
+		/* (larger value takes precedence -- 1=usable, 2,3,4,4+=unusable) */
+		current_type = 0;
+		for (i=0; i<overlap_entries; i++)
+			if (overlap_list[i]->type > current_type)
+				current_type = overlap_list[i]->type;
+		/* continue building up new bios map based on this information */
+		if (current_type != last_type)	{
+			if (last_type != 0)	 {
+				new_bios[new_bios_entry].size =
+					change_point[chgidx]->addr - last_addr;
+				/* move forward only if the new size was non-zero */
+				if (new_bios[new_bios_entry].size != 0)
+					if (++new_bios_entry >= E820MAX)
+						break; 	/* no more space left for new bios entries */
+			}
+			if (current_type != 0)	{
+				new_bios[new_bios_entry].addr = change_point[chgidx]->addr;
+				new_bios[new_bios_entry].type = current_type;
+				last_addr=change_point[chgidx]->addr;
+			}
+			last_type = current_type;
+		}
+	}
+	new_nr = new_bios_entry;   /* retain count for new bios entries */
+
+	/* copy new bios mapping into original location */
+	memcpy(biosmap, new_bios, new_nr*sizeof(struct e820entry));
+	*pnr_map = new_nr;
+
+	return 0;
+}
+
+/*
+ * Copy the BIOS e820 map into a safe place.
+ *
+ * Sanity-check it while we're at it..
+ *
+ * If we're lucky and live on a modern system, the setup code
+ * will have given us a memory map that we can use to properly
+ * set up memory.  If we aren't, we'll fake a memory map.
+ *
+ * We check to see that the memory map contains at least 2 elements
+ * before we'll use it, because the detection code in setup.S may
+ * not be perfect and most every PC known to man has two memory
+ * regions: one from 0 to 640k, and one from 1mb up.  (The IBM
+ * thinkpad 560x, for example, does not cooperate with the memory
+ * detection code.)
+ */
+static int __init copy_e820_map(struct e820entry * biosmap, int nr_map)
+{
+#ifndef CONFIG_XEN
+	/* Only one memory region (or negative)? Ignore it */
+	if (nr_map < 2)
+		return -1;
+#else
+	BUG_ON(nr_map < 1);
+#endif
+
+	do {
+		unsigned long start = biosmap->addr;
+		unsigned long size = biosmap->size;
+		unsigned long end = start + size;
+		unsigned long type = biosmap->type;
+
+		/* Overflow in 64 bits? Ignore the memory map. */
+		if (start > end)
+			return -1;
+
+#ifndef CONFIG_XEN
+		/*
+		 * Some BIOSes claim RAM in the 640k - 1M region.
+		 * Not right. Fix it up.
+		 * 
+		 * This should be removed on Hammer which is supposed to not
+		 * have non e820 covered ISA mappings there, but I had some strange
+		 * problems so it stays for now.  -AK
+		 */
+		if (type == E820_RAM) {
+			if (start < 0x100000ULL && end > 0xA0000ULL) {
+				if (start < 0xA0000ULL)
+					add_memory_region(start, 0xA0000ULL-start, type);
+				if (end <= 0x100000ULL)
+					continue;
+				start = 0x100000ULL;
+				size = end - start;
+			}
+		}
+#endif
+
+		add_memory_region(start, size, type);
+	} while (biosmap++,--nr_map);
+
+#ifdef CONFIG_XEN
+	if (is_initial_xendomain()) {
+		struct xen_memory_map memmap;
+
+		memmap.nr_entries = E820MAX;
+		set_xen_guest_handle(memmap.buffer, machine_e820.map);
+
+		if (HYPERVISOR_memory_op(XENMEM_machine_memory_map, &memmap))
+			BUG();
+		machine_e820.nr_map = memmap.nr_entries;
+	} else
+		machine_e820 = e820;
+#endif
+
+	return 0;
+}
+
+#ifndef CONFIG_XEN
+void __init setup_memory_region(void)
+{
+	char *who = "BIOS-e820";
+
+	/*
+	 * Try to copy the BIOS-supplied E820-map.
+	 *
+	 * Otherwise fake a memory map; one section from 0k->640k,
+	 * the next section from 1mb->appropriate_mem_k
+	 */
+	sanitize_e820_map(E820_MAP, &E820_MAP_NR);
+	if (copy_e820_map(E820_MAP, E820_MAP_NR) < 0) {
+		unsigned long mem_size;
+
+		/* compare results from other methods and take the greater */
+		if (ALT_MEM_K < EXT_MEM_K) {
+			mem_size = EXT_MEM_K;
+			who = "BIOS-88";
+		} else {
+			mem_size = ALT_MEM_K;
+			who = "BIOS-e801";
+		}
+
+		e820.nr_map = 0;
+		add_memory_region(0, LOWMEMSIZE(), E820_RAM);
+		add_memory_region(HIGH_MEMORY, mem_size << 10, E820_RAM);
+  	}
+	printk(KERN_INFO "BIOS-provided physical RAM map:\n");
+	e820_print_map(who);
+}
+
+#else  /* CONFIG_XEN */
+
+void __init setup_memory_region(void)
+{
+	int rc;
+	struct xen_memory_map memmap;
+	/*
+	 * This is rather large for a stack variable but this early in
+	 * the boot process we know we have plenty slack space.
+	 */
+	struct e820entry map[E820MAX];
+
+	memmap.nr_entries = E820MAX;
+	set_xen_guest_handle(memmap.buffer, map);
+
+	rc = HYPERVISOR_memory_op(XENMEM_memory_map, &memmap);
+	if ( rc == -ENOSYS ) {
+		memmap.nr_entries = 1;
+		map[0].addr = 0ULL;
+		map[0].size = xen_start_info->nr_pages << PAGE_SHIFT;
+		/* 8MB slack (to balance backend allocations). */
+		map[0].size += 8 << 20;
+		map[0].type = E820_RAM;
+		rc = 0;
+	}
+	BUG_ON(rc);
+
+	sanitize_e820_map(map, (char *)&memmap.nr_entries);
+
+	BUG_ON(copy_e820_map(map, (char)memmap.nr_entries) < 0);
+
+	printk(KERN_INFO "BIOS-provided physical RAM map:\n");
+	e820_print_map("Xen");
+}
+#endif
+
+void __init parse_memopt(char *p, char **from) 
+{ 
+	int i;
+	unsigned long current_end;
+	unsigned long end;
+
+	end_user_pfn = memparse(p, from);
+	end_user_pfn >>= PAGE_SHIFT;	
+
+	end = end_user_pfn<<PAGE_SHIFT;
+	i = e820.nr_map-1;
+	current_end = e820.map[i].addr + e820.map[i].size;
+
+	if (current_end < end) {
+		/*
+                 * The e820 map ends before our requested size so
+                 * extend the final entry to the requested address.
+                 */
+		if (e820.map[i].type == E820_RAM)
+			e820.map[i].size = end - e820.map[i].addr;
+		else
+			add_memory_region(current_end, end - current_end, E820_RAM);
+	}
+} 
+
+void __init parse_memmapopt(char *p, char **from)
+{
+	unsigned long long start_at, mem_size;
+
+	mem_size = memparse(p, from);
+	p = *from;
+	if (*p == '@') {
+		start_at = memparse(p+1, from);
+		add_memory_region(start_at, mem_size, E820_RAM);
+	} else if (*p == '#') {
+		start_at = memparse(p+1, from);
+		add_memory_region(start_at, mem_size, E820_ACPI);
+	} else if (*p == '$') {
+		start_at = memparse(p+1, from);
+		add_memory_region(start_at, mem_size, E820_RESERVED);
+	} else {
+		end_user_pfn = (mem_size >> PAGE_SHIFT);
+	}
+	p = *from;
+}
+
+unsigned long pci_mem_start = 0xaeedbabe;
+EXPORT_SYMBOL(pci_mem_start);
+
+/*
+ * Search for the biggest gap in the low 32 bits of the e820
+ * memory space.  We pass this space to PCI to assign MMIO resources
+ * for hotplug or unconfigured devices in.
+ * Hopefully the BIOS let enough space left.
+ */
+__init void e820_setup_gap(struct e820entry *e820, int nr_map)
+{
+	unsigned long gapstart, gapsize, round;
+	unsigned long last;
+	int i;
+	int found = 0;
+
+	last = 0x100000000ull;
+	gapstart = 0x10000000;
+	gapsize = 0x400000;
+	i = nr_map;
+	while (--i >= 0) {
+		unsigned long long start = e820[i].addr;
+		unsigned long long end = start + e820[i].size;
+
+		/*
+		 * Since "last" is at most 4GB, we know we'll
+		 * fit in 32 bits if this condition is true
+		 */
+		if (last > end) {
+			unsigned long gap = last - end;
+
+			if (gap > gapsize) {
+				gapsize = gap;
+				gapstart = end;
+				found = 1;
+			}
+		}
+		if (start < last)
+			last = start;
+	}
+
+	if (!found) {
+		gapstart = (end_pfn << PAGE_SHIFT) + 1024*1024;
+		printk(KERN_ERR "PCI: Warning: Cannot find a gap in the 32bit address range\n"
+		       KERN_ERR "PCI: Unassigned devices with 32bit resource registers may break!\n");
+	}
+
+	/*
+	 * See how much we want to round up: start off with
+	 * rounding to the next 1MB area.
+	 */
+	round = 0x100000;
+	while ((gapsize >> 4) > round)
+		round += round;
+	/* Fun with two's complement */
+	pci_mem_start = (gapstart + round) & -round;
+
+	printk(KERN_INFO "Allocating PCI resources starting at %lx (gap: %lx:%lx)\n",
+		pci_mem_start, gapstart, gapsize);
+}
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/early_printk-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/early_printk-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/early_printk-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/early_printk-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,302 @@
+#include <linux/console.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/string.h>
+#include <linux/screen_info.h>
+#include <asm/io.h>
+#include <asm/processor.h>
+#include <asm/fcntl.h>
+
+/* Simple VGA output */
+
+#ifdef __i386__
+#include <asm/setup.h>
+#define VGABASE		(__ISA_IO_base + 0xb8000)
+#else
+#include <asm/bootsetup.h>
+#define VGABASE		((void __iomem *)0xffffffff800b8000UL)
+#endif
+
+#ifndef CONFIG_XEN
+static int max_ypos = 25, max_xpos = 80;
+static int current_ypos = 25, current_xpos = 0; 
+
+static void early_vga_write(struct console *con, const char *str, unsigned n)
+{
+	char c;
+	int  i, k, j;
+
+	while ((c = *str++) != '\0' && n-- > 0) {
+		if (current_ypos >= max_ypos) {
+			/* scroll 1 line up */
+			for (k = 1, j = 0; k < max_ypos; k++, j++) {
+				for (i = 0; i < max_xpos; i++) {
+					writew(readw(VGABASE+2*(max_xpos*k+i)),
+					       VGABASE + 2*(max_xpos*j + i));
+				}
+			}
+			for (i = 0; i < max_xpos; i++)
+				writew(0x720, VGABASE + 2*(max_xpos*j + i));
+			current_ypos = max_ypos-1;
+		}
+		if (c == '\n') {
+			current_xpos = 0;
+			current_ypos++;
+		} else if (c != '\r')  {
+			writew(((0x7 << 8) | (unsigned short) c),
+			       VGABASE + 2*(max_xpos*current_ypos +
+						current_xpos++));
+			if (current_xpos >= max_xpos) {
+				current_xpos = 0;
+				current_ypos++;
+			}
+		}
+	}
+}
+
+static struct console early_vga_console = {
+	.name =		"earlyvga",
+	.write =	early_vga_write,
+	.flags =	CON_PRINTBUFFER,
+	.index =	-1,
+};
+
+/* Serial functions loosely based on a similar package from Klaus P. Gerlicher */
+
+static int early_serial_base = 0x3f8;  /* ttyS0 */
+
+#define XMTRDY          0x20
+
+#define DLAB		0x80
+
+#define TXR             0       /*  Transmit register (WRITE) */
+#define RXR             0       /*  Receive register  (READ)  */
+#define IER             1       /*  Interrupt Enable          */
+#define IIR             2       /*  Interrupt ID              */
+#define FCR             2       /*  FIFO control              */
+#define LCR             3       /*  Line control              */
+#define MCR             4       /*  Modem control             */
+#define LSR             5       /*  Line Status               */
+#define MSR             6       /*  Modem Status              */
+#define DLL             0       /*  Divisor Latch Low         */
+#define DLH             1       /*  Divisor latch High        */
+
+static int early_serial_putc(unsigned char ch)
+{
+	unsigned timeout = 0xffff;
+	while ((inb(early_serial_base + LSR) & XMTRDY) == 0 && --timeout)
+		cpu_relax();
+	outb(ch, early_serial_base + TXR);
+	return timeout ? 0 : -1;
+}
+
+static void early_serial_write(struct console *con, const char *s, unsigned n)
+{
+	while (*s && n-- > 0) {
+		early_serial_putc(*s);
+		if (*s == '\n')
+			early_serial_putc('\r');
+		s++;
+	}
+}
+
+#define DEFAULT_BAUD 9600
+
+static __init void early_serial_init(char *s)
+{
+	unsigned char c;
+	unsigned divisor;
+	unsigned baud = DEFAULT_BAUD;
+	char *e;
+
+	if (*s == ',')
+		++s;
+
+	if (*s) {
+		unsigned port;
+		if (!strncmp(s,"0x",2)) {
+			early_serial_base = simple_strtoul(s, &e, 16);
+		} else {
+			static int bases[] = { 0x3f8, 0x2f8 };
+
+			if (!strncmp(s,"ttyS",4))
+				s += 4;
+			port = simple_strtoul(s, &e, 10);
+			if (port > 1 || s == e)
+				port = 0;
+			early_serial_base = bases[port];
+		}
+		s += strcspn(s, ",");
+		if (*s == ',')
+			s++;
+	}
+
+	outb(0x3, early_serial_base + LCR);	/* 8n1 */
+	outb(0, early_serial_base + IER);	/* no interrupt */
+	outb(0, early_serial_base + FCR);	/* no fifo */
+	outb(0x3, early_serial_base + MCR);	/* DTR + RTS */
+
+	if (*s) {
+		baud = simple_strtoul(s, &e, 0);
+		if (baud == 0 || s == e)
+			baud = DEFAULT_BAUD;
+	}
+
+	divisor = 115200 / baud;
+	c = inb(early_serial_base + LCR);
+	outb(c | DLAB, early_serial_base + LCR);
+	outb(divisor & 0xff, early_serial_base + DLL);
+	outb((divisor >> 8) & 0xff, early_serial_base + DLH);
+	outb(c & ~DLAB, early_serial_base + LCR);
+}
+
+#else /* CONFIG_XEN */
+
+static void
+early_serial_write(struct console *con, const char *s, unsigned count)
+{
+	int n;
+
+	while (count > 0) {
+		n = HYPERVISOR_console_io(CONSOLEIO_write, count, (char *)s);
+		if (n <= 0)
+			break;
+		count -= n;
+		s += n;
+	}
+} 
+
+static __init void early_serial_init(char *s)
+{
+}
+
+/*
+ * No early VGA console on Xen, as we do not have convenient ISA-space
+ * mappings. Someone should fix this for domain 0. For now, use fake serial.
+ */
+#define early_vga_console early_serial_console
+
+#endif
+
+static struct console early_serial_console = {
+	.name =		"earlyser",
+	.write =	early_serial_write,
+	.flags =	CON_PRINTBUFFER,
+	.index =	-1,
+};
+
+/* Console interface to a host file on AMD's SimNow! */
+
+static int simnow_fd;
+
+enum {
+	MAGIC1 = 0xBACCD00A,
+	MAGIC2 = 0xCA110000,
+	XOPEN = 5,
+	XWRITE = 4,
+};
+
+static noinline long simnow(long cmd, long a, long b, long c)
+{
+	long ret;
+	asm volatile("cpuid" :
+		     "=a" (ret) :
+		     "b" (a), "c" (b), "d" (c), "0" (MAGIC1), "D" (cmd + MAGIC2));
+	return ret;
+}
+
+void __init simnow_init(char *str)
+{
+	char *fn = "klog";
+	if (*str == '=')
+		fn = ++str;
+	/* error ignored */
+	simnow_fd = simnow(XOPEN, (unsigned long)fn, O_WRONLY|O_APPEND|O_CREAT, 0644);
+}
+
+static void simnow_write(struct console *con, const char *s, unsigned n)
+{
+	simnow(XWRITE, simnow_fd, (unsigned long)s, n);
+}
+
+static struct console simnow_console = {
+	.name =		"simnow",
+	.write =	simnow_write,
+	.flags =	CON_PRINTBUFFER,
+	.index =	-1,
+};
+
+/* Direct interface for emergencies */
+struct console *early_console = &early_vga_console;
+static int early_console_initialized = 0;
+
+void early_printk(const char *fmt, ...)
+{
+	char buf[512];
+	int n;
+	va_list ap;
+
+	va_start(ap,fmt);
+	n = vscnprintf(buf,512,fmt,ap);
+	early_console->write(early_console,buf,n);
+	va_end(ap);
+}
+
+static int __initdata keep_early;
+
+int __init setup_early_printk(char *opt)
+{
+	char *space;
+	char buf[256];
+
+	if (early_console_initialized)
+		return 1;
+
+	strlcpy(buf,opt,sizeof(buf));
+	space = strchr(buf, ' ');
+	if (space)
+		*space = 0;
+
+	if (strstr(buf,"keep"))
+		keep_early = 1;
+
+	if (!strncmp(buf, "serial", 6)) {
+		early_serial_init(buf + 6);
+		early_console = &early_serial_console;
+	} else if (!strncmp(buf, "ttyS", 4)) {
+		early_serial_init(buf);
+		early_console = &early_serial_console;
+	} else if (!strncmp(buf, "vga", 3)
+#ifndef CONFIG_XEN
+	           && SCREEN_INFO.orig_video_isVGA == 1) {
+		max_xpos = SCREEN_INFO.orig_video_cols;
+		max_ypos = SCREEN_INFO.orig_video_lines;
+		current_ypos = SCREEN_INFO.orig_y;
+#else
+	           || !strncmp(buf, "xen", 3)) {
+#endif
+		early_console = &early_vga_console;
+ 	} else if (!strncmp(buf, "simnow", 6)) {
+ 		simnow_init(buf + 6);
+ 		early_console = &simnow_console;
+ 		keep_early = 1;
+	}
+	early_console_initialized = 1;
+	register_console(early_console);
+	return 0;
+}
+
+void __init disable_early_printk(void)
+{
+	if (!early_console_initialized || !early_console)
+		return;
+	if (!keep_early) {
+		printk("disabling early console\n");
+		unregister_console(early_console);
+		early_console_initialized = 0;
+	} else {
+		printk("keeping early console\n");
+	}
+}
+
+__setup("earlyprintk=", setup_early_printk);
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/entry-xen.S linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/entry-xen.S
--- linux-2.6.18.8/arch/x86_64/kernel/entry-xen.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/entry-xen.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1322 @@
+/*
+ *  linux/arch/x86_64/entry.S
+ *
+ *  Copyright (C) 1991, 1992  Linus Torvalds
+ *  Copyright (C) 2000, 2001, 2002  Andi Kleen SuSE Labs
+ *  Copyright (C) 2000  Pavel Machek <pavel@suse.cz>
+ * 
+ *  $Id$
+ *
+ *  Jun Nakajima <jun.nakajima@intel.com>
+ *  Asit Mallick <asit.k.mallick@intel.com>
+ *      Modified for Xen
+ */
+
+/*
+ * entry.S contains the system-call and fault low-level handling routines.
+ *
+ * NOTE: This code handles signal-recognition, which happens every time
+ * after an interrupt and after each system call.
+ * 
+ * Normal syscalls and interrupts don't save a full stack frame, this is 
+ * only done for syscall tracing, signals or fork/exec et.al.
+ * 
+ * A note on terminology:	 
+ * - top of stack: Architecture defined interrupt frame from SS to RIP 
+ * at the top of the kernel process stack.	
+ * - partial stack frame: partially saved registers upto R11.
+ * - full stack frame: Like partial stack frame, but all register saved. 
+ *	
+ * TODO:	 
+ * - schedule it carefully for the final hardware.
+ */
+
+#define ASSEMBLY 1
+#include <linux/linkage.h>
+#include <asm/segment.h>
+#include <asm/smp.h>
+#include <asm/cache.h>
+#include <asm/errno.h>
+#include <asm/dwarf2.h>
+#include <asm/calling.h>
+#include <asm/asm-offsets.h>
+#include <asm/msr.h>
+#include <asm/unistd.h>
+#include <asm/thread_info.h>
+#include <asm/hw_irq.h>
+#include <asm/page.h>
+#include <asm/irqflags.h>
+#include <asm/errno.h>
+#include <xen/interface/arch-x86_64.h>
+#include <xen/interface/features.h>
+
+#include "xen_entry.S"
+	
+	.code64
+
+#ifndef CONFIG_PREEMPT
+#define retint_kernel retint_restore_args
+#endif	
+
+
+.macro TRACE_IRQS_IRETQ offset=ARGOFFSET
+#ifdef CONFIG_TRACE_IRQFLAGS
+	bt   $9,EFLAGS-\offset(%rsp)	/* interrupts off? */
+	jnc  1f
+	TRACE_IRQS_ON
+1:
+#endif
+.endm
+
+NMI_MASK = 0x80000000
+	
+/*
+ * C code is not supposed to know about undefined top of stack. Every time 
+ * a C function with an pt_regs argument is called from the SYSCALL based 
+ * fast path FIXUP_TOP_OF_STACK is needed.
+ * RESTORE_TOP_OF_STACK syncs the syscall state after any possible ptregs
+ * manipulation.
+ */        	
+		
+	/* %rsp:at FRAMEEND */ 
+	.macro FIXUP_TOP_OF_STACK tmp
+	movq    $__USER_CS,CS(%rsp)
+	movq 	$-1,RCX(%rsp)
+	.endm
+
+	.macro RESTORE_TOP_OF_STACK tmp,offset=0
+	.endm
+
+	.macro FAKE_STACK_FRAME child_rip
+	/* push in order ss, rsp, eflags, cs, rip */
+	xorl %eax, %eax
+	pushq %rax /* ss */
+	CFI_ADJUST_CFA_OFFSET	8
+	/*CFI_REL_OFFSET	ss,0*/
+	pushq %rax /* rsp */
+	CFI_ADJUST_CFA_OFFSET	8
+	CFI_REL_OFFSET	rsp,0
+	pushq $(1<<9) /* eflags - interrupts on */
+	CFI_ADJUST_CFA_OFFSET	8
+	/*CFI_REL_OFFSET	rflags,0*/
+	pushq $__KERNEL_CS /* cs */
+	CFI_ADJUST_CFA_OFFSET	8
+	/*CFI_REL_OFFSET	cs,0*/
+	pushq \child_rip /* rip */
+	CFI_ADJUST_CFA_OFFSET	8
+	CFI_REL_OFFSET	rip,0
+	pushq	%rax /* orig rax */
+	CFI_ADJUST_CFA_OFFSET	8
+	.endm
+
+	.macro UNFAKE_STACK_FRAME
+	addq $8*6, %rsp
+	CFI_ADJUST_CFA_OFFSET	-(6*8)
+	.endm
+
+	.macro	CFI_DEFAULT_STACK start=1,adj=0
+	.if \start
+	CFI_STARTPROC	simple
+	CFI_DEF_CFA	rsp,SS+8-(\adj*ARGOFFSET)
+	.else
+	CFI_DEF_CFA_OFFSET SS+8-(\adj*ARGOFFSET)
+	.endif
+	.if \adj == 0
+	CFI_REL_OFFSET	r15,R15
+	CFI_REL_OFFSET	r14,R14
+	CFI_REL_OFFSET	r13,R13
+	CFI_REL_OFFSET	r12,R12
+	CFI_REL_OFFSET	rbp,RBP
+	CFI_REL_OFFSET	rbx,RBX
+	.endif
+	CFI_REL_OFFSET	r11,R11
+	CFI_REL_OFFSET	r10,R10
+	CFI_REL_OFFSET	r9,R9
+	CFI_REL_OFFSET	r8,R8
+	CFI_REL_OFFSET	rax,RAX
+	CFI_REL_OFFSET	rcx,RCX
+	CFI_REL_OFFSET	rdx,RDX
+	CFI_REL_OFFSET	rsi,RSI
+	CFI_REL_OFFSET	rdi,RDI
+	CFI_REL_OFFSET	rip,RIP
+	/*CFI_REL_OFFSET	cs,CS*/
+	/*CFI_REL_OFFSET	rflags,EFLAGS*/
+	CFI_REL_OFFSET	rsp,RSP
+	/*CFI_REL_OFFSET	ss,SS*/
+	.endm
+
+        /*
+         * Must be consistent with the definition in arch-x86/xen-x86_64.h:
+         *     struct iret_context {
+         *        u64 rax, r11, rcx, flags, rip, cs, rflags, rsp, ss;
+         *     };
+         * with rax, r11, and rcx being taken care of in the hypercall stub.
+         */
+	.macro HYPERVISOR_IRET flag
+	testb $3,1*8(%rsp)
+	jnz   2f
+	testl $NMI_MASK,2*8(%rsp)
+	jnz   2f
+
+	cmpb  $0,(xen_features+XENFEAT_supervisor_mode_kernel)(%rip)
+	jne   1f
+
+	/* Direct iret to kernel space. Correct CS and SS. */
+	orl   $3,1*8(%rsp)
+	orl   $3,4*8(%rsp)
+1:	iretq
+
+2:	/* Slow iret via hypervisor. */
+	andl  $~NMI_MASK, 2*8(%rsp)
+	pushq $\flag
+	jmp  hypercall_page + (__HYPERVISOR_iret * 32)
+	.endm
+
+/*
+ * A newly forked process directly context switches into this.
+ */ 	
+/* rdi:	prev */	
+ENTRY(ret_from_fork)
+	CFI_DEFAULT_STACK
+	push kernel_eflags(%rip)
+	CFI_ADJUST_CFA_OFFSET 4
+	popf				# reset kernel eflags
+	CFI_ADJUST_CFA_OFFSET -4
+	call schedule_tail
+	GET_THREAD_INFO(%rcx)
+	testl $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT),threadinfo_flags(%rcx)
+	jnz rff_trace
+rff_action:	
+	RESTORE_REST
+	testl $3,CS-ARGOFFSET(%rsp)	# from kernel_thread?
+	je   int_ret_from_sys_call
+	testl $_TIF_IA32,threadinfo_flags(%rcx)
+	jnz  int_ret_from_sys_call
+	RESTORE_TOP_OF_STACK %rdi,ARGOFFSET
+	jmp ret_from_sys_call
+rff_trace:
+	movq %rsp,%rdi
+	call syscall_trace_leave
+	GET_THREAD_INFO(%rcx)	
+	jmp rff_action
+	CFI_ENDPROC
+END(ret_from_fork)
+
+/*
+ * initial frame state for interrupts and exceptions
+ */
+	.macro _frame ref
+	CFI_STARTPROC simple
+	CFI_DEF_CFA rsp,SS+8-\ref
+	/*CFI_REL_OFFSET ss,SS-\ref*/
+	CFI_REL_OFFSET rsp,RSP-\ref
+	/*CFI_REL_OFFSET rflags,EFLAGS-\ref*/
+	/*CFI_REL_OFFSET cs,CS-\ref*/
+	CFI_REL_OFFSET rip,RIP-\ref
+	.endm
+
+/*
+ * System call entry. Upto 6 arguments in registers are supported.
+ *
+ * SYSCALL does not save anything on the stack and does not change the
+ * stack pointer.
+ */
+		
+/*
+ * Register setup:	
+ * rax  system call number
+ * rdi  arg0
+ * rcx  return address for syscall/sysret, C arg3 
+ * rsi  arg1
+ * rdx  arg2	
+ * r10  arg3 	(--> moved to rcx for C)
+ * r8   arg4
+ * r9   arg5
+ * r11  eflags for syscall/sysret, temporary for C
+ * r12-r15,rbp,rbx saved by C code, not touched. 		
+ * 
+ * Interrupts are enabled on entry.
+ * Only called from user space.
+ *
+ * XXX	if we had a free scratch register we could save the RSP into the stack frame
+ *      and report it properly in ps. Unfortunately we haven't.
+ *
+ * When user can change the frames always force IRET. That is because
+ * it deals with uncanonical addresses better. SYSRET has trouble
+ * with them due to bugs in both AMD and Intel CPUs.
+ */ 			 		
+
+ENTRY(system_call)
+	_frame (RIP-0x10)
+	SAVE_ARGS -8,0
+	movq  %rax,ORIG_RAX-ARGOFFSET(%rsp) 
+	GET_THREAD_INFO(%rcx)
+	testl $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SECCOMP),threadinfo_flags(%rcx)
+	CFI_REMEMBER_STATE
+	jnz tracesys
+	cmpq $__NR_syscall_max,%rax
+	ja badsys
+	movq %r10,%rcx
+	call *sys_call_table(,%rax,8)  # XXX:	 rip relative
+	movq %rax,RAX-ARGOFFSET(%rsp)
+/*
+ * Syscall return path ending with SYSRET (fast path)
+ * Has incomplete stack frame and undefined top of stack. 
+ */		
+	.globl ret_from_sys_call
+ret_from_sys_call:
+	movl $_TIF_ALLWORK_MASK,%edi
+	/* edi:	flagmask */
+sysret_check:		
+	GET_THREAD_INFO(%rcx)
+        XEN_BLOCK_EVENTS(%rsi)        
+	TRACE_IRQS_OFF
+	movl threadinfo_flags(%rcx),%edx
+	andl %edi,%edx
+	CFI_REMEMBER_STATE
+	jnz  sysret_careful 
+	/*
+	 * sysretq will re-enable interrupts:
+	 */
+	TRACE_IRQS_ON
+        XEN_UNBLOCK_EVENTS(%rsi)                
+	RESTORE_ARGS 0,8,0
+        HYPERVISOR_IRET VGCF_IN_SYSCALL
+
+	/* Handle reschedules */
+	/* edx:	work, edi: workmask */	
+sysret_careful:
+	CFI_RESTORE_STATE
+	bt $TIF_NEED_RESCHED,%edx
+	jnc sysret_signal
+	TRACE_IRQS_ON
+	XEN_UNBLOCK_EVENTS(%rsi)
+	pushq %rdi
+	CFI_ADJUST_CFA_OFFSET 8
+	call schedule
+	popq  %rdi
+	CFI_ADJUST_CFA_OFFSET -8
+	jmp sysret_check
+
+	/* Handle a signal */ 
+sysret_signal:
+	TRACE_IRQS_ON
+/*	sti */
+        XEN_UNBLOCK_EVENTS(%rsi)        
+	testl $(_TIF_SIGPENDING|_TIF_NOTIFY_RESUME|_TIF_SINGLESTEP),%edx
+	jz    1f
+
+	/* Really a signal */
+	/* edx:	work flags (arg3) */
+	leaq do_notify_resume(%rip),%rax
+	leaq -ARGOFFSET(%rsp),%rdi # &pt_regs -> arg1
+	xorl %esi,%esi # oldset -> arg2
+	call ptregscall_common
+1:	movl $_TIF_NEED_RESCHED,%edi
+	/* Use IRET because user could have changed frame. This
+	   works because ptregscall_common has called FIXUP_TOP_OF_STACK. */
+	XEN_BLOCK_EVENTS(%rsi)
+	TRACE_IRQS_OFF
+	jmp int_with_check
+	
+badsys:
+	movq $-ENOSYS,RAX-ARGOFFSET(%rsp)
+	jmp ret_from_sys_call
+
+	/* Do syscall tracing */
+tracesys:			 
+	CFI_RESTORE_STATE
+	SAVE_REST
+	movq $-ENOSYS,RAX(%rsp)
+	FIXUP_TOP_OF_STACK %rdi
+	movq %rsp,%rdi
+	call syscall_trace_enter
+	LOAD_ARGS ARGOFFSET  /* reload args from stack in case ptrace changed it */
+	RESTORE_REST
+	cmpq $__NR_syscall_max,%rax
+	ja  int_ret_from_sys_call
+	movq %r10,%rcx	/* fixup for C */
+	call *sys_call_table(,%rax,8)
+	movq %rax,RAX-ARGOFFSET(%rsp)
+	/* Use IRET because user could have changed frame */
+	jmp int_ret_from_sys_call
+	CFI_ENDPROC
+END(system_call)
+		
+/* 
+ * Syscall return path ending with IRET.
+ * Has correct top of stack, but partial stack frame.
+ */ 	
+ENTRY(int_ret_from_sys_call)
+	CFI_STARTPROC	simple
+	CFI_DEF_CFA	rsp,SS+8-ARGOFFSET
+	/*CFI_REL_OFFSET	ss,SS-ARGOFFSET*/
+	CFI_REL_OFFSET	rsp,RSP-ARGOFFSET
+	/*CFI_REL_OFFSET	rflags,EFLAGS-ARGOFFSET*/
+	/*CFI_REL_OFFSET	cs,CS-ARGOFFSET*/
+	CFI_REL_OFFSET	rip,RIP-ARGOFFSET
+	CFI_REL_OFFSET	rdx,RDX-ARGOFFSET
+	CFI_REL_OFFSET	rcx,RCX-ARGOFFSET
+	CFI_REL_OFFSET	rax,RAX-ARGOFFSET
+	CFI_REL_OFFSET	rdi,RDI-ARGOFFSET
+	CFI_REL_OFFSET	rsi,RSI-ARGOFFSET
+	CFI_REL_OFFSET	r8,R8-ARGOFFSET
+	CFI_REL_OFFSET	r9,R9-ARGOFFSET
+	CFI_REL_OFFSET	r10,R10-ARGOFFSET
+	CFI_REL_OFFSET	r11,R11-ARGOFFSET
+        XEN_BLOCK_EVENTS(%rsi)
+	TRACE_IRQS_OFF
+	testb $3,CS-ARGOFFSET(%rsp)
+        jnz 1f
+        /* Need to set the proper %ss (not NULL) for ring 3 iretq */
+        movl $__KERNEL_DS,SS-ARGOFFSET(%rsp)
+        jmp retint_restore_args   # retrun from ring3 kernel
+1:              
+	movl $_TIF_ALLWORK_MASK,%edi
+	/* edi:	mask to check */
+int_with_check:
+	GET_THREAD_INFO(%rcx)
+	movl threadinfo_flags(%rcx),%edx
+	andl %edi,%edx
+	jnz   int_careful
+	andl    $~TS_COMPAT,threadinfo_status(%rcx)
+	jmp   retint_restore_args
+
+	/* Either reschedule or signal or syscall exit tracking needed. */
+	/* First do a reschedule test. */
+	/* edx:	work, edi: workmask */
+int_careful:
+	bt $TIF_NEED_RESCHED,%edx
+	jnc  int_very_careful
+	TRACE_IRQS_ON
+/*	sti */
+        XEN_UNBLOCK_EVENTS(%rsi)
+	pushq %rdi
+	CFI_ADJUST_CFA_OFFSET 8
+	call schedule
+	popq %rdi
+	CFI_ADJUST_CFA_OFFSET -8
+	XEN_BLOCK_EVENTS(%rsi)
+	TRACE_IRQS_OFF
+	jmp int_with_check
+
+	/* handle signals and tracing -- both require a full stack frame */
+int_very_careful:
+	TRACE_IRQS_ON
+/*	sti */
+        XEN_UNBLOCK_EVENTS(%rsi)
+	SAVE_REST
+	/* Check for syscall exit trace */	
+	testl $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SINGLESTEP),%edx
+	jz int_signal
+	pushq %rdi
+	CFI_ADJUST_CFA_OFFSET 8
+	leaq 8(%rsp),%rdi	# &ptregs -> arg1	
+	call syscall_trace_leave
+	popq %rdi
+	CFI_ADJUST_CFA_OFFSET -8
+	andl $~(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SINGLESTEP),%edi
+	XEN_BLOCK_EVENTS(%rsi)
+	TRACE_IRQS_OFF
+	jmp int_restore_rest
+	
+int_signal:
+	testl $(_TIF_NOTIFY_RESUME|_TIF_SIGPENDING|_TIF_SINGLESTEP),%edx
+	jz 1f
+	movq %rsp,%rdi		# &ptregs -> arg1
+	xorl %esi,%esi		# oldset -> arg2
+	call do_notify_resume
+1:	movl $_TIF_NEED_RESCHED,%edi	
+int_restore_rest:
+	RESTORE_REST
+	XEN_BLOCK_EVENTS(%rsi)
+	TRACE_IRQS_OFF
+	jmp int_with_check
+	CFI_ENDPROC
+END(int_ret_from_sys_call)
+		
+/* 
+ * Certain special system calls that need to save a complete full stack frame.
+ */ 								
+	
+	.macro PTREGSCALL label,func,arg
+	.globl \label
+\label:
+	leaq	\func(%rip),%rax
+	leaq    -ARGOFFSET+8(%rsp),\arg /* 8 for return address */
+	jmp	ptregscall_common
+END(\label)
+	.endm
+
+	CFI_STARTPROC
+
+	PTREGSCALL stub_clone, sys_clone, %r8
+	PTREGSCALL stub_fork, sys_fork, %rdi
+	PTREGSCALL stub_vfork, sys_vfork, %rdi
+	PTREGSCALL stub_rt_sigsuspend, sys_rt_sigsuspend, %rdx
+	PTREGSCALL stub_sigaltstack, sys_sigaltstack, %rdx
+	PTREGSCALL stub_iopl, sys_iopl, %rsi
+
+ENTRY(ptregscall_common)
+	popq %r11
+	CFI_ADJUST_CFA_OFFSET -8
+	CFI_REGISTER rip, r11
+	SAVE_REST
+	movq %r11, %r15
+	CFI_REGISTER rip, r15
+	FIXUP_TOP_OF_STACK %r11
+	call *%rax
+	RESTORE_TOP_OF_STACK %r11
+	movq %r15, %r11
+	CFI_REGISTER rip, r11
+	RESTORE_REST
+	pushq %r11
+	CFI_ADJUST_CFA_OFFSET 8
+	CFI_REL_OFFSET rip, 0
+	ret
+	CFI_ENDPROC
+END(ptregscall_common)
+	
+ENTRY(stub_execve)
+	CFI_STARTPROC
+	popq %r11
+	CFI_ADJUST_CFA_OFFSET -8
+	CFI_REGISTER rip, r11
+	SAVE_REST
+	FIXUP_TOP_OF_STACK %r11
+	call sys_execve
+	RESTORE_TOP_OF_STACK %r11
+	movq %rax,RAX(%rsp)
+	RESTORE_REST
+	jmp int_ret_from_sys_call
+	CFI_ENDPROC
+END(stub_execve)
+	
+/*
+ * sigreturn is special because it needs to restore all registers on return.
+ * This cannot be done with SYSRET, so use the IRET return path instead.
+ */                
+ENTRY(stub_rt_sigreturn)
+	CFI_STARTPROC
+	addq $8, %rsp
+	CFI_ADJUST_CFA_OFFSET	-8
+	SAVE_REST
+	movq %rsp,%rdi
+	FIXUP_TOP_OF_STACK %r11
+	call sys_rt_sigreturn
+	movq %rax,RAX(%rsp) # fixme, this could be done at the higher layer
+	RESTORE_REST
+	jmp int_ret_from_sys_call
+	CFI_ENDPROC
+END(stub_rt_sigreturn)
+
+/* initial frame state for interrupts (and exceptions without error code) */
+#define INTR_FRAME _frame (RIP-0x10); \
+	CFI_REL_OFFSET rcx,0; \
+	CFI_REL_OFFSET r11,8
+
+/* initial frame state for exceptions with error code (and interrupts with
+   vector already pushed) */
+#define XCPT_FRAME _frame (RIP-0x18); \
+	CFI_REL_OFFSET rcx,0; \
+	CFI_REL_OFFSET r11,8
+
+/* 
+ * Interrupt exit.
+ *
+ */ 
+
+retint_check:
+	CFI_DEFAULT_STACK adj=1
+	movl threadinfo_flags(%rcx),%edx
+	andl %edi,%edx
+	CFI_REMEMBER_STATE
+	jnz  retint_careful
+retint_restore_args:
+	movl EFLAGS-REST_SKIP(%rsp), %eax
+	shr $9, %eax			# EAX[0] == IRET_EFLAGS.IF
+	XEN_GET_VCPU_INFO(%rsi)
+	andb evtchn_upcall_mask(%rsi),%al
+	andb $1,%al			# EAX[0] == IRET_EFLAGS.IF & event_mask
+	jnz restore_all_enable_events	#        != 0 => enable event delivery
+	XEN_PUT_VCPU_INFO(%rsi)
+		
+	RESTORE_ARGS 0,8,0
+	HYPERVISOR_IRET 0
+	
+	/* edi: workmask, edx: work */
+retint_careful:
+	CFI_RESTORE_STATE
+	bt    $TIF_NEED_RESCHED,%edx
+	jnc   retint_signal
+	TRACE_IRQS_ON
+	XEN_UNBLOCK_EVENTS(%rsi)
+/*	sti */        
+	pushq %rdi
+	CFI_ADJUST_CFA_OFFSET	8
+	call  schedule
+	popq %rdi		
+	CFI_ADJUST_CFA_OFFSET	-8
+	GET_THREAD_INFO(%rcx)
+	XEN_BLOCK_EVENTS(%rsi)		
+/*	cli */
+	TRACE_IRQS_OFF
+	jmp retint_check
+	
+retint_signal:
+	testl $(_TIF_SIGPENDING|_TIF_NOTIFY_RESUME|_TIF_SINGLESTEP),%edx
+	jz    retint_restore_args
+	TRACE_IRQS_ON
+        XEN_UNBLOCK_EVENTS(%rsi)
+	SAVE_REST
+	movq $-1,ORIG_RAX(%rsp) 			
+	xorl %esi,%esi		# oldset
+	movq %rsp,%rdi		# &pt_regs
+	call do_notify_resume
+	RESTORE_REST
+        XEN_BLOCK_EVENTS(%rsi)		
+	TRACE_IRQS_OFF
+	movl $_TIF_NEED_RESCHED,%edi
+	GET_THREAD_INFO(%rcx)
+	jmp retint_check
+
+#ifdef CONFIG_PREEMPT
+	/* Returning to kernel space. Check if we need preemption */
+	/* rcx:	 threadinfo. interrupts off. */
+	.p2align
+retint_kernel:	
+	cmpl $0,threadinfo_preempt_count(%rcx)
+	jnz  retint_restore_args
+	bt  $TIF_NEED_RESCHED,threadinfo_flags(%rcx)
+	jnc  retint_restore_args
+	bt   $9,EFLAGS-ARGOFFSET(%rsp)	/* interrupts off? */
+	jnc  retint_restore_args
+	call preempt_schedule_irq
+	jmp retint_kernel       /* check again */
+#endif	
+
+	CFI_ENDPROC
+END(retint_check)
+	
+#ifndef CONFIG_XEN
+/*
+ * APIC interrupts.
+ */		
+	.macro apicinterrupt num,func
+	INTR_FRAME
+	pushq $~(\num)
+	CFI_ADJUST_CFA_OFFSET 8
+	interrupt \func
+	jmp error_entry
+	CFI_ENDPROC
+	.endm
+
+ENTRY(thermal_interrupt)
+	apicinterrupt THERMAL_APIC_VECTOR,smp_thermal_interrupt
+END(thermal_interrupt)
+
+ENTRY(threshold_interrupt)
+	apicinterrupt THRESHOLD_APIC_VECTOR,mce_threshold_interrupt
+END(threshold_interrupt)
+
+#ifdef CONFIG_SMP	
+ENTRY(reschedule_interrupt)
+	apicinterrupt RESCHEDULE_VECTOR,smp_reschedule_interrupt
+END(reschedule_interrupt)
+
+	.macro INVALIDATE_ENTRY num
+ENTRY(invalidate_interrupt\num)
+	apicinterrupt INVALIDATE_TLB_VECTOR_START+\num,smp_invalidate_interrupt	
+END(invalidate_interrupt\num)
+	.endm
+
+	INVALIDATE_ENTRY 0
+	INVALIDATE_ENTRY 1
+	INVALIDATE_ENTRY 2
+	INVALIDATE_ENTRY 3
+	INVALIDATE_ENTRY 4
+	INVALIDATE_ENTRY 5
+	INVALIDATE_ENTRY 6
+	INVALIDATE_ENTRY 7
+
+ENTRY(call_function_interrupt)
+	apicinterrupt CALL_FUNCTION_VECTOR,smp_call_function_interrupt
+END(call_function_interrupt)
+#endif
+
+#ifdef CONFIG_X86_LOCAL_APIC	
+ENTRY(apic_timer_interrupt)
+	apicinterrupt LOCAL_TIMER_VECTOR,smp_apic_timer_interrupt
+END(apic_timer_interrupt)
+
+ENTRY(error_interrupt)
+	apicinterrupt ERROR_APIC_VECTOR,smp_error_interrupt
+END(error_interrupt)
+
+ENTRY(spurious_interrupt)
+	apicinterrupt SPURIOUS_APIC_VECTOR,smp_spurious_interrupt
+END(spurious_interrupt)
+#endif
+#endif /* !CONFIG_XEN */
+				
+/*
+ * Exception entry points.
+ */ 		
+	.macro zeroentry sym
+	INTR_FRAME
+        movq (%rsp),%rcx
+	CFI_RESTORE rcx
+        movq 8(%rsp),%r11
+	CFI_RESTORE r11
+        addq $0x10,%rsp /* skip rcx and r11 */
+	CFI_ADJUST_CFA_OFFSET -0x10
+	pushq $0	/* push error code/oldrax */ 
+	CFI_ADJUST_CFA_OFFSET 8
+	pushq %rax	/* push real oldrax to the rdi slot */ 
+	CFI_ADJUST_CFA_OFFSET 8
+	CFI_REL_OFFSET rax,0
+	leaq  \sym(%rip),%rax
+	jmp error_entry
+	CFI_ENDPROC
+	.endm	
+
+	.macro errorentry sym
+	XCPT_FRAME
+        movq (%rsp),%rcx
+	CFI_RESTORE rcx
+        movq 8(%rsp),%r11
+	CFI_RESTORE r11
+        addq $0x10,%rsp /* rsp points to the error code */
+	CFI_ADJUST_CFA_OFFSET -0x10
+	pushq %rax
+	CFI_ADJUST_CFA_OFFSET 8
+	CFI_REL_OFFSET rax,0
+	leaq  \sym(%rip),%rax
+	jmp error_entry
+	CFI_ENDPROC
+	.endm
+
+#if 0 /* not XEN */
+	/* error code is on the stack already */
+	/* handle NMI like exceptions that can happen everywhere */
+	.macro paranoidentry sym, ist=0, irqtrace=1
+        movq (%rsp),%rcx
+        movq 8(%rsp),%r11
+        addq $0x10,%rsp /* skip rcx and r11 */        
+	SAVE_ALL
+	cld
+#if 0 /* not XEN */
+	movl $1,%ebx
+	movl  $MSR_GS_BASE,%ecx
+	rdmsr
+	testl %edx,%edx
+	js    1f
+	swapgs
+	xorl  %ebx,%ebx
+1:
+#endif
+	.if \ist
+	movq	%gs:pda_data_offset, %rbp
+	.endif
+	movq %rsp,%rdi
+	movq ORIG_RAX(%rsp),%rsi
+	movq $-1,ORIG_RAX(%rsp)
+	.if \ist
+	subq	$EXCEPTION_STKSZ, per_cpu__init_tss + TSS_ist + (\ist - 1) * 8(%rbp)
+	.endif
+	call \sym
+	.if \ist
+	addq	$EXCEPTION_STKSZ, per_cpu__init_tss + TSS_ist + (\ist - 1) * 8(%rbp)
+	.endif
+/*	cli */
+	XEN_BLOCK_EVENTS(%rsi)		
+	.if \irqtrace
+	TRACE_IRQS_OFF
+	.endif
+	.endm
+
+	/*
+ 	 * "Paranoid" exit path from exception stack.
+  	 * Paranoid because this is used by NMIs and cannot take
+	 * any kernel state for granted.
+	 * We don't do kernel preemption checks here, because only
+	 * NMI should be common and it does not enable IRQs and
+	 * cannot get reschedule ticks.
+	 *
+	 * "trace" is 0 for the NMI handler only, because irq-tracing
+	 * is fundamentally NMI-unsafe. (we cannot change the soft and
+	 * hard flags at once, atomically)
+	 */
+	.macro paranoidexit trace=1
+	/* ebx:	no swapgs flag */
+paranoid_exit\trace:
+	testl %ebx,%ebx				/* swapgs needed? */
+	jnz paranoid_restore\trace
+	testl $3,CS(%rsp)
+	jnz   paranoid_userspace\trace
+paranoid_swapgs\trace:
+	TRACE_IRQS_IRETQ 0
+	swapgs
+paranoid_restore\trace:
+	RESTORE_ALL 8
+	iretq
+paranoid_userspace\trace:
+	GET_THREAD_INFO(%rcx)
+	movl threadinfo_flags(%rcx),%ebx
+	andl $_TIF_WORK_MASK,%ebx
+	jz paranoid_swapgs\trace
+	movq %rsp,%rdi			/* &pt_regs */
+	call sync_regs
+	movq %rax,%rsp			/* switch stack for scheduling */
+	testl $_TIF_NEED_RESCHED,%ebx
+	jnz paranoid_schedule\trace
+	movl %ebx,%edx			/* arg3: thread flags */
+	.if \trace
+	TRACE_IRQS_ON
+	.endif
+	sti
+	xorl %esi,%esi 			/* arg2: oldset */
+	movq %rsp,%rdi 			/* arg1: &pt_regs */
+	call do_notify_resume
+	cli
+	.if \trace
+	TRACE_IRQS_OFF
+	.endif
+	jmp paranoid_userspace\trace
+paranoid_schedule\trace:
+	.if \trace
+	TRACE_IRQS_ON
+	.endif
+	sti
+	call schedule
+	cli
+	.if \trace
+	TRACE_IRQS_OFF
+	.endif
+	jmp paranoid_userspace\trace
+	CFI_ENDPROC
+	.endm
+#endif
+
+/*
+ * Exception entry point. This expects an error code/orig_rax on the stack
+ * and the exception handler in %rax.	
+ */ 		  				
+ENTRY(error_entry)
+	_frame RDI
+	CFI_REL_OFFSET rax,0
+	/* rdi slot contains rax, oldrax contains error code */
+	cld	
+	subq  $14*8,%rsp
+	CFI_ADJUST_CFA_OFFSET	(14*8)
+	movq %rsi,13*8(%rsp)
+	CFI_REL_OFFSET	rsi,RSI
+	movq 14*8(%rsp),%rsi	/* load rax from rdi slot */
+	CFI_REGISTER	rax,rsi
+	movq %rdx,12*8(%rsp)
+	CFI_REL_OFFSET	rdx,RDX
+	movq %rcx,11*8(%rsp)
+	CFI_REL_OFFSET	rcx,RCX
+	movq %rsi,10*8(%rsp)	/* store rax */ 
+	CFI_REL_OFFSET	rax,RAX
+	movq %r8, 9*8(%rsp)
+	CFI_REL_OFFSET	r8,R8
+	movq %r9, 8*8(%rsp)
+	CFI_REL_OFFSET	r9,R9
+	movq %r10,7*8(%rsp)
+	CFI_REL_OFFSET	r10,R10
+	movq %r11,6*8(%rsp)
+	CFI_REL_OFFSET	r11,R11
+	movq %rbx,5*8(%rsp) 
+	CFI_REL_OFFSET	rbx,RBX
+	movq %rbp,4*8(%rsp) 
+	CFI_REL_OFFSET	rbp,RBP
+	movq %r12,3*8(%rsp) 
+	CFI_REL_OFFSET	r12,R12
+	movq %r13,2*8(%rsp) 
+	CFI_REL_OFFSET	r13,R13
+	movq %r14,1*8(%rsp) 
+	CFI_REL_OFFSET	r14,R14
+	movq %r15,(%rsp) 
+	CFI_REL_OFFSET	r15,R15
+#if 0        
+	cmpl $__KERNEL_CS,CS(%rsp)
+	CFI_REMEMBER_STATE
+	je  error_kernelspace
+#endif        
+error_call_handler:
+	movq %rdi, RDI(%rsp)            
+	CFI_REL_OFFSET	rdi,RDI
+	movq %rsp,%rdi
+	movq ORIG_RAX(%rsp),%rsi	# get error code 
+	movq $-1,ORIG_RAX(%rsp)
+	call *%rax
+error_exit:		
+	RESTORE_REST
+/*	cli */
+	XEN_BLOCK_EVENTS(%rsi)		
+	TRACE_IRQS_OFF
+	GET_THREAD_INFO(%rcx)	
+	testb $3,CS-ARGOFFSET(%rsp)
+	jz retint_kernel
+	movl  threadinfo_flags(%rcx),%edx
+	movl  $_TIF_WORK_MASK,%edi	
+	andl  %edi,%edx
+	jnz   retint_careful
+	/*
+	 * The iret might restore flags:
+	 */
+	TRACE_IRQS_IRETQ
+	jmp   retint_restore_args
+
+#if 0
+         /*
+         * We need to re-write the logic here because we don't do iretq to 
+         * to return to user mode. It's still possible that we get trap/fault
+         * in the kernel (when accessing buffers pointed to by system calls, 
+         * for example).
+         *
+         */           
+	CFI_RESTORE_STATE
+error_kernelspace:
+	incl %ebx
+       /* There are two places in the kernel that can potentially fault with
+          usergs. Handle them here. The exception handlers after
+	   iret run with kernel gs again, so don't set the user space flag.
+	   B stepping K8s sometimes report an truncated RIP for IRET 
+	   exceptions returning to compat mode. Check for these here too. */
+	leaq iret_label(%rip),%rbp
+	cmpq %rbp,RIP(%rsp) 
+	je   error_swapgs
+	movl %ebp,%ebp	/* zero extend */
+	cmpq %rbp,RIP(%rsp) 
+	je   error_swapgs
+	cmpq $gs_change,RIP(%rsp)
+        je   error_swapgs
+	jmp  error_sti
+#endif
+	CFI_ENDPROC
+END(error_entry)
+	
+ENTRY(hypervisor_callback)
+	zeroentry do_hypervisor_callback
+END(hypervisor_callback)
+        
+/*
+ * Copied from arch/xen/i386/kernel/entry.S
+ */               
+# A note on the "critical region" in our callback handler.
+# We want to avoid stacking callback handlers due to events occurring
+# during handling of the last event. To do this, we keep events disabled
+# until we've done all processing. HOWEVER, we must enable events before
+# popping the stack frame (can't be done atomically) and so it would still
+# be possible to get enough handler activations to overflow the stack.
+# Although unlikely, bugs of that kind are hard to track down, so we'd
+# like to avoid the possibility.
+# So, on entry to the handler we detect whether we interrupted an
+# existing activation in its critical region -- if so, we pop the current
+# activation and restart the handler using the previous one.
+ENTRY(do_hypervisor_callback)   # do_hypervisor_callback(struct *pt_regs)
+	CFI_STARTPROC
+# Since we don't modify %rdi, evtchn_do_upall(struct *pt_regs) will
+# see the correct pointer to the pt_regs
+	movq %rdi, %rsp            # we don't return, adjust the stack frame
+	CFI_ENDPROC
+	CFI_DEFAULT_STACK
+11:	incl %gs:pda_irqcount
+	movq %rsp,%rbp
+	CFI_DEF_CFA_REGISTER rbp
+	cmovzq %gs:pda_irqstackptr,%rsp
+	pushq %rbp			# backlink for old unwinder
+	call evtchn_do_upcall
+	popq %rsp
+	CFI_DEF_CFA_REGISTER rsp
+	decl %gs:pda_irqcount
+	jmp  error_exit
+	CFI_ENDPROC
+END(do_hypervisor_callback)
+
+#ifdef CONFIG_X86_LOCAL_APIC
+KPROBE_ENTRY(nmi)
+	zeroentry do_nmi_callback
+ENTRY(do_nmi_callback)
+	CFI_STARTPROC
+        addq $8, %rsp
+	CFI_ENDPROC
+	CFI_DEFAULT_STACK
+        call do_nmi
+        orl  $NMI_MASK,EFLAGS(%rsp)
+        RESTORE_REST
+        XEN_BLOCK_EVENTS(%rsi)
+	TRACE_IRQS_OFF
+        GET_THREAD_INFO(%rcx)
+        jmp  retint_restore_args
+	CFI_ENDPROC
+	.previous .text
+END(nmi)
+#endif
+
+        ALIGN
+restore_all_enable_events:  
+	CFI_DEFAULT_STACK adj=1
+	TRACE_IRQS_ON
+	XEN_UNBLOCK_EVENTS(%rsi)        # %rsi is already set up...
+
+scrit:	/**** START OF CRITICAL REGION ****/
+	XEN_TEST_PENDING(%rsi)
+	CFI_REMEMBER_STATE
+	jnz  14f			# process more events if necessary...
+	XEN_PUT_VCPU_INFO(%rsi)
+        RESTORE_ARGS 0,8,0
+        HYPERVISOR_IRET 0
+        
+	CFI_RESTORE_STATE
+14:	XEN_LOCKED_BLOCK_EVENTS(%rsi)
+	XEN_PUT_VCPU_INFO(%rsi)
+	SAVE_REST
+        movq %rsp,%rdi                  # set the argument again
+	jmp  11b
+	CFI_ENDPROC
+ecrit:  /**** END OF CRITICAL REGION ****/
+# At this point, unlike on x86-32, we don't do the fixup to simplify the 
+# code and the stack frame is more complex on x86-64.
+# When the kernel is interrupted in the critical section, the kernel 
+# will do IRET in that case, and everything will be restored at that point, 
+# i.e. it just resumes from the next instruction interrupted with the same context. 
+
+# Hypervisor uses this for application faults while it executes.
+# We get here for two reasons:
+#  1. Fault while reloading DS, ES, FS or GS
+#  2. Fault while executing IRET
+# Category 1 we do not need to fix up as Xen has already reloaded all segment
+# registers that could be reloaded and zeroed the others.
+# Category 2 we fix up by killing the current process. We cannot use the
+# normal Linux return path in this case because if we use the IRET hypercall
+# to pop the stack frame we end up in an infinite loop of failsafe callbacks.
+# We distinguish between categories by comparing each saved segment register
+# with its current contents: any discrepancy means we in category 1.
+ENTRY(failsafe_callback)
+	_frame (RIP-0x30)
+	CFI_REL_OFFSET rcx, 0
+	CFI_REL_OFFSET r11, 8
+	movw %ds,%cx
+	cmpw %cx,0x10(%rsp)
+	CFI_REMEMBER_STATE
+	jne 1f
+	movw %es,%cx
+	cmpw %cx,0x18(%rsp)
+	jne 1f
+	movw %fs,%cx
+	cmpw %cx,0x20(%rsp)
+	jne 1f
+	movw %gs,%cx
+	cmpw %cx,0x28(%rsp)
+	jne 1f
+	/* All segments match their saved values => Category 2 (Bad IRET). */
+	movq (%rsp),%rcx
+	CFI_RESTORE rcx
+	movq 8(%rsp),%r11
+	CFI_RESTORE r11
+	addq $0x30,%rsp
+	CFI_ADJUST_CFA_OFFSET -0x30
+	movq $11,%rdi	/* SIGSEGV */
+	jmp do_exit			
+	CFI_RESTORE_STATE
+1:	/* Segment mismatch => Category 1 (Bad segment). Retry the IRET. */
+	movq (%rsp),%rcx
+	CFI_RESTORE rcx
+	movq 8(%rsp),%r11
+	CFI_RESTORE r11
+	addq $0x30,%rsp
+	CFI_ADJUST_CFA_OFFSET -0x30
+	pushq $0
+	CFI_ADJUST_CFA_OFFSET 8
+	SAVE_ALL
+	jmp error_exit
+	CFI_ENDPROC
+#if 0	      
+        .section __ex_table,"a"
+        .align 8
+        .quad gs_change,bad_gs
+        .previous
+        .section .fixup,"ax"
+	/* running with kernelgs */
+bad_gs: 
+/*	swapgs		*/	/* switch back to user gs */
+	xorl %eax,%eax
+        movl %eax,%gs
+        jmp  2b
+        .previous       
+#endif
+	
+/*
+ * Create a kernel thread.
+ *
+ * C extern interface:
+ *	extern long kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
+ *
+ * asm input arguments:
+ *	rdi: fn, rsi: arg, rdx: flags
+ */
+ENTRY(kernel_thread)
+	CFI_STARTPROC
+	FAKE_STACK_FRAME $child_rip
+	SAVE_ALL
+
+	# rdi: flags, rsi: usp, rdx: will be &pt_regs
+	movq %rdx,%rdi
+	orq  kernel_thread_flags(%rip),%rdi
+	movq $-1, %rsi
+	movq %rsp, %rdx
+
+	xorl %r8d,%r8d
+	xorl %r9d,%r9d
+	
+	# clone now
+	call do_fork
+	movq %rax,RAX(%rsp)
+	xorl %edi,%edi
+
+	/*
+	 * It isn't worth to check for reschedule here,
+	 * so internally to the x86_64 port you can rely on kernel_thread()
+	 * not to reschedule the child before returning, this avoids the need
+	 * of hacks for example to fork off the per-CPU idle tasks.
+         * [Hopefully no generic code relies on the reschedule -AK]	
+	 */
+	RESTORE_ALL
+	UNFAKE_STACK_FRAME
+	ret
+	CFI_ENDPROC
+ENDPROC(kernel_thread)
+	
+child_rip:
+	pushq $0		# fake return address
+	CFI_STARTPROC
+	/*
+	 * Here we are in the child and the registers are set as they were
+	 * at kernel_thread() invocation in the parent.
+	 */
+	movq %rdi, %rax
+	movq %rsi, %rdi
+	call *%rax
+	# exit
+	xorl %edi, %edi
+	call do_exit
+	CFI_ENDPROC
+ENDPROC(child_rip)
+
+/*
+ * execve(). This function needs to use IRET, not SYSRET, to set up all state properly.
+ *
+ * C extern interface:
+ *	 extern long execve(char *name, char **argv, char **envp)
+ *
+ * asm input arguments:
+ *	rdi: name, rsi: argv, rdx: envp
+ *
+ * We want to fallback into:
+ *	extern long sys_execve(char *name, char **argv,char **envp, struct pt_regs regs)
+ *
+ * do_sys_execve asm fallback arguments:
+ *	rdi: name, rsi: argv, rdx: envp, fake frame on the stack
+ */
+ENTRY(execve)
+	CFI_STARTPROC
+	FAKE_STACK_FRAME $0
+	SAVE_ALL	
+	call sys_execve
+	movq %rax, RAX(%rsp)	
+	RESTORE_REST
+	testq %rax,%rax
+	jne 1f
+        jmp int_ret_from_sys_call
+1:      RESTORE_ARGS
+	UNFAKE_STACK_FRAME
+	ret
+	CFI_ENDPROC
+ENDPROC(execve)
+
+KPROBE_ENTRY(page_fault)
+	errorentry do_page_fault
+END(page_fault)
+	.previous .text
+
+ENTRY(coprocessor_error)
+	zeroentry do_coprocessor_error
+END(coprocessor_error)
+
+ENTRY(simd_coprocessor_error)
+	zeroentry do_simd_coprocessor_error	
+END(simd_coprocessor_error)
+
+ENTRY(device_not_available)
+	zeroentry math_state_restore
+END(device_not_available)
+
+	/* runs on exception stack */
+KPROBE_ENTRY(debug)
+/* 	INTR_FRAME
+	pushq $0
+	CFI_ADJUST_CFA_OFFSET 8	*/
+	zeroentry do_debug
+/*	paranoidexit
+	CFI_ENDPROC */
+END(debug)
+	.previous .text
+
+#if 0
+	/* runs on exception stack */	
+KPROBE_ENTRY(nmi)
+	INTR_FRAME
+	pushq $-1
+	CFI_ADJUST_CFA_OFFSET 8
+	paranoidentry do_nmi, 0, 0
+#ifdef CONFIG_TRACE_IRQFLAGS
+	paranoidexit 0
+#else
+	jmp paranoid_exit1
+ 	CFI_ENDPROC
+#endif
+END(nmi)
+	.previous .text
+#endif        
+
+KPROBE_ENTRY(int3)
+/* 	INTR_FRAME
+ 	pushq $0
+ 	CFI_ADJUST_CFA_OFFSET 8 */
+ 	zeroentry do_int3
+/* 	jmp paranoid_exit1
+ 	CFI_ENDPROC */
+END(int3)
+	.previous .text
+
+ENTRY(overflow)
+	zeroentry do_overflow
+END(overflow)
+
+ENTRY(bounds)
+	zeroentry do_bounds
+END(bounds)
+
+ENTRY(invalid_op)
+	zeroentry do_invalid_op	
+END(invalid_op)
+
+ENTRY(coprocessor_segment_overrun)
+	zeroentry do_coprocessor_segment_overrun
+END(coprocessor_segment_overrun)
+
+ENTRY(reserved)
+	zeroentry do_reserved
+END(reserved)
+
+#if 0
+	/* runs on exception stack */
+ENTRY(double_fault)
+	XCPT_FRAME
+	paranoidentry do_double_fault
+	jmp paranoid_exit1
+	CFI_ENDPROC
+END(double_fault)
+#endif
+
+ENTRY(invalid_TSS)
+	errorentry do_invalid_TSS
+END(invalid_TSS)
+
+ENTRY(segment_not_present)
+	errorentry do_segment_not_present
+END(segment_not_present)
+
+	/* runs on exception stack */
+ENTRY(stack_segment)
+/*	XCPT_FRAME
+	paranoidentry do_stack_segment */
+	errorentry do_stack_segment
+/*	jmp paranoid_exit1
+	CFI_ENDPROC */
+END(stack_segment)
+
+KPROBE_ENTRY(general_protection)
+	errorentry do_general_protection
+END(general_protection)
+	.previous .text
+
+ENTRY(alignment_check)
+	errorentry do_alignment_check
+END(alignment_check)
+
+ENTRY(divide_error)
+	zeroentry do_divide_error
+END(divide_error)
+
+ENTRY(spurious_interrupt_bug)
+	zeroentry do_spurious_interrupt_bug
+END(spurious_interrupt_bug)
+
+#ifdef CONFIG_X86_MCE
+	/* runs on exception stack */
+ENTRY(machine_check)
+	INTR_FRAME
+	pushq $0
+	CFI_ADJUST_CFA_OFFSET 8	
+	paranoidentry do_machine_check
+	jmp paranoid_exit1
+	CFI_ENDPROC
+END(machine_check)
+#endif
+
+/* Call softirq on interrupt stack. Interrupts are off. */
+ENTRY(call_softirq)
+	CFI_STARTPROC
+	push %rbp
+	CFI_ADJUST_CFA_OFFSET	8
+	CFI_REL_OFFSET rbp,0
+	mov  %rsp,%rbp
+	CFI_DEF_CFA_REGISTER rbp
+	incl %gs:pda_irqcount
+	cmove %gs:pda_irqstackptr,%rsp
+	push  %rbp			# backlink for old unwinder
+	call __do_softirq
+	leaveq
+	CFI_DEF_CFA_REGISTER	rsp
+	CFI_ADJUST_CFA_OFFSET   -8
+	decl %gs:pda_irqcount
+	ret
+	CFI_ENDPROC
+ENDPROC(call_softirq)
+
+#ifdef CONFIG_STACK_UNWIND
+ENTRY(arch_unwind_init_running)
+	CFI_STARTPROC
+	movq	%r15, R15(%rdi)
+	movq	%r14, R14(%rdi)
+	xchgq	%rsi, %rdx
+	movq	%r13, R13(%rdi)
+	movq	%r12, R12(%rdi)
+	xorl	%eax, %eax
+	movq	%rbp, RBP(%rdi)
+	movq	%rbx, RBX(%rdi)
+	movq	(%rsp), %rcx
+	movq	%rax, R11(%rdi)
+	movq	%rax, R10(%rdi)
+	movq	%rax, R9(%rdi)
+	movq	%rax, R8(%rdi)
+	movq	%rax, RAX(%rdi)
+	movq	%rax, RCX(%rdi)
+	movq	%rax, RDX(%rdi)
+	movq	%rax, RSI(%rdi)
+	movq	%rax, RDI(%rdi)
+	movq	%rax, ORIG_RAX(%rdi)
+	movq	%rcx, RIP(%rdi)
+	leaq	8(%rsp), %rcx
+	movq	$__KERNEL_CS, CS(%rdi)
+	movq	%rax, EFLAGS(%rdi)
+	movq	%rcx, RSP(%rdi)
+	movq	$__KERNEL_DS, SS(%rdi)
+	jmpq	*%rdx
+	CFI_ENDPROC
+ENDPROC(arch_unwind_init_running)
+#endif
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/genapic_xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/genapic_xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/genapic_xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/genapic_xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,161 @@
+/*
+ * Copyright 2004 James Cleverdon, IBM.
+ * Subject to the GNU Public License, v.2
+ *
+ * Xen APIC subarch code.  Maximum 8 CPUs, logical delivery.
+ *
+ * Hacked for x86-64 by James Cleverdon from i386 architecture code by
+ * Martin Bligh, Andi Kleen, James Bottomley, John Stultz, and
+ * James Cleverdon.
+ *
+ * Hacked to pieces for Xen by Chris Wright.
+ */
+#include <linux/threads.h>
+#include <linux/cpumask.h>
+#include <linux/string.h>
+#include <linux/kernel.h>
+#include <linux/ctype.h>
+#include <linux/init.h>
+#ifdef CONFIG_XEN_PRIVILEGED_GUEST
+#include <asm/smp.h>
+#include <asm/ipi.h>
+#else
+#include <asm/apic.h>
+#include <asm/apicdef.h>
+#include <asm/genapic.h>
+#endif
+#include <xen/evtchn.h>
+
+DECLARE_PER_CPU(int, ipi_to_irq[NR_IPIS]);
+
+static inline void __send_IPI_one(unsigned int cpu, int vector)
+{
+	int irq = per_cpu(ipi_to_irq, cpu)[vector];
+	BUG_ON(irq < 0);
+	notify_remote_via_irq(irq);
+}
+
+void xen_send_IPI_shortcut(unsigned int shortcut, int vector, unsigned int dest)
+{
+	int cpu;
+
+	switch (shortcut) {
+	case APIC_DEST_SELF:
+		__send_IPI_one(smp_processor_id(), vector);
+		break;
+	case APIC_DEST_ALLBUT:
+		for (cpu = 0; cpu < NR_CPUS; ++cpu) {
+			if (cpu == smp_processor_id())
+				continue;
+			if (cpu_isset(cpu, cpu_online_map)) {
+				__send_IPI_one(cpu, vector);
+			}
+		}
+		break;
+	case APIC_DEST_ALLINC:
+		for (cpu = 0; cpu < NR_CPUS; ++cpu) {
+			if (cpu_isset(cpu, cpu_online_map)) {
+				__send_IPI_one(cpu, vector);
+			}
+		}
+		break;
+	default:
+		printk("XXXXXX __send_IPI_shortcut %08x vector %d\n", shortcut,
+		       vector);
+		break;
+	}
+}
+
+static cpumask_t xen_target_cpus(void)
+{
+	return cpu_online_map;
+}
+
+/*
+ * Set up the logical destination ID.
+ * Do nothing, not called now.
+ */
+static void xen_init_apic_ldr(void)
+{
+	Dprintk("%s\n", __FUNCTION__);
+	return;
+}
+
+static void xen_send_IPI_allbutself(int vector)
+{
+	/*
+	 * if there are no other CPUs in the system then
+	 * we get an APIC send error if we try to broadcast.
+	 * thus we have to avoid sending IPIs in this case.
+	 */
+	Dprintk("%s\n", __FUNCTION__);
+	if (num_online_cpus() > 1)
+		xen_send_IPI_shortcut(APIC_DEST_ALLBUT, vector, APIC_DEST_LOGICAL);
+}
+
+static void xen_send_IPI_all(int vector)
+{
+	Dprintk("%s\n", __FUNCTION__);
+	xen_send_IPI_shortcut(APIC_DEST_ALLINC, vector, APIC_DEST_LOGICAL);
+}
+
+static void xen_send_IPI_mask(cpumask_t cpumask, int vector)
+{
+	unsigned long mask = cpus_addr(cpumask)[0];
+	unsigned int cpu;
+	unsigned long flags;
+
+	Dprintk("%s\n", __FUNCTION__);
+	local_irq_save(flags);
+	WARN_ON(mask & ~cpus_addr(cpu_online_map)[0]);
+
+	for (cpu = 0; cpu < NR_CPUS; ++cpu) {
+		if (cpu_isset(cpu, cpumask)) {
+			__send_IPI_one(cpu, vector);
+		}
+	}
+	local_irq_restore(flags);
+}
+
+#ifdef CONFIG_XEN_PRIVILEGED_GUEST
+static int xen_apic_id_registered(void)
+{
+	/* better be set */
+	Dprintk("%s\n", __FUNCTION__);
+	return physid_isset(smp_processor_id(), phys_cpu_present_map);
+}
+#endif
+
+static unsigned int xen_cpu_mask_to_apicid(cpumask_t cpumask)
+{
+	Dprintk("%s\n", __FUNCTION__);
+	return cpus_addr(cpumask)[0] & APIC_ALL_CPUS;
+}
+
+static unsigned int phys_pkg_id(int index_msb)
+{
+	u32 ebx;
+
+	Dprintk("%s\n", __FUNCTION__);
+	ebx = cpuid_ebx(1);
+	return ((ebx >> 24) & 0xFF) >> index_msb;
+}
+
+struct genapic apic_xen =  {
+	.name = "xen",
+#ifdef CONFIG_XEN_PRIVILEGED_GUEST
+	.int_delivery_mode = dest_LowestPrio,
+#endif
+	.int_dest_mode = (APIC_DEST_LOGICAL != 0),
+	.int_delivery_dest = APIC_DEST_LOGICAL | APIC_DM_LOWEST,
+	.target_cpus = xen_target_cpus,
+#ifdef CONFIG_XEN_PRIVILEGED_GUEST
+	.apic_id_registered = xen_apic_id_registered,
+#endif
+	.init_apic_ldr = xen_init_apic_ldr,
+	.send_IPI_all = xen_send_IPI_all,
+	.send_IPI_allbutself = xen_send_IPI_allbutself,
+	.send_IPI_mask = xen_send_IPI_mask,
+	.cpu_mask_to_apicid = xen_cpu_mask_to_apicid,
+	.phys_pkg_id = phys_pkg_id,
+};
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/genapic-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/genapic-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/genapic-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/genapic-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,143 @@
+/*
+ * Copyright 2004 James Cleverdon, IBM.
+ * Subject to the GNU Public License, v.2
+ *
+ * Generic APIC sub-arch probe layer.
+ *
+ * Hacked for x86-64 by James Cleverdon from i386 architecture code by
+ * Martin Bligh, Andi Kleen, James Bottomley, John Stultz, and
+ * James Cleverdon.
+ */
+#include <linux/threads.h>
+#include <linux/cpumask.h>
+#include <linux/string.h>
+#include <linux/kernel.h>
+#include <linux/ctype.h>
+#include <linux/init.h>
+#include <linux/module.h>
+
+#include <asm/smp.h>
+#include <asm/ipi.h>
+
+#if defined(CONFIG_ACPI)
+#include <acpi/acpi_bus.h>
+#endif
+
+/* which logical CPU number maps to which CPU (physical APIC ID) */
+u8 x86_cpu_to_apicid[NR_CPUS] __read_mostly = { [0 ... NR_CPUS-1] = BAD_APICID };
+EXPORT_SYMBOL(x86_cpu_to_apicid);
+u8 x86_cpu_to_log_apicid[NR_CPUS] = { [0 ... NR_CPUS-1] = BAD_APICID };
+
+extern struct genapic apic_cluster;
+extern struct genapic apic_flat;
+extern struct genapic apic_physflat;
+
+#ifndef CONFIG_XEN
+struct genapic *genapic = &apic_flat;
+#else
+extern struct genapic apic_xen;
+struct genapic *genapic = &apic_xen;
+#endif
+
+
+/*
+ * Check the APIC IDs in bios_cpu_apicid and choose the APIC mode.
+ */
+void __init clustered_apic_check(void)
+{
+#ifndef CONFIG_XEN
+	long i;
+	u8 clusters, max_cluster;
+	u8 id;
+	u8 cluster_cnt[NUM_APIC_CLUSTERS];
+	int max_apic = 0;
+
+#if defined(CONFIG_ACPI)
+	/*
+	 * Some x86_64 machines use physical APIC mode regardless of how many
+	 * procs/clusters are present (x86_64 ES7000 is an example).
+	 */
+	if (acpi_fadt.revision > FADT2_REVISION_ID)
+		if (acpi_fadt.force_apic_physical_destination_mode) {
+			genapic = &apic_cluster;
+			goto print;
+		}
+#endif
+
+	memset(cluster_cnt, 0, sizeof(cluster_cnt));
+	for (i = 0; i < NR_CPUS; i++) {
+		id = bios_cpu_apicid[i];
+		if (id == BAD_APICID)
+			continue;
+		if (id > max_apic)
+			max_apic = id;
+		cluster_cnt[APIC_CLUSTERID(id)]++;
+	}
+
+	/* Don't use clustered mode on AMD platforms. */
+ 	if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD) {
+		genapic = &apic_physflat;
+#ifndef CONFIG_HOTPLUG_CPU
+		/* In the CPU hotplug case we cannot use broadcast mode
+		   because that opens a race when a CPU is removed.
+		   Stay at physflat mode in this case.
+		   It is bad to do this unconditionally though. Once
+		   we have ACPI platform support for CPU hotplug
+		   we should detect hotplug capablity from ACPI tables and
+		   only do this when really needed. -AK */
+		if (max_apic <= 8)
+			genapic = &apic_flat;
+#endif
+ 		goto print;
+ 	}
+
+	clusters = 0;
+	max_cluster = 0;
+
+	for (i = 0; i < NUM_APIC_CLUSTERS; i++) {
+		if (cluster_cnt[i] > 0) {
+			++clusters;
+			if (cluster_cnt[i] > max_cluster)
+				max_cluster = cluster_cnt[i];
+		}
+	}
+
+	/*
+	 * If we have clusters <= 1 and CPUs <= 8 in cluster 0, then flat mode,
+	 * else if max_cluster <= 4 and cluster_cnt[15] == 0, clustered logical
+	 * else physical mode.
+	 * (We don't use lowest priority delivery + HW APIC IRQ steering, so
+	 * can ignore the clustered logical case and go straight to physical.)
+	 */
+	if (clusters <= 1 && max_cluster <= 8 && cluster_cnt[0] == max_cluster) {
+#ifdef CONFIG_HOTPLUG_CPU
+		/* Don't use APIC shortcuts in CPU hotplug to avoid races */
+		genapic = &apic_physflat;
+#else
+		genapic = &apic_flat;
+#endif
+	} else
+		genapic = &apic_cluster;
+
+print:
+#else
+	/* hardcode to xen apic functions */
+	genapic = &apic_xen;
+#endif
+	printk(KERN_INFO "Setting APIC routing to %s\n", genapic->name);
+}
+
+/* Same for both flat and clustered. */
+
+#ifdef CONFIG_XEN
+extern void xen_send_IPI_shortcut(unsigned int shortcut, int vector, unsigned int dest);
+#endif
+
+void send_IPI_self(int vector)
+{
+#ifndef CONFIG_XEN
+	__send_IPI_shortcut(APIC_DEST_SELF, vector, APIC_DEST_PHYSICAL);
+#else
+	xen_send_IPI_shortcut(APIC_DEST_SELF, vector, APIC_DEST_PHYSICAL);
+#endif
+}
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/head64-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/head64-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/head64-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/head64-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,162 @@
+/*
+ *  linux/arch/x86_64/kernel/head64.c -- prepare to run common code
+ *
+ *  Copyright (C) 2000 Andrea Arcangeli <andrea@suse.de> SuSE
+ *
+ *  Jun Nakajima <jun.nakajima@intel.com>
+ *	Modified for Xen.
+ */
+
+#include <linux/init.h>
+#include <linux/linkage.h>
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/string.h>
+#include <linux/percpu.h>
+#include <linux/module.h>
+
+#include <asm/processor.h>
+#include <asm/proto.h>
+#include <asm/smp.h>
+#include <asm/bootsetup.h>
+#include <asm/setup.h>
+#include <asm/desc.h>
+#include <asm/pgtable.h>
+#include <asm/sections.h>
+
+unsigned long start_pfn;
+
+/* Don't add a printk in there. printk relies on the PDA which is not initialized 
+   yet. */
+#if 0
+static void __init clear_bss(void)
+{
+	memset(__bss_start, 0,
+	       (unsigned long) __bss_stop - (unsigned long) __bss_start);
+}
+#endif
+
+#define NEW_CL_POINTER		0x228	/* Relative to real mode data */
+#define OLD_CL_MAGIC_ADDR	0x90020
+#define OLD_CL_MAGIC            0xA33F
+#define OLD_CL_BASE_ADDR        0x90000
+#define OLD_CL_OFFSET           0x90022
+
+extern char saved_command_line[];
+
+static void __init copy_bootdata(char *real_mode_data)
+{
+#ifndef CONFIG_XEN
+	int new_data;
+	char * command_line;
+
+	memcpy(x86_boot_params, real_mode_data, BOOT_PARAM_SIZE);
+	new_data = *(int *) (x86_boot_params + NEW_CL_POINTER);
+	if (!new_data) {
+		if (OLD_CL_MAGIC != * (u16 *) OLD_CL_MAGIC_ADDR) {
+			printk("so old bootloader that it does not support commandline?!\n");
+			return;
+		}
+		new_data = OLD_CL_BASE_ADDR + * (u16 *) OLD_CL_OFFSET;
+		printk("old bootloader convention, maybe loadlin?\n");
+	}
+	command_line = (char *) ((u64)(new_data));
+	memcpy(saved_command_line, command_line, COMMAND_LINE_SIZE);
+#else
+	int max_cmdline;
+	
+	if ((max_cmdline = MAX_GUEST_CMDLINE) > COMMAND_LINE_SIZE)
+		max_cmdline = COMMAND_LINE_SIZE;
+	memcpy(saved_command_line, xen_start_info->cmd_line, max_cmdline);
+	saved_command_line[max_cmdline-1] = '\0';
+#endif
+	printk("Bootdata ok (command line is %s)\n", saved_command_line);
+}
+
+static void __init setup_boot_cpu_data(void)
+{
+	unsigned int dummy, eax;
+
+	/* get vendor info */
+	cpuid(0, (unsigned int *)&boot_cpu_data.cpuid_level,
+	      (unsigned int *)&boot_cpu_data.x86_vendor_id[0],
+	      (unsigned int *)&boot_cpu_data.x86_vendor_id[8],
+	      (unsigned int *)&boot_cpu_data.x86_vendor_id[4]);
+
+	/* get cpu type */
+	cpuid(1, &eax, &dummy, &dummy,
+		(unsigned int *) &boot_cpu_data.x86_capability);
+	boot_cpu_data.x86 = (eax >> 8) & 0xf;
+	boot_cpu_data.x86_model = (eax >> 4) & 0xf;
+	boot_cpu_data.x86_mask = eax & 0xf;
+}
+
+#include <xen/interface/memory.h>
+unsigned long *machine_to_phys_mapping;
+EXPORT_SYMBOL(machine_to_phys_mapping);
+unsigned int machine_to_phys_order;
+EXPORT_SYMBOL(machine_to_phys_order);
+
+void __init x86_64_start_kernel(char * real_mode_data)
+{
+	struct xen_machphys_mapping mapping;
+	unsigned long machine_to_phys_nr_ents;
+	char *s;
+	int i;
+
+	setup_xen_features();
+
+	xen_start_info = (struct start_info *)real_mode_data;
+	if (!xen_feature(XENFEAT_auto_translated_physmap))
+		phys_to_machine_mapping =
+			(unsigned long *)xen_start_info->mfn_list;
+	start_pfn = (__pa(xen_start_info->pt_base) >> PAGE_SHIFT) +
+		xen_start_info->nr_pt_frames;
+
+	machine_to_phys_mapping = (unsigned long *)MACH2PHYS_VIRT_START;
+	machine_to_phys_nr_ents = MACH2PHYS_NR_ENTRIES;
+	if (HYPERVISOR_memory_op(XENMEM_machphys_mapping, &mapping) == 0) {
+		machine_to_phys_mapping = (unsigned long *)mapping.v_start;
+		machine_to_phys_nr_ents = mapping.max_mfn + 1;
+	}
+	while ((1UL << machine_to_phys_order) < machine_to_phys_nr_ents )
+		machine_to_phys_order++;
+
+#if 0
+	for (i = 0; i < 256; i++)
+		set_intr_gate(i, early_idt_handler);
+	asm volatile("lidt %0" :: "m" (idt_descr));
+#endif
+
+	/*
+	 * This must be called really, really early:
+	 */
+	lockdep_init();
+
+ 	for (i = 0; i < NR_CPUS; i++)
+ 		cpu_pda(i) = &boot_cpu_pda[i];
+
+	pda_init(0);
+	copy_bootdata(real_mode_data);
+#ifdef CONFIG_SMP
+	cpu_set(0, cpu_online_map);
+#endif
+	s = strstr(saved_command_line, "earlyprintk=");
+	if (s != NULL)
+		setup_early_printk(strchr(s, '=') + 1);
+#ifdef CONFIG_NUMA
+	s = strstr(saved_command_line, "numa=");
+	if (s != NULL)
+		numa_setup(s+5);
+#endif
+#ifdef CONFIG_X86_IO_APIC
+	if (strstr(saved_command_line, "disableapic"))
+		disable_apic = 1;
+#endif
+	/* You need early console to see that */
+	if (__pa_symbol(&_end) >= KERNEL_TEXT_SIZE)
+		panic("Kernel too big for kernel mapping\n");
+
+	setup_boot_cpu_data();
+	start_kernel();
+}
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/head-xen.S linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/head-xen.S
--- linux-2.6.18.8/arch/x86_64/kernel/head-xen.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/head-xen.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,214 @@
+/*
+ *  linux/arch/x86_64/kernel/head.S -- start in 32bit and switch to 64bit
+ *
+ *  Copyright (C) 2000 Andrea Arcangeli <andrea@suse.de> SuSE
+ *  Copyright (C) 2000 Pavel Machek <pavel@suse.cz>
+ *  Copyright (C) 2000 Karsten Keil <kkeil@suse.de>
+ *  Copyright (C) 2001,2002 Andi Kleen <ak@suse.de>
+ *
+ *  $Id: head.S,v 1.49 2002/03/19 17:39:25 ak Exp $
+ *
+ *  Jun Nakajima <jun.nakajima@intel.com>
+ *    Modified for Xen                                
+ */
+
+
+#include <linux/linkage.h>
+#include <linux/threads.h>
+#include <linux/init.h>
+#include <linux/elfnote.h>
+#include <asm/desc.h>
+#include <asm/segment.h>
+#include <asm/page.h>
+#include <asm/msr.h>
+#include <asm/cache.h>
+#include <asm/dwarf2.h>
+#include <xen/interface/elfnote.h>
+
+	.section .bootstrap.text, "ax", @progbits
+	.code64
+	.globl startup_64
+startup_64:
+	movq $(init_thread_union+THREAD_SIZE-8),%rsp
+
+	/* rsi is pointer to startup info structure.
+	   pass it to C */
+	movq %rsi,%rdi
+	pushq $0		# fake return address
+	jmp x86_64_start_kernel
+
+#ifdef CONFIG_ACPI_SLEEP
+.org 0xf00
+	.globl pGDT32
+pGDT32:
+	.word	gdt_end-cpu_gdt_table-1
+	.long	cpu_gdt_table-__START_KERNEL_map
+#endif
+ENTRY(stext)
+ENTRY(_stext)
+
+	$page = 0
+#define NEXT_PAGE(name) \
+	$page = $page + 1; \
+	.org $page * 0x1000; \
+	phys_##name = $page * 0x1000 + __PHYSICAL_START; \
+ENTRY(name)
+
+NEXT_PAGE(init_level4_pgt)
+	/* This gets initialized in x86_64_start_kernel */
+	.fill	512,8,0
+NEXT_PAGE(init_level4_user_pgt)
+        /*
+         * We update two pgd entries to make kernel and user pgd consistent
+         * at pgd_populate(). It can be used for kernel modules. So we place 
+         * this page here for those cases to avoid memory corruption.
+         * We also use this page to establish the initial mapping for the
+         * vsyscall area.
+         */
+	.fill	512,8,0
+
+NEXT_PAGE(level3_kernel_pgt)
+	.fill	512,8,0
+
+        /*
+         * This is used for vsyscall area mapping as we have a different
+         * level4 page table for user.
+         */
+NEXT_PAGE(level3_user_pgt)
+        .fill	512,8,0
+
+NEXT_PAGE(level2_kernel_pgt)
+	.fill	512,8,0
+
+NEXT_PAGE(hypercall_page)
+	CFI_STARTPROC
+	.rept 0x1000 / 0x20
+	.skip 1 /* push %rcx */
+	CFI_ADJUST_CFA_OFFSET	8
+	CFI_REL_OFFSET	rcx,0
+	.skip 2 /* push %r11 */
+	CFI_ADJUST_CFA_OFFSET	8
+	CFI_REL_OFFSET	rcx,0
+	.skip 5 /* mov $#,%eax */
+	.skip 2 /* syscall */
+	.skip 2 /* pop %r11 */
+	CFI_ADJUST_CFA_OFFSET -8
+	CFI_RESTORE r11
+	.skip 1 /* pop %rcx */
+	CFI_ADJUST_CFA_OFFSET -8
+	CFI_RESTORE rcx
+	.align 0x20,0 /* ret */
+	.endr
+	CFI_ENDPROC
+
+#undef NEXT_PAGE
+
+	.data
+/* Just dummy symbol to allow compilation. Not used in sleep path */
+#ifdef CONFIG_ACPI_SLEEP
+	.align PAGE_SIZE
+ENTRY(wakeup_level4_pgt)
+	.fill	512,8,0
+#endif
+
+	.data
+
+	.align 16
+	.globl cpu_gdt_descr
+cpu_gdt_descr:
+	.word	gdt_end-cpu_gdt_table-1
+gdt:
+	.quad	cpu_gdt_table
+#ifdef CONFIG_SMP
+	.rept	NR_CPUS-1
+	.word	0
+	.quad	0
+	.endr
+#endif
+
+/* We need valid kernel segments for data and code in long mode too
+ * IRET will check the segment types  kkeil 2000/10/28
+ * Also sysret mandates a special GDT layout 
+ */
+		 		
+	.section .data.page_aligned, "aw"
+	.align PAGE_SIZE
+
+/* The TLS descriptors are currently at a different place compared to i386.
+   Hopefully nobody expects them at a fixed place (Wine?) */
+
+ENTRY(cpu_gdt_table)
+	.quad	0x0000000000000000	/* NULL descriptor */
+	.quad	0x0			/* unused */
+	.quad	0x00af9a000000ffff	/* __KERNEL_CS */
+	.quad	0x00cf92000000ffff	/* __KERNEL_DS */
+	.quad	0x00cffa000000ffff	/* __USER32_CS */
+	.quad	0x00cff2000000ffff	/* __USER_DS, __USER32_DS  */
+	.quad	0x00affa000000ffff	/* __USER_CS */
+	.quad	0x00cf9a000000ffff	/* __KERNEL32_CS */
+	.quad	0,0			/* TSS */
+	.quad	0,0			/* LDT */
+	.quad   0,0,0			/* three TLS descriptors */
+	.quad	0			/* unused */
+gdt_end:
+	/* asm/segment.h:GDT_ENTRIES must match this */
+	/* This should be a multiple of the cache line size */
+	/* GDTs of other CPUs are now dynamically allocated */
+
+	/* zero the remaining page */
+	.fill PAGE_SIZE / 8 - GDT_ENTRIES,8,0
+
+	.section .bss.page_aligned, "aw", @nobits
+	.align PAGE_SIZE
+ENTRY(empty_zero_page)
+	.skip PAGE_SIZE
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+/*
+ * __xen_guest information
+ */
+.macro utoh value
+ .if (\value) < 0 || (\value) >= 0x10
+	utoh (((\value)>>4)&0x0fffffffffffffff)
+ .endif
+ .if ((\value) & 0xf) < 10
+  .byte '0' + ((\value) & 0xf)
+ .else
+  .byte 'A' + ((\value) & 0xf) - 10
+ .endif
+.endm
+
+.section __xen_guest
+	.ascii	"GUEST_OS=linux,GUEST_VER=2.6"
+	.ascii	",XEN_VER=xen-3.0"
+	.ascii	",VIRT_BASE=0x"
+		utoh __START_KERNEL_map
+	.ascii	",ELF_PADDR_OFFSET=0x"
+		utoh __START_KERNEL_map
+	.ascii	",VIRT_ENTRY=0x"
+		utoh (__START_KERNEL_map + __PHYSICAL_START)
+	.ascii	",HYPERCALL_PAGE=0x"
+		utoh (phys_hypercall_page >> PAGE_SHIFT)
+	.ascii  ",FEATURES=writable_page_tables"
+	.ascii		 "|writable_descriptor_tables"
+	.ascii		 "|auto_translated_physmap"
+	.ascii	         "|supervisor_mode_kernel"
+	.ascii	",LOADER=generic"
+	.byte	0
+#endif /* CONFIG_XEN_COMPAT <= 0x030002 */
+	
+	ELFNOTE(Xen, XEN_ELFNOTE_GUEST_OS,       .asciz, "linux")
+	ELFNOTE(Xen, XEN_ELFNOTE_GUEST_VERSION,  .asciz, "2.6")
+	ELFNOTE(Xen, XEN_ELFNOTE_XEN_VERSION,    .asciz, "xen-3.0")
+	ELFNOTE(Xen, XEN_ELFNOTE_VIRT_BASE,      .quad,  __START_KERNEL_map)
+#if CONFIG_XEN_COMPAT <= 0x030002
+	ELFNOTE(Xen, XEN_ELFNOTE_PADDR_OFFSET,   .quad,  __START_KERNEL_map)
+#else
+	ELFNOTE(Xen, XEN_ELFNOTE_PADDR_OFFSET,   .quad,  0)
+#endif
+	ELFNOTE(Xen, XEN_ELFNOTE_ENTRY,          .quad,  startup_64)
+	ELFNOTE(Xen, XEN_ELFNOTE_HYPERCALL_PAGE, .quad,  hypercall_page)
+	ELFNOTE(Xen, XEN_ELFNOTE_L1_MFN_VALID,   .quad,  _PAGE_PRESENT,_PAGE_PRESENT)
+	ELFNOTE(Xen, XEN_ELFNOTE_FEATURES,       .asciz, "writable_page_tables|writable_descriptor_tables|auto_translated_physmap|pae_pgdir_above_4gb|supervisor_mode_kernel")
+	ELFNOTE(Xen, XEN_ELFNOTE_LOADER,         .asciz, "generic")
+	ELFNOTE(Xen, XEN_ELFNOTE_SUSPEND_CANCEL, .long,  1)
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/init_task.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/init_task.c
--- linux-2.6.18.8/arch/x86_64/kernel/init_task.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/init_task.c	2008-08-21 11:36:07.000000000 +0200
@@ -37,6 +37,8 @@ union thread_union init_thread_union 
 struct task_struct init_task = INIT_TASK(init_task);
 
 EXPORT_SYMBOL(init_task);
+
+#ifndef CONFIG_X86_NO_TSS
 /*
  * per-CPU TSS segments. Threads are completely 'soft' on Linux,
  * no more per-task TSS's. The TSS size is kept cacheline-aligned
@@ -50,5 +52,6 @@ DEFINE_PER_CPU(struct tss_struct, init_t
  * debugging, no special alignment required.
  */
 DEFINE_PER_CPU(struct orig_ist, orig_ist);
+#endif
 
 #define ALIGN_TO_4K __attribute__((section(".data.init_task")))
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/io_apic-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/io_apic-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/io_apic-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/io_apic-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,2262 @@
+/*
+ *	Intel IO-APIC support for multi-Pentium hosts.
+ *
+ *	Copyright (C) 1997, 1998, 1999, 2000 Ingo Molnar, Hajnalka Szabo
+ *
+ *	Many thanks to Stig Venaas for trying out countless experimental
+ *	patches and reporting/debugging problems patiently!
+ *
+ *	(c) 1999, Multiple IO-APIC support, developed by
+ *	Ken-ichi Yaku <yaku@css1.kbnes.nec.co.jp> and
+ *      Hidemi Kishimoto <kisimoto@css1.kbnes.nec.co.jp>,
+ *	further tested and cleaned up by Zach Brown <zab@redhat.com>
+ *	and Ingo Molnar <mingo@redhat.com>
+ *
+ *	Fixes
+ *	Maciej W. Rozycki	:	Bits for genuine 82489DX APICs;
+ *					thanks to Eric Gilmore
+ *					and Rolf G. Tews
+ *					for testing these extensively
+ *	Paul Diefenbaugh	:	Added full ACPI support
+ */
+
+#include <linux/mm.h>
+#include <linux/interrupt.h>
+#include <linux/init.h>
+#include <linux/delay.h>
+#include <linux/sched.h>
+#include <linux/smp_lock.h>
+#include <linux/mc146818rtc.h>
+#include <linux/acpi.h>
+#include <linux/sysdev.h>
+#ifdef CONFIG_ACPI
+#include <acpi/acpi_bus.h>
+#endif
+
+#include <asm/io.h>
+#include <asm/smp.h>
+#include <asm/desc.h>
+#include <asm/proto.h>
+#include <asm/mach_apic.h>
+#include <asm/acpi.h>
+#include <asm/dma.h>
+#include <asm/nmi.h>
+
+#define __apicdebuginit  __init
+
+int sis_apic_bug; /* not actually supported, dummy for compile */
+
+static int no_timer_check;
+
+int disable_timer_pin_1 __initdata;
+
+#ifndef CONFIG_XEN
+int timer_over_8254 __initdata = 0;
+
+/* Where if anywhere is the i8259 connect in external int mode */
+static struct { int pin, apic; } ioapic_i8259 = { -1, -1 };
+#endif
+
+static DEFINE_SPINLOCK(ioapic_lock);
+static DEFINE_SPINLOCK(vector_lock);
+
+/*
+ * # of IRQ routing registers
+ */
+int nr_ioapic_registers[MAX_IO_APICS];
+
+/*
+ * Rough estimation of how many shared IRQs there are, can
+ * be changed anytime.
+ */
+#define MAX_PLUS_SHARED_IRQS NR_IRQ_VECTORS
+#define PIN_MAP_SIZE (MAX_PLUS_SHARED_IRQS + NR_IRQS)
+
+/*
+ * This is performance-critical, we want to do it O(1)
+ *
+ * the indexing order of this array favors 1:1 mappings
+ * between pins and IRQs.
+ */
+
+static struct irq_pin_list {
+	short apic, pin, next;
+} irq_2_pin[PIN_MAP_SIZE];
+
+int vector_irq[NR_VECTORS] __read_mostly = { [0 ... NR_VECTORS - 1] = -1};
+#ifdef CONFIG_PCI_MSI
+#define vector_to_irq(vector) 	\
+	(platform_legacy_irq(vector) ? vector : vector_irq[vector])
+#else
+#define vector_to_irq(vector)	(vector)
+#endif
+
+#ifdef CONFIG_XEN
+
+#include <xen/interface/xen.h>
+#include <xen/interface/physdev.h>
+
+/* Fake i8259 */
+#define make_8259A_irq(_irq)     (io_apic_irqs &= ~(1UL<<(_irq)))
+#define disable_8259A_irq(_irq)  ((void)0)
+#define i8259A_irq_pending(_irq) (0)
+
+unsigned long io_apic_irqs;
+
+static inline unsigned int xen_io_apic_read(unsigned int apic, unsigned int reg)
+{
+	struct physdev_apic apic_op;
+	int ret;
+
+	apic_op.apic_physbase = mp_ioapics[apic].mpc_apicaddr;
+	apic_op.reg = reg;
+	ret = HYPERVISOR_physdev_op(PHYSDEVOP_apic_read, &apic_op);
+	if (ret)
+		return ret;
+	return apic_op.value;
+}
+
+static inline void xen_io_apic_write(unsigned int apic, unsigned int reg, unsigned int value)
+{
+	struct physdev_apic apic_op;
+
+	apic_op.apic_physbase = mp_ioapics[apic].mpc_apicaddr;
+	apic_op.reg = reg;
+	apic_op.value = value;
+	WARN_ON(HYPERVISOR_physdev_op(PHYSDEVOP_apic_write, &apic_op));
+}
+
+#define io_apic_read(a,r)    xen_io_apic_read(a,r)
+#define io_apic_write(a,r,v) xen_io_apic_write(a,r,v)
+
+#define clear_IO_APIC() ((void)0)
+
+#else
+
+#ifdef CONFIG_SMP
+static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t mask)
+{
+	unsigned long flags;
+	unsigned int dest;
+	cpumask_t tmp;
+
+	cpus_and(tmp, mask, cpu_online_map);
+	if (cpus_empty(tmp))
+		tmp = TARGET_CPUS;
+
+	cpus_and(mask, tmp, CPU_MASK_ALL);
+
+	dest = cpu_mask_to_apicid(mask);
+
+	/*
+	 * Only the high 8 bits are valid.
+	 */
+	dest = SET_APIC_LOGICAL_ID(dest);
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	__DO_ACTION(1, = dest, )
+	set_irq_info(irq, mask);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+}
+#endif
+
+#endif /* !CONFIG_XEN */
+
+/*
+ * The common case is 1:1 IRQ<->pin mappings. Sometimes there are
+ * shared ISA-space IRQs, so we have to support them. We are super
+ * fast in the common case, and fast for shared ISA-space IRQs.
+ */
+static void add_pin_to_irq(unsigned int irq, int apic, int pin)
+{
+	static int first_free_entry = NR_IRQS;
+	struct irq_pin_list *entry = irq_2_pin + irq;
+
+	BUG_ON(irq >= NR_IRQS);
+	while (entry->next)
+		entry = irq_2_pin + entry->next;
+
+	if (entry->pin != -1) {
+		entry->next = first_free_entry;
+		entry = irq_2_pin + entry->next;
+		if (++first_free_entry >= PIN_MAP_SIZE)
+			panic("io_apic.c: ran out of irq_2_pin entries!");
+	}
+	entry->apic = apic;
+	entry->pin = pin;
+}
+
+#ifndef CONFIG_XEN
+#define __DO_ACTION(R, ACTION, FINAL)					\
+									\
+{									\
+	int pin;							\
+	struct irq_pin_list *entry = irq_2_pin + irq;			\
+									\
+	BUG_ON(irq >= NR_IRQS);						\
+	for (;;) {							\
+		unsigned int reg;					\
+		pin = entry->pin;					\
+		if (pin == -1)						\
+			break;						\
+		reg = io_apic_read(entry->apic, 0x10 + R + pin*2);	\
+		reg ACTION;						\
+		io_apic_modify(entry->apic, reg);			\
+		if (!entry->next)					\
+			break;						\
+		entry = irq_2_pin + entry->next;			\
+	}								\
+	FINAL;								\
+}
+
+#define DO_ACTION(name,R,ACTION, FINAL)					\
+									\
+	static void name##_IO_APIC_irq (unsigned int irq)		\
+	__DO_ACTION(R, ACTION, FINAL)
+
+DO_ACTION( __mask,             0, |= 0x00010000, io_apic_sync(entry->apic) )
+						/* mask = 1 */
+DO_ACTION( __unmask,           0, &= 0xfffeffff, )
+						/* mask = 0 */
+
+static void mask_IO_APIC_irq (unsigned int irq)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	__mask_IO_APIC_irq(irq);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+}
+
+static void unmask_IO_APIC_irq (unsigned int irq)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	__unmask_IO_APIC_irq(irq);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+}
+
+static void clear_IO_APIC_pin(unsigned int apic, unsigned int pin)
+{
+	struct IO_APIC_route_entry entry;
+	unsigned long flags;
+
+	/* Check delivery_mode to be sure we're not clearing an SMI pin */
+	spin_lock_irqsave(&ioapic_lock, flags);
+	*(((int*)&entry) + 0) = io_apic_read(apic, 0x10 + 2 * pin);
+	*(((int*)&entry) + 1) = io_apic_read(apic, 0x11 + 2 * pin);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+	if (entry.delivery_mode == dest_SMI)
+		return;
+	/*
+	 * Disable it in the IO-APIC irq-routing table:
+	 */
+	memset(&entry, 0, sizeof(entry));
+	entry.mask = 1;
+	spin_lock_irqsave(&ioapic_lock, flags);
+	io_apic_write(apic, 0x10 + 2 * pin, *(((int *)&entry) + 0));
+	io_apic_write(apic, 0x11 + 2 * pin, *(((int *)&entry) + 1));
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+}
+
+static void clear_IO_APIC (void)
+{
+	int apic, pin;
+
+	for (apic = 0; apic < nr_ioapics; apic++)
+		for (pin = 0; pin < nr_ioapic_registers[apic]; pin++)
+			clear_IO_APIC_pin(apic, pin);
+}
+
+#endif /* !CONFIG_XEN */
+
+static u8 gsi_2_irq[NR_IRQ_VECTORS] = { [0 ... NR_IRQ_VECTORS-1] = 0xFF };
+
+/*
+ * support for broken MP BIOSs, enables hand-redirection of PIRQ0-7 to
+ * specific CPU-side IRQs.
+ */
+
+#define MAX_PIRQS 8
+static int pirq_entries [MAX_PIRQS];
+static int pirqs_enabled;
+int skip_ioapic_setup;
+int ioapic_force;
+
+/* dummy parsing: see setup.c */
+
+static int __init disable_ioapic_setup(char *str)
+{
+	skip_ioapic_setup = 1;
+	return 1;
+}
+
+static int __init enable_ioapic_setup(char *str)
+{
+	ioapic_force = 1;
+	skip_ioapic_setup = 0;
+	return 1;
+}
+
+__setup("noapic", disable_ioapic_setup);
+__setup("apic", enable_ioapic_setup);
+
+#ifndef CONFIG_XEN
+static int __init setup_disable_8254_timer(char *s)
+{
+	timer_over_8254 = -1;
+	return 1;
+}
+static int __init setup_enable_8254_timer(char *s)
+{
+	timer_over_8254 = 2;
+	return 1;
+}
+
+__setup("disable_8254_timer", setup_disable_8254_timer);
+__setup("enable_8254_timer", setup_enable_8254_timer);
+#endif /* !CONFIG_XEN */
+
+#include <asm/pci-direct.h>
+#include <linux/pci_ids.h>
+#include <linux/pci.h>
+
+
+#ifdef CONFIG_ACPI
+
+static int nvidia_hpet_detected __initdata;
+
+static int __init nvidia_hpet_check(unsigned long phys, unsigned long size)
+{
+	nvidia_hpet_detected = 1;
+	return 0;
+}
+#endif
+
+/* Temporary Hack. Nvidia and VIA boards currently only work with IO-APIC
+   off. Check for an Nvidia or VIA PCI bridge and turn it off.
+   Use pci direct infrastructure because this runs before the PCI subsystem. 
+
+   Can be overwritten with "apic"
+
+   And another hack to disable the IOMMU on VIA chipsets.
+
+   ... and others. Really should move this somewhere else.
+
+   Kludge-O-Rama. */
+void __init check_ioapic(void) 
+{ 
+	int num,slot,func; 
+	/* Poor man's PCI discovery */
+	for (num = 0; num < 32; num++) { 
+		for (slot = 0; slot < 32; slot++) { 
+			for (func = 0; func < 8; func++) { 
+				u32 class;
+				u32 vendor;
+				u8 type;
+				class = read_pci_config(num,slot,func,
+							PCI_CLASS_REVISION);
+				if (class == 0xffffffff)
+					break; 
+
+		       		if ((class >> 16) != PCI_CLASS_BRIDGE_PCI)
+					continue; 
+
+				vendor = read_pci_config(num, slot, func, 
+							 PCI_VENDOR_ID);
+				vendor &= 0xffff;
+				switch (vendor) { 
+				case PCI_VENDOR_ID_VIA:
+#ifdef CONFIG_IOMMU
+					if ((end_pfn > MAX_DMA32_PFN ||
+					     force_iommu) &&
+					    !iommu_aperture_allowed) {
+						printk(KERN_INFO
+    "Looks like a VIA chipset. Disabling IOMMU. Override with \"iommu=allowed\"\n");
+						iommu_aperture_disabled = 1;
+					}
+#endif
+					return;
+				case PCI_VENDOR_ID_NVIDIA:
+#ifdef CONFIG_ACPI
+					/*
+					 * All timer overrides on Nvidia are
+					 * wrong unless HPET is enabled.
+					 */
+					nvidia_hpet_detected = 0;
+					acpi_table_parse(ACPI_HPET,
+							nvidia_hpet_check);
+					if (nvidia_hpet_detected == 0) {
+						acpi_skip_timer_override = 1;
+						printk(KERN_INFO "Nvidia board "
+						    "detected. Ignoring ACPI "
+						    "timer override.\n");
+					}
+#endif
+					/* RED-PEN skip them on mptables too? */
+					return;
+				case PCI_VENDOR_ID_ATI:
+
+				/* This should be actually default, but
+				   for 2.6.16 let's do it for ATI only where
+				   it's really needed. */
+#ifndef CONFIG_XEN
+					if (timer_over_8254 == 1) {	
+						timer_over_8254 = 0;	
+					printk(KERN_INFO
+		"ATI board detected. Disabling timer routing over 8254.\n");
+					}	
+#endif
+					return;
+				} 
+
+
+				/* No multi-function device? */
+				type = read_pci_config_byte(num,slot,func,
+							    PCI_HEADER_TYPE);
+				if (!(type & 0x80))
+					break;
+			} 
+		}
+	}
+} 
+
+static int __init ioapic_pirq_setup(char *str)
+{
+	int i, max;
+	int ints[MAX_PIRQS+1];
+
+	get_options(str, ARRAY_SIZE(ints), ints);
+
+	for (i = 0; i < MAX_PIRQS; i++)
+		pirq_entries[i] = -1;
+
+	pirqs_enabled = 1;
+	apic_printk(APIC_VERBOSE, "PIRQ redirection, working around broken MP-BIOS.\n");
+	max = MAX_PIRQS;
+	if (ints[0] < MAX_PIRQS)
+		max = ints[0];
+
+	for (i = 0; i < max; i++) {
+		apic_printk(APIC_VERBOSE, "... PIRQ%d -> IRQ %d\n", i, ints[i+1]);
+		/*
+		 * PIRQs are mapped upside down, usually.
+		 */
+		pirq_entries[MAX_PIRQS-i-1] = ints[i+1];
+	}
+	return 1;
+}
+
+__setup("pirq=", ioapic_pirq_setup);
+
+/*
+ * Find the IRQ entry number of a certain pin.
+ */
+static int find_irq_entry(int apic, int pin, int type)
+{
+	int i;
+
+	for (i = 0; i < mp_irq_entries; i++)
+		if (mp_irqs[i].mpc_irqtype == type &&
+		    (mp_irqs[i].mpc_dstapic == mp_ioapics[apic].mpc_apicid ||
+		     mp_irqs[i].mpc_dstapic == MP_APIC_ALL) &&
+		    mp_irqs[i].mpc_dstirq == pin)
+			return i;
+
+	return -1;
+}
+
+#ifndef CONFIG_XEN
+/*
+ * Find the pin to which IRQ[irq] (ISA) is connected
+ */
+static int __init find_isa_irq_pin(int irq, int type)
+{
+	int i;
+
+	for (i = 0; i < mp_irq_entries; i++) {
+		int lbus = mp_irqs[i].mpc_srcbus;
+
+		if ((mp_bus_id_to_type[lbus] == MP_BUS_ISA ||
+		     mp_bus_id_to_type[lbus] == MP_BUS_EISA ||
+		     mp_bus_id_to_type[lbus] == MP_BUS_MCA) &&
+		    (mp_irqs[i].mpc_irqtype == type) &&
+		    (mp_irqs[i].mpc_srcbusirq == irq))
+
+			return mp_irqs[i].mpc_dstirq;
+	}
+	return -1;
+}
+
+static int __init find_isa_irq_apic(int irq, int type)
+{
+	int i;
+
+	for (i = 0; i < mp_irq_entries; i++) {
+		int lbus = mp_irqs[i].mpc_srcbus;
+
+		if ((mp_bus_id_to_type[lbus] == MP_BUS_ISA ||
+		     mp_bus_id_to_type[lbus] == MP_BUS_EISA ||
+		     mp_bus_id_to_type[lbus] == MP_BUS_MCA) &&
+		    (mp_irqs[i].mpc_irqtype == type) &&
+		    (mp_irqs[i].mpc_srcbusirq == irq))
+			break;
+	}
+	if (i < mp_irq_entries) {
+		int apic;
+		for(apic = 0; apic < nr_ioapics; apic++) {
+			if (mp_ioapics[apic].mpc_apicid == mp_irqs[i].mpc_dstapic)
+				return apic;
+		}
+	}
+
+	return -1;
+}
+#endif
+
+/*
+ * Find a specific PCI IRQ entry.
+ * Not an __init, possibly needed by modules
+ */
+static int pin_2_irq(int idx, int apic, int pin);
+
+int IO_APIC_get_PCI_irq_vector(int bus, int slot, int pin)
+{
+	int apic, i, best_guess = -1;
+
+	apic_printk(APIC_DEBUG, "querying PCI -> IRQ mapping bus:%d, slot:%d, pin:%d.\n",
+		bus, slot, pin);
+	if (mp_bus_id_to_pci_bus[bus] == -1) {
+		apic_printk(APIC_VERBOSE, "PCI BIOS passed nonexistent PCI bus %d!\n", bus);
+		return -1;
+	}
+	for (i = 0; i < mp_irq_entries; i++) {
+		int lbus = mp_irqs[i].mpc_srcbus;
+
+		for (apic = 0; apic < nr_ioapics; apic++)
+			if (mp_ioapics[apic].mpc_apicid == mp_irqs[i].mpc_dstapic ||
+			    mp_irqs[i].mpc_dstapic == MP_APIC_ALL)
+				break;
+
+		if ((mp_bus_id_to_type[lbus] == MP_BUS_PCI) &&
+		    !mp_irqs[i].mpc_irqtype &&
+		    (bus == lbus) &&
+		    (slot == ((mp_irqs[i].mpc_srcbusirq >> 2) & 0x1f))) {
+			int irq = pin_2_irq(i,apic,mp_irqs[i].mpc_dstirq);
+
+			if (!(apic || IO_APIC_IRQ(irq)))
+				continue;
+
+			if (pin == (mp_irqs[i].mpc_srcbusirq & 3))
+				return irq;
+			/*
+			 * Use the first all-but-pin matching entry as a
+			 * best-guess fuzzy result for broken mptables.
+			 */
+			if (best_guess < 0)
+				best_guess = irq;
+		}
+	}
+	BUG_ON(best_guess >= NR_IRQS);
+	return best_guess;
+}
+
+/*
+ * EISA Edge/Level control register, ELCR
+ */
+static int EISA_ELCR(unsigned int irq)
+{
+	if (irq < 16) {
+		unsigned int port = 0x4d0 + (irq >> 3);
+		return (inb(port) >> (irq & 7)) & 1;
+	}
+	apic_printk(APIC_VERBOSE, "Broken MPtable reports ISA irq %d\n", irq);
+	return 0;
+}
+
+/* EISA interrupts are always polarity zero and can be edge or level
+ * trigger depending on the ELCR value.  If an interrupt is listed as
+ * EISA conforming in the MP table, that means its trigger type must
+ * be read in from the ELCR */
+
+#define default_EISA_trigger(idx)	(EISA_ELCR(mp_irqs[idx].mpc_srcbusirq))
+#define default_EISA_polarity(idx)	(0)
+
+/* ISA interrupts are always polarity zero edge triggered,
+ * when listed as conforming in the MP table. */
+
+#define default_ISA_trigger(idx)	(0)
+#define default_ISA_polarity(idx)	(0)
+
+/* PCI interrupts are always polarity one level triggered,
+ * when listed as conforming in the MP table. */
+
+#define default_PCI_trigger(idx)	(1)
+#define default_PCI_polarity(idx)	(1)
+
+/* MCA interrupts are always polarity zero level triggered,
+ * when listed as conforming in the MP table. */
+
+#define default_MCA_trigger(idx)	(1)
+#define default_MCA_polarity(idx)	(0)
+
+static int __init MPBIOS_polarity(int idx)
+{
+	int bus = mp_irqs[idx].mpc_srcbus;
+	int polarity;
+
+	/*
+	 * Determine IRQ line polarity (high active or low active):
+	 */
+	switch (mp_irqs[idx].mpc_irqflag & 3)
+	{
+		case 0: /* conforms, ie. bus-type dependent polarity */
+		{
+			switch (mp_bus_id_to_type[bus])
+			{
+				case MP_BUS_ISA: /* ISA pin */
+				{
+					polarity = default_ISA_polarity(idx);
+					break;
+				}
+				case MP_BUS_EISA: /* EISA pin */
+				{
+					polarity = default_EISA_polarity(idx);
+					break;
+				}
+				case MP_BUS_PCI: /* PCI pin */
+				{
+					polarity = default_PCI_polarity(idx);
+					break;
+				}
+				case MP_BUS_MCA: /* MCA pin */
+				{
+					polarity = default_MCA_polarity(idx);
+					break;
+				}
+				default:
+				{
+					printk(KERN_WARNING "broken BIOS!!\n");
+					polarity = 1;
+					break;
+				}
+			}
+			break;
+		}
+		case 1: /* high active */
+		{
+			polarity = 0;
+			break;
+		}
+		case 2: /* reserved */
+		{
+			printk(KERN_WARNING "broken BIOS!!\n");
+			polarity = 1;
+			break;
+		}
+		case 3: /* low active */
+		{
+			polarity = 1;
+			break;
+		}
+		default: /* invalid */
+		{
+			printk(KERN_WARNING "broken BIOS!!\n");
+			polarity = 1;
+			break;
+		}
+	}
+	return polarity;
+}
+
+static int MPBIOS_trigger(int idx)
+{
+	int bus = mp_irqs[idx].mpc_srcbus;
+	int trigger;
+
+	/*
+	 * Determine IRQ trigger mode (edge or level sensitive):
+	 */
+	switch ((mp_irqs[idx].mpc_irqflag>>2) & 3)
+	{
+		case 0: /* conforms, ie. bus-type dependent */
+		{
+			switch (mp_bus_id_to_type[bus])
+			{
+				case MP_BUS_ISA: /* ISA pin */
+				{
+					trigger = default_ISA_trigger(idx);
+					break;
+				}
+				case MP_BUS_EISA: /* EISA pin */
+				{
+					trigger = default_EISA_trigger(idx);
+					break;
+				}
+				case MP_BUS_PCI: /* PCI pin */
+				{
+					trigger = default_PCI_trigger(idx);
+					break;
+				}
+				case MP_BUS_MCA: /* MCA pin */
+				{
+					trigger = default_MCA_trigger(idx);
+					break;
+				}
+				default:
+				{
+					printk(KERN_WARNING "broken BIOS!!\n");
+					trigger = 1;
+					break;
+				}
+			}
+			break;
+		}
+		case 1: /* edge */
+		{
+			trigger = 0;
+			break;
+		}
+		case 2: /* reserved */
+		{
+			printk(KERN_WARNING "broken BIOS!!\n");
+			trigger = 1;
+			break;
+		}
+		case 3: /* level */
+		{
+			trigger = 1;
+			break;
+		}
+		default: /* invalid */
+		{
+			printk(KERN_WARNING "broken BIOS!!\n");
+			trigger = 0;
+			break;
+		}
+	}
+	return trigger;
+}
+
+static inline int irq_polarity(int idx)
+{
+	return MPBIOS_polarity(idx);
+}
+
+static inline int irq_trigger(int idx)
+{
+	return MPBIOS_trigger(idx);
+}
+
+static int next_irq = 16;
+
+/*
+ * gsi_irq_sharing -- Name overload!  "irq" can be either a legacy IRQ
+ * in the range 0-15, a linux IRQ in the range 0-223, or a GSI number
+ * from ACPI, which can reach 800 in large boxen.
+ *
+ * Compact the sparse GSI space into a sequential IRQ series and reuse
+ * vectors if possible.
+ */
+int gsi_irq_sharing(int gsi)
+{
+	int i, tries, vector;
+
+	BUG_ON(gsi >= NR_IRQ_VECTORS);
+
+	if (platform_legacy_irq(gsi))
+		return gsi;
+
+	if (gsi_2_irq[gsi] != 0xFF)
+		return (int)gsi_2_irq[gsi];
+
+	tries = NR_IRQS;
+  try_again:
+	vector = assign_irq_vector(gsi);
+
+	/*
+	 * Sharing vectors means sharing IRQs, so scan irq_vectors for previous
+	 * use of vector and if found, return that IRQ.  However, we never want
+	 * to share legacy IRQs, which usually have a different trigger mode
+	 * than PCI.
+	 */
+	for (i = 0; i < NR_IRQS; i++)
+		if (IO_APIC_VECTOR(i) == vector)
+			break;
+	if (platform_legacy_irq(i)) {
+		if (--tries >= 0) {
+			IO_APIC_VECTOR(i) = 0;
+			goto try_again;
+		}
+		panic("gsi_irq_sharing: didn't find an IRQ using vector 0x%02X for GSI %d", vector, gsi);
+	}
+	if (i < NR_IRQS) {
+		gsi_2_irq[gsi] = i;
+		printk(KERN_INFO "GSI %d sharing vector 0x%02X and IRQ %d\n",
+				gsi, vector, i);
+		return i;
+	}
+
+	i = next_irq++;
+	BUG_ON(i >= NR_IRQS);
+	gsi_2_irq[gsi] = i;
+	IO_APIC_VECTOR(i) = vector;
+	printk(KERN_INFO "GSI %d assigned vector 0x%02X and IRQ %d\n",
+			gsi, vector, i);
+	return i;
+}
+
+static int pin_2_irq(int idx, int apic, int pin)
+{
+	int irq, i;
+	int bus = mp_irqs[idx].mpc_srcbus;
+
+	/*
+	 * Debugging check, we are in big trouble if this message pops up!
+	 */
+	if (mp_irqs[idx].mpc_dstirq != pin)
+		printk(KERN_ERR "broken BIOS or MPTABLE parser, ayiee!!\n");
+
+	switch (mp_bus_id_to_type[bus])
+	{
+		case MP_BUS_ISA: /* ISA pin */
+		case MP_BUS_EISA:
+		case MP_BUS_MCA:
+		{
+			irq = mp_irqs[idx].mpc_srcbusirq;
+			break;
+		}
+		case MP_BUS_PCI: /* PCI pin */
+		{
+			/*
+			 * PCI IRQs are mapped in order
+			 */
+			i = irq = 0;
+			while (i < apic)
+				irq += nr_ioapic_registers[i++];
+			irq += pin;
+			irq = gsi_irq_sharing(irq);
+			break;
+		}
+		default:
+		{
+			printk(KERN_ERR "unknown bus type %d.\n",bus); 
+			irq = 0;
+			break;
+		}
+	}
+	BUG_ON(irq >= NR_IRQS);
+
+	/*
+	 * PCI IRQ command line redirection. Yes, limits are hardcoded.
+	 */
+	if ((pin >= 16) && (pin <= 23)) {
+		if (pirq_entries[pin-16] != -1) {
+			if (!pirq_entries[pin-16]) {
+				apic_printk(APIC_VERBOSE, "disabling PIRQ%d\n", pin-16);
+			} else {
+				irq = pirq_entries[pin-16];
+				apic_printk(APIC_VERBOSE, "using PIRQ%d -> IRQ %d\n",
+						pin-16, irq);
+			}
+		}
+	}
+	BUG_ON(irq >= NR_IRQS);
+	return irq;
+}
+
+static inline int IO_APIC_irq_trigger(int irq)
+{
+	int apic, idx, pin;
+
+	for (apic = 0; apic < nr_ioapics; apic++) {
+		for (pin = 0; pin < nr_ioapic_registers[apic]; pin++) {
+			idx = find_irq_entry(apic,pin,mp_INT);
+			if ((idx != -1) && (irq == pin_2_irq(idx,apic,pin)))
+				return irq_trigger(idx);
+		}
+	}
+	/*
+	 * nonexistent IRQs are edge default
+	 */
+	return 0;
+}
+
+/* irq_vectors is indexed by the sum of all RTEs in all I/O APICs. */
+u8 irq_vector[NR_IRQ_VECTORS] __read_mostly;
+
+int assign_irq_vector(int irq)
+{
+	unsigned long flags;
+	int vector;
+	struct physdev_irq irq_op;
+  
+  	BUG_ON(irq != AUTO_ASSIGN && (unsigned)irq >= NR_IRQ_VECTORS);
+
+	spin_lock_irqsave(&vector_lock, flags);
+
+  	if (irq != AUTO_ASSIGN && IO_APIC_VECTOR(irq) > 0) {
+		spin_unlock_irqrestore(&vector_lock, flags);
+  		return IO_APIC_VECTOR(irq);
+	}
+
+	irq_op.irq = irq;
+	if (HYPERVISOR_physdev_op(PHYSDEVOP_alloc_irq_vector, &irq_op)) {
+		spin_unlock_irqrestore(&vector_lock, flags);
+		return -ENOSPC;
+	}
+
+	vector = irq_op.vector;
+	vector_irq[vector] = irq;
+	if (irq != AUTO_ASSIGN)
+		IO_APIC_VECTOR(irq) = vector;
+
+	spin_unlock_irqrestore(&vector_lock, flags);
+
+	return vector;
+}
+
+extern void (*interrupt[NR_IRQS])(void);
+#ifndef CONFIG_XEN
+static struct hw_interrupt_type ioapic_level_type;
+static struct hw_interrupt_type ioapic_edge_type;
+
+#define IOAPIC_AUTO	-1
+#define IOAPIC_EDGE	0
+#define IOAPIC_LEVEL	1
+
+static void ioapic_register_intr(int irq, int vector, unsigned long trigger)
+{
+	unsigned idx;
+
+	idx = use_pci_vector() && !platform_legacy_irq(irq) ? vector : irq;
+
+	if ((trigger == IOAPIC_AUTO && IO_APIC_irq_trigger(irq)) ||
+			trigger == IOAPIC_LEVEL)
+		irq_desc[idx].chip = &ioapic_level_type;
+	else
+		irq_desc[idx].chip = &ioapic_edge_type;
+	set_intr_gate(vector, interrupt[idx]);
+}
+#else
+#define ioapic_register_intr(_irq,_vector,_trigger) ((void)0)
+#endif /* !CONFIG_XEN */
+
+static void __init setup_IO_APIC_irqs(void)
+{
+	struct IO_APIC_route_entry entry;
+	int apic, pin, idx, irq, first_notcon = 1, vector;
+	unsigned long flags;
+
+	apic_printk(APIC_VERBOSE, KERN_DEBUG "init IO_APIC IRQs\n");
+
+	for (apic = 0; apic < nr_ioapics; apic++) {
+	for (pin = 0; pin < nr_ioapic_registers[apic]; pin++) {
+
+		/*
+		 * add it to the IO-APIC irq-routing table:
+		 */
+		memset(&entry,0,sizeof(entry));
+
+		entry.delivery_mode = INT_DELIVERY_MODE;
+		entry.dest_mode = INT_DEST_MODE;
+		entry.mask = 0;				/* enable IRQ */
+		entry.dest.logical.logical_dest = cpu_mask_to_apicid(TARGET_CPUS);
+
+		idx = find_irq_entry(apic,pin,mp_INT);
+		if (idx == -1) {
+			if (first_notcon) {
+				apic_printk(APIC_VERBOSE, KERN_DEBUG " IO-APIC (apicid-pin) %d-%d", mp_ioapics[apic].mpc_apicid, pin);
+				first_notcon = 0;
+			} else
+				apic_printk(APIC_VERBOSE, ", %d-%d", mp_ioapics[apic].mpc_apicid, pin);
+			continue;
+		}
+
+		entry.trigger = irq_trigger(idx);
+		entry.polarity = irq_polarity(idx);
+
+		if (irq_trigger(idx)) {
+			entry.trigger = 1;
+			entry.mask = 1;
+			entry.dest.logical.logical_dest = cpu_mask_to_apicid(TARGET_CPUS);
+		}
+
+		irq = pin_2_irq(idx, apic, pin);
+		add_pin_to_irq(irq, apic, pin);
+
+		if (/* !apic && */ !IO_APIC_IRQ(irq))
+			continue;
+
+		if (IO_APIC_IRQ(irq)) {
+			vector = assign_irq_vector(irq);
+			entry.vector = vector;
+
+			ioapic_register_intr(irq, vector, IOAPIC_AUTO);
+			if (!apic && (irq < 16))
+				disable_8259A_irq(irq);
+		}
+		spin_lock_irqsave(&ioapic_lock, flags);
+		io_apic_write(apic, 0x11+2*pin, *(((int *)&entry)+1));
+		io_apic_write(apic, 0x10+2*pin, *(((int *)&entry)+0));
+		set_native_irq_info(irq, TARGET_CPUS);
+		spin_unlock_irqrestore(&ioapic_lock, flags);
+	}
+	}
+
+	if (!first_notcon)
+		apic_printk(APIC_VERBOSE," not connected.\n");
+}
+
+#ifndef CONFIG_XEN
+/*
+ * Set up the 8259A-master output pin as broadcast to all
+ * CPUs.
+ */
+static void __init setup_ExtINT_IRQ0_pin(unsigned int apic, unsigned int pin, int vector)
+{
+	struct IO_APIC_route_entry entry;
+	unsigned long flags;
+
+	memset(&entry,0,sizeof(entry));
+
+	disable_8259A_irq(0);
+
+	/* mask LVT0 */
+	apic_write(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_EXTINT);
+
+	/*
+	 * We use logical delivery to get the timer IRQ
+	 * to the first CPU.
+	 */
+	entry.dest_mode = INT_DEST_MODE;
+	entry.mask = 0;					/* unmask IRQ now */
+	entry.dest.logical.logical_dest = cpu_mask_to_apicid(TARGET_CPUS);
+	entry.delivery_mode = INT_DELIVERY_MODE;
+	entry.polarity = 0;
+	entry.trigger = 0;
+	entry.vector = vector;
+
+	/*
+	 * The timer IRQ doesn't have to know that behind the
+	 * scene we have a 8259A-master in AEOI mode ...
+	 */
+	irq_desc[0].chip = &ioapic_edge_type;
+
+	/*
+	 * Add it to the IO-APIC irq-routing table:
+	 */
+	spin_lock_irqsave(&ioapic_lock, flags);
+	io_apic_write(apic, 0x11+2*pin, *(((int *)&entry)+1));
+	io_apic_write(apic, 0x10+2*pin, *(((int *)&entry)+0));
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	enable_8259A_irq(0);
+}
+
+void __init UNEXPECTED_IO_APIC(void)
+{
+}
+
+void __apicdebuginit print_IO_APIC(void)
+{
+	int apic, i;
+	union IO_APIC_reg_00 reg_00;
+	union IO_APIC_reg_01 reg_01;
+	union IO_APIC_reg_02 reg_02;
+	unsigned long flags;
+
+	if (apic_verbosity == APIC_QUIET)
+		return;
+
+	printk(KERN_DEBUG "number of MP IRQ sources: %d.\n", mp_irq_entries);
+	for (i = 0; i < nr_ioapics; i++)
+		printk(KERN_DEBUG "number of IO-APIC #%d registers: %d.\n",
+		       mp_ioapics[i].mpc_apicid, nr_ioapic_registers[i]);
+
+	/*
+	 * We are a bit conservative about what we expect.  We have to
+	 * know about every hardware change ASAP.
+	 */
+	printk(KERN_INFO "testing the IO APIC.......................\n");
+
+	for (apic = 0; apic < nr_ioapics; apic++) {
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	reg_00.raw = io_apic_read(apic, 0);
+	reg_01.raw = io_apic_read(apic, 1);
+	if (reg_01.bits.version >= 0x10)
+		reg_02.raw = io_apic_read(apic, 2);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	printk("\n");
+	printk(KERN_DEBUG "IO APIC #%d......\n", mp_ioapics[apic].mpc_apicid);
+	printk(KERN_DEBUG ".... register #00: %08X\n", reg_00.raw);
+	printk(KERN_DEBUG ".......    : physical APIC id: %02X\n", reg_00.bits.ID);
+	if (reg_00.bits.__reserved_1 || reg_00.bits.__reserved_2)
+		UNEXPECTED_IO_APIC();
+
+	printk(KERN_DEBUG ".... register #01: %08X\n", *(int *)&reg_01);
+	printk(KERN_DEBUG ".......     : max redirection entries: %04X\n", reg_01.bits.entries);
+	if (	(reg_01.bits.entries != 0x0f) && /* older (Neptune) boards */
+		(reg_01.bits.entries != 0x17) && /* typical ISA+PCI boards */
+		(reg_01.bits.entries != 0x1b) && /* Compaq Proliant boards */
+		(reg_01.bits.entries != 0x1f) && /* dual Xeon boards */
+		(reg_01.bits.entries != 0x22) && /* bigger Xeon boards */
+		(reg_01.bits.entries != 0x2E) &&
+		(reg_01.bits.entries != 0x3F) &&
+		(reg_01.bits.entries != 0x03) 
+	)
+		UNEXPECTED_IO_APIC();
+
+	printk(KERN_DEBUG ".......     : PRQ implemented: %X\n", reg_01.bits.PRQ);
+	printk(KERN_DEBUG ".......     : IO APIC version: %04X\n", reg_01.bits.version);
+	if (	(reg_01.bits.version != 0x01) && /* 82489DX IO-APICs */
+		(reg_01.bits.version != 0x02) && /* 82801BA IO-APICs (ICH2) */
+		(reg_01.bits.version != 0x10) && /* oldest IO-APICs */
+		(reg_01.bits.version != 0x11) && /* Pentium/Pro IO-APICs */
+		(reg_01.bits.version != 0x13) && /* Xeon IO-APICs */
+		(reg_01.bits.version != 0x20)    /* Intel P64H (82806 AA) */
+	)
+		UNEXPECTED_IO_APIC();
+	if (reg_01.bits.__reserved_1 || reg_01.bits.__reserved_2)
+		UNEXPECTED_IO_APIC();
+
+	if (reg_01.bits.version >= 0x10) {
+		printk(KERN_DEBUG ".... register #02: %08X\n", reg_02.raw);
+		printk(KERN_DEBUG ".......     : arbitration: %02X\n", reg_02.bits.arbitration);
+		if (reg_02.bits.__reserved_1 || reg_02.bits.__reserved_2)
+			UNEXPECTED_IO_APIC();
+	}
+
+	printk(KERN_DEBUG ".... IRQ redirection table:\n");
+
+	printk(KERN_DEBUG " NR Log Phy Mask Trig IRR Pol"
+			  " Stat Dest Deli Vect:   \n");
+
+	for (i = 0; i <= reg_01.bits.entries; i++) {
+		struct IO_APIC_route_entry entry;
+
+		spin_lock_irqsave(&ioapic_lock, flags);
+		*(((int *)&entry)+0) = io_apic_read(apic, 0x10+i*2);
+		*(((int *)&entry)+1) = io_apic_read(apic, 0x11+i*2);
+		spin_unlock_irqrestore(&ioapic_lock, flags);
+
+		printk(KERN_DEBUG " %02x %03X %02X  ",
+			i,
+			entry.dest.logical.logical_dest,
+			entry.dest.physical.physical_dest
+		);
+
+		printk("%1d    %1d    %1d   %1d   %1d    %1d    %1d    %02X\n",
+			entry.mask,
+			entry.trigger,
+			entry.irr,
+			entry.polarity,
+			entry.delivery_status,
+			entry.dest_mode,
+			entry.delivery_mode,
+			entry.vector
+		);
+	}
+	}
+	if (use_pci_vector())
+		printk(KERN_INFO "Using vector-based indexing\n");
+	printk(KERN_DEBUG "IRQ to pin mappings:\n");
+	for (i = 0; i < NR_IRQS; i++) {
+		struct irq_pin_list *entry = irq_2_pin + i;
+		if (entry->pin < 0)
+			continue;
+ 		if (use_pci_vector() && !platform_legacy_irq(i))
+			printk(KERN_DEBUG "IRQ%d ", IO_APIC_VECTOR(i));
+		else
+			printk(KERN_DEBUG "IRQ%d ", i);
+		for (;;) {
+			printk("-> %d:%d", entry->apic, entry->pin);
+			if (!entry->next)
+				break;
+			entry = irq_2_pin + entry->next;
+		}
+		printk("\n");
+	}
+
+	printk(KERN_INFO ".................................... done.\n");
+
+	return;
+}
+
+static __apicdebuginit void print_APIC_bitfield (int base)
+{
+	unsigned int v;
+	int i, j;
+
+	if (apic_verbosity == APIC_QUIET)
+		return;
+
+	printk(KERN_DEBUG "0123456789abcdef0123456789abcdef\n" KERN_DEBUG);
+	for (i = 0; i < 8; i++) {
+		v = apic_read(base + i*0x10);
+		for (j = 0; j < 32; j++) {
+			if (v & (1<<j))
+				printk("1");
+			else
+				printk("0");
+		}
+		printk("\n");
+	}
+}
+
+void __apicdebuginit print_local_APIC(void * dummy)
+{
+	unsigned int v, ver, maxlvt;
+
+	if (apic_verbosity == APIC_QUIET)
+		return;
+
+	printk("\n" KERN_DEBUG "printing local APIC contents on CPU#%d/%d:\n",
+		smp_processor_id(), hard_smp_processor_id());
+	v = apic_read(APIC_ID);
+	printk(KERN_INFO "... APIC ID:      %08x (%01x)\n", v, GET_APIC_ID(v));
+	v = apic_read(APIC_LVR);
+	printk(KERN_INFO "... APIC VERSION: %08x\n", v);
+	ver = GET_APIC_VERSION(v);
+	maxlvt = get_maxlvt();
+
+	v = apic_read(APIC_TASKPRI);
+	printk(KERN_DEBUG "... APIC TASKPRI: %08x (%02x)\n", v, v & APIC_TPRI_MASK);
+
+	v = apic_read(APIC_ARBPRI);
+	printk(KERN_DEBUG "... APIC ARBPRI: %08x (%02x)\n", v,
+		v & APIC_ARBPRI_MASK);
+	v = apic_read(APIC_PROCPRI);
+	printk(KERN_DEBUG "... APIC PROCPRI: %08x\n", v);
+
+	v = apic_read(APIC_EOI);
+	printk(KERN_DEBUG "... APIC EOI: %08x\n", v);
+	v = apic_read(APIC_RRR);
+	printk(KERN_DEBUG "... APIC RRR: %08x\n", v);
+	v = apic_read(APIC_LDR);
+	printk(KERN_DEBUG "... APIC LDR: %08x\n", v);
+	v = apic_read(APIC_DFR);
+	printk(KERN_DEBUG "... APIC DFR: %08x\n", v);
+	v = apic_read(APIC_SPIV);
+	printk(KERN_DEBUG "... APIC SPIV: %08x\n", v);
+
+	printk(KERN_DEBUG "... APIC ISR field:\n");
+	print_APIC_bitfield(APIC_ISR);
+	printk(KERN_DEBUG "... APIC TMR field:\n");
+	print_APIC_bitfield(APIC_TMR);
+	printk(KERN_DEBUG "... APIC IRR field:\n");
+	print_APIC_bitfield(APIC_IRR);
+
+	v = apic_read(APIC_ESR);
+	printk(KERN_DEBUG "... APIC ESR: %08x\n", v);
+
+	v = apic_read(APIC_ICR);
+	printk(KERN_DEBUG "... APIC ICR: %08x\n", v);
+	v = apic_read(APIC_ICR2);
+	printk(KERN_DEBUG "... APIC ICR2: %08x\n", v);
+
+	v = apic_read(APIC_LVTT);
+	printk(KERN_DEBUG "... APIC LVTT: %08x\n", v);
+
+	if (maxlvt > 3) {                       /* PC is LVT#4. */
+		v = apic_read(APIC_LVTPC);
+		printk(KERN_DEBUG "... APIC LVTPC: %08x\n", v);
+	}
+	v = apic_read(APIC_LVT0);
+	printk(KERN_DEBUG "... APIC LVT0: %08x\n", v);
+	v = apic_read(APIC_LVT1);
+	printk(KERN_DEBUG "... APIC LVT1: %08x\n", v);
+
+	if (maxlvt > 2) {			/* ERR is LVT#3. */
+		v = apic_read(APIC_LVTERR);
+		printk(KERN_DEBUG "... APIC LVTERR: %08x\n", v);
+	}
+
+	v = apic_read(APIC_TMICT);
+	printk(KERN_DEBUG "... APIC TMICT: %08x\n", v);
+	v = apic_read(APIC_TMCCT);
+	printk(KERN_DEBUG "... APIC TMCCT: %08x\n", v);
+	v = apic_read(APIC_TDCR);
+	printk(KERN_DEBUG "... APIC TDCR: %08x\n", v);
+	printk("\n");
+}
+
+void print_all_local_APICs (void)
+{
+	on_each_cpu(print_local_APIC, NULL, 1, 1);
+}
+
+void __apicdebuginit print_PIC(void)
+{
+	unsigned int v;
+	unsigned long flags;
+
+	if (apic_verbosity == APIC_QUIET)
+		return;
+
+	printk(KERN_DEBUG "\nprinting PIC contents\n");
+
+	spin_lock_irqsave(&i8259A_lock, flags);
+
+	v = inb(0xa1) << 8 | inb(0x21);
+	printk(KERN_DEBUG "... PIC  IMR: %04x\n", v);
+
+	v = inb(0xa0) << 8 | inb(0x20);
+	printk(KERN_DEBUG "... PIC  IRR: %04x\n", v);
+
+	outb(0x0b,0xa0);
+	outb(0x0b,0x20);
+	v = inb(0xa0) << 8 | inb(0x20);
+	outb(0x0a,0xa0);
+	outb(0x0a,0x20);
+
+	spin_unlock_irqrestore(&i8259A_lock, flags);
+
+	printk(KERN_DEBUG "... PIC  ISR: %04x\n", v);
+
+	v = inb(0x4d1) << 8 | inb(0x4d0);
+	printk(KERN_DEBUG "... PIC ELCR: %04x\n", v);
+}
+#endif /* !CONFIG_XEN */
+
+static void __init enable_IO_APIC(void)
+{
+	union IO_APIC_reg_01 reg_01;
+#ifndef CONFIG_XEN
+	int i8259_apic, i8259_pin;
+#endif
+	int i, apic;
+	unsigned long flags;
+
+	for (i = 0; i < PIN_MAP_SIZE; i++) {
+		irq_2_pin[i].pin = -1;
+		irq_2_pin[i].next = 0;
+	}
+	if (!pirqs_enabled)
+		for (i = 0; i < MAX_PIRQS; i++)
+			pirq_entries[i] = -1;
+
+	/*
+	 * The number of IO-APIC IRQ registers (== #pins):
+	 */
+	for (apic = 0; apic < nr_ioapics; apic++) {
+		spin_lock_irqsave(&ioapic_lock, flags);
+		reg_01.raw = io_apic_read(apic, 1);
+		spin_unlock_irqrestore(&ioapic_lock, flags);
+		nr_ioapic_registers[apic] = reg_01.bits.entries+1;
+	}
+#ifndef CONFIG_XEN
+	for(apic = 0; apic < nr_ioapics; apic++) {
+		int pin;
+		/* See if any of the pins is in ExtINT mode */
+		for (pin = 0; pin < nr_ioapic_registers[apic]; pin++) {
+			struct IO_APIC_route_entry entry;
+			spin_lock_irqsave(&ioapic_lock, flags);
+			*(((int *)&entry) + 0) = io_apic_read(apic, 0x10 + 2 * pin);
+			*(((int *)&entry) + 1) = io_apic_read(apic, 0x11 + 2 * pin);
+			spin_unlock_irqrestore(&ioapic_lock, flags);
+
+
+			/* If the interrupt line is enabled and in ExtInt mode
+			 * I have found the pin where the i8259 is connected.
+			 */
+			if ((entry.mask == 0) && (entry.delivery_mode == dest_ExtINT)) {
+				ioapic_i8259.apic = apic;
+				ioapic_i8259.pin  = pin;
+				goto found_i8259;
+			}
+		}
+	}
+ found_i8259:
+	/* Look to see what if the MP table has reported the ExtINT */
+	i8259_pin  = find_isa_irq_pin(0, mp_ExtINT);
+	i8259_apic = find_isa_irq_apic(0, mp_ExtINT);
+	/* Trust the MP table if nothing is setup in the hardware */
+	if ((ioapic_i8259.pin == -1) && (i8259_pin >= 0)) {
+		printk(KERN_WARNING "ExtINT not setup in hardware but reported by MP table\n");
+		ioapic_i8259.pin  = i8259_pin;
+		ioapic_i8259.apic = i8259_apic;
+	}
+	/* Complain if the MP table and the hardware disagree */
+	if (((ioapic_i8259.apic != i8259_apic) || (ioapic_i8259.pin != i8259_pin)) &&
+		(i8259_pin >= 0) && (ioapic_i8259.pin >= 0))
+	{
+		printk(KERN_WARNING "ExtINT in hardware and MP table differ\n");
+	}
+#endif
+
+	/*
+	 * Do not trust the IO-APIC being empty at bootup
+	 */
+	clear_IO_APIC();
+}
+
+/*
+ * Not an __init, needed by the reboot code
+ */
+void disable_IO_APIC(void)
+{
+	/*
+	 * Clear the IO-APIC before rebooting:
+	 */
+	clear_IO_APIC();
+
+#ifndef CONFIG_XEN
+	/*
+	 * If the i8259 is routed through an IOAPIC
+	 * Put that IOAPIC in virtual wire mode
+	 * so legacy interrupts can be delivered.
+	 */
+	if (ioapic_i8259.pin != -1) {
+		struct IO_APIC_route_entry entry;
+		unsigned long flags;
+
+		memset(&entry, 0, sizeof(entry));
+		entry.mask            = 0; /* Enabled */
+		entry.trigger         = 0; /* Edge */
+		entry.irr             = 0;
+		entry.polarity        = 0; /* High */
+		entry.delivery_status = 0;
+		entry.dest_mode       = 0; /* Physical */
+		entry.delivery_mode   = dest_ExtINT; /* ExtInt */
+		entry.vector          = 0;
+		entry.dest.physical.physical_dest =
+					GET_APIC_ID(apic_read(APIC_ID));
+
+		/*
+		 * Add it to the IO-APIC irq-routing table:
+		 */
+		spin_lock_irqsave(&ioapic_lock, flags);
+		io_apic_write(ioapic_i8259.apic, 0x11+2*ioapic_i8259.pin,
+			*(((int *)&entry)+1));
+		io_apic_write(ioapic_i8259.apic, 0x10+2*ioapic_i8259.pin,
+			*(((int *)&entry)+0));
+		spin_unlock_irqrestore(&ioapic_lock, flags);
+	}
+
+	disconnect_bsp_APIC(ioapic_i8259.pin != -1);
+#endif
+}
+
+/*
+ * function to set the IO-APIC physical IDs based on the
+ * values stored in the MPC table.
+ *
+ * by Matt Domsch <Matt_Domsch@dell.com>  Tue Dec 21 12:25:05 CST 1999
+ */
+
+#ifndef CONFIG_XEN
+static void __init setup_ioapic_ids_from_mpc (void)
+{
+	union IO_APIC_reg_00 reg_00;
+	int apic;
+	int i;
+	unsigned char old_id;
+	unsigned long flags;
+
+	/*
+	 * Set the IOAPIC ID to the value stored in the MPC table.
+	 */
+	for (apic = 0; apic < nr_ioapics; apic++) {
+
+		/* Read the register 0 value */
+		spin_lock_irqsave(&ioapic_lock, flags);
+		reg_00.raw = io_apic_read(apic, 0);
+		spin_unlock_irqrestore(&ioapic_lock, flags);
+		
+		old_id = mp_ioapics[apic].mpc_apicid;
+
+
+		printk(KERN_INFO "Using IO-APIC %d\n", mp_ioapics[apic].mpc_apicid);
+
+
+		/*
+		 * We need to adjust the IRQ routing table
+		 * if the ID changed.
+		 */
+		if (old_id != mp_ioapics[apic].mpc_apicid)
+			for (i = 0; i < mp_irq_entries; i++)
+				if (mp_irqs[i].mpc_dstapic == old_id)
+					mp_irqs[i].mpc_dstapic
+						= mp_ioapics[apic].mpc_apicid;
+
+		/*
+		 * Read the right value from the MPC table and
+		 * write it into the ID register.
+	 	 */
+		apic_printk(APIC_VERBOSE,KERN_INFO "...changing IO-APIC physical APIC ID to %d ...",
+				mp_ioapics[apic].mpc_apicid);
+
+		reg_00.bits.ID = mp_ioapics[apic].mpc_apicid;
+		spin_lock_irqsave(&ioapic_lock, flags);
+		io_apic_write(apic, 0, reg_00.raw);
+		spin_unlock_irqrestore(&ioapic_lock, flags);
+
+		/*
+		 * Sanity check
+		 */
+		spin_lock_irqsave(&ioapic_lock, flags);
+		reg_00.raw = io_apic_read(apic, 0);
+		spin_unlock_irqrestore(&ioapic_lock, flags);
+		if (reg_00.bits.ID != mp_ioapics[apic].mpc_apicid)
+			printk("could not set ID!\n");
+		else
+			apic_printk(APIC_VERBOSE," ok.\n");
+	}
+}
+#else
+static void __init setup_ioapic_ids_from_mpc(void) { }
+#endif
+
+/*
+ * There is a nasty bug in some older SMP boards, their mptable lies
+ * about the timer IRQ. We do the following to work around the situation:
+ *
+ *	- timer IRQ defaults to IO-APIC IRQ
+ *	- if this function detects that timer IRQs are defunct, then we fall
+ *	  back to ISA timer IRQs
+ */
+#ifndef CONFIG_XEN
+static int __init timer_irq_works(void)
+{
+	unsigned long t1 = jiffies;
+
+	local_irq_enable();
+	/* Let ten ticks pass... */
+	mdelay((10 * 1000) / HZ);
+
+	/*
+	 * Expect a few ticks at least, to be sure some possible
+	 * glue logic does not lock up after one or two first
+	 * ticks in a non-ExtINT mode.  Also the local APIC
+	 * might have cached one ExtINT interrupt.  Finally, at
+	 * least one tick may be lost due to delays.
+	 */
+
+	/* jiffies wrap? */
+	if (jiffies - t1 > 4)
+		return 1;
+	return 0;
+}
+
+/*
+ * In the SMP+IOAPIC case it might happen that there are an unspecified
+ * number of pending IRQ events unhandled. These cases are very rare,
+ * so we 'resend' these IRQs via IPIs, to the same CPU. It's much
+ * better to do it this way as thus we do not have to be aware of
+ * 'pending' interrupts in the IRQ path, except at this point.
+ */
+/*
+ * Edge triggered needs to resend any interrupt
+ * that was delayed but this is now handled in the device
+ * independent code.
+ */
+
+/*
+ * Starting up a edge-triggered IO-APIC interrupt is
+ * nasty - we need to make sure that we get the edge.
+ * If it is already asserted for some reason, we need
+ * return 1 to indicate that is was pending.
+ *
+ * This is not complete - we should be able to fake
+ * an edge even if it isn't on the 8259A...
+ */
+
+static unsigned int startup_edge_ioapic_irq(unsigned int irq)
+{
+	int was_pending = 0;
+	unsigned long flags;
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	if (irq < 16) {
+		disable_8259A_irq(irq);
+		if (i8259A_irq_pending(irq))
+			was_pending = 1;
+	}
+	__unmask_IO_APIC_irq(irq);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	return was_pending;
+}
+
+/*
+ * Once we have recorded IRQ_PENDING already, we can mask the
+ * interrupt for real. This prevents IRQ storms from unhandled
+ * devices.
+ */
+static void ack_edge_ioapic_irq(unsigned int irq)
+{
+	move_irq(irq);
+	if ((irq_desc[irq].status & (IRQ_PENDING | IRQ_DISABLED))
+					== (IRQ_PENDING | IRQ_DISABLED))
+		mask_IO_APIC_irq(irq);
+	ack_APIC_irq();
+}
+
+/*
+ * Level triggered interrupts can just be masked,
+ * and shutting down and starting up the interrupt
+ * is the same as enabling and disabling them -- except
+ * with a startup need to return a "was pending" value.
+ *
+ * Level triggered interrupts are special because we
+ * do not touch any IO-APIC register while handling
+ * them. We ack the APIC in the end-IRQ handler, not
+ * in the start-IRQ-handler. Protection against reentrance
+ * from the same interrupt is still provided, both by the
+ * generic IRQ layer and by the fact that an unacked local
+ * APIC does not accept IRQs.
+ */
+static unsigned int startup_level_ioapic_irq (unsigned int irq)
+{
+	unmask_IO_APIC_irq(irq);
+
+	return 0; /* don't check for pending */
+}
+
+static void end_level_ioapic_irq (unsigned int irq)
+{
+	move_irq(irq);
+	ack_APIC_irq();
+}
+
+#ifdef CONFIG_PCI_MSI
+static unsigned int startup_edge_ioapic_vector(unsigned int vector)
+{
+	int irq = vector_to_irq(vector);
+
+	return startup_edge_ioapic_irq(irq);
+}
+
+static void ack_edge_ioapic_vector(unsigned int vector)
+{
+	int irq = vector_to_irq(vector);
+
+	move_native_irq(vector);
+	ack_edge_ioapic_irq(irq);
+}
+
+static unsigned int startup_level_ioapic_vector (unsigned int vector)
+{
+	int irq = vector_to_irq(vector);
+
+	return startup_level_ioapic_irq (irq);
+}
+
+static void end_level_ioapic_vector (unsigned int vector)
+{
+	int irq = vector_to_irq(vector);
+
+	move_native_irq(vector);
+	end_level_ioapic_irq(irq);
+}
+
+static void mask_IO_APIC_vector (unsigned int vector)
+{
+	int irq = vector_to_irq(vector);
+
+	mask_IO_APIC_irq(irq);
+}
+
+static void unmask_IO_APIC_vector (unsigned int vector)
+{
+	int irq = vector_to_irq(vector);
+
+	unmask_IO_APIC_irq(irq);
+}
+
+#ifdef CONFIG_SMP
+static void set_ioapic_affinity_vector (unsigned int vector,
+					cpumask_t cpu_mask)
+{
+	int irq = vector_to_irq(vector);
+
+	set_native_irq_info(vector, cpu_mask);
+	set_ioapic_affinity_irq(irq, cpu_mask);
+}
+#endif // CONFIG_SMP
+#endif // CONFIG_PCI_MSI
+
+static int ioapic_retrigger(unsigned int irq)
+{
+	send_IPI_self(IO_APIC_VECTOR(irq));
+
+	return 1;
+}
+
+/*
+ * Level and edge triggered IO-APIC interrupts need different handling,
+ * so we use two separate IRQ descriptors. Edge triggered IRQs can be
+ * handled with the level-triggered descriptor, but that one has slightly
+ * more overhead. Level-triggered interrupts cannot be handled with the
+ * edge-triggered handler, without risking IRQ storms and other ugly
+ * races.
+ */
+
+static struct hw_interrupt_type ioapic_edge_type __read_mostly = {
+	.typename = "IO-APIC-edge",
+	.startup 	= startup_edge_ioapic,
+	.shutdown 	= shutdown_edge_ioapic,
+	.enable 	= enable_edge_ioapic,
+	.disable 	= disable_edge_ioapic,
+	.ack 		= ack_edge_ioapic,
+	.end 		= end_edge_ioapic,
+#ifdef CONFIG_SMP
+	.set_affinity = set_ioapic_affinity,
+#endif
+	.retrigger	= ioapic_retrigger,
+};
+
+static struct hw_interrupt_type ioapic_level_type __read_mostly = {
+	.typename = "IO-APIC-level",
+	.startup 	= startup_level_ioapic,
+	.shutdown 	= shutdown_level_ioapic,
+	.enable 	= enable_level_ioapic,
+	.disable 	= disable_level_ioapic,
+	.ack 		= mask_and_ack_level_ioapic,
+	.end 		= end_level_ioapic,
+#ifdef CONFIG_SMP
+	.set_affinity = set_ioapic_affinity,
+#endif
+	.retrigger	= ioapic_retrigger,
+};
+#endif /* !CONFIG_XEN */
+
+static inline void init_IO_APIC_traps(void)
+{
+	int irq;
+
+	/*
+	 * NOTE! The local APIC isn't very good at handling
+	 * multiple interrupts at the same interrupt level.
+	 * As the interrupt level is determined by taking the
+	 * vector number and shifting that right by 4, we
+	 * want to spread these out a bit so that they don't
+	 * all fall in the same interrupt level.
+	 *
+	 * Also, we've got to be careful not to trash gate
+	 * 0x80, because int 0x80 is hm, kind of importantish. ;)
+	 */
+	for (irq = 0; irq < NR_IRQS ; irq++) {
+		int tmp = irq;
+		if (use_pci_vector()) {
+			if (!platform_legacy_irq(tmp))
+				if ((tmp = vector_to_irq(tmp)) == -1)
+					continue;
+		}
+		if (IO_APIC_IRQ(tmp) && !IO_APIC_VECTOR(tmp)) {
+			/*
+			 * Hmm.. We don't have an entry for this,
+			 * so default to an old-fashioned 8259
+			 * interrupt if we can..
+			 */
+			if (irq < 16)
+				make_8259A_irq(irq);
+#ifndef CONFIG_XEN
+			else
+				/* Strange. Oh, well.. */
+				irq_desc[irq].chip = &no_irq_type;
+#endif
+		}
+	}
+}
+
+#ifndef CONFIG_XEN
+static void enable_lapic_irq (unsigned int irq)
+{
+	unsigned long v;
+
+	v = apic_read(APIC_LVT0);
+	apic_write(APIC_LVT0, v & ~APIC_LVT_MASKED);
+}
+
+static void disable_lapic_irq (unsigned int irq)
+{
+	unsigned long v;
+
+	v = apic_read(APIC_LVT0);
+	apic_write(APIC_LVT0, v | APIC_LVT_MASKED);
+}
+
+static void ack_lapic_irq (unsigned int irq)
+{
+	ack_APIC_irq();
+}
+
+static void end_lapic_irq (unsigned int i) { /* nothing */ }
+
+static struct hw_interrupt_type lapic_irq_type __read_mostly = {
+	.typename = "local-APIC-edge",
+	.startup = NULL, /* startup_irq() not used for IRQ0 */
+	.shutdown = NULL, /* shutdown_irq() not used for IRQ0 */
+	.enable = enable_lapic_irq,
+	.disable = disable_lapic_irq,
+	.ack = ack_lapic_irq,
+	.end = end_lapic_irq,
+};
+
+static void setup_nmi (void)
+{
+	/*
+ 	 * Dirty trick to enable the NMI watchdog ...
+	 * We put the 8259A master into AEOI mode and
+	 * unmask on all local APICs LVT0 as NMI.
+	 *
+	 * The idea to use the 8259A in AEOI mode ('8259A Virtual Wire')
+	 * is from Maciej W. Rozycki - so we do not have to EOI from
+	 * the NMI handler or the timer interrupt.
+	 */ 
+	printk(KERN_INFO "activating NMI Watchdog ...");
+
+	enable_NMI_through_LVT0(NULL);
+
+	printk(" done.\n");
+}
+
+/*
+ * This looks a bit hackish but it's about the only one way of sending
+ * a few INTA cycles to 8259As and any associated glue logic.  ICR does
+ * not support the ExtINT mode, unfortunately.  We need to send these
+ * cycles as some i82489DX-based boards have glue logic that keeps the
+ * 8259A interrupt line asserted until INTA.  --macro
+ */
+static inline void unlock_ExtINT_logic(void)
+{
+	int apic, pin, i;
+	struct IO_APIC_route_entry entry0, entry1;
+	unsigned char save_control, save_freq_select;
+	unsigned long flags;
+
+	pin  = find_isa_irq_pin(8, mp_INT);
+	apic = find_isa_irq_apic(8, mp_INT);
+	if (pin == -1)
+		return;
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	*(((int *)&entry0) + 1) = io_apic_read(apic, 0x11 + 2 * pin);
+	*(((int *)&entry0) + 0) = io_apic_read(apic, 0x10 + 2 * pin);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+	clear_IO_APIC_pin(apic, pin);
+
+	memset(&entry1, 0, sizeof(entry1));
+
+	entry1.dest_mode = 0;			/* physical delivery */
+	entry1.mask = 0;			/* unmask IRQ now */
+	entry1.dest.physical.physical_dest = hard_smp_processor_id();
+	entry1.delivery_mode = dest_ExtINT;
+	entry1.polarity = entry0.polarity;
+	entry1.trigger = 0;
+	entry1.vector = 0;
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	io_apic_write(apic, 0x11 + 2 * pin, *(((int *)&entry1) + 1));
+	io_apic_write(apic, 0x10 + 2 * pin, *(((int *)&entry1) + 0));
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	save_control = CMOS_READ(RTC_CONTROL);
+	save_freq_select = CMOS_READ(RTC_FREQ_SELECT);
+	CMOS_WRITE((save_freq_select & ~RTC_RATE_SELECT) | 0x6,
+		   RTC_FREQ_SELECT);
+	CMOS_WRITE(save_control | RTC_PIE, RTC_CONTROL);
+
+	i = 100;
+	while (i-- > 0) {
+		mdelay(10);
+		if ((CMOS_READ(RTC_INTR_FLAGS) & RTC_PF) == RTC_PF)
+			i -= 10;
+	}
+
+	CMOS_WRITE(save_control, RTC_CONTROL);
+	CMOS_WRITE(save_freq_select, RTC_FREQ_SELECT);
+	clear_IO_APIC_pin(apic, pin);
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	io_apic_write(apic, 0x11 + 2 * pin, *(((int *)&entry0) + 1));
+	io_apic_write(apic, 0x10 + 2 * pin, *(((int *)&entry0) + 0));
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+}
+
+int timer_uses_ioapic_pin_0;
+
+/*
+ * This code may look a bit paranoid, but it's supposed to cooperate with
+ * a wide range of boards and BIOS bugs.  Fortunately only the timer IRQ
+ * is so screwy.  Thanks to Brian Perkins for testing/hacking this beast
+ * fanatically on his truly buggy board.
+ *
+ * FIXME: really need to revamp this for modern platforms only.
+ */
+static inline void check_timer(void)
+{
+	int apic1, pin1, apic2, pin2;
+	int vector;
+
+	/*
+	 * get/set the timer IRQ vector:
+	 */
+	disable_8259A_irq(0);
+	vector = assign_irq_vector(0);
+	set_intr_gate(vector, interrupt[0]);
+
+	/*
+	 * Subtle, code in do_timer_interrupt() expects an AEOI
+	 * mode for the 8259A whenever interrupts are routed
+	 * through I/O APICs.  Also IRQ0 has to be enabled in
+	 * the 8259A which implies the virtual wire has to be
+	 * disabled in the local APIC.
+	 */
+	apic_write(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_EXTINT);
+	init_8259A(1);
+	if (timer_over_8254 > 0)
+		enable_8259A_irq(0);
+
+	pin1  = find_isa_irq_pin(0, mp_INT);
+	apic1 = find_isa_irq_apic(0, mp_INT);
+	pin2  = ioapic_i8259.pin;
+	apic2 = ioapic_i8259.apic;
+
+	if (pin1 == 0)
+		timer_uses_ioapic_pin_0 = 1;
+
+	apic_printk(APIC_VERBOSE,KERN_INFO "..TIMER: vector=0x%02X apic1=%d pin1=%d apic2=%d pin2=%d\n",
+		vector, apic1, pin1, apic2, pin2);
+
+	if (pin1 != -1) {
+		/*
+		 * Ok, does IRQ0 through the IOAPIC work?
+		 */
+		unmask_IO_APIC_irq(0);
+		if (!no_timer_check && timer_irq_works()) {
+			nmi_watchdog_default();
+			if (nmi_watchdog == NMI_IO_APIC) {
+				disable_8259A_irq(0);
+				setup_nmi();
+				enable_8259A_irq(0);
+			}
+			if (disable_timer_pin_1 > 0)
+				clear_IO_APIC_pin(0, pin1);
+			return;
+		}
+		clear_IO_APIC_pin(apic1, pin1);
+		apic_printk(APIC_QUIET,KERN_ERR "..MP-BIOS bug: 8254 timer not "
+				"connected to IO-APIC\n");
+	}
+
+	apic_printk(APIC_VERBOSE,KERN_INFO "...trying to set up timer (IRQ0) "
+				"through the 8259A ... ");
+	if (pin2 != -1) {
+		apic_printk(APIC_VERBOSE,"\n..... (found apic %d pin %d) ...",
+			apic2, pin2);
+		/*
+		 * legacy devices should be connected to IO APIC #0
+		 */
+		setup_ExtINT_IRQ0_pin(apic2, pin2, vector);
+		if (timer_irq_works()) {
+			apic_printk(APIC_VERBOSE," works.\n");
+			nmi_watchdog_default();
+			if (nmi_watchdog == NMI_IO_APIC) {
+				setup_nmi();
+			}
+			return;
+		}
+		/*
+		 * Cleanup, just in case ...
+		 */
+		clear_IO_APIC_pin(apic2, pin2);
+	}
+	apic_printk(APIC_VERBOSE," failed.\n");
+
+	if (nmi_watchdog == NMI_IO_APIC) {
+		printk(KERN_WARNING "timer doesn't work through the IO-APIC - disabling NMI Watchdog!\n");
+		nmi_watchdog = 0;
+	}
+
+	apic_printk(APIC_VERBOSE, KERN_INFO "...trying to set up timer as Virtual Wire IRQ...");
+
+	disable_8259A_irq(0);
+	irq_desc[0].chip = &lapic_irq_type;
+	apic_write(APIC_LVT0, APIC_DM_FIXED | vector);	/* Fixed mode */
+	enable_8259A_irq(0);
+
+	if (timer_irq_works()) {
+		apic_printk(APIC_VERBOSE," works.\n");
+		return;
+	}
+	apic_write(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_FIXED | vector);
+	apic_printk(APIC_VERBOSE," failed.\n");
+
+	apic_printk(APIC_VERBOSE, KERN_INFO "...trying to set up timer as ExtINT IRQ...");
+
+	init_8259A(0);
+	make_8259A_irq(0);
+	apic_write(APIC_LVT0, APIC_DM_EXTINT);
+
+	unlock_ExtINT_logic();
+
+	if (timer_irq_works()) {
+		apic_printk(APIC_VERBOSE," works.\n");
+		return;
+	}
+	apic_printk(APIC_VERBOSE," failed :(.\n");
+	panic("IO-APIC + timer doesn't work! Try using the 'noapic' kernel parameter\n");
+}
+#else
+#define check_timer() ((void)0)
+int timer_uses_ioapic_pin_0 = 0;
+#endif /* !CONFIG_XEN */
+
+static int __init notimercheck(char *s)
+{
+	no_timer_check = 1;
+	return 1;
+}
+__setup("no_timer_check", notimercheck);
+
+/*
+ *
+ * IRQ's that are handled by the PIC in the MPS IOAPIC case.
+ * - IRQ2 is the cascade IRQ, and cannot be a io-apic IRQ.
+ *   Linux doesn't really care, as it's not actually used
+ *   for any interrupt handling anyway.
+ */
+#define PIC_IRQS	(1<<2)
+
+void __init setup_IO_APIC(void)
+{
+	enable_IO_APIC();
+
+	if (acpi_ioapic)
+		io_apic_irqs = ~0;	/* all IRQs go through IOAPIC */
+	else
+		io_apic_irqs = ~PIC_IRQS;
+
+	apic_printk(APIC_VERBOSE, "ENABLING IO-APIC IRQs\n");
+
+	/*
+	 * Set up the IO-APIC IRQ routing table.
+	 */
+	if (!acpi_ioapic)
+		setup_ioapic_ids_from_mpc();
+#ifndef CONFIG_XEN
+	sync_Arb_IDs();
+#endif /* !CONFIG_XEN */
+	setup_IO_APIC_irqs();
+	init_IO_APIC_traps();
+	check_timer();
+	if (!acpi_ioapic)
+		print_IO_APIC();
+}
+
+struct sysfs_ioapic_data {
+	struct sys_device dev;
+	struct IO_APIC_route_entry entry[0];
+};
+static struct sysfs_ioapic_data * mp_ioapic_data[MAX_IO_APICS];
+
+static int ioapic_suspend(struct sys_device *dev, pm_message_t state)
+{
+	struct IO_APIC_route_entry *entry;
+	struct sysfs_ioapic_data *data;
+	unsigned long flags;
+	int i;
+
+	data = container_of(dev, struct sysfs_ioapic_data, dev);
+	entry = data->entry;
+	spin_lock_irqsave(&ioapic_lock, flags);
+	for (i = 0; i < nr_ioapic_registers[dev->id]; i ++, entry ++ ) {
+		*(((int *)entry) + 1) = io_apic_read(dev->id, 0x11 + 2 * i);
+		*(((int *)entry) + 0) = io_apic_read(dev->id, 0x10 + 2 * i);
+	}
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	return 0;
+}
+
+static int ioapic_resume(struct sys_device *dev)
+{
+	struct IO_APIC_route_entry *entry;
+	struct sysfs_ioapic_data *data;
+	unsigned long flags;
+	union IO_APIC_reg_00 reg_00;
+	int i;
+
+	data = container_of(dev, struct sysfs_ioapic_data, dev);
+	entry = data->entry;
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	reg_00.raw = io_apic_read(dev->id, 0);
+	if (reg_00.bits.ID != mp_ioapics[dev->id].mpc_apicid) {
+		reg_00.bits.ID = mp_ioapics[dev->id].mpc_apicid;
+		io_apic_write(dev->id, 0, reg_00.raw);
+	}
+	for (i = 0; i < nr_ioapic_registers[dev->id]; i ++, entry ++ ) {
+		io_apic_write(dev->id, 0x11+2*i, *(((int *)entry)+1));
+		io_apic_write(dev->id, 0x10+2*i, *(((int *)entry)+0));
+	}
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	return 0;
+}
+
+static struct sysdev_class ioapic_sysdev_class = {
+	set_kset_name("ioapic"),
+	.suspend = ioapic_suspend,
+	.resume = ioapic_resume,
+};
+
+static int __init ioapic_init_sysfs(void)
+{
+	struct sys_device * dev;
+	int i, size, error = 0;
+
+	error = sysdev_class_register(&ioapic_sysdev_class);
+	if (error)
+		return error;
+
+	for (i = 0; i < nr_ioapics; i++ ) {
+		size = sizeof(struct sys_device) + nr_ioapic_registers[i]
+			* sizeof(struct IO_APIC_route_entry);
+		mp_ioapic_data[i] = kmalloc(size, GFP_KERNEL);
+		if (!mp_ioapic_data[i]) {
+			printk(KERN_ERR "Can't suspend/resume IOAPIC %d\n", i);
+			continue;
+		}
+		memset(mp_ioapic_data[i], 0, size);
+		dev = &mp_ioapic_data[i]->dev;
+		dev->id = i;
+		dev->cls = &ioapic_sysdev_class;
+		error = sysdev_register(dev);
+		if (error) {
+			kfree(mp_ioapic_data[i]);
+			mp_ioapic_data[i] = NULL;
+			printk(KERN_ERR "Can't suspend/resume IOAPIC %d\n", i);
+			continue;
+		}
+	}
+
+	return 0;
+}
+
+device_initcall(ioapic_init_sysfs);
+
+/* --------------------------------------------------------------------------
+                          ACPI-based IOAPIC Configuration
+   -------------------------------------------------------------------------- */
+
+#ifdef CONFIG_ACPI
+
+#define IO_APIC_MAX_ID		0xFE
+
+int __init io_apic_get_version (int ioapic)
+{
+	union IO_APIC_reg_01	reg_01;
+	unsigned long flags;
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	reg_01.raw = io_apic_read(ioapic, 1);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	return reg_01.bits.version;
+}
+
+
+int __init io_apic_get_redir_entries (int ioapic)
+{
+	union IO_APIC_reg_01	reg_01;
+	unsigned long flags;
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	reg_01.raw = io_apic_read(ioapic, 1);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	return reg_01.bits.entries;
+}
+
+
+int io_apic_set_pci_routing (int ioapic, int pin, int irq, int edge_level, int active_high_low)
+{
+	struct IO_APIC_route_entry entry;
+	unsigned long flags;
+
+	if (!IO_APIC_IRQ(irq)) {
+		apic_printk(APIC_QUIET,KERN_ERR "IOAPIC[%d]: Invalid reference to IRQ 0\n",
+			ioapic);
+		return -EINVAL;
+	}
+
+	/*
+	 * Generate a PCI IRQ routing entry and program the IOAPIC accordingly.
+	 * Note that we mask (disable) IRQs now -- these get enabled when the
+	 * corresponding device driver registers for this IRQ.
+	 */
+
+	memset(&entry,0,sizeof(entry));
+
+	entry.delivery_mode = INT_DELIVERY_MODE;
+	entry.dest_mode = INT_DEST_MODE;
+	entry.dest.logical.logical_dest = cpu_mask_to_apicid(TARGET_CPUS);
+	entry.trigger = edge_level;
+	entry.polarity = active_high_low;
+	entry.mask = 1;					 /* Disabled (masked) */
+
+	irq = gsi_irq_sharing(irq);
+	/*
+	 * IRQs < 16 are already in the irq_2_pin[] map
+	 */
+	if (irq >= 16)
+		add_pin_to_irq(irq, ioapic, pin);
+
+	entry.vector = assign_irq_vector(irq);
+
+	apic_printk(APIC_VERBOSE,KERN_DEBUG "IOAPIC[%d]: Set PCI routing entry (%d-%d -> 0x%x -> "
+		"IRQ %d Mode:%i Active:%i)\n", ioapic, 
+	       mp_ioapics[ioapic].mpc_apicid, pin, entry.vector, irq,
+	       edge_level, active_high_low);
+
+	ioapic_register_intr(irq, entry.vector, edge_level);
+
+	if (!ioapic && (irq < 16))
+		disable_8259A_irq(irq);
+
+	spin_lock_irqsave(&ioapic_lock, flags);
+	io_apic_write(ioapic, 0x11+2*pin, *(((int *)&entry)+1));
+	io_apic_write(ioapic, 0x10+2*pin, *(((int *)&entry)+0));
+	set_native_irq_info(use_pci_vector() ?  entry.vector : irq, TARGET_CPUS);
+	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	return 0;
+}
+
+#endif /* CONFIG_ACPI */
+
+
+#ifndef CONFIG_XEN
+/*
+ * This function currently is only a helper for the i386 smp boot process where
+ * we need to reprogram the ioredtbls to cater for the cpus which have come online
+ * so mask in all cases should simply be TARGET_CPUS
+ */
+#ifdef CONFIG_SMP
+void __init setup_ioapic_dest(void)
+{
+	int pin, ioapic, irq, irq_entry;
+
+	if (skip_ioapic_setup == 1)
+		return;
+
+	for (ioapic = 0; ioapic < nr_ioapics; ioapic++) {
+		for (pin = 0; pin < nr_ioapic_registers[ioapic]; pin++) {
+			irq_entry = find_irq_entry(ioapic, pin, mp_INT);
+			if (irq_entry == -1)
+				continue;
+			irq = pin_2_irq(irq_entry, ioapic, pin);
+			set_ioapic_affinity_irq(irq, TARGET_CPUS);
+		}
+
+	}
+}
+#endif
+#endif /* !CONFIG_XEN */
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/ioport-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/ioport-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/ioport-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/ioport-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,100 @@
+/*
+ *	linux/arch/x86_64/kernel/ioport.c
+ *
+ * This contains the io-permission bitmap code - written by obz, with changes
+ * by Linus.
+ */
+
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/capability.h>
+#include <linux/errno.h>
+#include <linux/types.h>
+#include <linux/ioport.h>
+#include <linux/mm.h>
+#include <linux/smp.h>
+#include <linux/smp_lock.h>
+#include <linux/stddef.h>
+#include <linux/slab.h>
+#include <linux/thread_info.h>
+#include <xen/interface/physdev.h>
+
+/* Set EXTENT bits starting at BASE in BITMAP to value TURN_ON. */
+static void set_bitmap(unsigned long *bitmap, unsigned int base, unsigned int extent, int new_value)
+{
+	int i;
+
+	if (new_value)
+		for (i = base; i < base + extent; i++)
+			__set_bit(i, bitmap);
+	else
+		for (i = base; i < base + extent; i++)
+			clear_bit(i, bitmap);
+}
+
+/*
+ * this changes the io permissions bitmap in the current task.
+ */
+asmlinkage long sys_ioperm(unsigned long from, unsigned long num, int turn_on)
+{
+	struct thread_struct * t = &current->thread;
+	unsigned long *bitmap;
+	struct physdev_set_iobitmap set_iobitmap;
+
+	if ((from + num <= from) || (from + num > IO_BITMAP_BITS))
+		return -EINVAL;
+	if (turn_on && !capable(CAP_SYS_RAWIO))
+		return -EPERM;
+
+	/*
+	 * If it's the first ioperm() call in this thread's lifetime, set the
+	 * IO bitmap up. ioperm() is much less timing critical than clone(),
+	 * this is why we delay this operation until now:
+	 */
+	if (!t->io_bitmap_ptr) {
+		bitmap = kmalloc(IO_BITMAP_BYTES, GFP_KERNEL);
+		if (!bitmap)
+			return -ENOMEM;
+
+		memset(bitmap, 0xff, IO_BITMAP_BYTES);
+		t->io_bitmap_ptr = bitmap;
+
+		set_xen_guest_handle(set_iobitmap.bitmap, (char *)bitmap);
+		set_iobitmap.nr_ports = IO_BITMAP_BITS;
+		WARN_ON(HYPERVISOR_physdev_op(PHYSDEVOP_set_iobitmap,
+					      &set_iobitmap));
+	}
+
+	set_bitmap(t->io_bitmap_ptr, from, num, !turn_on);
+
+	return 0;
+}
+
+/*
+ * sys_iopl has to be used when you want to access the IO ports
+ * beyond the 0x3ff range: to get the full 65536 ports bitmapped
+ * you'd need 8kB of bitmaps/process, which is a bit excessive.
+ *
+ */
+
+asmlinkage long sys_iopl(unsigned int new_iopl, struct pt_regs *regs)
+{
+	unsigned int old_iopl = current->thread.iopl;
+	struct physdev_set_iopl set_iopl;
+
+	if (new_iopl > 3)
+		return -EINVAL;
+
+	/* Need "raw I/O" privileges for direct port access. */
+	if ((new_iopl > old_iopl) && !capable(CAP_SYS_RAWIO))
+		return -EPERM;
+
+	/* Change our version of the privilege levels. */
+	current->thread.iopl = new_iopl;
+
+	/* Force the change at ring 0. */
+	set_iopl.iopl = (new_iopl == 0) ? 1 : new_iopl;
+	WARN_ON(HYPERVISOR_physdev_op(PHYSDEVOP_set_iopl, &set_iopl));
+
+	return 0;
+}
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/irq-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/irq-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/irq-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/irq-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,197 @@
+/*
+ *	linux/arch/x86_64/kernel/irq.c
+ *
+ *	Copyright (C) 1992, 1998 Linus Torvalds, Ingo Molnar
+ *
+ * This file contains the lowest level x86_64-specific interrupt
+ * entry and irq statistics code. All the remaining irq logic is
+ * done by the generic kernel/irq/ code and in the
+ * x86_64-specific irq controller code. (e.g. i8259.c and
+ * io_apic.c.)
+ */
+
+#include <linux/kernel_stat.h>
+#include <linux/interrupt.h>
+#include <linux/seq_file.h>
+#include <linux/module.h>
+#include <linux/delay.h>
+#include <asm/uaccess.h>
+#include <asm/io_apic.h>
+#include <asm/idle.h>
+
+atomic_t irq_err_count;
+#ifdef CONFIG_X86_IO_APIC
+#ifdef APIC_MISMATCH_DEBUG
+atomic_t irq_mis_count;
+#endif
+#endif
+
+#ifdef CONFIG_DEBUG_STACKOVERFLOW
+/*
+ * Probabilistic stack overflow check:
+ *
+ * Only check the stack in process context, because everything else
+ * runs on the big interrupt stacks. Checking reliably is too expensive,
+ * so we just check from interrupts.
+ */
+static inline void stack_overflow_check(struct pt_regs *regs)
+{
+	u64 curbase = (u64) current->thread_info;
+	static unsigned long warned = -60*HZ;
+
+	if (regs->rsp >= curbase && regs->rsp <= curbase + THREAD_SIZE &&
+	    regs->rsp <  curbase + sizeof(struct thread_info) + 128 &&
+	    time_after(jiffies, warned + 60*HZ)) {
+		printk("do_IRQ: %s near stack overflow (cur:%Lx,rsp:%lx)\n",
+		       current->comm, curbase, regs->rsp);
+		show_stack(NULL,NULL);
+		warned = jiffies;
+	}
+}
+#endif
+
+/*
+ * Generic, controller-independent functions:
+ */
+
+int show_interrupts(struct seq_file *p, void *v)
+{
+	int i = *(loff_t *) v, j;
+	struct irqaction * action;
+	unsigned long flags;
+
+	if (i == 0) {
+		seq_printf(p, "           ");
+		for_each_online_cpu(j)
+			seq_printf(p, "CPU%-8d",j);
+		seq_putc(p, '\n');
+	}
+
+	if (i < NR_IRQS) {
+		spin_lock_irqsave(&irq_desc[i].lock, flags);
+		action = irq_desc[i].action;
+		if (!action) 
+			goto skip;
+		seq_printf(p, "%3d: ",i);
+#ifndef CONFIG_SMP
+		seq_printf(p, "%10u ", kstat_irqs(i));
+#else
+		for_each_online_cpu(j)
+			seq_printf(p, "%10u ", kstat_cpu(j).irqs[i]);
+#endif
+		seq_printf(p, " %14s", irq_desc[i].chip->typename);
+
+		seq_printf(p, "  %s", action->name);
+		for (action=action->next; action; action = action->next)
+			seq_printf(p, ", %s", action->name);
+		seq_putc(p, '\n');
+skip:
+		spin_unlock_irqrestore(&irq_desc[i].lock, flags);
+	} else if (i == NR_IRQS) {
+		seq_printf(p, "NMI: ");
+		for_each_online_cpu(j)
+			seq_printf(p, "%10u ", cpu_pda(j)->__nmi_count);
+		seq_putc(p, '\n');
+#ifdef CONFIG_X86_LOCAL_APIC
+		seq_printf(p, "LOC: ");
+		for_each_online_cpu(j)
+			seq_printf(p, "%10u ", cpu_pda(j)->apic_timer_irqs);
+		seq_putc(p, '\n');
+#endif
+		seq_printf(p, "ERR: %10u\n", atomic_read(&irq_err_count));
+#ifdef CONFIG_X86_IO_APIC
+#ifdef APIC_MISMATCH_DEBUG
+		seq_printf(p, "MIS: %10u\n", atomic_read(&irq_mis_count));
+#endif
+#endif
+	}
+	return 0;
+}
+
+/*
+ * do_IRQ handles all normal device IRQ's (the special
+ * SMP cross-CPU interrupts have their own specific
+ * handlers).
+ */
+asmlinkage unsigned int do_IRQ(struct pt_regs *regs)
+{	
+	/* high bit used in ret_from_ code  */
+	unsigned irq = ~regs->orig_rax;
+
+	if (unlikely(irq >= NR_IRQS)) {
+		printk(KERN_EMERG "%s: cannot handle IRQ %d\n",
+					__FUNCTION__, irq);
+		BUG();
+	}
+
+	exit_idle();
+	irq_enter();
+#ifdef CONFIG_DEBUG_STACKOVERFLOW
+	stack_overflow_check(regs);
+#endif
+	__do_IRQ(irq, regs);
+	irq_exit();
+
+	return 1;
+}
+
+#ifdef CONFIG_HOTPLUG_CPU
+void fixup_irqs(cpumask_t map)
+{
+	unsigned int irq;
+	static int warned;
+
+	for (irq = 0; irq < NR_IRQS; irq++) {
+		cpumask_t mask;
+		if (irq == 2)
+			continue;
+
+		cpus_and(mask, irq_desc[irq].affinity, map);
+		if (any_online_cpu(mask) == NR_CPUS) {
+			/*printk("Breaking affinity for irq %i\n", irq);*/
+			mask = map;
+		}
+		if (irq_desc[irq].chip->set_affinity)
+			irq_desc[irq].chip->set_affinity(irq, mask);
+		else if (irq_desc[irq].action && !(warned++))
+			printk("Cannot set affinity for irq %i\n", irq);
+	}
+
+	/* That doesn't seem sufficient.  Give it 1ms. */
+	local_irq_enable();
+	mdelay(1);
+	local_irq_disable();
+}
+#endif
+
+extern void call_softirq(void);
+
+asmlinkage void do_softirq(void)
+{
+ 	__u32 pending;
+ 	unsigned long flags;
+
+ 	if (in_interrupt())
+ 		return;
+
+ 	local_irq_save(flags);
+ 	pending = local_softirq_pending();
+ 	/* Switch to interrupt stack */
+ 	if (pending) {
+		call_softirq();
+		WARN_ON_ONCE(softirq_count());
+	}
+ 	local_irq_restore(flags);
+}
+EXPORT_SYMBOL(do_softirq);
+
+#ifndef CONFIG_X86_LOCAL_APIC
+/*
+ * 'what should we do if we get a hw irq event on an illegal vector'.
+ * each architecture has to answer this themselves.
+ */
+void ack_bad_irq(unsigned int irq)
+{
+        printk("unexpected IRQ trap at vector %02x\n", irq);
+}
+#endif
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/ldt-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/ldt-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/ldt-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/ldt-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,282 @@
+/*
+ * linux/arch/x86_64/kernel/ldt.c
+ *
+ * Copyright (C) 1992 Krishna Balasubramanian and Linus Torvalds
+ * Copyright (C) 1999 Ingo Molnar <mingo@redhat.com>
+ * Copyright (C) 2002 Andi Kleen
+ * 
+ * This handles calls from both 32bit and 64bit mode.
+ */
+
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/string.h>
+#include <linux/mm.h>
+#include <linux/smp.h>
+#include <linux/smp_lock.h>
+#include <linux/vmalloc.h>
+#include <linux/slab.h>
+
+#include <asm/uaccess.h>
+#include <asm/system.h>
+#include <asm/ldt.h>
+#include <asm/desc.h>
+#include <asm/proto.h>
+#include <asm/pgalloc.h>
+
+#ifdef CONFIG_SMP /* avoids "defined but not used" warnig */
+static void flush_ldt(void *null)
+{
+	if (current->active_mm)
+               load_LDT(&current->active_mm->context);
+}
+#endif
+
+static int alloc_ldt(mm_context_t *pc, unsigned mincount, int reload)
+{
+	void *oldldt;
+	void *newldt;
+	unsigned oldsize;
+
+	if (mincount <= (unsigned)pc->size)
+		return 0;
+	oldsize = pc->size;
+	mincount = (mincount+511)&(~511);
+	if (mincount*LDT_ENTRY_SIZE > PAGE_SIZE)
+		newldt = vmalloc(mincount*LDT_ENTRY_SIZE);
+	else
+		newldt = kmalloc(mincount*LDT_ENTRY_SIZE, GFP_KERNEL);
+
+	if (!newldt)
+		return -ENOMEM;
+
+	if (oldsize)
+		memcpy(newldt, pc->ldt, oldsize*LDT_ENTRY_SIZE);
+	oldldt = pc->ldt;
+	memset(newldt+oldsize*LDT_ENTRY_SIZE, 0, (mincount-oldsize)*LDT_ENTRY_SIZE);
+	wmb();
+	pc->ldt = newldt;
+	wmb();
+	pc->size = mincount;
+	wmb();
+	if (reload) {
+#ifdef CONFIG_SMP
+		cpumask_t mask;
+
+		preempt_disable();
+#endif
+		make_pages_readonly(
+			pc->ldt,
+			(pc->size * LDT_ENTRY_SIZE) / PAGE_SIZE,
+			XENFEAT_writable_descriptor_tables);
+		load_LDT(pc);
+#ifdef CONFIG_SMP
+		mask = cpumask_of_cpu(smp_processor_id());
+		if (!cpus_equal(current->mm->cpu_vm_mask, mask))
+			smp_call_function(flush_ldt, NULL, 1, 1);
+		preempt_enable();
+#endif
+	}
+	if (oldsize) {
+		make_pages_writable(
+			oldldt,
+			(oldsize * LDT_ENTRY_SIZE) / PAGE_SIZE,
+			XENFEAT_writable_descriptor_tables);
+		if (oldsize*LDT_ENTRY_SIZE > PAGE_SIZE)
+			vfree(oldldt);
+		else
+			kfree(oldldt);
+	}
+	return 0;
+}
+
+static inline int copy_ldt(mm_context_t *new, mm_context_t *old)
+{
+	int err = alloc_ldt(new, old->size, 0);
+	if (err < 0)
+		return err;
+	memcpy(new->ldt, old->ldt, old->size*LDT_ENTRY_SIZE);
+	make_pages_readonly(
+		new->ldt,
+		(new->size * LDT_ENTRY_SIZE) / PAGE_SIZE,
+		XENFEAT_writable_descriptor_tables);
+	return 0;
+}
+
+/*
+ * we do not have to muck with descriptors here, that is
+ * done in switch_mm() as needed.
+ */
+int init_new_context(struct task_struct *tsk, struct mm_struct *mm)
+{
+	struct mm_struct * old_mm;
+	int retval = 0;
+
+	memset(&mm->context, 0, sizeof(mm->context));
+	init_MUTEX(&mm->context.sem);
+	old_mm = current->mm;
+	if (old_mm && old_mm->context.size > 0) {
+		down(&old_mm->context.sem);
+		retval = copy_ldt(&mm->context, &old_mm->context);
+		up(&old_mm->context.sem);
+	}
+	if (retval == 0) {
+		spin_lock(&mm_unpinned_lock);
+		list_add(&mm->context.unpinned, &mm_unpinned);
+		spin_unlock(&mm_unpinned_lock);
+	}
+	return retval;
+}
+
+/*
+ * 
+ * Don't touch the LDT register - we're already in the next thread.
+ */
+void destroy_context(struct mm_struct *mm)
+{
+	if (mm->context.size) {
+		if (mm == current->active_mm)
+			clear_LDT();
+		make_pages_writable(
+			mm->context.ldt,
+			(mm->context.size * LDT_ENTRY_SIZE) / PAGE_SIZE,
+			XENFEAT_writable_descriptor_tables);
+		if (mm->context.size*LDT_ENTRY_SIZE > PAGE_SIZE)
+			vfree(mm->context.ldt);
+		else
+			kfree(mm->context.ldt);
+		mm->context.size = 0;
+	}
+	if (!mm->context.pinned) {
+		spin_lock(&mm_unpinned_lock);
+		list_del(&mm->context.unpinned);
+		spin_unlock(&mm_unpinned_lock);
+	}
+}
+
+static int read_ldt(void __user * ptr, unsigned long bytecount)
+{
+	int err;
+	unsigned long size;
+	struct mm_struct * mm = current->mm;
+
+	if (!mm->context.size)
+		return 0;
+	if (bytecount > LDT_ENTRY_SIZE*LDT_ENTRIES)
+		bytecount = LDT_ENTRY_SIZE*LDT_ENTRIES;
+
+	down(&mm->context.sem);
+	size = mm->context.size*LDT_ENTRY_SIZE;
+	if (size > bytecount)
+		size = bytecount;
+
+	err = 0;
+	if (copy_to_user(ptr, mm->context.ldt, size))
+		err = -EFAULT;
+	up(&mm->context.sem);
+	if (err < 0)
+		goto error_return;
+	if (size != bytecount) {
+		/* zero-fill the rest */
+		if (clear_user(ptr+size, bytecount-size) != 0) {
+			err = -EFAULT;
+			goto error_return;
+		}
+	}
+	return bytecount;
+error_return:
+	return err;
+}
+
+static int read_default_ldt(void __user * ptr, unsigned long bytecount)
+{
+	/* Arbitrary number */ 
+	/* x86-64 default LDT is all zeros */
+	if (bytecount > 128) 
+		bytecount = 128; 	
+	if (clear_user(ptr, bytecount))
+		return -EFAULT;
+	return bytecount; 
+}
+
+static int write_ldt(void __user * ptr, unsigned long bytecount, int oldmode)
+{
+	struct task_struct *me = current;
+	struct mm_struct * mm = me->mm;
+	__u32 entry_1, entry_2, *lp;
+	unsigned long mach_lp;
+	int error;
+	struct user_desc ldt_info;
+
+	error = -EINVAL;
+
+	if (bytecount != sizeof(ldt_info))
+		goto out;
+	error = -EFAULT; 	
+	if (copy_from_user(&ldt_info, ptr, bytecount))
+		goto out;
+
+	error = -EINVAL;
+	if (ldt_info.entry_number >= LDT_ENTRIES)
+		goto out;
+	if (ldt_info.contents == 3) {
+		if (oldmode)
+			goto out;
+		if (ldt_info.seg_not_present == 0)
+			goto out;
+	}
+
+	down(&mm->context.sem);
+	if (ldt_info.entry_number >= (unsigned)mm->context.size) {
+		error = alloc_ldt(&current->mm->context, ldt_info.entry_number+1, 1);
+		if (error < 0)
+			goto out_unlock;
+	}
+
+	lp = (__u32 *) ((ldt_info.entry_number << 3) + (char *) mm->context.ldt);
+ 	mach_lp = arbitrary_virt_to_machine(lp);
+
+   	/* Allow LDTs to be cleared by the user. */
+   	if (ldt_info.base_addr == 0 && ldt_info.limit == 0) {
+		if (oldmode || LDT_empty(&ldt_info)) {
+			entry_1 = 0;
+			entry_2 = 0;
+			goto install;
+		}
+	}
+
+	entry_1 = LDT_entry_a(&ldt_info);
+	entry_2 = LDT_entry_b(&ldt_info);
+	if (oldmode)
+		entry_2 &= ~(1 << 20);
+
+	/* Install the new entry ...  */
+install:
+	error = HYPERVISOR_update_descriptor(mach_lp, (unsigned long)((entry_1 | (unsigned long) entry_2 << 32)));
+
+out_unlock:
+	up(&mm->context.sem);
+out:
+	return error;
+}
+
+asmlinkage int sys_modify_ldt(int func, void __user *ptr, unsigned long bytecount)
+{
+	int ret = -ENOSYS;
+
+	switch (func) {
+	case 0:
+		ret = read_ldt(ptr, bytecount);
+		break;
+	case 1:
+		ret = write_ldt(ptr, bytecount, 1);
+		break;
+	case 2:
+		ret = read_default_ldt(ptr, bytecount);
+		break;
+	case 0x11:
+		ret = write_ldt(ptr, bytecount, 0);
+		break;
+	}
+	return ret;
+}
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/machine_kexec.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/machine_kexec.c
--- linux-2.6.18.8/arch/x86_64/kernel/machine_kexec.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/machine_kexec.c	2008-08-21 11:36:07.000000000 +0200
@@ -15,6 +15,128 @@
 #include <asm/mmu_context.h>
 #include <asm/io.h>
 
+#define PAGE_ALIGNED __attribute__ ((__aligned__(PAGE_SIZE)))
+static u64 kexec_pgd[512] PAGE_ALIGNED;
+static u64 kexec_pud0[512] PAGE_ALIGNED;
+static u64 kexec_pmd0[512] PAGE_ALIGNED;
+static u64 kexec_pte0[512] PAGE_ALIGNED;
+static u64 kexec_pud1[512] PAGE_ALIGNED;
+static u64 kexec_pmd1[512] PAGE_ALIGNED;
+static u64 kexec_pte1[512] PAGE_ALIGNED;
+
+#ifdef CONFIG_XEN
+
+/* In the case of Xen, override hypervisor functions to be able to create
+ * a regular identity mapping page table...
+ */
+
+#include <xen/interface/kexec.h>
+#include <xen/interface/memory.h>
+
+#define x__pmd(x) ((pmd_t) { (x) } )
+#define x__pud(x) ((pud_t) { (x) } )
+#define x__pgd(x) ((pgd_t) { (x) } )
+
+#define x_pmd_val(x)   ((x).pmd)
+#define x_pud_val(x)   ((x).pud)
+#define x_pgd_val(x)   ((x).pgd)
+
+static inline void x_set_pmd(pmd_t *dst, pmd_t val)
+{
+	x_pmd_val(*dst) = x_pmd_val(val);
+}
+
+static inline void x_set_pud(pud_t *dst, pud_t val)
+{
+	x_pud_val(*dst) = phys_to_machine(x_pud_val(val));
+}
+
+static inline void x_pud_clear (pud_t *pud)
+{
+	x_pud_val(*pud) = 0;
+}
+
+static inline void x_set_pgd(pgd_t *dst, pgd_t val)
+{
+	x_pgd_val(*dst) = phys_to_machine(x_pgd_val(val));
+}
+
+static inline void x_pgd_clear (pgd_t * pgd)
+{
+	x_pgd_val(*pgd) = 0;
+}
+
+#define X__PAGE_KERNEL_LARGE_EXEC \
+         _PAGE_PRESENT | _PAGE_RW | _PAGE_DIRTY | _PAGE_ACCESSED | _PAGE_PSE
+#define X_KERNPG_TABLE _PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY
+
+#define __ma(x) (pfn_to_mfn(__pa((x)) >> PAGE_SHIFT) << PAGE_SHIFT)
+
+#if PAGES_NR > KEXEC_XEN_NO_PAGES
+#error PAGES_NR is greater than KEXEC_XEN_NO_PAGES - Xen support will break
+#endif
+
+#if PA_CONTROL_PAGE != 0
+#error PA_CONTROL_PAGE is non zero - Xen support will break
+#endif
+
+void machine_kexec_setup_load_arg(xen_kexec_image_t *xki, struct kimage *image)
+{
+	void *control_page;
+	void *table_page;
+
+	memset(xki->page_list, 0, sizeof(xki->page_list));
+
+	control_page = page_address(image->control_code_page) + PAGE_SIZE;
+	memcpy(control_page, relocate_kernel, PAGE_SIZE);
+
+	table_page = page_address(image->control_code_page);
+
+	xki->page_list[PA_CONTROL_PAGE] = __ma(control_page);
+	xki->page_list[PA_TABLE_PAGE] = __ma(table_page);
+
+	xki->page_list[PA_PGD] = __ma(kexec_pgd);
+	xki->page_list[PA_PUD_0] = __ma(kexec_pud0);
+	xki->page_list[PA_PUD_1] = __ma(kexec_pud1);
+	xki->page_list[PA_PMD_0] = __ma(kexec_pmd0);
+	xki->page_list[PA_PMD_1] = __ma(kexec_pmd1);
+	xki->page_list[PA_PTE_0] = __ma(kexec_pte0);
+	xki->page_list[PA_PTE_1] = __ma(kexec_pte1);
+}
+
+int __init machine_kexec_setup_resources(struct resource *hypervisor,
+					 struct resource *phys_cpus,
+					 int nr_phys_cpus)
+{
+	int k;
+
+	/* The per-cpu crash note resources belong to the hypervisor resource */
+	for (k = 0; k < nr_phys_cpus; k++)
+		request_resource(hypervisor, phys_cpus + k);
+
+	return 0;
+}
+
+void machine_kexec_register_resources(struct resource *res) { ; }
+
+#else /* CONFIG_XEN */
+
+#define x__pmd(x) __pmd(x)
+#define x__pud(x) __pud(x)
+#define x__pgd(x) __pgd(x)
+
+#define x_set_pmd(x, y) set_pmd(x, y)
+#define x_set_pud(x, y) set_pud(x, y)
+#define x_set_pgd(x, y) set_pgd(x, y)
+
+#define x_pud_clear(x) pud_clear(x)
+#define x_pgd_clear(x) pgd_clear(x)
+
+#define X__PAGE_KERNEL_LARGE_EXEC __PAGE_KERNEL_LARGE_EXEC
+#define X_KERNPG_TABLE _KERNPG_TABLE
+
+#endif /* CONFIG_XEN */
+
 static void init_level2_page(pmd_t *level2p, unsigned long addr)
 {
 	unsigned long end_addr;
@@ -22,7 +144,7 @@ static void init_level2_page(pmd_t *leve
 	addr &= PAGE_MASK;
 	end_addr = addr + PUD_SIZE;
 	while (addr < end_addr) {
-		set_pmd(level2p++, __pmd(addr | __PAGE_KERNEL_LARGE_EXEC));
+		x_set_pmd(level2p++, x__pmd(addr | X__PAGE_KERNEL_LARGE_EXEC));
 		addr += PMD_SIZE;
 	}
 }
@@ -47,12 +169,12 @@ static int init_level3_page(struct kimag
 		}
 		level2p = (pmd_t *)page_address(page);
 		init_level2_page(level2p, addr);
-		set_pud(level3p++, __pud(__pa(level2p) | _KERNPG_TABLE));
+		x_set_pud(level3p++, x__pud(__pa(level2p) | X_KERNPG_TABLE));
 		addr += PUD_SIZE;
 	}
 	/* clear the unused entries */
 	while (addr < end_addr) {
-		pud_clear(level3p++);
+		x_pud_clear(level3p++);
 		addr += PUD_SIZE;
 	}
 out:
@@ -83,12 +205,12 @@ static int init_level4_page(struct kimag
 		if (result) {
 			goto out;
 		}
-		set_pgd(level4p++, __pgd(__pa(level3p) | _KERNPG_TABLE));
+		x_set_pgd(level4p++, x__pgd(__pa(level3p) | X_KERNPG_TABLE));
 		addr += PGDIR_SIZE;
 	}
 	/* clear the unused entries */
 	while (addr < end_addr) {
-		pgd_clear(level4p++);
+		x_pgd_clear(level4p++);
 		addr += PGDIR_SIZE;
 	}
 out:
@@ -99,77 +221,29 @@ out:
 static int init_pgtable(struct kimage *image, unsigned long start_pgtable)
 {
 	pgd_t *level4p;
-	level4p = (pgd_t *)__va(start_pgtable);
- 	return init_level4_page(image, level4p, 0, end_pfn << PAGE_SHIFT);
-}
-
-static void set_idt(void *newidt, u16 limit)
-{
-	struct desc_ptr curidt;
-
-	/* x86-64 supports unaliged loads & stores */
-	curidt.size    = limit;
-	curidt.address = (unsigned long)newidt;
+	unsigned long x_end_pfn = end_pfn;
 
-	__asm__ __volatile__ (
-		"lidtq %0\n"
-		: : "m" (curidt)
-		);
-};
+#ifdef CONFIG_XEN
+	x_end_pfn = HYPERVISOR_memory_op(XENMEM_maximum_ram_page, NULL);
+#endif
 
-
-static void set_gdt(void *newgdt, u16 limit)
-{
-	struct desc_ptr curgdt;
-
-	/* x86-64 supports unaligned loads & stores */
-	curgdt.size    = limit;
-	curgdt.address = (unsigned long)newgdt;
-
-	__asm__ __volatile__ (
-		"lgdtq %0\n"
-		: : "m" (curgdt)
-		);
-};
-
-static void load_segments(void)
-{
-	__asm__ __volatile__ (
-		"\tmovl %0,%%ds\n"
-		"\tmovl %0,%%es\n"
-		"\tmovl %0,%%ss\n"
-		"\tmovl %0,%%fs\n"
-		"\tmovl %0,%%gs\n"
-		: : "a" (__KERNEL_DS) : "memory"
-		);
+	level4p = (pgd_t *)__va(start_pgtable);
+ 	return init_level4_page(image, level4p, 0, x_end_pfn << PAGE_SHIFT);
 }
 
-typedef NORET_TYPE void (*relocate_new_kernel_t)(unsigned long indirection_page,
-					unsigned long control_code_buffer,
-					unsigned long start_address,
-					unsigned long pgtable) ATTRIB_NORET;
-
-extern const unsigned char relocate_new_kernel[];
-extern const unsigned long relocate_new_kernel_size;
-
 int machine_kexec_prepare(struct kimage *image)
 {
-	unsigned long start_pgtable, control_code_buffer;
+	unsigned long start_pgtable;
 	int result;
 
 	/* Calculate the offsets */
 	start_pgtable = page_to_pfn(image->control_code_page) << PAGE_SHIFT;
-	control_code_buffer = start_pgtable + PAGE_SIZE;
 
 	/* Setup the identity mapped 64bit page table */
 	result = init_pgtable(image, start_pgtable);
 	if (result)
 		return result;
 
-	/* Place the code in the reboot code buffer */
-	memcpy(__va(control_code_buffer), relocate_new_kernel,
-						relocate_new_kernel_size);
-
 	return 0;
 }
 
@@ -178,51 +252,43 @@ void machine_kexec_cleanup(struct kimage
 	return;
 }
 
+#ifndef CONFIG_XEN
 /*
  * Do not allocate memory (or fail in any way) in machine_kexec().
  * We are past the point of no return, committed to rebooting now.
  */
 NORET_TYPE void machine_kexec(struct kimage *image)
 {
-	unsigned long page_list;
-	unsigned long control_code_buffer;
-	unsigned long start_pgtable;
-	relocate_new_kernel_t rnk;
+	unsigned long page_list[PAGES_NR];
+	void *control_page;
 
 	/* Interrupts aren't acceptable while we reboot */
 	local_irq_disable();
 
-	/* Calculate the offsets */
-	page_list = image->head;
-	start_pgtable = page_to_pfn(image->control_code_page) << PAGE_SHIFT;
-	control_code_buffer = start_pgtable + PAGE_SIZE;
+	control_page = page_address(image->control_code_page) + PAGE_SIZE;
+	memcpy(control_page, relocate_kernel, PAGE_SIZE);
+
+	page_list[PA_CONTROL_PAGE] = __pa(control_page);
+	page_list[VA_CONTROL_PAGE] = (unsigned long)relocate_kernel;
+	page_list[PA_PGD] = __pa(kexec_pgd);
+	page_list[VA_PGD] = (unsigned long)kexec_pgd;
+	page_list[PA_PUD_0] = __pa(kexec_pud0);
+	page_list[VA_PUD_0] = (unsigned long)kexec_pud0;
+	page_list[PA_PMD_0] = __pa(kexec_pmd0);
+	page_list[VA_PMD_0] = (unsigned long)kexec_pmd0;
+	page_list[PA_PTE_0] = __pa(kexec_pte0);
+	page_list[VA_PTE_0] = (unsigned long)kexec_pte0;
+	page_list[PA_PUD_1] = __pa(kexec_pud1);
+	page_list[VA_PUD_1] = (unsigned long)kexec_pud1;
+	page_list[PA_PMD_1] = __pa(kexec_pmd1);
+	page_list[VA_PMD_1] = (unsigned long)kexec_pmd1;
+	page_list[PA_PTE_1] = __pa(kexec_pte1);
+	page_list[VA_PTE_1] = (unsigned long)kexec_pte1;
+
+	page_list[PA_TABLE_PAGE] =
+	  (unsigned long)__pa(page_address(image->control_code_page));
 
-	/* Set the low half of the page table to my identity mapped
-	 * page table for kexec.  Leave the high half pointing at the
-	 * kernel pages.   Don't bother to flush the global pages
-	 * as that will happen when I fully switch to my identity mapped
-	 * page table anyway.
-	 */
-	memcpy(__va(read_cr3()), __va(start_pgtable), PAGE_SIZE/2);
-	__flush_tlb();
-
-
-	/* The segment registers are funny things, they have both a
-	 * visible and an invisible part.  Whenever the visible part is
-	 * set to a specific selector, the invisible part is loaded
-	 * with from a table in memory.  At no other time is the
-	 * descriptor table in memory accessed.
-	 *
-	 * I take advantage of this here by force loading the
-	 * segments, before I zap the gdt with an invalid value.
-	 */
-	load_segments();
-	/* The gdt & idt are now invalid.
-	 * If you want to load them you must set up your own idt & gdt.
-	 */
-	set_gdt(phys_to_virt(0),0);
-	set_idt(phys_to_virt(0),0);
-	/* now call it */
-	rnk = (relocate_new_kernel_t) control_code_buffer;
-	(*rnk)(page_list, control_code_buffer, image->start, start_pgtable);
+	relocate_kernel((unsigned long)image->head, (unsigned long)page_list,
+			image->start);
 }
+#endif
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/Makefile linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/Makefile
--- linux-2.6.18.8/arch/x86_64/kernel/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -21,11 +21,13 @@ obj-$(CONFIG_MICROCODE)		+= microcode.o
 obj-$(CONFIG_X86_CPUID)		+= cpuid.o
 obj-$(CONFIG_SMP)		+= smp.o smpboot.o trampoline.o
 obj-$(CONFIG_X86_LOCAL_APIC)	+= apic.o  nmi.o
+obj-$(CONFIG_X86_XEN_GENAPIC)	+= genapic.o genapic_xen.o
 obj-$(CONFIG_X86_IO_APIC)	+= io_apic.o mpparse.o \
 		genapic.o genapic_cluster.o genapic_flat.o
 obj-$(CONFIG_KEXEC)		+= machine_kexec.o relocate_kernel.o crash.o
 obj-$(CONFIG_CRASH_DUMP)	+= crash_dump.o
-obj-$(CONFIG_PM)		+= suspend.o
+obj-$(CONFIG_SOFTWARE_SUSPEND)	+= suspend.o
+obj-$(CONFIG_ACPI_SLEEP)	+= suspend.o
 obj-$(CONFIG_SOFTWARE_SUSPEND)	+= suspend_asm.o
 obj-$(CONFIG_CPU_FREQ)		+= cpufreq/
 obj-$(CONFIG_EARLY_PRINTK)	+= early_printk.o
@@ -55,3 +57,8 @@ i8237-y				+= ../../i386/kernel/i8237.o
 msr-$(subst m,y,$(CONFIG_X86_MSR))  += ../../i386/kernel/msr.o
 alternative-y			+= ../../i386/kernel/alternative.o
 
+time-$(CONFIG_XEN)		+= ../../i386/kernel/time.o
+pci-dma-$(CONFIG_XEN)		+= ../../i386/kernel/pci-dma.o
+
+disabled-obj-$(CONFIG_XEN)	:= i8259.o reboot.o smpboot.o trampoline.o
+%/head.o %/head.s: $(if $(CONFIG_XEN),EXTRA_AFLAGS,dummy) :=
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/mpparse-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/mpparse-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/mpparse-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/mpparse-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1011 @@
+/*
+ *	Intel Multiprocessor Specification 1.1 and 1.4
+ *	compliant MP-table parsing routines.
+ *
+ *	(c) 1995 Alan Cox, Building #3 <alan@redhat.com>
+ *	(c) 1998, 1999, 2000 Ingo Molnar <mingo@redhat.com>
+ *
+ *	Fixes
+ *		Erich Boleyn	:	MP v1.4 and additional changes.
+ *		Alan Cox	:	Added EBDA scanning
+ *		Ingo Molnar	:	various cleanups and rewrites
+ *		Maciej W. Rozycki:	Bits for default MP configurations
+ *		Paul Diefenbaugh:	Added full ACPI support
+ */
+
+#include <linux/mm.h>
+#include <linux/init.h>
+#include <linux/delay.h>
+#include <linux/bootmem.h>
+#include <linux/smp_lock.h>
+#include <linux/kernel_stat.h>
+#include <linux/mc146818rtc.h>
+#include <linux/acpi.h>
+#include <linux/module.h>
+
+#include <asm/smp.h>
+#include <asm/mtrr.h>
+#include <asm/mpspec.h>
+#include <asm/pgalloc.h>
+#include <asm/io_apic.h>
+#include <asm/proto.h>
+#include <asm/acpi.h>
+
+/* Have we found an MP table */
+int smp_found_config;
+unsigned int __initdata maxcpus = NR_CPUS;
+
+int acpi_found_madt;
+
+/*
+ * Various Linux-internal data structures created from the
+ * MP-table.
+ */
+unsigned char apic_version [MAX_APICS];
+unsigned char mp_bus_id_to_type [MAX_MP_BUSSES] = { [0 ... MAX_MP_BUSSES-1] = -1 };
+int mp_bus_id_to_pci_bus [MAX_MP_BUSSES] = { [0 ... MAX_MP_BUSSES-1] = -1 };
+
+static int mp_current_pci_id = 0;
+/* I/O APIC entries */
+struct mpc_config_ioapic mp_ioapics[MAX_IO_APICS];
+
+/* # of MP IRQ source entries */
+struct mpc_config_intsrc mp_irqs[MAX_IRQ_SOURCES];
+
+/* MP IRQ source entries */
+int mp_irq_entries;
+
+int nr_ioapics;
+int pic_mode;
+unsigned long mp_lapic_addr = 0;
+
+
+
+/* Processor that is doing the boot up */
+unsigned int boot_cpu_id = -1U;
+/* Internal processor count */
+unsigned int num_processors __initdata = 0;
+
+unsigned disabled_cpus __initdata;
+
+/* Bitmask of physically existing CPUs */
+physid_mask_t phys_cpu_present_map = PHYSID_MASK_NONE;
+
+/* ACPI MADT entry parsing functions */
+#ifdef CONFIG_ACPI
+extern struct acpi_boot_flags acpi_boot;
+#ifdef CONFIG_X86_LOCAL_APIC
+extern int acpi_parse_lapic (acpi_table_entry_header *header);
+extern int acpi_parse_lapic_addr_ovr (acpi_table_entry_header *header);
+extern int acpi_parse_lapic_nmi (acpi_table_entry_header *header);
+#endif /*CONFIG_X86_LOCAL_APIC*/
+#ifdef CONFIG_X86_IO_APIC
+extern int acpi_parse_ioapic (acpi_table_entry_header *header);
+#endif /*CONFIG_X86_IO_APIC*/
+#endif /*CONFIG_ACPI*/
+
+u8 bios_cpu_apicid[NR_CPUS] = { [0 ... NR_CPUS-1] = BAD_APICID };
+
+
+/*
+ * Intel MP BIOS table parsing routines:
+ */
+
+/*
+ * Checksum an MP configuration block.
+ */
+
+static int __init mpf_checksum(unsigned char *mp, int len)
+{
+	int sum = 0;
+
+	while (len--)
+		sum += *mp++;
+
+	return sum & 0xFF;
+}
+
+#ifndef CONFIG_XEN
+static void __cpuinit MP_processor_info (struct mpc_config_processor *m)
+{
+	int cpu;
+	unsigned char ver;
+	cpumask_t tmp_map;
+
+	if (!(m->mpc_cpuflag & CPU_ENABLED)) {
+		disabled_cpus++;
+		return;
+	}
+
+	printk(KERN_INFO "Processor #%d %d:%d APIC version %d\n",
+		m->mpc_apicid,
+	       (m->mpc_cpufeature & CPU_FAMILY_MASK)>>8,
+	       (m->mpc_cpufeature & CPU_MODEL_MASK)>>4,
+		m->mpc_apicver);
+
+	if (m->mpc_cpuflag & CPU_BOOTPROCESSOR) {
+		Dprintk("    Bootup CPU\n");
+		boot_cpu_id = m->mpc_apicid;
+	}
+	if (num_processors >= NR_CPUS) {
+		printk(KERN_WARNING "WARNING: NR_CPUS limit of %i reached."
+			" Processor ignored.\n", NR_CPUS);
+		return;
+	}
+
+	num_processors++;
+	cpus_complement(tmp_map, cpu_present_map);
+	cpu = first_cpu(tmp_map);
+
+#if MAX_APICS < 255	
+	if ((int)m->mpc_apicid > MAX_APICS) {
+		printk(KERN_ERR "Processor #%d INVALID. (Max ID: %d).\n",
+			m->mpc_apicid, MAX_APICS);
+		return;
+	}
+#endif
+	ver = m->mpc_apicver;
+
+	physid_set(m->mpc_apicid, phys_cpu_present_map);
+	/*
+	 * Validate version
+	 */
+	if (ver == 0x0) {
+		printk(KERN_ERR "BIOS bug, APIC version is 0 for CPU#%d! fixing up to 0x10. (tell your hw vendor)\n", m->mpc_apicid);
+		ver = 0x10;
+	}
+	apic_version[m->mpc_apicid] = ver;
+ 	if (m->mpc_cpuflag & CPU_BOOTPROCESSOR) {
+ 		/*
+ 		 * bios_cpu_apicid is required to have processors listed
+ 		 * in same order as logical cpu numbers. Hence the first
+ 		 * entry is BSP, and so on.
+ 		 */
+		cpu = 0;
+ 	}
+	bios_cpu_apicid[cpu] = m->mpc_apicid;
+	x86_cpu_to_apicid[cpu] = m->mpc_apicid;
+
+	cpu_set(cpu, cpu_possible_map);
+	cpu_set(cpu, cpu_present_map);
+}
+#else
+static void __cpuinit MP_processor_info (struct mpc_config_processor *m)
+{
+	num_processors++;
+}
+#endif /* CONFIG_XEN */
+
+static void __init MP_bus_info (struct mpc_config_bus *m)
+{
+	char str[7];
+
+	memcpy(str, m->mpc_bustype, 6);
+	str[6] = 0;
+	Dprintk("Bus #%d is %s\n", m->mpc_busid, str);
+
+	if (strncmp(str, "ISA", 3) == 0) {
+		mp_bus_id_to_type[m->mpc_busid] = MP_BUS_ISA;
+	} else if (strncmp(str, "EISA", 4) == 0) {
+		mp_bus_id_to_type[m->mpc_busid] = MP_BUS_EISA;
+	} else if (strncmp(str, "PCI", 3) == 0) {
+		mp_bus_id_to_type[m->mpc_busid] = MP_BUS_PCI;
+		mp_bus_id_to_pci_bus[m->mpc_busid] = mp_current_pci_id;
+		mp_current_pci_id++;
+	} else if (strncmp(str, "MCA", 3) == 0) {
+		mp_bus_id_to_type[m->mpc_busid] = MP_BUS_MCA;
+	} else {
+		printk(KERN_ERR "Unknown bustype %s\n", str);
+	}
+}
+
+static void __init MP_ioapic_info (struct mpc_config_ioapic *m)
+{
+	if (!(m->mpc_flags & MPC_APIC_USABLE))
+		return;
+
+	printk("I/O APIC #%d Version %d at 0x%X.\n",
+		m->mpc_apicid, m->mpc_apicver, m->mpc_apicaddr);
+	if (nr_ioapics >= MAX_IO_APICS) {
+		printk(KERN_ERR "Max # of I/O APICs (%d) exceeded (found %d).\n",
+			MAX_IO_APICS, nr_ioapics);
+		panic("Recompile kernel with bigger MAX_IO_APICS!.\n");
+	}
+	if (!m->mpc_apicaddr) {
+		printk(KERN_ERR "WARNING: bogus zero I/O APIC address"
+			" found in MP table, skipping!\n");
+		return;
+	}
+	mp_ioapics[nr_ioapics] = *m;
+	nr_ioapics++;
+}
+
+static void __init MP_intsrc_info (struct mpc_config_intsrc *m)
+{
+	mp_irqs [mp_irq_entries] = *m;
+	Dprintk("Int: type %d, pol %d, trig %d, bus %d,"
+		" IRQ %02x, APIC ID %x, APIC INT %02x\n",
+			m->mpc_irqtype, m->mpc_irqflag & 3,
+			(m->mpc_irqflag >> 2) & 3, m->mpc_srcbus,
+			m->mpc_srcbusirq, m->mpc_dstapic, m->mpc_dstirq);
+	if (++mp_irq_entries >= MAX_IRQ_SOURCES)
+		panic("Max # of irq sources exceeded!!\n");
+}
+
+static void __init MP_lintsrc_info (struct mpc_config_lintsrc *m)
+{
+	Dprintk("Lint: type %d, pol %d, trig %d, bus %d,"
+		" IRQ %02x, APIC ID %x, APIC LINT %02x\n",
+			m->mpc_irqtype, m->mpc_irqflag & 3,
+			(m->mpc_irqflag >> 2) &3, m->mpc_srcbusid,
+			m->mpc_srcbusirq, m->mpc_destapic, m->mpc_destapiclint);
+	/*
+	 * Well it seems all SMP boards in existence
+	 * use ExtINT/LVT1 == LINT0 and
+	 * NMI/LVT2 == LINT1 - the following check
+	 * will show us if this assumptions is false.
+	 * Until then we do not have to add baggage.
+	 */
+	if ((m->mpc_irqtype == mp_ExtINT) &&
+		(m->mpc_destapiclint != 0))
+			BUG();
+	if ((m->mpc_irqtype == mp_NMI) &&
+		(m->mpc_destapiclint != 1))
+			BUG();
+}
+
+/*
+ * Read/parse the MPC
+ */
+
+static int __init smp_read_mpc(struct mp_config_table *mpc)
+{
+	char str[16];
+	int count=sizeof(*mpc);
+	unsigned char *mpt=((unsigned char *)mpc)+count;
+
+	if (memcmp(mpc->mpc_signature,MPC_SIGNATURE,4)) {
+		printk("SMP mptable: bad signature [%c%c%c%c]!\n",
+			mpc->mpc_signature[0],
+			mpc->mpc_signature[1],
+			mpc->mpc_signature[2],
+			mpc->mpc_signature[3]);
+		return 0;
+	}
+	if (mpf_checksum((unsigned char *)mpc,mpc->mpc_length)) {
+		printk("SMP mptable: checksum error!\n");
+		return 0;
+	}
+	if (mpc->mpc_spec!=0x01 && mpc->mpc_spec!=0x04) {
+		printk(KERN_ERR "SMP mptable: bad table version (%d)!!\n",
+			mpc->mpc_spec);
+		return 0;
+	}
+	if (!mpc->mpc_lapic) {
+		printk(KERN_ERR "SMP mptable: null local APIC address!\n");
+		return 0;
+	}
+	memcpy(str,mpc->mpc_oem,8);
+	str[8]=0;
+	printk(KERN_INFO "OEM ID: %s ",str);
+
+	memcpy(str,mpc->mpc_productid,12);
+	str[12]=0;
+	printk("Product ID: %s ",str);
+
+	printk("APIC at: 0x%X\n",mpc->mpc_lapic);
+
+	/* save the local APIC address, it might be non-default */
+	if (!acpi_lapic)
+	mp_lapic_addr = mpc->mpc_lapic;
+
+	/*
+	 *	Now process the configuration blocks.
+	 */
+	while (count < mpc->mpc_length) {
+		switch(*mpt) {
+			case MP_PROCESSOR:
+			{
+				struct mpc_config_processor *m=
+					(struct mpc_config_processor *)mpt;
+				if (!acpi_lapic)
+				MP_processor_info(m);
+				mpt += sizeof(*m);
+				count += sizeof(*m);
+				break;
+			}
+			case MP_BUS:
+			{
+				struct mpc_config_bus *m=
+					(struct mpc_config_bus *)mpt;
+				MP_bus_info(m);
+				mpt += sizeof(*m);
+				count += sizeof(*m);
+				break;
+			}
+			case MP_IOAPIC:
+			{
+				struct mpc_config_ioapic *m=
+					(struct mpc_config_ioapic *)mpt;
+				MP_ioapic_info(m);
+				mpt+=sizeof(*m);
+				count+=sizeof(*m);
+				break;
+			}
+			case MP_INTSRC:
+			{
+				struct mpc_config_intsrc *m=
+					(struct mpc_config_intsrc *)mpt;
+
+				MP_intsrc_info(m);
+				mpt+=sizeof(*m);
+				count+=sizeof(*m);
+				break;
+			}
+			case MP_LINTSRC:
+			{
+				struct mpc_config_lintsrc *m=
+					(struct mpc_config_lintsrc *)mpt;
+				MP_lintsrc_info(m);
+				mpt+=sizeof(*m);
+				count+=sizeof(*m);
+				break;
+			}
+		}
+	}
+	clustered_apic_check();
+	if (!num_processors)
+		printk(KERN_ERR "SMP mptable: no processors registered!\n");
+	return num_processors;
+}
+
+static int __init ELCR_trigger(unsigned int irq)
+{
+	unsigned int port;
+
+	port = 0x4d0 + (irq >> 3);
+	return (inb(port) >> (irq & 7)) & 1;
+}
+
+static void __init construct_default_ioirq_mptable(int mpc_default_type)
+{
+	struct mpc_config_intsrc intsrc;
+	int i;
+	int ELCR_fallback = 0;
+
+	intsrc.mpc_type = MP_INTSRC;
+	intsrc.mpc_irqflag = 0;			/* conforming */
+	intsrc.mpc_srcbus = 0;
+	intsrc.mpc_dstapic = mp_ioapics[0].mpc_apicid;
+
+	intsrc.mpc_irqtype = mp_INT;
+
+	/*
+	 *  If true, we have an ISA/PCI system with no IRQ entries
+	 *  in the MP table. To prevent the PCI interrupts from being set up
+	 *  incorrectly, we try to use the ELCR. The sanity check to see if
+	 *  there is good ELCR data is very simple - IRQ0, 1, 2 and 13 can
+	 *  never be level sensitive, so we simply see if the ELCR agrees.
+	 *  If it does, we assume it's valid.
+	 */
+	if (mpc_default_type == 5) {
+		printk(KERN_INFO "ISA/PCI bus type with no IRQ information... falling back to ELCR\n");
+
+		if (ELCR_trigger(0) || ELCR_trigger(1) || ELCR_trigger(2) || ELCR_trigger(13))
+			printk(KERN_ERR "ELCR contains invalid data... not using ELCR\n");
+		else {
+			printk(KERN_INFO "Using ELCR to identify PCI interrupts\n");
+			ELCR_fallback = 1;
+		}
+	}
+
+	for (i = 0; i < 16; i++) {
+		switch (mpc_default_type) {
+		case 2:
+			if (i == 0 || i == 13)
+				continue;	/* IRQ0 & IRQ13 not connected */
+			/* fall through */
+		default:
+			if (i == 2)
+				continue;	/* IRQ2 is never connected */
+		}
+
+		if (ELCR_fallback) {
+			/*
+			 *  If the ELCR indicates a level-sensitive interrupt, we
+			 *  copy that information over to the MP table in the
+			 *  irqflag field (level sensitive, active high polarity).
+			 */
+			if (ELCR_trigger(i))
+				intsrc.mpc_irqflag = 13;
+			else
+				intsrc.mpc_irqflag = 0;
+		}
+
+		intsrc.mpc_srcbusirq = i;
+		intsrc.mpc_dstirq = i ? i : 2;		/* IRQ0 to INTIN2 */
+		MP_intsrc_info(&intsrc);
+	}
+
+	intsrc.mpc_irqtype = mp_ExtINT;
+	intsrc.mpc_srcbusirq = 0;
+	intsrc.mpc_dstirq = 0;				/* 8259A to INTIN0 */
+	MP_intsrc_info(&intsrc);
+}
+
+static inline void __init construct_default_ISA_mptable(int mpc_default_type)
+{
+	struct mpc_config_processor processor;
+	struct mpc_config_bus bus;
+	struct mpc_config_ioapic ioapic;
+	struct mpc_config_lintsrc lintsrc;
+	int linttypes[2] = { mp_ExtINT, mp_NMI };
+	int i;
+
+	/*
+	 * local APIC has default address
+	 */
+	mp_lapic_addr = APIC_DEFAULT_PHYS_BASE;
+
+	/*
+	 * 2 CPUs, numbered 0 & 1.
+	 */
+	processor.mpc_type = MP_PROCESSOR;
+	/* Either an integrated APIC or a discrete 82489DX. */
+	processor.mpc_apicver = mpc_default_type > 4 ? 0x10 : 0x01;
+	processor.mpc_cpuflag = CPU_ENABLED;
+	processor.mpc_cpufeature = (boot_cpu_data.x86 << 8) |
+				   (boot_cpu_data.x86_model << 4) |
+				   boot_cpu_data.x86_mask;
+	processor.mpc_featureflag = boot_cpu_data.x86_capability[0];
+	processor.mpc_reserved[0] = 0;
+	processor.mpc_reserved[1] = 0;
+	for (i = 0; i < 2; i++) {
+		processor.mpc_apicid = i;
+		MP_processor_info(&processor);
+	}
+
+	bus.mpc_type = MP_BUS;
+	bus.mpc_busid = 0;
+	switch (mpc_default_type) {
+		default:
+			printk(KERN_ERR "???\nUnknown standard configuration %d\n",
+				mpc_default_type);
+			/* fall through */
+		case 1:
+		case 5:
+			memcpy(bus.mpc_bustype, "ISA   ", 6);
+			break;
+		case 2:
+		case 6:
+		case 3:
+			memcpy(bus.mpc_bustype, "EISA  ", 6);
+			break;
+		case 4:
+		case 7:
+			memcpy(bus.mpc_bustype, "MCA   ", 6);
+	}
+	MP_bus_info(&bus);
+	if (mpc_default_type > 4) {
+		bus.mpc_busid = 1;
+		memcpy(bus.mpc_bustype, "PCI   ", 6);
+		MP_bus_info(&bus);
+	}
+
+	ioapic.mpc_type = MP_IOAPIC;
+	ioapic.mpc_apicid = 2;
+	ioapic.mpc_apicver = mpc_default_type > 4 ? 0x10 : 0x01;
+	ioapic.mpc_flags = MPC_APIC_USABLE;
+	ioapic.mpc_apicaddr = 0xFEC00000;
+	MP_ioapic_info(&ioapic);
+
+	/*
+	 * We set up most of the low 16 IO-APIC pins according to MPS rules.
+	 */
+	construct_default_ioirq_mptable(mpc_default_type);
+
+	lintsrc.mpc_type = MP_LINTSRC;
+	lintsrc.mpc_irqflag = 0;		/* conforming */
+	lintsrc.mpc_srcbusid = 0;
+	lintsrc.mpc_srcbusirq = 0;
+	lintsrc.mpc_destapic = MP_APIC_ALL;
+	for (i = 0; i < 2; i++) {
+		lintsrc.mpc_irqtype = linttypes[i];
+		lintsrc.mpc_destapiclint = i;
+		MP_lintsrc_info(&lintsrc);
+	}
+}
+
+static struct intel_mp_floating *mpf_found;
+
+/*
+ * Scan the memory blocks for an SMP configuration block.
+ */
+void __init get_smp_config (void)
+{
+	struct intel_mp_floating *mpf = mpf_found;
+
+	/*
+ 	 * ACPI supports both logical (e.g. Hyper-Threading) and physical 
+ 	 * processors, where MPS only supports physical.
+ 	 */
+ 	if (acpi_lapic && acpi_ioapic) {
+ 		printk(KERN_INFO "Using ACPI (MADT) for SMP configuration information\n");
+ 		return;
+	}
+ 	else if (acpi_lapic)
+ 		printk(KERN_INFO "Using ACPI for processor (LAPIC) configuration information\n");
+
+	printk("Intel MultiProcessor Specification v1.%d\n", mpf->mpf_specification);
+	if (mpf->mpf_feature2 & (1<<7)) {
+		printk(KERN_INFO "    IMCR and PIC compatibility mode.\n");
+		pic_mode = 1;
+	} else {
+		printk(KERN_INFO "    Virtual Wire compatibility mode.\n");
+		pic_mode = 0;
+	}
+
+	/*
+	 * Now see if we need to read further.
+	 */
+	if (mpf->mpf_feature1 != 0) {
+
+		printk(KERN_INFO "Default MP configuration #%d\n", mpf->mpf_feature1);
+		construct_default_ISA_mptable(mpf->mpf_feature1);
+
+	} else if (mpf->mpf_physptr) {
+
+		/*
+		 * Read the physical hardware table.  Anything here will
+		 * override the defaults.
+		 */
+ 		if (!smp_read_mpc(isa_bus_to_virt(mpf->mpf_physptr))) {
+			smp_found_config = 0;
+			printk(KERN_ERR "BIOS bug, MP table errors detected!...\n");
+			printk(KERN_ERR "... disabling SMP support. (tell your hw vendor)\n");
+			return;
+		}
+		/*
+		 * If there are no explicit MP IRQ entries, then we are
+		 * broken.  We set up most of the low 16 IO-APIC pins to
+		 * ISA defaults and hope it will work.
+		 */
+		if (!mp_irq_entries) {
+			struct mpc_config_bus bus;
+
+			printk(KERN_ERR "BIOS bug, no explicit IRQ entries, using default mptable. (tell your hw vendor)\n");
+
+			bus.mpc_type = MP_BUS;
+			bus.mpc_busid = 0;
+			memcpy(bus.mpc_bustype, "ISA   ", 6);
+			MP_bus_info(&bus);
+
+			construct_default_ioirq_mptable(0);
+		}
+
+	} else
+		BUG();
+
+	printk(KERN_INFO "Processors: %d\n", num_processors);
+	/*
+	 * Only use the first configuration found.
+	 */
+}
+
+static int __init smp_scan_config (unsigned long base, unsigned long length)
+{
+	extern void __bad_mpf_size(void); 
+	unsigned int *bp = isa_bus_to_virt(base);
+	struct intel_mp_floating *mpf;
+
+	Dprintk("Scan SMP from %p for %ld bytes.\n", bp,length);
+	if (sizeof(*mpf) != 16)
+		__bad_mpf_size();
+
+	while (length > 0) {
+		mpf = (struct intel_mp_floating *)bp;
+		if ((*bp == SMP_MAGIC_IDENT) &&
+			(mpf->mpf_length == 1) &&
+			!mpf_checksum((unsigned char *)bp, 16) &&
+			((mpf->mpf_specification == 1)
+				|| (mpf->mpf_specification == 4)) ) {
+
+			smp_found_config = 1;
+			mpf_found = mpf;
+			return 1;
+		}
+		bp += 4;
+		length -= 16;
+	}
+	return 0;
+}
+
+void __init find_intel_smp (void)
+{
+	unsigned int address;
+
+	/*
+	 * FIXME: Linux assumes you have 640K of base ram..
+	 * this continues the error...
+	 *
+	 * 1) Scan the bottom 1K for a signature
+	 * 2) Scan the top 1K of base RAM
+	 * 3) Scan the 64K of bios
+	 */
+	if (smp_scan_config(0x0,0x400) ||
+		smp_scan_config(639*0x400,0x400) ||
+			smp_scan_config(0xF0000,0x10000))
+		return;
+	/*
+	 * If it is an SMP machine we should know now, unless the
+	 * configuration is in an EISA/MCA bus machine with an
+	 * extended bios data area.
+	 *
+	 * there is a real-mode segmented pointer pointing to the
+	 * 4K EBDA area at 0x40E, calculate and scan it here.
+	 *
+	 * NOTE! There are Linux loaders that will corrupt the EBDA
+	 * area, and as such this kind of SMP config may be less
+	 * trustworthy, simply because the SMP table may have been
+	 * stomped on during early boot. These loaders are buggy and
+	 * should be fixed.
+	 */
+
+	address = *(unsigned short *)phys_to_virt(0x40E);
+	address <<= 4;
+	if (smp_scan_config(address, 0x1000))
+		return;
+
+	/* If we have come this far, we did not find an MP table  */
+	 printk(KERN_INFO "No mptable found.\n");
+}
+
+/*
+ * - Intel MP Configuration Table
+ */
+void __init find_smp_config (void)
+{
+#ifdef CONFIG_X86_LOCAL_APIC
+	find_intel_smp();
+#endif
+}
+
+
+/* --------------------------------------------------------------------------
+                            ACPI-based MP Configuration
+   -------------------------------------------------------------------------- */
+
+#ifdef CONFIG_ACPI
+
+void __init mp_register_lapic_address (
+	u64			address)
+{
+#ifndef CONFIG_XEN
+	mp_lapic_addr = (unsigned long) address;
+
+	set_fixmap_nocache(FIX_APIC_BASE, mp_lapic_addr);
+
+	if (boot_cpu_id == -1U)
+		boot_cpu_id = GET_APIC_ID(apic_read(APIC_ID));
+
+	Dprintk("Boot CPU = %d\n", boot_cpu_physical_apicid);
+#endif
+}
+
+
+void __cpuinit mp_register_lapic (
+	u8			id, 
+	u8			enabled)
+{
+	struct mpc_config_processor processor;
+	int			boot_cpu = 0;
+	
+	if (id >= MAX_APICS) {
+		printk(KERN_WARNING "Processor #%d invalid (max %d)\n",
+			id, MAX_APICS);
+		return;
+	}
+
+	if (id == boot_cpu_physical_apicid)
+		boot_cpu = 1;
+
+#ifndef CONFIG_XEN
+	processor.mpc_type = MP_PROCESSOR;
+	processor.mpc_apicid = id;
+	processor.mpc_apicver = GET_APIC_VERSION(apic_read(APIC_LVR));
+	processor.mpc_cpuflag = (enabled ? CPU_ENABLED : 0);
+	processor.mpc_cpuflag |= (boot_cpu ? CPU_BOOTPROCESSOR : 0);
+	processor.mpc_cpufeature = (boot_cpu_data.x86 << 8) | 
+		(boot_cpu_data.x86_model << 4) | boot_cpu_data.x86_mask;
+	processor.mpc_featureflag = boot_cpu_data.x86_capability[0];
+	processor.mpc_reserved[0] = 0;
+	processor.mpc_reserved[1] = 0;
+#endif
+
+	MP_processor_info(&processor);
+}
+
+#ifdef CONFIG_X86_IO_APIC
+
+#define MP_ISA_BUS		0
+#define MP_MAX_IOAPIC_PIN	127
+
+static struct mp_ioapic_routing {
+	int			apic_id;
+	int			gsi_start;
+	int			gsi_end;
+	u32			pin_programmed[4];
+} mp_ioapic_routing[MAX_IO_APICS];
+
+
+static int mp_find_ioapic (
+	int			gsi)
+{
+	int			i = 0;
+
+	/* Find the IOAPIC that manages this GSI. */
+	for (i = 0; i < nr_ioapics; i++) {
+		if ((gsi >= mp_ioapic_routing[i].gsi_start)
+			&& (gsi <= mp_ioapic_routing[i].gsi_end))
+			return i;
+	}
+
+	printk(KERN_ERR "ERROR: Unable to locate IOAPIC for GSI %d\n", gsi);
+
+	return -1;
+}
+	
+
+void __init mp_register_ioapic (
+	u8			id, 
+	u32			address,
+	u32			gsi_base)
+{
+	int			idx = 0;
+
+	if (nr_ioapics >= MAX_IO_APICS) {
+		printk(KERN_ERR "ERROR: Max # of I/O APICs (%d) exceeded "
+			"(found %d)\n", MAX_IO_APICS, nr_ioapics);
+		panic("Recompile kernel with bigger MAX_IO_APICS!\n");
+	}
+	if (!address) {
+		printk(KERN_ERR "WARNING: Bogus (zero) I/O APIC address"
+			" found in MADT table, skipping!\n");
+		return;
+	}
+
+	idx = nr_ioapics++;
+
+	mp_ioapics[idx].mpc_type = MP_IOAPIC;
+	mp_ioapics[idx].mpc_flags = MPC_APIC_USABLE;
+	mp_ioapics[idx].mpc_apicaddr = address;
+
+#ifndef CONFIG_XEN
+	set_fixmap_nocache(FIX_IO_APIC_BASE_0 + idx, address);
+#endif
+	mp_ioapics[idx].mpc_apicid = id;
+	mp_ioapics[idx].mpc_apicver = io_apic_get_version(idx);
+	
+	/* 
+	 * Build basic IRQ lookup table to facilitate gsi->io_apic lookups
+	 * and to prevent reprogramming of IOAPIC pins (PCI IRQs).
+	 */
+	mp_ioapic_routing[idx].apic_id = mp_ioapics[idx].mpc_apicid;
+	mp_ioapic_routing[idx].gsi_start = gsi_base;
+	mp_ioapic_routing[idx].gsi_end = gsi_base + 
+		io_apic_get_redir_entries(idx);
+
+	printk(KERN_INFO "IOAPIC[%d]: apic_id %d, version %d, address 0x%x, "
+		"GSI %d-%d\n", idx, mp_ioapics[idx].mpc_apicid, 
+		mp_ioapics[idx].mpc_apicver, mp_ioapics[idx].mpc_apicaddr,
+		mp_ioapic_routing[idx].gsi_start,
+		mp_ioapic_routing[idx].gsi_end);
+
+	return;
+}
+
+
+void __init mp_override_legacy_irq (
+	u8			bus_irq,
+	u8			polarity, 
+	u8			trigger, 
+	u32			gsi)
+{
+	struct mpc_config_intsrc intsrc;
+	int			ioapic = -1;
+	int			pin = -1;
+
+	/* 
+	 * Convert 'gsi' to 'ioapic.pin'.
+	 */
+	ioapic = mp_find_ioapic(gsi);
+	if (ioapic < 0)
+		return;
+	pin = gsi - mp_ioapic_routing[ioapic].gsi_start;
+
+	/*
+	 * TBD: This check is for faulty timer entries, where the override
+	 *      erroneously sets the trigger to level, resulting in a HUGE 
+	 *      increase of timer interrupts!
+	 */
+	if ((bus_irq == 0) && (trigger == 3))
+		trigger = 1;
+
+	intsrc.mpc_type = MP_INTSRC;
+	intsrc.mpc_irqtype = mp_INT;
+	intsrc.mpc_irqflag = (trigger << 2) | polarity;
+	intsrc.mpc_srcbus = MP_ISA_BUS;
+	intsrc.mpc_srcbusirq = bus_irq;				       /* IRQ */
+	intsrc.mpc_dstapic = mp_ioapics[ioapic].mpc_apicid;	   /* APIC ID */
+	intsrc.mpc_dstirq = pin;				    /* INTIN# */
+
+	Dprintk("Int: type %d, pol %d, trig %d, bus %d, irq %d, %d-%d\n", 
+		intsrc.mpc_irqtype, intsrc.mpc_irqflag & 3, 
+		(intsrc.mpc_irqflag >> 2) & 3, intsrc.mpc_srcbus, 
+		intsrc.mpc_srcbusirq, intsrc.mpc_dstapic, intsrc.mpc_dstirq);
+
+	mp_irqs[mp_irq_entries] = intsrc;
+	if (++mp_irq_entries == MAX_IRQ_SOURCES)
+		panic("Max # of irq sources exceeded!\n");
+
+	return;
+}
+
+
+void __init mp_config_acpi_legacy_irqs (void)
+{
+	struct mpc_config_intsrc intsrc;
+	int			i = 0;
+	int			ioapic = -1;
+
+	/* 
+	 * Fabricate the legacy ISA bus (bus #31).
+	 */
+	mp_bus_id_to_type[MP_ISA_BUS] = MP_BUS_ISA;
+	Dprintk("Bus #%d is ISA\n", MP_ISA_BUS);
+
+	/* 
+	 * Locate the IOAPIC that manages the ISA IRQs (0-15). 
+	 */
+	ioapic = mp_find_ioapic(0);
+	if (ioapic < 0)
+		return;
+
+	intsrc.mpc_type = MP_INTSRC;
+	intsrc.mpc_irqflag = 0;					/* Conforming */
+	intsrc.mpc_srcbus = MP_ISA_BUS;
+	intsrc.mpc_dstapic = mp_ioapics[ioapic].mpc_apicid;
+
+	/* 
+	 * Use the default configuration for the IRQs 0-15.  Unless
+	 * overridden by (MADT) interrupt source override entries.
+	 */
+	for (i = 0; i < 16; i++) {
+		int idx;
+
+		for (idx = 0; idx < mp_irq_entries; idx++) {
+			struct mpc_config_intsrc *irq = mp_irqs + idx;
+
+			/* Do we already have a mapping for this ISA IRQ? */
+			if (irq->mpc_srcbus == MP_ISA_BUS && irq->mpc_srcbusirq == i)
+				break;
+
+			/* Do we already have a mapping for this IOAPIC pin */
+			if ((irq->mpc_dstapic == intsrc.mpc_dstapic) &&
+				(irq->mpc_dstirq == i))
+				break;
+		}
+
+		if (idx != mp_irq_entries) {
+			printk(KERN_DEBUG "ACPI: IRQ%d used by override.\n", i);
+			continue;			/* IRQ already used */
+		}
+
+		intsrc.mpc_irqtype = mp_INT;
+		intsrc.mpc_srcbusirq = i;		   /* Identity mapped */
+		intsrc.mpc_dstirq = i;
+
+		Dprintk("Int: type %d, pol %d, trig %d, bus %d, irq %d, "
+			"%d-%d\n", intsrc.mpc_irqtype, intsrc.mpc_irqflag & 3, 
+			(intsrc.mpc_irqflag >> 2) & 3, intsrc.mpc_srcbus, 
+			intsrc.mpc_srcbusirq, intsrc.mpc_dstapic, 
+			intsrc.mpc_dstirq);
+
+		mp_irqs[mp_irq_entries] = intsrc;
+		if (++mp_irq_entries == MAX_IRQ_SOURCES)
+			panic("Max # of irq sources exceeded!\n");
+	}
+
+	return;
+}
+
+#define MAX_GSI_NUM	4096
+
+int mp_register_gsi(u32 gsi, int triggering, int polarity)
+{
+	int			ioapic = -1;
+	int			ioapic_pin = 0;
+	int			idx, bit = 0;
+	static int		pci_irq = 16;
+	/*
+	 * Mapping between Global System Interrupts, which
+	 * represent all possible interrupts, to the IRQs
+	 * assigned to actual devices.
+	 */
+	static int		gsi_to_irq[MAX_GSI_NUM];
+
+	if (acpi_irq_model != ACPI_IRQ_MODEL_IOAPIC)
+		return gsi;
+
+	/* Don't set up the ACPI SCI because it's already set up */
+	if (acpi_fadt.sci_int == gsi)
+		return gsi;
+
+	ioapic = mp_find_ioapic(gsi);
+	if (ioapic < 0) {
+		printk(KERN_WARNING "No IOAPIC for GSI %u\n", gsi);
+		return gsi;
+	}
+
+	ioapic_pin = gsi - mp_ioapic_routing[ioapic].gsi_start;
+
+	/* 
+	 * Avoid pin reprogramming.  PRTs typically include entries  
+	 * with redundant pin->gsi mappings (but unique PCI devices);
+	 * we only program the IOAPIC on the first.
+	 */
+	bit = ioapic_pin % 32;
+	idx = (ioapic_pin < 32) ? 0 : (ioapic_pin / 32);
+	if (idx > 3) {
+		printk(KERN_ERR "Invalid reference to IOAPIC pin "
+			"%d-%d\n", mp_ioapic_routing[ioapic].apic_id, 
+			ioapic_pin);
+		return gsi;
+	}
+	if ((1<<bit) & mp_ioapic_routing[ioapic].pin_programmed[idx]) {
+		Dprintk(KERN_DEBUG "Pin %d-%d already programmed\n",
+			mp_ioapic_routing[ioapic].apic_id, ioapic_pin);
+		return gsi_to_irq[gsi];
+	}
+
+	mp_ioapic_routing[ioapic].pin_programmed[idx] |= (1<<bit);
+
+	if (triggering == ACPI_LEVEL_SENSITIVE) {
+		/*
+		 * For PCI devices assign IRQs in order, avoiding gaps
+		 * due to unused I/O APIC pins.
+		 */
+		int irq = gsi;
+		if (gsi < MAX_GSI_NUM) {
+			/*
+			 * Retain the VIA chipset work-around (gsi > 15), but
+			 * avoid a problem where the 8254 timer (IRQ0) is setup
+			 * via an override (so it's not on pin 0 of the ioapic),
+			 * and at the same time, the pin 0 interrupt is a PCI
+			 * type.  The gsi > 15 test could cause these two pins
+			 * to be shared as IRQ0, and they are not shareable.
+			 * So test for this condition, and if necessary, avoid
+			 * the pin collision.
+			 */
+			if (gsi > 15 || (gsi == 0 && !timer_uses_ioapic_pin_0))
+				gsi = pci_irq++;
+			/*
+			 * Don't assign IRQ used by ACPI SCI
+			 */
+			if (gsi == acpi_fadt.sci_int)
+				gsi = pci_irq++;
+			gsi_to_irq[irq] = gsi;
+		} else {
+			printk(KERN_ERR "GSI %u is too high\n", gsi);
+			return gsi;
+		}
+	}
+
+	io_apic_set_pci_routing(ioapic, ioapic_pin, gsi,
+		triggering == ACPI_EDGE_SENSITIVE ? 0 : 1,
+		polarity == ACPI_ACTIVE_HIGH ? 0 : 1);
+	return gsi;
+}
+
+#endif /*CONFIG_X86_IO_APIC*/
+#endif /*CONFIG_ACPI*/
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/pci-swiotlb-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/pci-swiotlb-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/pci-swiotlb-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/pci-swiotlb-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,55 @@
+/* Glue code to lib/swiotlb.c */
+
+#include <linux/pci.h>
+#include <linux/cache.h>
+#include <linux/module.h>
+#include <asm/dma-mapping.h>
+#include <asm/proto.h>
+#include <asm/swiotlb.h>
+#include <asm/dma.h>
+
+#if 0
+int swiotlb __read_mostly;
+EXPORT_SYMBOL(swiotlb);
+#endif
+
+struct dma_mapping_ops swiotlb_dma_ops = {
+#if 0
+	.mapping_error = swiotlb_dma_mapping_error,
+	.alloc_coherent = swiotlb_alloc_coherent,
+	.free_coherent = swiotlb_free_coherent,
+	.map_single = swiotlb_map_single,
+	.unmap_single = swiotlb_unmap_single,
+	.sync_single_for_cpu = swiotlb_sync_single_for_cpu,
+	.sync_single_for_device = swiotlb_sync_single_for_device,
+	.sync_single_range_for_cpu = swiotlb_sync_single_range_for_cpu,
+	.sync_single_range_for_device = swiotlb_sync_single_range_for_device,
+	.sync_sg_for_cpu = swiotlb_sync_sg_for_cpu,
+	.sync_sg_for_device = swiotlb_sync_sg_for_device,
+	.map_sg = swiotlb_map_sg,
+	.unmap_sg = swiotlb_unmap_sg,
+	.dma_supported = NULL,
+#endif
+};
+
+void pci_swiotlb_init(void)
+{
+#if 0
+	/* don't initialize swiotlb if iommu=off (no_iommu=1) */
+	if (!iommu_detected && !no_iommu && end_pfn > MAX_DMA32_PFN)
+	       swiotlb = 1;
+	if (swiotlb_force)
+		swiotlb = 1;
+	if (swiotlb) {
+		printk(KERN_INFO "PCI-DMA: Using software bounce buffering for IO (SWIOTLB)\n");
+		swiotlb_init();
+		dma_ops = &swiotlb_dma_ops;
+	}
+#else
+	swiotlb_init();
+	if (swiotlb) {
+		printk(KERN_INFO "PCI-DMA: Using software bounce buffering for IO (SWIOTLB)\n");
+		dma_ops = &swiotlb_dma_ops;
+	}
+#endif
+}
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/process-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/process-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/process-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/process-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,848 @@
+/*
+ *  linux/arch/x86-64/kernel/process.c
+ *
+ *  Copyright (C) 1995  Linus Torvalds
+ *
+ *  Pentium III FXSR, SSE support
+ *	Gareth Hughes <gareth@valinux.com>, May 2000
+ * 
+ *  X86-64 port
+ *	Andi Kleen.
+ *
+ *	CPU hotplug support - ashok.raj@intel.com
+ * 
+ *  Jun Nakajima <jun.nakajima@intel.com> 
+ *     Modified for Xen
+ */
+
+/*
+ * This file handles the architecture-dependent parts of process handling..
+ */
+
+#include <stdarg.h>
+
+#include <linux/cpu.h>
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/elfcore.h>
+#include <linux/smp.h>
+#include <linux/slab.h>
+#include <linux/user.h>
+#include <linux/module.h>
+#include <linux/a.out.h>
+#include <linux/interrupt.h>
+#include <linux/delay.h>
+#include <linux/ptrace.h>
+#include <linux/utsname.h>
+#include <linux/random.h>
+#include <linux/notifier.h>
+#include <linux/kprobes.h>
+
+#include <asm/uaccess.h>
+#include <asm/pgtable.h>
+#include <asm/system.h>
+#include <asm/io.h>
+#include <asm/processor.h>
+#include <asm/i387.h>
+#include <asm/mmu_context.h>
+#include <asm/pda.h>
+#include <asm/prctl.h>
+#include <asm/kdebug.h>
+#include <xen/interface/platform.h>
+#include <xen/interface/physdev.h>
+#include <xen/interface/vcpu.h>
+#include <asm/desc.h>
+#include <asm/proto.h>
+#include <asm/hardirq.h>
+#include <asm/ia32.h>
+#include <asm/idle.h>
+
+#include <xen/cpu_hotplug.h>
+
+asmlinkage extern void ret_from_fork(void);
+
+unsigned long kernel_thread_flags = CLONE_VM | CLONE_UNTRACED;
+
+unsigned long boot_option_idle_override = 0;
+EXPORT_SYMBOL(boot_option_idle_override);
+
+/*
+ * Powermanagement idle function, if any..
+ */
+void (*pm_idle)(void);
+EXPORT_SYMBOL(pm_idle);
+static DEFINE_PER_CPU(unsigned int, cpu_idle_state);
+
+static ATOMIC_NOTIFIER_HEAD(idle_notifier);
+
+void idle_notifier_register(struct notifier_block *n)
+{
+	atomic_notifier_chain_register(&idle_notifier, n);
+}
+EXPORT_SYMBOL_GPL(idle_notifier_register);
+
+void idle_notifier_unregister(struct notifier_block *n)
+{
+	atomic_notifier_chain_unregister(&idle_notifier, n);
+}
+EXPORT_SYMBOL(idle_notifier_unregister);
+
+enum idle_state { CPU_IDLE, CPU_NOT_IDLE };
+static DEFINE_PER_CPU(enum idle_state, idle_state) = CPU_NOT_IDLE;
+
+void enter_idle(void)
+{
+	__get_cpu_var(idle_state) = CPU_IDLE;
+	atomic_notifier_call_chain(&idle_notifier, IDLE_START, NULL);
+}
+
+static void __exit_idle(void)
+{
+	__get_cpu_var(idle_state) = CPU_NOT_IDLE;
+	atomic_notifier_call_chain(&idle_notifier, IDLE_END, NULL);
+}
+
+/* Called from interrupts to signify idle end */
+void exit_idle(void)
+{
+	if (current->pid | read_pda(irqcount))
+		return;
+	__exit_idle();
+}
+
+/*
+ * On SMP it's slightly faster (but much more power-consuming!)
+ * to poll the ->need_resched flag instead of waiting for the
+ * cross-CPU IPI to arrive. Use this option with caution.
+ */
+static void poll_idle (void)
+{
+	local_irq_enable();
+
+	asm volatile(
+		"2:"
+		"testl %0,%1;"
+		"rep; nop;"
+		"je 2b;"
+		: :
+		"i" (_TIF_NEED_RESCHED),
+		"m" (current_thread_info()->flags));
+}
+
+static void xen_idle(void)
+{
+	local_irq_disable();
+
+	if (need_resched())
+		local_irq_enable();
+	else {
+		current_thread_info()->status &= ~TS_POLLING;
+		smp_mb__after_clear_bit();
+		safe_halt();
+		current_thread_info()->status |= TS_POLLING;
+	}
+}
+
+#ifdef CONFIG_HOTPLUG_CPU
+static inline void play_dead(void)
+{
+	idle_task_exit();
+	local_irq_disable();
+	cpu_clear(smp_processor_id(), cpu_initialized);
+	preempt_enable_no_resched();
+	VOID(HYPERVISOR_vcpu_op(VCPUOP_down, smp_processor_id(), NULL));
+	cpu_bringup();
+}
+#else
+static inline void play_dead(void)
+{
+	BUG();
+}
+#endif /* CONFIG_HOTPLUG_CPU */
+
+/*
+ * The idle thread. There's no useful work to be
+ * done, so just try to conserve power and have a
+ * low exit latency (ie sit in a loop waiting for
+ * somebody to say that they'd like to reschedule)
+ */
+void cpu_idle (void)
+{
+	current_thread_info()->status |= TS_POLLING;
+	/* endless idle loop with no priority at all */
+	while (1) {
+		while (!need_resched()) {
+			void (*idle)(void);
+
+			if (__get_cpu_var(cpu_idle_state))
+				__get_cpu_var(cpu_idle_state) = 0;
+			rmb();
+			idle = xen_idle; /* no alternatives */
+			if (cpu_is_offline(smp_processor_id()))
+				play_dead();
+			enter_idle();
+			idle();
+			__exit_idle();
+		}
+
+		preempt_enable_no_resched();
+		schedule();
+		preempt_disable();
+	}
+}
+
+void cpu_idle_wait(void)
+{
+	unsigned int cpu, this_cpu = get_cpu();
+	cpumask_t map;
+
+	set_cpus_allowed(current, cpumask_of_cpu(this_cpu));
+	put_cpu();
+
+ 	cpus_clear(map);
+	for_each_online_cpu(cpu) {
+		per_cpu(cpu_idle_state, cpu) = 1;
+		cpu_set(cpu, map);
+	}
+
+	__get_cpu_var(cpu_idle_state) = 0;
+
+	wmb();
+	do {
+		ssleep(1);
+		for_each_online_cpu(cpu) {
+			if (cpu_isset(cpu, map) &&
+					!per_cpu(cpu_idle_state, cpu))
+				cpu_clear(cpu, map);
+		}
+		cpus_and(map, map, cpu_online_map);
+	} while (!cpus_empty(map));
+}
+EXPORT_SYMBOL_GPL(cpu_idle_wait);
+
+void __cpuinit select_idle_routine(const struct cpuinfo_x86 *c) 
+{
+}
+
+static int __init idle_setup (char *str)
+{
+	if (!strncmp(str, "poll", 4)) {
+		printk("using polling idle threads.\n");
+		pm_idle = poll_idle;
+	}
+
+	boot_option_idle_override = 1;
+	return 1;
+}
+
+__setup("idle=", idle_setup);
+
+/* Prints also some state that isn't saved in the pt_regs */ 
+void __show_regs(struct pt_regs * regs)
+{
+	unsigned long fs, gs, shadowgs;
+	unsigned int fsindex,gsindex;
+	unsigned int ds,cs,es; 
+
+	printk("\n");
+	print_modules();
+	printk("Pid: %d, comm: %.20s %s %s %.*s\n",
+		current->pid, current->comm, print_tainted(),
+		system_utsname.release,
+		(int)strcspn(system_utsname.version, " "),
+		system_utsname.version);
+	printk("RIP: %04lx:[<%016lx>] ", regs->cs & 0xffff, regs->rip);
+	printk_address(regs->rip); 
+	printk("RSP: %04lx:%016lx  EFLAGS: %08lx\n", regs->ss, regs->rsp,
+		regs->eflags);
+	printk("RAX: %016lx RBX: %016lx RCX: %016lx\n",
+	       regs->rax, regs->rbx, regs->rcx);
+	printk("RDX: %016lx RSI: %016lx RDI: %016lx\n",
+	       regs->rdx, regs->rsi, regs->rdi); 
+	printk("RBP: %016lx R08: %016lx R09: %016lx\n",
+	       regs->rbp, regs->r8, regs->r9); 
+	printk("R10: %016lx R11: %016lx R12: %016lx\n",
+	       regs->r10, regs->r11, regs->r12); 
+	printk("R13: %016lx R14: %016lx R15: %016lx\n",
+	       regs->r13, regs->r14, regs->r15); 
+
+	asm("mov %%ds,%0" : "=r" (ds)); 
+	asm("mov %%cs,%0" : "=r" (cs)); 
+	asm("mov %%es,%0" : "=r" (es)); 
+	asm("mov %%fs,%0" : "=r" (fsindex));
+	asm("mov %%gs,%0" : "=r" (gsindex));
+
+	rdmsrl(MSR_FS_BASE, fs);
+	rdmsrl(MSR_GS_BASE, gs); 
+	rdmsrl(MSR_KERNEL_GS_BASE, shadowgs); 
+
+	printk("FS:  %016lx(%04x) GS:%016lx(%04x) knlGS:%016lx\n", 
+	       fs,fsindex,gs,gsindex,shadowgs); 
+	printk("CS:  %04x DS: %04x ES: %04x\n", cs, ds, es); 
+
+}
+
+void show_regs(struct pt_regs *regs)
+{
+	printk("CPU %d:", smp_processor_id());
+	__show_regs(regs);
+	show_trace(NULL, regs, (void *)(regs + 1));
+}
+
+/*
+ * Free current thread data structures etc..
+ */
+void exit_thread(void)
+{
+	struct task_struct *me = current;
+	struct thread_struct *t = &me->thread;
+
+	if (me->thread.io_bitmap_ptr) { 
+#ifndef CONFIG_X86_NO_TSS
+		struct tss_struct *tss = &per_cpu(init_tss, get_cpu());
+#endif
+#ifdef CONFIG_XEN
+		struct physdev_set_iobitmap iobmp_op;
+		memset(&iobmp_op, 0, sizeof(iobmp_op));
+#endif
+
+		kfree(t->io_bitmap_ptr);
+		t->io_bitmap_ptr = NULL;
+		/*
+		 * Careful, clear this in the TSS too:
+		 */
+#ifndef CONFIG_X86_NO_TSS
+		memset(tss->io_bitmap, 0xff, t->io_bitmap_max);
+		put_cpu();
+#endif
+#ifdef CONFIG_XEN
+		WARN_ON(HYPERVISOR_physdev_op(PHYSDEVOP_set_iobitmap,
+					      &iobmp_op));
+#endif
+		t->io_bitmap_max = 0;
+	}
+}
+
+void load_gs_index(unsigned gs)
+{
+	WARN_ON(HYPERVISOR_set_segment_base(SEGBASE_GS_USER_SEL, gs));
+}
+
+void flush_thread(void)
+{
+	struct task_struct *tsk = current;
+	struct thread_info *t = current_thread_info();
+
+	if (t->flags & _TIF_ABI_PENDING) {
+		t->flags ^= (_TIF_ABI_PENDING | _TIF_IA32);
+		if (t->flags & _TIF_IA32)
+			current_thread_info()->status |= TS_COMPAT;
+	}
+
+	tsk->thread.debugreg0 = 0;
+	tsk->thread.debugreg1 = 0;
+	tsk->thread.debugreg2 = 0;
+	tsk->thread.debugreg3 = 0;
+	tsk->thread.debugreg6 = 0;
+	tsk->thread.debugreg7 = 0;
+	memset(tsk->thread.tls_array, 0, sizeof(tsk->thread.tls_array));	
+	/*
+	 * Forget coprocessor state..
+	 */
+	clear_fpu(tsk);
+	clear_used_math();
+}
+
+void release_thread(struct task_struct *dead_task)
+{
+	if (dead_task->mm) {
+		if (dead_task->mm->context.size) {
+			printk("WARNING: dead process %8s still has LDT? <%p/%d>\n",
+					dead_task->comm,
+					dead_task->mm->context.ldt,
+					dead_task->mm->context.size);
+			BUG();
+		}
+	}
+}
+
+static inline void set_32bit_tls(struct task_struct *t, int tls, u32 addr)
+{
+	struct user_desc ud = { 
+		.base_addr = addr,
+		.limit = 0xfffff,
+		.seg_32bit = 1,
+		.limit_in_pages = 1,
+		.useable = 1,
+	};
+	struct n_desc_struct *desc = (void *)t->thread.tls_array;
+	desc += tls;
+	desc->a = LDT_entry_a(&ud); 
+	desc->b = LDT_entry_b(&ud); 
+}
+
+static inline u32 read_32bit_tls(struct task_struct *t, int tls)
+{
+	struct desc_struct *desc = (void *)t->thread.tls_array;
+	desc += tls;
+	return desc->base0 | 
+		(((u32)desc->base1) << 16) | 
+		(((u32)desc->base2) << 24);
+}
+
+/*
+ * This gets called before we allocate a new thread and copy
+ * the current task into it.
+ */
+void prepare_to_copy(struct task_struct *tsk)
+{
+	unlazy_fpu(tsk);
+}
+
+int copy_thread(int nr, unsigned long clone_flags, unsigned long rsp, 
+		unsigned long unused,
+	struct task_struct * p, struct pt_regs * regs)
+{
+	int err;
+	struct pt_regs * childregs;
+	struct task_struct *me = current;
+
+	childregs = ((struct pt_regs *)
+			(THREAD_SIZE + task_stack_page(p))) - 1;
+	*childregs = *regs;
+
+	childregs->rax = 0;
+	childregs->rsp = rsp;
+	if (rsp == ~0UL)
+		childregs->rsp = (unsigned long)childregs;
+
+	p->thread.rsp = (unsigned long) childregs;
+	p->thread.rsp0 = (unsigned long) (childregs+1);
+	p->thread.userrsp = me->thread.userrsp; 
+
+	set_tsk_thread_flag(p, TIF_FORK);
+
+	p->thread.fs = me->thread.fs;
+	p->thread.gs = me->thread.gs;
+
+	asm("mov %%gs,%0" : "=m" (p->thread.gsindex));
+	asm("mov %%fs,%0" : "=m" (p->thread.fsindex));
+	asm("mov %%es,%0" : "=m" (p->thread.es));
+	asm("mov %%ds,%0" : "=m" (p->thread.ds));
+
+	if (unlikely(me->thread.io_bitmap_ptr != NULL)) { 
+		p->thread.io_bitmap_ptr = kmalloc(IO_BITMAP_BYTES, GFP_KERNEL);
+		if (!p->thread.io_bitmap_ptr) {
+			p->thread.io_bitmap_max = 0;
+			return -ENOMEM;
+		}
+		memcpy(p->thread.io_bitmap_ptr, me->thread.io_bitmap_ptr,
+				IO_BITMAP_BYTES);
+	} 
+
+	/*
+	 * Set a new TLS for the child thread?
+	 */
+	if (clone_flags & CLONE_SETTLS) {
+#ifdef CONFIG_IA32_EMULATION
+		if (test_thread_flag(TIF_IA32))
+			err = ia32_child_tls(p, childregs); 
+		else 			
+#endif	 
+			err = do_arch_prctl(p, ARCH_SET_FS, childregs->r8); 
+		if (err) 
+			goto out;
+	}
+        p->thread.iopl = current->thread.iopl;
+
+	err = 0;
+out:
+	if (err && p->thread.io_bitmap_ptr) {
+		kfree(p->thread.io_bitmap_ptr);
+		p->thread.io_bitmap_max = 0;
+	}
+	return err;
+}
+
+static inline void __save_init_fpu( struct task_struct *tsk )
+{
+	asm volatile( "rex64 ; fxsave %0 ; fnclex"
+		      : "=m" (tsk->thread.i387.fxsave));
+	tsk->thread_info->status &= ~TS_USEDFPU;
+}
+
+/*
+ *	switch_to(x,y) should switch tasks from x to y.
+ *
+ * This could still be optimized: 
+ * - fold all the options into a flag word and test it with a single test.
+ * - could test fs/gs bitsliced
+ *
+ * Kprobes not supported here. Set the probe on schedule instead.
+ */
+__kprobes struct task_struct *
+__switch_to(struct task_struct *prev_p, struct task_struct *next_p)
+{
+	struct thread_struct *prev = &prev_p->thread,
+				 *next = &next_p->thread;
+	int cpu = smp_processor_id();  
+#ifndef CONFIG_X86_NO_TSS
+	struct tss_struct *tss = &per_cpu(init_tss, cpu);
+#endif
+#if CONFIG_XEN_COMPAT > 0x030002
+	struct physdev_set_iopl iopl_op;
+	struct physdev_set_iobitmap iobmp_op;
+#else
+	struct physdev_op _pdo[2], *pdo = _pdo;
+#define iopl_op pdo->u.set_iopl
+#define iobmp_op pdo->u.set_iobitmap
+#endif
+	multicall_entry_t _mcl[8], *mcl = _mcl;
+
+	/*
+	 * This is basically '__unlazy_fpu', except that we queue a
+	 * multicall to indicate FPU task switch, rather than
+	 * synchronously trapping to Xen.
+	 * The AMD workaround requires it to be after DS reload, or
+	 * after DS has been cleared, which we do in __prepare_arch_switch.
+	 */
+	if (prev_p->thread_info->status & TS_USEDFPU) {
+		__save_init_fpu(prev_p); /* _not_ save_init_fpu() */
+		mcl->op      = __HYPERVISOR_fpu_taskswitch;
+		mcl->args[0] = 1;
+		mcl++;
+	}
+
+	/*
+	 * Reload esp0, LDT and the page table pointer:
+	 */
+	mcl->op      = __HYPERVISOR_stack_switch;
+	mcl->args[0] = __KERNEL_DS;
+	mcl->args[1] = next->rsp0;
+	mcl++;
+
+	/*
+	 * Load the per-thread Thread-Local Storage descriptor.
+	 * This is load_TLS(next, cpu) with multicalls.
+	 */
+#define C(i) do {							\
+	if (unlikely(next->tls_array[i] != prev->tls_array[i])) {	\
+		mcl->op      = __HYPERVISOR_update_descriptor;		\
+		mcl->args[0] = virt_to_machine(				\
+			&cpu_gdt(cpu)[GDT_ENTRY_TLS_MIN + i]);		\
+		mcl->args[1] = next->tls_array[i];			\
+		mcl++;							\
+	}								\
+} while (0)
+	C(0); C(1); C(2);
+#undef C
+
+	if (unlikely(prev->iopl != next->iopl)) {
+		iopl_op.iopl = (next->iopl == 0) ? 1 : next->iopl;
+#if CONFIG_XEN_COMPAT > 0x030002
+		mcl->op      = __HYPERVISOR_physdev_op;
+		mcl->args[0] = PHYSDEVOP_set_iopl;
+		mcl->args[1] = (unsigned long)&iopl_op;
+#else
+		mcl->op      = __HYPERVISOR_physdev_op_compat;
+		pdo->cmd     = PHYSDEVOP_set_iopl;
+		mcl->args[0] = (unsigned long)pdo++;
+#endif
+		mcl++;
+	}
+
+	if (unlikely(prev->io_bitmap_ptr || next->io_bitmap_ptr)) {
+		set_xen_guest_handle(iobmp_op.bitmap,
+				     (char *)next->io_bitmap_ptr);
+		iobmp_op.nr_ports = next->io_bitmap_ptr ? IO_BITMAP_BITS : 0;
+#if CONFIG_XEN_COMPAT > 0x030002
+		mcl->op      = __HYPERVISOR_physdev_op;
+		mcl->args[0] = PHYSDEVOP_set_iobitmap;
+		mcl->args[1] = (unsigned long)&iobmp_op;
+#else
+		mcl->op      = __HYPERVISOR_physdev_op_compat;
+		pdo->cmd     = PHYSDEVOP_set_iobitmap;
+		mcl->args[0] = (unsigned long)pdo++;
+#endif
+		mcl++;
+	}
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+	BUG_ON(pdo > _pdo + ARRAY_SIZE(_pdo));
+#endif
+	BUG_ON(mcl > _mcl + ARRAY_SIZE(_mcl));
+	if (unlikely(HYPERVISOR_multicall_check(_mcl, mcl - _mcl, NULL)))
+		BUG();
+
+	/* 
+	 * Switch DS and ES.
+	 * This won't pick up thread selector changes, but I guess that is ok.
+	 */
+	if (unlikely(next->es))
+		loadsegment(es, next->es); 
+	
+	if (unlikely(next->ds))
+		loadsegment(ds, next->ds);
+
+	/* 
+	 * Switch FS and GS.
+	 */
+	if (unlikely(next->fsindex))
+		loadsegment(fs, next->fsindex);
+
+	if (next->fs)
+		WARN_ON(HYPERVISOR_set_segment_base(SEGBASE_FS, next->fs));
+	
+	if (unlikely(next->gsindex))
+		load_gs_index(next->gsindex);
+
+	if (next->gs)
+		WARN_ON(HYPERVISOR_set_segment_base(SEGBASE_GS_USER, next->gs));
+
+	/* 
+	 * Switch the PDA context.
+	 */
+	prev->userrsp = read_pda(oldrsp); 
+	write_pda(oldrsp, next->userrsp); 
+	write_pda(pcurrent, next_p); 
+	write_pda(kernelstack,
+		  task_stack_page(next_p) + THREAD_SIZE - PDA_STACKOFFSET);
+
+	/*
+	 * Now maybe reload the debug registers
+	 */
+	if (unlikely(next->debugreg7)) {
+		set_debugreg(next->debugreg0, 0);
+		set_debugreg(next->debugreg1, 1);
+		set_debugreg(next->debugreg2, 2);
+		set_debugreg(next->debugreg3, 3);
+		/* no 4 and 5 */
+		set_debugreg(next->debugreg6, 6);
+		set_debugreg(next->debugreg7, 7);
+	}
+
+	return prev_p;
+}
+
+/*
+ * sys_execve() executes a new program.
+ */
+asmlinkage 
+long sys_execve(char __user *name, char __user * __user *argv,
+		char __user * __user *envp, struct pt_regs regs)
+{
+	long error;
+	char * filename;
+
+	filename = getname(name);
+	error = PTR_ERR(filename);
+	if (IS_ERR(filename)) 
+		return error;
+	error = do_execve(filename, argv, envp, &regs); 
+	if (error == 0) {
+		task_lock(current);
+		current->ptrace &= ~PT_DTRACE;
+		task_unlock(current);
+	}
+	putname(filename);
+	return error;
+}
+
+void set_personality_64bit(void)
+{
+	/* inherit personality from parent */
+
+	/* Make sure to be in 64bit mode */
+	clear_thread_flag(TIF_IA32); 
+
+	/* TBD: overwrites user setup. Should have two bits.
+	   But 64bit processes have always behaved this way,
+	   so it's not too bad. The main problem is just that
+   	   32bit childs are affected again. */
+	current->personality &= ~READ_IMPLIES_EXEC;
+}
+
+asmlinkage long sys_fork(struct pt_regs *regs)
+{
+	return do_fork(SIGCHLD, regs->rsp, regs, 0, NULL, NULL);
+}
+
+asmlinkage long
+sys_clone(unsigned long clone_flags, unsigned long newsp,
+	  void __user *parent_tid, void __user *child_tid, struct pt_regs *regs)
+{
+	if (!newsp)
+		newsp = regs->rsp;
+	return do_fork(clone_flags, newsp, regs, 0, parent_tid, child_tid);
+}
+
+/*
+ * This is trivial, and on the face of it looks like it
+ * could equally well be done in user mode.
+ *
+ * Not so, for quite unobvious reasons - register pressure.
+ * In user mode vfork() cannot have a stack frame, and if
+ * done by calling the "clone()" system call directly, you
+ * do not have enough call-clobbered registers to hold all
+ * the information you need.
+ */
+asmlinkage long sys_vfork(struct pt_regs *regs)
+{
+	return do_fork(CLONE_VFORK | CLONE_VM | SIGCHLD, regs->rsp, regs, 0,
+		    NULL, NULL);
+}
+
+unsigned long get_wchan(struct task_struct *p)
+{
+	unsigned long stack;
+	u64 fp,rip;
+	int count = 0;
+
+	if (!p || p == current || p->state==TASK_RUNNING)
+		return 0; 
+	stack = (unsigned long)task_stack_page(p);
+	if (p->thread.rsp < stack || p->thread.rsp > stack+THREAD_SIZE)
+		return 0;
+	fp = *(u64 *)(p->thread.rsp);
+	do { 
+		if (fp < (unsigned long)stack ||
+		    fp > (unsigned long)stack+THREAD_SIZE)
+			return 0; 
+		rip = *(u64 *)(fp+8); 
+		if (!in_sched_functions(rip))
+			return rip; 
+		fp = *(u64 *)fp; 
+	} while (count++ < 16); 
+	return 0;
+}
+
+long do_arch_prctl(struct task_struct *task, int code, unsigned long addr)
+{ 
+	int ret = 0; 
+	int doit = task == current;
+	int cpu;
+
+	switch (code) { 
+	case ARCH_SET_GS:
+		if (addr >= TASK_SIZE_OF(task))
+			return -EPERM; 
+		cpu = get_cpu();
+		/* handle small bases via the GDT because that's faster to 
+		   switch. */
+		if (addr <= 0xffffffff) {  
+			set_32bit_tls(task, GS_TLS, addr); 
+			if (doit) { 
+				load_TLS(&task->thread, cpu);
+				load_gs_index(GS_TLS_SEL); 
+			}
+			task->thread.gsindex = GS_TLS_SEL; 
+			task->thread.gs = 0;
+		} else { 
+			task->thread.gsindex = 0;
+			task->thread.gs = addr;
+			if (doit) {
+				load_gs_index(0);
+				ret = HYPERVISOR_set_segment_base(
+					SEGBASE_GS_USER, addr);
+			} 
+		}
+		put_cpu();
+		break;
+	case ARCH_SET_FS:
+		/* Not strictly needed for fs, but do it for symmetry
+		   with gs */
+		if (addr >= TASK_SIZE_OF(task))
+			return -EPERM; 
+		cpu = get_cpu();
+		/* handle small bases via the GDT because that's faster to 
+		   switch. */
+		if (addr <= 0xffffffff) { 
+			set_32bit_tls(task, FS_TLS, addr);
+			if (doit) { 
+				load_TLS(&task->thread, cpu); 
+				asm volatile("movl %0,%%fs" :: "r"(FS_TLS_SEL));
+			}
+			task->thread.fsindex = FS_TLS_SEL;
+			task->thread.fs = 0;
+		} else { 
+			task->thread.fsindex = 0;
+			task->thread.fs = addr;
+			if (doit) {
+				/* set the selector to 0 to not confuse
+				   __switch_to */
+				asm volatile("movl %0,%%fs" :: "r" (0));
+                                ret = HYPERVISOR_set_segment_base(SEGBASE_FS,
+								  addr);
+			}
+		}
+		put_cpu();
+		break;
+	case ARCH_GET_FS: { 
+		unsigned long base; 
+		if (task->thread.fsindex == FS_TLS_SEL)
+			base = read_32bit_tls(task, FS_TLS);
+		else if (doit)
+			rdmsrl(MSR_FS_BASE, base);
+		else
+			base = task->thread.fs;
+		ret = put_user(base, (unsigned long __user *)addr); 
+		break; 
+	}
+	case ARCH_GET_GS: { 
+		unsigned long base;
+		unsigned gsindex;
+		if (task->thread.gsindex == GS_TLS_SEL)
+			base = read_32bit_tls(task, GS_TLS);
+		else if (doit) {
+ 			asm("movl %%gs,%0" : "=r" (gsindex));
+			if (gsindex)
+				rdmsrl(MSR_KERNEL_GS_BASE, base);
+			else
+				base = task->thread.gs;
+		}
+		else
+			base = task->thread.gs;
+		ret = put_user(base, (unsigned long __user *)addr); 
+		break;
+	}
+
+	default:
+		ret = -EINVAL;
+		break;
+	} 
+
+	return ret;	
+} 
+
+long sys_arch_prctl(int code, unsigned long addr)
+{
+	return do_arch_prctl(current, code, addr);
+} 
+
+/* 
+ * Capture the user space registers if the task is not running (in user space)
+ */
+int dump_task_regs(struct task_struct *tsk, elf_gregset_t *regs)
+{
+	struct pt_regs *pp, ptregs;
+
+	pp = task_pt_regs(tsk);
+
+	ptregs = *pp; 
+	ptregs.cs &= 0xffff;
+	ptregs.ss &= 0xffff;
+
+	elf_core_copy_regs(regs, &ptregs);
+ 
+        boot_option_idle_override = 1;
+	return 1;
+}
+
+unsigned long arch_align_stack(unsigned long sp)
+{
+	if (randomize_va_space)
+		sp -= get_random_int() % 8192;
+	return sp & ~0xf;
+}
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/relocate_kernel.S linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/relocate_kernel.S
--- linux-2.6.18.8/arch/x86_64/kernel/relocate_kernel.S	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/relocate_kernel.S	2008-08-21 11:36:07.000000000 +0200
@@ -7,31 +7,195 @@
  */
 
 #include <linux/linkage.h>
+#include <asm/page.h>
+#include <asm/kexec.h>
 
-	/*
-	 * Must be relocatable PIC code callable as a C function, that once
-	 * it starts can not use the previous processes stack.
-	 */
-	.globl relocate_new_kernel
+/*
+ * Must be relocatable PIC code callable as a C function
+ */
+
+#define PTR(x) (x << 3)
+#define PAGE_ALIGNED (1 << PAGE_SHIFT)
+#define PAGE_ATTR 0x63 /* _PAGE_PRESENT|_PAGE_RW|_PAGE_ACCESSED|_PAGE_DIRTY */
+
+	.text
+	.align PAGE_ALIGNED
 	.code64
+	.globl relocate_kernel
+relocate_kernel:
+	/* %rdi indirection_page
+	 * %rsi page_list
+	 * %rdx start address
+	 */
+
+	/* map the control page at its virtual address */
+
+	movq	$0x0000ff8000000000, %r10        /* mask */
+	mov	$(39 - 3), %cl                   /* bits to shift */
+	movq	PTR(VA_CONTROL_PAGE)(%rsi), %r11 /* address to map */
+
+	movq	%r11, %r9
+	andq	%r10, %r9
+	shrq	%cl, %r9
+
+	movq	PTR(VA_PGD)(%rsi), %r8
+	addq	%r8, %r9
+	movq	PTR(PA_PUD_0)(%rsi), %r8
+	orq	$PAGE_ATTR, %r8
+	movq	%r8, (%r9)
+
+	shrq	$9, %r10
+	sub	$9, %cl
+
+	movq	%r11, %r9
+	andq	%r10, %r9
+	shrq	%cl, %r9
+
+	movq	PTR(VA_PUD_0)(%rsi), %r8
+	addq	%r8, %r9
+	movq	PTR(PA_PMD_0)(%rsi), %r8
+	orq	$PAGE_ATTR, %r8
+	movq	%r8, (%r9)
+
+	shrq	$9, %r10
+	sub	$9, %cl
+
+	movq	%r11, %r9
+	andq	%r10, %r9
+	shrq	%cl, %r9
+
+	movq	PTR(VA_PMD_0)(%rsi), %r8
+	addq	%r8, %r9
+	movq	PTR(PA_PTE_0)(%rsi), %r8
+	orq	$PAGE_ATTR, %r8
+	movq	%r8, (%r9)
+
+	shrq	$9, %r10
+	sub	$9, %cl
+
+	movq	%r11, %r9
+	andq	%r10, %r9
+	shrq	%cl, %r9
+
+	movq	PTR(VA_PTE_0)(%rsi), %r8
+	addq	%r8, %r9
+	movq	PTR(PA_CONTROL_PAGE)(%rsi), %r8
+	orq	$PAGE_ATTR, %r8
+	movq	%r8, (%r9)
+
+	/* identity map the control page at its physical address */
+
+	movq	$0x0000ff8000000000, %r10        /* mask */
+	mov	$(39 - 3), %cl                   /* bits to shift */
+	movq	PTR(PA_CONTROL_PAGE)(%rsi), %r11 /* address to map */
+
+	movq	%r11, %r9
+	andq	%r10, %r9
+	shrq	%cl, %r9
+
+	movq	PTR(VA_PGD)(%rsi), %r8
+	addq	%r8, %r9
+	movq	PTR(PA_PUD_1)(%rsi), %r8
+	orq	$PAGE_ATTR, %r8
+	movq	%r8, (%r9)
+
+	shrq	$9, %r10
+	sub	$9, %cl
+
+	movq	%r11, %r9
+	andq	%r10, %r9
+	shrq	%cl, %r9
+
+	movq	PTR(VA_PUD_1)(%rsi), %r8
+	addq	%r8, %r9
+	movq	PTR(PA_PMD_1)(%rsi), %r8
+	orq	$PAGE_ATTR, %r8
+	movq	%r8, (%r9)
+
+	shrq	$9, %r10
+	sub	$9, %cl
+
+	movq	%r11, %r9
+	andq	%r10, %r9
+	shrq	%cl, %r9
+
+	movq	PTR(VA_PMD_1)(%rsi), %r8
+	addq	%r8, %r9
+	movq	PTR(PA_PTE_1)(%rsi), %r8
+	orq	$PAGE_ATTR, %r8
+	movq	%r8, (%r9)
+
+	shrq	$9, %r10
+	sub	$9, %cl
+
+	movq	%r11, %r9
+	andq	%r10, %r9
+	shrq	%cl, %r9
+
+	movq	PTR(VA_PTE_1)(%rsi), %r8
+	addq	%r8, %r9
+	movq	PTR(PA_CONTROL_PAGE)(%rsi), %r8
+	orq	$PAGE_ATTR, %r8
+	movq	%r8, (%r9)
+
 relocate_new_kernel:
-	/* %rdi page_list
-	 * %rsi reboot_code_buffer
+	/* %rdi indirection_page
+	 * %rsi page_list
 	 * %rdx start address
-	 * %rcx page_table
-	 * %r8  arg5
-	 * %r9  arg6
 	 */
 
 	/* zero out flags, and disable interrupts */
 	pushq $0
 	popfq
 
-	/* set a new stack at the bottom of our page... */
-	lea   4096(%rsi), %rsp
-
-	/* store the parameters back on the stack */
-	pushq	%rdx /* store the start address */
+	/* get physical address of control page now */
+	/* this is impossible after page table switch */
+	movq	PTR(PA_CONTROL_PAGE)(%rsi), %r8
+
+	/* get physical address of page table now too */
+	movq	PTR(PA_TABLE_PAGE)(%rsi), %rcx
+
+	/* switch to new set of page tables */
+	movq	PTR(PA_PGD)(%rsi), %r9
+	movq	%r9, %cr3
+
+	/* setup idt */
+	movq    %r8, %rax
+	addq    $(idt_80 - relocate_kernel), %rax
+	lidtq   (%rax)
+
+	/* setup gdt */
+	movq    %r8, %rax
+	addq    $(gdt - relocate_kernel), %rax
+	movq    %r8, %r9
+	addq    $((gdt_80 - relocate_kernel) + 2), %r9
+	movq    %rax, (%r9)
+
+	movq    %r8, %rax
+	addq    $(gdt_80 - relocate_kernel), %rax
+	lgdtq   (%rax)
+
+	/* setup data segment registers */
+	xorl	%eax, %eax
+	movl    %eax, %ds
+	movl    %eax, %es
+	movl    %eax, %fs
+	movl    %eax, %gs
+	movl    %eax, %ss
+	
+	/* setup a new stack at the end of the physical control page */
+	lea	4096(%r8), %rsp
+
+	/* load new code segment and jump to identity mapped page */
+	movq	%r8, %rax
+	addq    $(identity_mapped - relocate_kernel), %rax
+	pushq	$(gdt_cs - gdt)
+	pushq	%rax
+	lretq
+
+identity_mapped:
+	/* store the start address on the stack */
+	pushq   %rdx
 
 	/* Set cr0 to a known state:
 	 * 31 1 == Paging enabled
@@ -134,10 +298,19 @@ relocate_new_kernel:
 	xorq	%r13, %r13
 	xorq	%r14, %r14
 	xorq	%r15, %r15
-
 	ret
-relocate_new_kernel_end:
 
-	.globl relocate_new_kernel_size
-relocate_new_kernel_size:
-	.quad relocate_new_kernel_end - relocate_new_kernel
+	.align  16
+gdt:
+	.quad	0x0000000000000000	/* NULL descriptor */
+gdt_cs:
+	.quad   0x00af9a000000ffff
+gdt_end:
+
+gdt_80:
+	.word	gdt_end - gdt - 1	/* limit */
+	.quad	0			/* base - filled in by code above */
+
+idt_80:
+	.word	0			/* limit */
+	.quad	0			/* base */
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/setup64-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/setup64-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/setup64-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/setup64-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,367 @@
+/* 
+ * X86-64 specific CPU setup.
+ * Copyright (C) 1995  Linus Torvalds
+ * Copyright 2001, 2002, 2003 SuSE Labs / Andi Kleen.
+ * See setup.c for older changelog.
+ *
+ * Jun Nakajima <jun.nakajima@intel.com> 
+ *   Modified for Xen
+ *
+ */ 
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/string.h>
+#include <linux/bootmem.h>
+#include <linux/bitops.h>
+#include <linux/module.h>
+#include <asm/bootsetup.h>
+#include <asm/pda.h>
+#include <asm/pgtable.h>
+#include <asm/processor.h>
+#include <asm/desc.h>
+#include <asm/atomic.h>
+#include <asm/mmu_context.h>
+#include <asm/smp.h>
+#include <asm/i387.h>
+#include <asm/percpu.h>
+#include <asm/proto.h>
+#include <asm/sections.h>
+#ifdef CONFIG_XEN
+#include <asm/hypervisor.h>
+#endif
+
+char x86_boot_params[BOOT_PARAM_SIZE] __initdata = {0,};
+
+cpumask_t cpu_initialized __cpuinitdata = CPU_MASK_NONE;
+
+struct x8664_pda *_cpu_pda[NR_CPUS] __read_mostly;
+EXPORT_SYMBOL(_cpu_pda);
+struct x8664_pda boot_cpu_pda[NR_CPUS] __cacheline_aligned;
+
+#ifndef CONFIG_X86_NO_IDT
+struct desc_ptr idt_descr = { 256 * 16 - 1, (unsigned long) idt_table }; 
+#endif
+
+char boot_cpu_stack[IRQSTACKSIZE] __attribute__((section(".bss.page_aligned")));
+
+unsigned long __supported_pte_mask __read_mostly = ~0UL;
+EXPORT_SYMBOL(__supported_pte_mask);
+static int do_not_nx __cpuinitdata = 0;
+
+/* noexec=on|off
+Control non executable mappings for 64bit processes.
+
+on	Enable(default)
+off	Disable
+*/ 
+int __init nonx_setup(char *str)
+{
+	if (!strncmp(str, "on", 2)) {
+                __supported_pte_mask |= _PAGE_NX; 
+ 		do_not_nx = 0; 
+	} else if (!strncmp(str, "off", 3)) {
+		do_not_nx = 1;
+		__supported_pte_mask &= ~_PAGE_NX;
+        }
+	return 1;
+} 
+__setup("noexec=", nonx_setup);	/* parsed early actually */
+
+int force_personality32 = 0; 
+
+/* noexec32=on|off
+Control non executable heap for 32bit processes.
+To control the stack too use noexec=off
+
+on	PROT_READ does not imply PROT_EXEC for 32bit processes
+off	PROT_READ implies PROT_EXEC (default)
+*/
+static int __init nonx32_setup(char *str)
+{
+	if (!strcmp(str, "on"))
+		force_personality32 &= ~READ_IMPLIES_EXEC;
+	else if (!strcmp(str, "off"))
+		force_personality32 |= READ_IMPLIES_EXEC;
+	return 1;
+}
+__setup("noexec32=", nonx32_setup);
+
+/*
+ * Great future plan:
+ * Declare PDA itself and support (irqstack,tss,pgd) as per cpu data.
+ * Always point %gs to its beginning
+ */
+void __init setup_per_cpu_areas(void)
+{ 
+	int i;
+	unsigned long size;
+
+#ifdef CONFIG_HOTPLUG_CPU
+	prefill_possible_map();
+#endif
+
+	/* Copy section for each CPU (we discard the original) */
+	size = ALIGN(__per_cpu_end - __per_cpu_start, SMP_CACHE_BYTES);
+#ifdef CONFIG_MODULES
+	if (size < PERCPU_ENOUGH_ROOM)
+		size = PERCPU_ENOUGH_ROOM;
+#endif
+
+	for_each_cpu_mask (i, cpu_possible_map) {
+		char *ptr;
+
+		if (!NODE_DATA(cpu_to_node(i))) {
+			printk("cpu with no node %d, num_online_nodes %d\n",
+			       i, num_online_nodes());
+			ptr = alloc_bootmem(size);
+		} else { 
+			ptr = alloc_bootmem_node(NODE_DATA(cpu_to_node(i)), size);
+		}
+		if (!ptr)
+			panic("Cannot allocate cpu data for CPU %d\n", i);
+		cpu_pda(i)->data_offset = ptr - __per_cpu_start;
+		memcpy(ptr, __per_cpu_start, __per_cpu_end - __per_cpu_start);
+	}
+} 
+
+#ifdef CONFIG_XEN
+static void switch_pt(void)
+{
+	xen_pt_switch(__pa_symbol(init_level4_pgt));
+	xen_new_user_pt(__pa_symbol(__user_pgd(init_level4_pgt)));
+}
+
+static void __cpuinit cpu_gdt_init(const struct desc_ptr *gdt_descr)
+{
+	unsigned long frames[16];
+	unsigned long va;
+	int f;
+
+	for (va = gdt_descr->address, f = 0;
+	     va < gdt_descr->address + gdt_descr->size;
+	     va += PAGE_SIZE, f++) {
+		frames[f] = virt_to_mfn(va);
+		make_page_readonly(
+			(void *)va, XENFEAT_writable_descriptor_tables);
+	}
+	if (HYPERVISOR_set_gdt(frames, (gdt_descr->size + 1) /
+                               sizeof (struct desc_struct)))
+		BUG();
+}
+#else
+static void switch_pt(void)
+{
+	asm volatile("movq %0,%%cr3" :: "r" (__pa_symbol(&init_level4_pgt)));
+}
+
+static void __cpuinit cpu_gdt_init(const struct desc_ptr *gdt_descr)
+{
+	asm volatile("lgdt %0" :: "m" (*gdt_descr));
+	asm volatile("lidt %0" :: "m" (idt_descr));
+}
+#endif
+
+void pda_init(int cpu)
+{ 
+	struct x8664_pda *pda = cpu_pda(cpu);
+
+	/* Setup up data that may be needed in __get_free_pages early */
+	asm volatile("movl %0,%%fs ; movl %0,%%gs" :: "r" (0)); 
+#ifndef CONFIG_XEN
+	wrmsrl(MSR_GS_BASE, pda);
+#else
+	if (HYPERVISOR_set_segment_base(SEGBASE_GS_KERNEL,
+					(unsigned long)pda))
+		BUG();
+#endif
+	pda->cpunumber = cpu; 
+	pda->irqcount = -1;
+	pda->kernelstack = 
+		(unsigned long)stack_thread_info() - PDA_STACKOFFSET + THREAD_SIZE; 
+	pda->active_mm = &init_mm;
+	pda->mmu_state = 0;
+
+	if (cpu == 0) {
+#ifdef CONFIG_XEN
+		xen_init_pt();
+#endif
+		/* others are initialized in smpboot.c */
+		pda->pcurrent = &init_task;
+		pda->irqstackptr = boot_cpu_stack; 
+	} else {
+		pda->irqstackptr = (char *)
+			__get_free_pages(GFP_ATOMIC, IRQSTACK_ORDER);
+		if (!pda->irqstackptr)
+			panic("cannot allocate irqstack for cpu %d", cpu); 
+	}
+
+	switch_pt();
+
+	pda->irqstackptr += IRQSTACKSIZE-64;
+} 
+
+#ifndef CONFIG_X86_NO_TSS
+char boot_exception_stacks[(N_EXCEPTION_STACKS - 1) * EXCEPTION_STKSZ + DEBUG_STKSZ]
+__attribute__((section(".bss.page_aligned")));
+#endif
+
+/* May not be marked __init: used by software suspend */
+void syscall_init(void)
+{
+#ifndef CONFIG_XEN
+	/* 
+	 * LSTAR and STAR live in a bit strange symbiosis.
+	 * They both write to the same internal register. STAR allows to set CS/DS
+	 * but only a 32bit target. LSTAR sets the 64bit rip. 	 
+	 */ 
+	wrmsrl(MSR_STAR,  ((u64)__USER32_CS)<<48  | ((u64)__KERNEL_CS)<<32); 
+	wrmsrl(MSR_LSTAR, system_call); 
+
+	/* Flags to clear on syscall */
+	wrmsrl(MSR_SYSCALL_MASK, EF_TF|EF_DF|EF_IE|0x3000); 
+#endif
+#ifdef CONFIG_IA32_EMULATION   		
+	syscall32_cpu_init ();
+#endif
+}
+
+void __cpuinit check_efer(void)
+{
+	unsigned long efer;
+
+	rdmsrl(MSR_EFER, efer); 
+        if (!(efer & EFER_NX) || do_not_nx) { 
+                __supported_pte_mask &= ~_PAGE_NX; 
+        }       
+}
+
+unsigned long kernel_eflags;
+
+/*
+ * cpu_init() initializes state that is per-CPU. Some data is already
+ * initialized (naturally) in the bootstrap process, such as the GDT
+ * and IDT. We reload them nevertheless, this function acts as a
+ * 'CPU state barrier', nothing should get across.
+ * A lot of state is already set up in PDA init.
+ */
+void __cpuinit cpu_init (void)
+{
+	int cpu = stack_smp_processor_id();
+#ifndef CONFIG_X86_NO_TSS
+	struct tss_struct *t = &per_cpu(init_tss, cpu);
+	struct orig_ist *orig_ist = &per_cpu(orig_ist, cpu);
+	unsigned long v; 
+	char *estacks = NULL; 
+	unsigned i;
+#endif
+	struct task_struct *me;
+
+	/* CPU 0 is initialised in head64.c */
+	if (cpu != 0) {
+		pda_init(cpu);
+		zap_low_mappings(cpu);
+	}
+#ifndef CONFIG_X86_NO_TSS
+	else
+		estacks = boot_exception_stacks; 
+#endif
+
+	me = current;
+
+	if (cpu_test_and_set(cpu, cpu_initialized))
+		panic("CPU#%d already initialized!\n", cpu);
+
+	printk("Initializing CPU#%d\n", cpu);
+
+	clear_in_cr4(X86_CR4_VME|X86_CR4_PVI|X86_CR4_TSD|X86_CR4_DE);
+
+	/*
+	 * Initialize the per-CPU GDT with the boot GDT,
+	 * and set up the GDT descriptor:
+	 */
+#ifndef CONFIG_XEN 
+	if (cpu)
+ 		memcpy(cpu_gdt(cpu), cpu_gdt_table, GDT_SIZE);
+#endif
+
+	cpu_gdt_descr[cpu].size = GDT_SIZE;
+	cpu_gdt_init(&cpu_gdt_descr[cpu]);
+
+	memset(me->thread.tls_array, 0, GDT_ENTRY_TLS_ENTRIES * 8);
+	syscall_init();
+
+	wrmsrl(MSR_FS_BASE, 0);
+	wrmsrl(MSR_KERNEL_GS_BASE, 0);
+	barrier(); 
+
+	check_efer();
+
+#ifndef CONFIG_X86_NO_TSS
+	/*
+	 * set up and load the per-CPU TSS
+	 */
+	for (v = 0; v < N_EXCEPTION_STACKS; v++) {
+		if (cpu) {
+			static const unsigned int order[N_EXCEPTION_STACKS] = {
+				[0 ... N_EXCEPTION_STACKS - 1] = EXCEPTION_STACK_ORDER,
+				[DEBUG_STACK - 1] = DEBUG_STACK_ORDER
+			};
+
+			estacks = (char *)__get_free_pages(GFP_ATOMIC, order[v]);
+			if (!estacks)
+				panic("Cannot allocate exception stack %ld %d\n",
+				      v, cpu); 
+		}
+		switch (v + 1) {
+#if DEBUG_STKSZ > EXCEPTION_STKSZ
+		case DEBUG_STACK:
+			cpu_pda(cpu)->debugstack = (unsigned long)estacks;
+			estacks += DEBUG_STKSZ;
+			break;
+#endif
+		default:
+			estacks += EXCEPTION_STKSZ;
+			break;
+		}
+		orig_ist->ist[v] = t->ist[v] = (unsigned long)estacks;
+	}
+
+	t->io_bitmap_base = offsetof(struct tss_struct, io_bitmap);
+	/*
+	 * <= is required because the CPU will access up to
+	 * 8 bits beyond the end of the IO permission bitmap.
+	 */
+	for (i = 0; i <= IO_BITMAP_LONGS; i++)
+		t->io_bitmap[i] = ~0UL;
+#endif
+
+	atomic_inc(&init_mm.mm_count);
+	me->active_mm = &init_mm;
+	if (me->mm)
+		BUG();
+	enter_lazy_tlb(&init_mm, me);
+
+#ifndef CONFIG_X86_NO_TSS
+	set_tss_desc(cpu, t);
+#endif
+#ifndef CONFIG_XEN
+	load_TR_desc();
+#endif
+	load_LDT(&init_mm.context);
+
+	/*
+	 * Clear all 6 debug registers:
+	 */
+
+	set_debugreg(0UL, 0);
+	set_debugreg(0UL, 1);
+	set_debugreg(0UL, 2);
+	set_debugreg(0UL, 3);
+	set_debugreg(0UL, 6);
+	set_debugreg(0UL, 7);
+
+	fpu_init(); 
+
+	raw_local_save_flags(kernel_eflags);
+}
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/setup.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/setup.c
--- linux-2.6.18.8/arch/x86_64/kernel/setup.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/setup.c	2008-08-21 11:36:07.000000000 +0200
@@ -122,7 +122,9 @@ struct resource standard_io_resources[] 
 		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
 	{ .name = "timer1", .start = 0x50, .end = 0x53,
 		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
-	{ .name = "keyboard", .start = 0x60, .end = 0x6f,
+	{ .name = "keyboard", .start = 0x60, .end = 0x60,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "keyboard", .start = 0x64, .end = 0x64,
 		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
 	{ .name = "dma page reg", .start = 0x80, .end = 0x8f,
 		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/setup-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/setup-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/setup-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/setup-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1654 @@
+/*
+ *  linux/arch/x86-64/kernel/setup.c
+ *
+ *  Copyright (C) 1995  Linus Torvalds
+ *
+ *  Nov 2001 Dave Jones <davej@suse.de>
+ *  Forked from i386 setup code.
+ */
+
+/*
+ * This file handles the architecture-dependent parts of initialization
+ */
+
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/stddef.h>
+#include <linux/unistd.h>
+#include <linux/ptrace.h>
+#include <linux/slab.h>
+#include <linux/user.h>
+#include <linux/a.out.h>
+#include <linux/screen_info.h>
+#include <linux/ioport.h>
+#include <linux/delay.h>
+#include <linux/init.h>
+#include <linux/initrd.h>
+#include <linux/highmem.h>
+#include <linux/bootmem.h>
+#include <linux/module.h>
+#include <asm/processor.h>
+#include <linux/console.h>
+#include <linux/seq_file.h>
+#include <linux/crash_dump.h>
+#include <linux/root_dev.h>
+#include <linux/pci.h>
+#include <linux/acpi.h>
+#include <linux/kallsyms.h>
+#include <linux/edd.h>
+#include <linux/mmzone.h>
+#include <linux/kexec.h>
+#include <linux/cpufreq.h>
+#include <linux/dmi.h>
+#include <linux/dma-mapping.h>
+#include <linux/ctype.h>
+
+#include <asm/mtrr.h>
+#include <asm/uaccess.h>
+#include <asm/system.h>
+#include <asm/io.h>
+#include <asm/smp.h>
+#include <asm/msr.h>
+#include <asm/desc.h>
+#include <video/edid.h>
+#include <asm/e820.h>
+#include <asm/dma.h>
+#include <asm/mpspec.h>
+#include <asm/mmu_context.h>
+#include <asm/bootsetup.h>
+#include <asm/proto.h>
+#include <asm/setup.h>
+#include <asm/mach_apic.h>
+#include <asm/numa.h>
+#include <asm/sections.h>
+#include <asm/dmi.h>
+#ifdef CONFIG_XEN
+#include <linux/percpu.h>
+#include <xen/interface/physdev.h>
+#include "setup_arch_pre.h"
+#include <asm/hypervisor.h>
+#include <xen/interface/nmi.h>
+#include <xen/features.h>
+#include <xen/firmware.h>
+#include <xen/xencons.h>
+#define PFN_UP(x)       (((x) + PAGE_SIZE-1) >> PAGE_SHIFT)
+#define PFN_PHYS(x)     ((x) << PAGE_SHIFT)
+#include <asm/mach-xen/setup_arch_post.h>
+#include <xen/interface/memory.h>
+
+#ifdef CONFIG_XEN
+#include <xen/interface/kexec.h>
+#endif
+
+extern unsigned long start_pfn;
+extern struct edid_info edid_info;
+
+shared_info_t *HYPERVISOR_shared_info = (shared_info_t *)empty_zero_page;
+EXPORT_SYMBOL(HYPERVISOR_shared_info);
+
+extern char hypercall_page[PAGE_SIZE];
+EXPORT_SYMBOL(hypercall_page);
+
+static int xen_panic_event(struct notifier_block *, unsigned long, void *);
+static struct notifier_block xen_panic_block = {
+	xen_panic_event, NULL, 0 /* try to go last */
+};
+
+unsigned long *phys_to_machine_mapping;
+unsigned long *pfn_to_mfn_frame_list_list, *pfn_to_mfn_frame_list[512];
+
+EXPORT_SYMBOL(phys_to_machine_mapping);
+
+DEFINE_PER_CPU(multicall_entry_t, multicall_list[8]);
+DEFINE_PER_CPU(int, nr_multicall_ents);
+
+/* Raw start-of-day parameters from the hypervisor. */
+start_info_t *xen_start_info;
+EXPORT_SYMBOL(xen_start_info);
+#endif
+
+/*
+ * Machine setup..
+ */
+
+struct cpuinfo_x86 boot_cpu_data __read_mostly;
+EXPORT_SYMBOL(boot_cpu_data);
+
+unsigned long mmu_cr4_features;
+
+int acpi_disabled;
+EXPORT_SYMBOL(acpi_disabled);
+#ifdef	CONFIG_ACPI
+extern int __initdata acpi_ht;
+extern acpi_interrupt_flags	acpi_sci_flags;
+int __initdata acpi_force = 0;
+#endif
+
+int acpi_numa __initdata;
+
+/* Boot loader ID as an integer, for the benefit of proc_dointvec */
+int bootloader_type;
+
+unsigned long saved_video_mode;
+
+/* 
+ * Early DMI memory
+ */
+int dmi_alloc_index;
+char dmi_alloc_data[DMI_MAX_DATA];
+
+/*
+ * Setup options
+ */
+struct screen_info screen_info;
+EXPORT_SYMBOL(screen_info);
+struct sys_desc_table_struct {
+	unsigned short length;
+	unsigned char table[0];
+};
+
+struct edid_info edid_info;
+EXPORT_SYMBOL_GPL(edid_info);
+struct e820map e820;
+#ifdef CONFIG_XEN
+struct e820map machine_e820;
+#endif
+
+extern int root_mountflags;
+
+char command_line[COMMAND_LINE_SIZE];
+
+struct resource standard_io_resources[] = {
+	{ .name = "dma1", .start = 0x00, .end = 0x1f,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "pic1", .start = 0x20, .end = 0x21,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "timer0", .start = 0x40, .end = 0x43,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "timer1", .start = 0x50, .end = 0x53,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "keyboard", .start = 0x60, .end = 0x60,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "keyboard", .start = 0x64, .end = 0x64,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "dma page reg", .start = 0x80, .end = 0x8f,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "pic2", .start = 0xa0, .end = 0xa1,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "dma2", .start = 0xc0, .end = 0xdf,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO },
+	{ .name = "fpu", .start = 0xf0, .end = 0xff,
+		.flags = IORESOURCE_BUSY | IORESOURCE_IO }
+};
+
+#define STANDARD_IO_RESOURCES \
+	(sizeof standard_io_resources / sizeof standard_io_resources[0])
+
+#define IORESOURCE_RAM (IORESOURCE_BUSY | IORESOURCE_MEM)
+
+struct resource data_resource = {
+	.name = "Kernel data",
+	.start = 0,
+	.end = 0,
+	.flags = IORESOURCE_RAM,
+};
+struct resource code_resource = {
+	.name = "Kernel code",
+	.start = 0,
+	.end = 0,
+	.flags = IORESOURCE_RAM,
+};
+
+#define IORESOURCE_ROM (IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM)
+
+static struct resource system_rom_resource = {
+	.name = "System ROM",
+	.start = 0xf0000,
+	.end = 0xfffff,
+	.flags = IORESOURCE_ROM,
+};
+
+static struct resource extension_rom_resource = {
+	.name = "Extension ROM",
+	.start = 0xe0000,
+	.end = 0xeffff,
+	.flags = IORESOURCE_ROM,
+};
+
+static struct resource adapter_rom_resources[] = {
+	{ .name = "Adapter ROM", .start = 0xc8000, .end = 0,
+		.flags = IORESOURCE_ROM },
+	{ .name = "Adapter ROM", .start = 0, .end = 0,
+		.flags = IORESOURCE_ROM },
+	{ .name = "Adapter ROM", .start = 0, .end = 0,
+		.flags = IORESOURCE_ROM },
+	{ .name = "Adapter ROM", .start = 0, .end = 0,
+		.flags = IORESOURCE_ROM },
+	{ .name = "Adapter ROM", .start = 0, .end = 0,
+		.flags = IORESOURCE_ROM },
+	{ .name = "Adapter ROM", .start = 0, .end = 0,
+		.flags = IORESOURCE_ROM }
+};
+
+#define ADAPTER_ROM_RESOURCES \
+	(sizeof adapter_rom_resources / sizeof adapter_rom_resources[0])
+
+static struct resource video_rom_resource = {
+	.name = "Video ROM",
+	.start = 0xc0000,
+	.end = 0xc7fff,
+	.flags = IORESOURCE_ROM,
+};
+
+static struct resource video_ram_resource = {
+	.name = "Video RAM area",
+	.start = 0xa0000,
+	.end = 0xbffff,
+	.flags = IORESOURCE_RAM,
+};
+
+#define romsignature(x) (*(unsigned short *)(x) == 0xaa55)
+
+static int __init romchecksum(unsigned char *rom, unsigned long length)
+{
+	unsigned char *p, sum = 0;
+
+	for (p = rom; p < rom + length; p++)
+		sum += *p;
+	return sum == 0;
+}
+
+static void __init probe_roms(void)
+{
+	unsigned long start, length, upper;
+	unsigned char *rom;
+	int	      i;
+
+#ifdef CONFIG_XEN
+	/* Nothing to do if not running in dom0. */
+	if (!is_initial_xendomain())
+		return;
+#endif
+
+	/* video rom */
+	upper = adapter_rom_resources[0].start;
+	for (start = video_rom_resource.start; start < upper; start += 2048) {
+		rom = isa_bus_to_virt(start);
+		if (!romsignature(rom))
+			continue;
+
+		video_rom_resource.start = start;
+
+		/* 0 < length <= 0x7f * 512, historically */
+		length = rom[2] * 512;
+
+		/* if checksum okay, trust length byte */
+		if (length && romchecksum(rom, length))
+			video_rom_resource.end = start + length - 1;
+
+		request_resource(&iomem_resource, &video_rom_resource);
+		break;
+			}
+
+	start = (video_rom_resource.end + 1 + 2047) & ~2047UL;
+	if (start < upper)
+		start = upper;
+
+	/* system rom */
+	request_resource(&iomem_resource, &system_rom_resource);
+	upper = system_rom_resource.start;
+
+	/* check for extension rom (ignore length byte!) */
+	rom = isa_bus_to_virt(extension_rom_resource.start);
+	if (romsignature(rom)) {
+		length = extension_rom_resource.end - extension_rom_resource.start + 1;
+		if (romchecksum(rom, length)) {
+			request_resource(&iomem_resource, &extension_rom_resource);
+			upper = extension_rom_resource.start;
+		}
+	}
+
+	/* check for adapter roms on 2k boundaries */
+	for (i = 0; i < ADAPTER_ROM_RESOURCES && start < upper; start += 2048) {
+		rom = isa_bus_to_virt(start);
+		if (!romsignature(rom))
+			continue;
+
+		/* 0 < length <= 0x7f * 512, historically */
+		length = rom[2] * 512;
+
+		/* but accept any length that fits if checksum okay */
+		if (!length || start + length > upper || !romchecksum(rom, length))
+			continue;
+
+		adapter_rom_resources[i].start = start;
+		adapter_rom_resources[i].end = start + length - 1;
+		request_resource(&iomem_resource, &adapter_rom_resources[i]);
+
+		start = adapter_rom_resources[i++].end & ~2047UL;
+	}
+}
+
+/* Check for full argument with no trailing characters */
+static int fullarg(char *p, char *arg)
+{
+	int l = strlen(arg);
+	return !memcmp(p, arg, l) && (p[l] == 0 || isspace(p[l]));
+}
+
+static __init void parse_cmdline_early (char ** cmdline_p)
+{
+	char c = ' ', *to = command_line, *from = COMMAND_LINE;
+	int len = 0;
+	int userdef = 0;
+
+	for (;;) {
+		if (c != ' ') 
+			goto next_char; 
+
+#ifdef  CONFIG_SMP
+		/*
+		 * If the BIOS enumerates physical processors before logical,
+		 * maxcpus=N at enumeration-time can be used to disable HT.
+		 */
+		else if (!memcmp(from, "maxcpus=", 8)) {
+			extern unsigned int maxcpus;
+
+			maxcpus = simple_strtoul(from + 8, NULL, 0);
+		}
+#endif
+#ifdef CONFIG_ACPI
+		/* "acpi=off" disables both ACPI table parsing and interpreter init */
+		if (fullarg(from,"acpi=off"))
+			disable_acpi();
+
+		if (fullarg(from, "acpi=force")) { 
+			/* add later when we do DMI horrors: */
+			acpi_force = 1;
+			acpi_disabled = 0;
+		}
+
+		/* acpi=ht just means: do ACPI MADT parsing 
+		   at bootup, but don't enable the full ACPI interpreter */
+		if (fullarg(from, "acpi=ht")) { 
+			if (!acpi_force)
+				disable_acpi();
+			acpi_ht = 1; 
+		}
+                else if (fullarg(from, "pci=noacpi")) 
+			acpi_disable_pci();
+		else if (fullarg(from, "acpi=noirq"))
+			acpi_noirq_set();
+
+		else if (fullarg(from, "acpi_sci=edge"))
+			acpi_sci_flags.trigger =  1;
+		else if (fullarg(from, "acpi_sci=level"))
+			acpi_sci_flags.trigger = 3;
+		else if (fullarg(from, "acpi_sci=high"))
+			acpi_sci_flags.polarity = 1;
+		else if (fullarg(from, "acpi_sci=low"))
+			acpi_sci_flags.polarity = 3;
+
+		/* acpi=strict disables out-of-spec workarounds */
+		else if (fullarg(from, "acpi=strict")) {
+			acpi_strict = 1;
+		}
+#ifdef CONFIG_X86_IO_APIC
+		else if (fullarg(from, "acpi_skip_timer_override"))
+			acpi_skip_timer_override = 1;
+#endif
+#endif
+
+#ifndef CONFIG_XEN
+		if (fullarg(from, "nolapic") || fullarg(from, "disableapic")) {
+			clear_bit(X86_FEATURE_APIC, boot_cpu_data.x86_capability);
+			disable_apic = 1;
+		}
+
+		if (fullarg(from, "noapic"))
+			skip_ioapic_setup = 1;
+
+		if (fullarg(from,"apic")) {
+			skip_ioapic_setup = 0;
+			ioapic_force = 1;
+		}
+#endif
+			
+		if (!memcmp(from, "mem=", 4))
+			parse_memopt(from+4, &from); 
+
+		if (!memcmp(from, "memmap=", 7)) {
+			/* exactmap option is for used defined memory */
+			if (!memcmp(from+7, "exactmap", 8)) {
+#ifdef CONFIG_CRASH_DUMP
+				/* If we are doing a crash dump, we
+				 * still need to know the real mem
+				 * size before original memory map is
+				 * reset.
+				 */
+				saved_max_pfn = e820_end_of_ram();
+#endif
+				from += 8+7;
+				end_pfn_map = 0;
+				e820.nr_map = 0;
+				userdef = 1;
+			}
+			else {
+				parse_memmapopt(from+7, &from);
+				userdef = 1;
+			}
+		}
+
+#ifdef CONFIG_NUMA
+		if (!memcmp(from, "numa=", 5))
+			numa_setup(from+5); 
+#endif
+
+		if (!memcmp(from,"iommu=",6)) { 
+			iommu_setup(from+6); 
+		}
+
+		if (fullarg(from,"oops=panic"))
+			panic_on_oops = 1;
+
+		if (!memcmp(from, "noexec=", 7))
+			nonx_setup(from + 7);
+
+#ifdef CONFIG_KEXEC
+		/* crashkernel=size@addr specifies the location to reserve for
+		 * a crash kernel.  By reserving this memory we guarantee
+		 * that linux never set's it up as a DMA target.
+		 * Useful for holding code to do something appropriate
+		 * after a kernel panic.
+		 */
+		else if (!memcmp(from, "crashkernel=", 12)) {
+#ifndef CONFIG_XEN
+			unsigned long size, base;
+			size = memparse(from+12, &from);
+			if (*from == '@') {
+				base = memparse(from+1, &from);
+				/* FIXME: Do I want a sanity check
+				 * to validate the memory range?
+				 */
+				crashk_res.start = base;
+				crashk_res.end   = base + size - 1;
+			}
+#else
+			printk("Ignoring crashkernel command line, "
+			       "parameter will be supplied by xen\n");
+#endif
+		}
+#endif
+
+#ifdef CONFIG_PROC_VMCORE
+		/* elfcorehdr= specifies the location of elf core header
+		 * stored by the crashed kernel. This option will be passed
+		 * by kexec loader to the capture kernel.
+		 */
+		else if(!memcmp(from, "elfcorehdr=", 11))
+			elfcorehdr_addr = memparse(from+11, &from);
+#endif
+
+#if defined(CONFIG_HOTPLUG_CPU) && !defined(CONFIG_XEN)
+		else if (!memcmp(from, "additional_cpus=", 16))
+			setup_additional_cpus(from+16);
+#endif
+
+	next_char:
+		c = *(from++);
+		if (!c)
+			break;
+		if (COMMAND_LINE_SIZE <= ++len)
+			break;
+		*(to++) = c;
+	}
+	if (userdef) {
+		printk(KERN_INFO "user-defined physical RAM map:\n");
+		e820_print_map("user");
+	}
+	*to = '\0';
+	*cmdline_p = command_line;
+}
+
+#ifndef CONFIG_NUMA
+static void __init
+contig_initmem_init(unsigned long start_pfn, unsigned long end_pfn)
+{
+	unsigned long bootmap_size, bootmap;
+
+	bootmap_size = bootmem_bootmap_pages(end_pfn)<<PAGE_SHIFT;
+	bootmap = find_e820_area(0, end_pfn<<PAGE_SHIFT, bootmap_size);
+	if (bootmap == -1L)
+		panic("Cannot find bootmem map of size %ld\n",bootmap_size);
+	bootmap_size = init_bootmem(bootmap >> PAGE_SHIFT, end_pfn);
+#ifdef CONFIG_XEN
+	e820_bootmem_free(NODE_DATA(0), 0, xen_start_info->nr_pages<<PAGE_SHIFT);
+#else
+	e820_bootmem_free(NODE_DATA(0), 0, end_pfn << PAGE_SHIFT);
+#endif
+	reserve_bootmem(bootmap, bootmap_size);
+} 
+#endif
+
+#if defined(CONFIG_EDD) || defined(CONFIG_EDD_MODULE)
+struct edd edd;
+#ifdef CONFIG_EDD_MODULE
+EXPORT_SYMBOL(edd);
+#endif
+#ifndef CONFIG_XEN
+/**
+ * copy_edd() - Copy the BIOS EDD information
+ *              from boot_params into a safe place.
+ *
+ */
+static inline void copy_edd(void)
+{
+     memcpy(edd.mbr_signature, EDD_MBR_SIGNATURE, sizeof(edd.mbr_signature));
+     memcpy(edd.edd_info, EDD_BUF, sizeof(edd.edd_info));
+     edd.mbr_signature_nr = EDD_MBR_SIG_NR;
+     edd.edd_info_nr = EDD_NR;
+}
+#endif
+#else
+static inline void copy_edd(void)
+{
+}
+#endif
+
+#ifndef CONFIG_XEN
+#define EBDA_ADDR_POINTER 0x40E
+
+unsigned __initdata ebda_addr;
+unsigned __initdata ebda_size;
+
+static void discover_ebda(void)
+{
+	/*
+	 * there is a real-mode segmented pointer pointing to the 
+	 * 4K EBDA area at 0x40E
+	 */
+	ebda_addr = *(unsigned short *)EBDA_ADDR_POINTER;
+	ebda_addr <<= 4;
+
+	ebda_size = *(unsigned short *)(unsigned long)ebda_addr;
+
+	/* Round EBDA up to pages */
+	if (ebda_size == 0)
+		ebda_size = 1;
+	ebda_size <<= 10;
+	ebda_size = round_up(ebda_size + (ebda_addr & ~PAGE_MASK), PAGE_SIZE);
+	if (ebda_size > 64*1024)
+		ebda_size = 64*1024;
+}
+#else
+#define discover_ebda() ((void)0)
+#endif
+
+void __init setup_arch(char **cmdline_p)
+{
+#ifdef CONFIG_XEN
+	/* Register a call for panic conditions. */
+	atomic_notifier_chain_register(&panic_notifier_list, &xen_panic_block);
+
+ 	ROOT_DEV = MKDEV(RAMDISK_MAJOR,0); 
+ 	screen_info = SCREEN_INFO;
+
+	if (is_initial_xendomain()) {
+		const struct dom0_vga_console_info *info =
+			(void *)((char *)xen_start_info +
+			         xen_start_info->console.dom0.info_off);
+
+		dom0_init_screen_info(info,
+		                      xen_start_info->console.dom0.info_size);
+		xen_start_info->console.domU.mfn = 0;
+		xen_start_info->console.domU.evtchn = 0;
+	} else
+		screen_info.orig_video_isVGA = 0;
+
+	copy_edid();
+
+	WARN_ON(HYPERVISOR_vm_assist(VMASST_CMD_enable,
+				     VMASST_TYPE_writable_pagetables));
+
+	ARCH_SETUP
+#else
+ 	ROOT_DEV = old_decode_dev(ORIG_ROOT_DEV);
+ 	screen_info = SCREEN_INFO;
+	edid_info = EDID_INFO;
+#endif	/* !CONFIG_XEN */
+	saved_video_mode = SAVED_VIDEO_MODE;
+	bootloader_type = LOADER_TYPE;
+
+#ifdef CONFIG_BLK_DEV_RAM
+	rd_image_start = RAMDISK_FLAGS & RAMDISK_IMAGE_START_MASK;
+	rd_prompt = ((RAMDISK_FLAGS & RAMDISK_PROMPT_FLAG) != 0);
+	rd_doload = ((RAMDISK_FLAGS & RAMDISK_LOAD_FLAG) != 0);
+#endif
+	setup_memory_region();
+	copy_edd();
+
+	if (!MOUNT_ROOT_RDONLY)
+		root_mountflags &= ~MS_RDONLY;
+	init_mm.start_code = (unsigned long) &_text;
+	init_mm.end_code = (unsigned long) &_etext;
+	init_mm.end_data = (unsigned long) &_edata;
+	init_mm.brk = (unsigned long) &_end;
+
+	code_resource.start = virt_to_phys(&_text);
+	code_resource.end = virt_to_phys(&_etext)-1;
+	data_resource.start = virt_to_phys(&_etext);
+	data_resource.end = virt_to_phys(&_edata)-1;
+
+	parse_cmdline_early(cmdline_p);
+
+	early_identify_cpu(&boot_cpu_data);
+
+	/*
+	 * partially used pages are not usable - thus
+	 * we are rounding upwards:
+	 */
+	end_pfn = e820_end_of_ram();
+	num_physpages = end_pfn;		/* for pfn_valid */
+
+	check_efer();
+
+	discover_ebda();
+
+	init_memory_mapping(0, (end_pfn_map << PAGE_SHIFT));
+
+	if (is_initial_xendomain())
+		dmi_scan_machine();
+
+#ifdef CONFIG_ACPI_NUMA
+	/*
+	 * Parse SRAT to discover nodes.
+	 */
+	acpi_numa_init();
+#endif
+
+#ifdef CONFIG_NUMA
+	numa_initmem_init(0, end_pfn); 
+#else
+	contig_initmem_init(0, end_pfn);
+#endif
+
+#ifdef CONFIG_XEN
+	/*
+	 * Reserve kernel, physmap, start info, initial page tables, and
+	 * direct mapping.
+	 */
+	reserve_bootmem_generic(__pa_symbol(&_text),
+	                        (table_end << PAGE_SHIFT) - __pa_symbol(&_text));
+#else
+	/* Reserve direct mapping */
+	reserve_bootmem_generic(table_start << PAGE_SHIFT, 
+				(table_end - table_start) << PAGE_SHIFT);
+
+	/* reserve kernel */
+	reserve_bootmem_generic(__pa_symbol(&_text),
+				__pa_symbol(&_end) - __pa_symbol(&_text));
+
+	/*
+	 * reserve physical page 0 - it's a special BIOS page on many boxes,
+	 * enabling clean reboots, SMP operation, laptop functions.
+	 */
+	reserve_bootmem_generic(0, PAGE_SIZE);
+
+	/* reserve ebda region */
+	if (ebda_addr)
+		reserve_bootmem_generic(ebda_addr, ebda_size);
+
+#ifdef CONFIG_SMP
+	/*
+	 * But first pinch a few for the stack/trampoline stuff
+	 * FIXME: Don't need the extra page at 4K, but need to fix
+	 * trampoline before removing it. (see the GDT stuff)
+	 */
+	reserve_bootmem_generic(PAGE_SIZE, PAGE_SIZE);
+
+	/* Reserve SMP trampoline */
+	reserve_bootmem_generic(SMP_TRAMPOLINE_BASE, PAGE_SIZE);
+#endif
+#endif
+
+#ifdef CONFIG_ACPI_SLEEP
+       /*
+        * Reserve low memory region for sleep support.
+        */
+       acpi_reserve_bootmem();
+#endif
+#ifdef CONFIG_XEN
+#ifdef CONFIG_BLK_DEV_INITRD
+	if (xen_start_info->mod_start) {
+		if (INITRD_START + INITRD_SIZE <= (end_pfn << PAGE_SHIFT)) {
+			/*reserve_bootmem_generic(INITRD_START, INITRD_SIZE);*/
+			initrd_start = INITRD_START + PAGE_OFFSET;
+			initrd_end = initrd_start+INITRD_SIZE;
+			initrd_below_start_ok = 1;
+		} else {
+			printk(KERN_ERR "initrd extends beyond end of memory "
+				"(0x%08lx > 0x%08lx)\ndisabling initrd\n",
+				(unsigned long)(INITRD_START + INITRD_SIZE),
+				(unsigned long)(end_pfn << PAGE_SHIFT));
+			initrd_start = 0;
+		}
+	}
+#endif
+#else	/* CONFIG_XEN */
+#ifdef CONFIG_BLK_DEV_INITRD
+	if (LOADER_TYPE && INITRD_START) {
+		if (INITRD_START + INITRD_SIZE <= (end_pfn << PAGE_SHIFT)) {
+			reserve_bootmem_generic(INITRD_START, INITRD_SIZE);
+			initrd_start =
+				INITRD_START ? INITRD_START + PAGE_OFFSET : 0;
+			initrd_end = initrd_start+INITRD_SIZE;
+		}
+		else {
+			printk(KERN_ERR "initrd extends beyond end of memory "
+			    "(0x%08lx > 0x%08lx)\ndisabling initrd\n",
+			    (unsigned long)(INITRD_START + INITRD_SIZE),
+			    (unsigned long)(end_pfn << PAGE_SHIFT));
+			initrd_start = 0;
+		}
+	}
+#endif
+#endif	/* !CONFIG_XEN */
+#ifdef CONFIG_KEXEC
+#ifdef CONFIG_XEN
+	xen_machine_kexec_setup_resources();
+#else
+	if (crashk_res.start != crashk_res.end) {
+		reserve_bootmem_generic(crashk_res.start,
+			crashk_res.end - crashk_res.start + 1);
+	}
+#endif
+#endif
+
+	paging_init();
+#ifdef CONFIG_X86_LOCAL_APIC
+	/*
+	 * Find and reserve possible boot-time SMP configuration:
+	 */
+	find_smp_config();
+#endif
+#ifdef CONFIG_XEN
+	{
+		int i, j, k, fpp;
+		unsigned long p2m_pages;
+
+		p2m_pages = end_pfn;
+		if (xen_start_info->nr_pages > end_pfn) {
+			/*
+			 * the end_pfn was shrunk (probably by mem= or highmem=
+			 * kernel parameter); shrink reservation with the HV
+			 */
+			struct xen_memory_reservation reservation = {
+				.address_bits = 0,
+				.extent_order = 0,
+				.domid = DOMID_SELF
+			};
+			unsigned int difference;
+			int ret;
+			
+			difference = xen_start_info->nr_pages - end_pfn;
+			
+			set_xen_guest_handle(reservation.extent_start,
+					     ((unsigned long *)xen_start_info->mfn_list) + end_pfn);
+			reservation.nr_extents = difference;
+			ret = HYPERVISOR_memory_op(XENMEM_decrease_reservation,
+						   &reservation);
+			BUG_ON (ret != difference);
+		}
+		else if (end_pfn > xen_start_info->nr_pages)
+			p2m_pages = xen_start_info->nr_pages;
+
+		if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+			/* Make sure we have a large enough P->M table. */
+			phys_to_machine_mapping = alloc_bootmem_pages(
+				end_pfn * sizeof(unsigned long));
+			memset(phys_to_machine_mapping, ~0,
+			       end_pfn * sizeof(unsigned long));
+			memcpy(phys_to_machine_mapping,
+			       (unsigned long *)xen_start_info->mfn_list,
+			       p2m_pages * sizeof(unsigned long));
+			free_bootmem(
+				__pa(xen_start_info->mfn_list),
+				PFN_PHYS(PFN_UP(xen_start_info->nr_pages *
+						sizeof(unsigned long))));
+
+			/*
+			 * Initialise the list of the frames that specify the
+			 * list of frames that make up the p2m table. Used by
+                         * save/restore.
+			 */
+			pfn_to_mfn_frame_list_list = alloc_bootmem_pages(PAGE_SIZE);
+
+			fpp = PAGE_SIZE/sizeof(unsigned long);
+			for (i=0, j=0, k=-1; i< end_pfn; i+=fpp, j++) {
+				if ((j % fpp) == 0) {
+					k++;
+					BUG_ON(k>=fpp);
+					pfn_to_mfn_frame_list[k] =
+						alloc_bootmem_pages(PAGE_SIZE);
+					pfn_to_mfn_frame_list_list[k] =
+						virt_to_mfn(pfn_to_mfn_frame_list[k]);
+					j=0;
+				}
+				pfn_to_mfn_frame_list[k][j] =
+					virt_to_mfn(&phys_to_machine_mapping[i]);
+			}
+			HYPERVISOR_shared_info->arch.max_pfn = end_pfn;
+			HYPERVISOR_shared_info->arch.pfn_to_mfn_frame_list_list =
+				virt_to_mfn(pfn_to_mfn_frame_list_list);
+		}
+
+		/* Mark all ISA DMA channels in-use - using them wouldn't work. */
+		for (i = 0; i < MAX_DMA_CHANNELS; ++i)
+			if (i != 4 && request_dma(i, "xen") != 0)
+				BUG();
+	}
+
+	if (!is_initial_xendomain()) {
+		acpi_disabled = 1;
+#ifdef  CONFIG_ACPI
+		acpi_ht = 0;
+#endif
+	}
+#endif
+
+#ifndef CONFIG_XEN
+	check_ioapic();
+#endif
+
+	zap_low_mappings(0);
+
+	/*
+	 * set this early, so we dont allocate cpu0
+	 * if MADT list doesnt list BSP first
+	 * mpparse.c/MP_processor_info() allocates logical cpu numbers.
+	 */
+	cpu_set(0, cpu_present_map);
+#ifdef CONFIG_ACPI
+	/*
+	 * Initialize the ACPI boot-time table parser (gets the RSDP and SDT).
+	 * Call this early for SRAT node setup.
+	 */
+	acpi_boot_table_init();
+
+	/*
+	 * Read APIC and some other early information from ACPI tables.
+	 */
+	acpi_boot_init();
+#endif
+
+	init_cpu_to_node();
+
+#ifdef CONFIG_X86_LOCAL_APIC
+	/*
+	 * get boot-time SMP configuration:
+	 */
+	if (smp_found_config)
+		get_smp_config();
+#ifndef CONFIG_XEN
+	init_apic_mappings();
+#endif
+#endif
+#if defined(CONFIG_XEN) && defined(CONFIG_SMP) && !defined(CONFIG_HOTPLUG_CPU)
+	prefill_possible_map();
+#endif
+
+	/*
+	 * Request address space for all standard RAM and ROM resources
+	 * and also for regions reported as reserved by the e820.
+	 */
+	probe_roms();
+#ifdef CONFIG_XEN
+	if (is_initial_xendomain())
+		e820_reserve_resources(machine_e820.map, machine_e820.nr_map);
+#else
+	e820_reserve_resources(e820.map, e820.nr_map);
+#endif
+
+	request_resource(&iomem_resource, &video_ram_resource);
+
+	{
+	unsigned i;
+	/* request I/O space for devices used on all i[345]86 PCs */
+	for (i = 0; i < STANDARD_IO_RESOURCES; i++)
+		request_resource(&ioport_resource, &standard_io_resources[i]);
+	}
+
+#ifdef CONFIG_XEN
+	if (is_initial_xendomain())
+		e820_setup_gap(machine_e820.map, machine_e820.nr_map);
+#else
+	e820_setup_gap(e820.map, e820.nr_map);
+#endif
+
+#ifdef CONFIG_XEN
+	{
+		struct physdev_set_iopl set_iopl;
+
+		set_iopl.iopl = 1;
+		WARN_ON(HYPERVISOR_physdev_op(PHYSDEVOP_set_iopl, &set_iopl));
+
+		if (is_initial_xendomain()) {
+#ifdef CONFIG_VT
+#if defined(CONFIG_VGA_CONSOLE)
+			conswitchp = &vga_con;
+#elif defined(CONFIG_DUMMY_CONSOLE)
+			conswitchp = &dummy_con;
+#endif
+#endif
+		} else {
+#if defined(CONFIG_VT) && defined(CONFIG_DUMMY_CONSOLE)
+			conswitchp = &dummy_con;
+#endif
+                }
+	}
+#else	/* CONFIG_XEN */
+
+#ifdef CONFIG_VT
+#if defined(CONFIG_VGA_CONSOLE)
+	conswitchp = &vga_con;
+#elif defined(CONFIG_DUMMY_CONSOLE)
+	conswitchp = &dummy_con;
+#endif
+#endif
+
+#endif /* !CONFIG_XEN */
+}
+
+#ifdef CONFIG_XEN
+static int
+xen_panic_event(struct notifier_block *this, unsigned long event, void *ptr)
+{
+	HYPERVISOR_shutdown(SHUTDOWN_crash);
+	/* we're never actually going to get here... */
+	return NOTIFY_DONE;
+}
+#endif /* !CONFIG_XEN */
+
+
+static int __cpuinit get_model_name(struct cpuinfo_x86 *c)
+{
+	unsigned int *v;
+
+	if (c->extended_cpuid_level < 0x80000004)
+		return 0;
+
+	v = (unsigned int *) c->x86_model_id;
+	cpuid(0x80000002, &v[0], &v[1], &v[2], &v[3]);
+	cpuid(0x80000003, &v[4], &v[5], &v[6], &v[7]);
+	cpuid(0x80000004, &v[8], &v[9], &v[10], &v[11]);
+	c->x86_model_id[48] = 0;
+	return 1;
+}
+
+
+static void __cpuinit display_cacheinfo(struct cpuinfo_x86 *c)
+{
+	unsigned int n, dummy, eax, ebx, ecx, edx;
+
+	n = c->extended_cpuid_level;
+
+	if (n >= 0x80000005) {
+		cpuid(0x80000005, &dummy, &ebx, &ecx, &edx);
+		printk(KERN_INFO "CPU: L1 I Cache: %dK (%d bytes/line), D cache %dK (%d bytes/line)\n",
+			edx>>24, edx&0xFF, ecx>>24, ecx&0xFF);
+		c->x86_cache_size=(ecx>>24)+(edx>>24);
+		/* On K8 L1 TLB is inclusive, so don't count it */
+		c->x86_tlbsize = 0;
+	}
+
+	if (n >= 0x80000006) {
+		cpuid(0x80000006, &dummy, &ebx, &ecx, &edx);
+		ecx = cpuid_ecx(0x80000006);
+		c->x86_cache_size = ecx >> 16;
+		c->x86_tlbsize += ((ebx >> 16) & 0xfff) + (ebx & 0xfff);
+
+		printk(KERN_INFO "CPU: L2 Cache: %dK (%d bytes/line)\n",
+		c->x86_cache_size, ecx & 0xFF);
+	}
+
+	if (n >= 0x80000007)
+		cpuid(0x80000007, &dummy, &dummy, &dummy, &c->x86_power); 
+	if (n >= 0x80000008) {
+		cpuid(0x80000008, &eax, &dummy, &dummy, &dummy); 
+		c->x86_virt_bits = (eax >> 8) & 0xff;
+		c->x86_phys_bits = eax & 0xff;
+	}
+}
+
+#ifdef CONFIG_NUMA
+static int nearby_node(int apicid)
+{
+	int i;
+	for (i = apicid - 1; i >= 0; i--) {
+		int node = apicid_to_node[i];
+		if (node != NUMA_NO_NODE && node_online(node))
+			return node;
+	}
+	for (i = apicid + 1; i < MAX_LOCAL_APIC; i++) {
+		int node = apicid_to_node[i];
+		if (node != NUMA_NO_NODE && node_online(node))
+			return node;
+	}
+	return first_node(node_online_map); /* Shouldn't happen */
+}
+#endif
+
+/*
+ * On a AMD dual core setup the lower bits of the APIC id distingush the cores.
+ * Assumes number of cores is a power of two.
+ */
+static void __init amd_detect_cmp(struct cpuinfo_x86 *c)
+{
+#ifdef CONFIG_SMP
+	unsigned bits;
+#ifdef CONFIG_NUMA
+	int cpu = smp_processor_id();
+	int node = 0;
+	unsigned apicid = hard_smp_processor_id();
+#endif
+	unsigned ecx = cpuid_ecx(0x80000008);
+
+	c->x86_max_cores = (ecx & 0xff) + 1;
+
+	/* CPU telling us the core id bits shift? */
+	bits = (ecx >> 12) & 0xF;
+
+	/* Otherwise recompute */
+	if (bits == 0) {
+		while ((1 << bits) < c->x86_max_cores)
+			bits++;
+	}
+
+	/* Low order bits define the core id (index of core in socket) */
+	c->cpu_core_id = c->phys_proc_id & ((1 << bits)-1);
+	/* Convert the APIC ID into the socket ID */
+	c->phys_proc_id = phys_pkg_id(bits);
+
+#ifdef CONFIG_NUMA
+  	node = c->phys_proc_id;
+ 	if (apicid_to_node[apicid] != NUMA_NO_NODE)
+ 		node = apicid_to_node[apicid];
+ 	if (!node_online(node)) {
+ 		/* Two possibilities here:
+ 		   - The CPU is missing memory and no node was created.
+ 		   In that case try picking one from a nearby CPU
+ 		   - The APIC IDs differ from the HyperTransport node IDs
+ 		   which the K8 northbridge parsing fills in.
+ 		   Assume they are all increased by a constant offset,
+ 		   but in the same order as the HT nodeids.
+ 		   If that doesn't result in a usable node fall back to the
+ 		   path for the previous case.  */
+ 		int ht_nodeid = apicid - (cpu_data[0].phys_proc_id << bits);
+ 		if (ht_nodeid >= 0 &&
+ 		    apicid_to_node[ht_nodeid] != NUMA_NO_NODE)
+ 			node = apicid_to_node[ht_nodeid];
+ 		/* Pick a nearby node */
+ 		if (!node_online(node))
+ 			node = nearby_node(apicid);
+ 	}
+	numa_set_node(cpu, node);
+
+	printk(KERN_INFO "CPU %d/%x -> Node %d\n", cpu, apicid, node);
+#endif
+#endif
+}
+
+static void __init init_amd(struct cpuinfo_x86 *c)
+{
+	unsigned level;
+
+#ifdef CONFIG_SMP
+	unsigned long value;
+
+	/*
+	 * Disable TLB flush filter by setting HWCR.FFDIS on K8
+	 * bit 6 of msr C001_0015
+ 	 *
+	 * Errata 63 for SH-B3 steppings
+	 * Errata 122 for all steppings (F+ have it disabled by default)
+	 */
+	if (c->x86 == 15) {
+		rdmsrl(MSR_K8_HWCR, value);
+		value |= 1 << 6;
+		wrmsrl(MSR_K8_HWCR, value);
+	}
+#endif
+
+	/* Bit 31 in normal CPUID used for nonstandard 3DNow ID;
+	   3DNow is IDd by bit 31 in extended CPUID (1*32+31) anyway */
+	clear_bit(0*32+31, &c->x86_capability);
+	
+	/* On C+ stepping K8 rep microcode works well for copy/memset */
+	level = cpuid_eax(1);
+	if (c->x86 == 15 && ((level >= 0x0f48 && level < 0x0f50) || level >= 0x0f58))
+		set_bit(X86_FEATURE_REP_GOOD, &c->x86_capability);
+
+	/* Enable workaround for FXSAVE leak */
+	if (c->x86 >= 6)
+		set_bit(X86_FEATURE_FXSAVE_LEAK, &c->x86_capability);
+
+	level = get_model_name(c);
+	if (!level) {
+		switch (c->x86) { 
+		case 15:
+			/* Should distinguish Models here, but this is only
+			   a fallback anyways. */
+			strcpy(c->x86_model_id, "Hammer");
+			break; 
+		} 
+	} 
+	display_cacheinfo(c);
+
+	/* c->x86_power is 8000_0007 edx. Bit 8 is constant TSC */
+	if (c->x86_power & (1<<8))
+		set_bit(X86_FEATURE_CONSTANT_TSC, &c->x86_capability);
+
+	/* Multi core CPU? */
+	if (c->extended_cpuid_level >= 0x80000008)
+		amd_detect_cmp(c);
+
+	/* Fix cpuid4 emulation for more */
+	num_cache_leaves = 3;
+}
+
+static void __cpuinit detect_ht(struct cpuinfo_x86 *c)
+{
+#ifdef CONFIG_SMP
+	u32 	eax, ebx, ecx, edx;
+	int 	index_msb, core_bits;
+
+	cpuid(1, &eax, &ebx, &ecx, &edx);
+
+
+	if (!cpu_has(c, X86_FEATURE_HT))
+		return;
+ 	if (cpu_has(c, X86_FEATURE_CMP_LEGACY))
+		goto out;
+
+	smp_num_siblings = (ebx & 0xff0000) >> 16;
+
+	if (smp_num_siblings == 1) {
+		printk(KERN_INFO  "CPU: Hyper-Threading is disabled\n");
+	} else if (smp_num_siblings > 1 ) {
+
+		if (smp_num_siblings > NR_CPUS) {
+			printk(KERN_WARNING "CPU: Unsupported number of the siblings %d", smp_num_siblings);
+			smp_num_siblings = 1;
+			return;
+		}
+
+		index_msb = get_count_order(smp_num_siblings);
+		c->phys_proc_id = phys_pkg_id(index_msb);
+
+		smp_num_siblings = smp_num_siblings / c->x86_max_cores;
+
+		index_msb = get_count_order(smp_num_siblings) ;
+
+		core_bits = get_count_order(c->x86_max_cores);
+
+		c->cpu_core_id = phys_pkg_id(index_msb) &
+					       ((1 << core_bits) - 1);
+	}
+out:
+	if ((c->x86_max_cores * smp_num_siblings) > 1) {
+		printk(KERN_INFO  "CPU: Physical Processor ID: %d\n", c->phys_proc_id);
+		printk(KERN_INFO  "CPU: Processor Core ID: %d\n", c->cpu_core_id);
+	}
+
+#endif
+}
+
+/*
+ * find out the number of processor cores on the die
+ */
+static int __cpuinit intel_num_cpu_cores(struct cpuinfo_x86 *c)
+{
+	unsigned int eax, t;
+
+	if (c->cpuid_level < 4)
+		return 1;
+
+	cpuid_count(4, 0, &eax, &t, &t, &t);
+
+	if (eax & 0x1f)
+		return ((eax >> 26) + 1);
+	else
+		return 1;
+}
+
+static void srat_detect_node(void)
+{
+#ifdef CONFIG_NUMA
+	unsigned node;
+	int cpu = smp_processor_id();
+	int apicid = hard_smp_processor_id();
+
+	/* Don't do the funky fallback heuristics the AMD version employs
+	   for now. */
+	node = apicid_to_node[apicid];
+	if (node == NUMA_NO_NODE)
+		node = first_node(node_online_map);
+	numa_set_node(cpu, node);
+
+	if (acpi_numa > 0)
+		printk(KERN_INFO "CPU %d/%x -> Node %d\n", cpu, apicid, node);
+#endif
+}
+
+static void __cpuinit init_intel(struct cpuinfo_x86 *c)
+{
+	/* Cache sizes */
+	unsigned n;
+
+	init_intel_cacheinfo(c);
+	if (c->cpuid_level > 9 ) {
+		unsigned eax = cpuid_eax(10);
+		/* Check for version and the number of counters */
+		if ((eax & 0xff) && (((eax>>8) & 0xff) > 1))
+			set_bit(X86_FEATURE_ARCH_PERFMON, &c->x86_capability);
+	}
+
+	n = c->extended_cpuid_level;
+	if (n >= 0x80000008) {
+		unsigned eax = cpuid_eax(0x80000008);
+		c->x86_virt_bits = (eax >> 8) & 0xff;
+		c->x86_phys_bits = eax & 0xff;
+		/* CPUID workaround for Intel 0F34 CPU */
+		if (c->x86_vendor == X86_VENDOR_INTEL &&
+		    c->x86 == 0xF && c->x86_model == 0x3 &&
+		    c->x86_mask == 0x4)
+			c->x86_phys_bits = 36;
+	}
+
+	if (c->x86 == 15)
+		c->x86_cache_alignment = c->x86_clflush_size * 2;
+	if ((c->x86 == 0xf && c->x86_model >= 0x03) ||
+	    (c->x86 == 0x6 && c->x86_model >= 0x0e))
+		set_bit(X86_FEATURE_CONSTANT_TSC, &c->x86_capability);
+	set_bit(X86_FEATURE_SYNC_RDTSC, &c->x86_capability);
+ 	c->x86_max_cores = intel_num_cpu_cores(c);
+
+	srat_detect_node();
+}
+
+static void __cpuinit get_cpu_vendor(struct cpuinfo_x86 *c)
+{
+	char *v = c->x86_vendor_id;
+
+	if (!strcmp(v, "AuthenticAMD"))
+		c->x86_vendor = X86_VENDOR_AMD;
+	else if (!strcmp(v, "GenuineIntel"))
+		c->x86_vendor = X86_VENDOR_INTEL;
+	else
+		c->x86_vendor = X86_VENDOR_UNKNOWN;
+}
+
+struct cpu_model_info {
+	int vendor;
+	int family;
+	char *model_names[16];
+};
+
+/* Do some early cpuid on the boot CPU to get some parameter that are
+   needed before check_bugs. Everything advanced is in identify_cpu
+   below. */
+void __cpuinit early_identify_cpu(struct cpuinfo_x86 *c)
+{
+	u32 tfms;
+
+	c->loops_per_jiffy = loops_per_jiffy;
+	c->x86_cache_size = -1;
+	c->x86_vendor = X86_VENDOR_UNKNOWN;
+	c->x86_model = c->x86_mask = 0;	/* So far unknown... */
+	c->x86_vendor_id[0] = '\0'; /* Unset */
+	c->x86_model_id[0] = '\0';  /* Unset */
+	c->x86_clflush_size = 64;
+	c->x86_cache_alignment = c->x86_clflush_size;
+	c->x86_max_cores = 1;
+	c->extended_cpuid_level = 0;
+	memset(&c->x86_capability, 0, sizeof c->x86_capability);
+
+	/* Get vendor name */
+	cpuid(0x00000000, (unsigned int *)&c->cpuid_level,
+	      (unsigned int *)&c->x86_vendor_id[0],
+	      (unsigned int *)&c->x86_vendor_id[8],
+	      (unsigned int *)&c->x86_vendor_id[4]);
+		
+	get_cpu_vendor(c);
+
+	/* Initialize the standard set of capabilities */
+	/* Note that the vendor-specific code below might override */
+
+	/* Intel-defined flags: level 0x00000001 */
+	if (c->cpuid_level >= 0x00000001) {
+		__u32 misc;
+		cpuid(0x00000001, &tfms, &misc, &c->x86_capability[4],
+		      &c->x86_capability[0]);
+		c->x86 = (tfms >> 8) & 0xf;
+		c->x86_model = (tfms >> 4) & 0xf;
+		c->x86_mask = tfms & 0xf;
+		if (c->x86 == 0xf)
+			c->x86 += (tfms >> 20) & 0xff;
+		if (c->x86 >= 0x6)
+			c->x86_model += ((tfms >> 16) & 0xF) << 4;
+		if (c->x86_capability[0] & (1<<19)) 
+			c->x86_clflush_size = ((misc >> 8) & 0xff) * 8;
+	} else {
+		/* Have CPUID level 0 only - unheard of */
+		c->x86 = 4;
+	}
+
+#ifdef CONFIG_SMP
+	c->phys_proc_id = (cpuid_ebx(1) >> 24) & 0xff;
+#endif
+}
+
+/*
+ * This does the hard work of actually picking apart the CPU stuff...
+ */
+void __cpuinit identify_cpu(struct cpuinfo_x86 *c)
+{
+	int i;
+	u32 xlvl;
+
+	early_identify_cpu(c);
+
+	/* AMD-defined flags: level 0x80000001 */
+	xlvl = cpuid_eax(0x80000000);
+	c->extended_cpuid_level = xlvl;
+	if ((xlvl & 0xffff0000) == 0x80000000) {
+		if (xlvl >= 0x80000001) {
+			c->x86_capability[1] = cpuid_edx(0x80000001);
+			c->x86_capability[6] = cpuid_ecx(0x80000001);
+		}
+		if (xlvl >= 0x80000004)
+			get_model_name(c); /* Default name */
+	}
+
+	/* Transmeta-defined flags: level 0x80860001 */
+	xlvl = cpuid_eax(0x80860000);
+	if ((xlvl & 0xffff0000) == 0x80860000) {
+		/* Don't set x86_cpuid_level here for now to not confuse. */
+		if (xlvl >= 0x80860001)
+			c->x86_capability[2] = cpuid_edx(0x80860001);
+	}
+
+	c->apicid = phys_pkg_id(0);
+
+	/*
+	 * Vendor-specific initialization.  In this section we
+	 * canonicalize the feature flags, meaning if there are
+	 * features a certain CPU supports which CPUID doesn't
+	 * tell us, CPUID claiming incorrect flags, or other bugs,
+	 * we handle them here.
+	 *
+	 * At the end of this section, c->x86_capability better
+	 * indicate the features this CPU genuinely supports!
+	 */
+	switch (c->x86_vendor) {
+	case X86_VENDOR_AMD:
+		init_amd(c);
+		break;
+
+	case X86_VENDOR_INTEL:
+		init_intel(c);
+		break;
+
+	case X86_VENDOR_UNKNOWN:
+	default:
+		display_cacheinfo(c);
+		break;
+	}
+
+	select_idle_routine(c);
+	detect_ht(c); 
+
+	/*
+	 * On SMP, boot_cpu_data holds the common feature set between
+	 * all CPUs; so make sure that we indicate which features are
+	 * common between the CPUs.  The first time this routine gets
+	 * executed, c == &boot_cpu_data.
+	 */
+	if (c != &boot_cpu_data) {
+		/* AND the already accumulated flags with these */
+		for (i = 0 ; i < NCAPINTS ; i++)
+			boot_cpu_data.x86_capability[i] &= c->x86_capability[i];
+	}
+
+#ifdef CONFIG_X86_MCE
+	mcheck_init(c);
+#endif
+	if (c == &boot_cpu_data)
+		mtrr_bp_init();
+	else
+		mtrr_ap_init();
+#ifdef CONFIG_NUMA
+	numa_add_cpu(smp_processor_id());
+#endif
+}
+ 
+
+void __cpuinit print_cpu_info(struct cpuinfo_x86 *c)
+{
+	if (c->x86_model_id[0])
+		printk("%s", c->x86_model_id);
+
+	if (c->x86_mask || c->cpuid_level >= 0) 
+		printk(" stepping %02x\n", c->x86_mask);
+	else
+		printk("\n");
+}
+
+/*
+ *	Get CPU information for use by the procfs.
+ */
+
+static int show_cpuinfo(struct seq_file *m, void *v)
+{
+	struct cpuinfo_x86 *c = v;
+
+	/* 
+	 * These flag bits must match the definitions in <asm/cpufeature.h>.
+	 * NULL means this bit is undefined or reserved; either way it doesn't
+	 * have meaning as far as Linux is concerned.  Note that it's important
+	 * to realize there is a difference between this table and CPUID -- if
+	 * applications want to get the raw CPUID data, they should access
+	 * /dev/cpu/<cpu_nr>/cpuid instead.
+	 */
+	static char *x86_cap_flags[] = {
+		/* Intel-defined */
+	        "fpu", "vme", "de", "pse", "tsc", "msr", "pae", "mce",
+	        "cx8", "apic", NULL, "sep", "mtrr", "pge", "mca", "cmov",
+	        "pat", "pse36", "pn", "clflush", NULL, "dts", "acpi", "mmx",
+	        "fxsr", "sse", "sse2", "ss", "ht", "tm", "ia64", NULL,
+
+		/* AMD-defined */
+		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+		NULL, NULL, NULL, "syscall", NULL, NULL, NULL, NULL,
+		NULL, NULL, NULL, NULL, "nx", NULL, "mmxext", NULL,
+		NULL, "fxsr_opt", NULL, "rdtscp", NULL, "lm", "3dnowext", "3dnow",
+
+		/* Transmeta-defined */
+		"recovery", "longrun", NULL, "lrti", NULL, NULL, NULL, NULL,
+		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+
+		/* Other (Linux-defined) */
+		"cxmmx", NULL, "cyrix_arr", "centaur_mcr", NULL,
+		"constant_tsc", NULL, NULL,
+		"up", NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+
+		/* Intel-defined (#2) */
+		"pni", NULL, NULL, "monitor", "ds_cpl", "vmx", "smx", "est",
+		"tm2", NULL, "cid", NULL, NULL, "cx16", "xtpr", NULL,
+		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+
+		/* VIA/Cyrix/Centaur-defined */
+		NULL, NULL, "rng", "rng_en", NULL, NULL, "ace", "ace_en",
+		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+
+		/* AMD-defined (#2) */
+		"lahf_lm", "cmp_legacy", "svm", NULL, "cr8_legacy", NULL, NULL, NULL,
+		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+		NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+	};
+	static char *x86_power_flags[] = { 
+		"ts",	/* temperature sensor */
+		"fid",  /* frequency id control */
+		"vid",  /* voltage id control */
+		"ttp",  /* thermal trip */
+		"tm",
+		"stc",
+		NULL,
+		/* nothing */	/* constant_tsc - moved to flags */
+	};
+
+
+#ifdef CONFIG_SMP
+	if (!cpu_online(c-cpu_data))
+		return 0;
+#endif
+
+	seq_printf(m,"processor\t: %u\n"
+		     "vendor_id\t: %s\n"
+		     "cpu family\t: %d\n"
+		     "model\t\t: %d\n"
+		     "model name\t: %s\n",
+		     (unsigned)(c-cpu_data),
+		     c->x86_vendor_id[0] ? c->x86_vendor_id : "unknown",
+		     c->x86,
+		     (int)c->x86_model,
+		     c->x86_model_id[0] ? c->x86_model_id : "unknown");
+	
+	if (c->x86_mask || c->cpuid_level >= 0)
+		seq_printf(m, "stepping\t: %d\n", c->x86_mask);
+	else
+		seq_printf(m, "stepping\t: unknown\n");
+	
+	if (cpu_has(c,X86_FEATURE_TSC)) {
+		unsigned int freq = cpufreq_quick_get((unsigned)(c-cpu_data));
+		if (!freq)
+			freq = cpu_khz;
+		seq_printf(m, "cpu MHz\t\t: %u.%03u\n",
+			     freq / 1000, (freq % 1000));
+	}
+
+	/* Cache size */
+	if (c->x86_cache_size >= 0) 
+		seq_printf(m, "cache size\t: %d KB\n", c->x86_cache_size);
+	
+#ifdef CONFIG_SMP
+	if (smp_num_siblings * c->x86_max_cores > 1) {
+		int cpu = c - cpu_data;
+		seq_printf(m, "physical id\t: %d\n", c->phys_proc_id);
+		seq_printf(m, "siblings\t: %d\n", cpus_weight(cpu_core_map[cpu]));
+		seq_printf(m, "core id\t\t: %d\n", c->cpu_core_id);
+		seq_printf(m, "cpu cores\t: %d\n", c->booted_cores);
+	}
+#endif	
+
+	seq_printf(m,
+	        "fpu\t\t: yes\n"
+	        "fpu_exception\t: yes\n"
+	        "cpuid level\t: %d\n"
+	        "wp\t\t: yes\n"
+	        "flags\t\t:",
+		   c->cpuid_level);
+
+	{ 
+		int i; 
+		for ( i = 0 ; i < 32*NCAPINTS ; i++ )
+			if (cpu_has(c, i) && x86_cap_flags[i] != NULL)
+				seq_printf(m, " %s", x86_cap_flags[i]);
+	}
+		
+	seq_printf(m, "\nbogomips\t: %lu.%02lu\n",
+		   c->loops_per_jiffy/(500000/HZ),
+		   (c->loops_per_jiffy/(5000/HZ)) % 100);
+
+	if (c->x86_tlbsize > 0) 
+		seq_printf(m, "TLB size\t: %d 4K pages\n", c->x86_tlbsize);
+	seq_printf(m, "clflush size\t: %d\n", c->x86_clflush_size);
+	seq_printf(m, "cache_alignment\t: %d\n", c->x86_cache_alignment);
+
+	seq_printf(m, "address sizes\t: %u bits physical, %u bits virtual\n", 
+		   c->x86_phys_bits, c->x86_virt_bits);
+
+	seq_printf(m, "power management:");
+	{
+		unsigned i;
+		for (i = 0; i < 32; i++) 
+			if (c->x86_power & (1 << i)) {
+				if (i < ARRAY_SIZE(x86_power_flags) &&
+					x86_power_flags[i])
+					seq_printf(m, "%s%s",
+						x86_power_flags[i][0]?" ":"",
+						x86_power_flags[i]);
+				else
+					seq_printf(m, " [%d]", i);
+			}
+	}
+
+	seq_printf(m, "\n\n");
+
+	return 0;
+}
+
+static void *c_start(struct seq_file *m, loff_t *pos)
+{
+	return *pos < NR_CPUS ? cpu_data + *pos : NULL;
+}
+
+static void *c_next(struct seq_file *m, void *v, loff_t *pos)
+{
+	++*pos;
+	return c_start(m, pos);
+}
+
+static void c_stop(struct seq_file *m, void *v)
+{
+}
+
+struct seq_operations cpuinfo_op = {
+	.start =c_start,
+	.next =	c_next,
+	.stop =	c_stop,
+	.show =	show_cpuinfo,
+};
+
+#if defined(CONFIG_INPUT_PCSPKR) || defined(CONFIG_INPUT_PCSPKR_MODULE)
+#include <linux/platform_device.h>
+static __init int add_pcspkr(void)
+{
+	struct platform_device *pd;
+	int ret;
+
+	if (!is_initial_xendomain())
+		return 0;
+
+	pd = platform_device_alloc("pcspkr", -1);
+	if (!pd)
+		return -ENOMEM;
+
+	ret = platform_device_add(pd);
+	if (ret)
+		platform_device_put(pd);
+
+	return ret;
+}
+device_initcall(add_pcspkr);
+#endif
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/signal.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/signal.c
--- linux-2.6.18.8/arch/x86_64/kernel/signal.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/signal.c	2008-08-21 11:36:07.000000000 +0200
@@ -38,37 +38,6 @@ int ia32_setup_frame(int sig, struct k_s
             sigset_t *set, struct pt_regs * regs); 
 
 asmlinkage long
-sys_rt_sigsuspend(sigset_t __user *unewset, size_t sigsetsize, struct pt_regs *regs)
-{
-	sigset_t saveset, newset;
-
-	/* XXX: Don't preclude handling different sized sigset_t's.  */
-	if (sigsetsize != sizeof(sigset_t))
-		return -EINVAL;
-
-	if (copy_from_user(&newset, unewset, sizeof(newset)))
-		return -EFAULT;
-	sigdelsetmask(&newset, ~_BLOCKABLE);
-
-	spin_lock_irq(&current->sighand->siglock);
-	saveset = current->blocked;
-	current->blocked = newset;
-	recalc_sigpending();
-	spin_unlock_irq(&current->sighand->siglock);
-#ifdef DEBUG_SIG
-	printk("rt_sigsuspend savset(%lx) newset(%lx) regs(%p) rip(%lx)\n",
-		saveset, newset, regs, regs->rip);
-#endif 
-	regs->rax = -EINTR;
-	while (1) {
-		current->state = TASK_INTERRUPTIBLE;
-		schedule();
-		if (do_signal(regs, &saveset))
-			return -EINTR;
-	}
-}
-
-asmlinkage long
 sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss,
 		struct pt_regs *regs)
 {
@@ -341,11 +310,11 @@ static int setup_rt_frame(int sig, struc
 		current->comm, current->pid, frame, regs->rip, frame->pretcode);
 #endif
 
-	return 1;
+	return 0;
 
 give_sigsegv:
 	force_sigsegv(sig, current);
-	return 0;
+	return -EFAULT;
 }
 
 /*
@@ -408,7 +377,7 @@ handle_signal(unsigned long sig, siginfo
 #endif
 	ret = setup_rt_frame(sig, ka, info, oldset, regs);
 
-	if (ret) {
+	if (ret == 0) {
 		spin_lock_irq(&current->sighand->siglock);
 		sigorsets(&current->blocked,&current->blocked,&ka->sa.sa_mask);
 		if (!(ka->sa.sa_flags & SA_NODEFER))
@@ -425,11 +394,12 @@ handle_signal(unsigned long sig, siginfo
  * want to handle. Thus you cannot kill init even with a SIGKILL even by
  * mistake.
  */
-int do_signal(struct pt_regs *regs, sigset_t *oldset)
+static void do_signal(struct pt_regs *regs)
 {
 	struct k_sigaction ka;
 	siginfo_t info;
 	int signr;
+	sigset_t *oldset;
 
 	/*
 	 * We want the common case to go fast, which
@@ -438,9 +408,11 @@ int do_signal(struct pt_regs *regs, sigs
 	 * if so.
 	 */
 	if (!user_mode(regs))
-		return 1;
+		return;
 
-	if (!oldset)
+	if (test_thread_flag(TIF_RESTORE_SIGMASK))
+		oldset = &current->saved_sigmask;
+	else
 		oldset = &current->blocked;
 
 	signr = get_signal_to_deliver(&info, &ka, regs, NULL);
@@ -454,30 +426,46 @@ int do_signal(struct pt_regs *regs, sigs
 			set_debugreg(current->thread.debugreg7, 7);
 
 		/* Whee!  Actually deliver the signal.  */
-		return handle_signal(signr, &info, &ka, oldset, regs);
+		if (handle_signal(signr, &info, &ka, oldset, regs) == 0) {
+			/* a signal was successfully delivered; the saved
+			 * sigmask will have been stored in the signal frame,
+			 * and will be restored by sigreturn, so we can simply
+			 * clear the TIF_RESTORE_SIGMASK flag */
+			clear_thread_flag(TIF_RESTORE_SIGMASK);
+		}
+		return;
 	}
 
 	/* Did we come from a system call? */
 	if ((long)regs->orig_rax >= 0) {
 		/* Restart the system call - no handlers present */
 		long res = regs->rax;
-		if (res == -ERESTARTNOHAND ||
-		    res == -ERESTARTSYS ||
-		    res == -ERESTARTNOINTR) {
+		switch (res) {
+		case -ERESTARTNOHAND:
+		case -ERESTARTSYS:
+		case -ERESTARTNOINTR:
 			regs->rax = regs->orig_rax;
 			regs->rip -= 2;
-		}
-		if (regs->rax == (unsigned long)-ERESTART_RESTARTBLOCK) {
+			break;
+		case -ERESTART_RESTARTBLOCK:
 			regs->rax = test_thread_flag(TIF_IA32) ?
 					__NR_ia32_restart_syscall :
 					__NR_restart_syscall;
 			regs->rip -= 2;
+			break;
 		}
 	}
-	return 0;
+
+	/* if there's no signal to deliver, we just put the saved sigmask
+	   back. */
+	if (test_thread_flag(TIF_RESTORE_SIGMASK)) {
+		clear_thread_flag(TIF_RESTORE_SIGMASK);
+		sigprocmask(SIG_SETMASK, &current->saved_sigmask, NULL);
+	}
 }
 
-void do_notify_resume(struct pt_regs *regs, sigset_t *oldset, __u32 thread_info_flags)
+void
+do_notify_resume(struct pt_regs *regs, void *unused, __u32 thread_info_flags)
 {
 #ifdef DEBUG_SIG
 	printk("do_notify_resume flags:%x rip:%lx rsp:%lx caller:%lx pending:%lx\n",
@@ -491,8 +479,8 @@ void do_notify_resume(struct pt_regs *re
 	}
 
 	/* deal with pending signal delivery */
-	if (thread_info_flags & _TIF_SIGPENDING)
-		do_signal(regs,oldset);
+	if (thread_info_flags & (_TIF_SIGPENDING|_TIF_RESTORE_SIGMASK))
+		do_signal(regs);
 }
 
 void signal_fault(struct pt_regs *regs, void __user *frame, char *where)
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/smp-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/smp-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/smp-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/smp-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,575 @@
+/*
+ *	Intel SMP support routines.
+ *
+ *	(c) 1995 Alan Cox, Building #3 <alan@redhat.com>
+ *	(c) 1998-99, 2000 Ingo Molnar <mingo@redhat.com>
+ *      (c) 2002,2003 Andi Kleen, SuSE Labs.
+ *
+ *	This code is released under the GNU General Public License version 2 or
+ *	later.
+ */
+
+#include <linux/init.h>
+
+#include <linux/mm.h>
+#include <linux/delay.h>
+#include <linux/spinlock.h>
+#include <linux/smp_lock.h>
+#include <linux/smp.h>
+#include <linux/kernel_stat.h>
+#include <linux/mc146818rtc.h>
+#include <linux/interrupt.h>
+
+#include <asm/mtrr.h>
+#include <asm/pgalloc.h>
+#include <asm/tlbflush.h>
+#include <asm/mach_apic.h>
+#include <asm/mmu_context.h>
+#include <asm/proto.h>
+#include <asm/apicdef.h>
+#include <asm/idle.h>
+#ifdef CONFIG_XEN
+#include <xen/evtchn.h>
+#endif
+
+#ifndef CONFIG_XEN
+/*
+ *	Smarter SMP flushing macros. 
+ *		c/o Linus Torvalds.
+ *
+ *	These mean you can really definitely utterly forget about
+ *	writing to user space from interrupts. (Its not allowed anyway).
+ *
+ *	Optimizations Manfred Spraul <manfred@colorfullife.com>
+ *
+ * 	More scalable flush, from Andi Kleen
+ *
+ * 	To avoid global state use 8 different call vectors.
+ * 	Each CPU uses a specific vector to trigger flushes on other
+ * 	CPUs. Depending on the received vector the target CPUs look into
+ *	the right per cpu variable for the flush data.
+ *
+ * 	With more than 8 CPUs they are hashed to the 8 available
+ * 	vectors. The limited global vector space forces us to this right now.
+ *	In future when interrupts are split into per CPU domains this could be
+ *	fixed, at the cost of triggering multiple IPIs in some cases.
+ */
+
+union smp_flush_state {
+	struct {
+		cpumask_t flush_cpumask;
+		struct mm_struct *flush_mm;
+		unsigned long flush_va;
+#define FLUSH_ALL	-1ULL
+		spinlock_t tlbstate_lock;
+	};
+	char pad[SMP_CACHE_BYTES];
+} ____cacheline_aligned;
+
+/* State is put into the per CPU data section, but padded
+   to a full cache line because other CPUs can access it and we don't
+   want false sharing in the per cpu data segment. */
+static DEFINE_PER_CPU(union smp_flush_state, flush_state);
+
+/*
+ * We cannot call mmdrop() because we are in interrupt context, 
+ * instead update mm->cpu_vm_mask.
+ */
+static inline void leave_mm(unsigned long cpu)
+{
+	if (read_pda(mmu_state) == TLBSTATE_OK)
+		BUG();
+	cpu_clear(cpu, read_pda(active_mm)->cpu_vm_mask);
+	load_cr3(swapper_pg_dir);
+}
+
+/*
+ *
+ * The flush IPI assumes that a thread switch happens in this order:
+ * [cpu0: the cpu that switches]
+ * 1) switch_mm() either 1a) or 1b)
+ * 1a) thread switch to a different mm
+ * 1a1) cpu_clear(cpu, old_mm->cpu_vm_mask);
+ * 	Stop ipi delivery for the old mm. This is not synchronized with
+ * 	the other cpus, but smp_invalidate_interrupt ignore flush ipis
+ * 	for the wrong mm, and in the worst case we perform a superfluous
+ * 	tlb flush.
+ * 1a2) set cpu mmu_state to TLBSTATE_OK
+ * 	Now the smp_invalidate_interrupt won't call leave_mm if cpu0
+ *	was in lazy tlb mode.
+ * 1a3) update cpu active_mm
+ * 	Now cpu0 accepts tlb flushes for the new mm.
+ * 1a4) cpu_set(cpu, new_mm->cpu_vm_mask);
+ * 	Now the other cpus will send tlb flush ipis.
+ * 1a4) change cr3.
+ * 1b) thread switch without mm change
+ *	cpu active_mm is correct, cpu0 already handles
+ *	flush ipis.
+ * 1b1) set cpu mmu_state to TLBSTATE_OK
+ * 1b2) test_and_set the cpu bit in cpu_vm_mask.
+ * 	Atomically set the bit [other cpus will start sending flush ipis],
+ * 	and test the bit.
+ * 1b3) if the bit was 0: leave_mm was called, flush the tlb.
+ * 2) switch %%esp, ie current
+ *
+ * The interrupt must handle 2 special cases:
+ * - cr3 is changed before %%esp, ie. it cannot use current->{active_,}mm.
+ * - the cpu performs speculative tlb reads, i.e. even if the cpu only
+ *   runs in kernel space, the cpu could load tlb entries for user space
+ *   pages.
+ *
+ * The good news is that cpu mmu_state is local to each cpu, no
+ * write/read ordering problems.
+ */
+
+/*
+ * TLB flush IPI:
+ *
+ * 1) Flush the tlb entries if the cpu uses the mm that's being flushed.
+ * 2) Leave the mm if we are in the lazy tlb mode.
+ *
+ * Interrupts are disabled.
+ */
+
+asmlinkage void smp_invalidate_interrupt(struct pt_regs *regs)
+{
+	int cpu;
+	int sender;
+	union smp_flush_state *f;
+
+	cpu = smp_processor_id();
+	/*
+	 * orig_rax contains the negated interrupt vector.
+	 * Use that to determine where the sender put the data.
+	 */
+	sender = ~regs->orig_rax - INVALIDATE_TLB_VECTOR_START;
+	f = &per_cpu(flush_state, sender);
+
+	if (!cpu_isset(cpu, f->flush_cpumask))
+		goto out;
+		/* 
+		 * This was a BUG() but until someone can quote me the
+		 * line from the intel manual that guarantees an IPI to
+		 * multiple CPUs is retried _only_ on the erroring CPUs
+		 * its staying as a return
+		 *
+		 * BUG();
+		 */
+		 
+	if (f->flush_mm == read_pda(active_mm)) {
+		if (read_pda(mmu_state) == TLBSTATE_OK) {
+			if (f->flush_va == FLUSH_ALL)
+				local_flush_tlb();
+			else
+				__flush_tlb_one(f->flush_va);
+		} else
+			leave_mm(cpu);
+	}
+out:
+	ack_APIC_irq();
+	cpu_clear(cpu, f->flush_cpumask);
+}
+
+static void flush_tlb_others(cpumask_t cpumask, struct mm_struct *mm,
+						unsigned long va)
+{
+	int sender;
+	union smp_flush_state *f;
+
+	/* Caller has disabled preemption */
+	sender = smp_processor_id() % NUM_INVALIDATE_TLB_VECTORS;
+	f = &per_cpu(flush_state, sender);
+
+	/* Could avoid this lock when
+	   num_online_cpus() <= NUM_INVALIDATE_TLB_VECTORS, but it is
+	   probably not worth checking this for a cache-hot lock. */
+	spin_lock(&f->tlbstate_lock);
+
+	f->flush_mm = mm;
+	f->flush_va = va;
+	cpus_or(f->flush_cpumask, cpumask, f->flush_cpumask);
+
+	/*
+	 * We have to send the IPI only to
+	 * CPUs affected.
+	 */
+	send_IPI_mask(cpumask, INVALIDATE_TLB_VECTOR_START + sender);
+
+	while (!cpus_empty(f->flush_cpumask))
+		cpu_relax();
+
+	f->flush_mm = NULL;
+	f->flush_va = 0;
+	spin_unlock(&f->tlbstate_lock);
+}
+
+int __cpuinit init_smp_flush(void)
+{
+	int i;
+	for_each_cpu_mask(i, cpu_possible_map) {
+		spin_lock_init(&per_cpu(flush_state, i).tlbstate_lock);
+	}
+	return 0;
+}
+
+core_initcall(init_smp_flush);
+	
+void flush_tlb_current_task(void)
+{
+	struct mm_struct *mm = current->mm;
+	cpumask_t cpu_mask;
+
+	preempt_disable();
+	cpu_mask = mm->cpu_vm_mask;
+	cpu_clear(smp_processor_id(), cpu_mask);
+
+	local_flush_tlb();
+	if (!cpus_empty(cpu_mask))
+		flush_tlb_others(cpu_mask, mm, FLUSH_ALL);
+	preempt_enable();
+}
+EXPORT_SYMBOL(flush_tlb_current_task);
+
+void flush_tlb_mm (struct mm_struct * mm)
+{
+	cpumask_t cpu_mask;
+
+	preempt_disable();
+	cpu_mask = mm->cpu_vm_mask;
+	cpu_clear(smp_processor_id(), cpu_mask);
+
+	if (current->active_mm == mm) {
+		if (current->mm)
+			local_flush_tlb();
+		else
+			leave_mm(smp_processor_id());
+	}
+	if (!cpus_empty(cpu_mask))
+		flush_tlb_others(cpu_mask, mm, FLUSH_ALL);
+
+	preempt_enable();
+}
+EXPORT_SYMBOL(flush_tlb_mm);
+
+void flush_tlb_page(struct vm_area_struct * vma, unsigned long va)
+{
+	struct mm_struct *mm = vma->vm_mm;
+	cpumask_t cpu_mask;
+
+	preempt_disable();
+	cpu_mask = mm->cpu_vm_mask;
+	cpu_clear(smp_processor_id(), cpu_mask);
+
+	if (current->active_mm == mm) {
+		if(current->mm)
+			__flush_tlb_one(va);
+		 else
+		 	leave_mm(smp_processor_id());
+	}
+
+	if (!cpus_empty(cpu_mask))
+		flush_tlb_others(cpu_mask, mm, va);
+
+	preempt_enable();
+}
+EXPORT_SYMBOL(flush_tlb_page);
+
+static void do_flush_tlb_all(void* info)
+{
+	unsigned long cpu = smp_processor_id();
+
+	__flush_tlb_all();
+	if (read_pda(mmu_state) == TLBSTATE_LAZY)
+		leave_mm(cpu);
+}
+
+void flush_tlb_all(void)
+{
+	on_each_cpu(do_flush_tlb_all, NULL, 1, 1);
+}
+#endif /* Xen */
+
+/*
+ * this function sends a 'reschedule' IPI to another CPU.
+ * it goes straight through and wastes no time serializing
+ * anything. Worst case is that we lose a reschedule ...
+ */
+
+void smp_send_reschedule(int cpu)
+{
+	send_IPI_mask(cpumask_of_cpu(cpu), RESCHEDULE_VECTOR);
+}
+
+/*
+ * Structure and data for smp_call_function(). This is designed to minimise
+ * static memory requirements. It also looks cleaner.
+ */
+static DEFINE_SPINLOCK(call_lock);
+
+struct call_data_struct {
+	void (*func) (void *info);
+	void *info;
+	atomic_t started;
+	atomic_t finished;
+	int wait;
+};
+
+static struct call_data_struct * call_data;
+
+void lock_ipi_call_lock(void)
+{
+	spin_lock_irq(&call_lock);
+}
+
+void unlock_ipi_call_lock(void)
+{
+	spin_unlock_irq(&call_lock);
+}
+
+/*
+ * this function sends a 'generic call function' IPI to one other CPU
+ * in the system.
+ *
+ * cpu is a standard Linux logical CPU number.
+ */
+static void
+__smp_call_function_single(int cpu, void (*func) (void *info), void *info,
+				int nonatomic, int wait)
+{
+	struct call_data_struct data;
+	int cpus = 1;
+
+	data.func = func;
+	data.info = info;
+	atomic_set(&data.started, 0);
+	data.wait = wait;
+	if (wait)
+		atomic_set(&data.finished, 0);
+
+	call_data = &data;
+	wmb();
+	/* Send a message to all other CPUs and wait for them to respond */
+	send_IPI_mask(cpumask_of_cpu(cpu), CALL_FUNCTION_VECTOR);
+
+	/* Wait for response */
+	while (atomic_read(&data.started) != cpus)
+		cpu_relax();
+
+	if (!wait)
+		return;
+
+	while (atomic_read(&data.finished) != cpus)
+		cpu_relax();
+}
+
+/*
+ * smp_call_function_single - Run a function on another CPU
+ * @func: The function to run. This must be fast and non-blocking.
+ * @info: An arbitrary pointer to pass to the function.
+ * @nonatomic: Currently unused.
+ * @wait: If true, wait until function has completed on other CPUs.
+ *
+ * Retrurns 0 on success, else a negative status code.
+ *
+ * Does not return until the remote CPU is nearly ready to execute <func>
+ * or is or has executed.
+ */
+
+int smp_call_function_single (int cpu, void (*func) (void *info), void *info,
+	int nonatomic, int wait)
+{
+	/* prevent preemption and reschedule on another processor */
+	int me = get_cpu();
+	if (cpu == me) {
+		WARN_ON(1);
+		put_cpu();
+		return -EBUSY;
+	}
+	spin_lock_bh(&call_lock);
+	__smp_call_function_single(cpu, func, info, nonatomic, wait);
+	spin_unlock_bh(&call_lock);
+	put_cpu();
+	return 0;
+}
+
+/*
+ * this function sends a 'generic call function' IPI to all other CPUs
+ * in the system.
+ */
+static void __smp_call_function (void (*func) (void *info), void *info,
+				int nonatomic, int wait)
+{
+	struct call_data_struct data;
+	int cpus = num_online_cpus()-1;
+
+	if (!cpus)
+		return;
+
+	data.func = func;
+	data.info = info;
+	atomic_set(&data.started, 0);
+	data.wait = wait;
+	if (wait)
+		atomic_set(&data.finished, 0);
+
+	call_data = &data;
+	wmb();
+	/* Send a message to all other CPUs and wait for them to respond */
+	send_IPI_allbutself(CALL_FUNCTION_VECTOR);
+
+	/* Wait for response */
+	while (atomic_read(&data.started) != cpus)
+		cpu_relax();
+
+	if (!wait)
+		return;
+
+	while (atomic_read(&data.finished) != cpus)
+		cpu_relax();
+}
+
+/*
+ * smp_call_function - run a function on all other CPUs.
+ * @func: The function to run. This must be fast and non-blocking.
+ * @info: An arbitrary pointer to pass to the function.
+ * @nonatomic: currently unused.
+ * @wait: If true, wait (atomically) until function has completed on other
+ *        CPUs.
+ *
+ * Returns 0 on success, else a negative status code. Does not return until
+ * remote CPUs are nearly ready to execute func or are or have executed.
+ *
+ * You must not call this function with disabled interrupts or from a
+ * hardware interrupt handler or from a bottom half handler.
+ * Actually there are a few legal cases, like panic.
+ */
+int smp_call_function (void (*func) (void *info), void *info, int nonatomic,
+			int wait)
+{
+	spin_lock(&call_lock);
+	__smp_call_function(func,info,nonatomic,wait);
+	spin_unlock(&call_lock);
+	return 0;
+}
+EXPORT_SYMBOL(smp_call_function);
+
+void smp_stop_cpu(void)
+{
+	unsigned long flags;
+	/*
+	 * Remove this CPU:
+	 */
+	cpu_clear(smp_processor_id(), cpu_online_map);
+	local_irq_save(flags);
+	disable_all_local_evtchn();
+	local_irq_restore(flags); 
+}
+
+static void smp_really_stop_cpu(void *dummy)
+{
+	smp_stop_cpu(); 
+	for (;;) 
+		halt();
+} 
+
+void smp_send_stop(void)
+{
+	int nolock = 0;
+#ifndef CONFIG_XEN
+	if (reboot_force)
+		return;
+#endif
+	/* Don't deadlock on the call lock in panic */
+	if (!spin_trylock(&call_lock)) {
+		/* ignore locking because we have panicked anyways */
+		nolock = 1;
+	}
+	__smp_call_function(smp_really_stop_cpu, NULL, 0, 0);
+	if (!nolock)
+		spin_unlock(&call_lock);
+
+	local_irq_disable();
+	disable_all_local_evtchn();
+	local_irq_enable();
+}
+
+/*
+ * Reschedule call back. Nothing to do,
+ * all the work is done automatically when
+ * we return from the interrupt.
+ */
+#ifndef CONFIG_XEN
+asmlinkage void smp_reschedule_interrupt(void)
+#else
+asmlinkage irqreturn_t smp_reschedule_interrupt(void)
+#endif
+{
+#ifndef CONFIG_XEN
+	ack_APIC_irq();
+#else
+	return IRQ_HANDLED;
+#endif
+}
+
+#ifndef CONFIG_XEN
+asmlinkage void smp_call_function_interrupt(void)
+#else
+asmlinkage irqreturn_t smp_call_function_interrupt(void)
+#endif
+{
+	void (*func) (void *info) = call_data->func;
+	void *info = call_data->info;
+	int wait = call_data->wait;
+
+#ifndef CONFIG_XEN
+	ack_APIC_irq();
+#endif
+	/*
+	 * Notify initiating CPU that I've grabbed the data and am
+	 * about to execute the function
+	 */
+	mb();
+	atomic_inc(&call_data->started);
+	/*
+	 * At this point the info structure may be out of scope unless wait==1
+	 */
+	exit_idle();
+	irq_enter();
+	(*func)(info);
+	irq_exit();
+	if (wait) {
+		mb();
+		atomic_inc(&call_data->finished);
+	}
+#ifdef CONFIG_XEN
+	return IRQ_HANDLED;
+#endif
+}
+
+int safe_smp_processor_id(void)
+{
+#ifdef CONFIG_XEN
+	return smp_processor_id();
+#else
+	unsigned apicid, i;
+
+	if (disable_apic)
+		return 0;
+
+	apicid = hard_smp_processor_id();
+	if (apicid < NR_CPUS && x86_cpu_to_apicid[apicid] == apicid)
+		return apicid;
+
+	for (i = 0; i < NR_CPUS; ++i) {
+		if (x86_cpu_to_apicid[i] == apicid)
+			return i;
+	}
+
+	/* No entries in x86_cpu_to_apicid?  Either no MPS|ACPI,
+	 * or called too early.  Either way, we must be CPU 0. */
+      	if (x86_cpu_to_apicid[0] == BAD_APICID)
+		return 0;
+
+	return 0; /* Should not happen */
+#endif
+}
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/suspend.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/suspend.c
--- linux-2.6.18.8/arch/x86_64/kernel/suspend.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/suspend.c	2008-08-21 11:36:07.000000000 +0200
@@ -114,12 +114,14 @@ void restore_processor_state(void)
 
 void fix_processor_context(void)
 {
+#ifndef CONFIG_X86_NO_TSS    
 	int cpu = smp_processor_id();
 	struct tss_struct *t = &per_cpu(init_tss, cpu);
 
 	set_tss_desc(cpu,t);	/* This just modifies memory; should not be neccessary. But... This is neccessary, because 386 hardware has concept of busy TSS or some similar stupidity. */
 
 	cpu_gdt(cpu)[GDT_ENTRY_TSS].type = 9;
+#endif
 
 	syscall_init();                         /* This sets MSR_*STAR and related */
 	load_TR_desc();				/* This does ltr */
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/traps-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/traps-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/traps-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/traps-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1173 @@
+/*
+ *  linux/arch/x86-64/traps.c
+ *
+ *  Copyright (C) 1991, 1992  Linus Torvalds
+ *  Copyright (C) 2000, 2001, 2002 Andi Kleen, SuSE Labs
+ *
+ *  Pentium III FXSR, SSE support
+ *	Gareth Hughes <gareth@valinux.com>, May 2000
+ */
+
+/*
+ * 'Traps.c' handles hardware traps and faults after we have saved some
+ * state in 'entry.S'.
+ */
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/string.h>
+#include <linux/errno.h>
+#include <linux/ptrace.h>
+#include <linux/timer.h>
+#include <linux/mm.h>
+#include <linux/init.h>
+#include <linux/delay.h>
+#include <linux/spinlock.h>
+#include <linux/interrupt.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/nmi.h>
+#include <linux/kprobes.h>
+#include <linux/kexec.h>
+#include <linux/unwind.h>
+
+#include <asm/system.h>
+#include <asm/uaccess.h>
+#include <asm/io.h>
+#include <asm/atomic.h>
+#include <asm/debugreg.h>
+#include <asm/desc.h>
+#include <asm/i387.h>
+#include <asm/kdebug.h>
+#include <asm/processor.h>
+#include <asm/unwind.h>
+#include <asm/smp.h>
+#include <asm/pgalloc.h>
+#include <asm/pda.h>
+#include <asm/proto.h>
+#include <asm/nmi.h>
+
+asmlinkage void divide_error(void);
+asmlinkage void debug(void);
+asmlinkage void nmi(void);
+asmlinkage void int3(void);
+asmlinkage void overflow(void);
+asmlinkage void bounds(void);
+asmlinkage void invalid_op(void);
+asmlinkage void device_not_available(void);
+asmlinkage void double_fault(void);
+asmlinkage void coprocessor_segment_overrun(void);
+asmlinkage void invalid_TSS(void);
+asmlinkage void segment_not_present(void);
+asmlinkage void stack_segment(void);
+asmlinkage void general_protection(void);
+asmlinkage void page_fault(void);
+asmlinkage void coprocessor_error(void);
+asmlinkage void simd_coprocessor_error(void);
+asmlinkage void reserved(void);
+asmlinkage void alignment_check(void);
+asmlinkage void machine_check(void);
+asmlinkage void spurious_interrupt_bug(void);
+
+ATOMIC_NOTIFIER_HEAD(die_chain);
+EXPORT_SYMBOL(die_chain);
+
+int register_die_notifier(struct notifier_block *nb)
+{
+	vmalloc_sync_all();
+	return atomic_notifier_chain_register(&die_chain, nb);
+}
+EXPORT_SYMBOL(register_die_notifier); /* used modular by kdb */
+
+int unregister_die_notifier(struct notifier_block *nb)
+{
+	return atomic_notifier_chain_unregister(&die_chain, nb);
+}
+EXPORT_SYMBOL(unregister_die_notifier); /* used modular by kdb */
+
+static inline void conditional_sti(struct pt_regs *regs)
+{
+	if (regs->eflags & X86_EFLAGS_IF)
+		local_irq_enable();
+}
+
+static inline void preempt_conditional_sti(struct pt_regs *regs)
+{
+	preempt_disable();
+	if (regs->eflags & X86_EFLAGS_IF)
+		local_irq_enable();
+}
+
+static inline void preempt_conditional_cli(struct pt_regs *regs)
+{
+	if (regs->eflags & X86_EFLAGS_IF)
+		local_irq_disable();
+	/* Make sure to not schedule here because we could be running
+	   on an exception stack. */
+	preempt_enable_no_resched();
+}
+
+static int kstack_depth_to_print = 12;
+#ifdef CONFIG_STACK_UNWIND
+static int call_trace = 1;
+#else
+#define call_trace (-1)
+#endif
+
+#ifdef CONFIG_KALLSYMS
+# include <linux/kallsyms.h>
+void printk_address(unsigned long address)
+{
+	unsigned long offset = 0, symsize;
+	const char *symname;
+	char *modname;
+	char *delim = ":";
+	char namebuf[128];
+
+	symname = kallsyms_lookup(address, &symsize, &offset,
+					&modname, namebuf);
+	if (!symname) {
+		printk(" [<%016lx>]\n", address);
+		return;
+	}
+	if (!modname)
+		modname = delim = ""; 		
+	printk(" [<%016lx>] %s%s%s%s+0x%lx/0x%lx\n",
+		address, delim, modname, delim, symname, offset, symsize);
+}
+#else
+void printk_address(unsigned long address)
+{
+	printk(" [<%016lx>]\n", address);
+}
+#endif
+
+static unsigned long *in_exception_stack(unsigned cpu, unsigned long stack,
+					unsigned *usedp, const char **idp)
+{
+#ifndef CONFIG_X86_NO_TSS
+	static char ids[][8] = {
+		[DEBUG_STACK - 1] = "#DB",
+		[NMI_STACK - 1] = "NMI",
+		[DOUBLEFAULT_STACK - 1] = "#DF",
+		[STACKFAULT_STACK - 1] = "#SS",
+		[MCE_STACK - 1] = "#MC",
+#if DEBUG_STKSZ > EXCEPTION_STKSZ
+		[N_EXCEPTION_STACKS ... N_EXCEPTION_STACKS + DEBUG_STKSZ / EXCEPTION_STKSZ - 2] = "#DB[?]"
+#endif
+	};
+	unsigned k;
+
+	/*
+	 * Iterate over all exception stacks, and figure out whether
+	 * 'stack' is in one of them:
+	 */
+	for (k = 0; k < N_EXCEPTION_STACKS; k++) {
+		unsigned long end;
+
+		/*
+		 * set 'end' to the end of the exception stack.
+		 */
+		switch (k + 1) {
+		/*
+		 * TODO: this block is not needed i think, because
+		 * setup64.c:cpu_init() sets up t->ist[DEBUG_STACK]
+		 * properly too.
+		 */
+#if DEBUG_STKSZ > EXCEPTION_STKSZ
+		case DEBUG_STACK:
+			end = cpu_pda(cpu)->debugstack + DEBUG_STKSZ;
+			break;
+#endif
+		default:
+			end = per_cpu(orig_ist, cpu).ist[k];
+			break;
+		}
+		/*
+		 * Is 'stack' above this exception frame's end?
+		 * If yes then skip to the next frame.
+		 */
+		if (stack >= end)
+			continue;
+		/*
+		 * Is 'stack' above this exception frame's start address?
+		 * If yes then we found the right frame.
+		 */
+		if (stack >= end - EXCEPTION_STKSZ) {
+			/*
+			 * Make sure we only iterate through an exception
+			 * stack once. If it comes up for the second time
+			 * then there's something wrong going on - just
+			 * break out and return NULL:
+			 */
+			if (*usedp & (1U << k))
+				break;
+			*usedp |= 1U << k;
+			*idp = ids[k];
+			return (unsigned long *)end;
+		}
+		/*
+		 * If this is a debug stack, and if it has a larger size than
+		 * the usual exception stacks, then 'stack' might still
+		 * be within the lower portion of the debug stack:
+		 */
+#if DEBUG_STKSZ > EXCEPTION_STKSZ
+		if (k == DEBUG_STACK - 1 && stack >= end - DEBUG_STKSZ) {
+			unsigned j = N_EXCEPTION_STACKS - 1;
+
+			/*
+			 * Black magic. A large debug stack is composed of
+			 * multiple exception stack entries, which we
+			 * iterate through now. Dont look:
+			 */
+			do {
+				++j;
+				end -= EXCEPTION_STKSZ;
+				ids[j][4] = '1' + (j - N_EXCEPTION_STACKS);
+			} while (stack < end - EXCEPTION_STKSZ);
+			if (*usedp & (1U << j))
+				break;
+			*usedp |= 1U << j;
+			*idp = ids[j];
+			return (unsigned long *)end;
+		}
+#endif
+	}
+#endif
+	return NULL;
+}
+
+static int show_trace_unwind(struct unwind_frame_info *info, void *context)
+{
+	int n = 0;
+
+	while (unwind(info) == 0 && UNW_PC(info)) {
+		n++;
+		printk_address(UNW_PC(info));
+		if (arch_unw_user_mode(info))
+			break;
+	}
+	return n;
+}
+
+/*
+ * x86-64 can have upto three kernel stacks: 
+ * process stack
+ * interrupt stack
+ * severe exception (double fault, nmi, stack fault, debug, mce) hardware stack
+ */
+
+void show_trace(struct task_struct *tsk, struct pt_regs *regs, unsigned long * stack)
+{
+	const unsigned cpu = safe_smp_processor_id();
+	unsigned long *irqstack_end = (unsigned long *)cpu_pda(cpu)->irqstackptr;
+	unsigned used = 0;
+
+	printk("\nCall Trace:\n");
+
+	if (!tsk)
+		tsk = current;
+
+	if (call_trace >= 0) {
+		int unw_ret = 0;
+		struct unwind_frame_info info;
+
+		if (regs) {
+			if (unwind_init_frame_info(&info, tsk, regs) == 0)
+				unw_ret = show_trace_unwind(&info, NULL);
+		} else if (tsk == current)
+			unw_ret = unwind_init_running(&info, show_trace_unwind, NULL);
+		else {
+			if (unwind_init_blocked(&info, tsk) == 0)
+				unw_ret = show_trace_unwind(&info, NULL);
+		}
+		if (unw_ret > 0) {
+			if (call_trace == 1 && !arch_unw_user_mode(&info)) {
+				print_symbol("DWARF2 unwinder stuck at %s\n",
+					     UNW_PC(&info));
+				if ((long)UNW_SP(&info) < 0) {
+					printk("Leftover inexact backtrace:\n");
+					stack = (unsigned long *)UNW_SP(&info);
+				} else
+					printk("Full inexact backtrace again:\n");
+			} else if (call_trace >= 1)
+				return;
+			else
+				printk("Full inexact backtrace again:\n");
+		} else
+			printk("Inexact backtrace:\n");
+	}
+
+	/*
+	 * Print function call entries within a stack. 'cond' is the
+	 * "end of stackframe" condition, that the 'stack++'
+	 * iteration will eventually trigger.
+	 */
+#define HANDLE_STACK(cond) \
+	do while (cond) { \
+		unsigned long addr = *stack++; \
+		if (kernel_text_address(addr)) { \
+			/* \
+			 * If the address is either in the text segment of the \
+			 * kernel, or in the region which contains vmalloc'ed \
+			 * memory, it *may* be the address of a calling \
+			 * routine; if so, print it so that someone tracing \
+			 * down the cause of the crash will be able to figure \
+			 * out the call path that was taken. \
+			 */ \
+			printk_address(addr); \
+		} \
+	} while (0)
+
+	/*
+	 * Print function call entries in all stacks, starting at the
+	 * current stack address. If the stacks consist of nested
+	 * exceptions
+	 */
+	for ( ; ; ) {
+		const char *id;
+		unsigned long *estack_end;
+		estack_end = in_exception_stack(cpu, (unsigned long)stack,
+						&used, &id);
+
+		if (estack_end) {
+			printk(" <%s>", id);
+			HANDLE_STACK (stack < estack_end);
+			printk(" <EOE>");
+			/*
+			 * We link to the next stack via the
+			 * second-to-last pointer (index -2 to end) in the
+			 * exception stack:
+			 */
+			stack = (unsigned long *) estack_end[-2];
+			continue;
+		}
+		if (irqstack_end) {
+			unsigned long *irqstack;
+			irqstack = irqstack_end -
+				(IRQSTACKSIZE - 64) / sizeof(*irqstack);
+
+			if (stack >= irqstack && stack < irqstack_end) {
+				printk(" <IRQ>");
+				HANDLE_STACK (stack < irqstack_end);
+				/*
+				 * We link to the next stack (which would be
+				 * the process stack normally) the last
+				 * pointer (index -1 to end) in the IRQ stack:
+				 */
+				stack = (unsigned long *) (irqstack_end[-1]);
+				irqstack_end = NULL;
+				printk(" <EOI>");
+				continue;
+			}
+		}
+		break;
+	}
+
+	/*
+	 * This prints the process stack:
+	 */
+	HANDLE_STACK (((long) stack & (THREAD_SIZE-1)) != 0);
+#undef HANDLE_STACK
+
+	printk("\n");
+}
+
+static void _show_stack(struct task_struct *tsk, struct pt_regs *regs, unsigned long * rsp)
+{
+	unsigned long *stack;
+	int i;
+	const int cpu = safe_smp_processor_id();
+	unsigned long *irqstack_end = (unsigned long *) (cpu_pda(cpu)->irqstackptr);
+	unsigned long *irqstack = (unsigned long *) (cpu_pda(cpu)->irqstackptr - IRQSTACKSIZE);
+
+	// debugging aid: "show_stack(NULL, NULL);" prints the
+	// back trace for this cpu.
+
+	if (rsp == NULL) {
+		if (tsk)
+			rsp = (unsigned long *)tsk->thread.rsp;
+		else
+			rsp = (unsigned long *)&rsp;
+	}
+
+	stack = rsp;
+	for(i=0; i < kstack_depth_to_print; i++) {
+		if (stack >= irqstack && stack <= irqstack_end) {
+			if (stack == irqstack_end) {
+				stack = (unsigned long *) (irqstack_end[-1]);
+				printk(" <EOI> ");
+			}
+		} else {
+		if (((long) stack & (THREAD_SIZE-1)) == 0)
+			break;
+		}
+		if (i && ((i % 4) == 0))
+			printk("\n");
+		printk(" %016lx", *stack++);
+		touch_nmi_watchdog();
+	}
+	show_trace(tsk, regs, rsp);
+}
+
+void show_stack(struct task_struct *tsk, unsigned long * rsp)
+{
+	_show_stack(tsk, NULL, rsp);
+}
+
+/*
+ * The architecture-independent dump_stack generator
+ */
+void dump_stack(void)
+{
+	unsigned long dummy;
+	show_trace(NULL, NULL, &dummy);
+}
+
+EXPORT_SYMBOL(dump_stack);
+
+void show_registers(struct pt_regs *regs)
+{
+	int i;
+	int in_kernel = !user_mode(regs);
+	unsigned long rsp;
+	const int cpu = safe_smp_processor_id(); 
+	struct task_struct *cur = cpu_pda(cpu)->pcurrent;
+
+		rsp = regs->rsp;
+
+	printk("CPU %d ", cpu);
+	__show_regs(regs);
+	printk("Process %s (pid: %d, threadinfo %p, task %p)\n",
+		cur->comm, cur->pid, task_thread_info(cur), cur);
+
+	/*
+	 * When in-kernel, we also print out the stack and code at the
+	 * time of the fault..
+	 */
+	if (in_kernel) {
+
+		printk("Stack: ");
+		_show_stack(NULL, regs, (unsigned long*)rsp);
+
+		printk("\nCode: ");
+		if (regs->rip < PAGE_OFFSET)
+			goto bad;
+
+		for (i=0; i<20; i++) {
+			unsigned char c;
+			if (__get_user(c, &((unsigned char*)regs->rip)[i])) {
+bad:
+				printk(" Bad RIP value.");
+				break;
+			}
+			printk("%02x ", c);
+		}
+	}
+	printk("\n");
+}	
+
+void handle_BUG(struct pt_regs *regs)
+{ 
+	struct bug_frame f;
+	long len;
+	const char *prefix = "";
+
+	if (user_mode(regs))
+		return; 
+	if (__copy_from_user(&f, (const void __user *) regs->rip,
+			     sizeof(struct bug_frame)))
+		return; 
+	if (f.filename >= 0 ||
+	    f.ud2[0] != 0x0f || f.ud2[1] != 0x0b) 
+		return;
+	len = __strnlen_user((char *)(long)f.filename, PATH_MAX) - 1;
+	if (len < 0 || len >= PATH_MAX)
+		f.filename = (int)(long)"unmapped filename";
+	else if (len > 50) {
+		f.filename += len - 50;
+		prefix = "...";
+	}
+	printk("----------- [cut here ] --------- [please bite here ] ---------\n");
+	printk(KERN_ALERT "Kernel BUG at %s%.50s:%d\n", prefix, (char *)(long)f.filename, f.line);
+} 
+
+#ifdef CONFIG_BUG
+void out_of_line_bug(void)
+{ 
+	BUG(); 
+} 
+EXPORT_SYMBOL(out_of_line_bug);
+#endif
+
+static DEFINE_SPINLOCK(die_lock);
+static int die_owner = -1;
+static unsigned int die_nest_count;
+
+unsigned __kprobes long oops_begin(void)
+{
+	int cpu = safe_smp_processor_id();
+	unsigned long flags;
+
+	/* racy, but better than risking deadlock. */
+	local_irq_save(flags);
+	if (!spin_trylock(&die_lock)) { 
+		if (cpu == die_owner) 
+			/* nested oops. should stop eventually */;
+		else
+			spin_lock(&die_lock);
+	}
+	die_nest_count++;
+	die_owner = cpu;
+	console_verbose();
+	bust_spinlocks(1);
+	return flags;
+}
+
+void __kprobes oops_end(unsigned long flags)
+{ 
+	die_owner = -1;
+	bust_spinlocks(0);
+	die_nest_count--;
+	if (die_nest_count)
+		/* We still own the lock */
+		local_irq_restore(flags);
+	else
+		/* Nest count reaches zero, release the lock. */
+		spin_unlock_irqrestore(&die_lock, flags);
+	if (panic_on_oops)
+		panic("Fatal exception");
+}
+
+void __kprobes __die(const char * str, struct pt_regs * regs, long err)
+{
+	static int die_counter;
+	printk(KERN_EMERG "%s: %04lx [%u] ", str, err & 0xffff,++die_counter);
+#ifdef CONFIG_PREEMPT
+	printk("PREEMPT ");
+#endif
+#ifdef CONFIG_SMP
+	printk("SMP ");
+#endif
+#ifdef CONFIG_DEBUG_PAGEALLOC
+	printk("DEBUG_PAGEALLOC");
+#endif
+	printk("\n");
+	notify_die(DIE_OOPS, str, regs, err, current->thread.trap_no, SIGSEGV);
+	show_registers(regs);
+	/* Executive summary in case the oops scrolled away */
+	printk(KERN_ALERT "RIP ");
+	printk_address(regs->rip); 
+	printk(" RSP <%016lx>\n", regs->rsp); 
+	if (kexec_should_crash(current))
+		crash_kexec(regs);
+}
+
+void die(const char * str, struct pt_regs * regs, long err)
+{
+	unsigned long flags = oops_begin();
+
+	handle_BUG(regs);
+	__die(str, regs, err);
+	oops_end(flags);
+	do_exit(SIGSEGV); 
+}
+
+#ifdef CONFIG_X86_LOCAL_APIC
+void __kprobes die_nmi(char *str, struct pt_regs *regs)
+{
+	unsigned long flags = oops_begin();
+
+	/*
+	 * We are in trouble anyway, lets at least try
+	 * to get a message out.
+	 */
+	printk(str, safe_smp_processor_id());
+	show_registers(regs);
+	if (kexec_should_crash(current))
+		crash_kexec(regs);
+	if (panic_on_timeout || panic_on_oops)
+		panic("nmi watchdog");
+	printk("console shuts up ...\n");
+	oops_end(flags);
+	nmi_exit();
+	local_irq_enable();
+	do_exit(SIGSEGV);
+}
+#endif
+
+static void __kprobes do_trap(int trapnr, int signr, char *str,
+			      struct pt_regs * regs, long error_code,
+			      siginfo_t *info)
+{
+	struct task_struct *tsk = current;
+
+	tsk->thread.error_code = error_code;
+	tsk->thread.trap_no = trapnr;
+
+	if (user_mode(regs)) {
+		if (exception_trace && unhandled_signal(tsk, signr))
+			printk(KERN_INFO
+			       "%s[%d] trap %s rip:%lx rsp:%lx error:%lx\n",
+			       tsk->comm, tsk->pid, str,
+			       regs->rip, regs->rsp, error_code); 
+
+		if (info)
+			force_sig_info(signr, info, tsk);
+		else
+			force_sig(signr, tsk);
+		return;
+	}
+
+
+	/* kernel trap */ 
+	{	     
+		const struct exception_table_entry *fixup;
+		fixup = search_exception_tables(regs->rip);
+		if (fixup)
+			regs->rip = fixup->fixup;
+		else	
+			die(str, regs, error_code);
+		return;
+	}
+}
+
+#define DO_ERROR(trapnr, signr, str, name) \
+asmlinkage void do_##name(struct pt_regs * regs, long error_code) \
+{ \
+	if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) \
+							== NOTIFY_STOP) \
+		return; \
+	conditional_sti(regs);						\
+	do_trap(trapnr, signr, str, regs, error_code, NULL); \
+}
+
+#define DO_ERROR_INFO(trapnr, signr, str, name, sicode, siaddr) \
+asmlinkage void do_##name(struct pt_regs * regs, long error_code) \
+{ \
+	siginfo_t info; \
+	info.si_signo = signr; \
+	info.si_errno = 0; \
+	info.si_code = sicode; \
+	info.si_addr = (void __user *)siaddr; \
+	if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) \
+							== NOTIFY_STOP) \
+		return; \
+	conditional_sti(regs);						\
+	do_trap(trapnr, signr, str, regs, error_code, &info); \
+}
+
+DO_ERROR_INFO( 0, SIGFPE,  "divide error", divide_error, FPE_INTDIV, regs->rip)
+DO_ERROR( 4, SIGSEGV, "overflow", overflow)
+DO_ERROR( 5, SIGSEGV, "bounds", bounds)
+DO_ERROR_INFO( 6, SIGILL,  "invalid opcode", invalid_op, ILL_ILLOPN, regs->rip)
+DO_ERROR( 7, SIGSEGV, "device not available", device_not_available)
+DO_ERROR( 9, SIGFPE,  "coprocessor segment overrun", coprocessor_segment_overrun)
+DO_ERROR(10, SIGSEGV, "invalid TSS", invalid_TSS)
+DO_ERROR(11, SIGBUS,  "segment not present", segment_not_present)
+DO_ERROR_INFO(17, SIGBUS, "alignment check", alignment_check, BUS_ADRALN, 0)
+DO_ERROR(18, SIGSEGV, "reserved", reserved)
+
+/* Runs on IST stack */
+asmlinkage void do_stack_segment(struct pt_regs *regs, long error_code)
+{
+	if (notify_die(DIE_TRAP, "stack segment", regs, error_code,
+			12, SIGBUS) == NOTIFY_STOP)
+		return;
+	preempt_conditional_sti(regs);
+	do_trap(12, SIGBUS, "stack segment", regs, error_code, NULL);
+	preempt_conditional_cli(regs);
+}
+
+asmlinkage void do_double_fault(struct pt_regs * regs, long error_code)
+{
+	static const char str[] = "double fault";
+	struct task_struct *tsk = current;
+
+	/* Return not checked because double check cannot be ignored */
+	notify_die(DIE_TRAP, str, regs, error_code, 8, SIGSEGV);
+
+	tsk->thread.error_code = error_code;
+	tsk->thread.trap_no = 8;
+
+	/* This is always a kernel trap and never fixable (and thus must
+	   never return). */
+	for (;;)
+		die(str, regs, error_code);
+}
+
+asmlinkage void __kprobes do_general_protection(struct pt_regs * regs,
+						long error_code)
+{
+	struct task_struct *tsk = current;
+
+	conditional_sti(regs);
+
+	tsk->thread.error_code = error_code;
+	tsk->thread.trap_no = 13;
+
+	if (user_mode(regs)) {
+		if (exception_trace && unhandled_signal(tsk, SIGSEGV))
+			printk(KERN_INFO
+		       "%s[%d] general protection rip:%lx rsp:%lx error:%lx\n",
+			       tsk->comm, tsk->pid,
+			       regs->rip, regs->rsp, error_code); 
+
+		force_sig(SIGSEGV, tsk);
+		return;
+	} 
+
+	/* kernel gp */
+	{
+		const struct exception_table_entry *fixup;
+		fixup = search_exception_tables(regs->rip);
+		if (fixup) {
+			regs->rip = fixup->fixup;
+			return;
+		}
+		if (notify_die(DIE_GPF, "general protection fault", regs,
+					error_code, 13, SIGSEGV) == NOTIFY_STOP)
+			return;
+		die("general protection fault", regs, error_code);
+	}
+}
+
+static __kprobes void
+mem_parity_error(unsigned char reason, struct pt_regs * regs)
+{
+	printk("Uhhuh. NMI received. Dazed and confused, but trying to continue\n");
+	printk("You probably have a hardware problem with your RAM chips\n");
+
+#if 0 /* XEN */
+	/* Clear and disable the memory parity error line. */
+	reason = (reason & 0xf) | 4;
+	outb(reason, 0x61);
+#endif /* XEN */
+}
+
+static __kprobes void
+io_check_error(unsigned char reason, struct pt_regs * regs)
+{
+	printk("NMI: IOCK error (debug interrupt?)\n");
+	show_registers(regs);
+
+#if 0 /* XEN */
+	/* Re-enable the IOCK line, wait for a few seconds */
+	reason = (reason & 0xf) | 8;
+	outb(reason, 0x61);
+	mdelay(2000);
+	reason &= ~8;
+	outb(reason, 0x61);
+#endif /* XEN */
+}
+
+static __kprobes void
+unknown_nmi_error(unsigned char reason, struct pt_regs * regs)
+{	printk("Uhhuh. NMI received for unknown reason %02x.\n", reason);
+	printk("Dazed and confused, but trying to continue\n");
+	printk("Do you have a strange power saving mode enabled?\n");
+}
+
+/* Runs on IST stack. This code must keep interrupts off all the time.
+   Nested NMIs are prevented by the CPU. */
+asmlinkage __kprobes void default_do_nmi(struct pt_regs *regs)
+{
+	unsigned char reason = 0;
+	int cpu;
+
+	cpu = smp_processor_id();
+
+	/* Only the BSP gets external NMIs from the system.  */
+	if (!cpu)
+		reason = get_nmi_reason();
+
+	if (!(reason & 0xc0)) {
+		if (notify_die(DIE_NMI_IPI, "nmi_ipi", regs, reason, 2, SIGINT)
+								== NOTIFY_STOP)
+			return;
+#ifdef CONFIG_X86_LOCAL_APIC
+		/*
+		 * Ok, so this is none of the documented NMI sources,
+		 * so it must be the NMI watchdog.
+		 */
+		if (nmi_watchdog > 0) {
+			nmi_watchdog_tick(regs,reason);
+			return;
+		}
+#endif
+		unknown_nmi_error(reason, regs);
+		return;
+	}
+	if (notify_die(DIE_NMI, "nmi", regs, reason, 2, SIGINT) == NOTIFY_STOP)
+		return; 
+
+	/* AK: following checks seem to be broken on modern chipsets. FIXME */
+
+	if (reason & 0x80)
+		mem_parity_error(reason, regs);
+	if (reason & 0x40)
+		io_check_error(reason, regs);
+}
+
+/* runs on IST stack. */
+asmlinkage void __kprobes do_int3(struct pt_regs * regs, long error_code)
+{
+	if (notify_die(DIE_INT3, "int3", regs, error_code, 3, SIGTRAP) == NOTIFY_STOP) {
+		return;
+	}
+	preempt_conditional_sti(regs);
+	do_trap(3, SIGTRAP, "int3", regs, error_code, NULL);
+	preempt_conditional_cli(regs);
+}
+
+/* Help handler running on IST stack to switch back to user stack
+   for scheduling or signal handling. The actual stack switch is done in
+   entry.S */
+asmlinkage __kprobes struct pt_regs *sync_regs(struct pt_regs *eregs)
+{
+	struct pt_regs *regs = eregs;
+	/* Did already sync */
+	if (eregs == (struct pt_regs *)eregs->rsp)
+		;
+	/* Exception from user space */
+	else if (user_mode(eregs))
+		regs = task_pt_regs(current);
+	/* Exception from kernel and interrupts are enabled. Move to
+ 	   kernel process stack. */
+	else if (eregs->eflags & X86_EFLAGS_IF)
+		regs = (struct pt_regs *)(eregs->rsp -= sizeof(struct pt_regs));
+	if (eregs != regs)
+		*regs = *eregs;
+	return regs;
+}
+
+/* runs on IST stack. */
+asmlinkage void __kprobes do_debug(struct pt_regs * regs,
+				   unsigned long error_code)
+{
+	unsigned long condition;
+	struct task_struct *tsk = current;
+	siginfo_t info;
+
+	get_debugreg(condition, 6);
+
+	if (notify_die(DIE_DEBUG, "debug", regs, condition, error_code,
+						SIGTRAP) == NOTIFY_STOP)
+		return;
+
+	preempt_conditional_sti(regs);
+
+	/* Mask out spurious debug traps due to lazy DR7 setting */
+	if (condition & (DR_TRAP0|DR_TRAP1|DR_TRAP2|DR_TRAP3)) {
+		if (!tsk->thread.debugreg7) { 
+			goto clear_dr7;
+		}
+	}
+
+	tsk->thread.debugreg6 = condition;
+
+	/* Mask out spurious TF errors due to lazy TF clearing */
+	if (condition & DR_STEP) {
+		/*
+		 * The TF error should be masked out only if the current
+		 * process is not traced and if the TRAP flag has been set
+		 * previously by a tracing process (condition detected by
+		 * the PT_DTRACE flag); remember that the i386 TRAP flag
+		 * can be modified by the process itself in user mode,
+		 * allowing programs to debug themselves without the ptrace()
+		 * interface.
+		 */
+                if (!user_mode(regs))
+                       goto clear_TF_reenable;
+		/*
+		 * Was the TF flag set by a debugger? If so, clear it now,
+		 * so that register information is correct.
+		 */
+		if (tsk->ptrace & PT_DTRACE) {
+			regs->eflags &= ~TF_MASK;
+			tsk->ptrace &= ~PT_DTRACE;
+		}
+	}
+
+	/* Ok, finally something we can handle */
+	tsk->thread.trap_no = 1;
+	tsk->thread.error_code = error_code;
+	info.si_signo = SIGTRAP;
+	info.si_errno = 0;
+	info.si_code = TRAP_BRKPT;
+	info.si_addr = user_mode(regs) ? (void __user *)regs->rip : NULL;
+	force_sig_info(SIGTRAP, &info, tsk);
+
+clear_dr7:
+	set_debugreg(0UL, 7);
+	preempt_conditional_cli(regs);
+	return;
+
+clear_TF_reenable:
+	set_tsk_thread_flag(tsk, TIF_SINGLESTEP);
+	regs->eflags &= ~TF_MASK;
+	preempt_conditional_cli(regs);
+}
+
+static int kernel_math_error(struct pt_regs *regs, const char *str, int trapnr)
+{
+	const struct exception_table_entry *fixup;
+	fixup = search_exception_tables(regs->rip);
+	if (fixup) {
+		regs->rip = fixup->fixup;
+		return 1;
+	}
+	notify_die(DIE_GPF, str, regs, 0, trapnr, SIGFPE);
+	/* Illegal floating point operation in the kernel */
+	current->thread.trap_no = trapnr;
+	die(str, regs, 0);
+	return 0;
+}
+
+/*
+ * Note that we play around with the 'TS' bit in an attempt to get
+ * the correct behaviour even in the presence of the asynchronous
+ * IRQ13 behaviour
+ */
+asmlinkage void do_coprocessor_error(struct pt_regs *regs)
+{
+	void __user *rip = (void __user *)(regs->rip);
+	struct task_struct * task;
+	siginfo_t info;
+	unsigned short cwd, swd;
+
+	conditional_sti(regs);
+	if (!user_mode(regs) &&
+	    kernel_math_error(regs, "kernel x87 math error", 16))
+		return;
+
+	/*
+	 * Save the info for the exception handler and clear the error.
+	 */
+	task = current;
+	save_init_fpu(task);
+	task->thread.trap_no = 16;
+	task->thread.error_code = 0;
+	info.si_signo = SIGFPE;
+	info.si_errno = 0;
+	info.si_code = __SI_FAULT;
+	info.si_addr = rip;
+	/*
+	 * (~cwd & swd) will mask out exceptions that are not set to unmasked
+	 * status.  0x3f is the exception bits in these regs, 0x200 is the
+	 * C1 reg you need in case of a stack fault, 0x040 is the stack
+	 * fault bit.  We should only be taking one exception at a time,
+	 * so if this combination doesn't produce any single exception,
+	 * then we have a bad program that isn't synchronizing its FPU usage
+	 * and it will suffer the consequences since we won't be able to
+	 * fully reproduce the context of the exception
+	 */
+	cwd = get_fpu_cwd(task);
+	swd = get_fpu_swd(task);
+	switch (swd & ~cwd & 0x3f) {
+		case 0x000:
+		default:
+			break;
+		case 0x001: /* Invalid Op */
+			/*
+			 * swd & 0x240 == 0x040: Stack Underflow
+			 * swd & 0x240 == 0x240: Stack Overflow
+			 * User must clear the SF bit (0x40) if set
+			 */
+			info.si_code = FPE_FLTINV;
+			break;
+		case 0x002: /* Denormalize */
+		case 0x010: /* Underflow */
+			info.si_code = FPE_FLTUND;
+			break;
+		case 0x004: /* Zero Divide */
+			info.si_code = FPE_FLTDIV;
+			break;
+		case 0x008: /* Overflow */
+			info.si_code = FPE_FLTOVF;
+			break;
+		case 0x020: /* Precision */
+			info.si_code = FPE_FLTRES;
+			break;
+	}
+	force_sig_info(SIGFPE, &info, task);
+}
+
+asmlinkage void bad_intr(void)
+{
+	printk("bad interrupt"); 
+}
+
+asmlinkage void do_simd_coprocessor_error(struct pt_regs *regs)
+{
+	void __user *rip = (void __user *)(regs->rip);
+	struct task_struct * task;
+	siginfo_t info;
+	unsigned short mxcsr;
+
+	conditional_sti(regs);
+	if (!user_mode(regs) &&
+        	kernel_math_error(regs, "kernel simd math error", 19))
+		return;
+
+	/*
+	 * Save the info for the exception handler and clear the error.
+	 */
+	task = current;
+	save_init_fpu(task);
+	task->thread.trap_no = 19;
+	task->thread.error_code = 0;
+	info.si_signo = SIGFPE;
+	info.si_errno = 0;
+	info.si_code = __SI_FAULT;
+	info.si_addr = rip;
+	/*
+	 * The SIMD FPU exceptions are handled a little differently, as there
+	 * is only a single status/control register.  Thus, to determine which
+	 * unmasked exception was caught we must mask the exception mask bits
+	 * at 0x1f80, and then use these to mask the exception bits at 0x3f.
+	 */
+	mxcsr = get_fpu_mxcsr(task);
+	switch (~((mxcsr & 0x1f80) >> 7) & (mxcsr & 0x3f)) {
+		case 0x000:
+		default:
+			break;
+		case 0x001: /* Invalid Op */
+			info.si_code = FPE_FLTINV;
+			break;
+		case 0x002: /* Denormalize */
+		case 0x010: /* Underflow */
+			info.si_code = FPE_FLTUND;
+			break;
+		case 0x004: /* Zero Divide */
+			info.si_code = FPE_FLTDIV;
+			break;
+		case 0x008: /* Overflow */
+			info.si_code = FPE_FLTOVF;
+			break;
+		case 0x020: /* Precision */
+			info.si_code = FPE_FLTRES;
+			break;
+	}
+	force_sig_info(SIGFPE, &info, task);
+}
+
+asmlinkage void do_spurious_interrupt_bug(struct pt_regs * regs)
+{
+}
+
+#if 0
+asmlinkage void __attribute__((weak)) smp_thermal_interrupt(void)
+{
+}
+#endif
+
+asmlinkage void __attribute__((weak)) mce_threshold_interrupt(void)
+{
+}
+
+/*
+ *  'math_state_restore()' saves the current math information in the
+ * old math state array, and gets the new ones from the current task
+ *
+ * Careful.. There are problems with IBM-designed IRQ13 behaviour.
+ * Don't touch unless you *really* know how it works.
+ */
+asmlinkage void math_state_restore(void)
+{
+	struct task_struct *me = current;
+        /* clts(); */ /* 'clts' is done for us by Xen during virtual trap. */
+
+	if (!used_math())
+		init_fpu(me);
+	restore_fpu_checking(&me->thread.i387.fxsave);
+	task_thread_info(me)->status |= TS_USEDFPU;
+}
+
+
+/*
+ * NB. All these are "interrupt gates" (i.e. events_mask is set) because we
+ * specify <dpl>|4 in the second field.
+ */
+static trap_info_t __cpuinitdata trap_table[] = {
+        {  0, 0|4, __KERNEL_CS, (unsigned long)divide_error               },
+        {  1, 0|4, __KERNEL_CS, (unsigned long)debug                      },
+        {  3, 3|4, __KERNEL_CS, (unsigned long)int3                       },
+        {  4, 3|4, __KERNEL_CS, (unsigned long)overflow                   },
+        {  5, 0|4, __KERNEL_CS, (unsigned long)bounds                     },
+        {  6, 0|4, __KERNEL_CS, (unsigned long)invalid_op                 },
+        {  7, 0|4, __KERNEL_CS, (unsigned long)device_not_available       },
+        {  9, 0|4, __KERNEL_CS, (unsigned long)coprocessor_segment_overrun},
+        { 10, 0|4, __KERNEL_CS, (unsigned long)invalid_TSS                },
+        { 11, 0|4, __KERNEL_CS, (unsigned long)segment_not_present        },
+        { 12, 0|4, __KERNEL_CS, (unsigned long)stack_segment              },
+        { 13, 0|4, __KERNEL_CS, (unsigned long)general_protection         },
+        { 14, 0|4, __KERNEL_CS, (unsigned long)page_fault                 },
+        { 15, 0|4, __KERNEL_CS, (unsigned long)spurious_interrupt_bug     },
+        { 16, 0|4, __KERNEL_CS, (unsigned long)coprocessor_error          },
+        { 17, 0|4, __KERNEL_CS, (unsigned long)alignment_check            },
+#ifdef CONFIG_X86_MCE
+        { 18, 0|4, __KERNEL_CS, (unsigned long)machine_check              },
+#endif
+        { 19, 0|4, __KERNEL_CS, (unsigned long)simd_coprocessor_error     },
+#ifdef CONFIG_IA32_EMULATION
+	{ IA32_SYSCALL_VECTOR, 3, __KERNEL_CS, (unsigned long)ia32_syscall},
+#endif
+        {  0, 0,           0, 0                                              }
+};
+
+void __init trap_init(void)
+{
+        int ret;
+
+        ret = HYPERVISOR_set_trap_table(trap_table);
+        if (ret) 
+		printk("HYPERVISOR_set_trap_table failed: error %d\n", ret);
+
+	/*
+	 * Should be a barrier for any external CPU state.
+	 */
+	cpu_init();
+}
+
+void __cpuinit smp_trap_init(trap_info_t *trap_ctxt)
+{
+	const trap_info_t *t = trap_table;
+
+	for (t = trap_table; t->address; t++) {
+		trap_ctxt[t->vector].flags = t->flags;
+		trap_ctxt[t->vector].cs = t->cs;
+		trap_ctxt[t->vector].address = t->address;
+	}
+}
+
+
+/* Actual parsing is done early in setup.c. */
+static int __init oops_dummy(char *s)
+{ 
+	panic_on_oops = 1;
+	return 1;
+} 
+__setup("oops=", oops_dummy); 
+
+static int __init kstack_setup(char *s)
+{
+	kstack_depth_to_print = simple_strtoul(s,NULL,0);
+	return 1;
+}
+__setup("kstack=", kstack_setup);
+
+#ifdef CONFIG_STACK_UNWIND
+static int __init call_trace_setup(char *s)
+{
+	if (strcmp(s, "old") == 0)
+		call_trace = -1;
+	else if (strcmp(s, "both") == 0)
+		call_trace = 0;
+	else if (strcmp(s, "newfallback") == 0)
+		call_trace = 1;
+	else if (strcmp(s, "new") == 0)
+		call_trace = 2;
+	return 1;
+}
+__setup("call_trace=", call_trace_setup);
+#endif
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/vmlinux.lds.S linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/vmlinux.lds.S
--- linux-2.6.18.8/arch/x86_64/kernel/vmlinux.lds.S	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/vmlinux.lds.S	2008-08-21 11:36:07.000000000 +0200
@@ -13,6 +13,13 @@ OUTPUT_FORMAT("elf64-x86-64", "elf64-x86
 OUTPUT_ARCH(i386:x86-64)
 ENTRY(phys_startup_64)
 jiffies_64 = jiffies;
+PHDRS {
+	text PT_LOAD FLAGS(5);	/* R_E */
+	data PT_LOAD FLAGS(7);	/* RWE */
+	user PT_LOAD FLAGS(7);	/* RWE */
+	data.init PT_LOAD FLAGS(7);	/* RWE */
+	note PT_NOTE FLAGS(4);	/* R__ */
+}
 SECTIONS
 {
   . = __START_KERNEL;
@@ -31,7 +38,7 @@ SECTIONS
 	KPROBES_TEXT
 	*(.fixup)
 	*(.gnu.warning)
-	} = 0x9090
+	} :text = 0x9090
   				/* out-of-line lock text */
   .text.lock : AT(ADDR(.text.lock) - LOAD_OFFSET) { *(.text.lock) }
 
@@ -57,17 +64,10 @@ SECTIONS
   .data : AT(ADDR(.data) - LOAD_OFFSET) {
 	*(.data)
 	CONSTRUCTORS
-	}
+	} :data
 
   _edata = .;			/* End of data section */
 
-  __bss_start = .;		/* BSS */
-  .bss : AT(ADDR(.bss) - LOAD_OFFSET) {
-	*(.bss.page_aligned)	
-	*(.bss)
-	}
-  __bss_stop = .;
-
   . = ALIGN(PAGE_SIZE);
   . = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
   .data.cacheline_aligned : AT(ADDR(.data.cacheline_aligned) - LOAD_OFFSET) {
@@ -89,7 +89,7 @@ SECTIONS
 #define VVIRT(x) (ADDR(x) - VVIRT_OFFSET)
 
   . = VSYSCALL_ADDR;
-  .vsyscall_0 :	 AT(VSYSCALL_PHYS_ADDR) { *(.vsyscall_0) }
+  .vsyscall_0 :	 AT(VSYSCALL_PHYS_ADDR) { *(.vsyscall_0) } :user
   __vsyscall_0 = VSYSCALL_VIRT_ADDR;
 
   . = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
@@ -132,7 +132,7 @@ SECTIONS
   . = ALIGN(8192);		/* init_task */
   .data.init_task : AT(ADDR(.data.init_task) - LOAD_OFFSET) {
 	*(.data.init_task)
-  }
+  }:data.init
 
   . = ALIGN(4096);
   .data.page_aligned : AT(ADDR(.data.page_aligned) - LOAD_OFFSET) {
@@ -222,6 +222,14 @@ SECTIONS
   . = ALIGN(4096);
   __nosave_end = .;
 
+  __bss_start = .;		/* BSS */
+  . = ALIGN(4096);
+  .bss : AT(ADDR(.bss) - LOAD_OFFSET) {
+	*(.bss.page_aligned)
+	*(.bss)
+	}
+  __bss_stop = .;
+
   _end = . ;
 
   /* Sections to be discarded */
@@ -235,4 +243,6 @@ SECTIONS
   STABS_DEBUG
 
   DWARF_DEBUG
+
+  NOTES
 }
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/vsyscall-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/vsyscall-xen.c
--- linux-2.6.18.8/arch/x86_64/kernel/vsyscall-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/vsyscall-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,227 @@
+/*
+ *  linux/arch/x86_64/kernel/vsyscall.c
+ *
+ *  Copyright (C) 2001 Andrea Arcangeli <andrea@suse.de> SuSE
+ *  Copyright 2003 Andi Kleen, SuSE Labs.
+ *
+ *  Thanks to hpa@transmeta.com for some useful hint.
+ *  Special thanks to Ingo Molnar for his early experience with
+ *  a different vsyscall implementation for Linux/IA32 and for the name.
+ *
+ *  vsyscall 1 is located at -10Mbyte, vsyscall 2 is located
+ *  at virtual address -10Mbyte+1024bytes etc... There are at max 4
+ *  vsyscalls. One vsyscall can reserve more than 1 slot to avoid
+ *  jumping out of line if necessary. We cannot add more with this
+ *  mechanism because older kernels won't return -ENOSYS.
+ *  If we want more than four we need a vDSO.
+ *
+ *  Note: the concept clashes with user mode linux. If you use UML and
+ *  want per guest time just set the kernel.vsyscall64 sysctl to 0.
+ */
+
+#include <linux/time.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/timer.h>
+#include <linux/seqlock.h>
+#include <linux/jiffies.h>
+#include <linux/sysctl.h>
+
+#include <asm/vsyscall.h>
+#include <asm/pgtable.h>
+#include <asm/page.h>
+#include <asm/fixmap.h>
+#include <asm/errno.h>
+#include <asm/io.h>
+
+#define __vsyscall(nr) __attribute__ ((unused,__section__(".vsyscall_" #nr)))
+
+int __sysctl_vsyscall __section_sysctl_vsyscall = 1;
+seqlock_t __xtime_lock __section_xtime_lock = SEQLOCK_UNLOCKED;
+
+#include <asm/unistd.h>
+
+static __always_inline void timeval_normalize(struct timeval * tv)
+{
+	time_t __sec;
+
+	__sec = tv->tv_usec / 1000000;
+	if (__sec) {
+		tv->tv_usec %= 1000000;
+		tv->tv_sec += __sec;
+	}
+}
+
+static __always_inline void do_vgettimeofday(struct timeval * tv)
+{
+	long sequence, t;
+	unsigned long sec, usec;
+
+	do {
+		sequence = read_seqbegin(&__xtime_lock);
+		
+		sec = __xtime.tv_sec;
+		usec = (__xtime.tv_nsec / 1000) +
+			(__jiffies - __wall_jiffies) * (1000000 / HZ);
+
+		if (__vxtime.mode != VXTIME_HPET) {
+			t = get_cycles_sync();
+			if (t < __vxtime.last_tsc)
+				t = __vxtime.last_tsc;
+			usec += ((t - __vxtime.last_tsc) *
+				 __vxtime.tsc_quot) >> 32;
+			/* See comment in x86_64 do_gettimeofday. */
+		} else {
+			usec += ((readl((void *)fix_to_virt(VSYSCALL_HPET) + 0xf0) -
+				  __vxtime.last) * __vxtime.quot) >> 32;
+		}
+	} while (read_seqretry(&__xtime_lock, sequence));
+
+	tv->tv_sec = sec + usec / 1000000;
+	tv->tv_usec = usec % 1000000;
+}
+
+/* RED-PEN may want to readd seq locking, but then the variable should be write-once. */
+static __always_inline void do_get_tz(struct timezone * tz)
+{
+	*tz = __sys_tz;
+}
+
+static __always_inline int gettimeofday(struct timeval *tv, struct timezone *tz)
+{
+	int ret;
+	asm volatile("vsysc2: syscall"
+		: "=a" (ret)
+		: "0" (__NR_gettimeofday),"D" (tv),"S" (tz) : __syscall_clobber );
+	return ret;
+}
+
+static __always_inline long time_syscall(long *t)
+{
+	long secs;
+	asm volatile("vsysc1: syscall"
+		: "=a" (secs)
+		: "0" (__NR_time),"D" (t) : __syscall_clobber);
+	return secs;
+}
+
+int __vsyscall(0) vgettimeofday(struct timeval * tv, struct timezone * tz)
+{
+	if (!__sysctl_vsyscall)
+		return gettimeofday(tv,tz);
+	if (tv)
+		do_vgettimeofday(tv);
+	if (tz)
+		do_get_tz(tz);
+	return 0;
+}
+
+/* This will break when the xtime seconds get inaccurate, but that is
+ * unlikely */
+time_t __vsyscall(1) vtime(time_t *t)
+{
+	if (!__sysctl_vsyscall)
+		return time_syscall(t);
+	else if (t)
+		*t = __xtime.tv_sec;		
+	return __xtime.tv_sec;
+}
+
+long __vsyscall(2) venosys_0(void)
+{
+	return -ENOSYS;
+}
+
+long __vsyscall(3) venosys_1(void)
+{
+	return -ENOSYS;
+}
+
+#ifdef CONFIG_SYSCTL
+
+#define SYSCALL 0x050f
+#define NOP2    0x9090
+
+/*
+ * NOP out syscall in vsyscall page when not needed.
+ */
+static int vsyscall_sysctl_change(ctl_table *ctl, int write, struct file * filp,
+                        void __user *buffer, size_t *lenp, loff_t *ppos)
+{
+	extern u16 vsysc1, vsysc2;
+	u16 *map1, *map2;
+	int ret = proc_dointvec(ctl, write, filp, buffer, lenp, ppos);
+	if (!write)
+		return ret;
+	/* gcc has some trouble with __va(__pa()), so just do it this
+	   way. */
+	map1 = ioremap(__pa_symbol(&vsysc1), 2);
+	if (!map1)
+		return -ENOMEM;
+	map2 = ioremap(__pa_symbol(&vsysc2), 2);
+	if (!map2) {
+		ret = -ENOMEM;
+		goto out;
+	}
+	if (!sysctl_vsyscall) {
+		*map1 = SYSCALL;
+		*map2 = SYSCALL;
+	} else {
+		*map1 = NOP2;
+		*map2 = NOP2;
+	}
+	iounmap(map2);
+out:
+	iounmap(map1);
+	return ret;
+}
+
+static int vsyscall_sysctl_nostrat(ctl_table *t, int __user *name, int nlen,
+				void __user *oldval, size_t __user *oldlenp,
+				void __user *newval, size_t newlen,
+				void **context)
+{
+	return -ENOSYS;
+}
+
+static ctl_table kernel_table2[] = {
+	{ .ctl_name = 99, .procname = "vsyscall64",
+	  .data = &sysctl_vsyscall, .maxlen = sizeof(int), .mode = 0644,
+	  .strategy = vsyscall_sysctl_nostrat,
+	  .proc_handler = vsyscall_sysctl_change },
+	{ 0, }
+};
+
+static ctl_table kernel_root_table2[] = {
+	{ .ctl_name = CTL_KERN, .procname = "kernel", .mode = 0555,
+	  .child = kernel_table2 },
+	{ 0 },
+};
+
+#endif
+
+static void __init map_vsyscall(void)
+{
+	extern char __vsyscall_0;
+	unsigned long physaddr_page0 = __pa_symbol(&__vsyscall_0);
+
+	__set_fixmap(VSYSCALL_FIRST_PAGE, physaddr_page0, PAGE_KERNEL_VSYSCALL);
+}
+
+static int __init vsyscall_init(void)
+{
+	BUG_ON(((unsigned long) &vgettimeofday !=
+			VSYSCALL_ADDR(__NR_vgettimeofday)));
+	BUG_ON((unsigned long) &vtime != VSYSCALL_ADDR(__NR_vtime));
+	BUG_ON((VSYSCALL_ADDR(0) != __fix_to_virt(VSYSCALL_FIRST_PAGE)));
+	map_vsyscall();
+#ifdef CONFIG_XEN
+	sysctl_vsyscall = 0; /* disable vgettimeofay() */
+#endif
+#ifdef CONFIG_SYSCTL
+	register_sysctl_table(kernel_root_table2, 0);
+#endif
+	return 0;
+}
+
+__initcall(vsyscall_init);
diff -rpuN linux-2.6.18.8/arch/x86_64/kernel/xen_entry.S linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/xen_entry.S
--- linux-2.6.18.8/arch/x86_64/kernel/xen_entry.S	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/kernel/xen_entry.S	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,36 @@
+/*
+ * Copied from arch/xen/i386/kernel/entry.S
+ */                        
+/* Offsets into shared_info_t. */                
+#define evtchn_upcall_pending		/* 0 */
+#define evtchn_upcall_mask		1
+
+#define sizeof_vcpu_shift		6
+
+#ifdef CONFIG_SMP
+//#define preempt_disable(reg)	incl threadinfo_preempt_count(reg)
+//#define preempt_enable(reg)	decl threadinfo_preempt_count(reg)
+#define preempt_disable(reg)
+#define preempt_enable(reg)
+#define XEN_GET_VCPU_INFO(reg)	preempt_disable(%rbp)			; \
+				movq %gs:pda_cpunumber,reg		; \
+				shl  $32, reg				; \
+				shr  $32-sizeof_vcpu_shift,reg		; \
+				addq HYPERVISOR_shared_info,reg
+#define XEN_PUT_VCPU_INFO(reg)	preempt_enable(%rbp)			; \
+#define XEN_PUT_VCPU_INFO_fixup .byte 0xff,0xff,0xff
+#else
+#define XEN_GET_VCPU_INFO(reg)	movq HYPERVISOR_shared_info,reg
+#define XEN_PUT_VCPU_INFO(reg)
+#define XEN_PUT_VCPU_INFO_fixup
+#endif
+
+#define XEN_LOCKED_BLOCK_EVENTS(reg)	movb $1,evtchn_upcall_mask(reg)
+#define XEN_LOCKED_UNBLOCK_EVENTS(reg)	movb $0,evtchn_upcall_mask(reg)
+#define XEN_BLOCK_EVENTS(reg)	XEN_GET_VCPU_INFO(reg)			; \
+				XEN_LOCKED_BLOCK_EVENTS(reg)		; \
+    				XEN_PUT_VCPU_INFO(reg)
+#define XEN_UNBLOCK_EVENTS(reg)	XEN_GET_VCPU_INFO(reg)			; \
+				XEN_LOCKED_UNBLOCK_EVENTS(reg)		; \
+    				XEN_PUT_VCPU_INFO(reg)
+#define XEN_TEST_PENDING(reg)	testb $0xFF,evtchn_upcall_pending(reg)
diff -rpuN linux-2.6.18.8/arch/x86_64/lib/Makefile linux-2.6.18-xen-3.3.0/arch/x86_64/lib/Makefile
--- linux-2.6.18.8/arch/x86_64/lib/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/lib/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -10,3 +10,4 @@ lib-y := csum-partial.o csum-copy.o csum
 	usercopy.o getuser.o putuser.o  \
 	thunk.o clear_page.o copy_page.o bitstr.o bitops.o
 lib-y += memcpy.o memmove.o memset.o copy_user.o
+lib-$(CONFIG_XEN_SCRUB_PAGES) += scrub.o
diff -rpuN linux-2.6.18.8/arch/x86_64/lib/scrub.c linux-2.6.18-xen-3.3.0/arch/x86_64/lib/scrub.c
--- linux-2.6.18.8/arch/x86_64/lib/scrub.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/lib/scrub.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1 @@
+#include "../../i386/lib/scrub.c"
diff -rpuN linux-2.6.18.8/arch/x86_64/Makefile linux-2.6.18-xen-3.3.0/arch/x86_64/Makefile
--- linux-2.6.18.8/arch/x86_64/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -71,9 +71,22 @@ drivers-$(CONFIG_OPROFILE)		+= arch/x86_
 
 boot := arch/x86_64/boot
 
-PHONY += bzImage bzlilo install archmrproper \
+PHONY += bzImage bzlilo vmlinuz install archmrproper \
 	 fdimage fdimage144 fdimage288 isoimage archclean
 
+ifdef CONFIG_XEN
+CPPFLAGS := -D__XEN_INTERFACE_VERSION__=$(CONFIG_XEN_INTERFACE_VERSION) \
+	-Iinclude$(if $(KBUILD_SRC),2)/asm/mach-xen $(CPPFLAGS)
+LDFLAGS_vmlinux := -e startup_64
+#Default target when executing "make"
+all: vmlinuz
+
+BOOTIMAGE                     := $(boot)/vmlinuz
+KBUILD_IMAGE                  := $(BOOTIMAGE)
+
+vmlinuz: vmlinux
+	$(Q)$(MAKE) $(build)=$(boot) $(BOOTIMAGE)
+else
 #Default target when executing "make"
 all: bzImage
 
@@ -91,6 +104,7 @@ bzdisk: vmlinux
 
 fdimage fdimage144 fdimage288 isoimage: vmlinux
 	$(Q)$(MAKE) $(build)=$(boot) BOOTIMAGE=$(BOOTIMAGE) $@
+endif
 
 install:
 	$(Q)$(MAKE) $(build)=$(boot) BOOTIMAGE=$(BOOTIMAGE) $@ 
diff -rpuN linux-2.6.18.8/arch/x86_64/mm/fault-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/mm/fault-xen.c
--- linux-2.6.18.8/arch/x86_64/mm/fault-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/mm/fault-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,724 @@
+/*
+ *  linux/arch/x86-64/mm/fault.c
+ *
+ *  Copyright (C) 1995  Linus Torvalds
+ *  Copyright (C) 2001,2002 Andi Kleen, SuSE Labs.
+ */
+
+#include <linux/signal.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/ptrace.h>
+#include <linux/mman.h>
+#include <linux/mm.h>
+#include <linux/smp.h>
+#include <linux/smp_lock.h>
+#include <linux/interrupt.h>
+#include <linux/init.h>
+#include <linux/tty.h>
+#include <linux/vt_kern.h>		/* For unblank_screen() */
+#include <linux/compiler.h>
+#include <linux/module.h>
+#include <linux/kprobes.h>
+
+#include <asm/system.h>
+#include <asm/uaccess.h>
+#include <asm/pgalloc.h>
+#include <asm/smp.h>
+#include <asm/tlbflush.h>
+#include <asm/proto.h>
+#include <asm/kdebug.h>
+#include <asm-generic/sections.h>
+
+/* Page fault error code bits */
+#define PF_PROT	(1<<0)		/* or no page found */
+#define PF_WRITE	(1<<1)
+#define PF_USER	(1<<2)
+#define PF_RSVD	(1<<3)
+#define PF_INSTR	(1<<4)
+
+#ifdef CONFIG_KPROBES
+ATOMIC_NOTIFIER_HEAD(notify_page_fault_chain);
+
+/* Hook to register for page fault notifications */
+int register_page_fault_notifier(struct notifier_block *nb)
+{
+	vmalloc_sync_all();
+	return atomic_notifier_chain_register(&notify_page_fault_chain, nb);
+}
+
+int unregister_page_fault_notifier(struct notifier_block *nb)
+{
+	return atomic_notifier_chain_unregister(&notify_page_fault_chain, nb);
+}
+
+static inline int notify_page_fault(enum die_val val, const char *str,
+			struct pt_regs *regs, long err, int trap, int sig)
+{
+	struct die_args args = {
+		.regs = regs,
+		.str = str,
+		.err = err,
+		.trapnr = trap,
+		.signr = sig
+	};
+	return atomic_notifier_call_chain(&notify_page_fault_chain, val, &args);
+}
+#else
+static inline int notify_page_fault(enum die_val val, const char *str,
+			struct pt_regs *regs, long err, int trap, int sig)
+{
+	return NOTIFY_DONE;
+}
+#endif
+
+void bust_spinlocks(int yes)
+{
+	int loglevel_save = console_loglevel;
+	if (yes) {
+		oops_in_progress = 1;
+	} else {
+#ifdef CONFIG_VT
+		unblank_screen();
+#endif
+		oops_in_progress = 0;
+		/*
+		 * OK, the message is on the console.  Now we call printk()
+		 * without oops_in_progress set so that printk will give klogd
+		 * a poke.  Hold onto your hats...
+		 */
+		console_loglevel = 15;		/* NMI oopser may have shut the console up */
+		printk(" ");
+		console_loglevel = loglevel_save;
+	}
+}
+
+/* Sometimes the CPU reports invalid exceptions on prefetch.
+   Check that here and ignore.
+   Opcode checker based on code by Richard Brunner */
+static noinline int is_prefetch(struct pt_regs *regs, unsigned long addr,
+				unsigned long error_code)
+{ 
+	unsigned char *instr;
+	int scan_more = 1;
+	int prefetch = 0; 
+	unsigned char *max_instr;
+
+	/* If it was a exec fault ignore */
+	if (error_code & PF_INSTR)
+		return 0;
+	
+	instr = (unsigned char *)convert_rip_to_linear(current, regs);
+	max_instr = instr + 15;
+
+	if (user_mode(regs) && instr >= (unsigned char *)TASK_SIZE)
+		return 0;
+
+	while (scan_more && instr < max_instr) { 
+		unsigned char opcode;
+		unsigned char instr_hi;
+		unsigned char instr_lo;
+
+		if (__get_user(opcode, instr))
+			break; 
+
+		instr_hi = opcode & 0xf0; 
+		instr_lo = opcode & 0x0f; 
+		instr++;
+
+		switch (instr_hi) { 
+		case 0x20:
+		case 0x30:
+			/* Values 0x26,0x2E,0x36,0x3E are valid x86
+			   prefixes.  In long mode, the CPU will signal
+			   invalid opcode if some of these prefixes are
+			   present so we will never get here anyway */
+			scan_more = ((instr_lo & 7) == 0x6);
+			break;
+			
+		case 0x40:
+			/* In AMD64 long mode, 0x40 to 0x4F are valid REX prefixes
+			   Need to figure out under what instruction mode the
+			   instruction was issued ... */
+			/* Could check the LDT for lm, but for now it's good
+			   enough to assume that long mode only uses well known
+			   segments or kernel. */
+			scan_more = (!user_mode(regs)) || (regs->cs == __USER_CS);
+			break;
+			
+		case 0x60:
+			/* 0x64 thru 0x67 are valid prefixes in all modes. */
+			scan_more = (instr_lo & 0xC) == 0x4;
+			break;		
+		case 0xF0:
+			/* 0xF0, 0xF2, and 0xF3 are valid prefixes in all modes. */
+			scan_more = !instr_lo || (instr_lo>>1) == 1;
+			break;			
+		case 0x00:
+			/* Prefetch instruction is 0x0F0D or 0x0F18 */
+			scan_more = 0;
+			if (__get_user(opcode, instr)) 
+				break;
+			prefetch = (instr_lo == 0xF) &&
+				(opcode == 0x0D || opcode == 0x18);
+			break;			
+		default:
+			scan_more = 0;
+			break;
+		} 
+	}
+	return prefetch;
+}
+
+static int bad_address(void *p) 
+{ 
+	unsigned long dummy;
+	return __get_user(dummy, (unsigned long *)p);
+} 
+
+void dump_pagetable(unsigned long address)
+{
+	pgd_t *pgd;
+	pud_t *pud;
+	pmd_t *pmd;
+	pte_t *pte;
+
+	pgd = __va(read_cr3() & PHYSICAL_PAGE_MASK);
+	pgd += pgd_index(address);
+	if (bad_address(pgd)) goto bad;
+	printk("PGD %lx ", pgd_val(*pgd));
+	if (!pgd_present(*pgd)) goto ret; 
+
+	pud = pud_offset(pgd, address);
+	if (bad_address(pud)) goto bad;
+	printk("PUD %lx ", pud_val(*pud));
+	if (!pud_present(*pud))	goto ret;
+
+	pmd = pmd_offset(pud, address);
+	if (bad_address(pmd)) goto bad;
+	printk("PMD %lx ", pmd_val(*pmd));
+	if (!pmd_present(*pmd))	goto ret;	 
+
+	pte = pte_offset_kernel(pmd, address);
+	if (bad_address(pte)) goto bad;
+	printk("PTE %lx", pte_val(*pte)); 
+ret:
+	printk("\n");
+	return;
+bad:
+	printk("BAD\n");
+}
+
+static const char errata93_warning[] = 
+KERN_ERR "******* Your BIOS seems to not contain a fix for K8 errata #93\n"
+KERN_ERR "******* Working around it, but it may cause SEGVs or burn power.\n"
+KERN_ERR "******* Please consider a BIOS update.\n"
+KERN_ERR "******* Disabling USB legacy in the BIOS may also help.\n";
+
+/* Workaround for K8 erratum #93 & buggy BIOS.
+   BIOS SMM functions are required to use a specific workaround
+   to avoid corruption of the 64bit RIP register on C stepping K8. 
+   A lot of BIOS that didn't get tested properly miss this. 
+   The OS sees this as a page fault with the upper 32bits of RIP cleared.
+   Try to work around it here.
+   Note we only handle faults in kernel here. */
+
+static int is_errata93(struct pt_regs *regs, unsigned long address) 
+{
+	static int warned;
+	if (address != regs->rip)
+		return 0;
+	if ((address >> 32) != 0) 
+		return 0;
+	address |= 0xffffffffUL << 32;
+	if ((address >= (u64)_stext && address <= (u64)_etext) || 
+	    (address >= MODULES_VADDR && address <= MODULES_END)) { 
+		if (!warned) {
+			printk(errata93_warning); 		
+			warned = 1;
+		}
+		regs->rip = address;
+		return 1;
+	}
+	return 0;
+} 
+
+int unhandled_signal(struct task_struct *tsk, int sig)
+{
+	if (tsk->pid == 1)
+		return 1;
+	if (tsk->ptrace & PT_PTRACED)
+		return 0;
+	return (tsk->sighand->action[sig-1].sa.sa_handler == SIG_IGN) ||
+		(tsk->sighand->action[sig-1].sa.sa_handler == SIG_DFL);
+}
+
+static noinline void pgtable_bad(unsigned long address, struct pt_regs *regs,
+				 unsigned long error_code)
+{
+	unsigned long flags = oops_begin();
+	struct task_struct *tsk;
+
+	printk(KERN_ALERT "%s: Corrupted page table at address %lx\n",
+	       current->comm, address);
+	dump_pagetable(address);
+	tsk = current;
+	tsk->thread.cr2 = address;
+	tsk->thread.trap_no = 14;
+	tsk->thread.error_code = error_code;
+	__die("Bad pagetable", regs, error_code);
+	oops_end(flags);
+	do_exit(SIGKILL);
+}
+
+/*
+ * Handle a fault on the vmalloc area
+ *
+ * This assumes no large pages in there.
+ */
+static int vmalloc_fault(unsigned long address)
+{
+	pgd_t *pgd, *pgd_ref;
+	pud_t *pud, *pud_ref;
+	pmd_t *pmd, *pmd_ref;
+	pte_t *pte, *pte_ref;
+
+	/* Copy kernel mappings over when needed. This can also
+	   happen within a race in page table update. In the later
+	   case just flush. */
+
+	/* On Xen the line below does not always work. Needs investigating! */
+	/*pgd = pgd_offset(current->mm ?: &init_mm, address);*/
+	pgd = __va(read_cr3() & PHYSICAL_PAGE_MASK);
+	pgd += pgd_index(address);
+	pgd_ref = pgd_offset_k(address);
+	if (pgd_none(*pgd_ref))
+		return -1;
+	if (pgd_none(*pgd))
+		set_pgd(pgd, *pgd_ref);
+	else
+		BUG_ON(pgd_page(*pgd) != pgd_page(*pgd_ref));
+
+	/* Below here mismatches are bugs because these lower tables
+	   are shared */
+
+	pud = pud_offset(pgd, address);
+	pud_ref = pud_offset(pgd_ref, address);
+	if (pud_none(*pud_ref))
+		return -1;
+	if (pud_none(*pud) || pud_page(*pud) != pud_page(*pud_ref))
+		BUG();
+	pmd = pmd_offset(pud, address);
+	pmd_ref = pmd_offset(pud_ref, address);
+	if (pmd_none(*pmd_ref))
+		return -1;
+	if (pmd_none(*pmd) || pmd_page(*pmd) != pmd_page(*pmd_ref))
+		BUG();
+	pte_ref = pte_offset_kernel(pmd_ref, address);
+	if (!pte_present(*pte_ref))
+		return -1;
+	pte = pte_offset_kernel(pmd, address);
+	/* Don't use pte_page here, because the mappings can point
+	   outside mem_map, and the NUMA hash lookup cannot handle
+	   that. */
+	if (!pte_present(*pte) || pte_pfn(*pte) != pte_pfn(*pte_ref))
+		BUG();
+	return 0;
+}
+
+int page_fault_trace = 0;
+int exception_trace = 1;
+
+
+#define MEM_VERBOSE 1
+
+#ifdef MEM_VERBOSE
+#define MEM_LOG(_f, _a...)			\
+	printk("fault.c:[%d]-> " _f "\n",	\
+	__LINE__ , ## _a )
+#else
+#define MEM_LOG(_f, _a...) ((void)0)
+#endif
+
+static int spurious_fault(struct pt_regs *regs,
+			  unsigned long address,
+			  unsigned long error_code)
+{
+	pgd_t *pgd;
+	pud_t *pud;
+	pmd_t *pmd;
+	pte_t *pte;
+
+#ifdef CONFIG_XEN
+	/* Faults in hypervisor area are never spurious. */
+	if ((address >= HYPERVISOR_VIRT_START) &&
+	    (address < HYPERVISOR_VIRT_END))
+		return 0;
+#endif
+
+	/* Reserved-bit violation or user access to kernel space? */
+	if (error_code & (PF_RSVD|PF_USER))
+		return 0;
+
+	pgd = init_mm.pgd + pgd_index(address);
+	if (!pgd_present(*pgd))
+		return 0;
+
+	pud = pud_offset(pgd, address);
+	if (!pud_present(*pud))
+		return 0;
+
+	pmd = pmd_offset(pud, address);
+	if (!pmd_present(*pmd))
+		return 0;
+
+	pte = pte_offset_kernel(pmd, address);
+	if (!pte_present(*pte))
+		return 0;
+	if ((error_code & PF_WRITE) && !pte_write(*pte))
+		return 0;
+	if ((error_code & PF_INSTR) && (__pte_val(*pte) & _PAGE_NX))
+		return 0;
+
+	return 1;
+}
+
+/*
+ * This routine handles page faults.  It determines the address,
+ * and the problem, and then passes it off to one of the appropriate
+ * routines.
+ */
+asmlinkage void __kprobes do_page_fault(struct pt_regs *regs,
+					unsigned long error_code)
+{
+	struct task_struct *tsk;
+	struct mm_struct *mm;
+	struct vm_area_struct * vma;
+	unsigned long address;
+	const struct exception_table_entry *fixup;
+	int write;
+	unsigned long flags;
+	siginfo_t info;
+
+	if (!user_mode(regs))
+		error_code &= ~PF_USER; /* means kernel */
+
+	tsk = current;
+	mm = tsk->mm;
+	prefetchw(&mm->mmap_sem);
+
+	/* get the address */
+	address = current_vcpu_info()->arch.cr2;
+
+	info.si_code = SEGV_MAPERR;
+
+
+	/*
+	 * We fault-in kernel-space virtual memory on-demand. The
+	 * 'reference' page table is init_mm.pgd.
+	 *
+	 * NOTE! We MUST NOT take any locks for this case. We may
+	 * be in an interrupt or a critical region, and should
+	 * only copy the information from the master page table,
+	 * nothing more.
+	 *
+	 * This verifies that the fault happens in kernel space
+	 * (error_code & 4) == 0, and that the fault was not a
+	 * protection error (error_code & 9) == 0.
+	 */
+	if (unlikely(address >= TASK_SIZE64)) {
+		/*
+		 * Don't check for the module range here: its PML4
+		 * is always initialized because it's shared with the main
+		 * kernel text. Only vmalloc may need PML4 syncups.
+		 */
+		if (!(error_code & (PF_RSVD|PF_USER|PF_PROT)) &&
+		      ((address >= VMALLOC_START && address < VMALLOC_END))) {
+			if (vmalloc_fault(address) >= 0)
+				return;
+		}
+		/* Can take a spurious fault if mapping changes R/O -> R/W. */
+		if (spurious_fault(regs, address, error_code))
+			return;
+		if (notify_page_fault(DIE_PAGE_FAULT, "page fault", regs, error_code, 14,
+						SIGSEGV) == NOTIFY_STOP)
+			return;
+		/*
+		 * Don't take the mm semaphore here. If we fixup a prefetch
+		 * fault we could otherwise deadlock.
+		 */
+		goto bad_area_nosemaphore;
+	}
+
+	if (notify_page_fault(DIE_PAGE_FAULT, "page fault", regs, error_code, 14,
+					SIGSEGV) == NOTIFY_STOP)
+		return;
+
+	if (likely(regs->eflags & X86_EFLAGS_IF))
+		local_irq_enable();
+
+	if (unlikely(page_fault_trace))
+		printk("pagefault rip:%lx rsp:%lx cs:%lu ss:%lu address %lx error %lx\n",
+		       regs->rip,regs->rsp,regs->cs,regs->ss,address,error_code); 
+
+	if (unlikely(error_code & PF_RSVD))
+		pgtable_bad(address, regs, error_code);
+
+	/*
+	 * If we're in an interrupt or have no user
+	 * context, we must not take the fault..
+	 */
+	if (unlikely(in_atomic() || !mm))
+		goto bad_area_nosemaphore;
+
+ again:
+	/* When running in the kernel we expect faults to occur only to
+	 * addresses in user space.  All other faults represent errors in the
+	 * kernel and should generate an OOPS.  Unfortunatly, in the case of an
+	 * erroneous fault occurring in a code path which already holds mmap_sem
+	 * we will deadlock attempting to validate the fault against the
+	 * address space.  Luckily the kernel only validly references user
+	 * space from well defined areas of code, which are listed in the
+	 * exceptions table.
+	 *
+	 * As the vast majority of faults will be valid we will only perform
+	 * the source reference check when there is a possibilty of a deadlock.
+	 * Attempt to lock the address space, if we cannot we then validate the
+	 * source.  If this is invalid we can skip the address space check,
+	 * thus avoiding the deadlock.
+	 */
+	if (!down_read_trylock(&mm->mmap_sem)) {
+		if ((error_code & PF_USER) == 0 &&
+		    !search_exception_tables(regs->rip))
+			goto bad_area_nosemaphore;
+		down_read(&mm->mmap_sem);
+	}
+
+	vma = find_vma(mm, address);
+	if (!vma)
+		goto bad_area;
+	if (likely(vma->vm_start <= address))
+		goto good_area;
+	if (!(vma->vm_flags & VM_GROWSDOWN))
+		goto bad_area;
+	if (error_code & 4) {
+		/* Allow userspace just enough access below the stack pointer
+		 * to let the 'enter' instruction work.
+		 */
+		if (address + 65536 + 32 * sizeof(unsigned long) < regs->rsp)
+			goto bad_area;
+	}
+	if (expand_stack(vma, address))
+		goto bad_area;
+/*
+ * Ok, we have a good vm_area for this memory access, so
+ * we can handle it..
+ */
+good_area:
+	info.si_code = SEGV_ACCERR;
+	write = 0;
+	switch (error_code & (PF_PROT|PF_WRITE)) {
+		default:	/* 3: write, present */
+			/* fall through */
+		case PF_WRITE:		/* write, not present */
+			if (!(vma->vm_flags & VM_WRITE))
+				goto bad_area;
+			write++;
+			break;
+		case PF_PROT:		/* read, present */
+			goto bad_area;
+		case 0:			/* read, not present */
+			if (!(vma->vm_flags & (VM_READ | VM_EXEC)))
+				goto bad_area;
+	}
+
+	/*
+	 * If for any reason at all we couldn't handle the fault,
+	 * make sure we exit gracefully rather than endlessly redo
+	 * the fault.
+	 */
+	switch (handle_mm_fault(mm, vma, address, write)) {
+	case VM_FAULT_MINOR:
+		tsk->min_flt++;
+		break;
+	case VM_FAULT_MAJOR:
+		tsk->maj_flt++;
+		break;
+	case VM_FAULT_SIGBUS:
+		goto do_sigbus;
+	default:
+		goto out_of_memory;
+	}
+
+	up_read(&mm->mmap_sem);
+	return;
+
+/*
+ * Something tried to access memory that isn't in our memory map..
+ * Fix it, but check if it's kernel or user first..
+ */
+bad_area:
+	up_read(&mm->mmap_sem);
+
+bad_area_nosemaphore:
+	/* User mode accesses just cause a SIGSEGV */
+	if (error_code & PF_USER) {
+		if (is_prefetch(regs, address, error_code))
+			return;
+
+		/* Work around K8 erratum #100 K8 in compat mode
+		   occasionally jumps to illegal addresses >4GB.  We
+		   catch this here in the page fault handler because
+		   these addresses are not reachable. Just detect this
+		   case and return.  Any code segment in LDT is
+		   compatibility mode. */
+		if ((regs->cs == __USER32_CS || (regs->cs & (1<<2))) &&
+		    (address >> 32))
+			return;
+
+		if (exception_trace && unhandled_signal(tsk, SIGSEGV)) {
+			printk(
+		       "%s%s[%d]: segfault at %016lx rip %016lx rsp %016lx error %lx\n",
+					tsk->pid > 1 ? KERN_INFO : KERN_EMERG,
+					tsk->comm, tsk->pid, address, regs->rip,
+					regs->rsp, error_code);
+		}
+       
+		tsk->thread.cr2 = address;
+		/* Kernel addresses are always protection faults */
+		tsk->thread.error_code = error_code | (address >= TASK_SIZE);
+		tsk->thread.trap_no = 14;
+		info.si_signo = SIGSEGV;
+		info.si_errno = 0;
+		/* info.si_code has been set above */
+		info.si_addr = (void __user *)address;
+		force_sig_info(SIGSEGV, &info, tsk);
+		return;
+	}
+
+no_context:
+	
+	/* Are we prepared to handle this kernel fault?  */
+	fixup = search_exception_tables(regs->rip);
+	if (fixup) {
+		regs->rip = fixup->fixup;
+		return;
+	}
+
+	/* 
+	 * Hall of shame of CPU/BIOS bugs.
+	 */
+
+ 	if (is_prefetch(regs, address, error_code))
+ 		return;
+
+	if (is_errata93(regs, address))
+		return; 
+
+/*
+ * Oops. The kernel tried to access some bad page. We'll have to
+ * terminate things with extreme prejudice.
+ */
+
+	flags = oops_begin();
+
+	if (address < PAGE_SIZE)
+		printk(KERN_ALERT "Unable to handle kernel NULL pointer dereference");
+	else
+		printk(KERN_ALERT "Unable to handle kernel paging request");
+	printk(" at %016lx RIP: \n" KERN_ALERT,address);
+	printk_address(regs->rip);
+	dump_pagetable(address);
+	tsk->thread.cr2 = address;
+	tsk->thread.trap_no = 14;
+	tsk->thread.error_code = error_code;
+	__die("Oops", regs, error_code);
+	/* Executive summary in case the body of the oops scrolled away */
+	printk(KERN_EMERG "CR2: %016lx\n", address);
+	oops_end(flags);
+	do_exit(SIGKILL);
+
+/*
+ * We ran out of memory, or some other thing happened to us that made
+ * us unable to handle the page fault gracefully.
+ */
+out_of_memory:
+	up_read(&mm->mmap_sem);
+	if (current->pid == 1) { 
+		yield();
+		goto again;
+	}
+	printk("VM: killing process %s\n", tsk->comm);
+	if (error_code & 4)
+		do_exit(SIGKILL);
+	goto no_context;
+
+do_sigbus:
+	up_read(&mm->mmap_sem);
+
+	/* Kernel mode? Handle exceptions or die */
+	if (!(error_code & PF_USER))
+		goto no_context;
+
+	tsk->thread.cr2 = address;
+	tsk->thread.error_code = error_code;
+	tsk->thread.trap_no = 14;
+	info.si_signo = SIGBUS;
+	info.si_errno = 0;
+	info.si_code = BUS_ADRERR;
+	info.si_addr = (void __user *)address;
+	force_sig_info(SIGBUS, &info, tsk);
+	return;
+}
+
+DEFINE_SPINLOCK(pgd_lock);
+struct page *pgd_list;
+
+void vmalloc_sync_all(void)
+{
+	/* Note that races in the updates of insync and start aren't 
+	   problematic:
+	   insync can only get set bits added, and updates to start are only
+	   improving performance (without affecting correctness if undone). */
+	static DECLARE_BITMAP(insync, PTRS_PER_PGD);
+	static unsigned long start = VMALLOC_START & PGDIR_MASK;
+	unsigned long address;
+
+	for (address = start; address <= VMALLOC_END; address += PGDIR_SIZE) {
+		if (!test_bit(pgd_index(address), insync)) {
+			const pgd_t *pgd_ref = pgd_offset_k(address);
+			struct page *page;
+
+			if (pgd_none(*pgd_ref))
+				continue;
+			spin_lock(&pgd_lock);
+			for (page = pgd_list; page;
+			     page = (struct page *)page->index) {
+				pgd_t *pgd;
+				pgd = (pgd_t *)page_address(page) + pgd_index(address);
+				if (pgd_none(*pgd))
+					set_pgd(pgd, *pgd_ref);
+				else
+					BUG_ON(pgd_page(*pgd) != pgd_page(*pgd_ref));
+			}
+			spin_unlock(&pgd_lock);
+			set_bit(pgd_index(address), insync);
+		}
+		if (address == start)
+			start = address + PGDIR_SIZE;
+	}
+	/* Check that there is no need to do the same for the modules area. */
+	BUILD_BUG_ON(!(MODULES_VADDR > __START_KERNEL));
+	BUILD_BUG_ON(!(((MODULES_END - 1) & PGDIR_MASK) == 
+				(__START_KERNEL & PGDIR_MASK)));
+}
+
+static int __init enable_pagefaulttrace(char *str)
+{
+	page_fault_trace = 1;
+	return 1;
+}
+__setup("pagefaulttrace", enable_pagefaulttrace);
diff -rpuN linux-2.6.18.8/arch/x86_64/mm/init-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/mm/init-xen.c
--- linux-2.6.18.8/arch/x86_64/mm/init-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/mm/init-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1213 @@
+/*
+ *  linux/arch/x86_64/mm/init.c
+ *
+ *  Copyright (C) 1995  Linus Torvalds
+ *  Copyright (C) 2000  Pavel Machek <pavel@suse.cz>
+ *  Copyright (C) 2002,2003 Andi Kleen <ak@suse.de>
+ *
+ *  Jun Nakajima <jun.nakajima@intel.com>
+ *	Modified for Xen.
+ */
+
+#include <linux/signal.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/ptrace.h>
+#include <linux/mman.h>
+#include <linux/mm.h>
+#include <linux/swap.h>
+#include <linux/smp.h>
+#include <linux/init.h>
+#include <linux/pagemap.h>
+#include <linux/bootmem.h>
+#include <linux/proc_fs.h>
+#include <linux/pci.h>
+#include <linux/poison.h>
+#include <linux/dma-mapping.h>
+#include <linux/module.h>
+#include <linux/memory_hotplug.h>
+
+#include <asm/processor.h>
+#include <asm/system.h>
+#include <asm/uaccess.h>
+#include <asm/pgtable.h>
+#include <asm/pgalloc.h>
+#include <asm/dma.h>
+#include <asm/fixmap.h>
+#include <asm/e820.h>
+#include <asm/apic.h>
+#include <asm/tlb.h>
+#include <asm/mmu_context.h>
+#include <asm/proto.h>
+#include <asm/smp.h>
+#include <asm/sections.h>
+
+#include <xen/features.h>
+
+#ifndef Dprintk
+#define Dprintk(x...)
+#endif
+
+struct dma_mapping_ops* dma_ops;
+EXPORT_SYMBOL(dma_ops);
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+unsigned int __kernel_page_user;
+EXPORT_SYMBOL(__kernel_page_user);
+#endif
+
+int after_bootmem;
+
+extern unsigned long *contiguous_bitmap;
+
+static unsigned long dma_reserve __initdata;
+
+DEFINE_PER_CPU(struct mmu_gather, mmu_gathers);
+extern unsigned long start_pfn;
+
+/*
+ * Use this until direct mapping is established, i.e. before __va() is 
+ * available in init_memory_mapping().
+ */
+
+#define addr_to_page(addr, page)				\
+	(addr) &= PHYSICAL_PAGE_MASK;				\
+	(page) = ((unsigned long *) ((unsigned long)		\
+	(((mfn_to_pfn((addr) >> PAGE_SHIFT)) << PAGE_SHIFT) +	\
+	__START_KERNEL_map)))
+
+static void __meminit early_make_page_readonly(void *va, unsigned int feature)
+{
+	unsigned long addr, _va = (unsigned long)va;
+	pte_t pte, *ptep;
+	unsigned long *page = (unsigned long *) init_level4_pgt;
+
+	BUG_ON(after_bootmem);
+
+	if (xen_feature(feature))
+		return;
+
+	addr = (unsigned long) page[pgd_index(_va)];
+	addr_to_page(addr, page);
+
+	addr = page[pud_index(_va)];
+	addr_to_page(addr, page);
+
+	addr = page[pmd_index(_va)];
+	addr_to_page(addr, page);
+
+	ptep = (pte_t *) &page[pte_index(_va)];
+
+	pte.pte = ptep->pte & ~_PAGE_RW;
+	if (HYPERVISOR_update_va_mapping(_va, pte, 0))
+		BUG();
+}
+
+static void __make_page_readonly(void *va)
+{
+	pgd_t *pgd; pud_t *pud; pmd_t *pmd; pte_t pte, *ptep;
+	unsigned long addr = (unsigned long) va;
+
+	pgd = pgd_offset_k(addr);
+	pud = pud_offset(pgd, addr);
+	pmd = pmd_offset(pud, addr);
+	ptep = pte_offset_kernel(pmd, addr);
+
+	pte.pte = ptep->pte & ~_PAGE_RW;
+	if (HYPERVISOR_update_va_mapping(addr, pte, 0))
+		xen_l1_entry_update(ptep, pte); /* fallback */
+
+	if ((addr >= VMALLOC_START) && (addr < VMALLOC_END))
+		__make_page_readonly(__va(pte_pfn(pte) << PAGE_SHIFT));
+}
+
+static void __make_page_writable(void *va)
+{
+	pgd_t *pgd; pud_t *pud; pmd_t *pmd; pte_t pte, *ptep;
+	unsigned long addr = (unsigned long) va;
+
+	pgd = pgd_offset_k(addr);
+	pud = pud_offset(pgd, addr);
+	pmd = pmd_offset(pud, addr);
+	ptep = pte_offset_kernel(pmd, addr);
+
+	pte.pte = ptep->pte | _PAGE_RW;
+	if (HYPERVISOR_update_va_mapping(addr, pte, 0))
+		xen_l1_entry_update(ptep, pte); /* fallback */
+
+	if ((addr >= VMALLOC_START) && (addr < VMALLOC_END))
+		__make_page_writable(__va(pte_pfn(pte) << PAGE_SHIFT));
+}
+
+void make_page_readonly(void *va, unsigned int feature)
+{
+	if (!xen_feature(feature))
+		__make_page_readonly(va);
+}
+
+void make_page_writable(void *va, unsigned int feature)
+{
+	if (!xen_feature(feature))
+		__make_page_writable(va);
+}
+
+void make_pages_readonly(void *va, unsigned nr, unsigned int feature)
+{
+	if (xen_feature(feature))
+		return;
+
+	while (nr-- != 0) {
+		__make_page_readonly(va);
+		va = (void*)((unsigned long)va + PAGE_SIZE);
+	}
+}
+
+void make_pages_writable(void *va, unsigned nr, unsigned int feature)
+{
+	if (xen_feature(feature))
+		return;
+
+	while (nr-- != 0) {
+		__make_page_writable(va);
+		va = (void*)((unsigned long)va + PAGE_SIZE);
+	}
+}
+
+/*
+ * NOTE: pagetable_init alloc all the fixmap pagetables contiguous on the
+ * physical space so we can cache the place of the first one and move
+ * around without checking the pgd every time.
+ */
+
+void show_mem(void)
+{
+	long i, total = 0, reserved = 0;
+	long shared = 0, cached = 0;
+	pg_data_t *pgdat;
+	struct page *page;
+
+	printk(KERN_INFO "Mem-info:\n");
+	show_free_areas();
+	printk(KERN_INFO "Free swap:       %6ldkB\n", nr_swap_pages<<(PAGE_SHIFT-10));
+
+	for_each_online_pgdat(pgdat) {
+               for (i = 0; i < pgdat->node_spanned_pages; ++i) {
+			page = pfn_to_page(pgdat->node_start_pfn + i);
+			total++;
+			if (PageReserved(page))
+				reserved++;
+			else if (PageSwapCache(page))
+				cached++;
+			else if (page_count(page))
+				shared += page_count(page) - 1;
+               }
+	}
+	printk(KERN_INFO "%lu pages of RAM\n", total);
+	printk(KERN_INFO "%lu reserved pages\n",reserved);
+	printk(KERN_INFO "%lu pages shared\n",shared);
+	printk(KERN_INFO "%lu pages swap cached\n",cached);
+}
+
+
+static __init void *spp_getpage(void)
+{ 
+	void *ptr;
+	if (after_bootmem)
+		ptr = (void *) get_zeroed_page(GFP_ATOMIC); 
+	else if (start_pfn < table_end) {
+		ptr = __va(start_pfn << PAGE_SHIFT);
+		start_pfn++;
+		memset(ptr, 0, PAGE_SIZE);
+	} else
+		ptr = alloc_bootmem_pages(PAGE_SIZE);
+	if (!ptr || ((unsigned long)ptr & ~PAGE_MASK))
+		panic("set_pte_phys: cannot allocate page data %s\n", after_bootmem?"after bootmem":"");
+
+	Dprintk("spp_getpage %p\n", ptr);
+	return ptr;
+} 
+
+#define pgd_offset_u(address) (__user_pgd(init_level4_pgt) + pgd_index(address))
+#define pud_offset_u(address) (level3_user_pgt + pud_index(address))
+
+static __init void set_pte_phys(unsigned long vaddr,
+			 unsigned long phys, pgprot_t prot, int user_mode)
+{
+	pgd_t *pgd;
+	pud_t *pud;
+	pmd_t *pmd;
+	pte_t *pte, new_pte;
+
+	Dprintk("set_pte_phys %lx to %lx\n", vaddr, phys);
+
+	pgd = (user_mode ? pgd_offset_u(vaddr) : pgd_offset_k(vaddr));
+	if (pgd_none(*pgd)) {
+		printk("PGD FIXMAP MISSING, it should be setup in head.S!\n");
+		return;
+	}
+	pud = (user_mode ? pud_offset_u(vaddr) : pud_offset(pgd, vaddr));
+	if (pud_none(*pud)) {
+		pmd = (pmd_t *) spp_getpage(); 
+		make_page_readonly(pmd, XENFEAT_writable_page_tables);
+		set_pud(pud, __pud(__pa(pmd) | _KERNPG_TABLE | _PAGE_USER));
+		if (pmd != pmd_offset(pud, 0)) {
+			printk("PAGETABLE BUG #01! %p <-> %p\n", pmd, pmd_offset(pud,0));
+			return;
+		}
+	}
+	pmd = pmd_offset(pud, vaddr);
+	if (pmd_none(*pmd)) {
+		pte = (pte_t *) spp_getpage();
+		make_page_readonly(pte, XENFEAT_writable_page_tables);
+		set_pmd(pmd, __pmd(__pa(pte) | _KERNPG_TABLE | _PAGE_USER));
+		if (pte != pte_offset_kernel(pmd, 0)) {
+			printk("PAGETABLE BUG #02!\n");
+			return;
+		}
+	}
+	if (pgprot_val(prot))
+		new_pte = pfn_pte(phys >> PAGE_SHIFT, prot);
+	else
+		new_pte = __pte(0);
+
+	pte = pte_offset_kernel(pmd, vaddr);
+	if (!pte_none(*pte) &&
+	    __pte_val(*pte) != (__pte_val(new_pte) & __supported_pte_mask))
+		pte_ERROR(*pte);
+	set_pte(pte, new_pte);
+
+	/*
+	 * It's enough to flush this one mapping.
+	 * (PGE mappings get flushed as well)
+	 */
+	__flush_tlb_one(vaddr);
+}
+
+static __init void set_pte_phys_ma(unsigned long vaddr,
+				   unsigned long phys, pgprot_t prot)
+{
+	pgd_t *pgd;
+	pud_t *pud;
+	pmd_t *pmd;
+	pte_t *pte, new_pte;
+
+	Dprintk("set_pte_phys %lx to %lx\n", vaddr, phys);
+
+	pgd = pgd_offset_k(vaddr);
+	if (pgd_none(*pgd)) {
+		printk("PGD FIXMAP MISSING, it should be setup in head.S!\n");
+		return;
+	}
+	pud = pud_offset(pgd, vaddr);
+	if (pud_none(*pud)) {
+
+		pmd = (pmd_t *) spp_getpage(); 
+		make_page_readonly(pmd, XENFEAT_writable_page_tables);
+		set_pud(pud, __pud(__pa(pmd) | _KERNPG_TABLE | _PAGE_USER));
+		if (pmd != pmd_offset(pud, 0)) {
+			printk("PAGETABLE BUG #01! %p <-> %p\n", pmd, pmd_offset(pud,0));
+			return;
+		}
+	}
+	pmd = pmd_offset(pud, vaddr);
+	if (pmd_none(*pmd)) {
+		pte = (pte_t *) spp_getpage();
+		make_page_readonly(pte, XENFEAT_writable_page_tables);
+		set_pmd(pmd, __pmd(__pa(pte) | _KERNPG_TABLE | _PAGE_USER));
+		if (pte != pte_offset_kernel(pmd, 0)) {
+			printk("PAGETABLE BUG #02!\n");
+			return;
+		}
+	}
+	new_pte = pfn_pte_ma(phys >> PAGE_SHIFT, prot);
+
+	pte = pte_offset_kernel(pmd, vaddr);
+	if (!pte_none(*pte) &&
+#ifdef CONFIG_ACPI
+	    /* __acpi_map_table() fails to properly call clear_fixmap() */
+	    (vaddr < __fix_to_virt(FIX_ACPI_END) ||
+	     vaddr > __fix_to_virt(FIX_ACPI_BEGIN)) &&
+#endif
+	    __pte_val(*pte) != (__pte_val(new_pte) & __supported_pte_mask))
+		pte_ERROR(*pte);
+	set_pte(pte, new_pte);
+
+	/*
+	 * It's enough to flush this one mapping.
+	 * (PGE mappings get flushed as well)
+	 */
+	__flush_tlb_one(vaddr);
+}
+
+/* NOTE: this is meant to be run only at boot */
+void __init 
+__set_fixmap (enum fixed_addresses idx, unsigned long phys, pgprot_t prot)
+{
+	unsigned long address = __fix_to_virt(idx);
+
+	if (idx >= __end_of_fixed_addresses) {
+		printk("Invalid __set_fixmap\n");
+		return;
+	}
+	switch (idx) {
+	case VSYSCALL_LAST_PAGE ... VSYSCALL_FIRST_PAGE:
+		set_pte_phys(address, phys, prot, 0);
+		set_pte_phys(address, phys, prot, 1);
+		break;
+	default:
+		set_pte_phys_ma(address, phys, prot);
+		break;
+	}
+}
+
+unsigned long __initdata table_start, table_end; 
+
+static __meminit void *alloc_static_page(unsigned long *phys)
+{
+	unsigned long va = (start_pfn << PAGE_SHIFT) + __START_KERNEL_map;
+
+	if (after_bootmem) {
+		void *adr = (void *)get_zeroed_page(GFP_ATOMIC);
+
+		*phys = __pa(adr);
+		return adr;
+	}
+
+	*phys = start_pfn << PAGE_SHIFT;
+	start_pfn++;
+	memset((void *)va, 0, PAGE_SIZE);
+	return (void *)va;
+} 
+
+#define PTE_SIZE PAGE_SIZE
+
+static inline int make_readonly(unsigned long paddr)
+{
+	extern char __vsyscall_0;
+	int readonly = 0;
+
+	/* Make new page tables read-only. */
+	if (!xen_feature(XENFEAT_writable_page_tables)
+	    && (paddr >= (table_start << PAGE_SHIFT))
+	    && (paddr < (table_end << PAGE_SHIFT)))
+		readonly = 1;
+	/* Make old page tables read-only. */
+	if (!xen_feature(XENFEAT_writable_page_tables)
+	    && (paddr >= (xen_start_info->pt_base - __START_KERNEL_map))
+	    && (paddr < (start_pfn << PAGE_SHIFT)))
+		readonly = 1;
+
+	/*
+	 * No need for writable mapping of kernel image. This also ensures that
+	 * page and descriptor tables embedded inside don't have writable
+	 * mappings. Exclude the vsyscall area here, allowing alternative
+	 * instruction patching to work.
+	 */
+	if ((paddr >= __pa_symbol(&_text)) && (paddr < __pa_symbol(&_end))
+	    && !(paddr >= __pa_symbol(&__vsyscall_0)
+	         && paddr < __pa_symbol(&__vsyscall_0) + PAGE_SIZE))
+		readonly = 1;
+
+	return readonly;
+}
+
+#ifndef CONFIG_XEN
+/* Must run before zap_low_mappings */
+__init void *early_ioremap(unsigned long addr, unsigned long size)
+{
+	unsigned long map = round_down(addr, LARGE_PAGE_SIZE);
+
+	/* actually usually some more */
+	if (size >= LARGE_PAGE_SIZE) {
+		printk("SMBIOS area too long %lu\n", size);
+		return NULL;
+	}
+	set_pmd(temp_mappings[0].pmd,  __pmd(map | _KERNPG_TABLE | _PAGE_PSE));
+	map += LARGE_PAGE_SIZE;
+	set_pmd(temp_mappings[1].pmd,  __pmd(map | _KERNPG_TABLE | _PAGE_PSE));
+	__flush_tlb();
+	return temp_mappings[0].address + (addr & (LARGE_PAGE_SIZE-1));
+}
+
+/* To avoid virtual aliases later */
+__init void early_iounmap(void *addr, unsigned long size)
+{
+	if ((void *)round_down((unsigned long)addr, LARGE_PAGE_SIZE) != temp_mappings[0].address)
+		printk("early_iounmap: bad address %p\n", addr);
+	set_pmd(temp_mappings[0].pmd, __pmd(0));
+	set_pmd(temp_mappings[1].pmd, __pmd(0));
+	__flush_tlb();
+}
+#endif
+
+static void __meminit
+phys_pmd_init(pmd_t *pmd, unsigned long address, unsigned long end)
+{
+	int i, k;
+
+	for (i = 0; i < PTRS_PER_PMD; pmd++, i++) {
+		unsigned long pte_phys;
+		pte_t *pte, *pte_save;
+
+		if (address >= end)
+			break;
+		pte = alloc_static_page(&pte_phys);
+		pte_save = pte;
+		for (k = 0; k < PTRS_PER_PTE; pte++, k++, address += PTE_SIZE) {
+			unsigned long pteval = address | _PAGE_NX | _KERNPG_TABLE;
+
+			if (address >= (after_bootmem
+			                ? end
+			                : xen_start_info->nr_pages << PAGE_SHIFT))
+				pteval = 0;
+			else if (make_readonly(address))
+				pteval &= ~_PAGE_RW;
+			set_pte(pte, __pte(pteval & __supported_pte_mask));
+		}
+		if (!after_bootmem) {
+			early_make_page_readonly(pte_save, XENFEAT_writable_page_tables);
+			*pmd = __pmd(pte_phys | _KERNPG_TABLE);
+		} else {
+			make_page_readonly(pte_save, XENFEAT_writable_page_tables);
+			set_pmd(pmd, __pmd(pte_phys | _KERNPG_TABLE));
+		}
+	}
+}
+
+static void __meminit
+phys_pmd_update(pud_t *pud, unsigned long address, unsigned long end)
+{
+	pmd_t *pmd = pmd_offset(pud, (unsigned long)__va(address));
+
+	if (pmd_none(*pmd)) {
+		spin_lock(&init_mm.page_table_lock);
+		phys_pmd_init(pmd, address, end);
+		spin_unlock(&init_mm.page_table_lock);
+		__flush_tlb_all();
+	}
+}
+
+static void __meminit phys_pud_init(pud_t *pud, unsigned long address, unsigned long end)
+{ 
+	long i = pud_index(address);
+
+	pud = pud + i;
+
+	if (after_bootmem && pud_val(*pud)) {
+		phys_pmd_update(pud, address, end);
+		return;
+	}
+
+	for (; i < PTRS_PER_PUD; pud++, i++) {
+		unsigned long paddr, pmd_phys;
+		pmd_t *pmd;
+
+		paddr = (address & PGDIR_MASK) + i*PUD_SIZE;
+		if (paddr >= end)
+			break;
+
+		pmd = alloc_static_page(&pmd_phys);
+
+		spin_lock(&init_mm.page_table_lock);
+		*pud = __pud(pmd_phys | _KERNPG_TABLE);
+		phys_pmd_init(pmd, paddr, end);
+		spin_unlock(&init_mm.page_table_lock);
+
+		early_make_page_readonly(pmd, XENFEAT_writable_page_tables);
+	}
+	__flush_tlb();
+} 
+
+void __init xen_init_pt(void)
+{
+	unsigned long addr, *page;
+
+	/* Find the initial pte page that was built for us. */
+	page = (unsigned long *)xen_start_info->pt_base;
+	addr = page[pgd_index(__START_KERNEL_map)];
+	addr_to_page(addr, page);
+	addr = page[pud_index(__START_KERNEL_map)];
+	addr_to_page(addr, page);
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+	/* On Xen 3.0.2 and older we may need to explicitly specify _PAGE_USER
+	   in kernel PTEs. We check that here. */
+	if (HYPERVISOR_xen_version(XENVER_version, NULL) <= 0x30000) {
+		unsigned long *pg;
+		pte_t pte;
+
+		/* Mess with the initial mapping of page 0. It's not needed. */
+		BUILD_BUG_ON(__START_KERNEL <= __START_KERNEL_map);
+		addr = page[pmd_index(__START_KERNEL_map)];
+		addr_to_page(addr, pg);
+		pte.pte = pg[pte_index(__START_KERNEL_map)];
+		BUG_ON(!(pte.pte & _PAGE_PRESENT));
+
+		/* If _PAGE_USER isn't set, we obviously do not need it. */
+		if (pte.pte & _PAGE_USER) {
+			/* _PAGE_USER is needed, but is it set implicitly? */
+			pte.pte &= ~_PAGE_USER;
+			if ((HYPERVISOR_update_va_mapping(__START_KERNEL_map,
+							  pte, 0) != 0) ||
+			    !(pg[pte_index(__START_KERNEL_map)] & _PAGE_USER))
+				/* We need to explicitly specify _PAGE_USER. */
+				__kernel_page_user = _PAGE_USER;
+		}
+	}
+#endif
+
+	/* Construct mapping of initial pte page in our own directories. */
+	init_level4_pgt[pgd_index(__START_KERNEL_map)] = 
+		__pgd(__pa_symbol(level3_kernel_pgt) | _PAGE_TABLE);
+	level3_kernel_pgt[pud_index(__START_KERNEL_map)] = 
+		__pud(__pa_symbol(level2_kernel_pgt) | _PAGE_TABLE);
+	memcpy(level2_kernel_pgt, page, PAGE_SIZE);
+
+	__user_pgd(init_level4_pgt)[pgd_index(VSYSCALL_START)] =
+		__pgd(__pa_symbol(level3_user_pgt) | _PAGE_TABLE);
+
+	early_make_page_readonly(init_level4_pgt,
+				 XENFEAT_writable_page_tables);
+	early_make_page_readonly(__user_pgd(init_level4_pgt),
+				 XENFEAT_writable_page_tables);
+	early_make_page_readonly(level3_kernel_pgt,
+				 XENFEAT_writable_page_tables);
+	early_make_page_readonly(level3_user_pgt,
+				 XENFEAT_writable_page_tables);
+	early_make_page_readonly(level2_kernel_pgt,
+				 XENFEAT_writable_page_tables);
+
+	if (!xen_feature(XENFEAT_writable_page_tables)) {
+		xen_pgd_pin(__pa_symbol(init_level4_pgt));
+		xen_pgd_pin(__pa_symbol(__user_pgd(init_level4_pgt)));
+	}
+}
+
+static void __init extend_init_mapping(unsigned long tables_space)
+{
+	unsigned long va = __START_KERNEL_map;
+	unsigned long phys, addr, *pte_page;
+	pmd_t *pmd;
+	pte_t *pte, new_pte;
+	unsigned long *page = (unsigned long *)init_level4_pgt;
+
+	addr = page[pgd_index(va)];
+	addr_to_page(addr, page);
+	addr = page[pud_index(va)];
+	addr_to_page(addr, page);
+
+	/* Kill mapping of low 1MB. */
+	while (va < (unsigned long)&_text) {
+		if (HYPERVISOR_update_va_mapping(va, __pte_ma(0), 0))
+			BUG();
+		va += PAGE_SIZE;
+	}
+
+	/* Ensure init mappings cover kernel text/data and initial tables. */
+	while (va < (__START_KERNEL_map
+		     + (start_pfn << PAGE_SHIFT)
+		     + tables_space)) {
+		pmd = (pmd_t *)&page[pmd_index(va)];
+		if (pmd_none(*pmd)) {
+			pte_page = alloc_static_page(&phys);
+			early_make_page_readonly(
+				pte_page, XENFEAT_writable_page_tables);
+			set_pmd(pmd, __pmd(phys | _KERNPG_TABLE));
+		} else {
+			addr = page[pmd_index(va)];
+			addr_to_page(addr, pte_page);
+		}
+		pte = (pte_t *)&pte_page[pte_index(va)];
+		if (pte_none(*pte)) {
+			new_pte = pfn_pte(
+				(va - __START_KERNEL_map) >> PAGE_SHIFT, 
+				__pgprot(_KERNPG_TABLE));
+			xen_l1_entry_update(pte, new_pte);
+		}
+		va += PAGE_SIZE;
+	}
+
+	/* Finally, blow away any spurious initial mappings. */
+	while (1) {
+		pmd = (pmd_t *)&page[pmd_index(va)];
+		if (pmd_none(*pmd))
+			break;
+		if (HYPERVISOR_update_va_mapping(va, __pte_ma(0), 0))
+			BUG();
+		va += PAGE_SIZE;
+	}
+}
+
+static void __init find_early_table_space(unsigned long end)
+{
+	unsigned long puds, pmds, ptes, tables;
+
+	puds = (end + PUD_SIZE - 1) >> PUD_SHIFT;
+	pmds = (end + PMD_SIZE - 1) >> PMD_SHIFT;
+	ptes = (end + PTE_SIZE - 1) >> PAGE_SHIFT;
+
+	tables = round_up(puds * 8, PAGE_SIZE) + 
+		round_up(pmds * 8, PAGE_SIZE) + 
+		round_up(ptes * 8, PAGE_SIZE); 
+
+	extend_init_mapping(tables);
+
+	table_start = start_pfn;
+	table_end = table_start + (tables>>PAGE_SHIFT);
+
+	early_printk("kernel direct mapping tables up to %lx @ %lx-%lx\n",
+		end, table_start << PAGE_SHIFT,
+		(table_start << PAGE_SHIFT) + tables);
+}
+
+static void xen_finish_init_mapping(void)
+{
+	unsigned long i, start, end;
+
+	/* Re-vector virtual addresses pointing into the initial
+	   mapping to the just-established permanent ones. */
+	xen_start_info = __va(__pa(xen_start_info));
+	xen_start_info->pt_base = (unsigned long)
+		__va(__pa(xen_start_info->pt_base));
+	if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+		phys_to_machine_mapping =
+			__va(__pa(xen_start_info->mfn_list));
+		xen_start_info->mfn_list = (unsigned long)
+			phys_to_machine_mapping;
+	}
+	if (xen_start_info->mod_start)
+		xen_start_info->mod_start = (unsigned long)
+			__va(__pa(xen_start_info->mod_start));
+
+	/* Destroy the Xen-created mappings beyond the kernel image as
+	 * well as the temporary mappings created above. Prevents
+	 * overlap with modules area (if init mapping is very big).
+	 */
+	start = PAGE_ALIGN((unsigned long)_end);
+	end   = __START_KERNEL_map + (table_end << PAGE_SHIFT);
+	for (; start < end; start += PAGE_SIZE)
+		if (HYPERVISOR_update_va_mapping(start, __pte_ma(0), 0))
+			BUG();
+
+	/* Allocate pte's for initial fixmaps from 'start_pfn' allocator. */
+	table_end = ~0UL;
+
+	/*
+	 * Prefetch pte's for the bt_ioremap() area. It gets used before the
+	 * boot-time allocator is online, so allocate-on-demand would fail.
+	 */
+	for (i = FIX_BTMAP_END; i <= FIX_BTMAP_BEGIN; i++)
+		__set_fixmap(i, 0, __pgprot(0));
+
+	/* Switch to the real shared_info page, and clear the dummy page. */
+	set_fixmap(FIX_SHARED_INFO, xen_start_info->shared_info);
+	HYPERVISOR_shared_info = (shared_info_t *)fix_to_virt(FIX_SHARED_INFO);
+	memset(empty_zero_page, 0, sizeof(empty_zero_page));
+
+	/* Set up mapping of lowest 1MB of physical memory. */
+	for (i = 0; i < NR_FIX_ISAMAPS; i++)
+		if (is_initial_xendomain())
+			set_fixmap(FIX_ISAMAP_BEGIN - i, i * PAGE_SIZE);
+		else
+			__set_fixmap(FIX_ISAMAP_BEGIN - i,
+				     virt_to_mfn(empty_zero_page)
+				     << PAGE_SHIFT,
+				     PAGE_KERNEL_RO);
+
+	/* Disable the 'start_pfn' allocator. */
+	table_end = start_pfn;
+}
+
+/* Setup the direct mapping of the physical memory at PAGE_OFFSET.
+   This runs before bootmem is initialized and gets pages directly from the 
+   physical memory. To access them they are temporarily mapped. */
+void __meminit init_memory_mapping(unsigned long start, unsigned long end)
+{ 
+	unsigned long next;
+
+	Dprintk("init_memory_mapping\n");
+
+	/* 
+	 * Find space for the kernel direct mapping tables.
+	 * Later we should allocate these tables in the local node of the memory
+	 * mapped.  Unfortunately this is done currently before the nodes are 
+	 * discovered.
+	 */
+	if (!after_bootmem)
+		find_early_table_space(end);
+
+	start = (unsigned long)__va(start);
+	end = (unsigned long)__va(end);
+
+	for (; start < end; start = next) {
+		unsigned long pud_phys; 
+		pgd_t *pgd = pgd_offset_k(start);
+		pud_t *pud;
+
+		if (after_bootmem)
+			pud = pud_offset(pgd, start & PGDIR_MASK);
+		else
+			pud = alloc_static_page(&pud_phys);
+		next = start + PGDIR_SIZE;
+		if (next > end) 
+			next = end; 
+		phys_pud_init(pud, __pa(start), __pa(next));
+		if (!after_bootmem) {
+			early_make_page_readonly(pud, XENFEAT_writable_page_tables);
+			set_pgd(pgd_offset_k(start), mk_kernel_pgd(pud_phys));
+		}
+	}
+
+	if (!after_bootmem) {
+		BUG_ON(start_pfn != table_end);
+		xen_finish_init_mapping();
+	}
+
+	__flush_tlb_all();
+}
+
+void __cpuinit zap_low_mappings(int cpu)
+{
+	/* this is not required for Xen */
+#if 0
+	swap_low_mappings();
+#endif
+}
+
+/* Compute zone sizes for the DMA and DMA32 zones in a node. */
+__init void
+size_zones(unsigned long *z, unsigned long *h,
+	   unsigned long start_pfn, unsigned long end_pfn)
+{
+ 	int i;
+ 	unsigned long w;
+
+ 	for (i = 0; i < MAX_NR_ZONES; i++)
+ 		z[i] = 0;
+
+ 	if (start_pfn < MAX_DMA_PFN)
+ 		z[ZONE_DMA] = MAX_DMA_PFN - start_pfn;
+ 	if (start_pfn < MAX_DMA32_PFN) {
+ 		unsigned long dma32_pfn = MAX_DMA32_PFN;
+ 		if (dma32_pfn > end_pfn)
+ 			dma32_pfn = end_pfn;
+ 		z[ZONE_DMA32] = dma32_pfn - start_pfn;
+ 	}
+ 	z[ZONE_NORMAL] = end_pfn - start_pfn;
+
+ 	/* Remove lower zones from higher ones. */
+ 	w = 0;
+ 	for (i = 0; i < MAX_NR_ZONES; i++) {
+ 		if (z[i])
+ 			z[i] -= w;
+ 	        w += z[i];
+	}
+
+	/* Compute holes */
+	w = start_pfn;
+	for (i = 0; i < MAX_NR_ZONES; i++) {
+		unsigned long s = w;
+		w += z[i];
+		h[i] = e820_hole_size(s, w);
+	}
+
+	/* Add the space pace needed for mem_map to the holes too. */
+	for (i = 0; i < MAX_NR_ZONES; i++)
+		h[i] += (z[i] * sizeof(struct page)) / PAGE_SIZE;
+
+	/* The 16MB DMA zone has the kernel and other misc mappings.
+ 	   Account them too */
+	if (h[ZONE_DMA]) {
+		h[ZONE_DMA] += dma_reserve;
+		if (h[ZONE_DMA] >= z[ZONE_DMA]) {
+			printk(KERN_WARNING
+				"Kernel too large and filling up ZONE_DMA?\n");
+			h[ZONE_DMA] = z[ZONE_DMA];
+		}
+	}
+}
+
+#ifndef CONFIG_NUMA
+void __init paging_init(void)
+{
+	unsigned long zones[MAX_NR_ZONES], holes[MAX_NR_ZONES];
+
+	memory_present(0, 0, end_pfn);
+	sparse_init();
+	size_zones(zones, holes, 0, end_pfn);
+	free_area_init_node(0, NODE_DATA(0), zones,
+			    __pa(PAGE_OFFSET) >> PAGE_SHIFT, holes);
+
+	init_mm.context.pinned = 1;
+}
+#endif
+
+/* Unmap a kernel mapping if it exists. This is useful to avoid prefetches
+   from the CPU leading to inconsistent cache lines. address and size
+   must be aligned to 2MB boundaries. 
+   Does nothing when the mapping doesn't exist. */
+void __init clear_kernel_mapping(unsigned long address, unsigned long size) 
+{
+	unsigned long end = address + size;
+
+	BUG_ON(address & ~LARGE_PAGE_MASK);
+	BUG_ON(size & ~LARGE_PAGE_MASK); 
+	
+	for (; address < end; address += LARGE_PAGE_SIZE) { 
+		pgd_t *pgd = pgd_offset_k(address);
+		pud_t *pud;
+		pmd_t *pmd;
+		if (pgd_none(*pgd))
+			continue;
+		pud = pud_offset(pgd, address);
+		if (pud_none(*pud))
+			continue; 
+		pmd = pmd_offset(pud, address);
+		if (!pmd || pmd_none(*pmd))
+			continue; 
+		if (0 == (__pmd_val(*pmd) & _PAGE_PSE)) {
+			/* Could handle this, but it should not happen currently. */
+			printk(KERN_ERR 
+	       "clear_kernel_mapping: mapping has been split. will leak memory\n"); 
+			pmd_ERROR(*pmd); 
+		}
+		set_pmd(pmd, __pmd(0)); 		
+	}
+	__flush_tlb_all();
+} 
+
+/*
+ * Memory hotplug specific functions
+ */
+void online_page(struct page *page)
+{
+	ClearPageReserved(page);
+	init_page_count(page);
+	__free_page(page);
+	totalram_pages++;
+	num_physpages++;
+}
+
+#ifdef CONFIG_MEMORY_HOTPLUG
+/*
+ * XXX: memory_add_physaddr_to_nid() is to find node id from physical address
+ *	via probe interface of sysfs. If acpi notifies hot-add event, then it
+ *	can tell node id by searching dsdt. But, probe interface doesn't have
+ *	node id. So, return 0 as node id at this time.
+ */
+#ifdef CONFIG_NUMA
+int memory_add_physaddr_to_nid(u64 start)
+{
+	return 0;
+}
+#endif
+
+/*
+ * Memory is added always to NORMAL zone. This means you will never get
+ * additional DMA/DMA32 memory.
+ */
+int arch_add_memory(int nid, u64 start, u64 size)
+{
+	struct pglist_data *pgdat = NODE_DATA(nid);
+	struct zone *zone = pgdat->node_zones + MAX_NR_ZONES-2;
+	unsigned long start_pfn = start >> PAGE_SHIFT;
+	unsigned long nr_pages = size >> PAGE_SHIFT;
+	int ret;
+
+	ret = __add_pages(zone, start_pfn, nr_pages);
+	if (ret)
+		goto error;
+
+	init_memory_mapping(start, (start + size -1));
+
+	return ret;
+error:
+	printk("%s: Problem encountered in __add_pages!\n", __func__);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(arch_add_memory);
+
+int remove_memory(u64 start, u64 size)
+{
+	return -EINVAL;
+}
+EXPORT_SYMBOL_GPL(remove_memory);
+
+#else /* CONFIG_MEMORY_HOTPLUG */
+/*
+ * Memory Hotadd without sparsemem. The mem_maps have been allocated in advance,
+ * just online the pages.
+ */
+int __add_pages(struct zone *z, unsigned long start_pfn, unsigned long nr_pages)
+{
+	int err = -EIO;
+	unsigned long pfn;
+	unsigned long total = 0, mem = 0;
+	for (pfn = start_pfn; pfn < start_pfn + nr_pages; pfn++) {
+		if (pfn_valid(pfn)) {
+			online_page(pfn_to_page(pfn));
+			err = 0;
+			mem++;
+		}
+		total++;
+	}
+	if (!err) {
+		z->spanned_pages += total;
+		z->present_pages += mem;
+		z->zone_pgdat->node_spanned_pages += total;
+		z->zone_pgdat->node_present_pages += mem;
+	}
+	return err;
+}
+#endif /* CONFIG_MEMORY_HOTPLUG */
+
+static struct kcore_list kcore_mem, kcore_vmalloc, kcore_kernel, kcore_modules,
+			 kcore_vsyscall;
+
+void __init mem_init(void)
+{
+	long codesize, reservedpages, datasize, initsize;
+	unsigned long pfn;
+
+	contiguous_bitmap = alloc_bootmem_low_pages(
+		(end_pfn + 2*BITS_PER_LONG) >> 3);
+	BUG_ON(!contiguous_bitmap);
+	memset(contiguous_bitmap, 0, (end_pfn + 2*BITS_PER_LONG) >> 3);
+
+	pci_iommu_alloc();
+
+	/* How many end-of-memory variables you have, grandma! */
+	max_low_pfn = end_pfn;
+	max_pfn = end_pfn;
+	num_physpages = end_pfn;
+	high_memory = (void *) __va(end_pfn * PAGE_SIZE);
+
+	/* clear the zero-page */
+	memset(empty_zero_page, 0, PAGE_SIZE);
+
+	reservedpages = 0;
+
+	/* this will put all low memory onto the freelists */
+#ifdef CONFIG_NUMA
+	totalram_pages = numa_free_all_bootmem();
+#else
+	totalram_pages = free_all_bootmem();
+#endif
+	/* XEN: init and count pages outside initial allocation. */
+	for (pfn = xen_start_info->nr_pages; pfn < max_pfn; pfn++) {
+		ClearPageReserved(pfn_to_page(pfn));
+		init_page_count(pfn_to_page(pfn));
+		totalram_pages++;
+	}
+	reservedpages = end_pfn - totalram_pages - e820_hole_size(0, end_pfn);
+
+	after_bootmem = 1;
+
+	codesize =  (unsigned long) &_etext - (unsigned long) &_text;
+	datasize =  (unsigned long) &_edata - (unsigned long) &_etext;
+	initsize =  (unsigned long) &__init_end - (unsigned long) &__init_begin;
+
+	/* Register memory areas for /proc/kcore */
+	kclist_add(&kcore_mem, __va(0), max_low_pfn << PAGE_SHIFT); 
+	kclist_add(&kcore_vmalloc, (void *)VMALLOC_START, 
+		   VMALLOC_END-VMALLOC_START);
+	kclist_add(&kcore_kernel, &_stext, _end - _stext);
+	kclist_add(&kcore_modules, (void *)MODULES_VADDR, MODULES_LEN);
+	kclist_add(&kcore_vsyscall, (void *)VSYSCALL_START, 
+				 VSYSCALL_END - VSYSCALL_START);
+
+	printk("Memory: %luk/%luk available (%ldk kernel code, %ldk reserved, %ldk data, %ldk init)\n",
+		(unsigned long) nr_free_pages() << (PAGE_SHIFT-10),
+		end_pfn << (PAGE_SHIFT-10),
+		codesize >> 10,
+		reservedpages << (PAGE_SHIFT-10),
+		datasize >> 10,
+		initsize >> 10);
+
+#ifndef CONFIG_XEN
+#ifdef CONFIG_SMP
+	/*
+	 * Sync boot_level4_pgt mappings with the init_level4_pgt
+	 * except for the low identity mappings which are already zapped
+	 * in init_level4_pgt. This sync-up is essential for AP's bringup
+	 */
+	memcpy(boot_level4_pgt+1, init_level4_pgt+1, (PTRS_PER_PGD-1)*sizeof(pgd_t));
+#endif
+#endif
+}
+
+void free_init_pages(char *what, unsigned long begin, unsigned long end)
+{
+	unsigned long addr;
+
+	if (begin >= end)
+		return;
+
+	printk(KERN_INFO "Freeing %s: %ldk freed\n", what, (end - begin) >> 10);
+	for (addr = begin; addr < end; addr += PAGE_SIZE) {
+		ClearPageReserved(virt_to_page(addr));
+		init_page_count(virt_to_page(addr));
+		memset((void *)(addr & ~(PAGE_SIZE-1)),
+		       POISON_FREE_INITMEM, PAGE_SIZE); 
+		if (addr >= __START_KERNEL_map) {
+			/* make_readonly() reports all kernel addresses. */
+			__make_page_writable(__va(__pa(addr)));
+			if (HYPERVISOR_update_va_mapping(addr, __pte(0), 0)) {
+				pgd_t *pgd = pgd_offset_k(addr);
+				pud_t *pud = pud_offset(pgd, addr);
+				pmd_t *pmd = pmd_offset(pud, addr);
+				pte_t *pte = pte_offset_kernel(pmd, addr);
+
+				xen_l1_entry_update(pte, __pte(0)); /* fallback */
+			}
+		}
+		free_page(addr);
+		totalram_pages++;
+	}
+}
+
+void free_initmem(void)
+{
+	memset(__initdata_begin, POISON_FREE_INITDATA,
+		__initdata_end - __initdata_begin);
+	free_init_pages("unused kernel memory",
+			(unsigned long)(&__init_begin),
+			(unsigned long)(&__init_end));
+}
+
+#ifdef CONFIG_DEBUG_RODATA
+
+void mark_rodata_ro(void)
+{
+	unsigned long addr = (unsigned long)__start_rodata;
+
+	for (; addr < (unsigned long)__end_rodata; addr += PAGE_SIZE)
+		change_page_attr_addr(addr, 1, PAGE_KERNEL_RO);
+
+	printk ("Write protecting the kernel read-only data: %luk\n",
+			(__end_rodata - __start_rodata) >> 10);
+
+	/*
+	 * change_page_attr_addr() requires a global_flush_tlb() call after it.
+	 * We do this after the printk so that if something went wrong in the
+	 * change, the printk gets out at least to give a better debug hint
+	 * of who is the culprit.
+	 */
+	global_flush_tlb();
+}
+#endif
+
+#ifdef CONFIG_BLK_DEV_INITRD
+void free_initrd_mem(unsigned long start, unsigned long end)
+{
+	free_init_pages("initrd memory", start, end);
+}
+#endif
+
+void __init reserve_bootmem_generic(unsigned long phys, unsigned len) 
+{ 
+	/* Should check here against the e820 map to avoid double free */ 
+#ifdef CONFIG_NUMA
+	int nid = phys_to_nid(phys);
+  	reserve_bootmem_node(NODE_DATA(nid), phys, len);
+#else       		
+	reserve_bootmem(phys, len);    
+#endif
+	if (phys+len <= MAX_DMA_PFN*PAGE_SIZE)
+		dma_reserve += len / PAGE_SIZE;
+}
+
+int kern_addr_valid(unsigned long addr) 
+{ 
+	unsigned long above = ((long)addr) >> __VIRTUAL_MASK_SHIFT;
+       pgd_t *pgd;
+       pud_t *pud;
+       pmd_t *pmd;
+       pte_t *pte;
+
+	if (above != 0 && above != -1UL)
+		return 0; 
+	
+	pgd = pgd_offset_k(addr);
+	if (pgd_none(*pgd))
+		return 0;
+
+	pud = pud_offset(pgd, addr);
+	if (pud_none(*pud))
+		return 0; 
+
+	pmd = pmd_offset(pud, addr);
+	if (pmd_none(*pmd))
+		return 0;
+	if (pmd_large(*pmd))
+		return pfn_valid(pmd_pfn(*pmd));
+
+	pte = pte_offset_kernel(pmd, addr);
+	if (pte_none(*pte))
+		return 0;
+	return pfn_valid(pte_pfn(*pte));
+}
+
+#ifdef CONFIG_SYSCTL
+#include <linux/sysctl.h>
+
+extern int exception_trace, page_fault_trace;
+
+static ctl_table debug_table2[] = {
+	{ 99, "exception-trace", &exception_trace, sizeof(int), 0644, NULL,
+	  proc_dointvec },
+	{ 0, }
+}; 
+
+static ctl_table debug_root_table2[] = { 
+	{ .ctl_name = CTL_DEBUG, .procname = "debug", .mode = 0555, 
+	   .child = debug_table2 }, 
+	{ 0 }, 
+}; 
+
+static __init int x8664_sysctl_init(void)
+{ 
+	register_sysctl_table(debug_root_table2, 1);
+	return 0;
+}
+__initcall(x8664_sysctl_init);
+#endif
+
+/* A pseudo VMAs to allow ptrace access for the vsyscall page.   This only
+   covers the 64bit vsyscall page now. 32bit has a real VMA now and does
+   not need special handling anymore. */
+
+static struct vm_area_struct gate_vma = {
+	.vm_start = VSYSCALL_START,
+	.vm_end = VSYSCALL_END,
+	.vm_page_prot = PAGE_READONLY
+};
+
+struct vm_area_struct *get_gate_vma(struct task_struct *tsk)
+{
+#ifdef CONFIG_IA32_EMULATION
+	if (test_tsk_thread_flag(tsk, TIF_IA32))
+		return NULL;
+#endif
+	return &gate_vma;
+}
+
+int in_gate_area(struct task_struct *task, unsigned long addr)
+{
+	struct vm_area_struct *vma = get_gate_vma(task);
+	if (!vma)
+		return 0;
+	return (addr >= vma->vm_start) && (addr < vma->vm_end);
+}
+
+/* Use this when you have no reliable task/vma, typically from interrupt
+ * context.  It is less reliable than using the task's vma and may give
+ * false positives.
+ */
+int in_gate_area_no_task(unsigned long addr)
+{
+	return (addr >= VSYSCALL_START) && (addr < VSYSCALL_END);
+}
diff -rpuN linux-2.6.18.8/arch/x86_64/mm/Makefile linux-2.6.18-xen-3.3.0/arch/x86_64/mm/Makefile
--- linux-2.6.18.8/arch/x86_64/mm/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/mm/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -7,5 +7,8 @@ obj-$(CONFIG_HUGETLB_PAGE) += hugetlbpag
 obj-$(CONFIG_NUMA) += numa.o
 obj-$(CONFIG_K8_NUMA) += k8topology.o
 obj-$(CONFIG_ACPI_NUMA) += srat.o
+obj-$(CONFIG_XEN) += hypervisor.o
 
 hugetlbpage-y = ../../i386/mm/hugetlbpage.o
+ioremap-$(CONFIG_XEN) := ../../i386/mm/ioremap.o
+hypervisor-y := ../../i386/mm/hypervisor.o
diff -rpuN linux-2.6.18.8/arch/x86_64/mm/pageattr-xen.c linux-2.6.18-xen-3.3.0/arch/x86_64/mm/pageattr-xen.c
--- linux-2.6.18.8/arch/x86_64/mm/pageattr-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/mm/pageattr-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,502 @@
+/* 
+ * Copyright 2002 Andi Kleen, SuSE Labs. 
+ * Thanks to Ben LaHaise for precious feedback.
+ */ 
+
+#include <linux/mm.h>
+#include <linux/sched.h>
+#include <linux/highmem.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <asm/uaccess.h>
+#include <asm/processor.h>
+#include <asm/tlbflush.h>
+#include <asm/io.h>
+
+#ifdef CONFIG_XEN
+#include <asm/pgalloc.h>
+#include <asm/mmu_context.h>
+
+LIST_HEAD(mm_unpinned);
+DEFINE_SPINLOCK(mm_unpinned_lock);
+
+static void _pin_lock(struct mm_struct *mm, int lock) {
+	if (lock)
+		spin_lock(&mm->page_table_lock);
+#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
+	/* While mm->page_table_lock protects us against insertions and
+	 * removals of higher level page table pages, it doesn't protect
+	 * against updates of pte-s. Such updates, however, require the
+	 * pte pages to be in consistent state (unpinned+writable or
+	 * pinned+readonly). The pinning and attribute changes, however
+	 * cannot be done atomically, which is why such updates must be
+	 * prevented from happening concurrently.
+	 * Note that no pte lock can ever elsewhere be acquired nesting
+	 * with an already acquired one in the same mm, or with the mm's
+	 * page_table_lock already acquired, as that would break in the
+	 * non-split case (where all these are actually resolving to the
+	 * one page_table_lock). Thus acquiring all of them here is not
+	 * going to result in dead locks, and the order of acquires
+	 * doesn't matter.
+	 */
+	{
+		pgd_t *pgd = mm->pgd;
+		unsigned g;
+
+		for (g = 0; g <= ((TASK_SIZE64-1) / PGDIR_SIZE); g++, pgd++) {
+			pud_t *pud;
+			unsigned u;
+
+			if (pgd_none(*pgd))
+				continue;
+			pud = pud_offset(pgd, 0);
+			for (u = 0; u < PTRS_PER_PUD; u++, pud++) {
+				pmd_t *pmd;
+				unsigned m;
+
+				if (pud_none(*pud))
+					continue;
+				pmd = pmd_offset(pud, 0);
+				for (m = 0; m < PTRS_PER_PMD; m++, pmd++) {
+					spinlock_t *ptl;
+
+					if (pmd_none(*pmd))
+						continue;
+					ptl = pte_lockptr(0, pmd);
+					if (lock)
+						spin_lock(ptl);
+					else
+						spin_unlock(ptl);
+				}
+			}
+		}
+	}
+#endif
+	if (!lock)
+		spin_unlock(&mm->page_table_lock);
+}
+#define pin_lock(mm) _pin_lock(mm, 1)
+#define pin_unlock(mm) _pin_lock(mm, 0)
+
+#define PIN_BATCH 8
+static DEFINE_PER_CPU(multicall_entry_t[PIN_BATCH], pb_mcl);
+
+static inline unsigned int mm_walk_set_prot(void *pt, pgprot_t flags,
+                                            unsigned int cpu, unsigned int seq)
+{
+	struct page *page = virt_to_page(pt);
+	unsigned long pfn = page_to_pfn(page);
+
+	MULTI_update_va_mapping(per_cpu(pb_mcl, cpu) + seq,
+		(unsigned long)__va(pfn << PAGE_SHIFT),
+		pfn_pte(pfn, flags), 0);
+	if (unlikely(++seq == PIN_BATCH)) {
+		if (unlikely(HYPERVISOR_multicall_check(per_cpu(pb_mcl, cpu),
+	                                                PIN_BATCH, NULL)))
+			BUG();
+		seq = 0;
+	}
+
+	return seq;
+}
+
+static void mm_walk(struct mm_struct *mm, pgprot_t flags)
+{
+	pgd_t       *pgd;
+	pud_t       *pud;
+	pmd_t       *pmd;
+	pte_t       *pte;
+	int          g,u,m;
+	unsigned int cpu, seq;
+	multicall_entry_t *mcl;
+
+	pgd = mm->pgd;
+	cpu = get_cpu();
+
+	/*
+	 * Cannot iterate up to USER_PTRS_PER_PGD as these pagetables may not
+	 * be the 'current' task's pagetables (e.g., current may be 32-bit,
+	 * but the pagetables may be for a 64-bit task).
+	 * Subtracting 1 from TASK_SIZE64 means the loop limit is correct
+	 * regardless of whether TASK_SIZE64 is a multiple of PGDIR_SIZE.
+	 */
+	for (g = 0, seq = 0; g <= ((TASK_SIZE64-1) / PGDIR_SIZE); g++, pgd++) {
+		if (pgd_none(*pgd))
+			continue;
+		pud = pud_offset(pgd, 0);
+		if (PTRS_PER_PUD > 1) /* not folded */ 
+			seq = mm_walk_set_prot(pud,flags,cpu,seq);
+		for (u = 0; u < PTRS_PER_PUD; u++, pud++) {
+			if (pud_none(*pud))
+				continue;
+			pmd = pmd_offset(pud, 0);
+			if (PTRS_PER_PMD > 1) /* not folded */ 
+				seq = mm_walk_set_prot(pmd,flags,cpu,seq);
+			for (m = 0; m < PTRS_PER_PMD; m++, pmd++) {
+				if (pmd_none(*pmd))
+					continue;
+				pte = pte_offset_kernel(pmd,0);
+				seq = mm_walk_set_prot(pte,flags,cpu,seq);
+			}
+		}
+	}
+
+	mcl = per_cpu(pb_mcl, cpu);
+	if (unlikely(seq > PIN_BATCH - 2)) {
+		if (unlikely(HYPERVISOR_multicall_check(mcl, seq, NULL)))
+			BUG();
+		seq = 0;
+	}
+	MULTI_update_va_mapping(mcl + seq,
+	       (unsigned long)__user_pgd(mm->pgd),
+	       pfn_pte(virt_to_phys(__user_pgd(mm->pgd))>>PAGE_SHIFT, flags),
+	       0);
+	MULTI_update_va_mapping(mcl + seq + 1,
+	       (unsigned long)mm->pgd,
+	       pfn_pte(virt_to_phys(mm->pgd)>>PAGE_SHIFT, flags),
+	       UVMF_TLB_FLUSH);
+	if (unlikely(HYPERVISOR_multicall_check(mcl, seq + 2, NULL)))
+		BUG();
+
+	put_cpu();
+}
+
+void mm_pin(struct mm_struct *mm)
+{
+	if (xen_feature(XENFEAT_writable_page_tables))
+		return;
+
+	pin_lock(mm);
+
+	mm_walk(mm, PAGE_KERNEL_RO);
+	xen_pgd_pin(__pa(mm->pgd)); /* kernel */
+	xen_pgd_pin(__pa(__user_pgd(mm->pgd))); /* user */
+	mm->context.pinned = 1;
+	spin_lock(&mm_unpinned_lock);
+	list_del(&mm->context.unpinned);
+	spin_unlock(&mm_unpinned_lock);
+
+	pin_unlock(mm);
+}
+
+void mm_unpin(struct mm_struct *mm)
+{
+	if (xen_feature(XENFEAT_writable_page_tables))
+		return;
+
+	pin_lock(mm);
+
+	xen_pgd_unpin(__pa(mm->pgd));
+	xen_pgd_unpin(__pa(__user_pgd(mm->pgd)));
+	mm_walk(mm, PAGE_KERNEL);
+	mm->context.pinned = 0;
+	spin_lock(&mm_unpinned_lock);
+	list_add(&mm->context.unpinned, &mm_unpinned);
+	spin_unlock(&mm_unpinned_lock);
+
+	pin_unlock(mm);
+}
+
+void mm_pin_all(void)
+{
+	if (xen_feature(XENFEAT_writable_page_tables))
+		return;
+
+	/*
+	 * Allow uninterrupted access to the mm_unpinned list. We don't
+	 * actually take the mm_unpinned_lock as it is taken inside mm_pin().
+	 * All other CPUs must be at a safe point (e.g., in stop_machine
+	 * or offlined entirely).
+	 */
+	preempt_disable();
+	while (!list_empty(&mm_unpinned))	
+		mm_pin(list_entry(mm_unpinned.next, struct mm_struct,
+				  context.unpinned));
+	preempt_enable();
+}
+
+void _arch_dup_mmap(struct mm_struct *mm)
+{
+	if (!mm->context.pinned)
+		mm_pin(mm);
+}
+
+void _arch_exit_mmap(struct mm_struct *mm)
+{
+	struct task_struct *tsk = current;
+
+	task_lock(tsk);
+
+	/*
+	 * We aggressively remove defunct pgd from cr3. We execute unmap_vmas()
+	 * *much* faster this way, as no tlb flushes means bigger wrpt batches.
+	 */
+	if (tsk->active_mm == mm) {
+		tsk->active_mm = &init_mm;
+		atomic_inc(&init_mm.mm_count);
+
+		switch_mm(mm, &init_mm, tsk);
+
+		atomic_dec(&mm->mm_count);
+		BUG_ON(atomic_read(&mm->mm_count) == 0);
+	}
+
+	task_unlock(tsk);
+
+	if ( mm->context.pinned && (atomic_read(&mm->mm_count) == 1) &&
+	     !mm->context.has_foreign_mappings )
+		mm_unpin(mm);
+}
+
+struct page *pte_alloc_one(struct mm_struct *mm, unsigned long address)
+{
+	struct page *pte;
+
+	pte = alloc_pages(GFP_KERNEL|__GFP_REPEAT|__GFP_ZERO, 0);
+	if (pte) {
+		SetPageForeign(pte, pte_free);
+		init_page_count(pte);
+	}
+	return pte;
+}
+
+void pte_free(struct page *pte)
+{
+	unsigned long va = (unsigned long)__va(page_to_pfn(pte)<<PAGE_SHIFT);
+
+	if (!pte_write(*virt_to_ptep(va)))
+		if (HYPERVISOR_update_va_mapping(
+			va, pfn_pte(page_to_pfn(pte), PAGE_KERNEL), 0))
+			BUG();
+
+	ClearPageForeign(pte);
+	init_page_count(pte);
+
+	__free_page(pte);
+}
+#endif	/* CONFIG_XEN */
+
+pte_t *lookup_address(unsigned long address) 
+{ 
+	pgd_t *pgd = pgd_offset_k(address);
+	pud_t *pud;
+	pmd_t *pmd;
+	pte_t *pte;
+	if (pgd_none(*pgd))
+		return NULL;
+	pud = pud_offset(pgd, address);
+	if (!pud_present(*pud))
+		return NULL; 
+	pmd = pmd_offset(pud, address);
+	if (!pmd_present(*pmd))
+		return NULL; 
+	if (pmd_large(*pmd))
+		return (pte_t *)pmd;
+	pte = pte_offset_kernel(pmd, address);
+	if (pte && !pte_present(*pte))
+		pte = NULL; 
+	return pte;
+} 
+
+static struct page *split_large_page(unsigned long address, pgprot_t prot,
+				     pgprot_t ref_prot)
+{ 
+	int i; 
+	unsigned long addr;
+	struct page *base = alloc_pages(GFP_KERNEL, 0);
+	pte_t *pbase;
+	if (!base) 
+		return NULL;
+	/*
+	 * page_private is used to track the number of entries in
+	 * the page table page have non standard attributes.
+	 */
+	SetPagePrivate(base);
+	page_private(base) = 0;
+
+	address = __pa(address);
+	addr = address & LARGE_PAGE_MASK; 
+	pbase = (pte_t *)page_address(base);
+	for (i = 0; i < PTRS_PER_PTE; i++, addr += PAGE_SIZE) {
+		pbase[i] = pfn_pte(addr >> PAGE_SHIFT, 
+				   addr == address ? prot : ref_prot);
+	}
+	return base;
+} 
+
+
+static void flush_kernel_map(void *address) 
+{
+	if (0 && address && cpu_has_clflush) {
+		/* is this worth it? */ 
+		int i;
+		for (i = 0; i < PAGE_SIZE; i += boot_cpu_data.x86_clflush_size) 
+			asm volatile("clflush (%0)" :: "r" (address + i)); 
+	} else
+		asm volatile("wbinvd":::"memory"); 
+	if (address)
+		__flush_tlb_one(address);
+	else
+		__flush_tlb_all();
+}
+
+
+static inline void flush_map(unsigned long address)
+{	
+	on_each_cpu(flush_kernel_map, (void *)address, 1, 1);
+}
+
+static struct page *deferred_pages; /* protected by init_mm.mmap_sem */
+
+static inline void save_page(struct page *fpage)
+{
+	fpage->lru.next = (struct list_head *)deferred_pages;
+	deferred_pages = fpage;
+}
+
+/* 
+ * No more special protections in this 2/4MB area - revert to a
+ * large page again. 
+ */
+static void revert_page(unsigned long address, pgprot_t ref_prot)
+{
+	pgd_t *pgd;
+	pud_t *pud;
+	pmd_t *pmd;
+	pte_t large_pte;
+
+	pgd = pgd_offset_k(address);
+	BUG_ON(pgd_none(*pgd));
+	pud = pud_offset(pgd,address);
+	BUG_ON(pud_none(*pud));
+	pmd = pmd_offset(pud, address);
+	BUG_ON(__pmd_val(*pmd) & _PAGE_PSE);
+	pgprot_val(ref_prot) |= _PAGE_PSE;
+	large_pte = mk_pte_phys(__pa(address) & LARGE_PAGE_MASK, ref_prot);
+	set_pte((pte_t *)pmd, large_pte);
+}      
+
+static int
+__change_page_attr(unsigned long address, unsigned long pfn, pgprot_t prot,
+				   pgprot_t ref_prot)
+{ 
+	pte_t *kpte; 
+	struct page *kpte_page;
+	unsigned kpte_flags;
+	pgprot_t ref_prot2;
+	kpte = lookup_address(address);
+	if (!kpte) return 0;
+	kpte_page = virt_to_page(((unsigned long)kpte) & PAGE_MASK);
+	kpte_flags = pte_val(*kpte); 
+	if (pgprot_val(prot) != pgprot_val(ref_prot)) { 
+		if ((kpte_flags & _PAGE_PSE) == 0) { 
+			set_pte(kpte, pfn_pte(pfn, prot));
+		} else {
+ 			/*
+			 * split_large_page will take the reference for this
+			 * change_page_attr on the split page.
+ 			 */
+
+			struct page *split;
+			ref_prot2 = __pgprot(pgprot_val(pte_pgprot(*lookup_address(address))) & ~(1<<_PAGE_BIT_PSE));
+
+			split = split_large_page(address, prot, ref_prot2);
+			if (!split)
+				return -ENOMEM;
+			set_pte(kpte,mk_pte(split, ref_prot2));
+			kpte_page = split;
+		}	
+		page_private(kpte_page)++;
+	} else if ((kpte_flags & _PAGE_PSE) == 0) { 
+		set_pte(kpte, pfn_pte(pfn, ref_prot));
+		BUG_ON(page_private(kpte_page) == 0);
+		page_private(kpte_page)--;
+	} else
+		BUG();
+
+	/* on x86-64 the direct mapping set at boot is not using 4k pages */
+	/*
+	 * ..., but the XEN guest kernels (currently) do:
+	 * If the pte was reserved, it means it was created at boot
+	 * time (not via split_large_page) and in turn we must not
+	 * replace it with a large page.
+	 */
+#ifndef CONFIG_XEN
+ 	BUG_ON(PageReserved(kpte_page));
+#else
+	if (PageReserved(kpte_page))
+		return 0;
+#endif
+
+	if (page_private(kpte_page) == 0) {
+		save_page(kpte_page);
+		revert_page(address, ref_prot);
+	}
+	return 0;
+} 
+
+/*
+ * Change the page attributes of an page in the linear mapping.
+ *
+ * This should be used when a page is mapped with a different caching policy
+ * than write-back somewhere - some CPUs do not like it when mappings with
+ * different caching policies exist. This changes the page attributes of the
+ * in kernel linear mapping too.
+ * 
+ * The caller needs to ensure that there are no conflicting mappings elsewhere.
+ * This function only deals with the kernel linear map.
+ * 
+ * Caller must call global_flush_tlb() after this.
+ */
+int change_page_attr_addr(unsigned long address, int numpages, pgprot_t prot)
+{
+	int err = 0; 
+	int i; 
+
+	down_write(&init_mm.mmap_sem);
+	for (i = 0; i < numpages; i++, address += PAGE_SIZE) {
+		unsigned long pfn = __pa(address) >> PAGE_SHIFT;
+
+		err = __change_page_attr(address, pfn, prot, PAGE_KERNEL);
+		if (err) 
+			break; 
+		/* Handle kernel mapping too which aliases part of the
+		 * lowmem */
+		if (__pa(address) < KERNEL_TEXT_SIZE) {
+			unsigned long addr2;
+			pgprot_t prot2 = prot;
+			addr2 = __START_KERNEL_map + __pa(address);
+ 			pgprot_val(prot2) &= ~_PAGE_NX;
+			err = __change_page_attr(addr2, pfn, prot2, PAGE_KERNEL_EXEC);
+		} 
+	} 	
+	up_write(&init_mm.mmap_sem); 
+	return err;
+}
+
+/* Don't call this for MMIO areas that may not have a mem_map entry */
+int change_page_attr(struct page *page, int numpages, pgprot_t prot)
+{
+	unsigned long addr = (unsigned long)page_address(page);
+	return change_page_attr_addr(addr, numpages, prot);
+}
+
+void global_flush_tlb(void)
+{ 
+	struct page *dpage;
+
+	down_read(&init_mm.mmap_sem);
+	dpage = xchg(&deferred_pages, NULL);
+	up_read(&init_mm.mmap_sem);
+
+	flush_map((dpage && !dpage->lru.next) ? (unsigned long)page_address(dpage) : 0);
+	while (dpage) {
+		struct page *tmp = dpage;
+		dpage = (struct page *)dpage->lru.next;
+		ClearPagePrivate(tmp);
+		__free_page(tmp);
+	} 
+} 
+
+EXPORT_SYMBOL(change_page_attr);
+EXPORT_SYMBOL(global_flush_tlb);
diff -rpuN linux-2.6.18.8/arch/x86_64/oprofile/Makefile linux-2.6.18-xen-3.3.0/arch/x86_64/oprofile/Makefile
--- linux-2.6.18.8/arch/x86_64/oprofile/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/oprofile/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -11,9 +11,15 @@ DRIVER_OBJS = $(addprefix ../../../drive
 	oprofilefs.o oprofile_stats.o \
 	timer_int.o )
 
+ifdef CONFIG_XEN
+XENOPROF_COMMON_OBJS = $(addprefix ../../../drivers/xen/xenoprof/, \
+			 xenoprofile.o)
+OPROFILE-y := xenoprof.o
+else
 OPROFILE-y := init.o backtrace.o
 OPROFILE-$(CONFIG_X86_LOCAL_APIC) += nmi_int.o op_model_athlon.o op_model_p4.o \
 				     op_model_ppro.o
 OPROFILE-$(CONFIG_X86_IO_APIC)    += nmi_timer_int.o 
-
-oprofile-y = $(DRIVER_OBJS) $(addprefix ../../i386/oprofile/, $(OPROFILE-y))
+endif
+oprofile-y = $(DRIVER_OBJS) $(XENOPROF_COMMON_OBJS) \
+	     $(addprefix ../../i386/oprofile/, $(OPROFILE-y))
diff -rpuN linux-2.6.18.8/arch/x86_64/pci/Makefile linux-2.6.18-xen-3.3.0/arch/x86_64/pci/Makefile
--- linux-2.6.18.8/arch/x86_64/pci/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/arch/x86_64/pci/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -15,8 +15,13 @@ obj-$(CONFIG_PCI_MMCONFIG) += mmconfig.o
 
 obj-$(CONFIG_NUMA)	+= k8-bus.o
 
+# pcifront should be after mmconfig.o and direct.o as it should only
+# take over if direct access to the PCI bus is unavailable
+obj-$(CONFIG_XEN_PCIDEV_FRONTEND)	+= pcifront.o
+
 direct-y += ../../i386/pci/direct.o
 acpi-y   += ../../i386/pci/acpi.o
+pcifront-y += ../../i386/pci/pcifront.o
 legacy-y += ../../i386/pci/legacy.o
 irq-y    += ../../i386/pci/irq.o
 common-y += ../../i386/pci/common.o
diff -rpuN linux-2.6.18.8/block/elevator.c linux-2.6.18-xen-3.3.0/block/elevator.c
--- linux-2.6.18.8/block/elevator.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/block/elevator.c	2008-08-21 11:36:07.000000000 +0200
@@ -493,6 +493,16 @@ struct request *elv_next_request(request
 	int ret;
 
 	while ((rq = __elv_next_request(q)) != NULL) {
+		/*
+		 * Kill the empty barrier place holder, the driver must
+		 * not ever see it.
+		 */
+		if (blk_empty_barrier(rq)) {
+			blkdev_dequeue_request(rq);
+			end_that_request_chunk(rq, 1, 0);
+			end_that_request_last(rq, 1);
+			continue;
+		}
 		if (!(rq->flags & REQ_STARTED)) {
 			elevator_t *e = q->elevator;
 
diff -rpuN linux-2.6.18.8/block/ll_rw_blk.c linux-2.6.18-xen-3.3.0/block/ll_rw_blk.c
--- linux-2.6.18.8/block/ll_rw_blk.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/block/ll_rw_blk.c	2008-08-21 11:36:07.000000000 +0200
@@ -483,9 +483,12 @@ static inline struct request *start_orde
 	 * Queue ordered sequence.  As we stack them at the head, we
 	 * need to queue in reverse order.  Note that we rely on that
 	 * no fs request uses ELEVATOR_INSERT_FRONT and thus no fs
-	 * request gets inbetween ordered sequence.
+	 * request gets inbetween ordered sequence. If this request is
+	 * an empty barrier, we don't need to do a postflush ever since
+	 * there will be no data written between the pre and post flush.
+	 * Hence a single flush will suffice.
 	 */
-	if (q->ordered & QUEUE_ORDERED_POSTFLUSH)
+	if ((q->ordered & QUEUE_ORDERED_POSTFLUSH) && !blk_empty_barrier(rq))
 		queue_flush(q, QUEUE_ORDERED_POSTFLUSH);
 	else
 		q->ordseq |= QUEUE_ORDSEQ_POSTFLUSH;
@@ -2967,7 +2970,7 @@ static inline void blk_partition_remap(s
 {
 	struct block_device *bdev = bio->bi_bdev;
 
-	if (bdev != bdev->bd_contains) {
+	if (bio_sectors(bio) && bdev != bdev->bd_contains) {
 		struct hd_struct *p = bdev->bd_part;
 		const int rw = bio_data_dir(bio);
 
@@ -3028,7 +3031,7 @@ void generic_make_request(struct bio *bi
 	might_sleep();
 	/* Test device or partition size, when known. */
 	maxsector = bio->bi_bdev->bd_inode->i_size >> 9;
-	if (maxsector) {
+	if (maxsector && nr_sectors) {
 		sector_t sector = bio->bi_sector;
 
 		if (maxsector < nr_sectors || maxsector - nr_sectors < sector) {
@@ -3094,7 +3097,7 @@ end_io:
 		old_dev = bio->bi_bdev->bd_dev;
 
 		maxsector = bio->bi_bdev->bd_inode->i_size >> 9;
-		if (maxsector) {
+		if (maxsector && nr_sectors) {
 			sector_t sector = bio->bi_sector;
 
 			if (maxsector < nr_sectors || maxsector - nr_sectors < sector) {
@@ -3128,21 +3131,25 @@ void submit_bio(int rw, struct bio *bio)
 {
 	int count = bio_sectors(bio);
 
-	BIO_BUG_ON(!bio->bi_size);
-	BIO_BUG_ON(!bio->bi_io_vec);
 	bio->bi_rw |= rw;
-	if (rw & WRITE)
-		count_vm_events(PGPGOUT, count);
-	else
-		count_vm_events(PGPGIN, count);
 
-	if (unlikely(block_dump)) {
-		char b[BDEVNAME_SIZE];
-		printk(KERN_DEBUG "%s(%d): %s block %Lu on %s\n",
-			current->comm, current->pid,
-			(rw & WRITE) ? "WRITE" : "READ",
-			(unsigned long long)bio->bi_sector,
-			bdevname(bio->bi_bdev,b));
+	if (!bio_empty_barrier(bio)) {
+		BIO_BUG_ON(!bio->bi_size);
+		BIO_BUG_ON(!bio->bi_io_vec);
+
+		if (rw & WRITE)
+			count_vm_events(PGPGOUT, count);
+		else
+			count_vm_events(PGPGIN, count);
+
+		if (unlikely(block_dump)) {
+			char b[BDEVNAME_SIZE];
+			printk(KERN_DEBUG "%s(%d): %s block %Lu on %s\n",
+				current->comm, current->pid,
+				(rw & WRITE) ? "WRITE" : "READ",
+				(unsigned long long)bio->bi_sector,
+				bdevname(bio->bi_bdev,b));
+		}
 	}
 
 	generic_make_request(bio);
@@ -3260,6 +3267,13 @@ static int __end_that_request_first(stru
 	while ((bio = req->bio) != NULL) {
 		int nbytes;
 
+		/* For an empty barrier request, the low level driver must
+		 * store a potential error location in ->sector. We pass
+		 * that back up in ->bi_sector
+		 */
+		if (blk_empty_barrier(req))
+			bio->bi_sector = req->sector;
+
 		if (nr_bytes >= bio->bi_size) {
 			req->bio = bio->bi_next;
 			nbytes = bio->bi_size;
diff -rpuN linux-2.6.18.8/buildconfigs/conf.linux-native/00_xen_to_native linux-2.6.18-xen-3.3.0/buildconfigs/conf.linux-native/00_xen_to_native
--- linux-2.6.18.8/buildconfigs/conf.linux-native/00_xen_to_native	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/buildconfigs/conf.linux-native/00_xen_to_native	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,86 @@
+# Linux kernel version: 2.6.16.13-native
+# Mon May 15 10:59:54 2006
+#
+CONFIG_X86_PC=y
+# CONFIG_X86_XEN is not set
+# CONFIG_HPET_TIMER is not set
+# CONFIG_SCHED_SMT is not set
+# CONFIG_X86_MCE is not set
+# CONFIG_X86_MSR is not set
+# CONFIG_SWIOTLB is not set
+# CONFIG_EDD is not set
+CONFIG_ARCH_FLATMEM_ENABLE=y
+CONFIG_ARCH_SPARSEMEM_ENABLE=y
+CONFIG_ARCH_SELECT_MEMORY_MODEL=y
+CONFIG_SPARSEMEM_STATIC=y
+CONFIG_SPLIT_PTLOCK_CPUS=4
+# CONFIG_HIGHPTE is not set
+# CONFIG_MATH_EMULATION is not set
+# CONFIG_EFI is not set
+CONFIG_IRQBALANCE=y
+# CONFIG_KEXEC is not set
+CONFIG_DOUBLEFAULT=y
+CONFIG_PM_LEGACY=y
+# CONFIG_PM_DEBUG is not set
+# CONFIG_SOFTWARE_SUSPEND is not set
+CONFIG_SUSPEND_SMP=y
+CONFIG_ACPI_SLEEP=y
+CONFIG_ACPI_SLEEP_PROC_FS=y
+# CONFIG_ACPI_SLEEP_PROC_SLEEP is not set
+CONFIG_X86_PM_TIMER=y
+# APM (Advanced Power Management) BIOS Support
+#
+# CONFIG_APM is not set
+
+#
+CONFIG_PCI_BIOS=y
+# CONFIG_XEN_PCIDEV_FRONTEND is not set
+# CONFIG_XEN_PCIDEV_FE_DEBUG is not set
+# CONFIG_PCI_MSI is not set
+# CONFIG_ISA is not set
+# CONFIG_MCA is not set
+# CONFIG_HOTPLUG_PCI_COMPAQ is not set
+# CONFIG_HOTPLUG_PCI_IBM is not set
+# CONFIG_I2O_EXT_ADAPTEC_DMA64 is not set
+CONFIG_SERIAL_8250=y
+CONFIG_SERIAL_8250_CONSOLE=y
+CONFIG_SERIAL_CORE=y
+CONFIG_SERIAL_CORE_CONSOLE=y
+# CONFIG_TCG_XEN is not set
+# CONFIG_HUGETLBFS is not set
+# CONFIG_XEN is not set
+# CONFIG_XEN_INTERFACE_VERSION is not set
+
+#
+# XEN
+#
+# CONFIG_XEN_PRIVILEGED_GUEST is not set
+# CONFIG_XEN_UNPRIVILEGED_GUEST is not set
+# CONFIG_XEN_BACKEND is not set
+# CONFIG_XEN_PCIDEV_BACKEND is not set
+# CONFIG_XEN_PCIDEV_BACKEND_VPCI is not set
+# CONFIG_XEN_PCIDEV_BACKEND_PASS is not set
+# CONFIG_XEN_PCIDEV_BE_DEBUG is not set
+# CONFIG_XEN_BLKDEV_BACKEND is not set
+# CONFIG_XEN_BLKDEV_TAP_BE is not set
+# CONFIG_XEN_NETDEV_BACKEND is not set
+# CONFIG_XEN_NETDEV_PIPELINED_TRANSMITTER is not set
+# CONFIG_XEN_NETDEV_LOOPBACK is not set
+# CONFIG_XEN_TPMDEV_BACKEND is not set
+# CONFIG_XEN_BLKDEV_FRONTEND is not set
+# CONFIG_XEN_NETDEV_FRONTEND is not set
+# CONFIG_XEN_BLKDEV_TAP is not set
+# CONFIG_XEN_SCRUB_PAGES is not set
+# CONFIG_XEN_DISABLE_SERIAL is not set
+# CONFIG_XEN_SYSFS is not set
+# CONFIG_XEN_COMPAT_030002_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
+# CONFIG_XEN_COMPAT_030002 is not set
+# CONFIG_XEN_COMPAT_030004 is not set
+# CONFIG_HAVE_ARCH_ALLOC_SKB is not set
+# CONFIG_HAVE_ARCH_DEV_ALLOC_SKB is not set
+# CONFIG_NO_IDLE_HZ is not set
+CONFIG_X86_HT=y
+# CONFIG_X86_NO_TSS is not set
+# CONFIG_X86_NO_IDT is not set
diff -rpuN linux-2.6.18.8/buildconfigs/create_config.sh linux-2.6.18-xen-3.3.0/buildconfigs/create_config.sh
--- linux-2.6.18.8/buildconfigs/create_config.sh	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/buildconfigs/create_config.sh	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,54 @@
+#!/bin/sh
+set -e
+
+
+# Parse arguments
+#
+if [ $# -lt 1 -o $# -gt 4 ]; then
+    echo "Usage: $0 config-file EXTRAVERSION XEN_TARGET_ARCH XEN_SYSTYPE"
+    exit 1
+fi
+
+config_file=$1
+extraversion=$2
+target_arch=$3
+systype=$4
+
+
+# Start with initial config skeleton file, if any.
+# Derive from linux-defconfig_xen_x86_32 otherwise.
+#
+skeleton=buildconfigs/linux-defconfig_${extraversion#-}_${target_arch}${systype}
+[ -r $skeleton ] || skeleton=buildconfigs/linux-defconfig_xen_x86_32
+[ -r $skeleton.local ] && skeleton=$skeleton.local
+cp $skeleton $config_file
+
+echo "Using $skeleton as base config"
+
+# Update
+#
+filter_template="s/^#\{0,1\} *\(CONFIG[^= ]*\).*/\/^#\\\{0,1\\\} *\1[= ].*\/d/p"
+config_dirs="buildconfigs/conf.linux buildconfigs/conf.linux-${target_arch} buildconfigs/conf.linux${extraversion} buildconfigs/conf.linux-${target_arch}${extraversion}"
+
+for config_dir in $config_dirs
+do
+    if [ -d $config_dir ]; then
+        echo "Processing $config_dir..." 1>&2
+        # processing is done in alphanumeric order
+        find $config_dir -type f | sort | while read update
+        do
+            echo "   ... $update" 1>&2
+            # create the filter rules in a temp file
+            filter_rules=`mktemp -t xenupdateconf.XXXXXXXXXX`
+            sed -n "${filter_template}" < $update > $filter_rules
+
+            # filter the config file in place, removing any options that
+            # will be updated.
+            sed -f $filter_rules -i $config_file
+            cat $update >> $config_file
+
+            # clean up
+            rm -f $filter_rules
+        done
+    fi
+done
diff -rpuN linux-2.6.18.8/buildconfigs/linux-defconfig_xen0_ia64 linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xen0_ia64
--- linux-2.6.18.8/buildconfigs/linux-defconfig_xen0_ia64	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xen0_ia64	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1703 @@
+#
+# Automatically generated make config: don't edit
+# Linux kernel version: 2.6.18.8
+# Tue Feb 19 11:20:09 2008
+#
+CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
+
+#
+# Code maturity level options
+#
+CONFIG_EXPERIMENTAL=y
+CONFIG_LOCK_KERNEL=y
+CONFIG_INIT_ENV_ARG_LIMIT=32
+
+#
+# General setup
+#
+CONFIG_LOCALVERSION=""
+CONFIG_LOCALVERSION_AUTO=y
+CONFIG_SWAP=y
+CONFIG_SYSVIPC=y
+CONFIG_POSIX_MQUEUE=y
+CONFIG_BSD_PROCESS_ACCT=y
+# CONFIG_BSD_PROCESS_ACCT_V3 is not set
+# CONFIG_TASKSTATS is not set
+# CONFIG_AUDIT is not set
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+# CONFIG_CPUSETS is not set
+# CONFIG_RELAY is not set
+CONFIG_INITRAMFS_SOURCE=""
+CONFIG_CC_OPTIMIZE_FOR_SIZE=y
+# CONFIG_EMBEDDED is not set
+CONFIG_SYSCTL=y
+CONFIG_KALLSYMS=y
+CONFIG_KALLSYMS_ALL=y
+CONFIG_KALLSYMS_EXTRA_PASS=y
+CONFIG_HOTPLUG=y
+CONFIG_PRINTK=y
+CONFIG_BUG=y
+CONFIG_ELF_CORE=y
+CONFIG_BASE_FULL=y
+CONFIG_FUTEX=y
+CONFIG_EPOLL=y
+CONFIG_SHMEM=y
+CONFIG_SLAB=y
+CONFIG_VM_EVENT_COUNTERS=y
+CONFIG_RT_MUTEXES=y
+# CONFIG_TINY_SHMEM is not set
+CONFIG_BASE_SMALL=0
+# CONFIG_SLOB is not set
+
+#
+# Loadable module support
+#
+CONFIG_MODULES=y
+CONFIG_MODULE_UNLOAD=y
+# CONFIG_MODULE_FORCE_UNLOAD is not set
+CONFIG_MODVERSIONS=y
+CONFIG_MODULE_SRCVERSION_ALL=y
+CONFIG_KMOD=y
+CONFIG_STOP_MACHINE=y
+
+#
+# Block layer
+#
+# CONFIG_BLK_DEV_IO_TRACE is not set
+
+#
+# IO Schedulers
+#
+CONFIG_IOSCHED_NOOP=y
+CONFIG_IOSCHED_AS=y
+CONFIG_IOSCHED_DEADLINE=y
+CONFIG_IOSCHED_CFQ=y
+CONFIG_DEFAULT_AS=y
+# CONFIG_DEFAULT_DEADLINE is not set
+# CONFIG_DEFAULT_CFQ is not set
+# CONFIG_DEFAULT_NOOP is not set
+CONFIG_DEFAULT_IOSCHED="anticipatory"
+
+#
+# Processor type and features
+#
+CONFIG_IA64=y
+CONFIG_64BIT=y
+CONFIG_MMU=y
+CONFIG_SWIOTLB=y
+CONFIG_RWSEM_XCHGADD_ALGORITHM=y
+CONFIG_GENERIC_FIND_NEXT_BIT=y
+CONFIG_GENERIC_CALIBRATE_DELAY=y
+CONFIG_TIME_INTERPOLATION=y
+CONFIG_DMI=y
+CONFIG_EFI=y
+CONFIG_GENERIC_IOMAP=y
+CONFIG_XEN=y
+CONFIG_XEN_IA64_EXPOSE_P2M=y
+CONFIG_XEN_IA64_EXPOSE_P2M_USE_DTR=y
+CONFIG_SCHED_NO_NO_OMIT_FRAME_POINTER=y
+CONFIG_DMA_IS_DMA32=y
+CONFIG_AUDIT_ARCH=y
+CONFIG_IA64_GENERIC=y
+# CONFIG_IA64_DIG is not set
+# CONFIG_IA64_HP_ZX1 is not set
+# CONFIG_IA64_HP_ZX1_SWIOTLB is not set
+# CONFIG_IA64_SGI_SN2 is not set
+# CONFIG_IA64_HP_SIM is not set
+# CONFIG_IA64_XEN is not set
+# CONFIG_ITANIUM is not set
+CONFIG_MCKINLEY=y
+# CONFIG_IA64_PAGE_SIZE_4KB is not set
+# CONFIG_IA64_PAGE_SIZE_8KB is not set
+CONFIG_IA64_PAGE_SIZE_16KB=y
+# CONFIG_IA64_PAGE_SIZE_64KB is not set
+CONFIG_PGTABLE_3=y
+# CONFIG_PGTABLE_4 is not set
+CONFIG_HZ_100=y
+# CONFIG_HZ_250 is not set
+# CONFIG_HZ_1000 is not set
+CONFIG_HZ=100
+CONFIG_IA64_L1_CACHE_SHIFT=7
+CONFIG_IA64_CYCLONE=y
+CONFIG_IOSAPIC=y
+# CONFIG_IA64_SGI_SN_XP is not set
+CONFIG_FORCE_MAX_ZONEORDER=11
+CONFIG_SMP=y
+CONFIG_NR_CPUS=16
+CONFIG_HOTPLUG_CPU=y
+CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y
+# CONFIG_SCHED_SMT is not set
+# CONFIG_PERMIT_BSP_REMOVE is not set
+# CONFIG_PREEMPT is not set
+CONFIG_SELECT_MEMORY_MODEL=y
+# CONFIG_FLATMEM_MANUAL is not set
+CONFIG_DISCONTIGMEM_MANUAL=y
+# CONFIG_SPARSEMEM_MANUAL is not set
+CONFIG_DISCONTIGMEM=y
+CONFIG_FLAT_NODE_MEM_MAP=y
+CONFIG_NEED_MULTIPLE_NODES=y
+# CONFIG_SPARSEMEM_STATIC is not set
+CONFIG_SPLIT_PTLOCK_CPUS=4
+# CONFIG_MIGRATION is not set
+CONFIG_RESOURCES_64BIT=y
+CONFIG_ARCH_SELECT_MEMORY_MODEL=y
+CONFIG_ARCH_DISCONTIGMEM_ENABLE=y
+CONFIG_ARCH_FLATMEM_ENABLE=y
+CONFIG_ARCH_SPARSEMEM_ENABLE=y
+CONFIG_ARCH_DISCONTIGMEM_DEFAULT=y
+CONFIG_NUMA=y
+CONFIG_NODES_SHIFT=10
+CONFIG_VIRTUAL_MEM_MAP=y
+CONFIG_HOLES_IN_ZONE=y
+CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID=y
+CONFIG_HAVE_ARCH_NODEDATA_EXTENSION=y
+# CONFIG_IA32_SUPPORT is not set
+# CONFIG_IA64_MCA_RECOVERY is not set
+CONFIG_PERFMON=y
+CONFIG_IA64_PALINFO=y
+CONFIG_SGI_SN=y
+
+#
+# SN Devices
+#
+# CONFIG_SGI_IOC3 is not set
+CONFIG_KEXEC=y
+
+#
+# Firmware Drivers
+#
+CONFIG_EFI_VARS=y
+CONFIG_EFI_PCDP=y
+CONFIG_BINFMT_ELF=y
+CONFIG_BINFMT_MISC=y
+
+#
+# Power management and ACPI
+#
+CONFIG_PM=y
+CONFIG_PM_LEGACY=y
+# CONFIG_PM_DEBUG is not set
+
+#
+# ACPI (Advanced Configuration and Power Interface) Support
+#
+CONFIG_ACPI=y
+CONFIG_ACPI_BUTTON=y
+CONFIG_ACPI_FAN=y
+# CONFIG_ACPI_DOCK is not set
+CONFIG_ACPI_PROCESSOR=y
+CONFIG_ACPI_HOTPLUG_CPU=y
+CONFIG_ACPI_THERMAL=y
+CONFIG_ACPI_NUMA=y
+CONFIG_ACPI_BLACKLIST_YEAR=0
+# CONFIG_ACPI_DEBUG is not set
+CONFIG_ACPI_EC=y
+CONFIG_ACPI_POWER=y
+CONFIG_ACPI_SYSTEM=y
+CONFIG_ACPI_CONTAINER=y
+
+#
+# CPU Frequency scaling
+#
+# CONFIG_CPU_FREQ is not set
+
+#
+# Bus options (PCI, PCMCIA)
+#
+CONFIG_PCI=y
+CONFIG_PCI_DOMAINS=y
+CONFIG_XEN_PCIDEV_FRONTEND=y
+# CONFIG_XEN_PCIDEV_FE_DEBUG is not set
+# CONFIG_PCIEPORTBUS is not set
+# CONFIG_PCI_DEBUG is not set
+
+#
+# PCI Hotplug Support
+#
+CONFIG_HOTPLUG_PCI=y
+# CONFIG_HOTPLUG_PCI_FAKE is not set
+CONFIG_HOTPLUG_PCI_ACPI=y
+# CONFIG_HOTPLUG_PCI_ACPI_IBM is not set
+# CONFIG_HOTPLUG_PCI_CPCI is not set
+# CONFIG_HOTPLUG_PCI_SHPC is not set
+# CONFIG_HOTPLUG_PCI_SGI is not set
+
+#
+# PCCARD (PCMCIA/CardBus) support
+#
+# CONFIG_PCCARD is not set
+
+#
+# Networking
+#
+CONFIG_NET=y
+
+#
+# Networking options
+#
+# CONFIG_NETDEBUG is not set
+CONFIG_PACKET=y
+# CONFIG_PACKET_MMAP is not set
+CONFIG_UNIX=y
+CONFIG_XFRM=y
+# CONFIG_XFRM_USER is not set
+# CONFIG_NET_KEY is not set
+CONFIG_INET=y
+CONFIG_IP_MULTICAST=y
+# CONFIG_IP_ADVANCED_ROUTER is not set
+CONFIG_IP_FIB_HASH=y
+CONFIG_IP_PNP=y
+CONFIG_IP_PNP_DHCP=y
+# CONFIG_IP_PNP_BOOTP is not set
+# CONFIG_IP_PNP_RARP is not set
+# CONFIG_NET_IPIP is not set
+# CONFIG_NET_IPGRE is not set
+# CONFIG_IP_MROUTE is not set
+CONFIG_ARPD=y
+CONFIG_SYN_COOKIES=y
+# CONFIG_INET_AH is not set
+# CONFIG_INET_ESP is not set
+# CONFIG_INET_IPCOMP is not set
+# CONFIG_INET_XFRM_TUNNEL is not set
+# CONFIG_INET_TUNNEL is not set
+CONFIG_INET_XFRM_MODE_TRANSPORT=y
+CONFIG_INET_XFRM_MODE_TUNNEL=y
+CONFIG_INET_DIAG=y
+CONFIG_INET_TCP_DIAG=y
+# CONFIG_TCP_CONG_ADVANCED is not set
+CONFIG_TCP_CONG_BIC=y
+
+#
+# IP: Virtual Server Configuration
+#
+# CONFIG_IP_VS is not set
+# CONFIG_IPV6 is not set
+# CONFIG_INET6_XFRM_TUNNEL is not set
+# CONFIG_INET6_TUNNEL is not set
+# CONFIG_NETWORK_SECMARK is not set
+CONFIG_NETFILTER=y
+# CONFIG_NETFILTER_DEBUG is not set
+CONFIG_BRIDGE_NETFILTER=y
+
+#
+# Core Netfilter Configuration
+#
+# CONFIG_NETFILTER_NETLINK is not set
+# CONFIG_NF_CONNTRACK is not set
+# CONFIG_NETFILTER_XTABLES is not set
+
+#
+# IP: Netfilter Configuration
+#
+# CONFIG_IP_NF_CONNTRACK is not set
+# CONFIG_IP_NF_QUEUE is not set
+
+#
+# Bridge: Netfilter Configuration
+#
+# CONFIG_BRIDGE_NF_EBTABLES is not set
+
+#
+# DCCP Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP_DCCP is not set
+
+#
+# SCTP Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP_SCTP is not set
+
+#
+# TIPC Configuration (EXPERIMENTAL)
+#
+# CONFIG_TIPC is not set
+# CONFIG_ATM is not set
+CONFIG_BRIDGE=y
+# CONFIG_VLAN_8021Q is not set
+# CONFIG_DECNET is not set
+CONFIG_LLC=y
+# CONFIG_LLC2 is not set
+# CONFIG_IPX is not set
+# CONFIG_ATALK is not set
+# CONFIG_X25 is not set
+# CONFIG_LAPB is not set
+# CONFIG_ECONET is not set
+# CONFIG_WAN_ROUTER is not set
+
+#
+# QoS and/or fair queueing
+#
+# CONFIG_NET_SCHED is not set
+
+#
+# Network testing
+#
+# CONFIG_NET_PKTGEN is not set
+# CONFIG_HAMRADIO is not set
+# CONFIG_IRDA is not set
+# CONFIG_BT is not set
+# CONFIG_IEEE80211 is not set
+
+#
+# Device Drivers
+#
+
+#
+# Generic Driver Options
+#
+CONFIG_STANDALONE=y
+CONFIG_PREVENT_FIRMWARE_BUILD=y
+CONFIG_FW_LOADER=y
+# CONFIG_DEBUG_DRIVER is not set
+CONFIG_SYS_HYPERVISOR=y
+
+#
+# Connector - unified userspace <-> kernelspace linker
+#
+# CONFIG_CONNECTOR is not set
+
+#
+# Memory Technology Devices (MTD)
+#
+# CONFIG_MTD is not set
+
+#
+# Parallel port support
+#
+# CONFIG_PARPORT is not set
+
+#
+# Plug and Play support
+#
+# CONFIG_PNP is not set
+
+#
+# Block devices
+#
+# CONFIG_BLK_CPQ_DA is not set
+CONFIG_BLK_CPQ_CISS_DA=y
+# CONFIG_CISS_SCSI_TAPE is not set
+# CONFIG_BLK_DEV_DAC960 is not set
+# CONFIG_BLK_DEV_UMEM is not set
+# CONFIG_BLK_DEV_COW_COMMON is not set
+CONFIG_BLK_DEV_LOOP=y
+CONFIG_BLK_DEV_CRYPTOLOOP=y
+CONFIG_BLK_DEV_NBD=m
+# CONFIG_BLK_DEV_SX8 is not set
+# CONFIG_BLK_DEV_UB is not set
+CONFIG_BLK_DEV_RAM=y
+CONFIG_BLK_DEV_RAM_COUNT=16
+CONFIG_BLK_DEV_RAM_SIZE=4096
+CONFIG_BLK_DEV_RAM_BLOCKSIZE=1024
+CONFIG_BLK_DEV_INITRD=y
+# CONFIG_CDROM_PKTCDVD is not set
+# CONFIG_ATA_OVER_ETH is not set
+
+#
+# ATA/ATAPI/MFM/RLL support
+#
+CONFIG_IDE=y
+CONFIG_IDE_MAX_HWIFS=4
+CONFIG_BLK_DEV_IDE=y
+
+#
+# Please see Documentation/ide.txt for help/info on IDE drives
+#
+# CONFIG_BLK_DEV_IDE_SATA is not set
+CONFIG_BLK_DEV_IDEDISK=y
+# CONFIG_IDEDISK_MULTI_MODE is not set
+CONFIG_BLK_DEV_IDECD=y
+# CONFIG_BLK_DEV_IDETAPE is not set
+CONFIG_BLK_DEV_IDEFLOPPY=y
+CONFIG_BLK_DEV_IDESCSI=y
+# CONFIG_IDE_TASK_IOCTL is not set
+
+#
+# IDE chipset support/bugfixes
+#
+# CONFIG_IDE_GENERIC is not set
+CONFIG_BLK_DEV_IDEPCI=y
+# CONFIG_IDEPCI_SHARE_IRQ is not set
+# CONFIG_BLK_DEV_OFFBOARD is not set
+CONFIG_BLK_DEV_GENERIC=y
+# CONFIG_BLK_DEV_OPTI621 is not set
+CONFIG_BLK_DEV_IDEDMA_PCI=y
+# CONFIG_BLK_DEV_IDEDMA_FORCED is not set
+CONFIG_IDEDMA_PCI_AUTO=y
+# CONFIG_IDEDMA_ONLYDISK is not set
+# CONFIG_BLK_DEV_AEC62XX is not set
+# CONFIG_BLK_DEV_ALI15X3 is not set
+# CONFIG_BLK_DEV_AMD74XX is not set
+CONFIG_BLK_DEV_CMD64X=y
+# CONFIG_BLK_DEV_TRIFLEX is not set
+# CONFIG_BLK_DEV_CY82C693 is not set
+# CONFIG_BLK_DEV_CS5520 is not set
+# CONFIG_BLK_DEV_CS5530 is not set
+# CONFIG_BLK_DEV_HPT34X is not set
+# CONFIG_BLK_DEV_HPT366 is not set
+# CONFIG_BLK_DEV_SC1200 is not set
+CONFIG_BLK_DEV_PIIX=y
+# CONFIG_BLK_DEV_IT821X is not set
+# CONFIG_BLK_DEV_NS87415 is not set
+# CONFIG_BLK_DEV_PDC202XX_OLD is not set
+# CONFIG_BLK_DEV_PDC202XX_NEW is not set
+# CONFIG_BLK_DEV_SVWKS is not set
+# CONFIG_BLK_DEV_SIIMAGE is not set
+# CONFIG_BLK_DEV_SLC90E66 is not set
+# CONFIG_BLK_DEV_TRM290 is not set
+# CONFIG_BLK_DEV_VIA82CXXX is not set
+# CONFIG_IDE_ARM is not set
+CONFIG_BLK_DEV_IDEDMA=y
+# CONFIG_IDEDMA_IVB is not set
+CONFIG_IDEDMA_AUTO=y
+# CONFIG_BLK_DEV_HD is not set
+
+#
+# SCSI device support
+#
+# CONFIG_RAID_ATTRS is not set
+CONFIG_SCSI=y
+CONFIG_SCSI_PROC_FS=y
+
+#
+# SCSI support type (disk, tape, CD-ROM)
+#
+CONFIG_BLK_DEV_SD=y
+CONFIG_CHR_DEV_ST=y
+CONFIG_CHR_DEV_OSST=y
+CONFIG_BLK_DEV_SR=y
+CONFIG_BLK_DEV_SR_VENDOR=y
+CONFIG_CHR_DEV_SG=y
+# CONFIG_CHR_DEV_SCH is not set
+
+#
+# Some SCSI devices (e.g. CD jukebox) support multiple LUNs
+#
+CONFIG_SCSI_MULTI_LUN=y
+CONFIG_SCSI_CONSTANTS=y
+CONFIG_SCSI_LOGGING=y
+
+#
+# SCSI Transport Attributes
+#
+CONFIG_SCSI_SPI_ATTRS=y
+CONFIG_SCSI_FC_ATTRS=y
+# CONFIG_SCSI_ISCSI_ATTRS is not set
+CONFIG_SCSI_SAS_ATTRS=y
+
+#
+# SCSI low-level drivers
+#
+# CONFIG_ISCSI_TCP is not set
+# CONFIG_BLK_DEV_3W_XXXX_RAID is not set
+# CONFIG_SCSI_3W_9XXX is not set
+# CONFIG_SCSI_ACARD is not set
+# CONFIG_SCSI_AACRAID is not set
+# CONFIG_SCSI_AIC7XXX is not set
+# CONFIG_SCSI_AIC7XXX_OLD is not set
+# CONFIG_SCSI_AIC79XX is not set
+# CONFIG_MEGARAID_NEWGEN is not set
+# CONFIG_MEGARAID_LEGACY is not set
+# CONFIG_MEGARAID_SAS is not set
+# CONFIG_SCSI_SATA is not set
+# CONFIG_SCSI_HPTIOP is not set
+# CONFIG_SCSI_DMX3191D is not set
+# CONFIG_SCSI_FUTURE_DOMAIN is not set
+# CONFIG_SCSI_IPS is not set
+# CONFIG_SCSI_INITIO is not set
+# CONFIG_SCSI_INIA100 is not set
+CONFIG_SCSI_SYM53C8XX_2=y
+CONFIG_SCSI_SYM53C8XX_DMA_ADDRESSING_MODE=1
+CONFIG_SCSI_SYM53C8XX_DEFAULT_TAGS=16
+CONFIG_SCSI_SYM53C8XX_MAX_TAGS=64
+CONFIG_SCSI_SYM53C8XX_MMIO=y
+# CONFIG_SCSI_IPR is not set
+CONFIG_SCSI_QLOGIC_1280=y
+# CONFIG_SCSI_QLA_FC is not set
+# CONFIG_SCSI_LPFC is not set
+# CONFIG_SCSI_DC395x is not set
+# CONFIG_SCSI_DC390T is not set
+# CONFIG_SCSI_DEBUG is not set
+
+#
+# Multi-device support (RAID and LVM)
+#
+CONFIG_MD=y
+# CONFIG_BLK_DEV_MD is not set
+CONFIG_BLK_DEV_DM=y
+CONFIG_DM_CRYPT=m
+CONFIG_DM_SNAPSHOT=y
+CONFIG_DM_MIRROR=m
+CONFIG_DM_ZERO=m
+CONFIG_DM_MULTIPATH=m
+CONFIG_DM_MULTIPATH_EMC=m
+
+#
+# Fusion MPT device support
+#
+CONFIG_FUSION=y
+CONFIG_FUSION_SPI=y
+# CONFIG_FUSION_FC is not set
+CONFIG_FUSION_SAS=y
+CONFIG_FUSION_MAX_SGE=128
+# CONFIG_FUSION_CTL is not set
+
+#
+# IEEE 1394 (FireWire) support
+#
+# CONFIG_IEEE1394 is not set
+
+#
+# I2O device support
+#
+# CONFIG_I2O is not set
+
+#
+# Network device support
+#
+CONFIG_NETDEVICES=y
+CONFIG_DUMMY=y
+# CONFIG_BONDING is not set
+# CONFIG_EQUALIZER is not set
+CONFIG_TUN=y
+
+#
+# ARCnet devices
+#
+CONFIG_ARCNET=y
+# CONFIG_ARCNET_1201 is not set
+# CONFIG_ARCNET_1051 is not set
+# CONFIG_ARCNET_RAW is not set
+# CONFIG_ARCNET_CAP is not set
+# CONFIG_ARCNET_COM90xx is not set
+# CONFIG_ARCNET_COM90xxIO is not set
+# CONFIG_ARCNET_RIM_I is not set
+# CONFIG_ARCNET_COM20020 is not set
+
+#
+# PHY device support
+#
+# CONFIG_PHYLIB is not set
+
+#
+# Ethernet (10 or 100Mbit)
+#
+CONFIG_NET_ETHERNET=y
+CONFIG_MII=y
+# CONFIG_HAPPYMEAL is not set
+# CONFIG_SUNGEM is not set
+# CONFIG_CASSINI is not set
+# CONFIG_NET_VENDOR_3COM is not set
+
+#
+# Tulip family network device support
+#
+CONFIG_NET_TULIP=y
+# CONFIG_DE2104X is not set
+CONFIG_TULIP=y
+CONFIG_TULIP_MWI=y
+CONFIG_TULIP_MMIO=y
+CONFIG_TULIP_NAPI=y
+CONFIG_TULIP_NAPI_HW_MITIGATION=y
+# CONFIG_DE4X5 is not set
+# CONFIG_WINBOND_840 is not set
+# CONFIG_DM9102 is not set
+# CONFIG_ULI526X is not set
+# CONFIG_HP100 is not set
+CONFIG_NET_PCI=y
+# CONFIG_PCNET32 is not set
+# CONFIG_AMD8111_ETH is not set
+# CONFIG_ADAPTEC_STARFIRE is not set
+# CONFIG_B44 is not set
+# CONFIG_FORCEDETH is not set
+# CONFIG_DGRS is not set
+CONFIG_EEPRO100=y
+CONFIG_E100=y
+# CONFIG_FEALNX is not set
+# CONFIG_NATSEMI is not set
+# CONFIG_NE2K_PCI is not set
+# CONFIG_8139CP is not set
+# CONFIG_8139TOO is not set
+# CONFIG_SIS900 is not set
+# CONFIG_EPIC100 is not set
+# CONFIG_SUNDANCE is not set
+# CONFIG_VIA_RHINE is not set
+
+#
+# Ethernet (1000 Mbit)
+#
+# CONFIG_ACENIC is not set
+# CONFIG_DL2K is not set
+CONFIG_E1000=y
+# CONFIG_E1000_NAPI is not set
+# CONFIG_E1000_DISABLE_PACKET_SPLIT is not set
+# CONFIG_NS83820 is not set
+# CONFIG_HAMACHI is not set
+# CONFIG_YELLOWFIN is not set
+# CONFIG_R8169 is not set
+# CONFIG_SIS190 is not set
+# CONFIG_SKGE is not set
+# CONFIG_SKY2 is not set
+# CONFIG_SK98LIN is not set
+# CONFIG_VIA_VELOCITY is not set
+CONFIG_TIGON3=y
+# CONFIG_BNX2 is not set
+
+#
+# Ethernet (10000 Mbit)
+#
+# CONFIG_CHELSIO_T1 is not set
+# CONFIG_IXGB is not set
+# CONFIG_S2IO is not set
+# CONFIG_MYRI10GE is not set
+# CONFIG_SFC is not set
+
+#
+# Token Ring devices
+#
+# CONFIG_TR is not set
+
+#
+# Wireless LAN (non-hamradio)
+#
+# CONFIG_NET_RADIO is not set
+
+#
+# Wan interfaces
+#
+# CONFIG_WAN is not set
+# CONFIG_FDDI is not set
+# CONFIG_HIPPI is not set
+# CONFIG_PPP is not set
+# CONFIG_SLIP is not set
+# CONFIG_NET_FC is not set
+# CONFIG_SHAPER is not set
+CONFIG_NETCONSOLE=y
+CONFIG_NETPOLL=y
+# CONFIG_NETPOLL_RX is not set
+# CONFIG_NETPOLL_TRAP is not set
+CONFIG_NET_POLL_CONTROLLER=y
+
+#
+# ISDN subsystem
+#
+CONFIG_ISDN=m
+
+#
+# Old ISDN4Linux
+#
+# CONFIG_ISDN_I4L is not set
+
+#
+# CAPI subsystem
+#
+# CONFIG_ISDN_CAPI is not set
+
+#
+# Telephony Support
+#
+# CONFIG_PHONE is not set
+
+#
+# Input device support
+#
+CONFIG_INPUT=y
+
+#
+# Userland interfaces
+#
+CONFIG_INPUT_MOUSEDEV=y
+CONFIG_INPUT_MOUSEDEV_PSAUX=y
+CONFIG_INPUT_MOUSEDEV_SCREEN_X=1024
+CONFIG_INPUT_MOUSEDEV_SCREEN_Y=768
+CONFIG_INPUT_JOYDEV=y
+# CONFIG_INPUT_TSDEV is not set
+CONFIG_INPUT_EVDEV=y
+# CONFIG_INPUT_EVBUG is not set
+
+#
+# Input Device Drivers
+#
+CONFIG_INPUT_KEYBOARD=y
+CONFIG_KEYBOARD_ATKBD=y
+# CONFIG_KEYBOARD_SUNKBD is not set
+# CONFIG_KEYBOARD_LKKBD is not set
+# CONFIG_KEYBOARD_XTKBD is not set
+# CONFIG_KEYBOARD_NEWTON is not set
+CONFIG_INPUT_MOUSE=y
+CONFIG_MOUSE_PS2=y
+# CONFIG_MOUSE_SERIAL is not set
+# CONFIG_MOUSE_VSXXXAA is not set
+# CONFIG_INPUT_JOYSTICK is not set
+# CONFIG_INPUT_TOUCHSCREEN is not set
+# CONFIG_INPUT_MISC is not set
+
+#
+# Hardware I/O ports
+#
+CONFIG_SERIO=y
+CONFIG_SERIO_I8042=y
+# CONFIG_SERIO_SERPORT is not set
+# CONFIG_SERIO_PCIPS2 is not set
+CONFIG_SERIO_LIBPS2=y
+# CONFIG_SERIO_RAW is not set
+CONFIG_GAMEPORT=y
+# CONFIG_GAMEPORT_NS558 is not set
+# CONFIG_GAMEPORT_L4 is not set
+# CONFIG_GAMEPORT_EMU10K1 is not set
+# CONFIG_GAMEPORT_FM801 is not set
+
+#
+# Character devices
+#
+CONFIG_VT=y
+CONFIG_VT_CONSOLE=y
+CONFIG_HW_CONSOLE=y
+# CONFIG_VT_HW_CONSOLE_BINDING is not set
+CONFIG_SERIAL_NONSTANDARD=y
+# CONFIG_COMPUTONE is not set
+# CONFIG_ROCKETPORT is not set
+# CONFIG_CYCLADES is not set
+# CONFIG_DIGIEPCA is not set
+# CONFIG_MOXA_INTELLIO is not set
+# CONFIG_MOXA_SMARTIO is not set
+# CONFIG_ISI is not set
+# CONFIG_SYNCLINKMP is not set
+# CONFIG_SYNCLINK_GT is not set
+# CONFIG_N_HDLC is not set
+# CONFIG_SPECIALIX is not set
+# CONFIG_SX is not set
+# CONFIG_RIO is not set
+# CONFIG_STALDRV is not set
+# CONFIG_SGI_SNSC is not set
+# CONFIG_SGI_TIOCX is not set
+
+#
+# Serial drivers
+#
+
+#
+# Non-8250 serial port support
+#
+# CONFIG_SERIAL_SGI_L1_CONSOLE is not set
+# CONFIG_SERIAL_JSM is not set
+CONFIG_UNIX98_PTYS=y
+CONFIG_LEGACY_PTYS=y
+CONFIG_LEGACY_PTY_COUNT=256
+
+#
+# IPMI
+#
+# CONFIG_IPMI_HANDLER is not set
+
+#
+# Watchdog Cards
+#
+# CONFIG_WATCHDOG is not set
+# CONFIG_HW_RANDOM is not set
+CONFIG_EFI_RTC=y
+# CONFIG_DTLK is not set
+# CONFIG_R3964 is not set
+# CONFIG_APPLICOM is not set
+
+#
+# Ftape, the floppy tape device driver
+#
+CONFIG_AGP=y
+# CONFIG_AGP_SIS is not set
+# CONFIG_AGP_VIA is not set
+CONFIG_AGP_I460=y
+# CONFIG_AGP_HP_ZX1 is not set
+# CONFIG_AGP_SGI_TIOCA is not set
+CONFIG_DRM=y
+# CONFIG_DRM_TDFX is not set
+# CONFIG_DRM_R128 is not set
+# CONFIG_DRM_RADEON is not set
+# CONFIG_DRM_MGA is not set
+# CONFIG_DRM_SIS is not set
+# CONFIG_DRM_VIA is not set
+# CONFIG_DRM_SAVAGE is not set
+# CONFIG_RAW_DRIVER is not set
+# CONFIG_HPET is not set
+# CONFIG_HANGCHECK_TIMER is not set
+# CONFIG_MMTIMER is not set
+
+#
+# TPM devices
+#
+# CONFIG_TCG_TPM is not set
+# CONFIG_TELCLOCK is not set
+
+#
+# I2C support
+#
+CONFIG_I2C=y
+CONFIG_I2C_CHARDEV=y
+
+#
+# I2C Algorithms
+#
+CONFIG_I2C_ALGOBIT=y
+CONFIG_I2C_ALGOPCF=y
+# CONFIG_I2C_ALGOPCA is not set
+
+#
+# I2C Hardware Bus support
+#
+# CONFIG_I2C_ALI1535 is not set
+# CONFIG_I2C_ALI1563 is not set
+# CONFIG_I2C_ALI15X3 is not set
+# CONFIG_I2C_AMD756 is not set
+# CONFIG_I2C_AMD8111 is not set
+# CONFIG_I2C_I801 is not set
+# CONFIG_I2C_I810 is not set
+# CONFIG_I2C_PIIX4 is not set
+# CONFIG_I2C_NFORCE2 is not set
+# CONFIG_I2C_OCORES is not set
+# CONFIG_I2C_PARPORT_LIGHT is not set
+# CONFIG_I2C_PROSAVAGE is not set
+# CONFIG_I2C_SAVAGE4 is not set
+# CONFIG_I2C_SIS5595 is not set
+# CONFIG_I2C_SIS630 is not set
+# CONFIG_I2C_SIS96X is not set
+# CONFIG_I2C_STUB is not set
+# CONFIG_I2C_VIA is not set
+# CONFIG_I2C_VIAPRO is not set
+# CONFIG_I2C_VOODOO3 is not set
+# CONFIG_I2C_PCA_ISA is not set
+
+#
+# Miscellaneous I2C Chip support
+#
+# CONFIG_SENSORS_DS1337 is not set
+# CONFIG_SENSORS_DS1374 is not set
+# CONFIG_SENSORS_EEPROM is not set
+# CONFIG_SENSORS_PCF8574 is not set
+# CONFIG_SENSORS_PCA9539 is not set
+# CONFIG_SENSORS_PCF8591 is not set
+# CONFIG_SENSORS_MAX6875 is not set
+# CONFIG_I2C_DEBUG_CORE is not set
+# CONFIG_I2C_DEBUG_ALGO is not set
+# CONFIG_I2C_DEBUG_BUS is not set
+# CONFIG_I2C_DEBUG_CHIP is not set
+
+#
+# SPI support
+#
+# CONFIG_SPI is not set
+# CONFIG_SPI_MASTER is not set
+
+#
+# Dallas's 1-wire bus
+#
+
+#
+# Hardware Monitoring support
+#
+CONFIG_HWMON=y
+# CONFIG_HWMON_VID is not set
+# CONFIG_SENSORS_ABITUGURU is not set
+# CONFIG_SENSORS_ADM1021 is not set
+# CONFIG_SENSORS_ADM1025 is not set
+# CONFIG_SENSORS_ADM1026 is not set
+# CONFIG_SENSORS_ADM1031 is not set
+# CONFIG_SENSORS_ADM9240 is not set
+# CONFIG_SENSORS_ASB100 is not set
+# CONFIG_SENSORS_ATXP1 is not set
+# CONFIG_SENSORS_DS1621 is not set
+# CONFIG_SENSORS_F71805F is not set
+# CONFIG_SENSORS_FSCHER is not set
+# CONFIG_SENSORS_FSCPOS is not set
+# CONFIG_SENSORS_GL518SM is not set
+# CONFIG_SENSORS_GL520SM is not set
+# CONFIG_SENSORS_IT87 is not set
+# CONFIG_SENSORS_LM63 is not set
+# CONFIG_SENSORS_LM75 is not set
+# CONFIG_SENSORS_LM77 is not set
+# CONFIG_SENSORS_LM78 is not set
+# CONFIG_SENSORS_LM80 is not set
+# CONFIG_SENSORS_LM83 is not set
+# CONFIG_SENSORS_LM85 is not set
+# CONFIG_SENSORS_LM87 is not set
+# CONFIG_SENSORS_LM90 is not set
+# CONFIG_SENSORS_LM92 is not set
+# CONFIG_SENSORS_MAX1619 is not set
+# CONFIG_SENSORS_PC87360 is not set
+# CONFIG_SENSORS_SIS5595 is not set
+# CONFIG_SENSORS_SMSC47M1 is not set
+# CONFIG_SENSORS_SMSC47M192 is not set
+# CONFIG_SENSORS_SMSC47B397 is not set
+# CONFIG_SENSORS_VIA686A is not set
+# CONFIG_SENSORS_VT8231 is not set
+# CONFIG_SENSORS_W83781D is not set
+# CONFIG_SENSORS_W83791D is not set
+# CONFIG_SENSORS_W83792D is not set
+# CONFIG_SENSORS_W83L785TS is not set
+# CONFIG_SENSORS_W83627HF is not set
+# CONFIG_SENSORS_W83627EHF is not set
+# CONFIG_HWMON_DEBUG_CHIP is not set
+
+#
+# Misc devices
+#
+
+#
+# Multimedia devices
+#
+CONFIG_VIDEO_DEV=y
+CONFIG_VIDEO_V4L1=y
+CONFIG_VIDEO_V4L1_COMPAT=y
+CONFIG_VIDEO_V4L2=y
+
+#
+# Video Capture Adapters
+#
+
+#
+# Video Capture Adapters
+#
+# CONFIG_VIDEO_ADV_DEBUG is not set
+# CONFIG_VIDEO_VIVI is not set
+# CONFIG_VIDEO_BT848 is not set
+# CONFIG_VIDEO_CPIA is not set
+# CONFIG_VIDEO_CPIA2 is not set
+# CONFIG_VIDEO_SAA5246A is not set
+# CONFIG_VIDEO_SAA5249 is not set
+# CONFIG_TUNER_3036 is not set
+# CONFIG_VIDEO_STRADIS is not set
+# CONFIG_VIDEO_ZORAN is not set
+# CONFIG_VIDEO_SAA7134 is not set
+# CONFIG_VIDEO_MXB is not set
+# CONFIG_VIDEO_DPC is not set
+# CONFIG_VIDEO_HEXIUM_ORION is not set
+# CONFIG_VIDEO_HEXIUM_GEMINI is not set
+# CONFIG_VIDEO_CX88 is not set
+
+#
+# Encoders and Decoders
+#
+# CONFIG_VIDEO_MSP3400 is not set
+# CONFIG_VIDEO_CS53L32A is not set
+# CONFIG_VIDEO_TLV320AIC23B is not set
+# CONFIG_VIDEO_WM8775 is not set
+# CONFIG_VIDEO_WM8739 is not set
+# CONFIG_VIDEO_CX2341X is not set
+# CONFIG_VIDEO_CX25840 is not set
+# CONFIG_VIDEO_SAA711X is not set
+# CONFIG_VIDEO_SAA7127 is not set
+# CONFIG_VIDEO_UPD64031A is not set
+# CONFIG_VIDEO_UPD64083 is not set
+
+#
+# V4L USB devices
+#
+# CONFIG_VIDEO_PVRUSB2 is not set
+# CONFIG_VIDEO_EM28XX is not set
+# CONFIG_USB_VICAM is not set
+# CONFIG_USB_IBMCAM is not set
+# CONFIG_USB_KONICAWC is not set
+# CONFIG_USB_QUICKCAM_MESSENGER is not set
+# CONFIG_USB_ET61X251 is not set
+# CONFIG_VIDEO_OVCAMCHIP is not set
+# CONFIG_USB_W9968CF is not set
+# CONFIG_USB_OV511 is not set
+# CONFIG_USB_SE401 is not set
+# CONFIG_USB_SN9C102 is not set
+# CONFIG_USB_STV680 is not set
+# CONFIG_USB_ZC0301 is not set
+# CONFIG_USB_PWC is not set
+
+#
+# Radio Adapters
+#
+# CONFIG_RADIO_GEMTEK_PCI is not set
+# CONFIG_RADIO_MAXIRADIO is not set
+# CONFIG_RADIO_MAESTRO is not set
+# CONFIG_USB_DSBR is not set
+
+#
+# Digital Video Broadcasting Devices
+#
+# CONFIG_DVB is not set
+# CONFIG_USB_DABUSB is not set
+
+#
+# Graphics support
+#
+CONFIG_FIRMWARE_EDID=y
+CONFIG_FB=y
+CONFIG_FB_CFB_FILLRECT=y
+CONFIG_FB_CFB_COPYAREA=y
+CONFIG_FB_CFB_IMAGEBLIT=y
+# CONFIG_FB_MACMODES is not set
+# CONFIG_FB_BACKLIGHT is not set
+CONFIG_FB_MODE_HELPERS=y
+# CONFIG_FB_TILEBLITTING is not set
+# CONFIG_FB_CIRRUS is not set
+# CONFIG_FB_PM2 is not set
+# CONFIG_FB_CYBER2000 is not set
+# CONFIG_FB_ASILIANT is not set
+# CONFIG_FB_IMSTT is not set
+# CONFIG_FB_S1D13XXX is not set
+# CONFIG_FB_NVIDIA is not set
+# CONFIG_FB_RIVA is not set
+# CONFIG_FB_MATROX is not set
+# CONFIG_FB_RADEON is not set
+# CONFIG_FB_ATY128 is not set
+# CONFIG_FB_ATY is not set
+# CONFIG_FB_SAVAGE is not set
+# CONFIG_FB_SIS is not set
+# CONFIG_FB_NEOMAGIC is not set
+# CONFIG_FB_KYRO is not set
+# CONFIG_FB_3DFX is not set
+# CONFIG_FB_VOODOO1 is not set
+# CONFIG_FB_TRIDENT is not set
+# CONFIG_FB_VIRTUAL is not set
+
+#
+# Console display driver support
+#
+CONFIG_VGA_CONSOLE=y
+# CONFIG_VGACON_SOFT_SCROLLBACK is not set
+CONFIG_DUMMY_CONSOLE=y
+CONFIG_FRAMEBUFFER_CONSOLE=y
+# CONFIG_FRAMEBUFFER_CONSOLE_ROTATION is not set
+# CONFIG_FONTS is not set
+CONFIG_FONT_8x8=y
+CONFIG_FONT_8x16=y
+
+#
+# Logo configuration
+#
+CONFIG_LOGO=y
+# CONFIG_LOGO_LINUX_MONO is not set
+# CONFIG_LOGO_LINUX_VGA16 is not set
+CONFIG_LOGO_LINUX_CLUT224=y
+# CONFIG_BACKLIGHT_LCD_SUPPORT is not set
+
+#
+# Sound
+#
+CONFIG_SOUND=y
+
+#
+# Advanced Linux Sound Architecture
+#
+CONFIG_SND=y
+CONFIG_SND_TIMER=y
+CONFIG_SND_PCM=y
+CONFIG_SND_HWDEP=y
+CONFIG_SND_RAWMIDI=y
+CONFIG_SND_SEQUENCER=y
+CONFIG_SND_SEQ_DUMMY=y
+CONFIG_SND_OSSEMUL=y
+CONFIG_SND_MIXER_OSS=y
+CONFIG_SND_PCM_OSS=y
+CONFIG_SND_PCM_OSS_PLUGINS=y
+CONFIG_SND_SEQUENCER_OSS=y
+# CONFIG_SND_DYNAMIC_MINORS is not set
+CONFIG_SND_SUPPORT_OLD_API=y
+CONFIG_SND_VERBOSE_PROCFS=y
+# CONFIG_SND_VERBOSE_PRINTK is not set
+# CONFIG_SND_DEBUG is not set
+
+#
+# Generic devices
+#
+CONFIG_SND_MPU401_UART=y
+CONFIG_SND_OPL3_LIB=y
+CONFIG_SND_AC97_CODEC=y
+CONFIG_SND_AC97_BUS=y
+CONFIG_SND_DUMMY=y
+CONFIG_SND_VIRMIDI=y
+# CONFIG_SND_MTPAV is not set
+CONFIG_SND_SERIAL_U16550=y
+CONFIG_SND_MPU401=y
+
+#
+# PCI devices
+#
+# CONFIG_SND_AD1889 is not set
+# CONFIG_SND_ALS300 is not set
+# CONFIG_SND_ALI5451 is not set
+CONFIG_SND_ATIIXP=y
+# CONFIG_SND_ATIIXP_MODEM is not set
+# CONFIG_SND_AU8810 is not set
+# CONFIG_SND_AU8820 is not set
+# CONFIG_SND_AU8830 is not set
+# CONFIG_SND_AZT3328 is not set
+# CONFIG_SND_BT87X is not set
+# CONFIG_SND_CA0106 is not set
+# CONFIG_SND_CMIPCI is not set
+# CONFIG_SND_CS4281 is not set
+# CONFIG_SND_CS46XX is not set
+# CONFIG_SND_DARLA20 is not set
+# CONFIG_SND_GINA20 is not set
+# CONFIG_SND_LAYLA20 is not set
+# CONFIG_SND_DARLA24 is not set
+# CONFIG_SND_GINA24 is not set
+# CONFIG_SND_LAYLA24 is not set
+# CONFIG_SND_MONA is not set
+# CONFIG_SND_MIA is not set
+# CONFIG_SND_ECHO3G is not set
+# CONFIG_SND_INDIGO is not set
+# CONFIG_SND_INDIGOIO is not set
+# CONFIG_SND_INDIGODJ is not set
+# CONFIG_SND_EMU10K1 is not set
+# CONFIG_SND_EMU10K1X is not set
+# CONFIG_SND_ENS1370 is not set
+# CONFIG_SND_ENS1371 is not set
+# CONFIG_SND_ES1938 is not set
+# CONFIG_SND_ES1968 is not set
+CONFIG_SND_FM801=y
+# CONFIG_SND_FM801_TEA575X_BOOL is not set
+# CONFIG_SND_HDA_INTEL is not set
+# CONFIG_SND_HDSP is not set
+# CONFIG_SND_HDSPM is not set
+# CONFIG_SND_ICE1712 is not set
+# CONFIG_SND_ICE1724 is not set
+# CONFIG_SND_INTEL8X0 is not set
+# CONFIG_SND_INTEL8X0M is not set
+# CONFIG_SND_KORG1212 is not set
+# CONFIG_SND_MAESTRO3 is not set
+# CONFIG_SND_MIXART is not set
+# CONFIG_SND_NM256 is not set
+# CONFIG_SND_PCXHR is not set
+# CONFIG_SND_RIPTIDE is not set
+# CONFIG_SND_RME32 is not set
+# CONFIG_SND_RME96 is not set
+# CONFIG_SND_RME9652 is not set
+# CONFIG_SND_SONICVIBES is not set
+# CONFIG_SND_TRIDENT is not set
+# CONFIG_SND_VIA82XX is not set
+# CONFIG_SND_VIA82XX_MODEM is not set
+# CONFIG_SND_VX222 is not set
+# CONFIG_SND_YMFPCI is not set
+
+#
+# USB devices
+#
+# CONFIG_SND_USB_AUDIO is not set
+
+#
+# Open Sound System
+#
+CONFIG_SOUND_PRIME=y
+# CONFIG_OSS_OBSOLETE_DRIVER is not set
+# CONFIG_SOUND_BT878 is not set
+# CONFIG_SOUND_ES1371 is not set
+# CONFIG_SOUND_ICH is not set
+# CONFIG_SOUND_TRIDENT is not set
+# CONFIG_SOUND_MSNDCLAS is not set
+# CONFIG_SOUND_MSNDPIN is not set
+# CONFIG_SOUND_VIA82CXXX is not set
+# CONFIG_SOUND_TVMIXER is not set
+
+#
+# USB support
+#
+CONFIG_USB_ARCH_HAS_HCD=y
+CONFIG_USB_ARCH_HAS_OHCI=y
+CONFIG_USB_ARCH_HAS_EHCI=y
+CONFIG_USB=y
+# CONFIG_USB_DEBUG is not set
+
+#
+# Miscellaneous USB options
+#
+CONFIG_USB_DEVICEFS=y
+CONFIG_USB_BANDWIDTH=y
+# CONFIG_USB_DYNAMIC_MINORS is not set
+# CONFIG_USB_SUSPEND is not set
+# CONFIG_USB_OTG is not set
+
+#
+# USB Host Controller Drivers
+#
+CONFIG_USB_EHCI_HCD=y
+# CONFIG_USB_EHCI_SPLIT_ISO is not set
+# CONFIG_USB_EHCI_ROOT_HUB_TT is not set
+# CONFIG_USB_EHCI_TT_NEWSCHED is not set
+# CONFIG_USB_ISP116X_HCD is not set
+CONFIG_USB_OHCI_HCD=y
+# CONFIG_USB_OHCI_BIG_ENDIAN is not set
+CONFIG_USB_OHCI_LITTLE_ENDIAN=y
+CONFIG_USB_UHCI_HCD=y
+# CONFIG_USB_SL811_HCD is not set
+
+#
+# USB Device Class drivers
+#
+# CONFIG_USB_ACM is not set
+# CONFIG_USB_PRINTER is not set
+
+#
+# NOTE: USB_STORAGE enables SCSI, and 'SCSI disk support'
+#
+
+#
+# may also be needed; see USB_STORAGE Help for more information
+#
+CONFIG_USB_STORAGE=y
+# CONFIG_USB_STORAGE_DEBUG is not set
+# CONFIG_USB_STORAGE_DATAFAB is not set
+# CONFIG_USB_STORAGE_FREECOM is not set
+# CONFIG_USB_STORAGE_ISD200 is not set
+# CONFIG_USB_STORAGE_DPCM is not set
+# CONFIG_USB_STORAGE_USBAT is not set
+# CONFIG_USB_STORAGE_SDDR09 is not set
+# CONFIG_USB_STORAGE_SDDR55 is not set
+# CONFIG_USB_STORAGE_JUMPSHOT is not set
+# CONFIG_USB_STORAGE_ALAUDA is not set
+# CONFIG_USB_LIBUSUAL is not set
+
+#
+# USB Input Devices
+#
+CONFIG_USB_HID=y
+CONFIG_USB_HIDINPUT=y
+# CONFIG_USB_HIDINPUT_POWERBOOK is not set
+# CONFIG_HID_FF is not set
+CONFIG_USB_HIDDEV=y
+# CONFIG_USB_AIPTEK is not set
+# CONFIG_USB_WACOM is not set
+# CONFIG_USB_ACECAD is not set
+# CONFIG_USB_KBTAB is not set
+# CONFIG_USB_POWERMATE is not set
+# CONFIG_USB_TOUCHSCREEN is not set
+# CONFIG_USB_YEALINK is not set
+# CONFIG_USB_XPAD is not set
+# CONFIG_USB_ATI_REMOTE is not set
+# CONFIG_USB_ATI_REMOTE2 is not set
+# CONFIG_USB_KEYSPAN_REMOTE is not set
+# CONFIG_USB_APPLETOUCH is not set
+
+#
+# USB Imaging devices
+#
+# CONFIG_USB_MDC800 is not set
+# CONFIG_USB_MICROTEK is not set
+
+#
+# USB Network Adapters
+#
+# CONFIG_USB_CATC is not set
+# CONFIG_USB_KAWETH is not set
+# CONFIG_USB_PEGASUS is not set
+# CONFIG_USB_RTL8150 is not set
+# CONFIG_USB_USBNET is not set
+CONFIG_USB_MON=y
+
+#
+# USB port drivers
+#
+
+#
+# USB Serial Converter support
+#
+# CONFIG_USB_SERIAL is not set
+
+#
+# USB Miscellaneous drivers
+#
+# CONFIG_USB_EMI62 is not set
+# CONFIG_USB_EMI26 is not set
+# CONFIG_USB_AUERSWALD is not set
+# CONFIG_USB_RIO500 is not set
+# CONFIG_USB_LEGOTOWER is not set
+# CONFIG_USB_LCD is not set
+# CONFIG_USB_LED is not set
+# CONFIG_USB_CYPRESS_CY7C63 is not set
+# CONFIG_USB_CYTHERM is not set
+# CONFIG_USB_PHIDGETKIT is not set
+# CONFIG_USB_PHIDGETSERVO is not set
+# CONFIG_USB_IDMOUSE is not set
+# CONFIG_USB_APPLEDISPLAY is not set
+# CONFIG_USB_SISUSBVGA is not set
+# CONFIG_USB_LD is not set
+# CONFIG_USB_TEST is not set
+
+#
+# USB DSL modem support
+#
+
+#
+# USB Gadget Support
+#
+# CONFIG_USB_GADGET is not set
+
+#
+# MMC/SD Card support
+#
+# CONFIG_MMC is not set
+
+#
+# LED devices
+#
+# CONFIG_NEW_LEDS is not set
+
+#
+# LED drivers
+#
+
+#
+# LED Triggers
+#
+
+#
+# InfiniBand support
+#
+# CONFIG_INFINIBAND is not set
+
+#
+# EDAC - error detection and reporting (RAS) (EXPERIMENTAL)
+#
+
+#
+# Real Time Clock
+#
+# CONFIG_RTC_CLASS is not set
+
+#
+# DMA Engine support
+#
+# CONFIG_DMA_ENGINE is not set
+
+#
+# DMA Clients
+#
+
+#
+# DMA Devices
+#
+
+#
+# File systems
+#
+CONFIG_EXT2_FS=y
+CONFIG_EXT2_FS_XATTR=y
+CONFIG_EXT2_FS_POSIX_ACL=y
+CONFIG_EXT2_FS_SECURITY=y
+# CONFIG_EXT2_FS_XIP is not set
+CONFIG_EXT3_FS=y
+CONFIG_EXT3_FS_XATTR=y
+CONFIG_EXT3_FS_POSIX_ACL=y
+CONFIG_EXT3_FS_SECURITY=y
+CONFIG_JBD=y
+# CONFIG_JBD_DEBUG is not set
+CONFIG_FS_MBCACHE=y
+CONFIG_REISERFS_FS=y
+# CONFIG_REISERFS_CHECK is not set
+# CONFIG_REISERFS_PROC_INFO is not set
+CONFIG_REISERFS_FS_XATTR=y
+CONFIG_REISERFS_FS_POSIX_ACL=y
+CONFIG_REISERFS_FS_SECURITY=y
+# CONFIG_JFS_FS is not set
+CONFIG_FS_POSIX_ACL=y
+CONFIG_XFS_FS=y
+# CONFIG_XFS_QUOTA is not set
+# CONFIG_XFS_SECURITY is not set
+# CONFIG_XFS_POSIX_ACL is not set
+# CONFIG_XFS_RT is not set
+# CONFIG_OCFS2_FS is not set
+# CONFIG_MINIX_FS is not set
+# CONFIG_ROMFS_FS is not set
+CONFIG_INOTIFY=y
+CONFIG_INOTIFY_USER=y
+# CONFIG_QUOTA is not set
+CONFIG_DNOTIFY=y
+CONFIG_AUTOFS_FS=y
+CONFIG_AUTOFS4_FS=y
+# CONFIG_FUSE_FS is not set
+
+#
+# CD-ROM/DVD Filesystems
+#
+CONFIG_ISO9660_FS=y
+CONFIG_JOLIET=y
+# CONFIG_ZISOFS is not set
+CONFIG_UDF_FS=y
+CONFIG_UDF_NLS=y
+
+#
+# DOS/FAT/NT Filesystems
+#
+CONFIG_FAT_FS=y
+CONFIG_MSDOS_FS=y
+CONFIG_VFAT_FS=y
+CONFIG_FAT_DEFAULT_CODEPAGE=437
+CONFIG_FAT_DEFAULT_IOCHARSET="iso8859-1"
+# CONFIG_NTFS_FS is not set
+
+#
+# Pseudo filesystems
+#
+CONFIG_PROC_FS=y
+CONFIG_PROC_KCORE=y
+CONFIG_PROC_IOMEM_MACHINE=y
+CONFIG_SYSFS=y
+CONFIG_TMPFS=y
+# CONFIG_HUGETLB_PAGE is not set
+CONFIG_RAMFS=y
+# CONFIG_CONFIGFS_FS is not set
+
+#
+# Miscellaneous filesystems
+#
+# CONFIG_ADFS_FS is not set
+# CONFIG_AFFS_FS is not set
+# CONFIG_HFS_FS is not set
+# CONFIG_HFSPLUS_FS is not set
+# CONFIG_BEFS_FS is not set
+# CONFIG_BFS_FS is not set
+# CONFIG_EFS_FS is not set
+# CONFIG_CRAMFS is not set
+# CONFIG_VXFS_FS is not set
+# CONFIG_HPFS_FS is not set
+# CONFIG_QNX4FS_FS is not set
+# CONFIG_SYSV_FS is not set
+# CONFIG_UFS_FS is not set
+
+#
+# Network File Systems
+#
+CONFIG_NFS_FS=y
+CONFIG_NFS_V3=y
+# CONFIG_NFS_V3_ACL is not set
+CONFIG_NFS_V4=y
+CONFIG_NFS_DIRECTIO=y
+CONFIG_NFSD=y
+CONFIG_NFSD_V3=y
+# CONFIG_NFSD_V3_ACL is not set
+CONFIG_NFSD_V4=y
+CONFIG_NFSD_TCP=y
+CONFIG_ROOT_NFS=y
+CONFIG_LOCKD=y
+CONFIG_LOCKD_V4=y
+CONFIG_EXPORTFS=y
+CONFIG_NFS_COMMON=y
+CONFIG_SUNRPC=y
+CONFIG_SUNRPC_GSS=y
+CONFIG_RPCSEC_GSS_KRB5=y
+# CONFIG_RPCSEC_GSS_SPKM3 is not set
+CONFIG_SMB_FS=y
+CONFIG_SMB_NLS_DEFAULT=y
+CONFIG_SMB_NLS_REMOTE="cp437"
+CONFIG_CIFS=y
+# CONFIG_CIFS_STATS is not set
+# CONFIG_CIFS_WEAK_PW_HASH is not set
+# CONFIG_CIFS_XATTR is not set
+# CONFIG_CIFS_DEBUG2 is not set
+# CONFIG_CIFS_EXPERIMENTAL is not set
+# CONFIG_NCP_FS is not set
+# CONFIG_CODA_FS is not set
+# CONFIG_AFS_FS is not set
+# CONFIG_9P_FS is not set
+
+#
+# Partition Types
+#
+CONFIG_PARTITION_ADVANCED=y
+# CONFIG_ACORN_PARTITION is not set
+# CONFIG_OSF_PARTITION is not set
+# CONFIG_AMIGA_PARTITION is not set
+# CONFIG_ATARI_PARTITION is not set
+# CONFIG_MAC_PARTITION is not set
+CONFIG_MSDOS_PARTITION=y
+# CONFIG_BSD_DISKLABEL is not set
+# CONFIG_MINIX_SUBPARTITION is not set
+# CONFIG_SOLARIS_X86_PARTITION is not set
+# CONFIG_UNIXWARE_DISKLABEL is not set
+# CONFIG_LDM_PARTITION is not set
+CONFIG_SGI_PARTITION=y
+# CONFIG_ULTRIX_PARTITION is not set
+# CONFIG_SUN_PARTITION is not set
+# CONFIG_KARMA_PARTITION is not set
+CONFIG_EFI_PARTITION=y
+
+#
+# Native Language Support
+#
+CONFIG_NLS=y
+CONFIG_NLS_DEFAULT="iso8859-1"
+CONFIG_NLS_CODEPAGE_437=y
+CONFIG_NLS_CODEPAGE_737=y
+CONFIG_NLS_CODEPAGE_775=y
+CONFIG_NLS_CODEPAGE_850=y
+CONFIG_NLS_CODEPAGE_852=y
+CONFIG_NLS_CODEPAGE_855=y
+CONFIG_NLS_CODEPAGE_857=y
+CONFIG_NLS_CODEPAGE_860=y
+CONFIG_NLS_CODEPAGE_861=y
+CONFIG_NLS_CODEPAGE_862=y
+CONFIG_NLS_CODEPAGE_863=y
+CONFIG_NLS_CODEPAGE_864=y
+CONFIG_NLS_CODEPAGE_865=y
+CONFIG_NLS_CODEPAGE_866=y
+CONFIG_NLS_CODEPAGE_869=y
+CONFIG_NLS_CODEPAGE_936=y
+CONFIG_NLS_CODEPAGE_950=y
+CONFIG_NLS_CODEPAGE_932=y
+CONFIG_NLS_CODEPAGE_949=y
+CONFIG_NLS_CODEPAGE_874=y
+CONFIG_NLS_ISO8859_8=y
+# CONFIG_NLS_CODEPAGE_1250 is not set
+CONFIG_NLS_CODEPAGE_1251=y
+# CONFIG_NLS_ASCII is not set
+CONFIG_NLS_ISO8859_1=y
+CONFIG_NLS_ISO8859_2=y
+CONFIG_NLS_ISO8859_3=y
+CONFIG_NLS_ISO8859_4=y
+CONFIG_NLS_ISO8859_5=y
+CONFIG_NLS_ISO8859_6=y
+CONFIG_NLS_ISO8859_7=y
+CONFIG_NLS_ISO8859_9=y
+CONFIG_NLS_ISO8859_13=y
+CONFIG_NLS_ISO8859_14=y
+CONFIG_NLS_ISO8859_15=y
+CONFIG_NLS_KOI8_R=y
+CONFIG_NLS_KOI8_U=y
+CONFIG_NLS_UTF8=y
+
+#
+# Library routines
+#
+# CONFIG_CRC_CCITT is not set
+# CONFIG_CRC16 is not set
+CONFIG_CRC32=y
+# CONFIG_LIBCRC32C is not set
+CONFIG_PLIST=y
+CONFIG_GENERIC_HARDIRQS=y
+CONFIG_GENERIC_IRQ_PROBE=y
+CONFIG_GENERIC_PENDING_IRQ=y
+CONFIG_IRQ_PER_CPU=y
+
+#
+# HP Simulator drivers
+#
+# CONFIG_HP_SIMETH is not set
+# CONFIG_HP_SIMSERIAL is not set
+# CONFIG_HP_SIMSCSI is not set
+
+#
+# Instrumentation Support
+#
+# CONFIG_PROFILING is not set
+# CONFIG_KPROBES is not set
+
+#
+# Kernel hacking
+#
+# CONFIG_PRINTK_TIME is not set
+CONFIG_MAGIC_SYSRQ=y
+CONFIG_UNUSED_SYMBOLS=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_LOG_BUF_SHIFT=20
+CONFIG_DETECT_SOFTLOCKUP=y
+# CONFIG_SCHEDSTATS is not set
+# CONFIG_DEBUG_SLAB is not set
+# CONFIG_DEBUG_RT_MUTEXES is not set
+# CONFIG_RT_MUTEX_TESTER is not set
+# CONFIG_DEBUG_SPINLOCK is not set
+CONFIG_DEBUG_MUTEXES=y
+# CONFIG_DEBUG_RWSEMS is not set
+# CONFIG_DEBUG_SPINLOCK_SLEEP is not set
+# CONFIG_DEBUG_LOCKING_API_SELFTESTS is not set
+# CONFIG_DEBUG_KOBJECT is not set
+# CONFIG_DEBUG_INFO is not set
+# CONFIG_DEBUG_FS is not set
+# CONFIG_DEBUG_VM is not set
+CONFIG_FORCED_INLINING=y
+# CONFIG_RCU_TORTURE_TEST is not set
+CONFIG_IA64_GRANULE_16MB=y
+# CONFIG_IA64_GRANULE_64MB is not set
+CONFIG_IA64_PRINT_HAZARDS=y
+# CONFIG_DISABLE_VHPT is not set
+# CONFIG_IA64_DEBUG_CMPXCHG is not set
+# CONFIG_IA64_DEBUG_IRQ is not set
+
+#
+# Security options
+#
+# CONFIG_KEYS is not set
+# CONFIG_SECURITY is not set
+
+#
+# Cryptographic options
+#
+CONFIG_CRYPTO=y
+# CONFIG_CRYPTO_HMAC is not set
+# CONFIG_CRYPTO_NULL is not set
+# CONFIG_CRYPTO_MD4 is not set
+CONFIG_CRYPTO_MD5=y
+# CONFIG_CRYPTO_SHA1 is not set
+# CONFIG_CRYPTO_SHA256 is not set
+# CONFIG_CRYPTO_SHA512 is not set
+# CONFIG_CRYPTO_WP512 is not set
+# CONFIG_CRYPTO_TGR192 is not set
+CONFIG_CRYPTO_DES=y
+# CONFIG_CRYPTO_BLOWFISH is not set
+# CONFIG_CRYPTO_TWOFISH is not set
+# CONFIG_CRYPTO_SERPENT is not set
+# CONFIG_CRYPTO_AES is not set
+# CONFIG_CRYPTO_CAST5 is not set
+# CONFIG_CRYPTO_CAST6 is not set
+# CONFIG_CRYPTO_TEA is not set
+# CONFIG_CRYPTO_ARC4 is not set
+# CONFIG_CRYPTO_KHAZAD is not set
+# CONFIG_CRYPTO_ANUBIS is not set
+# CONFIG_CRYPTO_DEFLATE is not set
+# CONFIG_CRYPTO_MICHAEL_MIC is not set
+# CONFIG_CRYPTO_CRC32C is not set
+# CONFIG_CRYPTO_TEST is not set
+
+#
+# Hardware crypto devices
+#
+# CONFIG_XEN_SMPBOOT is not set
+# CONFIG_XEN_DEVMEM is not set
+CONFIG_XEN_INTERFACE_VERSION=0x00030207
+
+#
+# XEN
+#
+CONFIG_XEN_PRIVILEGED_GUEST=y
+# CONFIG_XEN_UNPRIVILEGED_GUEST is not set
+CONFIG_XEN_PRIVCMD=y
+CONFIG_XEN_XENBUS_DEV=y
+CONFIG_XEN_BACKEND=y
+CONFIG_XEN_BLKDEV_BACKEND=y
+CONFIG_XEN_BLKDEV_TAP=y
+CONFIG_XEN_NETDEV_BACKEND=y
+# CONFIG_XEN_NETDEV_PIPELINED_TRANSMITTER is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_UTIL is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_BACKEND is not set
+# CONFIG_XEN_NETDEV_LOOPBACK is not set
+CONFIG_XEN_PCIDEV_BACKEND=y
+# CONFIG_XEN_PCIDEV_BACKEND_VPCI is not set
+# CONFIG_XEN_PCIDEV_BACKEND_PASS is not set
+# CONFIG_XEN_PCIDEV_BACKEND_SLOT is not set
+CONFIG_XEN_PCIDEV_BACKEND_CONTROLLER=y
+# CONFIG_XEN_PCIDEV_BE_DEBUG is not set
+CONFIG_XEN_TPMDEV_BACKEND=m
+CONFIG_XEN_SCSI_BACKEND=m
+CONFIG_XEN_BLKDEV_FRONTEND=y
+CONFIG_XEN_NETDEV_FRONTEND=y
+CONFIG_XEN_SCSI_FRONTEND=m
+CONFIG_XEN_GRANT_DEV=y
+# CONFIG_XEN_NETDEV_ACCEL_SFC_FRONTEND is not set
+CONFIG_XEN_FRAMEBUFFER=y
+CONFIG_XEN_KEYBOARD=y
+# CONFIG_XEN_SCRUB_PAGES is not set
+CONFIG_XEN_DISABLE_SERIAL=y
+CONFIG_XEN_SYSFS=y
+CONFIG_XEN_COMPAT_030002_AND_LATER=y
+# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
+# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
+CONFIG_XEN_COMPAT=0x030002
+CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
+CONFIG_NO_IDLE_HZ=y
+CONFIG_XEN_BALLOON=y
+CONFIG_XEN_XENCOMM=y
diff -rpuN linux-2.6.18.8/buildconfigs/linux-defconfig_xen0_x86_32 linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xen0_x86_32
--- linux-2.6.18.8/buildconfigs/linux-defconfig_xen0_x86_32	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xen0_x86_32	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1458 @@
+#
+# Automatically generated make config: don't edit
+# Linux kernel version: 2.6.18.8
+# Tue Oct 16 09:31:19 2007
+#
+CONFIG_X86_32=y
+CONFIG_LOCKDEP_SUPPORT=y
+CONFIG_STACKTRACE_SUPPORT=y
+CONFIG_SEMAPHORE_SLEEPERS=y
+CONFIG_X86=y
+CONFIG_MMU=y
+CONFIG_GENERIC_ISA_DMA=y
+CONFIG_GENERIC_IOMAP=y
+CONFIG_GENERIC_HWEIGHT=y
+CONFIG_ARCH_MAY_HAVE_PC_FDC=y
+CONFIG_DMI=y
+CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
+
+#
+# Code maturity level options
+#
+CONFIG_EXPERIMENTAL=y
+CONFIG_LOCK_KERNEL=y
+CONFIG_INIT_ENV_ARG_LIMIT=32
+
+#
+# General setup
+#
+CONFIG_LOCALVERSION=""
+# CONFIG_LOCALVERSION_AUTO is not set
+CONFIG_SWAP=y
+CONFIG_SYSVIPC=y
+# CONFIG_POSIX_MQUEUE is not set
+# CONFIG_BSD_PROCESS_ACCT is not set
+# CONFIG_TASKSTATS is not set
+# CONFIG_AUDIT is not set
+# CONFIG_IKCONFIG is not set
+# CONFIG_CPUSETS is not set
+# CONFIG_RELAY is not set
+CONFIG_INITRAMFS_SOURCE=""
+# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set
+# CONFIG_EMBEDDED is not set
+CONFIG_UID16=y
+CONFIG_SYSCTL=y
+CONFIG_KALLSYMS=y
+# CONFIG_KALLSYMS_ALL is not set
+# CONFIG_KALLSYMS_EXTRA_PASS is not set
+CONFIG_HOTPLUG=y
+CONFIG_PRINTK=y
+CONFIG_BUG=y
+CONFIG_ELF_CORE=y
+CONFIG_BASE_FULL=y
+CONFIG_FUTEX=y
+CONFIG_EPOLL=y
+CONFIG_SHMEM=y
+CONFIG_SLAB=y
+CONFIG_VM_EVENT_COUNTERS=y
+CONFIG_RT_MUTEXES=y
+# CONFIG_TINY_SHMEM is not set
+CONFIG_BASE_SMALL=0
+# CONFIG_SLOB is not set
+
+#
+# Loadable module support
+#
+CONFIG_MODULES=y
+CONFIG_MODULE_UNLOAD=y
+# CONFIG_MODULE_FORCE_UNLOAD is not set
+# CONFIG_MODVERSIONS is not set
+# CONFIG_MODULE_SRCVERSION_ALL is not set
+CONFIG_KMOD=y
+CONFIG_STOP_MACHINE=y
+
+#
+# Block layer
+#
+# CONFIG_LBD is not set
+# CONFIG_BLK_DEV_IO_TRACE is not set
+CONFIG_LSF=y
+
+#
+# IO Schedulers
+#
+CONFIG_IOSCHED_NOOP=y
+CONFIG_IOSCHED_AS=y
+CONFIG_IOSCHED_DEADLINE=y
+CONFIG_IOSCHED_CFQ=y
+CONFIG_DEFAULT_AS=y
+# CONFIG_DEFAULT_DEADLINE is not set
+# CONFIG_DEFAULT_CFQ is not set
+# CONFIG_DEFAULT_NOOP is not set
+CONFIG_DEFAULT_IOSCHED="anticipatory"
+
+#
+# Processor type and features
+#
+CONFIG_SMP=y
+# CONFIG_X86_PC is not set
+CONFIG_X86_XEN=y
+# CONFIG_X86_ELAN is not set
+# CONFIG_X86_VOYAGER is not set
+# CONFIG_X86_NUMAQ is not set
+# CONFIG_X86_SUMMIT is not set
+# CONFIG_X86_BIGSMP is not set
+# CONFIG_X86_VISWS is not set
+# CONFIG_X86_GENERICARCH is not set
+# CONFIG_X86_ES7000 is not set
+# CONFIG_M386 is not set
+# CONFIG_M486 is not set
+# CONFIG_M586 is not set
+# CONFIG_M586TSC is not set
+# CONFIG_M586MMX is not set
+CONFIG_M686=y
+# CONFIG_MPENTIUMII is not set
+# CONFIG_MPENTIUMIII is not set
+# CONFIG_MPENTIUMM is not set
+# CONFIG_MPENTIUM4 is not set
+# CONFIG_MK6 is not set
+# CONFIG_MK7 is not set
+# CONFIG_MK8 is not set
+# CONFIG_MCRUSOE is not set
+# CONFIG_MEFFICEON is not set
+# CONFIG_MWINCHIPC6 is not set
+# CONFIG_MWINCHIP2 is not set
+# CONFIG_MWINCHIP3D is not set
+# CONFIG_MGEODEGX1 is not set
+# CONFIG_MGEODE_LX is not set
+# CONFIG_MCYRIXIII is not set
+# CONFIG_MVIAC3_2 is not set
+# CONFIG_X86_GENERIC is not set
+CONFIG_X86_CMPXCHG=y
+CONFIG_X86_XADD=y
+CONFIG_X86_L1_CACHE_SHIFT=5
+CONFIG_RWSEM_XCHGADD_ALGORITHM=y
+CONFIG_GENERIC_CALIBRATE_DELAY=y
+CONFIG_X86_PPRO_FENCE=y
+CONFIG_X86_WP_WORKS_OK=y
+CONFIG_X86_INVLPG=y
+CONFIG_X86_BSWAP=y
+CONFIG_X86_POPAD_OK=y
+CONFIG_X86_CMPXCHG64=y
+CONFIG_X86_GOOD_APIC=y
+CONFIG_X86_USE_PPRO_CHECKSUM=y
+CONFIG_NR_CPUS=8
+CONFIG_PREEMPT_NONE=y
+# CONFIG_PREEMPT_VOLUNTARY is not set
+# CONFIG_PREEMPT is not set
+CONFIG_PREEMPT_BKL=y
+CONFIG_X86_LOCAL_APIC=y
+CONFIG_X86_IO_APIC=y
+CONFIG_VM86=y
+# CONFIG_TOSHIBA is not set
+# CONFIG_I8K is not set
+# CONFIG_X86_REBOOTFIXUPS is not set
+CONFIG_MICROCODE=y
+CONFIG_X86_MSR=y
+CONFIG_X86_CPUID=y
+CONFIG_SWIOTLB=y
+
+#
+# Firmware Drivers
+#
+CONFIG_EDD=y
+# CONFIG_DELL_RBU is not set
+# CONFIG_DCDBAS is not set
+# CONFIG_NOHIGHMEM is not set
+CONFIG_HIGHMEM4G=y
+# CONFIG_HIGHMEM64G is not set
+CONFIG_PAGE_OFFSET=0xC0000000
+CONFIG_HIGHMEM=y
+CONFIG_SELECT_MEMORY_MODEL=y
+CONFIG_FLATMEM_MANUAL=y
+# CONFIG_DISCONTIGMEM_MANUAL is not set
+# CONFIG_SPARSEMEM_MANUAL is not set
+CONFIG_FLATMEM=y
+CONFIG_FLAT_NODE_MEM_MAP=y
+# CONFIG_SPARSEMEM_STATIC is not set
+CONFIG_SPLIT_PTLOCK_CPUS=4
+CONFIG_RESOURCES_64BIT=y
+# CONFIG_HIGHPTE is not set
+CONFIG_MTRR=y
+# CONFIG_REGPARM is not set
+CONFIG_SECCOMP=y
+CONFIG_HZ_100=y
+# CONFIG_HZ_250 is not set
+# CONFIG_HZ_1000 is not set
+CONFIG_HZ=100
+CONFIG_KEXEC=y
+# CONFIG_CRASH_DUMP is not set
+CONFIG_PHYSICAL_START=0x100000
+CONFIG_HOTPLUG_CPU=y
+CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y
+
+#
+# Power management options (ACPI, APM)
+#
+CONFIG_PM=y
+CONFIG_PM_LEGACY=y
+# CONFIG_PM_DEBUG is not set
+# CONFIG_SOFTWARE_SUSPEND is not set
+CONFIG_SUSPEND_SMP=y
+
+#
+# ACPI (Advanced Configuration and Power Interface) Support
+#
+CONFIG_ACPI=y
+CONFIG_ACPI_SLEEP=y
+CONFIG_ACPI_SLEEP_PROC_FS=y
+# CONFIG_ACPI_SLEEP_PROC_SLEEP is not set
+CONFIG_ACPI_AC=m
+CONFIG_ACPI_BATTERY=m
+CONFIG_ACPI_BUTTON=m
+CONFIG_ACPI_VIDEO=m
+CONFIG_ACPI_HOTKEY=m
+CONFIG_ACPI_FAN=m
+CONFIG_ACPI_DOCK=m
+CONFIG_ACPI_PROCESSOR=m
+CONFIG_ACPI_HOTPLUG_CPU=y
+CONFIG_ACPI_THERMAL=m
+CONFIG_ACPI_ASUS=m
+CONFIG_ACPI_IBM=m
+CONFIG_ACPI_TOSHIBA=m
+# CONFIG_ACPI_CUSTOM_DSDT is not set
+CONFIG_ACPI_BLACKLIST_YEAR=0
+# CONFIG_ACPI_DEBUG is not set
+CONFIG_ACPI_EC=y
+CONFIG_ACPI_POWER=y
+CONFIG_ACPI_SYSTEM=y
+CONFIG_ACPI_CONTAINER=m
+CONFIG_ACPI_PV_SLEEP=y
+
+#
+# CPU Frequency scaling
+#
+# CONFIG_CPU_FREQ is not set
+
+#
+# Bus options (PCI, PCMCIA, EISA, MCA, ISA)
+#
+CONFIG_PCI=y
+# CONFIG_PCI_GOBIOS is not set
+# CONFIG_PCI_GOMMCONFIG is not set
+# CONFIG_PCI_GODIRECT is not set
+# CONFIG_PCI_GOXEN_FE is not set
+CONFIG_PCI_GOANY=y
+CONFIG_PCI_DIRECT=y
+CONFIG_PCI_MMCONFIG=y
+CONFIG_XEN_PCIDEV_FRONTEND=y
+# CONFIG_XEN_PCIDEV_FE_DEBUG is not set
+# CONFIG_PCIEPORTBUS is not set
+# CONFIG_PCI_MSI is not set
+# CONFIG_PCI_DEBUG is not set
+CONFIG_ISA_DMA_API=y
+# CONFIG_SCx200 is not set
+CONFIG_K8_NB=y
+
+#
+# PCCARD (PCMCIA/CardBus) support
+#
+# CONFIG_PCCARD is not set
+
+#
+# PCI Hotplug Support
+#
+# CONFIG_HOTPLUG_PCI is not set
+
+#
+# Executable file formats
+#
+CONFIG_BINFMT_ELF=y
+# CONFIG_BINFMT_AOUT is not set
+# CONFIG_BINFMT_MISC is not set
+
+#
+# Networking
+#
+CONFIG_NET=y
+
+#
+# Networking options
+#
+# CONFIG_NETDEBUG is not set
+CONFIG_PACKET=y
+# CONFIG_PACKET_MMAP is not set
+CONFIG_UNIX=y
+CONFIG_XFRM=y
+# CONFIG_XFRM_USER is not set
+# CONFIG_NET_KEY is not set
+CONFIG_INET=y
+# CONFIG_IP_MULTICAST is not set
+# CONFIG_IP_ADVANCED_ROUTER is not set
+CONFIG_IP_FIB_HASH=y
+CONFIG_IP_PNP=y
+CONFIG_IP_PNP_DHCP=y
+# CONFIG_IP_PNP_BOOTP is not set
+# CONFIG_IP_PNP_RARP is not set
+# CONFIG_NET_IPIP is not set
+# CONFIG_NET_IPGRE is not set
+# CONFIG_ARPD is not set
+# CONFIG_SYN_COOKIES is not set
+# CONFIG_INET_AH is not set
+# CONFIG_INET_ESP is not set
+# CONFIG_INET_IPCOMP is not set
+# CONFIG_INET_XFRM_TUNNEL is not set
+# CONFIG_INET_TUNNEL is not set
+CONFIG_INET_XFRM_MODE_TRANSPORT=y
+CONFIG_INET_XFRM_MODE_TUNNEL=y
+# CONFIG_INET_DIAG is not set
+# CONFIG_TCP_CONG_ADVANCED is not set
+CONFIG_TCP_CONG_BIC=y
+
+#
+# IP: Virtual Server Configuration
+#
+# CONFIG_IP_VS is not set
+# CONFIG_IPV6 is not set
+# CONFIG_INET6_XFRM_TUNNEL is not set
+# CONFIG_INET6_TUNNEL is not set
+# CONFIG_NETWORK_SECMARK is not set
+CONFIG_NETFILTER=y
+# CONFIG_NETFILTER_DEBUG is not set
+CONFIG_BRIDGE_NETFILTER=y
+
+#
+# Core Netfilter Configuration
+#
+# CONFIG_NETFILTER_NETLINK is not set
+# CONFIG_NETFILTER_XTABLES is not set
+
+#
+# IP: Netfilter Configuration
+#
+CONFIG_IP_NF_CONNTRACK=m
+CONFIG_IP_NF_CT_ACCT=y
+# CONFIG_IP_NF_CONNTRACK_MARK is not set
+# CONFIG_IP_NF_CONNTRACK_EVENTS is not set
+# CONFIG_IP_NF_CT_PROTO_SCTP is not set
+CONFIG_IP_NF_FTP=m
+# CONFIG_IP_NF_IRC is not set
+# CONFIG_IP_NF_NETBIOS_NS is not set
+# CONFIG_IP_NF_TFTP is not set
+# CONFIG_IP_NF_AMANDA is not set
+# CONFIG_IP_NF_PPTP is not set
+# CONFIG_IP_NF_H323 is not set
+# CONFIG_IP_NF_SIP is not set
+# CONFIG_IP_NF_QUEUE is not set
+
+#
+# Bridge: Netfilter Configuration
+#
+# CONFIG_BRIDGE_NF_EBTABLES is not set
+
+#
+# DCCP Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP_DCCP is not set
+
+#
+# SCTP Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP_SCTP is not set
+
+#
+# TIPC Configuration (EXPERIMENTAL)
+#
+# CONFIG_TIPC is not set
+# CONFIG_ATM is not set
+CONFIG_BRIDGE=y
+# CONFIG_VLAN_8021Q is not set
+# CONFIG_DECNET is not set
+CONFIG_LLC=y
+# CONFIG_LLC2 is not set
+# CONFIG_IPX is not set
+# CONFIG_ATALK is not set
+# CONFIG_X25 is not set
+# CONFIG_LAPB is not set
+# CONFIG_ECONET is not set
+# CONFIG_WAN_ROUTER is not set
+
+#
+# QoS and/or fair queueing
+#
+# CONFIG_NET_SCHED is not set
+
+#
+# Network testing
+#
+# CONFIG_NET_PKTGEN is not set
+# CONFIG_HAMRADIO is not set
+# CONFIG_IRDA is not set
+# CONFIG_BT is not set
+# CONFIG_IEEE80211 is not set
+
+#
+# Device Drivers
+#
+
+#
+# Generic Driver Options
+#
+# CONFIG_STANDALONE is not set
+CONFIG_PREVENT_FIRMWARE_BUILD=y
+# CONFIG_FW_LOADER is not set
+# CONFIG_DEBUG_DRIVER is not set
+CONFIG_SYS_HYPERVISOR=y
+
+#
+# Connector - unified userspace <-> kernelspace linker
+#
+# CONFIG_CONNECTOR is not set
+
+#
+# Memory Technology Devices (MTD)
+#
+# CONFIG_MTD is not set
+
+#
+# Parallel port support
+#
+# CONFIG_PARPORT is not set
+
+#
+# Plug and Play support
+#
+CONFIG_PNP=y
+CONFIG_PNP_DEBUG=y
+
+#
+# Protocols
+#
+CONFIG_PNPACPI=y
+
+#
+# Block devices
+#
+CONFIG_BLK_DEV_FD=y
+# CONFIG_BLK_CPQ_DA is not set
+CONFIG_BLK_CPQ_CISS_DA=y
+# CONFIG_CISS_SCSI_TAPE is not set
+# CONFIG_BLK_DEV_DAC960 is not set
+# CONFIG_BLK_DEV_UMEM is not set
+# CONFIG_BLK_DEV_COW_COMMON is not set
+CONFIG_BLK_DEV_LOOP=y
+# CONFIG_BLK_DEV_CRYPTOLOOP is not set
+# CONFIG_BLK_DEV_NBD is not set
+# CONFIG_BLK_DEV_SX8 is not set
+# CONFIG_BLK_DEV_UB is not set
+CONFIG_BLK_DEV_RAM=y
+CONFIG_BLK_DEV_RAM_COUNT=16
+CONFIG_BLK_DEV_RAM_SIZE=4096
+CONFIG_BLK_DEV_RAM_BLOCKSIZE=1024
+CONFIG_BLK_DEV_INITRD=y
+# CONFIG_CDROM_PKTCDVD is not set
+# CONFIG_ATA_OVER_ETH is not set
+
+#
+# ATA/ATAPI/MFM/RLL support
+#
+CONFIG_IDE=y
+CONFIG_BLK_DEV_IDE=y
+
+#
+# Please see Documentation/ide.txt for help/info on IDE drives
+#
+# CONFIG_BLK_DEV_IDE_SATA is not set
+# CONFIG_BLK_DEV_HD_IDE is not set
+CONFIG_BLK_DEV_IDEDISK=y
+# CONFIG_IDEDISK_MULTI_MODE is not set
+CONFIG_BLK_DEV_IDECD=y
+# CONFIG_BLK_DEV_IDETAPE is not set
+# CONFIG_BLK_DEV_IDEFLOPPY is not set
+# CONFIG_BLK_DEV_IDESCSI is not set
+# CONFIG_IDE_TASK_IOCTL is not set
+
+#
+# IDE chipset support/bugfixes
+#
+CONFIG_IDE_GENERIC=y
+# CONFIG_BLK_DEV_CMD640 is not set
+# CONFIG_BLK_DEV_IDEPNP is not set
+CONFIG_BLK_DEV_IDEPCI=y
+# CONFIG_IDEPCI_SHARE_IRQ is not set
+# CONFIG_BLK_DEV_OFFBOARD is not set
+CONFIG_BLK_DEV_GENERIC=y
+# CONFIG_BLK_DEV_OPTI621 is not set
+# CONFIG_BLK_DEV_RZ1000 is not set
+CONFIG_BLK_DEV_IDEDMA_PCI=y
+# CONFIG_BLK_DEV_IDEDMA_FORCED is not set
+CONFIG_IDEDMA_PCI_AUTO=y
+# CONFIG_IDEDMA_ONLYDISK is not set
+# CONFIG_BLK_DEV_AEC62XX is not set
+# CONFIG_BLK_DEV_ALI15X3 is not set
+# CONFIG_BLK_DEV_AMD74XX is not set
+# CONFIG_BLK_DEV_ATIIXP is not set
+# CONFIG_BLK_DEV_CMD64X is not set
+# CONFIG_BLK_DEV_TRIFLEX is not set
+# CONFIG_BLK_DEV_CY82C693 is not set
+# CONFIG_BLK_DEV_CS5520 is not set
+# CONFIG_BLK_DEV_CS5530 is not set
+# CONFIG_BLK_DEV_CS5535 is not set
+# CONFIG_BLK_DEV_HPT34X is not set
+# CONFIG_BLK_DEV_HPT366 is not set
+# CONFIG_BLK_DEV_SC1200 is not set
+CONFIG_BLK_DEV_PIIX=y
+# CONFIG_BLK_DEV_IT821X is not set
+# CONFIG_BLK_DEV_NS87415 is not set
+# CONFIG_BLK_DEV_PDC202XX_OLD is not set
+# CONFIG_BLK_DEV_PDC202XX_NEW is not set
+CONFIG_BLK_DEV_SVWKS=y
+# CONFIG_BLK_DEV_SIIMAGE is not set
+# CONFIG_BLK_DEV_SIS5513 is not set
+# CONFIG_BLK_DEV_SLC90E66 is not set
+# CONFIG_BLK_DEV_TRM290 is not set
+# CONFIG_BLK_DEV_VIA82CXXX is not set
+# CONFIG_IDE_ARM is not set
+CONFIG_BLK_DEV_IDEDMA=y
+# CONFIG_IDEDMA_IVB is not set
+CONFIG_IDEDMA_AUTO=y
+# CONFIG_BLK_DEV_HD is not set
+
+#
+# SCSI device support
+#
+# CONFIG_RAID_ATTRS is not set
+CONFIG_SCSI=y
+CONFIG_SCSI_PROC_FS=y
+
+#
+# SCSI support type (disk, tape, CD-ROM)
+#
+CONFIG_BLK_DEV_SD=y
+# CONFIG_CHR_DEV_ST is not set
+# CONFIG_CHR_DEV_OSST is not set
+# CONFIG_BLK_DEV_SR is not set
+# CONFIG_CHR_DEV_SG is not set
+# CONFIG_CHR_DEV_SCH is not set
+
+#
+# Some SCSI devices (e.g. CD jukebox) support multiple LUNs
+#
+# CONFIG_SCSI_MULTI_LUN is not set
+# CONFIG_SCSI_CONSTANTS is not set
+# CONFIG_SCSI_LOGGING is not set
+
+#
+# SCSI Transport Attributes
+#
+CONFIG_SCSI_SPI_ATTRS=y
+# CONFIG_SCSI_FC_ATTRS is not set
+# CONFIG_SCSI_ISCSI_ATTRS is not set
+# CONFIG_SCSI_SAS_ATTRS is not set
+
+#
+# SCSI low-level drivers
+#
+# CONFIG_ISCSI_TCP is not set
+CONFIG_BLK_DEV_3W_XXXX_RAID=y
+# CONFIG_SCSI_3W_9XXX is not set
+# CONFIG_SCSI_ACARD is not set
+CONFIG_SCSI_AACRAID=y
+CONFIG_SCSI_AIC7XXX=y
+CONFIG_AIC7XXX_CMDS_PER_DEVICE=32
+CONFIG_AIC7XXX_RESET_DELAY_MS=15000
+CONFIG_AIC7XXX_DEBUG_ENABLE=y
+CONFIG_AIC7XXX_DEBUG_MASK=0
+CONFIG_AIC7XXX_REG_PRETTY_PRINT=y
+# CONFIG_SCSI_AIC7XXX_OLD is not set
+CONFIG_SCSI_AIC79XX=y
+CONFIG_AIC79XX_CMDS_PER_DEVICE=32
+CONFIG_AIC79XX_RESET_DELAY_MS=15000
+# CONFIG_AIC79XX_ENABLE_RD_STRM is not set
+CONFIG_AIC79XX_DEBUG_ENABLE=y
+CONFIG_AIC79XX_DEBUG_MASK=0
+CONFIG_AIC79XX_REG_PRETTY_PRINT=y
+# CONFIG_SCSI_DPT_I2O is not set
+CONFIG_SCSI_ADVANSYS=y
+CONFIG_MEGARAID_NEWGEN=y
+# CONFIG_MEGARAID_MM is not set
+# CONFIG_MEGARAID_LEGACY is not set
+# CONFIG_MEGARAID_SAS is not set
+CONFIG_SCSI_SATA=y
+CONFIG_SCSI_SATA_AHCI=y
+# CONFIG_SCSI_SATA_SVW is not set
+CONFIG_SCSI_ATA_PIIX=y
+# CONFIG_SCSI_SATA_MV is not set
+# CONFIG_SCSI_SATA_NV is not set
+# CONFIG_SCSI_PDC_ADMA is not set
+# CONFIG_SCSI_HPTIOP is not set
+# CONFIG_SCSI_SATA_QSTOR is not set
+CONFIG_SCSI_SATA_PROMISE=y
+CONFIG_SCSI_SATA_SX4=y
+CONFIG_SCSI_SATA_SIL=y
+CONFIG_SCSI_SATA_SIL24=y
+# CONFIG_SCSI_SATA_SIS is not set
+# CONFIG_SCSI_SATA_ULI is not set
+# CONFIG_SCSI_SATA_VIA is not set
+# CONFIG_SCSI_SATA_VITESSE is not set
+CONFIG_SCSI_SATA_INTEL_COMBINED=y
+# CONFIG_SCSI_BUSLOGIC is not set
+# CONFIG_SCSI_DMX3191D is not set
+# CONFIG_SCSI_EATA is not set
+# CONFIG_SCSI_FUTURE_DOMAIN is not set
+# CONFIG_SCSI_GDTH is not set
+# CONFIG_SCSI_IPS is not set
+# CONFIG_SCSI_INITIO is not set
+# CONFIG_SCSI_INIA100 is not set
+# CONFIG_SCSI_SYM53C8XX_2 is not set
+# CONFIG_SCSI_IPR is not set
+# CONFIG_SCSI_QLOGIC_1280 is not set
+# CONFIG_SCSI_QLA_FC is not set
+# CONFIG_SCSI_LPFC is not set
+# CONFIG_SCSI_DC395x is not set
+# CONFIG_SCSI_DC390T is not set
+# CONFIG_SCSI_NSP32 is not set
+# CONFIG_SCSI_DEBUG is not set
+
+#
+# Multi-device support (RAID and LVM)
+#
+CONFIG_MD=y
+CONFIG_BLK_DEV_MD=y
+# CONFIG_MD_LINEAR is not set
+CONFIG_MD_RAID0=y
+CONFIG_MD_RAID1=y
+# CONFIG_MD_RAID10 is not set
+# CONFIG_MD_RAID456 is not set
+# CONFIG_MD_MULTIPATH is not set
+# CONFIG_MD_FAULTY is not set
+CONFIG_BLK_DEV_DM=y
+# CONFIG_DM_CRYPT is not set
+CONFIG_DM_SNAPSHOT=y
+CONFIG_DM_MIRROR=y
+# CONFIG_DM_ZERO is not set
+# CONFIG_DM_MULTIPATH is not set
+
+#
+# Fusion MPT device support
+#
+CONFIG_FUSION=y
+CONFIG_FUSION_SPI=y
+# CONFIG_FUSION_FC is not set
+# CONFIG_FUSION_SAS is not set
+CONFIG_FUSION_MAX_SGE=128
+# CONFIG_FUSION_CTL is not set
+
+#
+# IEEE 1394 (FireWire) support
+#
+# CONFIG_IEEE1394 is not set
+
+#
+# I2O device support
+#
+# CONFIG_I2O is not set
+
+#
+# Network device support
+#
+CONFIG_NETDEVICES=y
+# CONFIG_DUMMY is not set
+# CONFIG_BONDING is not set
+# CONFIG_EQUALIZER is not set
+CONFIG_TUN=y
+# CONFIG_NET_SB1000 is not set
+
+#
+# ARCnet devices
+#
+# CONFIG_ARCNET is not set
+
+#
+# PHY device support
+#
+# CONFIG_PHYLIB is not set
+
+#
+# Ethernet (10 or 100Mbit)
+#
+CONFIG_NET_ETHERNET=y
+CONFIG_MII=y
+# CONFIG_HAPPYMEAL is not set
+# CONFIG_SUNGEM is not set
+# CONFIG_CASSINI is not set
+CONFIG_NET_VENDOR_3COM=y
+CONFIG_VORTEX=y
+# CONFIG_TYPHOON is not set
+
+#
+# Tulip family network device support
+#
+CONFIG_NET_TULIP=y
+# CONFIG_DE2104X is not set
+CONFIG_TULIP=y
+# CONFIG_TULIP_MWI is not set
+# CONFIG_TULIP_MMIO is not set
+# CONFIG_TULIP_NAPI is not set
+# CONFIG_DE4X5 is not set
+# CONFIG_WINBOND_840 is not set
+# CONFIG_DM9102 is not set
+# CONFIG_ULI526X is not set
+# CONFIG_HP100 is not set
+CONFIG_NET_PCI=y
+CONFIG_PCNET32=y
+# CONFIG_AMD8111_ETH is not set
+# CONFIG_ADAPTEC_STARFIRE is not set
+# CONFIG_B44 is not set
+# CONFIG_FORCEDETH is not set
+# CONFIG_DGRS is not set
+# CONFIG_EEPRO100 is not set
+CONFIG_E100=y
+# CONFIG_FEALNX is not set
+# CONFIG_NATSEMI is not set
+CONFIG_NE2K_PCI=y
+# CONFIG_8139CP is not set
+CONFIG_8139TOO=y
+CONFIG_8139TOO_PIO=y
+# CONFIG_8139TOO_TUNE_TWISTER is not set
+# CONFIG_8139TOO_8129 is not set
+# CONFIG_8139_OLD_RX_RESET is not set
+# CONFIG_SIS900 is not set
+# CONFIG_EPIC100 is not set
+# CONFIG_SUNDANCE is not set
+# CONFIG_TLAN is not set
+CONFIG_VIA_RHINE=y
+# CONFIG_VIA_RHINE_MMIO is not set
+# CONFIG_VIA_RHINE_NAPI is not set
+
+#
+# Ethernet (1000 Mbit)
+#
+CONFIG_ACENIC=y
+# CONFIG_ACENIC_OMIT_TIGON_I is not set
+# CONFIG_DL2K is not set
+CONFIG_E1000=y
+# CONFIG_E1000_NAPI is not set
+# CONFIG_E1000_DISABLE_PACKET_SPLIT is not set
+# CONFIG_NS83820 is not set
+# CONFIG_HAMACHI is not set
+# CONFIG_YELLOWFIN is not set
+# CONFIG_R8169 is not set
+# CONFIG_SIS190 is not set
+# CONFIG_SKGE is not set
+# CONFIG_SKY2 is not set
+CONFIG_SK98LIN=y
+# CONFIG_VIA_VELOCITY is not set
+CONFIG_TIGON3=y
+CONFIG_BNX2=y
+
+#
+# Ethernet (10000 Mbit)
+#
+# CONFIG_CHELSIO_T1 is not set
+# CONFIG_IXGB is not set
+# CONFIG_S2IO is not set
+# CONFIG_MYRI10GE is not set
+# CONFIG_SFC is not set
+
+#
+# Token Ring devices
+#
+# CONFIG_TR is not set
+
+#
+# Wireless LAN (non-hamradio)
+#
+# CONFIG_NET_RADIO is not set
+
+#
+# Wan interfaces
+#
+# CONFIG_WAN is not set
+# CONFIG_FDDI is not set
+# CONFIG_HIPPI is not set
+# CONFIG_PPP is not set
+# CONFIG_SLIP is not set
+# CONFIG_NET_FC is not set
+# CONFIG_SHAPER is not set
+# CONFIG_NETCONSOLE is not set
+# CONFIG_NETPOLL is not set
+# CONFIG_NET_POLL_CONTROLLER is not set
+
+#
+# ISDN subsystem
+#
+# CONFIG_ISDN is not set
+
+#
+# Telephony Support
+#
+# CONFIG_PHONE is not set
+
+#
+# Input device support
+#
+CONFIG_INPUT=y
+
+#
+# Userland interfaces
+#
+CONFIG_INPUT_MOUSEDEV=y
+CONFIG_INPUT_MOUSEDEV_PSAUX=y
+CONFIG_INPUT_MOUSEDEV_SCREEN_X=1024
+CONFIG_INPUT_MOUSEDEV_SCREEN_Y=768
+# CONFIG_INPUT_JOYDEV is not set
+# CONFIG_INPUT_TSDEV is not set
+# CONFIG_INPUT_EVDEV is not set
+# CONFIG_INPUT_EVBUG is not set
+
+#
+# Input Device Drivers
+#
+CONFIG_INPUT_KEYBOARD=y
+CONFIG_KEYBOARD_ATKBD=y
+# CONFIG_KEYBOARD_SUNKBD is not set
+# CONFIG_KEYBOARD_LKKBD is not set
+# CONFIG_KEYBOARD_XTKBD is not set
+# CONFIG_KEYBOARD_NEWTON is not set
+CONFIG_INPUT_MOUSE=y
+CONFIG_MOUSE_PS2=y
+# CONFIG_MOUSE_SERIAL is not set
+# CONFIG_MOUSE_VSXXXAA is not set
+# CONFIG_INPUT_JOYSTICK is not set
+# CONFIG_INPUT_TOUCHSCREEN is not set
+# CONFIG_INPUT_MISC is not set
+
+#
+# Hardware I/O ports
+#
+CONFIG_SERIO=y
+CONFIG_SERIO_I8042=y
+CONFIG_SERIO_SERPORT=y
+# CONFIG_SERIO_CT82C710 is not set
+# CONFIG_SERIO_PCIPS2 is not set
+CONFIG_SERIO_LIBPS2=y
+# CONFIG_SERIO_RAW is not set
+# CONFIG_GAMEPORT is not set
+
+#
+# Character devices
+#
+CONFIG_VT=y
+CONFIG_VT_CONSOLE=y
+CONFIG_HW_CONSOLE=y
+CONFIG_VT_HW_CONSOLE_BINDING=y
+# CONFIG_SERIAL_NONSTANDARD is not set
+
+#
+# Serial drivers
+#
+
+#
+# Non-8250 serial port support
+#
+# CONFIG_SERIAL_JSM is not set
+CONFIG_UNIX98_PTYS=y
+CONFIG_LEGACY_PTYS=y
+CONFIG_LEGACY_PTY_COUNT=256
+
+#
+# IPMI
+#
+# CONFIG_IPMI_HANDLER is not set
+
+#
+# Watchdog Cards
+#
+# CONFIG_WATCHDOG is not set
+# CONFIG_HW_RANDOM is not set
+# CONFIG_NVRAM is not set
+# CONFIG_RTC is not set
+# CONFIG_GEN_RTC is not set
+# CONFIG_DTLK is not set
+# CONFIG_R3964 is not set
+# CONFIG_APPLICOM is not set
+# CONFIG_SONYPI is not set
+
+#
+# Ftape, the floppy tape device driver
+#
+CONFIG_AGP=m
+CONFIG_AGP_ALI=m
+CONFIG_AGP_ATI=m
+CONFIG_AGP_AMD=m
+CONFIG_AGP_AMD64=m
+CONFIG_AGP_INTEL=m
+CONFIG_AGP_NVIDIA=m
+CONFIG_AGP_SIS=m
+CONFIG_AGP_SWORKS=m
+CONFIG_AGP_VIA=m
+# CONFIG_AGP_EFFICEON is not set
+CONFIG_DRM=m
+CONFIG_DRM_TDFX=m
+CONFIG_DRM_R128=m
+CONFIG_DRM_RADEON=m
+CONFIG_DRM_I810=m
+CONFIG_DRM_I830=m
+CONFIG_DRM_I915=m
+CONFIG_DRM_MGA=m
+CONFIG_DRM_SIS=m
+# CONFIG_DRM_VIA is not set
+# CONFIG_DRM_SAVAGE is not set
+# CONFIG_MWAVE is not set
+# CONFIG_PC8736x_GPIO is not set
+# CONFIG_NSC_GPIO is not set
+# CONFIG_CS5535_GPIO is not set
+# CONFIG_RAW_DRIVER is not set
+# CONFIG_HPET is not set
+# CONFIG_HANGCHECK_TIMER is not set
+
+#
+# TPM devices
+#
+# CONFIG_TCG_TPM is not set
+# CONFIG_TELCLOCK is not set
+
+#
+# I2C support
+#
+# CONFIG_I2C is not set
+
+#
+# SPI support
+#
+# CONFIG_SPI is not set
+# CONFIG_SPI_MASTER is not set
+
+#
+# Dallas's 1-wire bus
+#
+
+#
+# Hardware Monitoring support
+#
+# CONFIG_HWMON is not set
+# CONFIG_HWMON_VID is not set
+
+#
+# Misc devices
+#
+# CONFIG_IBM_ASM is not set
+
+#
+# Multimedia devices
+#
+# CONFIG_VIDEO_DEV is not set
+
+#
+# Digital Video Broadcasting Devices
+#
+# CONFIG_DVB is not set
+# CONFIG_USB_DABUSB is not set
+
+#
+# Graphics support
+#
+CONFIG_FIRMWARE_EDID=y
+# CONFIG_FB is not set
+
+#
+# Console display driver support
+#
+CONFIG_VGA_CONSOLE=y
+# CONFIG_VGACON_SOFT_SCROLLBACK is not set
+CONFIG_DUMMY_CONSOLE=y
+# CONFIG_BACKLIGHT_LCD_SUPPORT is not set
+
+#
+# Sound
+#
+# CONFIG_SOUND is not set
+
+#
+# USB support
+#
+CONFIG_USB_ARCH_HAS_HCD=y
+CONFIG_USB_ARCH_HAS_OHCI=y
+CONFIG_USB_ARCH_HAS_EHCI=y
+CONFIG_USB=y
+# CONFIG_USB_DEBUG is not set
+
+#
+# Miscellaneous USB options
+#
+# CONFIG_USB_DEVICEFS is not set
+# CONFIG_USB_BANDWIDTH is not set
+# CONFIG_USB_DYNAMIC_MINORS is not set
+# CONFIG_USB_SUSPEND is not set
+# CONFIG_USB_OTG is not set
+
+#
+# USB Host Controller Drivers
+#
+# CONFIG_USB_EHCI_HCD is not set
+# CONFIG_USB_ISP116X_HCD is not set
+CONFIG_USB_OHCI_HCD=y
+# CONFIG_USB_OHCI_BIG_ENDIAN is not set
+CONFIG_USB_OHCI_LITTLE_ENDIAN=y
+CONFIG_USB_UHCI_HCD=y
+# CONFIG_USB_SL811_HCD is not set
+
+#
+# USB Device Class drivers
+#
+# CONFIG_USB_ACM is not set
+# CONFIG_USB_PRINTER is not set
+
+#
+# NOTE: USB_STORAGE enables SCSI, and 'SCSI disk support'
+#
+
+#
+# may also be needed; see USB_STORAGE Help for more information
+#
+# CONFIG_USB_STORAGE is not set
+# CONFIG_USB_LIBUSUAL is not set
+
+#
+# USB Input Devices
+#
+CONFIG_USB_HID=y
+CONFIG_USB_HIDINPUT=y
+# CONFIG_USB_HIDINPUT_POWERBOOK is not set
+# CONFIG_HID_FF is not set
+# CONFIG_USB_HIDDEV is not set
+# CONFIG_USB_AIPTEK is not set
+# CONFIG_USB_WACOM is not set
+# CONFIG_USB_ACECAD is not set
+# CONFIG_USB_KBTAB is not set
+# CONFIG_USB_POWERMATE is not set
+# CONFIG_USB_TOUCHSCREEN is not set
+# CONFIG_USB_YEALINK is not set
+# CONFIG_USB_XPAD is not set
+# CONFIG_USB_ATI_REMOTE is not set
+# CONFIG_USB_ATI_REMOTE2 is not set
+# CONFIG_USB_KEYSPAN_REMOTE is not set
+# CONFIG_USB_APPLETOUCH is not set
+
+#
+# USB Imaging devices
+#
+# CONFIG_USB_MDC800 is not set
+# CONFIG_USB_MICROTEK is not set
+
+#
+# USB Network Adapters
+#
+# CONFIG_USB_CATC is not set
+# CONFIG_USB_KAWETH is not set
+# CONFIG_USB_PEGASUS is not set
+# CONFIG_USB_RTL8150 is not set
+# CONFIG_USB_USBNET is not set
+CONFIG_USB_MON=y
+
+#
+# USB port drivers
+#
+
+#
+# USB Serial Converter support
+#
+# CONFIG_USB_SERIAL is not set
+
+#
+# USB Miscellaneous drivers
+#
+# CONFIG_USB_EMI62 is not set
+# CONFIG_USB_EMI26 is not set
+# CONFIG_USB_AUERSWALD is not set
+# CONFIG_USB_RIO500 is not set
+# CONFIG_USB_LEGOTOWER is not set
+# CONFIG_USB_LCD is not set
+# CONFIG_USB_LED is not set
+# CONFIG_USB_CYPRESS_CY7C63 is not set
+# CONFIG_USB_CYTHERM is not set
+# CONFIG_USB_PHIDGETKIT is not set
+# CONFIG_USB_PHIDGETSERVO is not set
+# CONFIG_USB_IDMOUSE is not set
+# CONFIG_USB_APPLEDISPLAY is not set
+# CONFIG_USB_LD is not set
+
+#
+# USB DSL modem support
+#
+
+#
+# USB Gadget Support
+#
+# CONFIG_USB_GADGET is not set
+
+#
+# MMC/SD Card support
+#
+# CONFIG_MMC is not set
+
+#
+# LED devices
+#
+# CONFIG_NEW_LEDS is not set
+
+#
+# LED drivers
+#
+
+#
+# LED Triggers
+#
+
+#
+# InfiniBand support
+#
+# CONFIG_INFINIBAND is not set
+
+#
+# EDAC - error detection and reporting (RAS) (EXPERIMENTAL)
+#
+# CONFIG_EDAC is not set
+
+#
+# Real Time Clock
+#
+CONFIG_RTC_LIB=m
+CONFIG_RTC_CLASS=m
+
+#
+# RTC interfaces
+#
+CONFIG_RTC_INTF_SYSFS=m
+CONFIG_RTC_INTF_PROC=m
+CONFIG_RTC_INTF_DEV=m
+# CONFIG_RTC_INTF_DEV_UIE_EMUL is not set
+
+#
+# RTC drivers
+#
+# CONFIG_RTC_DRV_DS1553 is not set
+# CONFIG_RTC_DRV_DS1742 is not set
+CONFIG_RTC_DRV_M48T86=m
+CONFIG_RTC_DRV_TEST=m
+# CONFIG_RTC_DRV_V3020 is not set
+
+#
+# DMA Engine support
+#
+# CONFIG_DMA_ENGINE is not set
+
+#
+# DMA Clients
+#
+
+#
+# DMA Devices
+#
+
+#
+# File systems
+#
+CONFIG_EXT2_FS=y
+# CONFIG_EXT2_FS_XATTR is not set
+# CONFIG_EXT2_FS_XIP is not set
+CONFIG_EXT3_FS=y
+CONFIG_EXT3_FS_XATTR=y
+# CONFIG_EXT3_FS_POSIX_ACL is not set
+# CONFIG_EXT3_FS_SECURITY is not set
+CONFIG_JBD=y
+# CONFIG_JBD_DEBUG is not set
+CONFIG_FS_MBCACHE=y
+CONFIG_REISERFS_FS=y
+# CONFIG_REISERFS_CHECK is not set
+# CONFIG_REISERFS_PROC_INFO is not set
+# CONFIG_REISERFS_FS_XATTR is not set
+# CONFIG_JFS_FS is not set
+# CONFIG_FS_POSIX_ACL is not set
+# CONFIG_XFS_FS is not set
+# CONFIG_OCFS2_FS is not set
+# CONFIG_MINIX_FS is not set
+# CONFIG_ROMFS_FS is not set
+CONFIG_INOTIFY=y
+CONFIG_INOTIFY_USER=y
+# CONFIG_QUOTA is not set
+CONFIG_DNOTIFY=y
+CONFIG_AUTOFS_FS=y
+CONFIG_AUTOFS4_FS=y
+# CONFIG_FUSE_FS is not set
+
+#
+# CD-ROM/DVD Filesystems
+#
+CONFIG_ISO9660_FS=y
+CONFIG_JOLIET=y
+CONFIG_ZISOFS=y
+CONFIG_ZISOFS_FS=y
+# CONFIG_UDF_FS is not set
+
+#
+# DOS/FAT/NT Filesystems
+#
+CONFIG_FAT_FS=m
+CONFIG_MSDOS_FS=m
+CONFIG_VFAT_FS=m
+CONFIG_FAT_DEFAULT_CODEPAGE=437
+CONFIG_FAT_DEFAULT_IOCHARSET="iso8859-1"
+# CONFIG_NTFS_FS is not set
+
+#
+# Pseudo filesystems
+#
+CONFIG_PROC_FS=y
+CONFIG_PROC_KCORE=y
+CONFIG_SYSFS=y
+CONFIG_TMPFS=y
+# CONFIG_HUGETLB_PAGE is not set
+CONFIG_RAMFS=y
+# CONFIG_CONFIGFS_FS is not set
+
+#
+# Miscellaneous filesystems
+#
+# CONFIG_ADFS_FS is not set
+# CONFIG_AFFS_FS is not set
+# CONFIG_HFS_FS is not set
+# CONFIG_HFSPLUS_FS is not set
+# CONFIG_BEFS_FS is not set
+# CONFIG_BFS_FS is not set
+# CONFIG_EFS_FS is not set
+CONFIG_CRAMFS=y
+# CONFIG_VXFS_FS is not set
+# CONFIG_HPFS_FS is not set
+# CONFIG_QNX4FS_FS is not set
+# CONFIG_SYSV_FS is not set
+# CONFIG_UFS_FS is not set
+
+#
+# Network File Systems
+#
+CONFIG_NFS_FS=y
+CONFIG_NFS_V3=y
+# CONFIG_NFS_V3_ACL is not set
+# CONFIG_NFS_V4 is not set
+# CONFIG_NFS_DIRECTIO is not set
+CONFIG_NFSD=m
+CONFIG_NFSD_V3=y
+# CONFIG_NFSD_V3_ACL is not set
+# CONFIG_NFSD_V4 is not set
+CONFIG_NFSD_TCP=y
+CONFIG_ROOT_NFS=y
+CONFIG_LOCKD=y
+CONFIG_LOCKD_V4=y
+CONFIG_EXPORTFS=m
+CONFIG_NFS_COMMON=y
+CONFIG_SUNRPC=y
+# CONFIG_RPCSEC_GSS_KRB5 is not set
+# CONFIG_RPCSEC_GSS_SPKM3 is not set
+# CONFIG_SMB_FS is not set
+# CONFIG_CIFS is not set
+# CONFIG_NCP_FS is not set
+# CONFIG_CODA_FS is not set
+# CONFIG_AFS_FS is not set
+# CONFIG_9P_FS is not set
+
+#
+# Partition Types
+#
+# CONFIG_PARTITION_ADVANCED is not set
+CONFIG_MSDOS_PARTITION=y
+
+#
+# Native Language Support
+#
+CONFIG_NLS=y
+CONFIG_NLS_DEFAULT="iso8859-1"
+CONFIG_NLS_CODEPAGE_437=y
+# CONFIG_NLS_CODEPAGE_737 is not set
+# CONFIG_NLS_CODEPAGE_775 is not set
+# CONFIG_NLS_CODEPAGE_850 is not set
+# CONFIG_NLS_CODEPAGE_852 is not set
+# CONFIG_NLS_CODEPAGE_855 is not set
+# CONFIG_NLS_CODEPAGE_857 is not set
+# CONFIG_NLS_CODEPAGE_860 is not set
+# CONFIG_NLS_CODEPAGE_861 is not set
+# CONFIG_NLS_CODEPAGE_862 is not set
+# CONFIG_NLS_CODEPAGE_863 is not set
+# CONFIG_NLS_CODEPAGE_864 is not set
+# CONFIG_NLS_CODEPAGE_865 is not set
+# CONFIG_NLS_CODEPAGE_866 is not set
+# CONFIG_NLS_CODEPAGE_869 is not set
+# CONFIG_NLS_CODEPAGE_936 is not set
+# CONFIG_NLS_CODEPAGE_950 is not set
+# CONFIG_NLS_CODEPAGE_932 is not set
+# CONFIG_NLS_CODEPAGE_949 is not set
+# CONFIG_NLS_CODEPAGE_874 is not set
+# CONFIG_NLS_ISO8859_8 is not set
+# CONFIG_NLS_CODEPAGE_1250 is not set
+# CONFIG_NLS_CODEPAGE_1251 is not set
+# CONFIG_NLS_ASCII is not set
+CONFIG_NLS_ISO8859_1=y
+# CONFIG_NLS_ISO8859_2 is not set
+# CONFIG_NLS_ISO8859_3 is not set
+# CONFIG_NLS_ISO8859_4 is not set
+# CONFIG_NLS_ISO8859_5 is not set
+# CONFIG_NLS_ISO8859_6 is not set
+# CONFIG_NLS_ISO8859_7 is not set
+# CONFIG_NLS_ISO8859_9 is not set
+# CONFIG_NLS_ISO8859_13 is not set
+# CONFIG_NLS_ISO8859_14 is not set
+# CONFIG_NLS_ISO8859_15 is not set
+# CONFIG_NLS_KOI8_R is not set
+# CONFIG_NLS_KOI8_U is not set
+# CONFIG_NLS_UTF8 is not set
+
+#
+# Instrumentation Support
+#
+# CONFIG_PROFILING is not set
+# CONFIG_KPROBES is not set
+
+#
+# Kernel hacking
+#
+CONFIG_TRACE_IRQFLAGS_SUPPORT=y
+# CONFIG_PRINTK_TIME is not set
+CONFIG_MAGIC_SYSRQ=y
+CONFIG_UNUSED_SYMBOLS=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_LOG_BUF_SHIFT=14
+CONFIG_DETECT_SOFTLOCKUP=y
+# CONFIG_SCHEDSTATS is not set
+# CONFIG_DEBUG_SLAB is not set
+# CONFIG_DEBUG_RT_MUTEXES is not set
+# CONFIG_RT_MUTEX_TESTER is not set
+# CONFIG_DEBUG_SPINLOCK is not set
+# CONFIG_DEBUG_MUTEXES is not set
+# CONFIG_DEBUG_RWSEMS is not set
+# CONFIG_DEBUG_LOCK_ALLOC is not set
+# CONFIG_PROVE_LOCKING is not set
+# CONFIG_DEBUG_SPINLOCK_SLEEP is not set
+# CONFIG_DEBUG_LOCKING_API_SELFTESTS is not set
+# CONFIG_DEBUG_KOBJECT is not set
+# CONFIG_DEBUG_HIGHMEM is not set
+CONFIG_DEBUG_BUGVERBOSE=y
+CONFIG_DEBUG_INFO=y
+# CONFIG_DEBUG_FS is not set
+# CONFIG_DEBUG_VM is not set
+CONFIG_FRAME_POINTER=y
+# CONFIG_UNWIND_INFO is not set
+CONFIG_FORCED_INLINING=y
+# CONFIG_RCU_TORTURE_TEST is not set
+CONFIG_EARLY_PRINTK=y
+# CONFIG_DEBUG_STACKOVERFLOW is not set
+# CONFIG_DEBUG_STACK_USAGE is not set
+# CONFIG_DEBUG_PAGEALLOC is not set
+# CONFIG_DEBUG_RODATA is not set
+# CONFIG_4KSTACKS is not set
+CONFIG_X86_FIND_SMP_CONFIG=y
+CONFIG_X86_MPPARSE=y
+
+#
+# Security options
+#
+# CONFIG_KEYS is not set
+# CONFIG_SECURITY is not set
+
+#
+# Cryptographic options
+#
+CONFIG_CRYPTO=y
+CONFIG_CRYPTO_HMAC=y
+# CONFIG_CRYPTO_NULL is not set
+# CONFIG_CRYPTO_MD4 is not set
+CONFIG_CRYPTO_MD5=m
+CONFIG_CRYPTO_SHA1=m
+# CONFIG_CRYPTO_SHA256 is not set
+# CONFIG_CRYPTO_SHA512 is not set
+# CONFIG_CRYPTO_WP512 is not set
+# CONFIG_CRYPTO_TGR192 is not set
+CONFIG_CRYPTO_DES=m
+# CONFIG_CRYPTO_BLOWFISH is not set
+# CONFIG_CRYPTO_TWOFISH is not set
+# CONFIG_CRYPTO_SERPENT is not set
+# CONFIG_CRYPTO_AES is not set
+# CONFIG_CRYPTO_AES_586 is not set
+# CONFIG_CRYPTO_CAST5 is not set
+# CONFIG_CRYPTO_CAST6 is not set
+# CONFIG_CRYPTO_TEA is not set
+# CONFIG_CRYPTO_ARC4 is not set
+# CONFIG_CRYPTO_KHAZAD is not set
+# CONFIG_CRYPTO_ANUBIS is not set
+# CONFIG_CRYPTO_DEFLATE is not set
+# CONFIG_CRYPTO_MICHAEL_MIC is not set
+CONFIG_CRYPTO_CRC32C=m
+# CONFIG_CRYPTO_TEST is not set
+
+#
+# Hardware crypto devices
+#
+# CONFIG_CRYPTO_DEV_PADLOCK is not set
+CONFIG_XEN=y
+CONFIG_XEN_INTERFACE_VERSION=0x00030207
+
+#
+# XEN
+#
+CONFIG_XEN_PRIVILEGED_GUEST=y
+# CONFIG_XEN_UNPRIVILEGED_GUEST is not set
+CONFIG_XEN_PRIVCMD=y
+CONFIG_XEN_XENBUS_DEV=y
+CONFIG_XEN_BACKEND=y
+CONFIG_XEN_BLKDEV_BACKEND=y
+CONFIG_XEN_BLKDEV_TAP=y
+CONFIG_XEN_NETDEV_BACKEND=y
+# CONFIG_XEN_NETDEV_PIPELINED_TRANSMITTER is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_UTIL is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_BACKEND is not set
+# CONFIG_XEN_NETDEV_LOOPBACK is not set
+CONFIG_XEN_PCIDEV_BACKEND=y
+# CONFIG_XEN_PCIDEV_BACKEND_VPCI is not set
+CONFIG_XEN_PCIDEV_BACKEND_PASS=y
+# CONFIG_XEN_PCIDEV_BACKEND_SLOT is not set
+# CONFIG_XEN_PCIDEV_BACKEND_CONTROLLER is not set
+# CONFIG_XEN_PCIDEV_BE_DEBUG is not set
+CONFIG_XEN_TPMDEV_BACKEND=m
+CONFIG_XEN_SCSI_BACKEND=m
+CONFIG_XEN_BLKDEV_FRONTEND=y
+CONFIG_XEN_NETDEV_FRONTEND=y
+CONFIG_XEN_SCSI_FRONTEND=m
+CONFIG_XEN_GRANT_DEV=y
+# CONFIG_XEN_NETDEV_ACCEL_SFC_FRONTEND is not set
+CONFIG_XEN_SCRUB_PAGES=y
+CONFIG_XEN_DISABLE_SERIAL=y
+CONFIG_XEN_SYSFS=y
+CONFIG_XEN_COMPAT_030002_AND_LATER=y
+# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
+# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
+CONFIG_XEN_COMPAT=0x030002
+CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
+CONFIG_NO_IDLE_HZ=y
+CONFIG_XEN_SMPBOOT=y
+CONFIG_XEN_BALLOON=y
+CONFIG_XEN_DEVMEM=y
+
+#
+# Library routines
+#
+# CONFIG_CRC_CCITT is not set
+# CONFIG_CRC16 is not set
+CONFIG_CRC32=y
+CONFIG_LIBCRC32C=y
+CONFIG_ZLIB_INFLATE=y
+CONFIG_PLIST=y
+CONFIG_GENERIC_HARDIRQS=y
+CONFIG_GENERIC_IRQ_PROBE=y
+CONFIG_GENERIC_PENDING_IRQ=y
+CONFIG_X86_SMP=y
+CONFIG_X86_BIOS_REBOOT=y
+CONFIG_X86_TRAMPOLINE=y
+CONFIG_X86_NO_TSS=y
+CONFIG_X86_NO_IDT=y
+CONFIG_KTIME_SCALAR=y
diff -rpuN linux-2.6.18.8/buildconfigs/linux-defconfig_xen0_x86_64 linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xen0_x86_64
--- linux-2.6.18.8/buildconfigs/linux-defconfig_xen0_x86_64	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xen0_x86_64	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1390 @@
+#
+# Automatically generated make config: don't edit
+# Linux kernel version: 2.6.18.8
+# Mon Feb 18 10:41:04 2008
+#
+CONFIG_X86_64=y
+CONFIG_64BIT=y
+CONFIG_X86=y
+CONFIG_LOCKDEP_SUPPORT=y
+CONFIG_STACKTRACE_SUPPORT=y
+CONFIG_SEMAPHORE_SLEEPERS=y
+CONFIG_MMU=y
+CONFIG_RWSEM_GENERIC_SPINLOCK=y
+CONFIG_GENERIC_HWEIGHT=y
+CONFIG_GENERIC_CALIBRATE_DELAY=y
+CONFIG_X86_CMPXCHG=y
+CONFIG_EARLY_PRINTK=y
+CONFIG_GENERIC_ISA_DMA=y
+CONFIG_GENERIC_IOMAP=y
+CONFIG_ARCH_MAY_HAVE_PC_FDC=y
+CONFIG_DMI=y
+CONFIG_AUDIT_ARCH=y
+CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
+
+#
+# Code maturity level options
+#
+CONFIG_EXPERIMENTAL=y
+CONFIG_LOCK_KERNEL=y
+CONFIG_INIT_ENV_ARG_LIMIT=32
+
+#
+# General setup
+#
+CONFIG_LOCALVERSION=""
+# CONFIG_LOCALVERSION_AUTO is not set
+CONFIG_SWAP=y
+CONFIG_SYSVIPC=y
+# CONFIG_POSIX_MQUEUE is not set
+# CONFIG_BSD_PROCESS_ACCT is not set
+# CONFIG_TASKSTATS is not set
+# CONFIG_AUDIT is not set
+# CONFIG_IKCONFIG is not set
+# CONFIG_CPUSETS is not set
+# CONFIG_RELAY is not set
+CONFIG_INITRAMFS_SOURCE=""
+# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set
+# CONFIG_EMBEDDED is not set
+CONFIG_UID16=y
+CONFIG_SYSCTL=y
+CONFIG_KALLSYMS=y
+# CONFIG_KALLSYMS_ALL is not set
+# CONFIG_KALLSYMS_EXTRA_PASS is not set
+CONFIG_HOTPLUG=y
+CONFIG_PRINTK=y
+CONFIG_BUG=y
+CONFIG_ELF_CORE=y
+CONFIG_BASE_FULL=y
+CONFIG_FUTEX=y
+CONFIG_EPOLL=y
+CONFIG_SHMEM=y
+CONFIG_SLAB=y
+CONFIG_VM_EVENT_COUNTERS=y
+CONFIG_RT_MUTEXES=y
+# CONFIG_TINY_SHMEM is not set
+CONFIG_BASE_SMALL=0
+# CONFIG_SLOB is not set
+
+#
+# Loadable module support
+#
+CONFIG_MODULES=y
+CONFIG_MODULE_UNLOAD=y
+# CONFIG_MODULE_FORCE_UNLOAD is not set
+# CONFIG_MODVERSIONS is not set
+# CONFIG_MODULE_SRCVERSION_ALL is not set
+CONFIG_KMOD=y
+CONFIG_STOP_MACHINE=y
+
+#
+# Block layer
+#
+# CONFIG_LBD is not set
+# CONFIG_BLK_DEV_IO_TRACE is not set
+CONFIG_LSF=y
+
+#
+# IO Schedulers
+#
+CONFIG_IOSCHED_NOOP=y
+CONFIG_IOSCHED_AS=y
+CONFIG_IOSCHED_DEADLINE=y
+CONFIG_IOSCHED_CFQ=y
+CONFIG_DEFAULT_AS=y
+# CONFIG_DEFAULT_DEADLINE is not set
+# CONFIG_DEFAULT_CFQ is not set
+# CONFIG_DEFAULT_NOOP is not set
+CONFIG_DEFAULT_IOSCHED="anticipatory"
+
+#
+# Processor type and features
+#
+CONFIG_X86_PC=y
+# CONFIG_X86_VSMP is not set
+# CONFIG_MK8 is not set
+# CONFIG_MPSC is not set
+CONFIG_GENERIC_CPU=y
+CONFIG_X86_64_XEN=y
+CONFIG_X86_NO_TSS=y
+CONFIG_X86_NO_IDT=y
+CONFIG_X86_L1_CACHE_BYTES=128
+CONFIG_X86_L1_CACHE_SHIFT=7
+CONFIG_X86_INTERNODE_CACHE_BYTES=128
+CONFIG_X86_GOOD_APIC=y
+CONFIG_MICROCODE=y
+CONFIG_X86_MSR=y
+CONFIG_X86_CPUID=y
+CONFIG_X86_IO_APIC=y
+CONFIG_X86_XEN_GENAPIC=y
+CONFIG_X86_LOCAL_APIC=y
+CONFIG_MTRR=y
+CONFIG_SMP=y
+CONFIG_PREEMPT_NONE=y
+# CONFIG_PREEMPT_VOLUNTARY is not set
+# CONFIG_PREEMPT is not set
+CONFIG_PREEMPT_BKL=y
+CONFIG_ARCH_FLATMEM_ENABLE=y
+CONFIG_SELECT_MEMORY_MODEL=y
+CONFIG_FLATMEM_MANUAL=y
+# CONFIG_DISCONTIGMEM_MANUAL is not set
+# CONFIG_SPARSEMEM_MANUAL is not set
+CONFIG_FLATMEM=y
+CONFIG_FLAT_NODE_MEM_MAP=y
+# CONFIG_SPARSEMEM_STATIC is not set
+CONFIG_SPLIT_PTLOCK_CPUS=4
+CONFIG_RESOURCES_64BIT=y
+CONFIG_NR_CPUS=8
+CONFIG_HOTPLUG_CPU=y
+CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y
+CONFIG_SWIOTLB=y
+CONFIG_KEXEC=y
+# CONFIG_CRASH_DUMP is not set
+CONFIG_PHYSICAL_START=0x200000
+CONFIG_SECCOMP=y
+CONFIG_HZ_100=y
+# CONFIG_HZ_250 is not set
+# CONFIG_HZ_1000 is not set
+CONFIG_HZ=100
+# CONFIG_REORDER is not set
+CONFIG_GENERIC_HARDIRQS=y
+CONFIG_GENERIC_IRQ_PROBE=y
+CONFIG_ISA_DMA_API=y
+CONFIG_GENERIC_PENDING_IRQ=y
+
+#
+# Power management options
+#
+CONFIG_PM=y
+CONFIG_PM_LEGACY=y
+# CONFIG_PM_DEBUG is not set
+# CONFIG_SOFTWARE_SUSPEND is not set
+CONFIG_SUSPEND_SMP=y
+
+#
+# ACPI (Advanced Configuration and Power Interface) Support
+#
+CONFIG_ACPI=y
+CONFIG_ACPI_SLEEP=y
+CONFIG_ACPI_SLEEP_PROC_FS=y
+# CONFIG_ACPI_SLEEP_PROC_SLEEP is not set
+CONFIG_ACPI_AC=m
+CONFIG_ACPI_BATTERY=m
+CONFIG_ACPI_BUTTON=m
+CONFIG_ACPI_VIDEO=m
+CONFIG_ACPI_HOTKEY=m
+CONFIG_ACPI_FAN=m
+CONFIG_ACPI_DOCK=m
+CONFIG_ACPI_PROCESSOR=m
+CONFIG_ACPI_HOTPLUG_CPU=y
+CONFIG_ACPI_THERMAL=m
+CONFIG_ACPI_ASUS=m
+CONFIG_ACPI_IBM=m
+CONFIG_ACPI_TOSHIBA=m
+# CONFIG_ACPI_CUSTOM_DSDT is not set
+CONFIG_ACPI_BLACKLIST_YEAR=0
+# CONFIG_ACPI_DEBUG is not set
+CONFIG_ACPI_EC=y
+CONFIG_ACPI_POWER=y
+CONFIG_ACPI_SYSTEM=y
+CONFIG_ACPI_CONTAINER=m
+CONFIG_ACPI_PV_SLEEP=y
+
+#
+# CPU Frequency scaling
+#
+# CONFIG_CPU_FREQ is not set
+
+#
+# Bus options (PCI etc.)
+#
+CONFIG_PCI=y
+CONFIG_PCI_DIRECT=y
+CONFIG_PCI_MMCONFIG=y
+CONFIG_XEN_PCIDEV_FRONTEND=y
+# CONFIG_XEN_PCIDEV_FE_DEBUG is not set
+# CONFIG_PCIEPORTBUS is not set
+# CONFIG_PCI_MSI is not set
+# CONFIG_PCI_DEBUG is not set
+
+#
+# PCCARD (PCMCIA/CardBus) support
+#
+# CONFIG_PCCARD is not set
+
+#
+# PCI Hotplug Support
+#
+# CONFIG_HOTPLUG_PCI is not set
+
+#
+# Executable file formats / Emulations
+#
+CONFIG_BINFMT_ELF=y
+# CONFIG_BINFMT_MISC is not set
+CONFIG_IA32_EMULATION=y
+CONFIG_IA32_AOUT=y
+CONFIG_COMPAT=y
+CONFIG_SYSVIPC_COMPAT=y
+
+#
+# Networking
+#
+CONFIG_NET=y
+
+#
+# Networking options
+#
+# CONFIG_NETDEBUG is not set
+CONFIG_PACKET=y
+# CONFIG_PACKET_MMAP is not set
+CONFIG_UNIX=y
+CONFIG_XFRM=y
+# CONFIG_XFRM_USER is not set
+# CONFIG_NET_KEY is not set
+CONFIG_INET=y
+# CONFIG_IP_MULTICAST is not set
+# CONFIG_IP_ADVANCED_ROUTER is not set
+CONFIG_IP_FIB_HASH=y
+CONFIG_IP_PNP=y
+CONFIG_IP_PNP_DHCP=y
+# CONFIG_IP_PNP_BOOTP is not set
+# CONFIG_IP_PNP_RARP is not set
+# CONFIG_NET_IPIP is not set
+# CONFIG_NET_IPGRE is not set
+# CONFIG_ARPD is not set
+# CONFIG_SYN_COOKIES is not set
+# CONFIG_INET_AH is not set
+# CONFIG_INET_ESP is not set
+# CONFIG_INET_IPCOMP is not set
+# CONFIG_INET_XFRM_TUNNEL is not set
+# CONFIG_INET_TUNNEL is not set
+CONFIG_INET_XFRM_MODE_TRANSPORT=y
+CONFIG_INET_XFRM_MODE_TUNNEL=y
+# CONFIG_INET_DIAG is not set
+# CONFIG_TCP_CONG_ADVANCED is not set
+CONFIG_TCP_CONG_BIC=y
+
+#
+# IP: Virtual Server Configuration
+#
+# CONFIG_IP_VS is not set
+# CONFIG_IPV6 is not set
+# CONFIG_INET6_XFRM_TUNNEL is not set
+# CONFIG_INET6_TUNNEL is not set
+# CONFIG_NETWORK_SECMARK is not set
+CONFIG_NETFILTER=y
+# CONFIG_NETFILTER_DEBUG is not set
+CONFIG_BRIDGE_NETFILTER=y
+
+#
+# Core Netfilter Configuration
+#
+# CONFIG_NETFILTER_NETLINK is not set
+# CONFIG_NETFILTER_XTABLES is not set
+
+#
+# IP: Netfilter Configuration
+#
+CONFIG_IP_NF_CONNTRACK=m
+CONFIG_IP_NF_CT_ACCT=y
+# CONFIG_IP_NF_CONNTRACK_MARK is not set
+# CONFIG_IP_NF_CONNTRACK_EVENTS is not set
+# CONFIG_IP_NF_CT_PROTO_SCTP is not set
+CONFIG_IP_NF_FTP=m
+# CONFIG_IP_NF_IRC is not set
+# CONFIG_IP_NF_NETBIOS_NS is not set
+# CONFIG_IP_NF_TFTP is not set
+# CONFIG_IP_NF_AMANDA is not set
+# CONFIG_IP_NF_PPTP is not set
+# CONFIG_IP_NF_H323 is not set
+# CONFIG_IP_NF_SIP is not set
+# CONFIG_IP_NF_QUEUE is not set
+
+#
+# Bridge: Netfilter Configuration
+#
+# CONFIG_BRIDGE_NF_EBTABLES is not set
+
+#
+# DCCP Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP_DCCP is not set
+
+#
+# SCTP Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP_SCTP is not set
+
+#
+# TIPC Configuration (EXPERIMENTAL)
+#
+# CONFIG_TIPC is not set
+# CONFIG_ATM is not set
+CONFIG_BRIDGE=y
+# CONFIG_VLAN_8021Q is not set
+# CONFIG_DECNET is not set
+CONFIG_LLC=y
+# CONFIG_LLC2 is not set
+# CONFIG_IPX is not set
+# CONFIG_ATALK is not set
+# CONFIG_X25 is not set
+# CONFIG_LAPB is not set
+# CONFIG_ECONET is not set
+# CONFIG_WAN_ROUTER is not set
+
+#
+# QoS and/or fair queueing
+#
+# CONFIG_NET_SCHED is not set
+
+#
+# Network testing
+#
+# CONFIG_NET_PKTGEN is not set
+# CONFIG_HAMRADIO is not set
+# CONFIG_IRDA is not set
+# CONFIG_BT is not set
+# CONFIG_IEEE80211 is not set
+
+#
+# Device Drivers
+#
+
+#
+# Generic Driver Options
+#
+# CONFIG_STANDALONE is not set
+CONFIG_PREVENT_FIRMWARE_BUILD=y
+# CONFIG_FW_LOADER is not set
+# CONFIG_DEBUG_DRIVER is not set
+CONFIG_SYS_HYPERVISOR=y
+
+#
+# Connector - unified userspace <-> kernelspace linker
+#
+# CONFIG_CONNECTOR is not set
+
+#
+# Memory Technology Devices (MTD)
+#
+# CONFIG_MTD is not set
+
+#
+# Parallel port support
+#
+# CONFIG_PARPORT is not set
+
+#
+# Plug and Play support
+#
+CONFIG_PNP=y
+CONFIG_PNP_DEBUG=y
+
+#
+# Protocols
+#
+CONFIG_PNPACPI=y
+
+#
+# Block devices
+#
+CONFIG_BLK_DEV_FD=y
+# CONFIG_BLK_CPQ_DA is not set
+CONFIG_BLK_CPQ_CISS_DA=y
+# CONFIG_CISS_SCSI_TAPE is not set
+# CONFIG_BLK_DEV_DAC960 is not set
+# CONFIG_BLK_DEV_UMEM is not set
+# CONFIG_BLK_DEV_COW_COMMON is not set
+CONFIG_BLK_DEV_LOOP=y
+# CONFIG_BLK_DEV_CRYPTOLOOP is not set
+# CONFIG_BLK_DEV_NBD is not set
+# CONFIG_BLK_DEV_SX8 is not set
+# CONFIG_BLK_DEV_UB is not set
+CONFIG_BLK_DEV_RAM=y
+CONFIG_BLK_DEV_RAM_COUNT=16
+CONFIG_BLK_DEV_RAM_SIZE=4096
+CONFIG_BLK_DEV_RAM_BLOCKSIZE=1024
+CONFIG_BLK_DEV_INITRD=y
+# CONFIG_CDROM_PKTCDVD is not set
+# CONFIG_ATA_OVER_ETH is not set
+
+#
+# ATA/ATAPI/MFM/RLL support
+#
+CONFIG_IDE=y
+CONFIG_BLK_DEV_IDE=y
+
+#
+# Please see Documentation/ide.txt for help/info on IDE drives
+#
+# CONFIG_BLK_DEV_IDE_SATA is not set
+# CONFIG_BLK_DEV_HD_IDE is not set
+CONFIG_BLK_DEV_IDEDISK=y
+# CONFIG_IDEDISK_MULTI_MODE is not set
+CONFIG_BLK_DEV_IDECD=y
+# CONFIG_BLK_DEV_IDETAPE is not set
+# CONFIG_BLK_DEV_IDEFLOPPY is not set
+# CONFIG_BLK_DEV_IDESCSI is not set
+# CONFIG_IDE_TASK_IOCTL is not set
+
+#
+# IDE chipset support/bugfixes
+#
+CONFIG_IDE_GENERIC=y
+# CONFIG_BLK_DEV_CMD640 is not set
+# CONFIG_BLK_DEV_IDEPNP is not set
+CONFIG_BLK_DEV_IDEPCI=y
+# CONFIG_IDEPCI_SHARE_IRQ is not set
+# CONFIG_BLK_DEV_OFFBOARD is not set
+CONFIG_BLK_DEV_GENERIC=y
+# CONFIG_BLK_DEV_OPTI621 is not set
+# CONFIG_BLK_DEV_RZ1000 is not set
+CONFIG_BLK_DEV_IDEDMA_PCI=y
+# CONFIG_BLK_DEV_IDEDMA_FORCED is not set
+CONFIG_IDEDMA_PCI_AUTO=y
+# CONFIG_IDEDMA_ONLYDISK is not set
+# CONFIG_BLK_DEV_AEC62XX is not set
+# CONFIG_BLK_DEV_ALI15X3 is not set
+# CONFIG_BLK_DEV_AMD74XX is not set
+# CONFIG_BLK_DEV_ATIIXP is not set
+# CONFIG_BLK_DEV_CMD64X is not set
+# CONFIG_BLK_DEV_TRIFLEX is not set
+# CONFIG_BLK_DEV_CY82C693 is not set
+# CONFIG_BLK_DEV_CS5520 is not set
+# CONFIG_BLK_DEV_CS5530 is not set
+# CONFIG_BLK_DEV_HPT34X is not set
+# CONFIG_BLK_DEV_HPT366 is not set
+# CONFIG_BLK_DEV_SC1200 is not set
+CONFIG_BLK_DEV_PIIX=y
+# CONFIG_BLK_DEV_IT821X is not set
+# CONFIG_BLK_DEV_NS87415 is not set
+# CONFIG_BLK_DEV_PDC202XX_OLD is not set
+# CONFIG_BLK_DEV_PDC202XX_NEW is not set
+CONFIG_BLK_DEV_SVWKS=y
+# CONFIG_BLK_DEV_SIIMAGE is not set
+# CONFIG_BLK_DEV_SIS5513 is not set
+# CONFIG_BLK_DEV_SLC90E66 is not set
+# CONFIG_BLK_DEV_TRM290 is not set
+# CONFIG_BLK_DEV_VIA82CXXX is not set
+# CONFIG_IDE_ARM is not set
+CONFIG_BLK_DEV_IDEDMA=y
+# CONFIG_IDEDMA_IVB is not set
+CONFIG_IDEDMA_AUTO=y
+# CONFIG_BLK_DEV_HD is not set
+
+#
+# SCSI device support
+#
+# CONFIG_RAID_ATTRS is not set
+CONFIG_SCSI=y
+CONFIG_SCSI_PROC_FS=y
+
+#
+# SCSI support type (disk, tape, CD-ROM)
+#
+CONFIG_BLK_DEV_SD=y
+# CONFIG_CHR_DEV_ST is not set
+# CONFIG_CHR_DEV_OSST is not set
+# CONFIG_BLK_DEV_SR is not set
+# CONFIG_CHR_DEV_SG is not set
+# CONFIG_CHR_DEV_SCH is not set
+
+#
+# Some SCSI devices (e.g. CD jukebox) support multiple LUNs
+#
+# CONFIG_SCSI_MULTI_LUN is not set
+# CONFIG_SCSI_CONSTANTS is not set
+# CONFIG_SCSI_LOGGING is not set
+
+#
+# SCSI Transport Attributes
+#
+CONFIG_SCSI_SPI_ATTRS=y
+# CONFIG_SCSI_FC_ATTRS is not set
+# CONFIG_SCSI_ISCSI_ATTRS is not set
+# CONFIG_SCSI_SAS_ATTRS is not set
+
+#
+# SCSI low-level drivers
+#
+# CONFIG_ISCSI_TCP is not set
+CONFIG_BLK_DEV_3W_XXXX_RAID=y
+# CONFIG_SCSI_3W_9XXX is not set
+# CONFIG_SCSI_ACARD is not set
+CONFIG_SCSI_AACRAID=y
+CONFIG_SCSI_AIC7XXX=y
+CONFIG_AIC7XXX_CMDS_PER_DEVICE=32
+CONFIG_AIC7XXX_RESET_DELAY_MS=15000
+CONFIG_AIC7XXX_DEBUG_ENABLE=y
+CONFIG_AIC7XXX_DEBUG_MASK=0
+CONFIG_AIC7XXX_REG_PRETTY_PRINT=y
+# CONFIG_SCSI_AIC7XXX_OLD is not set
+CONFIG_SCSI_AIC79XX=y
+CONFIG_AIC79XX_CMDS_PER_DEVICE=32
+CONFIG_AIC79XX_RESET_DELAY_MS=15000
+# CONFIG_AIC79XX_ENABLE_RD_STRM is not set
+CONFIG_AIC79XX_DEBUG_ENABLE=y
+CONFIG_AIC79XX_DEBUG_MASK=0
+CONFIG_AIC79XX_REG_PRETTY_PRINT=y
+CONFIG_MEGARAID_NEWGEN=y
+# CONFIG_MEGARAID_MM is not set
+# CONFIG_MEGARAID_LEGACY is not set
+# CONFIG_MEGARAID_SAS is not set
+CONFIG_SCSI_SATA=y
+CONFIG_SCSI_SATA_AHCI=y
+# CONFIG_SCSI_SATA_SVW is not set
+CONFIG_SCSI_ATA_PIIX=y
+# CONFIG_SCSI_SATA_MV is not set
+# CONFIG_SCSI_SATA_NV is not set
+# CONFIG_SCSI_PDC_ADMA is not set
+# CONFIG_SCSI_HPTIOP is not set
+# CONFIG_SCSI_SATA_QSTOR is not set
+CONFIG_SCSI_SATA_PROMISE=y
+CONFIG_SCSI_SATA_SX4=y
+CONFIG_SCSI_SATA_SIL=y
+CONFIG_SCSI_SATA_SIL24=y
+# CONFIG_SCSI_SATA_SIS is not set
+# CONFIG_SCSI_SATA_ULI is not set
+# CONFIG_SCSI_SATA_VIA is not set
+# CONFIG_SCSI_SATA_VITESSE is not set
+CONFIG_SCSI_SATA_INTEL_COMBINED=y
+# CONFIG_SCSI_BUSLOGIC is not set
+# CONFIG_SCSI_DMX3191D is not set
+# CONFIG_SCSI_EATA is not set
+# CONFIG_SCSI_FUTURE_DOMAIN is not set
+# CONFIG_SCSI_GDTH is not set
+# CONFIG_SCSI_IPS is not set
+# CONFIG_SCSI_INITIO is not set
+# CONFIG_SCSI_INIA100 is not set
+# CONFIG_SCSI_SYM53C8XX_2 is not set
+# CONFIG_SCSI_IPR is not set
+# CONFIG_SCSI_QLOGIC_1280 is not set
+# CONFIG_SCSI_QLA_FC is not set
+# CONFIG_SCSI_LPFC is not set
+# CONFIG_SCSI_DC395x is not set
+# CONFIG_SCSI_DC390T is not set
+# CONFIG_SCSI_DEBUG is not set
+
+#
+# Multi-device support (RAID and LVM)
+#
+CONFIG_MD=y
+CONFIG_BLK_DEV_MD=y
+# CONFIG_MD_LINEAR is not set
+CONFIG_MD_RAID0=y
+CONFIG_MD_RAID1=y
+# CONFIG_MD_RAID10 is not set
+# CONFIG_MD_RAID456 is not set
+# CONFIG_MD_MULTIPATH is not set
+# CONFIG_MD_FAULTY is not set
+CONFIG_BLK_DEV_DM=y
+# CONFIG_DM_CRYPT is not set
+CONFIG_DM_SNAPSHOT=y
+CONFIG_DM_MIRROR=y
+# CONFIG_DM_ZERO is not set
+# CONFIG_DM_MULTIPATH is not set
+
+#
+# Fusion MPT device support
+#
+CONFIG_FUSION=y
+CONFIG_FUSION_SPI=y
+# CONFIG_FUSION_FC is not set
+# CONFIG_FUSION_SAS is not set
+CONFIG_FUSION_MAX_SGE=128
+# CONFIG_FUSION_CTL is not set
+
+#
+# IEEE 1394 (FireWire) support
+#
+# CONFIG_IEEE1394 is not set
+
+#
+# I2O device support
+#
+# CONFIG_I2O is not set
+
+#
+# Network device support
+#
+CONFIG_NETDEVICES=y
+# CONFIG_DUMMY is not set
+# CONFIG_BONDING is not set
+# CONFIG_EQUALIZER is not set
+CONFIG_TUN=y
+# CONFIG_NET_SB1000 is not set
+
+#
+# ARCnet devices
+#
+# CONFIG_ARCNET is not set
+
+#
+# PHY device support
+#
+# CONFIG_PHYLIB is not set
+
+#
+# Ethernet (10 or 100Mbit)
+#
+CONFIG_NET_ETHERNET=y
+CONFIG_MII=y
+# CONFIG_HAPPYMEAL is not set
+# CONFIG_SUNGEM is not set
+# CONFIG_CASSINI is not set
+CONFIG_NET_VENDOR_3COM=y
+CONFIG_VORTEX=y
+# CONFIG_TYPHOON is not set
+
+#
+# Tulip family network device support
+#
+CONFIG_NET_TULIP=y
+# CONFIG_DE2104X is not set
+CONFIG_TULIP=y
+# CONFIG_TULIP_MWI is not set
+# CONFIG_TULIP_MMIO is not set
+# CONFIG_TULIP_NAPI is not set
+# CONFIG_DE4X5 is not set
+# CONFIG_WINBOND_840 is not set
+# CONFIG_DM9102 is not set
+# CONFIG_ULI526X is not set
+# CONFIG_HP100 is not set
+CONFIG_NET_PCI=y
+CONFIG_PCNET32=y
+# CONFIG_AMD8111_ETH is not set
+# CONFIG_ADAPTEC_STARFIRE is not set
+# CONFIG_B44 is not set
+# CONFIG_FORCEDETH is not set
+# CONFIG_DGRS is not set
+# CONFIG_EEPRO100 is not set
+CONFIG_E100=y
+# CONFIG_FEALNX is not set
+# CONFIG_NATSEMI is not set
+CONFIG_NE2K_PCI=y
+# CONFIG_8139CP is not set
+CONFIG_8139TOO=y
+CONFIG_8139TOO_PIO=y
+# CONFIG_8139TOO_TUNE_TWISTER is not set
+# CONFIG_8139TOO_8129 is not set
+# CONFIG_8139_OLD_RX_RESET is not set
+# CONFIG_SIS900 is not set
+# CONFIG_EPIC100 is not set
+# CONFIG_SUNDANCE is not set
+CONFIG_VIA_RHINE=y
+# CONFIG_VIA_RHINE_MMIO is not set
+# CONFIG_VIA_RHINE_NAPI is not set
+
+#
+# Ethernet (1000 Mbit)
+#
+CONFIG_ACENIC=y
+# CONFIG_ACENIC_OMIT_TIGON_I is not set
+# CONFIG_DL2K is not set
+CONFIG_E1000=y
+# CONFIG_E1000_NAPI is not set
+# CONFIG_E1000_DISABLE_PACKET_SPLIT is not set
+# CONFIG_NS83820 is not set
+# CONFIG_HAMACHI is not set
+# CONFIG_YELLOWFIN is not set
+# CONFIG_R8169 is not set
+# CONFIG_SIS190 is not set
+# CONFIG_SKGE is not set
+# CONFIG_SKY2 is not set
+CONFIG_SK98LIN=y
+# CONFIG_VIA_VELOCITY is not set
+CONFIG_TIGON3=y
+CONFIG_BNX2=y
+
+#
+# Ethernet (10000 Mbit)
+#
+# CONFIG_CHELSIO_T1 is not set
+# CONFIG_IXGB is not set
+# CONFIG_S2IO is not set
+# CONFIG_MYRI10GE is not set
+# CONFIG_SFC is not set
+
+#
+# Token Ring devices
+#
+# CONFIG_TR is not set
+
+#
+# Wireless LAN (non-hamradio)
+#
+# CONFIG_NET_RADIO is not set
+
+#
+# Wan interfaces
+#
+# CONFIG_WAN is not set
+# CONFIG_FDDI is not set
+# CONFIG_HIPPI is not set
+# CONFIG_PPP is not set
+# CONFIG_SLIP is not set
+# CONFIG_NET_FC is not set
+# CONFIG_SHAPER is not set
+# CONFIG_NETCONSOLE is not set
+# CONFIG_NETPOLL is not set
+# CONFIG_NET_POLL_CONTROLLER is not set
+
+#
+# ISDN subsystem
+#
+# CONFIG_ISDN is not set
+
+#
+# Telephony Support
+#
+# CONFIG_PHONE is not set
+
+#
+# Input device support
+#
+CONFIG_INPUT=y
+
+#
+# Userland interfaces
+#
+CONFIG_INPUT_MOUSEDEV=y
+CONFIG_INPUT_MOUSEDEV_PSAUX=y
+CONFIG_INPUT_MOUSEDEV_SCREEN_X=1024
+CONFIG_INPUT_MOUSEDEV_SCREEN_Y=768
+# CONFIG_INPUT_JOYDEV is not set
+# CONFIG_INPUT_TSDEV is not set
+# CONFIG_INPUT_EVDEV is not set
+# CONFIG_INPUT_EVBUG is not set
+
+#
+# Input Device Drivers
+#
+CONFIG_INPUT_KEYBOARD=y
+CONFIG_KEYBOARD_ATKBD=y
+# CONFIG_KEYBOARD_SUNKBD is not set
+# CONFIG_KEYBOARD_LKKBD is not set
+# CONFIG_KEYBOARD_XTKBD is not set
+# CONFIG_KEYBOARD_NEWTON is not set
+CONFIG_INPUT_MOUSE=y
+CONFIG_MOUSE_PS2=y
+# CONFIG_MOUSE_SERIAL is not set
+# CONFIG_MOUSE_VSXXXAA is not set
+# CONFIG_INPUT_JOYSTICK is not set
+# CONFIG_INPUT_TOUCHSCREEN is not set
+# CONFIG_INPUT_MISC is not set
+
+#
+# Hardware I/O ports
+#
+CONFIG_SERIO=y
+CONFIG_SERIO_I8042=y
+CONFIG_SERIO_SERPORT=y
+# CONFIG_SERIO_CT82C710 is not set
+# CONFIG_SERIO_PCIPS2 is not set
+CONFIG_SERIO_LIBPS2=y
+# CONFIG_SERIO_RAW is not set
+# CONFIG_GAMEPORT is not set
+
+#
+# Character devices
+#
+CONFIG_VT=y
+CONFIG_VT_CONSOLE=y
+CONFIG_HW_CONSOLE=y
+CONFIG_VT_HW_CONSOLE_BINDING=y
+# CONFIG_SERIAL_NONSTANDARD is not set
+
+#
+# Serial drivers
+#
+
+#
+# Non-8250 serial port support
+#
+# CONFIG_SERIAL_JSM is not set
+CONFIG_UNIX98_PTYS=y
+CONFIG_LEGACY_PTYS=y
+CONFIG_LEGACY_PTY_COUNT=256
+
+#
+# IPMI
+#
+# CONFIG_IPMI_HANDLER is not set
+
+#
+# Watchdog Cards
+#
+# CONFIG_WATCHDOG is not set
+# CONFIG_HW_RANDOM is not set
+# CONFIG_NVRAM is not set
+# CONFIG_RTC is not set
+# CONFIG_GEN_RTC is not set
+# CONFIG_DTLK is not set
+# CONFIG_R3964 is not set
+# CONFIG_APPLICOM is not set
+
+#
+# Ftape, the floppy tape device driver
+#
+CONFIG_AGP=m
+# CONFIG_AGP_AMD64 is not set
+CONFIG_AGP_INTEL=m
+CONFIG_AGP_SIS=m
+CONFIG_AGP_VIA=m
+CONFIG_DRM=m
+CONFIG_DRM_TDFX=m
+CONFIG_DRM_R128=m
+CONFIG_DRM_RADEON=m
+CONFIG_DRM_I810=m
+CONFIG_DRM_I830=m
+CONFIG_DRM_I915=m
+CONFIG_DRM_MGA=m
+CONFIG_DRM_SIS=m
+# CONFIG_DRM_VIA is not set
+# CONFIG_DRM_SAVAGE is not set
+# CONFIG_MWAVE is not set
+# CONFIG_PC8736x_GPIO is not set
+# CONFIG_RAW_DRIVER is not set
+# CONFIG_HPET is not set
+# CONFIG_HANGCHECK_TIMER is not set
+
+#
+# TPM devices
+#
+# CONFIG_TCG_TPM is not set
+# CONFIG_TELCLOCK is not set
+
+#
+# I2C support
+#
+# CONFIG_I2C is not set
+
+#
+# SPI support
+#
+# CONFIG_SPI is not set
+# CONFIG_SPI_MASTER is not set
+
+#
+# Dallas's 1-wire bus
+#
+
+#
+# Hardware Monitoring support
+#
+# CONFIG_HWMON is not set
+# CONFIG_HWMON_VID is not set
+
+#
+# Misc devices
+#
+# CONFIG_IBM_ASM is not set
+
+#
+# Multimedia devices
+#
+# CONFIG_VIDEO_DEV is not set
+
+#
+# Digital Video Broadcasting Devices
+#
+# CONFIG_DVB is not set
+# CONFIG_USB_DABUSB is not set
+
+#
+# Graphics support
+#
+CONFIG_FIRMWARE_EDID=y
+# CONFIG_FB is not set
+
+#
+# Console display driver support
+#
+CONFIG_VGA_CONSOLE=y
+# CONFIG_VGACON_SOFT_SCROLLBACK is not set
+CONFIG_DUMMY_CONSOLE=y
+# CONFIG_BACKLIGHT_LCD_SUPPORT is not set
+
+#
+# Sound
+#
+# CONFIG_SOUND is not set
+
+#
+# USB support
+#
+CONFIG_USB_ARCH_HAS_HCD=y
+CONFIG_USB_ARCH_HAS_OHCI=y
+CONFIG_USB_ARCH_HAS_EHCI=y
+CONFIG_USB=y
+# CONFIG_USB_DEBUG is not set
+
+#
+# Miscellaneous USB options
+#
+# CONFIG_USB_DEVICEFS is not set
+# CONFIG_USB_BANDWIDTH is not set
+# CONFIG_USB_DYNAMIC_MINORS is not set
+# CONFIG_USB_SUSPEND is not set
+# CONFIG_USB_OTG is not set
+
+#
+# USB Host Controller Drivers
+#
+# CONFIG_USB_EHCI_HCD is not set
+# CONFIG_USB_ISP116X_HCD is not set
+CONFIG_USB_OHCI_HCD=y
+# CONFIG_USB_OHCI_BIG_ENDIAN is not set
+CONFIG_USB_OHCI_LITTLE_ENDIAN=y
+CONFIG_USB_UHCI_HCD=y
+# CONFIG_USB_SL811_HCD is not set
+
+#
+# USB Device Class drivers
+#
+# CONFIG_USB_ACM is not set
+# CONFIG_USB_PRINTER is not set
+
+#
+# NOTE: USB_STORAGE enables SCSI, and 'SCSI disk support'
+#
+
+#
+# may also be needed; see USB_STORAGE Help for more information
+#
+# CONFIG_USB_STORAGE is not set
+# CONFIG_USB_LIBUSUAL is not set
+
+#
+# USB Input Devices
+#
+CONFIG_USB_HID=y
+CONFIG_USB_HIDINPUT=y
+# CONFIG_USB_HIDINPUT_POWERBOOK is not set
+# CONFIG_HID_FF is not set
+# CONFIG_USB_HIDDEV is not set
+# CONFIG_USB_AIPTEK is not set
+# CONFIG_USB_WACOM is not set
+# CONFIG_USB_ACECAD is not set
+# CONFIG_USB_KBTAB is not set
+# CONFIG_USB_POWERMATE is not set
+# CONFIG_USB_TOUCHSCREEN is not set
+# CONFIG_USB_YEALINK is not set
+# CONFIG_USB_XPAD is not set
+# CONFIG_USB_ATI_REMOTE is not set
+# CONFIG_USB_ATI_REMOTE2 is not set
+# CONFIG_USB_KEYSPAN_REMOTE is not set
+# CONFIG_USB_APPLETOUCH is not set
+
+#
+# USB Imaging devices
+#
+# CONFIG_USB_MDC800 is not set
+# CONFIG_USB_MICROTEK is not set
+
+#
+# USB Network Adapters
+#
+# CONFIG_USB_CATC is not set
+# CONFIG_USB_KAWETH is not set
+# CONFIG_USB_PEGASUS is not set
+# CONFIG_USB_RTL8150 is not set
+# CONFIG_USB_USBNET is not set
+CONFIG_USB_MON=y
+
+#
+# USB port drivers
+#
+
+#
+# USB Serial Converter support
+#
+# CONFIG_USB_SERIAL is not set
+
+#
+# USB Miscellaneous drivers
+#
+# CONFIG_USB_EMI62 is not set
+# CONFIG_USB_EMI26 is not set
+# CONFIG_USB_AUERSWALD is not set
+# CONFIG_USB_RIO500 is not set
+# CONFIG_USB_LEGOTOWER is not set
+# CONFIG_USB_LCD is not set
+# CONFIG_USB_LED is not set
+# CONFIG_USB_CYPRESS_CY7C63 is not set
+# CONFIG_USB_CYTHERM is not set
+# CONFIG_USB_PHIDGETKIT is not set
+# CONFIG_USB_PHIDGETSERVO is not set
+# CONFIG_USB_IDMOUSE is not set
+# CONFIG_USB_APPLEDISPLAY is not set
+# CONFIG_USB_LD is not set
+
+#
+# USB DSL modem support
+#
+
+#
+# USB Gadget Support
+#
+# CONFIG_USB_GADGET is not set
+
+#
+# MMC/SD Card support
+#
+# CONFIG_MMC is not set
+
+#
+# LED devices
+#
+# CONFIG_NEW_LEDS is not set
+
+#
+# LED drivers
+#
+
+#
+# LED Triggers
+#
+
+#
+# InfiniBand support
+#
+# CONFIG_INFINIBAND is not set
+
+#
+# EDAC - error detection and reporting (RAS) (EXPERIMENTAL)
+#
+# CONFIG_EDAC is not set
+
+#
+# Real Time Clock
+#
+CONFIG_RTC_LIB=m
+CONFIG_RTC_CLASS=m
+
+#
+# RTC interfaces
+#
+CONFIG_RTC_INTF_SYSFS=m
+CONFIG_RTC_INTF_PROC=m
+CONFIG_RTC_INTF_DEV=m
+# CONFIG_RTC_INTF_DEV_UIE_EMUL is not set
+
+#
+# RTC drivers
+#
+# CONFIG_RTC_DRV_DS1553 is not set
+# CONFIG_RTC_DRV_DS1742 is not set
+CONFIG_RTC_DRV_M48T86=m
+CONFIG_RTC_DRV_TEST=m
+# CONFIG_RTC_DRV_V3020 is not set
+
+#
+# DMA Engine support
+#
+# CONFIG_DMA_ENGINE is not set
+
+#
+# DMA Clients
+#
+
+#
+# DMA Devices
+#
+
+#
+# Firmware Drivers
+#
+CONFIG_EDD=y
+# CONFIG_DELL_RBU is not set
+# CONFIG_DCDBAS is not set
+
+#
+# File systems
+#
+CONFIG_EXT2_FS=y
+# CONFIG_EXT2_FS_XATTR is not set
+# CONFIG_EXT2_FS_XIP is not set
+CONFIG_EXT3_FS=y
+CONFIG_EXT3_FS_XATTR=y
+# CONFIG_EXT3_FS_POSIX_ACL is not set
+# CONFIG_EXT3_FS_SECURITY is not set
+CONFIG_JBD=y
+# CONFIG_JBD_DEBUG is not set
+CONFIG_FS_MBCACHE=y
+CONFIG_REISERFS_FS=y
+# CONFIG_REISERFS_CHECK is not set
+# CONFIG_REISERFS_PROC_INFO is not set
+# CONFIG_REISERFS_FS_XATTR is not set
+# CONFIG_JFS_FS is not set
+# CONFIG_FS_POSIX_ACL is not set
+# CONFIG_XFS_FS is not set
+# CONFIG_OCFS2_FS is not set
+# CONFIG_MINIX_FS is not set
+# CONFIG_ROMFS_FS is not set
+CONFIG_INOTIFY=y
+CONFIG_INOTIFY_USER=y
+# CONFIG_QUOTA is not set
+CONFIG_DNOTIFY=y
+CONFIG_AUTOFS_FS=y
+CONFIG_AUTOFS4_FS=y
+# CONFIG_FUSE_FS is not set
+
+#
+# CD-ROM/DVD Filesystems
+#
+CONFIG_ISO9660_FS=y
+CONFIG_JOLIET=y
+CONFIG_ZISOFS=y
+CONFIG_ZISOFS_FS=y
+# CONFIG_UDF_FS is not set
+
+#
+# DOS/FAT/NT Filesystems
+#
+CONFIG_FAT_FS=m
+CONFIG_MSDOS_FS=m
+CONFIG_VFAT_FS=m
+CONFIG_FAT_DEFAULT_CODEPAGE=437
+CONFIG_FAT_DEFAULT_IOCHARSET="iso8859-1"
+# CONFIG_NTFS_FS is not set
+
+#
+# Pseudo filesystems
+#
+CONFIG_PROC_FS=y
+CONFIG_PROC_KCORE=y
+CONFIG_SYSFS=y
+CONFIG_TMPFS=y
+# CONFIG_HUGETLB_PAGE is not set
+CONFIG_RAMFS=y
+# CONFIG_CONFIGFS_FS is not set
+
+#
+# Miscellaneous filesystems
+#
+# CONFIG_ADFS_FS is not set
+# CONFIG_AFFS_FS is not set
+# CONFIG_HFS_FS is not set
+# CONFIG_HFSPLUS_FS is not set
+# CONFIG_BEFS_FS is not set
+# CONFIG_BFS_FS is not set
+# CONFIG_EFS_FS is not set
+CONFIG_CRAMFS=y
+# CONFIG_VXFS_FS is not set
+# CONFIG_HPFS_FS is not set
+# CONFIG_QNX4FS_FS is not set
+# CONFIG_SYSV_FS is not set
+# CONFIG_UFS_FS is not set
+
+#
+# Network File Systems
+#
+CONFIG_NFS_FS=y
+CONFIG_NFS_V3=y
+# CONFIG_NFS_V3_ACL is not set
+# CONFIG_NFS_V4 is not set
+# CONFIG_NFS_DIRECTIO is not set
+CONFIG_NFSD=m
+CONFIG_NFSD_V3=y
+# CONFIG_NFSD_V3_ACL is not set
+# CONFIG_NFSD_V4 is not set
+CONFIG_NFSD_TCP=y
+CONFIG_ROOT_NFS=y
+CONFIG_LOCKD=y
+CONFIG_LOCKD_V4=y
+CONFIG_EXPORTFS=m
+CONFIG_NFS_COMMON=y
+CONFIG_SUNRPC=y
+# CONFIG_RPCSEC_GSS_KRB5 is not set
+# CONFIG_RPCSEC_GSS_SPKM3 is not set
+# CONFIG_SMB_FS is not set
+# CONFIG_CIFS is not set
+# CONFIG_NCP_FS is not set
+# CONFIG_CODA_FS is not set
+# CONFIG_AFS_FS is not set
+# CONFIG_9P_FS is not set
+
+#
+# Partition Types
+#
+# CONFIG_PARTITION_ADVANCED is not set
+CONFIG_MSDOS_PARTITION=y
+
+#
+# Native Language Support
+#
+CONFIG_NLS=y
+CONFIG_NLS_DEFAULT="iso8859-1"
+CONFIG_NLS_CODEPAGE_437=y
+# CONFIG_NLS_CODEPAGE_737 is not set
+# CONFIG_NLS_CODEPAGE_775 is not set
+# CONFIG_NLS_CODEPAGE_850 is not set
+# CONFIG_NLS_CODEPAGE_852 is not set
+# CONFIG_NLS_CODEPAGE_855 is not set
+# CONFIG_NLS_CODEPAGE_857 is not set
+# CONFIG_NLS_CODEPAGE_860 is not set
+# CONFIG_NLS_CODEPAGE_861 is not set
+# CONFIG_NLS_CODEPAGE_862 is not set
+# CONFIG_NLS_CODEPAGE_863 is not set
+# CONFIG_NLS_CODEPAGE_864 is not set
+# CONFIG_NLS_CODEPAGE_865 is not set
+# CONFIG_NLS_CODEPAGE_866 is not set
+# CONFIG_NLS_CODEPAGE_869 is not set
+# CONFIG_NLS_CODEPAGE_936 is not set
+# CONFIG_NLS_CODEPAGE_950 is not set
+# CONFIG_NLS_CODEPAGE_932 is not set
+# CONFIG_NLS_CODEPAGE_949 is not set
+# CONFIG_NLS_CODEPAGE_874 is not set
+# CONFIG_NLS_ISO8859_8 is not set
+# CONFIG_NLS_CODEPAGE_1250 is not set
+# CONFIG_NLS_CODEPAGE_1251 is not set
+# CONFIG_NLS_ASCII is not set
+CONFIG_NLS_ISO8859_1=y
+# CONFIG_NLS_ISO8859_2 is not set
+# CONFIG_NLS_ISO8859_3 is not set
+# CONFIG_NLS_ISO8859_4 is not set
+# CONFIG_NLS_ISO8859_5 is not set
+# CONFIG_NLS_ISO8859_6 is not set
+# CONFIG_NLS_ISO8859_7 is not set
+# CONFIG_NLS_ISO8859_9 is not set
+# CONFIG_NLS_ISO8859_13 is not set
+# CONFIG_NLS_ISO8859_14 is not set
+# CONFIG_NLS_ISO8859_15 is not set
+# CONFIG_NLS_KOI8_R is not set
+# CONFIG_NLS_KOI8_U is not set
+# CONFIG_NLS_UTF8 is not set
+
+#
+# Instrumentation Support
+#
+# CONFIG_PROFILING is not set
+# CONFIG_KPROBES is not set
+
+#
+# Kernel hacking
+#
+CONFIG_TRACE_IRQFLAGS_SUPPORT=y
+# CONFIG_PRINTK_TIME is not set
+CONFIG_MAGIC_SYSRQ=y
+CONFIG_UNUSED_SYMBOLS=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_LOG_BUF_SHIFT=14
+CONFIG_DETECT_SOFTLOCKUP=y
+# CONFIG_SCHEDSTATS is not set
+# CONFIG_DEBUG_SLAB is not set
+# CONFIG_DEBUG_RT_MUTEXES is not set
+# CONFIG_RT_MUTEX_TESTER is not set
+# CONFIG_DEBUG_SPINLOCK is not set
+# CONFIG_DEBUG_MUTEXES is not set
+# CONFIG_DEBUG_RWSEMS is not set
+# CONFIG_DEBUG_LOCK_ALLOC is not set
+# CONFIG_PROVE_LOCKING is not set
+# CONFIG_DEBUG_SPINLOCK_SLEEP is not set
+# CONFIG_DEBUG_LOCKING_API_SELFTESTS is not set
+# CONFIG_DEBUG_KOBJECT is not set
+CONFIG_DEBUG_INFO=y
+# CONFIG_DEBUG_FS is not set
+# CONFIG_DEBUG_VM is not set
+CONFIG_FRAME_POINTER=y
+# CONFIG_UNWIND_INFO is not set
+CONFIG_FORCED_INLINING=y
+# CONFIG_RCU_TORTURE_TEST is not set
+# CONFIG_DEBUG_RODATA is not set
+# CONFIG_DEBUG_STACKOVERFLOW is not set
+# CONFIG_DEBUG_STACK_USAGE is not set
+
+#
+# Security options
+#
+# CONFIG_KEYS is not set
+# CONFIG_SECURITY is not set
+
+#
+# Cryptographic options
+#
+CONFIG_CRYPTO=y
+CONFIG_CRYPTO_HMAC=y
+# CONFIG_CRYPTO_NULL is not set
+# CONFIG_CRYPTO_MD4 is not set
+CONFIG_CRYPTO_MD5=m
+CONFIG_CRYPTO_SHA1=m
+# CONFIG_CRYPTO_SHA256 is not set
+# CONFIG_CRYPTO_SHA512 is not set
+# CONFIG_CRYPTO_WP512 is not set
+# CONFIG_CRYPTO_TGR192 is not set
+CONFIG_CRYPTO_DES=m
+# CONFIG_CRYPTO_BLOWFISH is not set
+# CONFIG_CRYPTO_TWOFISH is not set
+# CONFIG_CRYPTO_SERPENT is not set
+# CONFIG_CRYPTO_AES is not set
+# CONFIG_CRYPTO_AES_X86_64 is not set
+# CONFIG_CRYPTO_CAST5 is not set
+# CONFIG_CRYPTO_CAST6 is not set
+# CONFIG_CRYPTO_TEA is not set
+# CONFIG_CRYPTO_ARC4 is not set
+# CONFIG_CRYPTO_KHAZAD is not set
+# CONFIG_CRYPTO_ANUBIS is not set
+# CONFIG_CRYPTO_DEFLATE is not set
+# CONFIG_CRYPTO_MICHAEL_MIC is not set
+CONFIG_CRYPTO_CRC32C=m
+# CONFIG_CRYPTO_TEST is not set
+
+#
+# Hardware crypto devices
+#
+CONFIG_XEN=y
+CONFIG_XEN_INTERFACE_VERSION=0x00030207
+
+#
+# XEN
+#
+CONFIG_XEN_PRIVILEGED_GUEST=y
+# CONFIG_XEN_UNPRIVILEGED_GUEST is not set
+CONFIG_XEN_PRIVCMD=y
+CONFIG_XEN_XENBUS_DEV=y
+CONFIG_XEN_BACKEND=y
+CONFIG_XEN_BLKDEV_BACKEND=y
+CONFIG_XEN_BLKDEV_TAP=y
+CONFIG_XEN_NETDEV_BACKEND=y
+# CONFIG_XEN_NETDEV_PIPELINED_TRANSMITTER is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_UTIL is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_BACKEND is not set
+# CONFIG_XEN_NETDEV_LOOPBACK is not set
+CONFIG_XEN_PCIDEV_BACKEND=y
+# CONFIG_XEN_PCIDEV_BACKEND_VPCI is not set
+CONFIG_XEN_PCIDEV_BACKEND_PASS=y
+# CONFIG_XEN_PCIDEV_BACKEND_SLOT is not set
+# CONFIG_XEN_PCIDEV_BACKEND_CONTROLLER is not set
+# CONFIG_XEN_PCIDEV_BE_DEBUG is not set
+# CONFIG_XEN_TPMDEV_BACKEND is not set
+# CONFIG_XEN_SCSI_BACKEND is not set
+CONFIG_XEN_BLKDEV_FRONTEND=y
+CONFIG_XEN_NETDEV_FRONTEND=y
+CONFIG_XEN_SCSI_FRONTEND=m
+CONFIG_XEN_GRANT_DEV=y
+# CONFIG_XEN_NETDEV_ACCEL_SFC_FRONTEND is not set
+CONFIG_XEN_SCRUB_PAGES=y
+CONFIG_XEN_DISABLE_SERIAL=y
+CONFIG_XEN_SYSFS=y
+CONFIG_XEN_COMPAT_030002_AND_LATER=y
+# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
+# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
+CONFIG_XEN_COMPAT=0x030002
+CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
+CONFIG_NO_IDLE_HZ=y
+CONFIG_XEN_SMPBOOT=y
+CONFIG_XEN_BALLOON=y
+CONFIG_XEN_DEVMEM=y
+
+#
+# Library routines
+#
+# CONFIG_CRC_CCITT is not set
+# CONFIG_CRC16 is not set
+CONFIG_CRC32=y
+CONFIG_LIBCRC32C=y
+CONFIG_ZLIB_INFLATE=y
+CONFIG_PLIST=y
diff -rpuN linux-2.6.18.8/buildconfigs/linux-defconfig_xen_ia64 linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xen_ia64
--- linux-2.6.18.8/buildconfigs/linux-defconfig_xen_ia64	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xen_ia64	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1703 @@
+#
+# Automatically generated make config: don't edit
+# Linux kernel version: 2.6.18.8
+# Tue Feb 19 11:20:00 2008
+#
+CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
+
+#
+# Code maturity level options
+#
+CONFIG_EXPERIMENTAL=y
+CONFIG_LOCK_KERNEL=y
+CONFIG_INIT_ENV_ARG_LIMIT=32
+
+#
+# General setup
+#
+CONFIG_LOCALVERSION=""
+CONFIG_LOCALVERSION_AUTO=y
+CONFIG_SWAP=y
+CONFIG_SYSVIPC=y
+CONFIG_POSIX_MQUEUE=y
+CONFIG_BSD_PROCESS_ACCT=y
+# CONFIG_BSD_PROCESS_ACCT_V3 is not set
+# CONFIG_TASKSTATS is not set
+# CONFIG_AUDIT is not set
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+# CONFIG_CPUSETS is not set
+# CONFIG_RELAY is not set
+CONFIG_INITRAMFS_SOURCE=""
+CONFIG_CC_OPTIMIZE_FOR_SIZE=y
+# CONFIG_EMBEDDED is not set
+CONFIG_SYSCTL=y
+CONFIG_KALLSYMS=y
+CONFIG_KALLSYMS_ALL=y
+CONFIG_KALLSYMS_EXTRA_PASS=y
+CONFIG_HOTPLUG=y
+CONFIG_PRINTK=y
+CONFIG_BUG=y
+CONFIG_ELF_CORE=y
+CONFIG_BASE_FULL=y
+CONFIG_FUTEX=y
+CONFIG_EPOLL=y
+CONFIG_SHMEM=y
+CONFIG_SLAB=y
+CONFIG_VM_EVENT_COUNTERS=y
+CONFIG_RT_MUTEXES=y
+# CONFIG_TINY_SHMEM is not set
+CONFIG_BASE_SMALL=0
+# CONFIG_SLOB is not set
+
+#
+# Loadable module support
+#
+CONFIG_MODULES=y
+CONFIG_MODULE_UNLOAD=y
+# CONFIG_MODULE_FORCE_UNLOAD is not set
+CONFIG_MODVERSIONS=y
+CONFIG_MODULE_SRCVERSION_ALL=y
+CONFIG_KMOD=y
+CONFIG_STOP_MACHINE=y
+
+#
+# Block layer
+#
+# CONFIG_BLK_DEV_IO_TRACE is not set
+
+#
+# IO Schedulers
+#
+CONFIG_IOSCHED_NOOP=y
+CONFIG_IOSCHED_AS=y
+CONFIG_IOSCHED_DEADLINE=y
+CONFIG_IOSCHED_CFQ=y
+CONFIG_DEFAULT_AS=y
+# CONFIG_DEFAULT_DEADLINE is not set
+# CONFIG_DEFAULT_CFQ is not set
+# CONFIG_DEFAULT_NOOP is not set
+CONFIG_DEFAULT_IOSCHED="anticipatory"
+
+#
+# Processor type and features
+#
+CONFIG_IA64=y
+CONFIG_64BIT=y
+CONFIG_MMU=y
+CONFIG_SWIOTLB=y
+CONFIG_RWSEM_XCHGADD_ALGORITHM=y
+CONFIG_GENERIC_FIND_NEXT_BIT=y
+CONFIG_GENERIC_CALIBRATE_DELAY=y
+CONFIG_TIME_INTERPOLATION=y
+CONFIG_DMI=y
+CONFIG_EFI=y
+CONFIG_GENERIC_IOMAP=y
+CONFIG_XEN=y
+CONFIG_XEN_IA64_EXPOSE_P2M=y
+CONFIG_XEN_IA64_EXPOSE_P2M_USE_DTR=y
+CONFIG_SCHED_NO_NO_OMIT_FRAME_POINTER=y
+CONFIG_DMA_IS_DMA32=y
+CONFIG_AUDIT_ARCH=y
+CONFIG_IA64_GENERIC=y
+# CONFIG_IA64_DIG is not set
+# CONFIG_IA64_HP_ZX1 is not set
+# CONFIG_IA64_HP_ZX1_SWIOTLB is not set
+# CONFIG_IA64_SGI_SN2 is not set
+# CONFIG_IA64_HP_SIM is not set
+# CONFIG_IA64_XEN is not set
+# CONFIG_ITANIUM is not set
+CONFIG_MCKINLEY=y
+# CONFIG_IA64_PAGE_SIZE_4KB is not set
+# CONFIG_IA64_PAGE_SIZE_8KB is not set
+CONFIG_IA64_PAGE_SIZE_16KB=y
+# CONFIG_IA64_PAGE_SIZE_64KB is not set
+CONFIG_PGTABLE_3=y
+# CONFIG_PGTABLE_4 is not set
+CONFIG_HZ_100=y
+# CONFIG_HZ_250 is not set
+# CONFIG_HZ_1000 is not set
+CONFIG_HZ=100
+CONFIG_IA64_L1_CACHE_SHIFT=7
+CONFIG_IA64_CYCLONE=y
+CONFIG_IOSAPIC=y
+# CONFIG_IA64_SGI_SN_XP is not set
+CONFIG_FORCE_MAX_ZONEORDER=11
+CONFIG_SMP=y
+CONFIG_NR_CPUS=16
+CONFIG_HOTPLUG_CPU=y
+CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y
+# CONFIG_SCHED_SMT is not set
+# CONFIG_PERMIT_BSP_REMOVE is not set
+# CONFIG_PREEMPT is not set
+CONFIG_SELECT_MEMORY_MODEL=y
+# CONFIG_FLATMEM_MANUAL is not set
+CONFIG_DISCONTIGMEM_MANUAL=y
+# CONFIG_SPARSEMEM_MANUAL is not set
+CONFIG_DISCONTIGMEM=y
+CONFIG_FLAT_NODE_MEM_MAP=y
+CONFIG_NEED_MULTIPLE_NODES=y
+# CONFIG_SPARSEMEM_STATIC is not set
+CONFIG_SPLIT_PTLOCK_CPUS=4
+# CONFIG_MIGRATION is not set
+CONFIG_RESOURCES_64BIT=y
+CONFIG_ARCH_SELECT_MEMORY_MODEL=y
+CONFIG_ARCH_DISCONTIGMEM_ENABLE=y
+CONFIG_ARCH_FLATMEM_ENABLE=y
+CONFIG_ARCH_SPARSEMEM_ENABLE=y
+CONFIG_ARCH_DISCONTIGMEM_DEFAULT=y
+CONFIG_NUMA=y
+CONFIG_NODES_SHIFT=10
+CONFIG_VIRTUAL_MEM_MAP=y
+CONFIG_HOLES_IN_ZONE=y
+CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID=y
+CONFIG_HAVE_ARCH_NODEDATA_EXTENSION=y
+# CONFIG_IA32_SUPPORT is not set
+# CONFIG_IA64_MCA_RECOVERY is not set
+CONFIG_PERFMON=y
+CONFIG_IA64_PALINFO=y
+CONFIG_SGI_SN=y
+
+#
+# SN Devices
+#
+# CONFIG_SGI_IOC3 is not set
+CONFIG_KEXEC=y
+
+#
+# Firmware Drivers
+#
+CONFIG_EFI_VARS=y
+CONFIG_EFI_PCDP=y
+CONFIG_BINFMT_ELF=y
+CONFIG_BINFMT_MISC=y
+
+#
+# Power management and ACPI
+#
+CONFIG_PM=y
+CONFIG_PM_LEGACY=y
+# CONFIG_PM_DEBUG is not set
+
+#
+# ACPI (Advanced Configuration and Power Interface) Support
+#
+CONFIG_ACPI=y
+CONFIG_ACPI_BUTTON=y
+CONFIG_ACPI_FAN=y
+# CONFIG_ACPI_DOCK is not set
+CONFIG_ACPI_PROCESSOR=y
+CONFIG_ACPI_HOTPLUG_CPU=y
+CONFIG_ACPI_THERMAL=y
+CONFIG_ACPI_NUMA=y
+CONFIG_ACPI_BLACKLIST_YEAR=0
+# CONFIG_ACPI_DEBUG is not set
+CONFIG_ACPI_EC=y
+CONFIG_ACPI_POWER=y
+CONFIG_ACPI_SYSTEM=y
+CONFIG_ACPI_CONTAINER=y
+
+#
+# CPU Frequency scaling
+#
+# CONFIG_CPU_FREQ is not set
+
+#
+# Bus options (PCI, PCMCIA)
+#
+CONFIG_PCI=y
+CONFIG_PCI_DOMAINS=y
+CONFIG_XEN_PCIDEV_FRONTEND=y
+# CONFIG_XEN_PCIDEV_FE_DEBUG is not set
+# CONFIG_PCIEPORTBUS is not set
+# CONFIG_PCI_DEBUG is not set
+
+#
+# PCI Hotplug Support
+#
+CONFIG_HOTPLUG_PCI=y
+# CONFIG_HOTPLUG_PCI_FAKE is not set
+CONFIG_HOTPLUG_PCI_ACPI=y
+# CONFIG_HOTPLUG_PCI_ACPI_IBM is not set
+# CONFIG_HOTPLUG_PCI_CPCI is not set
+# CONFIG_HOTPLUG_PCI_SHPC is not set
+# CONFIG_HOTPLUG_PCI_SGI is not set
+
+#
+# PCCARD (PCMCIA/CardBus) support
+#
+# CONFIG_PCCARD is not set
+
+#
+# Networking
+#
+CONFIG_NET=y
+
+#
+# Networking options
+#
+# CONFIG_NETDEBUG is not set
+CONFIG_PACKET=y
+# CONFIG_PACKET_MMAP is not set
+CONFIG_UNIX=y
+CONFIG_XFRM=y
+# CONFIG_XFRM_USER is not set
+# CONFIG_NET_KEY is not set
+CONFIG_INET=y
+CONFIG_IP_MULTICAST=y
+# CONFIG_IP_ADVANCED_ROUTER is not set
+CONFIG_IP_FIB_HASH=y
+CONFIG_IP_PNP=y
+CONFIG_IP_PNP_DHCP=y
+# CONFIG_IP_PNP_BOOTP is not set
+# CONFIG_IP_PNP_RARP is not set
+# CONFIG_NET_IPIP is not set
+# CONFIG_NET_IPGRE is not set
+# CONFIG_IP_MROUTE is not set
+CONFIG_ARPD=y
+CONFIG_SYN_COOKIES=y
+# CONFIG_INET_AH is not set
+# CONFIG_INET_ESP is not set
+# CONFIG_INET_IPCOMP is not set
+# CONFIG_INET_XFRM_TUNNEL is not set
+# CONFIG_INET_TUNNEL is not set
+CONFIG_INET_XFRM_MODE_TRANSPORT=y
+CONFIG_INET_XFRM_MODE_TUNNEL=y
+CONFIG_INET_DIAG=y
+CONFIG_INET_TCP_DIAG=y
+# CONFIG_TCP_CONG_ADVANCED is not set
+CONFIG_TCP_CONG_BIC=y
+
+#
+# IP: Virtual Server Configuration
+#
+# CONFIG_IP_VS is not set
+# CONFIG_IPV6 is not set
+# CONFIG_INET6_XFRM_TUNNEL is not set
+# CONFIG_INET6_TUNNEL is not set
+# CONFIG_NETWORK_SECMARK is not set
+CONFIG_NETFILTER=y
+# CONFIG_NETFILTER_DEBUG is not set
+CONFIG_BRIDGE_NETFILTER=y
+
+#
+# Core Netfilter Configuration
+#
+# CONFIG_NETFILTER_NETLINK is not set
+# CONFIG_NF_CONNTRACK is not set
+# CONFIG_NETFILTER_XTABLES is not set
+
+#
+# IP: Netfilter Configuration
+#
+# CONFIG_IP_NF_CONNTRACK is not set
+# CONFIG_IP_NF_QUEUE is not set
+
+#
+# Bridge: Netfilter Configuration
+#
+# CONFIG_BRIDGE_NF_EBTABLES is not set
+
+#
+# DCCP Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP_DCCP is not set
+
+#
+# SCTP Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP_SCTP is not set
+
+#
+# TIPC Configuration (EXPERIMENTAL)
+#
+# CONFIG_TIPC is not set
+# CONFIG_ATM is not set
+CONFIG_BRIDGE=y
+# CONFIG_VLAN_8021Q is not set
+# CONFIG_DECNET is not set
+CONFIG_LLC=y
+# CONFIG_LLC2 is not set
+# CONFIG_IPX is not set
+# CONFIG_ATALK is not set
+# CONFIG_X25 is not set
+# CONFIG_LAPB is not set
+# CONFIG_ECONET is not set
+# CONFIG_WAN_ROUTER is not set
+
+#
+# QoS and/or fair queueing
+#
+# CONFIG_NET_SCHED is not set
+
+#
+# Network testing
+#
+# CONFIG_NET_PKTGEN is not set
+# CONFIG_HAMRADIO is not set
+# CONFIG_IRDA is not set
+# CONFIG_BT is not set
+# CONFIG_IEEE80211 is not set
+
+#
+# Device Drivers
+#
+
+#
+# Generic Driver Options
+#
+CONFIG_STANDALONE=y
+CONFIG_PREVENT_FIRMWARE_BUILD=y
+CONFIG_FW_LOADER=y
+# CONFIG_DEBUG_DRIVER is not set
+CONFIG_SYS_HYPERVISOR=y
+
+#
+# Connector - unified userspace <-> kernelspace linker
+#
+# CONFIG_CONNECTOR is not set
+
+#
+# Memory Technology Devices (MTD)
+#
+# CONFIG_MTD is not set
+
+#
+# Parallel port support
+#
+# CONFIG_PARPORT is not set
+
+#
+# Plug and Play support
+#
+# CONFIG_PNP is not set
+
+#
+# Block devices
+#
+# CONFIG_BLK_CPQ_DA is not set
+CONFIG_BLK_CPQ_CISS_DA=y
+# CONFIG_CISS_SCSI_TAPE is not set
+# CONFIG_BLK_DEV_DAC960 is not set
+# CONFIG_BLK_DEV_UMEM is not set
+# CONFIG_BLK_DEV_COW_COMMON is not set
+CONFIG_BLK_DEV_LOOP=y
+CONFIG_BLK_DEV_CRYPTOLOOP=y
+CONFIG_BLK_DEV_NBD=m
+# CONFIG_BLK_DEV_SX8 is not set
+# CONFIG_BLK_DEV_UB is not set
+CONFIG_BLK_DEV_RAM=y
+CONFIG_BLK_DEV_RAM_COUNT=16
+CONFIG_BLK_DEV_RAM_SIZE=4096
+CONFIG_BLK_DEV_RAM_BLOCKSIZE=1024
+CONFIG_BLK_DEV_INITRD=y
+# CONFIG_CDROM_PKTCDVD is not set
+# CONFIG_ATA_OVER_ETH is not set
+
+#
+# ATA/ATAPI/MFM/RLL support
+#
+CONFIG_IDE=y
+CONFIG_IDE_MAX_HWIFS=4
+CONFIG_BLK_DEV_IDE=y
+
+#
+# Please see Documentation/ide.txt for help/info on IDE drives
+#
+# CONFIG_BLK_DEV_IDE_SATA is not set
+CONFIG_BLK_DEV_IDEDISK=y
+# CONFIG_IDEDISK_MULTI_MODE is not set
+CONFIG_BLK_DEV_IDECD=y
+# CONFIG_BLK_DEV_IDETAPE is not set
+CONFIG_BLK_DEV_IDEFLOPPY=y
+CONFIG_BLK_DEV_IDESCSI=y
+# CONFIG_IDE_TASK_IOCTL is not set
+
+#
+# IDE chipset support/bugfixes
+#
+# CONFIG_IDE_GENERIC is not set
+CONFIG_BLK_DEV_IDEPCI=y
+# CONFIG_IDEPCI_SHARE_IRQ is not set
+# CONFIG_BLK_DEV_OFFBOARD is not set
+CONFIG_BLK_DEV_GENERIC=y
+# CONFIG_BLK_DEV_OPTI621 is not set
+CONFIG_BLK_DEV_IDEDMA_PCI=y
+# CONFIG_BLK_DEV_IDEDMA_FORCED is not set
+CONFIG_IDEDMA_PCI_AUTO=y
+# CONFIG_IDEDMA_ONLYDISK is not set
+# CONFIG_BLK_DEV_AEC62XX is not set
+# CONFIG_BLK_DEV_ALI15X3 is not set
+# CONFIG_BLK_DEV_AMD74XX is not set
+CONFIG_BLK_DEV_CMD64X=y
+# CONFIG_BLK_DEV_TRIFLEX is not set
+# CONFIG_BLK_DEV_CY82C693 is not set
+# CONFIG_BLK_DEV_CS5520 is not set
+# CONFIG_BLK_DEV_CS5530 is not set
+# CONFIG_BLK_DEV_HPT34X is not set
+# CONFIG_BLK_DEV_HPT366 is not set
+# CONFIG_BLK_DEV_SC1200 is not set
+CONFIG_BLK_DEV_PIIX=y
+# CONFIG_BLK_DEV_IT821X is not set
+# CONFIG_BLK_DEV_NS87415 is not set
+# CONFIG_BLK_DEV_PDC202XX_OLD is not set
+# CONFIG_BLK_DEV_PDC202XX_NEW is not set
+# CONFIG_BLK_DEV_SVWKS is not set
+# CONFIG_BLK_DEV_SIIMAGE is not set
+# CONFIG_BLK_DEV_SLC90E66 is not set
+# CONFIG_BLK_DEV_TRM290 is not set
+# CONFIG_BLK_DEV_VIA82CXXX is not set
+# CONFIG_IDE_ARM is not set
+CONFIG_BLK_DEV_IDEDMA=y
+# CONFIG_IDEDMA_IVB is not set
+CONFIG_IDEDMA_AUTO=y
+# CONFIG_BLK_DEV_HD is not set
+
+#
+# SCSI device support
+#
+# CONFIG_RAID_ATTRS is not set
+CONFIG_SCSI=y
+CONFIG_SCSI_PROC_FS=y
+
+#
+# SCSI support type (disk, tape, CD-ROM)
+#
+CONFIG_BLK_DEV_SD=y
+CONFIG_CHR_DEV_ST=y
+CONFIG_CHR_DEV_OSST=y
+CONFIG_BLK_DEV_SR=y
+CONFIG_BLK_DEV_SR_VENDOR=y
+CONFIG_CHR_DEV_SG=y
+# CONFIG_CHR_DEV_SCH is not set
+
+#
+# Some SCSI devices (e.g. CD jukebox) support multiple LUNs
+#
+CONFIG_SCSI_MULTI_LUN=y
+CONFIG_SCSI_CONSTANTS=y
+CONFIG_SCSI_LOGGING=y
+
+#
+# SCSI Transport Attributes
+#
+CONFIG_SCSI_SPI_ATTRS=y
+CONFIG_SCSI_FC_ATTRS=y
+# CONFIG_SCSI_ISCSI_ATTRS is not set
+CONFIG_SCSI_SAS_ATTRS=y
+
+#
+# SCSI low-level drivers
+#
+# CONFIG_ISCSI_TCP is not set
+# CONFIG_BLK_DEV_3W_XXXX_RAID is not set
+# CONFIG_SCSI_3W_9XXX is not set
+# CONFIG_SCSI_ACARD is not set
+# CONFIG_SCSI_AACRAID is not set
+# CONFIG_SCSI_AIC7XXX is not set
+# CONFIG_SCSI_AIC7XXX_OLD is not set
+# CONFIG_SCSI_AIC79XX is not set
+# CONFIG_MEGARAID_NEWGEN is not set
+# CONFIG_MEGARAID_LEGACY is not set
+# CONFIG_MEGARAID_SAS is not set
+# CONFIG_SCSI_SATA is not set
+# CONFIG_SCSI_HPTIOP is not set
+# CONFIG_SCSI_DMX3191D is not set
+# CONFIG_SCSI_FUTURE_DOMAIN is not set
+# CONFIG_SCSI_IPS is not set
+# CONFIG_SCSI_INITIO is not set
+# CONFIG_SCSI_INIA100 is not set
+CONFIG_SCSI_SYM53C8XX_2=y
+CONFIG_SCSI_SYM53C8XX_DMA_ADDRESSING_MODE=1
+CONFIG_SCSI_SYM53C8XX_DEFAULT_TAGS=16
+CONFIG_SCSI_SYM53C8XX_MAX_TAGS=64
+CONFIG_SCSI_SYM53C8XX_MMIO=y
+# CONFIG_SCSI_IPR is not set
+CONFIG_SCSI_QLOGIC_1280=y
+# CONFIG_SCSI_QLA_FC is not set
+# CONFIG_SCSI_LPFC is not set
+# CONFIG_SCSI_DC395x is not set
+# CONFIG_SCSI_DC390T is not set
+# CONFIG_SCSI_DEBUG is not set
+
+#
+# Multi-device support (RAID and LVM)
+#
+CONFIG_MD=y
+# CONFIG_BLK_DEV_MD is not set
+CONFIG_BLK_DEV_DM=y
+CONFIG_DM_CRYPT=m
+CONFIG_DM_SNAPSHOT=y
+CONFIG_DM_MIRROR=m
+CONFIG_DM_ZERO=m
+CONFIG_DM_MULTIPATH=m
+CONFIG_DM_MULTIPATH_EMC=m
+
+#
+# Fusion MPT device support
+#
+CONFIG_FUSION=y
+CONFIG_FUSION_SPI=y
+# CONFIG_FUSION_FC is not set
+CONFIG_FUSION_SAS=y
+CONFIG_FUSION_MAX_SGE=128
+# CONFIG_FUSION_CTL is not set
+
+#
+# IEEE 1394 (FireWire) support
+#
+# CONFIG_IEEE1394 is not set
+
+#
+# I2O device support
+#
+# CONFIG_I2O is not set
+
+#
+# Network device support
+#
+CONFIG_NETDEVICES=y
+CONFIG_DUMMY=y
+# CONFIG_BONDING is not set
+# CONFIG_EQUALIZER is not set
+CONFIG_TUN=y
+
+#
+# ARCnet devices
+#
+CONFIG_ARCNET=y
+# CONFIG_ARCNET_1201 is not set
+# CONFIG_ARCNET_1051 is not set
+# CONFIG_ARCNET_RAW is not set
+# CONFIG_ARCNET_CAP is not set
+# CONFIG_ARCNET_COM90xx is not set
+# CONFIG_ARCNET_COM90xxIO is not set
+# CONFIG_ARCNET_RIM_I is not set
+# CONFIG_ARCNET_COM20020 is not set
+
+#
+# PHY device support
+#
+# CONFIG_PHYLIB is not set
+
+#
+# Ethernet (10 or 100Mbit)
+#
+CONFIG_NET_ETHERNET=y
+CONFIG_MII=y
+# CONFIG_HAPPYMEAL is not set
+# CONFIG_SUNGEM is not set
+# CONFIG_CASSINI is not set
+# CONFIG_NET_VENDOR_3COM is not set
+
+#
+# Tulip family network device support
+#
+CONFIG_NET_TULIP=y
+# CONFIG_DE2104X is not set
+CONFIG_TULIP=y
+CONFIG_TULIP_MWI=y
+CONFIG_TULIP_MMIO=y
+CONFIG_TULIP_NAPI=y
+CONFIG_TULIP_NAPI_HW_MITIGATION=y
+# CONFIG_DE4X5 is not set
+# CONFIG_WINBOND_840 is not set
+# CONFIG_DM9102 is not set
+# CONFIG_ULI526X is not set
+# CONFIG_HP100 is not set
+CONFIG_NET_PCI=y
+# CONFIG_PCNET32 is not set
+# CONFIG_AMD8111_ETH is not set
+# CONFIG_ADAPTEC_STARFIRE is not set
+# CONFIG_B44 is not set
+# CONFIG_FORCEDETH is not set
+# CONFIG_DGRS is not set
+CONFIG_EEPRO100=y
+CONFIG_E100=y
+# CONFIG_FEALNX is not set
+# CONFIG_NATSEMI is not set
+# CONFIG_NE2K_PCI is not set
+# CONFIG_8139CP is not set
+# CONFIG_8139TOO is not set
+# CONFIG_SIS900 is not set
+# CONFIG_EPIC100 is not set
+# CONFIG_SUNDANCE is not set
+# CONFIG_VIA_RHINE is not set
+
+#
+# Ethernet (1000 Mbit)
+#
+# CONFIG_ACENIC is not set
+# CONFIG_DL2K is not set
+CONFIG_E1000=y
+# CONFIG_E1000_NAPI is not set
+# CONFIG_E1000_DISABLE_PACKET_SPLIT is not set
+# CONFIG_NS83820 is not set
+# CONFIG_HAMACHI is not set
+# CONFIG_YELLOWFIN is not set
+# CONFIG_R8169 is not set
+# CONFIG_SIS190 is not set
+# CONFIG_SKGE is not set
+# CONFIG_SKY2 is not set
+# CONFIG_SK98LIN is not set
+# CONFIG_VIA_VELOCITY is not set
+CONFIG_TIGON3=y
+# CONFIG_BNX2 is not set
+
+#
+# Ethernet (10000 Mbit)
+#
+# CONFIG_CHELSIO_T1 is not set
+# CONFIG_IXGB is not set
+# CONFIG_S2IO is not set
+# CONFIG_MYRI10GE is not set
+# CONFIG_SFC is not set
+
+#
+# Token Ring devices
+#
+# CONFIG_TR is not set
+
+#
+# Wireless LAN (non-hamradio)
+#
+# CONFIG_NET_RADIO is not set
+
+#
+# Wan interfaces
+#
+# CONFIG_WAN is not set
+# CONFIG_FDDI is not set
+# CONFIG_HIPPI is not set
+# CONFIG_PPP is not set
+# CONFIG_SLIP is not set
+# CONFIG_NET_FC is not set
+# CONFIG_SHAPER is not set
+CONFIG_NETCONSOLE=y
+CONFIG_NETPOLL=y
+# CONFIG_NETPOLL_RX is not set
+# CONFIG_NETPOLL_TRAP is not set
+CONFIG_NET_POLL_CONTROLLER=y
+
+#
+# ISDN subsystem
+#
+CONFIG_ISDN=m
+
+#
+# Old ISDN4Linux
+#
+# CONFIG_ISDN_I4L is not set
+
+#
+# CAPI subsystem
+#
+# CONFIG_ISDN_CAPI is not set
+
+#
+# Telephony Support
+#
+# CONFIG_PHONE is not set
+
+#
+# Input device support
+#
+CONFIG_INPUT=y
+
+#
+# Userland interfaces
+#
+CONFIG_INPUT_MOUSEDEV=y
+CONFIG_INPUT_MOUSEDEV_PSAUX=y
+CONFIG_INPUT_MOUSEDEV_SCREEN_X=1024
+CONFIG_INPUT_MOUSEDEV_SCREEN_Y=768
+CONFIG_INPUT_JOYDEV=y
+# CONFIG_INPUT_TSDEV is not set
+CONFIG_INPUT_EVDEV=y
+# CONFIG_INPUT_EVBUG is not set
+
+#
+# Input Device Drivers
+#
+CONFIG_INPUT_KEYBOARD=y
+CONFIG_KEYBOARD_ATKBD=y
+# CONFIG_KEYBOARD_SUNKBD is not set
+# CONFIG_KEYBOARD_LKKBD is not set
+# CONFIG_KEYBOARD_XTKBD is not set
+# CONFIG_KEYBOARD_NEWTON is not set
+CONFIG_INPUT_MOUSE=y
+CONFIG_MOUSE_PS2=y
+# CONFIG_MOUSE_SERIAL is not set
+# CONFIG_MOUSE_VSXXXAA is not set
+# CONFIG_INPUT_JOYSTICK is not set
+# CONFIG_INPUT_TOUCHSCREEN is not set
+# CONFIG_INPUT_MISC is not set
+
+#
+# Hardware I/O ports
+#
+CONFIG_SERIO=y
+CONFIG_SERIO_I8042=y
+# CONFIG_SERIO_SERPORT is not set
+# CONFIG_SERIO_PCIPS2 is not set
+CONFIG_SERIO_LIBPS2=y
+# CONFIG_SERIO_RAW is not set
+CONFIG_GAMEPORT=y
+# CONFIG_GAMEPORT_NS558 is not set
+# CONFIG_GAMEPORT_L4 is not set
+# CONFIG_GAMEPORT_EMU10K1 is not set
+# CONFIG_GAMEPORT_FM801 is not set
+
+#
+# Character devices
+#
+CONFIG_VT=y
+CONFIG_VT_CONSOLE=y
+CONFIG_HW_CONSOLE=y
+# CONFIG_VT_HW_CONSOLE_BINDING is not set
+CONFIG_SERIAL_NONSTANDARD=y
+# CONFIG_COMPUTONE is not set
+# CONFIG_ROCKETPORT is not set
+# CONFIG_CYCLADES is not set
+# CONFIG_DIGIEPCA is not set
+# CONFIG_MOXA_INTELLIO is not set
+# CONFIG_MOXA_SMARTIO is not set
+# CONFIG_ISI is not set
+# CONFIG_SYNCLINKMP is not set
+# CONFIG_SYNCLINK_GT is not set
+# CONFIG_N_HDLC is not set
+# CONFIG_SPECIALIX is not set
+# CONFIG_SX is not set
+# CONFIG_RIO is not set
+# CONFIG_STALDRV is not set
+# CONFIG_SGI_SNSC is not set
+# CONFIG_SGI_TIOCX is not set
+
+#
+# Serial drivers
+#
+
+#
+# Non-8250 serial port support
+#
+# CONFIG_SERIAL_SGI_L1_CONSOLE is not set
+# CONFIG_SERIAL_JSM is not set
+CONFIG_UNIX98_PTYS=y
+CONFIG_LEGACY_PTYS=y
+CONFIG_LEGACY_PTY_COUNT=256
+
+#
+# IPMI
+#
+# CONFIG_IPMI_HANDLER is not set
+
+#
+# Watchdog Cards
+#
+# CONFIG_WATCHDOG is not set
+# CONFIG_HW_RANDOM is not set
+CONFIG_EFI_RTC=y
+# CONFIG_DTLK is not set
+# CONFIG_R3964 is not set
+# CONFIG_APPLICOM is not set
+
+#
+# Ftape, the floppy tape device driver
+#
+CONFIG_AGP=y
+# CONFIG_AGP_SIS is not set
+# CONFIG_AGP_VIA is not set
+CONFIG_AGP_I460=y
+# CONFIG_AGP_HP_ZX1 is not set
+# CONFIG_AGP_SGI_TIOCA is not set
+CONFIG_DRM=y
+# CONFIG_DRM_TDFX is not set
+# CONFIG_DRM_R128 is not set
+# CONFIG_DRM_RADEON is not set
+# CONFIG_DRM_MGA is not set
+# CONFIG_DRM_SIS is not set
+# CONFIG_DRM_VIA is not set
+# CONFIG_DRM_SAVAGE is not set
+# CONFIG_RAW_DRIVER is not set
+# CONFIG_HPET is not set
+# CONFIG_HANGCHECK_TIMER is not set
+# CONFIG_MMTIMER is not set
+
+#
+# TPM devices
+#
+# CONFIG_TCG_TPM is not set
+# CONFIG_TELCLOCK is not set
+
+#
+# I2C support
+#
+CONFIG_I2C=y
+CONFIG_I2C_CHARDEV=y
+
+#
+# I2C Algorithms
+#
+CONFIG_I2C_ALGOBIT=y
+CONFIG_I2C_ALGOPCF=y
+# CONFIG_I2C_ALGOPCA is not set
+
+#
+# I2C Hardware Bus support
+#
+# CONFIG_I2C_ALI1535 is not set
+# CONFIG_I2C_ALI1563 is not set
+# CONFIG_I2C_ALI15X3 is not set
+# CONFIG_I2C_AMD756 is not set
+# CONFIG_I2C_AMD8111 is not set
+# CONFIG_I2C_I801 is not set
+# CONFIG_I2C_I810 is not set
+# CONFIG_I2C_PIIX4 is not set
+# CONFIG_I2C_NFORCE2 is not set
+# CONFIG_I2C_OCORES is not set
+# CONFIG_I2C_PARPORT_LIGHT is not set
+# CONFIG_I2C_PROSAVAGE is not set
+# CONFIG_I2C_SAVAGE4 is not set
+# CONFIG_I2C_SIS5595 is not set
+# CONFIG_I2C_SIS630 is not set
+# CONFIG_I2C_SIS96X is not set
+# CONFIG_I2C_STUB is not set
+# CONFIG_I2C_VIA is not set
+# CONFIG_I2C_VIAPRO is not set
+# CONFIG_I2C_VOODOO3 is not set
+# CONFIG_I2C_PCA_ISA is not set
+
+#
+# Miscellaneous I2C Chip support
+#
+# CONFIG_SENSORS_DS1337 is not set
+# CONFIG_SENSORS_DS1374 is not set
+# CONFIG_SENSORS_EEPROM is not set
+# CONFIG_SENSORS_PCF8574 is not set
+# CONFIG_SENSORS_PCA9539 is not set
+# CONFIG_SENSORS_PCF8591 is not set
+# CONFIG_SENSORS_MAX6875 is not set
+# CONFIG_I2C_DEBUG_CORE is not set
+# CONFIG_I2C_DEBUG_ALGO is not set
+# CONFIG_I2C_DEBUG_BUS is not set
+# CONFIG_I2C_DEBUG_CHIP is not set
+
+#
+# SPI support
+#
+# CONFIG_SPI is not set
+# CONFIG_SPI_MASTER is not set
+
+#
+# Dallas's 1-wire bus
+#
+
+#
+# Hardware Monitoring support
+#
+CONFIG_HWMON=y
+# CONFIG_HWMON_VID is not set
+# CONFIG_SENSORS_ABITUGURU is not set
+# CONFIG_SENSORS_ADM1021 is not set
+# CONFIG_SENSORS_ADM1025 is not set
+# CONFIG_SENSORS_ADM1026 is not set
+# CONFIG_SENSORS_ADM1031 is not set
+# CONFIG_SENSORS_ADM9240 is not set
+# CONFIG_SENSORS_ASB100 is not set
+# CONFIG_SENSORS_ATXP1 is not set
+# CONFIG_SENSORS_DS1621 is not set
+# CONFIG_SENSORS_F71805F is not set
+# CONFIG_SENSORS_FSCHER is not set
+# CONFIG_SENSORS_FSCPOS is not set
+# CONFIG_SENSORS_GL518SM is not set
+# CONFIG_SENSORS_GL520SM is not set
+# CONFIG_SENSORS_IT87 is not set
+# CONFIG_SENSORS_LM63 is not set
+# CONFIG_SENSORS_LM75 is not set
+# CONFIG_SENSORS_LM77 is not set
+# CONFIG_SENSORS_LM78 is not set
+# CONFIG_SENSORS_LM80 is not set
+# CONFIG_SENSORS_LM83 is not set
+# CONFIG_SENSORS_LM85 is not set
+# CONFIG_SENSORS_LM87 is not set
+# CONFIG_SENSORS_LM90 is not set
+# CONFIG_SENSORS_LM92 is not set
+# CONFIG_SENSORS_MAX1619 is not set
+# CONFIG_SENSORS_PC87360 is not set
+# CONFIG_SENSORS_SIS5595 is not set
+# CONFIG_SENSORS_SMSC47M1 is not set
+# CONFIG_SENSORS_SMSC47M192 is not set
+# CONFIG_SENSORS_SMSC47B397 is not set
+# CONFIG_SENSORS_VIA686A is not set
+# CONFIG_SENSORS_VT8231 is not set
+# CONFIG_SENSORS_W83781D is not set
+# CONFIG_SENSORS_W83791D is not set
+# CONFIG_SENSORS_W83792D is not set
+# CONFIG_SENSORS_W83L785TS is not set
+# CONFIG_SENSORS_W83627HF is not set
+# CONFIG_SENSORS_W83627EHF is not set
+# CONFIG_HWMON_DEBUG_CHIP is not set
+
+#
+# Misc devices
+#
+
+#
+# Multimedia devices
+#
+CONFIG_VIDEO_DEV=y
+CONFIG_VIDEO_V4L1=y
+CONFIG_VIDEO_V4L1_COMPAT=y
+CONFIG_VIDEO_V4L2=y
+
+#
+# Video Capture Adapters
+#
+
+#
+# Video Capture Adapters
+#
+# CONFIG_VIDEO_ADV_DEBUG is not set
+# CONFIG_VIDEO_VIVI is not set
+# CONFIG_VIDEO_BT848 is not set
+# CONFIG_VIDEO_CPIA is not set
+# CONFIG_VIDEO_CPIA2 is not set
+# CONFIG_VIDEO_SAA5246A is not set
+# CONFIG_VIDEO_SAA5249 is not set
+# CONFIG_TUNER_3036 is not set
+# CONFIG_VIDEO_STRADIS is not set
+# CONFIG_VIDEO_ZORAN is not set
+# CONFIG_VIDEO_SAA7134 is not set
+# CONFIG_VIDEO_MXB is not set
+# CONFIG_VIDEO_DPC is not set
+# CONFIG_VIDEO_HEXIUM_ORION is not set
+# CONFIG_VIDEO_HEXIUM_GEMINI is not set
+# CONFIG_VIDEO_CX88 is not set
+
+#
+# Encoders and Decoders
+#
+# CONFIG_VIDEO_MSP3400 is not set
+# CONFIG_VIDEO_CS53L32A is not set
+# CONFIG_VIDEO_TLV320AIC23B is not set
+# CONFIG_VIDEO_WM8775 is not set
+# CONFIG_VIDEO_WM8739 is not set
+# CONFIG_VIDEO_CX2341X is not set
+# CONFIG_VIDEO_CX25840 is not set
+# CONFIG_VIDEO_SAA711X is not set
+# CONFIG_VIDEO_SAA7127 is not set
+# CONFIG_VIDEO_UPD64031A is not set
+# CONFIG_VIDEO_UPD64083 is not set
+
+#
+# V4L USB devices
+#
+# CONFIG_VIDEO_PVRUSB2 is not set
+# CONFIG_VIDEO_EM28XX is not set
+# CONFIG_USB_VICAM is not set
+# CONFIG_USB_IBMCAM is not set
+# CONFIG_USB_KONICAWC is not set
+# CONFIG_USB_QUICKCAM_MESSENGER is not set
+# CONFIG_USB_ET61X251 is not set
+# CONFIG_VIDEO_OVCAMCHIP is not set
+# CONFIG_USB_W9968CF is not set
+# CONFIG_USB_OV511 is not set
+# CONFIG_USB_SE401 is not set
+# CONFIG_USB_SN9C102 is not set
+# CONFIG_USB_STV680 is not set
+# CONFIG_USB_ZC0301 is not set
+# CONFIG_USB_PWC is not set
+
+#
+# Radio Adapters
+#
+# CONFIG_RADIO_GEMTEK_PCI is not set
+# CONFIG_RADIO_MAXIRADIO is not set
+# CONFIG_RADIO_MAESTRO is not set
+# CONFIG_USB_DSBR is not set
+
+#
+# Digital Video Broadcasting Devices
+#
+# CONFIG_DVB is not set
+# CONFIG_USB_DABUSB is not set
+
+#
+# Graphics support
+#
+CONFIG_FIRMWARE_EDID=y
+CONFIG_FB=y
+CONFIG_FB_CFB_FILLRECT=y
+CONFIG_FB_CFB_COPYAREA=y
+CONFIG_FB_CFB_IMAGEBLIT=y
+# CONFIG_FB_MACMODES is not set
+# CONFIG_FB_BACKLIGHT is not set
+CONFIG_FB_MODE_HELPERS=y
+# CONFIG_FB_TILEBLITTING is not set
+# CONFIG_FB_CIRRUS is not set
+# CONFIG_FB_PM2 is not set
+# CONFIG_FB_CYBER2000 is not set
+# CONFIG_FB_ASILIANT is not set
+# CONFIG_FB_IMSTT is not set
+# CONFIG_FB_S1D13XXX is not set
+# CONFIG_FB_NVIDIA is not set
+# CONFIG_FB_RIVA is not set
+# CONFIG_FB_MATROX is not set
+# CONFIG_FB_RADEON is not set
+# CONFIG_FB_ATY128 is not set
+# CONFIG_FB_ATY is not set
+# CONFIG_FB_SAVAGE is not set
+# CONFIG_FB_SIS is not set
+# CONFIG_FB_NEOMAGIC is not set
+# CONFIG_FB_KYRO is not set
+# CONFIG_FB_3DFX is not set
+# CONFIG_FB_VOODOO1 is not set
+# CONFIG_FB_TRIDENT is not set
+# CONFIG_FB_VIRTUAL is not set
+
+#
+# Console display driver support
+#
+CONFIG_VGA_CONSOLE=y
+# CONFIG_VGACON_SOFT_SCROLLBACK is not set
+CONFIG_DUMMY_CONSOLE=y
+CONFIG_FRAMEBUFFER_CONSOLE=y
+# CONFIG_FRAMEBUFFER_CONSOLE_ROTATION is not set
+# CONFIG_FONTS is not set
+CONFIG_FONT_8x8=y
+CONFIG_FONT_8x16=y
+
+#
+# Logo configuration
+#
+CONFIG_LOGO=y
+# CONFIG_LOGO_LINUX_MONO is not set
+# CONFIG_LOGO_LINUX_VGA16 is not set
+CONFIG_LOGO_LINUX_CLUT224=y
+# CONFIG_BACKLIGHT_LCD_SUPPORT is not set
+
+#
+# Sound
+#
+CONFIG_SOUND=y
+
+#
+# Advanced Linux Sound Architecture
+#
+CONFIG_SND=y
+CONFIG_SND_TIMER=y
+CONFIG_SND_PCM=y
+CONFIG_SND_HWDEP=y
+CONFIG_SND_RAWMIDI=y
+CONFIG_SND_SEQUENCER=y
+CONFIG_SND_SEQ_DUMMY=y
+CONFIG_SND_OSSEMUL=y
+CONFIG_SND_MIXER_OSS=y
+CONFIG_SND_PCM_OSS=y
+CONFIG_SND_PCM_OSS_PLUGINS=y
+CONFIG_SND_SEQUENCER_OSS=y
+# CONFIG_SND_DYNAMIC_MINORS is not set
+CONFIG_SND_SUPPORT_OLD_API=y
+CONFIG_SND_VERBOSE_PROCFS=y
+# CONFIG_SND_VERBOSE_PRINTK is not set
+# CONFIG_SND_DEBUG is not set
+
+#
+# Generic devices
+#
+CONFIG_SND_MPU401_UART=y
+CONFIG_SND_OPL3_LIB=y
+CONFIG_SND_AC97_CODEC=y
+CONFIG_SND_AC97_BUS=y
+CONFIG_SND_DUMMY=y
+CONFIG_SND_VIRMIDI=y
+# CONFIG_SND_MTPAV is not set
+CONFIG_SND_SERIAL_U16550=y
+CONFIG_SND_MPU401=y
+
+#
+# PCI devices
+#
+# CONFIG_SND_AD1889 is not set
+# CONFIG_SND_ALS300 is not set
+# CONFIG_SND_ALI5451 is not set
+CONFIG_SND_ATIIXP=y
+# CONFIG_SND_ATIIXP_MODEM is not set
+# CONFIG_SND_AU8810 is not set
+# CONFIG_SND_AU8820 is not set
+# CONFIG_SND_AU8830 is not set
+# CONFIG_SND_AZT3328 is not set
+# CONFIG_SND_BT87X is not set
+# CONFIG_SND_CA0106 is not set
+# CONFIG_SND_CMIPCI is not set
+# CONFIG_SND_CS4281 is not set
+# CONFIG_SND_CS46XX is not set
+# CONFIG_SND_DARLA20 is not set
+# CONFIG_SND_GINA20 is not set
+# CONFIG_SND_LAYLA20 is not set
+# CONFIG_SND_DARLA24 is not set
+# CONFIG_SND_GINA24 is not set
+# CONFIG_SND_LAYLA24 is not set
+# CONFIG_SND_MONA is not set
+# CONFIG_SND_MIA is not set
+# CONFIG_SND_ECHO3G is not set
+# CONFIG_SND_INDIGO is not set
+# CONFIG_SND_INDIGOIO is not set
+# CONFIG_SND_INDIGODJ is not set
+# CONFIG_SND_EMU10K1 is not set
+# CONFIG_SND_EMU10K1X is not set
+# CONFIG_SND_ENS1370 is not set
+# CONFIG_SND_ENS1371 is not set
+# CONFIG_SND_ES1938 is not set
+# CONFIG_SND_ES1968 is not set
+CONFIG_SND_FM801=y
+# CONFIG_SND_FM801_TEA575X_BOOL is not set
+# CONFIG_SND_HDA_INTEL is not set
+# CONFIG_SND_HDSP is not set
+# CONFIG_SND_HDSPM is not set
+# CONFIG_SND_ICE1712 is not set
+# CONFIG_SND_ICE1724 is not set
+# CONFIG_SND_INTEL8X0 is not set
+# CONFIG_SND_INTEL8X0M is not set
+# CONFIG_SND_KORG1212 is not set
+# CONFIG_SND_MAESTRO3 is not set
+# CONFIG_SND_MIXART is not set
+# CONFIG_SND_NM256 is not set
+# CONFIG_SND_PCXHR is not set
+# CONFIG_SND_RIPTIDE is not set
+# CONFIG_SND_RME32 is not set
+# CONFIG_SND_RME96 is not set
+# CONFIG_SND_RME9652 is not set
+# CONFIG_SND_SONICVIBES is not set
+# CONFIG_SND_TRIDENT is not set
+# CONFIG_SND_VIA82XX is not set
+# CONFIG_SND_VIA82XX_MODEM is not set
+# CONFIG_SND_VX222 is not set
+# CONFIG_SND_YMFPCI is not set
+
+#
+# USB devices
+#
+# CONFIG_SND_USB_AUDIO is not set
+
+#
+# Open Sound System
+#
+CONFIG_SOUND_PRIME=y
+# CONFIG_OSS_OBSOLETE_DRIVER is not set
+# CONFIG_SOUND_BT878 is not set
+# CONFIG_SOUND_ES1371 is not set
+# CONFIG_SOUND_ICH is not set
+# CONFIG_SOUND_TRIDENT is not set
+# CONFIG_SOUND_MSNDCLAS is not set
+# CONFIG_SOUND_MSNDPIN is not set
+# CONFIG_SOUND_VIA82CXXX is not set
+# CONFIG_SOUND_TVMIXER is not set
+
+#
+# USB support
+#
+CONFIG_USB_ARCH_HAS_HCD=y
+CONFIG_USB_ARCH_HAS_OHCI=y
+CONFIG_USB_ARCH_HAS_EHCI=y
+CONFIG_USB=y
+# CONFIG_USB_DEBUG is not set
+
+#
+# Miscellaneous USB options
+#
+CONFIG_USB_DEVICEFS=y
+CONFIG_USB_BANDWIDTH=y
+# CONFIG_USB_DYNAMIC_MINORS is not set
+# CONFIG_USB_SUSPEND is not set
+# CONFIG_USB_OTG is not set
+
+#
+# USB Host Controller Drivers
+#
+CONFIG_USB_EHCI_HCD=y
+# CONFIG_USB_EHCI_SPLIT_ISO is not set
+# CONFIG_USB_EHCI_ROOT_HUB_TT is not set
+# CONFIG_USB_EHCI_TT_NEWSCHED is not set
+# CONFIG_USB_ISP116X_HCD is not set
+CONFIG_USB_OHCI_HCD=y
+# CONFIG_USB_OHCI_BIG_ENDIAN is not set
+CONFIG_USB_OHCI_LITTLE_ENDIAN=y
+CONFIG_USB_UHCI_HCD=y
+# CONFIG_USB_SL811_HCD is not set
+
+#
+# USB Device Class drivers
+#
+# CONFIG_USB_ACM is not set
+# CONFIG_USB_PRINTER is not set
+
+#
+# NOTE: USB_STORAGE enables SCSI, and 'SCSI disk support'
+#
+
+#
+# may also be needed; see USB_STORAGE Help for more information
+#
+CONFIG_USB_STORAGE=y
+# CONFIG_USB_STORAGE_DEBUG is not set
+# CONFIG_USB_STORAGE_DATAFAB is not set
+# CONFIG_USB_STORAGE_FREECOM is not set
+# CONFIG_USB_STORAGE_ISD200 is not set
+# CONFIG_USB_STORAGE_DPCM is not set
+# CONFIG_USB_STORAGE_USBAT is not set
+# CONFIG_USB_STORAGE_SDDR09 is not set
+# CONFIG_USB_STORAGE_SDDR55 is not set
+# CONFIG_USB_STORAGE_JUMPSHOT is not set
+# CONFIG_USB_STORAGE_ALAUDA is not set
+# CONFIG_USB_LIBUSUAL is not set
+
+#
+# USB Input Devices
+#
+CONFIG_USB_HID=y
+CONFIG_USB_HIDINPUT=y
+# CONFIG_USB_HIDINPUT_POWERBOOK is not set
+# CONFIG_HID_FF is not set
+CONFIG_USB_HIDDEV=y
+# CONFIG_USB_AIPTEK is not set
+# CONFIG_USB_WACOM is not set
+# CONFIG_USB_ACECAD is not set
+# CONFIG_USB_KBTAB is not set
+# CONFIG_USB_POWERMATE is not set
+# CONFIG_USB_TOUCHSCREEN is not set
+# CONFIG_USB_YEALINK is not set
+# CONFIG_USB_XPAD is not set
+# CONFIG_USB_ATI_REMOTE is not set
+# CONFIG_USB_ATI_REMOTE2 is not set
+# CONFIG_USB_KEYSPAN_REMOTE is not set
+# CONFIG_USB_APPLETOUCH is not set
+
+#
+# USB Imaging devices
+#
+# CONFIG_USB_MDC800 is not set
+# CONFIG_USB_MICROTEK is not set
+
+#
+# USB Network Adapters
+#
+# CONFIG_USB_CATC is not set
+# CONFIG_USB_KAWETH is not set
+# CONFIG_USB_PEGASUS is not set
+# CONFIG_USB_RTL8150 is not set
+# CONFIG_USB_USBNET is not set
+CONFIG_USB_MON=y
+
+#
+# USB port drivers
+#
+
+#
+# USB Serial Converter support
+#
+# CONFIG_USB_SERIAL is not set
+
+#
+# USB Miscellaneous drivers
+#
+# CONFIG_USB_EMI62 is not set
+# CONFIG_USB_EMI26 is not set
+# CONFIG_USB_AUERSWALD is not set
+# CONFIG_USB_RIO500 is not set
+# CONFIG_USB_LEGOTOWER is not set
+# CONFIG_USB_LCD is not set
+# CONFIG_USB_LED is not set
+# CONFIG_USB_CYPRESS_CY7C63 is not set
+# CONFIG_USB_CYTHERM is not set
+# CONFIG_USB_PHIDGETKIT is not set
+# CONFIG_USB_PHIDGETSERVO is not set
+# CONFIG_USB_IDMOUSE is not set
+# CONFIG_USB_APPLEDISPLAY is not set
+# CONFIG_USB_SISUSBVGA is not set
+# CONFIG_USB_LD is not set
+# CONFIG_USB_TEST is not set
+
+#
+# USB DSL modem support
+#
+
+#
+# USB Gadget Support
+#
+# CONFIG_USB_GADGET is not set
+
+#
+# MMC/SD Card support
+#
+# CONFIG_MMC is not set
+
+#
+# LED devices
+#
+# CONFIG_NEW_LEDS is not set
+
+#
+# LED drivers
+#
+
+#
+# LED Triggers
+#
+
+#
+# InfiniBand support
+#
+# CONFIG_INFINIBAND is not set
+
+#
+# EDAC - error detection and reporting (RAS) (EXPERIMENTAL)
+#
+
+#
+# Real Time Clock
+#
+# CONFIG_RTC_CLASS is not set
+
+#
+# DMA Engine support
+#
+# CONFIG_DMA_ENGINE is not set
+
+#
+# DMA Clients
+#
+
+#
+# DMA Devices
+#
+
+#
+# File systems
+#
+CONFIG_EXT2_FS=y
+CONFIG_EXT2_FS_XATTR=y
+CONFIG_EXT2_FS_POSIX_ACL=y
+CONFIG_EXT2_FS_SECURITY=y
+# CONFIG_EXT2_FS_XIP is not set
+CONFIG_EXT3_FS=y
+CONFIG_EXT3_FS_XATTR=y
+CONFIG_EXT3_FS_POSIX_ACL=y
+CONFIG_EXT3_FS_SECURITY=y
+CONFIG_JBD=y
+# CONFIG_JBD_DEBUG is not set
+CONFIG_FS_MBCACHE=y
+CONFIG_REISERFS_FS=y
+# CONFIG_REISERFS_CHECK is not set
+# CONFIG_REISERFS_PROC_INFO is not set
+CONFIG_REISERFS_FS_XATTR=y
+CONFIG_REISERFS_FS_POSIX_ACL=y
+CONFIG_REISERFS_FS_SECURITY=y
+# CONFIG_JFS_FS is not set
+CONFIG_FS_POSIX_ACL=y
+CONFIG_XFS_FS=y
+# CONFIG_XFS_QUOTA is not set
+# CONFIG_XFS_SECURITY is not set
+# CONFIG_XFS_POSIX_ACL is not set
+# CONFIG_XFS_RT is not set
+# CONFIG_OCFS2_FS is not set
+# CONFIG_MINIX_FS is not set
+# CONFIG_ROMFS_FS is not set
+CONFIG_INOTIFY=y
+CONFIG_INOTIFY_USER=y
+# CONFIG_QUOTA is not set
+CONFIG_DNOTIFY=y
+CONFIG_AUTOFS_FS=y
+CONFIG_AUTOFS4_FS=y
+# CONFIG_FUSE_FS is not set
+
+#
+# CD-ROM/DVD Filesystems
+#
+CONFIG_ISO9660_FS=y
+CONFIG_JOLIET=y
+# CONFIG_ZISOFS is not set
+CONFIG_UDF_FS=y
+CONFIG_UDF_NLS=y
+
+#
+# DOS/FAT/NT Filesystems
+#
+CONFIG_FAT_FS=y
+CONFIG_MSDOS_FS=y
+CONFIG_VFAT_FS=y
+CONFIG_FAT_DEFAULT_CODEPAGE=437
+CONFIG_FAT_DEFAULT_IOCHARSET="iso8859-1"
+# CONFIG_NTFS_FS is not set
+
+#
+# Pseudo filesystems
+#
+CONFIG_PROC_FS=y
+CONFIG_PROC_KCORE=y
+CONFIG_PROC_IOMEM_MACHINE=y
+CONFIG_SYSFS=y
+CONFIG_TMPFS=y
+# CONFIG_HUGETLB_PAGE is not set
+CONFIG_RAMFS=y
+# CONFIG_CONFIGFS_FS is not set
+
+#
+# Miscellaneous filesystems
+#
+# CONFIG_ADFS_FS is not set
+# CONFIG_AFFS_FS is not set
+# CONFIG_HFS_FS is not set
+# CONFIG_HFSPLUS_FS is not set
+# CONFIG_BEFS_FS is not set
+# CONFIG_BFS_FS is not set
+# CONFIG_EFS_FS is not set
+# CONFIG_CRAMFS is not set
+# CONFIG_VXFS_FS is not set
+# CONFIG_HPFS_FS is not set
+# CONFIG_QNX4FS_FS is not set
+# CONFIG_SYSV_FS is not set
+# CONFIG_UFS_FS is not set
+
+#
+# Network File Systems
+#
+CONFIG_NFS_FS=y
+CONFIG_NFS_V3=y
+# CONFIG_NFS_V3_ACL is not set
+CONFIG_NFS_V4=y
+CONFIG_NFS_DIRECTIO=y
+CONFIG_NFSD=y
+CONFIG_NFSD_V3=y
+# CONFIG_NFSD_V3_ACL is not set
+CONFIG_NFSD_V4=y
+CONFIG_NFSD_TCP=y
+CONFIG_ROOT_NFS=y
+CONFIG_LOCKD=y
+CONFIG_LOCKD_V4=y
+CONFIG_EXPORTFS=y
+CONFIG_NFS_COMMON=y
+CONFIG_SUNRPC=y
+CONFIG_SUNRPC_GSS=y
+CONFIG_RPCSEC_GSS_KRB5=y
+# CONFIG_RPCSEC_GSS_SPKM3 is not set
+CONFIG_SMB_FS=y
+CONFIG_SMB_NLS_DEFAULT=y
+CONFIG_SMB_NLS_REMOTE="cp437"
+CONFIG_CIFS=y
+# CONFIG_CIFS_STATS is not set
+# CONFIG_CIFS_WEAK_PW_HASH is not set
+# CONFIG_CIFS_XATTR is not set
+# CONFIG_CIFS_DEBUG2 is not set
+# CONFIG_CIFS_EXPERIMENTAL is not set
+# CONFIG_NCP_FS is not set
+# CONFIG_CODA_FS is not set
+# CONFIG_AFS_FS is not set
+# CONFIG_9P_FS is not set
+
+#
+# Partition Types
+#
+CONFIG_PARTITION_ADVANCED=y
+# CONFIG_ACORN_PARTITION is not set
+# CONFIG_OSF_PARTITION is not set
+# CONFIG_AMIGA_PARTITION is not set
+# CONFIG_ATARI_PARTITION is not set
+# CONFIG_MAC_PARTITION is not set
+CONFIG_MSDOS_PARTITION=y
+# CONFIG_BSD_DISKLABEL is not set
+# CONFIG_MINIX_SUBPARTITION is not set
+# CONFIG_SOLARIS_X86_PARTITION is not set
+# CONFIG_UNIXWARE_DISKLABEL is not set
+# CONFIG_LDM_PARTITION is not set
+CONFIG_SGI_PARTITION=y
+# CONFIG_ULTRIX_PARTITION is not set
+# CONFIG_SUN_PARTITION is not set
+# CONFIG_KARMA_PARTITION is not set
+CONFIG_EFI_PARTITION=y
+
+#
+# Native Language Support
+#
+CONFIG_NLS=y
+CONFIG_NLS_DEFAULT="iso8859-1"
+CONFIG_NLS_CODEPAGE_437=y
+CONFIG_NLS_CODEPAGE_737=y
+CONFIG_NLS_CODEPAGE_775=y
+CONFIG_NLS_CODEPAGE_850=y
+CONFIG_NLS_CODEPAGE_852=y
+CONFIG_NLS_CODEPAGE_855=y
+CONFIG_NLS_CODEPAGE_857=y
+CONFIG_NLS_CODEPAGE_860=y
+CONFIG_NLS_CODEPAGE_861=y
+CONFIG_NLS_CODEPAGE_862=y
+CONFIG_NLS_CODEPAGE_863=y
+CONFIG_NLS_CODEPAGE_864=y
+CONFIG_NLS_CODEPAGE_865=y
+CONFIG_NLS_CODEPAGE_866=y
+CONFIG_NLS_CODEPAGE_869=y
+CONFIG_NLS_CODEPAGE_936=y
+CONFIG_NLS_CODEPAGE_950=y
+CONFIG_NLS_CODEPAGE_932=y
+CONFIG_NLS_CODEPAGE_949=y
+CONFIG_NLS_CODEPAGE_874=y
+CONFIG_NLS_ISO8859_8=y
+# CONFIG_NLS_CODEPAGE_1250 is not set
+CONFIG_NLS_CODEPAGE_1251=y
+# CONFIG_NLS_ASCII is not set
+CONFIG_NLS_ISO8859_1=y
+CONFIG_NLS_ISO8859_2=y
+CONFIG_NLS_ISO8859_3=y
+CONFIG_NLS_ISO8859_4=y
+CONFIG_NLS_ISO8859_5=y
+CONFIG_NLS_ISO8859_6=y
+CONFIG_NLS_ISO8859_7=y
+CONFIG_NLS_ISO8859_9=y
+CONFIG_NLS_ISO8859_13=y
+CONFIG_NLS_ISO8859_14=y
+CONFIG_NLS_ISO8859_15=y
+CONFIG_NLS_KOI8_R=y
+CONFIG_NLS_KOI8_U=y
+CONFIG_NLS_UTF8=y
+
+#
+# Library routines
+#
+# CONFIG_CRC_CCITT is not set
+# CONFIG_CRC16 is not set
+CONFIG_CRC32=y
+# CONFIG_LIBCRC32C is not set
+CONFIG_PLIST=y
+CONFIG_GENERIC_HARDIRQS=y
+CONFIG_GENERIC_IRQ_PROBE=y
+CONFIG_GENERIC_PENDING_IRQ=y
+CONFIG_IRQ_PER_CPU=y
+
+#
+# HP Simulator drivers
+#
+# CONFIG_HP_SIMETH is not set
+# CONFIG_HP_SIMSERIAL is not set
+# CONFIG_HP_SIMSCSI is not set
+
+#
+# Instrumentation Support
+#
+# CONFIG_PROFILING is not set
+# CONFIG_KPROBES is not set
+
+#
+# Kernel hacking
+#
+# CONFIG_PRINTK_TIME is not set
+CONFIG_MAGIC_SYSRQ=y
+CONFIG_UNUSED_SYMBOLS=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_LOG_BUF_SHIFT=20
+CONFIG_DETECT_SOFTLOCKUP=y
+# CONFIG_SCHEDSTATS is not set
+# CONFIG_DEBUG_SLAB is not set
+# CONFIG_DEBUG_RT_MUTEXES is not set
+# CONFIG_RT_MUTEX_TESTER is not set
+# CONFIG_DEBUG_SPINLOCK is not set
+CONFIG_DEBUG_MUTEXES=y
+# CONFIG_DEBUG_RWSEMS is not set
+# CONFIG_DEBUG_SPINLOCK_SLEEP is not set
+# CONFIG_DEBUG_LOCKING_API_SELFTESTS is not set
+# CONFIG_DEBUG_KOBJECT is not set
+# CONFIG_DEBUG_INFO is not set
+# CONFIG_DEBUG_FS is not set
+# CONFIG_DEBUG_VM is not set
+CONFIG_FORCED_INLINING=y
+# CONFIG_RCU_TORTURE_TEST is not set
+CONFIG_IA64_GRANULE_16MB=y
+# CONFIG_IA64_GRANULE_64MB is not set
+CONFIG_IA64_PRINT_HAZARDS=y
+# CONFIG_DISABLE_VHPT is not set
+# CONFIG_IA64_DEBUG_CMPXCHG is not set
+# CONFIG_IA64_DEBUG_IRQ is not set
+
+#
+# Security options
+#
+# CONFIG_KEYS is not set
+# CONFIG_SECURITY is not set
+
+#
+# Cryptographic options
+#
+CONFIG_CRYPTO=y
+# CONFIG_CRYPTO_HMAC is not set
+# CONFIG_CRYPTO_NULL is not set
+# CONFIG_CRYPTO_MD4 is not set
+CONFIG_CRYPTO_MD5=y
+# CONFIG_CRYPTO_SHA1 is not set
+# CONFIG_CRYPTO_SHA256 is not set
+# CONFIG_CRYPTO_SHA512 is not set
+# CONFIG_CRYPTO_WP512 is not set
+# CONFIG_CRYPTO_TGR192 is not set
+CONFIG_CRYPTO_DES=y
+# CONFIG_CRYPTO_BLOWFISH is not set
+# CONFIG_CRYPTO_TWOFISH is not set
+# CONFIG_CRYPTO_SERPENT is not set
+# CONFIG_CRYPTO_AES is not set
+# CONFIG_CRYPTO_CAST5 is not set
+# CONFIG_CRYPTO_CAST6 is not set
+# CONFIG_CRYPTO_TEA is not set
+# CONFIG_CRYPTO_ARC4 is not set
+# CONFIG_CRYPTO_KHAZAD is not set
+# CONFIG_CRYPTO_ANUBIS is not set
+# CONFIG_CRYPTO_DEFLATE is not set
+# CONFIG_CRYPTO_MICHAEL_MIC is not set
+# CONFIG_CRYPTO_CRC32C is not set
+# CONFIG_CRYPTO_TEST is not set
+
+#
+# Hardware crypto devices
+#
+# CONFIG_XEN_SMPBOOT is not set
+# CONFIG_XEN_DEVMEM is not set
+CONFIG_XEN_INTERFACE_VERSION=0x00030207
+
+#
+# XEN
+#
+CONFIG_XEN_PRIVILEGED_GUEST=y
+# CONFIG_XEN_UNPRIVILEGED_GUEST is not set
+CONFIG_XEN_PRIVCMD=y
+CONFIG_XEN_XENBUS_DEV=y
+CONFIG_XEN_BACKEND=y
+CONFIG_XEN_BLKDEV_BACKEND=y
+CONFIG_XEN_BLKDEV_TAP=y
+CONFIG_XEN_NETDEV_BACKEND=y
+# CONFIG_XEN_NETDEV_PIPELINED_TRANSMITTER is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_UTIL is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_BACKEND is not set
+# CONFIG_XEN_NETDEV_LOOPBACK is not set
+CONFIG_XEN_PCIDEV_BACKEND=y
+# CONFIG_XEN_PCIDEV_BACKEND_VPCI is not set
+# CONFIG_XEN_PCIDEV_BACKEND_PASS is not set
+# CONFIG_XEN_PCIDEV_BACKEND_SLOT is not set
+CONFIG_XEN_PCIDEV_BACKEND_CONTROLLER=y
+# CONFIG_XEN_PCIDEV_BE_DEBUG is not set
+CONFIG_XEN_TPMDEV_BACKEND=m
+CONFIG_XEN_SCSI_BACKEND=m
+CONFIG_XEN_BLKDEV_FRONTEND=y
+CONFIG_XEN_NETDEV_FRONTEND=y
+CONFIG_XEN_SCSI_FRONTEND=m
+CONFIG_XEN_GRANT_DEV=y
+# CONFIG_XEN_NETDEV_ACCEL_SFC_FRONTEND is not set
+CONFIG_XEN_FRAMEBUFFER=y
+CONFIG_XEN_KEYBOARD=y
+# CONFIG_XEN_SCRUB_PAGES is not set
+CONFIG_XEN_DISABLE_SERIAL=y
+CONFIG_XEN_SYSFS=y
+CONFIG_XEN_COMPAT_030002_AND_LATER=y
+# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
+# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
+CONFIG_XEN_COMPAT=0x030002
+CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
+CONFIG_NO_IDLE_HZ=y
+CONFIG_XEN_BALLOON=y
+CONFIG_XEN_XENCOMM=y
diff -rpuN linux-2.6.18.8/buildconfigs/linux-defconfig_xenU_ia64 linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xenU_ia64
--- linux-2.6.18.8/buildconfigs/linux-defconfig_xenU_ia64	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xenU_ia64	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1511 @@
+#
+# Automatically generated make config: don't edit
+# Linux kernel version: 2.6.18.8
+# Tue Feb 19 11:20:14 2008
+#
+CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
+
+#
+# Code maturity level options
+#
+CONFIG_EXPERIMENTAL=y
+CONFIG_LOCK_KERNEL=y
+CONFIG_INIT_ENV_ARG_LIMIT=32
+
+#
+# General setup
+#
+CONFIG_LOCALVERSION=""
+CONFIG_LOCALVERSION_AUTO=y
+CONFIG_SWAP=y
+CONFIG_SYSVIPC=y
+# CONFIG_POSIX_MQUEUE is not set
+CONFIG_BSD_PROCESS_ACCT=y
+# CONFIG_BSD_PROCESS_ACCT_V3 is not set
+# CONFIG_TASKSTATS is not set
+# CONFIG_AUDIT is not set
+# CONFIG_IKCONFIG is not set
+# CONFIG_CPUSETS is not set
+# CONFIG_RELAY is not set
+CONFIG_INITRAMFS_SOURCE=""
+CONFIG_CC_OPTIMIZE_FOR_SIZE=y
+# CONFIG_EMBEDDED is not set
+CONFIG_SYSCTL=y
+CONFIG_KALLSYMS=y
+# CONFIG_KALLSYMS_ALL is not set
+# CONFIG_KALLSYMS_EXTRA_PASS is not set
+CONFIG_HOTPLUG=y
+CONFIG_PRINTK=y
+CONFIG_BUG=y
+CONFIG_ELF_CORE=y
+CONFIG_BASE_FULL=y
+CONFIG_FUTEX=y
+CONFIG_EPOLL=y
+CONFIG_SHMEM=y
+CONFIG_SLAB=y
+CONFIG_VM_EVENT_COUNTERS=y
+CONFIG_RT_MUTEXES=y
+# CONFIG_TINY_SHMEM is not set
+CONFIG_BASE_SMALL=0
+# CONFIG_SLOB is not set
+
+#
+# Loadable module support
+#
+CONFIG_MODULES=y
+# CONFIG_MODULE_UNLOAD is not set
+# CONFIG_MODVERSIONS is not set
+# CONFIG_MODULE_SRCVERSION_ALL is not set
+# CONFIG_KMOD is not set
+CONFIG_STOP_MACHINE=y
+
+#
+# Block layer
+#
+# CONFIG_BLK_DEV_IO_TRACE is not set
+
+#
+# IO Schedulers
+#
+CONFIG_IOSCHED_NOOP=y
+CONFIG_IOSCHED_AS=y
+CONFIG_IOSCHED_DEADLINE=y
+CONFIG_IOSCHED_CFQ=y
+CONFIG_DEFAULT_AS=y
+# CONFIG_DEFAULT_DEADLINE is not set
+# CONFIG_DEFAULT_CFQ is not set
+# CONFIG_DEFAULT_NOOP is not set
+CONFIG_DEFAULT_IOSCHED="anticipatory"
+
+#
+# Processor type and features
+#
+CONFIG_IA64=y
+CONFIG_64BIT=y
+CONFIG_MMU=y
+CONFIG_SWIOTLB=y
+CONFIG_RWSEM_XCHGADD_ALGORITHM=y
+CONFIG_GENERIC_FIND_NEXT_BIT=y
+CONFIG_GENERIC_CALIBRATE_DELAY=y
+CONFIG_TIME_INTERPOLATION=y
+CONFIG_DMI=y
+CONFIG_EFI=y
+CONFIG_GENERIC_IOMAP=y
+CONFIG_XEN=y
+CONFIG_XEN_IA64_EXPOSE_P2M=y
+CONFIG_XEN_IA64_EXPOSE_P2M_USE_DTR=y
+CONFIG_SCHED_NO_NO_OMIT_FRAME_POINTER=y
+CONFIG_DMA_IS_DMA32=y
+CONFIG_AUDIT_ARCH=y
+# CONFIG_IA64_GENERIC is not set
+# CONFIG_IA64_DIG is not set
+# CONFIG_IA64_HP_ZX1 is not set
+# CONFIG_IA64_HP_ZX1_SWIOTLB is not set
+# CONFIG_IA64_SGI_SN2 is not set
+# CONFIG_IA64_HP_SIM is not set
+CONFIG_IA64_XEN=y
+# CONFIG_ITANIUM is not set
+CONFIG_MCKINLEY=y
+# CONFIG_IA64_PAGE_SIZE_4KB is not set
+# CONFIG_IA64_PAGE_SIZE_8KB is not set
+CONFIG_IA64_PAGE_SIZE_16KB=y
+# CONFIG_IA64_PAGE_SIZE_64KB is not set
+CONFIG_PGTABLE_3=y
+# CONFIG_PGTABLE_4 is not set
+CONFIG_HZ_100=y
+# CONFIG_HZ_250 is not set
+# CONFIG_HZ_1000 is not set
+CONFIG_HZ=100
+CONFIG_IA64_L1_CACHE_SHIFT=7
+# CONFIG_IA64_CYCLONE is not set
+CONFIG_IOSAPIC=y
+CONFIG_FORCE_MAX_ZONEORDER=11
+CONFIG_SMP=y
+CONFIG_NR_CPUS=16
+CONFIG_HOTPLUG_CPU=y
+CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y
+# CONFIG_SCHED_SMT is not set
+# CONFIG_PERMIT_BSP_REMOVE is not set
+# CONFIG_PREEMPT is not set
+CONFIG_SELECT_MEMORY_MODEL=y
+CONFIG_FLATMEM_MANUAL=y
+# CONFIG_DISCONTIGMEM_MANUAL is not set
+# CONFIG_SPARSEMEM_MANUAL is not set
+CONFIG_FLATMEM=y
+CONFIG_FLAT_NODE_MEM_MAP=y
+# CONFIG_SPARSEMEM_STATIC is not set
+CONFIG_SPLIT_PTLOCK_CPUS=4
+CONFIG_RESOURCES_64BIT=y
+CONFIG_ARCH_SELECT_MEMORY_MODEL=y
+CONFIG_ARCH_DISCONTIGMEM_ENABLE=y
+CONFIG_ARCH_FLATMEM_ENABLE=y
+CONFIG_ARCH_SPARSEMEM_ENABLE=y
+CONFIG_VIRTUAL_MEM_MAP=y
+CONFIG_HOLES_IN_ZONE=y
+# CONFIG_IA32_SUPPORT is not set
+CONFIG_IA64_MCA_RECOVERY=y
+CONFIG_PERFMON=y
+CONFIG_IA64_PALINFO=y
+# CONFIG_KEXEC is not set
+# CONFIG_CRASH_DUMP is not set
+
+#
+# Firmware Drivers
+#
+CONFIG_EFI_VARS=y
+CONFIG_EFI_PCDP=y
+CONFIG_BINFMT_ELF=y
+CONFIG_BINFMT_MISC=y
+
+#
+# Power management and ACPI
+#
+CONFIG_PM=y
+CONFIG_PM_LEGACY=y
+# CONFIG_PM_DEBUG is not set
+
+#
+# ACPI (Advanced Configuration and Power Interface) Support
+#
+CONFIG_ACPI=y
+CONFIG_ACPI_BUTTON=y
+CONFIG_ACPI_FAN=y
+# CONFIG_ACPI_DOCK is not set
+CONFIG_ACPI_PROCESSOR=y
+CONFIG_ACPI_HOTPLUG_CPU=y
+CONFIG_ACPI_THERMAL=y
+CONFIG_ACPI_BLACKLIST_YEAR=0
+# CONFIG_ACPI_DEBUG is not set
+CONFIG_ACPI_EC=y
+CONFIG_ACPI_POWER=y
+CONFIG_ACPI_SYSTEM=y
+CONFIG_ACPI_CONTAINER=y
+
+#
+# CPU Frequency scaling
+#
+# CONFIG_CPU_FREQ is not set
+
+#
+# Bus options (PCI, PCMCIA)
+#
+CONFIG_PCI=y
+CONFIG_PCI_DOMAINS=y
+CONFIG_XEN_PCIDEV_FRONTEND=y
+# CONFIG_XEN_PCIDEV_FE_DEBUG is not set
+# CONFIG_PCIEPORTBUS is not set
+# CONFIG_PCI_DEBUG is not set
+
+#
+# PCI Hotplug Support
+#
+CONFIG_HOTPLUG_PCI=y
+# CONFIG_HOTPLUG_PCI_FAKE is not set
+CONFIG_HOTPLUG_PCI_ACPI=y
+# CONFIG_HOTPLUG_PCI_ACPI_IBM is not set
+# CONFIG_HOTPLUG_PCI_CPCI is not set
+# CONFIG_HOTPLUG_PCI_SHPC is not set
+
+#
+# PCCARD (PCMCIA/CardBus) support
+#
+# CONFIG_PCCARD is not set
+
+#
+# Networking
+#
+CONFIG_NET=y
+
+#
+# Networking options
+#
+# CONFIG_NETDEBUG is not set
+CONFIG_PACKET=y
+# CONFIG_PACKET_MMAP is not set
+CONFIG_UNIX=y
+CONFIG_XFRM=y
+# CONFIG_XFRM_USER is not set
+# CONFIG_NET_KEY is not set
+CONFIG_INET=y
+CONFIG_IP_MULTICAST=y
+# CONFIG_IP_ADVANCED_ROUTER is not set
+CONFIG_IP_FIB_HASH=y
+CONFIG_IP_PNP=y
+CONFIG_IP_PNP_DHCP=y
+# CONFIG_IP_PNP_BOOTP is not set
+# CONFIG_IP_PNP_RARP is not set
+# CONFIG_NET_IPIP is not set
+# CONFIG_NET_IPGRE is not set
+# CONFIG_IP_MROUTE is not set
+# CONFIG_ARPD is not set
+# CONFIG_SYN_COOKIES is not set
+# CONFIG_INET_AH is not set
+# CONFIG_INET_ESP is not set
+# CONFIG_INET_IPCOMP is not set
+# CONFIG_INET_XFRM_TUNNEL is not set
+# CONFIG_INET_TUNNEL is not set
+CONFIG_INET_XFRM_MODE_TRANSPORT=y
+CONFIG_INET_XFRM_MODE_TUNNEL=y
+CONFIG_INET_DIAG=y
+CONFIG_INET_TCP_DIAG=y
+# CONFIG_TCP_CONG_ADVANCED is not set
+CONFIG_TCP_CONG_BIC=y
+
+#
+# IP: Virtual Server Configuration
+#
+# CONFIG_IP_VS is not set
+# CONFIG_IPV6 is not set
+# CONFIG_INET6_XFRM_TUNNEL is not set
+# CONFIG_INET6_TUNNEL is not set
+# CONFIG_NETWORK_SECMARK is not set
+CONFIG_NETFILTER=y
+# CONFIG_NETFILTER_DEBUG is not set
+
+#
+# Core Netfilter Configuration
+#
+# CONFIG_NETFILTER_NETLINK is not set
+# CONFIG_NF_CONNTRACK is not set
+# CONFIG_NETFILTER_XTABLES is not set
+
+#
+# IP: Netfilter Configuration
+#
+# CONFIG_IP_NF_CONNTRACK is not set
+# CONFIG_IP_NF_QUEUE is not set
+
+#
+# DCCP Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP_DCCP is not set
+
+#
+# SCTP Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP_SCTP is not set
+
+#
+# TIPC Configuration (EXPERIMENTAL)
+#
+# CONFIG_TIPC is not set
+# CONFIG_ATM is not set
+# CONFIG_BRIDGE is not set
+# CONFIG_VLAN_8021Q is not set
+# CONFIG_DECNET is not set
+# CONFIG_LLC2 is not set
+# CONFIG_IPX is not set
+# CONFIG_ATALK is not set
+# CONFIG_X25 is not set
+# CONFIG_LAPB is not set
+# CONFIG_ECONET is not set
+# CONFIG_WAN_ROUTER is not set
+
+#
+# QoS and/or fair queueing
+#
+# CONFIG_NET_SCHED is not set
+
+#
+# Network testing
+#
+# CONFIG_NET_PKTGEN is not set
+# CONFIG_HAMRADIO is not set
+# CONFIG_IRDA is not set
+# CONFIG_BT is not set
+# CONFIG_IEEE80211 is not set
+
+#
+# Device Drivers
+#
+
+#
+# Generic Driver Options
+#
+CONFIG_STANDALONE=y
+CONFIG_PREVENT_FIRMWARE_BUILD=y
+# CONFIG_FW_LOADER is not set
+# CONFIG_DEBUG_DRIVER is not set
+CONFIG_SYS_HYPERVISOR=y
+
+#
+# Connector - unified userspace <-> kernelspace linker
+#
+# CONFIG_CONNECTOR is not set
+
+#
+# Memory Technology Devices (MTD)
+#
+# CONFIG_MTD is not set
+
+#
+# Parallel port support
+#
+# CONFIG_PARPORT is not set
+
+#
+# Plug and Play support
+#
+# CONFIG_PNP is not set
+
+#
+# Block devices
+#
+# CONFIG_BLK_CPQ_DA is not set
+# CONFIG_BLK_CPQ_CISS_DA is not set
+# CONFIG_BLK_DEV_DAC960 is not set
+# CONFIG_BLK_DEV_UMEM is not set
+# CONFIG_BLK_DEV_COW_COMMON is not set
+CONFIG_BLK_DEV_LOOP=y
+# CONFIG_BLK_DEV_CRYPTOLOOP is not set
+# CONFIG_BLK_DEV_NBD is not set
+# CONFIG_BLK_DEV_SX8 is not set
+# CONFIG_BLK_DEV_UB is not set
+CONFIG_BLK_DEV_RAM=y
+CONFIG_BLK_DEV_RAM_COUNT=16
+CONFIG_BLK_DEV_RAM_SIZE=4096
+CONFIG_BLK_DEV_RAM_BLOCKSIZE=1024
+CONFIG_BLK_DEV_INITRD=y
+# CONFIG_CDROM_PKTCDVD is not set
+# CONFIG_ATA_OVER_ETH is not set
+
+#
+# ATA/ATAPI/MFM/RLL support
+#
+# CONFIG_IDE is not set
+
+#
+# SCSI device support
+#
+# CONFIG_RAID_ATTRS is not set
+CONFIG_SCSI=y
+CONFIG_SCSI_PROC_FS=y
+
+#
+# SCSI support type (disk, tape, CD-ROM)
+#
+CONFIG_BLK_DEV_SD=y
+CONFIG_CHR_DEV_ST=y
+CONFIG_CHR_DEV_OSST=y
+CONFIG_BLK_DEV_SR=y
+CONFIG_BLK_DEV_SR_VENDOR=y
+CONFIG_CHR_DEV_SG=y
+# CONFIG_CHR_DEV_SCH is not set
+
+#
+# Some SCSI devices (e.g. CD jukebox) support multiple LUNs
+#
+CONFIG_SCSI_MULTI_LUN=y
+CONFIG_SCSI_CONSTANTS=y
+CONFIG_SCSI_LOGGING=y
+
+#
+# SCSI Transport Attributes
+#
+CONFIG_SCSI_SPI_ATTRS=y
+# CONFIG_SCSI_FC_ATTRS is not set
+# CONFIG_SCSI_ISCSI_ATTRS is not set
+CONFIG_SCSI_SAS_ATTRS=y
+
+#
+# SCSI low-level drivers
+#
+# CONFIG_ISCSI_TCP is not set
+# CONFIG_BLK_DEV_3W_XXXX_RAID is not set
+# CONFIG_SCSI_3W_9XXX is not set
+# CONFIG_SCSI_ACARD is not set
+# CONFIG_SCSI_AACRAID is not set
+# CONFIG_SCSI_AIC7XXX is not set
+# CONFIG_SCSI_AIC7XXX_OLD is not set
+# CONFIG_SCSI_AIC79XX is not set
+# CONFIG_MEGARAID_NEWGEN is not set
+# CONFIG_MEGARAID_LEGACY is not set
+# CONFIG_MEGARAID_SAS is not set
+# CONFIG_SCSI_SATA is not set
+# CONFIG_SCSI_HPTIOP is not set
+# CONFIG_SCSI_DMX3191D is not set
+# CONFIG_SCSI_FUTURE_DOMAIN is not set
+# CONFIG_SCSI_IPS is not set
+# CONFIG_SCSI_INITIO is not set
+# CONFIG_SCSI_INIA100 is not set
+CONFIG_SCSI_SYM53C8XX_2=y
+CONFIG_SCSI_SYM53C8XX_DMA_ADDRESSING_MODE=1
+CONFIG_SCSI_SYM53C8XX_DEFAULT_TAGS=16
+CONFIG_SCSI_SYM53C8XX_MAX_TAGS=64
+CONFIG_SCSI_SYM53C8XX_MMIO=y
+# CONFIG_SCSI_IPR is not set
+CONFIG_SCSI_QLOGIC_1280=y
+# CONFIG_SCSI_QLA_FC is not set
+# CONFIG_SCSI_LPFC is not set
+# CONFIG_SCSI_DC395x is not set
+# CONFIG_SCSI_DC390T is not set
+# CONFIG_SCSI_DEBUG is not set
+
+#
+# Multi-device support (RAID and LVM)
+#
+# CONFIG_MD is not set
+
+#
+# Fusion MPT device support
+#
+CONFIG_FUSION=y
+CONFIG_FUSION_SPI=y
+# CONFIG_FUSION_FC is not set
+CONFIG_FUSION_SAS=y
+CONFIG_FUSION_MAX_SGE=128
+# CONFIG_FUSION_CTL is not set
+
+#
+# IEEE 1394 (FireWire) support
+#
+# CONFIG_IEEE1394 is not set
+
+#
+# I2O device support
+#
+# CONFIG_I2O is not set
+
+#
+# Network device support
+#
+CONFIG_NETDEVICES=y
+CONFIG_DUMMY=y
+# CONFIG_BONDING is not set
+# CONFIG_EQUALIZER is not set
+# CONFIG_TUN is not set
+
+#
+# ARCnet devices
+#
+# CONFIG_ARCNET is not set
+
+#
+# PHY device support
+#
+# CONFIG_PHYLIB is not set
+
+#
+# Ethernet (10 or 100Mbit)
+#
+CONFIG_NET_ETHERNET=y
+CONFIG_MII=y
+# CONFIG_HAPPYMEAL is not set
+# CONFIG_SUNGEM is not set
+# CONFIG_CASSINI is not set
+# CONFIG_NET_VENDOR_3COM is not set
+
+#
+# Tulip family network device support
+#
+CONFIG_NET_TULIP=y
+# CONFIG_DE2104X is not set
+CONFIG_TULIP=y
+CONFIG_TULIP_MWI=y
+CONFIG_TULIP_MMIO=y
+CONFIG_TULIP_NAPI=y
+CONFIG_TULIP_NAPI_HW_MITIGATION=y
+# CONFIG_DE4X5 is not set
+# CONFIG_WINBOND_840 is not set
+# CONFIG_DM9102 is not set
+# CONFIG_ULI526X is not set
+# CONFIG_HP100 is not set
+CONFIG_NET_PCI=y
+# CONFIG_PCNET32 is not set
+# CONFIG_AMD8111_ETH is not set
+# CONFIG_ADAPTEC_STARFIRE is not set
+# CONFIG_B44 is not set
+# CONFIG_FORCEDETH is not set
+# CONFIG_DGRS is not set
+# CONFIG_EEPRO100 is not set
+CONFIG_E100=y
+# CONFIG_FEALNX is not set
+# CONFIG_NATSEMI is not set
+# CONFIG_NE2K_PCI is not set
+# CONFIG_8139CP is not set
+# CONFIG_8139TOO is not set
+# CONFIG_SIS900 is not set
+# CONFIG_EPIC100 is not set
+# CONFIG_SUNDANCE is not set
+# CONFIG_VIA_RHINE is not set
+
+#
+# Ethernet (1000 Mbit)
+#
+# CONFIG_ACENIC is not set
+# CONFIG_DL2K is not set
+CONFIG_E1000=y
+# CONFIG_E1000_NAPI is not set
+# CONFIG_E1000_DISABLE_PACKET_SPLIT is not set
+# CONFIG_NS83820 is not set
+# CONFIG_HAMACHI is not set
+# CONFIG_YELLOWFIN is not set
+# CONFIG_R8169 is not set
+# CONFIG_SIS190 is not set
+# CONFIG_SKGE is not set
+# CONFIG_SKY2 is not set
+# CONFIG_SK98LIN is not set
+# CONFIG_VIA_VELOCITY is not set
+CONFIG_TIGON3=y
+# CONFIG_BNX2 is not set
+
+#
+# Ethernet (10000 Mbit)
+#
+# CONFIG_CHELSIO_T1 is not set
+# CONFIG_IXGB is not set
+# CONFIG_S2IO is not set
+# CONFIG_MYRI10GE is not set
+# CONFIG_SFC is not set
+
+#
+# Token Ring devices
+#
+# CONFIG_TR is not set
+
+#
+# Wireless LAN (non-hamradio)
+#
+# CONFIG_NET_RADIO is not set
+
+#
+# Wan interfaces
+#
+# CONFIG_WAN is not set
+# CONFIG_FDDI is not set
+# CONFIG_HIPPI is not set
+# CONFIG_PPP is not set
+# CONFIG_SLIP is not set
+# CONFIG_NET_FC is not set
+# CONFIG_SHAPER is not set
+# CONFIG_NETCONSOLE is not set
+# CONFIG_NETPOLL is not set
+# CONFIG_NET_POLL_CONTROLLER is not set
+
+#
+# ISDN subsystem
+#
+# CONFIG_ISDN is not set
+
+#
+# Telephony Support
+#
+# CONFIG_PHONE is not set
+
+#
+# Input device support
+#
+CONFIG_INPUT=y
+
+#
+# Userland interfaces
+#
+CONFIG_INPUT_MOUSEDEV=y
+CONFIG_INPUT_MOUSEDEV_PSAUX=y
+CONFIG_INPUT_MOUSEDEV_SCREEN_X=1024
+CONFIG_INPUT_MOUSEDEV_SCREEN_Y=768
+CONFIG_INPUT_JOYDEV=y
+# CONFIG_INPUT_TSDEV is not set
+CONFIG_INPUT_EVDEV=y
+# CONFIG_INPUT_EVBUG is not set
+
+#
+# Input Device Drivers
+#
+# CONFIG_INPUT_KEYBOARD is not set
+# CONFIG_INPUT_MOUSE is not set
+# CONFIG_INPUT_JOYSTICK is not set
+# CONFIG_INPUT_TOUCHSCREEN is not set
+# CONFIG_INPUT_MISC is not set
+
+#
+# Hardware I/O ports
+#
+CONFIG_SERIO=y
+# CONFIG_SERIO_I8042 is not set
+# CONFIG_SERIO_SERPORT is not set
+# CONFIG_SERIO_PCIPS2 is not set
+# CONFIG_SERIO_RAW is not set
+# CONFIG_GAMEPORT is not set
+
+#
+# Character devices
+#
+CONFIG_VT=y
+CONFIG_VT_CONSOLE=y
+CONFIG_HW_CONSOLE=y
+# CONFIG_VT_HW_CONSOLE_BINDING is not set
+# CONFIG_SERIAL_NONSTANDARD is not set
+
+#
+# Serial drivers
+#
+CONFIG_SERIAL_8250=y
+CONFIG_SERIAL_8250_CONSOLE=y
+CONFIG_SERIAL_8250_PCI=y
+CONFIG_SERIAL_8250_NR_UARTS=8
+CONFIG_SERIAL_8250_RUNTIME_UARTS=4
+CONFIG_SERIAL_8250_EXTENDED=y
+CONFIG_SERIAL_8250_SHARE_IRQ=y
+# CONFIG_SERIAL_8250_DETECT_IRQ is not set
+# CONFIG_SERIAL_8250_RSA is not set
+
+#
+# Non-8250 serial port support
+#
+CONFIG_SERIAL_CORE=y
+CONFIG_SERIAL_CORE_CONSOLE=y
+# CONFIG_SERIAL_JSM is not set
+CONFIG_UNIX98_PTYS=y
+CONFIG_LEGACY_PTYS=y
+CONFIG_LEGACY_PTY_COUNT=256
+
+#
+# IPMI
+#
+# CONFIG_IPMI_HANDLER is not set
+
+#
+# Watchdog Cards
+#
+# CONFIG_WATCHDOG is not set
+# CONFIG_HW_RANDOM is not set
+CONFIG_EFI_RTC=y
+# CONFIG_DTLK is not set
+# CONFIG_R3964 is not set
+# CONFIG_APPLICOM is not set
+
+#
+# Ftape, the floppy tape device driver
+#
+CONFIG_AGP=y
+# CONFIG_AGP_SIS is not set
+# CONFIG_AGP_VIA is not set
+CONFIG_DRM=y
+# CONFIG_DRM_TDFX is not set
+# CONFIG_DRM_R128 is not set
+CONFIG_DRM_RADEON=y
+# CONFIG_DRM_MGA is not set
+# CONFIG_DRM_SIS is not set
+# CONFIG_DRM_VIA is not set
+# CONFIG_DRM_SAVAGE is not set
+# CONFIG_RAW_DRIVER is not set
+# CONFIG_HPET is not set
+# CONFIG_HANGCHECK_TIMER is not set
+
+#
+# TPM devices
+#
+# CONFIG_TCG_TPM is not set
+# CONFIG_TELCLOCK is not set
+
+#
+# I2C support
+#
+CONFIG_I2C=y
+CONFIG_I2C_CHARDEV=y
+
+#
+# I2C Algorithms
+#
+CONFIG_I2C_ALGOBIT=y
+CONFIG_I2C_ALGOPCF=y
+# CONFIG_I2C_ALGOPCA is not set
+
+#
+# I2C Hardware Bus support
+#
+# CONFIG_I2C_ALI1535 is not set
+# CONFIG_I2C_ALI1563 is not set
+# CONFIG_I2C_ALI15X3 is not set
+# CONFIG_I2C_AMD756 is not set
+# CONFIG_I2C_AMD8111 is not set
+# CONFIG_I2C_I801 is not set
+# CONFIG_I2C_I810 is not set
+# CONFIG_I2C_PIIX4 is not set
+# CONFIG_I2C_NFORCE2 is not set
+# CONFIG_I2C_OCORES is not set
+# CONFIG_I2C_PARPORT_LIGHT is not set
+# CONFIG_I2C_PROSAVAGE is not set
+# CONFIG_I2C_SAVAGE4 is not set
+# CONFIG_I2C_SIS5595 is not set
+# CONFIG_I2C_SIS630 is not set
+# CONFIG_I2C_SIS96X is not set
+# CONFIG_I2C_STUB is not set
+# CONFIG_I2C_VIA is not set
+# CONFIG_I2C_VIAPRO is not set
+# CONFIG_I2C_VOODOO3 is not set
+# CONFIG_I2C_PCA_ISA is not set
+
+#
+# Miscellaneous I2C Chip support
+#
+# CONFIG_SENSORS_DS1337 is not set
+# CONFIG_SENSORS_DS1374 is not set
+# CONFIG_SENSORS_EEPROM is not set
+# CONFIG_SENSORS_PCF8574 is not set
+# CONFIG_SENSORS_PCA9539 is not set
+# CONFIG_SENSORS_PCF8591 is not set
+# CONFIG_SENSORS_MAX6875 is not set
+# CONFIG_I2C_DEBUG_CORE is not set
+# CONFIG_I2C_DEBUG_ALGO is not set
+# CONFIG_I2C_DEBUG_BUS is not set
+# CONFIG_I2C_DEBUG_CHIP is not set
+
+#
+# SPI support
+#
+# CONFIG_SPI is not set
+# CONFIG_SPI_MASTER is not set
+
+#
+# Dallas's 1-wire bus
+#
+
+#
+# Hardware Monitoring support
+#
+CONFIG_HWMON=y
+# CONFIG_HWMON_VID is not set
+# CONFIG_SENSORS_ABITUGURU is not set
+# CONFIG_SENSORS_ADM1021 is not set
+# CONFIG_SENSORS_ADM1025 is not set
+# CONFIG_SENSORS_ADM1026 is not set
+# CONFIG_SENSORS_ADM1031 is not set
+# CONFIG_SENSORS_ADM9240 is not set
+# CONFIG_SENSORS_ASB100 is not set
+# CONFIG_SENSORS_ATXP1 is not set
+# CONFIG_SENSORS_DS1621 is not set
+# CONFIG_SENSORS_F71805F is not set
+# CONFIG_SENSORS_FSCHER is not set
+# CONFIG_SENSORS_FSCPOS is not set
+# CONFIG_SENSORS_GL518SM is not set
+# CONFIG_SENSORS_GL520SM is not set
+# CONFIG_SENSORS_IT87 is not set
+# CONFIG_SENSORS_LM63 is not set
+# CONFIG_SENSORS_LM75 is not set
+# CONFIG_SENSORS_LM77 is not set
+# CONFIG_SENSORS_LM78 is not set
+# CONFIG_SENSORS_LM80 is not set
+# CONFIG_SENSORS_LM83 is not set
+# CONFIG_SENSORS_LM85 is not set
+# CONFIG_SENSORS_LM87 is not set
+# CONFIG_SENSORS_LM90 is not set
+# CONFIG_SENSORS_LM92 is not set
+# CONFIG_SENSORS_MAX1619 is not set
+# CONFIG_SENSORS_PC87360 is not set
+# CONFIG_SENSORS_SIS5595 is not set
+# CONFIG_SENSORS_SMSC47M1 is not set
+# CONFIG_SENSORS_SMSC47M192 is not set
+# CONFIG_SENSORS_SMSC47B397 is not set
+# CONFIG_SENSORS_VIA686A is not set
+# CONFIG_SENSORS_VT8231 is not set
+# CONFIG_SENSORS_W83781D is not set
+# CONFIG_SENSORS_W83791D is not set
+# CONFIG_SENSORS_W83792D is not set
+# CONFIG_SENSORS_W83L785TS is not set
+# CONFIG_SENSORS_W83627HF is not set
+# CONFIG_SENSORS_W83627EHF is not set
+# CONFIG_HWMON_DEBUG_CHIP is not set
+
+#
+# Misc devices
+#
+
+#
+# Multimedia devices
+#
+CONFIG_VIDEO_DEV=y
+CONFIG_VIDEO_V4L1=y
+CONFIG_VIDEO_V4L1_COMPAT=y
+CONFIG_VIDEO_V4L2=y
+
+#
+# Video Capture Adapters
+#
+
+#
+# Video Capture Adapters
+#
+# CONFIG_VIDEO_ADV_DEBUG is not set
+# CONFIG_VIDEO_VIVI is not set
+# CONFIG_VIDEO_BT848 is not set
+# CONFIG_VIDEO_CPIA is not set
+# CONFIG_VIDEO_CPIA2 is not set
+# CONFIG_VIDEO_SAA5246A is not set
+# CONFIG_VIDEO_SAA5249 is not set
+# CONFIG_TUNER_3036 is not set
+# CONFIG_VIDEO_STRADIS is not set
+# CONFIG_VIDEO_ZORAN is not set
+# CONFIG_VIDEO_SAA7134 is not set
+# CONFIG_VIDEO_MXB is not set
+# CONFIG_VIDEO_DPC is not set
+# CONFIG_VIDEO_HEXIUM_ORION is not set
+# CONFIG_VIDEO_HEXIUM_GEMINI is not set
+# CONFIG_VIDEO_CX88 is not set
+
+#
+# Encoders and Decoders
+#
+# CONFIG_VIDEO_MSP3400 is not set
+# CONFIG_VIDEO_CS53L32A is not set
+# CONFIG_VIDEO_TLV320AIC23B is not set
+# CONFIG_VIDEO_WM8775 is not set
+# CONFIG_VIDEO_WM8739 is not set
+# CONFIG_VIDEO_CX2341X is not set
+# CONFIG_VIDEO_CX25840 is not set
+# CONFIG_VIDEO_SAA711X is not set
+# CONFIG_VIDEO_SAA7127 is not set
+# CONFIG_VIDEO_UPD64031A is not set
+# CONFIG_VIDEO_UPD64083 is not set
+
+#
+# V4L USB devices
+#
+# CONFIG_VIDEO_PVRUSB2 is not set
+# CONFIG_VIDEO_EM28XX is not set
+# CONFIG_USB_VICAM is not set
+# CONFIG_USB_IBMCAM is not set
+# CONFIG_USB_KONICAWC is not set
+# CONFIG_USB_QUICKCAM_MESSENGER is not set
+# CONFIG_USB_ET61X251 is not set
+# CONFIG_VIDEO_OVCAMCHIP is not set
+# CONFIG_USB_W9968CF is not set
+# CONFIG_USB_OV511 is not set
+# CONFIG_USB_SE401 is not set
+# CONFIG_USB_SN9C102 is not set
+# CONFIG_USB_STV680 is not set
+# CONFIG_USB_ZC0301 is not set
+# CONFIG_USB_PWC is not set
+
+#
+# Radio Adapters
+#
+# CONFIG_RADIO_GEMTEK_PCI is not set
+# CONFIG_RADIO_MAXIRADIO is not set
+# CONFIG_RADIO_MAESTRO is not set
+# CONFIG_USB_DSBR is not set
+
+#
+# Digital Video Broadcasting Devices
+#
+# CONFIG_DVB is not set
+# CONFIG_USB_DABUSB is not set
+
+#
+# Graphics support
+#
+CONFIG_FIRMWARE_EDID=y
+CONFIG_FB=y
+CONFIG_FB_CFB_FILLRECT=y
+CONFIG_FB_CFB_COPYAREA=y
+CONFIG_FB_CFB_IMAGEBLIT=y
+# CONFIG_FB_MACMODES is not set
+# CONFIG_FB_BACKLIGHT is not set
+CONFIG_FB_MODE_HELPERS=y
+# CONFIG_FB_TILEBLITTING is not set
+# CONFIG_FB_CIRRUS is not set
+# CONFIG_FB_PM2 is not set
+# CONFIG_FB_CYBER2000 is not set
+# CONFIG_FB_ASILIANT is not set
+# CONFIG_FB_IMSTT is not set
+# CONFIG_FB_S1D13XXX is not set
+# CONFIG_FB_NVIDIA is not set
+# CONFIG_FB_RIVA is not set
+# CONFIG_FB_MATROX is not set
+# CONFIG_FB_RADEON is not set
+# CONFIG_FB_ATY128 is not set
+# CONFIG_FB_ATY is not set
+# CONFIG_FB_SAVAGE is not set
+# CONFIG_FB_SIS is not set
+# CONFIG_FB_NEOMAGIC is not set
+# CONFIG_FB_KYRO is not set
+# CONFIG_FB_3DFX is not set
+# CONFIG_FB_VOODOO1 is not set
+# CONFIG_FB_TRIDENT is not set
+# CONFIG_FB_VIRTUAL is not set
+
+#
+# Console display driver support
+#
+CONFIG_VGA_CONSOLE=y
+# CONFIG_VGACON_SOFT_SCROLLBACK is not set
+CONFIG_DUMMY_CONSOLE=y
+CONFIG_FRAMEBUFFER_CONSOLE=y
+# CONFIG_FRAMEBUFFER_CONSOLE_ROTATION is not set
+# CONFIG_FONTS is not set
+CONFIG_FONT_8x8=y
+CONFIG_FONT_8x16=y
+
+#
+# Logo configuration
+#
+CONFIG_LOGO=y
+# CONFIG_LOGO_LINUX_MONO is not set
+# CONFIG_LOGO_LINUX_VGA16 is not set
+CONFIG_LOGO_LINUX_CLUT224=y
+# CONFIG_BACKLIGHT_LCD_SUPPORT is not set
+
+#
+# Sound
+#
+CONFIG_SOUND=y
+
+#
+# Advanced Linux Sound Architecture
+#
+CONFIG_SND=y
+CONFIG_SND_TIMER=y
+CONFIG_SND_PCM=y
+CONFIG_SND_HWDEP=y
+CONFIG_SND_RAWMIDI=y
+CONFIG_SND_SEQUENCER=y
+# CONFIG_SND_SEQ_DUMMY is not set
+CONFIG_SND_OSSEMUL=y
+CONFIG_SND_MIXER_OSS=y
+CONFIG_SND_PCM_OSS=y
+CONFIG_SND_PCM_OSS_PLUGINS=y
+CONFIG_SND_SEQUENCER_OSS=y
+# CONFIG_SND_DYNAMIC_MINORS is not set
+CONFIG_SND_SUPPORT_OLD_API=y
+CONFIG_SND_VERBOSE_PROCFS=y
+# CONFIG_SND_VERBOSE_PRINTK is not set
+# CONFIG_SND_DEBUG is not set
+
+#
+# Generic devices
+#
+CONFIG_SND_MPU401_UART=y
+CONFIG_SND_OPL3_LIB=y
+CONFIG_SND_AC97_CODEC=y
+CONFIG_SND_AC97_BUS=y
+# CONFIG_SND_DUMMY is not set
+# CONFIG_SND_VIRMIDI is not set
+# CONFIG_SND_MTPAV is not set
+# CONFIG_SND_SERIAL_U16550 is not set
+# CONFIG_SND_MPU401 is not set
+
+#
+# PCI devices
+#
+# CONFIG_SND_AD1889 is not set
+# CONFIG_SND_ALS300 is not set
+# CONFIG_SND_ALI5451 is not set
+# CONFIG_SND_ATIIXP is not set
+# CONFIG_SND_ATIIXP_MODEM is not set
+# CONFIG_SND_AU8810 is not set
+# CONFIG_SND_AU8820 is not set
+# CONFIG_SND_AU8830 is not set
+# CONFIG_SND_AZT3328 is not set
+# CONFIG_SND_BT87X is not set
+# CONFIG_SND_CA0106 is not set
+# CONFIG_SND_CMIPCI is not set
+# CONFIG_SND_CS4281 is not set
+# CONFIG_SND_CS46XX is not set
+# CONFIG_SND_EMU10K1 is not set
+# CONFIG_SND_EMU10K1X is not set
+# CONFIG_SND_ENS1370 is not set
+# CONFIG_SND_ENS1371 is not set
+# CONFIG_SND_ES1938 is not set
+# CONFIG_SND_ES1968 is not set
+CONFIG_SND_FM801=y
+# CONFIG_SND_FM801_TEA575X_BOOL is not set
+# CONFIG_SND_HDA_INTEL is not set
+# CONFIG_SND_HDSP is not set
+# CONFIG_SND_HDSPM is not set
+# CONFIG_SND_ICE1712 is not set
+# CONFIG_SND_ICE1724 is not set
+# CONFIG_SND_INTEL8X0 is not set
+# CONFIG_SND_INTEL8X0M is not set
+# CONFIG_SND_KORG1212 is not set
+# CONFIG_SND_MAESTRO3 is not set
+# CONFIG_SND_MIXART is not set
+# CONFIG_SND_NM256 is not set
+# CONFIG_SND_PCXHR is not set
+# CONFIG_SND_RME32 is not set
+# CONFIG_SND_RME96 is not set
+# CONFIG_SND_RME9652 is not set
+# CONFIG_SND_SONICVIBES is not set
+# CONFIG_SND_TRIDENT is not set
+# CONFIG_SND_VIA82XX is not set
+# CONFIG_SND_VIA82XX_MODEM is not set
+# CONFIG_SND_VX222 is not set
+# CONFIG_SND_YMFPCI is not set
+
+#
+# USB devices
+#
+# CONFIG_SND_USB_AUDIO is not set
+
+#
+# Open Sound System
+#
+# CONFIG_SOUND_PRIME is not set
+
+#
+# USB support
+#
+CONFIG_USB_ARCH_HAS_HCD=y
+CONFIG_USB_ARCH_HAS_OHCI=y
+CONFIG_USB_ARCH_HAS_EHCI=y
+CONFIG_USB=y
+# CONFIG_USB_DEBUG is not set
+
+#
+# Miscellaneous USB options
+#
+# CONFIG_USB_DEVICEFS is not set
+CONFIG_USB_BANDWIDTH=y
+# CONFIG_USB_DYNAMIC_MINORS is not set
+# CONFIG_USB_SUSPEND is not set
+# CONFIG_USB_OTG is not set
+
+#
+# USB Host Controller Drivers
+#
+CONFIG_USB_EHCI_HCD=y
+# CONFIG_USB_EHCI_SPLIT_ISO is not set
+# CONFIG_USB_EHCI_ROOT_HUB_TT is not set
+# CONFIG_USB_EHCI_TT_NEWSCHED is not set
+# CONFIG_USB_ISP116X_HCD is not set
+CONFIG_USB_OHCI_HCD=y
+# CONFIG_USB_OHCI_BIG_ENDIAN is not set
+CONFIG_USB_OHCI_LITTLE_ENDIAN=y
+CONFIG_USB_UHCI_HCD=y
+# CONFIG_USB_SL811_HCD is not set
+
+#
+# USB Device Class drivers
+#
+# CONFIG_USB_ACM is not set
+# CONFIG_USB_PRINTER is not set
+
+#
+# NOTE: USB_STORAGE enables SCSI, and 'SCSI disk support'
+#
+
+#
+# may also be needed; see USB_STORAGE Help for more information
+#
+CONFIG_USB_STORAGE=y
+# CONFIG_USB_STORAGE_DEBUG is not set
+# CONFIG_USB_STORAGE_DATAFAB is not set
+# CONFIG_USB_STORAGE_FREECOM is not set
+# CONFIG_USB_STORAGE_DPCM is not set
+# CONFIG_USB_STORAGE_USBAT is not set
+# CONFIG_USB_STORAGE_SDDR09 is not set
+# CONFIG_USB_STORAGE_SDDR55 is not set
+# CONFIG_USB_STORAGE_JUMPSHOT is not set
+# CONFIG_USB_STORAGE_ALAUDA is not set
+# CONFIG_USB_LIBUSUAL is not set
+
+#
+# USB Input Devices
+#
+CONFIG_USB_HID=y
+CONFIG_USB_HIDINPUT=y
+# CONFIG_USB_HIDINPUT_POWERBOOK is not set
+# CONFIG_HID_FF is not set
+CONFIG_USB_HIDDEV=y
+# CONFIG_USB_AIPTEK is not set
+# CONFIG_USB_WACOM is not set
+# CONFIG_USB_ACECAD is not set
+# CONFIG_USB_KBTAB is not set
+# CONFIG_USB_POWERMATE is not set
+# CONFIG_USB_TOUCHSCREEN is not set
+# CONFIG_USB_YEALINK is not set
+# CONFIG_USB_XPAD is not set
+# CONFIG_USB_ATI_REMOTE is not set
+# CONFIG_USB_ATI_REMOTE2 is not set
+# CONFIG_USB_KEYSPAN_REMOTE is not set
+# CONFIG_USB_APPLETOUCH is not set
+
+#
+# USB Imaging devices
+#
+# CONFIG_USB_MDC800 is not set
+# CONFIG_USB_MICROTEK is not set
+
+#
+# USB Network Adapters
+#
+# CONFIG_USB_CATC is not set
+# CONFIG_USB_KAWETH is not set
+# CONFIG_USB_PEGASUS is not set
+# CONFIG_USB_RTL8150 is not set
+# CONFIG_USB_USBNET is not set
+CONFIG_USB_MON=y
+
+#
+# USB port drivers
+#
+
+#
+# USB Serial Converter support
+#
+# CONFIG_USB_SERIAL is not set
+
+#
+# USB Miscellaneous drivers
+#
+# CONFIG_USB_EMI62 is not set
+# CONFIG_USB_EMI26 is not set
+# CONFIG_USB_AUERSWALD is not set
+# CONFIG_USB_RIO500 is not set
+# CONFIG_USB_LEGOTOWER is not set
+# CONFIG_USB_LCD is not set
+# CONFIG_USB_LED is not set
+# CONFIG_USB_CYPRESS_CY7C63 is not set
+# CONFIG_USB_CYTHERM is not set
+# CONFIG_USB_PHIDGETKIT is not set
+# CONFIG_USB_PHIDGETSERVO is not set
+# CONFIG_USB_IDMOUSE is not set
+# CONFIG_USB_APPLEDISPLAY is not set
+# CONFIG_USB_SISUSBVGA is not set
+# CONFIG_USB_LD is not set
+
+#
+# USB DSL modem support
+#
+
+#
+# USB Gadget Support
+#
+# CONFIG_USB_GADGET is not set
+
+#
+# MMC/SD Card support
+#
+# CONFIG_MMC is not set
+
+#
+# LED devices
+#
+# CONFIG_NEW_LEDS is not set
+
+#
+# LED drivers
+#
+
+#
+# LED Triggers
+#
+
+#
+# InfiniBand support
+#
+# CONFIG_INFINIBAND is not set
+
+#
+# EDAC - error detection and reporting (RAS) (EXPERIMENTAL)
+#
+
+#
+# Real Time Clock
+#
+# CONFIG_RTC_CLASS is not set
+
+#
+# DMA Engine support
+#
+# CONFIG_DMA_ENGINE is not set
+
+#
+# DMA Clients
+#
+
+#
+# DMA Devices
+#
+
+#
+# File systems
+#
+CONFIG_EXT2_FS=y
+CONFIG_EXT2_FS_XATTR=y
+# CONFIG_EXT2_FS_POSIX_ACL is not set
+# CONFIG_EXT2_FS_SECURITY is not set
+# CONFIG_EXT2_FS_XIP is not set
+CONFIG_EXT3_FS=y
+CONFIG_EXT3_FS_XATTR=y
+# CONFIG_EXT3_FS_POSIX_ACL is not set
+# CONFIG_EXT3_FS_SECURITY is not set
+CONFIG_JBD=y
+# CONFIG_JBD_DEBUG is not set
+CONFIG_FS_MBCACHE=y
+# CONFIG_REISERFS_FS is not set
+# CONFIG_JFS_FS is not set
+# CONFIG_FS_POSIX_ACL is not set
+# CONFIG_XFS_FS is not set
+# CONFIG_OCFS2_FS is not set
+# CONFIG_MINIX_FS is not set
+# CONFIG_ROMFS_FS is not set
+CONFIG_INOTIFY=y
+CONFIG_INOTIFY_USER=y
+# CONFIG_QUOTA is not set
+CONFIG_DNOTIFY=y
+CONFIG_AUTOFS_FS=y
+# CONFIG_AUTOFS4_FS is not set
+# CONFIG_FUSE_FS is not set
+
+#
+# CD-ROM/DVD Filesystems
+#
+CONFIG_ISO9660_FS=y
+CONFIG_JOLIET=y
+# CONFIG_ZISOFS is not set
+CONFIG_UDF_FS=y
+CONFIG_UDF_NLS=y
+
+#
+# DOS/FAT/NT Filesystems
+#
+CONFIG_FAT_FS=y
+CONFIG_MSDOS_FS=y
+CONFIG_VFAT_FS=y
+CONFIG_FAT_DEFAULT_CODEPAGE=437
+CONFIG_FAT_DEFAULT_IOCHARSET="iso8859-1"
+# CONFIG_NTFS_FS is not set
+
+#
+# Pseudo filesystems
+#
+CONFIG_PROC_FS=y
+CONFIG_PROC_KCORE=y
+CONFIG_SYSFS=y
+CONFIG_TMPFS=y
+# CONFIG_HUGETLB_PAGE is not set
+CONFIG_RAMFS=y
+# CONFIG_CONFIGFS_FS is not set
+
+#
+# Miscellaneous filesystems
+#
+# CONFIG_ADFS_FS is not set
+# CONFIG_AFFS_FS is not set
+# CONFIG_HFS_FS is not set
+# CONFIG_HFSPLUS_FS is not set
+# CONFIG_BEFS_FS is not set
+# CONFIG_BFS_FS is not set
+# CONFIG_EFS_FS is not set
+# CONFIG_CRAMFS is not set
+# CONFIG_VXFS_FS is not set
+# CONFIG_HPFS_FS is not set
+# CONFIG_QNX4FS_FS is not set
+# CONFIG_SYSV_FS is not set
+# CONFIG_UFS_FS is not set
+
+#
+# Network File Systems
+#
+CONFIG_NFS_FS=y
+CONFIG_NFS_V3=y
+# CONFIG_NFS_V3_ACL is not set
+CONFIG_NFS_V4=y
+# CONFIG_NFS_DIRECTIO is not set
+CONFIG_NFSD=y
+CONFIG_NFSD_V3=y
+# CONFIG_NFSD_V3_ACL is not set
+# CONFIG_NFSD_V4 is not set
+# CONFIG_NFSD_TCP is not set
+CONFIG_ROOT_NFS=y
+CONFIG_LOCKD=y
+CONFIG_LOCKD_V4=y
+CONFIG_EXPORTFS=y
+CONFIG_NFS_COMMON=y
+CONFIG_SUNRPC=y
+CONFIG_SUNRPC_GSS=y
+CONFIG_RPCSEC_GSS_KRB5=y
+# CONFIG_RPCSEC_GSS_SPKM3 is not set
+# CONFIG_SMB_FS is not set
+# CONFIG_CIFS is not set
+# CONFIG_NCP_FS is not set
+# CONFIG_CODA_FS is not set
+# CONFIG_AFS_FS is not set
+# CONFIG_9P_FS is not set
+
+#
+# Partition Types
+#
+CONFIG_PARTITION_ADVANCED=y
+# CONFIG_ACORN_PARTITION is not set
+# CONFIG_OSF_PARTITION is not set
+# CONFIG_AMIGA_PARTITION is not set
+# CONFIG_ATARI_PARTITION is not set
+# CONFIG_MAC_PARTITION is not set
+CONFIG_MSDOS_PARTITION=y
+# CONFIG_BSD_DISKLABEL is not set
+# CONFIG_MINIX_SUBPARTITION is not set
+# CONFIG_SOLARIS_X86_PARTITION is not set
+# CONFIG_UNIXWARE_DISKLABEL is not set
+# CONFIG_LDM_PARTITION is not set
+# CONFIG_SGI_PARTITION is not set
+# CONFIG_ULTRIX_PARTITION is not set
+# CONFIG_SUN_PARTITION is not set
+# CONFIG_KARMA_PARTITION is not set
+CONFIG_EFI_PARTITION=y
+
+#
+# Native Language Support
+#
+CONFIG_NLS=y
+CONFIG_NLS_DEFAULT="iso8859-1"
+CONFIG_NLS_CODEPAGE_437=y
+CONFIG_NLS_CODEPAGE_737=y
+CONFIG_NLS_CODEPAGE_775=y
+CONFIG_NLS_CODEPAGE_850=y
+CONFIG_NLS_CODEPAGE_852=y
+CONFIG_NLS_CODEPAGE_855=y
+CONFIG_NLS_CODEPAGE_857=y
+CONFIG_NLS_CODEPAGE_860=y
+CONFIG_NLS_CODEPAGE_861=y
+CONFIG_NLS_CODEPAGE_862=y
+CONFIG_NLS_CODEPAGE_863=y
+CONFIG_NLS_CODEPAGE_864=y
+CONFIG_NLS_CODEPAGE_865=y
+CONFIG_NLS_CODEPAGE_866=y
+CONFIG_NLS_CODEPAGE_869=y
+CONFIG_NLS_CODEPAGE_936=y
+CONFIG_NLS_CODEPAGE_950=y
+CONFIG_NLS_CODEPAGE_932=y
+CONFIG_NLS_CODEPAGE_949=y
+CONFIG_NLS_CODEPAGE_874=y
+CONFIG_NLS_ISO8859_8=y
+# CONFIG_NLS_CODEPAGE_1250 is not set
+CONFIG_NLS_CODEPAGE_1251=y
+# CONFIG_NLS_ASCII is not set
+CONFIG_NLS_ISO8859_1=y
+CONFIG_NLS_ISO8859_2=y
+CONFIG_NLS_ISO8859_3=y
+CONFIG_NLS_ISO8859_4=y
+CONFIG_NLS_ISO8859_5=y
+CONFIG_NLS_ISO8859_6=y
+CONFIG_NLS_ISO8859_7=y
+CONFIG_NLS_ISO8859_9=y
+CONFIG_NLS_ISO8859_13=y
+CONFIG_NLS_ISO8859_14=y
+CONFIG_NLS_ISO8859_15=y
+CONFIG_NLS_KOI8_R=y
+CONFIG_NLS_KOI8_U=y
+CONFIG_NLS_UTF8=y
+
+#
+# Library routines
+#
+# CONFIG_CRC_CCITT is not set
+# CONFIG_CRC16 is not set
+CONFIG_CRC32=y
+# CONFIG_LIBCRC32C is not set
+CONFIG_PLIST=y
+CONFIG_GENERIC_HARDIRQS=y
+CONFIG_GENERIC_IRQ_PROBE=y
+CONFIG_GENERIC_PENDING_IRQ=y
+CONFIG_IRQ_PER_CPU=y
+
+#
+# Instrumentation Support
+#
+# CONFIG_PROFILING is not set
+# CONFIG_KPROBES is not set
+
+#
+# Kernel hacking
+#
+# CONFIG_PRINTK_TIME is not set
+CONFIG_MAGIC_SYSRQ=y
+CONFIG_UNUSED_SYMBOLS=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_LOG_BUF_SHIFT=17
+CONFIG_DETECT_SOFTLOCKUP=y
+# CONFIG_SCHEDSTATS is not set
+# CONFIG_DEBUG_SLAB is not set
+# CONFIG_DEBUG_RT_MUTEXES is not set
+# CONFIG_RT_MUTEX_TESTER is not set
+# CONFIG_DEBUG_SPINLOCK is not set
+CONFIG_DEBUG_MUTEXES=y
+# CONFIG_DEBUG_RWSEMS is not set
+# CONFIG_DEBUG_SPINLOCK_SLEEP is not set
+# CONFIG_DEBUG_LOCKING_API_SELFTESTS is not set
+# CONFIG_DEBUG_KOBJECT is not set
+# CONFIG_DEBUG_INFO is not set
+# CONFIG_DEBUG_FS is not set
+# CONFIG_DEBUG_VM is not set
+CONFIG_FORCED_INLINING=y
+# CONFIG_RCU_TORTURE_TEST is not set
+CONFIG_IA64_GRANULE_16MB=y
+# CONFIG_IA64_GRANULE_64MB is not set
+CONFIG_IA64_PRINT_HAZARDS=y
+# CONFIG_DISABLE_VHPT is not set
+# CONFIG_IA64_DEBUG_CMPXCHG is not set
+# CONFIG_IA64_DEBUG_IRQ is not set
+
+#
+# Security options
+#
+# CONFIG_KEYS is not set
+# CONFIG_SECURITY is not set
+
+#
+# Cryptographic options
+#
+CONFIG_CRYPTO=y
+# CONFIG_CRYPTO_HMAC is not set
+# CONFIG_CRYPTO_NULL is not set
+# CONFIG_CRYPTO_MD4 is not set
+CONFIG_CRYPTO_MD5=y
+# CONFIG_CRYPTO_SHA1 is not set
+# CONFIG_CRYPTO_SHA256 is not set
+# CONFIG_CRYPTO_SHA512 is not set
+# CONFIG_CRYPTO_WP512 is not set
+# CONFIG_CRYPTO_TGR192 is not set
+CONFIG_CRYPTO_DES=y
+# CONFIG_CRYPTO_BLOWFISH is not set
+# CONFIG_CRYPTO_TWOFISH is not set
+# CONFIG_CRYPTO_SERPENT is not set
+# CONFIG_CRYPTO_AES is not set
+# CONFIG_CRYPTO_CAST5 is not set
+# CONFIG_CRYPTO_CAST6 is not set
+# CONFIG_CRYPTO_TEA is not set
+# CONFIG_CRYPTO_ARC4 is not set
+# CONFIG_CRYPTO_KHAZAD is not set
+# CONFIG_CRYPTO_ANUBIS is not set
+# CONFIG_CRYPTO_DEFLATE is not set
+# CONFIG_CRYPTO_MICHAEL_MIC is not set
+# CONFIG_CRYPTO_CRC32C is not set
+# CONFIG_CRYPTO_TEST is not set
+
+#
+# Hardware crypto devices
+#
+# CONFIG_XEN_SMPBOOT is not set
+# CONFIG_XEN_DEVMEM is not set
+CONFIG_XEN_INTERFACE_VERSION=0x00030207
+
+#
+# XEN
+#
+# CONFIG_XEN_PRIVILEGED_GUEST is not set
+CONFIG_XEN_UNPRIVILEGED_GUEST=y
+CONFIG_XEN_PRIVCMD=y
+CONFIG_XEN_XENBUS_DEV=y
+# CONFIG_XEN_BACKEND is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_UTIL is not set
+CONFIG_XEN_BLKDEV_FRONTEND=y
+CONFIG_XEN_NETDEV_FRONTEND=y
+CONFIG_XEN_SCSI_FRONTEND=m
+# CONFIG_XEN_GRANT_DEV is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_FRONTEND is not set
+CONFIG_XEN_FRAMEBUFFER=y
+CONFIG_XEN_KEYBOARD=y
+# CONFIG_XEN_SCRUB_PAGES is not set
+# CONFIG_XEN_DISABLE_SERIAL is not set
+CONFIG_XEN_SYSFS=y
+CONFIG_XEN_COMPAT_030002_AND_LATER=y
+# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
+# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
+CONFIG_XEN_COMPAT=0x030002
+CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
+CONFIG_NO_IDLE_HZ=y
+CONFIG_XEN_BALLOON=y
+CONFIG_XEN_XENCOMM=y
diff -rpuN linux-2.6.18.8/buildconfigs/linux-defconfig_xenU_x86_32 linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xenU_x86_32
--- linux-2.6.18.8/buildconfigs/linux-defconfig_xenU_x86_32	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xenU_x86_32	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,953 @@
+#
+# Automatically generated make config: don't edit
+# Linux kernel version: 2.6.18.8
+# Tue Oct 16 09:31:29 2007
+#
+CONFIG_X86_32=y
+CONFIG_LOCKDEP_SUPPORT=y
+CONFIG_STACKTRACE_SUPPORT=y
+CONFIG_SEMAPHORE_SLEEPERS=y
+CONFIG_X86=y
+CONFIG_MMU=y
+CONFIG_GENERIC_ISA_DMA=y
+CONFIG_GENERIC_IOMAP=y
+CONFIG_GENERIC_HWEIGHT=y
+CONFIG_ARCH_MAY_HAVE_PC_FDC=y
+CONFIG_DMI=y
+CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
+
+#
+# Code maturity level options
+#
+CONFIG_EXPERIMENTAL=y
+CONFIG_LOCK_KERNEL=y
+CONFIG_INIT_ENV_ARG_LIMIT=32
+
+#
+# General setup
+#
+CONFIG_LOCALVERSION=""
+# CONFIG_LOCALVERSION_AUTO is not set
+CONFIG_SWAP=y
+CONFIG_SYSVIPC=y
+# CONFIG_POSIX_MQUEUE is not set
+# CONFIG_BSD_PROCESS_ACCT is not set
+# CONFIG_TASKSTATS is not set
+# CONFIG_AUDIT is not set
+# CONFIG_IKCONFIG is not set
+# CONFIG_CPUSETS is not set
+# CONFIG_RELAY is not set
+CONFIG_INITRAMFS_SOURCE=""
+# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set
+# CONFIG_EMBEDDED is not set
+CONFIG_UID16=y
+CONFIG_SYSCTL=y
+CONFIG_KALLSYMS=y
+# CONFIG_KALLSYMS_ALL is not set
+# CONFIG_KALLSYMS_EXTRA_PASS is not set
+CONFIG_HOTPLUG=y
+CONFIG_PRINTK=y
+CONFIG_BUG=y
+CONFIG_ELF_CORE=y
+CONFIG_BASE_FULL=y
+CONFIG_FUTEX=y
+CONFIG_EPOLL=y
+CONFIG_SHMEM=y
+CONFIG_SLAB=y
+CONFIG_VM_EVENT_COUNTERS=y
+CONFIG_RT_MUTEXES=y
+# CONFIG_TINY_SHMEM is not set
+CONFIG_BASE_SMALL=0
+# CONFIG_SLOB is not set
+
+#
+# Loadable module support
+#
+CONFIG_MODULES=y
+CONFIG_MODULE_UNLOAD=y
+# CONFIG_MODULE_FORCE_UNLOAD is not set
+# CONFIG_MODVERSIONS is not set
+# CONFIG_MODULE_SRCVERSION_ALL is not set
+CONFIG_KMOD=y
+CONFIG_STOP_MACHINE=y
+
+#
+# Block layer
+#
+# CONFIG_LBD is not set
+# CONFIG_BLK_DEV_IO_TRACE is not set
+CONFIG_LSF=y
+
+#
+# IO Schedulers
+#
+CONFIG_IOSCHED_NOOP=y
+CONFIG_IOSCHED_AS=y
+CONFIG_IOSCHED_DEADLINE=y
+CONFIG_IOSCHED_CFQ=y
+CONFIG_DEFAULT_AS=y
+# CONFIG_DEFAULT_DEADLINE is not set
+# CONFIG_DEFAULT_CFQ is not set
+# CONFIG_DEFAULT_NOOP is not set
+CONFIG_DEFAULT_IOSCHED="anticipatory"
+
+#
+# Processor type and features
+#
+CONFIG_SMP=y
+# CONFIG_X86_PC is not set
+CONFIG_X86_XEN=y
+# CONFIG_X86_ELAN is not set
+# CONFIG_X86_VOYAGER is not set
+# CONFIG_X86_NUMAQ is not set
+# CONFIG_X86_SUMMIT is not set
+# CONFIG_X86_BIGSMP is not set
+# CONFIG_X86_VISWS is not set
+# CONFIG_X86_GENERICARCH is not set
+# CONFIG_X86_ES7000 is not set
+# CONFIG_M386 is not set
+# CONFIG_M486 is not set
+# CONFIG_M586 is not set
+# CONFIG_M586TSC is not set
+# CONFIG_M586MMX is not set
+CONFIG_M686=y
+# CONFIG_MPENTIUMII is not set
+# CONFIG_MPENTIUMIII is not set
+# CONFIG_MPENTIUMM is not set
+# CONFIG_MPENTIUM4 is not set
+# CONFIG_MK6 is not set
+# CONFIG_MK7 is not set
+# CONFIG_MK8 is not set
+# CONFIG_MCRUSOE is not set
+# CONFIG_MEFFICEON is not set
+# CONFIG_MWINCHIPC6 is not set
+# CONFIG_MWINCHIP2 is not set
+# CONFIG_MWINCHIP3D is not set
+# CONFIG_MGEODEGX1 is not set
+# CONFIG_MGEODE_LX is not set
+# CONFIG_MCYRIXIII is not set
+# CONFIG_MVIAC3_2 is not set
+# CONFIG_X86_GENERIC is not set
+CONFIG_X86_CMPXCHG=y
+CONFIG_X86_XADD=y
+CONFIG_X86_L1_CACHE_SHIFT=5
+CONFIG_RWSEM_XCHGADD_ALGORITHM=y
+CONFIG_GENERIC_CALIBRATE_DELAY=y
+CONFIG_X86_PPRO_FENCE=y
+CONFIG_X86_WP_WORKS_OK=y
+CONFIG_X86_INVLPG=y
+CONFIG_X86_BSWAP=y
+CONFIG_X86_POPAD_OK=y
+CONFIG_X86_CMPXCHG64=y
+CONFIG_X86_GOOD_APIC=y
+CONFIG_X86_USE_PPRO_CHECKSUM=y
+CONFIG_NR_CPUS=8
+CONFIG_PREEMPT_NONE=y
+# CONFIG_PREEMPT_VOLUNTARY is not set
+# CONFIG_PREEMPT is not set
+CONFIG_PREEMPT_BKL=y
+CONFIG_VM86=y
+# CONFIG_TOSHIBA is not set
+# CONFIG_I8K is not set
+# CONFIG_X86_REBOOTFIXUPS is not set
+# CONFIG_X86_MSR is not set
+CONFIG_X86_CPUID=y
+CONFIG_SWIOTLB=y
+
+#
+# Firmware Drivers
+#
+# CONFIG_EDD is not set
+# CONFIG_DELL_RBU is not set
+# CONFIG_DCDBAS is not set
+# CONFIG_NOHIGHMEM is not set
+CONFIG_HIGHMEM4G=y
+# CONFIG_HIGHMEM64G is not set
+CONFIG_PAGE_OFFSET=0xC0000000
+CONFIG_HIGHMEM=y
+CONFIG_SELECT_MEMORY_MODEL=y
+CONFIG_FLATMEM_MANUAL=y
+# CONFIG_DISCONTIGMEM_MANUAL is not set
+# CONFIG_SPARSEMEM_MANUAL is not set
+CONFIG_FLATMEM=y
+CONFIG_FLAT_NODE_MEM_MAP=y
+# CONFIG_SPARSEMEM_STATIC is not set
+CONFIG_SPLIT_PTLOCK_CPUS=4
+CONFIG_RESOURCES_64BIT=y
+# CONFIG_HIGHPTE is not set
+# CONFIG_REGPARM is not set
+CONFIG_SECCOMP=y
+CONFIG_HZ_100=y
+# CONFIG_HZ_250 is not set
+# CONFIG_HZ_1000 is not set
+CONFIG_HZ=100
+# CONFIG_CRASH_DUMP is not set
+CONFIG_PHYSICAL_START=0x100000
+CONFIG_HOTPLUG_CPU=y
+CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y
+
+#
+# Bus options (PCI, PCMCIA, EISA, MCA, ISA)
+#
+# CONFIG_PCI is not set
+CONFIG_ISA_DMA_API=y
+# CONFIG_SCx200 is not set
+
+#
+# PCCARD (PCMCIA/CardBus) support
+#
+# CONFIG_PCCARD is not set
+
+#
+# PCI Hotplug Support
+#
+
+#
+# Executable file formats
+#
+CONFIG_BINFMT_ELF=y
+# CONFIG_BINFMT_AOUT is not set
+# CONFIG_BINFMT_MISC is not set
+
+#
+# Networking
+#
+CONFIG_NET=y
+
+#
+# Networking options
+#
+# CONFIG_NETDEBUG is not set
+CONFIG_PACKET=y
+# CONFIG_PACKET_MMAP is not set
+CONFIG_UNIX=y
+CONFIG_XFRM=y
+# CONFIG_XFRM_USER is not set
+# CONFIG_NET_KEY is not set
+CONFIG_INET=y
+# CONFIG_IP_MULTICAST is not set
+# CONFIG_IP_ADVANCED_ROUTER is not set
+CONFIG_IP_FIB_HASH=y
+CONFIG_IP_PNP=y
+# CONFIG_IP_PNP_DHCP is not set
+# CONFIG_IP_PNP_BOOTP is not set
+# CONFIG_IP_PNP_RARP is not set
+# CONFIG_NET_IPIP is not set
+# CONFIG_NET_IPGRE is not set
+# CONFIG_ARPD is not set
+# CONFIG_SYN_COOKIES is not set
+# CONFIG_INET_AH is not set
+# CONFIG_INET_ESP is not set
+# CONFIG_INET_IPCOMP is not set
+# CONFIG_INET_XFRM_TUNNEL is not set
+# CONFIG_INET_TUNNEL is not set
+CONFIG_INET_XFRM_MODE_TRANSPORT=y
+CONFIG_INET_XFRM_MODE_TUNNEL=y
+# CONFIG_INET_DIAG is not set
+# CONFIG_TCP_CONG_ADVANCED is not set
+CONFIG_TCP_CONG_BIC=y
+# CONFIG_IPV6 is not set
+# CONFIG_INET6_XFRM_TUNNEL is not set
+# CONFIG_INET6_TUNNEL is not set
+# CONFIG_NETWORK_SECMARK is not set
+# CONFIG_NETFILTER is not set
+
+#
+# DCCP Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP_DCCP is not set
+
+#
+# SCTP Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP_SCTP is not set
+
+#
+# TIPC Configuration (EXPERIMENTAL)
+#
+# CONFIG_TIPC is not set
+# CONFIG_ATM is not set
+# CONFIG_BRIDGE is not set
+# CONFIG_VLAN_8021Q is not set
+# CONFIG_DECNET is not set
+# CONFIG_LLC2 is not set
+# CONFIG_IPX is not set
+# CONFIG_ATALK is not set
+# CONFIG_X25 is not set
+# CONFIG_LAPB is not set
+# CONFIG_ECONET is not set
+# CONFIG_WAN_ROUTER is not set
+
+#
+# QoS and/or fair queueing
+#
+# CONFIG_NET_SCHED is not set
+
+#
+# Network testing
+#
+# CONFIG_NET_PKTGEN is not set
+# CONFIG_HAMRADIO is not set
+# CONFIG_IRDA is not set
+# CONFIG_BT is not set
+# CONFIG_IEEE80211 is not set
+
+#
+# Device Drivers
+#
+
+#
+# Generic Driver Options
+#
+CONFIG_STANDALONE=y
+CONFIG_PREVENT_FIRMWARE_BUILD=y
+# CONFIG_FW_LOADER is not set
+# CONFIG_DEBUG_DRIVER is not set
+CONFIG_SYS_HYPERVISOR=y
+
+#
+# Connector - unified userspace <-> kernelspace linker
+#
+# CONFIG_CONNECTOR is not set
+
+#
+# Memory Technology Devices (MTD)
+#
+# CONFIG_MTD is not set
+
+#
+# Parallel port support
+#
+# CONFIG_PARPORT is not set
+
+#
+# Plug and Play support
+#
+
+#
+# Block devices
+#
+# CONFIG_BLK_DEV_FD is not set
+# CONFIG_BLK_DEV_COW_COMMON is not set
+CONFIG_BLK_DEV_LOOP=m
+# CONFIG_BLK_DEV_CRYPTOLOOP is not set
+CONFIG_BLK_DEV_NBD=m
+CONFIG_BLK_DEV_RAM=y
+CONFIG_BLK_DEV_RAM_COUNT=16
+CONFIG_BLK_DEV_RAM_SIZE=4096
+CONFIG_BLK_DEV_RAM_BLOCKSIZE=1024
+CONFIG_BLK_DEV_INITRD=y
+# CONFIG_CDROM_PKTCDVD is not set
+# CONFIG_ATA_OVER_ETH is not set
+
+#
+# ATA/ATAPI/MFM/RLL support
+#
+# CONFIG_IDE is not set
+
+#
+# SCSI device support
+#
+# CONFIG_RAID_ATTRS is not set
+CONFIG_SCSI=m
+CONFIG_SCSI_PROC_FS=y
+
+#
+# SCSI support type (disk, tape, CD-ROM)
+#
+CONFIG_BLK_DEV_SD=m
+# CONFIG_CHR_DEV_ST is not set
+# CONFIG_CHR_DEV_OSST is not set
+# CONFIG_BLK_DEV_SR is not set
+# CONFIG_CHR_DEV_SG is not set
+# CONFIG_CHR_DEV_SCH is not set
+
+#
+# Some SCSI devices (e.g. CD jukebox) support multiple LUNs
+#
+# CONFIG_SCSI_MULTI_LUN is not set
+# CONFIG_SCSI_CONSTANTS is not set
+# CONFIG_SCSI_LOGGING is not set
+
+#
+# SCSI Transport Attributes
+#
+# CONFIG_SCSI_SPI_ATTRS is not set
+# CONFIG_SCSI_FC_ATTRS is not set
+# CONFIG_SCSI_ISCSI_ATTRS is not set
+# CONFIG_SCSI_SAS_ATTRS is not set
+
+#
+# SCSI low-level drivers
+#
+# CONFIG_ISCSI_TCP is not set
+# CONFIG_SCSI_SATA is not set
+# CONFIG_SCSI_DEBUG is not set
+
+#
+# Multi-device support (RAID and LVM)
+#
+# CONFIG_MD is not set
+
+#
+# Fusion MPT device support
+#
+# CONFIG_FUSION is not set
+
+#
+# IEEE 1394 (FireWire) support
+#
+
+#
+# I2O device support
+#
+
+#
+# Network device support
+#
+CONFIG_NETDEVICES=y
+# CONFIG_DUMMY is not set
+# CONFIG_BONDING is not set
+# CONFIG_EQUALIZER is not set
+# CONFIG_TUN is not set
+
+#
+# PHY device support
+#
+
+#
+# Ethernet (10 or 100Mbit)
+#
+# CONFIG_NET_ETHERNET is not set
+
+#
+# Ethernet (1000 Mbit)
+#
+
+#
+# Ethernet (10000 Mbit)
+#
+
+#
+# Token Ring devices
+#
+
+#
+# Wireless LAN (non-hamradio)
+#
+# CONFIG_NET_RADIO is not set
+
+#
+# Wan interfaces
+#
+# CONFIG_WAN is not set
+# CONFIG_PPP is not set
+# CONFIG_SLIP is not set
+# CONFIG_SHAPER is not set
+# CONFIG_NETCONSOLE is not set
+# CONFIG_NETPOLL is not set
+# CONFIG_NET_POLL_CONTROLLER is not set
+
+#
+# ISDN subsystem
+#
+# CONFIG_ISDN is not set
+
+#
+# Telephony Support
+#
+# CONFIG_PHONE is not set
+
+#
+# Input device support
+#
+CONFIG_INPUT=y
+
+#
+# Userland interfaces
+#
+CONFIG_INPUT_MOUSEDEV=y
+# CONFIG_INPUT_MOUSEDEV_PSAUX is not set
+CONFIG_INPUT_MOUSEDEV_SCREEN_X=1024
+CONFIG_INPUT_MOUSEDEV_SCREEN_Y=768
+# CONFIG_INPUT_JOYDEV is not set
+# CONFIG_INPUT_TSDEV is not set
+# CONFIG_INPUT_EVDEV is not set
+# CONFIG_INPUT_EVBUG is not set
+
+#
+# Input Device Drivers
+#
+CONFIG_INPUT_KEYBOARD=y
+# CONFIG_KEYBOARD_ATKBD is not set
+# CONFIG_KEYBOARD_SUNKBD is not set
+# CONFIG_KEYBOARD_LKKBD is not set
+# CONFIG_KEYBOARD_XTKBD is not set
+# CONFIG_KEYBOARD_NEWTON is not set
+# CONFIG_INPUT_MOUSE is not set
+# CONFIG_INPUT_JOYSTICK is not set
+# CONFIG_INPUT_TOUCHSCREEN is not set
+# CONFIG_INPUT_MISC is not set
+
+#
+# Hardware I/O ports
+#
+CONFIG_SERIO=y
+CONFIG_SERIO_I8042=y
+CONFIG_SERIO_SERPORT=y
+# CONFIG_SERIO_CT82C710 is not set
+# CONFIG_SERIO_RAW is not set
+# CONFIG_GAMEPORT is not set
+
+#
+# Character devices
+#
+CONFIG_VT=y
+CONFIG_VT_CONSOLE=y
+CONFIG_HW_CONSOLE=y
+# CONFIG_VT_HW_CONSOLE_BINDING is not set
+# CONFIG_SERIAL_NONSTANDARD is not set
+
+#
+# Serial drivers
+#
+
+#
+# Non-8250 serial port support
+#
+CONFIG_UNIX98_PTYS=y
+CONFIG_LEGACY_PTYS=y
+CONFIG_LEGACY_PTY_COUNT=256
+
+#
+# IPMI
+#
+# CONFIG_IPMI_HANDLER is not set
+
+#
+# Watchdog Cards
+#
+# CONFIG_WATCHDOG is not set
+CONFIG_HW_RANDOM=y
+CONFIG_HW_RANDOM_VIA=y
+# CONFIG_NVRAM is not set
+# CONFIG_RTC is not set
+# CONFIG_GEN_RTC is not set
+# CONFIG_DTLK is not set
+# CONFIG_R3964 is not set
+
+#
+# Ftape, the floppy tape device driver
+#
+# CONFIG_MWAVE is not set
+# CONFIG_PC8736x_GPIO is not set
+# CONFIG_NSC_GPIO is not set
+# CONFIG_CS5535_GPIO is not set
+# CONFIG_RAW_DRIVER is not set
+# CONFIG_HANGCHECK_TIMER is not set
+
+#
+# TPM devices
+#
+# CONFIG_TCG_TPM is not set
+# CONFIG_TELCLOCK is not set
+
+#
+# I2C support
+#
+# CONFIG_I2C is not set
+
+#
+# SPI support
+#
+# CONFIG_SPI is not set
+# CONFIG_SPI_MASTER is not set
+
+#
+# Dallas's 1-wire bus
+#
+
+#
+# Hardware Monitoring support
+#
+# CONFIG_HWMON is not set
+# CONFIG_HWMON_VID is not set
+
+#
+# Misc devices
+#
+
+#
+# Multimedia devices
+#
+# CONFIG_VIDEO_DEV is not set
+
+#
+# Digital Video Broadcasting Devices
+#
+# CONFIG_DVB is not set
+
+#
+# Graphics support
+#
+CONFIG_FIRMWARE_EDID=y
+# CONFIG_FB is not set
+
+#
+# Console display driver support
+#
+CONFIG_VGA_CONSOLE=y
+# CONFIG_VGACON_SOFT_SCROLLBACK is not set
+CONFIG_DUMMY_CONSOLE=y
+# CONFIG_BACKLIGHT_LCD_SUPPORT is not set
+
+#
+# Sound
+#
+# CONFIG_SOUND is not set
+
+#
+# USB support
+#
+# CONFIG_USB_ARCH_HAS_HCD is not set
+# CONFIG_USB_ARCH_HAS_OHCI is not set
+# CONFIG_USB_ARCH_HAS_EHCI is not set
+
+#
+# NOTE: USB_STORAGE enables SCSI, and 'SCSI disk support'
+#
+
+#
+# USB Gadget Support
+#
+# CONFIG_USB_GADGET is not set
+
+#
+# MMC/SD Card support
+#
+# CONFIG_MMC is not set
+
+#
+# LED devices
+#
+# CONFIG_NEW_LEDS is not set
+
+#
+# LED drivers
+#
+
+#
+# LED Triggers
+#
+
+#
+# InfiniBand support
+#
+
+#
+# EDAC - error detection and reporting (RAS) (EXPERIMENTAL)
+#
+# CONFIG_EDAC is not set
+
+#
+# Real Time Clock
+#
+# CONFIG_RTC_CLASS is not set
+
+#
+# DMA Engine support
+#
+# CONFIG_DMA_ENGINE is not set
+
+#
+# DMA Clients
+#
+
+#
+# DMA Devices
+#
+
+#
+# File systems
+#
+CONFIG_EXT2_FS=y
+# CONFIG_EXT2_FS_XATTR is not set
+# CONFIG_EXT2_FS_XIP is not set
+CONFIG_EXT3_FS=y
+CONFIG_EXT3_FS_XATTR=y
+# CONFIG_EXT3_FS_POSIX_ACL is not set
+# CONFIG_EXT3_FS_SECURITY is not set
+CONFIG_JBD=y
+# CONFIG_JBD_DEBUG is not set
+CONFIG_FS_MBCACHE=y
+CONFIG_REISERFS_FS=y
+# CONFIG_REISERFS_CHECK is not set
+# CONFIG_REISERFS_PROC_INFO is not set
+# CONFIG_REISERFS_FS_XATTR is not set
+# CONFIG_JFS_FS is not set
+# CONFIG_FS_POSIX_ACL is not set
+# CONFIG_XFS_FS is not set
+# CONFIG_OCFS2_FS is not set
+# CONFIG_MINIX_FS is not set
+# CONFIG_ROMFS_FS is not set
+CONFIG_INOTIFY=y
+CONFIG_INOTIFY_USER=y
+# CONFIG_QUOTA is not set
+CONFIG_DNOTIFY=y
+CONFIG_AUTOFS_FS=y
+CONFIG_AUTOFS4_FS=y
+# CONFIG_FUSE_FS is not set
+
+#
+# CD-ROM/DVD Filesystems
+#
+CONFIG_ISO9660_FS=y
+CONFIG_JOLIET=y
+CONFIG_ZISOFS=y
+CONFIG_ZISOFS_FS=y
+# CONFIG_UDF_FS is not set
+
+#
+# DOS/FAT/NT Filesystems
+#
+CONFIG_FAT_FS=m
+CONFIG_MSDOS_FS=m
+CONFIG_VFAT_FS=m
+CONFIG_FAT_DEFAULT_CODEPAGE=437
+CONFIG_FAT_DEFAULT_IOCHARSET="iso8859-1"
+# CONFIG_NTFS_FS is not set
+
+#
+# Pseudo filesystems
+#
+CONFIG_PROC_FS=y
+CONFIG_PROC_KCORE=y
+CONFIG_SYSFS=y
+CONFIG_TMPFS=y
+# CONFIG_HUGETLB_PAGE is not set
+CONFIG_RAMFS=y
+# CONFIG_CONFIGFS_FS is not set
+
+#
+# Miscellaneous filesystems
+#
+# CONFIG_ADFS_FS is not set
+# CONFIG_AFFS_FS is not set
+# CONFIG_HFS_FS is not set
+# CONFIG_HFSPLUS_FS is not set
+# CONFIG_BEFS_FS is not set
+# CONFIG_BFS_FS is not set
+# CONFIG_EFS_FS is not set
+CONFIG_CRAMFS=y
+# CONFIG_VXFS_FS is not set
+# CONFIG_HPFS_FS is not set
+# CONFIG_QNX4FS_FS is not set
+# CONFIG_SYSV_FS is not set
+# CONFIG_UFS_FS is not set
+
+#
+# Network File Systems
+#
+CONFIG_NFS_FS=y
+CONFIG_NFS_V3=y
+# CONFIG_NFS_V3_ACL is not set
+# CONFIG_NFS_V4 is not set
+# CONFIG_NFS_DIRECTIO is not set
+# CONFIG_NFSD is not set
+CONFIG_ROOT_NFS=y
+CONFIG_LOCKD=y
+CONFIG_LOCKD_V4=y
+CONFIG_NFS_COMMON=y
+CONFIG_SUNRPC=y
+# CONFIG_RPCSEC_GSS_KRB5 is not set
+# CONFIG_RPCSEC_GSS_SPKM3 is not set
+# CONFIG_SMB_FS is not set
+# CONFIG_CIFS is not set
+# CONFIG_NCP_FS is not set
+# CONFIG_CODA_FS is not set
+# CONFIG_AFS_FS is not set
+# CONFIG_9P_FS is not set
+
+#
+# Partition Types
+#
+# CONFIG_PARTITION_ADVANCED is not set
+CONFIG_MSDOS_PARTITION=y
+
+#
+# Native Language Support
+#
+CONFIG_NLS=y
+CONFIG_NLS_DEFAULT="iso8859-1"
+CONFIG_NLS_CODEPAGE_437=y
+# CONFIG_NLS_CODEPAGE_737 is not set
+# CONFIG_NLS_CODEPAGE_775 is not set
+# CONFIG_NLS_CODEPAGE_850 is not set
+# CONFIG_NLS_CODEPAGE_852 is not set
+# CONFIG_NLS_CODEPAGE_855 is not set
+# CONFIG_NLS_CODEPAGE_857 is not set
+# CONFIG_NLS_CODEPAGE_860 is not set
+# CONFIG_NLS_CODEPAGE_861 is not set
+# CONFIG_NLS_CODEPAGE_862 is not set
+# CONFIG_NLS_CODEPAGE_863 is not set
+# CONFIG_NLS_CODEPAGE_864 is not set
+# CONFIG_NLS_CODEPAGE_865 is not set
+# CONFIG_NLS_CODEPAGE_866 is not set
+# CONFIG_NLS_CODEPAGE_869 is not set
+# CONFIG_NLS_CODEPAGE_936 is not set
+# CONFIG_NLS_CODEPAGE_950 is not set
+# CONFIG_NLS_CODEPAGE_932 is not set
+# CONFIG_NLS_CODEPAGE_949 is not set
+# CONFIG_NLS_CODEPAGE_874 is not set
+# CONFIG_NLS_ISO8859_8 is not set
+# CONFIG_NLS_CODEPAGE_1250 is not set
+# CONFIG_NLS_CODEPAGE_1251 is not set
+# CONFIG_NLS_ASCII is not set
+CONFIG_NLS_ISO8859_1=y
+# CONFIG_NLS_ISO8859_2 is not set
+# CONFIG_NLS_ISO8859_3 is not set
+# CONFIG_NLS_ISO8859_4 is not set
+# CONFIG_NLS_ISO8859_5 is not set
+# CONFIG_NLS_ISO8859_6 is not set
+# CONFIG_NLS_ISO8859_7 is not set
+# CONFIG_NLS_ISO8859_9 is not set
+# CONFIG_NLS_ISO8859_13 is not set
+# CONFIG_NLS_ISO8859_14 is not set
+# CONFIG_NLS_ISO8859_15 is not set
+# CONFIG_NLS_KOI8_R is not set
+# CONFIG_NLS_KOI8_U is not set
+# CONFIG_NLS_UTF8 is not set
+
+#
+# Instrumentation Support
+#
+# CONFIG_PROFILING is not set
+# CONFIG_KPROBES is not set
+
+#
+# Kernel hacking
+#
+CONFIG_TRACE_IRQFLAGS_SUPPORT=y
+# CONFIG_PRINTK_TIME is not set
+CONFIG_MAGIC_SYSRQ=y
+CONFIG_UNUSED_SYMBOLS=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_LOG_BUF_SHIFT=14
+CONFIG_DETECT_SOFTLOCKUP=y
+# CONFIG_SCHEDSTATS is not set
+# CONFIG_DEBUG_SLAB is not set
+# CONFIG_DEBUG_RT_MUTEXES is not set
+# CONFIG_RT_MUTEX_TESTER is not set
+# CONFIG_DEBUG_SPINLOCK is not set
+# CONFIG_DEBUG_MUTEXES is not set
+# CONFIG_DEBUG_RWSEMS is not set
+# CONFIG_DEBUG_LOCK_ALLOC is not set
+# CONFIG_PROVE_LOCKING is not set
+# CONFIG_DEBUG_SPINLOCK_SLEEP is not set
+# CONFIG_DEBUG_LOCKING_API_SELFTESTS is not set
+# CONFIG_DEBUG_KOBJECT is not set
+# CONFIG_DEBUG_HIGHMEM is not set
+CONFIG_DEBUG_BUGVERBOSE=y
+CONFIG_DEBUG_INFO=y
+# CONFIG_DEBUG_FS is not set
+# CONFIG_DEBUG_VM is not set
+CONFIG_FRAME_POINTER=y
+# CONFIG_UNWIND_INFO is not set
+CONFIG_FORCED_INLINING=y
+# CONFIG_RCU_TORTURE_TEST is not set
+CONFIG_EARLY_PRINTK=y
+# CONFIG_DEBUG_STACKOVERFLOW is not set
+# CONFIG_DEBUG_STACK_USAGE is not set
+# CONFIG_DEBUG_PAGEALLOC is not set
+# CONFIG_DEBUG_RODATA is not set
+# CONFIG_4KSTACKS is not set
+
+#
+# Security options
+#
+# CONFIG_KEYS is not set
+# CONFIG_SECURITY is not set
+
+#
+# Cryptographic options
+#
+CONFIG_CRYPTO=y
+# CONFIG_CRYPTO_HMAC is not set
+# CONFIG_CRYPTO_NULL is not set
+# CONFIG_CRYPTO_MD4 is not set
+CONFIG_CRYPTO_MD5=m
+# CONFIG_CRYPTO_SHA1 is not set
+# CONFIG_CRYPTO_SHA256 is not set
+# CONFIG_CRYPTO_SHA512 is not set
+# CONFIG_CRYPTO_WP512 is not set
+# CONFIG_CRYPTO_TGR192 is not set
+# CONFIG_CRYPTO_DES is not set
+# CONFIG_CRYPTO_BLOWFISH is not set
+# CONFIG_CRYPTO_TWOFISH is not set
+# CONFIG_CRYPTO_SERPENT is not set
+# CONFIG_CRYPTO_AES is not set
+# CONFIG_CRYPTO_AES_586 is not set
+# CONFIG_CRYPTO_CAST5 is not set
+# CONFIG_CRYPTO_CAST6 is not set
+# CONFIG_CRYPTO_TEA is not set
+# CONFIG_CRYPTO_ARC4 is not set
+# CONFIG_CRYPTO_KHAZAD is not set
+# CONFIG_CRYPTO_ANUBIS is not set
+# CONFIG_CRYPTO_DEFLATE is not set
+# CONFIG_CRYPTO_MICHAEL_MIC is not set
+CONFIG_CRYPTO_CRC32C=m
+# CONFIG_CRYPTO_TEST is not set
+
+#
+# Hardware crypto devices
+#
+# CONFIG_CRYPTO_DEV_PADLOCK is not set
+CONFIG_XEN=y
+CONFIG_XEN_INTERFACE_VERSION=0x00030207
+
+#
+# XEN
+#
+# CONFIG_XEN_PRIVILEGED_GUEST is not set
+CONFIG_XEN_UNPRIVILEGED_GUEST=y
+CONFIG_XEN_PRIVCMD=y
+CONFIG_XEN_XENBUS_DEV=y
+# CONFIG_XEN_BACKEND is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_UTIL is not set
+CONFIG_XEN_BLKDEV_FRONTEND=y
+CONFIG_XEN_NETDEV_FRONTEND=y
+CONFIG_XEN_SCSI_FRONTEND=m
+# CONFIG_XEN_GRANT_DEV is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_FRONTEND is not set
+CONFIG_XEN_SCRUB_PAGES=y
+CONFIG_XEN_DISABLE_SERIAL=y
+CONFIG_XEN_SYSFS=y
+CONFIG_XEN_COMPAT_030002_AND_LATER=y
+# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
+# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
+CONFIG_XEN_COMPAT=0x030002
+CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
+CONFIG_NO_IDLE_HZ=y
+CONFIG_XEN_SMPBOOT=y
+CONFIG_XEN_BALLOON=y
+CONFIG_XEN_DEVMEM=y
+
+#
+# Library routines
+#
+# CONFIG_CRC_CCITT is not set
+# CONFIG_CRC16 is not set
+# CONFIG_CRC32 is not set
+CONFIG_LIBCRC32C=m
+CONFIG_ZLIB_INFLATE=y
+CONFIG_PLIST=y
+CONFIG_GENERIC_HARDIRQS=y
+CONFIG_GENERIC_IRQ_PROBE=y
+CONFIG_GENERIC_PENDING_IRQ=y
+CONFIG_X86_SMP=y
+CONFIG_X86_BIOS_REBOOT=y
+CONFIG_X86_TRAMPOLINE=y
+CONFIG_X86_NO_TSS=y
+CONFIG_X86_NO_IDT=y
+CONFIG_KTIME_SCALAR=y
diff -rpuN linux-2.6.18.8/buildconfigs/linux-defconfig_xenU_x86_64 linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xenU_x86_64
--- linux-2.6.18.8/buildconfigs/linux-defconfig_xenU_x86_64	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xenU_x86_64	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,906 @@
+#
+# Automatically generated make config: don't edit
+# Linux kernel version: 2.6.18.8
+# Mon Feb 18 10:42:42 2008
+#
+CONFIG_X86_64=y
+CONFIG_64BIT=y
+CONFIG_X86=y
+CONFIG_LOCKDEP_SUPPORT=y
+CONFIG_STACKTRACE_SUPPORT=y
+CONFIG_SEMAPHORE_SLEEPERS=y
+CONFIG_MMU=y
+CONFIG_RWSEM_GENERIC_SPINLOCK=y
+CONFIG_GENERIC_HWEIGHT=y
+CONFIG_GENERIC_CALIBRATE_DELAY=y
+CONFIG_X86_CMPXCHG=y
+CONFIG_EARLY_PRINTK=y
+CONFIG_GENERIC_ISA_DMA=y
+CONFIG_GENERIC_IOMAP=y
+CONFIG_ARCH_MAY_HAVE_PC_FDC=y
+CONFIG_DMI=y
+CONFIG_AUDIT_ARCH=y
+CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
+
+#
+# Code maturity level options
+#
+CONFIG_EXPERIMENTAL=y
+CONFIG_LOCK_KERNEL=y
+CONFIG_INIT_ENV_ARG_LIMIT=32
+
+#
+# General setup
+#
+CONFIG_LOCALVERSION=""
+# CONFIG_LOCALVERSION_AUTO is not set
+CONFIG_SWAP=y
+CONFIG_SYSVIPC=y
+# CONFIG_POSIX_MQUEUE is not set
+# CONFIG_BSD_PROCESS_ACCT is not set
+# CONFIG_TASKSTATS is not set
+# CONFIG_AUDIT is not set
+# CONFIG_IKCONFIG is not set
+# CONFIG_CPUSETS is not set
+# CONFIG_RELAY is not set
+CONFIG_INITRAMFS_SOURCE=""
+# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set
+# CONFIG_EMBEDDED is not set
+CONFIG_UID16=y
+CONFIG_SYSCTL=y
+CONFIG_KALLSYMS=y
+# CONFIG_KALLSYMS_ALL is not set
+# CONFIG_KALLSYMS_EXTRA_PASS is not set
+CONFIG_HOTPLUG=y
+CONFIG_PRINTK=y
+CONFIG_BUG=y
+CONFIG_ELF_CORE=y
+CONFIG_BASE_FULL=y
+CONFIG_FUTEX=y
+CONFIG_EPOLL=y
+CONFIG_SHMEM=y
+CONFIG_SLAB=y
+CONFIG_VM_EVENT_COUNTERS=y
+CONFIG_RT_MUTEXES=y
+# CONFIG_TINY_SHMEM is not set
+CONFIG_BASE_SMALL=0
+# CONFIG_SLOB is not set
+
+#
+# Loadable module support
+#
+CONFIG_MODULES=y
+CONFIG_MODULE_UNLOAD=y
+# CONFIG_MODULE_FORCE_UNLOAD is not set
+# CONFIG_MODVERSIONS is not set
+# CONFIG_MODULE_SRCVERSION_ALL is not set
+CONFIG_KMOD=y
+CONFIG_STOP_MACHINE=y
+
+#
+# Block layer
+#
+# CONFIG_LBD is not set
+# CONFIG_BLK_DEV_IO_TRACE is not set
+CONFIG_LSF=y
+
+#
+# IO Schedulers
+#
+CONFIG_IOSCHED_NOOP=y
+CONFIG_IOSCHED_AS=y
+CONFIG_IOSCHED_DEADLINE=y
+CONFIG_IOSCHED_CFQ=y
+CONFIG_DEFAULT_AS=y
+# CONFIG_DEFAULT_DEADLINE is not set
+# CONFIG_DEFAULT_CFQ is not set
+# CONFIG_DEFAULT_NOOP is not set
+CONFIG_DEFAULT_IOSCHED="anticipatory"
+
+#
+# Processor type and features
+#
+CONFIG_X86_PC=y
+# CONFIG_X86_VSMP is not set
+# CONFIG_MK8 is not set
+# CONFIG_MPSC is not set
+CONFIG_GENERIC_CPU=y
+CONFIG_X86_64_XEN=y
+CONFIG_X86_NO_TSS=y
+CONFIG_X86_NO_IDT=y
+CONFIG_X86_L1_CACHE_BYTES=128
+CONFIG_X86_L1_CACHE_SHIFT=7
+CONFIG_X86_INTERNODE_CACHE_BYTES=128
+CONFIG_X86_GOOD_APIC=y
+# CONFIG_MICROCODE is not set
+# CONFIG_X86_MSR is not set
+CONFIG_X86_CPUID=y
+CONFIG_X86_XEN_GENAPIC=y
+CONFIG_SMP=y
+CONFIG_PREEMPT_NONE=y
+# CONFIG_PREEMPT_VOLUNTARY is not set
+# CONFIG_PREEMPT is not set
+CONFIG_PREEMPT_BKL=y
+CONFIG_ARCH_FLATMEM_ENABLE=y
+CONFIG_SELECT_MEMORY_MODEL=y
+CONFIG_FLATMEM_MANUAL=y
+# CONFIG_DISCONTIGMEM_MANUAL is not set
+# CONFIG_SPARSEMEM_MANUAL is not set
+CONFIG_FLATMEM=y
+CONFIG_FLAT_NODE_MEM_MAP=y
+# CONFIG_SPARSEMEM_STATIC is not set
+CONFIG_SPLIT_PTLOCK_CPUS=4
+CONFIG_RESOURCES_64BIT=y
+CONFIG_NR_CPUS=8
+CONFIG_HOTPLUG_CPU=y
+CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y
+CONFIG_SWIOTLB=y
+# CONFIG_CRASH_DUMP is not set
+CONFIG_PHYSICAL_START=0x200000
+CONFIG_SECCOMP=y
+CONFIG_HZ_100=y
+# CONFIG_HZ_250 is not set
+# CONFIG_HZ_1000 is not set
+CONFIG_HZ=100
+# CONFIG_REORDER is not set
+CONFIG_GENERIC_HARDIRQS=y
+CONFIG_GENERIC_IRQ_PROBE=y
+CONFIG_ISA_DMA_API=y
+CONFIG_GENERIC_PENDING_IRQ=y
+
+#
+# Bus options (PCI etc.)
+#
+# CONFIG_PCI is not set
+
+#
+# PCCARD (PCMCIA/CardBus) support
+#
+# CONFIG_PCCARD is not set
+
+#
+# PCI Hotplug Support
+#
+
+#
+# Executable file formats / Emulations
+#
+CONFIG_BINFMT_ELF=y
+# CONFIG_BINFMT_MISC is not set
+CONFIG_IA32_EMULATION=y
+CONFIG_IA32_AOUT=y
+CONFIG_COMPAT=y
+CONFIG_SYSVIPC_COMPAT=y
+
+#
+# Networking
+#
+CONFIG_NET=y
+
+#
+# Networking options
+#
+# CONFIG_NETDEBUG is not set
+CONFIG_PACKET=y
+# CONFIG_PACKET_MMAP is not set
+CONFIG_UNIX=y
+CONFIG_XFRM=y
+# CONFIG_XFRM_USER is not set
+# CONFIG_NET_KEY is not set
+CONFIG_INET=y
+# CONFIG_IP_MULTICAST is not set
+# CONFIG_IP_ADVANCED_ROUTER is not set
+CONFIG_IP_FIB_HASH=y
+CONFIG_IP_PNP=y
+# CONFIG_IP_PNP_DHCP is not set
+# CONFIG_IP_PNP_BOOTP is not set
+# CONFIG_IP_PNP_RARP is not set
+# CONFIG_NET_IPIP is not set
+# CONFIG_NET_IPGRE is not set
+# CONFIG_ARPD is not set
+# CONFIG_SYN_COOKIES is not set
+# CONFIG_INET_AH is not set
+# CONFIG_INET_ESP is not set
+# CONFIG_INET_IPCOMP is not set
+# CONFIG_INET_XFRM_TUNNEL is not set
+# CONFIG_INET_TUNNEL is not set
+CONFIG_INET_XFRM_MODE_TRANSPORT=y
+CONFIG_INET_XFRM_MODE_TUNNEL=y
+# CONFIG_INET_DIAG is not set
+# CONFIG_TCP_CONG_ADVANCED is not set
+CONFIG_TCP_CONG_BIC=y
+# CONFIG_IPV6 is not set
+# CONFIG_INET6_XFRM_TUNNEL is not set
+# CONFIG_INET6_TUNNEL is not set
+# CONFIG_NETWORK_SECMARK is not set
+# CONFIG_NETFILTER is not set
+
+#
+# DCCP Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP_DCCP is not set
+
+#
+# SCTP Configuration (EXPERIMENTAL)
+#
+# CONFIG_IP_SCTP is not set
+
+#
+# TIPC Configuration (EXPERIMENTAL)
+#
+# CONFIG_TIPC is not set
+# CONFIG_ATM is not set
+# CONFIG_BRIDGE is not set
+# CONFIG_VLAN_8021Q is not set
+# CONFIG_DECNET is not set
+# CONFIG_LLC2 is not set
+# CONFIG_IPX is not set
+# CONFIG_ATALK is not set
+# CONFIG_X25 is not set
+# CONFIG_LAPB is not set
+# CONFIG_ECONET is not set
+# CONFIG_WAN_ROUTER is not set
+
+#
+# QoS and/or fair queueing
+#
+# CONFIG_NET_SCHED is not set
+
+#
+# Network testing
+#
+# CONFIG_NET_PKTGEN is not set
+# CONFIG_HAMRADIO is not set
+# CONFIG_IRDA is not set
+# CONFIG_BT is not set
+# CONFIG_IEEE80211 is not set
+
+#
+# Device Drivers
+#
+
+#
+# Generic Driver Options
+#
+CONFIG_STANDALONE=y
+CONFIG_PREVENT_FIRMWARE_BUILD=y
+# CONFIG_FW_LOADER is not set
+# CONFIG_DEBUG_DRIVER is not set
+CONFIG_SYS_HYPERVISOR=y
+
+#
+# Connector - unified userspace <-> kernelspace linker
+#
+# CONFIG_CONNECTOR is not set
+
+#
+# Memory Technology Devices (MTD)
+#
+# CONFIG_MTD is not set
+
+#
+# Parallel port support
+#
+# CONFIG_PARPORT is not set
+
+#
+# Plug and Play support
+#
+
+#
+# Block devices
+#
+# CONFIG_BLK_DEV_FD is not set
+# CONFIG_BLK_DEV_COW_COMMON is not set
+CONFIG_BLK_DEV_LOOP=m
+# CONFIG_BLK_DEV_CRYPTOLOOP is not set
+CONFIG_BLK_DEV_NBD=m
+CONFIG_BLK_DEV_RAM=y
+CONFIG_BLK_DEV_RAM_COUNT=16
+CONFIG_BLK_DEV_RAM_SIZE=4096
+CONFIG_BLK_DEV_RAM_BLOCKSIZE=1024
+CONFIG_BLK_DEV_INITRD=y
+# CONFIG_CDROM_PKTCDVD is not set
+# CONFIG_ATA_OVER_ETH is not set
+
+#
+# ATA/ATAPI/MFM/RLL support
+#
+# CONFIG_IDE is not set
+
+#
+# SCSI device support
+#
+# CONFIG_RAID_ATTRS is not set
+CONFIG_SCSI=m
+CONFIG_SCSI_PROC_FS=y
+
+#
+# SCSI support type (disk, tape, CD-ROM)
+#
+CONFIG_BLK_DEV_SD=m
+# CONFIG_CHR_DEV_ST is not set
+# CONFIG_CHR_DEV_OSST is not set
+# CONFIG_BLK_DEV_SR is not set
+# CONFIG_CHR_DEV_SG is not set
+# CONFIG_CHR_DEV_SCH is not set
+
+#
+# Some SCSI devices (e.g. CD jukebox) support multiple LUNs
+#
+# CONFIG_SCSI_MULTI_LUN is not set
+# CONFIG_SCSI_CONSTANTS is not set
+# CONFIG_SCSI_LOGGING is not set
+
+#
+# SCSI Transport Attributes
+#
+# CONFIG_SCSI_SPI_ATTRS is not set
+# CONFIG_SCSI_FC_ATTRS is not set
+# CONFIG_SCSI_ISCSI_ATTRS is not set
+# CONFIG_SCSI_SAS_ATTRS is not set
+
+#
+# SCSI low-level drivers
+#
+# CONFIG_ISCSI_TCP is not set
+# CONFIG_SCSI_SATA is not set
+# CONFIG_SCSI_DEBUG is not set
+
+#
+# Multi-device support (RAID and LVM)
+#
+# CONFIG_MD is not set
+
+#
+# Fusion MPT device support
+#
+# CONFIG_FUSION is not set
+
+#
+# IEEE 1394 (FireWire) support
+#
+
+#
+# I2O device support
+#
+
+#
+# Network device support
+#
+CONFIG_NETDEVICES=y
+# CONFIG_DUMMY is not set
+# CONFIG_BONDING is not set
+# CONFIG_EQUALIZER is not set
+# CONFIG_TUN is not set
+
+#
+# PHY device support
+#
+
+#
+# Ethernet (10 or 100Mbit)
+#
+# CONFIG_NET_ETHERNET is not set
+
+#
+# Ethernet (1000 Mbit)
+#
+
+#
+# Ethernet (10000 Mbit)
+#
+
+#
+# Token Ring devices
+#
+
+#
+# Wireless LAN (non-hamradio)
+#
+# CONFIG_NET_RADIO is not set
+
+#
+# Wan interfaces
+#
+# CONFIG_WAN is not set
+# CONFIG_PPP is not set
+# CONFIG_SLIP is not set
+# CONFIG_SHAPER is not set
+# CONFIG_NETCONSOLE is not set
+# CONFIG_NETPOLL is not set
+# CONFIG_NET_POLL_CONTROLLER is not set
+
+#
+# ISDN subsystem
+#
+# CONFIG_ISDN is not set
+
+#
+# Telephony Support
+#
+# CONFIG_PHONE is not set
+
+#
+# Input device support
+#
+CONFIG_INPUT=y
+
+#
+# Userland interfaces
+#
+CONFIG_INPUT_MOUSEDEV=y
+# CONFIG_INPUT_MOUSEDEV_PSAUX is not set
+CONFIG_INPUT_MOUSEDEV_SCREEN_X=1024
+CONFIG_INPUT_MOUSEDEV_SCREEN_Y=768
+# CONFIG_INPUT_JOYDEV is not set
+# CONFIG_INPUT_TSDEV is not set
+# CONFIG_INPUT_EVDEV is not set
+# CONFIG_INPUT_EVBUG is not set
+
+#
+# Input Device Drivers
+#
+CONFIG_INPUT_KEYBOARD=y
+CONFIG_KEYBOARD_ATKBD=y
+# CONFIG_KEYBOARD_SUNKBD is not set
+# CONFIG_KEYBOARD_LKKBD is not set
+# CONFIG_KEYBOARD_XTKBD is not set
+# CONFIG_KEYBOARD_NEWTON is not set
+# CONFIG_INPUT_MOUSE is not set
+# CONFIG_INPUT_JOYSTICK is not set
+# CONFIG_INPUT_TOUCHSCREEN is not set
+# CONFIG_INPUT_MISC is not set
+
+#
+# Hardware I/O ports
+#
+CONFIG_SERIO=y
+CONFIG_SERIO_I8042=y
+CONFIG_SERIO_SERPORT=y
+# CONFIG_SERIO_CT82C710 is not set
+CONFIG_SERIO_LIBPS2=y
+# CONFIG_SERIO_RAW is not set
+# CONFIG_GAMEPORT is not set
+
+#
+# Character devices
+#
+CONFIG_VT=y
+CONFIG_VT_CONSOLE=y
+CONFIG_HW_CONSOLE=y
+# CONFIG_VT_HW_CONSOLE_BINDING is not set
+# CONFIG_SERIAL_NONSTANDARD is not set
+
+#
+# Serial drivers
+#
+
+#
+# Non-8250 serial port support
+#
+CONFIG_UNIX98_PTYS=y
+CONFIG_LEGACY_PTYS=y
+CONFIG_LEGACY_PTY_COUNT=256
+
+#
+# IPMI
+#
+# CONFIG_IPMI_HANDLER is not set
+
+#
+# Watchdog Cards
+#
+# CONFIG_WATCHDOG is not set
+CONFIG_HW_RANDOM=y
+# CONFIG_NVRAM is not set
+# CONFIG_RTC is not set
+# CONFIG_GEN_RTC is not set
+# CONFIG_DTLK is not set
+# CONFIG_R3964 is not set
+
+#
+# Ftape, the floppy tape device driver
+#
+# CONFIG_MWAVE is not set
+# CONFIG_PC8736x_GPIO is not set
+# CONFIG_RAW_DRIVER is not set
+# CONFIG_HANGCHECK_TIMER is not set
+
+#
+# TPM devices
+#
+# CONFIG_TCG_TPM is not set
+# CONFIG_TELCLOCK is not set
+
+#
+# I2C support
+#
+# CONFIG_I2C is not set
+
+#
+# SPI support
+#
+# CONFIG_SPI is not set
+# CONFIG_SPI_MASTER is not set
+
+#
+# Dallas's 1-wire bus
+#
+
+#
+# Hardware Monitoring support
+#
+# CONFIG_HWMON is not set
+# CONFIG_HWMON_VID is not set
+
+#
+# Misc devices
+#
+
+#
+# Multimedia devices
+#
+# CONFIG_VIDEO_DEV is not set
+
+#
+# Digital Video Broadcasting Devices
+#
+# CONFIG_DVB is not set
+
+#
+# Graphics support
+#
+CONFIG_FIRMWARE_EDID=y
+# CONFIG_FB is not set
+
+#
+# Console display driver support
+#
+CONFIG_VGA_CONSOLE=y
+# CONFIG_VGACON_SOFT_SCROLLBACK is not set
+CONFIG_DUMMY_CONSOLE=y
+# CONFIG_BACKLIGHT_LCD_SUPPORT is not set
+
+#
+# Sound
+#
+# CONFIG_SOUND is not set
+
+#
+# USB support
+#
+# CONFIG_USB_ARCH_HAS_HCD is not set
+# CONFIG_USB_ARCH_HAS_OHCI is not set
+# CONFIG_USB_ARCH_HAS_EHCI is not set
+
+#
+# NOTE: USB_STORAGE enables SCSI, and 'SCSI disk support'
+#
+
+#
+# USB Gadget Support
+#
+# CONFIG_USB_GADGET is not set
+
+#
+# MMC/SD Card support
+#
+# CONFIG_MMC is not set
+
+#
+# LED devices
+#
+# CONFIG_NEW_LEDS is not set
+
+#
+# LED drivers
+#
+
+#
+# LED Triggers
+#
+
+#
+# InfiniBand support
+#
+
+#
+# EDAC - error detection and reporting (RAS) (EXPERIMENTAL)
+#
+# CONFIG_EDAC is not set
+
+#
+# Real Time Clock
+#
+# CONFIG_RTC_CLASS is not set
+
+#
+# DMA Engine support
+#
+# CONFIG_DMA_ENGINE is not set
+
+#
+# DMA Clients
+#
+
+#
+# DMA Devices
+#
+
+#
+# Firmware Drivers
+#
+# CONFIG_EDD is not set
+# CONFIG_DELL_RBU is not set
+# CONFIG_DCDBAS is not set
+
+#
+# File systems
+#
+CONFIG_EXT2_FS=y
+# CONFIG_EXT2_FS_XATTR is not set
+# CONFIG_EXT2_FS_XIP is not set
+CONFIG_EXT3_FS=y
+CONFIG_EXT3_FS_XATTR=y
+# CONFIG_EXT3_FS_POSIX_ACL is not set
+# CONFIG_EXT3_FS_SECURITY is not set
+CONFIG_JBD=y
+# CONFIG_JBD_DEBUG is not set
+CONFIG_FS_MBCACHE=y
+CONFIG_REISERFS_FS=y
+# CONFIG_REISERFS_CHECK is not set
+# CONFIG_REISERFS_PROC_INFO is not set
+# CONFIG_REISERFS_FS_XATTR is not set
+# CONFIG_JFS_FS is not set
+# CONFIG_FS_POSIX_ACL is not set
+# CONFIG_XFS_FS is not set
+# CONFIG_OCFS2_FS is not set
+# CONFIG_MINIX_FS is not set
+# CONFIG_ROMFS_FS is not set
+CONFIG_INOTIFY=y
+CONFIG_INOTIFY_USER=y
+# CONFIG_QUOTA is not set
+CONFIG_DNOTIFY=y
+CONFIG_AUTOFS_FS=y
+CONFIG_AUTOFS4_FS=y
+# CONFIG_FUSE_FS is not set
+
+#
+# CD-ROM/DVD Filesystems
+#
+CONFIG_ISO9660_FS=y
+CONFIG_JOLIET=y
+CONFIG_ZISOFS=y
+CONFIG_ZISOFS_FS=y
+# CONFIG_UDF_FS is not set
+
+#
+# DOS/FAT/NT Filesystems
+#
+CONFIG_FAT_FS=m
+CONFIG_MSDOS_FS=m
+CONFIG_VFAT_FS=m
+CONFIG_FAT_DEFAULT_CODEPAGE=437
+CONFIG_FAT_DEFAULT_IOCHARSET="iso8859-1"
+# CONFIG_NTFS_FS is not set
+
+#
+# Pseudo filesystems
+#
+CONFIG_PROC_FS=y
+CONFIG_PROC_KCORE=y
+CONFIG_SYSFS=y
+CONFIG_TMPFS=y
+# CONFIG_HUGETLB_PAGE is not set
+CONFIG_RAMFS=y
+# CONFIG_CONFIGFS_FS is not set
+
+#
+# Miscellaneous filesystems
+#
+# CONFIG_ADFS_FS is not set
+# CONFIG_AFFS_FS is not set
+# CONFIG_HFS_FS is not set
+# CONFIG_HFSPLUS_FS is not set
+# CONFIG_BEFS_FS is not set
+# CONFIG_BFS_FS is not set
+# CONFIG_EFS_FS is not set
+CONFIG_CRAMFS=y
+# CONFIG_VXFS_FS is not set
+# CONFIG_HPFS_FS is not set
+# CONFIG_QNX4FS_FS is not set
+# CONFIG_SYSV_FS is not set
+# CONFIG_UFS_FS is not set
+
+#
+# Network File Systems
+#
+CONFIG_NFS_FS=y
+CONFIG_NFS_V3=y
+# CONFIG_NFS_V3_ACL is not set
+# CONFIG_NFS_V4 is not set
+# CONFIG_NFS_DIRECTIO is not set
+# CONFIG_NFSD is not set
+CONFIG_ROOT_NFS=y
+CONFIG_LOCKD=y
+CONFIG_LOCKD_V4=y
+CONFIG_NFS_COMMON=y
+CONFIG_SUNRPC=y
+# CONFIG_RPCSEC_GSS_KRB5 is not set
+# CONFIG_RPCSEC_GSS_SPKM3 is not set
+# CONFIG_SMB_FS is not set
+# CONFIG_CIFS is not set
+# CONFIG_NCP_FS is not set
+# CONFIG_CODA_FS is not set
+# CONFIG_AFS_FS is not set
+# CONFIG_9P_FS is not set
+
+#
+# Partition Types
+#
+# CONFIG_PARTITION_ADVANCED is not set
+CONFIG_MSDOS_PARTITION=y
+
+#
+# Native Language Support
+#
+CONFIG_NLS=y
+CONFIG_NLS_DEFAULT="iso8859-1"
+CONFIG_NLS_CODEPAGE_437=y
+# CONFIG_NLS_CODEPAGE_737 is not set
+# CONFIG_NLS_CODEPAGE_775 is not set
+# CONFIG_NLS_CODEPAGE_850 is not set
+# CONFIG_NLS_CODEPAGE_852 is not set
+# CONFIG_NLS_CODEPAGE_855 is not set
+# CONFIG_NLS_CODEPAGE_857 is not set
+# CONFIG_NLS_CODEPAGE_860 is not set
+# CONFIG_NLS_CODEPAGE_861 is not set
+# CONFIG_NLS_CODEPAGE_862 is not set
+# CONFIG_NLS_CODEPAGE_863 is not set
+# CONFIG_NLS_CODEPAGE_864 is not set
+# CONFIG_NLS_CODEPAGE_865 is not set
+# CONFIG_NLS_CODEPAGE_866 is not set
+# CONFIG_NLS_CODEPAGE_869 is not set
+# CONFIG_NLS_CODEPAGE_936 is not set
+# CONFIG_NLS_CODEPAGE_950 is not set
+# CONFIG_NLS_CODEPAGE_932 is not set
+# CONFIG_NLS_CODEPAGE_949 is not set
+# CONFIG_NLS_CODEPAGE_874 is not set
+# CONFIG_NLS_ISO8859_8 is not set
+# CONFIG_NLS_CODEPAGE_1250 is not set
+# CONFIG_NLS_CODEPAGE_1251 is not set
+# CONFIG_NLS_ASCII is not set
+CONFIG_NLS_ISO8859_1=y
+# CONFIG_NLS_ISO8859_2 is not set
+# CONFIG_NLS_ISO8859_3 is not set
+# CONFIG_NLS_ISO8859_4 is not set
+# CONFIG_NLS_ISO8859_5 is not set
+# CONFIG_NLS_ISO8859_6 is not set
+# CONFIG_NLS_ISO8859_7 is not set
+# CONFIG_NLS_ISO8859_9 is not set
+# CONFIG_NLS_ISO8859_13 is not set
+# CONFIG_NLS_ISO8859_14 is not set
+# CONFIG_NLS_ISO8859_15 is not set
+# CONFIG_NLS_KOI8_R is not set
+# CONFIG_NLS_KOI8_U is not set
+# CONFIG_NLS_UTF8 is not set
+
+#
+# Instrumentation Support
+#
+# CONFIG_PROFILING is not set
+# CONFIG_KPROBES is not set
+
+#
+# Kernel hacking
+#
+CONFIG_TRACE_IRQFLAGS_SUPPORT=y
+# CONFIG_PRINTK_TIME is not set
+CONFIG_MAGIC_SYSRQ=y
+CONFIG_UNUSED_SYMBOLS=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_LOG_BUF_SHIFT=14
+CONFIG_DETECT_SOFTLOCKUP=y
+# CONFIG_SCHEDSTATS is not set
+# CONFIG_DEBUG_SLAB is not set
+# CONFIG_DEBUG_RT_MUTEXES is not set
+# CONFIG_RT_MUTEX_TESTER is not set
+# CONFIG_DEBUG_SPINLOCK is not set
+# CONFIG_DEBUG_MUTEXES is not set
+# CONFIG_DEBUG_RWSEMS is not set
+# CONFIG_DEBUG_LOCK_ALLOC is not set
+# CONFIG_PROVE_LOCKING is not set
+# CONFIG_DEBUG_SPINLOCK_SLEEP is not set
+# CONFIG_DEBUG_LOCKING_API_SELFTESTS is not set
+# CONFIG_DEBUG_KOBJECT is not set
+CONFIG_DEBUG_INFO=y
+# CONFIG_DEBUG_FS is not set
+# CONFIG_DEBUG_VM is not set
+CONFIG_FRAME_POINTER=y
+# CONFIG_UNWIND_INFO is not set
+CONFIG_FORCED_INLINING=y
+# CONFIG_RCU_TORTURE_TEST is not set
+# CONFIG_DEBUG_RODATA is not set
+# CONFIG_DEBUG_STACKOVERFLOW is not set
+# CONFIG_DEBUG_STACK_USAGE is not set
+
+#
+# Security options
+#
+# CONFIG_KEYS is not set
+# CONFIG_SECURITY is not set
+
+#
+# Cryptographic options
+#
+CONFIG_CRYPTO=y
+# CONFIG_CRYPTO_HMAC is not set
+# CONFIG_CRYPTO_NULL is not set
+# CONFIG_CRYPTO_MD4 is not set
+CONFIG_CRYPTO_MD5=m
+# CONFIG_CRYPTO_SHA1 is not set
+# CONFIG_CRYPTO_SHA256 is not set
+# CONFIG_CRYPTO_SHA512 is not set
+# CONFIG_CRYPTO_WP512 is not set
+# CONFIG_CRYPTO_TGR192 is not set
+# CONFIG_CRYPTO_DES is not set
+# CONFIG_CRYPTO_BLOWFISH is not set
+# CONFIG_CRYPTO_TWOFISH is not set
+# CONFIG_CRYPTO_SERPENT is not set
+# CONFIG_CRYPTO_AES is not set
+# CONFIG_CRYPTO_AES_X86_64 is not set
+# CONFIG_CRYPTO_CAST5 is not set
+# CONFIG_CRYPTO_CAST6 is not set
+# CONFIG_CRYPTO_TEA is not set
+# CONFIG_CRYPTO_ARC4 is not set
+# CONFIG_CRYPTO_KHAZAD is not set
+# CONFIG_CRYPTO_ANUBIS is not set
+# CONFIG_CRYPTO_DEFLATE is not set
+# CONFIG_CRYPTO_MICHAEL_MIC is not set
+CONFIG_CRYPTO_CRC32C=m
+# CONFIG_CRYPTO_TEST is not set
+
+#
+# Hardware crypto devices
+#
+CONFIG_XEN=y
+CONFIG_XEN_INTERFACE_VERSION=0x00030207
+
+#
+# XEN
+#
+# CONFIG_XEN_PRIVILEGED_GUEST is not set
+CONFIG_XEN_UNPRIVILEGED_GUEST=y
+CONFIG_XEN_PRIVCMD=y
+CONFIG_XEN_XENBUS_DEV=y
+# CONFIG_XEN_BACKEND is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_UTIL is not set
+CONFIG_XEN_BLKDEV_FRONTEND=y
+CONFIG_XEN_NETDEV_FRONTEND=y
+CONFIG_XEN_SCSI_FRONTEND=m
+CONFIG_XEN_GRANT_DEV=y
+# CONFIG_XEN_NETDEV_ACCEL_SFC_FRONTEND is not set
+CONFIG_XEN_SCRUB_PAGES=y
+CONFIG_XEN_DISABLE_SERIAL=y
+CONFIG_XEN_SYSFS=y
+CONFIG_XEN_COMPAT_030002_AND_LATER=y
+# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
+# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
+CONFIG_XEN_COMPAT=0x030002
+CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
+CONFIG_NO_IDLE_HZ=y
+CONFIG_XEN_SMPBOOT=y
+CONFIG_XEN_BALLOON=y
+CONFIG_XEN_DEVMEM=y
+
+#
+# Library routines
+#
+# CONFIG_CRC_CCITT is not set
+# CONFIG_CRC16 is not set
+# CONFIG_CRC32 is not set
+CONFIG_LIBCRC32C=m
+CONFIG_ZLIB_INFLATE=y
+CONFIG_PLIST=y
diff -rpuN linux-2.6.18.8/buildconfigs/linux-defconfig_xen_x86_32 linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xen_x86_32
--- linux-2.6.18.8/buildconfigs/linux-defconfig_xen_x86_32	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xen_x86_32	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,3326 @@
+#
+# Automatically generated make config: don't edit
+# Linux kernel version: 2.6.18.8
+# Mon Oct 15 14:37:47 2007
+#
+CONFIG_X86_32=y
+CONFIG_LOCKDEP_SUPPORT=y
+CONFIG_STACKTRACE_SUPPORT=y
+CONFIG_SEMAPHORE_SLEEPERS=y
+CONFIG_X86=y
+CONFIG_MMU=y
+CONFIG_GENERIC_ISA_DMA=y
+CONFIG_GENERIC_IOMAP=y
+CONFIG_GENERIC_HWEIGHT=y
+CONFIG_ARCH_MAY_HAVE_PC_FDC=y
+CONFIG_DMI=y
+CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
+
+#
+# Code maturity level options
+#
+CONFIG_EXPERIMENTAL=y
+CONFIG_LOCK_KERNEL=y
+CONFIG_INIT_ENV_ARG_LIMIT=32
+
+#
+# General setup
+#
+CONFIG_LOCALVERSION=""
+# CONFIG_LOCALVERSION_AUTO is not set
+CONFIG_SWAP=y
+CONFIG_SYSVIPC=y
+CONFIG_POSIX_MQUEUE=y
+CONFIG_BSD_PROCESS_ACCT=y
+CONFIG_BSD_PROCESS_ACCT_V3=y
+CONFIG_TASKSTATS=y
+CONFIG_TASK_DELAY_ACCT=y
+CONFIG_AUDIT=y
+CONFIG_AUDITSYSCALL=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_CPUSETS=y
+# CONFIG_RELAY is not set
+CONFIG_INITRAMFS_SOURCE=""
+# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set
+# CONFIG_EMBEDDED is not set
+CONFIG_UID16=y
+CONFIG_SYSCTL=y
+CONFIG_KALLSYMS=y
+# CONFIG_KALLSYMS_ALL is not set
+# CONFIG_KALLSYMS_EXTRA_PASS is not set
+CONFIG_HOTPLUG=y
+CONFIG_PRINTK=y
+CONFIG_BUG=y
+CONFIG_ELF_CORE=y
+CONFIG_BASE_FULL=y
+CONFIG_FUTEX=y
+CONFIG_EPOLL=y
+CONFIG_SHMEM=y
+CONFIG_SLAB=y
+CONFIG_VM_EVENT_COUNTERS=y
+CONFIG_RT_MUTEXES=y
+# CONFIG_TINY_SHMEM is not set
+CONFIG_BASE_SMALL=0
+# CONFIG_SLOB is not set
+
+#
+# Loadable module support
+#
+CONFIG_MODULES=y
+CONFIG_MODULE_UNLOAD=y
+CONFIG_MODULE_FORCE_UNLOAD=y
+CONFIG_MODVERSIONS=y
+CONFIG_MODULE_SRCVERSION_ALL=y
+CONFIG_KMOD=y
+CONFIG_STOP_MACHINE=y
+
+#
+# Block layer
+#
+CONFIG_LBD=y
+# CONFIG_BLK_DEV_IO_TRACE is not set
+CONFIG_LSF=y
+
+#
+# IO Schedulers
+#
+CONFIG_IOSCHED_NOOP=y
+CONFIG_IOSCHED_AS=y
+CONFIG_IOSCHED_DEADLINE=y
+CONFIG_IOSCHED_CFQ=y
+# CONFIG_DEFAULT_AS is not set
+# CONFIG_DEFAULT_DEADLINE is not set
+CONFIG_DEFAULT_CFQ=y
+# CONFIG_DEFAULT_NOOP is not set
+CONFIG_DEFAULT_IOSCHED="cfq"
+
+#
+# Processor type and features
+#
+CONFIG_SMP=y
+# CONFIG_X86_PC is not set
+CONFIG_X86_XEN=y
+# CONFIG_X86_ELAN is not set
+# CONFIG_X86_VOYAGER is not set
+# CONFIG_X86_NUMAQ is not set
+# CONFIG_X86_SUMMIT is not set
+# CONFIG_X86_BIGSMP is not set
+# CONFIG_X86_VISWS is not set
+# CONFIG_X86_GENERICARCH is not set
+# CONFIG_X86_ES7000 is not set
+# CONFIG_M386 is not set
+# CONFIG_M486 is not set
+# CONFIG_M586 is not set
+# CONFIG_M586TSC is not set
+# CONFIG_M586MMX is not set
+CONFIG_M686=y
+# CONFIG_MPENTIUMII is not set
+# CONFIG_MPENTIUMIII is not set
+# CONFIG_MPENTIUMM is not set
+# CONFIG_MPENTIUM4 is not set
+# CONFIG_MK6 is not set
+# CONFIG_MK7 is not set
+# CONFIG_MK8 is not set
+# CONFIG_MCRUSOE is not set
+# CONFIG_MEFFICEON is not set
+# CONFIG_MWINCHIPC6 is not set
+# CONFIG_MWINCHIP2 is not set
+# CONFIG_MWINCHIP3D is not set
+# CONFIG_MGEODEGX1 is not set
+# CONFIG_MGEODE_LX is not set
+# CONFIG_MCYRIXIII is not set
+# CONFIG_MVIAC3_2 is not set
+CONFIG_X86_GENERIC=y
+CONFIG_X86_CMPXCHG=y
+CONFIG_X86_XADD=y
+CONFIG_X86_L1_CACHE_SHIFT=7
+CONFIG_RWSEM_XCHGADD_ALGORITHM=y
+CONFIG_GENERIC_CALIBRATE_DELAY=y
+CONFIG_X86_PPRO_FENCE=y
+CONFIG_X86_WP_WORKS_OK=y
+CONFIG_X86_INVLPG=y
+CONFIG_X86_BSWAP=y
+CONFIG_X86_POPAD_OK=y
+CONFIG_X86_CMPXCHG64=y
+CONFIG_X86_GOOD_APIC=y
+CONFIG_X86_INTEL_USERCOPY=y
+CONFIG_X86_USE_PPRO_CHECKSUM=y
+CONFIG_NR_CPUS=32
+# CONFIG_PREEMPT_NONE is not set
+CONFIG_PREEMPT_VOLUNTARY=y
+# CONFIG_PREEMPT is not set
+CONFIG_PREEMPT_BKL=y
+CONFIG_X86_LOCAL_APIC=y
+CONFIG_X86_IO_APIC=y
+CONFIG_VM86=y
+# CONFIG_TOSHIBA is not set
+# CONFIG_I8K is not set
+# CONFIG_X86_REBOOTFIXUPS is not set
+CONFIG_MICROCODE=y
+CONFIG_X86_MSR=y
+CONFIG_X86_CPUID=m
+CONFIG_SWIOTLB=y
+
+#
+# Firmware Drivers
+#
+CONFIG_EDD=y
+CONFIG_DELL_RBU=m
+CONFIG_DCDBAS=m
+# CONFIG_NOHIGHMEM is not set
+CONFIG_HIGHMEM4G=y
+# CONFIG_HIGHMEM64G is not set
+CONFIG_PAGE_OFFSET=0xC0000000
+CONFIG_HIGHMEM=y
+CONFIG_SELECT_MEMORY_MODEL=y
+CONFIG_FLATMEM_MANUAL=y
+# CONFIG_DISCONTIGMEM_MANUAL is not set
+# CONFIG_SPARSEMEM_MANUAL is not set
+CONFIG_FLATMEM=y
+CONFIG_FLAT_NODE_MEM_MAP=y
+# CONFIG_SPARSEMEM_STATIC is not set
+CONFIG_SPLIT_PTLOCK_CPUS=4
+CONFIG_RESOURCES_64BIT=y
+# CONFIG_HIGHPTE is not set
+CONFIG_MTRR=y
+CONFIG_REGPARM=y
+CONFIG_SECCOMP=y
+CONFIG_HZ_100=y
+# CONFIG_HZ_250 is not set
+# CONFIG_HZ_1000 is not set
+CONFIG_HZ=100
+CONFIG_KEXEC=y
+# CONFIG_CRASH_DUMP is not set
+CONFIG_PHYSICAL_START=0x100000
+CONFIG_HOTPLUG_CPU=y
+CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y
+
+#
+# Power management options (ACPI, APM)
+#
+CONFIG_PM=y
+# CONFIG_PM_LEGACY is not set
+CONFIG_PM_DEBUG=y
+# CONFIG_PM_TRACE is not set
+# CONFIG_SOFTWARE_SUSPEND is not set
+CONFIG_SUSPEND_SMP=y
+
+#
+# ACPI (Advanced Configuration and Power Interface) Support
+#
+CONFIG_ACPI=y
+CONFIG_ACPI_SLEEP=y
+CONFIG_ACPI_SLEEP_PROC_FS=y
+# CONFIG_ACPI_SLEEP_PROC_SLEEP is not set
+CONFIG_ACPI_AC=m
+CONFIG_ACPI_BATTERY=m
+CONFIG_ACPI_BUTTON=m
+CONFIG_ACPI_VIDEO=m
+CONFIG_ACPI_HOTKEY=m
+CONFIG_ACPI_FAN=m
+CONFIG_ACPI_DOCK=m
+CONFIG_ACPI_PROCESSOR=m
+CONFIG_ACPI_HOTPLUG_CPU=y
+CONFIG_ACPI_THERMAL=m
+CONFIG_ACPI_ASUS=m
+CONFIG_ACPI_IBM=m
+CONFIG_ACPI_TOSHIBA=m
+CONFIG_ACPI_BLACKLIST_YEAR=0
+# CONFIG_ACPI_DEBUG is not set
+CONFIG_ACPI_EC=y
+CONFIG_ACPI_POWER=y
+CONFIG_ACPI_SYSTEM=y
+CONFIG_ACPI_CONTAINER=m
+CONFIG_ACPI_SBS=m
+CONFIG_ACPI_PV_SLEEP=y
+
+#
+# CPU Frequency scaling
+#
+# CONFIG_CPU_FREQ is not set
+
+#
+# Bus options (PCI, PCMCIA, EISA, MCA, ISA)
+#
+CONFIG_PCI=y
+# CONFIG_PCI_GOBIOS is not set
+# CONFIG_PCI_GOMMCONFIG is not set
+# CONFIG_PCI_GODIRECT is not set
+# CONFIG_PCI_GOXEN_FE is not set
+CONFIG_PCI_GOANY=y
+CONFIG_PCI_DIRECT=y
+CONFIG_PCI_MMCONFIG=y
+CONFIG_XEN_PCIDEV_FRONTEND=y
+# CONFIG_XEN_PCIDEV_FE_DEBUG is not set
+# CONFIG_PCIEPORTBUS is not set
+# CONFIG_PCI_MSI is not set
+# CONFIG_PCI_DEBUG is not set
+CONFIG_ISA_DMA_API=y
+CONFIG_SCx200=m
+CONFIG_K8_NB=y
+
+#
+# PCCARD (PCMCIA/CardBus) support
+#
+CONFIG_PCCARD=m
+# CONFIG_PCMCIA_DEBUG is not set
+CONFIG_PCMCIA=m
+CONFIG_PCMCIA_LOAD_CIS=y
+CONFIG_PCMCIA_IOCTL=y
+CONFIG_CARDBUS=y
+
+#
+# PC-card bridges
+#
+CONFIG_YENTA=m
+CONFIG_YENTA_O2=y
+CONFIG_YENTA_RICOH=y
+CONFIG_YENTA_TI=y
+CONFIG_YENTA_ENE_TUNE=y
+CONFIG_YENTA_TOSHIBA=y
+CONFIG_PD6729=m
+CONFIG_I82092=m
+CONFIG_PCCARD_NONSTATIC=m
+
+#
+# PCI Hotplug Support
+#
+CONFIG_HOTPLUG_PCI=m
+CONFIG_HOTPLUG_PCI_FAKE=m
+CONFIG_HOTPLUG_PCI_ACPI=m
+CONFIG_HOTPLUG_PCI_ACPI_IBM=m
+CONFIG_HOTPLUG_PCI_CPCI=y
+CONFIG_HOTPLUG_PCI_CPCI_ZT5550=m
+CONFIG_HOTPLUG_PCI_CPCI_GENERIC=m
+CONFIG_HOTPLUG_PCI_SHPC=m
+# CONFIG_HOTPLUG_PCI_SHPC_POLL_EVENT_MODE is not set
+
+#
+# Executable file formats
+#
+CONFIG_BINFMT_ELF=y
+CONFIG_BINFMT_AOUT=m
+CONFIG_BINFMT_MISC=m
+
+#
+# Networking
+#
+CONFIG_NET=y
+
+#
+# Networking options
+#
+# CONFIG_NETDEBUG is not set
+CONFIG_PACKET=y
+CONFIG_PACKET_MMAP=y
+CONFIG_UNIX=y
+CONFIG_XFRM=y
+CONFIG_XFRM_USER=m
+CONFIG_NET_KEY=m
+CONFIG_INET=y
+CONFIG_IP_MULTICAST=y
+CONFIG_IP_ADVANCED_ROUTER=y
+CONFIG_ASK_IP_FIB_HASH=y
+# CONFIG_IP_FIB_TRIE is not set
+CONFIG_IP_FIB_HASH=y
+CONFIG_IP_MULTIPLE_TABLES=y
+CONFIG_IP_ROUTE_FWMARK=y
+CONFIG_IP_ROUTE_MULTIPATH=y
+# CONFIG_IP_ROUTE_MULTIPATH_CACHED is not set
+CONFIG_IP_ROUTE_VERBOSE=y
+# CONFIG_IP_PNP is not set
+CONFIG_NET_IPIP=m
+CONFIG_NET_IPGRE=m
+CONFIG_NET_IPGRE_BROADCAST=y
+CONFIG_IP_MROUTE=y
+CONFIG_IP_PIMSM_V1=y
+CONFIG_IP_PIMSM_V2=y
+# CONFIG_ARPD is not set
+CONFIG_SYN_COOKIES=y
+CONFIG_INET_AH=m
+CONFIG_INET_ESP=m
+CONFIG_INET_IPCOMP=m
+CONFIG_INET_XFRM_TUNNEL=m
+CONFIG_INET_TUNNEL=m
+CONFIG_INET_XFRM_MODE_TRANSPORT=m
+CONFIG_INET_XFRM_MODE_TUNNEL=m
+CONFIG_INET_DIAG=m
+CONFIG_INET_TCP_DIAG=m
+CONFIG_TCP_CONG_ADVANCED=y
+
+#
+# TCP congestion control
+#
+CONFIG_TCP_CONG_BIC=m
+CONFIG_TCP_CONG_CUBIC=m
+CONFIG_TCP_CONG_WESTWOOD=m
+CONFIG_TCP_CONG_HTCP=m
+CONFIG_TCP_CONG_HSTCP=m
+CONFIG_TCP_CONG_HYBLA=m
+CONFIG_TCP_CONG_VEGAS=m
+CONFIG_TCP_CONG_SCALABLE=m
+CONFIG_TCP_CONG_LP=m
+CONFIG_TCP_CONG_VENO=m
+
+#
+# IP: Virtual Server Configuration
+#
+CONFIG_IP_VS=m
+# CONFIG_IP_VS_DEBUG is not set
+CONFIG_IP_VS_TAB_BITS=12
+
+#
+# IPVS transport protocol load balancing support
+#
+CONFIG_IP_VS_PROTO_TCP=y
+CONFIG_IP_VS_PROTO_UDP=y
+CONFIG_IP_VS_PROTO_ESP=y
+CONFIG_IP_VS_PROTO_AH=y
+
+#
+# IPVS scheduler
+#
+CONFIG_IP_VS_RR=m
+CONFIG_IP_VS_WRR=m
+CONFIG_IP_VS_LC=m
+CONFIG_IP_VS_WLC=m
+CONFIG_IP_VS_LBLC=m
+CONFIG_IP_VS_LBLCR=m
+CONFIG_IP_VS_DH=m
+CONFIG_IP_VS_SH=m
+CONFIG_IP_VS_SED=m
+CONFIG_IP_VS_NQ=m
+
+#
+# IPVS application helper
+#
+CONFIG_IP_VS_FTP=m
+CONFIG_IPV6=m
+CONFIG_IPV6_PRIVACY=y
+# CONFIG_IPV6_ROUTER_PREF is not set
+CONFIG_INET6_AH=m
+CONFIG_INET6_ESP=m
+CONFIG_INET6_IPCOMP=m
+CONFIG_INET6_XFRM_TUNNEL=m
+CONFIG_INET6_TUNNEL=m
+CONFIG_INET6_XFRM_MODE_TRANSPORT=m
+CONFIG_INET6_XFRM_MODE_TUNNEL=m
+CONFIG_IPV6_TUNNEL=m
+CONFIG_NETWORK_SECMARK=y
+CONFIG_NETFILTER=y
+# CONFIG_NETFILTER_DEBUG is not set
+CONFIG_BRIDGE_NETFILTER=y
+
+#
+# Core Netfilter Configuration
+#
+CONFIG_NETFILTER_NETLINK=m
+CONFIG_NETFILTER_NETLINK_QUEUE=m
+CONFIG_NETFILTER_NETLINK_LOG=m
+CONFIG_NETFILTER_XTABLES=m
+CONFIG_NETFILTER_XT_TARGET_CLASSIFY=m
+CONFIG_NETFILTER_XT_TARGET_CONNMARK=m
+CONFIG_NETFILTER_XT_TARGET_MARK=m
+CONFIG_NETFILTER_XT_TARGET_NFQUEUE=m
+CONFIG_NETFILTER_XT_TARGET_NOTRACK=m
+CONFIG_NETFILTER_XT_TARGET_SECMARK=m
+# CONFIG_NETFILTER_XT_TARGET_CONNSECMARK is not set
+CONFIG_NETFILTER_XT_MATCH_COMMENT=m
+CONFIG_NETFILTER_XT_MATCH_CONNBYTES=m
+CONFIG_NETFILTER_XT_MATCH_CONNMARK=m
+CONFIG_NETFILTER_XT_MATCH_CONNTRACK=m
+CONFIG_NETFILTER_XT_MATCH_DCCP=m
+CONFIG_NETFILTER_XT_MATCH_ESP=m
+CONFIG_NETFILTER_XT_MATCH_HELPER=m
+CONFIG_NETFILTER_XT_MATCH_LENGTH=m
+CONFIG_NETFILTER_XT_MATCH_LIMIT=m
+CONFIG_NETFILTER_XT_MATCH_MAC=m
+CONFIG_NETFILTER_XT_MATCH_MARK=m
+CONFIG_NETFILTER_XT_MATCH_POLICY=m
+CONFIG_NETFILTER_XT_MATCH_MULTIPORT=m
+CONFIG_NETFILTER_XT_MATCH_PHYSDEV=m
+CONFIG_NETFILTER_XT_MATCH_PKTTYPE=m
+CONFIG_NETFILTER_XT_MATCH_QUOTA=m
+CONFIG_NETFILTER_XT_MATCH_REALM=m
+CONFIG_NETFILTER_XT_MATCH_SCTP=m
+CONFIG_NETFILTER_XT_MATCH_STATE=m
+CONFIG_NETFILTER_XT_MATCH_STATISTIC=m
+CONFIG_NETFILTER_XT_MATCH_STRING=m
+CONFIG_NETFILTER_XT_MATCH_TCPMSS=m
+
+#
+# IP: Netfilter Configuration
+#
+CONFIG_IP_NF_CONNTRACK=m
+CONFIG_IP_NF_CT_ACCT=y
+CONFIG_IP_NF_CONNTRACK_MARK=y
+CONFIG_IP_NF_CONNTRACK_SECMARK=y
+CONFIG_IP_NF_CONNTRACK_EVENTS=y
+CONFIG_IP_NF_CONNTRACK_NETLINK=m
+CONFIG_IP_NF_CT_PROTO_SCTP=m
+CONFIG_IP_NF_FTP=m
+CONFIG_IP_NF_IRC=m
+CONFIG_IP_NF_NETBIOS_NS=m
+CONFIG_IP_NF_TFTP=m
+CONFIG_IP_NF_AMANDA=m
+CONFIG_IP_NF_PPTP=m
+CONFIG_IP_NF_H323=m
+CONFIG_IP_NF_SIP=m
+CONFIG_IP_NF_QUEUE=m
+CONFIG_IP_NF_IPTABLES=m
+CONFIG_IP_NF_MATCH_IPRANGE=m
+CONFIG_IP_NF_MATCH_TOS=m
+CONFIG_IP_NF_MATCH_RECENT=m
+CONFIG_IP_NF_MATCH_ECN=m
+CONFIG_IP_NF_MATCH_DSCP=m
+CONFIG_IP_NF_MATCH_AH=m
+CONFIG_IP_NF_MATCH_TTL=m
+CONFIG_IP_NF_MATCH_OWNER=m
+CONFIG_IP_NF_MATCH_ADDRTYPE=m
+CONFIG_IP_NF_MATCH_HASHLIMIT=m
+CONFIG_IP_NF_FILTER=m
+CONFIG_IP_NF_TARGET_REJECT=m
+CONFIG_IP_NF_TARGET_LOG=m
+CONFIG_IP_NF_TARGET_ULOG=m
+CONFIG_IP_NF_TARGET_TCPMSS=m
+CONFIG_IP_NF_NAT=m
+CONFIG_IP_NF_NAT_NEEDED=y
+CONFIG_IP_NF_TARGET_MASQUERADE=m
+CONFIG_IP_NF_TARGET_REDIRECT=m
+CONFIG_IP_NF_TARGET_NETMAP=m
+CONFIG_IP_NF_TARGET_SAME=m
+CONFIG_IP_NF_NAT_SNMP_BASIC=m
+CONFIG_IP_NF_NAT_IRC=m
+CONFIG_IP_NF_NAT_FTP=m
+CONFIG_IP_NF_NAT_TFTP=m
+CONFIG_IP_NF_NAT_AMANDA=m
+CONFIG_IP_NF_NAT_PPTP=m
+CONFIG_IP_NF_NAT_H323=m
+CONFIG_IP_NF_NAT_SIP=m
+CONFIG_IP_NF_MANGLE=m
+CONFIG_IP_NF_TARGET_TOS=m
+CONFIG_IP_NF_TARGET_ECN=m
+CONFIG_IP_NF_TARGET_DSCP=m
+CONFIG_IP_NF_TARGET_TTL=m
+CONFIG_IP_NF_TARGET_CLUSTERIP=m
+CONFIG_IP_NF_RAW=m
+CONFIG_IP_NF_ARPTABLES=m
+CONFIG_IP_NF_ARPFILTER=m
+CONFIG_IP_NF_ARP_MANGLE=m
+
+#
+# IPv6: Netfilter Configuration (EXPERIMENTAL)
+#
+CONFIG_IP6_NF_QUEUE=m
+CONFIG_IP6_NF_IPTABLES=m
+CONFIG_IP6_NF_MATCH_RT=m
+CONFIG_IP6_NF_MATCH_OPTS=m
+CONFIG_IP6_NF_MATCH_FRAG=m
+CONFIG_IP6_NF_MATCH_HL=m
+CONFIG_IP6_NF_MATCH_OWNER=m
+CONFIG_IP6_NF_MATCH_IPV6HEADER=m
+CONFIG_IP6_NF_MATCH_AH=m
+CONFIG_IP6_NF_MATCH_EUI64=m
+CONFIG_IP6_NF_FILTER=m
+CONFIG_IP6_NF_TARGET_LOG=m
+CONFIG_IP6_NF_TARGET_REJECT=m
+CONFIG_IP6_NF_MANGLE=m
+CONFIG_IP6_NF_TARGET_HL=m
+CONFIG_IP6_NF_RAW=m
+
+#
+# DECnet: Netfilter Configuration
+#
+CONFIG_DECNET_NF_GRABULATOR=m
+
+#
+# Bridge: Netfilter Configuration
+#
+CONFIG_BRIDGE_NF_EBTABLES=m
+CONFIG_BRIDGE_EBT_BROUTE=m
+CONFIG_BRIDGE_EBT_T_FILTER=m
+CONFIG_BRIDGE_EBT_T_NAT=m
+CONFIG_BRIDGE_EBT_802_3=m
+CONFIG_BRIDGE_EBT_AMONG=m
+CONFIG_BRIDGE_EBT_ARP=m
+CONFIG_BRIDGE_EBT_IP=m
+CONFIG_BRIDGE_EBT_LIMIT=m
+CONFIG_BRIDGE_EBT_MARK=m
+CONFIG_BRIDGE_EBT_PKTTYPE=m
+CONFIG_BRIDGE_EBT_STP=m
+CONFIG_BRIDGE_EBT_VLAN=m
+CONFIG_BRIDGE_EBT_ARPREPLY=m
+CONFIG_BRIDGE_EBT_DNAT=m
+CONFIG_BRIDGE_EBT_MARK_T=m
+CONFIG_BRIDGE_EBT_REDIRECT=m
+CONFIG_BRIDGE_EBT_SNAT=m
+CONFIG_BRIDGE_EBT_LOG=m
+CONFIG_BRIDGE_EBT_ULOG=m
+
+#
+# DCCP Configuration (EXPERIMENTAL)
+#
+CONFIG_IP_DCCP=m
+CONFIG_INET_DCCP_DIAG=m
+CONFIG_IP_DCCP_ACKVEC=y
+
+#
+# DCCP CCIDs Configuration (EXPERIMENTAL)
+#
+CONFIG_IP_DCCP_CCID2=m
+CONFIG_IP_DCCP_CCID3=m
+CONFIG_IP_DCCP_TFRC_LIB=m
+
+#
+# DCCP Kernel Hacking
+#
+# CONFIG_IP_DCCP_DEBUG is not set
+
+#
+# SCTP Configuration (EXPERIMENTAL)
+#
+CONFIG_IP_SCTP=m
+# CONFIG_SCTP_DBG_MSG is not set
+# CONFIG_SCTP_DBG_OBJCNT is not set
+# CONFIG_SCTP_HMAC_NONE is not set
+# CONFIG_SCTP_HMAC_SHA1 is not set
+CONFIG_SCTP_HMAC_MD5=y
+
+#
+# TIPC Configuration (EXPERIMENTAL)
+#
+# CONFIG_TIPC is not set
+CONFIG_ATM=m
+CONFIG_ATM_CLIP=m
+CONFIG_ATM_CLIP_NO_ICMP=y
+CONFIG_ATM_LANE=m
+CONFIG_ATM_MPOA=m
+CONFIG_ATM_BR2684=m
+# CONFIG_ATM_BR2684_IPFILTER is not set
+CONFIG_BRIDGE=m
+CONFIG_VLAN_8021Q=m
+CONFIG_DECNET=m
+CONFIG_DECNET_ROUTER=y
+CONFIG_DECNET_ROUTE_FWMARK=y
+CONFIG_LLC=y
+CONFIG_LLC2=m
+CONFIG_IPX=m
+# CONFIG_IPX_INTERN is not set
+CONFIG_ATALK=m
+CONFIG_DEV_APPLETALK=m
+CONFIG_IPDDP=m
+CONFIG_IPDDP_ENCAP=y
+CONFIG_IPDDP_DECAP=y
+CONFIG_X25=m
+CONFIG_LAPB=m
+CONFIG_ECONET=m
+CONFIG_ECONET_AUNUDP=y
+CONFIG_ECONET_NATIVE=y
+CONFIG_WAN_ROUTER=m
+
+#
+# QoS and/or fair queueing
+#
+CONFIG_NET_SCHED=y
+CONFIG_NET_SCH_CLK_JIFFIES=y
+# CONFIG_NET_SCH_CLK_GETTIMEOFDAY is not set
+# CONFIG_NET_SCH_CLK_CPU is not set
+
+#
+# Queueing/Scheduling
+#
+CONFIG_NET_SCH_CBQ=m
+CONFIG_NET_SCH_HTB=m
+CONFIG_NET_SCH_HFSC=m
+CONFIG_NET_SCH_ATM=m
+CONFIG_NET_SCH_PRIO=m
+CONFIG_NET_SCH_RED=m
+CONFIG_NET_SCH_SFQ=m
+CONFIG_NET_SCH_TEQL=m
+CONFIG_NET_SCH_TBF=m
+CONFIG_NET_SCH_GRED=m
+CONFIG_NET_SCH_DSMARK=m
+CONFIG_NET_SCH_NETEM=m
+CONFIG_NET_SCH_INGRESS=m
+
+#
+# Classification
+#
+CONFIG_NET_CLS=y
+CONFIG_NET_CLS_BASIC=m
+CONFIG_NET_CLS_TCINDEX=m
+CONFIG_NET_CLS_ROUTE4=m
+CONFIG_NET_CLS_ROUTE=y
+CONFIG_NET_CLS_FW=m
+CONFIG_NET_CLS_U32=m
+CONFIG_CLS_U32_PERF=y
+CONFIG_CLS_U32_MARK=y
+CONFIG_NET_CLS_RSVP=m
+CONFIG_NET_CLS_RSVP6=m
+CONFIG_NET_EMATCH=y
+CONFIG_NET_EMATCH_STACK=32
+CONFIG_NET_EMATCH_CMP=m
+CONFIG_NET_EMATCH_NBYTE=m
+CONFIG_NET_EMATCH_U32=m
+CONFIG_NET_EMATCH_META=m
+CONFIG_NET_EMATCH_TEXT=m
+# CONFIG_NET_CLS_ACT is not set
+CONFIG_NET_CLS_POLICE=y
+# CONFIG_NET_CLS_IND is not set
+CONFIG_NET_ESTIMATOR=y
+
+#
+# Network testing
+#
+CONFIG_NET_PKTGEN=m
+CONFIG_HAMRADIO=y
+
+#
+# Packet Radio protocols
+#
+CONFIG_AX25=m
+# CONFIG_AX25_DAMA_SLAVE is not set
+CONFIG_NETROM=m
+CONFIG_ROSE=m
+
+#
+# AX.25 network device drivers
+#
+CONFIG_MKISS=m
+CONFIG_6PACK=m
+CONFIG_BPQETHER=m
+CONFIG_BAYCOM_SER_FDX=m
+CONFIG_BAYCOM_SER_HDX=m
+CONFIG_BAYCOM_PAR=m
+CONFIG_BAYCOM_EPP=m
+CONFIG_YAM=m
+CONFIG_IRDA=m
+
+#
+# IrDA protocols
+#
+CONFIG_IRLAN=m
+CONFIG_IRNET=m
+CONFIG_IRCOMM=m
+# CONFIG_IRDA_ULTRA is not set
+
+#
+# IrDA options
+#
+CONFIG_IRDA_CACHE_LAST_LSAP=y
+CONFIG_IRDA_FAST_RR=y
+CONFIG_IRDA_DEBUG=y
+
+#
+# Infrared-port device drivers
+#
+
+#
+# SIR device drivers
+#
+CONFIG_IRTTY_SIR=m
+
+#
+# Dongle support
+#
+CONFIG_DONGLE=y
+CONFIG_ESI_DONGLE=m
+CONFIG_ACTISYS_DONGLE=m
+CONFIG_TEKRAM_DONGLE=m
+CONFIG_TOIM3232_DONGLE=m
+CONFIG_LITELINK_DONGLE=m
+CONFIG_MA600_DONGLE=m
+CONFIG_GIRBIL_DONGLE=m
+CONFIG_MCP2120_DONGLE=m
+CONFIG_OLD_BELKIN_DONGLE=m
+CONFIG_ACT200L_DONGLE=m
+
+#
+# Old SIR device drivers
+#
+
+#
+# Old Serial dongle support
+#
+
+#
+# FIR device drivers
+#
+CONFIG_USB_IRDA=m
+CONFIG_SIGMATEL_FIR=m
+CONFIG_NSC_FIR=m
+CONFIG_WINBOND_FIR=m
+CONFIG_TOSHIBA_FIR=m
+CONFIG_SMC_IRCC_FIR=m
+CONFIG_ALI_FIR=m
+CONFIG_VLSI_FIR=m
+CONFIG_VIA_FIR=m
+CONFIG_MCS_FIR=m
+CONFIG_BT=m
+CONFIG_BT_L2CAP=m
+CONFIG_BT_SCO=m
+CONFIG_BT_RFCOMM=m
+CONFIG_BT_RFCOMM_TTY=y
+CONFIG_BT_BNEP=m
+CONFIG_BT_BNEP_MC_FILTER=y
+CONFIG_BT_BNEP_PROTO_FILTER=y
+CONFIG_BT_CMTP=m
+CONFIG_BT_HIDP=m
+
+#
+# Bluetooth device drivers
+#
+CONFIG_BT_HCIUSB=m
+CONFIG_BT_HCIUSB_SCO=y
+CONFIG_BT_HCIUART=m
+CONFIG_BT_HCIUART_H4=y
+CONFIG_BT_HCIUART_BCSP=y
+CONFIG_BT_HCIBCM203X=m
+CONFIG_BT_HCIBPA10X=m
+CONFIG_BT_HCIBFUSB=m
+CONFIG_BT_HCIDTL1=m
+CONFIG_BT_HCIBT3C=m
+CONFIG_BT_HCIBLUECARD=m
+CONFIG_BT_HCIBTUART=m
+CONFIG_BT_HCIVHCI=m
+CONFIG_IEEE80211=m
+# CONFIG_IEEE80211_DEBUG is not set
+CONFIG_IEEE80211_CRYPT_WEP=m
+CONFIG_IEEE80211_CRYPT_CCMP=m
+CONFIG_IEEE80211_CRYPT_TKIP=m
+CONFIG_IEEE80211_SOFTMAC=m
+# CONFIG_IEEE80211_SOFTMAC_DEBUG is not set
+CONFIG_WIRELESS_EXT=y
+
+#
+# Device Drivers
+#
+
+#
+# Generic Driver Options
+#
+CONFIG_STANDALONE=y
+CONFIG_PREVENT_FIRMWARE_BUILD=y
+CONFIG_FW_LOADER=m
+# CONFIG_DEBUG_DRIVER is not set
+CONFIG_SYS_HYPERVISOR=y
+
+#
+# Connector - unified userspace <-> kernelspace linker
+#
+CONFIG_CONNECTOR=m
+
+#
+# Memory Technology Devices (MTD)
+#
+CONFIG_MTD=m
+# CONFIG_MTD_DEBUG is not set
+CONFIG_MTD_CONCAT=m
+CONFIG_MTD_PARTITIONS=y
+CONFIG_MTD_REDBOOT_PARTS=m
+CONFIG_MTD_REDBOOT_DIRECTORY_BLOCK=-1
+# CONFIG_MTD_REDBOOT_PARTS_UNALLOCATED is not set
+# CONFIG_MTD_REDBOOT_PARTS_READONLY is not set
+# CONFIG_MTD_CMDLINE_PARTS is not set
+
+#
+# User Modules And Translation Layers
+#
+CONFIG_MTD_CHAR=m
+CONFIG_MTD_BLOCK=m
+CONFIG_MTD_BLOCK_RO=m
+CONFIG_FTL=m
+CONFIG_NFTL=m
+CONFIG_NFTL_RW=y
+CONFIG_INFTL=m
+CONFIG_RFD_FTL=m
+
+#
+# RAM/ROM/Flash chip drivers
+#
+CONFIG_MTD_CFI=m
+CONFIG_MTD_JEDECPROBE=m
+CONFIG_MTD_GEN_PROBE=m
+CONFIG_MTD_CFI_ADV_OPTIONS=y
+CONFIG_MTD_CFI_NOSWAP=y
+# CONFIG_MTD_CFI_BE_BYTE_SWAP is not set
+# CONFIG_MTD_CFI_LE_BYTE_SWAP is not set
+# CONFIG_MTD_CFI_GEOMETRY is not set
+CONFIG_MTD_MAP_BANK_WIDTH_1=y
+CONFIG_MTD_MAP_BANK_WIDTH_2=y
+CONFIG_MTD_MAP_BANK_WIDTH_4=y
+# CONFIG_MTD_MAP_BANK_WIDTH_8 is not set
+# CONFIG_MTD_MAP_BANK_WIDTH_16 is not set
+# CONFIG_MTD_MAP_BANK_WIDTH_32 is not set
+CONFIG_MTD_CFI_I1=y
+CONFIG_MTD_CFI_I2=y
+# CONFIG_MTD_CFI_I4 is not set
+# CONFIG_MTD_CFI_I8 is not set
+# CONFIG_MTD_OTP is not set
+CONFIG_MTD_CFI_INTELEXT=m
+CONFIG_MTD_CFI_AMDSTD=m
+CONFIG_MTD_CFI_STAA=m
+CONFIG_MTD_CFI_UTIL=m
+CONFIG_MTD_RAM=m
+CONFIG_MTD_ROM=m
+CONFIG_MTD_ABSENT=m
+# CONFIG_MTD_OBSOLETE_CHIPS is not set
+
+#
+# Mapping drivers for chip access
+#
+CONFIG_MTD_COMPLEX_MAPPINGS=y
+CONFIG_MTD_PHYSMAP=m
+CONFIG_MTD_PHYSMAP_START=0x8000000
+CONFIG_MTD_PHYSMAP_LEN=0x4000000
+CONFIG_MTD_PHYSMAP_BANKWIDTH=2
+CONFIG_MTD_PNC2000=m
+CONFIG_MTD_SC520CDP=m
+CONFIG_MTD_NETSC520=m
+CONFIG_MTD_TS5500=m
+CONFIG_MTD_SBC_GXX=m
+CONFIG_MTD_SCx200_DOCFLASH=m
+CONFIG_MTD_AMD76XROM=m
+CONFIG_MTD_ICHXROM=m
+CONFIG_MTD_SCB2_FLASH=m
+CONFIG_MTD_NETtel=m
+CONFIG_MTD_DILNETPC=m
+CONFIG_MTD_DILNETPC_BOOTSIZE=0x80000
+CONFIG_MTD_L440GX=m
+CONFIG_MTD_PCI=m
+CONFIG_MTD_PLATRAM=m
+
+#
+# Self-contained MTD device drivers
+#
+CONFIG_MTD_PMC551=m
+# CONFIG_MTD_PMC551_BUGFIX is not set
+# CONFIG_MTD_PMC551_DEBUG is not set
+CONFIG_MTD_DATAFLASH=m
+CONFIG_MTD_M25P80=m
+CONFIG_MTD_SLRAM=m
+CONFIG_MTD_PHRAM=m
+CONFIG_MTD_MTDRAM=m
+CONFIG_MTDRAM_TOTAL_SIZE=4096
+CONFIG_MTDRAM_ERASE_SIZE=128
+CONFIG_MTD_BLOCK2MTD=m
+
+#
+# Disk-On-Chip Device Drivers
+#
+CONFIG_MTD_DOC2000=m
+CONFIG_MTD_DOC2001=m
+CONFIG_MTD_DOC2001PLUS=m
+CONFIG_MTD_DOCPROBE=m
+CONFIG_MTD_DOCECC=m
+# CONFIG_MTD_DOCPROBE_ADVANCED is not set
+CONFIG_MTD_DOCPROBE_ADDRESS=0
+
+#
+# NAND Flash Device Drivers
+#
+CONFIG_MTD_NAND=m
+# CONFIG_MTD_NAND_VERIFY_WRITE is not set
+# CONFIG_MTD_NAND_ECC_SMC is not set
+CONFIG_MTD_NAND_IDS=m
+CONFIG_MTD_NAND_DISKONCHIP=m
+# CONFIG_MTD_NAND_DISKONCHIP_PROBE_ADVANCED is not set
+CONFIG_MTD_NAND_DISKONCHIP_PROBE_ADDRESS=0
+CONFIG_MTD_NAND_DISKONCHIP_BBTWRITE=y
+CONFIG_MTD_NAND_NANDSIM=m
+
+#
+# OneNAND Flash Device Drivers
+#
+CONFIG_MTD_ONENAND=m
+# CONFIG_MTD_ONENAND_VERIFY_WRITE is not set
+# CONFIG_MTD_ONENAND_OTP is not set
+
+#
+# Parallel port support
+#
+CONFIG_PARPORT=m
+CONFIG_PARPORT_PC=m
+CONFIG_PARPORT_SERIAL=m
+CONFIG_PARPORT_PC_FIFO=y
+CONFIG_PARPORT_PC_SUPERIO=y
+CONFIG_PARPORT_PC_PCMCIA=m
+CONFIG_PARPORT_NOT_PC=y
+# CONFIG_PARPORT_GSC is not set
+CONFIG_PARPORT_AX88796=m
+CONFIG_PARPORT_1284=y
+
+#
+# Plug and Play support
+#
+CONFIG_PNP=y
+# CONFIG_PNP_DEBUG is not set
+
+#
+# Protocols
+#
+CONFIG_PNPACPI=y
+
+#
+# Block devices
+#
+CONFIG_BLK_DEV_FD=y
+CONFIG_PARIDE=m
+CONFIG_PARIDE_PARPORT=m
+
+#
+# Parallel IDE high-level drivers
+#
+CONFIG_PARIDE_PD=m
+CONFIG_PARIDE_PCD=m
+CONFIG_PARIDE_PF=m
+CONFIG_PARIDE_PT=m
+CONFIG_PARIDE_PG=m
+
+#
+# Parallel IDE protocol modules
+#
+CONFIG_PARIDE_ATEN=m
+CONFIG_PARIDE_BPCK=m
+CONFIG_PARIDE_BPCK6=m
+CONFIG_PARIDE_COMM=m
+CONFIG_PARIDE_DSTR=m
+CONFIG_PARIDE_FIT2=m
+CONFIG_PARIDE_FIT3=m
+CONFIG_PARIDE_EPAT=m
+CONFIG_PARIDE_EPATC8=y
+CONFIG_PARIDE_EPIA=m
+CONFIG_PARIDE_FRIQ=m
+CONFIG_PARIDE_FRPW=m
+CONFIG_PARIDE_KBIC=m
+CONFIG_PARIDE_KTTI=m
+CONFIG_PARIDE_ON20=m
+CONFIG_PARIDE_ON26=m
+CONFIG_BLK_CPQ_DA=m
+CONFIG_BLK_CPQ_CISS_DA=m
+CONFIG_CISS_SCSI_TAPE=y
+CONFIG_BLK_DEV_DAC960=m
+CONFIG_BLK_DEV_UMEM=m
+# CONFIG_BLK_DEV_COW_COMMON is not set
+CONFIG_BLK_DEV_LOOP=y
+CONFIG_BLK_DEV_CRYPTOLOOP=m
+CONFIG_BLK_DEV_NBD=m
+CONFIG_BLK_DEV_SX8=m
+# CONFIG_BLK_DEV_UB is not set
+CONFIG_BLK_DEV_RAM=y
+CONFIG_BLK_DEV_RAM_COUNT=16
+CONFIG_BLK_DEV_RAM_SIZE=16384
+CONFIG_BLK_DEV_RAM_BLOCKSIZE=1024
+CONFIG_BLK_DEV_INITRD=y
+CONFIG_CDROM_PKTCDVD=m
+CONFIG_CDROM_PKTCDVD_BUFFERS=8
+CONFIG_CDROM_PKTCDVD_WCACHE=y
+CONFIG_ATA_OVER_ETH=m
+
+#
+# ATA/ATAPI/MFM/RLL support
+#
+CONFIG_IDE=y
+CONFIG_BLK_DEV_IDE=y
+
+#
+# Please see Documentation/ide.txt for help/info on IDE drives
+#
+# CONFIG_BLK_DEV_IDE_SATA is not set
+# CONFIG_BLK_DEV_HD_IDE is not set
+CONFIG_BLK_DEV_IDEDISK=m
+CONFIG_IDEDISK_MULTI_MODE=y
+CONFIG_BLK_DEV_IDECS=m
+CONFIG_BLK_DEV_IDECD=m
+CONFIG_BLK_DEV_IDETAPE=m
+CONFIG_BLK_DEV_IDEFLOPPY=m
+CONFIG_BLK_DEV_IDESCSI=m
+# CONFIG_IDE_TASK_IOCTL is not set
+
+#
+# IDE chipset support/bugfixes
+#
+CONFIG_IDE_GENERIC=m
+CONFIG_BLK_DEV_CMD640=y
+CONFIG_BLK_DEV_CMD640_ENHANCED=y
+CONFIG_BLK_DEV_IDEPNP=y
+CONFIG_BLK_DEV_IDEPCI=y
+CONFIG_IDEPCI_SHARE_IRQ=y
+CONFIG_BLK_DEV_OFFBOARD=y
+CONFIG_BLK_DEV_GENERIC=y
+CONFIG_BLK_DEV_OPTI621=m
+CONFIG_BLK_DEV_RZ1000=m
+CONFIG_BLK_DEV_IDEDMA_PCI=y
+# CONFIG_BLK_DEV_IDEDMA_FORCED is not set
+CONFIG_IDEDMA_PCI_AUTO=y
+# CONFIG_IDEDMA_ONLYDISK is not set
+CONFIG_BLK_DEV_AEC62XX=m
+CONFIG_BLK_DEV_ALI15X3=m
+# CONFIG_WDC_ALI15X3 is not set
+CONFIG_BLK_DEV_AMD74XX=m
+CONFIG_BLK_DEV_ATIIXP=m
+CONFIG_BLK_DEV_CMD64X=m
+CONFIG_BLK_DEV_TRIFLEX=m
+CONFIG_BLK_DEV_CY82C693=m
+CONFIG_BLK_DEV_CS5520=m
+CONFIG_BLK_DEV_CS5530=m
+CONFIG_BLK_DEV_CS5535=m
+CONFIG_BLK_DEV_HPT34X=m
+CONFIG_HPT34X_AUTODMA=y
+CONFIG_BLK_DEV_HPT366=m
+CONFIG_BLK_DEV_SC1200=m
+CONFIG_BLK_DEV_PIIX=m
+CONFIG_BLK_DEV_IT821X=m
+CONFIG_BLK_DEV_NS87415=m
+CONFIG_BLK_DEV_PDC202XX_OLD=m
+CONFIG_PDC202XX_BURST=y
+CONFIG_BLK_DEV_PDC202XX_NEW=m
+CONFIG_BLK_DEV_SVWKS=m
+CONFIG_BLK_DEV_SIIMAGE=m
+CONFIG_BLK_DEV_SIS5513=m
+CONFIG_BLK_DEV_SLC90E66=m
+CONFIG_BLK_DEV_TRM290=m
+CONFIG_BLK_DEV_VIA82CXXX=m
+# CONFIG_IDE_ARM is not set
+CONFIG_BLK_DEV_IDEDMA=y
+# CONFIG_IDEDMA_IVB is not set
+CONFIG_IDEDMA_AUTO=y
+# CONFIG_BLK_DEV_HD is not set
+
+#
+# SCSI device support
+#
+CONFIG_RAID_ATTRS=m
+CONFIG_SCSI=m
+CONFIG_SCSI_PROC_FS=y
+
+#
+# SCSI support type (disk, tape, CD-ROM)
+#
+CONFIG_BLK_DEV_SD=m
+CONFIG_CHR_DEV_ST=m
+CONFIG_CHR_DEV_OSST=m
+CONFIG_BLK_DEV_SR=m
+# CONFIG_BLK_DEV_SR_VENDOR is not set
+CONFIG_CHR_DEV_SG=m
+CONFIG_CHR_DEV_SCH=m
+
+#
+# Some SCSI devices (e.g. CD jukebox) support multiple LUNs
+#
+CONFIG_SCSI_MULTI_LUN=y
+CONFIG_SCSI_CONSTANTS=y
+CONFIG_SCSI_LOGGING=y
+
+#
+# SCSI Transport Attributes
+#
+CONFIG_SCSI_SPI_ATTRS=m
+CONFIG_SCSI_FC_ATTRS=m
+CONFIG_SCSI_ISCSI_ATTRS=m
+CONFIG_SCSI_SAS_ATTRS=m
+
+#
+# SCSI low-level drivers
+#
+CONFIG_ISCSI_TCP=m
+CONFIG_BLK_DEV_3W_XXXX_RAID=m
+CONFIG_SCSI_3W_9XXX=m
+CONFIG_SCSI_ACARD=m
+CONFIG_SCSI_AACRAID=m
+CONFIG_SCSI_AIC7XXX=m
+CONFIG_AIC7XXX_CMDS_PER_DEVICE=8
+CONFIG_AIC7XXX_RESET_DELAY_MS=15000
+CONFIG_AIC7XXX_DEBUG_ENABLE=y
+CONFIG_AIC7XXX_DEBUG_MASK=0
+CONFIG_AIC7XXX_REG_PRETTY_PRINT=y
+CONFIG_SCSI_AIC7XXX_OLD=m
+CONFIG_SCSI_AIC79XX=m
+CONFIG_AIC79XX_CMDS_PER_DEVICE=32
+CONFIG_AIC79XX_RESET_DELAY_MS=15000
+CONFIG_AIC79XX_ENABLE_RD_STRM=y
+CONFIG_AIC79XX_DEBUG_ENABLE=y
+CONFIG_AIC79XX_DEBUG_MASK=0
+CONFIG_AIC79XX_REG_PRETTY_PRINT=y
+CONFIG_SCSI_DPT_I2O=m
+CONFIG_SCSI_ADVANSYS=m
+CONFIG_MEGARAID_NEWGEN=y
+CONFIG_MEGARAID_MM=m
+CONFIG_MEGARAID_MAILBOX=m
+CONFIG_MEGARAID_LEGACY=m
+CONFIG_MEGARAID_SAS=m
+CONFIG_SCSI_SATA=m
+CONFIG_SCSI_SATA_AHCI=m
+CONFIG_SCSI_SATA_SVW=m
+CONFIG_SCSI_ATA_PIIX=m
+CONFIG_SCSI_SATA_MV=m
+CONFIG_SCSI_SATA_NV=m
+CONFIG_SCSI_PDC_ADMA=m
+CONFIG_SCSI_HPTIOP=m
+CONFIG_SCSI_SATA_QSTOR=m
+CONFIG_SCSI_SATA_PROMISE=m
+CONFIG_SCSI_SATA_SX4=m
+CONFIG_SCSI_SATA_SIL=m
+CONFIG_SCSI_SATA_SIL24=m
+CONFIG_SCSI_SATA_SIS=m
+CONFIG_SCSI_SATA_ULI=m
+CONFIG_SCSI_SATA_VIA=m
+CONFIG_SCSI_SATA_VITESSE=m
+CONFIG_SCSI_SATA_INTEL_COMBINED=y
+CONFIG_SCSI_BUSLOGIC=m
+# CONFIG_SCSI_OMIT_FLASHPOINT is not set
+CONFIG_SCSI_DMX3191D=m
+CONFIG_SCSI_EATA=m
+CONFIG_SCSI_EATA_TAGGED_QUEUE=y
+CONFIG_SCSI_EATA_LINKED_COMMANDS=y
+CONFIG_SCSI_EATA_MAX_TAGS=16
+CONFIG_SCSI_FUTURE_DOMAIN=m
+CONFIG_SCSI_GDTH=m
+CONFIG_SCSI_IPS=m
+CONFIG_SCSI_INITIO=m
+CONFIG_SCSI_INIA100=m
+CONFIG_SCSI_PPA=m
+CONFIG_SCSI_IMM=m
+# CONFIG_SCSI_IZIP_EPP16 is not set
+# CONFIG_SCSI_IZIP_SLOW_CTR is not set
+CONFIG_SCSI_SYM53C8XX_2=m
+CONFIG_SCSI_SYM53C8XX_DMA_ADDRESSING_MODE=1
+CONFIG_SCSI_SYM53C8XX_DEFAULT_TAGS=16
+CONFIG_SCSI_SYM53C8XX_MAX_TAGS=64
+CONFIG_SCSI_SYM53C8XX_MMIO=y
+CONFIG_SCSI_IPR=m
+CONFIG_SCSI_IPR_TRACE=y
+CONFIG_SCSI_IPR_DUMP=y
+CONFIG_SCSI_QLOGIC_1280=m
+CONFIG_SCSI_QLA_FC=m
+CONFIG_SCSI_LPFC=m
+CONFIG_SCSI_DC395x=m
+CONFIG_SCSI_DC390T=m
+CONFIG_SCSI_NSP32=m
+CONFIG_SCSI_DEBUG=m
+
+#
+# PCMCIA SCSI adapter support
+#
+CONFIG_PCMCIA_AHA152X=m
+CONFIG_PCMCIA_FDOMAIN=m
+CONFIG_PCMCIA_NINJA_SCSI=m
+CONFIG_PCMCIA_QLOGIC=m
+CONFIG_PCMCIA_SYM53C500=m
+
+#
+# Multi-device support (RAID and LVM)
+#
+CONFIG_MD=y
+CONFIG_BLK_DEV_MD=y
+CONFIG_MD_LINEAR=m
+CONFIG_MD_RAID0=m
+CONFIG_MD_RAID1=m
+CONFIG_MD_RAID10=m
+CONFIG_MD_RAID456=m
+CONFIG_MD_RAID5_RESHAPE=y
+CONFIG_MD_MULTIPATH=m
+CONFIG_MD_FAULTY=m
+CONFIG_BLK_DEV_DM=m
+CONFIG_DM_CRYPT=m
+CONFIG_DM_SNAPSHOT=m
+CONFIG_DM_MIRROR=m
+CONFIG_DM_ZERO=m
+CONFIG_DM_MULTIPATH=m
+CONFIG_DM_MULTIPATH_EMC=m
+
+#
+# Fusion MPT device support
+#
+CONFIG_FUSION=y
+CONFIG_FUSION_SPI=m
+CONFIG_FUSION_FC=m
+CONFIG_FUSION_SAS=m
+CONFIG_FUSION_MAX_SGE=128
+CONFIG_FUSION_CTL=m
+CONFIG_FUSION_LAN=m
+
+#
+# IEEE 1394 (FireWire) support
+#
+CONFIG_IEEE1394=m
+
+#
+# Subsystem Options
+#
+# CONFIG_IEEE1394_VERBOSEDEBUG is not set
+# CONFIG_IEEE1394_OUI_DB is not set
+CONFIG_IEEE1394_EXTRA_CONFIG_ROMS=y
+CONFIG_IEEE1394_CONFIG_ROM_IP1394=y
+CONFIG_IEEE1394_EXPORT_FULL_API=y
+
+#
+# Device Drivers
+#
+CONFIG_IEEE1394_PCILYNX=m
+CONFIG_IEEE1394_OHCI1394=m
+
+#
+# Protocol Drivers
+#
+CONFIG_IEEE1394_VIDEO1394=m
+CONFIG_IEEE1394_SBP2=m
+# CONFIG_IEEE1394_SBP2_PHYS_DMA is not set
+CONFIG_IEEE1394_ETH1394=m
+CONFIG_IEEE1394_DV1394=m
+CONFIG_IEEE1394_RAWIO=m
+
+#
+# I2O device support
+#
+CONFIG_I2O=m
+CONFIG_I2O_LCT_NOTIFY_ON_CHANGES=y
+CONFIG_I2O_EXT_ADAPTEC=y
+CONFIG_I2O_EXT_ADAPTEC_DMA64=y
+CONFIG_I2O_CONFIG=m
+CONFIG_I2O_CONFIG_OLD_IOCTL=y
+CONFIG_I2O_BUS=m
+CONFIG_I2O_BLOCK=m
+CONFIG_I2O_SCSI=m
+CONFIG_I2O_PROC=m
+
+#
+# Network device support
+#
+CONFIG_NETDEVICES=y
+CONFIG_DUMMY=m
+CONFIG_BONDING=m
+CONFIG_EQUALIZER=m
+CONFIG_TUN=m
+CONFIG_NET_SB1000=m
+
+#
+# ARCnet devices
+#
+CONFIG_ARCNET=m
+CONFIG_ARCNET_1201=m
+CONFIG_ARCNET_1051=m
+CONFIG_ARCNET_RAW=m
+CONFIG_ARCNET_CAP=m
+CONFIG_ARCNET_COM90xx=m
+CONFIG_ARCNET_COM90xxIO=m
+CONFIG_ARCNET_RIM_I=m
+# CONFIG_ARCNET_COM20020 is not set
+
+#
+# PHY device support
+#
+CONFIG_PHYLIB=m
+
+#
+# MII PHY device drivers
+#
+CONFIG_MARVELL_PHY=m
+CONFIG_DAVICOM_PHY=m
+CONFIG_QSEMI_PHY=m
+CONFIG_LXT_PHY=m
+CONFIG_CICADA_PHY=m
+CONFIG_VITESSE_PHY=m
+CONFIG_SMSC_PHY=m
+CONFIG_FIXED_PHY=m
+# CONFIG_FIXED_MII_10_FDX is not set
+# CONFIG_FIXED_MII_100_FDX is not set
+
+#
+# Ethernet (10 or 100Mbit)
+#
+CONFIG_NET_ETHERNET=y
+CONFIG_MII=m
+CONFIG_HAPPYMEAL=m
+CONFIG_SUNGEM=m
+CONFIG_CASSINI=m
+CONFIG_NET_VENDOR_3COM=y
+CONFIG_VORTEX=m
+CONFIG_TYPHOON=m
+
+#
+# Tulip family network device support
+#
+CONFIG_NET_TULIP=y
+CONFIG_DE2104X=m
+CONFIG_TULIP=m
+# CONFIG_TULIP_MWI is not set
+# CONFIG_TULIP_MMIO is not set
+CONFIG_TULIP_NAPI=y
+CONFIG_TULIP_NAPI_HW_MITIGATION=y
+CONFIG_DE4X5=m
+CONFIG_WINBOND_840=m
+CONFIG_DM9102=m
+CONFIG_ULI526X=m
+CONFIG_PCMCIA_XIRCOM=m
+CONFIG_HP100=m
+CONFIG_NET_PCI=y
+CONFIG_PCNET32=m
+CONFIG_AMD8111_ETH=m
+# CONFIG_AMD8111E_NAPI is not set
+CONFIG_ADAPTEC_STARFIRE=m
+CONFIG_ADAPTEC_STARFIRE_NAPI=y
+CONFIG_B44=m
+CONFIG_FORCEDETH=m
+CONFIG_DGRS=m
+CONFIG_EEPRO100=m
+CONFIG_E100=m
+CONFIG_FEALNX=m
+CONFIG_NATSEMI=m
+CONFIG_NE2K_PCI=m
+CONFIG_8139CP=m
+CONFIG_8139TOO=m
+# CONFIG_8139TOO_PIO is not set
+# CONFIG_8139TOO_TUNE_TWISTER is not set
+CONFIG_8139TOO_8129=y
+# CONFIG_8139_OLD_RX_RESET is not set
+CONFIG_SIS900=m
+CONFIG_EPIC100=m
+CONFIG_SUNDANCE=m
+# CONFIG_SUNDANCE_MMIO is not set
+CONFIG_TLAN=m
+CONFIG_VIA_RHINE=m
+# CONFIG_VIA_RHINE_MMIO is not set
+# CONFIG_VIA_RHINE_NAPI is not set
+CONFIG_NET_POCKET=y
+CONFIG_ATP=m
+CONFIG_DE600=m
+CONFIG_DE620=m
+
+#
+# Ethernet (1000 Mbit)
+#
+CONFIG_ACENIC=m
+# CONFIG_ACENIC_OMIT_TIGON_I is not set
+CONFIG_DL2K=m
+CONFIG_E1000=m
+CONFIG_E1000_NAPI=y
+# CONFIG_E1000_DISABLE_PACKET_SPLIT is not set
+CONFIG_NS83820=m
+CONFIG_HAMACHI=m
+CONFIG_YELLOWFIN=m
+CONFIG_R8169=m
+# CONFIG_R8169_NAPI is not set
+CONFIG_R8169_VLAN=y
+CONFIG_SIS190=m
+CONFIG_SKGE=m
+CONFIG_SKY2=m
+CONFIG_SK98LIN=m
+CONFIG_VIA_VELOCITY=m
+CONFIG_TIGON3=m
+CONFIG_BNX2=m
+
+#
+# Ethernet (10000 Mbit)
+#
+CONFIG_CHELSIO_T1=m
+CONFIG_IXGB=m
+CONFIG_IXGB_NAPI=y
+CONFIG_S2IO=m
+CONFIG_S2IO_NAPI=y
+CONFIG_MYRI10GE=m
+# CONFIG_SFC is not set
+
+#
+# Token Ring devices
+#
+CONFIG_TR=y
+CONFIG_IBMOL=m
+CONFIG_IBMLS=m
+CONFIG_3C359=m
+CONFIG_TMS380TR=m
+CONFIG_TMSPCI=m
+CONFIG_ABYSS=m
+
+#
+# Wireless LAN (non-hamradio)
+#
+CONFIG_NET_RADIO=y
+CONFIG_NET_WIRELESS_RTNETLINK=y
+
+#
+# Obsolete Wireless cards support (pre-802.11)
+#
+CONFIG_STRIP=m
+CONFIG_PCMCIA_WAVELAN=m
+CONFIG_PCMCIA_NETWAVE=m
+
+#
+# Wireless 802.11 Frequency Hopping cards support
+#
+CONFIG_PCMCIA_RAYCS=m
+
+#
+# Wireless 802.11b ISA/PCI cards support
+#
+CONFIG_IPW2100=m
+CONFIG_IPW2100_MONITOR=y
+# CONFIG_IPW2100_DEBUG is not set
+CONFIG_IPW2200=m
+CONFIG_IPW2200_MONITOR=y
+CONFIG_IPW2200_RADIOTAP=y
+CONFIG_IPW2200_PROMISCUOUS=y
+CONFIG_IPW2200_QOS=y
+# CONFIG_IPW2200_DEBUG is not set
+CONFIG_AIRO=m
+CONFIG_HERMES=m
+CONFIG_PLX_HERMES=m
+CONFIG_TMD_HERMES=m
+CONFIG_NORTEL_HERMES=m
+CONFIG_PCI_HERMES=m
+CONFIG_ATMEL=m
+CONFIG_PCI_ATMEL=m
+
+#
+# Wireless 802.11b Pcmcia/Cardbus cards support
+#
+CONFIG_PCMCIA_HERMES=m
+CONFIG_PCMCIA_SPECTRUM=m
+CONFIG_AIRO_CS=m
+CONFIG_PCMCIA_ATMEL=m
+CONFIG_PCMCIA_WL3501=m
+
+#
+# Prism GT/Duette 802.11(a/b/g) PCI/Cardbus support
+#
+CONFIG_PRISM54=m
+CONFIG_USB_ZD1201=m
+CONFIG_HOSTAP=m
+CONFIG_HOSTAP_FIRMWARE=y
+CONFIG_HOSTAP_FIRMWARE_NVRAM=y
+CONFIG_HOSTAP_PLX=m
+CONFIG_HOSTAP_PCI=m
+CONFIG_HOSTAP_CS=m
+CONFIG_BCM43XX=m
+CONFIG_BCM43XX_DEBUG=y
+CONFIG_BCM43XX_DMA=y
+CONFIG_BCM43XX_PIO=y
+CONFIG_BCM43XX_DMA_AND_PIO_MODE=y
+# CONFIG_BCM43XX_DMA_MODE is not set
+# CONFIG_BCM43XX_PIO_MODE is not set
+CONFIG_ZD1211RW=m
+# CONFIG_ZD1211RW_DEBUG is not set
+CONFIG_NET_WIRELESS=y
+
+#
+# PCMCIA network device support
+#
+CONFIG_NET_PCMCIA=y
+CONFIG_PCMCIA_3C589=m
+CONFIG_PCMCIA_3C574=m
+CONFIG_PCMCIA_FMVJ18X=m
+CONFIG_PCMCIA_PCNET=m
+CONFIG_PCMCIA_NMCLAN=m
+CONFIG_PCMCIA_SMC91C92=m
+CONFIG_PCMCIA_XIRC2PS=m
+CONFIG_PCMCIA_AXNET=m
+CONFIG_PCMCIA_IBMTR=m
+
+#
+# Wan interfaces
+#
+CONFIG_WAN=y
+CONFIG_DSCC4=m
+CONFIG_DSCC4_PCISYNC=y
+CONFIG_DSCC4_PCI_RST=y
+CONFIG_LANMEDIA=m
+CONFIG_HDLC=m
+CONFIG_HDLC_RAW=y
+CONFIG_HDLC_RAW_ETH=y
+CONFIG_HDLC_CISCO=y
+CONFIG_HDLC_FR=y
+CONFIG_HDLC_PPP=y
+CONFIG_HDLC_X25=y
+CONFIG_PCI200SYN=m
+CONFIG_WANXL=m
+CONFIG_PC300=m
+CONFIG_PC300_MLPPP=y
+CONFIG_FARSYNC=m
+CONFIG_DLCI=m
+CONFIG_DLCI_COUNT=24
+CONFIG_DLCI_MAX=8
+CONFIG_WAN_ROUTER_DRIVERS=y
+CONFIG_CYCLADES_SYNC=m
+CONFIG_CYCLOMX_X25=y
+CONFIG_LAPBETHER=m
+CONFIG_X25_ASY=m
+CONFIG_SBNI=m
+# CONFIG_SBNI_MULTILINE is not set
+
+#
+# ATM drivers
+#
+CONFIG_ATM_DUMMY=m
+CONFIG_ATM_TCP=m
+CONFIG_ATM_LANAI=m
+CONFIG_ATM_ENI=m
+# CONFIG_ATM_ENI_DEBUG is not set
+# CONFIG_ATM_ENI_TUNE_BURST is not set
+CONFIG_ATM_FIRESTREAM=m
+CONFIG_ATM_ZATM=m
+# CONFIG_ATM_ZATM_DEBUG is not set
+CONFIG_ATM_NICSTAR=m
+CONFIG_ATM_NICSTAR_USE_SUNI=y
+CONFIG_ATM_NICSTAR_USE_IDT77105=y
+CONFIG_ATM_IDT77252=m
+# CONFIG_ATM_IDT77252_DEBUG is not set
+CONFIG_ATM_IDT77252_RCV_ALL=y
+CONFIG_ATM_IDT77252_USE_SUNI=y
+CONFIG_ATM_AMBASSADOR=m
+# CONFIG_ATM_AMBASSADOR_DEBUG is not set
+CONFIG_ATM_HORIZON=m
+# CONFIG_ATM_HORIZON_DEBUG is not set
+CONFIG_ATM_IA=m
+# CONFIG_ATM_IA_DEBUG is not set
+CONFIG_ATM_FORE200E_MAYBE=m
+CONFIG_ATM_FORE200E_PCA=y
+CONFIG_ATM_FORE200E_PCA_DEFAULT_FW=y
+CONFIG_ATM_FORE200E_USE_TASKLET=y
+CONFIG_ATM_FORE200E_TX_RETRY=16
+CONFIG_ATM_FORE200E_DEBUG=0
+CONFIG_ATM_FORE200E=m
+CONFIG_ATM_HE=m
+CONFIG_ATM_HE_USE_SUNI=y
+CONFIG_FDDI=y
+# CONFIG_DEFXX is not set
+CONFIG_SKFP=m
+CONFIG_HIPPI=y
+CONFIG_ROADRUNNER=m
+CONFIG_ROADRUNNER_LARGE_RINGS=y
+CONFIG_PLIP=m
+CONFIG_PPP=m
+CONFIG_PPP_MULTILINK=y
+CONFIG_PPP_FILTER=y
+CONFIG_PPP_ASYNC=m
+CONFIG_PPP_SYNC_TTY=m
+CONFIG_PPP_DEFLATE=m
+CONFIG_PPP_BSDCOMP=m
+CONFIG_PPP_MPPE=m
+CONFIG_PPPOE=m
+CONFIG_PPPOATM=m
+CONFIG_SLIP=m
+CONFIG_SLIP_COMPRESSED=y
+CONFIG_SLIP_SMART=y
+CONFIG_SLIP_MODE_SLIP6=y
+CONFIG_NET_FC=y
+CONFIG_SHAPER=m
+CONFIG_NETCONSOLE=m
+CONFIG_NETPOLL=y
+CONFIG_NETPOLL_RX=y
+CONFIG_NETPOLL_TRAP=y
+CONFIG_NET_POLL_CONTROLLER=y
+
+#
+# ISDN subsystem
+#
+CONFIG_ISDN=m
+
+#
+# Old ISDN4Linux
+#
+CONFIG_ISDN_I4L=m
+CONFIG_ISDN_PPP=y
+CONFIG_ISDN_PPP_VJ=y
+CONFIG_ISDN_MPP=y
+CONFIG_IPPP_FILTER=y
+CONFIG_ISDN_PPP_BSDCOMP=m
+CONFIG_ISDN_AUDIO=y
+CONFIG_ISDN_TTY_FAX=y
+CONFIG_ISDN_X25=y
+
+#
+# ISDN feature submodules
+#
+CONFIG_ISDN_DIVERSION=m
+
+#
+# ISDN4Linux hardware drivers
+#
+
+#
+# Passive cards
+#
+CONFIG_ISDN_DRV_HISAX=m
+
+#
+# D-channel protocol features
+#
+CONFIG_HISAX_EURO=y
+CONFIG_DE_AOC=y
+# CONFIG_HISAX_NO_SENDCOMPLETE is not set
+# CONFIG_HISAX_NO_LLC is not set
+# CONFIG_HISAX_NO_KEYPAD is not set
+CONFIG_HISAX_1TR6=y
+CONFIG_HISAX_NI1=y
+CONFIG_HISAX_MAX_CARDS=8
+
+#
+# HiSax supported cards
+#
+CONFIG_HISAX_16_3=y
+CONFIG_HISAX_TELESPCI=y
+CONFIG_HISAX_S0BOX=y
+CONFIG_HISAX_FRITZPCI=y
+CONFIG_HISAX_AVM_A1_PCMCIA=y
+CONFIG_HISAX_ELSA=y
+CONFIG_HISAX_DIEHLDIVA=y
+CONFIG_HISAX_SEDLBAUER=y
+CONFIG_HISAX_NETJET=y
+CONFIG_HISAX_NETJET_U=y
+CONFIG_HISAX_NICCY=y
+CONFIG_HISAX_BKM_A4T=y
+CONFIG_HISAX_SCT_QUADRO=y
+CONFIG_HISAX_GAZEL=y
+CONFIG_HISAX_HFC_PCI=y
+CONFIG_HISAX_W6692=y
+CONFIG_HISAX_HFC_SX=y
+CONFIG_HISAX_ENTERNOW_PCI=y
+# CONFIG_HISAX_DEBUG is not set
+
+#
+# HiSax PCMCIA card service modules
+#
+CONFIG_HISAX_SEDLBAUER_CS=m
+CONFIG_HISAX_ELSA_CS=m
+CONFIG_HISAX_AVM_A1_CS=m
+CONFIG_HISAX_TELES_CS=m
+
+#
+# HiSax sub driver modules
+#
+CONFIG_HISAX_ST5481=m
+CONFIG_HISAX_HFCUSB=m
+CONFIG_HISAX_HFC4S8S=m
+CONFIG_HISAX_FRITZ_PCIPNP=m
+CONFIG_HISAX_HDLC=y
+
+#
+# Active cards
+#
+
+#
+# Siemens Gigaset
+#
+CONFIG_ISDN_DRV_GIGASET=m
+CONFIG_GIGASET_BASE=m
+CONFIG_GIGASET_M105=m
+# CONFIG_GIGASET_DEBUG is not set
+# CONFIG_GIGASET_UNDOCREQ is not set
+
+#
+# CAPI subsystem
+#
+CONFIG_ISDN_CAPI=m
+CONFIG_ISDN_DRV_AVMB1_VERBOSE_REASON=y
+CONFIG_ISDN_CAPI_MIDDLEWARE=y
+CONFIG_ISDN_CAPI_CAPI20=m
+CONFIG_ISDN_CAPI_CAPIFS_BOOL=y
+CONFIG_ISDN_CAPI_CAPIFS=m
+CONFIG_ISDN_CAPI_CAPIDRV=m
+
+#
+# CAPI hardware drivers
+#
+
+#
+# Active AVM cards
+#
+CONFIG_CAPI_AVM=y
+CONFIG_ISDN_DRV_AVMB1_B1PCI=m
+CONFIG_ISDN_DRV_AVMB1_B1PCIV4=y
+CONFIG_ISDN_DRV_AVMB1_B1PCMCIA=m
+CONFIG_ISDN_DRV_AVMB1_AVM_CS=m
+CONFIG_ISDN_DRV_AVMB1_T1PCI=m
+CONFIG_ISDN_DRV_AVMB1_C4=m
+
+#
+# Active Eicon DIVA Server cards
+#
+CONFIG_CAPI_EICON=y
+CONFIG_ISDN_DIVAS=m
+CONFIG_ISDN_DIVAS_BRIPCI=y
+CONFIG_ISDN_DIVAS_PRIPCI=y
+CONFIG_ISDN_DIVAS_DIVACAPI=m
+CONFIG_ISDN_DIVAS_USERIDI=m
+CONFIG_ISDN_DIVAS_MAINT=m
+
+#
+# Telephony Support
+#
+CONFIG_PHONE=m
+CONFIG_PHONE_IXJ=m
+CONFIG_PHONE_IXJ_PCMCIA=m
+
+#
+# Input device support
+#
+CONFIG_INPUT=y
+
+#
+# Userland interfaces
+#
+CONFIG_INPUT_MOUSEDEV=y
+CONFIG_INPUT_MOUSEDEV_PSAUX=y
+CONFIG_INPUT_MOUSEDEV_SCREEN_X=1024
+CONFIG_INPUT_MOUSEDEV_SCREEN_Y=768
+CONFIG_INPUT_JOYDEV=m
+CONFIG_INPUT_TSDEV=m
+CONFIG_INPUT_TSDEV_SCREEN_X=240
+CONFIG_INPUT_TSDEV_SCREEN_Y=320
+CONFIG_INPUT_EVDEV=m
+# CONFIG_INPUT_EVBUG is not set
+
+#
+# Input Device Drivers
+#
+CONFIG_INPUT_KEYBOARD=y
+CONFIG_KEYBOARD_ATKBD=y
+CONFIG_KEYBOARD_SUNKBD=m
+CONFIG_KEYBOARD_LKKBD=m
+CONFIG_KEYBOARD_XTKBD=m
+CONFIG_KEYBOARD_NEWTON=m
+CONFIG_INPUT_MOUSE=y
+CONFIG_MOUSE_PS2=y
+CONFIG_MOUSE_SERIAL=m
+CONFIG_MOUSE_VSXXXAA=m
+CONFIG_INPUT_JOYSTICK=y
+CONFIG_JOYSTICK_ANALOG=m
+CONFIG_JOYSTICK_A3D=m
+CONFIG_JOYSTICK_ADI=m
+CONFIG_JOYSTICK_COBRA=m
+CONFIG_JOYSTICK_GF2K=m
+CONFIG_JOYSTICK_GRIP=m
+CONFIG_JOYSTICK_GRIP_MP=m
+CONFIG_JOYSTICK_GUILLEMOT=m
+CONFIG_JOYSTICK_INTERACT=m
+CONFIG_JOYSTICK_SIDEWINDER=m
+CONFIG_JOYSTICK_TMDC=m
+CONFIG_JOYSTICK_IFORCE=m
+CONFIG_JOYSTICK_IFORCE_USB=y
+CONFIG_JOYSTICK_IFORCE_232=y
+CONFIG_JOYSTICK_WARRIOR=m
+CONFIG_JOYSTICK_MAGELLAN=m
+CONFIG_JOYSTICK_SPACEORB=m
+CONFIG_JOYSTICK_SPACEBALL=m
+CONFIG_JOYSTICK_STINGER=m
+CONFIG_JOYSTICK_TWIDJOY=m
+CONFIG_JOYSTICK_DB9=m
+CONFIG_JOYSTICK_GAMECON=m
+CONFIG_JOYSTICK_TURBOGRAFX=m
+CONFIG_JOYSTICK_JOYDUMP=m
+CONFIG_INPUT_TOUCHSCREEN=y
+CONFIG_TOUCHSCREEN_ADS7846=m
+CONFIG_TOUCHSCREEN_GUNZE=m
+CONFIG_TOUCHSCREEN_ELO=m
+CONFIG_TOUCHSCREEN_MTOUCH=m
+CONFIG_TOUCHSCREEN_MK712=m
+CONFIG_INPUT_MISC=y
+CONFIG_INPUT_PCSPKR=m
+CONFIG_INPUT_WISTRON_BTNS=m
+CONFIG_INPUT_UINPUT=m
+
+#
+# Hardware I/O ports
+#
+CONFIG_SERIO=y
+CONFIG_SERIO_I8042=y
+CONFIG_SERIO_SERPORT=m
+CONFIG_SERIO_CT82C710=m
+CONFIG_SERIO_PARKBD=m
+CONFIG_SERIO_PCIPS2=m
+CONFIG_SERIO_LIBPS2=y
+CONFIG_SERIO_RAW=m
+CONFIG_GAMEPORT=m
+CONFIG_GAMEPORT_NS558=m
+CONFIG_GAMEPORT_L4=m
+CONFIG_GAMEPORT_EMU10K1=m
+CONFIG_GAMEPORT_FM801=m
+
+#
+# Character devices
+#
+CONFIG_VT=y
+CONFIG_VT_CONSOLE=y
+CONFIG_HW_CONSOLE=y
+CONFIG_VT_HW_CONSOLE_BINDING=y
+# CONFIG_SERIAL_NONSTANDARD is not set
+
+#
+# Serial drivers
+#
+CONFIG_SERIAL_8250=m
+CONFIG_SERIAL_8250_PCI=m
+CONFIG_SERIAL_8250_PNP=m
+# CONFIG_SERIAL_8250_CS is not set
+CONFIG_SERIAL_8250_NR_UARTS=4
+CONFIG_SERIAL_8250_RUNTIME_UARTS=4
+# CONFIG_SERIAL_8250_EXTENDED is not set
+
+#
+# Non-8250 serial port support
+#
+CONFIG_SERIAL_CORE=m
+CONFIG_SERIAL_JSM=m
+CONFIG_UNIX98_PTYS=y
+CONFIG_LEGACY_PTYS=y
+CONFIG_LEGACY_PTY_COUNT=256
+CONFIG_PRINTER=m
+# CONFIG_LP_CONSOLE is not set
+CONFIG_PPDEV=m
+CONFIG_TIPAR=m
+
+#
+# IPMI
+#
+CONFIG_IPMI_HANDLER=m
+# CONFIG_IPMI_PANIC_EVENT is not set
+CONFIG_IPMI_DEVICE_INTERFACE=m
+CONFIG_IPMI_SI=m
+CONFIG_IPMI_WATCHDOG=m
+CONFIG_IPMI_POWEROFF=m
+
+#
+# Watchdog Cards
+#
+CONFIG_WATCHDOG=y
+# CONFIG_WATCHDOG_NOWAYOUT is not set
+
+#
+# Watchdog Device Drivers
+#
+CONFIG_SOFT_WATCHDOG=m
+CONFIG_ACQUIRE_WDT=m
+CONFIG_ADVANTECH_WDT=m
+CONFIG_ALIM1535_WDT=m
+CONFIG_ALIM7101_WDT=m
+CONFIG_SC520_WDT=m
+CONFIG_EUROTECH_WDT=m
+CONFIG_IB700_WDT=m
+CONFIG_IBMASR=m
+CONFIG_WAFER_WDT=m
+CONFIG_I6300ESB_WDT=m
+CONFIG_I8XX_TCO=m
+CONFIG_SC1200_WDT=m
+CONFIG_SCx200_WDT=m
+CONFIG_60XX_WDT=m
+CONFIG_SBC8360_WDT=m
+CONFIG_CPU5_WDT=m
+CONFIG_W83627HF_WDT=m
+CONFIG_W83877F_WDT=m
+CONFIG_W83977F_WDT=m
+CONFIG_MACHZ_WDT=m
+CONFIG_SBC_EPX_C3_WATCHDOG=m
+
+#
+# PCI-based Watchdog Cards
+#
+CONFIG_PCIPCWATCHDOG=m
+CONFIG_WDTPCI=m
+CONFIG_WDT_501_PCI=y
+
+#
+# USB-based Watchdog Cards
+#
+CONFIG_USBPCWATCHDOG=m
+CONFIG_HW_RANDOM=y
+CONFIG_HW_RANDOM_INTEL=m
+CONFIG_HW_RANDOM_AMD=m
+CONFIG_HW_RANDOM_GEODE=m
+CONFIG_HW_RANDOM_VIA=m
+CONFIG_NVRAM=m
+CONFIG_RTC=m
+CONFIG_GEN_RTC=m
+CONFIG_GEN_RTC_X=y
+CONFIG_DTLK=m
+CONFIG_R3964=m
+CONFIG_APPLICOM=m
+CONFIG_SONYPI=m
+
+#
+# Ftape, the floppy tape device driver
+#
+CONFIG_AGP=m
+CONFIG_AGP_ALI=m
+CONFIG_AGP_ATI=m
+CONFIG_AGP_AMD=m
+CONFIG_AGP_AMD64=m
+CONFIG_AGP_INTEL=m
+CONFIG_AGP_NVIDIA=m
+CONFIG_AGP_SIS=m
+CONFIG_AGP_SWORKS=m
+CONFIG_AGP_VIA=m
+CONFIG_AGP_EFFICEON=m
+CONFIG_DRM=m
+CONFIG_DRM_TDFX=m
+CONFIG_DRM_R128=m
+CONFIG_DRM_RADEON=m
+CONFIG_DRM_I810=m
+CONFIG_DRM_I830=m
+CONFIG_DRM_I915=m
+CONFIG_DRM_MGA=m
+CONFIG_DRM_SIS=m
+CONFIG_DRM_VIA=m
+CONFIG_DRM_SAVAGE=m
+
+#
+# PCMCIA character devices
+#
+CONFIG_SYNCLINK_CS=m
+CONFIG_CARDMAN_4000=m
+CONFIG_CARDMAN_4040=m
+CONFIG_MWAVE=m
+CONFIG_SCx200_GPIO=m
+CONFIG_PC8736x_GPIO=m
+CONFIG_NSC_GPIO=m
+CONFIG_CS5535_GPIO=m
+CONFIG_RAW_DRIVER=m
+CONFIG_MAX_RAW_DEVS=256
+# CONFIG_HPET is not set
+CONFIG_HANGCHECK_TIMER=m
+
+#
+# TPM devices
+#
+CONFIG_TCG_TPM=m
+CONFIG_TCG_TIS=m
+CONFIG_TCG_NSC=m
+CONFIG_TCG_ATMEL=m
+CONFIG_TCG_INFINEON=m
+CONFIG_TCG_XEN=m
+CONFIG_TELCLOCK=m
+
+#
+# I2C support
+#
+CONFIG_I2C=m
+CONFIG_I2C_CHARDEV=m
+
+#
+# I2C Algorithms
+#
+CONFIG_I2C_ALGOBIT=m
+CONFIG_I2C_ALGOPCF=m
+CONFIG_I2C_ALGOPCA=m
+
+#
+# I2C Hardware Bus support
+#
+CONFIG_I2C_ALI1535=m
+CONFIG_I2C_ALI1563=m
+CONFIG_I2C_ALI15X3=m
+CONFIG_I2C_AMD756=m
+CONFIG_I2C_AMD756_S4882=m
+CONFIG_I2C_AMD8111=m
+CONFIG_I2C_I801=m
+CONFIG_I2C_I810=m
+CONFIG_I2C_PIIX4=m
+CONFIG_I2C_ISA=m
+CONFIG_I2C_NFORCE2=m
+CONFIG_I2C_OCORES=m
+CONFIG_I2C_PARPORT=m
+CONFIG_I2C_PARPORT_LIGHT=m
+CONFIG_I2C_PROSAVAGE=m
+CONFIG_I2C_SAVAGE4=m
+CONFIG_SCx200_I2C=m
+CONFIG_SCx200_I2C_SCL=12
+CONFIG_SCx200_I2C_SDA=13
+CONFIG_SCx200_ACB=m
+CONFIG_I2C_SIS5595=m
+CONFIG_I2C_SIS630=m
+CONFIG_I2C_SIS96X=m
+CONFIG_I2C_STUB=m
+CONFIG_I2C_VIA=m
+CONFIG_I2C_VIAPRO=m
+CONFIG_I2C_VOODOO3=m
+CONFIG_I2C_PCA_ISA=m
+
+#
+# Miscellaneous I2C Chip support
+#
+CONFIG_SENSORS_DS1337=m
+CONFIG_SENSORS_DS1374=m
+CONFIG_SENSORS_EEPROM=m
+CONFIG_SENSORS_PCF8574=m
+CONFIG_SENSORS_PCA9539=m
+CONFIG_SENSORS_PCF8591=m
+CONFIG_SENSORS_MAX6875=m
+# CONFIG_I2C_DEBUG_CORE is not set
+# CONFIG_I2C_DEBUG_ALGO is not set
+# CONFIG_I2C_DEBUG_BUS is not set
+# CONFIG_I2C_DEBUG_CHIP is not set
+
+#
+# SPI support
+#
+CONFIG_SPI=y
+# CONFIG_SPI_DEBUG is not set
+CONFIG_SPI_MASTER=y
+
+#
+# SPI Master Controller Drivers
+#
+CONFIG_SPI_BITBANG=m
+CONFIG_SPI_BUTTERFLY=m
+
+#
+# SPI Protocol Masters
+#
+
+#
+# Dallas's 1-wire bus
+#
+CONFIG_W1=m
+CONFIG_W1_CON=y
+
+#
+# 1-wire Bus Masters
+#
+CONFIG_W1_MASTER_MATROX=m
+CONFIG_W1_MASTER_DS2490=m
+CONFIG_W1_MASTER_DS2482=m
+
+#
+# 1-wire Slaves
+#
+CONFIG_W1_SLAVE_THERM=m
+CONFIG_W1_SLAVE_SMEM=m
+CONFIG_W1_SLAVE_DS2433=m
+# CONFIG_W1_SLAVE_DS2433_CRC is not set
+
+#
+# Hardware Monitoring support
+#
+CONFIG_HWMON=m
+CONFIG_HWMON_VID=m
+CONFIG_SENSORS_ABITUGURU=m
+CONFIG_SENSORS_ADM1021=m
+CONFIG_SENSORS_ADM1025=m
+CONFIG_SENSORS_ADM1026=m
+CONFIG_SENSORS_ADM1031=m
+CONFIG_SENSORS_ADM9240=m
+CONFIG_SENSORS_ASB100=m
+CONFIG_SENSORS_ATXP1=m
+CONFIG_SENSORS_DS1621=m
+CONFIG_SENSORS_F71805F=m
+CONFIG_SENSORS_FSCHER=m
+CONFIG_SENSORS_FSCPOS=m
+CONFIG_SENSORS_GL518SM=m
+CONFIG_SENSORS_GL520SM=m
+CONFIG_SENSORS_IT87=m
+CONFIG_SENSORS_LM63=m
+CONFIG_SENSORS_LM70=m
+CONFIG_SENSORS_LM75=m
+CONFIG_SENSORS_LM77=m
+CONFIG_SENSORS_LM78=m
+CONFIG_SENSORS_LM80=m
+CONFIG_SENSORS_LM83=m
+CONFIG_SENSORS_LM85=m
+CONFIG_SENSORS_LM87=m
+CONFIG_SENSORS_LM90=m
+CONFIG_SENSORS_LM92=m
+CONFIG_SENSORS_MAX1619=m
+CONFIG_SENSORS_PC87360=m
+CONFIG_SENSORS_SIS5595=m
+CONFIG_SENSORS_SMSC47M1=m
+CONFIG_SENSORS_SMSC47M192=m
+CONFIG_SENSORS_SMSC47B397=m
+CONFIG_SENSORS_VIA686A=m
+CONFIG_SENSORS_VT8231=m
+CONFIG_SENSORS_W83781D=m
+CONFIG_SENSORS_W83791D=m
+CONFIG_SENSORS_W83792D=m
+CONFIG_SENSORS_W83L785TS=m
+CONFIG_SENSORS_W83627HF=m
+CONFIG_SENSORS_W83627EHF=m
+CONFIG_SENSORS_HDAPS=m
+# CONFIG_HWMON_DEBUG_CHIP is not set
+
+#
+# Misc devices
+#
+CONFIG_IBM_ASM=m
+
+#
+# Multimedia devices
+#
+CONFIG_VIDEO_DEV=m
+CONFIG_VIDEO_V4L1=y
+CONFIG_VIDEO_V4L1_COMPAT=y
+CONFIG_VIDEO_V4L2=y
+
+#
+# Video Capture Adapters
+#
+
+#
+# Video Capture Adapters
+#
+# CONFIG_VIDEO_ADV_DEBUG is not set
+CONFIG_VIDEO_VIVI=m
+CONFIG_VIDEO_BT848=m
+CONFIG_VIDEO_BT848_DVB=y
+CONFIG_VIDEO_SAA6588=m
+CONFIG_VIDEO_BWQCAM=m
+CONFIG_VIDEO_CQCAM=m
+CONFIG_VIDEO_W9966=m
+CONFIG_VIDEO_CPIA=m
+CONFIG_VIDEO_CPIA_PP=m
+CONFIG_VIDEO_CPIA_USB=m
+CONFIG_VIDEO_CPIA2=m
+CONFIG_VIDEO_SAA5246A=m
+CONFIG_VIDEO_SAA5249=m
+CONFIG_TUNER_3036=m
+CONFIG_VIDEO_STRADIS=m
+CONFIG_VIDEO_ZORAN=m
+CONFIG_VIDEO_ZORAN_BUZ=m
+CONFIG_VIDEO_ZORAN_DC10=m
+CONFIG_VIDEO_ZORAN_DC30=m
+CONFIG_VIDEO_ZORAN_LML33=m
+CONFIG_VIDEO_ZORAN_LML33R10=m
+CONFIG_VIDEO_ZORAN_AVS6EYES=m
+CONFIG_VIDEO_MEYE=m
+CONFIG_VIDEO_SAA7134=m
+CONFIG_VIDEO_SAA7134_ALSA=m
+# CONFIG_VIDEO_SAA7134_OSS is not set
+CONFIG_VIDEO_SAA7134_DVB=m
+CONFIG_VIDEO_SAA7134_DVB_ALL_FRONTENDS=y
+CONFIG_VIDEO_MXB=m
+CONFIG_VIDEO_DPC=m
+CONFIG_VIDEO_HEXIUM_ORION=m
+CONFIG_VIDEO_HEXIUM_GEMINI=m
+CONFIG_VIDEO_CX88_VP3054=m
+CONFIG_VIDEO_CX88=m
+CONFIG_VIDEO_CX88_ALSA=m
+CONFIG_VIDEO_CX88_BLACKBIRD=m
+CONFIG_VIDEO_CX88_DVB=m
+CONFIG_VIDEO_CX88_DVB_ALL_FRONTENDS=y
+
+#
+# Encoders and Decoders
+#
+CONFIG_VIDEO_MSP3400=m
+CONFIG_VIDEO_CS53L32A=m
+CONFIG_VIDEO_TLV320AIC23B=m
+CONFIG_VIDEO_WM8775=m
+CONFIG_VIDEO_WM8739=m
+CONFIG_VIDEO_CX2341X=m
+CONFIG_VIDEO_CX25840=m
+CONFIG_VIDEO_SAA711X=m
+CONFIG_VIDEO_SAA7127=m
+CONFIG_VIDEO_UPD64031A=m
+CONFIG_VIDEO_UPD64083=m
+
+#
+# V4L USB devices
+#
+CONFIG_VIDEO_PVRUSB2=m
+CONFIG_VIDEO_PVRUSB2_24XXX=y
+CONFIG_VIDEO_PVRUSB2_SYSFS=y
+# CONFIG_VIDEO_PVRUSB2_DEBUGIFC is not set
+CONFIG_VIDEO_EM28XX=m
+CONFIG_VIDEO_USBVIDEO=m
+CONFIG_USB_VICAM=m
+CONFIG_USB_IBMCAM=m
+CONFIG_USB_KONICAWC=m
+CONFIG_USB_QUICKCAM_MESSENGER=m
+CONFIG_USB_ET61X251=m
+CONFIG_VIDEO_OVCAMCHIP=m
+CONFIG_USB_W9968CF=m
+CONFIG_USB_OV511=m
+CONFIG_USB_SE401=m
+CONFIG_USB_SN9C102=m
+CONFIG_USB_STV680=m
+CONFIG_USB_ZC0301=m
+CONFIG_USB_PWC=m
+# CONFIG_USB_PWC_DEBUG is not set
+
+#
+# Radio Adapters
+#
+CONFIG_RADIO_GEMTEK_PCI=m
+CONFIG_RADIO_MAXIRADIO=m
+CONFIG_RADIO_MAESTRO=m
+CONFIG_USB_DSBR=m
+
+#
+# Digital Video Broadcasting Devices
+#
+CONFIG_DVB=y
+CONFIG_DVB_CORE=m
+
+#
+# Supported SAA7146 based PCI Adapters
+#
+CONFIG_DVB_AV7110=m
+CONFIG_DVB_AV7110_OSD=y
+CONFIG_DVB_BUDGET=m
+CONFIG_DVB_BUDGET_CI=m
+CONFIG_DVB_BUDGET_AV=m
+CONFIG_DVB_BUDGET_PATCH=m
+
+#
+# Supported USB Adapters
+#
+CONFIG_DVB_USB=m
+# CONFIG_DVB_USB_DEBUG is not set
+CONFIG_DVB_USB_A800=m
+CONFIG_DVB_USB_DIBUSB_MB=m
+# CONFIG_DVB_USB_DIBUSB_MB_FAULTY is not set
+CONFIG_DVB_USB_DIBUSB_MC=m
+CONFIG_DVB_USB_UMT_010=m
+CONFIG_DVB_USB_CXUSB=m
+CONFIG_DVB_USB_DIGITV=m
+CONFIG_DVB_USB_VP7045=m
+CONFIG_DVB_USB_VP702X=m
+CONFIG_DVB_USB_GP8PSK=m
+CONFIG_DVB_USB_NOVA_T_USB2=m
+CONFIG_DVB_USB_DTT200U=m
+CONFIG_DVB_TTUSB_BUDGET=m
+CONFIG_DVB_TTUSB_DEC=m
+CONFIG_DVB_CINERGYT2=m
+# CONFIG_DVB_CINERGYT2_TUNING is not set
+
+#
+# Supported FlexCopII (B2C2) Adapters
+#
+CONFIG_DVB_B2C2_FLEXCOP=m
+CONFIG_DVB_B2C2_FLEXCOP_PCI=m
+CONFIG_DVB_B2C2_FLEXCOP_USB=m
+# CONFIG_DVB_B2C2_FLEXCOP_DEBUG is not set
+
+#
+# Supported BT878 Adapters
+#
+CONFIG_DVB_BT8XX=m
+
+#
+# Supported Pluto2 Adapters
+#
+CONFIG_DVB_PLUTO2=m
+
+#
+# Supported DVB Frontends
+#
+
+#
+# Customise DVB Frontends
+#
+
+#
+# DVB-S (satellite) frontends
+#
+CONFIG_DVB_STV0299=m
+CONFIG_DVB_CX24110=m
+CONFIG_DVB_CX24123=m
+CONFIG_DVB_TDA8083=m
+CONFIG_DVB_MT312=m
+CONFIG_DVB_VES1X93=m
+CONFIG_DVB_S5H1420=m
+
+#
+# DVB-T (terrestrial) frontends
+#
+CONFIG_DVB_SP8870=m
+CONFIG_DVB_SP887X=m
+CONFIG_DVB_CX22700=m
+CONFIG_DVB_CX22702=m
+CONFIG_DVB_L64781=m
+CONFIG_DVB_TDA1004X=m
+CONFIG_DVB_NXT6000=m
+CONFIG_DVB_MT352=m
+CONFIG_DVB_ZL10353=m
+CONFIG_DVB_DIB3000MB=m
+CONFIG_DVB_DIB3000MC=m
+
+#
+# DVB-C (cable) frontends
+#
+CONFIG_DVB_VES1820=m
+CONFIG_DVB_TDA10021=m
+CONFIG_DVB_STV0297=m
+
+#
+# ATSC (North American/Korean Terrestrial/Cable DTV) frontends
+#
+CONFIG_DVB_NXT200X=m
+CONFIG_DVB_OR51211=m
+CONFIG_DVB_OR51132=m
+CONFIG_DVB_BCM3510=m
+CONFIG_DVB_LGDT330X=m
+
+#
+# Miscellaneous devices
+#
+CONFIG_DVB_PLL=m
+CONFIG_DVB_LNBP21=m
+CONFIG_DVB_ISL6421=m
+CONFIG_VIDEO_SAA7146=m
+CONFIG_VIDEO_SAA7146_VV=m
+CONFIG_VIDEO_VIDEOBUF=m
+CONFIG_VIDEO_TUNER=m
+CONFIG_VIDEO_BUF=m
+CONFIG_VIDEO_BUF_DVB=m
+CONFIG_VIDEO_BTCX=m
+CONFIG_VIDEO_IR=m
+CONFIG_VIDEO_TVEEPROM=m
+CONFIG_USB_DABUSB=m
+
+#
+# Graphics support
+#
+CONFIG_FIRMWARE_EDID=y
+CONFIG_FB=y
+CONFIG_FB_CFB_FILLRECT=y
+CONFIG_FB_CFB_COPYAREA=y
+CONFIG_FB_CFB_IMAGEBLIT=y
+# CONFIG_FB_MACMODES is not set
+# CONFIG_FB_BACKLIGHT is not set
+CONFIG_FB_MODE_HELPERS=y
+CONFIG_FB_TILEBLITTING=y
+CONFIG_FB_CIRRUS=m
+CONFIG_FB_PM2=m
+CONFIG_FB_PM2_FIFO_DISCONNECT=y
+CONFIG_FB_CYBER2000=m
+CONFIG_FB_ARC=m
+# CONFIG_FB_ASILIANT is not set
+# CONFIG_FB_IMSTT is not set
+CONFIG_FB_VGA16=m
+CONFIG_FB_VESA=y
+CONFIG_FB_HGA=m
+# CONFIG_FB_HGA_ACCEL is not set
+CONFIG_FB_S1D13XXX=m
+CONFIG_FB_NVIDIA=m
+CONFIG_FB_NVIDIA_I2C=y
+CONFIG_FB_RIVA=m
+CONFIG_FB_RIVA_I2C=y
+CONFIG_FB_RIVA_DEBUG=y
+CONFIG_FB_I810=m
+# CONFIG_FB_I810_GTF is not set
+CONFIG_FB_INTEL=m
+# CONFIG_FB_INTEL_DEBUG is not set
+CONFIG_FB_MATROX=m
+CONFIG_FB_MATROX_MILLENIUM=y
+CONFIG_FB_MATROX_MYSTIQUE=y
+CONFIG_FB_MATROX_G=y
+# CONFIG_FB_MATROX_I2C is not set
+CONFIG_FB_MATROX_MULTIHEAD=y
+CONFIG_FB_RADEON=m
+CONFIG_FB_RADEON_I2C=y
+# CONFIG_FB_RADEON_DEBUG is not set
+CONFIG_FB_ATY128=m
+CONFIG_FB_ATY=m
+CONFIG_FB_ATY_CT=y
+CONFIG_FB_ATY_GENERIC_LCD=y
+CONFIG_FB_ATY_GX=y
+CONFIG_FB_SAVAGE=m
+CONFIG_FB_SAVAGE_I2C=y
+CONFIG_FB_SAVAGE_ACCEL=y
+CONFIG_FB_SIS=m
+CONFIG_FB_SIS_300=y
+CONFIG_FB_SIS_315=y
+CONFIG_FB_NEOMAGIC=m
+CONFIG_FB_KYRO=m
+CONFIG_FB_3DFX=m
+# CONFIG_FB_3DFX_ACCEL is not set
+CONFIG_FB_VOODOO1=m
+CONFIG_FB_CYBLA=m
+CONFIG_FB_TRIDENT=m
+# CONFIG_FB_TRIDENT_ACCEL is not set
+CONFIG_FB_GEODE=y
+CONFIG_FB_GEODE_GX=m
+CONFIG_FB_GEODE_GX1=m
+CONFIG_FB_VIRTUAL=m
+
+#
+# Console display driver support
+#
+CONFIG_VGA_CONSOLE=y
+# CONFIG_VGACON_SOFT_SCROLLBACK is not set
+CONFIG_VIDEO_SELECT=y
+CONFIG_DUMMY_CONSOLE=y
+CONFIG_FRAMEBUFFER_CONSOLE=m
+# CONFIG_FRAMEBUFFER_CONSOLE_ROTATION is not set
+# CONFIG_FONTS is not set
+CONFIG_FONT_8x8=y
+CONFIG_FONT_8x16=y
+
+#
+# Logo configuration
+#
+# CONFIG_LOGO is not set
+CONFIG_BACKLIGHT_LCD_SUPPORT=y
+CONFIG_BACKLIGHT_CLASS_DEVICE=m
+CONFIG_BACKLIGHT_DEVICE=y
+CONFIG_LCD_CLASS_DEVICE=m
+CONFIG_LCD_DEVICE=y
+
+#
+# Sound
+#
+CONFIG_SOUND=m
+
+#
+# Advanced Linux Sound Architecture
+#
+CONFIG_SND=m
+CONFIG_SND_TIMER=m
+CONFIG_SND_PCM=m
+CONFIG_SND_HWDEP=m
+CONFIG_SND_RAWMIDI=m
+CONFIG_SND_SEQUENCER=m
+CONFIG_SND_SEQ_DUMMY=m
+CONFIG_SND_OSSEMUL=y
+CONFIG_SND_MIXER_OSS=m
+CONFIG_SND_PCM_OSS=m
+CONFIG_SND_PCM_OSS_PLUGINS=y
+CONFIG_SND_SEQUENCER_OSS=y
+CONFIG_SND_RTCTIMER=m
+CONFIG_SND_SEQ_RTCTIMER_DEFAULT=y
+# CONFIG_SND_DYNAMIC_MINORS is not set
+CONFIG_SND_SUPPORT_OLD_API=y
+CONFIG_SND_VERBOSE_PROCFS=y
+# CONFIG_SND_VERBOSE_PRINTK is not set
+# CONFIG_SND_DEBUG is not set
+
+#
+# Generic devices
+#
+CONFIG_SND_MPU401_UART=m
+CONFIG_SND_OPL3_LIB=m
+CONFIG_SND_VX_LIB=m
+CONFIG_SND_AC97_CODEC=m
+CONFIG_SND_AC97_BUS=m
+CONFIG_SND_DUMMY=m
+CONFIG_SND_VIRMIDI=m
+CONFIG_SND_MTPAV=m
+CONFIG_SND_SERIAL_U16550=m
+CONFIG_SND_MPU401=m
+
+#
+# PCI devices
+#
+CONFIG_SND_AD1889=m
+CONFIG_SND_ALS300=m
+CONFIG_SND_ALS4000=m
+CONFIG_SND_ALI5451=m
+CONFIG_SND_ATIIXP=m
+CONFIG_SND_ATIIXP_MODEM=m
+CONFIG_SND_AU8810=m
+CONFIG_SND_AU8820=m
+CONFIG_SND_AU8830=m
+CONFIG_SND_AZT3328=m
+CONFIG_SND_BT87X=m
+# CONFIG_SND_BT87X_OVERCLOCK is not set
+CONFIG_SND_CA0106=m
+CONFIG_SND_CMIPCI=m
+CONFIG_SND_CS4281=m
+CONFIG_SND_CS46XX=m
+CONFIG_SND_CS46XX_NEW_DSP=y
+CONFIG_SND_CS5535AUDIO=m
+CONFIG_SND_DARLA20=m
+CONFIG_SND_GINA20=m
+CONFIG_SND_LAYLA20=m
+CONFIG_SND_DARLA24=m
+CONFIG_SND_GINA24=m
+CONFIG_SND_LAYLA24=m
+CONFIG_SND_MONA=m
+CONFIG_SND_MIA=m
+CONFIG_SND_ECHO3G=m
+CONFIG_SND_INDIGO=m
+CONFIG_SND_INDIGOIO=m
+CONFIG_SND_INDIGODJ=m
+CONFIG_SND_EMU10K1=m
+CONFIG_SND_EMU10K1X=m
+CONFIG_SND_ENS1370=m
+CONFIG_SND_ENS1371=m
+CONFIG_SND_ES1938=m
+CONFIG_SND_ES1968=m
+CONFIG_SND_FM801=m
+# CONFIG_SND_FM801_TEA575X_BOOL is not set
+CONFIG_SND_HDA_INTEL=m
+CONFIG_SND_HDSP=m
+CONFIG_SND_HDSPM=m
+CONFIG_SND_ICE1712=m
+CONFIG_SND_ICE1724=m
+CONFIG_SND_INTEL8X0=m
+CONFIG_SND_INTEL8X0M=m
+CONFIG_SND_KORG1212=m
+CONFIG_SND_MAESTRO3=m
+CONFIG_SND_MIXART=m
+CONFIG_SND_NM256=m
+CONFIG_SND_PCXHR=m
+CONFIG_SND_RIPTIDE=m
+CONFIG_SND_RME32=m
+CONFIG_SND_RME96=m
+CONFIG_SND_RME9652=m
+CONFIG_SND_SONICVIBES=m
+CONFIG_SND_TRIDENT=m
+CONFIG_SND_VIA82XX=m
+# CONFIG_SND_VIA82XX_MODEM is not set
+CONFIG_SND_VX222=m
+CONFIG_SND_YMFPCI=m
+
+#
+# USB devices
+#
+CONFIG_SND_USB_AUDIO=m
+CONFIG_SND_USB_USX2Y=m
+
+#
+# PCMCIA devices
+#
+CONFIG_SND_VXPOCKET=m
+CONFIG_SND_PDAUDIOCF=m
+
+#
+# Open Sound System
+#
+CONFIG_SOUND_PRIME=m
+# CONFIG_OSS_OBSOLETE_DRIVER is not set
+CONFIG_SOUND_BT878=m
+CONFIG_SOUND_ES1371=m
+CONFIG_SOUND_ICH=m
+CONFIG_SOUND_TRIDENT=m
+# CONFIG_SOUND_MSNDCLAS is not set
+# CONFIG_SOUND_MSNDPIN is not set
+CONFIG_SOUND_VIA82CXXX=m
+# CONFIG_MIDI_VIA82CXXX is not set
+# CONFIG_SOUND_OSS is not set
+CONFIG_SOUND_TVMIXER=m
+
+#
+# USB support
+#
+CONFIG_USB_ARCH_HAS_HCD=y
+CONFIG_USB_ARCH_HAS_OHCI=y
+CONFIG_USB_ARCH_HAS_EHCI=y
+CONFIG_USB=m
+# CONFIG_USB_DEBUG is not set
+
+#
+# Miscellaneous USB options
+#
+CONFIG_USB_DEVICEFS=y
+CONFIG_USB_BANDWIDTH=y
+# CONFIG_USB_DYNAMIC_MINORS is not set
+# CONFIG_USB_SUSPEND is not set
+# CONFIG_USB_OTG is not set
+
+#
+# USB Host Controller Drivers
+#
+CONFIG_USB_EHCI_HCD=m
+CONFIG_USB_EHCI_SPLIT_ISO=y
+CONFIG_USB_EHCI_ROOT_HUB_TT=y
+CONFIG_USB_EHCI_TT_NEWSCHED=y
+CONFIG_USB_ISP116X_HCD=m
+CONFIG_USB_OHCI_HCD=m
+# CONFIG_USB_OHCI_BIG_ENDIAN is not set
+CONFIG_USB_OHCI_LITTLE_ENDIAN=y
+CONFIG_USB_UHCI_HCD=m
+CONFIG_USB_SL811_HCD=m
+CONFIG_USB_SL811_CS=m
+
+#
+# USB Device Class drivers
+#
+CONFIG_USB_ACM=m
+CONFIG_USB_PRINTER=m
+
+#
+# NOTE: USB_STORAGE enables SCSI, and 'SCSI disk support'
+#
+
+#
+# may also be needed; see USB_STORAGE Help for more information
+#
+CONFIG_USB_STORAGE=m
+# CONFIG_USB_STORAGE_DEBUG is not set
+CONFIG_USB_STORAGE_DATAFAB=y
+CONFIG_USB_STORAGE_FREECOM=y
+CONFIG_USB_STORAGE_ISD200=y
+CONFIG_USB_STORAGE_DPCM=y
+CONFIG_USB_STORAGE_USBAT=y
+CONFIG_USB_STORAGE_SDDR09=y
+CONFIG_USB_STORAGE_SDDR55=y
+CONFIG_USB_STORAGE_JUMPSHOT=y
+CONFIG_USB_STORAGE_ALAUDA=y
+CONFIG_USB_LIBUSUAL=y
+
+#
+# USB Input Devices
+#
+CONFIG_USB_HID=m
+CONFIG_USB_HIDINPUT=y
+# CONFIG_USB_HIDINPUT_POWERBOOK is not set
+CONFIG_HID_FF=y
+CONFIG_HID_PID=y
+CONFIG_LOGITECH_FF=y
+CONFIG_THRUSTMASTER_FF=y
+CONFIG_USB_HIDDEV=y
+
+#
+# USB HID Boot Protocol drivers
+#
+# CONFIG_USB_KBD is not set
+# CONFIG_USB_MOUSE is not set
+CONFIG_USB_AIPTEK=m
+CONFIG_USB_WACOM=m
+CONFIG_USB_ACECAD=m
+CONFIG_USB_KBTAB=m
+CONFIG_USB_POWERMATE=m
+CONFIG_USB_TOUCHSCREEN=m
+CONFIG_USB_TOUCHSCREEN_EGALAX=y
+CONFIG_USB_TOUCHSCREEN_PANJIT=y
+CONFIG_USB_TOUCHSCREEN_3M=y
+CONFIG_USB_TOUCHSCREEN_ITM=y
+CONFIG_USB_YEALINK=m
+CONFIG_USB_XPAD=m
+CONFIG_USB_ATI_REMOTE=m
+CONFIG_USB_ATI_REMOTE2=m
+CONFIG_USB_KEYSPAN_REMOTE=m
+CONFIG_USB_APPLETOUCH=m
+
+#
+# USB Imaging devices
+#
+CONFIG_USB_MDC800=m
+CONFIG_USB_MICROTEK=m
+
+#
+# USB Network Adapters
+#
+CONFIG_USB_CATC=m
+CONFIG_USB_KAWETH=m
+CONFIG_USB_PEGASUS=m
+CONFIG_USB_RTL8150=m
+CONFIG_USB_USBNET=m
+CONFIG_USB_NET_AX8817X=m
+CONFIG_USB_NET_CDCETHER=m
+CONFIG_USB_NET_GL620A=m
+CONFIG_USB_NET_NET1080=m
+CONFIG_USB_NET_PLUSB=m
+CONFIG_USB_NET_RNDIS_HOST=m
+CONFIG_USB_NET_CDC_SUBSET=m
+CONFIG_USB_ALI_M5632=y
+CONFIG_USB_AN2720=y
+CONFIG_USB_BELKIN=y
+CONFIG_USB_ARMLINUX=y
+# CONFIG_USB_EPSON2888 is not set
+CONFIG_USB_NET_ZAURUS=m
+CONFIG_USB_MON=y
+
+#
+# USB port drivers
+#
+CONFIG_USB_USS720=m
+
+#
+# USB Serial Converter support
+#
+CONFIG_USB_SERIAL=m
+CONFIG_USB_SERIAL_GENERIC=y
+CONFIG_USB_SERIAL_AIRPRIME=m
+CONFIG_USB_SERIAL_ARK3116=m
+CONFIG_USB_SERIAL_BELKIN=m
+CONFIG_USB_SERIAL_WHITEHEAT=m
+CONFIG_USB_SERIAL_DIGI_ACCELEPORT=m
+CONFIG_USB_SERIAL_CP2101=m
+CONFIG_USB_SERIAL_CYPRESS_M8=m
+CONFIG_USB_SERIAL_EMPEG=m
+CONFIG_USB_SERIAL_FTDI_SIO=m
+CONFIG_USB_SERIAL_FUNSOFT=m
+CONFIG_USB_SERIAL_VISOR=m
+CONFIG_USB_SERIAL_IPAQ=m
+CONFIG_USB_SERIAL_IR=m
+CONFIG_USB_SERIAL_EDGEPORT=m
+CONFIG_USB_SERIAL_EDGEPORT_TI=m
+CONFIG_USB_SERIAL_GARMIN=m
+CONFIG_USB_SERIAL_IPW=m
+CONFIG_USB_SERIAL_KEYSPAN_PDA=m
+CONFIG_USB_SERIAL_KEYSPAN=m
+CONFIG_USB_SERIAL_KEYSPAN_MPR=y
+CONFIG_USB_SERIAL_KEYSPAN_USA28=y
+CONFIG_USB_SERIAL_KEYSPAN_USA28X=y
+CONFIG_USB_SERIAL_KEYSPAN_USA28XA=y
+CONFIG_USB_SERIAL_KEYSPAN_USA28XB=y
+CONFIG_USB_SERIAL_KEYSPAN_USA19=y
+CONFIG_USB_SERIAL_KEYSPAN_USA18X=y
+CONFIG_USB_SERIAL_KEYSPAN_USA19W=y
+CONFIG_USB_SERIAL_KEYSPAN_USA19QW=y
+CONFIG_USB_SERIAL_KEYSPAN_USA19QI=y
+CONFIG_USB_SERIAL_KEYSPAN_USA49W=y
+CONFIG_USB_SERIAL_KEYSPAN_USA49WLC=y
+CONFIG_USB_SERIAL_KLSI=m
+CONFIG_USB_SERIAL_KOBIL_SCT=m
+CONFIG_USB_SERIAL_MCT_U232=m
+CONFIG_USB_SERIAL_NAVMAN=m
+CONFIG_USB_SERIAL_PL2303=m
+CONFIG_USB_SERIAL_HP4X=m
+CONFIG_USB_SERIAL_SAFE=m
+CONFIG_USB_SERIAL_SAFE_PADDED=y
+CONFIG_USB_SERIAL_SIERRAWIRELESS=m
+CONFIG_USB_SERIAL_TI=m
+CONFIG_USB_SERIAL_CYBERJACK=m
+CONFIG_USB_SERIAL_XIRCOM=m
+CONFIG_USB_SERIAL_OPTION=m
+CONFIG_USB_SERIAL_OMNINET=m
+CONFIG_USB_EZUSB=y
+
+#
+# USB Miscellaneous drivers
+#
+CONFIG_USB_EMI62=m
+CONFIG_USB_EMI26=m
+CONFIG_USB_AUERSWALD=m
+CONFIG_USB_RIO500=m
+CONFIG_USB_LEGOTOWER=m
+CONFIG_USB_LCD=m
+CONFIG_USB_LED=m
+CONFIG_USB_CYPRESS_CY7C63=m
+CONFIG_USB_CYTHERM=m
+CONFIG_USB_PHIDGETKIT=m
+CONFIG_USB_PHIDGETSERVO=m
+CONFIG_USB_IDMOUSE=m
+CONFIG_USB_APPLEDISPLAY=m
+CONFIG_USB_SISUSBVGA=m
+CONFIG_USB_SISUSBVGA_CON=y
+CONFIG_USB_LD=m
+CONFIG_USB_TEST=m
+
+#
+# USB DSL modem support
+#
+CONFIG_USB_ATM=m
+CONFIG_USB_SPEEDTOUCH=m
+CONFIG_USB_CXACRU=m
+CONFIG_USB_UEAGLEATM=m
+CONFIG_USB_XUSBATM=m
+
+#
+# USB Gadget Support
+#
+CONFIG_USB_GADGET=m
+# CONFIG_USB_GADGET_DEBUG_FILES is not set
+CONFIG_USB_GADGET_SELECTED=y
+CONFIG_USB_GADGET_NET2280=y
+CONFIG_USB_NET2280=m
+# CONFIG_USB_GADGET_PXA2XX is not set
+# CONFIG_USB_GADGET_GOKU is not set
+# CONFIG_USB_GADGET_LH7A40X is not set
+# CONFIG_USB_GADGET_OMAP is not set
+# CONFIG_USB_GADGET_AT91 is not set
+# CONFIG_USB_GADGET_DUMMY_HCD is not set
+CONFIG_USB_GADGET_DUALSPEED=y
+CONFIG_USB_ZERO=m
+CONFIG_USB_ETH=m
+CONFIG_USB_ETH_RNDIS=y
+CONFIG_USB_GADGETFS=m
+CONFIG_USB_FILE_STORAGE=m
+# CONFIG_USB_FILE_STORAGE_TEST is not set
+CONFIG_USB_G_SERIAL=m
+
+#
+# MMC/SD Card support
+#
+CONFIG_MMC=m
+# CONFIG_MMC_DEBUG is not set
+CONFIG_MMC_BLOCK=m
+CONFIG_MMC_SDHCI=m
+CONFIG_MMC_WBSD=m
+
+#
+# LED devices
+#
+CONFIG_NEW_LEDS=y
+CONFIG_LEDS_CLASS=m
+
+#
+# LED drivers
+#
+CONFIG_LEDS_NET48XX=m
+
+#
+# LED Triggers
+#
+CONFIG_LEDS_TRIGGERS=y
+CONFIG_LEDS_TRIGGER_TIMER=m
+CONFIG_LEDS_TRIGGER_IDE_DISK=y
+CONFIG_LEDS_TRIGGER_HEARTBEAT=m
+
+#
+# InfiniBand support
+#
+CONFIG_INFINIBAND=m
+CONFIG_INFINIBAND_USER_MAD=m
+CONFIG_INFINIBAND_USER_ACCESS=m
+CONFIG_INFINIBAND_ADDR_TRANS=y
+CONFIG_INFINIBAND_MTHCA=m
+CONFIG_INFINIBAND_MTHCA_DEBUG=y
+CONFIG_INFINIBAND_IPOIB=m
+CONFIG_INFINIBAND_IPOIB_DEBUG=y
+# CONFIG_INFINIBAND_IPOIB_DEBUG_DATA is not set
+CONFIG_INFINIBAND_SRP=m
+CONFIG_INFINIBAND_ISER=m
+
+#
+# EDAC - error detection and reporting (RAS) (EXPERIMENTAL)
+#
+CONFIG_EDAC=m
+
+#
+# Reporting subsystems
+#
+# CONFIG_EDAC_DEBUG is not set
+CONFIG_EDAC_MM_EDAC=m
+CONFIG_EDAC_AMD76X=m
+CONFIG_EDAC_E7XXX=m
+CONFIG_EDAC_E752X=m
+CONFIG_EDAC_I82875P=m
+CONFIG_EDAC_I82860=m
+CONFIG_EDAC_R82600=m
+CONFIG_EDAC_POLL=y
+
+#
+# Real Time Clock
+#
+CONFIG_RTC_LIB=m
+CONFIG_RTC_CLASS=m
+
+#
+# RTC interfaces
+#
+CONFIG_RTC_INTF_SYSFS=m
+CONFIG_RTC_INTF_PROC=m
+CONFIG_RTC_INTF_DEV=m
+CONFIG_RTC_INTF_DEV_UIE_EMUL=y
+
+#
+# RTC drivers
+#
+CONFIG_RTC_DRV_X1205=m
+CONFIG_RTC_DRV_DS1307=m
+CONFIG_RTC_DRV_DS1553=m
+CONFIG_RTC_DRV_ISL1208=m
+CONFIG_RTC_DRV_DS1672=m
+CONFIG_RTC_DRV_DS1742=m
+CONFIG_RTC_DRV_PCF8563=m
+CONFIG_RTC_DRV_PCF8583=m
+CONFIG_RTC_DRV_RS5C348=m
+CONFIG_RTC_DRV_RS5C372=m
+CONFIG_RTC_DRV_M48T86=m
+CONFIG_RTC_DRV_TEST=m
+CONFIG_RTC_DRV_MAX6902=m
+CONFIG_RTC_DRV_V3020=m
+
+#
+# DMA Engine support
+#
+CONFIG_DMA_ENGINE=y
+
+#
+# DMA Clients
+#
+CONFIG_NET_DMA=y
+
+#
+# DMA Devices
+#
+CONFIG_INTEL_IOATDMA=m
+
+#
+# File systems
+#
+CONFIG_EXT2_FS=y
+CONFIG_EXT2_FS_XATTR=y
+CONFIG_EXT2_FS_POSIX_ACL=y
+CONFIG_EXT2_FS_SECURITY=y
+CONFIG_EXT2_FS_XIP=y
+CONFIG_FS_XIP=y
+CONFIG_EXT3_FS=m
+CONFIG_EXT3_FS_XATTR=y
+CONFIG_EXT3_FS_POSIX_ACL=y
+CONFIG_EXT3_FS_SECURITY=y
+CONFIG_JBD=m
+# CONFIG_JBD_DEBUG is not set
+CONFIG_FS_MBCACHE=y
+CONFIG_REISERFS_FS=m
+# CONFIG_REISERFS_CHECK is not set
+# CONFIG_REISERFS_PROC_INFO is not set
+CONFIG_REISERFS_FS_XATTR=y
+CONFIG_REISERFS_FS_POSIX_ACL=y
+CONFIG_REISERFS_FS_SECURITY=y
+CONFIG_JFS_FS=m
+CONFIG_JFS_POSIX_ACL=y
+# CONFIG_JFS_SECURITY is not set
+# CONFIG_JFS_DEBUG is not set
+CONFIG_JFS_STATISTICS=y
+CONFIG_FS_POSIX_ACL=y
+CONFIG_XFS_FS=m
+CONFIG_XFS_QUOTA=y
+CONFIG_XFS_SECURITY=y
+CONFIG_XFS_POSIX_ACL=y
+CONFIG_XFS_RT=y
+CONFIG_OCFS2_FS=m
+CONFIG_OCFS2_DEBUG_MASKLOG=y
+CONFIG_MINIX_FS=m
+CONFIG_ROMFS_FS=m
+CONFIG_INOTIFY=y
+CONFIG_INOTIFY_USER=y
+CONFIG_QUOTA=y
+CONFIG_QFMT_V1=m
+CONFIG_QFMT_V2=m
+CONFIG_QUOTACTL=y
+CONFIG_DNOTIFY=y
+CONFIG_AUTOFS_FS=m
+CONFIG_AUTOFS4_FS=m
+CONFIG_FUSE_FS=m
+
+#
+# CD-ROM/DVD Filesystems
+#
+CONFIG_ISO9660_FS=m
+CONFIG_JOLIET=y
+CONFIG_ZISOFS=y
+CONFIG_ZISOFS_FS=m
+CONFIG_UDF_FS=m
+CONFIG_UDF_NLS=y
+
+#
+# DOS/FAT/NT Filesystems
+#
+CONFIG_FAT_FS=m
+CONFIG_MSDOS_FS=m
+CONFIG_VFAT_FS=m
+CONFIG_FAT_DEFAULT_CODEPAGE=437
+CONFIG_FAT_DEFAULT_IOCHARSET="iso8859-1"
+CONFIG_NTFS_FS=m
+# CONFIG_NTFS_DEBUG is not set
+# CONFIG_NTFS_RW is not set
+
+#
+# Pseudo filesystems
+#
+CONFIG_PROC_FS=y
+CONFIG_PROC_KCORE=y
+CONFIG_SYSFS=y
+CONFIG_TMPFS=y
+# CONFIG_HUGETLB_PAGE is not set
+CONFIG_RAMFS=y
+CONFIG_CONFIGFS_FS=m
+
+#
+# Miscellaneous filesystems
+#
+CONFIG_ADFS_FS=m
+# CONFIG_ADFS_FS_RW is not set
+CONFIG_AFFS_FS=m
+CONFIG_HFS_FS=m
+CONFIG_HFSPLUS_FS=m
+CONFIG_BEFS_FS=m
+# CONFIG_BEFS_DEBUG is not set
+CONFIG_BFS_FS=m
+CONFIG_EFS_FS=m
+CONFIG_JFFS_FS=m
+CONFIG_JFFS_FS_VERBOSE=0
+CONFIG_JFFS_PROC_FS=y
+CONFIG_JFFS2_FS=m
+CONFIG_JFFS2_FS_DEBUG=0
+CONFIG_JFFS2_FS_WRITEBUFFER=y
+# CONFIG_JFFS2_SUMMARY is not set
+# CONFIG_JFFS2_FS_XATTR is not set
+# CONFIG_JFFS2_COMPRESSION_OPTIONS is not set
+CONFIG_JFFS2_ZLIB=y
+CONFIG_JFFS2_RTIME=y
+# CONFIG_JFFS2_RUBIN is not set
+CONFIG_CRAMFS=m
+CONFIG_VXFS_FS=m
+CONFIG_HPFS_FS=m
+CONFIG_QNX4FS_FS=m
+CONFIG_SYSV_FS=m
+CONFIG_UFS_FS=m
+# CONFIG_UFS_FS_WRITE is not set
+# CONFIG_UFS_DEBUG is not set
+
+#
+# Network File Systems
+#
+CONFIG_NFS_FS=m
+CONFIG_NFS_V3=y
+CONFIG_NFS_V3_ACL=y
+CONFIG_NFS_V4=y
+CONFIG_NFS_DIRECTIO=y
+CONFIG_NFSD=m
+CONFIG_NFSD_V2_ACL=y
+CONFIG_NFSD_V3=y
+CONFIG_NFSD_V3_ACL=y
+CONFIG_NFSD_V4=y
+CONFIG_NFSD_TCP=y
+CONFIG_LOCKD=m
+CONFIG_LOCKD_V4=y
+CONFIG_EXPORTFS=m
+CONFIG_NFS_ACL_SUPPORT=m
+CONFIG_NFS_COMMON=y
+CONFIG_SUNRPC=m
+CONFIG_SUNRPC_GSS=m
+CONFIG_RPCSEC_GSS_KRB5=m
+CONFIG_RPCSEC_GSS_SPKM3=m
+CONFIG_SMB_FS=m
+CONFIG_SMB_NLS_DEFAULT=y
+CONFIG_SMB_NLS_REMOTE="cp850"
+CONFIG_CIFS=m
+CONFIG_CIFS_STATS=y
+CONFIG_CIFS_STATS2=y
+# CONFIG_CIFS_WEAK_PW_HASH is not set
+CONFIG_CIFS_XATTR=y
+# CONFIG_CIFS_POSIX is not set
+# CONFIG_CIFS_DEBUG2 is not set
+# CONFIG_CIFS_EXPERIMENTAL is not set
+CONFIG_NCP_FS=m
+CONFIG_NCPFS_PACKET_SIGNING=y
+CONFIG_NCPFS_IOCTL_LOCKING=y
+CONFIG_NCPFS_STRONG=y
+CONFIG_NCPFS_NFS_NS=y
+CONFIG_NCPFS_OS2_NS=y
+# CONFIG_NCPFS_SMALLDOS is not set
+CONFIG_NCPFS_NLS=y
+CONFIG_NCPFS_EXTRAS=y
+CONFIG_CODA_FS=m
+# CONFIG_CODA_FS_OLD_API is not set
+CONFIG_AFS_FS=m
+CONFIG_RXRPC=m
+CONFIG_9P_FS=m
+
+#
+# Partition Types
+#
+CONFIG_PARTITION_ADVANCED=y
+# CONFIG_ACORN_PARTITION is not set
+CONFIG_OSF_PARTITION=y
+# CONFIG_AMIGA_PARTITION is not set
+CONFIG_ATARI_PARTITION=y
+CONFIG_MAC_PARTITION=y
+CONFIG_MSDOS_PARTITION=y
+CONFIG_BSD_DISKLABEL=y
+# CONFIG_MINIX_SUBPARTITION is not set
+CONFIG_SOLARIS_X86_PARTITION=y
+CONFIG_UNIXWARE_DISKLABEL=y
+CONFIG_LDM_PARTITION=y
+# CONFIG_LDM_DEBUG is not set
+CONFIG_SGI_PARTITION=y
+CONFIG_ULTRIX_PARTITION=y
+CONFIG_SUN_PARTITION=y
+CONFIG_KARMA_PARTITION=y
+CONFIG_EFI_PARTITION=y
+
+#
+# Native Language Support
+#
+CONFIG_NLS=y
+CONFIG_NLS_DEFAULT="utf8"
+CONFIG_NLS_CODEPAGE_437=m
+CONFIG_NLS_CODEPAGE_737=m
+CONFIG_NLS_CODEPAGE_775=m
+CONFIG_NLS_CODEPAGE_850=m
+CONFIG_NLS_CODEPAGE_852=m
+CONFIG_NLS_CODEPAGE_855=m
+CONFIG_NLS_CODEPAGE_857=m
+CONFIG_NLS_CODEPAGE_860=m
+CONFIG_NLS_CODEPAGE_861=m
+CONFIG_NLS_CODEPAGE_862=m
+CONFIG_NLS_CODEPAGE_863=m
+CONFIG_NLS_CODEPAGE_864=m
+CONFIG_NLS_CODEPAGE_865=m
+CONFIG_NLS_CODEPAGE_866=m
+CONFIG_NLS_CODEPAGE_869=m
+CONFIG_NLS_CODEPAGE_936=m
+CONFIG_NLS_CODEPAGE_950=m
+CONFIG_NLS_CODEPAGE_932=m
+CONFIG_NLS_CODEPAGE_949=m
+CONFIG_NLS_CODEPAGE_874=m
+CONFIG_NLS_ISO8859_8=m
+CONFIG_NLS_CODEPAGE_1250=m
+CONFIG_NLS_CODEPAGE_1251=m
+CONFIG_NLS_ASCII=m
+CONFIG_NLS_ISO8859_1=m
+CONFIG_NLS_ISO8859_2=m
+CONFIG_NLS_ISO8859_3=m
+CONFIG_NLS_ISO8859_4=m
+CONFIG_NLS_ISO8859_5=m
+CONFIG_NLS_ISO8859_6=m
+CONFIG_NLS_ISO8859_7=m
+CONFIG_NLS_ISO8859_9=m
+CONFIG_NLS_ISO8859_13=m
+CONFIG_NLS_ISO8859_14=m
+CONFIG_NLS_ISO8859_15=m
+CONFIG_NLS_KOI8_R=m
+CONFIG_NLS_KOI8_U=m
+CONFIG_NLS_UTF8=m
+
+#
+# Instrumentation Support
+#
+# CONFIG_PROFILING is not set
+# CONFIG_KPROBES is not set
+
+#
+# Kernel hacking
+#
+CONFIG_TRACE_IRQFLAGS_SUPPORT=y
+# CONFIG_PRINTK_TIME is not set
+CONFIG_MAGIC_SYSRQ=y
+CONFIG_UNUSED_SYMBOLS=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_LOG_BUF_SHIFT=14
+CONFIG_DETECT_SOFTLOCKUP=y
+# CONFIG_SCHEDSTATS is not set
+# CONFIG_DEBUG_SLAB is not set
+# CONFIG_DEBUG_RT_MUTEXES is not set
+# CONFIG_RT_MUTEX_TESTER is not set
+# CONFIG_DEBUG_SPINLOCK is not set
+# CONFIG_DEBUG_MUTEXES is not set
+# CONFIG_DEBUG_RWSEMS is not set
+# CONFIG_DEBUG_LOCK_ALLOC is not set
+# CONFIG_PROVE_LOCKING is not set
+# CONFIG_DEBUG_SPINLOCK_SLEEP is not set
+# CONFIG_DEBUG_LOCKING_API_SELFTESTS is not set
+# CONFIG_DEBUG_KOBJECT is not set
+# CONFIG_DEBUG_HIGHMEM is not set
+CONFIG_DEBUG_BUGVERBOSE=y
+# CONFIG_DEBUG_INFO is not set
+# CONFIG_DEBUG_FS is not set
+# CONFIG_DEBUG_VM is not set
+# CONFIG_FRAME_POINTER is not set
+# CONFIG_UNWIND_INFO is not set
+CONFIG_FORCED_INLINING=y
+# CONFIG_RCU_TORTURE_TEST is not set
+CONFIG_EARLY_PRINTK=y
+CONFIG_DEBUG_STACKOVERFLOW=y
+# CONFIG_DEBUG_STACK_USAGE is not set
+# CONFIG_DEBUG_PAGEALLOC is not set
+# CONFIG_DEBUG_RODATA is not set
+# CONFIG_4KSTACKS is not set
+CONFIG_X86_FIND_SMP_CONFIG=y
+CONFIG_X86_MPPARSE=y
+
+#
+# Security options
+#
+CONFIG_KEYS=y
+# CONFIG_KEYS_DEBUG_PROC_KEYS is not set
+CONFIG_SECURITY=y
+CONFIG_SECURITY_NETWORK=y
+# CONFIG_SECURITY_NETWORK_XFRM is not set
+CONFIG_SECURITY_CAPABILITIES=y
+CONFIG_SECURITY_ROOTPLUG=m
+CONFIG_SECURITY_SECLVL=m
+# CONFIG_SECURITY_SELINUX is not set
+
+#
+# Cryptographic options
+#
+CONFIG_CRYPTO=y
+CONFIG_CRYPTO_HMAC=y
+CONFIG_CRYPTO_NULL=m
+CONFIG_CRYPTO_MD4=m
+CONFIG_CRYPTO_MD5=y
+CONFIG_CRYPTO_SHA1=m
+CONFIG_CRYPTO_SHA256=m
+CONFIG_CRYPTO_SHA512=m
+CONFIG_CRYPTO_WP512=m
+CONFIG_CRYPTO_TGR192=m
+CONFIG_CRYPTO_DES=m
+CONFIG_CRYPTO_BLOWFISH=m
+CONFIG_CRYPTO_TWOFISH=m
+CONFIG_CRYPTO_SERPENT=m
+CONFIG_CRYPTO_AES=m
+CONFIG_CRYPTO_AES_586=m
+CONFIG_CRYPTO_CAST5=m
+CONFIG_CRYPTO_CAST6=m
+CONFIG_CRYPTO_TEA=m
+CONFIG_CRYPTO_ARC4=m
+CONFIG_CRYPTO_KHAZAD=m
+CONFIG_CRYPTO_ANUBIS=m
+CONFIG_CRYPTO_DEFLATE=m
+CONFIG_CRYPTO_MICHAEL_MIC=m
+CONFIG_CRYPTO_CRC32C=m
+CONFIG_CRYPTO_TEST=m
+
+#
+# Hardware crypto devices
+#
+# CONFIG_CRYPTO_DEV_PADLOCK is not set
+CONFIG_XEN=y
+CONFIG_XEN_INTERFACE_VERSION=0x00030207
+
+#
+# XEN
+#
+CONFIG_XEN_PRIVILEGED_GUEST=y
+# CONFIG_XEN_UNPRIVILEGED_GUEST is not set
+CONFIG_XEN_PRIVCMD=y
+CONFIG_XEN_XENBUS_DEV=y
+CONFIG_XEN_BACKEND=y
+CONFIG_XEN_BLKDEV_BACKEND=y
+CONFIG_XEN_BLKDEV_TAP=y
+CONFIG_XEN_NETDEV_BACKEND=y
+# CONFIG_XEN_NETDEV_PIPELINED_TRANSMITTER is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_UTIL is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_BACKEND is not set
+# CONFIG_XEN_NETDEV_LOOPBACK is not set
+CONFIG_XEN_PCIDEV_BACKEND=m
+CONFIG_XEN_PCIDEV_BACKEND_VPCI=y
+# CONFIG_XEN_PCIDEV_BACKEND_PASS is not set
+# CONFIG_XEN_PCIDEV_BACKEND_SLOT is not set
+# CONFIG_XEN_PCIDEV_BACKEND_CONTROLLER is not set
+# CONFIG_XEN_PCIDEV_BE_DEBUG is not set
+CONFIG_XEN_TPMDEV_BACKEND=m
+CONFIG_XEN_SCSI_BACKEND=m
+CONFIG_XEN_BLKDEV_FRONTEND=y
+CONFIG_XEN_NETDEV_FRONTEND=y
+CONFIG_XEN_SCSI_FRONTEND=m
+CONFIG_XEN_GRANT_DEV=y
+# CONFIG_XEN_NETDEV_ACCEL_SFC_FRONTEND is not set
+CONFIG_XEN_FRAMEBUFFER=y
+CONFIG_XEN_KEYBOARD=y
+CONFIG_XEN_SCRUB_PAGES=y
+CONFIG_XEN_DISABLE_SERIAL=y
+CONFIG_XEN_SYSFS=y
+CONFIG_XEN_COMPAT_030002_AND_LATER=y
+# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
+# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
+CONFIG_XEN_COMPAT=0x030002
+CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
+CONFIG_NO_IDLE_HZ=y
+CONFIG_XEN_SMPBOOT=y
+CONFIG_XEN_BALLOON=y
+CONFIG_XEN_DEVMEM=y
+
+#
+# Library routines
+#
+CONFIG_CRC_CCITT=m
+CONFIG_CRC16=m
+CONFIG_CRC32=y
+CONFIG_LIBCRC32C=m
+CONFIG_AUDIT_GENERIC=y
+CONFIG_ZLIB_INFLATE=m
+CONFIG_ZLIB_DEFLATE=m
+CONFIG_REED_SOLOMON=m
+CONFIG_REED_SOLOMON_DEC16=y
+CONFIG_TEXTSEARCH=y
+CONFIG_TEXTSEARCH_KMP=m
+CONFIG_TEXTSEARCH_BM=m
+CONFIG_TEXTSEARCH_FSM=m
+CONFIG_PLIST=y
+CONFIG_GENERIC_HARDIRQS=y
+CONFIG_GENERIC_IRQ_PROBE=y
+CONFIG_GENERIC_PENDING_IRQ=y
+CONFIG_X86_SMP=y
+CONFIG_X86_BIOS_REBOOT=y
+CONFIG_X86_TRAMPOLINE=y
+CONFIG_X86_NO_TSS=y
+CONFIG_X86_NO_IDT=y
+CONFIG_KTIME_SCALAR=y
diff -rpuN linux-2.6.18.8/buildconfigs/linux-defconfig_xen_x86_64 linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xen_x86_64
--- linux-2.6.18.8/buildconfigs/linux-defconfig_xen_x86_64	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/buildconfigs/linux-defconfig_xen_x86_64	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,3146 @@
+#
+# Automatically generated make config: don't edit
+# Linux kernel version: 2.6.18.8
+# Mon Feb 18 10:38:24 2008
+#
+CONFIG_X86_64=y
+CONFIG_64BIT=y
+CONFIG_X86=y
+CONFIG_LOCKDEP_SUPPORT=y
+CONFIG_STACKTRACE_SUPPORT=y
+CONFIG_SEMAPHORE_SLEEPERS=y
+CONFIG_MMU=y
+CONFIG_RWSEM_GENERIC_SPINLOCK=y
+CONFIG_GENERIC_HWEIGHT=y
+CONFIG_GENERIC_CALIBRATE_DELAY=y
+CONFIG_X86_CMPXCHG=y
+CONFIG_EARLY_PRINTK=y
+CONFIG_GENERIC_ISA_DMA=y
+CONFIG_GENERIC_IOMAP=y
+CONFIG_ARCH_MAY_HAVE_PC_FDC=y
+CONFIG_DMI=y
+CONFIG_AUDIT_ARCH=y
+CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
+
+#
+# Code maturity level options
+#
+CONFIG_EXPERIMENTAL=y
+CONFIG_LOCK_KERNEL=y
+CONFIG_INIT_ENV_ARG_LIMIT=32
+
+#
+# General setup
+#
+CONFIG_LOCALVERSION=""
+# CONFIG_LOCALVERSION_AUTO is not set
+CONFIG_SWAP=y
+CONFIG_SYSVIPC=y
+CONFIG_POSIX_MQUEUE=y
+CONFIG_BSD_PROCESS_ACCT=y
+CONFIG_BSD_PROCESS_ACCT_V3=y
+CONFIG_TASKSTATS=y
+CONFIG_TASK_DELAY_ACCT=y
+CONFIG_AUDIT=y
+CONFIG_AUDITSYSCALL=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_CPUSETS=y
+# CONFIG_RELAY is not set
+CONFIG_INITRAMFS_SOURCE=""
+CONFIG_CC_OPTIMIZE_FOR_SIZE=y
+# CONFIG_EMBEDDED is not set
+CONFIG_UID16=y
+CONFIG_SYSCTL=y
+CONFIG_KALLSYMS=y
+# CONFIG_KALLSYMS_ALL is not set
+CONFIG_KALLSYMS_EXTRA_PASS=y
+CONFIG_HOTPLUG=y
+CONFIG_PRINTK=y
+CONFIG_BUG=y
+CONFIG_ELF_CORE=y
+CONFIG_BASE_FULL=y
+CONFIG_FUTEX=y
+CONFIG_EPOLL=y
+CONFIG_SHMEM=y
+CONFIG_SLAB=y
+CONFIG_VM_EVENT_COUNTERS=y
+CONFIG_RT_MUTEXES=y
+# CONFIG_TINY_SHMEM is not set
+CONFIG_BASE_SMALL=0
+# CONFIG_SLOB is not set
+
+#
+# Loadable module support
+#
+CONFIG_MODULES=y
+CONFIG_MODULE_UNLOAD=y
+CONFIG_MODULE_FORCE_UNLOAD=y
+CONFIG_MODVERSIONS=y
+CONFIG_MODULE_SRCVERSION_ALL=y
+CONFIG_KMOD=y
+CONFIG_STOP_MACHINE=y
+
+#
+# Block layer
+#
+CONFIG_LBD=y
+# CONFIG_BLK_DEV_IO_TRACE is not set
+CONFIG_LSF=y
+
+#
+# IO Schedulers
+#
+CONFIG_IOSCHED_NOOP=y
+CONFIG_IOSCHED_AS=y
+CONFIG_IOSCHED_DEADLINE=y
+CONFIG_IOSCHED_CFQ=y
+# CONFIG_DEFAULT_AS is not set
+# CONFIG_DEFAULT_DEADLINE is not set
+CONFIG_DEFAULT_CFQ=y
+# CONFIG_DEFAULT_NOOP is not set
+CONFIG_DEFAULT_IOSCHED="cfq"
+
+#
+# Processor type and features
+#
+CONFIG_X86_PC=y
+# CONFIG_X86_VSMP is not set
+# CONFIG_MK8 is not set
+# CONFIG_MPSC is not set
+CONFIG_GENERIC_CPU=y
+CONFIG_X86_64_XEN=y
+CONFIG_X86_NO_TSS=y
+CONFIG_X86_NO_IDT=y
+CONFIG_X86_L1_CACHE_BYTES=128
+CONFIG_X86_L1_CACHE_SHIFT=7
+CONFIG_X86_INTERNODE_CACHE_BYTES=128
+CONFIG_X86_GOOD_APIC=y
+CONFIG_MICROCODE=y
+CONFIG_X86_MSR=m
+CONFIG_X86_CPUID=m
+CONFIG_X86_IO_APIC=y
+CONFIG_X86_XEN_GENAPIC=y
+CONFIG_X86_LOCAL_APIC=y
+CONFIG_MTRR=y
+CONFIG_SMP=y
+# CONFIG_PREEMPT_NONE is not set
+CONFIG_PREEMPT_VOLUNTARY=y
+# CONFIG_PREEMPT is not set
+CONFIG_PREEMPT_BKL=y
+CONFIG_ARCH_FLATMEM_ENABLE=y
+CONFIG_SELECT_MEMORY_MODEL=y
+CONFIG_FLATMEM_MANUAL=y
+# CONFIG_DISCONTIGMEM_MANUAL is not set
+# CONFIG_SPARSEMEM_MANUAL is not set
+CONFIG_FLATMEM=y
+CONFIG_FLAT_NODE_MEM_MAP=y
+# CONFIG_SPARSEMEM_STATIC is not set
+CONFIG_SPLIT_PTLOCK_CPUS=4
+CONFIG_RESOURCES_64BIT=y
+CONFIG_NR_CPUS=32
+CONFIG_HOTPLUG_CPU=y
+CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y
+CONFIG_SWIOTLB=y
+CONFIG_KEXEC=y
+# CONFIG_CRASH_DUMP is not set
+CONFIG_PHYSICAL_START=0x200000
+CONFIG_SECCOMP=y
+CONFIG_HZ_100=y
+# CONFIG_HZ_250 is not set
+# CONFIG_HZ_1000 is not set
+CONFIG_HZ=100
+# CONFIG_REORDER is not set
+CONFIG_K8_NB=y
+CONFIG_GENERIC_HARDIRQS=y
+CONFIG_GENERIC_IRQ_PROBE=y
+CONFIG_ISA_DMA_API=y
+CONFIG_GENERIC_PENDING_IRQ=y
+
+#
+# Power management options
+#
+CONFIG_PM=y
+# CONFIG_PM_LEGACY is not set
+CONFIG_PM_DEBUG=y
+# CONFIG_SOFTWARE_SUSPEND is not set
+CONFIG_SUSPEND_SMP=y
+
+#
+# ACPI (Advanced Configuration and Power Interface) Support
+#
+CONFIG_ACPI=y
+CONFIG_ACPI_SLEEP=y
+CONFIG_ACPI_SLEEP_PROC_FS=y
+# CONFIG_ACPI_SLEEP_PROC_SLEEP is not set
+CONFIG_ACPI_AC=m
+CONFIG_ACPI_BATTERY=m
+CONFIG_ACPI_BUTTON=m
+CONFIG_ACPI_VIDEO=m
+CONFIG_ACPI_HOTKEY=m
+CONFIG_ACPI_FAN=m
+CONFIG_ACPI_DOCK=m
+CONFIG_ACPI_PROCESSOR=m
+CONFIG_ACPI_HOTPLUG_CPU=y
+CONFIG_ACPI_THERMAL=m
+CONFIG_ACPI_ASUS=m
+CONFIG_ACPI_IBM=m
+CONFIG_ACPI_TOSHIBA=m
+CONFIG_ACPI_BLACKLIST_YEAR=0
+# CONFIG_ACPI_DEBUG is not set
+CONFIG_ACPI_EC=y
+CONFIG_ACPI_POWER=y
+CONFIG_ACPI_SYSTEM=y
+CONFIG_ACPI_CONTAINER=m
+CONFIG_ACPI_SBS=m
+CONFIG_ACPI_PV_SLEEP=y
+
+#
+# CPU Frequency scaling
+#
+# CONFIG_CPU_FREQ is not set
+
+#
+# Bus options (PCI etc.)
+#
+CONFIG_PCI=y
+CONFIG_PCI_DIRECT=y
+# CONFIG_PCI_MMCONFIG is not set
+CONFIG_XEN_PCIDEV_FRONTEND=y
+# CONFIG_XEN_PCIDEV_FE_DEBUG is not set
+# CONFIG_PCIEPORTBUS is not set
+# CONFIG_PCI_MSI is not set
+# CONFIG_PCI_DEBUG is not set
+
+#
+# PCCARD (PCMCIA/CardBus) support
+#
+CONFIG_PCCARD=m
+# CONFIG_PCMCIA_DEBUG is not set
+CONFIG_PCMCIA=m
+CONFIG_PCMCIA_LOAD_CIS=y
+CONFIG_PCMCIA_IOCTL=y
+CONFIG_CARDBUS=y
+
+#
+# PC-card bridges
+#
+CONFIG_YENTA=m
+CONFIG_YENTA_O2=y
+CONFIG_YENTA_RICOH=y
+CONFIG_YENTA_TI=y
+CONFIG_YENTA_ENE_TUNE=y
+CONFIG_YENTA_TOSHIBA=y
+CONFIG_PD6729=m
+CONFIG_I82092=m
+CONFIG_PCCARD_NONSTATIC=m
+
+#
+# PCI Hotplug Support
+#
+CONFIG_HOTPLUG_PCI=m
+# CONFIG_HOTPLUG_PCI_FAKE is not set
+# CONFIG_HOTPLUG_PCI_ACPI is not set
+# CONFIG_HOTPLUG_PCI_CPCI is not set
+# CONFIG_HOTPLUG_PCI_SHPC is not set
+
+#
+# Executable file formats / Emulations
+#
+CONFIG_BINFMT_ELF=y
+CONFIG_BINFMT_MISC=m
+CONFIG_IA32_EMULATION=y
+CONFIG_IA32_AOUT=y
+CONFIG_COMPAT=y
+CONFIG_SYSVIPC_COMPAT=y
+
+#
+# Networking
+#
+CONFIG_NET=y
+
+#
+# Networking options
+#
+# CONFIG_NETDEBUG is not set
+CONFIG_PACKET=y
+CONFIG_PACKET_MMAP=y
+CONFIG_UNIX=y
+CONFIG_XFRM=y
+CONFIG_XFRM_USER=m
+CONFIG_NET_KEY=m
+CONFIG_INET=y
+CONFIG_IP_MULTICAST=y
+CONFIG_IP_ADVANCED_ROUTER=y
+CONFIG_ASK_IP_FIB_HASH=y
+# CONFIG_IP_FIB_TRIE is not set
+CONFIG_IP_FIB_HASH=y
+CONFIG_IP_MULTIPLE_TABLES=y
+CONFIG_IP_ROUTE_FWMARK=y
+CONFIG_IP_ROUTE_MULTIPATH=y
+# CONFIG_IP_ROUTE_MULTIPATH_CACHED is not set
+CONFIG_IP_ROUTE_VERBOSE=y
+CONFIG_IP_PNP=y
+CONFIG_IP_PNP_DHCP=y
+CONFIG_IP_PNP_BOOTP=y
+CONFIG_IP_PNP_RARP=y
+CONFIG_NET_IPIP=m
+CONFIG_NET_IPGRE=m
+CONFIG_NET_IPGRE_BROADCAST=y
+CONFIG_IP_MROUTE=y
+CONFIG_IP_PIMSM_V1=y
+CONFIG_IP_PIMSM_V2=y
+# CONFIG_ARPD is not set
+CONFIG_SYN_COOKIES=y
+CONFIG_INET_AH=m
+CONFIG_INET_ESP=m
+CONFIG_INET_IPCOMP=m
+CONFIG_INET_XFRM_TUNNEL=m
+CONFIG_INET_TUNNEL=m
+CONFIG_INET_XFRM_MODE_TRANSPORT=m
+CONFIG_INET_XFRM_MODE_TUNNEL=m
+CONFIG_INET_DIAG=m
+CONFIG_INET_TCP_DIAG=m
+CONFIG_TCP_CONG_ADVANCED=y
+
+#
+# TCP congestion control
+#
+CONFIG_TCP_CONG_BIC=m
+CONFIG_TCP_CONG_CUBIC=m
+CONFIG_TCP_CONG_WESTWOOD=m
+CONFIG_TCP_CONG_HTCP=m
+CONFIG_TCP_CONG_HSTCP=m
+CONFIG_TCP_CONG_HYBLA=m
+CONFIG_TCP_CONG_VEGAS=m
+CONFIG_TCP_CONG_SCALABLE=m
+CONFIG_TCP_CONG_LP=m
+CONFIG_TCP_CONG_VENO=m
+
+#
+# IP: Virtual Server Configuration
+#
+CONFIG_IP_VS=m
+# CONFIG_IP_VS_DEBUG is not set
+CONFIG_IP_VS_TAB_BITS=12
+
+#
+# IPVS transport protocol load balancing support
+#
+CONFIG_IP_VS_PROTO_TCP=y
+CONFIG_IP_VS_PROTO_UDP=y
+CONFIG_IP_VS_PROTO_ESP=y
+CONFIG_IP_VS_PROTO_AH=y
+
+#
+# IPVS scheduler
+#
+CONFIG_IP_VS_RR=m
+CONFIG_IP_VS_WRR=m
+CONFIG_IP_VS_LC=m
+CONFIG_IP_VS_WLC=m
+CONFIG_IP_VS_LBLC=m
+CONFIG_IP_VS_LBLCR=m
+CONFIG_IP_VS_DH=m
+CONFIG_IP_VS_SH=m
+CONFIG_IP_VS_SED=m
+CONFIG_IP_VS_NQ=m
+
+#
+# IPVS application helper
+#
+CONFIG_IP_VS_FTP=m
+CONFIG_IPV6=m
+CONFIG_IPV6_PRIVACY=y
+# CONFIG_IPV6_ROUTER_PREF is not set
+CONFIG_INET6_AH=m
+CONFIG_INET6_ESP=m
+CONFIG_INET6_IPCOMP=m
+CONFIG_INET6_XFRM_TUNNEL=m
+CONFIG_INET6_TUNNEL=m
+CONFIG_INET6_XFRM_MODE_TRANSPORT=m
+CONFIG_INET6_XFRM_MODE_TUNNEL=m
+CONFIG_IPV6_TUNNEL=m
+CONFIG_NETWORK_SECMARK=y
+CONFIG_NETFILTER=y
+# CONFIG_NETFILTER_DEBUG is not set
+CONFIG_BRIDGE_NETFILTER=y
+
+#
+# Core Netfilter Configuration
+#
+CONFIG_NETFILTER_NETLINK=m
+CONFIG_NETFILTER_NETLINK_QUEUE=m
+CONFIG_NETFILTER_NETLINK_LOG=m
+CONFIG_NETFILTER_XTABLES=m
+CONFIG_NETFILTER_XT_TARGET_CLASSIFY=m
+CONFIG_NETFILTER_XT_TARGET_CONNMARK=m
+CONFIG_NETFILTER_XT_TARGET_MARK=m
+CONFIG_NETFILTER_XT_TARGET_NFQUEUE=m
+CONFIG_NETFILTER_XT_TARGET_NOTRACK=m
+CONFIG_NETFILTER_XT_TARGET_SECMARK=m
+CONFIG_NETFILTER_XT_TARGET_CONNSECMARK=m
+CONFIG_NETFILTER_XT_MATCH_COMMENT=m
+CONFIG_NETFILTER_XT_MATCH_CONNBYTES=m
+CONFIG_NETFILTER_XT_MATCH_CONNMARK=m
+CONFIG_NETFILTER_XT_MATCH_CONNTRACK=m
+CONFIG_NETFILTER_XT_MATCH_DCCP=m
+CONFIG_NETFILTER_XT_MATCH_ESP=m
+CONFIG_NETFILTER_XT_MATCH_HELPER=m
+CONFIG_NETFILTER_XT_MATCH_LENGTH=m
+CONFIG_NETFILTER_XT_MATCH_LIMIT=m
+CONFIG_NETFILTER_XT_MATCH_MAC=m
+CONFIG_NETFILTER_XT_MATCH_MARK=m
+CONFIG_NETFILTER_XT_MATCH_POLICY=m
+CONFIG_NETFILTER_XT_MATCH_MULTIPORT=m
+CONFIG_NETFILTER_XT_MATCH_PHYSDEV=m
+CONFIG_NETFILTER_XT_MATCH_PKTTYPE=m
+CONFIG_NETFILTER_XT_MATCH_QUOTA=m
+CONFIG_NETFILTER_XT_MATCH_REALM=m
+CONFIG_NETFILTER_XT_MATCH_SCTP=m
+CONFIG_NETFILTER_XT_MATCH_STATE=m
+CONFIG_NETFILTER_XT_MATCH_STATISTIC=m
+CONFIG_NETFILTER_XT_MATCH_STRING=m
+CONFIG_NETFILTER_XT_MATCH_TCPMSS=m
+
+#
+# IP: Netfilter Configuration
+#
+CONFIG_IP_NF_CONNTRACK=m
+CONFIG_IP_NF_CT_ACCT=y
+CONFIG_IP_NF_CONNTRACK_MARK=y
+CONFIG_IP_NF_CONNTRACK_SECMARK=y
+CONFIG_IP_NF_CONNTRACK_EVENTS=y
+CONFIG_IP_NF_CONNTRACK_NETLINK=m
+CONFIG_IP_NF_CT_PROTO_SCTP=m
+CONFIG_IP_NF_FTP=m
+CONFIG_IP_NF_IRC=m
+CONFIG_IP_NF_NETBIOS_NS=m
+CONFIG_IP_NF_TFTP=m
+CONFIG_IP_NF_AMANDA=m
+CONFIG_IP_NF_PPTP=m
+CONFIG_IP_NF_H323=m
+CONFIG_IP_NF_SIP=m
+CONFIG_IP_NF_QUEUE=m
+CONFIG_IP_NF_IPTABLES=m
+CONFIG_IP_NF_MATCH_IPRANGE=m
+CONFIG_IP_NF_MATCH_TOS=m
+CONFIG_IP_NF_MATCH_RECENT=m
+CONFIG_IP_NF_MATCH_ECN=m
+CONFIG_IP_NF_MATCH_DSCP=m
+CONFIG_IP_NF_MATCH_AH=m
+CONFIG_IP_NF_MATCH_TTL=m
+CONFIG_IP_NF_MATCH_OWNER=m
+CONFIG_IP_NF_MATCH_ADDRTYPE=m
+CONFIG_IP_NF_MATCH_HASHLIMIT=m
+CONFIG_IP_NF_FILTER=m
+CONFIG_IP_NF_TARGET_REJECT=m
+CONFIG_IP_NF_TARGET_LOG=m
+CONFIG_IP_NF_TARGET_ULOG=m
+CONFIG_IP_NF_TARGET_TCPMSS=m
+CONFIG_IP_NF_NAT=m
+CONFIG_IP_NF_NAT_NEEDED=y
+CONFIG_IP_NF_TARGET_MASQUERADE=m
+CONFIG_IP_NF_TARGET_REDIRECT=m
+CONFIG_IP_NF_TARGET_NETMAP=m
+CONFIG_IP_NF_TARGET_SAME=m
+CONFIG_IP_NF_NAT_SNMP_BASIC=m
+CONFIG_IP_NF_NAT_IRC=m
+CONFIG_IP_NF_NAT_FTP=m
+CONFIG_IP_NF_NAT_TFTP=m
+CONFIG_IP_NF_NAT_AMANDA=m
+CONFIG_IP_NF_NAT_PPTP=m
+CONFIG_IP_NF_NAT_H323=m
+CONFIG_IP_NF_NAT_SIP=m
+CONFIG_IP_NF_MANGLE=m
+CONFIG_IP_NF_TARGET_TOS=m
+CONFIG_IP_NF_TARGET_ECN=m
+CONFIG_IP_NF_TARGET_DSCP=m
+CONFIG_IP_NF_TARGET_TTL=m
+CONFIG_IP_NF_TARGET_CLUSTERIP=m
+CONFIG_IP_NF_RAW=m
+CONFIG_IP_NF_ARPTABLES=m
+CONFIG_IP_NF_ARPFILTER=m
+CONFIG_IP_NF_ARP_MANGLE=m
+
+#
+# IPv6: Netfilter Configuration (EXPERIMENTAL)
+#
+CONFIG_IP6_NF_QUEUE=m
+CONFIG_IP6_NF_IPTABLES=m
+CONFIG_IP6_NF_MATCH_RT=m
+CONFIG_IP6_NF_MATCH_OPTS=m
+CONFIG_IP6_NF_MATCH_FRAG=m
+CONFIG_IP6_NF_MATCH_HL=m
+CONFIG_IP6_NF_MATCH_OWNER=m
+CONFIG_IP6_NF_MATCH_IPV6HEADER=m
+CONFIG_IP6_NF_MATCH_AH=m
+CONFIG_IP6_NF_MATCH_EUI64=m
+CONFIG_IP6_NF_FILTER=m
+CONFIG_IP6_NF_TARGET_LOG=m
+CONFIG_IP6_NF_TARGET_REJECT=m
+CONFIG_IP6_NF_MANGLE=m
+CONFIG_IP6_NF_TARGET_HL=m
+CONFIG_IP6_NF_RAW=m
+
+#
+# DECnet: Netfilter Configuration
+#
+# CONFIG_DECNET_NF_GRABULATOR is not set
+
+#
+# Bridge: Netfilter Configuration
+#
+CONFIG_BRIDGE_NF_EBTABLES=m
+CONFIG_BRIDGE_EBT_BROUTE=m
+CONFIG_BRIDGE_EBT_T_FILTER=m
+CONFIG_BRIDGE_EBT_T_NAT=m
+CONFIG_BRIDGE_EBT_802_3=m
+CONFIG_BRIDGE_EBT_AMONG=m
+CONFIG_BRIDGE_EBT_ARP=m
+CONFIG_BRIDGE_EBT_IP=m
+CONFIG_BRIDGE_EBT_LIMIT=m
+CONFIG_BRIDGE_EBT_MARK=m
+CONFIG_BRIDGE_EBT_PKTTYPE=m
+CONFIG_BRIDGE_EBT_STP=m
+CONFIG_BRIDGE_EBT_VLAN=m
+CONFIG_BRIDGE_EBT_ARPREPLY=m
+CONFIG_BRIDGE_EBT_DNAT=m
+CONFIG_BRIDGE_EBT_MARK_T=m
+CONFIG_BRIDGE_EBT_REDIRECT=m
+CONFIG_BRIDGE_EBT_SNAT=m
+CONFIG_BRIDGE_EBT_LOG=m
+CONFIG_BRIDGE_EBT_ULOG=m
+
+#
+# DCCP Configuration (EXPERIMENTAL)
+#
+CONFIG_IP_DCCP=m
+CONFIG_INET_DCCP_DIAG=m
+CONFIG_IP_DCCP_ACKVEC=y
+
+#
+# DCCP CCIDs Configuration (EXPERIMENTAL)
+#
+CONFIG_IP_DCCP_CCID2=m
+CONFIG_IP_DCCP_CCID3=m
+CONFIG_IP_DCCP_TFRC_LIB=m
+
+#
+# DCCP Kernel Hacking
+#
+# CONFIG_IP_DCCP_DEBUG is not set
+
+#
+# SCTP Configuration (EXPERIMENTAL)
+#
+CONFIG_IP_SCTP=m
+# CONFIG_SCTP_DBG_MSG is not set
+# CONFIG_SCTP_DBG_OBJCNT is not set
+# CONFIG_SCTP_HMAC_NONE is not set
+# CONFIG_SCTP_HMAC_SHA1 is not set
+CONFIG_SCTP_HMAC_MD5=y
+
+#
+# TIPC Configuration (EXPERIMENTAL)
+#
+CONFIG_TIPC=m
+# CONFIG_TIPC_ADVANCED is not set
+# CONFIG_TIPC_DEBUG is not set
+CONFIG_ATM=m
+CONFIG_ATM_CLIP=m
+# CONFIG_ATM_CLIP_NO_ICMP is not set
+CONFIG_ATM_LANE=m
+# CONFIG_ATM_MPOA is not set
+CONFIG_ATM_BR2684=m
+# CONFIG_ATM_BR2684_IPFILTER is not set
+CONFIG_BRIDGE=m
+CONFIG_VLAN_8021Q=m
+CONFIG_DECNET=m
+# CONFIG_DECNET_ROUTER is not set
+CONFIG_LLC=y
+CONFIG_LLC2=m
+CONFIG_IPX=m
+CONFIG_IPX_INTERN=y
+CONFIG_ATALK=m
+CONFIG_DEV_APPLETALK=m
+CONFIG_IPDDP=m
+CONFIG_IPDDP_ENCAP=y
+CONFIG_IPDDP_DECAP=y
+CONFIG_X25=m
+CONFIG_LAPB=m
+CONFIG_ECONET=m
+# CONFIG_ECONET_AUNUDP is not set
+# CONFIG_ECONET_NATIVE is not set
+CONFIG_WAN_ROUTER=m
+
+#
+# QoS and/or fair queueing
+#
+CONFIG_NET_SCHED=y
+CONFIG_NET_SCH_CLK_JIFFIES=y
+# CONFIG_NET_SCH_CLK_GETTIMEOFDAY is not set
+# CONFIG_NET_SCH_CLK_CPU is not set
+
+#
+# Queueing/Scheduling
+#
+CONFIG_NET_SCH_CBQ=m
+CONFIG_NET_SCH_HTB=m
+CONFIG_NET_SCH_HFSC=m
+CONFIG_NET_SCH_ATM=m
+CONFIG_NET_SCH_PRIO=m
+CONFIG_NET_SCH_RED=m
+CONFIG_NET_SCH_SFQ=m
+CONFIG_NET_SCH_TEQL=m
+CONFIG_NET_SCH_TBF=m
+CONFIG_NET_SCH_GRED=m
+CONFIG_NET_SCH_DSMARK=m
+CONFIG_NET_SCH_NETEM=m
+CONFIG_NET_SCH_INGRESS=m
+
+#
+# Classification
+#
+CONFIG_NET_CLS=y
+CONFIG_NET_CLS_BASIC=m
+CONFIG_NET_CLS_TCINDEX=m
+CONFIG_NET_CLS_ROUTE4=m
+CONFIG_NET_CLS_ROUTE=y
+CONFIG_NET_CLS_FW=m
+CONFIG_NET_CLS_U32=m
+CONFIG_CLS_U32_PERF=y
+CONFIG_CLS_U32_MARK=y
+CONFIG_NET_CLS_RSVP=m
+CONFIG_NET_CLS_RSVP6=m
+CONFIG_NET_EMATCH=y
+CONFIG_NET_EMATCH_STACK=32
+CONFIG_NET_EMATCH_CMP=m
+CONFIG_NET_EMATCH_NBYTE=m
+CONFIG_NET_EMATCH_U32=m
+CONFIG_NET_EMATCH_META=m
+CONFIG_NET_EMATCH_TEXT=m
+# CONFIG_NET_CLS_ACT is not set
+CONFIG_NET_CLS_POLICE=y
+CONFIG_NET_CLS_IND=y
+CONFIG_NET_ESTIMATOR=y
+
+#
+# Network testing
+#
+CONFIG_NET_PKTGEN=m
+CONFIG_HAMRADIO=y
+
+#
+# Packet Radio protocols
+#
+CONFIG_AX25=m
+CONFIG_AX25_DAMA_SLAVE=y
+CONFIG_NETROM=m
+CONFIG_ROSE=m
+
+#
+# AX.25 network device drivers
+#
+CONFIG_MKISS=m
+CONFIG_6PACK=m
+CONFIG_BPQETHER=m
+CONFIG_BAYCOM_SER_FDX=m
+CONFIG_BAYCOM_SER_HDX=m
+CONFIG_BAYCOM_PAR=m
+CONFIG_YAM=m
+CONFIG_IRDA=m
+
+#
+# IrDA protocols
+#
+CONFIG_IRLAN=m
+CONFIG_IRNET=m
+CONFIG_IRCOMM=m
+CONFIG_IRDA_ULTRA=y
+
+#
+# IrDA options
+#
+CONFIG_IRDA_CACHE_LAST_LSAP=y
+CONFIG_IRDA_FAST_RR=y
+# CONFIG_IRDA_DEBUG is not set
+
+#
+# Infrared-port device drivers
+#
+
+#
+# SIR device drivers
+#
+CONFIG_IRTTY_SIR=m
+
+#
+# Dongle support
+#
+CONFIG_DONGLE=y
+CONFIG_ESI_DONGLE=m
+CONFIG_ACTISYS_DONGLE=m
+CONFIG_TEKRAM_DONGLE=m
+CONFIG_TOIM3232_DONGLE=m
+CONFIG_LITELINK_DONGLE=m
+CONFIG_MA600_DONGLE=m
+CONFIG_GIRBIL_DONGLE=m
+CONFIG_MCP2120_DONGLE=m
+CONFIG_OLD_BELKIN_DONGLE=m
+CONFIG_ACT200L_DONGLE=m
+
+#
+# Old SIR device drivers
+#
+
+#
+# Old Serial dongle support
+#
+
+#
+# FIR device drivers
+#
+CONFIG_USB_IRDA=m
+CONFIG_SIGMATEL_FIR=m
+CONFIG_NSC_FIR=m
+CONFIG_WINBOND_FIR=m
+CONFIG_SMC_IRCC_FIR=m
+CONFIG_ALI_FIR=m
+CONFIG_VLSI_FIR=m
+CONFIG_VIA_FIR=m
+CONFIG_MCS_FIR=m
+CONFIG_BT=m
+CONFIG_BT_L2CAP=m
+CONFIG_BT_SCO=m
+CONFIG_BT_RFCOMM=m
+CONFIG_BT_RFCOMM_TTY=y
+CONFIG_BT_BNEP=m
+CONFIG_BT_BNEP_MC_FILTER=y
+CONFIG_BT_BNEP_PROTO_FILTER=y
+CONFIG_BT_CMTP=m
+CONFIG_BT_HIDP=m
+
+#
+# Bluetooth device drivers
+#
+CONFIG_BT_HCIUSB=m
+CONFIG_BT_HCIUSB_SCO=y
+CONFIG_BT_HCIUART=m
+CONFIG_BT_HCIUART_H4=y
+CONFIG_BT_HCIUART_BCSP=y
+CONFIG_BT_HCIBCM203X=m
+CONFIG_BT_HCIBPA10X=m
+CONFIG_BT_HCIBFUSB=m
+# CONFIG_BT_HCIDTL1 is not set
+# CONFIG_BT_HCIBT3C is not set
+# CONFIG_BT_HCIBLUECARD is not set
+# CONFIG_BT_HCIBTUART is not set
+CONFIG_BT_HCIVHCI=m
+CONFIG_IEEE80211=m
+# CONFIG_IEEE80211_DEBUG is not set
+CONFIG_IEEE80211_CRYPT_WEP=m
+CONFIG_IEEE80211_CRYPT_CCMP=m
+CONFIG_IEEE80211_CRYPT_TKIP=m
+CONFIG_IEEE80211_SOFTMAC=m
+# CONFIG_IEEE80211_SOFTMAC_DEBUG is not set
+CONFIG_WIRELESS_EXT=y
+
+#
+# Device Drivers
+#
+
+#
+# Generic Driver Options
+#
+CONFIG_STANDALONE=y
+CONFIG_PREVENT_FIRMWARE_BUILD=y
+CONFIG_FW_LOADER=m
+# CONFIG_DEBUG_DRIVER is not set
+CONFIG_SYS_HYPERVISOR=y
+
+#
+# Connector - unified userspace <-> kernelspace linker
+#
+CONFIG_CONNECTOR=y
+CONFIG_PROC_EVENTS=y
+
+#
+# Memory Technology Devices (MTD)
+#
+CONFIG_MTD=m
+# CONFIG_MTD_DEBUG is not set
+CONFIG_MTD_CONCAT=m
+CONFIG_MTD_PARTITIONS=y
+CONFIG_MTD_REDBOOT_PARTS=m
+CONFIG_MTD_REDBOOT_DIRECTORY_BLOCK=-1
+# CONFIG_MTD_REDBOOT_PARTS_UNALLOCATED is not set
+# CONFIG_MTD_REDBOOT_PARTS_READONLY is not set
+CONFIG_MTD_CMDLINE_PARTS=y
+
+#
+# User Modules And Translation Layers
+#
+CONFIG_MTD_CHAR=m
+CONFIG_MTD_BLOCK=m
+CONFIG_MTD_BLOCK_RO=m
+CONFIG_FTL=m
+CONFIG_NFTL=m
+CONFIG_NFTL_RW=y
+CONFIG_INFTL=m
+CONFIG_RFD_FTL=m
+
+#
+# RAM/ROM/Flash chip drivers
+#
+CONFIG_MTD_CFI=m
+CONFIG_MTD_JEDECPROBE=m
+CONFIG_MTD_GEN_PROBE=m
+CONFIG_MTD_CFI_ADV_OPTIONS=y
+CONFIG_MTD_CFI_NOSWAP=y
+# CONFIG_MTD_CFI_BE_BYTE_SWAP is not set
+# CONFIG_MTD_CFI_LE_BYTE_SWAP is not set
+# CONFIG_MTD_CFI_GEOMETRY is not set
+CONFIG_MTD_MAP_BANK_WIDTH_1=y
+CONFIG_MTD_MAP_BANK_WIDTH_2=y
+CONFIG_MTD_MAP_BANK_WIDTH_4=y
+# CONFIG_MTD_MAP_BANK_WIDTH_8 is not set
+# CONFIG_MTD_MAP_BANK_WIDTH_16 is not set
+# CONFIG_MTD_MAP_BANK_WIDTH_32 is not set
+CONFIG_MTD_CFI_I1=y
+CONFIG_MTD_CFI_I2=y
+# CONFIG_MTD_CFI_I4 is not set
+# CONFIG_MTD_CFI_I8 is not set
+# CONFIG_MTD_OTP is not set
+CONFIG_MTD_CFI_INTELEXT=m
+CONFIG_MTD_CFI_AMDSTD=m
+CONFIG_MTD_CFI_STAA=m
+CONFIG_MTD_CFI_UTIL=m
+CONFIG_MTD_RAM=m
+# CONFIG_MTD_ROM is not set
+CONFIG_MTD_ABSENT=m
+# CONFIG_MTD_OBSOLETE_CHIPS is not set
+
+#
+# Mapping drivers for chip access
+#
+CONFIG_MTD_COMPLEX_MAPPINGS=y
+CONFIG_MTD_PHYSMAP=m
+CONFIG_MTD_PHYSMAP_START=0x8000000
+CONFIG_MTD_PHYSMAP_LEN=0x4000000
+CONFIG_MTD_PHYSMAP_BANKWIDTH=2
+# CONFIG_MTD_PNC2000 is not set
+CONFIG_MTD_SC520CDP=m
+CONFIG_MTD_NETSC520=m
+CONFIG_MTD_TS5500=m
+CONFIG_MTD_SBC_GXX=m
+# CONFIG_MTD_AMD76XROM is not set
+# CONFIG_MTD_ICHXROM is not set
+CONFIG_MTD_SCB2_FLASH=m
+# CONFIG_MTD_NETtel is not set
+# CONFIG_MTD_DILNETPC is not set
+# CONFIG_MTD_L440GX is not set
+CONFIG_MTD_PCI=m
+CONFIG_MTD_PLATRAM=m
+
+#
+# Self-contained MTD device drivers
+#
+CONFIG_MTD_PMC551=m
+# CONFIG_MTD_PMC551_BUGFIX is not set
+# CONFIG_MTD_PMC551_DEBUG is not set
+CONFIG_MTD_DATAFLASH=m
+CONFIG_MTD_M25P80=m
+# CONFIG_MTD_SLRAM is not set
+# CONFIG_MTD_PHRAM is not set
+CONFIG_MTD_MTDRAM=m
+CONFIG_MTDRAM_TOTAL_SIZE=4096
+CONFIG_MTDRAM_ERASE_SIZE=128
+CONFIG_MTD_BLOCK2MTD=m
+
+#
+# Disk-On-Chip Device Drivers
+#
+CONFIG_MTD_DOC2000=m
+CONFIG_MTD_DOC2001=m
+CONFIG_MTD_DOC2001PLUS=m
+CONFIG_MTD_DOCPROBE=m
+CONFIG_MTD_DOCECC=m
+CONFIG_MTD_DOCPROBE_ADVANCED=y
+CONFIG_MTD_DOCPROBE_ADDRESS=0x0000
+CONFIG_MTD_DOCPROBE_HIGH=y
+CONFIG_MTD_DOCPROBE_55AA=y
+
+#
+# NAND Flash Device Drivers
+#
+CONFIG_MTD_NAND=m
+# CONFIG_MTD_NAND_VERIFY_WRITE is not set
+# CONFIG_MTD_NAND_ECC_SMC is not set
+CONFIG_MTD_NAND_IDS=m
+CONFIG_MTD_NAND_DISKONCHIP=m
+# CONFIG_MTD_NAND_DISKONCHIP_PROBE_ADVANCED is not set
+CONFIG_MTD_NAND_DISKONCHIP_PROBE_ADDRESS=0
+CONFIG_MTD_NAND_DISKONCHIP_BBTWRITE=y
+CONFIG_MTD_NAND_NANDSIM=m
+
+#
+# OneNAND Flash Device Drivers
+#
+CONFIG_MTD_ONENAND=m
+# CONFIG_MTD_ONENAND_VERIFY_WRITE is not set
+# CONFIG_MTD_ONENAND_OTP is not set
+
+#
+# Parallel port support
+#
+CONFIG_PARPORT=m
+CONFIG_PARPORT_PC=m
+# CONFIG_PARPORT_PC_FIFO is not set
+# CONFIG_PARPORT_PC_SUPERIO is not set
+# CONFIG_PARPORT_PC_PCMCIA is not set
+CONFIG_PARPORT_NOT_PC=y
+# CONFIG_PARPORT_GSC is not set
+CONFIG_PARPORT_AX88796=m
+CONFIG_PARPORT_1284=y
+
+#
+# Plug and Play support
+#
+CONFIG_PNP=y
+CONFIG_PNP_DEBUG=y
+
+#
+# Protocols
+#
+CONFIG_PNPACPI=y
+
+#
+# Block devices
+#
+CONFIG_BLK_DEV_FD=m
+CONFIG_PARIDE=m
+CONFIG_PARIDE_PARPORT=m
+
+#
+# Parallel IDE high-level drivers
+#
+CONFIG_PARIDE_PD=m
+CONFIG_PARIDE_PCD=m
+CONFIG_PARIDE_PF=m
+CONFIG_PARIDE_PT=m
+CONFIG_PARIDE_PG=m
+
+#
+# Parallel IDE protocol modules
+#
+CONFIG_PARIDE_ATEN=m
+CONFIG_PARIDE_BPCK=m
+CONFIG_PARIDE_COMM=m
+CONFIG_PARIDE_DSTR=m
+CONFIG_PARIDE_FIT2=m
+CONFIG_PARIDE_FIT3=m
+CONFIG_PARIDE_EPAT=m
+CONFIG_PARIDE_EPATC8=y
+CONFIG_PARIDE_EPIA=m
+CONFIG_PARIDE_FRIQ=m
+CONFIG_PARIDE_FRPW=m
+CONFIG_PARIDE_KBIC=m
+CONFIG_PARIDE_KTTI=m
+CONFIG_PARIDE_ON20=m
+CONFIG_PARIDE_ON26=m
+CONFIG_BLK_CPQ_DA=m
+CONFIG_BLK_CPQ_CISS_DA=m
+CONFIG_CISS_SCSI_TAPE=y
+CONFIG_BLK_DEV_DAC960=m
+CONFIG_BLK_DEV_UMEM=m
+# CONFIG_BLK_DEV_COW_COMMON is not set
+CONFIG_BLK_DEV_LOOP=y
+CONFIG_BLK_DEV_CRYPTOLOOP=m
+CONFIG_BLK_DEV_NBD=m
+CONFIG_BLK_DEV_SX8=m
+# CONFIG_BLK_DEV_UB is not set
+CONFIG_BLK_DEV_RAM=y
+CONFIG_BLK_DEV_RAM_COUNT=16
+CONFIG_BLK_DEV_RAM_SIZE=16384
+CONFIG_BLK_DEV_RAM_BLOCKSIZE=1024
+CONFIG_BLK_DEV_INITRD=y
+CONFIG_CDROM_PKTCDVD=m
+CONFIG_CDROM_PKTCDVD_BUFFERS=8
+# CONFIG_CDROM_PKTCDVD_WCACHE is not set
+CONFIG_ATA_OVER_ETH=m
+
+#
+# ATA/ATAPI/MFM/RLL support
+#
+CONFIG_IDE=y
+CONFIG_BLK_DEV_IDE=y
+
+#
+# Please see Documentation/ide.txt for help/info on IDE drives
+#
+# CONFIG_BLK_DEV_IDE_SATA is not set
+# CONFIG_BLK_DEV_HD_IDE is not set
+CONFIG_BLK_DEV_IDEDISK=m
+CONFIG_IDEDISK_MULTI_MODE=y
+# CONFIG_BLK_DEV_IDECS is not set
+CONFIG_BLK_DEV_IDECD=m
+CONFIG_BLK_DEV_IDETAPE=m
+CONFIG_BLK_DEV_IDEFLOPPY=m
+CONFIG_BLK_DEV_IDESCSI=m
+# CONFIG_IDE_TASK_IOCTL is not set
+
+#
+# IDE chipset support/bugfixes
+#
+CONFIG_IDE_GENERIC=m
+CONFIG_BLK_DEV_CMD640=y
+CONFIG_BLK_DEV_CMD640_ENHANCED=y
+CONFIG_BLK_DEV_IDEPNP=y
+CONFIG_BLK_DEV_IDEPCI=y
+CONFIG_IDEPCI_SHARE_IRQ=y
+CONFIG_BLK_DEV_OFFBOARD=y
+CONFIG_BLK_DEV_GENERIC=y
+# CONFIG_BLK_DEV_OPTI621 is not set
+CONFIG_BLK_DEV_RZ1000=m
+CONFIG_BLK_DEV_IDEDMA_PCI=y
+# CONFIG_BLK_DEV_IDEDMA_FORCED is not set
+CONFIG_IDEDMA_PCI_AUTO=y
+# CONFIG_IDEDMA_ONLYDISK is not set
+CONFIG_BLK_DEV_AEC62XX=m
+CONFIG_BLK_DEV_ALI15X3=m
+# CONFIG_WDC_ALI15X3 is not set
+CONFIG_BLK_DEV_AMD74XX=m
+CONFIG_BLK_DEV_ATIIXP=m
+CONFIG_BLK_DEV_CMD64X=m
+CONFIG_BLK_DEV_TRIFLEX=m
+CONFIG_BLK_DEV_CY82C693=m
+CONFIG_BLK_DEV_CS5520=m
+CONFIG_BLK_DEV_CS5530=m
+CONFIG_BLK_DEV_HPT34X=m
+CONFIG_HPT34X_AUTODMA=y
+CONFIG_BLK_DEV_HPT366=m
+CONFIG_BLK_DEV_SC1200=m
+CONFIG_BLK_DEV_PIIX=m
+CONFIG_BLK_DEV_IT821X=m
+CONFIG_BLK_DEV_NS87415=m
+CONFIG_BLK_DEV_PDC202XX_OLD=m
+CONFIG_PDC202XX_BURST=y
+CONFIG_BLK_DEV_PDC202XX_NEW=m
+CONFIG_BLK_DEV_SVWKS=m
+CONFIG_BLK_DEV_SIIMAGE=m
+CONFIG_BLK_DEV_SIS5513=m
+CONFIG_BLK_DEV_SLC90E66=m
+# CONFIG_BLK_DEV_TRM290 is not set
+CONFIG_BLK_DEV_VIA82CXXX=m
+# CONFIG_IDE_ARM is not set
+CONFIG_BLK_DEV_IDEDMA=y
+# CONFIG_IDEDMA_IVB is not set
+CONFIG_IDEDMA_AUTO=y
+# CONFIG_BLK_DEV_HD is not set
+
+#
+# SCSI device support
+#
+CONFIG_RAID_ATTRS=m
+CONFIG_SCSI=m
+CONFIG_SCSI_PROC_FS=y
+
+#
+# SCSI support type (disk, tape, CD-ROM)
+#
+CONFIG_BLK_DEV_SD=m
+CONFIG_CHR_DEV_ST=m
+CONFIG_CHR_DEV_OSST=m
+CONFIG_BLK_DEV_SR=m
+CONFIG_BLK_DEV_SR_VENDOR=y
+CONFIG_CHR_DEV_SG=m
+CONFIG_CHR_DEV_SCH=m
+
+#
+# Some SCSI devices (e.g. CD jukebox) support multiple LUNs
+#
+CONFIG_SCSI_MULTI_LUN=y
+CONFIG_SCSI_CONSTANTS=y
+CONFIG_SCSI_LOGGING=y
+
+#
+# SCSI Transport Attributes
+#
+CONFIG_SCSI_SPI_ATTRS=m
+CONFIG_SCSI_FC_ATTRS=m
+CONFIG_SCSI_ISCSI_ATTRS=m
+CONFIG_SCSI_SAS_ATTRS=m
+
+#
+# SCSI low-level drivers
+#
+CONFIG_ISCSI_TCP=m
+CONFIG_BLK_DEV_3W_XXXX_RAID=m
+CONFIG_SCSI_3W_9XXX=m
+CONFIG_SCSI_ACARD=m
+CONFIG_SCSI_AACRAID=m
+CONFIG_SCSI_AIC7XXX=m
+CONFIG_AIC7XXX_CMDS_PER_DEVICE=4
+CONFIG_AIC7XXX_RESET_DELAY_MS=15000
+# CONFIG_AIC7XXX_DEBUG_ENABLE is not set
+CONFIG_AIC7XXX_DEBUG_MASK=0
+# CONFIG_AIC7XXX_REG_PRETTY_PRINT is not set
+CONFIG_SCSI_AIC7XXX_OLD=m
+CONFIG_SCSI_AIC79XX=m
+CONFIG_AIC79XX_CMDS_PER_DEVICE=4
+CONFIG_AIC79XX_RESET_DELAY_MS=15000
+# CONFIG_AIC79XX_ENABLE_RD_STRM is not set
+# CONFIG_AIC79XX_DEBUG_ENABLE is not set
+CONFIG_AIC79XX_DEBUG_MASK=0
+# CONFIG_AIC79XX_REG_PRETTY_PRINT is not set
+CONFIG_MEGARAID_NEWGEN=y
+CONFIG_MEGARAID_MM=m
+CONFIG_MEGARAID_MAILBOX=m
+CONFIG_MEGARAID_LEGACY=m
+CONFIG_MEGARAID_SAS=m
+CONFIG_SCSI_SATA=m
+CONFIG_SCSI_SATA_AHCI=m
+CONFIG_SCSI_SATA_SVW=m
+CONFIG_SCSI_ATA_PIIX=m
+CONFIG_SCSI_SATA_MV=m
+CONFIG_SCSI_SATA_NV=m
+CONFIG_SCSI_PDC_ADMA=m
+CONFIG_SCSI_HPTIOP=m
+CONFIG_SCSI_SATA_QSTOR=m
+CONFIG_SCSI_SATA_PROMISE=m
+CONFIG_SCSI_SATA_SX4=m
+CONFIG_SCSI_SATA_SIL=m
+CONFIG_SCSI_SATA_SIL24=m
+CONFIG_SCSI_SATA_SIS=m
+CONFIG_SCSI_SATA_ULI=m
+CONFIG_SCSI_SATA_VIA=m
+CONFIG_SCSI_SATA_VITESSE=m
+CONFIG_SCSI_SATA_INTEL_COMBINED=y
+CONFIG_SCSI_BUSLOGIC=m
+# CONFIG_SCSI_OMIT_FLASHPOINT is not set
+CONFIG_SCSI_DMX3191D=m
+CONFIG_SCSI_EATA=m
+CONFIG_SCSI_EATA_TAGGED_QUEUE=y
+CONFIG_SCSI_EATA_LINKED_COMMANDS=y
+CONFIG_SCSI_EATA_MAX_TAGS=16
+CONFIG_SCSI_FUTURE_DOMAIN=m
+CONFIG_SCSI_GDTH=m
+CONFIG_SCSI_IPS=m
+CONFIG_SCSI_INITIO=m
+CONFIG_SCSI_INIA100=m
+CONFIG_SCSI_PPA=m
+CONFIG_SCSI_IMM=m
+# CONFIG_SCSI_IZIP_EPP16 is not set
+# CONFIG_SCSI_IZIP_SLOW_CTR is not set
+CONFIG_SCSI_SYM53C8XX_2=m
+CONFIG_SCSI_SYM53C8XX_DMA_ADDRESSING_MODE=1
+CONFIG_SCSI_SYM53C8XX_DEFAULT_TAGS=16
+CONFIG_SCSI_SYM53C8XX_MAX_TAGS=64
+CONFIG_SCSI_SYM53C8XX_MMIO=y
+# CONFIG_SCSI_IPR is not set
+CONFIG_SCSI_QLOGIC_1280=m
+CONFIG_SCSI_QLA_FC=m
+CONFIG_SCSI_LPFC=m
+CONFIG_SCSI_DC395x=m
+CONFIG_SCSI_DC390T=m
+# CONFIG_SCSI_DEBUG is not set
+
+#
+# PCMCIA SCSI adapter support
+#
+# CONFIG_PCMCIA_FDOMAIN is not set
+# CONFIG_PCMCIA_QLOGIC is not set
+# CONFIG_PCMCIA_SYM53C500 is not set
+
+#
+# Multi-device support (RAID and LVM)
+#
+CONFIG_MD=y
+CONFIG_BLK_DEV_MD=y
+CONFIG_MD_LINEAR=m
+CONFIG_MD_RAID0=m
+CONFIG_MD_RAID1=m
+CONFIG_MD_RAID10=m
+CONFIG_MD_RAID456=m
+CONFIG_MD_RAID5_RESHAPE=y
+CONFIG_MD_MULTIPATH=m
+CONFIG_MD_FAULTY=m
+CONFIG_BLK_DEV_DM=m
+CONFIG_DM_CRYPT=m
+CONFIG_DM_SNAPSHOT=m
+CONFIG_DM_MIRROR=m
+CONFIG_DM_ZERO=m
+CONFIG_DM_MULTIPATH=m
+CONFIG_DM_MULTIPATH_EMC=m
+
+#
+# Fusion MPT device support
+#
+CONFIG_FUSION=y
+CONFIG_FUSION_SPI=m
+CONFIG_FUSION_FC=m
+CONFIG_FUSION_SAS=m
+CONFIG_FUSION_MAX_SGE=128
+CONFIG_FUSION_CTL=m
+CONFIG_FUSION_LAN=m
+
+#
+# IEEE 1394 (FireWire) support
+#
+CONFIG_IEEE1394=m
+
+#
+# Subsystem Options
+#
+# CONFIG_IEEE1394_VERBOSEDEBUG is not set
+CONFIG_IEEE1394_OUI_DB=y
+CONFIG_IEEE1394_EXTRA_CONFIG_ROMS=y
+CONFIG_IEEE1394_CONFIG_ROM_IP1394=y
+# CONFIG_IEEE1394_EXPORT_FULL_API is not set
+
+#
+# Device Drivers
+#
+CONFIG_IEEE1394_PCILYNX=m
+CONFIG_IEEE1394_OHCI1394=m
+
+#
+# Protocol Drivers
+#
+CONFIG_IEEE1394_VIDEO1394=m
+CONFIG_IEEE1394_SBP2=m
+CONFIG_IEEE1394_ETH1394=m
+CONFIG_IEEE1394_DV1394=m
+CONFIG_IEEE1394_RAWIO=m
+
+#
+# I2O device support
+#
+CONFIG_I2O=m
+CONFIG_I2O_LCT_NOTIFY_ON_CHANGES=y
+CONFIG_I2O_EXT_ADAPTEC=y
+CONFIG_I2O_EXT_ADAPTEC_DMA64=y
+CONFIG_I2O_CONFIG=m
+CONFIG_I2O_CONFIG_OLD_IOCTL=y
+CONFIG_I2O_BUS=m
+CONFIG_I2O_BLOCK=m
+CONFIG_I2O_SCSI=m
+CONFIG_I2O_PROC=m
+
+#
+# Network device support
+#
+CONFIG_NETDEVICES=y
+CONFIG_DUMMY=m
+CONFIG_BONDING=m
+CONFIG_EQUALIZER=m
+CONFIG_TUN=m
+CONFIG_NET_SB1000=m
+
+#
+# ARCnet devices
+#
+CONFIG_ARCNET=m
+CONFIG_ARCNET_1201=m
+CONFIG_ARCNET_1051=m
+CONFIG_ARCNET_RAW=m
+CONFIG_ARCNET_CAP=m
+CONFIG_ARCNET_COM90xx=m
+CONFIG_ARCNET_COM90xxIO=m
+CONFIG_ARCNET_RIM_I=m
+# CONFIG_ARCNET_COM20020 is not set
+
+#
+# PHY device support
+#
+CONFIG_PHYLIB=m
+
+#
+# MII PHY device drivers
+#
+CONFIG_MARVELL_PHY=m
+CONFIG_DAVICOM_PHY=m
+CONFIG_QSEMI_PHY=m
+CONFIG_LXT_PHY=m
+CONFIG_CICADA_PHY=m
+CONFIG_VITESSE_PHY=m
+CONFIG_SMSC_PHY=m
+CONFIG_FIXED_PHY=m
+# CONFIG_FIXED_MII_10_FDX is not set
+# CONFIG_FIXED_MII_100_FDX is not set
+
+#
+# Ethernet (10 or 100Mbit)
+#
+CONFIG_NET_ETHERNET=y
+CONFIG_MII=m
+CONFIG_HAPPYMEAL=m
+CONFIG_SUNGEM=m
+CONFIG_CASSINI=m
+CONFIG_NET_VENDOR_3COM=y
+CONFIG_VORTEX=m
+CONFIG_TYPHOON=m
+
+#
+# Tulip family network device support
+#
+CONFIG_NET_TULIP=y
+CONFIG_DE2104X=m
+CONFIG_TULIP=m
+# CONFIG_TULIP_MWI is not set
+# CONFIG_TULIP_MMIO is not set
+CONFIG_TULIP_NAPI=y
+CONFIG_TULIP_NAPI_HW_MITIGATION=y
+CONFIG_DE4X5=m
+CONFIG_WINBOND_840=m
+CONFIG_DM9102=m
+CONFIG_ULI526X=m
+CONFIG_PCMCIA_XIRCOM=m
+CONFIG_HP100=m
+CONFIG_NET_PCI=y
+CONFIG_PCNET32=m
+CONFIG_AMD8111_ETH=m
+CONFIG_AMD8111E_NAPI=y
+CONFIG_ADAPTEC_STARFIRE=m
+CONFIG_ADAPTEC_STARFIRE_NAPI=y
+CONFIG_B44=m
+CONFIG_FORCEDETH=m
+CONFIG_DGRS=m
+CONFIG_EEPRO100=m
+CONFIG_E100=m
+CONFIG_FEALNX=m
+CONFIG_NATSEMI=m
+CONFIG_NE2K_PCI=m
+CONFIG_8139CP=m
+CONFIG_8139TOO=m
+# CONFIG_8139TOO_PIO is not set
+# CONFIG_8139TOO_TUNE_TWISTER is not set
+CONFIG_8139TOO_8129=y
+# CONFIG_8139_OLD_RX_RESET is not set
+CONFIG_SIS900=m
+CONFIG_EPIC100=m
+CONFIG_SUNDANCE=m
+# CONFIG_SUNDANCE_MMIO is not set
+CONFIG_VIA_RHINE=m
+# CONFIG_VIA_RHINE_MMIO is not set
+# CONFIG_VIA_RHINE_NAPI is not set
+# CONFIG_NET_POCKET is not set
+
+#
+# Ethernet (1000 Mbit)
+#
+CONFIG_ACENIC=m
+# CONFIG_ACENIC_OMIT_TIGON_I is not set
+CONFIG_DL2K=m
+CONFIG_E1000=m
+CONFIG_E1000_NAPI=y
+# CONFIG_E1000_DISABLE_PACKET_SPLIT is not set
+CONFIG_NS83820=m
+CONFIG_HAMACHI=m
+CONFIG_YELLOWFIN=m
+CONFIG_R8169=m
+CONFIG_R8169_NAPI=y
+CONFIG_R8169_VLAN=y
+CONFIG_SIS190=m
+CONFIG_SKGE=m
+CONFIG_SKY2=m
+CONFIG_SK98LIN=m
+CONFIG_VIA_VELOCITY=m
+CONFIG_TIGON3=m
+CONFIG_BNX2=m
+
+#
+# Ethernet (10000 Mbit)
+#
+CONFIG_CHELSIO_T1=m
+CONFIG_IXGB=m
+CONFIG_IXGB_NAPI=y
+CONFIG_S2IO=m
+CONFIG_S2IO_NAPI=y
+CONFIG_MYRI10GE=m
+# CONFIG_SFC is not set
+
+#
+# Token Ring devices
+#
+CONFIG_TR=y
+CONFIG_IBMOL=m
+CONFIG_3C359=m
+CONFIG_TMS380TR=m
+CONFIG_TMSPCI=m
+CONFIG_ABYSS=m
+
+#
+# Wireless LAN (non-hamradio)
+#
+CONFIG_NET_RADIO=y
+CONFIG_NET_WIRELESS_RTNETLINK=y
+
+#
+# Obsolete Wireless cards support (pre-802.11)
+#
+CONFIG_STRIP=m
+CONFIG_PCMCIA_WAVELAN=m
+CONFIG_PCMCIA_NETWAVE=m
+
+#
+# Wireless 802.11 Frequency Hopping cards support
+#
+CONFIG_PCMCIA_RAYCS=m
+
+#
+# Wireless 802.11b ISA/PCI cards support
+#
+CONFIG_IPW2100=m
+CONFIG_IPW2100_MONITOR=y
+# CONFIG_IPW2100_DEBUG is not set
+CONFIG_IPW2200=m
+CONFIG_IPW2200_MONITOR=y
+# CONFIG_IPW2200_RADIOTAP is not set
+# CONFIG_IPW2200_PROMISCUOUS is not set
+CONFIG_IPW2200_QOS=y
+# CONFIG_IPW2200_DEBUG is not set
+CONFIG_AIRO=m
+CONFIG_HERMES=m
+CONFIG_PLX_HERMES=m
+CONFIG_TMD_HERMES=m
+CONFIG_NORTEL_HERMES=m
+CONFIG_PCI_HERMES=m
+CONFIG_ATMEL=m
+CONFIG_PCI_ATMEL=m
+
+#
+# Wireless 802.11b Pcmcia/Cardbus cards support
+#
+# CONFIG_PCMCIA_HERMES is not set
+# CONFIG_PCMCIA_SPECTRUM is not set
+# CONFIG_AIRO_CS is not set
+# CONFIG_PCMCIA_ATMEL is not set
+# CONFIG_PCMCIA_WL3501 is not set
+
+#
+# Prism GT/Duette 802.11(a/b/g) PCI/Cardbus support
+#
+CONFIG_PRISM54=m
+CONFIG_USB_ZD1201=m
+CONFIG_HOSTAP=m
+CONFIG_HOSTAP_FIRMWARE=y
+CONFIG_HOSTAP_FIRMWARE_NVRAM=y
+CONFIG_HOSTAP_PLX=m
+CONFIG_HOSTAP_PCI=m
+# CONFIG_HOSTAP_CS is not set
+CONFIG_BCM43XX=m
+CONFIG_BCM43XX_DEBUG=y
+CONFIG_BCM43XX_DMA=y
+CONFIG_BCM43XX_PIO=y
+CONFIG_BCM43XX_DMA_AND_PIO_MODE=y
+# CONFIG_BCM43XX_DMA_MODE is not set
+# CONFIG_BCM43XX_PIO_MODE is not set
+CONFIG_ZD1211RW=m
+# CONFIG_ZD1211RW_DEBUG is not set
+CONFIG_NET_WIRELESS=y
+
+#
+# PCMCIA network device support
+#
+# CONFIG_NET_PCMCIA is not set
+
+#
+# Wan interfaces
+#
+# CONFIG_WAN is not set
+
+#
+# ATM drivers
+#
+CONFIG_ATM_DUMMY=m
+CONFIG_ATM_TCP=m
+CONFIG_ATM_LANAI=m
+CONFIG_ATM_ENI=m
+# CONFIG_ATM_ENI_DEBUG is not set
+# CONFIG_ATM_ENI_TUNE_BURST is not set
+CONFIG_ATM_FIRESTREAM=m
+# CONFIG_ATM_ZATM is not set
+CONFIG_ATM_IDT77252=m
+# CONFIG_ATM_IDT77252_DEBUG is not set
+# CONFIG_ATM_IDT77252_RCV_ALL is not set
+CONFIG_ATM_IDT77252_USE_SUNI=y
+CONFIG_ATM_AMBASSADOR=m
+# CONFIG_ATM_AMBASSADOR_DEBUG is not set
+CONFIG_ATM_HORIZON=m
+# CONFIG_ATM_HORIZON_DEBUG is not set
+CONFIG_ATM_FORE200E_MAYBE=m
+# CONFIG_ATM_FORE200E_PCA is not set
+CONFIG_ATM_HE=m
+# CONFIG_ATM_HE_USE_SUNI is not set
+CONFIG_FDDI=y
+# CONFIG_DEFXX is not set
+CONFIG_SKFP=m
+CONFIG_HIPPI=y
+CONFIG_ROADRUNNER=m
+CONFIG_ROADRUNNER_LARGE_RINGS=y
+CONFIG_PLIP=m
+CONFIG_PPP=m
+CONFIG_PPP_MULTILINK=y
+CONFIG_PPP_FILTER=y
+CONFIG_PPP_ASYNC=m
+CONFIG_PPP_SYNC_TTY=m
+CONFIG_PPP_DEFLATE=m
+CONFIG_PPP_BSDCOMP=m
+CONFIG_PPP_MPPE=m
+CONFIG_PPPOE=m
+CONFIG_PPPOATM=m
+CONFIG_SLIP=m
+CONFIG_SLIP_COMPRESSED=y
+CONFIG_SLIP_SMART=y
+CONFIG_SLIP_MODE_SLIP6=y
+CONFIG_NET_FC=y
+CONFIG_SHAPER=m
+CONFIG_NETCONSOLE=m
+CONFIG_NETPOLL=y
+CONFIG_NETPOLL_RX=y
+CONFIG_NETPOLL_TRAP=y
+CONFIG_NET_POLL_CONTROLLER=y
+
+#
+# ISDN subsystem
+#
+CONFIG_ISDN=m
+
+#
+# Old ISDN4Linux
+#
+CONFIG_ISDN_I4L=m
+CONFIG_ISDN_PPP=y
+CONFIG_ISDN_PPP_VJ=y
+CONFIG_ISDN_MPP=y
+CONFIG_IPPP_FILTER=y
+CONFIG_ISDN_PPP_BSDCOMP=m
+CONFIG_ISDN_AUDIO=y
+CONFIG_ISDN_TTY_FAX=y
+CONFIG_ISDN_X25=y
+
+#
+# ISDN feature submodules
+#
+CONFIG_ISDN_DIVERSION=m
+
+#
+# ISDN4Linux hardware drivers
+#
+
+#
+# Passive cards
+#
+CONFIG_ISDN_DRV_HISAX=m
+
+#
+# D-channel protocol features
+#
+CONFIG_HISAX_EURO=y
+CONFIG_DE_AOC=y
+CONFIG_HISAX_NO_SENDCOMPLETE=y
+CONFIG_HISAX_NO_LLC=y
+CONFIG_HISAX_NO_KEYPAD=y
+CONFIG_HISAX_1TR6=y
+CONFIG_HISAX_NI1=y
+CONFIG_HISAX_MAX_CARDS=8
+
+#
+# HiSax supported cards
+#
+CONFIG_HISAX_16_3=y
+CONFIG_HISAX_TELESPCI=y
+CONFIG_HISAX_S0BOX=y
+CONFIG_HISAX_FRITZPCI=y
+CONFIG_HISAX_AVM_A1_PCMCIA=y
+CONFIG_HISAX_ELSA=y
+CONFIG_HISAX_DIEHLDIVA=y
+CONFIG_HISAX_SEDLBAUER=y
+CONFIG_HISAX_NETJET=y
+CONFIG_HISAX_NETJET_U=y
+CONFIG_HISAX_NICCY=y
+CONFIG_HISAX_BKM_A4T=y
+CONFIG_HISAX_SCT_QUADRO=y
+CONFIG_HISAX_GAZEL=y
+CONFIG_HISAX_HFC_PCI=y
+CONFIG_HISAX_W6692=y
+CONFIG_HISAX_HFC_SX=y
+CONFIG_HISAX_ENTERNOW_PCI=y
+# CONFIG_HISAX_DEBUG is not set
+
+#
+# HiSax PCMCIA card service modules
+#
+# CONFIG_HISAX_SEDLBAUER_CS is not set
+# CONFIG_HISAX_ELSA_CS is not set
+# CONFIG_HISAX_AVM_A1_CS is not set
+# CONFIG_HISAX_TELES_CS is not set
+
+#
+# HiSax sub driver modules
+#
+CONFIG_HISAX_ST5481=m
+CONFIG_HISAX_HFCUSB=m
+CONFIG_HISAX_HFC4S8S=m
+CONFIG_HISAX_FRITZ_PCIPNP=m
+CONFIG_HISAX_HDLC=y
+
+#
+# Active cards
+#
+
+#
+# Siemens Gigaset
+#
+CONFIG_ISDN_DRV_GIGASET=m
+CONFIG_GIGASET_BASE=m
+CONFIG_GIGASET_M105=m
+# CONFIG_GIGASET_DEBUG is not set
+# CONFIG_GIGASET_UNDOCREQ is not set
+
+#
+# CAPI subsystem
+#
+CONFIG_ISDN_CAPI=m
+CONFIG_ISDN_DRV_AVMB1_VERBOSE_REASON=y
+CONFIG_ISDN_CAPI_MIDDLEWARE=y
+CONFIG_ISDN_CAPI_CAPI20=m
+CONFIG_ISDN_CAPI_CAPIFS_BOOL=y
+CONFIG_ISDN_CAPI_CAPIFS=m
+CONFIG_ISDN_CAPI_CAPIDRV=m
+
+#
+# CAPI hardware drivers
+#
+
+#
+# Active AVM cards
+#
+CONFIG_CAPI_AVM=y
+CONFIG_ISDN_DRV_AVMB1_B1PCI=m
+CONFIG_ISDN_DRV_AVMB1_B1PCIV4=y
+CONFIG_ISDN_DRV_AVMB1_B1PCMCIA=m
+# CONFIG_ISDN_DRV_AVMB1_AVM_CS is not set
+CONFIG_ISDN_DRV_AVMB1_T1PCI=m
+CONFIG_ISDN_DRV_AVMB1_C4=m
+
+#
+# Active Eicon DIVA Server cards
+#
+# CONFIG_CAPI_EICON is not set
+
+#
+# Telephony Support
+#
+CONFIG_PHONE=m
+CONFIG_PHONE_IXJ=m
+CONFIG_PHONE_IXJ_PCMCIA=m
+
+#
+# Input device support
+#
+CONFIG_INPUT=y
+
+#
+# Userland interfaces
+#
+CONFIG_INPUT_MOUSEDEV=y
+CONFIG_INPUT_MOUSEDEV_PSAUX=y
+CONFIG_INPUT_MOUSEDEV_SCREEN_X=1024
+CONFIG_INPUT_MOUSEDEV_SCREEN_Y=768
+CONFIG_INPUT_JOYDEV=m
+CONFIG_INPUT_TSDEV=m
+CONFIG_INPUT_TSDEV_SCREEN_X=240
+CONFIG_INPUT_TSDEV_SCREEN_Y=320
+CONFIG_INPUT_EVDEV=y
+# CONFIG_INPUT_EVBUG is not set
+
+#
+# Input Device Drivers
+#
+CONFIG_INPUT_KEYBOARD=y
+CONFIG_KEYBOARD_ATKBD=y
+CONFIG_KEYBOARD_SUNKBD=m
+# CONFIG_KEYBOARD_LKKBD is not set
+CONFIG_KEYBOARD_XTKBD=m
+CONFIG_KEYBOARD_NEWTON=m
+CONFIG_INPUT_MOUSE=y
+CONFIG_MOUSE_PS2=y
+CONFIG_MOUSE_SERIAL=m
+# CONFIG_MOUSE_VSXXXAA is not set
+CONFIG_INPUT_JOYSTICK=y
+CONFIG_JOYSTICK_ANALOG=m
+CONFIG_JOYSTICK_A3D=m
+CONFIG_JOYSTICK_ADI=m
+CONFIG_JOYSTICK_COBRA=m
+CONFIG_JOYSTICK_GF2K=m
+CONFIG_JOYSTICK_GRIP=m
+CONFIG_JOYSTICK_GRIP_MP=m
+CONFIG_JOYSTICK_GUILLEMOT=m
+CONFIG_JOYSTICK_INTERACT=m
+CONFIG_JOYSTICK_SIDEWINDER=m
+CONFIG_JOYSTICK_TMDC=m
+CONFIG_JOYSTICK_IFORCE=m
+CONFIG_JOYSTICK_IFORCE_USB=y
+CONFIG_JOYSTICK_IFORCE_232=y
+CONFIG_JOYSTICK_WARRIOR=m
+CONFIG_JOYSTICK_MAGELLAN=m
+CONFIG_JOYSTICK_SPACEORB=m
+CONFIG_JOYSTICK_SPACEBALL=m
+CONFIG_JOYSTICK_STINGER=m
+CONFIG_JOYSTICK_TWIDJOY=m
+CONFIG_JOYSTICK_DB9=m
+CONFIG_JOYSTICK_GAMECON=m
+CONFIG_JOYSTICK_TURBOGRAFX=m
+CONFIG_JOYSTICK_JOYDUMP=m
+CONFIG_INPUT_TOUCHSCREEN=y
+CONFIG_TOUCHSCREEN_ADS7846=m
+CONFIG_TOUCHSCREEN_GUNZE=m
+CONFIG_TOUCHSCREEN_ELO=m
+CONFIG_TOUCHSCREEN_MTOUCH=m
+CONFIG_TOUCHSCREEN_MK712=m
+CONFIG_INPUT_MISC=y
+CONFIG_INPUT_PCSPKR=m
+CONFIG_INPUT_UINPUT=m
+
+#
+# Hardware I/O ports
+#
+CONFIG_SERIO=y
+CONFIG_SERIO_I8042=y
+CONFIG_SERIO_SERPORT=m
+CONFIG_SERIO_CT82C710=m
+CONFIG_SERIO_PARKBD=m
+CONFIG_SERIO_PCIPS2=m
+CONFIG_SERIO_LIBPS2=y
+CONFIG_SERIO_RAW=m
+CONFIG_GAMEPORT=m
+CONFIG_GAMEPORT_NS558=m
+CONFIG_GAMEPORT_L4=m
+CONFIG_GAMEPORT_EMU10K1=m
+CONFIG_GAMEPORT_FM801=m
+
+#
+# Character devices
+#
+CONFIG_VT=y
+CONFIG_VT_CONSOLE=y
+CONFIG_HW_CONSOLE=y
+CONFIG_VT_HW_CONSOLE_BINDING=y
+# CONFIG_SERIAL_NONSTANDARD is not set
+
+#
+# Serial drivers
+#
+
+#
+# Non-8250 serial port support
+#
+# CONFIG_SERIAL_JSM is not set
+CONFIG_UNIX98_PTYS=y
+CONFIG_LEGACY_PTYS=y
+CONFIG_LEGACY_PTY_COUNT=64
+CONFIG_PRINTER=m
+CONFIG_LP_CONSOLE=y
+CONFIG_PPDEV=m
+CONFIG_TIPAR=m
+
+#
+# IPMI
+#
+CONFIG_IPMI_HANDLER=m
+CONFIG_IPMI_PANIC_EVENT=y
+# CONFIG_IPMI_PANIC_STRING is not set
+CONFIG_IPMI_DEVICE_INTERFACE=m
+CONFIG_IPMI_SI=m
+CONFIG_IPMI_WATCHDOG=m
+CONFIG_IPMI_POWEROFF=m
+
+#
+# Watchdog Cards
+#
+CONFIG_WATCHDOG=y
+# CONFIG_WATCHDOG_NOWAYOUT is not set
+
+#
+# Watchdog Device Drivers
+#
+CONFIG_SOFT_WATCHDOG=m
+CONFIG_ACQUIRE_WDT=m
+CONFIG_ADVANTECH_WDT=m
+CONFIG_ALIM1535_WDT=m
+CONFIG_ALIM7101_WDT=m
+CONFIG_SC520_WDT=m
+CONFIG_EUROTECH_WDT=m
+CONFIG_IB700_WDT=m
+CONFIG_IBMASR=m
+CONFIG_WAFER_WDT=m
+CONFIG_I6300ESB_WDT=m
+CONFIG_I8XX_TCO=m
+CONFIG_SC1200_WDT=m
+CONFIG_60XX_WDT=m
+CONFIG_SBC8360_WDT=m
+CONFIG_CPU5_WDT=m
+CONFIG_W83627HF_WDT=m
+CONFIG_W83877F_WDT=m
+CONFIG_W83977F_WDT=m
+CONFIG_MACHZ_WDT=m
+CONFIG_SBC_EPX_C3_WATCHDOG=m
+
+#
+# PCI-based Watchdog Cards
+#
+CONFIG_PCIPCWATCHDOG=m
+CONFIG_WDTPCI=m
+CONFIG_WDT_501_PCI=y
+
+#
+# USB-based Watchdog Cards
+#
+CONFIG_USBPCWATCHDOG=m
+CONFIG_HW_RANDOM=y
+CONFIG_HW_RANDOM_INTEL=m
+CONFIG_HW_RANDOM_AMD=m
+CONFIG_HW_RANDOM_GEODE=m
+CONFIG_NVRAM=y
+CONFIG_RTC=y
+CONFIG_DTLK=m
+CONFIG_R3964=m
+CONFIG_APPLICOM=m
+
+#
+# Ftape, the floppy tape device driver
+#
+CONFIG_AGP=m
+CONFIG_AGP_AMD64=m
+CONFIG_AGP_INTEL=m
+CONFIG_AGP_SIS=m
+CONFIG_AGP_VIA=m
+CONFIG_DRM=m
+CONFIG_DRM_TDFX=m
+CONFIG_DRM_R128=m
+CONFIG_DRM_RADEON=m
+CONFIG_DRM_I810=m
+# CONFIG_DRM_I830 is not set
+CONFIG_DRM_I915=m
+CONFIG_DRM_MGA=m
+# CONFIG_DRM_SIS is not set
+CONFIG_DRM_VIA=m
+CONFIG_DRM_SAVAGE=m
+
+#
+# PCMCIA character devices
+#
+# CONFIG_SYNCLINK_CS is not set
+# CONFIG_CARDMAN_4000 is not set
+# CONFIG_CARDMAN_4040 is not set
+# CONFIG_MWAVE is not set
+CONFIG_PC8736x_GPIO=m
+CONFIG_NSC_GPIO=m
+# CONFIG_RAW_DRIVER is not set
+# CONFIG_HPET is not set
+CONFIG_HANGCHECK_TIMER=m
+
+#
+# TPM devices
+#
+CONFIG_TCG_TPM=m
+CONFIG_TCG_TIS=m
+CONFIG_TCG_NSC=m
+CONFIG_TCG_ATMEL=m
+CONFIG_TCG_INFINEON=m
+CONFIG_TCG_XEN=m
+CONFIG_TELCLOCK=m
+
+#
+# I2C support
+#
+CONFIG_I2C=m
+CONFIG_I2C_CHARDEV=m
+
+#
+# I2C Algorithms
+#
+CONFIG_I2C_ALGOBIT=m
+CONFIG_I2C_ALGOPCF=m
+CONFIG_I2C_ALGOPCA=m
+
+#
+# I2C Hardware Bus support
+#
+CONFIG_I2C_ALI1535=m
+CONFIG_I2C_ALI1563=m
+CONFIG_I2C_ALI15X3=m
+CONFIG_I2C_AMD756=m
+CONFIG_I2C_AMD756_S4882=m
+CONFIG_I2C_AMD8111=m
+CONFIG_I2C_I801=m
+CONFIG_I2C_I810=m
+CONFIG_I2C_PIIX4=m
+CONFIG_I2C_ISA=m
+CONFIG_I2C_NFORCE2=m
+CONFIG_I2C_OCORES=m
+CONFIG_I2C_PARPORT=m
+CONFIG_I2C_PARPORT_LIGHT=m
+CONFIG_I2C_PROSAVAGE=m
+CONFIG_I2C_SAVAGE4=m
+CONFIG_I2C_SIS5595=m
+CONFIG_I2C_SIS630=m
+CONFIG_I2C_SIS96X=m
+CONFIG_I2C_STUB=m
+CONFIG_I2C_VIA=m
+CONFIG_I2C_VIAPRO=m
+CONFIG_I2C_VOODOO3=m
+CONFIG_I2C_PCA_ISA=m
+
+#
+# Miscellaneous I2C Chip support
+#
+CONFIG_SENSORS_DS1337=m
+CONFIG_SENSORS_DS1374=m
+CONFIG_SENSORS_EEPROM=m
+CONFIG_SENSORS_PCF8574=m
+CONFIG_SENSORS_PCA9539=m
+CONFIG_SENSORS_PCF8591=m
+CONFIG_SENSORS_MAX6875=m
+# CONFIG_I2C_DEBUG_CORE is not set
+# CONFIG_I2C_DEBUG_ALGO is not set
+# CONFIG_I2C_DEBUG_BUS is not set
+# CONFIG_I2C_DEBUG_CHIP is not set
+
+#
+# SPI support
+#
+CONFIG_SPI=y
+# CONFIG_SPI_DEBUG is not set
+CONFIG_SPI_MASTER=y
+
+#
+# SPI Master Controller Drivers
+#
+CONFIG_SPI_BITBANG=m
+CONFIG_SPI_BUTTERFLY=m
+
+#
+# SPI Protocol Masters
+#
+
+#
+# Dallas's 1-wire bus
+#
+CONFIG_W1=m
+CONFIG_W1_CON=y
+
+#
+# 1-wire Bus Masters
+#
+CONFIG_W1_MASTER_MATROX=m
+CONFIG_W1_MASTER_DS2490=m
+CONFIG_W1_MASTER_DS2482=m
+
+#
+# 1-wire Slaves
+#
+CONFIG_W1_SLAVE_THERM=m
+CONFIG_W1_SLAVE_SMEM=m
+CONFIG_W1_SLAVE_DS2433=m
+# CONFIG_W1_SLAVE_DS2433_CRC is not set
+
+#
+# Hardware Monitoring support
+#
+CONFIG_HWMON=m
+CONFIG_HWMON_VID=m
+CONFIG_SENSORS_ABITUGURU=m
+CONFIG_SENSORS_ADM1021=m
+CONFIG_SENSORS_ADM1025=m
+CONFIG_SENSORS_ADM1026=m
+CONFIG_SENSORS_ADM1031=m
+CONFIG_SENSORS_ADM9240=m
+CONFIG_SENSORS_ASB100=m
+CONFIG_SENSORS_ATXP1=m
+CONFIG_SENSORS_DS1621=m
+CONFIG_SENSORS_F71805F=m
+CONFIG_SENSORS_FSCHER=m
+CONFIG_SENSORS_FSCPOS=m
+CONFIG_SENSORS_GL518SM=m
+CONFIG_SENSORS_GL520SM=m
+CONFIG_SENSORS_IT87=m
+CONFIG_SENSORS_LM63=m
+CONFIG_SENSORS_LM70=m
+CONFIG_SENSORS_LM75=m
+CONFIG_SENSORS_LM77=m
+CONFIG_SENSORS_LM78=m
+CONFIG_SENSORS_LM80=m
+CONFIG_SENSORS_LM83=m
+CONFIG_SENSORS_LM85=m
+CONFIG_SENSORS_LM87=m
+CONFIG_SENSORS_LM90=m
+CONFIG_SENSORS_LM92=m
+CONFIG_SENSORS_MAX1619=m
+CONFIG_SENSORS_PC87360=m
+CONFIG_SENSORS_SIS5595=m
+CONFIG_SENSORS_SMSC47M1=m
+CONFIG_SENSORS_SMSC47M192=m
+CONFIG_SENSORS_SMSC47B397=m
+CONFIG_SENSORS_VIA686A=m
+CONFIG_SENSORS_VT8231=m
+CONFIG_SENSORS_W83781D=m
+CONFIG_SENSORS_W83791D=m
+CONFIG_SENSORS_W83792D=m
+CONFIG_SENSORS_W83L785TS=m
+CONFIG_SENSORS_W83627HF=m
+CONFIG_SENSORS_W83627EHF=m
+CONFIG_SENSORS_HDAPS=m
+# CONFIG_HWMON_DEBUG_CHIP is not set
+
+#
+# Misc devices
+#
+CONFIG_IBM_ASM=m
+
+#
+# Multimedia devices
+#
+CONFIG_VIDEO_DEV=m
+CONFIG_VIDEO_V4L1=y
+CONFIG_VIDEO_V4L1_COMPAT=y
+CONFIG_VIDEO_V4L2=y
+
+#
+# Video Capture Adapters
+#
+
+#
+# Video Capture Adapters
+#
+# CONFIG_VIDEO_ADV_DEBUG is not set
+CONFIG_VIDEO_VIVI=m
+CONFIG_VIDEO_BT848=m
+CONFIG_VIDEO_BT848_DVB=y
+CONFIG_VIDEO_SAA6588=m
+CONFIG_VIDEO_BWQCAM=m
+CONFIG_VIDEO_CQCAM=m
+CONFIG_VIDEO_W9966=m
+CONFIG_VIDEO_CPIA=m
+CONFIG_VIDEO_CPIA_PP=m
+CONFIG_VIDEO_CPIA_USB=m
+CONFIG_VIDEO_CPIA2=m
+CONFIG_VIDEO_SAA5246A=m
+CONFIG_VIDEO_SAA5249=m
+CONFIG_TUNER_3036=m
+CONFIG_VIDEO_STRADIS=m
+CONFIG_VIDEO_ZORAN=m
+CONFIG_VIDEO_ZORAN_BUZ=m
+CONFIG_VIDEO_ZORAN_DC10=m
+CONFIG_VIDEO_ZORAN_DC30=m
+CONFIG_VIDEO_ZORAN_LML33=m
+CONFIG_VIDEO_ZORAN_LML33R10=m
+CONFIG_VIDEO_ZORAN_AVS6EYES=m
+CONFIG_VIDEO_SAA7134=m
+CONFIG_VIDEO_SAA7134_ALSA=m
+CONFIG_VIDEO_SAA7134_DVB=m
+CONFIG_VIDEO_SAA7134_DVB_ALL_FRONTENDS=y
+CONFIG_VIDEO_MXB=m
+CONFIG_VIDEO_DPC=m
+CONFIG_VIDEO_HEXIUM_ORION=m
+CONFIG_VIDEO_HEXIUM_GEMINI=m
+CONFIG_VIDEO_CX88_VP3054=m
+CONFIG_VIDEO_CX88=m
+CONFIG_VIDEO_CX88_ALSA=m
+CONFIG_VIDEO_CX88_BLACKBIRD=m
+CONFIG_VIDEO_CX88_DVB=m
+CONFIG_VIDEO_CX88_DVB_ALL_FRONTENDS=y
+
+#
+# Encoders and Decoders
+#
+CONFIG_VIDEO_MSP3400=m
+CONFIG_VIDEO_CS53L32A=m
+CONFIG_VIDEO_TLV320AIC23B=m
+CONFIG_VIDEO_WM8775=m
+CONFIG_VIDEO_WM8739=m
+CONFIG_VIDEO_CX2341X=m
+CONFIG_VIDEO_CX25840=m
+CONFIG_VIDEO_SAA711X=m
+CONFIG_VIDEO_SAA7127=m
+CONFIG_VIDEO_UPD64031A=m
+CONFIG_VIDEO_UPD64083=m
+
+#
+# V4L USB devices
+#
+CONFIG_VIDEO_PVRUSB2=m
+CONFIG_VIDEO_PVRUSB2_24XXX=y
+CONFIG_VIDEO_PVRUSB2_SYSFS=y
+# CONFIG_VIDEO_PVRUSB2_DEBUGIFC is not set
+CONFIG_VIDEO_EM28XX=m
+CONFIG_VIDEO_USBVIDEO=m
+CONFIG_USB_VICAM=m
+CONFIG_USB_IBMCAM=m
+CONFIG_USB_KONICAWC=m
+CONFIG_USB_QUICKCAM_MESSENGER=m
+CONFIG_USB_ET61X251=m
+CONFIG_VIDEO_OVCAMCHIP=m
+CONFIG_USB_W9968CF=m
+CONFIG_USB_OV511=m
+CONFIG_USB_SE401=m
+CONFIG_USB_SN9C102=m
+CONFIG_USB_STV680=m
+CONFIG_USB_ZC0301=m
+CONFIG_USB_PWC=m
+# CONFIG_USB_PWC_DEBUG is not set
+
+#
+# Radio Adapters
+#
+CONFIG_RADIO_GEMTEK_PCI=m
+CONFIG_RADIO_MAXIRADIO=m
+CONFIG_RADIO_MAESTRO=m
+CONFIG_USB_DSBR=m
+
+#
+# Digital Video Broadcasting Devices
+#
+CONFIG_DVB=y
+CONFIG_DVB_CORE=m
+
+#
+# Supported SAA7146 based PCI Adapters
+#
+CONFIG_DVB_AV7110=m
+CONFIG_DVB_AV7110_OSD=y
+CONFIG_DVB_BUDGET=m
+CONFIG_DVB_BUDGET_CI=m
+CONFIG_DVB_BUDGET_AV=m
+CONFIG_DVB_BUDGET_PATCH=m
+
+#
+# Supported USB Adapters
+#
+CONFIG_DVB_USB=m
+# CONFIG_DVB_USB_DEBUG is not set
+CONFIG_DVB_USB_A800=m
+CONFIG_DVB_USB_DIBUSB_MB=m
+# CONFIG_DVB_USB_DIBUSB_MB_FAULTY is not set
+CONFIG_DVB_USB_DIBUSB_MC=m
+CONFIG_DVB_USB_UMT_010=m
+CONFIG_DVB_USB_CXUSB=m
+CONFIG_DVB_USB_DIGITV=m
+CONFIG_DVB_USB_VP7045=m
+CONFIG_DVB_USB_VP702X=m
+CONFIG_DVB_USB_GP8PSK=m
+CONFIG_DVB_USB_NOVA_T_USB2=m
+CONFIG_DVB_USB_DTT200U=m
+CONFIG_DVB_TTUSB_BUDGET=m
+CONFIG_DVB_TTUSB_DEC=m
+CONFIG_DVB_CINERGYT2=m
+CONFIG_DVB_CINERGYT2_TUNING=y
+CONFIG_DVB_CINERGYT2_STREAM_URB_COUNT=32
+CONFIG_DVB_CINERGYT2_STREAM_BUF_SIZE=512
+CONFIG_DVB_CINERGYT2_QUERY_INTERVAL=250
+CONFIG_DVB_CINERGYT2_ENABLE_RC_INPUT_DEVICE=y
+CONFIG_DVB_CINERGYT2_RC_QUERY_INTERVAL=100
+
+#
+# Supported FlexCopII (B2C2) Adapters
+#
+CONFIG_DVB_B2C2_FLEXCOP=m
+CONFIG_DVB_B2C2_FLEXCOP_PCI=m
+CONFIG_DVB_B2C2_FLEXCOP_USB=m
+# CONFIG_DVB_B2C2_FLEXCOP_DEBUG is not set
+
+#
+# Supported BT878 Adapters
+#
+CONFIG_DVB_BT8XX=m
+
+#
+# Supported Pluto2 Adapters
+#
+CONFIG_DVB_PLUTO2=m
+
+#
+# Supported DVB Frontends
+#
+
+#
+# Customise DVB Frontends
+#
+
+#
+# DVB-S (satellite) frontends
+#
+CONFIG_DVB_STV0299=m
+CONFIG_DVB_CX24110=m
+CONFIG_DVB_CX24123=m
+CONFIG_DVB_TDA8083=m
+CONFIG_DVB_MT312=m
+CONFIG_DVB_VES1X93=m
+CONFIG_DVB_S5H1420=m
+
+#
+# DVB-T (terrestrial) frontends
+#
+CONFIG_DVB_SP8870=m
+CONFIG_DVB_SP887X=m
+CONFIG_DVB_CX22700=m
+CONFIG_DVB_CX22702=m
+CONFIG_DVB_L64781=m
+CONFIG_DVB_TDA1004X=m
+CONFIG_DVB_NXT6000=m
+CONFIG_DVB_MT352=m
+CONFIG_DVB_ZL10353=m
+CONFIG_DVB_DIB3000MB=m
+CONFIG_DVB_DIB3000MC=m
+
+#
+# DVB-C (cable) frontends
+#
+CONFIG_DVB_VES1820=m
+CONFIG_DVB_TDA10021=m
+CONFIG_DVB_STV0297=m
+
+#
+# ATSC (North American/Korean Terrestrial/Cable DTV) frontends
+#
+CONFIG_DVB_NXT200X=m
+CONFIG_DVB_OR51211=m
+CONFIG_DVB_OR51132=m
+CONFIG_DVB_BCM3510=m
+CONFIG_DVB_LGDT330X=m
+
+#
+# Miscellaneous devices
+#
+CONFIG_DVB_PLL=m
+CONFIG_DVB_LNBP21=m
+CONFIG_DVB_ISL6421=m
+CONFIG_VIDEO_SAA7146=m
+CONFIG_VIDEO_SAA7146_VV=m
+CONFIG_VIDEO_VIDEOBUF=m
+CONFIG_VIDEO_TUNER=m
+CONFIG_VIDEO_BUF=m
+CONFIG_VIDEO_BUF_DVB=m
+CONFIG_VIDEO_BTCX=m
+CONFIG_VIDEO_IR=m
+CONFIG_VIDEO_TVEEPROM=m
+CONFIG_USB_DABUSB=m
+
+#
+# Graphics support
+#
+CONFIG_FIRMWARE_EDID=y
+CONFIG_FB=y
+CONFIG_FB_CFB_FILLRECT=y
+CONFIG_FB_CFB_COPYAREA=y
+CONFIG_FB_CFB_IMAGEBLIT=y
+# CONFIG_FB_MACMODES is not set
+# CONFIG_FB_BACKLIGHT is not set
+CONFIG_FB_MODE_HELPERS=y
+CONFIG_FB_TILEBLITTING=y
+CONFIG_FB_CIRRUS=m
+# CONFIG_FB_PM2 is not set
+CONFIG_FB_CYBER2000=m
+CONFIG_FB_ARC=m
+# CONFIG_FB_ASILIANT is not set
+# CONFIG_FB_IMSTT is not set
+CONFIG_FB_VGA16=m
+CONFIG_FB_VESA=y
+CONFIG_FB_HGA=m
+CONFIG_FB_HGA_ACCEL=y
+CONFIG_FB_S1D13XXX=m
+CONFIG_FB_NVIDIA=m
+CONFIG_FB_NVIDIA_I2C=y
+CONFIG_FB_RIVA=m
+CONFIG_FB_RIVA_I2C=y
+# CONFIG_FB_RIVA_DEBUG is not set
+CONFIG_FB_INTEL=m
+# CONFIG_FB_INTEL_DEBUG is not set
+CONFIG_FB_MATROX=m
+CONFIG_FB_MATROX_MILLENIUM=y
+CONFIG_FB_MATROX_MYSTIQUE=y
+CONFIG_FB_MATROX_G=y
+CONFIG_FB_MATROX_I2C=m
+CONFIG_FB_MATROX_MAVEN=m
+CONFIG_FB_MATROX_MULTIHEAD=y
+CONFIG_FB_RADEON=m
+CONFIG_FB_RADEON_I2C=y
+# CONFIG_FB_RADEON_DEBUG is not set
+CONFIG_FB_ATY128=m
+CONFIG_FB_ATY=m
+CONFIG_FB_ATY_CT=y
+CONFIG_FB_ATY_GENERIC_LCD=y
+CONFIG_FB_ATY_GX=y
+CONFIG_FB_SAVAGE=m
+CONFIG_FB_SAVAGE_I2C=y
+CONFIG_FB_SAVAGE_ACCEL=y
+CONFIG_FB_SIS=m
+CONFIG_FB_SIS_300=y
+CONFIG_FB_SIS_315=y
+CONFIG_FB_NEOMAGIC=m
+CONFIG_FB_KYRO=m
+CONFIG_FB_3DFX=m
+CONFIG_FB_3DFX_ACCEL=y
+CONFIG_FB_VOODOO1=m
+CONFIG_FB_TRIDENT=m
+CONFIG_FB_TRIDENT_ACCEL=y
+CONFIG_FB_GEODE=y
+CONFIG_FB_GEODE_GX=m
+CONFIG_FB_GEODE_GX1=m
+CONFIG_FB_VIRTUAL=m
+
+#
+# Console display driver support
+#
+CONFIG_VGA_CONSOLE=y
+# CONFIG_VGACON_SOFT_SCROLLBACK is not set
+CONFIG_DUMMY_CONSOLE=y
+CONFIG_FRAMEBUFFER_CONSOLE=y
+CONFIG_FRAMEBUFFER_CONSOLE_ROTATION=y
+# CONFIG_FONTS is not set
+CONFIG_FONT_8x8=y
+CONFIG_FONT_8x16=y
+
+#
+# Logo configuration
+#
+CONFIG_LOGO=y
+# CONFIG_LOGO_LINUX_MONO is not set
+# CONFIG_LOGO_LINUX_VGA16 is not set
+CONFIG_LOGO_LINUX_CLUT224=y
+CONFIG_BACKLIGHT_LCD_SUPPORT=y
+CONFIG_BACKLIGHT_CLASS_DEVICE=m
+CONFIG_BACKLIGHT_DEVICE=y
+CONFIG_LCD_CLASS_DEVICE=m
+CONFIG_LCD_DEVICE=y
+
+#
+# Sound
+#
+CONFIG_SOUND=m
+
+#
+# Advanced Linux Sound Architecture
+#
+CONFIG_SND=m
+CONFIG_SND_TIMER=m
+CONFIG_SND_PCM=m
+CONFIG_SND_HWDEP=m
+CONFIG_SND_RAWMIDI=m
+CONFIG_SND_SEQUENCER=m
+CONFIG_SND_SEQ_DUMMY=m
+CONFIG_SND_OSSEMUL=y
+CONFIG_SND_MIXER_OSS=m
+CONFIG_SND_PCM_OSS=m
+CONFIG_SND_PCM_OSS_PLUGINS=y
+CONFIG_SND_SEQUENCER_OSS=y
+CONFIG_SND_RTCTIMER=m
+CONFIG_SND_SEQ_RTCTIMER_DEFAULT=y
+CONFIG_SND_DYNAMIC_MINORS=y
+CONFIG_SND_SUPPORT_OLD_API=y
+CONFIG_SND_VERBOSE_PROCFS=y
+# CONFIG_SND_VERBOSE_PRINTK is not set
+# CONFIG_SND_DEBUG is not set
+
+#
+# Generic devices
+#
+CONFIG_SND_MPU401_UART=m
+CONFIG_SND_OPL3_LIB=m
+CONFIG_SND_VX_LIB=m
+CONFIG_SND_AC97_CODEC=m
+CONFIG_SND_AC97_BUS=m
+CONFIG_SND_DUMMY=m
+CONFIG_SND_VIRMIDI=m
+CONFIG_SND_MTPAV=m
+CONFIG_SND_SERIAL_U16550=m
+CONFIG_SND_MPU401=m
+
+#
+# PCI devices
+#
+CONFIG_SND_AD1889=m
+CONFIG_SND_ALS300=m
+CONFIG_SND_ALS4000=m
+CONFIG_SND_ALI5451=m
+CONFIG_SND_ATIIXP=m
+CONFIG_SND_ATIIXP_MODEM=m
+CONFIG_SND_AU8810=m
+CONFIG_SND_AU8820=m
+CONFIG_SND_AU8830=m
+CONFIG_SND_AZT3328=m
+CONFIG_SND_BT87X=m
+# CONFIG_SND_BT87X_OVERCLOCK is not set
+CONFIG_SND_CA0106=m
+CONFIG_SND_CMIPCI=m
+CONFIG_SND_CS4281=m
+CONFIG_SND_CS46XX=m
+CONFIG_SND_CS46XX_NEW_DSP=y
+CONFIG_SND_DARLA20=m
+CONFIG_SND_GINA20=m
+CONFIG_SND_LAYLA20=m
+CONFIG_SND_DARLA24=m
+CONFIG_SND_GINA24=m
+CONFIG_SND_LAYLA24=m
+CONFIG_SND_MONA=m
+CONFIG_SND_MIA=m
+CONFIG_SND_ECHO3G=m
+CONFIG_SND_INDIGO=m
+CONFIG_SND_INDIGOIO=m
+CONFIG_SND_INDIGODJ=m
+CONFIG_SND_EMU10K1=m
+CONFIG_SND_EMU10K1X=m
+CONFIG_SND_ENS1370=m
+CONFIG_SND_ENS1371=m
+CONFIG_SND_ES1938=m
+CONFIG_SND_ES1968=m
+CONFIG_SND_FM801=m
+CONFIG_SND_FM801_TEA575X_BOOL=y
+CONFIG_SND_FM801_TEA575X=m
+CONFIG_SND_HDA_INTEL=m
+CONFIG_SND_HDSP=m
+CONFIG_SND_HDSPM=m
+CONFIG_SND_ICE1712=m
+CONFIG_SND_ICE1724=m
+CONFIG_SND_INTEL8X0=m
+CONFIG_SND_INTEL8X0M=m
+CONFIG_SND_KORG1212=m
+CONFIG_SND_MAESTRO3=m
+CONFIG_SND_MIXART=m
+CONFIG_SND_NM256=m
+CONFIG_SND_PCXHR=m
+CONFIG_SND_RIPTIDE=m
+CONFIG_SND_RME32=m
+CONFIG_SND_RME96=m
+CONFIG_SND_RME9652=m
+CONFIG_SND_SONICVIBES=m
+CONFIG_SND_TRIDENT=m
+CONFIG_SND_VIA82XX=m
+CONFIG_SND_VIA82XX_MODEM=m
+CONFIG_SND_VX222=m
+CONFIG_SND_YMFPCI=m
+
+#
+# USB devices
+#
+CONFIG_SND_USB_AUDIO=m
+CONFIG_SND_USB_USX2Y=m
+
+#
+# PCMCIA devices
+#
+CONFIG_SND_VXPOCKET=m
+CONFIG_SND_PDAUDIOCF=m
+
+#
+# Open Sound System
+#
+# CONFIG_SOUND_PRIME is not set
+
+#
+# USB support
+#
+CONFIG_USB_ARCH_HAS_HCD=y
+CONFIG_USB_ARCH_HAS_OHCI=y
+CONFIG_USB_ARCH_HAS_EHCI=y
+CONFIG_USB=m
+# CONFIG_USB_DEBUG is not set
+
+#
+# Miscellaneous USB options
+#
+CONFIG_USB_DEVICEFS=y
+# CONFIG_USB_BANDWIDTH is not set
+# CONFIG_USB_DYNAMIC_MINORS is not set
+# CONFIG_USB_SUSPEND is not set
+# CONFIG_USB_OTG is not set
+
+#
+# USB Host Controller Drivers
+#
+CONFIG_USB_EHCI_HCD=m
+CONFIG_USB_EHCI_SPLIT_ISO=y
+CONFIG_USB_EHCI_ROOT_HUB_TT=y
+CONFIG_USB_EHCI_TT_NEWSCHED=y
+CONFIG_USB_ISP116X_HCD=m
+CONFIG_USB_OHCI_HCD=m
+# CONFIG_USB_OHCI_BIG_ENDIAN is not set
+CONFIG_USB_OHCI_LITTLE_ENDIAN=y
+CONFIG_USB_UHCI_HCD=m
+CONFIG_USB_SL811_HCD=m
+# CONFIG_USB_SL811_CS is not set
+
+#
+# USB Device Class drivers
+#
+CONFIG_USB_ACM=m
+CONFIG_USB_PRINTER=m
+
+#
+# NOTE: USB_STORAGE enables SCSI, and 'SCSI disk support'
+#
+
+#
+# may also be needed; see USB_STORAGE Help for more information
+#
+CONFIG_USB_STORAGE=m
+# CONFIG_USB_STORAGE_DEBUG is not set
+CONFIG_USB_STORAGE_DATAFAB=y
+CONFIG_USB_STORAGE_FREECOM=y
+CONFIG_USB_STORAGE_ISD200=y
+CONFIG_USB_STORAGE_DPCM=y
+CONFIG_USB_STORAGE_USBAT=y
+CONFIG_USB_STORAGE_SDDR09=y
+CONFIG_USB_STORAGE_SDDR55=y
+CONFIG_USB_STORAGE_JUMPSHOT=y
+# CONFIG_USB_STORAGE_ALAUDA is not set
+# CONFIG_USB_LIBUSUAL is not set
+
+#
+# USB Input Devices
+#
+CONFIG_USB_HID=m
+CONFIG_USB_HIDINPUT=y
+# CONFIG_USB_HIDINPUT_POWERBOOK is not set
+CONFIG_HID_FF=y
+CONFIG_HID_PID=y
+CONFIG_LOGITECH_FF=y
+CONFIG_THRUSTMASTER_FF=y
+CONFIG_USB_HIDDEV=y
+
+#
+# USB HID Boot Protocol drivers
+#
+CONFIG_USB_KBD=m
+CONFIG_USB_MOUSE=m
+CONFIG_USB_AIPTEK=m
+CONFIG_USB_WACOM=m
+CONFIG_USB_ACECAD=m
+CONFIG_USB_KBTAB=m
+CONFIG_USB_POWERMATE=m
+CONFIG_USB_TOUCHSCREEN=m
+CONFIG_USB_TOUCHSCREEN_EGALAX=y
+CONFIG_USB_TOUCHSCREEN_PANJIT=y
+CONFIG_USB_TOUCHSCREEN_3M=y
+CONFIG_USB_TOUCHSCREEN_ITM=y
+CONFIG_USB_YEALINK=m
+CONFIG_USB_XPAD=m
+CONFIG_USB_ATI_REMOTE=m
+CONFIG_USB_ATI_REMOTE2=m
+CONFIG_USB_KEYSPAN_REMOTE=m
+CONFIG_USB_APPLETOUCH=m
+
+#
+# USB Imaging devices
+#
+CONFIG_USB_MDC800=m
+CONFIG_USB_MICROTEK=m
+
+#
+# USB Network Adapters
+#
+CONFIG_USB_CATC=m
+CONFIG_USB_KAWETH=m
+CONFIG_USB_PEGASUS=m
+CONFIG_USB_RTL8150=m
+CONFIG_USB_USBNET=m
+CONFIG_USB_NET_AX8817X=m
+CONFIG_USB_NET_CDCETHER=m
+CONFIG_USB_NET_GL620A=m
+CONFIG_USB_NET_NET1080=m
+CONFIG_USB_NET_PLUSB=m
+CONFIG_USB_NET_RNDIS_HOST=m
+CONFIG_USB_NET_CDC_SUBSET=m
+CONFIG_USB_ALI_M5632=y
+CONFIG_USB_AN2720=y
+CONFIG_USB_BELKIN=y
+CONFIG_USB_ARMLINUX=y
+CONFIG_USB_EPSON2888=y
+CONFIG_USB_NET_ZAURUS=m
+CONFIG_USB_MON=y
+
+#
+# USB port drivers
+#
+CONFIG_USB_USS720=m
+
+#
+# USB Serial Converter support
+#
+CONFIG_USB_SERIAL=m
+CONFIG_USB_SERIAL_GENERIC=y
+CONFIG_USB_SERIAL_AIRPRIME=m
+CONFIG_USB_SERIAL_ARK3116=m
+CONFIG_USB_SERIAL_BELKIN=m
+CONFIG_USB_SERIAL_WHITEHEAT=m
+CONFIG_USB_SERIAL_DIGI_ACCELEPORT=m
+CONFIG_USB_SERIAL_CP2101=m
+CONFIG_USB_SERIAL_CYPRESS_M8=m
+CONFIG_USB_SERIAL_EMPEG=m
+CONFIG_USB_SERIAL_FTDI_SIO=m
+CONFIG_USB_SERIAL_FUNSOFT=m
+CONFIG_USB_SERIAL_VISOR=m
+CONFIG_USB_SERIAL_IPAQ=m
+CONFIG_USB_SERIAL_IR=m
+CONFIG_USB_SERIAL_EDGEPORT=m
+CONFIG_USB_SERIAL_EDGEPORT_TI=m
+CONFIG_USB_SERIAL_GARMIN=m
+CONFIG_USB_SERIAL_IPW=m
+CONFIG_USB_SERIAL_KEYSPAN_PDA=m
+CONFIG_USB_SERIAL_KEYSPAN=m
+CONFIG_USB_SERIAL_KEYSPAN_MPR=y
+CONFIG_USB_SERIAL_KEYSPAN_USA28=y
+CONFIG_USB_SERIAL_KEYSPAN_USA28X=y
+CONFIG_USB_SERIAL_KEYSPAN_USA28XA=y
+CONFIG_USB_SERIAL_KEYSPAN_USA28XB=y
+CONFIG_USB_SERIAL_KEYSPAN_USA19=y
+CONFIG_USB_SERIAL_KEYSPAN_USA18X=y
+CONFIG_USB_SERIAL_KEYSPAN_USA19W=y
+CONFIG_USB_SERIAL_KEYSPAN_USA19QW=y
+CONFIG_USB_SERIAL_KEYSPAN_USA19QI=y
+CONFIG_USB_SERIAL_KEYSPAN_USA49W=y
+CONFIG_USB_SERIAL_KEYSPAN_USA49WLC=y
+CONFIG_USB_SERIAL_KLSI=m
+CONFIG_USB_SERIAL_KOBIL_SCT=m
+CONFIG_USB_SERIAL_MCT_U232=m
+CONFIG_USB_SERIAL_NAVMAN=m
+CONFIG_USB_SERIAL_PL2303=m
+CONFIG_USB_SERIAL_HP4X=m
+CONFIG_USB_SERIAL_SAFE=m
+CONFIG_USB_SERIAL_SAFE_PADDED=y
+CONFIG_USB_SERIAL_SIERRAWIRELESS=m
+CONFIG_USB_SERIAL_TI=m
+CONFIG_USB_SERIAL_CYBERJACK=m
+CONFIG_USB_SERIAL_XIRCOM=m
+# CONFIG_USB_SERIAL_OPTION is not set
+CONFIG_USB_SERIAL_OMNINET=m
+CONFIG_USB_EZUSB=y
+
+#
+# USB Miscellaneous drivers
+#
+CONFIG_USB_EMI62=m
+CONFIG_USB_EMI26=m
+CONFIG_USB_AUERSWALD=m
+CONFIG_USB_RIO500=m
+CONFIG_USB_LEGOTOWER=m
+CONFIG_USB_LCD=m
+CONFIG_USB_LED=m
+CONFIG_USB_CYPRESS_CY7C63=m
+CONFIG_USB_CYTHERM=m
+CONFIG_USB_PHIDGETKIT=m
+CONFIG_USB_PHIDGETSERVO=m
+CONFIG_USB_IDMOUSE=m
+CONFIG_USB_APPLEDISPLAY=m
+CONFIG_USB_SISUSBVGA=m
+CONFIG_USB_SISUSBVGA_CON=y
+CONFIG_USB_LD=m
+CONFIG_USB_TEST=m
+
+#
+# USB DSL modem support
+#
+CONFIG_USB_ATM=m
+CONFIG_USB_SPEEDTOUCH=m
+CONFIG_USB_CXACRU=m
+CONFIG_USB_UEAGLEATM=m
+CONFIG_USB_XUSBATM=m
+
+#
+# USB Gadget Support
+#
+# CONFIG_USB_GADGET is not set
+
+#
+# MMC/SD Card support
+#
+CONFIG_MMC=m
+# CONFIG_MMC_DEBUG is not set
+CONFIG_MMC_BLOCK=m
+CONFIG_MMC_SDHCI=m
+CONFIG_MMC_WBSD=m
+
+#
+# LED devices
+#
+CONFIG_NEW_LEDS=y
+CONFIG_LEDS_CLASS=m
+
+#
+# LED drivers
+#
+
+#
+# LED Triggers
+#
+CONFIG_LEDS_TRIGGERS=y
+CONFIG_LEDS_TRIGGER_TIMER=m
+CONFIG_LEDS_TRIGGER_IDE_DISK=y
+CONFIG_LEDS_TRIGGER_HEARTBEAT=m
+
+#
+# InfiniBand support
+#
+CONFIG_INFINIBAND=m
+CONFIG_INFINIBAND_USER_MAD=m
+CONFIG_INFINIBAND_USER_ACCESS=m
+CONFIG_INFINIBAND_ADDR_TRANS=y
+CONFIG_INFINIBAND_MTHCA=m
+CONFIG_INFINIBAND_MTHCA_DEBUG=y
+CONFIG_INFINIBAND_IPOIB=m
+CONFIG_INFINIBAND_IPOIB_DEBUG=y
+# CONFIG_INFINIBAND_IPOIB_DEBUG_DATA is not set
+CONFIG_INFINIBAND_SRP=m
+CONFIG_INFINIBAND_ISER=m
+
+#
+# EDAC - error detection and reporting (RAS) (EXPERIMENTAL)
+#
+CONFIG_EDAC=m
+
+#
+# Reporting subsystems
+#
+# CONFIG_EDAC_DEBUG is not set
+CONFIG_EDAC_MM_EDAC=m
+CONFIG_EDAC_E752X=m
+CONFIG_EDAC_POLL=y
+
+#
+# Real Time Clock
+#
+CONFIG_RTC_LIB=m
+CONFIG_RTC_CLASS=m
+
+#
+# RTC interfaces
+#
+CONFIG_RTC_INTF_SYSFS=m
+CONFIG_RTC_INTF_PROC=m
+CONFIG_RTC_INTF_DEV=m
+CONFIG_RTC_INTF_DEV_UIE_EMUL=y
+
+#
+# RTC drivers
+#
+CONFIG_RTC_DRV_X1205=m
+CONFIG_RTC_DRV_DS1307=m
+CONFIG_RTC_DRV_DS1553=m
+CONFIG_RTC_DRV_ISL1208=m
+CONFIG_RTC_DRV_DS1672=m
+CONFIG_RTC_DRV_DS1742=m
+CONFIG_RTC_DRV_PCF8563=m
+CONFIG_RTC_DRV_PCF8583=m
+CONFIG_RTC_DRV_RS5C348=m
+CONFIG_RTC_DRV_RS5C372=m
+CONFIG_RTC_DRV_M48T86=m
+CONFIG_RTC_DRV_TEST=m
+CONFIG_RTC_DRV_MAX6902=m
+CONFIG_RTC_DRV_V3020=m
+
+#
+# DMA Engine support
+#
+CONFIG_DMA_ENGINE=y
+
+#
+# DMA Clients
+#
+CONFIG_NET_DMA=y
+
+#
+# DMA Devices
+#
+CONFIG_INTEL_IOATDMA=m
+
+#
+# Firmware Drivers
+#
+CONFIG_EDD=y
+CONFIG_DELL_RBU=m
+CONFIG_DCDBAS=m
+
+#
+# File systems
+#
+CONFIG_EXT2_FS=y
+CONFIG_EXT2_FS_XATTR=y
+CONFIG_EXT2_FS_POSIX_ACL=y
+CONFIG_EXT2_FS_SECURITY=y
+CONFIG_EXT2_FS_XIP=y
+CONFIG_FS_XIP=y
+CONFIG_EXT3_FS=m
+CONFIG_EXT3_FS_XATTR=y
+CONFIG_EXT3_FS_POSIX_ACL=y
+CONFIG_EXT3_FS_SECURITY=y
+CONFIG_JBD=m
+# CONFIG_JBD_DEBUG is not set
+CONFIG_FS_MBCACHE=y
+CONFIG_REISERFS_FS=m
+# CONFIG_REISERFS_CHECK is not set
+CONFIG_REISERFS_PROC_INFO=y
+CONFIG_REISERFS_FS_XATTR=y
+CONFIG_REISERFS_FS_POSIX_ACL=y
+CONFIG_REISERFS_FS_SECURITY=y
+CONFIG_JFS_FS=m
+CONFIG_JFS_POSIX_ACL=y
+CONFIG_JFS_SECURITY=y
+# CONFIG_JFS_DEBUG is not set
+# CONFIG_JFS_STATISTICS is not set
+CONFIG_FS_POSIX_ACL=y
+CONFIG_XFS_FS=m
+CONFIG_XFS_QUOTA=y
+CONFIG_XFS_SECURITY=y
+CONFIG_XFS_POSIX_ACL=y
+CONFIG_XFS_RT=y
+CONFIG_OCFS2_FS=m
+CONFIG_OCFS2_DEBUG_MASKLOG=y
+CONFIG_MINIX_FS=m
+CONFIG_ROMFS_FS=m
+CONFIG_INOTIFY=y
+CONFIG_INOTIFY_USER=y
+CONFIG_QUOTA=y
+# CONFIG_QFMT_V1 is not set
+CONFIG_QFMT_V2=m
+CONFIG_QUOTACTL=y
+CONFIG_DNOTIFY=y
+CONFIG_AUTOFS_FS=m
+CONFIG_AUTOFS4_FS=m
+CONFIG_FUSE_FS=m
+
+#
+# CD-ROM/DVD Filesystems
+#
+CONFIG_ISO9660_FS=y
+CONFIG_JOLIET=y
+CONFIG_ZISOFS=y
+CONFIG_ZISOFS_FS=y
+CONFIG_UDF_FS=m
+CONFIG_UDF_NLS=y
+
+#
+# DOS/FAT/NT Filesystems
+#
+CONFIG_FAT_FS=m
+CONFIG_MSDOS_FS=m
+CONFIG_VFAT_FS=m
+CONFIG_FAT_DEFAULT_CODEPAGE=437
+CONFIG_FAT_DEFAULT_IOCHARSET="iso8859-1"
+CONFIG_NTFS_FS=m
+# CONFIG_NTFS_DEBUG is not set
+# CONFIG_NTFS_RW is not set
+
+#
+# Pseudo filesystems
+#
+CONFIG_PROC_FS=y
+CONFIG_PROC_KCORE=y
+CONFIG_SYSFS=y
+CONFIG_TMPFS=y
+# CONFIG_HUGETLB_PAGE is not set
+CONFIG_RAMFS=y
+CONFIG_CONFIGFS_FS=m
+
+#
+# Miscellaneous filesystems
+#
+CONFIG_ADFS_FS=m
+# CONFIG_ADFS_FS_RW is not set
+CONFIG_AFFS_FS=m
+CONFIG_HFS_FS=m
+CONFIG_HFSPLUS_FS=m
+CONFIG_BEFS_FS=m
+# CONFIG_BEFS_DEBUG is not set
+CONFIG_BFS_FS=m
+CONFIG_EFS_FS=m
+CONFIG_JFFS_FS=m
+CONFIG_JFFS_FS_VERBOSE=0
+CONFIG_JFFS_PROC_FS=y
+CONFIG_JFFS2_FS=m
+CONFIG_JFFS2_FS_DEBUG=0
+CONFIG_JFFS2_FS_WRITEBUFFER=y
+CONFIG_JFFS2_SUMMARY=y
+CONFIG_JFFS2_FS_XATTR=y
+CONFIG_JFFS2_FS_POSIX_ACL=y
+CONFIG_JFFS2_FS_SECURITY=y
+CONFIG_JFFS2_COMPRESSION_OPTIONS=y
+CONFIG_JFFS2_ZLIB=y
+CONFIG_JFFS2_RTIME=y
+# CONFIG_JFFS2_RUBIN is not set
+# CONFIG_JFFS2_CMODE_NONE is not set
+CONFIG_JFFS2_CMODE_PRIORITY=y
+# CONFIG_JFFS2_CMODE_SIZE is not set
+CONFIG_CRAMFS=m
+CONFIG_VXFS_FS=m
+CONFIG_HPFS_FS=m
+CONFIG_QNX4FS_FS=m
+CONFIG_SYSV_FS=m
+CONFIG_UFS_FS=m
+# CONFIG_UFS_FS_WRITE is not set
+# CONFIG_UFS_DEBUG is not set
+
+#
+# Network File Systems
+#
+CONFIG_NFS_FS=m
+CONFIG_NFS_V3=y
+CONFIG_NFS_V3_ACL=y
+CONFIG_NFS_V4=y
+CONFIG_NFS_DIRECTIO=y
+CONFIG_NFSD=m
+CONFIG_NFSD_V2_ACL=y
+CONFIG_NFSD_V3=y
+CONFIG_NFSD_V3_ACL=y
+CONFIG_NFSD_V4=y
+CONFIG_NFSD_TCP=y
+CONFIG_LOCKD=m
+CONFIG_LOCKD_V4=y
+CONFIG_EXPORTFS=m
+CONFIG_NFS_ACL_SUPPORT=m
+CONFIG_NFS_COMMON=y
+CONFIG_SUNRPC=m
+CONFIG_SUNRPC_GSS=m
+CONFIG_RPCSEC_GSS_KRB5=m
+CONFIG_RPCSEC_GSS_SPKM3=m
+CONFIG_SMB_FS=m
+CONFIG_SMB_NLS_DEFAULT=y
+CONFIG_SMB_NLS_REMOTE="cp850"
+CONFIG_CIFS=m
+CONFIG_CIFS_STATS=y
+# CONFIG_CIFS_STATS2 is not set
+# CONFIG_CIFS_WEAK_PW_HASH is not set
+CONFIG_CIFS_XATTR=y
+CONFIG_CIFS_POSIX=y
+# CONFIG_CIFS_DEBUG2 is not set
+# CONFIG_CIFS_EXPERIMENTAL is not set
+CONFIG_NCP_FS=m
+CONFIG_NCPFS_PACKET_SIGNING=y
+CONFIG_NCPFS_IOCTL_LOCKING=y
+CONFIG_NCPFS_STRONG=y
+CONFIG_NCPFS_NFS_NS=y
+CONFIG_NCPFS_OS2_NS=y
+CONFIG_NCPFS_SMALLDOS=y
+CONFIG_NCPFS_NLS=y
+CONFIG_NCPFS_EXTRAS=y
+CONFIG_CODA_FS=m
+# CONFIG_CODA_FS_OLD_API is not set
+CONFIG_AFS_FS=m
+CONFIG_RXRPC=m
+CONFIG_9P_FS=m
+
+#
+# Partition Types
+#
+CONFIG_PARTITION_ADVANCED=y
+# CONFIG_ACORN_PARTITION is not set
+CONFIG_OSF_PARTITION=y
+CONFIG_AMIGA_PARTITION=y
+# CONFIG_ATARI_PARTITION is not set
+CONFIG_MAC_PARTITION=y
+CONFIG_MSDOS_PARTITION=y
+CONFIG_BSD_DISKLABEL=y
+CONFIG_MINIX_SUBPARTITION=y
+CONFIG_SOLARIS_X86_PARTITION=y
+CONFIG_UNIXWARE_DISKLABEL=y
+# CONFIG_LDM_PARTITION is not set
+CONFIG_SGI_PARTITION=y
+# CONFIG_ULTRIX_PARTITION is not set
+CONFIG_SUN_PARTITION=y
+# CONFIG_KARMA_PARTITION is not set
+CONFIG_EFI_PARTITION=y
+
+#
+# Native Language Support
+#
+CONFIG_NLS=y
+CONFIG_NLS_DEFAULT="utf8"
+CONFIG_NLS_CODEPAGE_437=m
+CONFIG_NLS_CODEPAGE_737=m
+CONFIG_NLS_CODEPAGE_775=m
+CONFIG_NLS_CODEPAGE_850=m
+CONFIG_NLS_CODEPAGE_852=m
+CONFIG_NLS_CODEPAGE_855=m
+CONFIG_NLS_CODEPAGE_857=m
+CONFIG_NLS_CODEPAGE_860=m
+CONFIG_NLS_CODEPAGE_861=m
+CONFIG_NLS_CODEPAGE_862=m
+CONFIG_NLS_CODEPAGE_863=m
+CONFIG_NLS_CODEPAGE_864=m
+CONFIG_NLS_CODEPAGE_865=m
+CONFIG_NLS_CODEPAGE_866=m
+CONFIG_NLS_CODEPAGE_869=m
+CONFIG_NLS_CODEPAGE_936=m
+CONFIG_NLS_CODEPAGE_950=m
+CONFIG_NLS_CODEPAGE_932=m
+CONFIG_NLS_CODEPAGE_949=m
+CONFIG_NLS_CODEPAGE_874=m
+CONFIG_NLS_ISO8859_8=m
+CONFIG_NLS_CODEPAGE_1250=m
+CONFIG_NLS_CODEPAGE_1251=m
+CONFIG_NLS_ASCII=m
+CONFIG_NLS_ISO8859_1=m
+CONFIG_NLS_ISO8859_2=m
+CONFIG_NLS_ISO8859_3=m
+CONFIG_NLS_ISO8859_4=m
+CONFIG_NLS_ISO8859_5=m
+CONFIG_NLS_ISO8859_6=m
+CONFIG_NLS_ISO8859_7=m
+CONFIG_NLS_ISO8859_9=m
+CONFIG_NLS_ISO8859_13=m
+CONFIG_NLS_ISO8859_14=m
+CONFIG_NLS_ISO8859_15=m
+CONFIG_NLS_KOI8_R=m
+CONFIG_NLS_KOI8_U=m
+CONFIG_NLS_UTF8=m
+
+#
+# Instrumentation Support
+#
+# CONFIG_PROFILING is not set
+# CONFIG_KPROBES is not set
+
+#
+# Kernel hacking
+#
+CONFIG_TRACE_IRQFLAGS_SUPPORT=y
+# CONFIG_PRINTK_TIME is not set
+CONFIG_MAGIC_SYSRQ=y
+CONFIG_UNUSED_SYMBOLS=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_LOG_BUF_SHIFT=15
+CONFIG_DETECT_SOFTLOCKUP=y
+# CONFIG_SCHEDSTATS is not set
+# CONFIG_DEBUG_SLAB is not set
+# CONFIG_DEBUG_RT_MUTEXES is not set
+# CONFIG_RT_MUTEX_TESTER is not set
+# CONFIG_DEBUG_SPINLOCK is not set
+# CONFIG_DEBUG_MUTEXES is not set
+# CONFIG_DEBUG_RWSEMS is not set
+# CONFIG_DEBUG_LOCK_ALLOC is not set
+# CONFIG_PROVE_LOCKING is not set
+# CONFIG_DEBUG_SPINLOCK_SLEEP is not set
+# CONFIG_DEBUG_LOCKING_API_SELFTESTS is not set
+# CONFIG_DEBUG_KOBJECT is not set
+# CONFIG_DEBUG_INFO is not set
+# CONFIG_DEBUG_FS is not set
+# CONFIG_DEBUG_VM is not set
+# CONFIG_FRAME_POINTER is not set
+# CONFIG_UNWIND_INFO is not set
+CONFIG_FORCED_INLINING=y
+# CONFIG_RCU_TORTURE_TEST is not set
+# CONFIG_DEBUG_RODATA is not set
+# CONFIG_DEBUG_STACKOVERFLOW is not set
+# CONFIG_DEBUG_STACK_USAGE is not set
+
+#
+# Security options
+#
+CONFIG_KEYS=y
+CONFIG_KEYS_DEBUG_PROC_KEYS=y
+CONFIG_SECURITY=y
+CONFIG_SECURITY_NETWORK=y
+# CONFIG_SECURITY_NETWORK_XFRM is not set
+CONFIG_SECURITY_CAPABILITIES=y
+CONFIG_SECURITY_ROOTPLUG=m
+CONFIG_SECURITY_SECLVL=m
+# CONFIG_SECURITY_SELINUX is not set
+
+#
+# Cryptographic options
+#
+CONFIG_CRYPTO=y
+CONFIG_CRYPTO_HMAC=y
+CONFIG_CRYPTO_NULL=m
+CONFIG_CRYPTO_MD4=m
+CONFIG_CRYPTO_MD5=y
+CONFIG_CRYPTO_SHA1=m
+CONFIG_CRYPTO_SHA256=m
+CONFIG_CRYPTO_SHA512=m
+CONFIG_CRYPTO_WP512=m
+CONFIG_CRYPTO_TGR192=m
+CONFIG_CRYPTO_DES=m
+CONFIG_CRYPTO_BLOWFISH=m
+CONFIG_CRYPTO_TWOFISH=m
+CONFIG_CRYPTO_SERPENT=m
+CONFIG_CRYPTO_AES=m
+CONFIG_CRYPTO_AES_X86_64=m
+CONFIG_CRYPTO_CAST5=m
+CONFIG_CRYPTO_CAST6=m
+CONFIG_CRYPTO_TEA=m
+CONFIG_CRYPTO_ARC4=m
+CONFIG_CRYPTO_KHAZAD=m
+CONFIG_CRYPTO_ANUBIS=m
+CONFIG_CRYPTO_DEFLATE=m
+CONFIG_CRYPTO_MICHAEL_MIC=m
+CONFIG_CRYPTO_CRC32C=m
+CONFIG_CRYPTO_TEST=m
+
+#
+# Hardware crypto devices
+#
+CONFIG_XEN=y
+CONFIG_XEN_INTERFACE_VERSION=0x00030207
+
+#
+# XEN
+#
+CONFIG_XEN_PRIVILEGED_GUEST=y
+# CONFIG_XEN_UNPRIVILEGED_GUEST is not set
+CONFIG_XEN_PRIVCMD=y
+CONFIG_XEN_XENBUS_DEV=y
+CONFIG_XEN_BACKEND=y
+CONFIG_XEN_BLKDEV_BACKEND=y
+CONFIG_XEN_BLKDEV_TAP=y
+CONFIG_XEN_NETDEV_BACKEND=y
+# CONFIG_XEN_NETDEV_PIPELINED_TRANSMITTER is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_UTIL is not set
+# CONFIG_XEN_NETDEV_ACCEL_SFC_BACKEND is not set
+# CONFIG_XEN_NETDEV_LOOPBACK is not set
+CONFIG_XEN_PCIDEV_BACKEND=m
+# CONFIG_XEN_PCIDEV_BACKEND_VPCI is not set
+CONFIG_XEN_PCIDEV_BACKEND_PASS=y
+# CONFIG_XEN_PCIDEV_BACKEND_SLOT is not set
+# CONFIG_XEN_PCIDEV_BACKEND_CONTROLLER is not set
+# CONFIG_XEN_PCIDEV_BE_DEBUG is not set
+CONFIG_XEN_TPMDEV_BACKEND=m
+CONFIG_XEN_SCSI_BACKEND=m
+CONFIG_XEN_BLKDEV_FRONTEND=y
+CONFIG_XEN_NETDEV_FRONTEND=y
+CONFIG_XEN_SCSI_FRONTEND=m
+CONFIG_XEN_GRANT_DEV=y
+# CONFIG_XEN_NETDEV_ACCEL_SFC_FRONTEND is not set
+CONFIG_XEN_FRAMEBUFFER=y
+CONFIG_XEN_KEYBOARD=y
+CONFIG_XEN_SCRUB_PAGES=y
+CONFIG_XEN_DISABLE_SERIAL=y
+CONFIG_XEN_SYSFS=y
+CONFIG_XEN_COMPAT_030002_AND_LATER=y
+# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
+# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
+CONFIG_XEN_COMPAT=0x030002
+CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
+CONFIG_NO_IDLE_HZ=y
+CONFIG_XEN_SMPBOOT=y
+CONFIG_XEN_BALLOON=y
+CONFIG_XEN_DEVMEM=y
+
+#
+# Library routines
+#
+CONFIG_CRC_CCITT=m
+CONFIG_CRC16=m
+CONFIG_CRC32=y
+CONFIG_LIBCRC32C=m
+CONFIG_ZLIB_INFLATE=y
+CONFIG_ZLIB_DEFLATE=m
+CONFIG_REED_SOLOMON=m
+CONFIG_REED_SOLOMON_DEC16=y
+CONFIG_TEXTSEARCH=y
+CONFIG_TEXTSEARCH_KMP=m
+CONFIG_TEXTSEARCH_BM=m
+CONFIG_TEXTSEARCH_FSM=m
+CONFIG_PLIST=y
diff -rpuN linux-2.6.18.8/Documentation/i2c/busses/i2c-i801 linux-2.6.18-xen-3.3.0/Documentation/i2c/busses/i2c-i801
--- linux-2.6.18.8/Documentation/i2c/busses/i2c-i801	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/Documentation/i2c/busses/i2c-i801	2008-08-21 11:36:07.000000000 +0200
@@ -10,6 +10,11 @@ Supported adapters:
   * Intel 6300ESB
   * Intel 82801FB/FR/FW/FRW (ICH6)
   * Intel ICH7
+  * Intel 82801G (ICH7)
+  * Intel 82801H (ICH9)
+  * Intel 82801I (ICH9)
+  * Intel Tolapai
+  * Intel ICH10
     Datasheets: Publicly available at the Intel website
 
 Authors: 
diff -rpuN linux-2.6.18.8/drivers/acpi/dispatcher/dsobject.c linux-2.6.18-xen-3.3.0/drivers/acpi/dispatcher/dsobject.c
--- linux-2.6.18.8/drivers/acpi/dispatcher/dsobject.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/acpi/dispatcher/dsobject.c	2008-08-21 11:36:07.000000000 +0200
@@ -137,6 +137,71 @@ acpi_ds_build_internal_object(struct acp
 				return_ACPI_STATUS(status);
 			}
 		}
+
+		/* Special object resolution for elements of a package */
+
+		if ((op->common.parent->common.aml_opcode == AML_PACKAGE_OP) ||
+		    (op->common.parent->common.aml_opcode ==
+		     AML_VAR_PACKAGE_OP)) {
+			/*
+			 * Attempt to resolve the node to a value before we insert it into
+			 * the package. If this is a reference to a common data type,
+			 * resolve it immediately. According to the ACPI spec, package
+			 * elements can only be "data objects" or method references.
+			 * Attempt to resolve to an Integer, Buffer, String or Package.
+			 * If cannot, return the named reference (for things like Devices,
+			 * Methods, etc.) Buffer Fields and Fields will resolve to simple
+			 * objects (int/buf/str/pkg).
+			 *
+			 * NOTE: References to things like Devices, Methods, Mutexes, etc.
+			 * will remain as named references. This behavior is not described
+			 * in the ACPI spec, but it appears to be an oversight.
+			 */
+			obj_desc = (union acpi_operand_object *)op->common.node;
+
+			status =
+			    acpi_ex_resolve_node_to_value(ACPI_CAST_INDIRECT_PTR
+							  (struct
+							   acpi_namespace_node,
+							   &obj_desc),
+							  walk_state);
+			if (ACPI_FAILURE(status)) {
+				return_ACPI_STATUS(status);
+			}
+
+			switch (op->common.node->type) {
+				/*
+				 * For these types, we need the actual node, not the subobject.
+				 * However, the subobject got an extra reference count above.
+				 */
+			case ACPI_TYPE_MUTEX:
+			case ACPI_TYPE_METHOD:
+			case ACPI_TYPE_POWER:
+			case ACPI_TYPE_PROCESSOR:
+			case ACPI_TYPE_EVENT:
+			case ACPI_TYPE_REGION:
+			case ACPI_TYPE_DEVICE:
+			case ACPI_TYPE_THERMAL:
+
+				obj_desc =
+				    (union acpi_operand_object *)op->common.
+				    node;
+				break;
+
+			default:
+				break;
+			}
+
+			/*
+			 * If above resolved to an operand object, we are done. Otherwise,
+			 * we have a NS node, we must create the package entry as a named
+			 * reference.
+			 */
+			if (ACPI_GET_DESCRIPTOR_TYPE(obj_desc) !=
+			    ACPI_DESC_TYPE_NAMED) {
+				goto exit;
+			}
+		}
 	}
 
 	/* Create and init a new internal ACPI object */
@@ -156,6 +221,7 @@ acpi_ds_build_internal_object(struct acp
 		return_ACPI_STATUS(status);
 	}
 
+      exit:
 	*obj_desc_ptr = obj_desc;
 	return_ACPI_STATUS(AE_OK);
 }
@@ -358,12 +424,25 @@ acpi_ds_build_internal_package_obj(struc
 	arg = arg->common.next;
 	for (i = 0; arg; i++) {
 		if (arg->common.aml_opcode == AML_INT_RETURN_VALUE_OP) {
-
-			/* Object (package or buffer) is already built */
-
-			obj_desc->package.elements[i] =
-			    ACPI_CAST_PTR(union acpi_operand_object,
-					  arg->common.node);
+			if (arg->common.node->type == ACPI_TYPE_METHOD) {
+				/*
+				 * A method reference "looks" to the parser to be a method
+				 * invocation, so we special case it here
+				 */
+				arg->common.aml_opcode = AML_INT_NAMEPATH_OP;
+				status =
+				    acpi_ds_build_internal_object(walk_state,
+								  arg,
+								  &obj_desc->
+								  package.
+								  elements[i]);
+			} else {
+				/* This package element is already built, just get it */
+
+				obj_desc->package.elements[i] =
+				    ACPI_CAST_PTR(union acpi_operand_object,
+						  arg->common.node);
+			}
 		} else {
 			status = acpi_ds_build_internal_object(walk_state, arg,
 							       &obj_desc->
diff -rpuN linux-2.6.18.8/drivers/acpi/hardware/hwregs.c linux-2.6.18-xen-3.3.0/drivers/acpi/hardware/hwregs.c
--- linux-2.6.18.8/drivers/acpi/hardware/hwregs.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/acpi/hardware/hwregs.c	2008-08-21 11:36:07.000000000 +0200
@@ -665,8 +665,6 @@ acpi_status acpi_hw_register_write(u8 us
 
 		/*
 		 * Perform a read first to preserve certain bits (per ACPI spec)
-		 *
-		 * Note: This includes SCI_EN, we never want to change this bit
 		 */
 		status = acpi_hw_register_read(ACPI_MTX_DO_NOT_LOCK,
 					       ACPI_REGISTER_PM1_CONTROL,
diff -rpuN linux-2.6.18.8/drivers/acpi/hardware/hwsleep.c linux-2.6.18-xen-3.3.0/drivers/acpi/hardware/hwsleep.c
--- linux-2.6.18.8/drivers/acpi/hardware/hwsleep.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/acpi/hardware/hwsleep.c	2008-08-21 11:36:07.000000000 +0200
@@ -227,7 +227,11 @@ acpi_status asmlinkage acpi_enter_sleep_
 	u32 PM1Bcontrol;
 	struct acpi_bit_register_info *sleep_type_reg_info;
 	struct acpi_bit_register_info *sleep_enable_reg_info;
+#if !(defined(CONFIG_XEN) && defined(CONFIG_X86))
 	u32 in_value;
+#else
+	int err;
+#endif
 	acpi_status status;
 
 	ACPI_FUNCTION_TRACE(acpi_enter_sleep_state);
@@ -327,6 +331,7 @@ acpi_status asmlinkage acpi_enter_sleep_
 
 	ACPI_FLUSH_CPU_CACHE();
 
+#if !(defined(CONFIG_XEN) && defined(CONFIG_X86))
 	status = acpi_hw_register_write(ACPI_MTX_DO_NOT_LOCK,
 					ACPI_REGISTER_PM1A_CONTROL,
 					PM1Acontrol);
@@ -376,6 +381,16 @@ acpi_status asmlinkage acpi_enter_sleep_
 		/* Spin until we wake */
 
 	} while (!in_value);
+#else
+	/* PV ACPI just need check hypercall return value */
+	err = acpi_notify_hypervisor_state(sleep_state,
+			PM1Acontrol, PM1Bcontrol);
+	if (err) {
+		ACPI_DEBUG_PRINT((ACPI_DB_ERROR,
+				  "Hypervisor failure [%d]\n", err));
+		return_ACPI_STATUS(AE_ERROR);
+	}
+#endif
 
 	return_ACPI_STATUS(AE_OK);
 }
diff -rpuN linux-2.6.18.8/drivers/acpi/Kconfig linux-2.6.18-xen-3.3.0/drivers/acpi/Kconfig
--- linux-2.6.18.8/drivers/acpi/Kconfig	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/acpi/Kconfig	2008-08-21 11:36:07.000000000 +0200
@@ -305,6 +305,7 @@ config ACPI_SYSTEM
 config X86_PM_TIMER
 	bool "Power Management Timer Support" if EMBEDDED
 	depends on X86
+	depends on !XEN
 	default y
 	help
 	  The Power Management Timer is available on all ACPI-capable,
@@ -362,6 +363,15 @@ config ACPI_SBS
 	  A "Smart Battery" is quite old and quite rare compared
 	  to today's ACPI "Control Method" battery.
 
+config ACPI_PV_SLEEP
+	bool
+	depends on X86 && XEN && ACPI_SLEEP
+	default y
+
+config PROCESSOR_EXTERNAL_CONTROL
+	bool
+	depends on X86 && XEN
+	default y
 endif	# ACPI
 
 endmenu
diff -rpuN linux-2.6.18.8/drivers/acpi/Makefile linux-2.6.18-xen-3.3.0/drivers/acpi/Makefile
--- linux-2.6.18.8/drivers/acpi/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/acpi/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -34,6 +34,9 @@ processor-objs	+= processor_core.o proce
 ifdef CONFIG_CPU_FREQ
 processor-objs	+= processor_perflib.o			
 endif
+ifdef CONFIG_PROCESSOR_EXTERNAL_CONTROL
+processor-objs	+= processor_perflib.o processor_extcntl.o
+endif
 
 obj-y				+= sleep/
 obj-y				+= bus.o glue.o
diff -rpuN linux-2.6.18.8/drivers/acpi/processor_core.c linux-2.6.18-xen-3.3.0/drivers/acpi/processor_core.c
--- linux-2.6.18.8/drivers/acpi/processor_core.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/acpi/processor_core.c	2008-08-21 11:36:07.000000000 +0200
@@ -475,7 +475,8 @@ static int acpi_processor_get_info(struc
 	 */
 	if (cpu_index == -1) {
 		if (ACPI_FAILURE
-		    (acpi_processor_hotadd_init(pr->handle, &pr->id))) {
+		    (acpi_processor_hotadd_init(pr->handle, &pr->id)) &&
+		    !processor_cntl_external()) {
 			printk(KERN_ERR PREFIX
 				    "Getting cpuindex for acpiid 0x%x\n",
 				    pr->acpi_id);
@@ -508,7 +509,7 @@ static int acpi_processor_get_info(struc
 		request_region(pr->throttling.address, 6, "ACPI CPU throttle");
 	}
 
-#ifdef CONFIG_CPU_FREQ
+#if defined(CONFIG_CPU_FREQ) || defined(CONFIG_PROCESSOR_EXTERNAL_CONTROL)
 	acpi_processor_ppc_has_changed(pr);
 #endif
 	acpi_processor_get_throttling_info(pr);
@@ -517,7 +518,7 @@ static int acpi_processor_get_info(struc
 	return 0;
 }
 
-static void *processor_device_array[NR_CPUS];
+static void *processor_device_array[NR_ACPI_CPUS];
 
 static int acpi_processor_start(struct acpi_device *device)
 {
@@ -525,31 +526,46 @@ static int acpi_processor_start(struct a
 	acpi_status status = AE_OK;
 	struct acpi_processor *pr;
 
+	processor_extcntl_init();
 
 	pr = acpi_driver_data(device);
 
 	result = acpi_processor_get_info(pr);
-	if (result) {
+	if (result || 
+	    ((pr->id == -1) && !processor_cntl_external())) {
 		/* Processor is physically not present */
 		return 0;
 	}
 
-	BUG_ON((pr->id >= NR_CPUS) || (pr->id < 0));
+	BUG_ON(!processor_cntl_external() &&
+	       ((pr->id >= NR_CPUS) || (pr->id < 0)));
 
 	/*
 	 * Buggy BIOS check
 	 * ACPI id of processors can be reported wrongly by the BIOS.
 	 * Don't trust it blindly
 	 */
+#ifdef CONFIG_XEN
+	BUG_ON(pr->acpi_id >= NR_ACPI_CPUS);
+	if (processor_device_array[pr->acpi_id] != NULL &&
+	    processor_device_array[pr->acpi_id] != (void *)device) {
+#else
 	if (processor_device_array[pr->id] != NULL &&
 	    processor_device_array[pr->id] != (void *)device) {
+#endif /* CONFIG_XEN */
 		printk(KERN_WARNING "BIOS reported wrong ACPI id"
 			"for the processor\n");
 		return -ENODEV;
 	}
+#ifdef CONFIG_XEN
+	processor_device_array[pr->acpi_id] = (void *)device;
+	if (pr->id != -1)
+		processors[pr->id] = pr;
+#else
 	processor_device_array[pr->id] = (void *)device;
 
 	processors[pr->id] = pr;
+#endif /* CONFIG_XEN */
 
 	result = acpi_processor_add_fs(device);
 	if (result)
@@ -564,6 +580,10 @@ static int acpi_processor_start(struct a
 
 	acpi_processor_power_init(pr, device);
 
+	result = processor_extcntl_prepare(pr);
+	if (result)
+		goto end;
+
 	if (pr->flags.throttling) {
 		printk(KERN_INFO PREFIX "%s [%s] (supports",
 		       acpi_device_name(device), acpi_device_bid(device));
@@ -656,7 +676,13 @@ static int acpi_processor_remove(struct 
 
 	acpi_processor_remove_fs(device);
 
+#ifdef CONFIG_XEN
+	if (pr->id != -1)
+		processors[pr->id] = NULL;
+#else
 	processors[pr->id] = NULL;
+#endif /* CONFIG_XEN */
+
 
 	kfree(pr);
 
@@ -710,6 +736,10 @@ int acpi_processor_device_add(acpi_handl
 	if (!pr)
 		return -ENODEV;
 
+	if (processor_cntl_external())
+		processor_notify_external(pr,
+			PROCESSOR_HOTPLUG, HOTPLUG_TYPE_ADD);
+
 	if ((pr->id >= 0) && (pr->id < NR_CPUS)) {
 		kobject_uevent(&(*device)->kobj, KOBJ_ONLINE);
 	}
@@ -748,6 +778,10 @@ acpi_processor_hotplug_notify(acpi_handl
 			break;
 		}
 
+		if (processor_cntl_external())
+			processor_notify_external(pr,
+					PROCESSOR_HOTPLUG, HOTPLUG_TYPE_ADD);
+
 		if (pr->id >= 0 && (pr->id < NR_CPUS)) {
 			kobject_uevent(&device->kobj, KOBJ_OFFLINE);
 			break;
@@ -777,8 +811,18 @@ acpi_processor_hotplug_notify(acpi_handl
 			return;
 		}
 
+#ifdef CONFIG_XEN
+		if ((pr->id >= 0) && (pr->id < NR_CPUS)
+		    && (cpu_present(pr->id)))
+#else
 		if ((pr->id < NR_CPUS) && (cpu_present(pr->id)))
+#endif /* CONFIG_XEN */
 			kobject_uevent(&device->kobj, KOBJ_OFFLINE);
+
+		if (processor_cntl_external())
+			processor_notify_external(pr, PROCESSOR_HOTPLUG,
+							HOTPLUG_TYPE_REMOVE);
+
 		break;
 	default:
 		ACPI_DEBUG_PRINT((ACPI_DB_INFO,
@@ -843,6 +887,11 @@ static acpi_status acpi_processor_hotadd
 
 static int acpi_processor_handle_eject(struct acpi_processor *pr)
 {
+#ifdef CONFIG_XEN
+	if (pr->id == -1)
+		return (0);
+#endif /* CONFIG_XEN */
+
 	if (cpu_online(pr->id)) {
 		return (-EINVAL);
 	}
diff -rpuN linux-2.6.18.8/drivers/acpi/processor_extcntl.c linux-2.6.18-xen-3.3.0/drivers/acpi/processor_extcntl.c
--- linux-2.6.18.8/drivers/acpi/processor_extcntl.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/acpi/processor_extcntl.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,232 @@
+/*
+ * processor_extcntl.c - channel to external control logic
+ *
+ *  Copyright (C) 2008, Intel corporation
+ *
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or (at
+ *  your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful, but
+ *  WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ *  General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License along
+ *  with this program; if not, write to the Free Software Foundation, Inc.,
+ *  59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
+ *
+ */
+
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/types.h>
+#include <linux/acpi.h>
+#include <linux/pm.h>
+#include <linux/cpu.h>
+
+#include <acpi/processor.h>
+
+#define ACPI_PROCESSOR_COMPONENT        0x01000000
+#define ACPI_PROCESSOR_CLASS            "processor"
+#define ACPI_PROCESSOR_DRIVER_NAME      "ACPI Processor Driver"
+#define _COMPONENT              ACPI_PROCESSOR_COMPONENT
+ACPI_MODULE_NAME("acpi_processor")
+
+static int processor_extcntl_parse_csd(struct acpi_processor *pr);
+static int processor_extcntl_get_performance(struct acpi_processor *pr);
+/*
+ * External processor control logic may register with its own set of
+ * ops to get ACPI related notification. One example is like VMM.
+ */
+const struct processor_extcntl_ops *processor_extcntl_ops;
+EXPORT_SYMBOL(processor_extcntl_ops);
+
+static int processor_notify_smm(void)
+{
+	acpi_status status;
+	static int is_done = 0;
+
+	/* only need successfully notify BIOS once */
+	/* avoid double notification which may lead to unexpected result */
+	if (is_done)
+		return 0;
+
+	/* Can't write pstate_cnt to smi_cmd if either value is zero */
+	if ((!acpi_fadt.smi_cmd) || (!acpi_fadt.pstate_cnt)) {
+		ACPI_DEBUG_PRINT((ACPI_DB_INFO,"No SMI port or pstate_cnt\n"));
+		return 0;
+	}
+
+	ACPI_DEBUG_PRINT((ACPI_DB_INFO,
+		"Writing pstate_cnt [0x%x] to smi_cmd [0x%x]\n",
+		acpi_fadt.pstate_cnt, acpi_fadt.smi_cmd));
+
+	/* FADT v1 doesn't support pstate_cnt, many BIOS vendors use
+	 * it anyway, so we need to support it... */
+	if (acpi_fadt_is_v1) {
+		ACPI_DEBUG_PRINT((ACPI_DB_INFO,
+			"Using v1.0 FADT reserved value for pstate_cnt\n"));
+	}
+
+	status = acpi_os_write_port(acpi_fadt.smi_cmd,
+				    (u32) acpi_fadt.pstate_cnt, 8);
+	if (ACPI_FAILURE(status)) 
+		return status;
+
+	is_done = 1;
+
+	return 0;
+}
+
+int processor_notify_external(struct acpi_processor *pr, int event, int type)
+{
+	int ret = -EINVAL;
+
+	if (!processor_cntl_external())
+		return -EINVAL;
+
+	switch (event) {
+	case PROCESSOR_PM_INIT:
+	case PROCESSOR_PM_CHANGE:
+		if ((type >= PM_TYPE_MAX) ||
+			!processor_extcntl_ops->pm_ops[type])
+			break;
+
+		ret = processor_extcntl_ops->pm_ops[type](pr, event);
+		break;
+	case PROCESSOR_HOTPLUG:
+		if (processor_extcntl_ops->hotplug)
+			ret = processor_extcntl_ops->hotplug(pr, type);
+		break;
+	default:
+		printk(KERN_ERR "Unsupport processor events %d.\n", event);
+		break;
+	}
+
+	return ret;
+}
+
+/*
+ * External control logic can decide to grab full or part of physical
+ * processor control bits. Take a VMM for example, physical processors
+ * are owned by VMM and thus existence information like hotplug is
+ * always required to be notified to VMM. Similar is processor idle
+ * state which is also necessarily controlled by VMM. But for other
+ * control bits like performance/throttle states, VMM may choose to
+ * control or not upon its own policy.
+ */
+void processor_extcntl_init(void)
+{
+	if (!processor_extcntl_ops)
+		arch_acpi_processor_init_extcntl(&processor_extcntl_ops);
+}
+
+/*
+ * This is called from ACPI processor init, and targeted to hold
+ * some tricky housekeeping jobs to satisfy external control model.
+ * For example, we may put dependency parse stub here for idle
+ * and performance state. Those information may be not available
+ * if splitting from dom0 control logic like cpufreq driver.
+ */
+int processor_extcntl_prepare(struct acpi_processor *pr)
+{
+	/* parse cstate dependency information */
+	if (processor_pm_external())
+		processor_extcntl_parse_csd(pr);
+
+	/* Initialize performance states */
+	if (processor_pmperf_external())
+		processor_extcntl_get_performance(pr);
+
+	return 0;
+}
+
+/*
+ * Currently no _CSD is implemented which is why existing ACPI code
+ * doesn't parse _CSD at all. But to keep interface complete with
+ * external control logic, we put a placeholder here for future
+ * compatibility.
+ */
+static int processor_extcntl_parse_csd(struct acpi_processor *pr)
+{
+	int i;
+
+	for (i = 0; i < pr->power.count; i++) {
+		if (!pr->power.states[i].valid)
+			continue;
+
+		/* No dependency by default */
+		pr->power.states[i].domain_info = NULL;
+		pr->power.states[i].csd_count = 0;
+	}
+
+	return 0;
+}
+
+/*
+ * Existing ACPI module does parse performance states at some point,
+ * when acpi-cpufreq driver is loaded which however is something
+ * we'd like to disable to avoid confliction with external control
+ * logic. So we have to collect raw performance information here 
+ * when ACPI processor object is found and started.
+ */
+static int processor_extcntl_get_performance(struct acpi_processor *pr)
+{
+	int ret;
+	struct acpi_processor_performance *perf;
+	struct acpi_psd_package *pdomain;
+
+	if (pr->performance)
+		return -EBUSY;
+
+	perf = kzalloc(sizeof(struct acpi_processor_performance), GFP_KERNEL);
+	if (!perf)
+		return -ENOMEM;
+
+	pr->performance = perf;
+	/* Get basic performance state information */
+	ret = acpi_processor_get_performance_info(pr);
+	if (ret < 0)
+		goto err_out;
+
+	/*
+	 * Well, here we need retrieve performance dependency information
+	 * from _PSD object. The reason why existing interface is not used
+	 * is due to the reason that existing interface sticks to Linux cpu
+	 * id to construct some bitmap, however we want to split ACPI 
+	 * processor objects from Linux cpu id logic. For example, even
+	 * when Linux is configured as UP, we still want to parse all ACPI
+	 * processor objects to external logic. In this case, it's preferred
+	 * to use ACPI ID instead.
+	 */
+	pr->performance->domain_info.num_processors = 0;
+	ret = acpi_processor_get_psd(pr);
+	if (ret < 0)
+		goto err_out;
+
+	/* Some sanity check */
+	pdomain = &pr->performance->domain_info;
+	if ((pdomain->revision != ACPI_PSD_REV0_REVISION) ||
+	    (pdomain->num_entries != ACPI_PSD_REV0_ENTRIES) ||
+	    ((pdomain->coord_type != DOMAIN_COORD_TYPE_SW_ALL) &&
+	     (pdomain->coord_type != DOMAIN_COORD_TYPE_SW_ANY) &&
+	     (pdomain->coord_type != DOMAIN_COORD_TYPE_HW_ALL))) {
+		ret = -EINVAL;
+		goto err_out;
+	}
+
+	/* Last step is to notify BIOS that external logic exists */
+	processor_notify_smm();
+
+	processor_notify_external(pr, PROCESSOR_PM_INIT, PM_TYPE_PERF);
+
+	return 0;
+err_out:
+	pr->performance = NULL;
+	kfree(perf);
+	return ret;
+}
diff -rpuN linux-2.6.18.8/drivers/acpi/processor_idle.c linux-2.6.18-xen-3.3.0/drivers/acpi/processor_idle.c
--- linux-2.6.18.8/drivers/acpi/processor_idle.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/acpi/processor_idle.c	2008-08-21 11:36:07.000000000 +0200
@@ -714,8 +714,12 @@ static int acpi_processor_get_power_info
 		    (reg->space_id != ACPI_ADR_SPACE_FIXED_HARDWARE))
 			continue;
 
-		cx.address = (reg->space_id == ACPI_ADR_SPACE_FIXED_HARDWARE) ?
-		    0 : reg->address;
+		if (!processor_pm_external())
+			cx.address = (reg->space_id ==
+				      ACPI_ADR_SPACE_FIXED_HARDWARE) ?
+		    		      0 : reg->address;
+		else
+			cx.address = reg->address;
 
 		/* There should be an easy way to extract an integer... */
 		obj = (union acpi_object *)&(element->package.elements[1]);
@@ -724,7 +728,9 @@ static int acpi_processor_get_power_info
 
 		cx.type = obj->integer.value;
 
-		if ((cx.type != ACPI_STATE_C1) &&
+		/* Following check doesn't apply to external control case */
+		if (!processor_pm_external() &&
+		    (cx.type != ACPI_STATE_C1) &&
 		    (reg->space_id != ACPI_ADR_SPACE_SYSTEM_IO))
 			continue;
 
@@ -743,6 +749,12 @@ static int acpi_processor_get_power_info
 
 		cx.power = obj->integer.value;
 
+#ifdef CONFIG_PROCESSOR_EXTERNAL_CONTROL
+		/* cache control methods to notify external logic */
+		if (processor_pm_external())
+			memcpy(&cx.reg, reg, sizeof(*reg));
+#endif
+
 		current_count++;
 		memcpy(&(pr->power.states[current_count]), &cx, sizeof(cx));
 
@@ -985,12 +997,16 @@ int acpi_processor_cst_has_changed(struc
 		return -ENODEV;
 
 	/* Fall back to the default idle loop */
-	pm_idle = pm_idle_save;
+	if (!processor_pm_external())
+		pm_idle = pm_idle_save;
 	synchronize_sched();	/* Relies on interrupts forcing exit from idle. */
 
 	pr->flags.power = 0;
 	result = acpi_processor_get_power_info(pr);
-	if ((pr->flags.power == 1) && (pr->flags.power_setup_done))
+	if (processor_pm_external())
+		processor_notify_external(pr,
+			PROCESSOR_PM_CHANGE, PM_TYPE_IDLE);
+	else if ((pr->flags.power == 1) && (pr->flags.power_setup_done))
 		pm_idle = acpi_processor_idle;
 
 	return result;
@@ -1122,7 +1138,7 @@ int acpi_processor_power_init(struct acp
 				       pr->power.states[i].type);
 		printk(")\n");
 
-		if (pr->id == 0) {
+		if (!processor_pm_external() && (pr->id == 0)) {
 			pm_idle_save = pm_idle;
 			pm_idle = acpi_processor_idle;
 		}
@@ -1141,6 +1157,9 @@ int acpi_processor_power_init(struct acp
 
 	pr->flags.power_setup_done = 1;
 
+	if (processor_pm_external())
+		processor_notify_external(pr,
+			PROCESSOR_PM_INIT, PM_TYPE_IDLE);
 	return 0;
 }
 
diff -rpuN linux-2.6.18.8/drivers/acpi/processor_perflib.c linux-2.6.18-xen-3.3.0/drivers/acpi/processor_perflib.c
--- linux-2.6.18.8/drivers/acpi/processor_perflib.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/acpi/processor_perflib.c	2008-08-21 11:36:07.000000000 +0200
@@ -66,6 +66,7 @@ static DEFINE_MUTEX(performance_mutex);
 
 static int acpi_processor_ppc_status = 0;
 
+#ifdef CONFIG_CPU_FREQ
 static int acpi_processor_ppc_notifier(struct notifier_block *nb,
 				       unsigned long event, void *data)
 {
@@ -102,6 +103,7 @@ static int acpi_processor_ppc_notifier(s
 static struct notifier_block acpi_ppc_notifier_block = {
 	.notifier_call = acpi_processor_ppc_notifier,
 };
+#endif	/* CONFIG_CPU_FREQ */
 
 static int acpi_processor_get_platform_limit(struct acpi_processor *pr)
 {
@@ -137,9 +139,15 @@ int acpi_processor_ppc_has_changed(struc
 	if (ret < 0)
 		return (ret);
 	else
+#ifdef CONFIG_CPU_FREQ
 		return cpufreq_update_policy(pr->id);
+#elif CONFIG_PROCESSOR_EXTERNAL_CONTROL
+		return processor_notify_external(pr,
+				PROCESSOR_PM_CHANGE, PM_TYPE_PERF);
+#endif
 }
 
+#ifdef CONFIG_CPU_FREQ
 void acpi_processor_ppc_init(void)
 {
 	if (!cpufreq_register_notifier
@@ -158,6 +166,7 @@ void acpi_processor_ppc_exit(void)
 
 	acpi_processor_ppc_status &= ~PPC_REGISTERED;
 }
+#endif	/* CONFIG_CPU_FREQ */
 
 static int acpi_processor_get_performance_control(struct acpi_processor *pr)
 {
@@ -299,7 +308,10 @@ static int acpi_processor_get_performanc
 	return result;
 }
 
-static int acpi_processor_get_performance_info(struct acpi_processor *pr)
+#ifndef CONFIG_PROCESSOR_EXTERNAL_CONTROL
+static
+#endif
+int acpi_processor_get_performance_info(struct acpi_processor *pr)
 {
 	int result = 0;
 	acpi_status status = AE_OK;
@@ -331,6 +343,7 @@ static int acpi_processor_get_performanc
 	return 0;
 }
 
+#ifdef CONFIG_CPU_FREQ
 int acpi_processor_notify_smm(struct module *calling_module)
 {
 	acpi_status status;
@@ -398,6 +411,7 @@ int acpi_processor_notify_smm(struct mod
 }
 
 EXPORT_SYMBOL(acpi_processor_notify_smm);
+#endif	/* CONFIG_CPU_FREQ */
 
 #ifdef CONFIG_X86_ACPI_CPUFREQ_PROC_INTF
 /* /proc/acpi/processor/../performance interface (DEPRECATED) */
@@ -538,7 +552,10 @@ static void acpi_cpufreq_remove_file(str
 }
 #endif				/* CONFIG_X86_ACPI_CPUFREQ_PROC_INTF */
 
-static int acpi_processor_get_psd(struct acpi_processor	*pr)
+#ifndef CONFIG_PROCESSOR_EXTERNAL_CONTROL
+static
+#endif
+int acpi_processor_get_psd(struct acpi_processor *pr)
 {
 	int result = 0;
 	acpi_status status = AE_OK;
diff -rpuN linux-2.6.18.8/drivers/acpi/resources/rsxface.c linux-2.6.18-xen-3.3.0/drivers/acpi/resources/rsxface.c
--- linux-2.6.18.8/drivers/acpi/resources/rsxface.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/acpi/resources/rsxface.c	2008-08-21 11:36:07.000000000 +0200
@@ -476,8 +476,6 @@ acpi_rs_match_vendor_resource(struct acp
 	return (AE_CTRL_TERMINATE);
 }
 
-ACPI_EXPORT_SYMBOL(acpi_rs_match_vendor_resource)
-
 /*******************************************************************************
  *
  * FUNCTION:    acpi_walk_resources
diff -rpuN linux-2.6.18.8/drivers/acpi/sleep/main.c linux-2.6.18-xen-3.3.0/drivers/acpi/sleep/main.c
--- linux-2.6.18.8/drivers/acpi/sleep/main.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/acpi/sleep/main.c	2008-08-21 11:36:07.000000000 +0200
@@ -91,7 +91,14 @@ static int acpi_pm_enter(suspend_state_t
 		break;
 
 	case PM_SUSPEND_MEM:
+#ifdef CONFIG_ACPI_PV_SLEEP
+		/* Hyperviosr will save and restore CPU context
+		 * and then we can skip low level housekeeping here.
+		 */
+		acpi_enter_sleep_state(acpi_state);
+#else
 		do_suspend_lowlevel();
+#endif
 		break;
 
 	case PM_SUSPEND_DISK:
@@ -145,10 +152,12 @@ static int acpi_pm_finish(suspend_state_
 	/* reset firmware waking vector */
 	acpi_set_firmware_waking_vector((acpi_physical_address) 0);
 
+#ifndef CONFIG_ACPI_PV_SLEEP
 	if (init_8259A_after_S1) {
 		printk("Broken toshiba laptop -> kicking interrupts\n");
 		init_8259A(0);
 	}
+#endif
 	return 0;
 }
 
diff -rpuN linux-2.6.18.8/drivers/acpi/sleep/poweroff.c linux-2.6.18-xen-3.3.0/drivers/acpi/sleep/poweroff.c
--- linux-2.6.18.8/drivers/acpi/sleep/poweroff.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/acpi/sleep/poweroff.c	2008-08-21 11:36:07.000000000 +0200
@@ -20,6 +20,7 @@
 int acpi_sleep_prepare(u32 acpi_state)
 {
 #ifdef CONFIG_ACPI_SLEEP
+#ifndef CONFIG_ACPI_PV_SLEEP
 	/* do we have a wakeup address for S2 and S3? */
 	if (acpi_state == ACPI_STATE_S3) {
 		if (!acpi_wakeup_address) {
@@ -30,6 +31,7 @@ int acpi_sleep_prepare(u32 acpi_state)
 							     acpi_wakeup_address));
 
 	}
+#endif
 	ACPI_FLUSH_CPU_CACHE();
 	acpi_enable_wakeup_device_prep(acpi_state);
 #endif
diff -rpuN linux-2.6.18.8/drivers/acpi/sleep/proc.c linux-2.6.18-xen-3.3.0/drivers/acpi/sleep/proc.c
--- linux-2.6.18.8/drivers/acpi/sleep/proc.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/acpi/sleep/proc.c	2008-08-21 11:36:07.000000000 +0200
@@ -251,6 +251,8 @@ acpi_system_write_alarm(struct file *fil
 
 	if (adjust) {
 		yr += CMOS_READ(RTC_YEAR);
+		if (acpi_gbl_FADT->century)
+			yr += CMOS_READ(acpi_gbl_FADT->century) * 100;
 		mo += CMOS_READ(RTC_MONTH);
 		day += CMOS_READ(RTC_DAY_OF_MONTH);
 		hr += CMOS_READ(RTC_HOURS);
diff -rpuN linux-2.6.18.8/drivers/block/floppy.c linux-2.6.18-xen-3.3.0/drivers/block/floppy.c
--- linux-2.6.18.8/drivers/block/floppy.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/block/floppy.c	2008-08-21 11:36:07.000000000 +0200
@@ -4392,11 +4392,15 @@ static int floppy_grab_irq_and_dma(void)
 	if (fd_request_dma()) {
 		DPRINT("Unable to grab DMA%d for the floppy driver\n",
 		       FLOPPY_DMA);
-		fd_free_irq();
-		spin_lock_irqsave(&floppy_usage_lock, flags);
-		usage_count--;
-		spin_unlock_irqrestore(&floppy_usage_lock, flags);
-		return -1;
+		if (can_use_virtual_dma & 2)
+			use_virtual_dma = can_use_virtual_dma = 1;
+		if (!(can_use_virtual_dma & 1)) {
+			fd_free_irq();
+			spin_lock_irqsave(&floppy_usage_lock, flags);
+			usage_count--;
+			spin_unlock_irqrestore(&floppy_usage_lock, flags);
+			return -1;
+		}
 	}
 
 	for (fdc = 0; fdc < N_FDC; fdc++) {
diff -rpuN linux-2.6.18.8/drivers/char/agp/amd64-agp.c linux-2.6.18-xen-3.3.0/drivers/char/agp/amd64-agp.c
--- linux-2.6.18.8/drivers/char/agp/amd64-agp.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/char/agp/amd64-agp.c	2008-08-21 11:36:07.000000000 +0200
@@ -15,6 +15,7 @@
 #include <linux/mmzone.h>
 #include <asm/page.h>		/* PAGE_SIZE */
 #include <asm/k8.h>
+#include <asm/e820.h>
 #include "agp.h"
 
 /* PTE bits. */
@@ -252,7 +253,6 @@ static struct agp_bridge_driver amd_8151
 /* Some basic sanity checks for the aperture. */
 static int __devinit aperture_valid(u64 aper, u32 size)
 {
-	u32 pfn, c;
 	if (aper == 0) {
 		printk(KERN_ERR PFX "No aperture\n");
 		return 0;
@@ -265,14 +265,9 @@ static int __devinit aperture_valid(u64 
 		printk(KERN_ERR PFX "Aperture out of bounds\n");
 		return 0;
 	}
-	pfn = aper >> PAGE_SHIFT;
-	for (c = 0; c < size/PAGE_SIZE; c++) {
-		if (!pfn_valid(pfn + c))
-			break;
-		if (!PageReserved(pfn_to_page(pfn + c))) {
-			printk(KERN_ERR PFX "Aperture pointing to RAM\n");
-			return 0;
-		}
+	if (e820_any_mapped(aper, aper + size, E820_RAM)) {
+		printk(KERN_ERR PFX "Aperture pointing to RAM\n");
+		return 0;
 	}
 
 	/* Request the Aperture. This catches cases when someone else
diff -rpuN linux-2.6.18.8/drivers/char/agp/generic.c linux-2.6.18-xen-3.3.0/drivers/char/agp/generic.c
--- linux-2.6.18.8/drivers/char/agp/generic.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/char/agp/generic.c	2008-08-21 11:36:07.000000000 +0200
@@ -51,28 +51,6 @@ int agp_memory_reserved;
  */
 EXPORT_SYMBOL_GPL(agp_memory_reserved);
 
-#if defined(CONFIG_X86)
-int map_page_into_agp(struct page *page)
-{
-	int i;
-	i = change_page_attr(page, 1, PAGE_KERNEL_NOCACHE);
-	/* Caller's responsibility to call global_flush_tlb() for
-	 * performance reasons */
-	return i;
-}
-EXPORT_SYMBOL_GPL(map_page_into_agp);
-
-int unmap_page_from_agp(struct page *page)
-{
-	int i;
-	i = change_page_attr(page, 1, PAGE_KERNEL);
-	/* Caller's responsibility to call global_flush_tlb() for
-	 * performance reasons */
-	return i;
-}
-EXPORT_SYMBOL_GPL(unmap_page_from_agp);
-#endif
-
 /*
  * Generic routines for handling agp_memory structures -
  * They use the basic page allocation routines to do the brunt of the work.
diff -rpuN linux-2.6.18.8/drivers/char/agp/intel-agp.c linux-2.6.18-xen-3.3.0/drivers/char/agp/intel-agp.c
--- linux-2.6.18.8/drivers/char/agp/intel-agp.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/char/agp/intel-agp.c	2008-08-21 11:36:07.000000000 +0200
@@ -164,9 +164,17 @@ static void *i8xx_alloc_pages(void)
 	if (page == NULL)
 		return NULL;
 
+#ifdef CONFIG_XEN
+	if (xen_create_contiguous_region((unsigned long)page_address(page), 2, 32)) {
+		__free_pages(page, 2);
+		return NULL;
+	}
+#endif
+
 	if (change_page_attr(page, 4, PAGE_KERNEL_NOCACHE) < 0) {
+		change_page_attr(page, 4, PAGE_KERNEL);
 		global_flush_tlb();
-		__free_page(page);
+		__free_pages(page, 2);
 		return NULL;
 	}
 	global_flush_tlb();
@@ -186,9 +194,12 @@ static void i8xx_destroy_pages(void *add
 	page = virt_to_page(addr);
 	change_page_attr(page, 4, PAGE_KERNEL);
 	global_flush_tlb();
+#ifdef CONFIG_XEN
+	xen_destroy_contiguous_region((unsigned long)page_address(page), 2);
+#endif
 	put_page(page);
 	unlock_page(page);
-	free_pages((unsigned long)addr, 2);
+	__free_pages(page, 2);
 	atomic_dec(&agp_bridge->current_memory_agp);
 }
 
diff -rpuN linux-2.6.18.8/drivers/char/Kconfig linux-2.6.18-xen-3.3.0/drivers/char/Kconfig
--- linux-2.6.18.8/drivers/char/Kconfig	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/char/Kconfig	2008-08-21 11:36:07.000000000 +0200
@@ -1005,7 +1005,7 @@ config MAX_RAW_DEVS
 config HPET
 	bool "HPET - High Precision Event Timer" if (X86 || IA64)
 	default n
-	depends on ACPI
+	depends on ACPI && !XEN
 	help
 	  If you say Y here, you will have a miscdevice named "/dev/hpet/".  Each
 	  open selects one of the timers supported by the HPET.  The timers are
diff -rpuN linux-2.6.18.8/drivers/char/mem.c linux-2.6.18-xen-3.3.0/drivers/char/mem.c
--- linux-2.6.18.8/drivers/char/mem.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/char/mem.c	2008-08-21 11:36:07.000000000 +0200
@@ -101,6 +101,7 @@ static inline int valid_mmap_phys_addr_r
 }
 #endif
 
+#ifndef ARCH_HAS_DEV_MEM
 /*
  * This funcion reads the *physical* memory. The f_pos points directly to the 
  * memory location. 
@@ -223,6 +224,7 @@ static ssize_t write_mem(struct file * f
 	*ppos += written;
 	return written;
 }
+#endif
 
 #ifndef __HAVE_PHYS_MEM_ACCESS_PROT
 static pgprot_t phys_mem_access_prot(struct file *file, unsigned long pfn,
@@ -262,6 +264,9 @@ static int mmap_mem(struct file * file, 
 static int mmap_kmem(struct file * file, struct vm_area_struct * vma)
 {
 	unsigned long pfn;
+#ifdef CONFIG_XEN
+	unsigned long i, count;
+#endif
 
 	/* Turn a kernel-virtual address into a physical page frame */
 	pfn = __pa((u64)vma->vm_pgoff << PAGE_SHIFT) >> PAGE_SHIFT;
@@ -276,6 +281,13 @@ static int mmap_kmem(struct file * file,
 	if (!pfn_valid(pfn))
 		return -EIO;
 
+#ifdef CONFIG_XEN
+	count = (vma->vm_end - vma->vm_start) >> PAGE_SHIFT;
+	for (i = 0; i < count; i++)
+		if ((pfn + i) != mfn_to_local_pfn(pfn_to_mfn(pfn + i)))
+			return -EIO;
+#endif
+
 	vma->vm_pgoff = pfn;
 	return mmap_mem(file, vma);
 }
@@ -780,6 +792,7 @@ static int open_port(struct inode * inod
 #define open_kmem	open_mem
 #define open_oldmem	open_mem
 
+#ifndef ARCH_HAS_DEV_MEM
 static const struct file_operations mem_fops = {
 	.llseek		= memory_lseek,
 	.read		= read_mem,
@@ -787,6 +800,9 @@ static const struct file_operations mem_
 	.mmap		= mmap_mem,
 	.open		= open_mem,
 };
+#else
+extern const struct file_operations mem_fops;
+#endif
 
 static const struct file_operations kmem_fops = {
 	.llseek		= memory_lseek,
diff -rpuN linux-2.6.18.8/drivers/char/tpm/Kconfig linux-2.6.18-xen-3.3.0/drivers/char/tpm/Kconfig
--- linux-2.6.18.8/drivers/char/tpm/Kconfig	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/char/tpm/Kconfig	2008-08-21 11:36:07.000000000 +0200
@@ -58,5 +58,13 @@ config TCG_INFINEON
 	  Further information on this driver and the supported hardware
 	  can be found at http://www.prosec.rub.de/tpm
 
-endmenu
+config TCG_XEN
+	tristate "XEN TPM Interface"
+	depends on TCG_TPM && XEN
+	---help---
+	  If you want to make TPM support available to a Xen user domain,
+	  say Yes and it will be accessible from within Linux.
+	  To compile this driver as a module, choose M here; the module
+	  will be called tpm_xenu.
 
+endmenu
diff -rpuN linux-2.6.18.8/drivers/char/tpm/Makefile linux-2.6.18-xen-3.3.0/drivers/char/tpm/Makefile
--- linux-2.6.18.8/drivers/char/tpm/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/char/tpm/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -9,3 +9,5 @@ obj-$(CONFIG_TCG_TIS) += tpm_tis.o
 obj-$(CONFIG_TCG_NSC) += tpm_nsc.o
 obj-$(CONFIG_TCG_ATMEL) += tpm_atmel.o
 obj-$(CONFIG_TCG_INFINEON) += tpm_infineon.o
+obj-$(CONFIG_TCG_XEN) += tpm_xenu.o
+tpm_xenu-y = tpm_xen.o tpm_vtpm.o
diff -rpuN linux-2.6.18.8/drivers/char/tpm/tpm.h linux-2.6.18-xen-3.3.0/drivers/char/tpm/tpm.h
--- linux-2.6.18.8/drivers/char/tpm/tpm.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/char/tpm/tpm.h	2008-08-21 11:36:07.000000000 +0200
@@ -105,6 +105,9 @@ struct tpm_chip {
 	struct dentry **bios_dir;
 
 	struct list_head list;
+#ifdef CONFIG_XEN
+	void *priv;
+#endif
 };
 
 #define to_tpm_chip(n) container_of(n, struct tpm_chip, vendor)
@@ -121,6 +124,18 @@ static inline void tpm_write_index(int b
 	outb(value & 0xFF, base+1);
 }
 
+#ifdef CONFIG_XEN
+static inline void *chip_get_private(const struct tpm_chip *chip)
+{
+	return chip->priv;
+}
+
+static inline void chip_set_private(struct tpm_chip *chip, void *priv)
+{
+	chip->priv = priv;
+}
+#endif
+
 extern void tpm_get_timeouts(struct tpm_chip *);
 extern void tpm_gen_interrupt(struct tpm_chip *);
 extern void tpm_continue_selftest(struct tpm_chip *);
diff -rpuN linux-2.6.18.8/drivers/char/tpm/tpm_vtpm.c linux-2.6.18-xen-3.3.0/drivers/char/tpm/tpm_vtpm.c
--- linux-2.6.18.8/drivers/char/tpm/tpm_vtpm.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/char/tpm/tpm_vtpm.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,542 @@
+/*
+ * Copyright (C) 2006 IBM Corporation
+ *
+ * Authors:
+ * Stefan Berger <stefanb@us.ibm.com>
+ *
+ * Generic device driver part for device drivers in a virtualized
+ * environment.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation, version 2 of the
+ * License.
+ *
+ */
+
+#include <asm/uaccess.h>
+#include <linux/list.h>
+#include <linux/device.h>
+#include <linux/interrupt.h>
+#include <linux/platform_device.h>
+#include "tpm.h"
+#include "tpm_vtpm.h"
+
+/* read status bits */
+enum {
+	STATUS_BUSY = 0x01,
+	STATUS_DATA_AVAIL = 0x02,
+	STATUS_READY = 0x04
+};
+
+struct transmission {
+	struct list_head next;
+
+	unsigned char *request;
+	size_t  request_len;
+	size_t  request_buflen;
+
+	unsigned char *response;
+	size_t  response_len;
+	size_t  response_buflen;
+
+	unsigned int flags;
+};
+
+enum {
+	TRANSMISSION_FLAG_WAS_QUEUED = 0x1
+};
+
+
+enum {
+	DATAEX_FLAG_QUEUED_ONLY = 0x1
+};
+
+
+/* local variables */
+
+/* local function prototypes */
+static int _vtpm_send_queued(struct tpm_chip *chip);
+
+
+/* =============================================================
+ * Some utility functions
+ * =============================================================
+ */
+static void vtpm_state_init(struct vtpm_state *vtpms)
+{
+	vtpms->current_request = NULL;
+	spin_lock_init(&vtpms->req_list_lock);
+	init_waitqueue_head(&vtpms->req_wait_queue);
+	INIT_LIST_HEAD(&vtpms->queued_requests);
+
+	vtpms->current_response = NULL;
+	spin_lock_init(&vtpms->resp_list_lock);
+	init_waitqueue_head(&vtpms->resp_wait_queue);
+
+	vtpms->disconnect_time = jiffies;
+}
+
+
+static inline struct transmission *transmission_alloc(void)
+{
+	return kzalloc(sizeof(struct transmission), GFP_ATOMIC);
+}
+
+static unsigned char *
+transmission_set_req_buffer(struct transmission *t,
+                            unsigned char *buffer, size_t len)
+{
+	if (t->request_buflen < len) {
+		kfree(t->request);
+		t->request = kmalloc(len, GFP_KERNEL);
+		if (!t->request) {
+			t->request_buflen = 0;
+			return NULL;
+		}
+		t->request_buflen = len;
+	}
+
+	memcpy(t->request, buffer, len);
+	t->request_len = len;
+
+	return t->request;
+}
+
+static unsigned char *
+transmission_set_res_buffer(struct transmission *t,
+                            const unsigned char *buffer, size_t len)
+{
+	if (t->response_buflen < len) {
+		kfree(t->response);
+		t->response = kmalloc(len, GFP_ATOMIC);
+		if (!t->response) {
+			t->response_buflen = 0;
+			return NULL;
+		}
+		t->response_buflen = len;
+	}
+
+	memcpy(t->response, buffer, len);
+	t->response_len = len;
+
+	return t->response;
+}
+
+static inline void transmission_free(struct transmission *t)
+{
+	kfree(t->request);
+	kfree(t->response);
+	kfree(t);
+}
+
+/* =============================================================
+ * Interface with the lower layer driver
+ * =============================================================
+ */
+/*
+ * Lower layer uses this function to make a response available.
+ */
+int vtpm_vd_recv(const struct tpm_chip *chip,
+                 const unsigned char *buffer, size_t count,
+                 void *ptr)
+{
+	unsigned long flags;
+	int ret_size = 0;
+	struct transmission *t;
+	struct vtpm_state *vtpms;
+
+	vtpms = (struct vtpm_state *)chip_get_private(chip);
+
+	/*
+	 * The list with requests must contain one request
+	 * only and the element there must be the one that
+	 * was passed to me from the front-end.
+	 */
+	spin_lock_irqsave(&vtpms->resp_list_lock, flags);
+	if (vtpms->current_request != ptr) {
+		spin_unlock_irqrestore(&vtpms->resp_list_lock, flags);
+		return 0;
+	}
+
+	if ((t = vtpms->current_request)) {
+		transmission_free(t);
+		vtpms->current_request = NULL;
+	}
+
+	t = transmission_alloc();
+	if (t) {
+		if (!transmission_set_res_buffer(t, buffer, count)) {
+			transmission_free(t);
+			spin_unlock_irqrestore(&vtpms->resp_list_lock, flags);
+			return -ENOMEM;
+		}
+		ret_size = count;
+		vtpms->current_response = t;
+		wake_up_interruptible(&vtpms->resp_wait_queue);
+	}
+	spin_unlock_irqrestore(&vtpms->resp_list_lock, flags);
+
+	return ret_size;
+}
+
+
+/*
+ * Lower layer indicates its status (connected/disconnected)
+ */
+void vtpm_vd_status(const struct tpm_chip *chip, u8 vd_status)
+{
+	struct vtpm_state *vtpms;
+
+	vtpms = (struct vtpm_state *)chip_get_private(chip);
+
+	vtpms->vd_status = vd_status;
+	if ((vtpms->vd_status & TPM_VD_STATUS_CONNECTED) == 0) {
+		vtpms->disconnect_time = jiffies;
+	}
+}
+
+/* =============================================================
+ * Interface with the generic TPM driver
+ * =============================================================
+ */
+static int vtpm_recv(struct tpm_chip *chip, u8 *buf, size_t count)
+{
+	int rc = 0;
+	unsigned long flags;
+	struct vtpm_state *vtpms;
+
+	vtpms = (struct vtpm_state *)chip_get_private(chip);
+
+	/*
+	 * Check if the previous operation only queued the command
+	 * In this case there won't be a response, so I just
+	 * return from here and reset that flag. In any other
+	 * case I should receive a response from the back-end.
+	 */
+	spin_lock_irqsave(&vtpms->resp_list_lock, flags);
+	if ((vtpms->flags & DATAEX_FLAG_QUEUED_ONLY) != 0) {
+		vtpms->flags &= ~DATAEX_FLAG_QUEUED_ONLY;
+		spin_unlock_irqrestore(&vtpms->resp_list_lock, flags);
+		/*
+		 * The first few commands (measurements) must be
+		 * queued since it might not be possible to talk to the
+		 * TPM, yet.
+		 * Return a response of up to 30 '0's.
+		 */
+
+		count = min_t(size_t, count, 30);
+		memset(buf, 0x0, count);
+		return count;
+	}
+	/*
+	 * Check whether something is in the responselist and if
+	 * there's nothing in the list wait for something to appear.
+	 */
+
+	if (!vtpms->current_response) {
+		spin_unlock_irqrestore(&vtpms->resp_list_lock, flags);
+		interruptible_sleep_on_timeout(&vtpms->resp_wait_queue,
+		                               1000);
+		spin_lock_irqsave(&vtpms->resp_list_lock ,flags);
+	}
+
+	if (vtpms->current_response) {
+		struct transmission *t = vtpms->current_response;
+		vtpms->current_response = NULL;
+		rc = min(count, t->response_len);
+		memcpy(buf, t->response, rc);
+		transmission_free(t);
+	}
+
+	spin_unlock_irqrestore(&vtpms->resp_list_lock, flags);
+	return rc;
+}
+
+static int vtpm_send(struct tpm_chip *chip, u8 *buf, size_t count)
+{
+	int rc = 0;
+	unsigned long flags;
+	struct transmission *t = transmission_alloc();
+	struct vtpm_state *vtpms;
+
+	vtpms = (struct vtpm_state *)chip_get_private(chip);
+
+	if (!t)
+		return -ENOMEM;
+	/*
+	 * If there's a current request, it must be the
+	 * previous request that has timed out.
+	 */
+	spin_lock_irqsave(&vtpms->req_list_lock, flags);
+	if (vtpms->current_request != NULL) {
+		printk("WARNING: Sending although there is a request outstanding.\n"
+		       "         Previous request must have timed out.\n");
+		transmission_free(vtpms->current_request);
+		vtpms->current_request = NULL;
+	}
+	spin_unlock_irqrestore(&vtpms->req_list_lock, flags);
+
+	/*
+	 * Queue the packet if the driver below is not
+	 * ready, yet, or there is any packet already
+	 * in the queue.
+	 * If the driver below is ready, unqueue all
+	 * packets first before sending our current
+	 * packet.
+	 * For each unqueued packet, except for the
+	 * last (=current) packet, call the function
+	 * tpm_xen_recv to wait for the response to come
+	 * back.
+	 */
+	if ((vtpms->vd_status & TPM_VD_STATUS_CONNECTED) == 0) {
+		if (time_after(jiffies,
+		               vtpms->disconnect_time + HZ * 10)) {
+			rc = -ENOENT;
+		} else {
+			goto queue_it;
+		}
+	} else {
+		/*
+		 * Send all queued packets.
+		 */
+		if (_vtpm_send_queued(chip) == 0) {
+
+			vtpms->current_request = t;
+
+			rc = vtpm_vd_send(vtpms->tpm_private,
+			                  buf,
+			                  count,
+			                  t);
+			/*
+			 * The generic TPM driver will call
+			 * the function to receive the response.
+			 */
+			if (rc < 0) {
+				vtpms->current_request = NULL;
+				goto queue_it;
+			}
+		} else {
+queue_it:
+			if (!transmission_set_req_buffer(t, buf, count)) {
+				transmission_free(t);
+				rc = -ENOMEM;
+				goto exit;
+			}
+			/*
+			 * An error occurred. Don't event try
+			 * to send the current request. Just
+			 * queue it.
+			 */
+			spin_lock_irqsave(&vtpms->req_list_lock, flags);
+			vtpms->flags |= DATAEX_FLAG_QUEUED_ONLY;
+			list_add_tail(&t->next, &vtpms->queued_requests);
+			spin_unlock_irqrestore(&vtpms->req_list_lock, flags);
+		}
+	}
+
+exit:
+	return rc;
+}
+
+
+/*
+ * Send all queued requests.
+ */
+static int _vtpm_send_queued(struct tpm_chip *chip)
+{
+	int rc;
+	int error = 0;
+	long flags;
+	unsigned char buffer[1];
+	struct vtpm_state *vtpms;
+	vtpms = (struct vtpm_state *)chip_get_private(chip);
+
+	spin_lock_irqsave(&vtpms->req_list_lock, flags);
+
+	while (!list_empty(&vtpms->queued_requests)) {
+		/*
+		 * Need to dequeue them.
+		 * Read the result into a dummy buffer.
+		 */
+		struct transmission *qt = (struct transmission *)
+		                          vtpms->queued_requests.next;
+		list_del(&qt->next);
+		vtpms->current_request = qt;
+		spin_unlock_irqrestore(&vtpms->req_list_lock, flags);
+
+		rc = vtpm_vd_send(vtpms->tpm_private,
+		                  qt->request,
+		                  qt->request_len,
+		                  qt);
+
+		if (rc < 0) {
+			spin_lock_irqsave(&vtpms->req_list_lock, flags);
+			if ((qt = vtpms->current_request) != NULL) {
+				/*
+				 * requeue it at the beginning
+				 * of the list
+				 */
+				list_add(&qt->next,
+				         &vtpms->queued_requests);
+			}
+			vtpms->current_request = NULL;
+			error = 1;
+			break;
+		}
+		/*
+		 * After this point qt is not valid anymore!
+		 * It is freed when the front-end is delivering
+		 * the data by calling tpm_recv
+		 */
+		/*
+		 * Receive response into provided dummy buffer
+		 */
+		rc = vtpm_recv(chip, buffer, sizeof(buffer));
+		spin_lock_irqsave(&vtpms->req_list_lock, flags);
+	}
+
+	spin_unlock_irqrestore(&vtpms->req_list_lock, flags);
+
+	return error;
+}
+
+static void vtpm_cancel(struct tpm_chip *chip)
+{
+	unsigned long flags;
+	struct vtpm_state *vtpms = (struct vtpm_state *)chip_get_private(chip);
+
+	spin_lock_irqsave(&vtpms->resp_list_lock,flags);
+
+	if (!vtpms->current_response && vtpms->current_request) {
+		spin_unlock_irqrestore(&vtpms->resp_list_lock, flags);
+		interruptible_sleep_on(&vtpms->resp_wait_queue);
+		spin_lock_irqsave(&vtpms->resp_list_lock,flags);
+	}
+
+	if (vtpms->current_response) {
+		struct transmission *t = vtpms->current_response;
+		vtpms->current_response = NULL;
+		transmission_free(t);
+	}
+
+	spin_unlock_irqrestore(&vtpms->resp_list_lock,flags);
+}
+
+static u8 vtpm_status(struct tpm_chip *chip)
+{
+	u8 rc = 0;
+	unsigned long flags;
+	struct vtpm_state *vtpms;
+
+	vtpms = (struct vtpm_state *)chip_get_private(chip);
+
+	spin_lock_irqsave(&vtpms->resp_list_lock, flags);
+	/*
+	 * Data are available if:
+	 *  - there's a current response
+	 *  - the last packet was queued only (this is fake, but necessary to
+	 *      get the generic TPM layer to call the receive function.)
+	 */
+	if (vtpms->current_response ||
+	    0 != (vtpms->flags & DATAEX_FLAG_QUEUED_ONLY)) {
+		rc = STATUS_DATA_AVAIL;
+	} else if (!vtpms->current_response && !vtpms->current_request) {
+		rc = STATUS_READY;
+	}
+
+	spin_unlock_irqrestore(&vtpms->resp_list_lock, flags);
+	return rc;
+}
+
+static struct file_operations vtpm_ops = {
+	.owner = THIS_MODULE,
+	.llseek = no_llseek,
+	.open = tpm_open,
+	.read = tpm_read,
+	.write = tpm_write,
+	.release = tpm_release,
+};
+
+static DEVICE_ATTR(pubek, S_IRUGO, tpm_show_pubek, NULL);
+static DEVICE_ATTR(pcrs, S_IRUGO, tpm_show_pcrs, NULL);
+static DEVICE_ATTR(enabled, S_IRUGO, tpm_show_enabled, NULL);
+static DEVICE_ATTR(active, S_IRUGO, tpm_show_active, NULL);
+static DEVICE_ATTR(owned, S_IRUGO, tpm_show_owned, NULL);
+static DEVICE_ATTR(temp_deactivated, S_IRUGO, tpm_show_temp_deactivated,
+		   NULL);
+static DEVICE_ATTR(caps, S_IRUGO, tpm_show_caps, NULL);
+static DEVICE_ATTR(cancel, S_IWUSR |S_IWGRP, NULL, tpm_store_cancel);
+
+static struct attribute *vtpm_attrs[] = {
+	&dev_attr_pubek.attr,
+	&dev_attr_pcrs.attr,
+	&dev_attr_enabled.attr,
+	&dev_attr_active.attr,
+	&dev_attr_owned.attr,
+	&dev_attr_temp_deactivated.attr,
+	&dev_attr_caps.attr,
+	&dev_attr_cancel.attr,
+	NULL,
+};
+
+static struct attribute_group vtpm_attr_grp = { .attrs = vtpm_attrs };
+
+#define TPM_LONG_TIMEOUT   (10 * 60 * HZ)
+
+static struct tpm_vendor_specific tpm_vtpm = {
+	.recv = vtpm_recv,
+	.send = vtpm_send,
+	.cancel = vtpm_cancel,
+	.status = vtpm_status,
+	.req_complete_mask = STATUS_BUSY | STATUS_DATA_AVAIL,
+	.req_complete_val  = STATUS_DATA_AVAIL,
+	.req_canceled = STATUS_READY,
+	.attr_group = &vtpm_attr_grp,
+	.miscdev = {
+		.fops = &vtpm_ops,
+	},
+	.duration = {
+		TPM_LONG_TIMEOUT,
+		TPM_LONG_TIMEOUT,
+		TPM_LONG_TIMEOUT,
+	},
+};
+
+struct tpm_chip *init_vtpm(struct device *dev,
+                           struct tpm_private *tp)
+{
+	long rc;
+	struct tpm_chip *chip;
+	struct vtpm_state *vtpms;
+
+	vtpms = kzalloc(sizeof(struct vtpm_state), GFP_KERNEL);
+	if (!vtpms)
+		return ERR_PTR(-ENOMEM);
+
+	vtpm_state_init(vtpms);
+	vtpms->tpm_private = tp;
+
+	chip = tpm_register_hardware(dev, &tpm_vtpm);
+	if (!chip) {
+		rc = -ENODEV;
+		goto err_free_mem;
+	}
+
+	chip_set_private(chip, vtpms);
+
+	return chip;
+
+err_free_mem:
+	kfree(vtpms);
+
+	return ERR_PTR(rc);
+}
+
+void cleanup_vtpm(struct device *dev)
+{
+	struct tpm_chip *chip = dev_get_drvdata(dev);
+	struct vtpm_state *vtpms = (struct vtpm_state*)chip_get_private(chip);
+	tpm_remove_hardware(dev);
+	kfree(vtpms);
+}
diff -rpuN linux-2.6.18.8/drivers/char/tpm/tpm_vtpm.h linux-2.6.18-xen-3.3.0/drivers/char/tpm/tpm_vtpm.h
--- linux-2.6.18.8/drivers/char/tpm/tpm_vtpm.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/char/tpm/tpm_vtpm.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,55 @@
+#ifndef TPM_VTPM_H
+#define TPM_VTPM_H
+
+struct tpm_chip;
+struct tpm_private;
+
+struct vtpm_state {
+	struct transmission *current_request;
+	spinlock_t           req_list_lock;
+	wait_queue_head_t    req_wait_queue;
+
+	struct list_head     queued_requests;
+
+	struct transmission *current_response;
+	spinlock_t           resp_list_lock;
+	wait_queue_head_t    resp_wait_queue;     // processes waiting for responses
+
+	u8                   vd_status;
+	u8                   flags;
+
+	unsigned long        disconnect_time;
+
+	/*
+	 * The following is a private structure of the underlying
+	 * driver. It is passed as parameter in the send function.
+	 */
+	struct tpm_private *tpm_private;
+};
+
+
+enum vdev_status {
+	TPM_VD_STATUS_DISCONNECTED = 0x0,
+	TPM_VD_STATUS_CONNECTED = 0x1
+};
+
+/* this function is called from tpm_vtpm.c */
+int vtpm_vd_send(struct tpm_private * tp,
+                 const u8 * buf, size_t count, void *ptr);
+
+/* these functions are offered by tpm_vtpm.c */
+struct tpm_chip *init_vtpm(struct device *,
+                           struct tpm_private *);
+void cleanup_vtpm(struct device *);
+int vtpm_vd_recv(const struct tpm_chip* chip,
+                 const unsigned char *buffer, size_t count, void *ptr);
+void vtpm_vd_status(const struct tpm_chip *, u8 status);
+
+static inline struct tpm_private *tpm_private_from_dev(struct device *dev)
+{
+	struct tpm_chip *chip = dev_get_drvdata(dev);
+	struct vtpm_state *vtpms = chip_get_private(chip);
+	return vtpms->tpm_private;
+}
+
+#endif
diff -rpuN linux-2.6.18.8/drivers/char/tpm/tpm_xen.c linux-2.6.18-xen-3.3.0/drivers/char/tpm/tpm_xen.c
--- linux-2.6.18.8/drivers/char/tpm/tpm_xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/char/tpm/tpm_xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,722 @@
+/*
+ * Copyright (c) 2005, IBM Corporation
+ *
+ * Author: Stefan Berger, stefanb@us.ibm.com
+ * Grant table support: Mahadevan Gomathisankaran
+ *
+ * This code has been derived from drivers/xen/netfront/netfront.c
+ *
+ * Copyright (c) 2002-2004, K A Fraser
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/errno.h>
+#include <linux/err.h>
+#include <linux/interrupt.h>
+#include <linux/mutex.h>
+#include <asm/uaccess.h>
+#include <xen/evtchn.h>
+#include <xen/interface/grant_table.h>
+#include <xen/interface/io/tpmif.h>
+#include <xen/gnttab.h>
+#include <xen/xenbus.h>
+#include "tpm.h"
+#include "tpm_vtpm.h"
+
+#undef DEBUG
+
+/* local structures */
+struct tpm_private {
+	struct tpm_chip *chip;
+
+	tpmif_tx_interface_t *tx;
+	atomic_t refcnt;
+	unsigned int irq;
+	u8 is_connected;
+	u8 is_suspended;
+
+	spinlock_t tx_lock;
+
+	struct tx_buffer *tx_buffers[TPMIF_TX_RING_SIZE];
+
+	atomic_t tx_busy;
+	void *tx_remember;
+
+	domid_t backend_id;
+	wait_queue_head_t wait_q;
+
+	struct xenbus_device *dev;
+	int ring_ref;
+};
+
+struct tx_buffer {
+	unsigned int size;	// available space in data
+	unsigned int len;	// used space in data
+	unsigned char *data;	// pointer to a page
+};
+
+
+/* locally visible variables */
+static grant_ref_t gref_head;
+static struct tpm_private *my_priv;
+
+/* local function prototypes */
+static irqreturn_t tpmif_int(int irq,
+                             void *tpm_priv,
+                             struct pt_regs *ptregs);
+static void tpmif_rx_action(unsigned long unused);
+static int tpmif_connect(struct xenbus_device *dev,
+                         struct tpm_private *tp,
+                         domid_t domid);
+static DECLARE_TASKLET(tpmif_rx_tasklet, tpmif_rx_action, 0);
+static int tpmif_allocate_tx_buffers(struct tpm_private *tp);
+static void tpmif_free_tx_buffers(struct tpm_private *tp);
+static void tpmif_set_connected_state(struct tpm_private *tp,
+                                      u8 newstate);
+static int tpm_xmit(struct tpm_private *tp,
+                    const u8 * buf, size_t count, int userbuffer,
+                    void *remember);
+static void destroy_tpmring(struct tpm_private *tp);
+void __exit tpmif_exit(void);
+
+#define DPRINTK(fmt, args...) \
+    pr_debug("xen_tpm_fr (%s:%d) " fmt, __FUNCTION__, __LINE__, ##args)
+#define IPRINTK(fmt, args...) \
+    printk(KERN_INFO "xen_tpm_fr: " fmt, ##args)
+#define WPRINTK(fmt, args...) \
+    printk(KERN_WARNING "xen_tpm_fr: " fmt, ##args)
+
+#define GRANT_INVALID_REF	0
+
+
+static inline int
+tx_buffer_copy(struct tx_buffer *txb, const u8 *src, int len,
+               int isuserbuffer)
+{
+	int copied = len;
+
+	if (len > txb->size)
+		copied = txb->size;
+	if (isuserbuffer) {
+		if (copy_from_user(txb->data, src, copied))
+			return -EFAULT;
+	} else {
+		memcpy(txb->data, src, copied);
+	}
+	txb->len = len;
+	return copied;
+}
+
+static inline struct tx_buffer *tx_buffer_alloc(void)
+{
+	struct tx_buffer *txb;
+
+	txb = kzalloc(sizeof(struct tx_buffer), GFP_KERNEL);
+	if (!txb)
+		return NULL;
+
+	txb->len = 0;
+	txb->size = PAGE_SIZE;
+	txb->data = (unsigned char *)__get_free_page(GFP_KERNEL);
+	if (txb->data == NULL) {
+		kfree(txb);
+		txb = NULL;
+	}
+
+	return txb;
+}
+
+
+static inline void tx_buffer_free(struct tx_buffer *txb)
+{
+	if (txb) {
+		free_page((long)txb->data);
+		kfree(txb);
+	}
+}
+
+/**************************************************************
+ Utility function for the tpm_private structure
+**************************************************************/
+static void tpm_private_init(struct tpm_private *tp)
+{
+	spin_lock_init(&tp->tx_lock);
+	init_waitqueue_head(&tp->wait_q);
+	atomic_set(&tp->refcnt, 1);
+}
+
+static void tpm_private_put(void)
+{
+	if (!atomic_dec_and_test(&my_priv->refcnt))
+		return;
+
+	tpmif_free_tx_buffers(my_priv);
+	kfree(my_priv);
+	my_priv = NULL;
+}
+
+static struct tpm_private *tpm_private_get(void)
+{
+	int err;
+
+	if (my_priv) {
+		atomic_inc(&my_priv->refcnt);
+		return my_priv;
+	}
+
+	my_priv = kzalloc(sizeof(struct tpm_private), GFP_KERNEL);
+	if (!my_priv)
+		return NULL;
+
+	tpm_private_init(my_priv);
+	err = tpmif_allocate_tx_buffers(my_priv);
+	if (err < 0)
+		tpm_private_put();
+
+	return my_priv;
+}
+
+/**************************************************************
+
+ The interface to let the tpm plugin register its callback
+ function and send data to another partition using this module
+
+**************************************************************/
+
+static DEFINE_MUTEX(suspend_lock);
+/*
+ * Send data via this module by calling this function
+ */
+int vtpm_vd_send(struct tpm_private *tp,
+                 const u8 * buf, size_t count, void *ptr)
+{
+	int sent;
+
+	mutex_lock(&suspend_lock);
+	sent = tpm_xmit(tp, buf, count, 0, ptr);
+	mutex_unlock(&suspend_lock);
+
+	return sent;
+}
+
+/**************************************************************
+ XENBUS support code
+**************************************************************/
+
+static int setup_tpmring(struct xenbus_device *dev,
+                         struct tpm_private *tp)
+{
+	tpmif_tx_interface_t *sring;
+	int err;
+
+	tp->ring_ref = GRANT_INVALID_REF;
+
+	sring = (void *)__get_free_page(GFP_KERNEL);
+	if (!sring) {
+		xenbus_dev_fatal(dev, -ENOMEM, "allocating shared ring");
+		return -ENOMEM;
+	}
+	tp->tx = sring;
+
+	err = xenbus_grant_ring(dev, virt_to_mfn(tp->tx));
+	if (err < 0) {
+		free_page((unsigned long)sring);
+		tp->tx = NULL;
+		xenbus_dev_fatal(dev, err, "allocating grant reference");
+		goto fail;
+	}
+	tp->ring_ref = err;
+
+	err = tpmif_connect(dev, tp, dev->otherend_id);
+	if (err)
+		goto fail;
+
+	return 0;
+fail:
+	destroy_tpmring(tp);
+	return err;
+}
+
+
+static void destroy_tpmring(struct tpm_private *tp)
+{
+	tpmif_set_connected_state(tp, 0);
+
+	if (tp->ring_ref != GRANT_INVALID_REF) {
+		gnttab_end_foreign_access(tp->ring_ref, (unsigned long)tp->tx);
+		tp->ring_ref = GRANT_INVALID_REF;
+		tp->tx = NULL;
+	}
+
+	if (tp->irq)
+		unbind_from_irqhandler(tp->irq, tp);
+
+	tp->irq = 0;
+}
+
+
+static int talk_to_backend(struct xenbus_device *dev,
+                           struct tpm_private *tp)
+{
+	const char *message = NULL;
+	int err;
+	struct xenbus_transaction xbt;
+
+	err = setup_tpmring(dev, tp);
+	if (err) {
+		xenbus_dev_fatal(dev, err, "setting up ring");
+		goto out;
+	}
+
+again:
+	err = xenbus_transaction_start(&xbt);
+	if (err) {
+		xenbus_dev_fatal(dev, err, "starting transaction");
+		goto destroy_tpmring;
+	}
+
+	err = xenbus_printf(xbt, dev->nodename,
+	                    "ring-ref","%u", tp->ring_ref);
+	if (err) {
+		message = "writing ring-ref";
+		goto abort_transaction;
+	}
+
+	err = xenbus_printf(xbt, dev->nodename, "event-channel", "%u",
+			    irq_to_evtchn_port(tp->irq));
+	if (err) {
+		message = "writing event-channel";
+		goto abort_transaction;
+	}
+
+	err = xenbus_transaction_end(xbt, 0);
+	if (err == -EAGAIN)
+		goto again;
+	if (err) {
+		xenbus_dev_fatal(dev, err, "completing transaction");
+		goto destroy_tpmring;
+	}
+
+	xenbus_switch_state(dev, XenbusStateConnected);
+
+	return 0;
+
+abort_transaction:
+	xenbus_transaction_end(xbt, 1);
+	if (message)
+		xenbus_dev_error(dev, err, "%s", message);
+destroy_tpmring:
+	destroy_tpmring(tp);
+out:
+	return err;
+}
+
+/**
+ * Callback received when the backend's state changes.
+ */
+static void backend_changed(struct xenbus_device *dev,
+			    enum xenbus_state backend_state)
+{
+	struct tpm_private *tp = tpm_private_from_dev(&dev->dev);
+	DPRINTK("\n");
+
+	switch (backend_state) {
+	case XenbusStateInitialising:
+	case XenbusStateInitWait:
+	case XenbusStateInitialised:
+	case XenbusStateReconfiguring:
+	case XenbusStateReconfigured:
+	case XenbusStateUnknown:
+		break;
+
+	case XenbusStateConnected:
+		tpmif_set_connected_state(tp, 1);
+		break;
+
+	case XenbusStateClosing:
+		tpmif_set_connected_state(tp, 0);
+		xenbus_frontend_closed(dev);
+		break;
+
+	case XenbusStateClosed:
+		tpmif_set_connected_state(tp, 0);
+		if (tp->is_suspended == 0)
+			device_unregister(&dev->dev);
+		xenbus_frontend_closed(dev);
+		break;
+	}
+}
+
+static int tpmfront_probe(struct xenbus_device *dev,
+                          const struct xenbus_device_id *id)
+{
+	int err;
+	int handle;
+	struct tpm_private *tp = tpm_private_get();
+
+	if (!tp)
+		return -ENOMEM;
+
+	tp->chip = init_vtpm(&dev->dev, tp);
+	if (IS_ERR(tp->chip))
+		return PTR_ERR(tp->chip);
+
+	err = xenbus_scanf(XBT_NIL, dev->nodename,
+	                   "handle", "%i", &handle);
+	if (XENBUS_EXIST_ERR(err))
+		return err;
+
+	if (err < 0) {
+		xenbus_dev_fatal(dev,err,"reading virtual-device");
+		return err;
+	}
+
+	tp->dev = dev;
+
+	err = talk_to_backend(dev, tp);
+	if (err) {
+		tpm_private_put();
+		return err;
+	}
+
+	return 0;
+}
+
+
+static int tpmfront_remove(struct xenbus_device *dev)
+{
+	struct tpm_private *tp = tpm_private_from_dev(&dev->dev);
+	destroy_tpmring(tp);
+	cleanup_vtpm(&dev->dev);
+	return 0;
+}
+
+static int tpmfront_suspend(struct xenbus_device *dev)
+{
+	struct tpm_private *tp = tpm_private_from_dev(&dev->dev);
+	u32 ctr;
+
+	/* Take the lock, preventing any application from sending. */
+	mutex_lock(&suspend_lock);
+	tp->is_suspended = 1;
+
+	for (ctr = 0; atomic_read(&tp->tx_busy); ctr++) {
+		if ((ctr % 10) == 0)
+			printk("TPM-FE [INFO]: Waiting for outstanding "
+			       "request.\n");
+		/* Wait for a request to be responded to. */
+		interruptible_sleep_on_timeout(&tp->wait_q, 100);
+	}
+
+	return 0;
+}
+
+static int tpmfront_suspend_finish(struct tpm_private *tp)
+{
+	tp->is_suspended = 0;
+	/* Allow applications to send again. */
+	mutex_unlock(&suspend_lock);
+	return 0;
+}
+
+static int tpmfront_suspend_cancel(struct xenbus_device *dev)
+{
+	struct tpm_private *tp = tpm_private_from_dev(&dev->dev);
+	return tpmfront_suspend_finish(tp);
+}
+
+static int tpmfront_resume(struct xenbus_device *dev)
+{
+	struct tpm_private *tp = tpm_private_from_dev(&dev->dev);
+	destroy_tpmring(tp);
+	return talk_to_backend(dev, tp);
+}
+
+static int tpmif_connect(struct xenbus_device *dev,
+                         struct tpm_private *tp,
+                         domid_t domid)
+{
+	int err;
+
+	tp->backend_id = domid;
+
+	err = bind_listening_port_to_irqhandler(
+		domid, tpmif_int, SA_SAMPLE_RANDOM, "tpmif", tp);
+	if (err <= 0) {
+		WPRINTK("bind_listening_port_to_irqhandler failed "
+			"(err=%d)\n", err);
+		return err;
+	}
+	tp->irq = err;
+
+	return 0;
+}
+
+static struct xenbus_device_id tpmfront_ids[] = {
+	{ "vtpm" },
+	{ "" }
+};
+
+static struct xenbus_driver tpmfront = {
+	.name = "vtpm",
+	.owner = THIS_MODULE,
+	.ids = tpmfront_ids,
+	.probe = tpmfront_probe,
+	.remove =  tpmfront_remove,
+	.resume = tpmfront_resume,
+	.otherend_changed = backend_changed,
+	.suspend = tpmfront_suspend,
+	.suspend_cancel = tpmfront_suspend_cancel,
+};
+
+static void __init init_tpm_xenbus(void)
+{
+	xenbus_register_frontend(&tpmfront);
+}
+
+static int tpmif_allocate_tx_buffers(struct tpm_private *tp)
+{
+	unsigned int i;
+
+	for (i = 0; i < TPMIF_TX_RING_SIZE; i++) {
+		tp->tx_buffers[i] = tx_buffer_alloc();
+		if (!tp->tx_buffers[i]) {
+			tpmif_free_tx_buffers(tp);
+			return -ENOMEM;
+		}
+	}
+	return 0;
+}
+
+static void tpmif_free_tx_buffers(struct tpm_private *tp)
+{
+	unsigned int i;
+
+	for (i = 0; i < TPMIF_TX_RING_SIZE; i++)
+		tx_buffer_free(tp->tx_buffers[i]);
+}
+
+static void tpmif_rx_action(unsigned long priv)
+{
+	struct tpm_private *tp = (struct tpm_private *)priv;
+	int i = 0;
+	unsigned int received;
+	unsigned int offset = 0;
+	u8 *buffer;
+	tpmif_tx_request_t *tx = &tp->tx->ring[i].req;
+
+	atomic_set(&tp->tx_busy, 0);
+	wake_up_interruptible(&tp->wait_q);
+
+	received = tx->size;
+
+	buffer = kmalloc(received, GFP_ATOMIC);
+	if (!buffer)
+		return;
+
+	for (i = 0; i < TPMIF_TX_RING_SIZE && offset < received; i++) {
+		struct tx_buffer *txb = tp->tx_buffers[i];
+		tpmif_tx_request_t *tx;
+		unsigned int tocopy;
+
+		tx = &tp->tx->ring[i].req;
+		tocopy = tx->size;
+		if (tocopy > PAGE_SIZE)
+			tocopy = PAGE_SIZE;
+
+		memcpy(&buffer[offset], txb->data, tocopy);
+
+		gnttab_release_grant_reference(&gref_head, tx->ref);
+
+		offset += tocopy;
+	}
+
+	vtpm_vd_recv(tp->chip, buffer, received, tp->tx_remember);
+	kfree(buffer);
+}
+
+
+static irqreturn_t tpmif_int(int irq, void *tpm_priv, struct pt_regs *ptregs)
+{
+	struct tpm_private *tp = tpm_priv;
+	unsigned long flags;
+
+	spin_lock_irqsave(&tp->tx_lock, flags);
+	tpmif_rx_tasklet.data = (unsigned long)tp;
+	tasklet_schedule(&tpmif_rx_tasklet);
+	spin_unlock_irqrestore(&tp->tx_lock, flags);
+
+	return IRQ_HANDLED;
+}
+
+
+static int tpm_xmit(struct tpm_private *tp,
+                    const u8 * buf, size_t count, int isuserbuffer,
+                    void *remember)
+{
+	tpmif_tx_request_t *tx;
+	TPMIF_RING_IDX i;
+	unsigned int offset = 0;
+
+	spin_lock_irq(&tp->tx_lock);
+
+	if (unlikely(atomic_read(&tp->tx_busy))) {
+		printk("tpm_xmit: There's an outstanding request/response "
+		       "on the way!\n");
+		spin_unlock_irq(&tp->tx_lock);
+		return -EBUSY;
+	}
+
+	if (tp->is_connected != 1) {
+		spin_unlock_irq(&tp->tx_lock);
+		return -EIO;
+	}
+
+	for (i = 0; count > 0 && i < TPMIF_TX_RING_SIZE; i++) {
+		struct tx_buffer *txb = tp->tx_buffers[i];
+		int copied;
+
+		if (!txb) {
+			DPRINTK("txb (i=%d) is NULL. buffers initilized?\n"
+				"Not transmitting anything!\n", i);
+			spin_unlock_irq(&tp->tx_lock);
+			return -EFAULT;
+		}
+
+		copied = tx_buffer_copy(txb, &buf[offset], count,
+		                        isuserbuffer);
+		if (copied < 0) {
+			/* An error occurred */
+			spin_unlock_irq(&tp->tx_lock);
+			return copied;
+		}
+		count -= copied;
+		offset += copied;
+
+		tx = &tp->tx->ring[i].req;
+		tx->addr = virt_to_machine(txb->data);
+		tx->size = txb->len;
+		tx->unused = 0;
+
+		DPRINTK("First 4 characters sent by TPM-FE are "
+			"0x%02x 0x%02x 0x%02x 0x%02x\n",
+		        txb->data[0],txb->data[1],txb->data[2],txb->data[3]);
+
+		/* Get the granttable reference for this page. */
+		tx->ref = gnttab_claim_grant_reference(&gref_head);
+		if (tx->ref == -ENOSPC) {
+			spin_unlock_irq(&tp->tx_lock);
+			DPRINTK("Grant table claim reference failed in "
+				"func:%s line:%d file:%s\n",
+				__FUNCTION__, __LINE__, __FILE__);
+			return -ENOSPC;
+		}
+		gnttab_grant_foreign_access_ref(tx->ref,
+						tp->backend_id,
+						virt_to_mfn(txb->data),
+						0 /*RW*/);
+		wmb();
+	}
+
+	atomic_set(&tp->tx_busy, 1);
+	tp->tx_remember = remember;
+
+	mb();
+
+	notify_remote_via_irq(tp->irq);
+
+	spin_unlock_irq(&tp->tx_lock);
+	return offset;
+}
+
+
+static void tpmif_notify_upperlayer(struct tpm_private *tp)
+{
+	/* Notify upper layer about the state of the connection to the BE. */
+	vtpm_vd_status(tp->chip, (tp->is_connected
+				  ? TPM_VD_STATUS_CONNECTED
+				  : TPM_VD_STATUS_DISCONNECTED));
+}
+
+
+static void tpmif_set_connected_state(struct tpm_private *tp, u8 is_connected)
+{
+	/*
+	 * Don't notify upper layer if we are in suspend mode and
+	 * should disconnect - assumption is that we will resume
+	 * The mutex keeps apps from sending.
+	 */
+	if (is_connected == 0 && tp->is_suspended == 1)
+		return;
+
+	/*
+	 * Unlock the mutex if we are connected again
+	 * after being suspended - now resuming.
+	 * This also removes the suspend state.
+	 */
+	if (is_connected == 1 && tp->is_suspended == 1)
+		tpmfront_suspend_finish(tp);
+
+	if (is_connected != tp->is_connected) {
+		tp->is_connected = is_connected;
+		tpmif_notify_upperlayer(tp);
+	}
+}
+
+
+
+/* =================================================================
+ * Initialization function.
+ * =================================================================
+ */
+
+
+static int __init tpmif_init(void)
+{
+	struct tpm_private *tp;
+
+	if (is_initial_xendomain())
+		return -EPERM;
+
+	tp = tpm_private_get();
+	if (!tp)
+		return -ENOMEM;
+
+	IPRINTK("Initialising the vTPM driver.\n");
+	if (gnttab_alloc_grant_references(TPMIF_TX_RING_SIZE,
+					  &gref_head) < 0) {
+		tpm_private_put();
+		return -EFAULT;
+	}
+
+	init_tpm_xenbus();
+	return 0;
+}
+
+
+module_init(tpmif_init);
+
+MODULE_LICENSE("Dual BSD/GPL");
diff -rpuN linux-2.6.18.8/drivers/char/tty_io.c linux-2.6.18-xen-3.3.0/drivers/char/tty_io.c
--- linux-2.6.18.8/drivers/char/tty_io.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/char/tty_io.c	2008-08-21 11:36:07.000000000 +0200
@@ -130,6 +130,8 @@ LIST_HEAD(tty_drivers);			/* linked list
    vt.c for deeply disgusting hack reasons */
 DEFINE_MUTEX(tty_mutex);
 
+int console_use_vt = 1;
+
 #ifdef CONFIG_UNIX98_PTYS
 extern struct tty_driver *ptm_driver;	/* Unix98 pty masters; for /dev/ptmx */
 extern int pty_limit;		/* Config limit on Unix98 ptys */
@@ -2483,7 +2485,7 @@ retry_open:
 		goto got_driver;
 	}
 #ifdef CONFIG_VT
-	if (device == MKDEV(TTY_MAJOR,0)) {
+	if (console_use_vt && (device == MKDEV(TTY_MAJOR,0))) {
 		extern struct tty_driver *console_driver;
 		driver = console_driver;
 		index = fg_console;
@@ -3909,6 +3911,8 @@ static int __init tty_init(void)
 #endif
 
 #ifdef CONFIG_VT
+	if (!console_use_vt)
+		goto out_vt;
 	cdev_init(&vc0_cdev, &console_fops);
 	if (cdev_add(&vc0_cdev, MKDEV(TTY_MAJOR, 0), 1) ||
 	    register_chrdev_region(MKDEV(TTY_MAJOR, 0), 1, "/dev/vc/0") < 0)
@@ -3916,6 +3920,7 @@ static int __init tty_init(void)
 	class_device_create(tty_class, NULL, MKDEV(TTY_MAJOR, 0), NULL, "tty0");
 
 	vty_init();
+ out_vt:
 #endif
 	return 0;
 }
diff -rpuN linux-2.6.18.8/drivers/cpufreq/cpufreq_ondemand.c linux-2.6.18-xen-3.3.0/drivers/cpufreq/cpufreq_ondemand.c
--- linux-2.6.18.8/drivers/cpufreq/cpufreq_ondemand.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/cpufreq/cpufreq_ondemand.c	2008-08-21 11:36:07.000000000 +0200
@@ -96,6 +96,7 @@ static inline cputime64_t get_cpu_idle_t
 	return retval;
 }
 
+
 /************************** sysfs interface ************************/
 static ssize_t show_sampling_rate_max(struct cpufreq_policy *policy, char *buf)
 {
@@ -222,17 +223,14 @@ static struct attribute_group dbs_attr_g
 
 /************************** sysfs end ************************/
 
-static void dbs_check_cpu(struct cpu_dbs_info_s *this_dbs_info)
+#ifndef CONFIG_XEN
+static int dbs_calc_load(struct cpu_dbs_info_s *this_dbs_info)
 {
-	unsigned int idle_ticks, total_ticks;
-	unsigned int load;
-	cputime64_t cur_jiffies;
-
 	struct cpufreq_policy *policy;
+	cputime64_t cur_jiffies;
+	cputime64_t total_ticks, idle_ticks;
 	unsigned int j;
-
-	if (!this_dbs_info->enable)
-		return;
+	int load;
 
 	policy = this_dbs_info->cur_policy;
 	cur_jiffies = jiffies64_to_cputime64(get_jiffies_64());
@@ -240,7 +238,7 @@ static void dbs_check_cpu(struct cpu_dbs
 			this_dbs_info->prev_cpu_wall);
 	this_dbs_info->prev_cpu_wall = cur_jiffies;
 	if (!total_ticks)
-		return;
+		return 200;
 	/*
 	 * Every sampling_rate, we check, if current idle time is less
 	 * than 20% (default), then we try to increase frequency
@@ -270,6 +268,81 @@ static void dbs_check_cpu(struct cpu_dbs
 			idle_ticks = tmp_idle_ticks;
 	}
 	load = (100 * (total_ticks - idle_ticks)) / total_ticks;
+	return load;
+}
+#else
+
+#include <xen/interface/platform.h>
+static int dbs_calc_load(struct cpu_dbs_info_s *this_dbs_info)
+{
+	int load = 0;
+	struct xen_platform_op op;
+	uint64_t idletime[NR_CPUS];
+	struct cpufreq_policy *policy;
+	unsigned int j;
+	cpumask_t cpumap;
+
+	policy = this_dbs_info->cur_policy;
+	cpumap = policy->cpus;
+
+	op.cmd = XENPF_getidletime;
+	set_xen_guest_handle(op.u.getidletime.cpumap_bitmap, (uint8_t *) cpus_addr(cpumap));
+	op.u.getidletime.cpumap_nr_cpus = NR_CPUS;
+	set_xen_guest_handle(op.u.getidletime.idletime, idletime);
+	if (HYPERVISOR_platform_op(&op))
+		return 200;
+
+	for_each_cpu_mask(j, cpumap) {
+		cputime64_t total_idle_nsecs, tmp_idle_nsecs;
+		cputime64_t total_wall_nsecs, tmp_wall_nsecs;
+		struct cpu_dbs_info_s *j_dbs_info;
+		unsigned long tmp_load, tmp_wall_msecs, tmp_idle_msecs;
+
+		j_dbs_info = &per_cpu(cpu_dbs_info, j);
+		total_idle_nsecs = idletime[j];
+		tmp_idle_nsecs = cputime64_sub(total_idle_nsecs,
+				j_dbs_info->prev_cpu_idle);
+		total_wall_nsecs = op.u.getidletime.now;
+		tmp_wall_nsecs = cputime64_sub(total_wall_nsecs,
+				j_dbs_info->prev_cpu_wall);
+
+		if (tmp_wall_nsecs == 0)
+			return 200;
+
+		j_dbs_info->prev_cpu_wall = total_wall_nsecs;
+		j_dbs_info->prev_cpu_idle = total_idle_nsecs;
+
+		/* Convert nsecs to msecs and clamp times to sane values. */
+		do_div(tmp_wall_nsecs, 1000000);
+		tmp_wall_msecs = tmp_wall_nsecs;
+		do_div(tmp_idle_nsecs, 1000000);
+		tmp_idle_msecs = tmp_idle_nsecs;
+		if (tmp_wall_msecs == 0)
+			tmp_wall_msecs = 1;
+		if (tmp_idle_msecs > tmp_wall_msecs)
+			tmp_idle_msecs = tmp_wall_msecs;
+
+		tmp_load = (100 * (tmp_wall_msecs - tmp_idle_msecs)) /
+				tmp_wall_msecs;
+		load = max(load, min(100, (int) tmp_load));
+	}
+	return load;
+}
+#endif
+
+static void dbs_check_cpu(struct cpu_dbs_info_s *this_dbs_info)
+{
+	int load;
+
+	struct cpufreq_policy *policy;
+
+	if (!this_dbs_info->enable)
+		return;
+
+	policy = this_dbs_info->cur_policy;
+	load = dbs_calc_load(this_dbs_info);
+	if (load > 100) 
+		return;
 
 	/* Check for frequency increase */
 	if (load > dbs_tuners_ins.up_threshold) {
diff -rpuN linux-2.6.18.8/drivers/cpufreq/cpufreq_stats.c linux-2.6.18-xen-3.3.0/drivers/cpufreq/cpufreq_stats.c
--- linux-2.6.18.8/drivers/cpufreq/cpufreq_stats.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/cpufreq/cpufreq_stats.c	2008-08-21 11:36:07.000000000 +0200
@@ -292,6 +292,9 @@ cpufreq_stat_notifier_trans (struct noti
 	if (old_index == new_index)
 		return 0;
 
+        if ((old_index < 0) || (new_index < 0))
+                return 0;
+
 	spin_lock(&cpufreq_stats_lock);
 	stat->last_index = new_index;
 #ifdef CONFIG_CPU_FREQ_STAT_DETAILS
diff -rpuN linux-2.6.18.8/drivers/cpufreq/Kconfig linux-2.6.18-xen-3.3.0/drivers/cpufreq/Kconfig
--- linux-2.6.18.8/drivers/cpufreq/Kconfig	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/cpufreq/Kconfig	2008-08-21 11:36:07.000000000 +0200
@@ -1,5 +1,6 @@
 config CPU_FREQ
 	bool "CPU Frequency scaling"
+	depends on !PROCESSOR_EXTERNAL_CONTROL
 	help
 	  CPU Frequency scaling allows you to change the clock speed of 
 	  CPUs on the fly. This is a nice method to save power, because 
diff -rpuN linux-2.6.18.8/drivers/i2c/busses/i2c-i801.c linux-2.6.18-xen-3.3.0/drivers/i2c/busses/i2c-i801.c
--- linux-2.6.18.8/drivers/i2c/busses/i2c-i801.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/i2c/busses/i2c-i801.c	2008-08-21 11:36:07.000000000 +0200
@@ -33,6 +33,9 @@
     ICH7		27DA
     ESB2		269B
     ICH8		283E
+    ICH9		2930
+    ICH10		3A30
+    ICH10		3A60
     This driver supports several versions of Intel's I/O Controller Hubs (ICH).
     For SMBus support, they are similar to the PIIX4 and are part
     of Intel's '810' and other chipsets.
@@ -457,6 +460,9 @@ static struct pci_device_id i801_ids[] =
 	{ PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ICH7_17) },
 	{ PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ESB2_17) },
 	{ PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ICH8_5) },
+	{ PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ICH9_6) },
+	{ PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ICH10_4) },
+	{ PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_ICH10_5) },
 	{ 0, }
 };
 
diff -rpuN linux-2.6.18.8/drivers/i2c/busses/Kconfig linux-2.6.18-xen-3.3.0/drivers/i2c/busses/Kconfig
--- linux-2.6.18.8/drivers/i2c/busses/Kconfig	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/i2c/busses/Kconfig	2008-08-21 11:36:07.000000000 +0200
@@ -125,6 +125,8 @@ config I2C_I801
 	    ICH7
 	    ESB2
 	    ICH8
+	    ICH9
+	    ICH10
 
 	  This driver can also be built as a module.  If so, the module
 	  will be called i2c-i801.
diff -rpuN linux-2.6.18.8/drivers/ide/ide-lib.c linux-2.6.18-xen-3.3.0/drivers/ide/ide-lib.c
--- linux-2.6.18.8/drivers/ide/ide-lib.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/ide/ide-lib.c	2008-08-21 11:36:07.000000000 +0200
@@ -408,10 +408,10 @@ void ide_toggle_bounce(ide_drive_t *driv
 {
 	u64 addr = BLK_BOUNCE_HIGH;	/* dma64_addr_t */
 
-	if (!PCI_DMA_BUS_IS_PHYS) {
-		addr = BLK_BOUNCE_ANY;
-	} else if (on && drive->media == ide_disk) {
-		if (HWIF(drive)->pci_dev)
+	if (on && drive->media == ide_disk) {
+		if (!PCI_DMA_BUS_IS_PHYS)
+			addr = BLK_BOUNCE_ANY;
+		else if (HWIF(drive)->pci_dev)
 			addr = HWIF(drive)->pci_dev->dma_mask;
 	}
 
diff -rpuN linux-2.6.18.8/drivers/ide/pci/atiixp.c linux-2.6.18-xen-3.3.0/drivers/ide/pci/atiixp.c
--- linux-2.6.18.8/drivers/ide/pci/atiixp.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/ide/pci/atiixp.c	2008-08-21 11:36:07.000000000 +0200
@@ -348,6 +348,7 @@ static struct pci_device_id atiixp_pci_t
 	{ PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_IXP300_IDE, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0},
 	{ PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_IXP400_IDE, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0},
 	{ PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_IXP600_IDE, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0},
+	{ PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_IXP700_IDE, PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0},
 	{ 0, },
 };
 MODULE_DEVICE_TABLE(pci, atiixp_pci_tbl);
diff -rpuN linux-2.6.18.8/drivers/input/gameport/gameport.c linux-2.6.18-xen-3.3.0/drivers/input/gameport/gameport.c
--- linux-2.6.18.8/drivers/input/gameport/gameport.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/input/gameport/gameport.c	2008-08-21 11:36:07.000000000 +0200
@@ -37,8 +37,6 @@ EXPORT_SYMBOL(gameport_unregister_driver
 EXPORT_SYMBOL(gameport_open);
 EXPORT_SYMBOL(gameport_close);
 EXPORT_SYMBOL(gameport_rescan);
-EXPORT_SYMBOL(gameport_cooked_read);
-EXPORT_SYMBOL(gameport_set_name);
 EXPORT_SYMBOL(gameport_set_phys);
 EXPORT_SYMBOL(gameport_start_polling);
 EXPORT_SYMBOL(gameport_stop_polling);
diff -rpuN linux-2.6.18.8/drivers/Makefile linux-2.6.18-xen-3.3.0/drivers/Makefile
--- linux-2.6.18.8/drivers/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -31,6 +31,7 @@ obj-y				+= base/ block/ misc/ mfd/ net/
 obj-$(CONFIG_NUBUS)		+= nubus/
 obj-$(CONFIG_ATM)		+= atm/
 obj-$(CONFIG_PPC_PMAC)		+= macintosh/
+obj-$(CONFIG_XEN)		+= xen/
 obj-$(CONFIG_IDE)		+= ide/
 obj-$(CONFIG_FC4)		+= fc4/
 obj-$(CONFIG_SCSI)		+= scsi/
diff -rpuN linux-2.6.18.8/drivers/net/Kconfig linux-2.6.18-xen-3.3.0/drivers/net/Kconfig
--- linux-2.6.18.8/drivers/net/Kconfig	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/Kconfig	2008-08-21 11:36:07.000000000 +0200
@@ -2399,6 +2399,8 @@ config MYRI10GE
 	  <file:Documentation/networking/net-modules.txt>.  The module
 	  will be called myri10ge.
 
+source "drivers/net/sfc/Kconfig"
+
 endmenu
 
 source "drivers/net/tokenring/Kconfig"
diff -rpuN linux-2.6.18.8/drivers/net/Makefile linux-2.6.18-xen-3.3.0/drivers/net/Makefile
--- linux-2.6.18.8/drivers/net/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -219,3 +219,5 @@ obj-$(CONFIG_NETCONSOLE) += netconsole.o
 
 obj-$(CONFIG_FS_ENET) += fs_enet/
 
+obj-$(CONFIG_SFC) += sfc/
+
diff -rpuN linux-2.6.18.8/drivers/net/r8169.c linux-2.6.18-xen-3.3.0/drivers/net/r8169.c
--- linux-2.6.18.8/drivers/net/r8169.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/r8169.c	2008-08-21 11:36:07.000000000 +0200
@@ -6,26 +6,26 @@
  History:
  Feb  4 2002	- created initially by ShuChen <shuchen@realtek.com.tw>.
  May 20 2002	- Add link status force-mode and TBI mode support.
-        2004	- Massive updates. See kernel SCM system for details.
+	2004	- Massive updates. See kernel SCM system for details.
 =========================================================================
   1. [DEPRECATED: use ethtool instead] The media can be forced in 5 modes.
 	 Command: 'insmod r8169 media = SET_MEDIA'
 	 Ex:	  'insmod r8169 media = 0x04' will force PHY to operate in 100Mpbs Half-duplex.
-	
+
 	 SET_MEDIA can be:
  		_10_Half	= 0x01
  		_10_Full	= 0x02
  		_100_Half	= 0x04
  		_100_Full	= 0x08
  		_1000_Full	= 0x10
-  
+
   2. Support TBI mode.
 =========================================================================
 VERSION 1.1	<2002/10/4>
 
 	The bit4:0 of MII register 4 is called "selector field", and have to be
 	00001b to indicate support of IEEE std 802.3 during NWay process of
-	exchanging Link Code Word (FLP). 
+	exchanging Link Code Word (FLP).
 
 VERSION 1.2	<2002/11/30>
 
@@ -81,10 +81,10 @@ VERSION 2.2LK	<2005/01/25>
 
 #ifdef RTL8169_DEBUG
 #define assert(expr) \
-        if(!(expr)) {					\
-	        printk( "Assertion failed! %s,%s,%s,line=%d\n",	\
-        	#expr,__FILE__,__FUNCTION__,__LINE__);		\
-        }
+	if (!(expr)) {					\
+		printk( "Assertion failed! %s,%s,%s,line=%d\n",	\
+		#expr,__FILE__,__FUNCTION__,__LINE__);		\
+	}
 #define dprintk(fmt, args...)	do { printk(PFX fmt, ## args); } while (0)
 #else
 #define assert(expr) do {} while (0)
@@ -150,11 +150,16 @@ static const int multicast_filter_limit 
 #define RTL_R32(reg)		((unsigned long) readl (ioaddr + (reg)))
 
 enum mac_version {
-	RTL_GIGA_MAC_VER_B = 0x00,
-	/* RTL_GIGA_MAC_VER_C = 0x03, */
-	RTL_GIGA_MAC_VER_D = 0x01,
-	RTL_GIGA_MAC_VER_E = 0x02,
-	RTL_GIGA_MAC_VER_X = 0x04	/* Greater than RTL_GIGA_MAC_VER_E */
+	RTL_GIGA_MAC_VER_01 = 0x00,
+	RTL_GIGA_MAC_VER_02 = 0x01,
+	RTL_GIGA_MAC_VER_03 = 0x02,
+	RTL_GIGA_MAC_VER_04 = 0x03,
+	RTL_GIGA_MAC_VER_05 = 0x04,
+	RTL_GIGA_MAC_VER_11 = 0x0b,
+	RTL_GIGA_MAC_VER_12 = 0x0c,
+	RTL_GIGA_MAC_VER_13 = 0x0d,
+	RTL_GIGA_MAC_VER_14 = 0x0e,
+	RTL_GIGA_MAC_VER_15 = 0x0f
 };
 
 enum phy_version {
@@ -166,7 +171,6 @@ enum phy_version {
 	RTL_GIGA_PHY_VER_H = 0x08, /* PHY Reg 0x03 bit0-3 == 0x0003 */
 };
 
-
 #define _R(NAME,MAC,MASK) \
 	{ .name = NAME, .mac_version = MAC, .RxConfigMask = MASK }
 
@@ -175,19 +179,45 @@ static const struct {
 	u8 mac_version;
 	u32 RxConfigMask;	/* Clears the bits supported by this chip */
 } rtl_chip_info[] = {
-	_R("RTL8169",		RTL_GIGA_MAC_VER_B, 0xff7e1880),
-	_R("RTL8169s/8110s",	RTL_GIGA_MAC_VER_D, 0xff7e1880),
-	_R("RTL8169s/8110s",	RTL_GIGA_MAC_VER_E, 0xff7e1880),
-	_R("RTL8169s/8110s",	RTL_GIGA_MAC_VER_X, 0xff7e1880),
+	_R("RTL8169",		RTL_GIGA_MAC_VER_01, 0xff7e1880),
+	_R("RTL8169s/8110s",	RTL_GIGA_MAC_VER_02, 0xff7e1880),
+	_R("RTL8169s/8110s",	RTL_GIGA_MAC_VER_03, 0xff7e1880),
+	_R("RTL8169sb/8110sb",	RTL_GIGA_MAC_VER_04, 0xff7e1880),
+	_R("RTL8169sc/8110sc",	RTL_GIGA_MAC_VER_05, 0xff7e1880),
+	_R("RTL8168b/8111b",	RTL_GIGA_MAC_VER_11, 0xff7e1880), // PCI-E
+	_R("RTL8168b/8111b",	RTL_GIGA_MAC_VER_12, 0xff7e1880), // PCI-E
+	_R("RTL8101e",		RTL_GIGA_MAC_VER_13, 0xff7e1880), // PCI-E 8139
+	_R("RTL8100e",		RTL_GIGA_MAC_VER_14, 0xff7e1880), // PCI-E 8139
+	_R("RTL8100e",		RTL_GIGA_MAC_VER_15, 0xff7e1880)  // PCI-E 8139
 };
 #undef _R
 
+enum cfg_version {
+	RTL_CFG_0 = 0x00,
+	RTL_CFG_1,
+	RTL_CFG_2
+};
+
+static const struct {
+	unsigned int region;
+	unsigned int align;
+} rtl_cfg_info[] = {
+	[RTL_CFG_0] = { 1, NET_IP_ALIGN },
+	[RTL_CFG_1] = { 2, NET_IP_ALIGN },
+	[RTL_CFG_2] = { 2, 8 }
+};
+
 static struct pci_device_id rtl8169_pci_tbl[] = {
-	{ PCI_DEVICE(PCI_VENDOR_ID_REALTEK,	0x8169), },
-	{ PCI_DEVICE(PCI_VENDOR_ID_REALTEK,	0x8129), },
-	{ PCI_DEVICE(PCI_VENDOR_ID_DLINK,	0x4300), },
-	{ PCI_DEVICE(0x16ec,			0x0116), },
-	{ PCI_VENDOR_ID_LINKSYS,		0x1032, PCI_ANY_ID, 0x0024, },
+	{ PCI_DEVICE(PCI_VENDOR_ID_REALTEK,	0x8129), 0, 0, RTL_CFG_0 },
+	{ PCI_DEVICE(PCI_VENDOR_ID_REALTEK,	0x8136), 0, 0, RTL_CFG_2 },
+	{ PCI_DEVICE(PCI_VENDOR_ID_REALTEK,	0x8167), 0, 0, RTL_CFG_0 },
+	{ PCI_DEVICE(PCI_VENDOR_ID_REALTEK,	0x8168), 0, 0, RTL_CFG_2 },
+	{ PCI_DEVICE(PCI_VENDOR_ID_REALTEK,	0x8169), 0, 0, RTL_CFG_0 },
+	{ PCI_DEVICE(PCI_VENDOR_ID_DLINK,	0x4300), 0, 0, RTL_CFG_0 },
+	{ PCI_DEVICE(0x1259,			0xc107), 0, 0, RTL_CFG_0 },
+	{ PCI_DEVICE(0x16ec,			0x0116), 0, 0, RTL_CFG_0 },
+	{ PCI_VENDOR_ID_LINKSYS,		0x1032,
+		PCI_ANY_ID, 0x0024, 0, 0, RTL_CFG_0 },
 	{0,},
 };
 
@@ -257,10 +287,11 @@ enum RTL8169_register_content {
 	RxOK = 0x01,
 
 	/* RxStatusDesc */
-	RxRES = 0x00200000,
-	RxCRC = 0x00080000,
-	RxRUNT = 0x00100000,
-	RxRWT = 0x00400000,
+	RxFOVF	= (1 << 23),
+	RxRWT	= (1 << 22),
+	RxRES	= (1 << 21),
+	RxRUNT	= (1 << 20),
+	RxCRC	= (1 << 19),
 
 	/* ChipCmdBits */
 	CmdReset = 0x10,
@@ -326,30 +357,6 @@ enum RTL8169_register_content {
 	LinkStatus = 0x02,
 	FullDup = 0x01,
 
-	/* GIGABIT_PHY_registers */
-	PHY_CTRL_REG = 0,
-	PHY_STAT_REG = 1,
-	PHY_AUTO_NEGO_REG = 4,
-	PHY_1000_CTRL_REG = 9,
-
-	/* GIGABIT_PHY_REG_BIT */
-	PHY_Restart_Auto_Nego = 0x0200,
-	PHY_Enable_Auto_Nego = 0x1000,
-
-	/* PHY_STAT_REG = 1 */
-	PHY_Auto_Neco_Comp = 0x0020,
-
-	/* PHY_AUTO_NEGO_REG = 4 */
-	PHY_Cap_10_Half = 0x0020,
-	PHY_Cap_10_Full = 0x0040,
-	PHY_Cap_100_Half = 0x0080,
-	PHY_Cap_100_Full = 0x0100,
-
-	/* PHY_1000_CTRL_REG = 9 */
-	PHY_Cap_1000_Full = 0x0200,
-
-	PHY_Cap_Null = 0x0,
-
 	/* _MediaType */
 	_10_Half = 0x01,
 	_10_Full = 0x02,
@@ -417,6 +424,7 @@ struct ring_info {
 struct rtl8169_private {
 	void __iomem *mmio_addr;	/* memory map physical address */
 	struct pci_dev *pci_dev;	/* Index of PCI device */
+	struct net_device *dev;
 	struct net_device_stats stats;	/* statistics of net device */
 	spinlock_t lock;		/* spin lock flag */
 	u32 msg_enable;
@@ -433,6 +441,7 @@ struct rtl8169_private {
 	dma_addr_t RxPhyAddr;
 	struct sk_buff *Rx_skbuff[NUM_RX_DESC];	/* Rx data buffers */
 	struct ring_info tx_skb[NUM_TX_DESC];	/* Tx data buffers */
+	unsigned align;
 	unsigned rx_buf_sz;
 	struct timer_list timer;
 	u16 cp_cmd;
@@ -466,8 +475,7 @@ MODULE_VERSION(RTL8169_VERSION);
 
 static int rtl8169_open(struct net_device *dev);
 static int rtl8169_start_xmit(struct sk_buff *skb, struct net_device *dev);
-static irqreturn_t rtl8169_interrupt(int irq, void *dev_instance,
-			      struct pt_regs *regs);
+static irqreturn_t rtl8169_interrupt(int irq, void *dev_instance, struct pt_regs *regs);
 static int rtl8169_init_ring(struct net_device *dev);
 static void rtl8169_hw_start(struct net_device *dev);
 static int rtl8169_close(struct net_device *dev);
@@ -488,12 +496,7 @@ static const u16 rtl8169_intr_mask =
 static const u16 rtl8169_napi_event =
 	RxOK | RxOverflow | RxFIFOOver | TxOK | TxErr;
 static const unsigned int rtl8169_rx_config =
-    (RX_FIFO_THRESH << RxCfgFIFOShift) | (RX_DMA_BURST << RxCfgDMAShift);
-
-#define PHY_Cap_10_Half_Or_Less PHY_Cap_10_Half
-#define PHY_Cap_10_Full_Or_Less PHY_Cap_10_Full | PHY_Cap_10_Half_Or_Less
-#define PHY_Cap_100_Half_Or_Less PHY_Cap_100_Half | PHY_Cap_10_Full_Or_Less
-#define PHY_Cap_100_Full_Or_Less PHY_Cap_100_Full | PHY_Cap_100_Half_Or_Less
+	(RX_FIFO_THRESH << RxCfgFIFOShift) | (RX_DMA_BURST << RxCfgDMAShift);
 
 static void mdio_write(void __iomem *ioaddr, int RegAddr, int value)
 {
@@ -503,7 +506,7 @@ static void mdio_write(void __iomem *ioa
 
 	for (i = 20; i > 0; i--) {
 		/* Check if the RTL8169 has completed writing to the specified MII register */
-		if (!(RTL_R32(PHYAR) & 0x80000000)) 
+		if (!(RTL_R32(PHYAR) & 0x80000000))
 			break;
 		udelay(25);
 	}
@@ -547,7 +550,7 @@ static unsigned int rtl8169_tbi_reset_pe
 
 static unsigned int rtl8169_xmii_reset_pending(void __iomem *ioaddr)
 {
-	return mdio_read(ioaddr, 0) & 0x8000;
+	return mdio_read(ioaddr, MII_BMCR) & BMCR_RESET;
 }
 
 static unsigned int rtl8169_tbi_link_ok(void __iomem *ioaddr)
@@ -569,8 +572,8 @@ static void rtl8169_xmii_reset_enable(vo
 {
 	unsigned int val;
 
-	val = (mdio_read(ioaddr, PHY_CTRL_REG) | 0x8000) & 0xffff;
-	mdio_write(ioaddr, PHY_CTRL_REG, val);
+	mdio_write(ioaddr, MII_BMCR, BMCR_RESET);
+	val = mdio_read(ioaddr, MII_BMCR);
 }
 
 static void rtl8169_check_link_status(struct net_device *dev,
@@ -608,7 +611,7 @@ static void rtl8169_link_option(int idx,
 		{ SPEED_1000,	DUPLEX_FULL, AUTONEG_ENABLE,	0xff }
 	}, *p;
 	unsigned char option;
-	
+
 	option = ((idx < MAX_UNITS) && (idx >= 0)) ? media[idx] : 0xff;
 
 	if ((option != 0xff) && !idx && netif_msg_drv(&debug))
@@ -650,9 +653,9 @@ static void rtl8169_get_wol(struct net_d
 	if (options & UWF)
 		wol->wolopts |= WAKE_UCAST;
 	if (options & BWF)
-	        wol->wolopts |= WAKE_BCAST;
+		wol->wolopts |= WAKE_BCAST;
 	if (options & MWF)
-	        wol->wolopts |= WAKE_MCAST;
+		wol->wolopts |= WAKE_MCAST;
 
 out_unlock:
 	spin_unlock_irq(&tp->lock);
@@ -745,38 +748,57 @@ static int rtl8169_set_speed_xmii(struct
 	void __iomem *ioaddr = tp->mmio_addr;
 	int auto_nego, giga_ctrl;
 
-	auto_nego = mdio_read(ioaddr, PHY_AUTO_NEGO_REG);
-	auto_nego &= ~(PHY_Cap_10_Half | PHY_Cap_10_Full |
-		       PHY_Cap_100_Half | PHY_Cap_100_Full);
-	giga_ctrl = mdio_read(ioaddr, PHY_1000_CTRL_REG);
-	giga_ctrl &= ~(PHY_Cap_1000_Full | PHY_Cap_Null);
+	auto_nego = mdio_read(ioaddr, MII_ADVERTISE);
+	auto_nego &= ~(ADVERTISE_10HALF | ADVERTISE_10FULL |
+		       ADVERTISE_100HALF | ADVERTISE_100FULL);
+	giga_ctrl = mdio_read(ioaddr, MII_CTRL1000);
+	giga_ctrl &= ~(ADVERTISE_1000FULL | ADVERTISE_1000HALF);
 
 	if (autoneg == AUTONEG_ENABLE) {
-		auto_nego |= (PHY_Cap_10_Half | PHY_Cap_10_Full |
-			      PHY_Cap_100_Half | PHY_Cap_100_Full);
-		giga_ctrl |= PHY_Cap_1000_Full;
+		auto_nego |= (ADVERTISE_10HALF | ADVERTISE_10FULL |
+			      ADVERTISE_100HALF | ADVERTISE_100FULL);
+		giga_ctrl |= ADVERTISE_1000FULL | ADVERTISE_1000HALF;
 	} else {
 		if (speed == SPEED_10)
-			auto_nego |= PHY_Cap_10_Half | PHY_Cap_10_Full;
+			auto_nego |= ADVERTISE_10HALF | ADVERTISE_10FULL;
 		else if (speed == SPEED_100)
-			auto_nego |= PHY_Cap_100_Half | PHY_Cap_100_Full;
+			auto_nego |= ADVERTISE_100HALF | ADVERTISE_100FULL;
 		else if (speed == SPEED_1000)
-			giga_ctrl |= PHY_Cap_1000_Full;
+			giga_ctrl |= ADVERTISE_1000FULL | ADVERTISE_1000HALF;
 
 		if (duplex == DUPLEX_HALF)
-			auto_nego &= ~(PHY_Cap_10_Full | PHY_Cap_100_Full);
+			auto_nego &= ~(ADVERTISE_10FULL | ADVERTISE_100FULL);
 
 		if (duplex == DUPLEX_FULL)
-			auto_nego &= ~(PHY_Cap_10_Half | PHY_Cap_100_Half);
+			auto_nego &= ~(ADVERTISE_10HALF | ADVERTISE_100HALF);
+
+		/* This tweak comes straight from Realtek's driver. */
+		if ((speed == SPEED_100) && (duplex == DUPLEX_HALF) &&
+		    (tp->mac_version == RTL_GIGA_MAC_VER_13)) {
+			auto_nego = ADVERTISE_100HALF | ADVERTISE_CSMA;
+		}
 	}
 
+	/* The 8100e/8101e do Fast Ethernet only. */
+	if ((tp->mac_version == RTL_GIGA_MAC_VER_13) ||
+	    (tp->mac_version == RTL_GIGA_MAC_VER_14) ||
+	    (tp->mac_version == RTL_GIGA_MAC_VER_15)) {
+		if ((giga_ctrl & (ADVERTISE_1000FULL | ADVERTISE_1000HALF)) &&
+		    netif_msg_link(tp)) {
+			printk(KERN_INFO "%s: PHY does not support 1000Mbps.\n",
+			       dev->name);
+		}
+		giga_ctrl &= ~(ADVERTISE_1000FULL | ADVERTISE_1000HALF);
+	}
+
+	auto_nego |= ADVERTISE_PAUSE_CAP | ADVERTISE_PAUSE_ASYM;
+
 	tp->phy_auto_nego_reg = auto_nego;
 	tp->phy_1000_ctrl_reg = giga_ctrl;
 
-	mdio_write(ioaddr, PHY_AUTO_NEGO_REG, auto_nego);
-	mdio_write(ioaddr, PHY_1000_CTRL_REG, giga_ctrl);
-	mdio_write(ioaddr, PHY_CTRL_REG, PHY_Enable_Auto_Nego |
-					 PHY_Restart_Auto_Nego);
+	mdio_write(ioaddr, MII_ADVERTISE, auto_nego);
+	mdio_write(ioaddr, MII_CTRL1000, giga_ctrl);
+	mdio_write(ioaddr, MII_BMCR, BMCR_ANENABLE | BMCR_ANRESTART);
 	return 0;
 }
 
@@ -788,7 +810,7 @@ static int rtl8169_set_speed(struct net_
 
 	ret = tp->set_speed(dev, autoneg, speed, duplex);
 
-	if (netif_running(dev) && (tp->phy_1000_ctrl_reg & PHY_Cap_1000_Full))
+	if (netif_running(dev) && (tp->phy_1000_ctrl_reg & ADVERTISE_1000FULL))
 		mod_timer(&tp->timer, jiffies + RTL8169_PHY_TIMEOUT);
 
 	return ret;
@@ -803,7 +825,7 @@ static int rtl8169_set_settings(struct n
 	spin_lock_irqsave(&tp->lock, flags);
 	ret = rtl8169_set_speed(dev, cmd->autoneg, cmd->speed, cmd->duplex);
 	spin_unlock_irqrestore(&tp->lock, flags);
-	
+
 	return ret;
 }
 
@@ -936,20 +958,20 @@ static void rtl8169_gset_xmii(struct net
 			 SUPPORTED_100baseT_Full |
 			 SUPPORTED_1000baseT_Full |
 			 SUPPORTED_Autoneg |
-		         SUPPORTED_TP;
+			 SUPPORTED_TP;
 
 	cmd->autoneg = 1;
 	cmd->advertising = ADVERTISED_TP | ADVERTISED_Autoneg;
 
-	if (tp->phy_auto_nego_reg & PHY_Cap_10_Half)
+	if (tp->phy_auto_nego_reg & ADVERTISE_10HALF)
 		cmd->advertising |= ADVERTISED_10baseT_Half;
-	if (tp->phy_auto_nego_reg & PHY_Cap_10_Full)
+	if (tp->phy_auto_nego_reg & ADVERTISE_10FULL)
 		cmd->advertising |= ADVERTISED_10baseT_Full;
-	if (tp->phy_auto_nego_reg & PHY_Cap_100_Half)
+	if (tp->phy_auto_nego_reg & ADVERTISE_100HALF)
 		cmd->advertising |= ADVERTISED_100baseT_Half;
-	if (tp->phy_auto_nego_reg & PHY_Cap_100_Full)
+	if (tp->phy_auto_nego_reg & ADVERTISE_100FULL)
 		cmd->advertising |= ADVERTISED_100baseT_Full;
-	if (tp->phy_1000_ctrl_reg & PHY_Cap_1000_Full)
+	if (tp->phy_1000_ctrl_reg & ADVERTISE_1000FULL)
 		cmd->advertising |= ADVERTISED_1000baseT_Full;
 
 	status = RTL_R8(PHYstatus);
@@ -961,6 +983,11 @@ static void rtl8169_gset_xmii(struct net
 	else if (status & _10bps)
 		cmd->speed = SPEED_10;
 
+	if (status & TxFlowCtrl)
+		cmd->advertising |= ADVERTISED_Asym_Pause;
+	if (status & RxFlowCtrl)
+		cmd->advertising |= ADVERTISED_Pause;
+
 	cmd->duplex = ((status & _1000bpsF) || (status & FullDup)) ?
 		      DUPLEX_FULL : DUPLEX_HALF;
 }
@@ -981,15 +1008,15 @@ static int rtl8169_get_settings(struct n
 static void rtl8169_get_regs(struct net_device *dev, struct ethtool_regs *regs,
 			     void *p)
 {
-        struct rtl8169_private *tp = netdev_priv(dev);
-        unsigned long flags;
+	struct rtl8169_private *tp = netdev_priv(dev);
+	unsigned long flags;
 
-        if (regs->len > R8169_REGS_SIZE)
-        	regs->len = R8169_REGS_SIZE;
+	if (regs->len > R8169_REGS_SIZE)
+		regs->len = R8169_REGS_SIZE;
 
-        spin_lock_irqsave(&tp->lock, flags);
-        memcpy_fromio(p, tp->mmio_addr, regs->len);
-        spin_unlock_irqrestore(&tp->lock, flags);
+	spin_lock_irqsave(&tp->lock, flags);
+	memcpy_fromio(p, tp->mmio_addr, regs->len);
+	spin_unlock_irqrestore(&tp->lock, flags);
 }
 
 static u32 rtl8169_get_msglevel(struct net_device *dev)
@@ -1071,7 +1098,7 @@ static void rtl8169_get_ethtool_stats(st
 	RTL_W32(CounterAddrLow, 0);
 	RTL_W32(CounterAddrHigh, 0);
 
-	data[0]	= le64_to_cpu(counters->tx_packets);
+	data[0] = le64_to_cpu(counters->tx_packets);
 	data[1] = le64_to_cpu(counters->rx_packets);
 	data[2] = le64_to_cpu(counters->tx_errors);
 	data[3] = le32_to_cpu(counters->rx_errors);
@@ -1098,7 +1125,7 @@ static void rtl8169_get_strings(struct n
 }
 
 
-static struct ethtool_ops rtl8169_ethtool_ops = {
+static /*const*/ struct ethtool_ops rtl8169_ethtool_ops = {
 	.get_drvinfo		= rtl8169_get_drvinfo,
 	.get_regs_len		= rtl8169_get_regs_len,
 	.get_link		= ethtool_op_get_link,
@@ -1131,7 +1158,7 @@ static void rtl8169_write_gmii_reg_bit(v
 	val = mdio_read(ioaddr, reg);
 	val = (bitval == 1) ?
 		val | (bitval << bitnum) :  val & ~(0x0001 << bitnum);
-	mdio_write(ioaddr, reg, val & 0xffff); 
+	mdio_write(ioaddr, reg, val & 0xffff);
 }
 
 static void rtl8169_get_mac_version(struct rtl8169_private *tp, void __iomem *ioaddr)
@@ -1140,10 +1167,16 @@ static void rtl8169_get_mac_version(stru
 		u32 mask;
 		int mac_version;
 	} mac_info[] = {
-		{ 0x1 << 28,	RTL_GIGA_MAC_VER_X },
-		{ 0x1 << 26,	RTL_GIGA_MAC_VER_E },
-		{ 0x1 << 23,	RTL_GIGA_MAC_VER_D }, 
-		{ 0x00000000,	RTL_GIGA_MAC_VER_B } /* Catch-all */
+		{ 0x38800000,	RTL_GIGA_MAC_VER_15 },
+		{ 0x38000000,	RTL_GIGA_MAC_VER_12 },
+		{ 0x34000000,	RTL_GIGA_MAC_VER_13 },
+		{ 0x30800000,	RTL_GIGA_MAC_VER_14 },
+		{ 0x30000000,	RTL_GIGA_MAC_VER_11 },
+		{ 0x18000000,	RTL_GIGA_MAC_VER_05 },
+		{ 0x10000000,	RTL_GIGA_MAC_VER_04 },
+		{ 0x04000000,	RTL_GIGA_MAC_VER_03 },
+		{ 0x00800000,	RTL_GIGA_MAC_VER_02 },
+		{ 0x00000000,	RTL_GIGA_MAC_VER_01 }	/* Catch-all */
 	}, *p = mac_info;
 	u32 reg;
 
@@ -1155,24 +1188,7 @@ static void rtl8169_get_mac_version(stru
 
 static void rtl8169_print_mac_version(struct rtl8169_private *tp)
 {
-	struct {
-		int version;
-		char *msg;
-	} mac_print[] = {
-		{ RTL_GIGA_MAC_VER_E, "RTL_GIGA_MAC_VER_E" },
-		{ RTL_GIGA_MAC_VER_D, "RTL_GIGA_MAC_VER_D" },
-		{ RTL_GIGA_MAC_VER_B, "RTL_GIGA_MAC_VER_B" },
-		{ 0, NULL }
-	}, *p;
-
-	for (p = mac_print; p->msg; p++) {
-		if (tp->mac_version == p->version) {
-			dprintk("mac_version == %s (%04d)\n", p->msg,
-				  p->version);
-			return;
-		}
-	}
-	dprintk("mac_version == Unknown\n");
+	dprintk("mac_version = 0x%02x\n", tp->mac_version);
 }
 
 static void rtl8169_get_phy_version(struct rtl8169_private *tp, void __iomem *ioaddr)
@@ -1189,7 +1205,7 @@ static void rtl8169_get_phy_version(stru
 	}, *p = phy_info;
 	u16 reg;
 
-	reg = mdio_read(ioaddr, 3) & 0xffff;
+	reg = mdio_read(ioaddr, MII_PHYSID2) & 0xffff;
 	while ((reg & p->mask) != p->set)
 		p++;
 	tp->phy_version = p->phy_version;
@@ -1257,7 +1273,7 @@ static void rtl8169_hw_phy_config(struct
 	rtl8169_print_mac_version(tp);
 	rtl8169_print_phy_version(tp);
 
-	if (tp->mac_version <= RTL_GIGA_MAC_VER_B)
+	if (tp->mac_version <= RTL_GIGA_MAC_VER_01)
 		return;
 	if (tp->phy_version >= RTL_GIGA_PHY_VER_H)
 		return;
@@ -1267,12 +1283,7 @@ static void rtl8169_hw_phy_config(struct
 
 	/* Shazam ! */
 
-	if (tp->mac_version == RTL_GIGA_MAC_VER_X) {
-		mdio_write(ioaddr, 31, 0x0001);
-		mdio_write(ioaddr,  9, 0x273a);
-		mdio_write(ioaddr, 14, 0x7bfb);
-		mdio_write(ioaddr, 27, 0x841e);
-
+	if (tp->mac_version == RTL_GIGA_MAC_VER_04) {
 		mdio_write(ioaddr, 31, 0x0002);
 		mdio_write(ioaddr,  1, 0x90d0);
 		mdio_write(ioaddr, 31, 0x0000);
@@ -1306,16 +1317,16 @@ static void rtl8169_phy_timer(unsigned l
 	void __iomem *ioaddr = tp->mmio_addr;
 	unsigned long timeout = RTL8169_PHY_TIMEOUT;
 
-	assert(tp->mac_version > RTL_GIGA_MAC_VER_B);
+	assert(tp->mac_version > RTL_GIGA_MAC_VER_01);
 	assert(tp->phy_version < RTL_GIGA_PHY_VER_H);
 
-	if (!(tp->phy_1000_ctrl_reg & PHY_Cap_1000_Full))
+	if (!(tp->phy_1000_ctrl_reg & ADVERTISE_1000FULL))
 		return;
 
 	spin_lock_irq(&tp->lock);
 
 	if (tp->phy_reset_pending(ioaddr)) {
-		/* 
+		/*
 		 * A busy loop could burn quite a few cycles on nowadays CPU.
 		 * Let's delay the execution of the timer for a few ticks.
 		 */
@@ -1342,7 +1353,7 @@ static inline void rtl8169_delete_timer(
 	struct rtl8169_private *tp = netdev_priv(dev);
 	struct timer_list *timer = &tp->timer;
 
-	if ((tp->mac_version <= RTL_GIGA_MAC_VER_B) ||
+	if ((tp->mac_version <= RTL_GIGA_MAC_VER_01) ||
 	    (tp->phy_version >= RTL_GIGA_PHY_VER_H))
 		return;
 
@@ -1354,7 +1365,7 @@ static inline void rtl8169_request_timer
 	struct rtl8169_private *tp = netdev_priv(dev);
 	struct timer_list *timer = &tp->timer;
 
-	if ((tp->mac_version <= RTL_GIGA_MAC_VER_B) ||
+	if ((tp->mac_version <= RTL_GIGA_MAC_VER_01) ||
 	    (tp->phy_version >= RTL_GIGA_PHY_VER_H))
 		return;
 
@@ -1391,28 +1402,111 @@ static void rtl8169_release_board(struct
 	free_netdev(dev);
 }
 
+static void rtl8169_phy_reset(struct net_device *dev,
+			      struct rtl8169_private *tp)
+{
+	void __iomem *ioaddr = tp->mmio_addr;
+	int i;
+
+	tp->phy_reset_enable(ioaddr);
+	for (i = 0; i < 100; i++) {
+		if (!tp->phy_reset_pending(ioaddr))
+			return;
+		msleep(1);
+	}
+	if (netif_msg_link(tp))
+		printk(KERN_ERR "%s: PHY reset failed.\n", dev->name);
+}
+
+static void rtl8169_init_phy(struct net_device *dev, struct rtl8169_private *tp)
+{
+	void __iomem *ioaddr = tp->mmio_addr;
+	static int board_idx = -1;
+	u8 autoneg, duplex;
+	u16 speed;
+
+	board_idx++;
+
+	rtl8169_hw_phy_config(dev);
+
+	dprintk("Set MAC Reg C+CR Offset 0x82h = 0x01h\n");
+	RTL_W8(0x82, 0x01);
+
+	if (tp->mac_version < RTL_GIGA_MAC_VER_03) {
+		dprintk("Set PCI Latency=0x40\n");
+		pci_write_config_byte(tp->pci_dev, PCI_LATENCY_TIMER, 0x40);
+	}
+
+	if (tp->mac_version == RTL_GIGA_MAC_VER_02) {
+		dprintk("Set MAC Reg C+CR Offset 0x82h = 0x01h\n");
+		RTL_W8(0x82, 0x01);
+		dprintk("Set PHY Reg 0x0bh = 0x00h\n");
+		mdio_write(ioaddr, 0x0b, 0x0000); //w 0x0b 15 0 0
+	}
+
+	rtl8169_link_option(board_idx, &autoneg, &speed, &duplex);
+
+	rtl8169_phy_reset(dev, tp);
+
+	rtl8169_set_speed(dev, autoneg, speed, duplex);
+
+	if ((RTL_R8(PHYstatus) & TBI_Enable) && netif_msg_link(tp))
+		printk(KERN_INFO PFX "%s: TBI auto-negotiating\n", dev->name);
+}
+
+static int rtl8169_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
+{
+	struct rtl8169_private *tp = netdev_priv(dev);
+	struct mii_ioctl_data *data = if_mii(ifr);
+
+	if (!netif_running(dev))
+		return -ENODEV;
+
+	switch (cmd) {
+	case SIOCGMIIPHY:
+		data->phy_id = 32; /* Internal PHY */
+		return 0;
+
+	case SIOCGMIIREG:
+		data->val_out = mdio_read(tp->mmio_addr, data->reg_num & 0x1f);
+		return 0;
+
+	case SIOCSMIIREG:
+		if (!capable(CAP_NET_ADMIN))
+			return -EPERM;
+		mdio_write(tp->mmio_addr, data->reg_num & 0x1f, data->val_in);
+		return 0;
+	}
+	return -EOPNOTSUPP;
+}
+
 static int __devinit
-rtl8169_init_board(struct pci_dev *pdev, struct net_device **dev_out,
-		   void __iomem **ioaddr_out)
+rtl8169_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 {
-	void __iomem *ioaddr;
-	struct net_device *dev;
+	const unsigned int region = rtl_cfg_info[ent->driver_data].region;
 	struct rtl8169_private *tp;
-	int rc = -ENOMEM, i, acpi_idle_state = 0, pm_cap;
+	struct net_device *dev;
+	void __iomem *ioaddr;
+	unsigned int pm_cap;
+	int i, rc;
 
-	assert(ioaddr_out != NULL);
+	if (netif_msg_drv(&debug)) {
+		printk(KERN_INFO "%s Gigabit Ethernet driver %s loaded\n",
+		       MODULENAME, RTL8169_VERSION);
+	}
 
-	/* dev zeroed in alloc_etherdev */
 	dev = alloc_etherdev(sizeof (*tp));
-	if (dev == NULL) {
+	if (!dev) {
 		if (netif_msg_drv(&debug))
 			dev_err(&pdev->dev, "unable to alloc new ethernet\n");
-		goto err_out;
+		rc = -ENOMEM;
+		goto out;
 	}
 
 	SET_MODULE_OWNER(dev);
 	SET_NETDEV_DEV(dev, &pdev->dev);
 	tp = netdev_priv(dev);
+	tp->dev = dev;
 	tp->msg_enable = netif_msg_init(debug.msg_enable, R8169_MSG_DEFAULT);
 
 	/* enable device (incl. PCI PM wakeup and hotplug setup) */
@@ -1420,48 +1514,53 @@ rtl8169_init_board(struct pci_dev *pdev,
 	if (rc < 0) {
 		if (netif_msg_probe(tp))
 			dev_err(&pdev->dev, "enable failure\n");
-		goto err_out_free_dev;
+		goto err_out_free_dev_1;
 	}
 
 	rc = pci_set_mwi(pdev);
 	if (rc < 0)
-		goto err_out_disable;
+		goto err_out_disable_2;
 
 	/* save power state before pci_enable_device overwrites it */
 	pm_cap = pci_find_capability(pdev, PCI_CAP_ID_PM);
 	if (pm_cap) {
-		u16 pwr_command;
+		u16 pwr_command, acpi_idle_state;
 
 		pci_read_config_word(pdev, pm_cap + PCI_PM_CTRL, &pwr_command);
 		acpi_idle_state = pwr_command & PCI_PM_CTRL_STATE_MASK;
 	} else {
-		if (netif_msg_probe(tp))
+		if (netif_msg_probe(tp)) {
 			dev_err(&pdev->dev,
-			       "PowerManagement capability not found.\n");
+				"PowerManagement capability not found.\n");
+		}
 	}
 
 	/* make sure PCI base addr 1 is MMIO */
-	if (!(pci_resource_flags(pdev, 1) & IORESOURCE_MEM)) {
-		if (netif_msg_probe(tp))
+	if (!(pci_resource_flags(pdev, region) & IORESOURCE_MEM)) {
+		if (netif_msg_probe(tp)) {
 			dev_err(&pdev->dev,
-			       "region #1 not an MMIO resource, aborting\n");
+				"region #%d not an MMIO resource, aborting\n",
+				region);
+		}
 		rc = -ENODEV;
-		goto err_out_mwi;
+		goto err_out_mwi_3;
 	}
+
 	/* check for weird/broken PCI region reporting */
-	if (pci_resource_len(pdev, 1) < R8169_REGS_SIZE) {
-		if (netif_msg_probe(tp))
+	if (pci_resource_len(pdev, region) < R8169_REGS_SIZE) {
+		if (netif_msg_probe(tp)) {
 			dev_err(&pdev->dev,
-			       "Invalid PCI region size(s), aborting\n");
+				"Invalid PCI region size(s), aborting\n");
+		}
 		rc = -ENODEV;
-		goto err_out_mwi;
+		goto err_out_mwi_3;
 	}
 
 	rc = pci_request_regions(pdev, MODULENAME);
 	if (rc < 0) {
 		if (netif_msg_probe(tp))
 			dev_err(&pdev->dev, "could not request regions.\n");
-		goto err_out_mwi;
+		goto err_out_mwi_3;
 	}
 
 	tp->cp_cmd = PCIMulRW | RxChkSum;
@@ -1473,22 +1572,23 @@ rtl8169_init_board(struct pci_dev *pdev,
 	} else {
 		rc = pci_set_dma_mask(pdev, DMA_32BIT_MASK);
 		if (rc < 0) {
-			if (netif_msg_probe(tp))
+			if (netif_msg_probe(tp)) {
 				dev_err(&pdev->dev,
-				       "DMA configuration failed.\n");
-			goto err_out_free_res;
+					"DMA configuration failed.\n");
+			}
+			goto err_out_free_res_4;
 		}
 	}
 
 	pci_set_master(pdev);
 
 	/* ioremap MMIO region */
-	ioaddr = ioremap(pci_resource_start(pdev, 1), R8169_REGS_SIZE);
-	if (ioaddr == NULL) {
+	ioaddr = ioremap(pci_resource_start(pdev, region), R8169_REGS_SIZE);
+	if (!ioaddr) {
 		if (netif_msg_probe(tp))
 			dev_err(&pdev->dev, "cannot remap MMIO, aborting\n");
 		rc = -EIO;
-		goto err_out_free_res;
+		goto err_out_free_res_4;
 	}
 
 	/* Unneeded ? Don't mess with Mrs. Murphy. */
@@ -1498,10 +1598,10 @@ rtl8169_init_board(struct pci_dev *pdev,
 	RTL_W8(ChipCmd, CmdReset);
 
 	/* Check that the chip has finished the reset. */
-	for (i = 1000; i > 0; i--) {
+	for (i = 100; i > 0; i--) {
 		if ((RTL_R8(ChipCmd) & CmdReset) == 0)
 			break;
-		udelay(10);
+		msleep_interruptible(1);
 	}
 
 	/* Identify chip attached to board */
@@ -1519,8 +1619,8 @@ rtl8169_init_board(struct pci_dev *pdev,
 		/* Unknown chip: assume array element #0, original RTL-8169 */
 		if (netif_msg_probe(tp)) {
 			dev_printk(KERN_DEBUG, &pdev->dev,
-			       "unknown chip version, assuming %s\n",
-			       rtl_chip_info[0].name);
+				"unknown chip version, assuming %s\n",
+				rtl_chip_info[0].name);
 		}
 		i++;
 	}
@@ -1531,56 +1631,6 @@ rtl8169_init_board(struct pci_dev *pdev,
 	RTL_W8(Config5, RTL_R8(Config5) & PMEStatus);
 	RTL_W8(Cfg9346, Cfg9346_Lock);
 
-	*ioaddr_out = ioaddr;
-	*dev_out = dev;
-out:
-	return rc;
-
-err_out_free_res:
-	pci_release_regions(pdev);
-
-err_out_mwi:
-	pci_clear_mwi(pdev);
-
-err_out_disable:
-	pci_disable_device(pdev);
-
-err_out_free_dev:
-	free_netdev(dev);
-err_out:
-	*ioaddr_out = NULL;
-	*dev_out = NULL;
-	goto out;
-}
-
-static int __devinit
-rtl8169_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
-{
-	struct net_device *dev = NULL;
-	struct rtl8169_private *tp;
-	void __iomem *ioaddr = NULL;
-	static int board_idx = -1;
-	u8 autoneg, duplex;
-	u16 speed;
-	int i, rc;
-
-	assert(pdev != NULL);
-	assert(ent != NULL);
-
-	board_idx++;
-
-	if (netif_msg_drv(&debug)) {
-		printk(KERN_INFO "%s Gigabit Ethernet driver %s loaded\n",
-		       MODULENAME, RTL8169_VERSION);
-	}
-
-	rc = rtl8169_init_board(pdev, &dev, &ioaddr);
-	if (rc)
-		return rc;
-
-	tp = netdev_priv(dev);
-	assert(ioaddr != NULL);
-
 	if (RTL_R8(PHYstatus) & TBI_Enable) {
 		tp->set_speed = rtl8169_set_speed_tbi;
 		tp->get_settings = rtl8169_gset_tbi;
@@ -1588,13 +1638,15 @@ rtl8169_init_one(struct pci_dev *pdev, c
 		tp->phy_reset_pending = rtl8169_tbi_reset_pending;
 		tp->link_ok = rtl8169_tbi_link_ok;
 
-		tp->phy_1000_ctrl_reg = PHY_Cap_1000_Full; /* Implied by TBI */
+		tp->phy_1000_ctrl_reg = ADVERTISE_1000FULL; /* Implied by TBI */
 	} else {
 		tp->set_speed = rtl8169_set_speed_xmii;
 		tp->get_settings = rtl8169_gset_xmii;
 		tp->phy_reset_enable = rtl8169_xmii_reset_enable;
 		tp->phy_reset_pending = rtl8169_xmii_reset_pending;
 		tp->link_ok = rtl8169_xmii_link_ok;
+
+		dev->do_ioctl = rtl8169_ioctl;
 	}
 
 	/* Get MAC address.  FIXME: read EEPROM */
@@ -1632,19 +1684,13 @@ rtl8169_init_one(struct pci_dev *pdev, c
 	tp->intr_mask = 0xffff;
 	tp->pci_dev = pdev;
 	tp->mmio_addr = ioaddr;
+	tp->align = rtl_cfg_info[ent->driver_data].align;
 
 	spin_lock_init(&tp->lock);
 
 	rc = register_netdev(dev);
-	if (rc) {
-		rtl8169_release_board(pdev, dev, ioaddr);
-		return rc;
-	}
-
-	if (netif_msg_probe(tp)) {
-		printk(KERN_DEBUG "%s: Identified chip type is '%s'.\n",
-		       dev->name, rtl_chip_info[tp->chipset].name);
-	}
+	if (rc < 0)
+		goto err_out_unmap_5;
 
 	pci_set_drvdata(pdev, dev);
 
@@ -1653,38 +1699,29 @@ rtl8169_init_one(struct pci_dev *pdev, c
 		       "%2.2x:%2.2x:%2.2x:%2.2x:%2.2x:%2.2x, "
 		       "IRQ %d\n",
 		       dev->name,
-		       rtl_chip_info[ent->driver_data].name,
+		       rtl_chip_info[tp->chipset].name,
 		       dev->base_addr,
 		       dev->dev_addr[0], dev->dev_addr[1],
 		       dev->dev_addr[2], dev->dev_addr[3],
 		       dev->dev_addr[4], dev->dev_addr[5], dev->irq);
 	}
 
-	rtl8169_hw_phy_config(dev);
-
-	dprintk("Set MAC Reg C+CR Offset 0x82h = 0x01h\n");
-	RTL_W8(0x82, 0x01);
-
-	if (tp->mac_version < RTL_GIGA_MAC_VER_E) {
-		dprintk("Set PCI Latency=0x40\n");
-		pci_write_config_byte(pdev, PCI_LATENCY_TIMER, 0x40);
-	}
-
-	if (tp->mac_version == RTL_GIGA_MAC_VER_D) {
-		dprintk("Set MAC Reg C+CR Offset 0x82h = 0x01h\n");
-		RTL_W8(0x82, 0x01);
-		dprintk("Set PHY Reg 0x0bh = 0x00h\n");
-		mdio_write(ioaddr, 0x0b, 0x0000); //w 0x0b 15 0 0
-	}
-
-	rtl8169_link_option(board_idx, &autoneg, &speed, &duplex);
+	rtl8169_init_phy(dev, tp);
 
-	rtl8169_set_speed(dev, autoneg, speed, duplex);
-	
-	if ((RTL_R8(PHYstatus) & TBI_Enable) && netif_msg_link(tp))
-		printk(KERN_INFO PFX "%s: TBI auto-negotiating\n", dev->name);
+out:
+	return rc;
 
-	return 0;
+err_out_unmap_5:
+	iounmap(ioaddr);
+err_out_free_res_4:
+	pci_release_regions(pdev);
+err_out_mwi_3:
+	pci_clear_mwi(pdev);
+err_out_disable_2:
+	pci_disable_device(pdev);
+err_out_free_dev_1:
+	free_netdev(dev);
+	goto out;
 }
 
 static void __devexit
@@ -1775,62 +1812,120 @@ static void rtl8169_hw_reset(void __iome
 	RTL_R8(ChipCmd);
 }
 
-static void
-rtl8169_hw_start(struct net_device *dev)
+static void rtl8169_set_rx_tx_config_registers(struct rtl8169_private *tp)
+{
+	void __iomem *ioaddr = tp->mmio_addr;
+	u32 cfg = rtl8169_rx_config;
+
+	cfg |= (RTL_R32(RxConfig) & rtl_chip_info[tp->chipset].RxConfigMask);
+	RTL_W32(RxConfig, cfg);
+
+	/* Set DMA burst size and Interframe Gap Time */
+	RTL_W32(TxConfig, (TX_DMA_BURST << TxDMAShift) |
+		(InterFrameGap << TxInterFrameGapShift));
+}
+
+static void rtl8169_hw_start(struct net_device *dev)
 {
 	struct rtl8169_private *tp = netdev_priv(dev);
 	void __iomem *ioaddr = tp->mmio_addr;
+	struct pci_dev *pdev = tp->pci_dev;
+	u16 cmd;
 	u32 i;
 
 	/* Soft reset the chip. */
 	RTL_W8(ChipCmd, CmdReset);
 
 	/* Check that the chip has finished the reset. */
-	for (i = 1000; i > 0; i--) {
+	for (i = 100; i > 0; i--) {
 		if ((RTL_R8(ChipCmd) & CmdReset) == 0)
 			break;
-		udelay(10);
+		msleep_interruptible(1);
+	}
+
+	if (tp->mac_version == RTL_GIGA_MAC_VER_05) {
+		RTL_W16(CPlusCmd, RTL_R16(CPlusCmd) | PCIMulRW);
+		pci_write_config_byte(pdev, PCI_CACHE_LINE_SIZE, 0x08);
+	}
+
+	if (tp->mac_version == RTL_GIGA_MAC_VER_13) {
+		pci_write_config_word(pdev, 0x68, 0x00);
+		pci_write_config_word(pdev, 0x69, 0x08);
+	}
+
+	/* Undocumented stuff. */
+	if (tp->mac_version == RTL_GIGA_MAC_VER_05) {
+		/* Realtek's r1000_n.c driver uses '&& 0x01' here. Well... */
+		if ((RTL_R8(Config2) & 0x07) & 0x01)
+			RTL_W32(0x7c, 0x0007ffff);
+
+		RTL_W32(0x7c, 0x0007ff00);
+
+		pci_read_config_word(pdev, PCI_COMMAND, &cmd);
+		cmd = cmd & 0xef;
+		pci_write_config_word(pdev, PCI_COMMAND, cmd);
 	}
 
 	RTL_W8(Cfg9346, Cfg9346_Unlock);
-	RTL_W8(ChipCmd, CmdTxEnb | CmdRxEnb);
+	if ((tp->mac_version == RTL_GIGA_MAC_VER_01) ||
+	    (tp->mac_version == RTL_GIGA_MAC_VER_02) ||
+	    (tp->mac_version == RTL_GIGA_MAC_VER_03) ||
+	    (tp->mac_version == RTL_GIGA_MAC_VER_04))
+		RTL_W8(ChipCmd, CmdTxEnb | CmdRxEnb);
+
 	RTL_W8(EarlyTxThres, EarlyTxThld);
 
 	/* Low hurts. Let's disable the filtering. */
 	RTL_W16(RxMaxSize, 16383);
 
-	/* Set Rx Config register */
-	i = rtl8169_rx_config |
-		(RTL_R32(RxConfig) & rtl_chip_info[tp->chipset].RxConfigMask);
-	RTL_W32(RxConfig, i);
+	if ((tp->mac_version == RTL_GIGA_MAC_VER_01) ||
+	    (tp->mac_version == RTL_GIGA_MAC_VER_02) ||
+	    (tp->mac_version == RTL_GIGA_MAC_VER_03) ||
+	    (tp->mac_version == RTL_GIGA_MAC_VER_04))
+		rtl8169_set_rx_tx_config_registers(tp);
 
-	/* Set DMA burst size and Interframe Gap Time */
-	RTL_W32(TxConfig,
-		(TX_DMA_BURST << TxDMAShift) | (InterFrameGap <<
-						TxInterFrameGapShift));
-	tp->cp_cmd |= RTL_R16(CPlusCmd);
-	RTL_W16(CPlusCmd, tp->cp_cmd);
+	cmd = RTL_R16(CPlusCmd);
+	RTL_W16(CPlusCmd, cmd);
 
-	if ((tp->mac_version == RTL_GIGA_MAC_VER_D) ||
-	    (tp->mac_version == RTL_GIGA_MAC_VER_E)) {
+	tp->cp_cmd |= cmd | PCIMulRW;
+
+	if ((tp->mac_version == RTL_GIGA_MAC_VER_02) ||
+	    (tp->mac_version == RTL_GIGA_MAC_VER_03)) {
 		dprintk(KERN_INFO PFX "Set MAC Reg C+CR Offset 0xE0. "
 			"Bit-3 and bit-14 MUST be 1\n");
-		tp->cp_cmd |= (1 << 14) | PCIMulRW;
-		RTL_W16(CPlusCmd, tp->cp_cmd);
+		tp->cp_cmd |= (1 << 14);
 	}
 
+	RTL_W16(CPlusCmd, tp->cp_cmd);
+
 	/*
 	 * Undocumented corner. Supposedly:
 	 * (TxTimer << 12) | (TxPackets << 8) | (RxTimer << 4) | RxPackets
 	 */
 	RTL_W16(IntrMitigate, 0x0000);
 
-	RTL_W32(TxDescStartAddrLow, ((u64) tp->TxPhyAddr & DMA_32BIT_MASK));
+	/*
+	 * Magic spell: some iop3xx ARM board needs the TxDescAddrHigh
+	 * register to be written before TxDescAddrLow to work.
+	 * Switching from MMIO to I/O access fixes the issue as well.
+	 */
 	RTL_W32(TxDescStartAddrHigh, ((u64) tp->TxPhyAddr >> 32));
-	RTL_W32(RxDescAddrLow, ((u64) tp->RxPhyAddr & DMA_32BIT_MASK));
+	RTL_W32(TxDescStartAddrLow, ((u64) tp->TxPhyAddr & DMA_32BIT_MASK));
 	RTL_W32(RxDescAddrHigh, ((u64) tp->RxPhyAddr >> 32));
+	RTL_W32(RxDescAddrLow, ((u64) tp->RxPhyAddr & DMA_32BIT_MASK));
+
+	if ((tp->mac_version != RTL_GIGA_MAC_VER_01) &&
+	    (tp->mac_version != RTL_GIGA_MAC_VER_02) &&
+	    (tp->mac_version != RTL_GIGA_MAC_VER_03) &&
+	    (tp->mac_version != RTL_GIGA_MAC_VER_04)) {
+		RTL_W8(ChipCmd, CmdTxEnb | CmdRxEnb);
+		rtl8169_set_rx_tx_config_registers(tp);
+	}
+
 	RTL_W8(Cfg9346, Cfg9346_Lock);
-	udelay(10);
+
+	/* Initially a 10 us delay. Turned it into a PCI commit. - FR */
+	RTL_R8(IntrMask);
 
 	RTL_W32(RxMissed, 0);
 
@@ -1910,17 +2005,18 @@ static inline void rtl8169_map_to_asic(s
 }
 
 static int rtl8169_alloc_rx_skb(struct pci_dev *pdev, struct sk_buff **sk_buff,
-				struct RxDesc *desc, int rx_buf_sz)
+				struct RxDesc *desc, int rx_buf_sz,
+				unsigned int align)
 {
 	struct sk_buff *skb;
 	dma_addr_t mapping;
 	int ret = 0;
 
-	skb = dev_alloc_skb(rx_buf_sz + NET_IP_ALIGN);
+	skb = dev_alloc_skb(rx_buf_sz + align);
 	if (!skb)
 		goto err_out;
 
-	skb_reserve(skb, NET_IP_ALIGN);
+	skb_reserve(skb, (align - 1) & (u32)skb->data);
 	*sk_buff = skb;
 
 	mapping = pci_map_single(pdev, skb->data, rx_buf_sz,
@@ -1953,15 +2049,15 @@ static u32 rtl8169_rx_fill(struct rtl816
 			   u32 start, u32 end)
 {
 	u32 cur;
-	
+
 	for (cur = start; end - cur > 0; cur++) {
 		int ret, i = cur % NUM_RX_DESC;
 
 		if (tp->Rx_skbuff[i])
 			continue;
-			
+
 		ret = rtl8169_alloc_rx_skb(tp->pci_dev, tp->Rx_skbuff + i,
-					   tp->RxDescArray + i, tp->rx_buf_sz);
+			tp->RxDescArray + i, tp->rx_buf_sz, tp->align);
 		if (ret < 0)
 			break;
 	}
@@ -2190,8 +2286,8 @@ static int rtl8169_start_xmit(struct sk_
 	dma_addr_t mapping;
 	u32 status, len;
 	u32 opts1;
-	int ret = 0;
-	
+	int ret = NETDEV_TX_OK;
+
 	if (unlikely(TX_BUFFS_AVAIL(tp) < skb_shinfo(skb)->nr_frags)) {
 		if (netif_msg_drv(tp)) {
 			printk(KERN_ERR
@@ -2255,7 +2351,7 @@ out:
 
 err_stop:
 	netif_stop_queue(dev);
-	ret = 1;
+	ret = NETDEV_TX_BUSY;
 err_update_stats:
 	tp->stats.tx_dropped++;
 	goto out;
@@ -2280,12 +2376,17 @@ static void rtl8169_pcierr_interrupt(str
 	/*
 	 * The recovery sequence below admits a very elaborated explanation:
 	 * - it seems to work;
-	 * - I did not see what else could be done.
+	 * - I did not see what else could be done;
+	 * - it makes iop3xx happy.
 	 *
 	 * Feel free to adjust to your needs.
 	 */
-	pci_write_config_word(pdev, PCI_COMMAND,
-			      pci_cmd | PCI_COMMAND_SERR | PCI_COMMAND_PARITY);
+	if (pdev->broken_parity_status)
+		pci_cmd &= ~PCI_COMMAND_PARITY;
+	else
+		pci_cmd |= PCI_COMMAND_SERR | PCI_COMMAND_PARITY;
+
+	pci_write_config_word(pdev, PCI_COMMAND, pci_cmd);
 
 	pci_write_config_word(pdev, PCI_STATUS,
 		pci_status & (PCI_STATUS_DETECTED_PARITY |
@@ -2299,10 +2400,11 @@ static void rtl8169_pcierr_interrupt(str
 		tp->cp_cmd &= ~PCIDAC;
 		RTL_W16(CPlusCmd, tp->cp_cmd);
 		dev->features &= ~NETIF_F_HIGHDMA;
-		rtl8169_schedule_work(dev, rtl8169_reinit_task);
 	}
 
 	rtl8169_hw_reset(ioaddr);
+
+	rtl8169_schedule_work(dev, rtl8169_reinit_task);
 }
 
 static void
@@ -2372,16 +2474,17 @@ static inline void rtl8169_rx_csum(struc
 }
 
 static inline int rtl8169_try_rx_copy(struct sk_buff **sk_buff, int pkt_size,
-				      struct RxDesc *desc, int rx_buf_sz)
+				      struct RxDesc *desc, int rx_buf_sz,
+				      unsigned int align)
 {
 	int ret = -1;
 
 	if (pkt_size < rx_copybreak) {
 		struct sk_buff *skb;
 
-		skb = dev_alloc_skb(pkt_size + NET_IP_ALIGN);
+		skb = dev_alloc_skb(pkt_size + align);
 		if (skb) {
-			skb_reserve(skb, NET_IP_ALIGN);
+			skb_reserve(skb, (align - 1) & (u32)skb->data);
 			eth_copy_and_sum(skb, sk_buff[0]->data, pkt_size, 0);
 			*sk_buff = skb;
 			rtl8169_mark_to_asic(desc, rx_buf_sz);
@@ -2427,6 +2530,10 @@ rtl8169_rx_interrupt(struct net_device *
 				tp->stats.rx_length_errors++;
 			if (status & RxCRC)
 				tp->stats.rx_crc_errors++;
+			if (status & RxFOVF) {
+				rtl8169_schedule_work(dev, rtl8169_reset_task);
+				tp->stats.rx_fifo_errors++;
+			}
 			rtl8169_mark_to_asic(desc, tp->rx_buf_sz);
 		} else {
 			struct sk_buff *skb = tp->Rx_skbuff[entry];
@@ -2447,13 +2554,13 @@ rtl8169_rx_interrupt(struct net_device *
 			}
 
 			rtl8169_rx_csum(skb, desc);
-			
+
 			pci_dma_sync_single_for_cpu(tp->pci_dev,
 				le64_to_cpu(desc->addr), tp->rx_buf_sz,
 				PCI_DMA_FROMDEVICE);
 
 			if (rtl8169_try_rx_copy(&skb, pkt_size, desc,
-						tp->rx_buf_sz)) {
+						tp->rx_buf_sz, tp->align)) {
 				pci_action = pci_unmap_single;
 				tp->Rx_skbuff[entry] = NULL;
 			}
@@ -2543,7 +2650,7 @@ rtl8169_interrupt(int irq, void *dev_ins
 			__netif_rx_schedule(dev);
 		else if (netif_msg_intr(tp)) {
 			printk(KERN_INFO "%s: interrupt %04x taken in poll\n",
-			       dev->name, status);	
+			       dev->name, status);
 		}
 		break;
 #else
@@ -2606,6 +2713,7 @@ static void rtl8169_down(struct net_devi
 	struct rtl8169_private *tp = netdev_priv(dev);
 	void __iomem *ioaddr = tp->mmio_addr;
 	unsigned int poll_locked = 0;
+	unsigned int intrmask;
 
 	rtl8169_delete_timer(dev);
 
@@ -2644,8 +2752,11 @@ core_down:
 	 * 2) dev->change_mtu
 	 *    -> rtl8169_poll can not be issued again and re-enable the
 	 *       interruptions. Let's simply issue the IRQ down sequence again.
+	 *
+	 * No loop if hotpluged or major error (0xffff).
 	 */
-	if (RTL_R16(IntrMask))
+	intrmask = RTL_R16(IntrMask);
+	if (intrmask && (intrmask != 0xffff))
 		goto core_down;
 
 	rtl8169_tx_clear(tp);
@@ -2716,6 +2827,15 @@ rtl8169_set_rx_mode(struct net_device *d
 	tmp = rtl8169_rx_config | rx_mode |
 	      (RTL_R32(RxConfig) & rtl_chip_info[tp->chipset].RxConfigMask);
 
+	if ((tp->mac_version == RTL_GIGA_MAC_VER_11) ||
+	    (tp->mac_version == RTL_GIGA_MAC_VER_12) ||
+	    (tp->mac_version == RTL_GIGA_MAC_VER_13) ||
+	    (tp->mac_version == RTL_GIGA_MAC_VER_14) ||
+	    (tp->mac_version == RTL_GIGA_MAC_VER_15)) {
+		mc_filter[0] = 0xffffffff;
+		mc_filter[1] = 0xffffffff;
+	}
+
 	RTL_W32(RxConfig, tmp);
 	RTL_W32(MAR0 + 0, mc_filter[0]);
 	RTL_W32(MAR0 + 4, mc_filter[1]);
@@ -2741,7 +2861,7 @@ static struct net_device_stats *rtl8169_
 		RTL_W32(RxMissed, 0);
 		spin_unlock_irqrestore(&tp->lock, flags);
 	}
-		
+
 	return &tp->stats;
 }
 
@@ -2809,7 +2929,7 @@ static struct pci_driver rtl8169_pci_dri
 static int __init
 rtl8169_init_module(void)
 {
-	return pci_module_init(&rtl8169_pci_driver);
+	return pci_register_driver(&rtl8169_pci_driver);
 }
 
 static void __exit
diff -rpuN linux-2.6.18.8/drivers/net/sfc/alaska.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/alaska.c
--- linux-2.6.18.8/drivers/net/sfc/alaska.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/alaska.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,159 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005:      Fen Systems Ltd.
+ * Copyright 2006-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include "net_driver.h"
+#include <linux/ethtool.h>
+#include "gmii.h"
+#include "phy.h"
+
+/* Marvell 88E1111 "Alaska" PHY control */
+#define ALASKA_PHY_SPECIFIC 16
+#define ALASKA_ALLOW_SLEEP 0x0200
+
+#define ALASKA_EXTENDED_CONTROL 20
+#define EXTENDED_LINE_LOOPBACK 0x8000
+
+#define ALASKA_LED_CONTROL 24
+#define LED_BLINK_MASK 0x0700
+#define LED_BLINK_FAST 0x0100
+#define LED_BLINK_SLOW 0x0300
+#define LED_TX_CTRL_MASK 0x0041
+#define LED_TX_CTRL_LINK_AND_ACTIVITY 0x0001
+
+#define ALASKA_LED_OVERRIDE 25
+#define LED_LINK1000_MASK 0x0030
+#define LED_LINK1000_BLINK 0x0010
+#define LED_TX_MASK 0x0003
+#define LED_TX_BLINK 0x0001
+
+static void alaska_reconfigure(struct efx_nic *efx)
+{
+	struct mii_if_info *gmii = &efx->mii;
+	u32 bmcr, phy_ext;
+
+	/* Configure line loopback if requested */
+	phy_ext = gmii->mdio_read(gmii->dev, gmii->phy_id,
+				  ALASKA_EXTENDED_CONTROL);
+	if (efx->loopback_mode == LOOPBACK_NETWORK)
+		phy_ext |= EXTENDED_LINE_LOOPBACK;
+	else
+		phy_ext &= ~EXTENDED_LINE_LOOPBACK;
+	gmii->mdio_write(gmii->dev, gmii->phy_id, ALASKA_EXTENDED_CONTROL,
+			 phy_ext);
+
+	/* Configure PHY loopback if requested */
+	bmcr = gmii->mdio_read(gmii->dev, gmii->phy_id, MII_BMCR);
+	if (efx->loopback_mode == LOOPBACK_PHY)
+		bmcr |= BMCR_LOOPBACK;
+	else
+		bmcr &= ~BMCR_LOOPBACK;
+	gmii->mdio_write(gmii->dev, gmii->phy_id, MII_BMCR, bmcr);
+
+	/* Read link up status */
+	if (efx->loopback_mode == LOOPBACK_NONE)
+		efx->link_up = mii_link_ok(gmii);
+	else
+		efx->link_up = 1;
+
+	/* Determine link options from PHY */
+	if (gmii->force_media) {
+		efx->link_options = gmii_forced_result(bmcr);
+	} else {
+		int lpa = gmii_lpa(gmii);
+		int adv = gmii_advertised(gmii);
+		efx->link_options = gmii_nway_result(adv & lpa);
+	}
+}
+
+static void alaska_clear_interrupt(struct efx_nic *efx)
+{
+	struct mii_if_info *gmii = &efx->mii;
+
+	/* Read interrupt status register to clear */
+	gmii->mdio_read(gmii->dev, gmii->phy_id, GMII_ISR);
+}
+
+static int alaska_init(struct efx_nic *efx)
+{
+	struct mii_if_info *gmii = &efx->mii;
+	u32 ier, leds, ctrl_1g, phy_spec;
+
+	/* Read ISR to clear any outstanding PHY interrupts */
+	gmii->mdio_read(gmii->dev, gmii->phy_id, GMII_ISR);
+
+	/* Enable PHY interrupts */
+	ier = gmii->mdio_read(gmii->dev, gmii->phy_id, GMII_IER);
+	ier |= IER_LINK_CHG;
+	gmii->mdio_write(gmii->dev, gmii->phy_id, GMII_IER, ier);
+
+	/* Remove 1G half-duplex as unsupported in Mentor MAC */
+	ctrl_1g = gmii->mdio_read(gmii->dev, gmii->phy_id, MII_CTRL1000);
+	ctrl_1g &= ~(ADVERTISE_1000HALF);
+	gmii->mdio_write(gmii->dev, gmii->phy_id, MII_CTRL1000, ctrl_1g);
+
+	/*
+	 * The PHY can save power when there is no external connection
+	 * (sleep mode).  However, this is incompatible with PHY
+	 * loopback, and if enable and disable it quickly the PHY can
+	 * go to sleep even when sleep mode is disabled.  (SFC bug
+	 * 9309.)  Therefore we disable it all the time.
+	 */
+	phy_spec = gmii->mdio_read(gmii->dev, gmii->phy_id,
+				   ALASKA_PHY_SPECIFIC);
+	phy_spec &= ~ALASKA_ALLOW_SLEEP;
+	gmii->mdio_write(gmii->dev, gmii->phy_id, ALASKA_PHY_SPECIFIC,
+			 phy_spec);
+
+	/* Configure LEDs */
+	leds = gmii->mdio_read(gmii->dev, gmii->phy_id, ALASKA_LED_CONTROL);
+	leds &= ~(LED_BLINK_MASK | LED_TX_CTRL_MASK);
+	leds |= (LED_BLINK_FAST | LED_TX_CTRL_LINK_AND_ACTIVITY);
+	gmii->mdio_write(gmii->dev, gmii->phy_id, ALASKA_LED_CONTROL, leds);
+
+	return 0;
+}
+
+static void alaska_fini(struct efx_nic *efx)
+{
+	struct mii_if_info *gmii = &efx->mii;
+	u32 ier;
+
+	/* Disable PHY interrupts */
+	ier = gmii->mdio_read(gmii->dev, gmii->phy_id, GMII_IER);
+	ier &= ~IER_LINK_CHG;
+	gmii->mdio_write(gmii->dev, gmii->phy_id, GMII_IER, ier);
+}
+
+
+struct efx_phy_operations alaska_phy_operations = {
+	.init            = alaska_init,
+	.fini            = alaska_fini,
+	.reconfigure     = alaska_reconfigure,
+	.clear_interrupt = alaska_clear_interrupt,
+	.loopbacks       = (1 << LOOPBACK_PHY) | (1 << LOOPBACK_NETWORK),
+	.startup_loopback = LOOPBACK_PHY,
+};
diff -rpuN linux-2.6.18.8/drivers/net/sfc/bitfield.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/bitfield.h
--- linux-2.6.18.8/drivers/net/sfc/bitfield.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/bitfield.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,544 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_BITFIELD_H
+#define EFX_BITFIELD_H
+
+/*
+ * Efx bitfield access
+ *
+ * Efx NICs make extensive use of bitfields up to 128 bits
+ * wide.  Since there is no native 128-bit datatype on most systems,
+ * and since 64-bit datatypes are inefficient on 32-bit systems and
+ * vice versa, we wrap accesses in a way that uses the most efficient
+ * datatype.
+ *
+ * The NICs are PCI devices and therefore little-endian.  Since most
+ * of the quantities that we deal with are DMAed to/from host memory,
+ * we define our datatypes (efx_oword_t, efx_qword_t and
+ * efx_dword_t) to be little-endian.
+ */
+
+/* Lowest bit numbers and widths */
+#define EFX_DUMMY_FIELD_LBN 0
+#define EFX_DUMMY_FIELD_WIDTH 0
+#define EFX_DWORD_0_LBN 0
+#define EFX_DWORD_0_WIDTH 32
+#define EFX_DWORD_1_LBN 32
+#define EFX_DWORD_1_WIDTH 32
+#define EFX_DWORD_2_LBN 64
+#define EFX_DWORD_2_WIDTH 32
+#define EFX_DWORD_3_LBN 96
+#define EFX_DWORD_3_WIDTH 32
+
+#define EFX_BYTE  1
+#define EFX_WORD  2
+#define EFX_DWORD 4
+#define EFX_OWORD 8
+
+/* Specified attribute (e.g. LBN) of the specified field */
+#define EFX_VAL(field, attribute) field ## _ ## attribute
+/* Low bit number of the specified field */
+#define EFX_LOW_BIT(field) EFX_VAL(field, LBN)
+/* Bit width of the specified field */
+#define EFX_WIDTH(field) EFX_VAL(field, WIDTH)
+/* High bit number of the specified field */
+#define EFX_HIGH_BIT(field) (EFX_LOW_BIT(field) + EFX_WIDTH(field) - 1)
+/* Mask equal in width to the specified field.
+ *
+ * For example, a field with width 5 would have a mask of 0x1f.
+ *
+ * The maximum width mask that can be generated is 64 bits.
+ */
+#define EFX_MASK64(field)					\
+	(EFX_WIDTH(field) == 64 ? ~((u64) 0) :		\
+	 (((((u64) 1) << EFX_WIDTH(field))) - 1))
+
+/* Mask equal in width to the specified field.
+ *
+ * For example, a field with width 5 would have a mask of 0x1f.
+ *
+ * The maximum width mask that can be generated is 32 bits.  Use
+ * EFX_MASK64 for higher width fields.
+ */
+#define EFX_MASK32(field)					\
+	(EFX_WIDTH(field) == 32 ? ~((u32) 0) :		\
+	 (((((u32) 1) << EFX_WIDTH(field))) - 1))
+
+/* A doubleword (i.e. 4 byte) datatype - little-endian in HW */
+typedef union efx_dword {
+	__le32 u32[1];
+} efx_dword_t;
+
+/* A quadword (i.e. 8 byte) datatype - little-endian in HW */
+typedef union efx_qword {
+	__le64 u64[1];
+	__le32 u32[2];
+	efx_dword_t dword[2];
+} efx_qword_t;
+
+/* An octword (eight-word, i.e. 16 byte) datatype - little-endian in HW */
+typedef union efx_oword {
+	__le64 u64[2];
+	efx_qword_t qword[2];
+	__le32 u32[4];
+	efx_dword_t dword[4];
+} efx_oword_t;
+
+/* Format string and value expanders for printk */
+#define EFX_DWORD_FMT "%08x"
+#define EFX_QWORD_FMT "%08x:%08x"
+#define EFX_OWORD_FMT "%08x:%08x:%08x:%08x"
+#define EFX_DWORD_VAL(dword)				\
+	((unsigned int) le32_to_cpu((dword).u32[0]))
+#define EFX_QWORD_VAL(qword)				\
+	((unsigned int) le32_to_cpu((qword).u32[1])),	\
+	((unsigned int) le32_to_cpu((qword).u32[0]))
+#define EFX_OWORD_VAL(oword)				\
+	((unsigned int) le32_to_cpu((oword).u32[3])),	\
+	((unsigned int) le32_to_cpu((oword).u32[2])),	\
+	((unsigned int) le32_to_cpu((oword).u32[1])),	\
+	((unsigned int) le32_to_cpu((oword).u32[0]))
+
+/*
+ * Extract bit field portion [low,high) from the native-endian element
+ * which contains bits [min,max).
+ *
+ * For example, suppose "element" represents the high 32 bits of a
+ * 64-bit value, and we wish to extract the bits belonging to the bit
+ * field occupying bits 28-45 of this 64-bit value.
+ *
+ * Then EFX_EXTRACT ( element, 32, 63, 28, 45 ) would give
+ *
+ *   ( element ) << 4
+ *
+ * The result will contain the relevant bits filled in in the range
+ * [0,high-low), with garbage in bits [high-low+1,...).
+ */
+#define EFX_EXTRACT_NATIVE(native_element, min, max, low, high)		\
+	(((low > max) || (high < min)) ? 0 :				\
+	 ((low > min) ?							\
+	  ((native_element) >> (low - min)) :				\
+	  ((native_element) << (min - low))))
+
+/*
+ * Extract bit field portion [low,high) from the 64-bit little-endian
+ * element which contains bits [min,max)
+ */
+#define EFX_EXTRACT64(element, min, max, low, high)			\
+	EFX_EXTRACT_NATIVE(le64_to_cpu(element), min, max, low, high)
+
+/*
+ * Extract bit field portion [low,high) from the 32-bit little-endian
+ * element which contains bits [min,max)
+ */
+#define EFX_EXTRACT32(element, min, max, low, high)			\
+	EFX_EXTRACT_NATIVE(le32_to_cpu(element), min, max, low, high)
+
+#define EFX_EXTRACT_OWORD64(oword, low, high)				\
+	(EFX_EXTRACT64((oword).u64[0], 0, 63, low, high) |		\
+	 EFX_EXTRACT64((oword).u64[1], 64, 127, low, high))
+
+#define EFX_EXTRACT_QWORD64(qword, low, high)				\
+	EFX_EXTRACT64((qword).u64[0], 0, 63, low, high)
+
+#define EFX_EXTRACT_OWORD32(oword, low, high)				\
+	(EFX_EXTRACT32((oword).u32[0], 0, 31, low, high) |		\
+	 EFX_EXTRACT32((oword).u32[1], 32, 63, low, high) |		\
+	 EFX_EXTRACT32((oword).u32[2], 64, 95, low, high) |		\
+	 EFX_EXTRACT32((oword).u32[3], 96, 127, low, high))
+
+#define EFX_EXTRACT_QWORD32(qword, low, high)				\
+	(EFX_EXTRACT32((qword).u32[0], 0, 31, low, high) |		\
+	 EFX_EXTRACT32((qword).u32[1], 32, 63, low, high))
+
+#define EFX_EXTRACT_DWORD(dword, low, high)				\
+	EFX_EXTRACT32((dword).u32[0], 0, 31, low, high)
+
+#define EFX_OWORD_FIELD64(oword, field)					\
+	(EFX_EXTRACT_OWORD64(oword, EFX_LOW_BIT(field), EFX_HIGH_BIT(field)) \
+	 & EFX_MASK64(field))
+
+#define EFX_QWORD_FIELD64(qword, field)					\
+	(EFX_EXTRACT_QWORD64(qword, EFX_LOW_BIT(field), EFX_HIGH_BIT(field)) \
+	 & EFX_MASK64(field))
+
+#define EFX_OWORD_FIELD32(oword, field)					\
+	(EFX_EXTRACT_OWORD32(oword, EFX_LOW_BIT(field), EFX_HIGH_BIT(field)) \
+	 & EFX_MASK32(field))
+
+#define EFX_QWORD_FIELD32(qword, field)					\
+	(EFX_EXTRACT_QWORD32(qword, EFX_LOW_BIT(field), EFX_HIGH_BIT(field)) \
+	 & EFX_MASK32(field))
+
+#define EFX_DWORD_FIELD(dword, field)					   \
+	(EFX_EXTRACT_DWORD(dword, EFX_LOW_BIT(field), EFX_HIGH_BIT(field)) \
+	 & EFX_MASK32(field))
+
+#define EFX_OWORD_IS_ZERO64(oword)					\
+	(((oword).u64[0] | (oword).u64[1]) == (__force __le64) 0)
+
+#define EFX_QWORD_IS_ZERO64(qword)					\
+	(((qword).u64[0]) == (__force __le64) 0)
+
+#define EFX_OWORD_IS_ZERO32(oword)					     \
+	(((oword).u32[0] | (oword).u32[1] | (oword).u32[2] | (oword).u32[3]) \
+	 == (__force __le32) 0)
+
+#define EFX_QWORD_IS_ZERO32(qword)					\
+	(((qword).u32[0] | (qword).u32[1]) == (__force __le32) 0)
+
+#define EFX_DWORD_IS_ZERO(dword)					\
+	(((dword).u32[0]) == (__force __le32) 0)
+
+#define EFX_OWORD_IS_ALL_ONES64(oword)					\
+	(((oword).u64[0] & (oword).u64[1]) == ~((__force __le64) 0))
+
+#define EFX_QWORD_IS_ALL_ONES64(qword)					\
+	((qword).u64[0] == ~((__force __le64) 0))
+
+#define EFX_OWORD_IS_ALL_ONES32(oword)					\
+	(((oword).u32[0] & (oword).u32[1] & (oword).u32[2] & (oword).u32[3]) \
+	 == ~((__force __le32) 0))
+
+#define EFX_QWORD_IS_ALL_ONES32(qword)					\
+	(((qword).u32[0] & (qword).u32[1]) == ~((__force __le32) 0))
+
+#define EFX_DWORD_IS_ALL_ONES(dword)					\
+	((dword).u32[0] == ~((__force __le32) 0))
+
+#if BITS_PER_LONG == 64
+#define EFX_OWORD_FIELD		EFX_OWORD_FIELD64
+#define EFX_QWORD_FIELD		EFX_QWORD_FIELD64
+#define EFX_OWORD_IS_ZERO	EFX_OWORD_IS_ZERO64
+#define EFX_QWORD_IS_ZERO	EFX_QWORD_IS_ZERO64
+#define EFX_OWORD_IS_ALL_ONES	EFX_OWORD_IS_ALL_ONES64
+#define EFX_QWORD_IS_ALL_ONES	EFX_QWORD_IS_ALL_ONES64
+#else
+#define EFX_OWORD_FIELD		EFX_OWORD_FIELD32
+#define EFX_QWORD_FIELD		EFX_QWORD_FIELD32
+#define EFX_OWORD_IS_ZERO	EFX_OWORD_IS_ZERO32
+#define EFX_QWORD_IS_ZERO	EFX_QWORD_IS_ZERO32
+#define EFX_OWORD_IS_ALL_ONES	EFX_OWORD_IS_ALL_ONES32
+#define EFX_QWORD_IS_ALL_ONES	EFX_QWORD_IS_ALL_ONES32
+#endif
+
+/*
+ * Construct bit field portion
+ *
+ * Creates the portion of the bit field [low,high) that lies within
+ * the range [min,max).
+ */
+#define EFX_INSERT_NATIVE64(min, max, low, high, value)		\
+	(((low > max) || (high < min)) ? 0 :			\
+	 ((low > min) ?						\
+	  (((u64) (value)) << (low - min)) :		\
+	  (((u64) (value)) >> (min - low))))
+
+#define EFX_INSERT_NATIVE32(min, max, low, high, value)		\
+	(((low > max) || (high < min)) ? 0 :			\
+	 ((low > min) ?						\
+	  (((u32) (value)) << (low - min)) :		\
+	  (((u32) (value)) >> (min - low))))
+
+#define EFX_INSERT_NATIVE(min, max, low, high, value)		\
+	((((max - min) >= 32) || ((high - low) >= 32)) ?	\
+	 EFX_INSERT_NATIVE64(min, max, low, high, value) :	\
+	 EFX_INSERT_NATIVE32(min, max, low, high, value))
+
+/*
+ * Construct bit field portion
+ *
+ * Creates the portion of the named bit field that lies within the
+ * range [min,max).
+ */
+#define EFX_INSERT_FIELD_NATIVE(min, max, field, value)		\
+	EFX_INSERT_NATIVE(min, max, EFX_LOW_BIT(field),		\
+			  EFX_HIGH_BIT(field), value)
+
+/*
+ * Construct bit field
+ *
+ * Creates the portion of the named bit fields that lie within the
+ * range [min,max).
+ */
+#define EFX_INSERT_FIELDS_NATIVE(min, max,				\
+				 field1, value1,			\
+				 field2, value2,			\
+				 field3, value3,			\
+				 field4, value4,			\
+				 field5, value5,			\
+				 field6, value6,			\
+				 field7, value7,			\
+				 field8, value8,			\
+				 field9, value9,			\
+				 field10, value10)			\
+	(EFX_INSERT_FIELD_NATIVE((min), (max), field1, (value1)) |	\
+	 EFX_INSERT_FIELD_NATIVE((min), (max), field2, (value2)) |	\
+	 EFX_INSERT_FIELD_NATIVE((min), (max), field3, (value3)) |	\
+	 EFX_INSERT_FIELD_NATIVE((min), (max), field4, (value4)) |	\
+	 EFX_INSERT_FIELD_NATIVE((min), (max), field5, (value5)) |	\
+	 EFX_INSERT_FIELD_NATIVE((min), (max), field6, (value6)) |	\
+	 EFX_INSERT_FIELD_NATIVE((min), (max), field7, (value7)) |	\
+	 EFX_INSERT_FIELD_NATIVE((min), (max), field8, (value8)) |	\
+	 EFX_INSERT_FIELD_NATIVE((min), (max), field9, (value9)) |	\
+	 EFX_INSERT_FIELD_NATIVE((min), (max), field10, (value10)))
+
+#define EFX_INSERT_FIELDS64(...)				\
+	cpu_to_le64(EFX_INSERT_FIELDS_NATIVE(__VA_ARGS__))
+
+#define EFX_INSERT_FIELDS32(...)				\
+	cpu_to_le32(EFX_INSERT_FIELDS_NATIVE(__VA_ARGS__))
+
+#define EFX_POPULATE_OWORD64(oword, ...) do {				\
+	(oword).u64[0] = EFX_INSERT_FIELDS64(0, 63, __VA_ARGS__);	\
+	(oword).u64[1] = EFX_INSERT_FIELDS64(64, 127, __VA_ARGS__);	\
+	} while (0)
+
+#define EFX_POPULATE_QWORD64(qword, ...) do {				\
+	(qword).u64[0] = EFX_INSERT_FIELDS64(0, 63, __VA_ARGS__);	\
+	} while (0)
+
+#define EFX_POPULATE_OWORD32(oword, ...) do {				\
+	(oword).u32[0] = EFX_INSERT_FIELDS32(0, 31, __VA_ARGS__);	\
+	(oword).u32[1] = EFX_INSERT_FIELDS32(32, 63, __VA_ARGS__);	\
+	(oword).u32[2] = EFX_INSERT_FIELDS32(64, 95, __VA_ARGS__);	\
+	(oword).u32[3] = EFX_INSERT_FIELDS32(96, 127, __VA_ARGS__);	\
+	} while (0)
+
+#define EFX_POPULATE_QWORD32(qword, ...) do {				\
+	(qword).u32[0] = EFX_INSERT_FIELDS32(0, 31, __VA_ARGS__);	\
+	(qword).u32[1] = EFX_INSERT_FIELDS32(32, 63, __VA_ARGS__);	\
+	} while (0)
+
+#define EFX_POPULATE_DWORD(dword, ...) do {				\
+	(dword).u32[0] = EFX_INSERT_FIELDS32(0, 31, __VA_ARGS__);	\
+	} while (0)
+
+#if BITS_PER_LONG == 64
+#define EFX_POPULATE_OWORD EFX_POPULATE_OWORD64
+#define EFX_POPULATE_QWORD EFX_POPULATE_QWORD64
+#else
+#define EFX_POPULATE_OWORD EFX_POPULATE_OWORD32
+#define EFX_POPULATE_QWORD EFX_POPULATE_QWORD32
+#endif
+
+/* Populate an octword field with various numbers of arguments */
+#define EFX_POPULATE_OWORD_10 EFX_POPULATE_OWORD
+#define EFX_POPULATE_OWORD_9(oword, ...) \
+	EFX_POPULATE_OWORD_10(oword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_OWORD_8(oword, ...) \
+	EFX_POPULATE_OWORD_9(oword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_OWORD_7(oword, ...) \
+	EFX_POPULATE_OWORD_8(oword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_OWORD_6(oword, ...) \
+	EFX_POPULATE_OWORD_7(oword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_OWORD_5(oword, ...) \
+	EFX_POPULATE_OWORD_6(oword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_OWORD_4(oword, ...) \
+	EFX_POPULATE_OWORD_5(oword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_OWORD_3(oword, ...) \
+	EFX_POPULATE_OWORD_4(oword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_OWORD_2(oword, ...) \
+	EFX_POPULATE_OWORD_3(oword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_OWORD_1(oword, ...) \
+	EFX_POPULATE_OWORD_2(oword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_ZERO_OWORD(oword) \
+	EFX_POPULATE_OWORD_1(oword, EFX_DUMMY_FIELD, 0)
+#define EFX_SET_OWORD(oword) \
+	EFX_POPULATE_OWORD_4(oword, \
+			     EFX_DWORD_0, 0xffffffff, \
+			     EFX_DWORD_1, 0xffffffff, \
+			     EFX_DWORD_2, 0xffffffff, \
+			     EFX_DWORD_3, 0xffffffff)
+
+/* Populate a quadword field with various numbers of arguments */
+#define EFX_POPULATE_QWORD_10 EFX_POPULATE_QWORD
+#define EFX_POPULATE_QWORD_9(qword, ...) \
+	EFX_POPULATE_QWORD_10(qword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_QWORD_8(qword, ...) \
+	EFX_POPULATE_QWORD_9(qword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_QWORD_7(qword, ...) \
+	EFX_POPULATE_QWORD_8(qword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_QWORD_6(qword, ...) \
+	EFX_POPULATE_QWORD_7(qword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_QWORD_5(qword, ...) \
+	EFX_POPULATE_QWORD_6(qword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_QWORD_4(qword, ...) \
+	EFX_POPULATE_QWORD_5(qword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_QWORD_3(qword, ...) \
+	EFX_POPULATE_QWORD_4(qword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_QWORD_2(qword, ...) \
+	EFX_POPULATE_QWORD_3(qword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_QWORD_1(qword, ...) \
+	EFX_POPULATE_QWORD_2(qword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_ZERO_QWORD(qword) \
+	EFX_POPULATE_QWORD_1(qword, EFX_DUMMY_FIELD, 0)
+#define EFX_SET_QWORD(qword) \
+	EFX_POPULATE_QWORD_2(qword, \
+			     EFX_DWORD_0, 0xffffffff, \
+			     EFX_DWORD_1, 0xffffffff)
+
+/* Populate a dword field with various numbers of arguments */
+#define EFX_POPULATE_DWORD_10 EFX_POPULATE_DWORD
+#define EFX_POPULATE_DWORD_9(dword, ...) \
+	EFX_POPULATE_DWORD_10(dword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_DWORD_8(dword, ...) \
+	EFX_POPULATE_DWORD_9(dword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_DWORD_7(dword, ...) \
+	EFX_POPULATE_DWORD_8(dword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_DWORD_6(dword, ...) \
+	EFX_POPULATE_DWORD_7(dword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_DWORD_5(dword, ...) \
+	EFX_POPULATE_DWORD_6(dword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_DWORD_4(dword, ...) \
+	EFX_POPULATE_DWORD_5(dword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_DWORD_3(dword, ...) \
+	EFX_POPULATE_DWORD_4(dword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_DWORD_2(dword, ...) \
+	EFX_POPULATE_DWORD_3(dword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_POPULATE_DWORD_1(dword, ...) \
+	EFX_POPULATE_DWORD_2(dword, EFX_DUMMY_FIELD, 0, __VA_ARGS__)
+#define EFX_ZERO_DWORD(dword) \
+	EFX_POPULATE_DWORD_1(dword, EFX_DUMMY_FIELD, 0)
+#define EFX_SET_DWORD(dword) \
+	EFX_POPULATE_DWORD_1(dword, EFX_DWORD_0, 0xffffffff)
+
+/*
+ * Modify a named field within an already-populated structure.  Used
+ * for read-modify-write operations.
+ *
+ */
+
+#define EFX_INVERT_OWORD(oword) do {		\
+	(oword).u64[0] = ~((oword).u64[0]);	\
+	(oword).u64[1] = ~((oword).u64[1]);	\
+	} while (0)
+
+#define EFX_INSERT_FIELD64(...)					\
+	cpu_to_le64(EFX_INSERT_FIELD_NATIVE(__VA_ARGS__))
+
+#define EFX_INSERT_FIELD32(...)					\
+	cpu_to_le32(EFX_INSERT_FIELD_NATIVE(__VA_ARGS__))
+
+#define EFX_INPLACE_MASK64(min, max, field)			\
+	EFX_INSERT_FIELD64(min, max, field, EFX_MASK64(field))
+
+#define EFX_INPLACE_MASK32(min, max, field)			\
+	EFX_INSERT_FIELD32(min, max, field, EFX_MASK32(field))
+
+#define EFX_SET_OWORD_FIELD64(oword, field, value) do {			\
+	(oword).u64[0] = (((oword).u64[0] 				\
+			   & ~EFX_INPLACE_MASK64(0,  63, field))	\
+			  | EFX_INSERT_FIELD64(0,  63, field, value));  \
+	(oword).u64[1] = (((oword).u64[1] 				\
+			   & ~EFX_INPLACE_MASK64(64, 127, field))	\
+			  | EFX_INSERT_FIELD64(64, 127, field, value)); \
+	} while (0)
+
+#define EFX_SET_QWORD_FIELD64(qword, field, value) do {			\
+	(qword).u64[0] = (((qword).u64[0] 				\
+			   & ~EFX_INPLACE_MASK64(0, 63, field))		\
+			  | EFX_INSERT_FIELD64(0, 63, field, value));	\
+	} while (0)
+
+#define EFX_SET_OWORD_FIELD32(oword, field, value) do {			\
+	(oword).u32[0] = (((oword).u32[0] 				\
+			   & ~EFX_INPLACE_MASK32(0, 31, field))		\
+			  | EFX_INSERT_FIELD32(0, 31, field, value));	\
+	(oword).u32[1] = (((oword).u32[1] 				\
+			   & ~EFX_INPLACE_MASK32(32, 63, field))	\
+			  | EFX_INSERT_FIELD32(32, 63, field, value));	\
+	(oword).u32[2] = (((oword).u32[2] 				\
+			   & ~EFX_INPLACE_MASK32(64, 95, field))	\
+			  | EFX_INSERT_FIELD32(64, 95, field, value));	\
+	(oword).u32[3] = (((oword).u32[3] 				\
+			   & ~EFX_INPLACE_MASK32(96, 127, field))	\
+			  | EFX_INSERT_FIELD32(96, 127, field, value));	\
+	} while (0)
+
+#define EFX_SET_QWORD_FIELD32(qword, field, value) do {			\
+	(qword).u32[0] = (((qword).u32[0] 				\
+			   & ~EFX_INPLACE_MASK32(0, 31, field))		\
+			  | EFX_INSERT_FIELD32(0, 31, field, value));	\
+	(qword).u32[1] = (((qword).u32[1] 				\
+			   & ~EFX_INPLACE_MASK32(32, 63, field))	\
+			  | EFX_INSERT_FIELD32(32, 63, field, value));	\
+	} while (0)
+
+#define EFX_SET_DWORD_FIELD(dword, field, value) do {			\
+	(dword).u32[0] = (((dword).u32[0] 				\
+			   & ~EFX_INPLACE_MASK32(0, 31, field))		\
+			  | EFX_INSERT_FIELD32(0, 31, field, value));	\
+	} while (0)
+
+#if BITS_PER_LONG == 64
+#define EFX_SET_OWORD_FIELD EFX_SET_OWORD_FIELD64
+#define EFX_SET_QWORD_FIELD EFX_SET_QWORD_FIELD64
+#else
+#define EFX_SET_OWORD_FIELD EFX_SET_OWORD_FIELD32
+#define EFX_SET_QWORD_FIELD EFX_SET_QWORD_FIELD32
+#endif
+
+#define EFX_SET_OWORD_FIELD_VER(efx, oword, field, value) do { \
+	if (FALCON_REV(efx) == FALCON_REV_B0) {			   \
+		EFX_SET_OWORD_FIELD((oword), field##_B0, (value)); \
+	} else { \
+		EFX_SET_OWORD_FIELD((oword), field##_A1, (value)); \
+	} \
+} while (0)
+
+#define EFX_QWORD_FIELD_VER(efx, qword, field)	\
+	(FALCON_REV(efx) == FALCON_REV_B0 ?	\
+	 EFX_QWORD_FIELD((qword), field##_B0) :	\
+	 EFX_QWORD_FIELD((qword), field##_A1))
+
+/* Used to avoid compiler warnings about shift range exceeding width
+ * of the data types when dma_addr_t is only 32 bits wide.
+ */
+#define DMA_ADDR_T_WIDTH	(8 * sizeof(dma_addr_t))
+#define EFX_DMA_TYPE_WIDTH(width) \
+	(((width) < DMA_ADDR_T_WIDTH) ? (width) : DMA_ADDR_T_WIDTH)
+#define EFX_DMA_MAX_MASK ((DMA_ADDR_T_WIDTH == 64) ? \
+			  ~((u64) 0) : ~((u32) 0))
+#define EFX_DMA_MASK(mask) ((mask) & EFX_DMA_MAX_MASK)
+
+/*
+ * Determine if a DMA address is over the 4GB threshold
+ *
+ * Defined in a slightly tortuous way to avoid compiler warnings.
+ */
+static inline int efx_is_over_4gb(dma_addr_t address)
+{
+	if (DMA_ADDR_T_WIDTH > 32)
+		return (((u64) address) >> 32) ? 1 : 0;
+	else
+		/* Can never be true */
+		return 0;
+}
+
+#endif /* EFX_BITFIELD_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/boards.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/boards.c
--- linux-2.6.18.8/drivers/net/sfc/boards.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/boards.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,528 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2007:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include "net_driver.h"
+#include "phy.h"
+#include "lm87_support.h"
+#include "boards.h"
+#include "efx.h"
+
+/* Macros for unpacking the board revision */
+/* The revision info is in host byte order. */
+#define BOARD_TYPE(_rev) (_rev >> 8)
+#define BOARD_MAJOR(_rev) ((_rev >> 4) & 0xf)
+#define BOARD_MINOR(_rev) (_rev & 0xf)
+
+/* Blink support. If the PHY has no auto-blink mode so we hang it off a timer */
+#define BLINK_INTERVAL (HZ/2)
+
+static void blink_led_timer(unsigned long context)
+{
+	struct efx_nic *efx = (struct efx_nic *)context;
+	struct efx_blinker *bl = &efx->board_info.blinker;
+	efx->board_info.set_fault_led(efx, bl->state);
+	bl->state = !bl->state;
+	if (bl->resubmit) {
+		bl->timer.expires = jiffies + BLINK_INTERVAL;
+		add_timer(&bl->timer);
+	}
+}
+
+static void board_blink(struct efx_nic *efx, int blink)
+{
+	struct efx_blinker *blinker = &efx->board_info.blinker;
+
+	/* The rtnl mutex serialises all ethtool ioctls, so
+	 * nothing special needs doing here. */
+	if (blink) {
+		blinker->resubmit = 1;
+		blinker->state = 0;
+		setup_timer(&blinker->timer, blink_led_timer,
+			    (unsigned long)efx);
+		blinker->timer.expires = jiffies + BLINK_INTERVAL;
+		add_timer(&blinker->timer);
+	} else {
+		blinker->resubmit = 0;
+		if (blinker->timer.function)
+			del_timer_sync(&blinker->timer);
+		efx->board_info.set_fault_led(efx, 0);
+	}
+}
+
+
+struct sensor_conf {
+	const char *name;
+	const unsigned high;
+	const unsigned low;
+};
+
+#define NO_LIMIT	((unsigned)-1)
+
+#define LM87_SENSOR_BYTES	(18)
+
+static int sensor_limits_to_bytes(const struct sensor_conf *limits,
+				  int nlimits, u8 *bytes, int maxbytes)
+{
+	int i, nbytes;
+	nbytes = 0;
+	for (i = 0; i < nlimits; i++) {
+		bytes[nbytes++] = limits[i].high;
+		if (limits[i].low != NO_LIMIT)
+			bytes[nbytes++] = limits[i].low;
+		/* We may have overrun by one at this point, but this test
+		 * should only trigger in development drivers as the sizes
+		 * are not dynamic. */
+		if (nbytes > maxbytes) {
+			printk(KERN_ERR "%s: out of space!\n", __func__);
+			break;
+		}
+	}
+	return nbytes;
+}
+
+/*****************************************************************************
+ * Support for the SFE4002
+ *
+ */
+/* LM87 configuration data for the sensor on the SFE4002 board */
+static const struct sensor_conf sfe4002_lm87_limits[] = {
+	{"1.8V line", 0x91, 0x83},	/* 2.5V sensor, scaled for 1.8V */
+	{"1.2V line", 0x5a, 0x51},	/* Vccp1 */
+	{"3.3V line", 0xca, 0xb6},
+	{"5V line", 0xc9, 0xb6},
+	{"12V line", 0xe0, 0xb0},
+	{"1V line", 0x4b, 0x44},	/* vccp2 */
+	{"Ext. temp.", 0x46, 0x0a},	/* ASIC temp. */
+	{"Int. temp.", 0x3c, 0x0a},	/* Board temp. */
+	{"1.66V line", 0xb2, NO_LIMIT},	/* AIN1 only takes 1 value */
+	{"1.5V line", 0xa1, NO_LIMIT}	/* AIN2 only takes 1 value */
+};
+
+static const int sfe4002_lm87_nlimits = ARRAY_SIZE(sfe4002_lm87_limits);
+
+static u16 sfe4002_lm87_irq_mask = EFX_LM87_NO_INTS;
+
+/* I2C ID of the onboard LM87 chip. This is board-specific as the bottom two
+ * bits are set by strap pins */
+#define SFE4002_LM87_I2C_ID (0x2e)
+
+/****************************************************************************/
+/* LED allocations. Note that on rev A0 boards the schematic and the reality
+ * differ: red and green are swapped. Below is the fixed (A1) layout (there
+ * are only 3 A0 boards in existence, so no real reason to make this
+ * conditional).
+ */
+#define SFE4002_FAULT_LED (2)	/* Red */
+#define SFE4002_RX_LED    (0)	/* Green */
+#define SFE4002_TX_LED    (1)	/* Amber */
+
+static int sfe4002_init_leds(struct efx_nic *efx)
+{
+	/* Set the TX and RX LEDs to reflect status and activity, and the
+	 * fault LED off */
+	xfp_set_led(efx, SFE4002_TX_LED,
+		    QUAKE_LED_TXLINK | QUAKE_LED_LINK_ACTSTAT);
+	xfp_set_led(efx, SFE4002_RX_LED,
+		    QUAKE_LED_RXLINK | QUAKE_LED_LINK_ACTSTAT);
+	xfp_set_led(efx, SFE4002_FAULT_LED, QUAKE_LED_OFF);
+	efx->board_info.blinker.led_num = SFE4002_FAULT_LED;
+	return 0;
+}
+
+static void sfe4002_fault_led(struct efx_nic *efx, int state)
+{
+	xfp_set_led(efx, SFE4002_FAULT_LED, state ? QUAKE_LED_ON :
+			QUAKE_LED_OFF);
+}
+
+static int sfe4002_sensor_meaning(struct efx_nic *efx, int limit_num,
+				  unsigned val)
+{
+	const struct sensor_conf *lim = &sfe4002_lm87_limits[limit_num];
+	if (lim->low == NO_LIMIT)
+		EFX_ERR(efx, "%10s  0x%02x (nominal value 0x%02x)\n", lim->name,
+			val, lim->high);
+	else
+		EFX_ERR(efx, "%10s  0x%02x (nominal range 0x%02x - 0x%02x)\n",
+			lim->name, val, lim->high, lim->low);
+	return 1;
+}
+
+static int sfe4002_check_hw(struct efx_nic *efx)
+{
+	int rc;
+
+	/* A0 board rev. 4002s  report a temperature fault the whole time
+	 * (bad sensor) so we mask it out. */
+	unsigned alarm_mask = (efx->board_info.minor > 0) ?
+		0 : ~EFX_LM87_ETMP_INT;
+
+	/* Check the sensor (NOP if not present). */
+	rc = efx_check_lm87(efx, alarm_mask);
+
+	/* We treat both lm87 interrupts and failure to talk to the lm87
+	 * as problems (since failure will only be reported if we did
+	 * find the sensor at probe time. */
+	if (rc)
+		EFX_ERR(efx, "sensor alert!\n");
+	return rc;
+}
+
+static int sfe4002_init(struct efx_nic *efx)
+{
+	u8 lm87_bytes[LM87_SENSOR_BYTES];
+	int nbytes;
+	int rc;
+
+	efx->board_info.monitor = sfe4002_check_hw;
+	efx->board_info.interpret_sensor = sfe4002_sensor_meaning;
+	efx->board_info.init_leds = sfe4002_init_leds;
+	efx->board_info.set_fault_led = sfe4002_fault_led;
+	efx->board_info.blink = board_blink;
+	/* To clean up shut down the lm87 (NOP if not present) */
+	efx->board_info.fini = efx_remove_lm87;
+
+	nbytes = sensor_limits_to_bytes(sfe4002_lm87_limits,
+					sfe4002_lm87_nlimits, lm87_bytes,
+					LM87_SENSOR_BYTES);
+
+	/* Activate the lm87 sensor if present (succeeds if nothing there) */
+	rc = efx_probe_lm87(efx, SFE4002_LM87_I2C_ID,
+			    lm87_bytes, nbytes, sfe4002_lm87_irq_mask);
+
+	return rc;
+}
+
+/*****************************************************************************
+ * Support for the SFE4003
+ *
+ */
+/* LM87 configuration data for the sensor on the SFE4003 board */
+static const struct sensor_conf sfe4003_lm87_limits[] = {
+	{"1.5V line", 0x78, 0x6d},	/* 2.5V input, values scaled for 1.5V */
+	{"1.2V line", 0x5a, 0x51},	/* Vccp1 */
+	{"3.3V line", 0xca, 0xb6},
+	{"5V line", 0xc0, 0x00},	/* Sensor not connected. */
+	{"12V line", 0xe0, 0xb0},
+	{"1V line", 0x4b, 0x44},	/* Vccp2 */
+	{"Ext. temp.", 0x46, 0x0a},	/* ASIC temp. */
+	{"Int. temp.", 0x3c, 0x0a},	/* Board temp. */
+	{"", 0xff, NO_LIMIT},		/* FAN1/AIN1 unused */
+	{"", 0xff, NO_LIMIT}		/* FAN2/AIN2 unused */
+};
+
+static const int sfe4003_lm87_nlimits = ARRAY_SIZE(sfe4003_lm87_limits);
+
+static u16 sfe4003_lm87_irq_mask = EFX_LM87_NO_INTS;
+
+
+static int sfe4003_sensor_meaning(struct efx_nic *efx, int limit_num,
+				  unsigned val)
+{
+	const struct sensor_conf *lim = &sfe4003_lm87_limits[limit_num];
+	if (lim->low == NO_LIMIT)
+		return 0; /* Neither AIN1 nor AIN2 mean anything to us */
+	else
+		EFX_ERR(efx, "%10s  0x%02x (nominal range 0x%02x - 0x%02x)\n",
+			lim->name, val, lim->high, lim->low);
+	return 1;
+}
+
+/* I2C ID of the onboard LM87 chip. This is board-specific as the bottom two
+ * bits are set by strap pins */
+#define SFE4003_LM87_I2C_ID (0x2e)
+
+/* Board-specific LED info. */
+#define SFE4003_RED_LED_GPIO	(11)
+#define SFE4003_LED_ON		(1)
+#define SFE4003_LED_OFF		(0)
+
+static void sfe4003_fault_led(struct efx_nic *efx, int state)
+{
+	/* The LEDs were not wired to GPIOs before A3 */
+	if (efx->board_info.minor < 3 && efx->board_info.major == 0)
+		return;
+
+	txc_set_gpio_val(efx, SFE4003_RED_LED_GPIO,
+			 state ? SFE4003_LED_ON : SFE4003_LED_OFF);
+}
+
+static int sfe4003_init_leds(struct efx_nic *efx)
+{
+	/* The LEDs were not wired to GPIOs before A3 */
+	if (efx->board_info.minor < 3 && efx->board_info.major == 0)
+		return 0;
+
+	txc_set_gpio_dir(efx, SFE4003_RED_LED_GPIO, TXC_GPIO_DIR_OUTPUT);
+	txc_set_gpio_val(efx, SFE4003_RED_LED_GPIO, SFE4003_LED_OFF);
+	return 0;
+}
+
+static int sfe4003_check_hw(struct efx_nic *efx)
+{
+	int rc;
+	/* A0/A1/A2 board rev. 4003s  report a temperature fault the whole time
+	 * (bad sensor) so we mask it out. */
+	unsigned alarm_mask =
+		~(EFX_LM87_ETMP_INT | EFX_LM87_FAN1_INT | EFX_LM87_FAN2_INT);
+
+	/* Check the sensor (NOP if not present). */
+
+	rc = efx_check_lm87(efx, alarm_mask);
+	/* We treat both lm87 interrupts and failure to talk to the lm87
+	 * as problems (since failure will only be reported if we did
+	 * find the sensor at probe time. */
+	if (rc)
+		EFX_ERR(efx, "sensor alert!\n");
+
+	return rc;
+}
+
+static int sfe4003_init(struct efx_nic *efx)
+{
+	u8 lm87_bytes[LM87_SENSOR_BYTES];
+	int nbytes;
+	int rc;
+	efx->board_info.monitor = sfe4003_check_hw;
+	efx->board_info.interpret_sensor = sfe4003_sensor_meaning;
+	efx->board_info.init_leds = sfe4003_init_leds;
+	efx->board_info.set_fault_led = sfe4003_fault_led;
+	efx->board_info.blink = board_blink;
+	/* To clean up shut down the lm87 (NOP if not present) */
+	efx->board_info.fini = efx_remove_lm87;
+
+	nbytes = sensor_limits_to_bytes(sfe4003_lm87_limits,
+					sfe4003_lm87_nlimits, lm87_bytes,
+					LM87_SENSOR_BYTES);
+
+	/* Activate the lm87 sensor if present (succeeds if nothing there) */
+	rc = efx_probe_lm87(efx, SFE4003_LM87_I2C_ID,
+			    lm87_bytes, nbytes, sfe4003_lm87_irq_mask);
+
+	if (rc < 0)
+		EFX_ERR(efx, "Temperature sensor probe failure: "
+			"please check the jumper position\n");
+	return rc;
+}
+
+/*****************************************************************************
+ * Support for the SFE4005
+ *
+ */
+/* LM87 configuration data for the sensor on the SFE4005 board */
+static const u8 sfe4005_lm87_limits[] = {
+	0x51, /* 2.5V high lim. (actually monitor 1.0V line, so 1050mV)  */
+	0x49, /* 2.5V low lim. (950mV) */
+	0xf6, /* Vccp1 high lim. (3.3V rail, 3465 mV) */
+	0xde, /* Vcpp1 low lim. (3.3V rail, 3135 mV) */
+	0xca, /* 3.3V AUX high lim. (3465 mV)  */
+	0xb6, /* 3.3V AUX low lim. (3135mV) */
+	0xc0, /* 5V high lim. not connected) */
+	0x00, /* 5V low lim. (not connected) */
+	0xd0, /* 12V high lim. (13000mV) */
+	0xb0, /* 12V low lim. (11000mV) */
+	0xc0, /* Vccp2 high lim. (unused) */
+	0x00, /* Vccp2 low lim. (unused) */
+	0x46, /* Ext temp 1 (ASIC) high lim. */
+	0x0a, /* Ext temp 1 low lim. */
+	0x3c, /* Int temp (board) high lim. */
+	0x0a, /* Int temp 1 low lim. */
+	0xff, /* Fan 1 high (unused) */
+	0xff, /* Fan 2 high (unused) */
+};
+
+#define SFE4005_LM87_I2C_ID (0x2e)
+
+/* Until the LM87 monitoring is interrupt driven. */
+#define SFE4005_LM87_IRQMASK	EFX_LM87_NO_INTS
+
+#define SFE4005_PCF8575_I2C_ID	(0x20)
+/* Definitions for the I/O expander that controls the CX4 chip:
+ * which PCF8575 pin maps to which function */
+#define SFE4005_PORT0_EXTLOOP	(1 << 0)
+#define SFE4005_PORT1_EXTLOOP	(1 << 1)
+#define SFE4005_HOSTPROT_LOOP	(1 << 2)
+#define SFE4005_BCAST		(1 << 3) /* TX on both ports */
+#define SFE4005_PORT0_EQ	(1 << 4)
+#define SFE4005_PORT1_EQ	(1 << 5)
+#define SFE4005_HOSTPORT_EQ	(1 << 6)
+#define	SFE4005_PORTSEL		(1 << 7) /* Which port (for RX in BCAST mode) */
+#define SFE4005_PORT0_PRE_LBN	(8)      /* Preemphasis on port 0 (2 bits)*/
+#define SFE4005_PORT1_PRE_LBN	(10)     /* Preemphasis on port 1 (2 bits)*/
+#define SFE4005_HOSTPORT_PRE_LBN (12)    /* Preemphasis on host port (2 bits) */
+#define SFE4005_UNUSED		(1 << 14)
+#define SFE4005_CX4uC_nRESET	(1 << 15) /* Reset the controller on CX4 chip */
+
+
+/* By default only turn on host port EQ. Can also OR in SFE4005_PORT0_EQ,
+ * SFE4005_PORT1_EQ but this hasn't been seen to make a difference. */
+#define SFE4005_CX4_DEFAULTS (SFE4005_CX4uC_nRESET | SFE4005_HOSTPORT_EQ)
+
+static int sfe4005_write_ioexpander(struct efx_nic *efx)
+{
+	unsigned long iobits = (unsigned long)efx->phy_data;
+	struct efx_i2c_interface *i2c = &efx->i2c;
+	u8 send[2], check[2];
+	int rc;
+	/* Do not, EVER, deassert nRESET as that will reset Falcon too,
+	 * and the driver won't know to repush the configuration, so
+	 * nothing will work until the next power cycle. */
+	BUG_ON(!(iobits & SFE4005_CX4uC_nRESET));
+	send[0] = (iobits & 0xff);
+	send[1] = ((iobits >> 8) & 0xff);
+	rc = efx_i2c_send_bytes(i2c, SFE4005_PCF8575_I2C_ID, send, 2);
+	if (rc) {
+		EFX_ERR(efx, "failed to write to I/O expander: %d\n", rc);
+		return rc;
+	}
+	/* Paranoia: just check what the I/O expander reads back */
+	rc = efx_i2c_recv_bytes(i2c, SFE4005_PCF8575_I2C_ID, check, 2);
+	if (rc)
+		EFX_ERR(efx, "failed to read back from I/O expander: %d\n", rc);
+	else if (check[0] != send[0] || check[1] != send[1])
+		EFX_ERR(efx, "read back wrong value from I/O expander: "
+			"wanted %.2x%.2x, got %.2x%.2x\n",
+			send[1], send[0], check[1], check[0]);
+	return rc;
+}
+
+static int sfe4005_init(struct efx_nic *efx)
+{
+	unsigned long iobits = SFE4005_CX4_DEFAULTS;
+	int rc;
+
+	/* There is no PHY as such on the SFE4005 so phy_data is ours. */
+	efx->phy_data = (void *)iobits;
+
+	/* Push the values */
+	rc = sfe4005_write_ioexpander(efx);
+	if (rc)
+		return rc;
+
+	/* Activate the lm87 sensor if present (succeeds if nothing there) */
+	rc = efx_probe_lm87(efx, SFE4005_LM87_I2C_ID,
+			    sfe4005_lm87_limits,
+			    sizeof(sfe4005_lm87_limits), SFE4005_LM87_IRQMASK);
+
+	/* To clean up shut down the lm87 (NOP if not present) */
+	efx->board_info.fini = efx_remove_lm87;
+
+	return rc;
+}
+
+/* This will get expanded as board-specific details get moved out of the
+ * PHY drivers. */
+struct efx_board_data {
+	const char *ref_model;
+	const char *gen_type;
+	int (*init) (struct efx_nic *nic);
+	unsigned mwatts;
+};
+
+static void dummy_fini(struct efx_nic *nic)
+{
+}
+
+static int dummy_init(struct efx_nic *nic)
+{
+	nic->board_info.fini = dummy_fini;
+	return 0;
+}
+
+/* Maximum board power (mW)
+ * Falcon controller ASIC accounts for 2.2W
+ * 10Xpress PHY accounts for 12W
+ *
+ */
+#define SFE4001_POWER 18000
+#define SFE4002_POWER 7500
+#define SFE4003_POWER 4500
+#define SFE4005_POWER 4500
+
+static struct efx_board_data board_data[] = {
+	[EFX_BOARD_INVALID] =
+	{NULL,	    NULL,                  dummy_init,      0},
+	[EFX_BOARD_SFE4001] =
+	{"SFE4001", "10GBASE-T adapter",   sfe4001_poweron, SFE4001_POWER },
+	[EFX_BOARD_SFE4002] =
+	{"SFE4002", "XFP adapter",         sfe4002_init,    SFE4002_POWER },
+	[EFX_BOARD_SFE4003] =
+	{"SFE4003", "10GBASE-CX4 adapter", sfe4003_init,    SFE4003_POWER },
+	[EFX_BOARD_SFE4005] =
+	{"SFE4005", "10G blade adapter",   sfe4005_init,    SFE4005_POWER },
+};
+
+int efx_set_board_info(struct efx_nic *efx, u16 revision_info)
+{
+	int rc = 0;
+	struct efx_board_data *data;
+
+	if (BOARD_TYPE(revision_info) >= EFX_BOARD_MAX) {
+		EFX_ERR(efx, "squashing unknown board type %d\n",
+			BOARD_TYPE(revision_info));
+		revision_info = 0;
+	}
+
+	if (BOARD_TYPE(revision_info) == 0) {
+		efx->board_info.major = 0;
+		efx->board_info.minor = 0;
+		/* For early boards that don't have revision info. there is
+		 * only 1 board for each PHY type, so we can work it out, with
+		 * the exception of the PHY-less boards. */
+		switch (efx->phy_type) {
+		case PHY_TYPE_10XPRESS:
+			efx->board_info.type = EFX_BOARD_SFE4001;
+			break;
+		case PHY_TYPE_XFP:
+			efx->board_info.type = EFX_BOARD_SFE4002;
+			break;
+		case PHY_TYPE_CX4_RTMR:
+			efx->board_info.type = EFX_BOARD_SFE4003;
+			break;
+		default:
+			efx->board_info.type = 0;
+			break;
+		}
+	} else {
+		efx->board_info.type = BOARD_TYPE(revision_info);
+		efx->board_info.major = BOARD_MAJOR(revision_info);
+		efx->board_info.minor = BOARD_MINOR(revision_info);
+	}
+
+	data = &board_data[efx->board_info.type];
+
+	/* Report the board model number or generic type for recognisable
+	 * boards. */
+	if (efx->board_info.type != 0)
+		EFX_INFO(efx, "board is %s rev %c%d\n",
+			 (efx->pci_dev->subsystem_vendor == EFX_VENDID_SFC)
+			 ? data->ref_model : data->gen_type,
+			 'A' + efx->board_info.major, efx->board_info.minor);
+
+	efx->board_info.init = data->init;
+	efx->board_info.mwatts = data->mwatts;
+
+	return rc;
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/boards.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/boards.h
--- linux-2.6.18.8/drivers/net/sfc/boards.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/boards.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,51 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2007:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_BOARDS_H
+#define EFX_BOARDS_H
+
+/* Board IDs (must fit in 8 bits). Note that 0 must never be assigned because
+ * on early boards it means there is no revision info. Board types pre 400x
+ * are not covered here, but this is not a problem because:
+ * - the early Falcon boards (FPGA, 401, 403) don't have any extra H/W we
+ * need care about and aren't being updated.
+ */
+enum efx_board_type {
+	EFX_BOARD_INVALID = 0, /* Early boards do not have board rev. info. */
+	EFX_BOARD_SFE4001 = 1,
+	EFX_BOARD_SFE4002 = 2,
+	EFX_BOARD_SFE4003 = 3,
+	EFX_BOARD_SFE4005 = 4,
+	/* Insert new types before here */
+	EFX_BOARD_MAX
+};
+
+extern int efx_set_board_info(struct efx_nic *efx, u16 revision_info);
+
+/* SFE4001 (10GBASE-T) */
+extern int sfe4001_poweron(struct efx_nic *efx);
+extern void sfe4001_poweroff(struct efx_nic *efx);
+
+#endif
diff -rpuN linux-2.6.18.8/drivers/net/sfc/config.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/config.h
--- linux-2.6.18.8/drivers/net/sfc/config.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/config.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1 @@
+/* SFC config options can go here */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/debugfs.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/debugfs.c
--- linux-2.6.18.8/drivers/net/sfc/debugfs.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/debugfs.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,924 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/module.h>
+#include <linux/pci.h>
+/* For out-of-tree builds we always need procfs, if only for a compatibility
+ * symlink.
+ */
+#include <linux/proc_fs.h>
+#include <linux/dcache.h>
+#include <linux/seq_file.h>
+#include "net_driver.h"
+#include "efx.h"
+#include "debugfs.h"
+#include "falcon.h"
+
+/* EFX_USE_DEBUGFS is defined by kernel_compat.h so we can't decide whether to
+ * include this earlier.
+ */
+#ifdef EFX_USE_DEBUGFS
+#include <linux/debugfs.h>
+#endif
+
+#ifndef PRIu64
+#	if (BITS_PER_LONG == 64)
+#		define PRIu64 "lu"
+#	else
+#		define PRIu64 "llu"
+#	endif
+#endif
+
+#ifndef EFX_USE_DEBUGFS
+
+static void efx_debugfs_remove(struct proc_dir_entry *entry)
+{
+	if (entry)
+		remove_proc_entry(entry->name, entry->parent);
+}
+#define debugfs_remove efx_debugfs_remove
+
+#define debugfs_create_dir proc_mkdir
+#define debugfs_create_symlink proc_symlink
+
+#endif /* !EFX_USE_DEBUGFS */
+
+/* Parameter definition bound to a structure - each file has one of these */
+struct efx_debugfs_bound_param {
+	const struct efx_debugfs_parameter *param;
+	void *structure;
+};
+
+
+/* Maximum length for a name component or symlink target */
+#define EFX_DEBUGFS_NAME_LEN 32
+
+
+/* Top-level debug directory ([/sys/kernel]/debug/sfc) */
+static struct dentry *efx_debug_root;
+
+/* "cards" directory ([/sys/kernel]/debug/sfc/cards) */
+static struct dentry *efx_debug_cards;
+
+
+/* Sequential file interface to bound parameters */
+
+#if defined(EFX_USE_DEBUGFS)
+
+static int efx_debugfs_seq_show(struct seq_file *file, void *v)
+{
+	struct efx_debugfs_bound_param *binding =
+		(struct efx_debugfs_bound_param *)file->private;
+
+	return binding->param->reader(file,
+				      binding->structure +
+				      binding->param->offset);
+}
+
+static int efx_debugfs_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, efx_debugfs_seq_show, inode->i_private);
+}
+
+#else /* EFX_NOT_UPSTREAM && !EFX_USE_DEBUGFS */
+
+static int efx_debugfs_seq_show(struct seq_file *file, void *v)
+{
+	struct proc_dir_entry *entry = (struct proc_dir_entry *)file->private;
+	struct efx_debugfs_parameter *param =
+		(struct efx_debugfs_parameter *)entry->data;
+	void *structure = (void *)entry->read_proc;
+
+	if (!structure)
+		return -EIO;
+
+	return param->reader(file, structure + param->offset);
+}
+
+static int efx_debugfs_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, efx_debugfs_seq_show, PROC_I(inode)->pde);
+}
+
+#endif /* !EFX_NOT_UPSTREAM || EFX_USE_DEBUGFS */
+
+
+static struct file_operations efx_debugfs_file_ops = {
+	.owner   = THIS_MODULE,
+	.open    = efx_debugfs_open,
+	.read    = seq_read,
+	.llseek  = seq_lseek,
+	.release = seq_release
+};
+
+
+#if defined(EFX_USE_DEBUGFS)
+
+/**
+ * efx_fini_debugfs_child - remove a named child of a debugfs directory
+ * @dir:		Directory
+ * @name:		Name of child
+ *
+ * This removes the named child from the directory, if it exists.
+ */
+void efx_fini_debugfs_child(struct dentry *dir, const char *name)
+{
+	struct qstr child_name;
+	struct dentry *child;
+
+	child_name.len = strlen(name);
+	child_name.name = name;
+	child_name.hash = full_name_hash(child_name.name, child_name.len);
+	child = d_lookup(dir, &child_name);
+	if (child) {
+		/* If it's a "regular" file, free its parameter binding */
+		if (S_ISREG(child->d_inode->i_mode))
+			kfree(child->d_inode->i_private);
+		debugfs_remove(child);
+		dput(child);
+	}
+}
+
+#else /* EFX_NOT_UPSTREAM && !EFX_USE_DEBUGFS */
+
+void efx_fini_debugfs_child(struct proc_dir_entry *dir, const char *name)
+{
+	remove_proc_entry(name, dir);
+}
+
+#endif /* !EFX_NOT_UPSTREAM || EFX_USE_DEBUGFS */
+
+/*
+ * Remove a debugfs directory.
+ *
+ * This removes the named parameter-files and sym-links from the
+ * directory, and the directory itself.  It does not do any recursion
+ * to subdirectories.
+ */
+static void efx_fini_debugfs_dir(struct dentry *dir,
+				 struct efx_debugfs_parameter *params,
+				 const char *const *symlink_names)
+{
+	if (!dir)
+		return;
+
+	while (params->name) {
+		efx_fini_debugfs_child(dir, params->name);
+		params++;
+	}
+	while (symlink_names && *symlink_names) {
+		efx_fini_debugfs_child(dir, *symlink_names);
+		symlink_names++;
+	}
+	debugfs_remove(dir);
+}
+
+/* Functions for printing various types of parameter. */
+
+int efx_debugfs_read_uint(struct seq_file *file, void *data)
+{
+	return seq_printf(file, "%#x\n", *(unsigned int *)data);
+}
+
+int efx_debugfs_read_int(struct seq_file *file, void *data)
+{
+	return seq_printf(file, "%d\n", *(int *)data);
+}
+
+int efx_debugfs_read_atomic(struct seq_file *file, void *data)
+{
+	unsigned int value = atomic_read((atomic_t *) data);
+
+	return seq_printf(file, "%#x\n", value);
+}
+
+int efx_debugfs_read_dword(struct seq_file *file, void *data)
+{
+	unsigned int value = EFX_DWORD_FIELD(*(efx_dword_t *) data,
+					     EFX_DWORD_0);
+
+	return seq_printf(file, "%#x\n", value);
+}
+
+static int efx_debugfs_read_int_mode(struct seq_file *file, void *data)
+{
+	unsigned int value = *(enum efx_int_mode *) data;
+
+	return seq_printf(file, "%d => %s\n", value,
+			  STRING_TABLE_LOOKUP(value, efx_interrupt_mode));
+}
+
+#define EFX_INT_MODE_PARAMETER(container_type, parameter)		\
+	EFX_PARAMETER(container_type, parameter,			\
+		      enum efx_int_mode, efx_debugfs_read_int_mode)
+
+static int efx_debugfs_read_loop_mode(struct seq_file *file, void *data)
+{
+	unsigned int value = *(enum efx_loopback_mode *)data;
+
+	return seq_printf(file, "%d => %s\n", value,
+			  STRING_TABLE_LOOKUP(value, efx_loopback_mode));
+}
+
+#define EFX_LOOPBACK_MODE_PARAMETER(container_type, parameter)		\
+	EFX_PARAMETER(container_type, parameter,			\
+		      enum efx_loopback_mode, efx_debugfs_read_loop_mode)
+
+static int efx_debugfs_read_phy_type(struct seq_file *file, void *data)
+{
+	unsigned int value = *(enum phy_type *) data;
+
+	return seq_printf(file, "%d => %s\n", value,
+			  STRING_TABLE_LOOKUP(value, efx_phy_type));
+}
+
+#define EFX_PHY_TYPE_PARAMETER(container_type, parameter)		\
+	EFX_PARAMETER(container_type, parameter,			\
+		      enum phy_type, efx_debugfs_read_phy_type)
+
+int efx_debugfs_read_string(struct seq_file *file, void *data)
+{
+	return seq_puts(file, (const char *)data);
+}
+
+
+/**
+ * efx_init_debugfs_files - create parameter-files in a debugfs directory
+ * @parent:		Containing directory
+ * @params:		Pointer to zero-terminated parameter definition array
+ * @structure:		Structure containing parameters
+ *
+ * Add parameter-files to the given debugfs directory.  Return a
+ * negative error code or 0 on success.
+ */
+static int efx_init_debugfs_files(struct dentry *parent,
+				  struct efx_debugfs_parameter *params,
+				  void *structure)
+{
+	struct efx_debugfs_parameter *param = params;
+
+	while (param->name) {
+		struct dentry *entry;
+#if defined(EFX_USE_DEBUGFS)
+		struct efx_debugfs_bound_param *binding;
+
+		binding = kmalloc(sizeof(*binding), GFP_KERNEL);
+		if (!binding)
+			goto err;
+		binding->param = param;
+		binding->structure = structure;
+
+		entry = debugfs_create_file(param->name, S_IRUGO, parent,
+					    binding, &efx_debugfs_file_ops);
+		if (!entry) {
+			kfree(binding);
+			goto err;
+		}
+#else
+		entry = create_proc_entry(param->name, S_IRUGO, parent);
+		if (!entry)
+			goto err;
+		/*
+		 * We have no good way to free a binding created here.
+		 * However, once we install our file_operations the
+		 * read_proc pointer becomes redundant and we can
+		 * abuse it as a structure pointer.
+		 */
+		entry->data = param;
+		entry->read_proc = NULL;
+		smp_wmb();
+		entry->proc_fops = &efx_debugfs_file_ops;
+		smp_wmb();
+		entry->read_proc = (read_proc_t *) structure;
+#endif
+
+		param++;
+	}
+
+	return 0;
+
+ err:
+	while (param != params) {
+		param--;
+		efx_fini_debugfs_child(parent, param->name);
+	}
+	return -ENOMEM;
+}
+
+/**
+ * efx_init_debugfs_netdev - create debugfs sym-links for net device
+ * @net_dev:		Net device
+ *
+ * Create sym-links named after @net_dev to the debugfs directories for
+ * the corresponding NIC and  port.  Return a negative error code or 0 on
+ * success.  The sym-links must be cleaned up using
+ * efx_fini_debugfs_netdev().
+ */
+int efx_init_debugfs_netdev(struct net_device *net_dev)
+{
+	struct efx_nic *efx = net_dev->priv;
+	char name[EFX_DEBUGFS_NAME_LEN];
+	char target[EFX_DEBUGFS_NAME_LEN];
+	size_t len;
+
+	if (snprintf(name, sizeof(name), "nic_%s", net_dev->name) >=
+	    sizeof(name))
+		return -ENAMETOOLONG;
+	if (snprintf(target, sizeof(target), "cards/%s", pci_name(efx->pci_dev))
+	    >= sizeof(target))
+		return -ENAMETOOLONG;
+	efx->debug_symlink = debugfs_create_symlink(name,
+						    efx_debug_root, target);
+	if (!efx->debug_symlink)
+		return -ENOMEM;
+
+	if (snprintf(name, sizeof(name), "if_%s", net_dev->name) >=
+	    sizeof(name))
+		return -ENAMETOOLONG;
+	len = snprintf(target, sizeof(target),
+		       "cards/%s/port0", pci_name(efx->pci_dev));
+	if (len >= sizeof(target))
+		return -ENAMETOOLONG;
+	efx->debug_port_symlink = debugfs_create_symlink(name,
+							 efx_debug_root,
+							 target);
+	if (!efx->debug_port_symlink)
+		return -ENOMEM;
+
+	return 0;
+}
+
+/**
+ * efx_fini_debugfs_netdev - remove debugfs sym-links for net device
+ * @net_dev:		Net device
+ *
+ * Remove sym-links created for @net_dev by efx_init_debugfs_netdev().
+ */
+void efx_fini_debugfs_netdev(struct net_device *net_dev)
+{
+	struct efx_nic *efx = net_dev->priv;
+
+	debugfs_remove(efx->debug_port_symlink);
+	efx->debug_port_symlink = NULL;
+	debugfs_remove(efx->debug_symlink);
+	efx->debug_symlink = NULL;
+}
+
+/* Per-port parameters */
+static struct efx_debugfs_parameter efx_debugfs_port_parameters[] = {
+	EFX_NAMED_PARAMETER(enabled, struct efx_nic, port_enabled,
+			    int, efx_debugfs_read_int),
+	EFX_INT_PARAMETER(struct efx_nic, net_dev_registered),
+	EFX_INT_PARAMETER(struct efx_nic, rx_checksum_enabled),
+	EFX_ATOMIC_PARAMETER(struct efx_nic, netif_stop_count),
+	EFX_INT_PARAMETER(struct efx_nic, link_up),
+	EFX_UINT_PARAMETER(struct efx_nic, link_options),
+	EFX_INT_PARAMETER(struct efx_nic, promiscuous),
+	EFX_UINT_PARAMETER(struct efx_nic, loopback_modes),
+	EFX_LOOPBACK_MODE_PARAMETER(struct efx_nic, loopback_mode),
+	EFX_PHY_TYPE_PARAMETER(struct efx_nic, phy_type),
+	EFX_NAMED_PARAMETER(phy_id, struct efx_nic, mii.phy_id,
+			    int, efx_debugfs_read_int),
+	EFX_UINT_PARAMETER(struct efx_nic, n_link_state_changes),
+	{NULL},
+};
+
+/**
+ * efx_init_debugfs_port - create debugfs directory for port
+ * @efx:		Efx NIC
+ *
+ * Create a debugfs directory containing parameter-files for @efx.
+ * Return a negative error code or 0 on success.  The directory must be
+ * cleaned up using efx_fini_debugfs_port().
+ */
+int efx_init_debugfs_port(struct efx_nic *efx)
+{
+	int rc;
+
+	/* Create directory */
+	efx->debug_port_dir = debugfs_create_dir("port0", efx->debug_dir);
+	if (!efx->debug_port_dir)
+		return -ENOMEM;
+
+	/* Create files */
+	rc = efx_init_debugfs_files(efx->debug_port_dir,
+				    efx_debugfs_port_parameters,
+				    (void *)efx);
+	if (rc)
+		efx_fini_debugfs_port(efx);
+
+	return rc;
+}
+
+/**
+ * efx_fini_debugfs_port - remove debugfs directory for port
+ * @efx:		Efx NIC
+ *
+ * Remove directory created for @efx by efx_init_debugfs_port().
+ */
+void efx_fini_debugfs_port(struct efx_nic *efx)
+{
+	efx_fini_debugfs_dir(efx->debug_port_dir,
+			     efx_debugfs_port_parameters, NULL);
+	efx->debug_port_dir = NULL;
+}
+
+/**
+ * efx_extend_debugfs_port - add parameter-files to directory for port
+ * @efx:		Efx NIC
+ * @structure:		Structure containing parameters
+ * @params:		Pointer to zero-terminated parameter definition array
+ *
+ * Add parameter-files to the debugfs directory for @efx.  Return
+ * a negative error code or 0 on success.  This is intended for
+ * PHY-specific parameters.  The files must be cleaned up using
+ * efx_trim_debugfs_port().
+ */
+int efx_extend_debugfs_port(struct efx_nic *efx,
+			    void *structure,
+			    struct efx_debugfs_parameter *params)
+{
+	return efx_init_debugfs_files(efx->debug_port_dir, params, structure);
+}
+
+/**
+ * efx_trim_debugfs_port - remove parameter-files from directory for port
+ * @efx:		Efx NIC
+ * @params:		Pointer to zero-terminated parameter definition array
+ *
+ * Remove parameter-files previously added to the debugfs directory
+ * for @efx using efx_extend_debugfs_port().
+ */
+void efx_trim_debugfs_port(struct efx_nic *efx,
+			   struct efx_debugfs_parameter *params)
+{
+	struct dentry *dir = efx->debug_port_dir;
+
+	if (dir) {
+		struct efx_debugfs_parameter *field;
+		for (field = params; field->name; field++)
+			efx_fini_debugfs_child(dir, field->name);
+	}
+}
+
+/* Per-TX-queue parameters */
+static struct efx_debugfs_parameter efx_debugfs_tx_queue_parameters[] = {
+	EFX_UINT_PARAMETER(struct efx_tx_queue, insert_count),
+	EFX_UINT_PARAMETER(struct efx_tx_queue, write_count),
+	EFX_UINT_PARAMETER(struct efx_tx_queue, read_count),
+	EFX_INT_PARAMETER(struct efx_tx_queue, stopped),
+	{NULL},
+};
+
+static void efx_fini_debugfs_tx_queue(struct efx_tx_queue *tx_queue);
+
+/**
+ * efx_init_debugfs_tx_queue - create debugfs directory for TX queue
+ * @tx_queue:		Efx TX queue
+ *
+ * Create a debugfs directory containing parameter-files for @tx_queue.
+ * Return a negative error code or 0 on success.  The directory must be
+ * cleaned up using efx_fini_debugfs_tx_queue().
+ */
+static int efx_init_debugfs_tx_queue(struct efx_tx_queue *tx_queue)
+{
+	char name[EFX_DEBUGFS_NAME_LEN];
+	char target[EFX_DEBUGFS_NAME_LEN];
+	int rc;
+
+	/* Create directory */
+	if (snprintf(name, sizeof(name), EFX_TX_QUEUE_NAME(tx_queue))
+	    >= sizeof(name))
+		goto err_len;
+	tx_queue->debug_dir = debugfs_create_dir(name,
+						 tx_queue->efx->debug_dir);
+	if (!tx_queue->debug_dir)
+		goto err_mem;
+
+	/* Create files */
+	rc = efx_init_debugfs_files(tx_queue->debug_dir,
+				    efx_debugfs_tx_queue_parameters,
+				    (void *)tx_queue);
+	if (rc)
+		goto err;
+
+	/* Create symlink to channel */
+	if (snprintf(target, sizeof(target),
+		     "../" EFX_CHANNEL_NAME(tx_queue->channel)) >=
+	    sizeof(target))
+		goto err_len;
+	if (!debugfs_create_symlink("channel", tx_queue->debug_dir, target))
+		goto err_mem;
+
+	/* Create symlink to port */
+	if (!debugfs_create_symlink("port", tx_queue->debug_dir, "../port0"))
+		goto err_mem;
+
+	return 0;
+
+ err_len:
+	rc = -ENAMETOOLONG;
+	goto err;
+ err_mem:
+	rc = -ENOMEM;
+ err:
+	efx_fini_debugfs_tx_queue(tx_queue);
+	return rc;
+}
+
+/**
+ * efx_fini_debugfs_tx_queue - remove debugfs directory for TX queue
+ * @tx_queue:		Efx TX queue
+ *
+ * Remove directory created for @tx_queue by efx_init_debugfs_tx_queue().
+ */
+static void efx_fini_debugfs_tx_queue(struct efx_tx_queue *tx_queue)
+{
+	static const char *const symlink_names[] = {
+		"channel", "port", NULL
+	};
+
+	efx_fini_debugfs_dir(tx_queue->debug_dir,
+			     efx_debugfs_tx_queue_parameters, symlink_names);
+	tx_queue->debug_dir = NULL;
+}
+
+/* Per-RX-queue parameters */
+static struct efx_debugfs_parameter efx_debugfs_rx_queue_parameters[] = {
+	EFX_INT_PARAMETER(struct efx_rx_queue, added_count),
+	EFX_INT_PARAMETER(struct efx_rx_queue, removed_count),
+	EFX_UINT_PARAMETER(struct efx_rx_queue, max_fill),
+	EFX_UINT_PARAMETER(struct efx_rx_queue, fast_fill_trigger),
+	EFX_UINT_PARAMETER(struct efx_rx_queue, fast_fill_limit),
+	EFX_UINT_PARAMETER(struct efx_rx_queue, min_fill),
+	EFX_UINT_PARAMETER(struct efx_rx_queue, min_overfill),
+	EFX_UINT_PARAMETER(struct efx_rx_queue, alloc_page_count),
+	EFX_UINT_PARAMETER(struct efx_rx_queue, alloc_skb_count),
+	EFX_UINT_PARAMETER(struct efx_rx_queue, slow_fill_count),
+	{NULL},
+};
+
+static void efx_fini_debugfs_rx_queue(struct efx_rx_queue *rx_queue);
+
+/**
+ * efx_init_debugfs_rx_queue - create debugfs directory for RX queue
+ * @rx_queue:		Efx RX queue
+ *
+ * Create a debugfs directory containing parameter-files for @rx_queue.
+ * Return a negative error code or 0 on success.  The directory must be
+ * cleaned up using efx_fini_debugfs_rx_queue().
+ */
+static int efx_init_debugfs_rx_queue(struct efx_rx_queue *rx_queue)
+{
+	char name[EFX_DEBUGFS_NAME_LEN];
+	char target[EFX_DEBUGFS_NAME_LEN];
+	int rc;
+
+	/* Create directory */
+	if (snprintf(name, sizeof(name), EFX_RX_QUEUE_NAME(rx_queue))
+	    >= sizeof(name))
+		goto err_len;
+	rx_queue->debug_dir = debugfs_create_dir(name,
+						 rx_queue->efx->debug_dir);
+	if (!rx_queue->debug_dir)
+		goto err_mem;
+
+	/* Create files */
+	rc = efx_init_debugfs_files(rx_queue->debug_dir,
+				    efx_debugfs_rx_queue_parameters,
+				    (void *)rx_queue);
+	if (rc)
+		goto err;
+
+	/* Create symlink to channel */
+	if (snprintf(target, sizeof(target),
+		     "../" EFX_CHANNEL_NAME(rx_queue->channel)) >=
+	    sizeof(target))
+		goto err_len;
+	if (!debugfs_create_symlink("channel", rx_queue->debug_dir, target))
+		goto err_mem;
+
+	return 0;
+
+ err_len:
+	rc = -ENAMETOOLONG;
+	goto err;
+ err_mem:
+	rc = -ENOMEM;
+ err:
+	efx_fini_debugfs_rx_queue(rx_queue);
+	return rc;
+}
+
+/**
+ * efx_fini_debugfs_rx_queue - remove debugfs directory for RX queue
+ * @rx_queue:		Efx RX queue
+ *
+ * Remove directory created for @rx_queue by efx_init_debugfs_rx_queue().
+ */
+static void efx_fini_debugfs_rx_queue(struct efx_rx_queue *rx_queue)
+{
+	const char *const symlink_names[] = {
+		"channel", NULL
+	};
+
+	efx_fini_debugfs_dir(rx_queue->debug_dir,
+			     efx_debugfs_rx_queue_parameters, symlink_names);
+	rx_queue->debug_dir = NULL;
+}
+
+/* Per-channel parameters */
+static struct efx_debugfs_parameter efx_debugfs_channel_parameters[] = {
+	EFX_INT_PARAMETER(struct efx_channel, enabled),
+	EFX_INT_PARAMETER(struct efx_channel, irq),
+	EFX_UINT_PARAMETER(struct efx_channel, has_interrupt),
+	EFX_UINT_PARAMETER(struct efx_channel, irq_moderation),
+	EFX_UINT_PARAMETER(struct efx_channel, eventq_read_ptr),
+	EFX_UINT_PARAMETER(struct efx_channel, n_rx_tobe_disc),
+	EFX_UINT_PARAMETER(struct efx_channel, n_rx_ip_frag_err),
+	EFX_UINT_PARAMETER(struct efx_channel, n_rx_ip_hdr_chksum_err),
+	EFX_UINT_PARAMETER(struct efx_channel, n_rx_tcp_udp_chksum_err),
+	EFX_UINT_PARAMETER(struct efx_channel, n_rx_frm_trunc),
+	EFX_UINT_PARAMETER(struct efx_channel, n_rx_overlength),
+	EFX_UINT_PARAMETER(struct efx_channel, n_skbuff_leaks),
+	EFX_INT_PARAMETER(struct efx_channel, rx_alloc_level),
+	EFX_INT_PARAMETER(struct efx_channel, rx_alloc_push_pages),
+	EFX_INT_PARAMETER(struct efx_channel, rx_alloc_pop_pages),
+	{NULL},
+};
+
+static void efx_fini_debugfs_channel(struct efx_channel *channel);
+
+/**
+ * efx_init_debugfs_channel - create debugfs directory for channel
+ * @channel:		Efx channel
+ *
+ * Create a debugfs directory containing parameter-files for @channel.
+ * Return a negative error code or 0 on success.  The directory must be
+ * cleaned up using efx_fini_debugfs_channel().
+ */
+static int efx_init_debugfs_channel(struct efx_channel *channel)
+{
+	char name[EFX_DEBUGFS_NAME_LEN];
+	int rc;
+
+	/* Create directory */
+	if (snprintf(name, sizeof(name), EFX_CHANNEL_NAME(channel))
+	    >= sizeof(name))
+		goto err_len;
+	channel->debug_dir = debugfs_create_dir(name, channel->efx->debug_dir);
+	if (!channel->debug_dir)
+		goto err_mem;
+
+	/* Create files */
+	rc = efx_init_debugfs_files(channel->debug_dir,
+				    efx_debugfs_channel_parameters,
+				    (void *)channel);
+	if (rc)
+		goto err;
+
+	return 0;
+
+ err_len:
+	rc = -ENAMETOOLONG;
+	goto err;
+ err_mem:
+	rc = -ENOMEM;
+ err:
+	efx_fini_debugfs_channel(channel);
+	return rc;
+}
+
+/**
+ * efx_fini_debugfs_channel - remove debugfs directory for channel
+ * @channel:		Efx channel
+ *
+ * Remove directory created for @channel by efx_init_debugfs_channel().
+ */
+static void efx_fini_debugfs_channel(struct efx_channel *channel)
+{
+	efx_fini_debugfs_dir(channel->debug_dir,
+			     efx_debugfs_channel_parameters, NULL);
+	channel->debug_dir = NULL;
+}
+
+/* Per-NIC parameters */
+static struct efx_debugfs_parameter efx_debugfs_nic_parameters[] = {
+	EFX_INT_PARAMETER(struct efx_nic, legacy_irq),
+	EFX_INT_PARAMETER(struct efx_nic, rss_queues),
+	EFX_UINT_PARAMETER(struct efx_nic, rx_buffer_len),
+	EFX_INT_MODE_PARAMETER(struct efx_nic, interrupt_mode),
+	{.name = "hardware_desc",
+	 .offset = 0,
+	 .reader = falcon_debugfs_read_hardware_desc},
+	{NULL},
+};
+
+/* Per-NIC error counts */
+static struct efx_debugfs_parameter efx_debugfs_nic_error_parameters[] = {
+	EFX_ATOMIC_PARAMETER(struct efx_nic_errors, missing_event),
+	EFX_ATOMIC_PARAMETER(struct efx_nic_errors, rx_reset),
+	EFX_ATOMIC_PARAMETER(struct efx_nic_errors, rx_desc_fetch),
+	EFX_ATOMIC_PARAMETER(struct efx_nic_errors, tx_desc_fetch),
+	EFX_ATOMIC_PARAMETER(struct efx_nic_errors, spurious_tx),
+	{NULL},
+};
+
+/**
+ * efx_init_debugfs_channels - create debugfs directories for NIC channels
+ * @efx:		Efx NIC
+ *
+ * Create subdirectories of @efx's debugfs directory for all the
+ * channels, RX queues and TX queues used by this driver.  Return a
+ * negative error code or 0 on success.  The subdirectories must be
+ * cleaned up using efx_fini_debugfs_channels().
+ */
+int efx_init_debugfs_channels(struct efx_nic *efx)
+{
+	struct efx_channel *channel;
+	struct efx_rx_queue *rx_queue;
+	struct efx_tx_queue *tx_queue;
+	int rc;
+
+	efx_for_each_channel(channel, efx) {
+		rc = efx_init_debugfs_channel(channel);
+		if (rc)
+			goto err;
+	}
+
+	efx_for_each_rx_queue(rx_queue, efx) {
+		rc = efx_init_debugfs_rx_queue(rx_queue);
+		if (rc)
+			goto err;
+	}
+
+	efx_for_each_tx_queue(tx_queue, efx) {
+		rc = efx_init_debugfs_tx_queue(tx_queue);
+		if (rc)
+			goto err;
+	}
+
+	return 0;
+
+ err:
+	efx_fini_debugfs_channels(efx);
+	return rc;
+}
+
+/**
+ * efx_fini_debugfs_channels - remove debugfs directories for NIC queues
+ * @efx:		Efx NIC
+ *
+ * Remove subdirectories of @efx's debugfs directory created by
+ * efx_init_debugfs_channels().
+ */
+void efx_fini_debugfs_channels(struct efx_nic *efx)
+{
+	struct efx_channel *channel;
+	struct efx_rx_queue *rx_queue;
+	struct efx_tx_queue *tx_queue;
+
+	efx_for_each_tx_queue(tx_queue, efx)
+		efx_fini_debugfs_tx_queue(tx_queue);
+
+	efx_for_each_rx_queue(rx_queue, efx)
+		efx_fini_debugfs_rx_queue(rx_queue);
+
+	efx_for_each_channel(channel, efx)
+		efx_fini_debugfs_channel(channel);
+}
+
+/**
+ * efx_init_debugfs_nic - create debugfs directory for NIC
+ * @efx:		Efx NIC
+ *
+ * Create debugfs directory containing parameter-files for @efx,
+ * and a subdirectory "errors" containing per-NIC error counts.
+ * Return a negative error code or 0 on success.  The directories
+ * must be cleaned up using efx_fini_debugfs_nic().
+ */
+int efx_init_debugfs_nic(struct efx_nic *efx)
+{
+	int rc;
+
+	/* Create directory */
+	efx->debug_dir = debugfs_create_dir(pci_name(efx->pci_dev),
+					    efx_debug_cards);
+	if (!efx->debug_dir)
+		goto err_mem;
+
+	/* Create errors directory */
+	efx->errors.debug_dir = debugfs_create_dir("errors", efx->debug_dir);
+	if (!efx->errors.debug_dir)
+		goto err_mem;
+
+	/* Create files */
+	rc = efx_init_debugfs_files(efx->debug_dir,
+				    efx_debugfs_nic_parameters, (void *)efx);
+	if (rc)
+		goto err;
+	rc = efx_init_debugfs_files(efx->errors.debug_dir,
+				    efx_debugfs_nic_error_parameters,
+				    (void *)&efx->errors);
+	if (rc)
+		goto err;
+
+	return 0;
+
+ err_mem:
+	rc = -ENOMEM;
+ err:
+	efx_fini_debugfs_nic(efx);
+	return rc;
+}
+
+/**
+ * efx_fini_debugfs_nic - remove debugfs directories for NIC
+ * @efx:		Efx NIC
+ *
+ * Remove debugfs directories created for @efx by efx_init_debugfs_nic().
+ */
+void efx_fini_debugfs_nic(struct efx_nic *efx)
+{
+	efx_fini_debugfs_dir(efx->errors.debug_dir,
+			     efx_debugfs_nic_error_parameters, NULL);
+	efx->errors.debug_dir = NULL;
+	efx_fini_debugfs_dir(efx->debug_dir, efx_debugfs_nic_parameters, NULL);
+	efx->debug_dir = NULL;
+}
+
+/**
+ * efx_init_debugfs - create debugfs directories for sfc driver
+ *
+ * Create debugfs directories "sfc" and "sfc/cards".  This must be
+ * called before any of the other functions that create debugfs
+ * directories.  Return a negative error code or 0 on success.  The
+ * directories must be cleaned up using efx_fini_debugfs().
+ */
+int efx_init_debugfs(void)
+{
+	/* Create top-level directory */
+#if defined(EFX_USE_DEBUGFS)
+	efx_debug_root = debugfs_create_dir("sfc", NULL);
+#else
+	efx_debug_root = proc_mkdir("sfc", proc_root_driver);
+#endif
+	if (!efx_debug_root)
+		goto err;
+
+	/* Create "cards" directory */
+	efx_debug_cards = debugfs_create_dir("cards", efx_debug_root);
+	if (!efx_debug_cards)
+		goto err;
+
+#if defined(EFX_USE_DEBUGFS)
+	/* Create compatibility sym-link */
+	if (!proc_symlink("sfc", proc_root_driver, "/sys/kernel/debug/sfc"))
+		goto err;
+#endif
+	return 0;
+
+ err:
+	efx_fini_debugfs();
+	return -ENOMEM;
+}
+
+/**
+ * efx_fini_debugfs - remove debugfs directories for sfc driver
+ *
+ * Remove directories created by efx_init_debugfs().
+ */
+void efx_fini_debugfs(void)
+{
+#if defined(EFX_USE_DEBUGFS)
+	remove_proc_entry("sfc", proc_root_driver);
+#endif
+	debugfs_remove(efx_debug_cards);
+	efx_debug_cards = NULL;
+	debugfs_remove(efx_debug_root);
+	efx_debug_root = NULL;
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/debugfs.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/debugfs.h
--- linux-2.6.18.8/drivers/net/sfc/debugfs.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/debugfs.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,172 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_DEBUGFS_H
+#define EFX_DEBUGFS_H
+
+#ifdef CONFIG_SFC_DEBUGFS
+
+struct seq_file;
+
+struct efx_debugfs_parameter {
+	const char *name;
+	size_t offset;
+	int (*reader)(struct seq_file *, void *);
+};
+
+extern void efx_fini_debugfs_child(struct dentry *dir, const char *name);
+extern int efx_init_debugfs_netdev(struct net_device *net_dev);
+extern void efx_fini_debugfs_netdev(struct net_device *net_dev);
+extern int efx_init_debugfs_port(struct efx_nic *efx);
+extern void efx_fini_debugfs_port(struct efx_nic *efx);
+extern int efx_init_debugfs_nic(struct efx_nic *efx);
+extern void efx_fini_debugfs_nic(struct efx_nic *efx);
+extern int efx_init_debugfs_channels(struct efx_nic *efx);
+extern void efx_fini_debugfs_channels(struct efx_nic *efx);
+extern int efx_init_debugfs(void);
+extern void efx_fini_debugfs(void);
+extern int efx_extend_debugfs_port(struct efx_nic *efx,
+				   void *context,
+				   struct efx_debugfs_parameter *params);
+extern void efx_trim_debugfs_port(struct efx_nic *efx,
+				  struct efx_debugfs_parameter *params);
+
+/* Helpers for handling debugfs entry reads */
+extern int efx_debugfs_read_uint(struct seq_file *, void *);
+extern int efx_debugfs_read_string(struct seq_file *, void *);
+extern int efx_debugfs_read_int(struct seq_file *, void *);
+extern int efx_debugfs_read_atomic(struct seq_file *, void *);
+extern int efx_debugfs_read_dword(struct seq_file *, void *);
+
+/* Handy macros for filling out parameters */
+
+/* Initialiser for a struct efx_debugfs_parameter with type-checking */
+#define EFX_PARAMETER(container_type, parameter, field_type,		\
+			reader_function) {				\
+	.name = #parameter,						\
+	.offset = ((((field_type *) 0) ==				\
+		    &((container_type *) 0)->parameter) ?		\
+		   offsetof(container_type, parameter) :		\
+		   offsetof(container_type, parameter)),		\
+	.reader = reader_function,					\
+}
+
+/* Likewise, but the file name is not taken from the field name */
+#define EFX_NAMED_PARAMETER(_name, container_type, parameter, field_type, \
+				reader_function) {			\
+	.name = #_name,							\
+	.offset = ((((field_type *) 0) ==				\
+		    &((container_type *) 0)->parameter) ?		\
+		   offsetof(container_type, parameter) :		\
+		   offsetof(container_type, parameter)),		\
+	.reader = reader_function,					\
+}
+
+/* Likewise, but with one file for each of 4 lanes */
+#define EFX_PER_LANE_PARAMETER(prefix, suffix, container_type, parameter, \
+				field_type, reader_function) {		\
+	.name = prefix "0" suffix,					\
+	.offset = ((((field_type *) 0) ==				\
+		      ((container_type *) 0)->parameter) ?		\
+		    offsetof(container_type, parameter[0]) :		\
+		    offsetof(container_type, parameter[0])),		\
+	.reader = reader_function,					\
+},  {									\
+	.name = prefix "1" suffix,					\
+	.offset = offsetof(container_type, parameter[1]),		\
+	.reader = reader_function,					\
+}, {									\
+	.name = prefix "2" suffix,					\
+	.offset = offsetof(container_type, parameter[2]),		\
+	.reader = reader_function,					\
+}, {									\
+	.name = prefix "3" suffix,					\
+	.offset = offsetof(container_type, parameter[3]),		\
+	.reader = reader_function,					\
+}
+
+/* A string parameter (string embedded in the structure) */
+#define EFX_STRING_PARAMETER(container_type, parameter) {	\
+	.name = #parameter,					\
+	.offset = ((((char *) 0) ==				\
+		    ((container_type *) 0)->parameter) ?	\
+		   offsetof(container_type, parameter) :	\
+		   offsetof(container_type, parameter)),	\
+	.reader = efx_debugfs_read_string,			\
+}
+
+/* An unsigned integer parameter */
+#define EFX_UINT_PARAMETER(container_type, parameter)		\
+	EFX_PARAMETER(container_type, parameter,		\
+		      unsigned int, efx_debugfs_read_uint)
+
+/* A dword parameter */
+#define EFX_DWORD_PARAMETER(container_type, parameter)		\
+	EFX_PARAMETER(container_type, parameter,		\
+		      efx_dword_t, efx_debugfs_read_dword)
+
+/* An atomic_t parameter */
+#define EFX_ATOMIC_PARAMETER(container_type, parameter)		\
+	EFX_PARAMETER(container_type, parameter,		\
+		      atomic_t, efx_debugfs_read_atomic)
+
+/* An integer parameter */
+#define EFX_INT_PARAMETER(container_type, parameter)		\
+	EFX_PARAMETER(container_type, parameter,		\
+		      int, efx_debugfs_read_int)
+
+#else /* !CONFIG_SFC_DEBUGFS */
+
+static inline int efx_init_debugfs_netdev(struct net_device *net_dev)
+{
+	return 0;
+}
+static inline void efx_fini_debugfs_netdev(struct net_device *net_dev) {}
+static inline int efx_init_debugfs_port(struct efx_nic *efx)
+{
+	return 0;
+}
+static inline void efx_fini_debugfs_port(struct efx_nic *efx) {}
+static inline int efx_init_debugfs_nic(struct efx_nic *efx)
+{
+	return 0;
+}
+static inline void efx_fini_debugfs_nic(struct efx_nic *efx) {}
+static inline int efx_init_debugfs_channels(struct efx_nic *efx)
+{
+	return 0;
+}
+static inline void efx_fini_debugfs_channels(struct efx_nic *efx) {}
+static inline int efx_init_debugfs(void)
+{
+	return 0;
+}
+static inline void efx_fini_debugfs(void) {}
+
+#endif /* CONFIG_SFC_DEBUGFS */
+
+#endif /* EFX_DEBUGFS_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/driverlink_api.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/driverlink_api.h
--- linux-2.6.18.8/drivers/net/sfc/driverlink_api.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/driverlink_api.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,612 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_DRIVERLINK_API_H
+#define EFX_DRIVERLINK_API_H
+
+#include <linux/list.h> /* for struct list_head */
+#if !defined(EFX_USE_FASTCALL)
+	#include <linux/version.h>
+	#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,20)
+		#define EFX_USE_FASTCALL yes
+		#include <linux/linkage.h>
+	#endif
+#endif
+
+/**
+ * DOC: Efx driverlink API
+ *
+ * This file must be included by any driver that wishes to attach to
+ * devices claimed by the Solarflare NIC driver (sfc). It allows separate
+ * kernel modules to expose other functionality offered by the NIC, with
+ * the sfc driver remaining in overall control.
+ *
+ * Overview:
+ *
+ * Driverlink clients define a &struct efx_dl_driver, and register
+ * this structure with the driverlink layer using
+ * efx_dl_register_driver(), which is exported by the sfc driver.
+ *
+ * The probe() routine of each driverlink client driver is called by
+ * the driverlink layer for each physical port in the system, after
+ * the sfc driver has performed start-of-day hardware initialisation
+ * and self-test. If ports are added or removed via pci hotplug then
+ * the &struct efx_dl_driver probe() or remove() routines are called
+ * as appropriate.
+ *
+ * If the port doesn't provide the necessary hardware resources for a
+ * client, then that client can return failure from its probe()
+ * routine. Information provided to the client driver at probe time
+ * includes
+ *
+ * Each probe() routine is given a unique &struct efx_dl_device per
+ * port, which means it can safely use the @priv member to store any
+ * useful state it needs. The probe routine also has the opportunity
+ * to provide a &struct efx_dl_callbacks via
+ * efx_dl_register_callbacks(), which allows the client to intercept
+ * the sfc driver's operations at strategic points.
+ *
+ * Occasionally, the underlying Efx device may need to be reset to
+ * recover from an error condition.  The client's reset_suspend() and
+ * reset_resume() methods [if provided] will be called to enable the
+ * client to suspend operations and preserve any state before the
+ * reset.  The client can itself request a reset using efx_dl_reset()
+ * or efx_dl_schedule_reset(), should it detect an error condition
+ * necessitating a reset.
+ *
+ * Example:
+ *
+ * The MTD driver (mtd.c) uses the driverlink layer.
+ */
+
+/* Forward declarations */
+struct pci_dev;
+struct net_device;
+struct sk_buff;
+struct efx_dl_device;
+struct efx_dl_device_info;
+
+/*
+ * This is used to guard against the registration of driverlink
+ * clients using an incorrect version of the API.
+ */
+#define EFX_DRIVERLINK_API_VERSION 1
+
+
+/**
+ * struct efx_dl_driver - An Efx driverlink device driver
+ *
+ * This is the analogue of a struct pci_driver for a normal PCI
+ * driver.  Driverlink clients should register themselves using
+ * efx_dl_register_driver() at module initialisation, and deregister
+ * themselves using efx_dl_unregister_driver() at module exit.
+ *
+ * All calls to members of efx_dl_driver are serialised by a single
+ * semaphore, so you are allowed to sleep in these functions. Take care
+ * to not call driverlink methods from within these callbacks, otherwise
+ * a deadlock is possible.
+ *
+ * @name: Name of the driver
+ * @probe: Called when device added
+ * @remove: Called when device removed
+ * @reset_suspend: Called before device is reset
+ * @reset_resume: Called after device is reset
+ */
+struct efx_dl_driver {
+	const char *name;
+
+	/*
+	 * probe - Handle device addition.
+	 * @efx_dev:		Efx driverlink device
+	 * @net_dev:		The net_dev relevant to this port
+	 * @dev_info:		A linked list of device information.
+	 * @silicon_rev:	Silicon revision name.
+	 *
+	 * This will be called after driverlink client registration for
+	 * every port on the system, and for every port that appears
+	 * thereafter via hotplug.
+	 *
+	 * The client may use either @efx_dev->pci_dev, the dev_info linked
+	 * list of available driver information, or the silicon revision
+	 * name to determine if they can support this port. If they can,
+	 * they should return 0 to indicate the probe was successful. Any
+	 * other return code indicates that the probe failed, and the
+	 * @efx_dl_dev will be invalidated.
+	 *
+	 * The client should perform whatever initialisation it
+	 * requires, and store a pointer to its private data in
+	 * @efx_dl_dev->priv (which is not shared between clients).
+	 * It may also wish to hook in a callbacks table using
+	 * efx_dl_register_callbacks().
+	 *
+	 * Return a negative error code or 0 on success.
+	 */
+	int (*probe) (struct efx_dl_device *efx_dl_dev,
+		      const struct net_device *net_dev,
+		      const struct efx_dl_device_info *dev_info,
+		      const char *silicon_rev);
+
+	/*
+	 * remove - Handle device removal.
+	 * @efx_dev:		Efx driverlink device
+	 *
+	 * This will be called at driver exit (or hotplug removal) for
+	 * each registered driverlink client.
+	 *
+	 * The client must ensure that it has finished all operations
+	 * using this device before returning from this method.  If it
+	 * has hooked in a callbacks table using
+	 * efx_dl_register_callbacks(), it must unhook it using
+	 * efx_dl_unregister_callbacks(), and then ensure that all
+	 * callback-triggered operations (e.g. scheduled tasklets)
+	 * have completed before returning.  (It does not need to
+	 * explicitly wait for callback methods to finish executing,
+	 * since efx_dl_unregister_callbacks() will sleep until all
+	 * callbacks have returned anyway.)
+	 *
+	 * Note that the device itself may not have been removed; it
+	 * may be simply that the client is being unloaded
+	 * via efx_dl_unregister_driver(). In this case other clients
+	 * (and the sfc driver itself) will still be using the device,
+	 * so the client cannot assume that the device itself is quiescent.
+	 * In particular, callbacks may continue to be triggered at any
+	 * point until efx_dl_unregister_callbacks() is called.
+	 */
+	void (*remove) (struct efx_dl_device *efx_dev);
+
+	/*
+	 * reset_suspend - Suspend ready for reset.
+	 * @efx_dev:		Efx driverlink device
+	 *
+	 * This method will be called immediately before a hardware
+	 * reset (which may or may not have been initiated by the
+	 * driverlink client).  This client must save any state that it
+	 * will need to restore after the reset, and suspend all
+	 * operations that might access the hardware.  It must not
+	 * return until the client can guarantee to have stopped
+	 * touching the hardware.
+	 *
+	 * It is guaranteed that callbacks will be inactive by the
+	 * time this method is called; the driverlink layer will
+	 * already have prevented new callbacks being made and waited
+	 * for all callbacks functions to return before calling
+	 * reset_suspend().  However, any delayed work scheduled by
+	 * the callback functions (e.g. tasklets) may not yet have
+	 * completed.
+	 *
+	 * This method is allowed to sleep, so waiting on tasklets,
+	 * work queues etc. is permitted.  There will always be a
+	 * corresponding call to the reset_resume() method, so it is
+	 * safe to e.g. down a semaphore within reset_suspend() and up
+	 * it within reset_resume().  (However, you obviously cannot
+	 * do the same with a spinlock).
+	 *
+	 * Note that the reset operation may be being carried out in
+	 * the context of scheduled work, so you cannot use
+	 * flush_scheduled_work() to ensure that any work you may have
+	 * scheduled has completed.
+	 *
+	 * During hardware reset, there is a chance of receiving
+	 * spurious interrupts, so the client's ISR (if any) should be
+	 * unhooked or otherwise disabled.
+	 */
+	void (*reset_suspend) (struct efx_dl_device *efx_dev);
+
+	/*
+	 * reset_resume - Restore after a reset.
+	 * @efx_dev:		Efx driverlink device
+	 * @ok:			Reset success indicator
+	 *
+	 * This method will be called after a hardware reset.  There
+	 * will always have been a corresponding call to the
+	 * reset_suspend() method beforehand.
+	 *
+	 * If @ok is non-zero, the client should restore the state
+	 * that it saved during the call to reset_suspend() and resume
+	 * normal operations.
+	 *
+	 * If @ok is zero, the reset operation has failed and the
+	 * hardware is currently in an unusable state.  In this case,
+	 * the client should release any locks taken out by
+	 * reset_suspend(), but should not take any other action; in
+	 * particular, it must not access the hardware, nor resume
+	 * normal operations.  The hardware is effectively dead at
+	 * this point, and our sole aim is to avoid deadlocking or
+	 * crashing the host.
+	 *
+	 * The driverlink layer will still be locked when
+	 * reset_resume() is called, so the client may not call
+	 * driverlink functions.  In particular, if the reset failed,
+	 * the client must not call efx_dl_unregister_callbacks() at
+	 * this point; it should wait until remove() is called.
+	 */
+	void (*reset_resume) (struct efx_dl_device *efx_dev, int ok);
+
+/* private: */
+	struct list_head node;
+	struct list_head device_list;
+};
+
+/**
+ * DOC: Efx driverlink device information
+ *
+ * Each &struct efx_dl_device makes certain hardware resources visible
+ * to driverlink clients, and they describe which resources are
+ * available by passing a linked list of &struct efx_dl_device_info
+ * into the probe() routine.
+ *
+ * The driverlink client's probe function can iterate through the linked list,
+ * and provided that it understands the resources that are exported, it can
+ * choose to make use of them through an external interface.
+ */
+
+/**
+ * enum efx_dl_device_info_type - Device information identifier.
+ *
+ * Each distinct hardware resource API will have a member in this
+ * enumeration.
+ *
+ * @EFX_DL_FALCON_RESOURCES: Information type is &struct efx_dl_falcon_resources
+ */
+enum efx_dl_device_info_type {
+	/** Falcon resources available for export */
+	EFX_DL_FALCON_RESOURCES = 0,
+};
+
+/**
+ * struct efx_dl_device_info - device information structure
+ * @next: Link to next structure, if any
+ * @type: Type code for this structure
+ *
+ * This structure is embedded in other structures provided by the
+ * driverlink device provider, and implements a linked list of
+ * resources pertinent to a driverlink client.
+ *
+ * Example: &struct efx_dl_falcon_resources
+ */
+struct efx_dl_device_info {
+	struct efx_dl_device_info *next;
+	enum efx_dl_device_info_type type;
+};
+
+/**
+ * enum efx_dl_falcon_resource_flags - Falcon resource information flags.
+ *
+ * Flags that describe hardware variations for the described Falcon based port.
+ *
+ * @EFX_DL_FALCON_DUAL_FUNC: Port is dual-function.
+ *	Certain silicon revisions have two pci functions, and require
+ *	certain hardware resources to be accessed via the secondary
+ *	function. See the discussion of @pci_dev in &struct efx_dl_device
+ *	below.
+ * @EFX_DL_FALCON_USE_MSI: Port is initialised to use MSI/MSI-X interrupts.
+ *	Falcon supports traditional legacy interrupts and MSI/MSI-X
+ *	interrupts. Since the sfc driver supports either, as a run
+ *	time configuration, driverlink drivers need to be aware of which
+ *	one to use for their interrupting resources.
+ */
+enum efx_dl_falcon_resource_flags {
+	EFX_DL_FALCON_DUAL_FUNC = 0x1,
+	EFX_DL_FALCON_USE_MSI = 0x2,
+};
+
+/**
+ * struct efx_dl_falcon_resources - Falcon resource information.
+ *
+ * This structure describes Falcon hardware resources available for
+ * use by a driverlink driver.
+ *
+ * @hdr: Resource linked list header
+ * @biu_lock: Register access lock.
+ *	Some Falcon revisions require register access for configuration
+ *	registers to be serialised between ports and PCI functions.
+ *	The sfc driver will provide the appropriate lock semantics for
+ *	the underlying hardware.
+ * @buffer_table_min: First available buffer table entry
+ * @buffer_table_max: Last available buffer table entry + 1
+ * @evq_timer_min: First available event queue with timer
+ * @evq_timer_max: Last available event queue with timer + 1
+ * @evq_int_min: First available event queue with interrupt
+ * @evq_int_max: Last available event queue with interrupt + 1
+ * @rxq_min: First available RX queue
+ * @rxq_max: Last available RX queue + 1
+ * @txq_min: First available TX queue
+ * @txq_max: Last available TX queue + 1
+ * @flags: Hardware variation flags
+ */
+struct efx_dl_falcon_resources {
+	struct efx_dl_device_info hdr;
+	spinlock_t *biu_lock;
+	unsigned buffer_table_min, buffer_table_max;
+	unsigned evq_timer_min, evq_timer_max;
+	unsigned evq_int_min, evq_int_max;
+	unsigned rxq_min, rxq_max;
+	unsigned txq_min, txq_max;
+	enum efx_dl_falcon_resource_flags flags;
+};
+
+/**
+ * struct efx_dl_device - An Efx driverlink device.
+ *
+ * @pci_dev: Underlying PCI device.
+ *	This is the PCI device used by the sfc driver.  It will
+ *	already have been enabled for bus-mastering DMA etc.
+ * @priv: Driver private data
+ *	Driverlink clients can use this to store a pointer to their
+ *	internal per-device data structure. Each (driver, device)
+ *	tuple has a separate &struct efx_dl_device, so clients can use
+ *	this @priv field independently.
+ * @driver: Efx driverlink driver for this device
+ */
+struct efx_dl_device {
+	struct pci_dev *pci_dev;
+	void *priv;
+	struct efx_dl_driver *driver;
+};
+
+/**
+ * enum efx_veto - Packet veto request flag.
+ *
+ * This is the return type for the rx_packet() and tx_packet() methods
+ * in &struct efx_dl_callbacks.
+ *
+ * @EFX_ALLOW_PACKET: Packet may be transmitted/received
+ * @EFX_VETO_PACKET: Packet must not be transmitted/received
+ */
+enum efx_veto {
+	EFX_ALLOW_PACKET = 0,
+	EFX_VETO_PACKET = 1,
+};
+
+/**
+ * struct efx_dl_callbacks - Efx callbacks
+ *
+ * These methods can be hooked in to the sfc driver via
+ * efx_dl_register_callbacks().  They allow clients to intercept and/or
+ * modify the behaviour of the sfc driver at predetermined points.
+ *
+ * For efficiency, only one client can hook each callback.
+ *
+ * Since these callbacks are called on packet transmit and reception
+ * paths, clients should avoid acquiring locks or allocating memory.
+ *
+ * @tx_packet: Called when packet is about to be transmitted
+ * @rx_packet: Called when packet is received
+ * @link_change: Called when link status has changed
+ * @request_mtu: Called to request MTU change
+ * @mtu_changed: Called when MTU has been changed
+ * @event: Called when NIC event is not handled by the sfc driver
+ */
+struct efx_dl_callbacks {
+	/*
+	 * tx_packet - Packet about to be transmitted.
+	 * @efx_dev:		Efx driverlink device
+	 * @skb:		Socket buffer containing the packet to be sent
+	 *
+	 * This method is called for every packet about to be
+	 * transmitted.  It allows the client to snoop on traffic sent
+	 * via the kernel queues.
+	 *
+	 * The method may return %EFX_VETO_PACKET in order to prevent
+	 * the sfc driver from transmitting the packet.  The net
+	 * driver will then discard the packet.  If the client wishes
+	 * to retain a reference to the packet data after returning
+	 * %EFX_VETO_PACKET, it must obtain its own copy of the
+	 * packet (e.g. by calling skb_get(), or by copying out the
+	 * packet data to an external buffer).
+	 *
+	 * This method must return quickly, since it will have a
+	 * direct performance impact upon the sfc driver.  It will be
+	 * called with interrupts disabled (and may be called in
+	 * interrupt context), so may not sleep. Since the sfc driver
+	 * may have multiple TX queues, running in parallel, please avoid
+	 * the need for locking if it all possible.
+	 */
+#if defined(EFX_USE_FASTCALL)
+	enum efx_veto fastcall (*tx_packet) (struct efx_dl_device *efx_dev,
+					     struct sk_buff *skb);
+#else
+	enum efx_veto (*tx_packet) (struct efx_dl_device *efx_dev,
+				    struct sk_buff *skb);
+#endif
+
+	/*
+	 * rx_packet - Packet received.
+	 * @efx_dev:		Efx driverlink device
+	 * @pkt_hdr:		Pointer to received packet
+	 * @pkt_len:		Length of received packet
+	 *
+	 * This method is called for every received packet.  It allows
+	 * the client to snoop on traffic received by the kernel
+	 * queues.
+	 *
+	 * The method may return %EFX_VETO_PACKET in order to prevent
+	 * the sfc driver from passing the packet to the kernel.  The net
+	 * driver will then discard the packet.
+	 *
+	 * This method must return quickly, since it will have a
+	 * direct performance impact upon the sfc driver.  It is
+	 * called in tasklet context, so may not sleep.  Note that
+	 * there are per-channel tasklets in the sfc driver, so
+	 * rx_packet() may be called simultaneously on different CPUs
+	 * and must lock appropriately.  The design of the sfc driver
+	 * allows for lockless operation between receive channels, so
+	 * please avoid the need for locking if at all possible.
+	 */
+#if defined(EFX_USE_FASTCALL)
+	enum efx_veto fastcall (*rx_packet) (struct efx_dl_device *efx_dev,
+					     const char *pkt_hdr, int pkt_len);
+#else
+	enum efx_veto (*rx_packet) (struct efx_dl_device *efx_dev,
+				    const char *pkt_hdr, int pkt_len);
+#endif
+
+	/*
+	 * link_change - Link status change.
+	 * @efx_dev:		Efx driverlink device
+	 * @link_up:		Link up indicator
+	 *
+	 * This method is called to inform the driverlink client
+	 * whenever the PHY link status changes.  By the time this
+	 * function is called, the MAC has already been reconfigured
+	 * with the new autonegotiation settings from the PHY.
+	 *
+	 * This method is called from tasklet context and may not
+	 * sleep.
+	 */
+	void (*link_change) (struct efx_dl_device *efx_dev, int link_up);
+
+	/*
+	 * request_mtu: Request MTU change.
+	 * @efx_dev:		Efx driverlink device
+	 * @new_mtu:		Requested new MTU
+	 *
+	 * This method is called whenever the user requests an MTU
+	 * change on an interface.  The client may return an error, in
+	 * which case the MTU change request will be denied.  If the
+	 * client returns success, the MAC will be reconfigured with a
+	 * new maxmimum frame length equal to
+	 * EFX_MAX_FRAME_LEN(new_mtu).  The client will be notified
+	 * via the mtu_changed() method once the MAC has been
+	 * reconfigured.
+	 *
+	 * The current MTU for the port can be obtained via
+	 * efx_dl_get_netdev(efx_dl_device)->mtu.
+	 *
+	 * The sfc driver guarantees that no other callback functions
+	 * are in progress when this method is called.  This function
+	 * is called in process context and may sleep.
+	 *
+	 * Return a negative error code or 0 on success.
+	 */
+	int (*request_mtu) (struct efx_dl_device *efx_dev, int new_mtu);
+
+	/*
+	 * mtu_changed - MTU has been changed.
+	 * @efx_dev:		Efx driverlink device
+	 * @mtu:		The new MTU
+	 *
+	 * This method is called once the MAC has been reconfigured
+	 * with a new MTU.  There will have been a preceding call to
+	 * request_mtu().
+	 *
+	 * The sfc driver guarantees that no other callback functions
+	 * are in progress when this method is called.  This function
+	 * is called in process context and may sleep.
+	 */
+	void (*mtu_changed) (struct efx_dl_device *efx_dev, int mtu);
+
+	/*
+	 * event - Event callback.
+	 * @efx_dev:		Efx driverlink device
+	 * @p_event:		Pointer to event
+	 *
+	 * This method is called for each event that is not handled by the
+	 * sfc driver.
+	 */
+	void (*event) (struct efx_dl_device *efx_dev, void *p_event);
+};
+
+/* Include API version number in symbol used for efx_dl_register_driver */
+#define efx_dl_stringify_1(x, y) x ## y
+#define efx_dl_stringify_2(x, y) efx_dl_stringify_1(x, y)
+#define efx_dl_register_driver					\
+	efx_dl_stringify_2(efx_dl_register_driver_api_ver_,	\
+			   EFX_DRIVERLINK_API_VERSION)
+
+extern int efx_dl_register_driver(struct efx_dl_driver *driver);
+
+extern void efx_dl_unregister_driver(struct efx_dl_driver *driver);
+
+extern int efx_dl_register_callbacks(struct efx_dl_device *efx_dev,
+				     struct efx_dl_callbacks *callbacks);
+
+extern void efx_dl_unregister_callbacks(struct efx_dl_device *efx_dev,
+					struct efx_dl_callbacks *callbacks);
+
+extern void efx_dl_schedule_reset(struct efx_dl_device *efx_dev);
+
+/**
+ * efx_dl_for_each_device_info_matching - iterate an efx_dl_device_info list
+ * @_dev_info: Pointer to first &struct efx_dl_device_info
+ * @_type: Type code to look for
+ * @_info_type: Structure type corresponding to type code
+ * @_field: Name of &struct efx_dl_device_info field in the type
+ * @_p: Iterator variable
+ *
+ * Example:
+ *
+ * static int driver_dl_probe(... const struct efx_dl_device_info *dev_info ...)
+ * {
+ *        struct efx_dl_falcon_resources *res;
+ *
+ *        efx_dl_for_each_device_info_matching(dev_info,EFX_DL_FALCON_RESOURCES,
+ *                                             struct efx_dl_falcon_resources,
+ *                                             hdr, res) {
+ *                if (res->flags & EFX_DL_FALCON_DUAL_FUNC) {
+ *                          .....
+ *                }
+ *        }
+ * }
+ */
+#define efx_dl_for_each_device_info_matching(_dev_info, _type,		\
+					     _info_type, _field, _p)	\
+	for ((_p) = container_of((_dev_info), _info_type, _field);	\
+	     (_p) != NULL;						\
+	     (_p) = container_of((_p)->_field.next, _info_type, _field))\
+		if ((_p)->_field.type != _type)				\
+			continue;					\
+		else
+
+/**
+ * efx_dl_search_device_info - search an efx_dl_device_info list
+ * @_dev_info: Pointer to first &struct efx_dl_device_info
+ * @_type: Type code to look for
+ * @_info_type: Structure type corresponding to type code
+ * @_field: Name of &struct efx_dl_device_info member in this type
+ * @_p: Result variable
+ *
+ * Example:
+ *
+ * static int driver_dl_probe(... const struct efx_dl_device_info *dev_info ...)
+ * {
+ *        struct efx_dl_falcon_resources *res;
+ *
+ *        efx_dl_search_device_info(dev_info, EFX_DL_FALCON_RESOURCES,
+ *                                  struct efx_dl_falcon_resources, hdr, res);
+ *        if (res != NULL) {
+ *                 ....
+ *        }
+ * }
+ */
+#define efx_dl_search_device_info(_dev_info, _type, _info_type,		\
+				  _field, _p)				\
+	efx_dl_for_each_device_info_matching((_dev_info), (_type),	\
+					     _info_type, _field, (_p))	\
+		break;
+
+#endif /* EFX_DRIVERLINK_API_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/driverlink.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/driverlink.c
--- linux-2.6.18.8/drivers/net/sfc/driverlink.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/driverlink.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,544 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005:      Fen Systems Ltd.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/module.h>
+#include <linux/list.h>
+#include <linux/skbuff.h>
+#include <linux/rtnetlink.h>
+#include "net_driver.h"
+#include "efx.h"
+#include "driverlink.h"
+
+/* Driverlink semaphore
+ * This semaphore must be held for any operation that modifies any of
+ * the driverlink lists.
+ */
+static DEFINE_MUTEX(efx_driverlink_lock);
+
+/* List of all registered drivers */
+static LIST_HEAD(efx_driver_list);
+
+/* List of all registered Efx ports */
+static LIST_HEAD(efx_port_list);
+
+/* Driver link handle used internally to track devices */
+struct efx_dl_handle {
+	/* The efx_dl_device consumers see */
+	struct efx_dl_device efx_dev;
+	/* The efx_nic providers provide */
+	struct efx_nic *efx;
+	/* Per-device list */
+	struct list_head port_node;
+	/* Per-driver list */
+	struct list_head driver_node;
+};
+
+/* Get the handle for an efx_dl_device */
+static struct efx_dl_handle *efx_dl_handle(struct efx_dl_device *efx_dev)
+{
+	return container_of(efx_dev, struct efx_dl_handle, efx_dev);
+}
+
+/* Remove an Efx device
+ * You must hold the efx_driverlink_lock before calling this
+ * function.
+ */
+static void efx_dl_del_device(struct efx_dl_device *efx_dev)
+{
+	struct efx_dl_handle *efx_handle = efx_dl_handle(efx_dev);
+
+	EFX_INFO(efx_handle->efx, "%s driverlink client unregistering\n",
+		 efx_dev->driver->name);
+
+	/* Call driver's remove() routine */
+	if (efx_dev->driver->remove)
+		efx_dev->driver->remove(efx_dev);
+
+	/* Remove handle from per-driver and per-NIC lists */
+	list_del(&efx_handle->driver_node);
+	list_del(&efx_handle->port_node);
+
+	/* Free efx_handle structure */
+	kfree(efx_handle);
+}
+
+/* Try to add an Efx device
+ * Attempt to probe the given device with the driver, creating a
+ * new efx_dl_device. If the probe routine fails, because the driver
+ * doesn't support this port, then the efx_dl_device is destroyed,
+ */
+static void efx_dl_try_add_device(struct efx_nic *efx,
+				  struct efx_dl_driver *driver)
+{
+	struct efx_dl_handle *efx_handle;
+	struct efx_dl_device *efx_dev;
+	int rc;
+
+	/* Allocate and initialise new efx_dl_device structure */
+	efx_handle = kzalloc(sizeof(*efx_handle), GFP_KERNEL);
+	efx_dev = &efx_handle->efx_dev;
+	efx_handle->efx = efx;
+	efx_dev->driver = driver;
+	efx_dev->pci_dev = efx->pci_dev;
+	INIT_LIST_HEAD(&efx_handle->port_node);
+	INIT_LIST_HEAD(&efx_handle->driver_node);
+
+	/* Attempt driver probe */
+	rc = driver->probe(efx_dev, efx->net_dev,
+			   efx->dl_info, efx->silicon_rev);
+	if (rc)
+		goto fail;
+
+	/* Add device to per-driver and per-NIC lists */
+	list_add_tail(&efx_handle->driver_node, &driver->device_list);
+	list_add_tail(&efx_handle->port_node, &efx->dl_device_list);
+
+	EFX_INFO(efx, "%s driverlink client registered\n", driver->name);
+	return;
+
+ fail:
+	EFX_INFO(efx, "%s driverlink client skipped\n", driver->name);
+
+	kfree(efx_dev);
+}
+
+/**
+ * efx_dl_unregister_driver - unregister an Efx device driver
+ * @driver:		Efx driverlink driver
+ *
+ * Unregisters an Efx driver.  The driver's remove() method will be
+ * called for all Efx devices currently claimed by the driver.
+ */
+void efx_dl_unregister_driver(struct efx_dl_driver *driver)
+{
+	struct efx_dl_handle *efx_handle, *efx_handle_n;
+
+	printk(KERN_INFO "Efx driverlink unregistering %s driver\n",
+		 driver->name);
+
+	/* Acquire lock.  We can't return failure, so have to use
+	 * down() instead of down_interruptible()
+	 */
+	mutex_lock(&efx_driverlink_lock);
+
+	/* Remove all devices claimed by the driver */
+	list_for_each_entry_safe(efx_handle, efx_handle_n,
+				 &driver->device_list, driver_node)
+		efx_dl_del_device(&efx_handle->efx_dev);
+
+	/* Remove driver from driver list */
+	list_del(&driver->node);
+
+	/* Release lock */
+	mutex_unlock(&efx_driverlink_lock);
+}
+EXPORT_SYMBOL(efx_dl_unregister_driver);
+
+/**
+ * efx_dl_register_driver - register an Efx device driver
+ * @driver:		Efx driverlink driver
+ *
+ * Registers a new Efx driver.  The driver's probe() method will be
+ * called for all Efx NICs currently registered.
+ *
+ * Return a negative error code or 0 on success.
+ */
+int efx_dl_register_driver(struct efx_dl_driver *driver)
+{
+	struct efx_nic *efx;
+	int rc;
+
+	printk(KERN_INFO "Efx driverlink registering %s driver\n",
+		 driver->name);
+
+	/* Initialise driver list structures */
+	INIT_LIST_HEAD(&driver->node);
+	INIT_LIST_HEAD(&driver->device_list);
+
+	/* Acquire lock */
+	rc = mutex_lock_interruptible(&efx_driverlink_lock);
+	if (rc)
+		return rc;
+
+	/* Add driver to driver list */
+	list_add_tail(&driver->node, &efx_driver_list);
+
+	/* Feed all existing devices to driver */
+	list_for_each_entry(efx, &efx_port_list, dl_node)
+		efx_dl_try_add_device(efx, driver);
+
+	/* Release locks */
+	mutex_unlock(&efx_driverlink_lock);
+
+	return 0;
+}
+EXPORT_SYMBOL(efx_dl_register_driver);
+
+void efx_dl_unregister_nic(struct efx_nic *efx)
+{
+	struct efx_dl_handle *efx_handle, *efx_handle_n;
+
+	if (!efx)
+		return;
+
+	/* Acquire lock.  We can't return failure, so have to use
+	 * down() instead of down_interruptible()
+	 */
+	mutex_lock(&efx_driverlink_lock);
+
+	/* Remove all devices related to this NIC */
+	list_for_each_entry_safe_reverse(efx_handle, efx_handle_n,
+					 &efx->dl_device_list,
+					 port_node)
+		efx_dl_del_device(&efx_handle->efx_dev);
+
+	/* Remove port from port list */
+	list_del(&efx->dl_node);
+
+	/* Release lock */
+	mutex_unlock(&efx_driverlink_lock);
+}
+
+int efx_dl_register_nic(struct efx_nic *efx)
+{
+	struct efx_dl_driver *driver;
+	int rc;
+
+	/* Acquire lock */
+	rc = mutex_lock_interruptible(&efx_driverlink_lock);
+	if (rc)
+		return rc;
+
+	/* Add port to port list */
+	list_add_tail(&efx->dl_node, &efx_port_list);
+
+	/* Feed port to all existing drivers */
+	list_for_each_entry(driver, &efx_driver_list, node)
+		efx_dl_try_add_device(efx, driver);
+
+	/* Release lock */
+	mutex_unlock(&efx_driverlink_lock);
+
+	return 0;
+}
+
+/*
+ * Dummy callback implementations.
+ *
+ * To avoid a branch point on the fast-path, the callbacks are always
+ * implemented - they are never NULL.
+ */
+#if defined(EFX_USE_FASTCALL)
+static enum efx_veto fastcall
+#else
+static enum efx_veto
+#endif
+efx_dummy_tx_packet_callback(struct efx_dl_device *efx_dev, struct sk_buff *skb)
+{
+	/* Never veto the packet */
+	return EFX_ALLOW_PACKET;
+}
+
+#if defined(EFX_USE_FASTCALL)
+static enum efx_veto fastcall
+#else
+static enum efx_veto
+#endif
+efx_dummy_rx_packet_callback(struct efx_dl_device *efx_dev,
+			     const char *pkt_buf, int len)
+{
+	/* Never veto the packet */
+	return EFX_ALLOW_PACKET;
+}
+
+static void
+efx_dummy_link_change_callback(struct efx_dl_device *efx_dev, int link_up)
+{
+}
+
+static int
+efx_dummy_request_mtu_callback(struct efx_dl_device *efx_dev, int new_mtu)
+{
+	/* Always allow */
+	return 0;
+}
+
+static void
+efx_dummy_mtu_changed_callback(struct efx_dl_device *efx_dev, int mtu)
+{
+	return;
+}
+
+static void efx_dummy_event_callback(struct efx_dl_device *efx_dev, void *event)
+{
+	return;
+}
+
+struct efx_dl_callbacks efx_default_callbacks = {
+	.tx_packet	= efx_dummy_tx_packet_callback,
+	.rx_packet	= efx_dummy_rx_packet_callback,
+	.link_change	= efx_dummy_link_change_callback,
+	.request_mtu	= efx_dummy_request_mtu_callback,
+	.mtu_changed	= efx_dummy_mtu_changed_callback,
+	.event		= efx_dummy_event_callback,
+};
+
+#define EFX_DL_UNREGISTER_CALLBACK(_port, _dev, _member)		\
+	do {								\
+		BUG_ON((_port)->dl_cb_dev._member != (_dev));		\
+		(_port)->dl_cb._member =				\
+			efx_default_callbacks._member;			\
+		(_port)->dl_cb_dev._member = NULL;			\
+	} while (0)
+
+
+#define EFX_DL_REGISTER_CALLBACK(_port, _dev, _from, _member)		\
+	if ((_from)->_member) {						\
+		BUG_ON((_port)->dl_cb_dev._member != NULL);		\
+		(_port)->dl_cb._member = (_from)->_member;		\
+		(_port)->dl_cb_dev._member = _dev;			\
+	}
+
+/**
+ * efx_dl_unregister_callbacks - unregister callbacks for an Efx NIC
+ * @efx_dev:		Efx driverlink device
+ * @callbacks:		Callback list
+ *
+ * This removes a set of callbacks registered with
+ * efx_dl_register_callbacks().  It should be called as part of the
+ * client's remove() method.
+ *
+ * The net driver will ensure that all callback functions have
+ * returned to the net driver before efx_dl_unregister_callbacks()
+ * returns.  Note that the device itself may still be running when the
+ * client's remove() method is called.  The client must therefore
+ * unhook its callbacks using efx_dl_unregister_callbacks() and only
+ * then ensure that any delayed tasks triggered by callback methods
+ * (e.g. scheduled tasklets) have completed.
+ */
+void efx_dl_unregister_callbacks(struct efx_dl_device *efx_dev,
+				 struct efx_dl_callbacks *callbacks)
+{
+	struct efx_dl_handle *efx_handle = efx_dl_handle(efx_dev);
+	struct efx_nic *efx = efx_handle->efx;
+
+	/* Suspend net driver operations */
+	efx_suspend(efx);
+
+	EFX_INFO(efx, "removing callback hooks into %s driver\n",
+		 efx_dev->driver->name);
+
+	if (callbacks->tx_packet)
+		EFX_DL_UNREGISTER_CALLBACK(efx, efx_dev, tx_packet);
+
+	if (callbacks->rx_packet)
+		EFX_DL_UNREGISTER_CALLBACK(efx, efx_dev, rx_packet);
+
+	if (callbacks->link_change)
+		EFX_DL_UNREGISTER_CALLBACK(efx, efx_dev, link_change);
+
+	if (callbacks->request_mtu)
+		EFX_DL_UNREGISTER_CALLBACK(efx, efx_dev, request_mtu);
+
+	if (callbacks->mtu_changed)
+		EFX_DL_UNREGISTER_CALLBACK(efx, efx_dev, mtu_changed);
+
+	if (callbacks->event)
+		EFX_DL_UNREGISTER_CALLBACK(efx, efx_dev, event);
+
+	/* Resume net driver operations */
+	efx_resume(efx);
+}
+EXPORT_SYMBOL(efx_dl_unregister_callbacks);
+
+/**
+ * efx_dl_register_callbacks - register callbacks for an Efx NIC
+ * @efx_dev:		Efx driverlink device
+ * @callbacks:		Callback list
+ *
+ * This registers a set of callback functions with the net driver.
+ * These functions will be called at various key points to allow
+ * external code to monitor and/or modify the behaviour of the network
+ * driver.  Any of the callback function pointers may be %NULL if a
+ * callback is not required.  The intended user of this mechanism is
+ * the SFC char driver.
+ *
+ * This client should call efx_dl_register_callbacks() during its
+ * probe() method.  The client must ensure that it also calls
+ * efx_dl_unregister_callbacks() as part of its remove() method.
+ *
+ * Only one function may be registered for each callback per NIC.
+ * If a requested callback is already registered for this NIC, this
+ * function will return -%EBUSY.
+ *
+ * The device may already be running, so the client must be prepared
+ * for callbacks to be triggered immediately after calling
+ * efx_dl_register_callbacks().
+ *
+ * Return a negative error code or 0 on success.
+ */
+int efx_dl_register_callbacks(struct efx_dl_device *efx_dev,
+			      struct efx_dl_callbacks *callbacks)
+{
+	struct efx_dl_handle *efx_handle = efx_dl_handle(efx_dev);
+	struct efx_nic *efx = efx_handle->efx;
+	int rc = 0;
+
+	/* Suspend net driver operations */
+	efx_suspend(efx);
+
+	/* Check that the requested callbacks are not already hooked. */
+	if ((callbacks->tx_packet && efx->dl_cb_dev.tx_packet) ||
+	    (callbacks->rx_packet && efx->dl_cb_dev.rx_packet) ||
+	    (callbacks->link_change && efx->dl_cb_dev.link_change) ||
+	    (callbacks->request_mtu && efx->dl_cb_dev.request_mtu) ||
+	    (callbacks->mtu_changed && efx->dl_cb_dev.mtu_changed) ||
+	    (callbacks->event && efx->dl_cb_dev.event)) {
+		rc = -EBUSY;
+		goto out;
+	}
+
+	EFX_INFO(efx, "adding callback hooks to %s driver\n",
+		 efx_dev->driver->name);
+
+	/* Hook in callbacks.  For maximum speed, we never check to
+	 * see whether these are NULL before calling; therefore we
+	 * must ensure that they are never NULL.  If the set we're
+	 * being asked to hook in is sparse, we leave the default
+	 * values in place for the empty hooks.
+	 */
+	EFX_DL_REGISTER_CALLBACK(efx, efx_dev, callbacks, tx_packet);
+	EFX_DL_REGISTER_CALLBACK(efx, efx_dev, callbacks, rx_packet);
+	EFX_DL_REGISTER_CALLBACK(efx, efx_dev, callbacks, link_change);
+	EFX_DL_REGISTER_CALLBACK(efx, efx_dev, callbacks, request_mtu);
+	EFX_DL_REGISTER_CALLBACK(efx, efx_dev, callbacks, mtu_changed);
+	EFX_DL_REGISTER_CALLBACK(efx, efx_dev, callbacks, event);
+
+ out:
+	/* Resume net driver operations */
+	efx_resume(efx);
+
+	return rc;
+}
+EXPORT_SYMBOL(efx_dl_register_callbacks);
+
+/**
+ * efx_dl_schedule_reset - schedule an Efx NIC reset
+ * @efx_dev:		Efx driverlink device
+ *
+ * This schedules a hardware reset for a short time in the future.  It
+ * can be called from any context, and so can be used when
+ * efx_dl_reset() cannot be called.
+ */
+void efx_dl_schedule_reset(struct efx_dl_device *efx_dev)
+{
+	struct efx_dl_handle *efx_handle = efx_dl_handle(efx_dev);
+	struct efx_nic *efx = efx_handle->efx;
+
+	efx_schedule_reset(efx, RESET_TYPE_ALL);
+}
+EXPORT_SYMBOL(efx_dl_schedule_reset);
+
+/*
+ * Lock the driverlink layer before a reset
+ * To avoid deadlock, efx_driverlink_lock needs to be acquired before
+ * efx->suspend_lock.
+ */
+void efx_dl_reset_lock(void)
+{
+	/* Acquire lock */
+	mutex_lock(&efx_driverlink_lock);
+}
+
+/*
+ * Unlock the driverlink layer after a reset
+ * This call must be matched against efx_dl_reset_lock.
+ */
+void efx_dl_reset_unlock(void)
+{
+	/* Acquire lock */
+	mutex_unlock(&efx_driverlink_lock);
+}
+
+/*
+ * Suspend ready for reset
+ * This calls the reset_suspend method of all drivers registered to
+ * the specified NIC.  It must only be called between
+ * efx_dl_reset_lock and efx_dl_reset_unlock.
+ */
+void efx_dl_reset_suspend(struct efx_nic *efx)
+{
+	struct efx_dl_handle *efx_handle;
+	struct efx_dl_device *efx_dev;
+
+	BUG_ON(!mutex_is_locked(&efx_driverlink_lock));
+
+	/* Call suspend method of each driver in turn */
+	list_for_each_entry_reverse(efx_handle,
+				    &efx->dl_device_list,
+				    port_node) {
+		efx_dev = &efx_handle->efx_dev;
+		if (efx_dev->driver->reset_suspend)
+			efx_dev->driver->reset_suspend(efx_dev);
+	}
+}
+
+/*
+ * Resume after a reset
+ * This calls the reset_resume method of all drivers registered to the
+ * specified NIC.  It must only be called between efx_dl_reset_lock
+ * and efx_dl_reset_unlock.
+ */
+void efx_dl_reset_resume(struct efx_nic *efx, int ok)
+{
+	struct efx_dl_handle *efx_handle;
+	struct efx_dl_device *efx_dev;
+
+	BUG_ON(!mutex_is_locked(&efx_driverlink_lock));
+
+	/* Call resume method of each driver in turn */
+	list_for_each_entry(efx_handle, &efx->dl_device_list,
+			    port_node) {
+		efx_dev = &efx_handle->efx_dev;
+		if (efx_dev->driver->reset_resume)
+			efx_dev->driver->reset_resume(efx_dev, ok);
+	}
+}
+
+/**
+ * efx_dl_get_nic - obtain the Efx NIC for the given driverlink device
+ * @efx_dev:		Efx driverlink device
+ *
+ * Get a pointer to the &struct efx_nic corresponding to
+ * @efx_dev.  This can be used by driverlink clients built along with
+ * the sfc driver, which may have intimate knowledge of its internals.
+ */
+struct efx_nic *efx_dl_get_nic(struct efx_dl_device *efx_dev)
+{
+	return efx_dl_handle(efx_dev)->efx;
+}
+EXPORT_SYMBOL(efx_dl_get_nic);
diff -rpuN linux-2.6.18.8/drivers/net/sfc/driverlink.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/driverlink.h
--- linux-2.6.18.8/drivers/net/sfc/driverlink.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/driverlink.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,93 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005:      Fen Systems Ltd.
+ * Copyright 2006:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_DRIVERLINK_H
+#define EFX_DRIVERLINK_H
+
+/* Forward declarations */
+struct efx_dl_device;
+struct efx_nic;
+
+/*
+ * Efx driverlink
+ *
+ * This header file defines the portions of the Efx driverlink
+ * interface that are used only within the sfc module.  It also
+ * declares efx_dl_get_nic(), which may be used by sfc_mtd
+ * and any other module built along with sfc.
+ */
+
+
+/* Efx callback devices
+ *
+ * A list of the devices that own each callback. The partner to
+ * struct efx_dl_callbacks
+ */
+struct efx_dl_cb_devices {
+	/* Device owning the tx_packet callback */
+	struct efx_dl_device *tx_packet;
+	/* Device owning the rx_packet callback */
+	struct efx_dl_device *rx_packet;
+	/* Device owning the link_change callback. */
+	struct efx_dl_device *link_change;
+	/* Device owning the request_mtu callback. */
+	struct efx_dl_device *request_mtu;
+	/* Device owning the mtu_changed callback. */
+	struct efx_dl_device *mtu_changed;
+	/* Device owning the event callback. */
+	struct efx_dl_device *event;
+};
+
+/* No-op callbacks used for initialisation */
+extern struct efx_dl_callbacks efx_default_callbacks;
+
+/* Macro used to invoke callbacks */
+#define EFX_DL_CALLBACK(_port, _name, ...)				\
+	(_port)->dl_cb._name((_port)->dl_cb_dev._name, __VA_ARGS__)
+
+/* Register an Efx NIC */
+extern int efx_dl_register_nic(struct efx_nic *efx);
+
+/* Unregister an Efx NIC */
+extern void efx_dl_unregister_nic(struct efx_nic *efx);
+
+/* Lock the driverlink layer prior to a reset */
+extern void efx_dl_reset_lock(void);
+
+/* Unlock the driverlink layer following a reset */
+extern void efx_dl_reset_unlock(void);
+
+/* Suspend all drivers prior to a hardware reset */
+extern void efx_dl_reset_suspend(struct efx_nic *efx);
+
+/* Resume all drivers after a hardware reset */
+extern void efx_dl_reset_resume(struct efx_nic *efx, int ok);
+
+/* Obtain the Efx NIC for the given driverlink device. */
+extern struct efx_nic *efx_dl_get_nic(struct efx_dl_device *efx_dev);
+
+#endif /* EFX_DRIVERLINK_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/efx.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/efx.c
--- linux-2.6.18.8/drivers/net/sfc/efx.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/efx.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,2783 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/delay.h>
+#include <linux/notifier.h>
+#include <linux/ip.h>
+#include <linux/tcp.h>
+#include <linux/in.h>
+#include <linux/crc32.h>
+#include <linux/ethtool.h>
+#include <asm/uaccess.h>
+#include "net_driver.h"
+#include "gmii.h"
+#include "driverlink.h"
+#include "selftest.h"
+#include "debugfs.h"
+#include "ethtool.h"
+#include "tx.h"
+#include "rx.h"
+#include "efx.h"
+#include "mdio_10g.h"
+#include "falcon.h"
+#include "workarounds.h"
+
+/**************************************************************************
+ *
+ * Type name strings
+ *
+ **************************************************************************
+ */
+
+/* Loopback mode names (see LOOPBACK_MODE()) */
+const unsigned int efx_loopback_mode_max = LOOPBACK_MAX;
+const char *efx_loopback_mode_names[] = {
+	[LOOPBACK_NONE]	   = "NONE",
+	[LOOPBACK_MAC]	   = "MAC",
+	[LOOPBACK_XGMII]   = "XGMII",
+	[LOOPBACK_XGXS]	   = "XGXS",
+	[LOOPBACK_XAUI]    = "XAUI",
+	[LOOPBACK_PHY]	   = "PHY",
+	[LOOPBACK_PHYXS]   = "PHY(XS)",
+	[LOOPBACK_PCS]     = "PHY(PCS)",
+	[LOOPBACK_PMAPMD]  = "PHY(PMAPMD)",
+	[LOOPBACK_NETWORK] = "NETWORK",
+};
+
+/* Interrupt mode names (see INT_MODE())) */
+const unsigned int efx_interrupt_mode_max = EFX_INT_MODE_MAX;
+const char *efx_interrupt_mode_names[] = {
+	[EFX_INT_MODE_MSIX]   = "MSI-X",
+	[EFX_INT_MODE_MSI]    = "MSI",
+	[EFX_INT_MODE_LEGACY] = "legacy",
+};
+
+/* PHY type names (see PHY_TYPE())) */
+const unsigned int efx_phy_type_max = PHY_TYPE_MAX;
+const char *efx_phy_type_names[] = {
+	[PHY_TYPE_NONE]        = "none",
+	[PHY_TYPE_CX4_RTMR]    = "Mysticom CX4",
+	[PHY_TYPE_1G_ALASKA]   = "1G Alaska",
+	[PHY_TYPE_10XPRESS]    = "SFC 10Xpress",
+	[PHY_TYPE_XFP]         = "Quake XFP",
+	[PHY_TYPE_PM8358]      = "PM8358 XAUI",
+};
+
+const unsigned int efx_reset_type_max = RESET_TYPE_MAX;
+const char *efx_reset_type_names[] = {
+	[RESET_TYPE_INVISIBLE]    = "INVISIBLE",
+	[RESET_TYPE_ALL]          = "ALL",
+	[RESET_TYPE_WORLD]        = "WORLD",
+	[RESET_TYPE_DISABLE]      = "DISABLE",
+	[RESET_TYPE_MONITOR]      = "MONITOR",
+	[RESET_TYPE_INT_ERROR]    = "INT_ERROR",
+	[RESET_TYPE_RX_RECOVERY]  = "RX_RECOVERY",
+};
+
+const unsigned int efx_nic_state_max = STATE_MAX;
+const char *efx_nic_state_names[] = {
+	[STATE_INIT]          = "INIT",
+	[STATE_RUNNING]       = "RUNNING",
+	[STATE_FINI]          = "FINI",
+	[STATE_RESETTING]     = "RESETTING",
+	[STATE_DISABLED]      = "DISABLED",
+};
+
+#define EFX_MAX_MTU (9 * 1024)
+
+
+/**************************************************************************
+ *
+ * Configurable values
+ *
+ *************************************************************************/
+
+/*
+ * Use separate channels for TX and RX events
+ *
+ * Set this to 1 to use separate channels for TX and RX. It allows us to
+ * apply a higher level of interrupt moderation to TX events.
+ *
+ * This is forced to 0 for MSI interrupt mode as the interrupt vector
+ * is not written
+ */
+static unsigned int separate_tx_and_rx_channels = 1;
+
+/* This is the weight assigned to each of the (per-channel) virtual
+ * NAPI devices.
+ */
+static int napi_weight = 64;
+
+/* This is the time (in jiffies) between invocations of the hardware
+ * monitor, which checks for known hardware bugs and resets the
+ * hardware and driver as necessary.
+ */
+unsigned int efx_monitor_interval = 1 * HZ;
+
+/* This controls whether or not the hardware monitor will trigger a
+ * reset when it detects an error condition.
+ */
+static unsigned int monitor_reset = 1;
+
+/* This controls whether or not the driver will initialise devices
+ * with invalid MAC addresses stored in the EEPROM or flash.  If true,
+ * such devices will be initialised with a random locally-generated
+ * MAC address.  This allows for loading the efx_mtd driver to
+ * reprogram the flash, even if the flash contents (including the MAC
+ * address) have previously been erased.
+ */
+static unsigned int allow_bad_hwaddr;
+
+/* Initial interrupt moderation settings.  They can be modified after
+ * module load with ethtool.
+ *
+ * The default for RX should strike a balance between increasing the
+ * round-trip latency and reducing overhead.
+ */
+static unsigned int rx_irq_mod_usec = 60;
+
+/* Initial interrupt moderation settings.  They can be modified after
+ * module load with ethtool.
+ *
+ * This default is chosen to ensure that a 10G link does not go idle
+ * while a TX queue is stopped after it has become full.  A queue is
+ * restarted when it drops below half full.  The time this takes (assuming
+ * worst case 3 descriptors per packet and 1024 descriptors) is
+ *   512 / 3 * 1.2 = 205 usec.
+ */
+static unsigned int tx_irq_mod_usec = 150;
+
+/* Ignore online self-test failures at load
+ *
+ * If set to 1, then the driver will not fail to load
+ * if the online self-test fails. Useful only during testing
+ */
+static unsigned int allow_load_on_failure;
+
+/* Set to 1 to enable the use of Message-Signalled Interrupts (MSI).
+ * MSI will not work on some motherboards due to limitations of the
+ * chipset, so the default is off.
+ *
+ * This is the highest capability interrupt mode to use
+ * 0 => MSI-X
+ * 1 => MSI
+ * 2 => legacy
+ */
+static unsigned int interrupt_mode;
+
+/* If set to 1, then the driver will perform an offline self test
+ * when each interface first comes up. This will appear like the
+ * interface bounces up and down
+ */
+static unsigned int onload_offline_selftest = 1;
+
+/* This is the requested number of CPUs to use for Receive-Side Scaling (RSS),
+ * i.e. the number of CPUs among which we may distribute simultaneous
+ * interrupt handling.
+ *
+ * Cards without MSI-X will only target one CPU
+ *
+ * Default (0) means to use all CPUs in the system.  This parameter
+ * can be set using "rss_cpus=xxx" when loading the module.
+ */
+static unsigned int rss_cpus;
+module_param(rss_cpus, uint, 0444);
+MODULE_PARM_DESC(rss_cpus, "Number of CPUs to use for Receive-Side Scaling");
+
+/**************************************************************************
+ *
+ * Utility functions and prototypes
+ *
+ *************************************************************************/
+static void efx_remove_channel(struct efx_channel *channel);
+static void efx_remove_port(struct efx_nic *efx);
+static void efx_fini_napi(struct efx_nic *efx);
+static void efx_fini_channels(struct efx_nic *efx);
+
+/**************************************************************************
+ *
+ * Event queue processing
+ *
+ *************************************************************************/
+
+/* Process channel's event queue
+ *
+ * This function is responsible for processing the event queue of a
+ * single channel.  The caller must guarantee that this function will
+ * never be concurrently called more than once on the same channel,
+ * though different channels may be being processed concurrently.
+ */
+static inline int efx_process_channel(struct efx_channel *channel, int rx_quota)
+{
+	int rxdmaqs;
+	struct efx_rx_queue *rx_queue;
+
+	if (unlikely(channel->efx->reset_pending != RESET_TYPE_NONE ||
+		     !channel->enabled))
+		return rx_quota;
+
+	rxdmaqs = falcon_process_eventq(channel, &rx_quota);
+
+	/* Deliver last RX packet. */
+	if (channel->rx_pkt) {
+		__efx_rx_packet(channel, channel->rx_pkt,
+				channel->rx_pkt_csummed);
+		channel->rx_pkt = NULL;
+	}
+
+	efx_rx_strategy(channel);
+
+	/* Refill descriptor rings as necessary */
+	rx_queue = &channel->efx->rx_queue[0];
+	while (rxdmaqs) {
+		if (rxdmaqs & 0x01)
+			efx_fast_push_rx_descriptors(rx_queue);
+		rx_queue++;
+		rxdmaqs >>= 1;
+	}
+
+	return rx_quota;
+}
+
+/* Mark channel as finished processing
+ *
+ * Note that since we will not receive further interrupts for this
+ * channel before we finish processing and call the eventq_read_ack()
+ * method, there is no need to use the interrupt hold-off timers.
+ */
+static inline void efx_channel_processed(struct efx_channel *channel)
+{
+	/* Write to EVQ_RPTR_REG.  If a new event arrived in a race
+	 * with finishing processing, a new interrupt will be raised.
+	 */
+	channel->work_pending = 0;
+	smp_wmb(); /* Ensure channel updated before any new interrupt. */
+	falcon_eventq_read_ack(channel);
+}
+
+/* NAPI poll handler
+ *
+ * NAPI guarantees serialisation of polls of the same device, which
+ * provides the guarantee required by efx_process_channel().
+ */
+#if !defined(EFX_HAVE_OLD_NAPI)
+static int efx_poll(struct napi_struct *napi, int budget)
+{
+	struct efx_channel *channel =
+		container_of(napi, struct efx_channel, napi_str);
+	struct net_device *napi_dev = channel->napi_dev;
+#else
+static int efx_poll(struct net_device *napi, int *budget_ret)
+{
+	struct net_device *napi_dev = napi;
+	struct efx_channel *channel = napi_dev->priv;
+	int budget = min(napi_dev->quota, *budget_ret);
+#endif
+	int unused;
+	int rx_packets;
+
+	EFX_TRACE(channel->efx, "channel %d NAPI poll executing on CPU %d\n",
+		  channel->channel, raw_smp_processor_id());
+
+	unused = efx_process_channel(channel, budget);
+	rx_packets = (budget - unused);
+#if defined(EFX_HAVE_OLD_NAPI)
+	napi_dev->quota -= rx_packets;
+	*budget_ret -= rx_packets;
+#endif
+
+	if (rx_packets < budget) {
+		/* There is no race here; although napi_disable() will
+		 * only wait for netif_rx_complete(), this isn't a problem
+		 * since efx_channel_processed() will have no effect if
+		 * interrupts have already been disabled.
+		 */
+		netif_rx_complete(napi_dev, napi);
+		efx_channel_processed(channel);
+	}
+
+#if !defined(EFX_HAVE_OLD_NAPI)
+	return rx_packets;
+#else
+	return (rx_packets >= budget);
+#endif
+}
+
+/* Process the eventq of the specified channel immediately on this CPU
+ *
+ * Disable hardware generated interrupts, wait for any existing
+ * processing to finish, then directly poll (and ack ) the eventq.
+ * Finally reenable NAPI and interrupts.
+ *
+ * Since we are touching interrupts the caller should hold the suspend lock
+ */
+void efx_process_channel_now(struct efx_channel *channel)
+{
+	struct efx_nic *efx = channel->efx;
+
+	BUG_ON(!channel->used_flags);
+	BUG_ON(!channel->enabled);
+
+	/* Disable interrupts and wait for ISRs to complete */
+	falcon_disable_interrupts(efx);
+	if (efx->legacy_irq)
+		synchronize_irq(efx->legacy_irq);
+	if (channel->has_interrupt && channel->irq)
+		synchronize_irq(channel->irq);
+
+	/* Wait for any NAPI processing to complete */
+	napi_disable(&channel->napi_str);
+
+	/* Poll the channel */
+	(void) efx_process_channel(channel, efx->type->evq_size);
+
+	/* Ack the eventq. This may cause an interrupt to be generated
+	 * when they are reenabled */
+	efx_channel_processed(channel);
+
+	/* Reenable NAPI polling */
+	napi_enable(&channel->napi_str);
+
+	/* Reenable interrupts */
+	falcon_enable_interrupts(efx);
+}
+
+/* Create event queue
+ * Event queue memory allocations are done only once.  If the channel
+ * is reset, the memory buffer will be reused; this guards against
+ * errors during channel reset and also simplifies interrupt handling.
+ */
+static int efx_probe_eventq(struct efx_channel *channel)
+{
+	EFX_LOG(channel->efx, "chan %d create event queue\n", channel->channel);
+
+	return falcon_probe_eventq(channel);
+}
+
+/* Prepare channel's event queue */
+static int efx_init_eventq(struct efx_channel *channel)
+{
+	EFX_LOG(channel->efx, "chan %d init event queue\n", channel->channel);
+
+	ASSERT_RTNL();
+
+	/* Initialise fields */
+	channel->eventq_read_ptr = 0;
+
+	return falcon_init_eventq(channel);
+}
+
+static void efx_fini_eventq(struct efx_channel *channel)
+{
+	EFX_LOG(channel->efx, "chan %d fini event queue\n", channel->channel);
+
+	ASSERT_RTNL();
+
+	falcon_fini_eventq(channel);
+}
+
+static void efx_remove_eventq(struct efx_channel *channel)
+{
+	EFX_LOG(channel->efx, "chan %d remove event queue\n", channel->channel);
+
+	falcon_remove_eventq(channel);
+}
+
+/**************************************************************************
+ *
+ * Channel handling
+ *
+ *************************************************************************/
+
+/* Setup per-NIC RX buffer parameters.
+ * Calculate the rx buffer allocation parameters required to support
+ * the current MTU, including padding for header alignment and overruns.
+ */
+static void efx_calc_rx_buffer_params(struct efx_nic *efx)
+{
+	unsigned int order, len;
+
+	len = (max(EFX_PAGE_IP_ALIGN, NET_IP_ALIGN) +
+	       EFX_MAX_FRAME_LEN(efx->net_dev->mtu) +
+	       efx->type->rx_buffer_padding);
+
+	/* Page-based allocation page-order */
+	for (order = 0; ((1u << order) * PAGE_SIZE) < len; ++order)
+		;
+
+	efx->rx_buffer_len = len;
+	efx->rx_buffer_order = order;
+}
+
+static int efx_probe_channel(struct efx_channel *channel)
+{
+	struct efx_tx_queue *tx_queue;
+	struct efx_rx_queue *rx_queue;
+	int rc;
+
+	EFX_LOG(channel->efx, "creating channel %d\n", channel->channel);
+
+	rc = efx_probe_eventq(channel);
+	if (rc)
+		goto fail1;
+
+	efx_for_each_channel_tx_queue(tx_queue, channel) {
+		rc = efx_probe_tx_queue(tx_queue);
+		if (rc)
+			goto fail2;
+	}
+
+	efx_for_each_channel_rx_queue(rx_queue, channel) {
+		rc = efx_probe_rx_queue(rx_queue);
+		if (rc)
+			goto fail3;
+	}
+
+	channel->n_rx_frm_trunc = 0;
+
+	return 0;
+
+ fail3:
+	efx_for_each_channel_rx_queue(rx_queue, channel)
+		efx_remove_rx_queue(rx_queue);
+ fail2:
+	efx_for_each_channel_tx_queue(tx_queue, channel)
+		efx_remove_tx_queue(tx_queue);
+ fail1:
+	return rc;
+}
+
+
+/* Channels are shutdown and reinitialised whilst the NIC is running
+ * to propagate configuration changes (mtu, checksum offload), or
+ * to clear hardware error conditions
+ */
+static int efx_init_channels(struct efx_nic *efx)
+{
+	struct efx_tx_queue *tx_queue;
+	struct efx_rx_queue *rx_queue;
+	struct efx_channel *channel;
+	int rc = 0;
+
+	/* Recalculate the rx buffer parameters */
+	efx_calc_rx_buffer_params(efx);
+
+	/* Initialise the channels */
+	efx_for_each_channel(channel, efx) {
+		EFX_LOG(channel->efx, "init chan %d\n", channel->channel);
+
+		rc = efx_init_eventq(channel);
+		if (rc)
+			goto err;
+
+		efx_for_each_channel_tx_queue(tx_queue, channel) {
+			rc = efx_init_tx_queue(tx_queue);
+			if (rc)
+				goto err;
+		}
+
+		/* The rx buffer allocation strategy is MTU dependent */
+		efx_rx_strategy(channel);
+
+		efx_for_each_channel_rx_queue(rx_queue, channel) {
+			rc = efx_init_rx_queue(rx_queue);
+			if (rc)
+				goto err;
+		}
+
+		WARN_ON(channel->rx_pkt != NULL);
+		efx_rx_strategy(channel);
+	}
+
+	return 0;
+
+ err:
+	EFX_ERR(efx, "failed to initialise channel %d\n",
+		channel ? channel->channel : -1);
+	efx_fini_channels(efx);
+	return rc;
+}
+
+/* This enables event queue processing and packet transmission.
+ *
+ * Note that this function is not allowed to fail, since that would
+ * introduce too much complexity into the suspend/resume path.
+ */
+static void efx_start_channel(struct efx_channel *channel)
+{
+	struct efx_rx_queue *rx_queue;
+
+	EFX_LOG(channel->efx, "starting chan %d\n", channel->channel);
+
+	if (!(channel->efx->net_dev->flags & IFF_UP))
+		netif_napi_add(channel->napi_dev, &channel->napi_str,
+			       efx_poll, napi_weight);
+
+	/* Mark channel as enabled */
+	channel->work_pending = 0;
+	channel->enabled = 1;
+	smp_wmb(); /* ensure channel updated before first interrupt */
+
+	/* Enable NAPI poll handler */
+	napi_enable(&channel->napi_str);
+
+	/* Load up RX descriptors */
+	efx_for_each_channel_rx_queue(rx_queue, channel)
+		efx_fast_push_rx_descriptors(rx_queue);
+}
+
+/* This disables event queue processing and packet transmission.
+ * This function does not guarantee that all queue processing
+ * (e.g. RX refill) is complete.
+ */
+static void efx_stop_channel(struct efx_channel *channel)
+{
+	struct efx_rx_queue *rx_queue;
+
+	if (!channel->enabled)
+		return;
+
+	EFX_LOG(channel->efx, "stop chan %d\n", channel->channel);
+
+	/* Mark channel as disabled */
+	channel->enabled = 0;
+
+	/* Wait for any NAPI processing to complete */
+	napi_disable(&channel->napi_str);
+
+	/* Ensure that any worker threads have exited or will be
+	 * no-ops.
+	 */
+	efx_for_each_channel_rx_queue(rx_queue, channel) {
+		spin_lock_bh(&rx_queue->add_lock);
+		spin_unlock_bh(&rx_queue->add_lock);
+	}
+}
+
+static void efx_fini_channels(struct efx_nic *efx)
+{
+	struct efx_channel *channel;
+	struct efx_tx_queue *tx_queue;
+	struct efx_rx_queue *rx_queue;
+
+	ASSERT_RTNL();
+
+	efx_for_each_channel(channel, efx) {
+		EFX_LOG(channel->efx, "shut down chan %d\n", channel->channel);
+
+		efx_for_each_channel_rx_queue(rx_queue, channel)
+			efx_fini_rx_queue(rx_queue);
+		efx_for_each_channel_tx_queue(tx_queue, channel)
+			efx_fini_tx_queue(tx_queue);
+	}
+
+	/* Do the event queues last so that we can handle flush events
+	 * for all DMA queues. */
+	efx_for_each_channel(channel, efx) {
+		EFX_LOG(channel->efx, "shut down evq %d\n", channel->channel);
+
+		efx_fini_eventq(channel);
+	}
+}
+
+static void efx_remove_channel(struct efx_channel *channel)
+{
+	struct efx_tx_queue *tx_queue;
+	struct efx_rx_queue *rx_queue;
+
+	EFX_LOG(channel->efx, "destroy chan %d\n", channel->channel);
+
+	efx_for_each_channel_rx_queue(rx_queue, channel)
+		efx_remove_rx_queue(rx_queue);
+	efx_for_each_channel_tx_queue(tx_queue, channel)
+		efx_remove_tx_queue(tx_queue);
+	efx_remove_eventq(channel);
+
+	channel->used_flags = 0;
+}
+
+/**************************************************************************
+ *
+ * Port handling
+ *
+ **************************************************************************/
+
+/* This ensures that the kernel is kept informed (via
+ * netif_carrier_on/off) of the link status, and also maintains the
+ * link status's stop on the port's TX queue.
+ */
+static void efx_link_status_changed(struct efx_nic *efx)
+{
+	unsigned long flags __attribute__ ((unused));
+	int carrier_ok;
+
+	/* Ensure no link status notifications get sent to the OS after the net
+	 * device has been unregistered. */
+	if (!efx->net_dev_registered)
+		return;
+
+	carrier_ok = netif_carrier_ok(efx->net_dev) ? 1 : 0;
+	if (efx->link_up != carrier_ok) {
+		efx->n_link_state_changes++;
+
+		if (efx->link_up)
+			netif_carrier_on(efx->net_dev);
+		else
+			netif_carrier_off(efx->net_dev);
+	}
+
+	/* Inform driverlink client */
+	EFX_DL_CALLBACK(efx, link_change, efx->link_up);
+
+	/* Status message for kernel log */
+	if (efx->link_up) {
+		struct mii_if_info *gmii = &efx->mii;
+		unsigned adv, lpa;
+		/* NONE here means direct XAUI from the controller, with no
+		 * MDIO-attached device we can query. */
+		if (efx->phy_type != PHY_TYPE_NONE) {
+			adv = gmii_advertised(gmii);
+			lpa = gmii_lpa(gmii);
+		} else {
+			lpa = GM_LPA_10000 | LPA_DUPLEX;
+			adv = lpa;
+		}
+		EFX_INFO(efx, "link up at %dMbps %s-duplex "
+			 "(adv %04x lpa %04x) (MTU %d)%s%s%s%s\n",
+			 (efx->link_options & GM_LPA_10000 ? 10000 :
+			  (efx->link_options & GM_LPA_1000 ? 1000 :
+			   (efx->link_options & GM_LPA_100 ? 100 :
+			    10))),
+			 (efx->link_options & GM_LPA_DUPLEX ?
+			  "full" : "half"),
+			 adv, lpa,
+			 efx->net_dev->mtu,
+			 (efx->loopback_mode ? " [" : ""),
+			 (efx->loopback_mode ? LOOPBACK_MODE(efx) : ""),
+			 (efx->loopback_mode ? " LOOPBACK]" : ""),
+			 (efx->promiscuous ? " [PROMISC]" : ""));
+	} else {
+		EFX_INFO(efx, "link down\n");
+	}
+
+}
+
+/* This call reinitialises the MAC to pick up new PHY settings
+ * To call from a context that cannot sleep use reconfigure_work work item
+ * For on_disabled=1 the caller must be serialised against efx_reset,
+ * ideally by holding the rtnl lock.
+ */
+void efx_reconfigure_port(struct efx_nic *efx, int on_disabled)
+{
+	mutex_lock(&efx->mac_lock);
+
+	EFX_LOG(efx, "reconfiguring MAC from PHY settings\n");
+
+	if (on_disabled)
+		ASSERT_RTNL();
+	else if (!efx->port_enabled)
+		goto out;
+
+	efx->mac_op->reconfigure(efx);
+
+out:
+	/* Inform kernel of loss/gain of carrier */
+	efx_link_status_changed(efx);
+
+	mutex_unlock(&efx->mac_lock);
+}
+
+static void efx_reconfigure_work(struct work_struct *data)
+{
+	struct efx_nic *efx = container_of(data, struct efx_nic,
+					   reconfigure_work);
+
+	EFX_LOG(efx, "MAC reconfigure executing on CPU %d\n",
+		raw_smp_processor_id());
+
+	/* Reinitialise MAC to activate new PHY parameters */
+	efx_reconfigure_port(efx, 0);
+}
+
+static int efx_probe_port(struct efx_nic *efx)
+{
+	unsigned char *dev_addr;
+	int rc;
+
+	EFX_LOG(efx, "create port\n");
+
+	/* Connect up MAC/PHY operations table and read MAC address */
+	rc = falcon_probe_port(efx);
+	if (rc)
+		goto err;
+
+	/* Sanity check MAC address */
+	dev_addr = efx->mac_address;
+	if (!is_valid_ether_addr(dev_addr)) {
+		DECLARE_MAC_BUF(mac);
+
+		EFX_ERR(efx, "invalid MAC address %s\n",
+			print_mac(mac, dev_addr));
+		if (!allow_bad_hwaddr) {
+			rc = -EINVAL;
+			goto err;
+		}
+		random_ether_addr(dev_addr);
+		EFX_INFO(efx, "using locally-generated MAC %s\n",
+			 print_mac(mac, dev_addr));
+	}
+
+	/* Register debugfs entries */
+	rc = efx_init_debugfs_port(efx);
+	if (rc)
+		goto err;
+
+	return 0;
+
+ err:
+	efx_remove_port(efx);
+	return rc;
+}
+
+static int efx_init_port(struct efx_nic *efx)
+{
+	int rc;
+
+	EFX_LOG(efx, "init port\n");
+
+	/* The default power state is ON */
+	efx->phy_powered = 1;
+
+	/* Initialise the MAC and PHY */
+	rc = efx->mac_op->init(efx);
+	if (rc)
+		return rc;
+
+	efx->port_initialized = 1;
+
+	/* Reconfigure port to program MAC registers */
+	efx->mac_op->reconfigure(efx);
+
+	return 0;
+}
+
+/* Allow efx_reconfigure_port() to run, and propagate delayed changes
+ * to the promiscuous flag to the MAC if needed */
+static void efx_start_port(struct efx_nic *efx)
+{
+	EFX_LOG(efx, "start port\n");
+	ASSERT_RTNL();
+
+	BUG_ON(efx->port_enabled);
+
+	mutex_lock(&efx->mac_lock);
+	efx->port_enabled = 1;
+	mutex_unlock(&efx->mac_lock);
+
+	if (efx->net_dev_registered) {
+		int promiscuous;
+
+		netif_tx_lock_bh(efx->net_dev);
+		promiscuous = (efx->net_dev->flags & IFF_PROMISC) ? 1 : 0;
+		if (efx->promiscuous != promiscuous) {
+			efx->promiscuous = promiscuous;
+			queue_work(efx->workqueue, &efx->reconfigure_work);
+		}
+		netif_tx_unlock_bh(efx->net_dev);
+	}
+}
+
+/* Prevents efx_reconfigure_port() from executing, and prevents
+ * efx_set_multicast_list() from scheduling efx_reconfigure_work.
+ * efx_reconfigure_work can still be scheduled via NAPI processing
+ * until efx_flush_all() is called */
+static void efx_stop_port(struct efx_nic *efx)
+{
+	EFX_LOG(efx, "stop port\n");
+	ASSERT_RTNL();
+
+	mutex_lock(&efx->mac_lock);
+	efx->port_enabled = 0;
+	mutex_unlock(&efx->mac_lock);
+
+	/* Serialise against efx_set_multicast_list() */
+	if (efx->net_dev_registered) {
+		netif_tx_lock_bh(efx->net_dev);
+		netif_tx_unlock_bh(efx->net_dev);
+	}
+}
+
+static void efx_fini_port(struct efx_nic *efx)
+{
+	EFX_LOG(efx, "shut down port\n");
+
+	if (!efx->port_initialized)
+		return;
+
+	efx->mac_op->fini(efx);
+	efx->port_initialized = 0;
+
+	/* Mark the link down */
+	efx->link_up = 0;
+	efx_link_status_changed(efx);
+}
+
+static void efx_remove_port(struct efx_nic *efx)
+{
+	EFX_LOG(efx, "destroying port\n");
+
+	efx_fini_debugfs_port(efx);
+	falcon_remove_port(efx);
+}
+
+/**************************************************************************
+ *
+ * NIC handling
+ *
+ **************************************************************************/
+
+/* This configures the PCI device to enable I/O and DMA. */
+static int efx_init_io(struct efx_nic *efx)
+{
+	struct pci_dev *pci_dev = efx->pci_dev;
+	int rc;
+
+	EFX_LOG(efx, "initialising I/O\n");
+
+	/* Generic device-enabling code */
+	rc = pci_enable_device(pci_dev);
+	if (rc) {
+		EFX_ERR(efx, "failed to enable PCI device\n");
+		goto fail1;
+	}
+
+	pci_set_master(pci_dev);
+
+	/* Set the PCI DMA mask.  Try all possibilities from our
+	 * genuine mask down to 32 bits, because some architectures
+	 * (e.g. x86_64 with iommu_sac_force set) will allow 40 bit
+	 * masks event though they reject 46 bit masks.
+	 */
+	efx->dma_mask = efx->type->max_dma_mask;
+	while (efx->dma_mask > 0x7fffffffUL) {
+		if (pci_dma_supported(pci_dev, efx->dma_mask) &&
+		    ((rc = pci_set_dma_mask(pci_dev, efx->dma_mask)) == 0))
+			break;
+		efx->dma_mask >>= 1;
+	}
+	if (rc) {
+		EFX_ERR(efx, "could not find a suitable DMA mask\n");
+		goto fail2;
+	}
+	EFX_LOG(efx, "using DMA mask %llx\n",
+		(unsigned long long)efx->dma_mask);
+	rc = pci_set_consistent_dma_mask(pci_dev, efx->dma_mask);
+	if (rc) {
+		/* pci_set_consistent_dma_mask() is not *allowed* to
+		 * fail with a mask that pci_set_dma_mask() accepted,
+		 * but just in case...
+		 */
+		EFX_ERR(efx, "failed to set consistent DMA mask\n");
+		goto fail2;
+	}
+
+	/* Get memory base address */
+	efx->membase_phys = pci_resource_start(efx->pci_dev,
+					       efx->type->mem_bar);
+#if !defined(EFX_HAVE_MSIX_TABLE_RESERVED)
+	rc = pci_request_region(pci_dev, efx->type->mem_bar, "sfc");
+#else
+	if (!request_mem_region(efx->membase_phys, efx->type->mem_map_size,
+				"sfc"))
+		rc = -EIO;
+#endif
+	if (rc) {
+		EFX_ERR(efx, "request for memory BAR failed\n");
+		rc = -EIO;
+		goto fail3;
+	}
+	efx->membase = ioremap_nocache(efx->membase_phys,
+				       efx->type->mem_map_size);
+	if (!efx->membase) {
+		EFX_ERR(efx, "could not map memory BAR %d at %lx+%x\n",
+			efx->type->mem_bar, efx->membase_phys,
+			efx->type->mem_map_size);
+		rc = -ENOMEM;
+		goto fail4;
+	}
+	EFX_LOG(efx, "memory BAR %u at %lx+%x (virtual %p)\n",
+		efx->type->mem_bar, efx->membase_phys, efx->type->mem_map_size,
+		efx->membase);
+
+	return 0;
+
+ fail4:
+	release_mem_region(efx->membase_phys, efx->type->mem_map_size);
+ fail3:
+	efx->membase_phys = 0UL;
+	/* fall-thru */
+ fail2:
+	pci_disable_device(efx->pci_dev);
+ fail1:
+	return rc;
+}
+
+static void efx_fini_io(struct efx_nic *efx)
+{
+	EFX_LOG(efx, "shutting down I/O\n");
+
+	if (efx->membase) {
+		iounmap(efx->membase);
+		efx->membase = NULL;
+	}
+
+	if (efx->membase_phys) {
+#if !defined(EFX_HAVE_MSIX_TABLE_RESERVED)
+		pci_release_region(efx->pci_dev, efx->type->mem_bar);
+#else
+		release_mem_region(efx->membase_phys, efx->type->mem_map_size);
+#endif
+		efx->membase_phys = 0UL;
+	}
+
+	pci_disable_device(efx->pci_dev);
+}
+
+/* Probe the number and type of interrupts we are able to obtain. */
+static int efx_probe_interrupts(struct efx_nic *efx)
+{
+	struct msix_entry xentries[EFX_MAX_CHANNELS];
+	int rc, i;
+
+	/* Select number of used RSS queues */
+	/* TODO: Can we react to CPU hotplug? */
+	if (rss_cpus == 0)
+		rss_cpus = num_online_cpus();
+
+	efx->rss_queues = 1;
+	if (efx->interrupt_mode == EFX_INT_MODE_MSIX) {
+		unsigned int max_channel = efx->type->phys_addr_channels - 1;
+
+		BUG_ON(!pci_find_capability(efx->pci_dev, PCI_CAP_ID_MSIX));
+		efx->rss_queues = min(max_channel + 1, rss_cpus);
+		efx->rss_queues = min(efx->rss_queues, EFX_MAX_CHANNELS);
+	}
+
+	/* Determine how many RSS queues we can use, and mark channels
+	 * with the appropriate interrupt state */
+	if (efx->interrupt_mode == EFX_INT_MODE_MSIX) {
+		/* Build MSI request structure */
+		for (i = 0; i < efx->rss_queues; i++)
+			xentries[i].entry = i;
+
+		/* Request maximum number of MSI interrupts */
+		rc = pci_enable_msix(efx->pci_dev, xentries, efx->rss_queues);
+		if (rc > 0) {
+			EFX_BUG_ON_PARANOID(rc >= efx->rss_queues);
+			efx->rss_queues = rc;
+			rc = pci_enable_msix(efx->pci_dev, xentries,
+					     efx->rss_queues);
+		}
+		if (rc == 0) {
+			for (i = 0; i < efx->rss_queues; i++) {
+				efx->channel[i].has_interrupt = 1;
+				efx->channel[i].irq = xentries[i].vector;
+			}
+		} else {
+			/* Fall back to single channel MSI */
+			efx->interrupt_mode = EFX_INT_MODE_MSI;
+			EFX_ERR(efx, "could not enable MSI-X\n");
+		}
+	}
+
+	/* Try single interrupt MSI */
+	if (efx->interrupt_mode == EFX_INT_MODE_MSI) {
+		efx->rss_queues = 1;
+		rc = pci_enable_msi(efx->pci_dev);
+		if (rc == 0) {
+			efx->channel[0].irq = efx->pci_dev->irq;
+			efx->channel[0].has_interrupt = 1;
+		} else {
+			EFX_ERR(efx, "could not enable MSI\n");
+			efx->interrupt_mode = EFX_INT_MODE_LEGACY;
+		}
+	}
+
+	/* Assume legacy interrupts */
+	if (efx->interrupt_mode == EFX_INT_MODE_LEGACY) {
+		/* Every channel is interruptible */
+		for (i = 0; i < EFX_MAX_CHANNELS; i++)
+			efx->channel[i].has_interrupt = 1;
+		efx->legacy_irq = efx->pci_dev->irq;
+	}
+
+	return 0;
+}
+
+static void efx_remove_interrupts(struct efx_nic *efx)
+{
+	struct efx_channel *channel;
+
+	/* Remove MSI/MSI-X interrupts */
+	efx_for_each_channel_with_interrupt(channel, efx)
+		channel->irq = 0;
+	pci_disable_msi(efx->pci_dev);
+	pci_disable_msix(efx->pci_dev);
+
+	/* Remove legacy interrupt */
+	efx->legacy_irq = 0;
+}
+
+/* Select number of used resources
+ * Should be called after probe_interrupts()
+ */
+static int efx_select_used(struct efx_nic *efx)
+{
+	struct efx_tx_queue *tx_queue;
+	struct efx_rx_queue *rx_queue;
+	int i;
+
+	/* TX queues.  One per port per channel with TX capability
+	 * (more than one per port won't work on Linux, due to out
+	 *  of order issues... but will be fine on Solaris)
+	 */
+	tx_queue = &efx->tx_queue[0];
+
+	/* Perform this for each channel with TX capabilities.
+	 * At the moment, we only support a single TX queue
+	 */
+	tx_queue->used = 1;
+	if ((!EFX_INT_MODE_USE_MSI(efx)) && separate_tx_and_rx_channels)
+		tx_queue->channel = &efx->channel[1];
+	else
+		tx_queue->channel = &efx->channel[0];
+	tx_queue->channel->used_flags |= EFX_USED_BY_TX;
+	tx_queue++;
+
+	/* RX queues.  Each has a dedicated channel. */
+	for (i = 0; i < EFX_MAX_RX_QUEUES; i++) {
+		rx_queue = &efx->rx_queue[i];
+
+		if (i < efx->rss_queues) {
+			rx_queue->used = 1;
+			/* If we allow multiple RX queues per channel
+			 * we need to decide that here
+			 */
+			rx_queue->channel = &efx->channel[rx_queue->queue];
+			rx_queue->channel->used_flags |= EFX_USED_BY_RX;
+			rx_queue++;
+		}
+	}
+	return 0;
+}
+
+static int efx_probe_nic(struct efx_nic *efx)
+{
+	int rc;
+
+	EFX_LOG(efx, "creating NIC\n");
+
+	/* Carry out hardware-type specific initialisation */
+	rc = falcon_probe_nic(efx);
+	if (rc)
+		goto fail1;
+
+	/* Determine the number of channels and RX queues by trying to hook
+	 * in MSI-X interrupts. */
+	rc = efx_probe_interrupts(efx);
+	if (rc)
+		goto fail2;
+
+	/* Determine number of RX queues and TX queues */
+	rc = efx_select_used(efx);
+	if (rc)
+		goto fail3;
+
+	/* Register debugfs entries */
+	rc = efx_init_debugfs_nic(efx);
+	if (rc)
+		goto fail4;
+	/* Initialise the interrupt moderation settings */
+	efx_init_irq_moderation(efx, tx_irq_mod_usec, rx_irq_mod_usec);
+
+	return 0;
+
+ fail4:
+	/* fall-thru */
+ fail3:
+	efx_remove_interrupts(efx);
+ fail2:
+	falcon_remove_nic(efx);
+ fail1:
+	return rc;
+}
+
+static void efx_remove_nic(struct efx_nic *efx)
+{
+	EFX_LOG(efx, "destroying NIC\n");
+
+	efx_remove_interrupts(efx);
+	falcon_remove_nic(efx);
+
+	efx_fini_debugfs_nic(efx);
+}
+
+/**************************************************************************
+ *
+ * NIC startup/shutdown
+ *
+ *************************************************************************/
+
+static int efx_probe_all(struct efx_nic *efx)
+{
+	struct efx_channel *channel;
+	int rc;
+
+	/* Create NIC */
+	rc = efx_probe_nic(efx);
+	if (rc) {
+		EFX_ERR(efx, "failed to create NIC\n");
+		goto fail1;
+	}
+
+	/* Create port */
+	rc = efx_probe_port(efx);
+	if (rc) {
+		EFX_ERR(efx, "failed to create port\n");
+		goto fail2;
+	}
+
+	/* Create channels */
+	efx_for_each_channel(channel, efx) {
+		rc = efx_probe_channel(channel);
+		if (rc) {
+			EFX_ERR(efx, "failed to create channel %d\n",
+				channel->channel);
+			goto fail3;
+		}
+	}
+
+	return 0;
+
+ fail3:
+	efx_for_each_channel(channel, efx)
+		efx_remove_channel(channel);
+ fail2:
+	efx_remove_port(efx);
+ fail1:
+	return rc;
+}
+
+/* Called after previous invocation(s) of efx_stop_all, restarts the
+ * port, kernel transmit queue, NAPI processing and hardware interrupts.
+ * This function is safe to call multiple times when the NIC is in any
+ * state. */
+static void efx_start_all(struct efx_nic *efx)
+{
+	struct efx_channel *channel;
+
+	ASSERT_RTNL();
+
+	/* Check that it is appropriate to restart the interface. All
+	 * of these flags are safe to read under just the rtnl lock */
+	if (efx->port_enabled)
+		return;
+	if ((efx->state != STATE_RUNNING) && (efx->state != STATE_INIT))
+		return;
+	if (efx->net_dev_registered && !netif_running(efx->net_dev))
+		return;
+
+	/* Mark the port as enabled so port reconfigurations can start, then
+	 * restart the transmit interface early so the watchdog timer stops */
+	efx_start_port(efx);
+	efx_wake_queue(efx);
+
+	efx_for_each_channel(channel, efx)
+		efx_start_channel(channel);
+
+	falcon_enable_interrupts(efx);
+
+	/* Start hardware monitor if we're in RUNNING */
+	if (efx->state == STATE_RUNNING)
+		queue_delayed_work(efx->workqueue, &efx->monitor_work,
+				   efx_monitor_interval);
+}
+
+/* Flush all delayed work. Should only be called when no more delayed work
+ * will be scheduled. This doesn't flush pending online resets (efx_reset),
+ * since we're holding the rtnl_lock at this point. */
+static void efx_flush_all(struct efx_nic *efx)
+{
+#if defined(EFX_USE_CANCEL_DELAYED_WORK_SYNC)
+	struct efx_rx_queue *rx_queue;
+
+	/* Make sure the hardware monitor is stopped */
+	cancel_delayed_work_sync(&efx->monitor_work);
+
+	/* Ensure that all RX slow refills are complete. */
+	efx_for_each_rx_queue(rx_queue, efx) {
+		cancel_delayed_work_sync(&rx_queue->work);
+	}
+#endif
+
+#if defined(EFX_USE_CANCEL_WORK_SYNC)
+	/* Stop scheduled port reconfigurations */
+	cancel_work_sync(&efx->reconfigure_work);
+#endif
+
+#if !defined(EFX_USE_CANCEL_DELAYED_WORK_SYNC)
+	/* Ensure that the hardware monitor and asynchronous port
+	 * reconfigurations are complete, which are the only two consumers
+	 * of efx->workqueue. Since the hardware monitor runs on a long period,
+	 * we put in some effort to cancel the delayed work safely rather
+	 * than just flushing the queue twice (which is guaranteed to flush
+	 * all the work since both efx_monitor and efx_reconfigure_work disarm
+	 * if !efx->port_enabled. */
+	if (timer_pending(&efx->monitor_work.timer))
+		cancel_delayed_work(&efx->monitor_work);
+	flush_workqueue(efx->workqueue);
+	if (timer_pending(&efx->monitor_work.timer))
+		cancel_delayed_work(&efx->monitor_work);
+	flush_workqueue(efx->workqueue);
+
+	/* efx_rx_work will disarm if !channel->enabled, so we can just
+	 * flush the refill workqueue twice as well. */
+	flush_workqueue(efx->refill_workqueue);
+	flush_workqueue(efx->refill_workqueue);
+#endif
+}
+
+/* Quiesce hardware and software without bringing the link down.
+ * Safe to call multiple times, when the nic and interface is in any
+ * state. The caller is guaranteed to subsequently be in a position
+ * to modify any hardware and software state they see fit without
+ * taking locks. */
+static void efx_stop_all(struct efx_nic *efx)
+{
+	struct efx_channel *channel;
+
+	ASSERT_RTNL();
+
+	/* port_enabled can be read safely under the rtnl lock */
+	if (!efx->port_enabled)
+		return;
+
+	/* Disable interrupts and wait for ISR to complete */
+	falcon_disable_interrupts(efx);
+	if (efx->legacy_irq)
+		synchronize_irq(efx->legacy_irq);
+	efx_for_each_channel_with_interrupt(channel, efx)
+		if (channel->irq)
+			synchronize_irq(channel->irq);
+
+	/* Stop all synchronous port reconfigurations. */
+	efx_stop_port(efx);
+
+	/* Stop all NAPI processing and synchronous rx refills */
+	efx_for_each_channel(channel, efx)
+		efx_stop_channel(channel);
+
+	/* Flush reconfigure_work, refill_workqueue, monitor_work */
+	efx_flush_all(efx);
+
+	/* Stop the kernel transmit interface late, so the watchdog
+	 * timer isn't ticking over the flush */
+	efx_stop_queue(efx);
+	if (efx->net_dev_registered) {
+		netif_tx_lock_bh(efx->net_dev);
+		netif_tx_unlock_bh(efx->net_dev);
+	}
+}
+
+static void efx_remove_all(struct efx_nic *efx)
+{
+	struct efx_channel *channel;
+
+	efx_for_each_channel(channel, efx)
+		efx_remove_channel(channel);
+	efx_remove_port(efx);
+	efx_remove_nic(efx);
+}
+
+static int efx_run_selftests(struct efx_nic *efx)
+{
+	struct efx_self_tests tests;
+	unsigned modes = efx->startup_loopbacks & efx->loopback_modes;
+	int rc;
+
+	rc = efx_online_test(efx, &tests);
+	if (rc) {
+		EFX_ERR(efx, "failed self-tests with interrupt_mode of %s\n",
+			INT_MODE(efx));
+		goto fail;
+	}
+
+	if (onload_offline_selftest && modes) {
+		/* Run offline self test */
+		EFX_LOG(efx, "performing on-load offline self-tests\n");
+		rc = efx_offline_test(efx, &tests, modes);
+		EFX_LOG(efx, "%s on-load offline self-tests\n",
+			rc ? "FAILED" : "PASSED");
+		if (rc)
+			goto fail;
+	}
+
+	return 0;
+
+ fail:
+	EFX_ERR(efx, "self-tests failed. Given up!\n");
+	if (allow_load_on_failure)
+		rc = 0;
+
+	return rc;
+}
+
+int efx_flush_queues(struct efx_nic *efx)
+{
+	int rc;
+
+	ASSERT_RTNL();
+
+	efx_stop_all(efx);
+
+	/* We can't just flush the tx queues because the event queues
+	 * may contain tx completions from that queue. Just flush everything */
+	efx_fini_channels(efx);
+	rc = efx_init_channels(efx);
+	if (rc) {
+		efx_schedule_reset(efx, RESET_TYPE_DISABLE);
+		return rc;
+	}
+
+	efx_start_all(efx);
+
+	return 0;
+}
+
+/**************************************************************************
+ *
+ * Interrupt moderation
+ *
+ **************************************************************************/
+
+/* Set interrupt moderation parameters */
+void efx_init_irq_moderation(struct efx_nic *efx, int tx_usecs, int rx_usecs)
+{
+	struct efx_tx_queue *tx_queue;
+	struct efx_rx_queue *rx_queue;
+
+	ASSERT_RTNL();
+
+	efx_for_each_tx_queue(tx_queue, efx)
+		tx_queue->channel->irq_moderation = tx_usecs;
+
+	efx_for_each_rx_queue(rx_queue, efx)
+		rx_queue->channel->irq_moderation = rx_usecs;
+}
+
+/**************************************************************************
+ *
+ * Hardware monitor
+ *
+ **************************************************************************/
+
+/* Run periodically off the general workqueue. Serialised against
+ * efx_reconfigure_port via the mac_lock */
+static void efx_monitor(struct work_struct *data)
+{
+#if !defined(EFX_NEED_WORK_API_WRAPPERS)
+	struct efx_nic *efx = container_of(data, struct efx_nic,
+					   monitor_work.work);
+#else
+	struct efx_nic *efx = container_of(data, struct efx_nic,
+					   monitor_work);
+#endif
+	int rc = 0;
+
+	EFX_TRACE(efx, "hardware monitor executing on CPU %d\n",
+		  raw_smp_processor_id());
+
+
+#if !defined(EFX_USE_CANCEL_DELAYED_WORK_SYNC)
+	/* Without cancel_delayed_work_sync(), we have to make sure that
+	 * we don't rearm when port_enabled == 0 */
+	mutex_lock(&efx->mac_lock);
+	if (!efx->port_enabled) {
+		mutex_unlock(&efx->mac_lock);
+		return;
+	}
+
+	rc = efx->mac_op->check_hw(efx);
+#else
+	/* If the mac_lock is already held then it is likely a port
+	 * reconfiguration is already in place, which will likely do
+	 * most of the work of check_hw() anyway. */
+	if (!mutex_trylock(&efx->mac_lock)) {
+		queue_delayed_work(efx->workqueue, &efx->monitor_work,
+				   efx_monitor_interval);
+		return;
+	}
+
+	if (efx->port_enabled)
+		rc = efx->mac_op->check_hw(efx);
+#endif
+	mutex_unlock(&efx->mac_lock);
+
+	if (rc) {
+		if (monitor_reset) {
+			EFX_ERR(efx, "hardware monitor detected a fault: "
+				"triggering reset\n");
+			efx_schedule_reset(efx, RESET_TYPE_MONITOR);
+		} else {
+			EFX_ERR(efx, "hardware monitor detected a fault, "
+				"skipping reset\n");
+		}
+	}
+
+	queue_delayed_work(efx->workqueue, &efx->monitor_work,
+			   efx_monitor_interval);
+}
+
+/**************************************************************************
+ *
+ * ioctls
+ *
+ *************************************************************************/
+
+/* Net device ioctl
+ * Context: process, rtnl_lock() held.
+ */
+static int efx_ioctl(struct net_device *net_dev, struct ifreq *ifr, int cmd)
+{
+	struct efx_nic *efx = net_dev->priv;
+	int rc;
+
+	ASSERT_RTNL();
+
+	switch (cmd) {
+	case SIOCGMIIPHY:
+	case SIOCGMIIREG:
+		rc = generic_mii_ioctl(&efx->mii, if_mii(ifr), cmd, NULL);
+		break;
+	case SIOCSMIIREG:
+		rc = generic_mii_ioctl(&efx->mii, if_mii(ifr), cmd, NULL);
+		efx_reconfigure_port(efx, 0);
+		break;
+	default:
+		rc = -EOPNOTSUPP;
+	}
+
+	return rc;
+}
+
+/**************************************************************************
+ *
+ * NAPI interface
+ *
+ **************************************************************************/
+
+/* Allocate the NAPI dev's.
+ * Called after we know how many channels there are.
+ */
+static int efx_init_napi(struct efx_nic *efx)
+{
+	struct efx_channel *channel;
+	int rc;
+
+	ASSERT_RTNL();
+
+	/* Allocate the NAPI dev for the port */
+	efx->net_dev = alloc_etherdev(0);
+	if (!efx->net_dev) {
+		rc = -ENOMEM;
+		goto err;
+	}
+	efx->net_dev->priv = efx;
+	efx->mii.dev = efx->net_dev;
+
+	/* Set features based on module parameters and DMA mask.
+	 * Enable DMA to ZONE_HIGHMEM if the NIC can access all memory
+	 * directly.  This only has an effect on 32-bit systems and
+	 * PAE on x86 limits memory to 64GB so 40 bits is plenty to
+	 * address everything.  If the device can't address 40 bits
+	 * then it's safest to turn NETIF_F_HIGHDMA off because this
+	 * might be a PAE system with more than 4G of RAM and a 32-bit
+	 * NIC.  The use of EFX_DMA_MASK is to eliminate compiler
+	 * warnings on platforms where dma_addr_t is 32-bit.  We
+	 * assume that in those cases we can access all memory
+	 * directly if our DMA mask is all ones. */
+	efx->net_dev->features |= NETIF_F_IP_CSUM | NETIF_F_SG;
+	if (efx->dma_mask >= EFX_DMA_MASK(DMA_40BIT_MASK))
+		efx->net_dev->features |= NETIF_F_HIGHDMA;
+
+	/* Copy MAC address */
+	memcpy(&efx->net_dev->dev_addr, efx->mac_address, ETH_ALEN);
+
+	/* Allocate the per channel devs */
+	efx_for_each_channel(channel, efx) {
+#if !defined(EFX_HAVE_OLD_NAPI)
+		channel->napi_dev = efx->net_dev;
+#else
+		channel->napi_dev = alloc_etherdev(0);
+		if (!channel->napi_dev) {
+			rc = -ENOMEM;
+			goto err;
+		}
+		channel->napi_dev->priv = channel;
+		atomic_set(&channel->napi_dev->refcnt, 1);
+#endif
+	}
+
+	return 0;
+ err:
+	efx_fini_napi(efx);
+	return rc;
+}
+
+/* Free the NAPI state for the port and channels */
+static void efx_fini_napi(struct efx_nic *efx)
+{
+	struct efx_channel *channel;
+
+	ASSERT_RTNL();
+
+	efx_for_each_channel(channel, efx) {
+		/* Finish per channel NAPI */
+#if defined(EFX_HAVE_OLD_NAPI)
+		if (channel->napi_dev) {
+			channel->napi_dev->priv = NULL;
+			free_netdev(channel->napi_dev);
+		}
+#endif
+		channel->napi_dev = NULL;
+	}
+
+	/* Finish port NAPI */
+	if (efx->net_dev) {
+		efx->net_dev->priv = NULL;
+		free_netdev(efx->net_dev);
+		efx->net_dev = NULL;
+	}
+}
+
+/**************************************************************************
+ *
+ * Kernel netpoll interface
+ *
+ *************************************************************************/
+
+#ifdef CONFIG_NET_POLL_CONTROLLER
+
+/* Although in the common case interrupts will be disabled, this is not
+ * guaranteed. However, all our work happens inside the NAPI callback,
+ * so no locking is required.
+ */
+static void efx_netpoll(struct net_device *net_dev)
+{
+	struct efx_nic *efx = net_dev->priv;
+	struct efx_channel *channel;
+
+	efx_for_each_channel_with_interrupt(channel, efx)
+		efx_schedule_channel(channel);
+}
+
+#endif
+
+/**************************************************************************
+ *
+ * Kernel net device interface
+ *
+ *************************************************************************/
+
+/* Context: process, rtnl_lock() held. */
+static int efx_net_open(struct net_device *net_dev)
+{
+	struct efx_nic *efx = net_dev->priv;
+	ASSERT_RTNL();
+
+	EFX_LOG(efx, "opening device %s on CPU %d\n", net_dev->name,
+		raw_smp_processor_id());
+	efx_start_all(efx);
+	return 0;
+}
+
+/* Context: process, rtnl_lock() held.
+ * Note that the kernel will ignore our return code; this method
+ * should really be a void.
+ */
+static int efx_net_stop(struct net_device *net_dev)
+{
+	struct efx_nic *efx = net_dev->priv;
+	int rc;
+
+	EFX_LOG(efx, "closing %s on CPU %d\n", net_dev->name,
+		raw_smp_processor_id());
+
+	/* Stop device and flush all the channels */
+	efx_stop_all(efx);
+	efx_fini_channels(efx);
+	rc = efx_init_channels(efx);
+	if (rc)
+		efx_schedule_reset(efx, RESET_TYPE_DISABLE);
+
+	return 0;
+}
+
+/* Context: process, dev_base_lock held, non-blocking.
+ * Statistics are taken directly from the MAC.
+ */
+static struct net_device_stats *efx_net_stats(struct net_device *net_dev)
+{
+	struct efx_nic *efx = net_dev->priv;
+	struct efx_mac_stats *mac_stats = &efx->mac_stats;
+	struct net_device_stats *stats = &efx->stats;
+
+	if (!spin_trylock(&efx->stats_lock))
+		return stats;
+	if (efx->state == STATE_RUNNING)
+		efx->mac_op->update_stats(efx);
+	spin_unlock(&efx->stats_lock);
+
+	stats->rx_packets = mac_stats->rx_packets;
+	stats->tx_packets = mac_stats->tx_packets;
+	stats->rx_bytes = mac_stats->rx_bytes;
+	stats->tx_bytes = mac_stats->tx_bytes;
+	stats->tx_errors = mac_stats->tx_bad;
+	stats->multicast = mac_stats->rx_multicast;
+	stats->collisions = mac_stats->tx_collision;
+	stats->rx_length_errors = mac_stats->rx_gtjumbo;
+	stats->rx_over_errors = mac_stats->rx_overflow;
+	stats->rx_crc_errors = mac_stats->rx_bad;
+	stats->rx_frame_errors = mac_stats->rx_align_error;
+	stats->rx_fifo_errors = 0;
+	stats->rx_missed_errors = mac_stats->rx_missed;
+	stats->rx_errors = (stats->rx_length_errors +
+			    stats->rx_over_errors +
+			    stats->rx_crc_errors +
+			    stats->rx_frame_errors +
+			    stats->rx_fifo_errors +
+			    stats->rx_missed_errors +
+			    mac_stats->rx_symbol_error);
+	stats->tx_aborted_errors = 0;
+	stats->tx_carrier_errors = 0;
+	stats->tx_fifo_errors = 0;
+	stats->tx_heartbeat_errors = 0;
+	stats->tx_window_errors = 0;
+
+	return stats;
+}
+
+/* Context: netif_tx_lock held, BHs disabled. */
+static void efx_watchdog(struct net_device *net_dev)
+{
+	struct efx_nic *efx = net_dev->priv;
+
+	EFX_ERR(efx, "TX stuck with stop_count=%d port_enabled=%d: %s\n",
+		atomic_read(&efx->netif_stop_count), efx->port_enabled,
+		monitor_reset ? "resetting channels" : "skipping reset");
+
+	if (monitor_reset)
+		efx_schedule_reset(efx, RESET_TYPE_MONITOR);
+}
+
+
+/* Context: process, rtnl_lock() held. */
+static int efx_change_mtu(struct net_device *net_dev, int new_mtu)
+{
+	struct efx_nic *efx = net_dev->priv;
+	int rc = 0;
+
+	ASSERT_RTNL();
+
+	if (new_mtu > EFX_MAX_MTU)
+		return -EINVAL;
+
+	efx_stop_all(efx);
+
+	/* Ask driverlink client if we can change MTU */
+	rc = EFX_DL_CALLBACK(efx, request_mtu, new_mtu);
+	if (rc) {
+		EFX_ERR(efx, "MTU change vetoed by driverlink %s driver\n",
+			efx->dl_cb_dev.request_mtu->driver->name);
+		goto out;
+	}
+
+	EFX_LOG(efx, "changing MTU to %d\n", new_mtu);
+
+	efx_fini_channels(efx);
+	net_dev->mtu = new_mtu;
+	rc = efx_init_channels(efx);
+	if (rc)
+		goto fail;
+
+	/* Reconfigure the MAC */
+	efx_reconfigure_port(efx, 1);
+
+	/* Notify driverlink client of new MTU */
+	EFX_DL_CALLBACK(efx, mtu_changed, new_mtu);
+
+	efx_start_all(efx);
+
+ out:
+	return rc;
+
+ fail:
+	efx_schedule_reset(efx, RESET_TYPE_DISABLE);
+	return rc;
+}
+
+static int efx_set_mac_address(struct net_device *net_dev, void *data)
+{
+	struct efx_nic *efx = net_dev->priv;
+	struct sockaddr *addr = data;
+	char *new_addr = addr->sa_data;
+
+	ASSERT_RTNL();
+
+	if (!is_valid_ether_addr(new_addr)) {
+		DECLARE_MAC_BUF(mac);
+		EFX_ERR(efx, "invalid ethernet MAC address requested: %s\n",
+			print_mac(mac, new_addr));
+		return -EINVAL;
+	}
+
+	memcpy(net_dev->dev_addr, new_addr, net_dev->addr_len);
+
+	/* Reconfigure the MAC */
+	efx_reconfigure_port(efx, 1);
+
+	return 0;
+}
+
+/* Context: netif_tx_lock held, BHs disabled. */
+static void efx_set_multicast_list(struct net_device *net_dev)
+{
+	struct efx_nic *efx = net_dev->priv;
+	struct dev_mc_list *mc_list = net_dev->mc_list;
+	union efx_multicast_hash *mc_hash = &efx->multicast_hash;
+	unsigned long flags __attribute__ ((unused));
+	int promiscuous;
+	u32 crc;
+	int bit;
+	int i;
+
+	/* Set per-MAC promiscuity flag and reconfigure MAC if necessary */
+	promiscuous = (net_dev->flags & IFF_PROMISC) ? 1 : 0;
+	if (efx->promiscuous != promiscuous) {
+		if (efx->port_enabled) {
+			efx->promiscuous = promiscuous;
+			queue_work(efx->workqueue, &efx->reconfigure_work);
+		}
+	}
+
+	/* Build multicast hash table */
+	if (promiscuous || (net_dev->flags & IFF_ALLMULTI)) {
+		memset(mc_hash, 0xff, sizeof(*mc_hash));
+	} else {
+		memset(mc_hash, 0x00, sizeof(*mc_hash));
+		for (i = 0; i < net_dev->mc_count; i++) {
+			crc = ether_crc_le(ETH_ALEN, mc_list->dmi_addr);
+			bit = (crc & ((1 << EFX_MCAST_HASH_BITS) - 1));
+			set_bit_le(bit, (void *)mc_hash);
+			mc_list = mc_list->next;
+		}
+	}
+
+	/* Create and activate new global multicast hash table */
+	falcon_set_multicast_hash(efx);
+}
+
+/* Handle net device notifier events */
+static int efx_netdev_event(struct notifier_block *this,
+			    unsigned long event, void *ptr)
+{
+	struct net_device *net_dev = (struct net_device *)ptr;
+
+	if (net_dev->open == efx_net_open && event == NETDEV_CHANGENAME) {
+		struct efx_nic *efx = net_dev->priv;
+
+		strcpy(efx->name, net_dev->name);
+		efx_fini_debugfs_netdev(net_dev);
+		efx_init_debugfs_netdev(net_dev);
+	}
+
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block efx_netdev_notifier = {
+	.notifier_call = efx_netdev_event,
+};
+
+static int efx_register_netdev(struct efx_nic *efx)
+{
+	struct net_device *net_dev = efx->net_dev;
+	int rc;
+
+	net_dev->watchdog_timeo = 5 * HZ;
+	net_dev->irq = efx->pci_dev->irq;
+	net_dev->open = efx_net_open;
+	net_dev->stop = efx_net_stop;
+	net_dev->get_stats = efx_net_stats;
+	net_dev->tx_timeout = &efx_watchdog;
+	net_dev->hard_start_xmit = efx_hard_start_xmit;
+	net_dev->do_ioctl = efx_ioctl;
+	net_dev->change_mtu = efx_change_mtu;
+	net_dev->set_mac_address = efx_set_mac_address;
+	net_dev->set_multicast_list = efx_set_multicast_list;
+#ifdef CONFIG_NET_POLL_CONTROLLER
+	net_dev->poll_controller = efx_netpoll;
+#endif
+	SET_NETDEV_DEV(net_dev, &efx->pci_dev->dev);
+	SET_ETHTOOL_OPS(net_dev, &efx_ethtool_ops);
+
+	/* Always start with carrier off; PHY events will detect the link */
+	netif_carrier_off(efx->net_dev);
+
+	BUG_ON(efx->net_dev_registered);
+
+	/* Clear MAC statistics */
+	efx->mac_op->update_stats(efx);
+	memset(&efx->mac_stats, 0, sizeof(efx->mac_stats));
+
+	rc = register_netdev(net_dev);
+	if (rc) {
+		EFX_ERR(efx, "could not register net dev\n");
+		return rc;
+	}
+	strcpy(efx->name, net_dev->name);
+
+	/* Create debugfs symlinks */
+	rc = efx_init_debugfs_netdev(net_dev);
+	if (rc) {
+		EFX_ERR(efx, "failed to init net dev debugfs\n");
+		unregister_netdev(efx->net_dev);
+		return rc;
+	}
+
+	/* Allow link change notifications to be sent to the operating
+	 * system.  The must happen after register_netdev so that
+	 * there are no outstanding link changes if that call fails.
+	 * It must happen before efx_reconfigure_port so that the
+	 * initial state of the link is reported. */
+	mutex_lock(&efx->mac_lock);
+	efx->net_dev_registered = 1;
+	mutex_unlock(&efx->mac_lock);
+
+	/* Safety net: in case we don't get a PHY event */
+	rtnl_lock();
+	efx_reconfigure_port(efx, 1);
+	rtnl_unlock();
+
+	EFX_LOG(efx, "registered\n");
+
+	return 0;
+}
+
+static void efx_unregister_netdev(struct efx_nic *efx)
+{
+	int was_registered = efx->net_dev_registered;
+	struct efx_tx_queue *tx_queue;
+
+	if (!efx->net_dev)
+		return;
+
+	BUG_ON(efx->net_dev->priv != efx);
+
+	/* SFC Bug 5356: Ensure that no more link status notifications get
+	 * sent to the stack.  Bad things happen if there's an
+	 * outstanding notification after the net device is freed, but
+	 * they only get flushed out by unregister_netdev, not by
+	 * free_netdev. */
+	mutex_lock(&efx->mac_lock);
+	efx->net_dev_registered = 0;
+	mutex_unlock(&efx->mac_lock);
+
+	/* Free up any skbs still remaining. This has to happen before
+	 * we try to unregister the netdev as running their destructors
+	 * may be needed to get the device ref. count to 0. */
+	efx_for_each_tx_queue(tx_queue, efx)
+		efx_release_tx_buffers(tx_queue);
+
+	if (was_registered) {
+		strlcpy(efx->name, pci_name(efx->pci_dev), sizeof(efx->name));
+		efx_fini_debugfs_netdev(efx->net_dev);
+		unregister_netdev(efx->net_dev);
+	}
+}
+
+/**************************************************************************
+ *
+ * Device reset and suspend
+ *
+ **************************************************************************/
+
+/* This suspends the device (and acquires the suspend lock) without
+ * flushing the descriptor queues.  It is included for the convenience
+ * of the driverlink layer.
+ */
+void efx_suspend(struct efx_nic *efx)
+{
+	EFX_LOG(efx, "suspending operations\n");
+
+	down(&efx->suspend_lock);
+
+	rtnl_lock();
+	efx_stop_all(efx);
+}
+
+void efx_resume(struct efx_nic *efx)
+{
+	EFX_LOG(efx, "resuming operations\n");
+
+	efx_start_all(efx);
+	rtnl_unlock();
+
+	up(&efx->suspend_lock);
+}
+
+/* The final hardware and software finalisation before reset.
+ * This function does not handle serialisation with the kernel, it
+ * assumes the caller has done this */
+static int efx_reset_down(struct efx_nic *efx, struct ethtool_cmd *ecmd)
+{
+	int rc;
+
+	ASSERT_RTNL();
+
+	rc = efx->mac_op->get_settings(efx, ecmd);
+	if (rc) {
+		EFX_ERR(efx, "could not back up PHY settings\n");
+		goto fail;
+	}
+
+	efx_fini_channels(efx);
+	return 0;
+
+ fail:
+	return rc;
+}
+
+/* The first part of software initialisation after a hardware reset
+ * This function does not handle serialisation with the kernel, it
+ * assumes the caller has done this */
+static int efx_reset_up(struct efx_nic *efx, struct ethtool_cmd *ecmd)
+{
+	int rc;
+
+	rc = efx_init_channels(efx);
+	if (rc)
+		goto fail1;
+
+	/* In an INVISIBLE_RESET there might not be a link state transition,
+	 * so we push the multicast list here. */
+	falcon_set_multicast_hash(efx);
+
+	/* Restore MAC and PHY settings. */
+	rc = efx->mac_op->set_settings(efx, ecmd);
+	if (rc) {
+		EFX_ERR(efx, "could not restore PHY settings\n");
+		goto fail2;
+	}
+
+	return 0;
+
+ fail2:
+	efx_fini_channels(efx);
+ fail1:
+	return rc;
+}
+
+/* Reset the NIC as transparently as possible. Do not reset the PHY
+ * Note that the reset may fail, in which case the card will be left
+ * in a most-probably-unusable state.
+ *
+ * This function will sleep.  You cannot reset from within an atomic
+ * state; use efx_schedule_reset() instead.
+ */
+static int efx_reset(struct efx_nic *efx)
+{
+	struct ethtool_cmd ecmd;
+	unsigned long flags __attribute__ ((unused));
+	enum reset_type method = efx->reset_pending;
+	int rc;
+
+	efx_dl_reset_lock();
+
+	rc = down_interruptible(&efx->suspend_lock);
+	if (rc) {
+		EFX_ERR(efx, "reset aborted by signal\n");
+		goto unlock_dl_lock;
+	}
+
+	/* We've got suspend_lock, which means we can only be in
+	 * STATE_RUNNING or STATE_FINI. Don't clear
+	 * efx->reset_pending, since this flag indicates that we
+	 * should retry device initialisation.
+	 */
+	if (efx->state != STATE_RUNNING) {
+		EFX_INFO(efx, "scheduled reset quenched. NIC not RUNNING\n");
+		goto unlock_suspend_lock;
+	}
+
+	/* Notify driverlink clients of imminent reset. */
+	efx_dl_reset_suspend(efx);
+	rtnl_lock();
+
+	efx->state = STATE_RESETTING;
+	EFX_INFO(efx, "resetting (%s)\n", RESET_TYPE(method));
+
+	/* The net_dev->get_stats handler is quite slow, and will fail
+	 * if a fetch is pending over reset. Serialise against it. */
+	spin_lock(&efx->stats_lock);
+	spin_unlock(&efx->stats_lock);
+
+	efx_stop_all(efx);
+	mutex_lock(&efx->mac_lock);
+
+	rc = efx_reset_down(efx, &ecmd);
+	if (rc)
+		goto fail1;
+	falcon_fini_nic(efx);
+
+	rc = falcon_reset_hw(efx, method);
+	if (rc) {
+		EFX_ERR(efx, "failed to reset hardware\n");
+		goto fail2;
+	}
+
+	/* Allow resets to be rescheduled. */
+	efx->reset_pending = RESET_TYPE_NONE;
+
+	/* Reinitialise bus-mastering, which may have been turned off before
+	 * the reset was scheduled. This is still appropriate, even in the
+	 * RESET_TYPE_DISABLE since this driver generally assumes the hardware
+	 * can respond to requests. */
+	pci_set_master(efx->pci_dev);
+
+	/* Reinitialise device. This is appropriate in the RESET_TYPE_DISABLE
+	 * case so the driver can talk to external SRAM */
+	rc = falcon_init_nic(efx);
+	if (rc) {
+		EFX_ERR(efx, "failed to initialise NIC\n");
+		goto fail3;
+	}
+
+	/* Leave device stopped if necessary */
+	if (method == RESET_TYPE_DISABLE) {
+		/* Reinitialise the device anyway so the driver unload sequence
+		 * can talk to the external SRAM */
+		(void) falcon_init_nic(efx);
+		rc = -EIO;
+		goto fail4;
+	}
+
+	rc = efx_reset_up(efx, &ecmd);
+	if (rc)
+		goto fail5;
+
+	mutex_unlock(&efx->mac_lock);
+	efx_reconfigure_port(efx, 1);
+	EFX_LOG(efx, "reset complete\n");
+
+	efx->state = STATE_RUNNING;
+	efx_start_all(efx);
+
+	rtnl_unlock();
+
+	goto notify;
+
+ fail5:
+ fail4:
+ fail3:
+ fail2:
+ fail1:
+	EFX_ERR(efx, "has been disabled\n");
+	efx->state = STATE_DISABLED;
+
+	/* Remove the net_dev */
+	mutex_unlock(&efx->mac_lock);
+	rtnl_unlock();
+	efx_unregister_netdev(efx);
+	efx_fini_port(efx);
+
+ notify:
+	/* Notify driverlink clients of completed reset */
+	efx_dl_reset_resume(efx, (rc == 0));
+
+ unlock_suspend_lock:
+	up(&efx->suspend_lock);
+
+ unlock_dl_lock:
+	efx_dl_reset_unlock();
+
+	return rc;
+}
+
+/* The worker thread exists so that code that cannot sleep can
+ * schedule a reset for later.
+ */
+static void efx_reset_work(struct work_struct *data)
+{
+	struct efx_nic *nic = container_of(data, struct efx_nic, reset_work);
+
+	efx_reset(nic);
+}
+
+void efx_schedule_reset(struct efx_nic *efx, enum reset_type type)
+{
+	enum reset_type method;
+
+	if (efx->reset_pending != RESET_TYPE_NONE) {
+		EFX_INFO(efx, "quenching already scheduled reset\n");
+		return;
+	}
+
+	switch (type) {
+	case RESET_TYPE_INVISIBLE:
+	case RESET_TYPE_ALL:
+	case RESET_TYPE_WORLD:
+	case RESET_TYPE_DISABLE:
+		method = type;
+		break;
+	case RESET_TYPE_RX_RECOVERY:
+	case RESET_TYPE_RX_DESC_FETCH:
+	case RESET_TYPE_TX_DESC_FETCH:
+		method = RESET_TYPE_INVISIBLE;
+		break;
+	default:
+		method = RESET_TYPE_ALL;
+		break;
+	}
+
+	if (method != type)
+		EFX_LOG(efx, "scheduling %s reset for %s\n",
+			RESET_TYPE(method), RESET_TYPE(type));
+	else
+		EFX_LOG(efx, "scheduling %s reset\n", RESET_TYPE(method));
+
+	efx->reset_pending = method;
+
+#if !defined(EFX_USE_CANCEL_DELAYED_WORK_SYNC)
+	queue_work(efx->reset_workqueue, &efx->reset_work);
+#else
+	queue_work(efx->workqueue, &efx->reset_work);
+#endif
+}
+
+/**************************************************************************
+ *
+ * List of NICs we support
+ *
+ **************************************************************************/
+
+enum efx_type_index {
+	EFX_TYPE_FALCON_A = 0,
+	EFX_TYPE_FALCON_B = 1,
+};
+
+static struct efx_nic_type *efx_nic_types[] = {
+	[EFX_TYPE_FALCON_A] = &falcon_a_nic_type,
+	[EFX_TYPE_FALCON_B] = &falcon_b_nic_type,
+};
+
+
+/* PCI device ID table */
+static struct pci_device_id efx_pci_table[] __devinitdata = {
+	{EFX_VENDID_SFC, FALCON_A_P_DEVID, PCI_ANY_ID, PCI_ANY_ID,
+	 0, 0, EFX_TYPE_FALCON_A},
+	{EFX_VENDID_SFC, FALCON_B_P_DEVID, PCI_ANY_ID, PCI_ANY_ID,
+	 0, 0, EFX_TYPE_FALCON_B},
+	{0}			/* end of list */
+};
+
+/**************************************************************************
+ *
+ * Dummy PHY/MAC/Board operations
+ *
+ * Can be used where the MAC does not implement this operation
+ * Needed so all function pointers are valid and do not have to be tested
+ * before use
+ *
+ **************************************************************************/
+int efx_port_dummy_op_int(struct efx_nic *efx)
+{
+	return 0;
+}
+void efx_port_dummy_op_void(struct efx_nic *efx) {}
+void efx_port_dummy_op_blink(struct efx_nic *efx, int blink) {}
+
+static struct efx_mac_operations efx_dummy_mac_operations = {
+	.init		= efx_port_dummy_op_int,
+	.reconfigure	= efx_port_dummy_op_void,
+	.fini		= efx_port_dummy_op_void,
+};
+
+static struct efx_phy_operations efx_dummy_phy_operations = {
+	.init		 = efx_port_dummy_op_int,
+	.reconfigure	 = efx_port_dummy_op_void,
+	.check_hw        = efx_port_dummy_op_int,
+	.fini		 = efx_port_dummy_op_void,
+	.clear_interrupt = efx_port_dummy_op_void,
+	.reset_xaui      = efx_port_dummy_op_void,
+};
+
+/* Dummy board operations */
+static int efx_nic_dummy_op_int(struct efx_nic *nic)
+{
+	return 0;
+}
+
+static void efx_nic_dummy_op_void(struct efx_nic *nic) {}
+
+static struct efx_board efx_dummy_board_info = {
+	.init    = efx_nic_dummy_op_int,
+	.init_leds = efx_port_dummy_op_int,
+	.set_fault_led = efx_port_dummy_op_blink,
+	.monitor = efx_nic_dummy_op_int,
+	.blink = efx_port_dummy_op_blink,
+	.fini    = efx_nic_dummy_op_void,
+};
+
+/**************************************************************************
+ *
+ * Data housekeeping
+ *
+ **************************************************************************/
+
+/* This zeroes out and then fills in the invariants in a struct
+ * efx_nic (including all sub-structures).
+ */
+static int efx_init_struct(struct efx_nic *efx, enum efx_type_index type,
+			   struct pci_dev *pci_dev)
+{
+	struct efx_channel *channel;
+	struct efx_tx_queue *tx_queue;
+	struct efx_rx_queue *rx_queue;
+	int i, rc;
+
+	/* Initialise common structures */
+	memset(efx, 0, sizeof(*efx));
+	spin_lock_init(&efx->biu_lock);
+	spin_lock_init(&efx->phy_lock);
+	mutex_init(&efx->spi_lock);
+	sema_init(&efx->suspend_lock, 1);
+	INIT_WORK(&efx->reset_work, efx_reset_work);
+	INIT_DELAYED_WORK(&efx->monitor_work, efx_monitor);
+	efx->pci_dev = pci_dev;
+	efx->state = STATE_INIT;
+	efx->reset_pending = RESET_TYPE_NONE;
+	strlcpy(efx->name, pci_name(pci_dev), sizeof(efx->name));
+	efx->board_info = efx_dummy_board_info;
+
+	efx->rx_checksum_enabled = 1;
+	spin_lock_init(&efx->netif_stop_lock);
+	spin_lock_init(&efx->stats_lock);
+	mutex_init(&efx->mac_lock);
+	efx->mac_op = &efx_dummy_mac_operations;
+	efx->phy_op = &efx_dummy_phy_operations;
+	INIT_LIST_HEAD(&efx->dl_node);
+	INIT_LIST_HEAD(&efx->dl_device_list);
+	efx->dl_cb = efx_default_callbacks;
+	INIT_WORK(&efx->reconfigure_work, efx_reconfigure_work);
+	atomic_set(&efx->netif_stop_count, 1);
+
+	for (i = 0; i < EFX_MAX_CHANNELS; i++) {
+		channel = &efx->channel[i];
+		channel->efx = efx;
+		channel->channel = i;
+		channel->evqnum = i;
+		channel->work_pending = 0;
+	}
+	for (i = 0; i < EFX_MAX_TX_QUEUES; i++) {
+		tx_queue = &efx->tx_queue[i];
+		tx_queue->efx = efx;
+		tx_queue->queue = i;
+		tx_queue->buffer = NULL;
+		tx_queue->channel = &efx->channel[0]; /* for safety */
+	}
+	for (i = 0; i < EFX_MAX_RX_QUEUES; i++) {
+		rx_queue = &efx->rx_queue[i];
+		rx_queue->efx = efx;
+		rx_queue->queue = i;
+		rx_queue->channel = &efx->channel[0]; /* for safety */
+		rx_queue->buffer = NULL;
+		spin_lock_init(&rx_queue->add_lock);
+		INIT_DELAYED_WORK(&rx_queue->work, efx_rx_work);
+	}
+
+	efx->type = efx_nic_types[type];
+
+	/* Sanity-check NIC type */
+	EFX_BUG_ON_PARANOID(efx->type->txd_ring_mask &
+			    (efx->type->txd_ring_mask + 1));
+	EFX_BUG_ON_PARANOID(efx->type->rxd_ring_mask &
+			    (efx->type->rxd_ring_mask + 1));
+	EFX_BUG_ON_PARANOID(efx->type->evq_size &
+			    (efx->type->evq_size - 1));
+	/* As close as we can get to guaranteeing that we don't overflow */
+	EFX_BUG_ON_PARANOID(efx->type->evq_size <
+			    (efx->type->txd_ring_mask + 1 +
+			     efx->type->rxd_ring_mask + 1));
+
+	EFX_BUG_ON_PARANOID(efx->type->phys_addr_channels > EFX_MAX_CHANNELS);
+
+	/* Higher numbered interrupt modes are less capable! */
+	efx->interrupt_mode = max(efx->type->max_interrupt_mode,
+				  interrupt_mode);
+#if defined(EFX_NEED_DUMMY_MSIX)
+	if (efx->interrupt_mode == EFX_INT_MODE_MSIX)
+		efx->interrupt_mode = EFX_INT_MODE_MSI;
+#endif
+
+	/* Tasks that can fail are last */
+	efx->refill_workqueue = create_workqueue("sfc_refill");
+	if (!efx->refill_workqueue) {
+		rc = -ENOMEM;
+		goto fail1;
+	}
+
+	efx->workqueue = create_singlethread_workqueue("sfc_work");
+	if (!efx->workqueue) {
+		rc = -ENOMEM;
+		goto fail2;
+	}
+
+#if !defined(EFX_USE_CANCEL_DELAYED_WORK_SYNC)
+	efx->reset_workqueue = create_singlethread_workqueue("sfc_reset");
+	if (!efx->reset_workqueue) {
+		rc = -ENOMEM;
+		goto fail3;
+	}
+#endif
+
+	return 0;
+
+#if !defined(EFX_USE_CANCEL_DELAYED_WORK_SYNC)
+ fail3:
+	destroy_workqueue(efx->workqueue);
+	efx->workqueue = NULL;
+#endif
+
+ fail2:
+	destroy_workqueue(efx->refill_workqueue);
+	efx->refill_workqueue = NULL;
+ fail1:
+	return rc;
+}
+
+static void efx_fini_struct(struct efx_nic *efx)
+{
+#if !defined(EFX_USE_CANCEL_DELAYED_WORK_SYNC)
+	if (efx->reset_workqueue) {
+		destroy_workqueue(efx->reset_workqueue);
+		efx->reset_workqueue = NULL;
+	}
+#endif
+	if (efx->workqueue) {
+		destroy_workqueue(efx->workqueue);
+		efx->workqueue = NULL;
+	}
+	if (efx->refill_workqueue) {
+		destroy_workqueue(efx->refill_workqueue);
+		efx->refill_workqueue = NULL;
+	}
+}
+
+/**************************************************************************
+ *
+ * PCI interface
+ *
+ **************************************************************************/
+
+/* Main body of final NIC shutdown code
+ * This is called only at module unload (or hotplug removal).
+ */
+static void efx_pci_remove_main(struct efx_nic *efx)
+{
+	ASSERT_RTNL();
+
+	/* Skip everything if we never obtained a valid membase */
+	if (!efx->membase)
+		return;
+
+	efx_fini_channels(efx);
+	efx_fini_port(efx);
+
+	/* Shutdown the board, then the NIC and board state */
+	efx->board_info.fini(efx);
+	falcon_fini_nic(efx);
+	falcon_fini_interrupt(efx);
+	efx->board_info.fini(efx);
+
+	/* Tear down NAPI and LRO */
+	efx_fini_napi(efx);
+	efx_remove_all(efx);
+}
+
+/* Final NIC shutdown
+ * This is called only at module unload (or hotplug removal).
+ */
+static void efx_pci_remove(struct pci_dev *pci_dev)
+{
+	struct efx_nic *efx;
+
+	efx = pci_get_drvdata(pci_dev);
+	if (!efx)
+		return;
+
+	/* Unregister driver from driverlink layer */
+	efx_dl_unregister_nic(efx);
+
+	/* Mark the NIC as fini under both suspend_lock and
+	 * rtnl_lock */
+	down(&efx->suspend_lock);
+	rtnl_lock();
+	efx->state = STATE_FINI;
+	up(&efx->suspend_lock);
+
+	if (efx->membase) {
+		/* Stop the NIC. Since we're in STATE_FINI, this
+		 * won't be reversed. */
+		if (efx->net_dev_registered)
+			dev_close(efx->net_dev);
+
+		/* Release the rtnl lock. Any queued efx_resets()
+		 * can now return early [we're in STATE_FINI]. */
+		rtnl_unlock();
+
+		efx_unregister_netdev(efx);
+		efx_fini_debugfs_channels(efx);
+
+		/* Wait for any scheduled resets to complete. No more will be
+		 * scheduled from this point because efx_stop_all() has been
+		 * called, we are no longer registered with driverlink, and
+		 * the net_device's have been removed. */
+#if !defined(EFX_USE_CANCEL_DELAYED_WORK_SYNC)
+		flush_workqueue(efx->reset_workqueue);
+#else
+		flush_workqueue(efx->workqueue);
+#endif
+
+		/* Fini and remove all the software state */
+		rtnl_lock();
+		efx_pci_remove_main(efx);
+	}
+
+	rtnl_unlock();
+
+	efx_fini_io(efx);
+	EFX_LOG(efx, "shutdown successful\n");
+
+	pci_set_drvdata(pci_dev, NULL);
+	efx_fini_struct(efx);
+	kfree(efx);
+};
+
+/* Main body of NIC initialisation
+ * This is called at module load (or hotplug insertion, theoretically).
+ */
+static int efx_pci_probe_main(struct efx_nic *efx)
+{
+	int rc;
+
+	/* Do start-of-day initialisation */
+	rc = efx_probe_all(efx);
+	if (rc)
+		goto fail1;
+
+	/* Initialise port/channel net_dev's  */
+	rc = efx_init_napi(efx);
+	if (rc)
+		goto fail2;
+
+	/* Initialise the board */
+	rc = efx->board_info.init(efx);
+	if (rc) {
+		EFX_ERR(efx, "failed to initialise board\n");
+		goto fail3;
+	}
+
+	/* Initialise device */
+	rc = falcon_init_nic(efx);
+	if (rc) {
+		EFX_ERR(efx, "failed to initialise NIC\n");
+		goto fail4;
+	}
+
+	/* Initialise port */
+	rc = efx_init_port(efx);
+	if (rc) {
+		EFX_ERR(efx, "failed to initialise port\n");
+		goto fail5;
+	}
+
+	/* Initialise channels */
+	rc = efx_init_channels(efx);
+	if (rc)
+		goto fail6;
+
+	rc = falcon_init_interrupt(efx);
+	if (rc)
+		goto fail7;
+
+	/* Start up device - interrupts can occur from this point */
+	efx_start_all(efx);
+
+	/* Check basic functionality and set interrupt mode */
+	rc = efx_run_selftests(efx);
+	if (rc)
+		goto fail8;
+
+	/* Stop the NIC */
+	efx_stop_all(efx);
+
+	return 0;
+
+ fail8:
+	efx_stop_all(efx);
+	falcon_fini_interrupt(efx);
+ fail7:
+	efx_fini_channels(efx);
+ fail6:
+	efx_fini_port(efx);
+ fail5:
+	falcon_fini_nic(efx);
+ fail4:
+	efx->board_info.fini(efx);
+ fail3:
+	efx_fini_napi(efx);
+ fail2:
+	efx_remove_all(efx);
+ fail1:
+	return rc;
+}
+
+/* NIC initialisation
+ *
+ * This is called at module load (or hotplug insertion,
+ * theoretically).  It sets up PCI mappings, tests and resets the NIC,
+ * sets up and registers the network devices with the kernel and hooks
+ * the interrupt service routine.  It does not prepare the device for
+ * transmission; this is left to the first time one of the network
+ * interfaces is brought up (i.e. efx_net_open).
+ */
+static int __devinit efx_pci_probe(struct pci_dev *pci_dev,
+				   const struct pci_device_id *entry)
+{
+	struct efx_nic *efx;
+	enum efx_type_index type = entry->driver_data;
+	int i, rc;
+
+	/* Allocate and initialise a struct efx_nic */
+	efx = kmalloc(sizeof(*efx), GFP_KERNEL);
+	if (!efx) {
+		rc = -ENOMEM;
+		goto fail1;
+	}
+	pci_set_drvdata(pci_dev, efx);
+	rc = efx_init_struct(efx, type, pci_dev);
+	if (rc)
+		goto fail2;
+
+	EFX_INFO(efx, "Solarflare Communications NIC detected\n");
+
+	/* Set up basic I/O (BAR mappings etc) */
+	rc = efx_init_io(efx);
+	if (rc)
+		goto fail3;
+
+	/* From this point on we begin to expose the driver to the OS
+	 * to varying degrees, so lets grab the suspend_lock and
+	 * rtnl_lock to serialise against efx_reset() and
+	 * friends. efx->state is not STATE_RUNNING yet, but we don't
+	 * want these tasks to fail, just to block until we drop the
+	 * lock
+	 */
+	rc = down_interruptible(&efx->suspend_lock);
+	if (rc) {
+		EFX_ERR(efx, "suspend interrupted - aborting\n");
+		goto fail4;
+	}
+
+	rtnl_lock();
+
+	/* Probe, initialise and start everything. Run self-test */
+	for (i = 0; i < 5; i++) {
+		rc = efx_pci_probe_main(efx);
+		if (rc == 0)
+			break;
+
+		/* Retry if a recoverably reset event has been scheduled */
+		if ((efx->reset_pending != RESET_TYPE_INVISIBLE) &&
+		    (efx->reset_pending != RESET_TYPE_ALL))
+			goto fail5;
+
+		/* Serialise against efx_reset(). No more resets will be
+		 * scheduled since efx_stop_all() has been called, and we
+		 * have not and never have been registered with either
+		 * the rtnetlink or driverlink layers. */
+		rtnl_unlock();
+		up(&efx->suspend_lock);
+
+#if defined(EFX_USE_CANCEL_WORK_SYNC)
+		cancel_work_sync(&efx->reset_work);
+#else
+		flush_workqueue(efx->reset_workqueue);
+#endif
+
+		down(&efx->suspend_lock);
+		rtnl_lock();
+
+		efx->reset_pending = RESET_TYPE_NONE;
+	};
+	if (rc) {
+		EFX_ERR(efx, "Could not reset NIC\n");
+		goto fail5;
+	}
+
+	/* Self-tests have all passed */
+	rc = efx_init_debugfs_channels(efx);
+	if (rc)
+		goto fail6;
+
+	/* Switch to the running state before we expose the device to
+	 * the OS.  This is to ensure that the initial gathering of
+	 * MAC stats succeeds. */
+	efx->state = STATE_RUNNING;
+
+	rtnl_unlock();
+
+	rc = efx_register_netdev(efx);
+	if (rc)
+		goto fail7;
+
+	up(&efx->suspend_lock);
+
+	EFX_LOG(efx, "initialisation successful\n");
+
+	/* Register with driverlink layer */
+	rc = efx_dl_register_nic(efx);
+	if (rc)
+		goto fail8;
+
+	return 0;
+
+ fail8:
+	down(&efx->suspend_lock);
+	efx_unregister_netdev(efx);
+ fail7:
+	/* Re-acquire the rtnl lock around pci_remove_main() */
+	rtnl_lock();
+	efx_fini_debugfs_channels(efx);
+ fail6:
+	efx_pci_remove_main(efx);
+ fail5:
+	/* Drop the locks before fini */
+	rtnl_unlock();
+	up(&efx->suspend_lock);
+ fail4:
+	efx_fini_io(efx);
+ fail3:
+	efx_fini_struct(efx);
+ fail2:
+	kfree(efx);
+ fail1:
+	EFX_LOG(efx, "initialisation failed. rc=%d\n", rc);
+	return rc;
+}
+
+/* PCI driver definition */
+static struct pci_driver efx_pci_driver = {
+	.name		= EFX_DRIVER_NAME,
+	.id_table	= efx_pci_table,
+	.probe		= efx_pci_probe,
+	.remove		= efx_pci_remove,
+};
+
+/**************************************************************************
+ *
+ * Kernel module interface
+ *
+ *************************************************************************/
+
+module_param(interrupt_mode, uint, 0444);
+MODULE_PARM_DESC(interrupt_mode,
+		 "Interrupt mode (0=>MSIX 1=>MSI 2=>legacy)");
+
+module_param(onload_offline_selftest, uint, 0444);
+MODULE_PARM_DESC(onload_offline_selftest, "Perform offline selftest on load");
+
+static int __init efx_init_module(void)
+{
+	int rc;
+
+	printk(KERN_INFO "Solarflare NET driver v" EFX_DRIVER_VERSION "\n");
+
+	rc = efx_init_debugfs();
+	if (rc)
+		goto err_debugfs;
+
+	rc = register_netdevice_notifier(&efx_netdev_notifier);
+	if (rc)
+		goto err_notifier;
+
+	rc = pci_register_driver(&efx_pci_driver);
+	if (rc < 0)
+		goto err_pci;
+
+	return 0;
+
+ err_pci:
+	unregister_netdevice_notifier(&efx_netdev_notifier);
+ err_notifier:
+	efx_fini_debugfs();
+ err_debugfs:
+	return rc;
+}
+
+static void __exit efx_exit_module(void)
+{
+	printk(KERN_INFO "Solarflare NET driver unloading\n");
+
+	pci_unregister_driver(&efx_pci_driver);
+	unregister_netdevice_notifier(&efx_netdev_notifier);
+	efx_fini_debugfs();
+
+}
+
+module_init(efx_init_module);
+module_exit(efx_exit_module);
+
+MODULE_AUTHOR("Michael Brown <mbrown@fensystems.co.uk> and "
+	      "Solarflare Communications");
+MODULE_DESCRIPTION("Solarflare Communications network driver");
+MODULE_LICENSE("GPL");
+MODULE_DEVICE_TABLE(pci, efx_pci_table);
diff -rpuN linux-2.6.18.8/drivers/net/sfc/efx.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/efx.h
--- linux-2.6.18.8/drivers/net/sfc/efx.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/efx.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,103 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_EFX_H
+#define EFX_EFX_H
+
+#include "net_driver.h"
+
+/* PCI IDs */
+#define EFX_VENDID_SFC	        0x1924
+#define FALCON_A_P_DEVID	0x0703
+#define FALCON_A_S_DEVID        0x6703
+#define FALCON_B_P_DEVID        0x0710
+
+/* TX */
+extern int efx_xmit(struct efx_nic *efx,
+		    struct efx_tx_queue *tx_queue, struct sk_buff *skb);
+extern void efx_stop_queue(struct efx_nic *efx);
+extern void efx_wake_queue(struct efx_nic *efx);
+
+/* RX */
+#if defined(EFX_USE_FASTCALL)
+extern void fastcall efx_xmit_done(struct efx_tx_queue *tx_queue,
+				   unsigned int index);
+#else
+extern void efx_xmit_done(struct efx_tx_queue *tx_queue, unsigned int index);
+#endif
+#if defined(EFX_USE_FASTCALL)
+extern void fastcall efx_rx_packet(struct efx_rx_queue *rx_queue,
+				   unsigned int index, unsigned int len,
+				   int checksummed, int discard);
+#else
+extern void efx_rx_packet(struct efx_rx_queue *rx_queue, unsigned int index,
+			  unsigned int len, int checksummed, int discard);
+#endif
+extern void efx_fini_rx_buffer(struct efx_rx_queue *rx_queue,
+			       struct efx_rx_buffer *rx_buf);
+
+/* Channels */
+extern void efx_process_channel_now(struct efx_channel *channel);
+extern int efx_flush_queues(struct efx_nic *efx);
+
+/* Ports */
+extern void efx_reconfigure_port(struct efx_nic *efx,
+				 int on_disabled);
+
+/* Global */
+extern void efx_schedule_reset(struct efx_nic *efx, enum reset_type type);
+extern void efx_suspend(struct efx_nic *efx);
+extern void efx_resume(struct efx_nic *efx);
+extern void efx_init_irq_moderation(struct efx_nic *efx, int tx_usecs,
+				    int rx_usecs);
+extern int efx_request_power(struct efx_nic *efx, int mw, const char *name);
+extern void efx_hex_dump(const u8 *, unsigned int, const char *);
+
+/* Dummy PHY ops for PHY drivers */
+extern int efx_port_dummy_op_int(struct efx_nic *efx);
+extern void efx_port_dummy_op_void(struct efx_nic *efx);
+extern void efx_port_dummy_op_blink(struct efx_nic *efx, int blink);
+
+
+extern unsigned int efx_monitor_interval;
+
+static inline void efx_schedule_channel(struct efx_channel *channel)
+{
+	EFX_TRACE(channel->efx, "channel %d scheduling NAPI poll on CPU%d\n",
+		  channel->channel, raw_smp_processor_id());
+	channel->work_pending = 1;
+
+#if defined(EFX_HAVE_OLD_NAPI)
+	if (!test_and_set_bit(__LINK_STATE_RX_SCHED, &channel->napi_dev->state))
+		__netif_rx_schedule(channel->napi_dev);
+#else
+	netif_rx_schedule(channel->napi_dev, &channel->napi_str);
+#endif
+}
+
+
+#endif /* EFX_EFX_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/enum.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/enum.h
--- linux-2.6.18.8/drivers/net/sfc/enum.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/enum.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,117 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2007:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_ENUM_H
+#define EFX_ENUM_H
+
+/**
+ * enum efx_loopback_mode - loopback modes
+ * @LOOPBACK_NONE: no loopback
+ * @LOOPBACK_NEAR: loopback nearest to bus
+ * @LOOPBACK_MAC: loopback within MAC unspecified level
+ * @LOOPBACK_XGMII: loopback within MAC at XGMII level
+ * @LOOPBACK_XGXS: loopback within MAC at XGXS level
+ * @LOOPBACK_XAUI: loopback within MAC at XAUI level
+ * @LOOPBACK_PHY: loopback within PHY unspecified level
+ * @LOOPBACK_PHYXS: loopback within PHY at PHYXS level
+ * @LOOPBACK_PCS: loopback within PHY at PCS level
+ * @LOOPBACK_PMAPMD: loopback within PHY at PMAPMD level
+ * @LOOPBACK_FAR: loopback furthest from bus
+ * @LOOPBACK_NETWORK: reflecting loopback (even further than furthest!)
+ */
+/* Please keep in order and up-to-date w.r.t the following two #defines */
+enum efx_loopback_mode {
+	LOOPBACK_NONE = 0,
+	LOOPBACK_NEAR = 1,
+	LOOPBACK_MAC = 2,
+	LOOPBACK_XGMII = 3,
+	LOOPBACK_XGXS = 4,
+	LOOPBACK_XAUI = 5,
+	LOOPBACK_PHY = 6,
+	LOOPBACK_PHYXS = 7,
+	LOOPBACK_PCS = 8,
+	LOOPBACK_PMAPMD = 9,
+	LOOPBACK_FAR = 10,
+	LOOPBACK_NETWORK = 11,
+	LOOPBACK_MAX
+};
+#define LOOPBACK_TEST_MAX LOOPBACK_FAR
+
+/* These loopbacks occur within the controller */
+#define LOOPBACKS_10G_INTERNAL ((1 << LOOPBACK_XGMII)| \
+				(1 << LOOPBACK_XGXS) | \
+				(1 << LOOPBACK_XAUI))
+
+#define LOOPBACKS_1G_INTERNAL (1 << LOOPBACK_MAC)
+
+#define LOOPBACK_MASK(_efx)			\
+	(1 << (_efx)->loopback_mode)
+
+#define LOOPBACK_INTERNAL(_efx)					\
+	(((LOOPBACKS_10G_INTERNAL | LOOPBACKS_1G_INTERNAL) &	\
+	  LOOPBACK_MASK(_efx)) ? 1 : 0)
+
+#define LOOPBACK_CHANGED(_from, _to, _mask)		\
+	((LOOPBACK_MASK(_from) ^ LOOPBACK_MASK(_to)) &	\
+	 (_mask) ? 1 : 0)
+
+#define LOOPBACK_OUT_OF(_from, _to, _mask)		\
+	(((LOOPBACK_MASK(_from) & (_mask)) &&		\
+	  ((LOOPBACK_MASK(_to) & (_mask)) == 0)) ? 1 : 0)
+
+/*****************************************************************************/
+
+/**
+ * enum reset_type - reset types
+ *
+ * %RESET_TYPE_INVSIBLE, %RESET_TYPE_ALL, %RESET_TYPE_WORLD and
+ * %RESET_TYPE_DISABLE specify the method/scope of the reset.  The
+ * other valuesspecify reasons, which efx_schedule_reset() will choose
+ * a method for.
+ *
+ * @RESET_TYPE_INVISIBLE: don't reset the PHYs or interrupts
+ * @RESET_TYPE_ALL: reset everything but PCI core blocks
+ * @RESET_TYPE_WORLD: reset everything, save & restore PCI config
+ * @RESET_TYPE_DISABLE: disable NIC
+ * @RESET_TYPE_MONITOR: reset due to hardware monitor
+ * @RESET_TYPE_INT_ERROR: reset due to internal error
+ * @RESET_TYPE_RX_RECOVERY: reset to recover from RX datapath errors
+ */
+enum reset_type {
+	RESET_TYPE_NONE = -1,
+	RESET_TYPE_INVISIBLE = 0,
+	RESET_TYPE_ALL = 1,
+	RESET_TYPE_WORLD = 2,
+	RESET_TYPE_DISABLE = 3,
+	RESET_TYPE_MAX_METHOD,
+	RESET_TYPE_MONITOR,
+	RESET_TYPE_INT_ERROR,
+	RESET_TYPE_RX_RECOVERY,
+	RESET_TYPE_RX_DESC_FETCH,
+	RESET_TYPE_TX_DESC_FETCH,
+	RESET_TYPE_MAX,
+};
+
+#endif /* EFX_ENUM_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/ethtool.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/ethtool.c
--- linux-2.6.18.8/drivers/net/sfc/ethtool.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/ethtool.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,734 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/netdevice.h>
+#include <linux/ethtool.h>
+#include <linux/rtnetlink.h>
+#include <asm/uaccess.h>
+#include "net_driver.h"
+#include "selftest.h"
+#include "efx.h"
+#include "ethtool.h"
+#include "falcon.h"
+#include "gmii.h"
+
+static int efx_ethtool_set_tx_csum(struct net_device *net_dev, u32 enable);
+
+struct ethtool_string {
+	char name[ETH_GSTRING_LEN];
+};
+
+struct efx_ethtool_stat {
+	const char *name;
+	enum {
+		EFX_ETHTOOL_STAT_SOURCE_mac_stats,
+		EFX_ETHTOOL_STAT_SOURCE_nic,
+		EFX_ETHTOOL_STAT_SOURCE_channel
+	} source;
+	unsigned offset;
+	u64(*get_stat) (void *field); /* Reader function */
+};
+
+/* Initialiser for a struct #efx_ethtool_stat with type-checking */
+#define EFX_ETHTOOL_STAT(stat_name, source_name, field, field_type, \
+				get_stat_function) {			\
+	.name = #stat_name,						\
+	.source = EFX_ETHTOOL_STAT_SOURCE_##source_name,		\
+	.offset = ((((field_type *) 0) ==				\
+		      &((struct efx_##source_name *)0)->field) ?	\
+		    offsetof(struct efx_##source_name, field) :		\
+		    offsetof(struct efx_##source_name, field)),		\
+	.get_stat = get_stat_function,					\
+}
+
+static u64 efx_get_uint_stat(void *field)
+{
+	return *(unsigned int *)field;
+}
+
+static u64 efx_get_ulong_stat(void *field)
+{
+	return *(unsigned long *)field;
+}
+
+static u64 efx_get_u64_stat(void *field)
+{
+	return *(u64 *) field;
+}
+
+static u64 efx_get_atomic_stat(void *field)
+{
+	return atomic_read((atomic_t *) field);
+}
+
+#define EFX_ETHTOOL_ULONG_MAC_STAT(field)			\
+	EFX_ETHTOOL_STAT(field, mac_stats, field, 		\
+			  unsigned long, efx_get_ulong_stat)
+
+#define EFX_ETHTOOL_U64_MAC_STAT(field)				\
+	EFX_ETHTOOL_STAT(field, mac_stats, field, 		\
+			  u64, efx_get_u64_stat)
+
+#define EFX_ETHTOOL_UINT_NIC_STAT(name)				\
+	EFX_ETHTOOL_STAT(name, nic, n_##name,			\
+			 unsigned int, efx_get_uint_stat)
+
+#define EFX_ETHTOOL_ATOMIC_NIC_ERROR_STAT(field)		\
+	EFX_ETHTOOL_STAT(field, nic, errors.field,		\
+			 atomic_t, efx_get_atomic_stat)
+
+#define EFX_ETHTOOL_UINT_CHANNEL_STAT(field)			\
+	EFX_ETHTOOL_STAT(field, channel, n_##field,		\
+			 unsigned int, efx_get_uint_stat)
+
+static struct efx_ethtool_stat efx_ethtool_stats[] = {
+	EFX_ETHTOOL_U64_MAC_STAT(tx_bytes),
+	EFX_ETHTOOL_U64_MAC_STAT(tx_good_bytes),
+	EFX_ETHTOOL_U64_MAC_STAT(tx_bad_bytes),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_packets),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_bad),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_pause),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_control),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_unicast),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_multicast),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_broadcast),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_lt64),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_64),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_65_to_127),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_128_to_255),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_256_to_511),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_512_to_1023),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_1024_to_15xx),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_15xx_to_jumbo),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_gtjumbo),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_collision),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_single_collision),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_multiple_collision),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_excessive_collision),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_deferred),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_late_collision),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_excessive_deferred),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_non_tcpudp),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_mac_src_error),
+	EFX_ETHTOOL_ULONG_MAC_STAT(tx_ip_src_error),
+	EFX_ETHTOOL_U64_MAC_STAT(rx_bytes),
+	EFX_ETHTOOL_U64_MAC_STAT(rx_good_bytes),
+	EFX_ETHTOOL_U64_MAC_STAT(rx_bad_bytes),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_packets),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_good),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_bad),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_pause),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_control),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_unicast),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_multicast),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_broadcast),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_lt64),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_64),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_65_to_127),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_128_to_255),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_256_to_511),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_512_to_1023),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_1024_to_15xx),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_15xx_to_jumbo),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_gtjumbo),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_bad_lt64),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_bad_64_to_15xx),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_bad_15xx_to_jumbo),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_bad_gtjumbo),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_overflow),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_missed),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_false_carrier),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_symbol_error),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_align_error),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_length_error),
+	EFX_ETHTOOL_ULONG_MAC_STAT(rx_internal_error),
+	EFX_ETHTOOL_UINT_NIC_STAT(rx_nodesc_drop_cnt),
+	EFX_ETHTOOL_ATOMIC_NIC_ERROR_STAT(rx_reset),
+	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_tobe_disc),
+	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_ip_hdr_chksum_err),
+	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_tcp_udp_chksum_err),
+	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_frm_trunc),
+};
+
+/* Number of ethtool statistics */
+#define EFX_ETHTOOL_NUM_STATS ARRAY_SIZE(efx_ethtool_stats)
+
+/**************************************************************************
+ *
+ * Ethtool operations
+ *
+ **************************************************************************
+ */
+
+/* Identify device by flashing LEDs */
+static int efx_ethtool_phys_id(struct net_device *net_dev, u32 seconds)
+{
+	struct efx_nic *efx = net_dev->priv;
+
+	efx->board_info.blink(efx, 1);
+	schedule_timeout_interruptible(seconds * HZ);
+	efx->board_info.blink(efx, 0);
+	return 0;
+}
+
+/* This must be called with rtnl_lock held. */
+int efx_ethtool_get_settings(struct net_device *net_dev,
+			     struct ethtool_cmd *ecmd)
+{
+	struct efx_nic *efx = net_dev->priv;
+
+	return efx->mac_op->get_settings(efx, ecmd);
+}
+
+/* This must be called with rtnl_lock held. */
+int efx_ethtool_set_settings(struct net_device *net_dev,
+			     struct ethtool_cmd *ecmd)
+{
+	struct efx_nic *efx = net_dev->priv;
+	int rc;
+
+	rc = efx->mac_op->set_settings(efx, ecmd);
+	if (rc)
+		return rc;
+
+	/* Push the settings to the MAC */
+	efx_reconfigure_port(efx, 0);
+
+	return 0;
+}
+
+static void efx_ethtool_get_drvinfo(struct net_device *net_dev,
+				    struct ethtool_drvinfo *info)
+{
+	struct efx_nic *efx = net_dev->priv;
+
+	strlcpy(info->driver, EFX_DRIVER_NAME, sizeof(info->driver));
+	strlcpy(info->version, EFX_DRIVER_VERSION, sizeof(info->version));
+	strlcpy(info->bus_info, pci_name(efx->pci_dev), sizeof(info->bus_info));
+}
+
+/**
+ * efx_fill_test - fill in an individual self-test entry
+ * @test_index:		Index of the test
+ * @strings:		Ethtool strings, or %NULL
+ * @data:		Ethtool test results, or %NULL
+ * @test:		Pointer to test result (used only if data != %NULL)
+ * @unit_format:	Unit name format (e.g. "channel\%d")
+ * @unit_id:		Unit id (e.g. 0 for "channel0")
+ * @test_format:	Test name format (e.g. "loopback.\%s.tx.sent")
+ * @test_id:		Test id (e.g. "PHY" for "loopback.PHY.tx_sent")
+ *
+ * Fill in an individual self-test entry.
+ */
+static void efx_fill_test(unsigned int test_index,
+			  struct ethtool_string *strings, u64 *data,
+			  int *test, const char *unit_format, int unit_id,
+			  const char *test_format, const char *test_id)
+{
+	struct ethtool_string unit_str, test_str;
+
+	/* Fill data value, if applicable */
+	if (data)
+		data[test_index] = *test;
+
+	/* Fill string, if applicable */
+	if (strings) {
+		snprintf(unit_str.name, sizeof(unit_str.name),
+			 unit_format, unit_id);
+		snprintf(test_str.name, sizeof(test_str.name),
+			 test_format, test_id);
+		snprintf(strings[test_index].name,
+			 sizeof(strings[test_index].name),
+			 "%-9s%-17s", unit_str.name, test_str.name);
+	}
+}
+
+#define EFX_PORT_NAME "port%d", 0
+
+/**
+ * efx_fill_loopback_test - fill in a block of loopback self-test entries
+ * @efx:		Efx NIC
+ * @lb_tests:		Efx loopback self-test results structure
+ * @mode:		Loopback test mode
+ * @test_index:		Starting index of the test
+ * @strings:		Ethtool strings, or %NULL
+ * @data:		Ethtool test results, or %NULL
+ *
+ * Fill in a block of loopback self-test entries.  Return new test
+ * index.
+ */
+static int efx_fill_loopback_test(struct efx_nic *efx,
+				  struct efx_loopback_self_tests *lb_tests,
+				  enum efx_loopback_mode mode,
+				  unsigned int test_index,
+				  struct ethtool_string *strings, u64 *data)
+{
+	struct efx_tx_queue *tx_queue;
+
+	efx_for_each_tx_queue(tx_queue, efx) {
+		efx_fill_test(test_index++, strings, data,
+			      &lb_tests->tx_sent[tx_queue->queue],
+			      EFX_TX_QUEUE_NAME(tx_queue),
+			      "loopback.%s.tx_sent",
+			      efx_loopback_mode_names[mode]);
+		efx_fill_test(test_index++, strings, data,
+			      &lb_tests->tx_done[tx_queue->queue],
+			      EFX_TX_QUEUE_NAME(tx_queue),
+			      "loopback.%s.tx_done",
+			      efx_loopback_mode_names[mode]);
+	}
+	efx_fill_test(test_index++, strings, data,
+		      &lb_tests->rx_good,
+		      EFX_PORT_NAME,
+		      "loopback.%s.rx_good",
+		      efx_loopback_mode_names[mode]);
+	efx_fill_test(test_index++, strings, data,
+		      &lb_tests->rx_bad,
+		      EFX_PORT_NAME,
+		      "loopback.%s.rx_bad",
+		      efx_loopback_mode_names[mode]);
+
+	return test_index;
+}
+
+/**
+ * efx_ethtool_fill_self_tests - get self-test details
+ * @efx:		Efx NIC
+ * @tests:		Efx self-test results structure, or %NULL
+ * @strings:		Ethtool strings, or %NULL
+ * @data:		Ethtool test results, or %NULL
+ *
+ * Get self-test number of strings, strings, and/or test results.
+ * Return number of strings (== number of test results).
+ *
+ * The reason for merging these three functions is to make sure that
+ * they can never be inconsistent.
+ */
+static int efx_ethtool_fill_self_tests(struct efx_nic *efx,
+				       struct efx_self_tests *tests,
+				       struct ethtool_string *strings,
+				       u64 *data)
+{
+	struct efx_channel *channel;
+	unsigned int n = 0;
+	enum efx_loopback_mode mode;
+
+	/* Interrupt */
+	efx_fill_test(n++, strings, data, &tests->interrupt,
+		      "core", 0, "interrupt", NULL);
+
+	/* Event queues */
+	efx_for_each_channel(channel, efx) {
+		efx_fill_test(n++, strings, data,
+			      &tests->eventq_dma[channel->channel],
+			      EFX_CHANNEL_NAME(channel),
+			      "eventq.dma", NULL);
+		efx_fill_test(n++, strings, data,
+			      &tests->eventq_int[channel->channel],
+			      EFX_CHANNEL_NAME(channel),
+			      "eventq.int", NULL);
+		efx_fill_test(n++, strings, data,
+			      &tests->eventq_poll[channel->channel],
+			      EFX_CHANNEL_NAME(channel),
+			      "eventq.poll", NULL);
+	}
+
+	/* PHY presence */
+	efx_fill_test(n++, strings, data, &tests->phy_ok,
+		      EFX_PORT_NAME, "phy_ok", NULL);
+
+	/* Loopback tests */
+	efx_fill_test(n++, strings, data, &tests->loopback_speed,
+		      EFX_PORT_NAME, "loopback.speed", NULL);
+	efx_fill_test(n++, strings, data, &tests->loopback_full_duplex,
+		      EFX_PORT_NAME, "loopback.full_duplex", NULL);
+	for (mode = LOOPBACK_NONE; mode < LOOPBACK_TEST_MAX; mode++) {
+		if (!(efx->loopback_modes & (1 << mode)))
+			continue;
+		n = efx_fill_loopback_test(efx,
+					   &tests->loopback[mode], mode, n,
+					   strings, data);
+	}
+
+	return n;
+}
+
+static int efx_ethtool_get_stats_count(struct net_device *net_dev)
+{
+	return EFX_ETHTOOL_NUM_STATS;
+}
+
+static int efx_ethtool_self_test_count(struct net_device *net_dev)
+{
+	struct efx_nic *efx = net_dev->priv;
+
+	return efx_ethtool_fill_self_tests(efx, NULL, NULL, NULL);
+}
+
+static void efx_ethtool_get_strings(struct net_device *net_dev,
+				    u32 string_set, u8 *strings)
+{
+	struct efx_nic *efx = net_dev->priv;
+	struct ethtool_string *ethtool_strings =
+		(struct ethtool_string *)strings;
+	int i;
+
+	switch (string_set) {
+	case ETH_SS_STATS:
+		for (i = 0; i < EFX_ETHTOOL_NUM_STATS; i++)
+			strncpy(ethtool_strings[i].name,
+				efx_ethtool_stats[i].name,
+				sizeof(ethtool_strings[i].name));
+		break;
+	case ETH_SS_TEST:
+		efx_ethtool_fill_self_tests(efx, NULL,
+					    ethtool_strings, NULL);
+		break;
+	default:
+		/* No other string sets */
+		break;
+	}
+}
+
+static void efx_ethtool_get_stats(struct net_device *net_dev,
+				  struct ethtool_stats *stats
+				  __attribute__ ((unused)), u64 *data)
+{
+	unsigned long flags __attribute__ ((unused));
+	struct efx_nic *efx = net_dev->priv;
+	struct efx_mac_stats *mac_stats = &efx->mac_stats;
+	struct efx_ethtool_stat *stat;
+	struct efx_channel *channel;
+	int i;
+
+	EFX_BUG_ON_PARANOID(stats->n_stats != EFX_ETHTOOL_NUM_STATS);
+
+	/* Update MAC and NIC statistics */
+	net_dev->get_stats(net_dev);
+	falcon_update_nic_stats(efx);
+
+	/* Fill detailed statistics buffer */
+	for (i = 0; i < EFX_ETHTOOL_NUM_STATS; i++) {
+		stat = &efx_ethtool_stats[i];
+		switch (stat->source) {
+		case EFX_ETHTOOL_STAT_SOURCE_mac_stats:
+			data[i] = stat->get_stat((void *)mac_stats +
+						 stat->offset);
+			break;
+		case EFX_ETHTOOL_STAT_SOURCE_nic:
+			data[i] = stat->get_stat((void *)efx + stat->offset);
+			break;
+		case EFX_ETHTOOL_STAT_SOURCE_channel:
+			data[i] = 0;
+			efx_for_each_channel(channel, efx)
+				data[i] += stat->get_stat((void *)channel +
+							  stat->offset);
+			break;
+		}
+	}
+}
+
+static int efx_ethtool_set_tx_csum(struct net_device *net_dev, u32 enable)
+{
+	struct efx_nic *efx = net_dev->priv;
+	int rc;
+
+	rc = ethtool_op_set_tx_csum(net_dev, enable);
+	if (rc)
+		return rc;
+
+
+	efx_flush_queues(efx);
+
+	return 0;
+}
+
+static int efx_ethtool_set_rx_csum(struct net_device *net_dev, u32 enable)
+{
+	struct efx_nic *efx = net_dev->priv;
+
+	/* No way to stop the hardware doing the checks; we just
+	 * ignore the result.
+	 */
+	efx->rx_checksum_enabled = (enable ? 1 : 0);
+
+	return 0;
+}
+
+static u32 efx_ethtool_get_rx_csum(struct net_device *net_dev)
+{
+	struct efx_nic *efx = net_dev->priv;
+
+	return efx->rx_checksum_enabled;
+}
+
+static void efx_ethtool_self_test(struct net_device *net_dev,
+				  struct ethtool_test *test, u64 *data)
+{
+	struct efx_nic *efx = net_dev->priv;
+	struct efx_self_tests efx_tests;
+	int offline, already_up;
+	int rc;
+
+	/* Make sure we've got rtnl lock since we're playing with
+	 * interrupts, and calling efx_process_channel_now and others
+	 */
+	ASSERT_RTNL();
+
+	/* If the NIC isn't in the RUNNING state then exit */
+	if (efx->state != STATE_RUNNING) {
+		rc = -EIO;
+		goto fail1;
+	}
+
+	/* Make sure the interface is up. We need interrupts, NAPI
+	 * and some RX buffers so this is helpful.  NB. The caller has
+	 * rtnl_lock so nobody else can call dev_open. */
+	already_up = (efx->net_dev->flags & IFF_UP);
+	if (!already_up) {
+		rc = dev_open(efx->net_dev);
+		if (rc) {
+			EFX_ERR(efx, "failed opening device.\n");
+			goto fail2;
+		}
+	}
+
+	memset(&efx_tests, 0, sizeof(efx_tests));
+	offline = (test->flags & ETH_TEST_FL_OFFLINE);
+
+	/* Perform online self tests first */
+	rc = efx_online_test(efx, &efx_tests);
+	if (rc)
+		goto out;
+
+	/* Perform offline tests only if online tests passed */
+	if (offline) {
+		/* Stop the kernel from sending packets during the test. The
+		 * selftest will be consistently bringing the port up and down
+		 * as it moves between loopback modes, so the watchdog timer
+		 * probably won't run anyway */
+		efx_stop_queue(efx);
+
+		rc = efx_flush_queues(efx);
+		if (rc != 0)
+			goto out_offline;
+
+		rc = efx_offline_test(efx, &efx_tests,
+				      efx->loopback_modes);
+ out_offline:
+		efx_wake_queue(efx);
+	}
+
+	/* fall-thru */
+ out:
+	if (!already_up)
+		dev_close(efx->net_dev);
+
+	EFX_LOG(efx, "%s all %sline self-tests\n",
+		rc == 0 ? "passed" : "failed", offline ? "off" : "on");
+
+ fail2:
+ fail1:
+	/* Fill ethtool results structures */
+	efx_ethtool_fill_self_tests(efx, &efx_tests, NULL, data);
+	if (rc)
+		test->flags |= ETH_TEST_FL_FAILED;
+}
+
+/* Restart autonegotiation */
+static int efx_ethtool_nway_reset(struct net_device *net_dev)
+{
+	struct efx_nic *efx = net_dev->priv;
+
+	return mii_nway_restart(&efx->mii);
+}
+
+static u32 efx_ethtool_get_link(struct net_device *net_dev)
+{
+	struct efx_nic *efx = net_dev->priv;
+
+	return efx->link_up;
+}
+
+static int efx_ethtool_get_coalesce(struct net_device *net_dev,
+				    struct ethtool_coalesce *coalesce)
+{
+	struct efx_nic *efx = net_dev->priv;
+	struct efx_tx_queue *tx_queue;
+	struct efx_rx_queue *rx_queue;
+	struct efx_channel *channel;
+
+	memset(coalesce, 0, sizeof(*coalesce));
+
+	/* Find lowest IRQ moderation across all used TX queues */
+	coalesce->tx_coalesce_usecs_irq = ~((u32) 0);
+	efx_for_each_tx_queue(tx_queue, efx) {
+		channel = tx_queue->channel;
+		if (channel->irq_moderation < coalesce->tx_coalesce_usecs_irq) {
+			if (channel->used_flags != EFX_USED_BY_RX_TX)
+				coalesce->tx_coalesce_usecs_irq =
+					channel->irq_moderation;
+			else
+				coalesce->tx_coalesce_usecs_irq = 0;
+		}
+	}
+
+	/* Find lowest IRQ moderation across all used RX queues */
+	coalesce->rx_coalesce_usecs_irq = ~((u32) 0);
+	efx_for_each_rx_queue(rx_queue, efx) {
+		channel = rx_queue->channel;
+		if (channel->irq_moderation < coalesce->rx_coalesce_usecs_irq)
+			coalesce->rx_coalesce_usecs_irq =
+				channel->irq_moderation;
+	}
+
+	return 0;
+}
+
+/* Set coalescing parameters
+ * The difficulties occur for shared channels
+ */
+static int efx_ethtool_set_coalesce(struct net_device *net_dev,
+				    struct ethtool_coalesce *coalesce)
+{
+	struct efx_nic *efx = net_dev->priv;
+	struct efx_channel *channel;
+	struct efx_tx_queue *tx_queue;
+	unsigned tx_usecs, rx_usecs;
+
+	if (coalesce->use_adaptive_rx_coalesce ||
+	    coalesce->use_adaptive_tx_coalesce)
+		return -EOPNOTSUPP;
+
+	if (coalesce->rx_coalesce_usecs || coalesce->tx_coalesce_usecs) {
+		EFX_ERR(efx, "invalid coalescing setting. "
+			"Only rx/tx_coalesce_usecs_irq are supported\n");
+		return -EOPNOTSUPP;
+	}
+
+	rx_usecs = coalesce->rx_coalesce_usecs_irq;
+	tx_usecs = coalesce->tx_coalesce_usecs_irq;
+
+	/* If the channel is shared only allow RX parameters to be set */
+	efx_for_each_tx_queue(tx_queue, efx) {
+		if ((tx_queue->channel->used_flags == EFX_USED_BY_RX_TX) &&
+		    tx_usecs) {
+			EFX_ERR(efx, "Channel is shared. "
+				"Only RX coalescing may be set\n");
+			return -EOPNOTSUPP;
+		}
+	}
+
+	efx_init_irq_moderation(efx, tx_usecs, rx_usecs);
+
+	/* Reset channel to pick up new moderation value.  Note that
+	 * this may change the value of the irq_moderation field
+	 * (e.g. to allow for hardware timer granularity).
+	 */
+	efx_for_each_channel(channel, efx)
+		falcon_set_int_moderation(channel);
+
+	return 0;
+}
+
+static int efx_ethtool_set_pauseparam(struct net_device *net_dev,
+				      struct ethtool_pauseparam *pause)
+{
+	struct efx_nic *efx = net_dev->priv;
+	enum efx_fc_type flow_control = efx->flow_control;
+	int rc;
+
+	flow_control &= ~(EFX_FC_RX | EFX_FC_TX | EFX_FC_AUTO);
+	flow_control |= pause->rx_pause ? EFX_FC_RX : 0;
+	flow_control |= pause->tx_pause ? EFX_FC_TX : 0;
+	flow_control |= pause->autoneg ? EFX_FC_AUTO : 0;
+
+	/* Try to push the pause parameters */
+	rc = efx->mac_op->set_pause(efx, flow_control);
+	if (rc)
+		return rc;
+
+	/* Push the settings to the MAC */
+	efx_reconfigure_port(efx, 0);
+
+	return 0;
+}
+
+static void efx_ethtool_get_pauseparam(struct net_device *net_dev,
+				       struct ethtool_pauseparam *pause)
+{
+	struct efx_nic *efx = net_dev->priv;
+
+	pause->rx_pause = (efx->flow_control & EFX_FC_RX) ? 1 : 0;
+	pause->tx_pause = (efx->flow_control & EFX_FC_TX) ? 1 : 0;
+	pause->autoneg = (efx->flow_control & EFX_FC_AUTO) ? 1 : 0;
+}
+
+
+#if defined(EFX_USE_ETHTOOL_GET_PERM_ADDR)
+static int efx_ethtool_op_get_perm_addr(struct net_device *net_dev,
+					struct ethtool_perm_addr *addr,
+					u8 *data)
+{
+	struct efx_nic *efx = net_dev->priv;
+
+	memcpy(data, efx->mac_address, ETH_ALEN);
+
+	return 0;
+}
+#endif
+
+struct ethtool_ops efx_ethtool_ops = {
+	.get_settings		= efx_ethtool_get_settings,
+	.set_settings		= efx_ethtool_set_settings,
+	.get_drvinfo		= efx_ethtool_get_drvinfo,
+	.nway_reset		= efx_ethtool_nway_reset,
+	.get_link		= efx_ethtool_get_link,
+	.get_coalesce		= efx_ethtool_get_coalesce,
+	.set_coalesce		= efx_ethtool_set_coalesce,
+	.get_pauseparam         = efx_ethtool_get_pauseparam,
+	.set_pauseparam         = efx_ethtool_set_pauseparam,
+	.get_rx_csum		= efx_ethtool_get_rx_csum,
+	.set_rx_csum		= efx_ethtool_set_rx_csum,
+	.get_tx_csum		= ethtool_op_get_tx_csum,
+	.set_tx_csum		= efx_ethtool_set_tx_csum,
+	.get_sg			= ethtool_op_get_sg,
+	.set_sg			= ethtool_op_set_sg,
+#if defined(EFX_USE_ETHTOOL_FLAGS)
+	.get_flags		= ethtool_op_get_flags,
+	.set_flags		= ethtool_op_set_flags,
+#endif
+	.self_test_count	= efx_ethtool_self_test_count,
+	.self_test		= efx_ethtool_self_test,
+	.get_strings		= efx_ethtool_get_strings,
+	.phys_id		= efx_ethtool_phys_id,
+	.get_stats_count	= efx_ethtool_get_stats_count,
+	.get_ethtool_stats	= efx_ethtool_get_stats,
+#if defined(EFX_USE_ETHTOOL_GET_PERM_ADDR)
+	.get_perm_addr          = efx_ethtool_op_get_perm_addr,
+#endif
+};
diff -rpuN linux-2.6.18.8/drivers/net/sfc/ethtool.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/ethtool.h
--- linux-2.6.18.8/drivers/net/sfc/ethtool.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/ethtool.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,44 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005:      Fen Systems Ltd.
+ * Copyright 2006:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_ETHTOOL_H
+#define EFX_ETHTOOL_H
+
+#include "net_driver.h"
+
+/*
+ * Ethtool support
+ */
+
+extern int efx_ethtool_get_settings(struct net_device *net_dev,
+				    struct ethtool_cmd *ecmd);
+extern int efx_ethtool_set_settings(struct net_device *net_dev,
+				    struct ethtool_cmd *ecmd);
+
+extern struct ethtool_ops efx_ethtool_ops;
+
+#endif /* EFX_ETHTOOL_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/extraversion.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/extraversion.h
--- linux-2.6.18.8/drivers/net/sfc/extraversion.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/extraversion.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,4 @@
+/*
+ * If compiling on kernels with backported features you may need to
+ * define EFX_DIST_KVER_ symbols here
+ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/falcon.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/falcon.c
--- linux-2.6.18.8/drivers/net/sfc/falcon.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/falcon.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,3708 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <asm/io.h>
+#include <asm/bitops.h>
+#include <linux/delay.h>
+#include <linux/pci.h>
+#include <linux/module.h>
+#include <linux/seq_file.h>
+#include "net_driver.h"
+#include "bitfield.h"
+#include "efx.h"
+#include "mac.h"
+#include "gmii.h"
+#include "spi.h"
+#include "falcon.h"
+#include "falcon_hwdefs.h"
+#include "falcon_io.h"
+#include "mdio_10g.h"
+#include "phy.h"
+#include "boards.h"
+#include "driverlink.h"
+#include "workarounds.h"
+
+/* Falcon hardware control.
+ * Falcon is the internal codename for the SFC4000 controller that is
+ * present in SFE400X evaluation boards
+ */
+
+struct falcon_nic_data {
+	/* Number of entries in each TX queue descriptor cache. */
+	unsigned tx_dc_entries;
+	/* Number of entries in each RX queue descriptor cache. */
+	unsigned rx_dc_entries;
+	/* Base address in SRAM of TX queue descriptor caches. */
+	unsigned tx_dc_base;
+	/* Base address in SRAM of RX queue descriptor caches. */
+	unsigned rx_dc_base;
+
+	/* Previous loopback mode used in deconfigure_mac_wrapper */
+	enum efx_loopback_mode old_loopback_mode;
+
+	/* Driverlink parameters */
+	struct efx_dl_falcon_resources resources;
+};
+
+/**************************************************************************
+ *
+ * Configurable values
+ *
+ **************************************************************************
+ */
+
+static int disable_dma_stats;
+
+/* Specify the size of the RX descriptor cache */
+static int descriptor_cache_size = 64;
+
+/*
+ * Override EEPROM/flash type from non-volatile configuration or GPIO;
+ * may need to be specified if bootstrapping from blank flash.
+ */
+static unsigned int eeprom_type = -1;
+static unsigned int flash_type = -1;
+
+/* RX FIFO XOFF watermark
+ *
+ * When the amount of the RX FIFO increases used increases past this
+ * watermark send XOFF. Only used if RX flow control is enabled (ethtool -A)
+ * This also has an effect on RX/TX arbitration
+ */
+static int rx_xoff_thresh_bytes = -1;
+module_param(rx_xoff_thresh_bytes, int, 0644);
+MODULE_PARM_DESC(rx_xoff_thresh_bytes, "RX fifo XOFF threshold");
+
+/* RX FIFO XON watermark
+ *
+ * When the amount of the RX FIFO used decreases below this
+ * watermark send XON. Only used if TX flow control is enabled (ethtool -A)
+ * This also has an effect on RX/TX arbitration
+ */
+static int rx_xon_thresh_bytes = -1;
+module_param(rx_xon_thresh_bytes, int, 0644);
+MODULE_PARM_DESC(rx_xon_thresh_bytes, "RX fifo XON threshold");
+
+/* TX descriptor ring size - min 512 max 4k */
+#define FALCON_TXD_RING_ORDER TX_DESCQ_SIZE_1K
+#define FALCON_TXD_RING_SIZE 1024
+#define FALCON_TXD_RING_MASK (FALCON_TXD_RING_SIZE - 1)
+
+/* RX descriptor ring size - min 512 max 4k */
+#define FALCON_RXD_RING_ORDER RX_DESCQ_SIZE_1K
+#define FALCON_RXD_RING_SIZE 1024
+#define FALCON_RXD_RING_MASK (FALCON_RXD_RING_SIZE - 1)
+
+/* Event queue size - max 32k */
+#define FALCON_EVQ_ORDER EVQ_SIZE_4K
+#define FALCON_EVQ_SIZE 4096
+#define FALCON_EVQ_MASK (FALCON_EVQ_SIZE - 1)
+
+/* Max number of internal errors. After this resets will not be performed */
+#define FALCON_MAX_INT_ERRORS 4
+
+/* Maximum period that we wait for flush events. If the flush event
+ * doesn't arrive in this period of time then we check if the queue
+ * was disabled anyway. */
+#define FALCON_FLUSH_TIMEOUT 10 /* 10ms */
+
+/**************************************************************************
+ *
+ * Falcon constants
+ *
+ **************************************************************************
+ */
+
+/* DMA address mask (up to 46-bit, avoiding compiler warnings)
+ *
+ * Note that it is possible to have a platform with 64-bit longs and
+ * 32-bit DMA addresses, or vice versa.  EFX_DMA_MASK takes care of the
+ * platform DMA mask.
+ */
+#if BITS_PER_LONG == 64
+#define FALCON_DMA_MASK EFX_DMA_MASK(0x00003fffffffffffUL)
+#else
+#define FALCON_DMA_MASK EFX_DMA_MASK(0x00003fffffffffffULL)
+#endif
+
+/* TX DMA length mask (13-bit) */
+#define FALCON_TX_DMA_MASK (8192 - 1)
+
+/* Alignment of special buffers (4KB) */
+#define FALCON_BUF_ALIGN 4096
+
+/* Dummy SRAM size code */
+#define SRM_NB_BSZ_ONCHIP_ONLY (-1)
+
+/* Be nice if these (or equiv.) were in linux/pci_regs.h, but they're not. */
+#define PCI_EXP_DEVCAP_PWR_VAL_LBN	(18)
+/* This field takes up bits 26 and 27. */
+#define PCI_EXP_DEVCAP_PWR_SCL_LBN	(26)
+#define PCI_EXP_LNKSTA_LNK_WID		(0x3f0)
+#define PCI_EXP_LNKSTA_LNK_WID_LBN	(4)
+
+
+/**************************************************************************
+ *
+ * Falcon hardware access
+ *
+ **************************************************************************/
+
+/* Read the current event from the event queue */
+static inline efx_qword_t *falcon_event(struct efx_channel *channel,
+					unsigned int index)
+{
+	return (((efx_qword_t *) (channel->eventq.addr)) + index);
+}
+
+/* See if an event is present
+ *
+ * We check both the high and low dword of the event for all ones.  We
+ * wrote all ones when we cleared the event, and no valid event can
+ * have all ones in either its high or low dwords.  This approach is
+ * robust against reordering.
+ *
+ * Note that using a single 64-bit comparison is incorrect; even
+ * though the CPU read will be atomic, the DMA write may not be.
+ */
+static inline int falcon_event_present(efx_qword_t *event)
+{
+	return (!(EFX_DWORD_IS_ALL_ONES(event->dword[0]) |
+		  EFX_DWORD_IS_ALL_ONES(event->dword[1])));
+}
+
+/* Read dword from a Falcon PCIE core register */
+static void falcon_pcie_core_read_reg(struct efx_nic *efx, int address,
+				      efx_dword_t *result)
+{
+	efx_oword_t temp;
+
+	BUG_ON(FALCON_REV(efx) < FALCON_REV_B0);
+	BUG_ON(address & 3 || address < 0);
+
+	EFX_POPULATE_OWORD_1(temp, PCIE_CORE_ADDR, address);
+
+	falcon_write(efx, &temp, PCIE_CORE_INDIRECT_REG);
+	falcon_read(efx, &temp, PCIE_CORE_INDIRECT_REG);
+	/* Extract PCIE_CORE_VALUE without byte-swapping */
+	BUILD_BUG_ON(PCIE_CORE_VALUE_LBN != 32 ||
+		     PCIE_CORE_VALUE_WIDTH != 32);
+	result->u32[0] = temp.u32[1];
+}
+
+/* Write dword to a Falcon PCIE core register */
+static void falcon_pcie_core_write_reg(struct efx_nic *efx, int address,
+				       efx_dword_t value)
+{
+	efx_oword_t temp;
+
+	BUG_ON(FALCON_REV(efx) < FALCON_REV_B0);
+	BUG_ON(address & 0x3 || address < 0);
+
+	EFX_POPULATE_OWORD_2(temp,
+			     PCIE_CORE_ADDR, address,
+			     PCIE_CORE_RW, 1);
+	/* Fill PCIE_CORE_VALUE without byte-swapping */
+	BUILD_BUG_ON(PCIE_CORE_VALUE_LBN != 32 ||
+		     PCIE_CORE_VALUE_WIDTH != 32);
+	temp.u32[1] = value.u32[0];
+	falcon_write(efx, &temp, PCIE_CORE_INDIRECT_REG);
+}
+
+/**************************************************************************
+ *
+ * I2C bus - this is a bit-bashing interface using GPIO pins
+ * Note that it uses the output enables to tristate the outputs
+ * SDA is the data pin and SCL is the clock
+ *
+ **************************************************************************
+ */
+static void falcon_setsdascl(struct efx_i2c_interface *i2c)
+{
+	efx_oword_t reg;
+
+	falcon_read(i2c->efx, &reg, GPIO_CTL_REG_KER);
+	EFX_SET_OWORD_FIELD(reg, GPIO0_OEN, (i2c->scl ? 0 : 1));
+	EFX_SET_OWORD_FIELD(reg, GPIO3_OEN, (i2c->sda ? 0 : 1));
+	falcon_write(i2c->efx, &reg, GPIO_CTL_REG_KER);
+}
+
+static int falcon_getsda(struct efx_i2c_interface *i2c)
+{
+	efx_oword_t reg;
+
+	falcon_read(i2c->efx, &reg, GPIO_CTL_REG_KER);
+	return EFX_OWORD_FIELD(reg, GPIO3_IN);
+}
+
+static int falcon_getscl(struct efx_i2c_interface *i2c)
+{
+	efx_oword_t reg;
+
+	falcon_read(i2c->efx, &reg, GPIO_CTL_REG_KER);
+	return EFX_DWORD_FIELD(reg, GPIO0_IN);
+}
+
+static struct efx_i2c_bit_operations falcon_i2c_bit_operations = {
+	.setsda		= falcon_setsdascl,
+	.setscl		= falcon_setsdascl,
+	.getsda		= falcon_getsda,
+	.getscl		= falcon_getscl,
+	.udelay		= 100,
+	.mdelay		= 10,
+};
+
+/**************************************************************************
+ *
+ * Falcon special buffer handling
+ * Special buffers are used for event queues and the TX and RX
+ * descriptor rings.
+ *
+ *************************************************************************/
+
+/* Adds the relevant entries to the full-mode buffer table. */
+static int
+falcon_pin_special_buffer_full(struct efx_nic *efx,
+			       struct efx_special_buffer *buffer)
+{
+	efx_qword_t buf_desc;
+	int index;
+	dma_addr_t dma_addr;
+	int i;
+
+	/* Write buffer descriptors to NIC */
+	for (i = 0; i < buffer->entries; i++) {
+		index = buffer->index + i;
+		dma_addr = buffer->dma_addr + (i * 4096);
+		EFX_LOG(efx, "mapping special buffer %d at %llx\n",
+			index, (unsigned long long)dma_addr);
+		EFX_POPULATE_QWORD_4(buf_desc,
+				     IP_DAT_BUF_SIZE, IP_DAT_BUF_SIZE_4K,
+				     BUF_ADR_REGION, 0,
+				     BUF_ADR_FBUF, (dma_addr >> 12),
+				     BUF_OWNER_ID_FBUF, 0);
+		falcon_write_sram(efx, &buf_desc, index);
+	}
+
+	return 0;
+}
+
+/* Clears the relevant entries from the buffer table */
+static void
+falcon_clear_special_buffer_full(struct efx_nic *efx,
+				 struct efx_special_buffer *buffer)
+{
+	efx_oword_t buf_tbl_upd;
+	unsigned int start = buffer->index;
+	unsigned int end = (buffer->index + buffer->entries - 1);
+
+	EFX_LOG(efx, "unmapping special buffers %d-%d\n",
+		buffer->index, buffer->index + buffer->entries - 1);
+
+	EFX_POPULATE_OWORD_4(buf_tbl_upd,
+			     BUF_UPD_CMD, 0,
+			     BUF_CLR_CMD, 1,
+			     BUF_CLR_END_ID, end,
+			     BUF_CLR_START_ID, start);
+	falcon_write(efx, &buf_tbl_upd, BUF_TBL_UPD_REG_KER);
+}
+
+/*
+ * Allocate a new Falcon special buffer
+ *
+ * This allocates memory for a new buffer, clears it and allocates a
+ * new buffer ID range.  It does not write into Falcon's buffer table.
+ *
+ * This call will allocate 4kB buffers, since Falcon can't use 8kB
+ * buffers for event queues and descriptor rings.  It will always
+ * allocate an even number of 4kB buffers, since when we're in
+ * half-entry mode for the buffer table we can only deal with pairs of
+ * buffers.
+ */
+static int falcon_alloc_special_buffer(struct efx_nic *efx,
+				       struct efx_special_buffer *buffer,
+				       unsigned int len)
+{
+	struct falcon_nic_data *nic_data = efx->nic_data;
+
+	/* Round size up to an 8kB boundary (i.e. pairs of 4kB buffers) */
+	len = (len + 8192 - 1) & ~(8192 - 1);
+
+	/* Allocate buffer as consistent PCI DMA space */
+	buffer->addr = pci_alloc_consistent(efx->pci_dev, len,
+					    &buffer->dma_addr);
+	if (!buffer->addr)
+		return -ENOMEM;
+	buffer->len = len;
+	buffer->entries = len / 4096;
+	BUG_ON(buffer->dma_addr & (FALCON_BUF_ALIGN - 1));
+
+	/* All zeros is a potentially valid event so memset to 0xff */
+	memset(buffer->addr, 0xff, len);
+
+	/* Select new buffer ID */
+	buffer->index = nic_data->resources.buffer_table_min;
+	nic_data->resources.buffer_table_min += buffer->entries;
+
+	EFX_LOG(efx, "allocating special buffers %d-%d at %llx+%x "
+		"(virt %p phys %lx)\n", buffer->index,
+		buffer->index + buffer->entries - 1,
+		(unsigned long long)buffer->dma_addr, len,
+		buffer->addr, virt_to_phys(buffer->addr));
+
+	return 0;
+}
+
+/*
+ * Initialise a Falcon special buffer
+ *
+ * This will define a buffer (previously allocated via
+ * falcon_alloc_special_buffer()) in Falcon's buffer table, allowing
+ * it to be used for event queues, descriptor rings etc.
+ */
+static int falcon_init_special_buffer(struct efx_nic *efx,
+				      struct efx_special_buffer *buffer)
+{
+	EFX_BUG_ON_PARANOID(!buffer->addr);
+
+	/* Write buffer descriptors to NIC */
+	return falcon_pin_special_buffer_full(efx, buffer);
+}
+
+/* Unmaps a buffer from Falcon and clears the buffer table
+ * entries */
+static void falcon_fini_special_buffer(struct efx_nic *efx,
+				       struct efx_special_buffer *buffer)
+{
+
+	if (!buffer->entries)
+		return;
+
+	falcon_clear_special_buffer_full(efx, buffer);
+}
+
+/* Release the buffer memory. */
+static void falcon_free_special_buffer(struct efx_nic *efx,
+				       struct efx_special_buffer *buffer)
+{
+	if (!buffer->addr)
+		return;
+
+	EFX_LOG(efx, "deallocating special buffers %d-%d at %llx+%x "
+		"(virt %p phys %lx)\n", buffer->index,
+		buffer->index + buffer->entries - 1,
+		(unsigned long long)buffer->dma_addr, buffer->len,
+		buffer->addr, virt_to_phys(buffer->addr));
+
+	pci_free_consistent(efx->pci_dev, buffer->len, buffer->addr,
+			    buffer->dma_addr);
+	buffer->addr = NULL;
+	buffer->entries = 0;
+}
+
+/**************************************************************************
+ *
+ * Falcon generic buffer handling
+ * These buffers are used for interrupt status and MAC stats
+ *
+ **************************************************************************/
+
+static int falcon_alloc_buffer(struct efx_nic *efx,
+			       struct efx_buffer *buffer, unsigned int len)
+{
+	buffer->addr = pci_alloc_consistent(efx->pci_dev, len,
+					    &buffer->dma_addr);
+	if (!buffer->addr)
+		return -ENOMEM;
+	buffer->len = len;
+	memset(buffer->addr, 0, len);
+	return 0;
+}
+
+static void falcon_free_buffer(struct efx_nic *efx, struct efx_buffer *buffer)
+{
+	if (buffer->addr) {
+		pci_free_consistent(efx->pci_dev, buffer->len,
+				    buffer->addr, buffer->dma_addr);
+		buffer->addr = NULL;
+	}
+}
+
+/**************************************************************************
+ *
+ * Falcon TX path
+ *
+ **************************************************************************/
+
+/* Returns a pointer to the specified transmit descriptor in the TX
+ * descriptor queue belonging to the specified channel.
+ */
+static inline efx_qword_t *falcon_tx_desc(struct efx_tx_queue *tx_queue,
+					       unsigned int index)
+{
+	return (((efx_qword_t *) (tx_queue->txd.addr)) + index);
+}
+
+/* Update TX descriptor write pointer
+ * This writes to the TX_DESC_WPTR register for the specified
+ * channel's transmit descriptor ring.
+ */
+static inline void falcon_notify_tx_desc(struct efx_tx_queue *tx_queue)
+{
+	unsigned write_ptr;
+	efx_dword_t reg;
+
+	write_ptr = tx_queue->write_count & FALCON_TXD_RING_MASK;
+	EFX_POPULATE_DWORD_1(reg, TX_DESC_WPTR_DWORD, write_ptr);
+	falcon_writel_page(tx_queue->efx, &reg,
+			   TX_DESC_UPD_REG_KER_DWORD, tx_queue->queue);
+}
+
+
+/* For each entry inserted into the software descriptor ring, create a
+ * descriptor in the hardware TX descriptor ring (in host memory), and
+ * write a doorbell.
+ */
+#if defined(EFX_USE_FASTCALL)
+void fastcall falcon_push_buffers(struct efx_tx_queue *tx_queue)
+#else
+void falcon_push_buffers(struct efx_tx_queue *tx_queue)
+#endif
+{
+
+	struct efx_tx_buffer *buffer;
+	efx_qword_t *txd;
+	unsigned write_ptr;
+
+	BUG_ON(tx_queue->write_count == tx_queue->insert_count);
+
+	do {
+		write_ptr = tx_queue->write_count & FALCON_TXD_RING_MASK;
+		buffer = &tx_queue->buffer[write_ptr];
+		txd = falcon_tx_desc(tx_queue, write_ptr);
+		++tx_queue->write_count;
+
+		/* Create TX descriptor ring entry */
+		EFX_POPULATE_QWORD_5(*txd,
+				     TX_KER_PORT, 0,
+				     TX_KER_CONT, buffer->continuation,
+				     TX_KER_BYTE_CNT, buffer->len,
+				     TX_KER_BUF_REGION, 0,
+				     TX_KER_BUF_ADR, buffer->dma_addr);
+	} while (tx_queue->write_count != tx_queue->insert_count);
+
+	wmb(); /* Ensure descriptors are written before they are fetched */
+	falcon_notify_tx_desc(tx_queue);
+}
+
+/* Allocate hardware resources for a TX queue */
+int falcon_probe_tx(struct efx_tx_queue *tx_queue)
+{
+	struct efx_nic *efx = tx_queue->efx;
+	struct falcon_nic_data *nic_data = efx->nic_data;
+	int rc;
+
+	rc = falcon_alloc_special_buffer(efx, &tx_queue->txd,
+					 FALCON_TXD_RING_SIZE *
+					 sizeof(efx_qword_t));
+	if (rc)
+		return rc;
+
+	nic_data->resources.txq_min = max(nic_data->resources.txq_min,
+					  (unsigned)tx_queue->queue + 1);
+
+	return 0;
+}
+
+/* Prepare channel's TX datapath. */
+int falcon_init_tx(struct efx_tx_queue *tx_queue)
+{
+	efx_oword_t tx_desc_ptr;
+	struct efx_nic *efx = tx_queue->efx;
+	int rc;
+
+	/* Pin TX descriptor ring */
+	rc = falcon_init_special_buffer(efx, &tx_queue->txd);
+	if (rc)
+		return rc;
+
+	/* Push TX descriptor ring to card */
+	EFX_POPULATE_OWORD_10(tx_desc_ptr,
+			      TX_DESCQ_EN, 1,
+			      TX_ISCSI_DDIG_EN, 0,
+			      TX_ISCSI_HDIG_EN, 0,
+			      TX_DESCQ_BUF_BASE_ID, tx_queue->txd.index,
+			      TX_DESCQ_EVQ_ID, tx_queue->channel->evqnum,
+			      TX_DESCQ_OWNER_ID, 0,
+			      TX_DESCQ_LABEL, tx_queue->queue,
+			      TX_DESCQ_SIZE, FALCON_TXD_RING_ORDER,
+			      TX_DESCQ_TYPE, 0,	/* kernel queue */
+			      TX_NON_IP_DROP_DIS_B0, 1);
+
+	if (FALCON_REV(efx) >= FALCON_REV_B0) {
+		int csum = !(efx->net_dev->features & NETIF_F_IP_CSUM);
+		EFX_SET_OWORD_FIELD(tx_desc_ptr, TX_IP_CHKSM_DIS_B0, csum);
+		EFX_SET_OWORD_FIELD(tx_desc_ptr, TX_TCP_CHKSM_DIS_B0, csum);
+	}
+
+	falcon_write_table(efx, &tx_desc_ptr, efx->type->txd_ptr_tbl_base,
+			   tx_queue->queue);
+
+	if (FALCON_REV(efx) < FALCON_REV_B0) {
+		efx_oword_t reg;
+
+		/* Only 128 bits in this register */
+		BUG_ON(tx_queue->queue >= 128);
+
+		falcon_read(efx, &reg, TX_CHKSM_CFG_REG_KER_A1);
+		if (efx->net_dev->features & NETIF_F_IP_CSUM)
+			clear_bit_le(tx_queue->queue, (void *)&reg);
+		else
+			set_bit_le(tx_queue->queue, (void *)&reg);
+		falcon_write(efx, &reg, TX_CHKSM_CFG_REG_KER_A1);
+	}
+
+	return 0;
+}
+
+static int falcon_flush_tx_queue(struct efx_tx_queue *tx_queue)
+{
+	struct efx_nic *efx = tx_queue->efx;
+	struct efx_channel *channel = &efx->channel[0];
+	efx_oword_t tx_flush_descq;
+	unsigned int read_ptr, i;
+
+	/* Post a flush command */
+	EFX_POPULATE_OWORD_2(tx_flush_descq,
+			     TX_FLUSH_DESCQ_CMD, 1,
+			     TX_FLUSH_DESCQ, tx_queue->queue);
+	falcon_write(efx, &tx_flush_descq, TX_FLUSH_DESCQ_REG_KER);
+	msleep(FALCON_FLUSH_TIMEOUT);
+
+	/* If the NIC is resetting then don't bother checking */
+	if (EFX_WORKAROUND_7803(efx) || (efx->state == STATE_RESETTING))
+		return 0;
+
+	/* Look for a flush completed event */
+	read_ptr = channel->eventq_read_ptr;
+	for (i = 0; i < FALCON_EVQ_SIZE; ++i) {
+		efx_qword_t *event = falcon_event(channel, read_ptr);
+		int ev_code, ev_sub_code, ev_queue;
+		if (!falcon_event_present(event))
+			break;
+
+		ev_code = EFX_QWORD_FIELD(*event, EV_CODE);
+		ev_sub_code = EFX_QWORD_FIELD(*event, DRIVER_EV_SUB_CODE);
+		ev_queue = EFX_QWORD_FIELD(*event, DRIVER_EV_TX_DESCQ_ID);
+		if ((ev_sub_code == TX_DESCQ_FLS_DONE_EV_DECODE) &&
+		    (ev_queue == tx_queue->queue)) {
+			EFX_LOG(efx, "tx queue %d flush command succesful\n",
+				tx_queue->queue);
+			return 0;
+		}
+
+		read_ptr = (read_ptr + 1) & FALCON_EVQ_MASK;
+	}
+
+	if (EFX_WORKAROUND_11557(efx)) {
+		efx_oword_t reg;
+		int enabled;
+
+		falcon_read_table(efx, &reg, efx->type->txd_ptr_tbl_base,
+				  tx_queue->queue);
+		enabled = EFX_OWORD_FIELD(reg, TX_DESCQ_EN);
+		if (!enabled) {
+			EFX_LOG(efx, "tx queue %d disabled without a "
+				"flush event seen\n", tx_queue->queue);
+			return 0;
+		}
+	}
+
+	EFX_ERR(efx, "tx queue %d flush command timed out\n", tx_queue->queue);
+	return -ETIMEDOUT;
+}
+
+void falcon_fini_tx(struct efx_tx_queue *tx_queue)
+{
+	struct efx_nic *efx = tx_queue->efx;
+	efx_oword_t tx_desc_ptr;
+
+	/* Stop the hardware using the queue */
+	if (falcon_flush_tx_queue(tx_queue))
+		EFX_ERR(efx, "failed to flush tx queue %d\n", tx_queue->queue);
+
+	/* Remove TX descriptor ring from card */
+	EFX_ZERO_OWORD(tx_desc_ptr);
+	falcon_write_table(efx, &tx_desc_ptr, efx->type->txd_ptr_tbl_base,
+			   tx_queue->queue);
+
+	/* Unpin TX descriptor ring */
+	falcon_fini_special_buffer(efx, &tx_queue->txd);
+}
+
+/* Free buffers backing TX queue */
+void falcon_remove_tx(struct efx_tx_queue *tx_queue)
+{
+	falcon_free_special_buffer(tx_queue->efx, &tx_queue->txd);
+}
+
+/**************************************************************************
+ *
+ * Falcon RX path
+ *
+ **************************************************************************/
+
+/* Returns a pointer to the specified transmit descriptor in the RX
+ * descriptor queue.
+ */
+static inline efx_qword_t *falcon_rx_desc(struct efx_rx_queue *rx_queue,
+					       unsigned int index)
+{
+	return (((efx_qword_t *) (rx_queue->rxd.addr)) + index);
+}
+
+/* This creates an entry in the RX descriptor queue corresponding to
+ * the receive buffer.
+ */
+static inline void falcon_build_rx_desc(struct efx_rx_queue *rx_queue,
+					unsigned index)
+{
+	struct efx_rx_buffer *rx_buf;
+	efx_qword_t *rxd;
+
+	rxd = falcon_rx_desc(rx_queue, index);
+	rx_buf = efx_rx_buffer(rx_queue, index);
+	EFX_POPULATE_QWORD_3(*rxd,
+			     RX_KER_BUF_SIZE,
+			     rx_buf->len -
+			     rx_queue->efx->type->rx_buffer_padding,
+			     RX_KER_BUF_REGION, 0,
+			     RX_KER_BUF_ADR, rx_buf->dma_addr);
+}
+
+/* This writes to the RX_DESC_WPTR register for the specified receive
+ * descriptor ring.
+ */
+#if defined(EFX_USE_FASTCALL)
+void fastcall falcon_notify_rx_desc(struct efx_rx_queue *rx_queue)
+#else
+void falcon_notify_rx_desc(struct efx_rx_queue *rx_queue)
+#endif
+{
+	efx_dword_t reg;
+	unsigned write_ptr;
+
+	while (rx_queue->notified_count != rx_queue->added_count) {
+		falcon_build_rx_desc(rx_queue,
+				     rx_queue->notified_count &
+				     FALCON_RXD_RING_MASK);
+		++rx_queue->notified_count;
+	}
+
+	wmb();
+	write_ptr = rx_queue->added_count & FALCON_RXD_RING_MASK;
+	EFX_POPULATE_DWORD_1(reg, RX_DESC_WPTR_DWORD, write_ptr);
+	falcon_writel_page(rx_queue->efx, &reg,
+			   RX_DESC_UPD_REG_KER_DWORD, rx_queue->queue);
+}
+
+int falcon_probe_rx(struct efx_rx_queue *rx_queue)
+{
+	struct efx_nic *efx = rx_queue->efx;
+	struct falcon_nic_data *nic_data = efx->nic_data;
+	int rc;
+
+	rc = falcon_alloc_special_buffer(efx, &rx_queue->rxd,
+					 FALCON_RXD_RING_SIZE *
+					 sizeof(efx_qword_t));
+	if (rc)
+		return rc;
+
+	/* Increment the rxq_min counter */
+	nic_data->resources.rxq_min = max(nic_data->resources.rxq_min,
+					  (unsigned)rx_queue->queue + 1);
+
+	return 0;
+}
+
+int falcon_init_rx(struct efx_rx_queue *rx_queue)
+{
+	efx_oword_t rx_desc_ptr;
+	struct efx_nic *efx = rx_queue->efx;
+	int rc;
+	int is_b0 = FALCON_REV(efx) >= FALCON_REV_B0;
+	int iscsi_digest_en = is_b0;
+
+	EFX_LOG(efx, "RX queue %d ring in special buffers %d-%d\n",
+		rx_queue->queue, rx_queue->rxd.index,
+		rx_queue->rxd.index + rx_queue->rxd.entries - 1);
+
+	/* Pin RX descriptor ring */
+	rc = falcon_init_special_buffer(efx, &rx_queue->rxd);
+	if (rc)
+		return rc;
+
+	/* Push RX descriptor ring to card */
+	EFX_POPULATE_OWORD_10(rx_desc_ptr,
+			      RX_ISCSI_DDIG_EN, iscsi_digest_en,
+			      RX_ISCSI_HDIG_EN, iscsi_digest_en,
+			      RX_DESCQ_BUF_BASE_ID, rx_queue->rxd.index,
+			      RX_DESCQ_EVQ_ID, rx_queue->channel->evqnum,
+			      RX_DESCQ_OWNER_ID, 0,
+			      RX_DESCQ_LABEL, rx_queue->queue,
+			      RX_DESCQ_SIZE, FALCON_RXD_RING_ORDER,
+			      RX_DESCQ_TYPE, 0 /* kernel queue */ ,
+			      /* For >=B0 this is scatter so disable */
+			      RX_DESCQ_JUMBO, !is_b0,
+			      RX_DESCQ_EN, 1);
+	falcon_write_table(efx, &rx_desc_ptr, efx->type->rxd_ptr_tbl_base,
+			   rx_queue->queue);
+	return 0;
+}
+
+static int falcon_flush_rx_queue(struct efx_rx_queue *rx_queue)
+{
+	struct efx_nic *efx = rx_queue->efx;
+	struct efx_channel *channel = &efx->channel[0];
+	unsigned int read_ptr, i;
+	efx_oword_t rx_flush_descq;
+
+	/* Post a flush command */
+	EFX_POPULATE_OWORD_2(rx_flush_descq,
+			     RX_FLUSH_DESCQ_CMD, 1,
+			     RX_FLUSH_DESCQ, rx_queue->queue);
+
+	falcon_write(efx, &rx_flush_descq, RX_FLUSH_DESCQ_REG_KER);
+	msleep(FALCON_FLUSH_TIMEOUT);
+
+	/* If the NIC is resetting then don't bother checking */
+	if (EFX_WORKAROUND_7803(efx) || (efx->state == STATE_RESETTING))
+		return 0;
+
+	/* Look for a flush completed event */
+	read_ptr = channel->eventq_read_ptr;
+	for (i = 0; i < FALCON_EVQ_SIZE; ++i) {
+		efx_qword_t *event = falcon_event(channel, read_ptr);
+		int ev_code, ev_sub_code, ev_queue, ev_failed;
+		if (!falcon_event_present(event))
+			break;
+
+		ev_code = EFX_QWORD_FIELD(*event, EV_CODE);
+		ev_sub_code = EFX_QWORD_FIELD(*event, DRIVER_EV_SUB_CODE);
+		ev_queue = EFX_QWORD_FIELD(*event, DRIVER_EV_RX_DESCQ_ID);
+		ev_failed = EFX_QWORD_FIELD(*event, DRIVER_EV_RX_FLUSH_FAIL);
+
+		if ((ev_sub_code == RX_DESCQ_FLS_DONE_EV_DECODE) &&
+		    (ev_queue == rx_queue->queue)) {
+			if (ev_failed) {
+				EFX_INFO(efx, "rx queue %d flush command "
+					 "failed\n", rx_queue->queue);
+				return -EAGAIN;
+			} else {
+				EFX_LOG(efx, "rx queue %d flush command "
+					"succesful\n", rx_queue->queue);
+				return 0;
+			}
+		}
+
+		read_ptr = (read_ptr + 1) & FALCON_EVQ_MASK;
+	}
+
+	if (EFX_WORKAROUND_11557(efx)) {
+		efx_oword_t reg;
+		int enabled;
+
+		falcon_read_table(efx, &reg, efx->type->rxd_ptr_tbl_base,
+				  rx_queue->queue);
+		enabled = EFX_OWORD_FIELD(reg, RX_DESCQ_EN);
+		if (!enabled) {
+			EFX_LOG(efx, "rx queue %d disabled without a "
+				"flush event seen\n", rx_queue->queue);
+			return 0;
+		}
+	}
+
+	EFX_ERR(efx, "rx queue %d flush command timed out\n", rx_queue->queue);
+	return -ETIMEDOUT;
+}
+
+void falcon_fini_rx(struct efx_rx_queue *rx_queue)
+{
+	efx_oword_t rx_desc_ptr;
+	struct efx_nic *efx = rx_queue->efx;
+	int i, rc;
+
+	/* Try and flush the rx queue. This may need to be repeated */
+	for (i = 0; i < 5; i++) {
+		rc = falcon_flush_rx_queue(rx_queue);
+		if (rc == -EAGAIN)
+			continue;
+		break;
+	}
+	if (rc)
+		EFX_ERR(efx, "failed to flush rx queue %d\n", rx_queue->queue);
+
+	/* Remove RX descriptor ring from card */
+	EFX_ZERO_OWORD(rx_desc_ptr);
+	falcon_write_table(efx, &rx_desc_ptr, efx->type->rxd_ptr_tbl_base,
+			   rx_queue->queue);
+
+	/* Unpin RX descriptor ring */
+	falcon_fini_special_buffer(efx, &rx_queue->rxd);
+}
+
+/* Free buffers backing RX queue */
+void falcon_remove_rx(struct efx_rx_queue *rx_queue)
+{
+	falcon_free_special_buffer(rx_queue->efx, &rx_queue->rxd);
+}
+
+/**************************************************************************
+ *
+ * Falcon event queue processing
+ * Event queues are processed by per-channel tasklets.
+ *
+ **************************************************************************/
+
+/* Update a channel's event queue's read pointer (RPTR) register
+ *
+ * This writes the EVQ_RPTR_REG register for the specified channel's
+ * event queue.
+ *
+ * Note that EVQ_RPTR_REG contains the index of the "last read" event,
+ * whereas channel->eventq_read_ptr contains the index of the "next to
+ * read" event.
+ */
+#if defined(EFX_USE_FASTCALL)
+void fastcall falcon_eventq_read_ack(struct efx_channel *channel)
+#else
+void falcon_eventq_read_ack(struct efx_channel *channel)
+#endif
+{
+	efx_dword_t reg;
+	struct efx_nic *efx = channel->efx;
+
+	EFX_POPULATE_DWORD_1(reg, EVQ_RPTR_DWORD, channel->eventq_read_ptr);
+	falcon_writel_table(efx, &reg, efx->type->evq_rptr_tbl_base,
+			    channel->evqnum);
+}
+
+/* Use HW to insert a SW defined event */
+void falcon_generate_event(struct efx_channel *channel, efx_qword_t *event)
+{
+	efx_oword_t drv_ev_reg;
+
+	EFX_POPULATE_OWORD_2(drv_ev_reg,
+			     DRV_EV_QID, channel->evqnum,
+			     DRV_EV_DATA,
+			     EFX_QWORD_FIELD64(*event, WHOLE_EVENT));
+	falcon_write(channel->efx, &drv_ev_reg, DRV_EV_REG_KER);
+}
+
+/* Handle a transmit completion event
+ *
+ * Falcon batches TX completion events; the message we receive is of
+ * the form "complete all TX events up to this index".
+ */
+static inline void falcon_handle_tx_event(struct efx_channel *channel,
+					  efx_qword_t *event)
+{
+	unsigned int tx_ev_desc_ptr;
+	unsigned int tx_ev_q_label;
+	struct efx_tx_queue *tx_queue;
+	struct efx_nic *efx = channel->efx;
+
+	if (likely(EFX_QWORD_FIELD(*event, TX_EV_COMP))) {
+		/* Transmit completion */
+		tx_ev_desc_ptr = EFX_QWORD_FIELD(*event, TX_EV_DESC_PTR);
+		tx_ev_q_label = EFX_QWORD_FIELD(*event, TX_EV_Q_LABEL);
+		tx_queue = &efx->tx_queue[tx_ev_q_label];
+		efx_xmit_done(tx_queue, tx_ev_desc_ptr);
+	} else if (EFX_QWORD_FIELD(*event, TX_EV_WQ_FF_FULL)) {
+		/* Rewrite the FIFO write pointer */
+		tx_ev_q_label = EFX_QWORD_FIELD(*event, TX_EV_Q_LABEL);
+		tx_queue = &efx->tx_queue[tx_ev_q_label];
+
+		if (efx->net_dev_registered)
+			netif_tx_lock(efx->net_dev);
+		falcon_notify_tx_desc(tx_queue);
+		if (efx->net_dev_registered)
+			netif_tx_unlock(efx->net_dev);
+	} else if (EFX_QWORD_FIELD(*event, TX_EV_PKT_ERR) &&
+		   EFX_WORKAROUND_10727(efx)) {
+		efx_schedule_reset(efx, RESET_TYPE_TX_DESC_FETCH);
+	} else {
+		EFX_ERR(efx, "channel %d unexpected TX event "
+			EFX_QWORD_FMT"\n", channel->channel,
+			EFX_QWORD_VAL(*event));
+	}
+}
+
+/* Check received packet's destination MAC address. */
+static int check_dest_mac(struct efx_rx_queue *rx_queue,
+			  const efx_qword_t *event)
+{
+	struct efx_rx_buffer *rx_buf;
+	struct efx_nic *efx = rx_queue->efx;
+	int rx_ev_desc_ptr;
+	struct ethhdr *eh;
+
+	if (efx->promiscuous)
+		return 1;
+
+	rx_ev_desc_ptr = EFX_QWORD_FIELD(*event, RX_EV_DESC_PTR);
+	rx_buf = efx_rx_buffer(rx_queue, rx_ev_desc_ptr);
+	eh = (struct ethhdr *)rx_buf->data;
+	if (memcmp(eh->h_dest, efx->net_dev->dev_addr, ETH_ALEN))
+		return 0;
+	return 1;
+}
+
+/* Detect errors included in the rx_evt_pkt_ok bit. */
+static void falcon_handle_rx_not_ok(struct efx_rx_queue *rx_queue,
+				    const efx_qword_t *event,
+				    unsigned *rx_ev_pkt_ok,
+				    int *discard, int byte_count)
+{
+	struct efx_nic *efx = rx_queue->efx;
+	unsigned rx_ev_buf_owner_id_err, rx_ev_ip_hdr_chksum_err;
+	unsigned rx_ev_tcp_udp_chksum_err, rx_ev_eth_crc_err;
+	unsigned rx_ev_frm_trunc, rx_ev_drib_nib, rx_ev_tobe_disc;
+	unsigned rx_ev_pkt_type, rx_ev_other_err, rx_ev_pause_frm;
+	unsigned rx_ev_ip_frag_err, rx_ev_hdr_type, rx_ev_mcast_pkt;
+	int snap, non_ip;
+
+	rx_ev_hdr_type = EFX_QWORD_FIELD(*event, RX_EV_HDR_TYPE);
+	rx_ev_mcast_pkt = EFX_QWORD_FIELD(*event, RX_EV_MCAST_PKT);
+	rx_ev_tobe_disc = EFX_QWORD_FIELD(*event, RX_EV_TOBE_DISC);
+	rx_ev_pkt_type = EFX_QWORD_FIELD(*event, RX_EV_PKT_TYPE);
+	rx_ev_buf_owner_id_err = EFX_QWORD_FIELD(*event,
+						 RX_EV_BUF_OWNER_ID_ERR);
+	rx_ev_ip_frag_err = EFX_QWORD_FIELD(*event, RX_EV_IF_FRAG_ERR);
+	rx_ev_ip_hdr_chksum_err = EFX_QWORD_FIELD(*event,
+						  RX_EV_IP_HDR_CHKSUM_ERR);
+	rx_ev_tcp_udp_chksum_err = EFX_QWORD_FIELD(*event,
+						   RX_EV_TCP_UDP_CHKSUM_ERR);
+	rx_ev_eth_crc_err = EFX_QWORD_FIELD(*event, RX_EV_ETH_CRC_ERR);
+	rx_ev_frm_trunc = EFX_QWORD_FIELD(*event, RX_EV_FRM_TRUNC);
+	rx_ev_drib_nib = ((FALCON_REV(efx) >= FALCON_REV_B0) ?
+			  0 : EFX_QWORD_FIELD(*event, RX_EV_DRIB_NIB));
+	rx_ev_pause_frm = EFX_QWORD_FIELD(*event, RX_EV_PAUSE_FRM_ERR);
+
+	/* Every error apart from tobe_disc and pause_frm */
+	rx_ev_other_err = (rx_ev_drib_nib | rx_ev_tcp_udp_chksum_err |
+			   rx_ev_buf_owner_id_err | rx_ev_eth_crc_err |
+			   rx_ev_frm_trunc | rx_ev_ip_hdr_chksum_err);
+
+	snap = (rx_ev_pkt_type == RX_EV_PKT_TYPE_LLC_DECODE) ||
+		(rx_ev_pkt_type == RX_EV_PKT_TYPE_VLAN_LLC_DECODE);
+	non_ip = (rx_ev_hdr_type == RX_EV_HDR_TYPE_NON_IP_DECODE);
+
+	/* SFC bug 5475/8970: The Falcon XMAC incorrectly calculates the
+	 * length field of an LLC frame, which sets TOBE_DISC. We could set
+	 * PASS_LEN_ERR, but we want the MAC to filter out short frames (to
+	 * protect the RX block).
+	 *
+	 * bug5475 - LLC/SNAP: Falcon identifies SNAP packets.
+	 * bug8970 - LLC/noSNAP: Falcon does not provide an LLC flag.
+	 *                       LLC can't encapsulate IP, so by definition
+	 *                       these packets are NON_IP.
+	 *
+	 * Unicast mismatch will also cause TOBE_DISC, so the driver needs
+	 * to check this.
+	 */
+	if (EFX_WORKAROUND_5475(efx) && rx_ev_tobe_disc && (snap || non_ip)) {
+		/* If all the other flags are zero then we can state the
+		 * entire packet is ok, which will flag to the kernel not
+		 * to recalculate checksums.
+		 */
+		if (!(non_ip | rx_ev_other_err | rx_ev_pause_frm))
+			*rx_ev_pkt_ok = 1;
+
+		rx_ev_tobe_disc = 0;
+
+		/* TOBE_DISC is set for unicast mismatch.  But given that
+		 * we can't trust TOBE_DISC here, we must validate the dest
+		 * MAC address ourselves.
+		 */
+		if (!rx_ev_mcast_pkt && !check_dest_mac(rx_queue, event))
+			rx_ev_tobe_disc = 1;
+	}
+
+	/* Count errors that are not in MAC stats. */
+	if (rx_ev_frm_trunc)
+		++rx_queue->channel->n_rx_frm_trunc;
+	else if (rx_ev_tobe_disc)
+		++rx_queue->channel->n_rx_tobe_disc;
+	else if (rx_ev_ip_hdr_chksum_err)
+		++rx_queue->channel->n_rx_ip_hdr_chksum_err;
+	else if (rx_ev_tcp_udp_chksum_err)
+		++rx_queue->channel->n_rx_tcp_udp_chksum_err;
+	if (rx_ev_ip_frag_err)
+		++rx_queue->channel->n_rx_ip_frag_err;
+
+	/* The frame must be discarded if any of these are true. */
+	*discard = (rx_ev_eth_crc_err | rx_ev_frm_trunc | rx_ev_drib_nib |
+		    rx_ev_tobe_disc | rx_ev_pause_frm);
+
+	/* TOBE_DISC is expected on unicast mismatches; don't print out an
+	 * error message.  FRM_TRUNC indicates RXDP dropped the packet due
+	 * to a FIFO overflow.
+	 */
+#ifdef EFX_ENABLE_DEBUG
+	if (rx_ev_other_err) {
+		EFX_INFO_RL(efx, " RX queue %d unexpected RX event "
+			    EFX_QWORD_FMT "%s%s%s%s%s%s%s%s%s\n",
+			    rx_queue->queue, EFX_QWORD_VAL(*event),
+			    rx_ev_buf_owner_id_err ? " [OWNER_ID_ERR]" : "",
+			    rx_ev_ip_hdr_chksum_err ?
+			    " [IP_HDR_CHKSUM_ERR]" : "",
+			    rx_ev_tcp_udp_chksum_err ?
+			    " [TCP_UDP_CHKSUM_ERR]" : "",
+			    rx_ev_eth_crc_err ? " [ETH_CRC_ERR]" : "",
+			    rx_ev_frm_trunc ? " [FRM_TRUNC]" : "",
+			    rx_ev_drib_nib ? " [DRIB_NIB]" : "",
+			    rx_ev_tobe_disc ? " [TOBE_DISC]" : "",
+			    rx_ev_pause_frm ? " [PAUSE]" : "",
+			    snap ? " [SNAP/LLC]" : "");
+	}
+#endif
+
+	if (unlikely(rx_ev_eth_crc_err && EFX_WORKAROUND_10750(efx) &&
+		     efx->phy_type == PHY_TYPE_10XPRESS))
+		tenxpress_crc_err(efx);
+}
+
+
+/* Handle receive events that are not in-order. */
+static void falcon_handle_rx_bad_index(struct efx_rx_queue *rx_queue,
+				       unsigned index)
+{
+	struct efx_nic *efx = rx_queue->efx;
+	unsigned expected, dropped;
+
+	expected = rx_queue->removed_count & FALCON_RXD_RING_MASK;
+	dropped = ((index + FALCON_RXD_RING_SIZE - expected) &
+		   FALCON_RXD_RING_MASK);
+	EFX_INFO(efx, "dropped %d events (index=%d expected=%d)\n",
+		dropped, index, expected);
+
+	atomic_inc(&efx->errors.missing_event);
+	efx_schedule_reset(efx, EFX_WORKAROUND_5676(efx) ?
+			   RESET_TYPE_RX_RECOVERY : RESET_TYPE_DISABLE);
+}
+
+
+/* Handle a packet received event
+ *
+ * Falcon silicon gives a "discard" flag if it's a unicast packet with the
+ * wrong destination address
+ * Also "is multicast" and "matches multicast filter" flags can be used to
+ * discard non-matching multicast packets.
+ */
+static inline int falcon_handle_rx_event(struct efx_channel *channel,
+					 const efx_qword_t *event)
+{
+	unsigned int rx_ev_q_label, rx_ev_desc_ptr, rx_ev_byte_cnt;
+	unsigned int rx_ev_pkt_ok, rx_ev_hdr_type, rx_ev_mcast_pkt;
+	unsigned expected_ptr;
+	int discard = 0, checksummed;
+	struct efx_rx_queue *rx_queue;
+	struct efx_nic *efx = channel->efx;
+
+	/* Basic packet information */
+	rx_ev_byte_cnt = EFX_QWORD_FIELD(*event, RX_EV_BYTE_CNT);
+	rx_ev_pkt_ok = EFX_QWORD_FIELD(*event, RX_EV_PKT_OK);
+	rx_ev_hdr_type = EFX_QWORD_FIELD(*event, RX_EV_HDR_TYPE);
+	WARN_ON(EFX_QWORD_FIELD(*event, RX_EV_JUMBO_CONT));
+	WARN_ON(EFX_QWORD_FIELD(*event, RX_EV_SOP) != 1);
+
+	rx_ev_q_label = EFX_QWORD_FIELD(*event, RX_EV_Q_LABEL);
+	rx_queue = &efx->rx_queue[rx_ev_q_label];
+
+	rx_ev_desc_ptr = EFX_QWORD_FIELD(*event, RX_EV_DESC_PTR);
+	expected_ptr = rx_queue->removed_count & FALCON_RXD_RING_MASK;
+	if (unlikely(rx_ev_desc_ptr != expected_ptr)) {
+		falcon_handle_rx_bad_index(rx_queue, rx_ev_desc_ptr);
+		return rx_ev_q_label;
+	}
+
+	if (likely(rx_ev_pkt_ok)) {
+		/* If packet is marked as OK and packet type is TCP/IPv4 or
+		 * UDP/IPv4, then we can rely on the hardware checksum.
+		 */
+		checksummed = RX_EV_HDR_TYPE_HAS_CHECKSUMS(rx_ev_hdr_type);
+	} else {
+		falcon_handle_rx_not_ok(rx_queue, event, &rx_ev_pkt_ok,
+					&discard, rx_ev_byte_cnt);
+		checksummed = 0;
+	}
+
+	/* Detect multicast packets that didn't match the filter */
+	rx_ev_mcast_pkt = EFX_QWORD_FIELD(*event, RX_EV_MCAST_PKT);
+	if (rx_ev_mcast_pkt) {
+		unsigned int rx_ev_mcast_hash_match =
+			EFX_QWORD_FIELD(*event, RX_EV_MCAST_HASH_MATCH);
+
+		if (unlikely(!rx_ev_mcast_hash_match))
+			discard = 1;
+	}
+
+	/* Handle received packet */
+	efx_rx_packet(rx_queue, rx_ev_desc_ptr, rx_ev_byte_cnt,
+		      checksummed, discard);
+
+	return rx_ev_q_label;
+}
+
+/* Global events are basically PHY events */
+static void falcon_handle_global_event(struct efx_channel *channel,
+				       efx_qword_t *event)
+{
+	struct efx_nic *efx = channel->efx;
+	int is_phy_event = 0, handled = 0;
+
+	/* Check for interrupt on either port.  Some boards have a
+	 * single PHY wired to the interrupt line for port 1. */
+	if (EFX_QWORD_FIELD(*event, G_PHY0_INTR) ||
+	    EFX_QWORD_FIELD(*event, G_PHY1_INTR) ||
+	    EFX_QWORD_FIELD(*event, XG_PHY_INTR))
+		is_phy_event = 1;
+
+	if ((FALCON_REV(efx) >= FALCON_REV_B0) &&
+	    EFX_OWORD_FIELD(*event, XG_MNT_INTR_B0))
+		is_phy_event = 1;
+
+	if (is_phy_event) {
+		efx->phy_op->clear_interrupt(efx);
+		queue_work(efx->workqueue, &efx->reconfigure_work);
+		handled = 1;
+	}
+
+	if (EFX_QWORD_FIELD_VER(efx, *event, RX_RECOVERY)) {
+		EFX_ERR(efx, "channel %d seen global RX_RESET "
+			"event. Resetting.\n", channel->channel);
+
+		atomic_inc(&efx->errors.rx_reset);
+		efx_schedule_reset(efx, EFX_WORKAROUND_6555(efx) ?
+				   RESET_TYPE_RX_RECOVERY : RESET_TYPE_DISABLE);
+		handled = 1;
+	}
+
+	if (!handled)
+		EFX_ERR(efx, "channel %d unknown global event "
+			EFX_QWORD_FMT "\n", channel->channel,
+			EFX_QWORD_VAL(*event));
+}
+
+static void falcon_handle_driver_event(struct efx_channel *channel,
+				       efx_qword_t *event)
+{
+	struct efx_nic *efx = channel->efx;
+	unsigned int ev_sub_code;
+	unsigned int ev_sub_data;
+
+	ev_sub_code = EFX_QWORD_FIELD(*event, DRIVER_EV_SUB_CODE);
+	ev_sub_data = EFX_QWORD_FIELD(*event, DRIVER_EV_SUB_DATA);
+
+	switch (ev_sub_code) {
+	case TX_DESCQ_FLS_DONE_EV_DECODE:
+		EFX_TRACE(efx, "channel %d TXQ %d flushed\n",
+			  channel->channel, ev_sub_data);
+		EFX_DL_CALLBACK(efx, event, event);
+		break;
+	case RX_DESCQ_FLS_DONE_EV_DECODE:
+		EFX_TRACE(efx, "channel %d RXQ %d flushed\n",
+			  channel->channel, ev_sub_data);
+		EFX_DL_CALLBACK(efx, event, event);
+		break;
+	case EVQ_INIT_DONE_EV_DECODE:
+		EFX_LOG(efx, "channel %d EVQ %d initialised\n",
+			channel->channel, ev_sub_data);
+		break;
+	case SRM_UPD_DONE_EV_DECODE:
+		EFX_TRACE(efx, "channel %d SRAM update done\n",
+			  channel->channel);
+		EFX_DL_CALLBACK(efx, event, event);
+		break;
+	case WAKE_UP_EV_DECODE:
+		EFX_TRACE(efx, "channel %d RXQ %d wakeup event\n",
+			  channel->channel, ev_sub_data);
+		EFX_DL_CALLBACK(efx, event, event);
+		break;
+	case TIMER_EV_DECODE:
+		EFX_TRACE(efx, "channel %d RX queue %d timer expired\n",
+			  channel->channel, ev_sub_data);
+		EFX_DL_CALLBACK(efx, event, event);
+		break;
+	case RX_RECOVERY_EV_DECODE:
+		EFX_ERR(efx, "channel %d seen DRIVER RX_RESET event. "
+			"Resetting.\n", channel->channel);
+
+		atomic_inc(&efx->errors.rx_reset);
+		efx_schedule_reset(efx,
+				   EFX_WORKAROUND_6555(efx) ?
+				   RESET_TYPE_RX_RECOVERY :
+				   RESET_TYPE_DISABLE);
+		break;
+	case RX_DSC_ERROR_EV_DECODE:
+		EFX_ERR(efx, "RX DMA Q %d reports descriptor fetch error."
+			" RX Q %d is disabled.\n", ev_sub_data, ev_sub_data);
+		atomic_inc(&efx->errors.rx_desc_fetch);
+		efx_schedule_reset(efx, RESET_TYPE_RX_DESC_FETCH);
+		break;
+	case TX_DSC_ERROR_EV_DECODE:
+		EFX_ERR(efx, "TX DMA Q %d reports descriptor fetch error."
+			" TX Q %d is disabled.\n", ev_sub_data, ev_sub_data);
+		atomic_inc(&efx->errors.tx_desc_fetch);
+		efx_schedule_reset(efx, RESET_TYPE_TX_DESC_FETCH);
+		break;
+	default:
+		EFX_TRACE(efx, "channel %d unknown driver event code %d "
+			  "data %04x\n", channel->channel, ev_sub_code,
+			  ev_sub_data);
+		EFX_DL_CALLBACK(efx, event, event);
+		break;
+	}
+}
+
+#if defined(EFX_USE_FASTCALL)
+int fastcall falcon_process_eventq(struct efx_channel *channel, int *rx_quota)
+#else
+int falcon_process_eventq(struct efx_channel *channel, int *rx_quota)
+#endif
+{
+	unsigned int read_ptr;
+	efx_qword_t event, *p_event;
+	int ev_code;
+	int rxq;
+	int rxdmaqs = 0;
+
+	read_ptr = channel->eventq_read_ptr;
+
+	do {
+		p_event = falcon_event(channel, read_ptr);
+		event = *p_event;
+
+		if (!falcon_event_present(&event))
+			/* End of events */
+			break;
+
+		EFX_TRACE(channel->efx, "channel %d event is "EFX_QWORD_FMT"\n",
+			  channel->channel, EFX_QWORD_VAL(event));
+
+		/* Clear this event by marking it all ones */
+		EFX_SET_QWORD(*p_event);
+
+		ev_code = EFX_QWORD_FIELD(event, EV_CODE);
+
+		switch (ev_code) {
+		case RX_IP_EV_DECODE:
+			rxq = falcon_handle_rx_event(channel, &event);
+			rxdmaqs |= (1 << rxq);
+			(*rx_quota)--;
+			break;
+		case TX_IP_EV_DECODE:
+			falcon_handle_tx_event(channel, &event);
+			break;
+		case DRV_GEN_EV_DECODE:
+			channel->eventq_magic
+				= EFX_QWORD_FIELD(event, EVQ_MAGIC);
+			EFX_LOG(channel->efx, "channel %d received generated "
+				"event "EFX_QWORD_FMT"\n", channel->channel,
+				EFX_QWORD_VAL(event));
+			break;
+		case GLOBAL_EV_DECODE:
+			falcon_handle_global_event(channel, &event);
+			break;
+		case DRIVER_EV_DECODE:
+			falcon_handle_driver_event(channel, &event);
+			break;
+		default:
+			EFX_ERR(channel->efx, "channel %d unknown event type %d"
+				" (data " EFX_QWORD_FMT ")\n", channel->channel,
+				ev_code, EFX_QWORD_VAL(event));
+		}
+
+		/* Increment read pointer */
+		read_ptr = (read_ptr + 1) & FALCON_EVQ_MASK;
+
+	} while (*rx_quota);
+
+	channel->eventq_read_ptr = read_ptr;
+	return rxdmaqs;
+}
+
+void falcon_set_int_moderation(struct efx_channel *channel)
+{
+	efx_dword_t timer_cmd;
+	struct efx_nic *efx = channel->efx;
+
+	/* Set timer register */
+	if (channel->irq_moderation) {
+		/* Round to resolution supported by hardware.  The value we
+		 * program is based at 0.  So actual interrupt moderation
+		 * achieved is ((x + 1) * res).
+		 */
+		unsigned int res = 5;
+		channel->irq_moderation -= (channel->irq_moderation % res);
+		if (channel->irq_moderation < res)
+			channel->irq_moderation = res;
+		EFX_POPULATE_DWORD_2(timer_cmd,
+				     TIMER_MODE, TIMER_MODE_INT_HLDOFF,
+				     TIMER_VAL,
+				     (channel->irq_moderation / res) - 1);
+	} else {
+		EFX_POPULATE_DWORD_2(timer_cmd,
+				     TIMER_MODE, TIMER_MODE_DIS,
+				     TIMER_VAL, 0);
+	}
+	falcon_writel_page_locked(efx, &timer_cmd, TIMER_CMD_REG_KER,
+				  channel->evqnum);
+
+}
+
+/* Allocate buffer table entries for event queue */
+int falcon_probe_eventq(struct efx_channel *channel)
+{
+	struct efx_nic *efx = channel->efx;
+	struct falcon_nic_data *nic_data = efx->nic_data;
+	unsigned int evq_size;
+	int rc;
+
+	evq_size = FALCON_EVQ_SIZE * sizeof(efx_qword_t);
+	rc = falcon_alloc_special_buffer(efx, &channel->eventq, evq_size);
+	if (rc)
+		return rc;
+
+	nic_data->resources.evq_int_min = max(nic_data->resources.evq_int_min,
+					      (unsigned)channel->evqnum + 1);
+
+	return 0;
+}
+
+int falcon_init_eventq(struct efx_channel *channel)
+{
+	efx_oword_t evq_ptr;
+	struct efx_nic *efx = channel->efx;
+	int rc;
+
+	EFX_LOG(efx, "channel %d event queue in special buffers %d-%d\n",
+		channel->channel, channel->eventq.index,
+		channel->eventq.index + channel->eventq.entries - 1);
+
+	/* Pin event queue buffer */
+	rc = falcon_init_special_buffer(efx, &channel->eventq);
+	if (rc)
+		return rc;
+
+	/* Fill event queue with all ones (i.e. empty events) */
+	memset(channel->eventq.addr, 0xff, channel->eventq.len);
+
+	/* Push event queue to card */
+	EFX_POPULATE_OWORD_3(evq_ptr,
+			     EVQ_EN, 1,
+			     EVQ_SIZE, FALCON_EVQ_ORDER,
+			     EVQ_BUF_BASE_ID, channel->eventq.index);
+	falcon_write_table(efx, &evq_ptr, efx->type->evq_ptr_tbl_base,
+			   channel->evqnum);
+
+	falcon_set_int_moderation(channel);
+
+	return 0;
+}
+
+void falcon_fini_eventq(struct efx_channel *channel)
+{
+	efx_oword_t eventq_ptr;
+	struct efx_nic *efx = channel->efx;
+
+	/* Remove event queue from card */
+	EFX_ZERO_OWORD(eventq_ptr);
+	falcon_write_table(efx, &eventq_ptr, efx->type->evq_ptr_tbl_base,
+			   channel->evqnum);
+
+	/* Unpin event queue */
+	falcon_fini_special_buffer(efx, &channel->eventq);
+}
+
+/* Free buffers backing event queue */
+void falcon_remove_eventq(struct efx_channel *channel)
+{
+	falcon_free_special_buffer(channel->efx, &channel->eventq);
+}
+
+
+/* Generates a test event on the event queue.  A subsequent call to
+ * process_eventq() should pick up the event and place the value of
+ * "magic" into channel->eventq_magic;
+ */
+void falcon_generate_test_event(struct efx_channel *channel, unsigned int magic)
+{
+	efx_qword_t test_event;
+
+	EFX_POPULATE_QWORD_2(test_event,
+			     EV_CODE, DRV_GEN_EV_DECODE,
+			     EVQ_MAGIC, magic);
+	falcon_generate_event(channel, &test_event);
+}
+
+
+/**************************************************************************
+ *
+ * Falcon hardware interrupts
+ * The hardware interrupt handler does very little work; all the event
+ * queue processing is carried out by per-channel tasklets.
+ *
+ **************************************************************************/
+
+/* Enable/disable/generate Falcon interrupts */
+static inline void falcon_interrupts(struct efx_nic *efx, int enabled,
+				     int force)
+{
+	efx_oword_t int_en_reg_ker;
+
+	EFX_POPULATE_OWORD_2(int_en_reg_ker,
+			     KER_INT_KER, force,
+			     DRV_INT_EN_KER, enabled);
+	falcon_write(efx, &int_en_reg_ker, INT_EN_REG_KER);
+}
+
+void falcon_enable_interrupts(struct efx_nic *efx)
+{
+	efx_oword_t int_adr_reg_ker;
+	struct efx_channel *channel;
+
+	/* Zero INT_KER */
+	EFX_ZERO_OWORD(*((efx_oword_t *) efx->irq_status.addr));
+	wmb(); /* Ensure interrupt vector is clear before interrupts enabled */
+
+	/* Program INT_ADR_KER_REG */
+	EFX_POPULATE_OWORD_2(int_adr_reg_ker,
+			     NORM_INT_VEC_DIS_KER, EFX_INT_MODE_USE_MSI(efx),
+			     INT_ADR_KER, efx->irq_status.dma_addr);
+	falcon_write(efx, &int_adr_reg_ker, INT_ADR_REG_KER);
+
+	/* Enable interrupts */
+	falcon_interrupts(efx, 1, 0);
+
+	/* Force processing of all the channels to get the EVQ RPTRs up to
+	   date */
+	efx_for_each_channel_with_interrupt(channel, efx)
+		efx_schedule_channel(channel);
+}
+
+void falcon_disable_interrupts(struct efx_nic *efx)
+{
+	/* Disable interrupts */
+	falcon_interrupts(efx, 0, 0);
+}
+
+/* Generate a Falcon test interrupt
+ * Interrupt must already have been enabled, otherwise nasty things
+ * may happen.
+ */
+void falcon_generate_interrupt(struct efx_nic *efx)
+{
+	falcon_interrupts(efx, 1, 1);
+}
+
+/* Acknowledge a legacy interrupt from Falcon
+ *
+ * This acknowledges a legacy (not MSI) interrupt via INT_ACK_KER_REG.
+ *
+ * Due to SFC bug 3706 (silicon revision <=A1) reads can be duplicated in the
+ * BIU. Interrupt acknowledge is read sensitive so must write instead
+ * (then read to ensure the BIU collector is flushed)
+ *
+ * NB most hardware supports MSI interrupts
+ */
+static inline void falcon_irq_ack_a1(struct efx_nic *efx)
+{
+	efx_dword_t reg;
+
+	EFX_POPULATE_DWORD_1(reg, INT_ACK_DUMMY_DATA, 0xb7eb7e);
+	falcon_writel(efx, &reg, INT_ACK_REG_KER_A1);
+	falcon_readl(efx, &reg, WORK_AROUND_BROKEN_PCI_READS_REG_KER_A1);
+}
+
+/* Process a fatal interrupt
+ * Disable bus mastering ASAP and schedule a reset
+ */
+static irqreturn_t falcon_fatal_interrupt(struct efx_nic *efx)
+{
+	efx_oword_t *int_ker = (efx_oword_t *) efx->irq_status.addr;
+	efx_oword_t fatal_intr;
+	int error, mem_perr;
+	static int n_int_errors;
+
+	falcon_read(efx, &fatal_intr, FATAL_INTR_REG_KER);
+	error = EFX_OWORD_FIELD(fatal_intr, INT_KER_ERROR);
+
+	EFX_ERR(efx, "SYSTEM ERROR " EFX_OWORD_FMT " status "
+		EFX_OWORD_FMT ": %s\n", EFX_OWORD_VAL(*int_ker),
+		EFX_OWORD_VAL(fatal_intr),
+		error ? "disabling bus mastering" : "no recognised error");
+	if (error == 0)
+		goto out;
+
+	/* If this is a memory parity error dump which blocks are offending */
+	mem_perr = EFX_OWORD_FIELD(fatal_intr, MEM_PERR_INT_KER);
+	if (mem_perr) {
+		efx_oword_t reg;
+		falcon_read(efx, &reg, MEM_STAT_REG_KER);
+		EFX_ERR(efx, "SYSTEM ERROR: memory parity error "
+			EFX_OWORD_FMT "\n", EFX_OWORD_VAL(reg));
+	}
+
+	/* Disable DMA bus mastering on both devices */
+	pci_disable_device(efx->pci_dev);
+	if (efx->type->is_dual_func)
+		pci_disable_device(efx->pci_dev2);
+
+	if (++n_int_errors < FALCON_MAX_INT_ERRORS) {
+		EFX_ERR(efx, "SYSTEM ERROR - reset scheduled\n");
+		efx_schedule_reset(efx, RESET_TYPE_INT_ERROR);
+	} else {
+		EFX_ERR(efx, "SYSTEM ERROR - max number of errors seen."
+			"NIC will be disabled\n");
+		efx_schedule_reset(efx, RESET_TYPE_DISABLE);
+	}
+out:
+	return IRQ_HANDLED;
+}
+
+/* Handle a legacy interrupt from Falcon
+ * Acknowledges the interrupt and schedule event queue processing.
+ *
+ * This routine must guarantee not to touch the hardware when
+ * interrupts are disabled, to allow for correct semantics of
+ * efx_suspend() and efx_resume().
+ */
+#if !defined(EFX_HAVE_IRQ_HANDLER_REGS)
+static irqreturn_t falcon_legacy_interrupt_b0(int irq, void *dev_id)
+#else
+static irqreturn_t falcon_legacy_interrupt_b0(int irq, void *dev_id,
+					      struct pt_regs *regs
+					      __attribute__ ((unused)))
+#endif
+{
+	struct efx_nic *efx = (struct efx_nic *)dev_id;
+	efx_oword_t *int_ker = (efx_oword_t *) efx->irq_status.addr;
+	struct efx_channel *channel;
+	efx_dword_t reg;
+	u32 queues;
+	int syserr;
+
+	/* Read the ISR which also ACKs the interrupts */
+	falcon_readl(efx, &reg, INT_ISR0_B0);
+	queues = EFX_EXTRACT_DWORD(reg, 0, 31);
+
+	/* Check to see if we have a serious error condition */
+	syserr = EFX_OWORD_FIELD(*int_ker, FATAL_INT);
+	if (unlikely(syserr))
+		return falcon_fatal_interrupt(efx);
+
+	if (queues == 0)
+		return IRQ_NONE;
+
+	efx->last_irq_cpu = raw_smp_processor_id();
+	EFX_TRACE(efx, "IRQ %d on CPU %d status " EFX_DWORD_FMT "\n",
+		  irq, raw_smp_processor_id(), EFX_DWORD_VAL(reg));
+
+	/* Schedule processing of any interrupting queues */
+	channel = &efx->channel[0];
+	while (queues) {
+		if (queues & 0x01)
+			efx_schedule_channel(channel);
+		channel++;
+		queues >>= 1;
+	}
+
+	return IRQ_HANDLED;
+}
+
+
+#if !defined(EFX_HAVE_IRQ_HANDLER_REGS)
+static irqreturn_t falcon_legacy_interrupt_a1(int irq, void *dev_id)
+#else
+static irqreturn_t falcon_legacy_interrupt_a1(int irq, void *dev_id,
+					      struct pt_regs *regs
+					      __attribute__ ((unused)))
+#endif
+{
+	struct efx_nic *efx = (struct efx_nic *)dev_id;
+	efx_oword_t *int_ker = (efx_oword_t *) efx->irq_status.addr;
+	struct efx_channel *channel;
+	int syserr;
+	int queues;
+
+	/* Check to see if this is our interrupt.  If it isn't, we
+	 * exit without having touched the hardware.
+	 */
+	if (unlikely(EFX_OWORD_IS_ZERO(*int_ker))) {
+		EFX_TRACE(efx, "IRQ %d on CPU %d not for me\n", irq,
+			  raw_smp_processor_id());
+		return IRQ_NONE;
+	}
+	efx->last_irq_cpu = raw_smp_processor_id();
+	EFX_TRACE(efx, "IRQ %d on CPU %d status " EFX_OWORD_FMT "\n",
+		  irq, raw_smp_processor_id(), EFX_OWORD_VAL(*int_ker));
+
+	/* Check to see if we have a serious error condition */
+	syserr = EFX_OWORD_FIELD(*int_ker, FATAL_INT);
+	if (unlikely(syserr))
+		return falcon_fatal_interrupt(efx);
+
+	/* Determine interrupting queues, clear interrupt status
+	 * register and acknowledge the device interrupt.
+	 */
+	BUILD_BUG_ON(INT_EVQS_WIDTH > EFX_MAX_CHANNELS);
+	queues = EFX_OWORD_FIELD(*int_ker, INT_EVQS);
+	EFX_ZERO_OWORD(*int_ker);
+	wmb(); /* Ensure the vector is cleared before interrupt ack */
+	falcon_irq_ack_a1(efx);
+
+	/* Schedule processing of any interrupting queues */
+	channel = &efx->channel[0];
+	while (queues) {
+		if (queues & 0x01)
+			efx_schedule_channel(channel);
+		channel++;
+		queues >>= 1;
+	}
+
+	return IRQ_HANDLED;
+}
+
+/* Handle an MSI interrupt from Falcon
+ *
+ * Handle an MSI hardware interrupt.  This routine schedules event
+ * queue processing.  No interrupt acknowledgement cycle is necessary.
+ * Also, we never need to check that the interrupt is for us, since
+ * MSI interrupts cannot be shared.
+ *
+ * This routine must guarantee not to touch the hardware when
+ * interrupts are disabled, to allow for correct semantics of
+ * efx_suspend() and efx_resume().
+ */
+#if !defined(EFX_HAVE_IRQ_HANDLER_REGS)
+static irqreturn_t falcon_msi_interrupt(int irq, void *dev_id)
+#else
+static irqreturn_t falcon_msi_interrupt(int irq, void *dev_id,
+					struct pt_regs *regs
+					__attribute__ ((unused)))
+#endif
+{
+	struct efx_channel *channel = (struct efx_channel *)dev_id;
+	struct efx_nic *efx = channel->efx;
+	efx_oword_t *int_ker = (efx_oword_t *) efx->irq_status.addr;
+	int syserr;
+
+	efx->last_irq_cpu = raw_smp_processor_id();
+	EFX_TRACE(efx, "IRQ %d on CPU %d status " EFX_OWORD_FMT "\n",
+		  irq, raw_smp_processor_id(), EFX_OWORD_VAL(*int_ker));
+
+	/* Check to see if we have a serious error condition */
+	syserr = EFX_OWORD_FIELD(*int_ker, FATAL_INT);
+	if (unlikely(syserr))
+		return falcon_fatal_interrupt(efx);
+
+	/* Schedule processing of the channel */
+	efx_schedule_channel(channel);
+
+	return IRQ_HANDLED;
+}
+
+
+/* Setup RSS indirection table.
+ * This maps from the hash value of the packet to RXQ
+ */
+static void falcon_setup_rss_indir_table(struct efx_nic *efx)
+{
+	int i = 0;
+	unsigned long offset;
+	unsigned long flags __attribute__ ((unused));
+	efx_dword_t dword;
+
+	if (FALCON_REV(efx) < FALCON_REV_B0)
+		return;
+
+	for (offset = RX_RSS_INDIR_TBL_B0;
+	     offset < RX_RSS_INDIR_TBL_B0 + 0x800;
+	     offset += 0x10) {
+		EFX_POPULATE_DWORD_1(dword, RX_RSS_INDIR_ENT_B0,
+				     i % efx->rss_queues);
+		falcon_writel(efx, &dword, offset);
+		i++;
+	}
+}
+
+/* Hook interrupt handler(s)
+ * Try MSI and then legacy interrupts.
+ */
+int falcon_init_interrupt(struct efx_nic *efx)
+{
+	struct efx_channel *channel;
+	int rc;
+
+	if (!EFX_INT_MODE_USE_MSI(efx)) {
+		irq_handler_t handler;
+		if (FALCON_REV(efx) >= FALCON_REV_B0)
+			handler = falcon_legacy_interrupt_b0;
+		else
+			handler = falcon_legacy_interrupt_a1;
+
+		rc = request_irq(efx->legacy_irq, handler, IRQF_SHARED,
+				 efx->name, efx);
+		if (rc) {
+			EFX_ERR(efx, "failed to hook legacy IRQ %d\n",
+				efx->pci_dev->irq);
+			goto fail1;
+		}
+		return 0;
+	}
+
+	/* Hook MSI or MSI-X interrupt */
+	efx_for_each_channel_with_interrupt(channel, efx) {
+		rc = request_irq(channel->irq, falcon_msi_interrupt,
+				 IRQF_PROBE_SHARED, /* Not shared */
+				 efx->name, channel);
+		if (rc) {
+			EFX_ERR(efx, "failed to hook IRQ %d\n", channel->irq);
+			goto fail2;
+		}
+	}
+
+	return 0;
+
+ fail2:
+	efx_for_each_channel_with_interrupt(channel, efx)
+		free_irq(channel->irq, channel);
+ fail1:
+	return rc;
+}
+
+void falcon_fini_interrupt(struct efx_nic *efx)
+{
+	struct efx_channel *channel;
+	efx_oword_t reg;
+
+	/* Disable MSI/MSI-X interrupts */
+	efx_for_each_channel_with_interrupt(channel, efx)
+		if (channel->irq)
+			free_irq(channel->irq, channel);
+
+	/* ACK legacy interrupt */
+	if (FALCON_REV(efx) >= FALCON_REV_B0)
+		falcon_read(efx, &reg, INT_ISR0_B0);
+	else
+		falcon_irq_ack_a1(efx);
+
+	/* Disable legacy interrupt */
+	if (efx->legacy_irq)
+		free_irq(efx->legacy_irq, efx);
+}
+
+/**************************************************************************
+ *
+ * EEPROM/flash
+ *
+ **************************************************************************
+ */
+
+/* Wait for SPI command completion */
+static int falcon_spi_wait(struct efx_nic *efx)
+{
+	efx_oword_t reg;
+	int cmd_en, timer_active;
+	int count;
+
+	count = 0;
+	do {
+		falcon_read(efx, &reg, EE_SPI_HCMD_REG_KER);
+		cmd_en = EFX_OWORD_FIELD(reg, EE_SPI_HCMD_CMD_EN);
+		timer_active = EFX_OWORD_FIELD(reg, EE_WR_TIMER_ACTIVE);
+		if (!cmd_en && !timer_active)
+			return 0;
+		udelay(10);
+	} while (++count < 10000); /* wait upto 100msec */
+	EFX_ERR(efx, "timed out waiting for SPI\n");
+	return -ETIMEDOUT;
+}
+
+static int
+falcon_spi_read(const struct efx_spi_device *spi, struct efx_nic *efx,
+		unsigned int command, int address, void *data, unsigned int len)
+{
+	int addressed = (address >= 0);
+	efx_oword_t reg;
+	int rc;
+
+	/* Input validation */
+	if (len > FALCON_SPI_MAX_LEN)
+		return -EINVAL;
+
+	/* Acquire SPI lock */
+	mutex_lock(&efx->spi_lock);
+
+	/* Check SPI not currently being accessed */
+	rc = falcon_spi_wait(efx);
+	if (rc)
+		goto out;
+
+	/* Program address register, if we have an address */
+	if (addressed) {
+		EFX_POPULATE_OWORD_1(reg, EE_SPI_HADR_ADR, address);
+		falcon_write(efx, &reg, EE_SPI_HADR_REG_KER);
+	}
+
+	/* Issue read command */
+	EFX_POPULATE_OWORD_7(reg,
+			     EE_SPI_HCMD_CMD_EN, 1,
+			     EE_SPI_HCMD_SF_SEL, spi->device_id,
+			     EE_SPI_HCMD_DABCNT, len,
+			     EE_SPI_HCMD_READ, EE_SPI_READ,
+			     EE_SPI_HCMD_DUBCNT, 0,
+			     EE_SPI_HCMD_ADBCNT,
+			     (addressed ? spi->addr_len : 0),
+			     EE_SPI_HCMD_ENC, command);
+	falcon_write(efx, &reg, EE_SPI_HCMD_REG_KER);
+
+	/* Wait for read to complete */
+	rc = falcon_spi_wait(efx);
+	if (rc)
+		goto out;
+
+	/* Read data */
+	falcon_read(efx, &reg, EE_SPI_HDATA_REG_KER);
+	memcpy(data, &reg, len);
+
+ out:
+	/* Release SPI lock */
+	mutex_unlock(&efx->spi_lock);
+
+	return rc;
+}
+
+static int
+falcon_spi_write(const struct efx_spi_device *spi, struct efx_nic *efx,
+		 unsigned int command, int address, const void *data,
+		 unsigned int len)
+{
+	int addressed = (address >= 0);
+	efx_oword_t reg;
+	int rc;
+
+	/* Input validation */
+	if (len > (addressed ? efx_spi_write_limit(spi, address)
+		   : FALCON_SPI_MAX_LEN))
+		return -EINVAL;
+
+	/* Acquire SPI lock */
+	mutex_lock(&efx->spi_lock);
+
+	/* Check SPI not currently being accessed */
+	rc = falcon_spi_wait(efx);
+	if (rc)
+		goto out;
+
+	/* Program address register, if we have an address */
+	if (addressed) {
+		EFX_POPULATE_OWORD_1(reg, EE_SPI_HADR_ADR, address);
+		falcon_write(efx, &reg, EE_SPI_HADR_REG_KER);
+	}
+
+	/* Program data register, if we have data */
+	if (data) {
+		memcpy(&reg, data, len);
+		falcon_write(efx, &reg, EE_SPI_HDATA_REG_KER);
+	}
+
+	/* Issue write command */
+	EFX_POPULATE_OWORD_7(reg,
+			     EE_SPI_HCMD_CMD_EN, 1,
+			     EE_SPI_HCMD_SF_SEL, spi->device_id,
+			     EE_SPI_HCMD_DABCNT, len,
+			     EE_SPI_HCMD_READ, EE_SPI_WRITE,
+			     EE_SPI_HCMD_DUBCNT, 0,
+			     EE_SPI_HCMD_ADBCNT,
+			     (addressed ? spi->addr_len : 0),
+			     EE_SPI_HCMD_ENC, command);
+	falcon_write(efx, &reg, EE_SPI_HCMD_REG_KER);
+
+	/* Wait for write to complete */
+	rc = falcon_spi_wait(efx);
+	if (rc)
+		goto out;
+
+ out:
+	/* Release SPI lock */
+	mutex_unlock(&efx->spi_lock);
+
+	return rc;
+}
+
+/**************************************************************************
+ *
+ * MAC wrapper
+ *
+ **************************************************************************
+ */
+void falcon_drain_tx_fifo(struct efx_nic *efx)
+{
+	efx_oword_t temp;
+	efx_oword_t mcast_reg0;
+	efx_oword_t mcast_reg1;
+	int count;
+
+	if (FALCON_REV(efx) < FALCON_REV_B0)
+		return;
+
+	falcon_read(efx, &temp, MAC0_CTRL_REG_KER);
+	/* There is no point in draining more than once */
+	if (EFX_OWORD_FIELD(temp, TXFIFO_DRAIN_EN_B0))
+		return;
+
+	/* MAC stats will fail whilst the TX fifo is draining. Serialise
+	 * the drain sequence with the statistics fetch */
+	spin_lock(&efx->stats_lock);
+
+	EFX_SET_OWORD_FIELD(temp, TXFIFO_DRAIN_EN_B0, 1);
+	falcon_write(efx, &temp, MAC0_CTRL_REG_KER);
+
+	falcon_read(efx, &mcast_reg0, MAC_MCAST_HASH_REG0_KER);
+	falcon_read(efx, &mcast_reg1, MAC_MCAST_HASH_REG1_KER);
+
+	/* Reset the MAC and EM block. */
+	falcon_read(efx, &temp, GLB_CTL_REG_KER);
+	EFX_SET_OWORD_FIELD(temp, RST_XGTX, 1);
+	EFX_SET_OWORD_FIELD(temp, RST_XGRX, 1);
+	EFX_SET_OWORD_FIELD(temp, RST_EM, 1);
+	falcon_write(efx, &temp, GLB_CTL_REG_KER);
+
+	count = 0;
+	while (1) {
+		falcon_read(efx, &temp, GLB_CTL_REG_KER);
+		if (!EFX_OWORD_FIELD(temp, RST_XGTX) &&
+		    !EFX_OWORD_FIELD(temp, RST_XGRX) &&
+		    !EFX_OWORD_FIELD(temp, RST_EM)) {
+			EFX_LOG(efx, "Completed MAC reset after %d loops\n",
+				count);
+			break;
+		}
+		if (count > 20) {
+			EFX_ERR(efx, "MAC reset failed\n");
+			break;
+		}
+		count++;
+		udelay(10);
+	}
+
+	spin_unlock(&efx->stats_lock);
+
+	/* Restore the multicast hash registers. */
+	falcon_write(efx, &mcast_reg0, MAC_MCAST_HASH_REG0_KER);
+	falcon_write(efx, &mcast_reg1, MAC_MCAST_HASH_REG1_KER);
+
+	/* If we've reset the EM block and the link is up, then
+	 * we'll have to kick the XAUI link so the PHY can recover */
+	if (efx->link_up && EFX_IS10G(efx) && EFX_WORKAROUND_5147(efx))
+		falcon_reset_xaui(efx);
+}
+
+void falcon_deconfigure_mac_wrapper(struct efx_nic *efx)
+{
+	struct falcon_nic_data *nic_data = efx->nic_data;
+	efx_oword_t temp;
+	int changing_loopback;
+
+	if (FALCON_REV(efx) < FALCON_REV_B0)
+		return;
+
+	/* Isolate the MAC -> RX */
+	falcon_read(efx, &temp, RX_CFG_REG_KER);
+	EFX_SET_OWORD_FIELD(temp, RX_INGR_EN_B0, 0);
+	falcon_write(efx, &temp, RX_CFG_REG_KER);
+
+	/* Synchronise the EM block against any loopback mode changes by
+	 * draining the TX fifo and resetting. */
+	changing_loopback = (efx->loopback_mode != nic_data->old_loopback_mode);
+	nic_data->old_loopback_mode = efx->loopback_mode;
+	if (changing_loopback || !efx->link_up)
+		falcon_drain_tx_fifo(efx);
+}
+
+void falcon_reconfigure_mac_wrapper(struct efx_nic *efx)
+{
+	efx_oword_t reg;
+	int link_speed;
+	unsigned int tx_fc;
+
+	if (efx->link_options & GM_LPA_10000)
+		link_speed = 0x3;
+	else if (efx->link_options & GM_LPA_1000)
+		link_speed = 0x2;
+	else if (efx->link_options & GM_LPA_100)
+		link_speed = 0x1;
+	else
+		link_speed = 0x0;
+	/* MAC_LINK_STATUS controls MAC backpressure but doesn't work
+	 * as advertised.  Disable to ensure packets are not
+	 * indefinitely held and TX queue can be flushed at any point
+	 * while the link is down.
+	 */
+	EFX_POPULATE_OWORD_5(reg,
+			     MAC_XOFF_VAL, 0xffff /* max pause time */,
+			     MAC_BCAD_ACPT, 1,
+			     MAC_UC_PROM, efx->promiscuous,
+			     MAC_LINK_STATUS, 1, /* always set */
+			     MAC_SPEED, link_speed);
+	/* On B0, MAC backpressure can be disabled and packets get
+	 * discarded. */
+	if (FALCON_REV(efx) >= FALCON_REV_B0) {
+		EFX_SET_OWORD_FIELD(reg, TXFIFO_DRAIN_EN_B0,
+				    !efx->link_up);
+	}
+
+	falcon_write(efx, &reg, MAC0_CTRL_REG_KER);
+
+	/*
+	 * Transmission of pause frames when RX crosses the threshold is
+	 * covered by RX_XOFF_MAC_EN and XM_TX_CFG_REG:XM_FCNTL.
+	 *
+	 * Action on receipt of pause frames is controller by XM_DIS_FCNTL
+	 */
+	tx_fc = (efx->flow_control & EFX_FC_TX) ? 1 : 0;
+	falcon_read(efx, &reg, RX_CFG_REG_KER);
+	EFX_SET_OWORD_FIELD_VER(efx, reg, RX_XOFF_MAC_EN, tx_fc);
+
+	/* Unisolate the MAC -> RX */
+	if (FALCON_REV(efx) >= FALCON_REV_B0)
+		EFX_SET_OWORD_FIELD(reg, RX_INGR_EN_B0, 1);
+	falcon_write(efx, &reg, RX_CFG_REG_KER);
+}
+
+int falcon_dma_stats(struct efx_nic *efx, unsigned int done_offset)
+{
+	efx_oword_t reg;
+	u32 *dma_done;
+	int i;
+
+	if (disable_dma_stats)
+		return 0;
+
+	/* Statistics fetch will fail if the MAC is in TX drain */
+	if (FALCON_REV(efx) >= FALCON_REV_B0) {
+		efx_oword_t temp;
+		falcon_read(efx, &temp, MAC0_CTRL_REG_KER);
+		if (EFX_OWORD_FIELD(temp, TXFIFO_DRAIN_EN_B0))
+			return 0;
+	}
+
+	/* Clear completion pointer */
+	dma_done = (efx->stats_buffer.addr + done_offset);
+	*dma_done = FALCON_STATS_NOT_DONE;
+	wmb(); /* ensure done flag is clear */
+
+	/* Initiate DMA transfer of stats */
+	EFX_POPULATE_OWORD_2(reg,
+			     MAC_STAT_DMA_CMD, 1,
+			     MAC_STAT_DMA_ADR,
+			     efx->stats_buffer.dma_addr);
+	falcon_write(efx, &reg, MAC0_STAT_DMA_REG_KER);
+
+	/* Wait for transfer to complete */
+	for (i = 0; i < 400; i++) {
+		if (*(volatile u32 *)dma_done == FALCON_STATS_DONE)
+			return 0;
+		udelay(10);
+	}
+
+	if (EFX_WORKAROUND_8419(efx)) {
+		disable_dma_stats = 1;
+		EFX_INFO(efx, "MAC stats DMA disabled\n");
+	} else {
+		EFX_ERR(efx, "timed out waiting for statistics\n");
+	}
+
+	return -ETIMEDOUT;
+}
+
+/**************************************************************************
+ *
+ * PHY access via GMII
+ *
+ **************************************************************************
+ */
+
+/* Use the top bit of the MII PHY id to indicate the PHY type
+ * (1G/10G), with the remaining bits as the actual PHY id.
+ *
+ * This allows us to avoid leaking information from the mii_if_info
+ * structure into other data structures.
+ */
+#define FALCON_PHY_ID_ID_WIDTH  EFX_WIDTH(MD_PRT_DEV_ADR)
+#define FALCON_PHY_ID_ID_MASK   ((1 << FALCON_PHY_ID_ID_WIDTH) - 1)
+#define FALCON_PHY_ID_WIDTH     (FALCON_PHY_ID_ID_WIDTH + 1)
+#define FALCON_PHY_ID_MASK      ((1 << FALCON_PHY_ID_WIDTH) - 1)
+#define FALCON_PHY_ID_10G       (1 << (FALCON_PHY_ID_WIDTH - 1))
+
+
+/* Packing the clause 45 port and device fields into a single value */
+#define MD_PRT_ADR_COMP_LBN   (MD_PRT_ADR_LBN - MD_DEV_ADR_LBN)
+#define MD_PRT_ADR_COMP_WIDTH  MD_PRT_ADR_WIDTH
+#define MD_DEV_ADR_COMP_LBN    0
+#define MD_DEV_ADR_COMP_WIDTH  MD_DEV_ADR_WIDTH
+
+
+/* Wait for GMII access to complete */
+static int falcon_gmii_wait(struct efx_nic *efx)
+{
+	efx_dword_t md_stat;
+	int count;
+
+	for (count = 0; count < 1000; count++) {	/* wait upto 10ms */
+		falcon_readl(efx, &md_stat, MD_STAT_REG_KER);
+		if (EFX_DWORD_FIELD(md_stat, MD_BSY) == 0) {
+			if (EFX_DWORD_FIELD(md_stat, MD_LNFL) != 0 ||
+			    EFX_DWORD_FIELD(md_stat, MD_BSERR) != 0) {
+				EFX_ERR(efx, "error from GMII access "
+					EFX_DWORD_FMT"\n",
+					EFX_DWORD_VAL(md_stat));
+				return -EIO;
+			}
+			return 0;
+		}
+		udelay(10);
+	}
+	EFX_ERR(efx, "timed out waiting for GMII\n");
+	return -ETIMEDOUT;
+}
+
+/* Writes a GMII register of a PHY connected to Falcon using MDIO. */
+static void falcon_mdio_write(struct net_device *net_dev, int phy_id,
+			      int addr, int value)
+{
+	struct efx_nic *efx = (struct efx_nic *)net_dev->priv;
+	unsigned int phy_id2 = phy_id & FALCON_PHY_ID_ID_MASK;
+	unsigned int phy_10g = phy_id & FALCON_PHY_ID_10G;
+	efx_oword_t reg;
+
+	/* The 'generic' prt/dev packing in mdio_10g.h is conveniently
+	 * chosen so that the only current user, Falcon, can take the
+	 * packed value and use them directly.
+	 * Fail to build if this assumption is broken.
+	 */
+	BUILD_BUG_ON(FALCON_PHY_ID_10G != MDIO45_XPRT_ID_IS10G);
+	BUILD_BUG_ON(FALCON_PHY_ID_ID_WIDTH != MDIO45_PRT_DEV_WIDTH);
+	BUILD_BUG_ON(MD_PRT_ADR_COMP_LBN != MDIO45_PRT_ID_COMP_LBN);
+	BUILD_BUG_ON(MD_DEV_ADR_COMP_LBN != MDIO45_DEV_ID_COMP_LBN);
+
+	if (phy_id2 == PHY_ADDR_INVALID)
+		return;
+
+	/* See falcon_mdio_read for an explanation. */
+	if (EFX_ISCLAUSE45(efx) && !phy_10g) {
+		int mmd = ffs(efx->phy_op->mmds) - 1;
+		EFX_TRACE(efx, "Fixing erroneous clause22 write\n");
+		phy_id2 = mdio_clause45_pack(phy_id2, mmd)
+			& FALCON_PHY_ID_ID_MASK;
+		phy_10g = 1;
+	}
+
+	EFX_REGDUMP(efx, "writing GMII %d register %02x with %04x\n", phy_id,
+		    addr, value);
+
+	/* Obtain PHY lock */
+	spin_lock_bh(&efx->phy_lock);
+
+	/* Check MII not currently being accessed */
+	if (falcon_gmii_wait(efx) != 0)
+		goto out;
+
+	/* Write the address/ID register */
+	EFX_POPULATE_OWORD_1(reg, MD_PHY_ADR, addr);
+	falcon_write(efx, &reg, MD_PHY_ADR_REG_KER);
+
+	if (phy_10g)
+		EFX_POPULATE_OWORD_1(reg, MD_PRT_DEV_ADR, phy_id2);
+	else
+		/* MDIO clause 22 */
+		EFX_POPULATE_OWORD_2(reg,
+				     MD_PRT_ADR, phy_id2,
+				     MD_DEV_ADR, addr);
+	falcon_write(efx, &reg, MD_ID_REG_KER);
+
+	/* Write data */
+	EFX_POPULATE_OWORD_1(reg, MD_TXD, value);
+	falcon_write(efx, &reg, MD_TXD_REG_KER);
+
+	EFX_POPULATE_OWORD_2(reg,
+			     MD_WRC, 1,
+			     MD_GC, (phy_10g ? 0 : 1));
+	falcon_write(efx, &reg, MD_CS_REG_KER);
+
+	/* Wait for data to be written */
+	if (falcon_gmii_wait(efx) != 0) {
+		/* Abort the write operation */
+		EFX_POPULATE_OWORD_2(reg,
+				     MD_WRC, 0,
+				     MD_GC, 1);
+		falcon_write(efx, &reg, MD_CS_REG_KER);
+		udelay(10);
+	}
+
+ out:
+	/* Release PHY lock */
+	spin_unlock_bh(&efx->phy_lock);
+}
+
+/* Reads a GMII register from a PHY connected to Falcon.  If no value
+ * could be read, -1 will be returned. */
+static int falcon_mdio_read(struct net_device *net_dev, int phy_id, int addr)
+{
+	struct efx_nic *efx = (struct efx_nic *)net_dev->priv;
+	unsigned int phy_addr = phy_id & FALCON_PHY_ID_ID_MASK;
+	unsigned int phy_10g = phy_id & FALCON_PHY_ID_10G;
+	efx_oword_t reg;
+	int value = -1;
+	unsigned long flags __attribute__ ((unused));
+
+	if (phy_addr == PHY_ADDR_INVALID)
+		return -1;
+
+	/* Our PHY code knows whether it needs to talk clause 22(1G) or 45(10G)
+	 * but the generic Linux code does not make any distinction or have
+	 * any state for this.
+	 * We spot the case where someone tried to talk 22 to a 45 PHY and
+	 * redirect the request to the lowest numbered MMD as a clause45
+	 * request. This is enough to allow simple queries like id and link
+	 * state to succeed. TODO: We may need to do more in future.
+	 */
+	if (EFX_ISCLAUSE45(efx) && !phy_10g) {
+		int mmd = ffs(efx->phy_op->mmds) - 1;
+		EFX_TRACE(efx, "Fixing erroneous clause22 read\n");
+		phy_addr = mdio_clause45_pack(phy_addr, mmd)
+			& FALCON_PHY_ID_ID_MASK;
+		phy_10g = 1;
+	}
+
+	/* Obtain PHY lock */
+	spin_lock_bh(&efx->phy_lock);
+
+	/* Check MII not currently being accessed */
+	if (falcon_gmii_wait(efx) != 0)
+		goto out;
+
+	if (!phy_10g) {
+		/* Write the address registers */
+		EFX_POPULATE_OWORD_2(reg,
+				     MD_PRT_ADR, phy_addr,
+				     MD_DEV_ADR, addr);
+		falcon_write(efx, &reg, MD_ID_REG_KER);
+		/* Request data to be read */
+		EFX_POPULATE_OWORD_2(reg,
+				     MD_RIC, 1,
+				     MD_GC, 1);
+	} else {
+		EFX_POPULATE_OWORD_1(reg, MD_PHY_ADR, addr);
+		falcon_write(efx, &reg, MD_PHY_ADR_REG_KER);
+
+		EFX_POPULATE_OWORD_1(reg, MD_PRT_DEV_ADR, phy_addr);
+		falcon_write(efx, &reg, MD_ID_REG_KER);
+
+		/* Request data to be read */
+		EFX_POPULATE_OWORD_2(reg,
+				     MD_RDC, 1,
+				     MD_GC, 0);
+	}
+	falcon_write(efx, &reg, MD_CS_REG_KER);
+
+	/* Wait for data to become available */
+	value = falcon_gmii_wait(efx);
+	if (value == 0) {
+		falcon_read(efx, &reg, MD_RXD_REG_KER);
+		value = EFX_OWORD_FIELD(reg, MD_RXD);
+		EFX_REGDUMP(efx, "read from GMII %d register %02x, got %04x\n",
+			    phy_id, addr, value);
+	} else {
+		/* Abort the read operation */
+		EFX_POPULATE_OWORD_2(reg,
+				     MD_RIC, 0,
+				     MD_GC, 1);
+		falcon_write(efx, &reg, MD_CS_REG_KER);
+
+		EFX_LOG(efx, "read from GMII 0x%x register %02x, got "
+			"error %d\n", phy_id, addr, value);
+	}
+
+ out:
+	/* Release PHY lock */
+	spin_unlock_bh(&efx->phy_lock);
+
+	return value;
+}
+
+static void falcon_init_mdio(struct mii_if_info *gmii)
+{
+	gmii->mdio_read = falcon_mdio_read;
+	gmii->mdio_write = falcon_mdio_write;
+	gmii->phy_id_mask = FALCON_PHY_ID_MASK;
+	gmii->reg_num_mask = ((1 << EFX_WIDTH(MD_DEV_ADR)) - 1);
+}
+
+static int falcon_probe_gmac_port(struct efx_nic *efx)
+{
+	struct efx_phy_operations *phy_op = efx->phy_op;
+
+	efx->mac_op = &falcon_gmac_operations;
+	efx->loopback_modes = LOOPBACKS_1G_INTERNAL | phy_op->loopbacks;
+	efx->startup_loopbacks = ((1 << LOOPBACK_MAC) |
+				  (1 << phy_op->startup_loopback));
+	return 0;
+}
+
+static int falcon_probe_xmac_port(struct efx_nic *efx)
+{
+	struct efx_phy_operations *phy_op = efx->phy_op;
+
+	efx->mac_op = &falcon_xmac_operations;
+
+	/* The Falcon B0 FPGA only supports XGMII loopback */
+	if (FALCON_REV(efx) >= FALCON_REV_B0 && !efx->is_asic)
+		efx->loopback_modes = (1 << LOOPBACK_XGMII);
+	else
+		efx->loopback_modes = LOOPBACKS_10G_INTERNAL;
+	efx->loopback_modes |= phy_op->loopbacks;
+
+	efx->startup_loopbacks = ((1 << LOOPBACK_XGMII) |
+				  (1 << phy_op->startup_loopback));
+	return 0;
+}
+
+static int falcon_probe_phy(struct efx_nic *efx)
+{
+	switch (efx->phy_type) {
+	case PHY_TYPE_1G_ALASKA:
+		efx->phy_op = &alaska_phy_operations;
+		break;
+	case PHY_TYPE_10XPRESS:
+		efx->phy_op = &falcon_tenxpress_phy_ops;
+		break;
+	case PHY_TYPE_NONE:
+		efx->phy_op = &falcon_null_phy_ops;
+		break;
+	case PHY_TYPE_XFP:
+		efx->phy_op = &falcon_xfp_phy_ops;
+		break;
+	case PHY_TYPE_CX4_RTMR:
+		efx->phy_op = &falcon_txc_phy_ops;
+		break;
+	case PHY_TYPE_PM8358:
+		efx->phy_op = &falcon_pm8358_phy_ops;
+		break;
+	default:
+		EFX_ERR(efx, "Unknown PHY type %d\n",
+			efx->phy_type);
+		return -1;
+	}
+	return 0;
+}
+
+/* This call is responsible for hooking in the MAC and PHY operations */
+int falcon_probe_port(struct efx_nic *efx)
+{
+	int rc;
+
+	/* Hook in PHY operations table */
+	rc = falcon_probe_phy(efx);
+	if (rc)
+		return rc;
+
+	/* Hook in MAC operations table */
+	if (EFX_IS10G(efx))
+		rc = falcon_probe_xmac_port(efx);
+	else
+		rc = falcon_probe_gmac_port(efx);
+	if (rc)
+		return rc;
+
+	EFX_LOG(efx, "created port using %cMAC\n",
+		EFX_IS10G(efx) ? 'X' : 'G');
+
+	/* Set up GMII structure for PHY */
+	efx->mii.supports_gmii = 1;
+	falcon_init_mdio(&efx->mii);
+
+	/* Hardware flow ctrl. FalconA RX FIFO too small for pause generation */
+	if (FALCON_REV(efx) >= FALCON_REV_B0)
+		efx->flow_control = EFX_FC_RX | EFX_FC_TX;
+	else
+		efx->flow_control = EFX_FC_RX;
+
+	/* Allocate buffer for stats */
+	rc = falcon_alloc_buffer(efx, &efx->stats_buffer,
+				 FALCON_MAC_STATS_SIZE);
+	if (rc)
+		return rc;
+	EFX_LOG(efx, "stats buffer at %llx (virt %p phys %lx)\n",
+		(unsigned long long)efx->stats_buffer.dma_addr,
+		efx->stats_buffer.addr,
+		virt_to_phys(efx->stats_buffer.addr));
+
+	return 0;
+}
+
+void falcon_remove_port(struct efx_nic *efx)
+{
+	/* Free stats buffer */
+	falcon_free_buffer(efx, &efx->stats_buffer);
+}
+
+/**************************************************************************
+ *
+ * Multicast filtering
+ *
+ **************************************************************************
+ */
+
+void falcon_set_multicast_hash(struct efx_nic *efx)
+{
+	union efx_multicast_hash falcon_mc_hash;
+
+	/* Broadcast packets go through the multicast hash filter.
+	 * ether_crc_le() of the broadcast address is 0xbe2612ff
+	 * so we always add bit 0xff to the mask we are given.
+	 */
+	memcpy(&falcon_mc_hash, &efx->multicast_hash, sizeof(falcon_mc_hash));
+	set_bit_le(0xff, (void *)&falcon_mc_hash);
+
+	falcon_write(efx, &falcon_mc_hash.oword[0], MAC_MCAST_HASH_REG0_KER);
+	falcon_write(efx, &falcon_mc_hash.oword[1], MAC_MCAST_HASH_REG1_KER);
+}
+
+/**************************************************************************
+ *
+ * Device reset
+ *
+ **************************************************************************
+ */
+
+static int falcon_clear_b0_memories(struct efx_nic *efx)
+{
+	/* Need to clear memories after a reset. On B0 we can do this
+	 * via the net function.
+	 */
+	int rc = 0, offset;
+	efx_oword_t blanko;
+	efx_dword_t blankd;
+	unsigned long membase_phys, membase_len;
+	void __iomem *membase_orig;
+	unsigned long flags __attribute__ ((unused));
+
+	EFX_ZERO_OWORD(blanko);
+	EFX_ZERO_DWORD(blankd);
+	membase_orig = efx->membase;
+	membase_phys = pci_resource_start(efx->pci_dev, efx->type->mem_bar);
+
+	for (offset = RX_FILTER_TBL0;
+	     offset < RX_RSS_INDIR_TBL_B0;
+	     offset += 0x10)
+		falcon_write(efx, &blanko, offset);
+
+	/* Clear RSS indirection table */
+	for (offset = RX_RSS_INDIR_TBL_B0;
+	     offset < RX_RSS_INDIR_TBL_B0 + 0x800;
+	     offset += 0x10)
+		/* Clear 6 bits every 16 bytes */
+		falcon_writel(efx, &blankd, offset);
+
+	/* Need to split this into several mappings so MSI-X table and PBA
+	 * never get mapped
+	 */
+	membase_phys = membase_phys + 0x2800000;
+	membase_len = 0x3000000 - 0x2800000;
+
+	efx->membase = ioremap_nocache(membase_phys, membase_len);
+	if (efx->membase == NULL) {
+		EFX_ERR(efx, "could not map memory BAR %d at %lx+%lx\n",
+			efx->type->mem_bar, membase_phys, membase_len);
+		rc = -ENOMEM;
+		goto out;
+	}
+	/* Clear the buffer table.  The first 7/8 of it is a duplicate
+	 * of the mapping at 0x800000 and must be accessed 2 DWORDs at
+	 * a time.  The final 1/8 must be accessed 4 DWORDs at a time.
+	 * We make sure to obey both rules at the same time.
+	 */
+	for (offset = 0; offset < membase_len; offset += 0x10) {
+		spin_lock_irqsave(&efx->biu_lock, flags);
+		_falcon_writel(efx, 0, offset + 0x0);
+		wmb();
+		_falcon_writel(efx, 0, offset + 0x4);
+		wmb();
+		_falcon_writel(efx, 0, offset + 0x8);
+		wmb();
+		_falcon_writel(efx, 0, offset + 0xc);
+		mmiowb();
+		spin_unlock_irqrestore(&efx->biu_lock, flags);
+	}
+
+	iounmap(efx->membase);
+
+out:
+	/* Restore */
+	efx->membase = membase_orig;
+
+	return rc;
+}
+
+
+/* Resets NIC to known state.  This routine must be called in process
+ * context and is allowed to sleep. */
+int falcon_reset_hw(struct efx_nic *efx, enum reset_type method)
+{
+	efx_oword_t glb_ctl_reg_ker;
+	int rc;
+
+	EFX_LOG(efx, "performing %s hardware reset\n", RESET_TYPE(method));
+
+	/* Initiate device reset */
+	if (method == RESET_TYPE_WORLD) {
+		/* Save PCI config space */
+		rc = pci_save_state(efx->pci_dev);
+		if (rc) {
+			EFX_ERR(efx, "failed to backup PCI state of primary "
+				"function prior to hardware reset\n");
+			goto fail1;
+		}
+		if (efx->type->is_dual_func) {
+			rc = pci_save_state(efx->pci_dev2);
+			if (rc) {
+				EFX_ERR(efx, "failed to backup PCI state of "
+					"secondary function prior to "
+					"hardware reset\n");
+				goto fail2;
+			}
+		}
+
+		EFX_POPULATE_OWORD_2(glb_ctl_reg_ker,
+				     EXT_PHY_RST_DUR, 0x7,
+				     SWRST, 1);
+	} else {
+		int reset_phy = (method == RESET_TYPE_INVISIBLE ?
+				 EXCLUDE_FROM_RESET : 0);
+
+		EFX_POPULATE_OWORD_7(glb_ctl_reg_ker,
+				     EXT_PHY_RST_CTL, reset_phy,
+				     PCIE_CORE_RST_CTL, EXCLUDE_FROM_RESET,
+				     PCIE_NSTCK_RST_CTL, EXCLUDE_FROM_RESET,
+				     PCIE_SD_RST_CTL, EXCLUDE_FROM_RESET,
+				     EE_RST_CTL, EXCLUDE_FROM_RESET,
+				     EXT_PHY_RST_DUR, 0x7 /* 10ms */,
+				     SWRST, 1);
+	}
+	falcon_write(efx, &glb_ctl_reg_ker, GLB_CTL_REG_KER);
+
+	/* Wait for 50ms for the chip to come out of reset */
+	EFX_LOG(efx, "waiting for hardware reset\n");
+	schedule_timeout_uninterruptible(HZ / 20);
+
+	/* Restore PCI configuration if needed */
+	if (method == RESET_TYPE_WORLD) {
+		if (efx->type->is_dual_func) {
+			rc = pci_restore_state(efx->pci_dev2);
+			if (rc) {
+				EFX_ERR(efx, "failed to restore PCI config for "
+					"the secondary function\n");
+				goto fail3;
+			}
+		}
+		rc = pci_restore_state(efx->pci_dev);
+		if (rc) {
+			EFX_ERR(efx, "failed to restore PCI config for the "
+				"primary function\n");
+			goto fail4;
+		}
+		EFX_LOG(efx, "successfully restored PCI config\n");
+	}
+
+	/* Assert that reset complete */
+	falcon_read(efx, &glb_ctl_reg_ker, GLB_CTL_REG_KER);
+	if (EFX_OWORD_FIELD(glb_ctl_reg_ker, SWRST) != 0) {
+		rc = -ETIMEDOUT;
+		EFX_ERR(efx, "timed out waiting for hardware reset\n");
+		goto fail5;
+	}
+	EFX_LOG(efx, "hardware reset complete\n");
+
+	if (EFX_WORKAROUND_8202(efx)) {
+		rc = falcon_clear_b0_memories(efx);
+		if (rc)
+			goto fail6;
+	}
+
+	return 0;
+
+	/* pci_save_state() and pci_restore_state() MUST be called in pairs */
+fail2:
+fail3:
+	pci_restore_state(efx->pci_dev);
+	/* fall-thru */
+fail1:
+fail4:
+fail5:
+fail6:
+	return rc;
+}
+
+/* Zeroes out the SRAM contents.  This routine must be called in
+ * process context and is allowed to sleep.
+ */
+static int falcon_reset_sram(struct efx_nic *efx)
+{
+	efx_oword_t srm_cfg_reg_ker, gpio_cfg_reg_ker;
+	int count, onchip, sram_cfg_val;
+
+	/* Set the SRAM wake/sleep GPIO appropriately. */
+	onchip = (efx->external_sram_cfg == SRM_NB_BSZ_ONCHIP_ONLY);
+	falcon_read(efx, &gpio_cfg_reg_ker, GPIO_CTL_REG_KER);
+	EFX_SET_OWORD_FIELD(gpio_cfg_reg_ker, GPIO1_OEN, 1);
+	EFX_SET_OWORD_FIELD(gpio_cfg_reg_ker, GPIO1_OUT, onchip ? 1 : 0);
+	falcon_write(efx, &gpio_cfg_reg_ker, GPIO_CTL_REG_KER);
+
+	/* Initiate SRAM reset */
+	sram_cfg_val = (efx->external_sram_cfg == SRM_NB_BSZ_ONCHIP_ONLY) ?
+		0 : efx->external_sram_cfg;
+
+	EFX_POPULATE_OWORD_2(srm_cfg_reg_ker,
+			     SRAM_OOB_BT_INIT_EN, 1,
+			     SRM_NUM_BANKS_AND_BANK_SIZE, sram_cfg_val);
+	falcon_write(efx, &srm_cfg_reg_ker, SRM_CFG_REG_KER);
+
+	/* Wait for SRAM reset to complete */
+	count = 0;
+	do {
+		EFX_LOG(efx, "waiting for SRAM reset (attempt %d)...\n", count);
+
+		/* SRAM reset is slow; expect around 16ms */
+		schedule_timeout_uninterruptible(HZ / 50);
+
+		/* Check for reset complete */
+		falcon_read(efx, &srm_cfg_reg_ker, SRM_CFG_REG_KER);
+		if (!EFX_OWORD_FIELD(srm_cfg_reg_ker, SRAM_OOB_BT_INIT_EN)) {
+			EFX_LOG(efx, "SRAM reset complete\n");
+
+			return 0;
+		}
+	} while (++count < 20);	/* wait upto 0.4 sec */
+
+	EFX_ERR(efx, "timed out waiting for SRAM reset\n");
+	return -ETIMEDOUT;
+}
+
+static void falcon_spi_device_init(struct efx_spi_device **spi_device_ret,
+				   unsigned int device_id, u32 device_type)
+{
+	struct efx_spi_device *spi_device;
+
+	if (device_type != 0) {
+		spi_device = kmalloc(sizeof(*spi_device), GFP_KERNEL);
+		spi_device->device_id = device_id;
+		spi_device->size =
+			1 << SPI_DEV_TYPE_FIELD(device_type, SPI_DEV_TYPE_SIZE);
+		spi_device->addr_len =
+			SPI_DEV_TYPE_FIELD(device_type, SPI_DEV_TYPE_ADDR_LEN);
+		spi_device->munge_address = (spi_device->size == 1 << 9 &&
+					     spi_device->addr_len == 1);
+		spi_device->erase_command =
+			SPI_DEV_TYPE_FIELD(device_type, SPI_DEV_TYPE_ERASE_CMD);
+		spi_device->erase_size =
+			1 << SPI_DEV_TYPE_FIELD(device_type,
+						SPI_DEV_TYPE_ERASE_SIZE);
+		spi_device->block_size =
+			1 << SPI_DEV_TYPE_FIELD(device_type,
+						SPI_DEV_TYPE_BLOCK_SIZE);
+		spi_device->read = falcon_spi_read;
+		spi_device->write = falcon_spi_write;
+	} else {
+		spi_device = NULL;
+	}
+
+	kfree(*spi_device_ret);
+	*spi_device_ret = spi_device;
+}
+
+/* Extract non-volatile configuration */
+static int falcon_probe_nvconfig(struct efx_nic *efx)
+{
+	int rc;
+	struct falcon_nvconfig *nvconfig;
+	struct efx_spi_device *spi;
+	size_t offset, len;
+	int magic_num, struct_ver, board_rev, onchip_sram;
+
+	nvconfig = kmalloc(sizeof(*nvconfig), GFP_KERNEL);
+
+	/* Read the whole configuration structure into memory.  It's
+	 * in Falcon's boot device, which may be either flash or
+	 * EEPROM, but if both are present Falcon prefers flash.  The
+	 * boot device is always too large for 9-bit addressing, so we
+	 * don't have to munge commands.
+	 */
+	spi = efx->spi_flash ? efx->spi_flash : efx->spi_eeprom;
+	for (offset = 0; offset < sizeof(*nvconfig); offset += len) {
+		len = min(sizeof(*nvconfig) - offset,
+			  (size_t) FALCON_SPI_MAX_LEN);
+		rc = falcon_spi_read(spi, efx, SPI_READ,
+				     NVCONFIG_BASE + offset,
+				     (char *)nvconfig + offset, len);
+		if (rc)
+			goto out;
+	}
+
+	/* Read the MAC addresses */
+	memcpy(efx->mac_address, nvconfig->mac_address[0], ETH_ALEN);
+
+	/* Read the board configuration. */
+	magic_num = le16_to_cpu(nvconfig->board_magic_num);
+	struct_ver = le16_to_cpu(nvconfig->board_struct_ver);
+
+	if (magic_num != NVCONFIG_BOARD_MAGIC_NUM || struct_ver < 2) {
+		EFX_ERR(efx, "Non volatile memory bad magic=%x ver=%x "
+			"therefore using defaults\n", magic_num, struct_ver);
+		efx->phy_type = PHY_TYPE_NONE;
+		efx->mii.phy_id = PHY_ADDR_INVALID;
+		board_rev = 0;
+		onchip_sram = 1;
+
+	} else {
+		struct falcon_nvconfig_board_v2 *v2 = &nvconfig->board_v2;
+		struct falcon_nvconfig_board_v3 *v3 = &nvconfig->board_v3;
+
+		efx->phy_type = v2->port0_phy_type;
+		efx->mii.phy_id = v2->port0_phy_addr;
+		board_rev = le16_to_cpu(v2->board_revision);
+		onchip_sram = EFX_OWORD_FIELD(nvconfig->nic_stat_reg,
+					      ONCHIP_SRAM);
+
+		if (struct_ver >= 3) {
+			__le32 fl = v3->spi_device_type[EE_SPI_FLASH];
+			__le32 ee = v3->spi_device_type[EE_SPI_EEPROM];
+			falcon_spi_device_init(&efx->spi_flash, EE_SPI_FLASH,
+					       le32_to_cpu(fl));
+			falcon_spi_device_init(&efx->spi_eeprom, EE_SPI_EEPROM,
+					       le32_to_cpu(ee));
+		}
+	}
+
+	EFX_LOG(efx, "PHY is %s(%d) phy_id %d\n",
+		PHY_TYPE(efx), efx->phy_type,
+		efx->mii.phy_id);
+
+	efx_set_board_info(efx, board_rev);
+
+	/* Read the SRAM configuration.  The register is initialised
+	 * automatically but might may been reset since boot.
+	 */
+	if (onchip_sram) {
+		efx->external_sram_cfg = SRM_NB_BSZ_ONCHIP_ONLY;
+	} else {
+		efx->external_sram_cfg =
+		    EFX_OWORD_FIELD(nvconfig->srm_cfg_reg,
+				    SRM_NUM_BANKS_AND_BANK_SIZE);
+		WARN_ON(efx->external_sram_cfg == SRM_NB_BSZ_RESERVED);
+		/* Replace invalid setting with the smallest defaults */
+		if (efx->external_sram_cfg == SRM_NB_BSZ_DEFAULT)
+			efx->external_sram_cfg = SRM_NB_BSZ_1BANKS_2M;
+	}
+	EFX_LOG(efx, "external_sram_cfg=%d (>=0 is external)\n",
+		efx->external_sram_cfg);
+
+ out:
+	kfree(nvconfig);
+	return rc;
+}
+
+/* Looks at available SRAM resources and silicon revision, and works out
+ * how many queues we can support, and where things like descriptor caches
+ * should live. */
+static int falcon_dimension_resources(struct efx_nic *efx)
+{
+	unsigned buffer_entry_bytes, internal_dcs_entries, dcs;
+	struct falcon_nic_data *nic_data = efx->nic_data;
+	struct efx_dl_falcon_resources *res = &nic_data->resources;
+
+	/* Fill out the driverlink resource list */
+	res->hdr.type = EFX_DL_FALCON_RESOURCES;
+	res->biu_lock = &efx->biu_lock;
+	efx->dl_info = &res->hdr;
+
+	/* This is set to 16 for a good reason.  In summary, if larger than
+	 * 16, the descriptor cache holds more than a default socket
+	 * buffer's worth of packets (for UDP we can only have at most one
+	 * socket buffer's worth outstanding).  This combined with the fact
+	 * that we only get 1 TX event per descriptor cache means the NIC
+	 * goes idle.
+	 * 16 gives us up to 256 TXQs on Falcon B in internal-SRAM mode,
+	 * and up to 512 on Falcon A.
+	 */
+	nic_data->tx_dc_entries = 16;
+
+	/* Set the RX descriptor cache size.  Values 16, 32 and 64 are
+	 * supported (8 won't work).  Bigger is better, especially on B
+	 * silicon.
+	 */
+	nic_data->rx_dc_entries = descriptor_cache_size;
+	dcs = ffs(nic_data->rx_dc_entries);
+	if ((dcs < 5) || (dcs > 7) ||
+	    ((1 << (dcs - 1)) != nic_data->rx_dc_entries)) {
+		EFX_ERR(efx, "bad descriptor_cache_size=%d (dcs=%d)\n",
+			nic_data->rx_dc_entries, dcs);
+		return -EINVAL;
+	}
+
+	/* NB. The minimum values get increased as this driver initialises
+	 * its resources, so this should prevent any overlap.
+	 */
+	switch (FALCON_REV(efx)) {
+	case FALCON_REV_A1:
+		res->rxq_min = res->txq_min = 16;
+		res->evq_int_min = res->evq_int_max = 4;
+		res->evq_timer_min = 5;
+		res->evq_timer_max = 4096;
+		internal_dcs_entries = 8192;
+		break;
+	case FALCON_REV_B0:
+	default:
+		res->rxq_min = res->txq_min = res->evq_int_min = 0;
+		res->evq_int_max = 64;
+		res->evq_timer_min = 64;
+		res->evq_timer_max = 4096;
+		internal_dcs_entries = 4096;
+		break;
+	}
+
+	buffer_entry_bytes = 8;
+
+	if (efx->external_sram_cfg == SRM_NB_BSZ_ONCHIP_ONLY) {
+		res->rxq_max = internal_dcs_entries / nic_data->rx_dc_entries;
+		res->txq_max = internal_dcs_entries / nic_data->tx_dc_entries;
+		/* Prog model says 8K entries for buffer table in internal
+		 * mode.  But does this not depend on full/half mode?
+		 */
+		res->buffer_table_max = 8192;
+		nic_data->tx_dc_base = 0x130000;
+		nic_data->rx_dc_base = 0x100000;
+	} else {
+		unsigned sram_bytes, vnic_bytes, max_vnics, n_vnics;
+
+		/* Determine how much SRAM we have to play with.  We have
+		 * to fit buffer table and descriptor caches in.
+		 */
+		switch (efx->external_sram_cfg) {
+		case SRM_NB_BSZ_1BANKS_2M:
+		default:
+			sram_bytes = 2 * 1024 * 1024;
+			break;
+		case SRM_NB_BSZ_1BANKS_4M:
+		case SRM_NB_BSZ_2BANKS_4M:
+			sram_bytes = 4 * 1024 * 1024;
+			break;
+		case SRM_NB_BSZ_1BANKS_8M:
+		case SRM_NB_BSZ_2BANKS_8M:
+			sram_bytes = 8 * 1024 * 1024;
+			break;
+		case SRM_NB_BSZ_2BANKS_16M:
+			sram_bytes = 16 * 1024 * 1024;
+			break;
+		}
+		/* For each VNIC allow at least 512 buffer table entries
+		 * and descriptor cache for an rxq and txq.  Buffer table
+		 * space for evqs and dmaqs is relatively trivial, so not
+		 * considered in this calculation.
+		 */
+		vnic_bytes = (512 * buffer_entry_bytes
+			      + nic_data->rx_dc_entries * 8
+			      + nic_data->tx_dc_entries * 8);
+		max_vnics = sram_bytes / vnic_bytes;
+		for (n_vnics = 1; n_vnics < res->evq_timer_min + max_vnics;)
+			n_vnics *= 2;
+		res->rxq_max = n_vnics;
+		res->txq_max = n_vnics;
+
+		dcs = n_vnics * nic_data->tx_dc_entries * 8;
+		nic_data->tx_dc_base = sram_bytes - dcs;
+		dcs = n_vnics * nic_data->rx_dc_entries * 8;
+		nic_data->rx_dc_base = nic_data->tx_dc_base - dcs;
+		res->buffer_table_max = nic_data->rx_dc_base / 8;
+	}
+
+	if (efx->type->is_dual_func)
+		res->flags |= EFX_DL_FALCON_DUAL_FUNC;
+
+	if (EFX_INT_MODE_USE_MSI(efx))
+		res->flags |= EFX_DL_FALCON_USE_MSI;
+
+	return 0;
+}
+
+/* Probe the NIC variant (revision, ASIC vs FPGA, function count, port
+ * count, port speed).  Set workaround and feature flags accordingly.
+ */
+static int falcon_probe_nic_variant(struct efx_nic *efx)
+{
+	efx_oword_t altera_build;
+
+	falcon_read(efx, &altera_build, ALTERA_BUILD_REG_KER);
+	efx->is_asic = EFX_OWORD_FIELD(altera_build, VER_ALL) == 0;
+
+#if !defined(EFX_USE_PCI_DEV_REVISION)
+	{
+		int rc;
+		rc = pci_read_config_byte(efx->pci_dev, PCI_CLASS_REVISION,
+					  &efx->revision);
+		if (rc)
+			return rc;
+	}
+#endif
+	switch (FALCON_REV(efx)) {
+	case FALCON_REV_A0:
+	case 0xff:
+		EFX_ERR(efx, "Falcon rev A0 not supported\n");
+		return -ENODEV;
+
+	case FALCON_REV_A1:{
+		efx_oword_t nic_stat;
+
+		falcon_read(efx, &nic_stat, NIC_STAT_REG);
+
+		if (!efx->is_asic) {
+			EFX_ERR(efx, "Falcon rev A1 FPGA not supported\n");
+			return -ENODEV;
+		}
+		if (EFX_OWORD_FIELD(nic_stat, STRAP_PCIE) == 0) {
+			EFX_ERR(efx, "Falcon rev A1 PCI-X not supported\n");
+			return -ENODEV;
+		}
+		efx->is_10g = EFX_OWORD_FIELD(nic_stat, STRAP_10G);
+		efx->silicon_rev = "falcon/a1";
+		break;
+	}
+
+	case FALCON_REV_B0:{
+		efx->is_10g = 1;
+		efx->silicon_rev = "falcon/b0";
+		break;
+	}
+
+	default:
+		EFX_ERR(efx, "Unknown Falcon rev %d\n", FALCON_REV(efx));
+		return -ENODEV;
+	}
+
+	return 0;
+}
+
+/* Probe all SPI devices on the NIC */
+static void falcon_probe_spi_devices(struct efx_nic *efx)
+{
+	efx_oword_t nic_stat, gpio_ctl, ee_vpd_cfg;
+	unsigned int has_flash, has_eeprom, boot_is_external;
+
+	falcon_read(efx, &gpio_ctl, GPIO_CTL_REG_KER);
+	falcon_read(efx, &nic_stat, NIC_STAT_REG);
+	falcon_read(efx, &ee_vpd_cfg, EE_VPD_CFG_REG_KER);
+
+	has_flash = EFX_OWORD_FIELD(nic_stat, SF_PRST);
+	has_eeprom = EFX_OWORD_FIELD(nic_stat, EE_PRST);
+	boot_is_external = EFX_OWORD_FIELD(gpio_ctl, BOOTED_USING_NVDEVICE);
+
+	if (has_flash) {
+		u32 flash_device_type;
+
+		if (flash_type == -1) {
+			/* Default flash SPI device: Atmel AT25F1024
+			 * 128 KB, 24-bit address, 32 KB erase block,
+			 * 256 B write block
+			 */
+			flash_device_type =
+				(17 << SPI_DEV_TYPE_SIZE_LBN)
+				| (3 << SPI_DEV_TYPE_ADDR_LEN_LBN)
+				| (0x52 << SPI_DEV_TYPE_ERASE_CMD_LBN)
+				| (15 << SPI_DEV_TYPE_ERASE_SIZE_LBN)
+				| (8 << SPI_DEV_TYPE_BLOCK_SIZE_LBN);
+		} else {
+			flash_device_type = flash_type;
+		}
+
+		falcon_spi_device_init(&efx->spi_flash, EE_SPI_FLASH,
+				       flash_device_type);
+
+		if (!boot_is_external) {
+			/* Disable VPD and set clock dividers to safe
+			 * values for initial programming.
+			 */
+			EFX_LOG(efx, "Booted from internal ASIC settings;"
+				" setting SPI config\n");
+			EFX_POPULATE_OWORD_3(ee_vpd_cfg, EE_VPD_EN, 0,
+					     /* 125 MHz / 7 ~= 20 MHz */
+					     EE_SF_CLOCK_DIV, 7,
+					     /* 125 MHz / 63 ~= 2 MHz */
+					     EE_EE_CLOCK_DIV, 63);
+			falcon_write(efx, &ee_vpd_cfg, EE_VPD_CFG_REG_KER);
+		}
+	}
+
+	if (has_eeprom) {
+		u32 eeprom_device_type;
+
+		/* eeprom_type may be -1 (default) for automatic detection,
+		 * 0 or 1 to select the default or large EEPROM, or
+		 * some larger number to specify the precise configuration
+		 */
+		if (eeprom_type == -1 || eeprom_type <= 1) {
+			/* If it has no flash, it must have a large EEPROM
+			 * for chip config; otherwise check whether 9-bit
+			 * addressing is used for VPD configuration
+			 */
+			if (eeprom_type == 0 ||
+			    (eeprom_type == -1 && has_flash &&
+			     (!boot_is_external ||
+			      EFX_OWORD_FIELD(ee_vpd_cfg,
+					      EE_VPD_EN_AD9_MODE)))) {
+				/* Default SPI device: Atmel AT25040 or similar
+				 * 512 B, 9-bit address, 8 B write block
+				 */
+				eeprom_device_type =
+					(9 << SPI_DEV_TYPE_SIZE_LBN)
+					| (1 << SPI_DEV_TYPE_ADDR_LEN_LBN)
+					| (3 << SPI_DEV_TYPE_BLOCK_SIZE_LBN);
+			} else {
+				/* "Large" SPI device: Atmel AT25640 or similar
+				 * 8 KB, 16-bit address, 32 B write block
+				 */
+				eeprom_device_type =
+					(13 << SPI_DEV_TYPE_SIZE_LBN)
+					| (2 << SPI_DEV_TYPE_ADDR_LEN_LBN)
+					| (5 << SPI_DEV_TYPE_BLOCK_SIZE_LBN);
+			}
+		} else {
+			eeprom_device_type = eeprom_type;
+		}
+
+		falcon_spi_device_init(&efx->spi_eeprom, EE_SPI_EEPROM,
+				       eeprom_device_type);
+	}
+
+	EFX_LOG(efx, "flash is %s, EEPROM is %s\n",
+		(has_flash ? "present" : "absent"),
+		(has_eeprom ? "present" : "absent"));
+}
+
+static void falcon_remove_spi_devices(struct efx_nic *efx)
+{
+	kfree(efx->spi_eeprom);
+	efx->spi_eeprom = NULL;
+	kfree(efx->spi_flash);
+	efx->spi_flash = NULL;
+}
+
+#ifdef CONFIG_SFC_DEBUGFS
+
+/* Generate a hardware revision string */
+int falcon_debugfs_read_hardware_desc(struct seq_file *file, void *data)
+{
+	struct efx_nic *efx = data;
+	efx_oword_t altera_build;
+	int major, minor, build;
+	int rc, len;
+
+	if (efx->is_asic) {
+		rc = seq_puts(file, "Falcon ASIC");
+	} else {
+		falcon_read(efx, &altera_build, ALTERA_BUILD_REG_KER);
+
+		major = EFX_OWORD_FIELD(altera_build, VER_MAJOR);
+		minor = EFX_OWORD_FIELD(altera_build, VER_MINOR);
+		build = EFX_OWORD_FIELD(altera_build, VER_BUILD);
+		rc = seq_printf(file, "Falcon FPGA v%x.%x.%x",
+				major, minor, build);
+	}
+	len = rc;
+
+	switch (FALCON_REV(efx)) {
+	case FALCON_REV_A1:
+		rc = seq_puts(file, " rev A1 ");
+		break;
+	case FALCON_REV_B0:
+		rc = seq_puts(file, " rev B0 ");
+		break;
+	default:
+		rc = seq_puts(file, " rev ?? ");
+		break;
+	}
+	len += rc;
+
+	rc = seq_printf(file, "%s %s\n",
+			efx->is_10g ? "10G" : "1G", PHY_TYPE(efx));
+	len += rc;
+
+	return rc < 0 ? rc : len;
+}
+
+#endif /* CONFIG_SFC_DEBUGFS */
+
+int falcon_probe_nic(struct efx_nic *efx)
+{
+	struct falcon_nic_data *nic_data;
+	int rc;
+
+	/* Initialise I2C interface state */
+	efx->i2c.efx = efx;
+	efx->i2c.op = &falcon_i2c_bit_operations;
+	efx->i2c.sda = 1;
+	efx->i2c.scl = 1;
+
+	/* Determine number of ports etc. */
+	rc = falcon_probe_nic_variant(efx);
+	if (rc)
+		goto fail1;
+
+	/* Probe secondary function if expected */
+	if (efx->type->is_dual_func) {
+		struct pci_dev *dev = pci_dev_get(efx->pci_dev);
+
+		while ((dev = pci_get_device(EFX_VENDID_SFC, FALCON_A_S_DEVID,
+					     dev))) {
+			if (dev->bus == efx->pci_dev->bus &&
+			    dev->devfn == efx->pci_dev->devfn + 1) {
+				efx->pci_dev2 = dev;
+				break;
+			}
+		}
+		if (!efx->pci_dev2) {
+			EFX_ERR(efx, "failed to find secondary function\n");
+			rc = -ENODEV;
+			goto fail2;
+		}
+	}
+
+	/* Now we can reset the NIC */
+	rc = falcon_reset_hw(efx, RESET_TYPE_ALL);
+	if (rc) {
+		EFX_ERR(efx, "failed to reset NIC\n");
+		goto fail3;
+	}
+
+	/* Allocate memory for INT_KER */
+	rc = falcon_alloc_buffer(efx, &efx->irq_status, sizeof(efx_oword_t));
+	if (rc)
+		goto fail4;
+	BUG_ON(efx->irq_status.dma_addr & 0x0f);
+
+	EFX_LOG(efx, "INT_KER at %llx (virt %p phys %lx)\n",
+		(unsigned long long)efx->irq_status.dma_addr,
+		efx->irq_status.addr, virt_to_phys(efx->irq_status.addr));
+
+	/* Determine attached SPI devices */
+	falcon_probe_spi_devices(efx);
+
+	/* Read in the non-volatile configuration */
+	rc = falcon_probe_nvconfig(efx);
+	if (rc)
+		goto fail5;
+
+	if (!efx->is_10g && efx->phy_type != PHY_TYPE_1G_ALASKA) {
+		/* Actually using 1G port, not 10G port */
+		efx->phy_type = PHY_TYPE_1G_ALASKA;
+		efx->mii.phy_id = 2;
+	}
+
+	/* Decide how many resources we can allocate, to ourselves
+	 * and to driverlink clients */
+	nic_data = kzalloc(sizeof(*nic_data), GFP_KERNEL);
+	efx->nic_data = (void *) nic_data;
+
+	rc = falcon_dimension_resources(efx);
+	if (rc)
+		goto fail6;
+
+	return 0;
+
+ fail6:
+	kfree(nic_data);
+	efx->nic_data = efx->dl_info = NULL;
+ fail5:
+	falcon_remove_spi_devices(efx);
+	falcon_free_buffer(efx, &efx->irq_status);
+ fail4:
+	/* fall-thru */
+ fail3:
+	if (efx->pci_dev2) {
+		pci_dev_put(efx->pci_dev2);
+		efx->pci_dev2 = NULL;
+	}
+ fail2:
+	/* fall-thru */
+ fail1:
+	return rc;
+}
+
+static int falcon_check_power_limit(struct efx_nic *efx)
+{
+	int pciecap_offset = pci_find_capability(efx->pci_dev, PCI_CAP_ID_EXP);
+	u32 pcie_devcap;
+	unsigned val, scale;
+	int rc;
+
+	if (!pciecap_offset)
+		return -EIO;
+	rc = pci_read_config_dword(efx->pci_dev,
+				   (pciecap_offset + PCI_EXP_DEVCAP),
+				   &pcie_devcap);
+	if (rc)
+		return rc;
+
+	val = ((pcie_devcap & PCI_EXP_DEVCAP_PWR_VAL) >>
+	       PCI_EXP_DEVCAP_PWR_VAL_LBN);
+	scale = ((pcie_devcap & PCI_EXP_DEVCAP_PWR_SCL) >>
+		 PCI_EXP_DEVCAP_PWR_SCL_LBN);
+
+	/* Re-scale to milliwatts if necessary */
+	while (scale != 3) {
+		val *= 10;
+		scale++;
+	}
+
+	if (val != 0 && efx->board_info.mwatts > val) {
+		EFX_ERR(efx, "board needs %d mW but only %d mW available\n",
+			efx->board_info.mwatts, val);
+		return -EIO;
+	}
+
+	return 0;
+}
+
+static void falcon_init_ack_repl_timer(struct efx_nic *efx, int num_lanes)
+{
+	unsigned tlp_size;
+	efx_dword_t pcie_ack_rpl_reg;
+	efx_dword_t pcie_ack_freq_reg;
+	efx_dword_t pcie_ctrl_stat_reg;
+	u16 pcie_devicectrl;
+	int lut_index, tlp_size_decoded;
+	int current_replay, expected_replay;
+	int current_ack_timer, current_ack_freq;
+
+	static struct efx_tlp_ack_factor {
+		int tlp;
+		int replay[4]; /* 0=1x, 1=2x, 3=4x, 4=8x (see pcie docs) */
+	} tlp_ack_factor_lut[4] = {
+		{ 128,  { 421, 257, 174, 166 } },
+		{ 256,  { 689, 391, 241, 225 } },
+		{ 512,  { 903, 498, 295, 193 } },
+		{ 1024, { 1670, 881, 487, 290 } }
+	};
+	struct efx_tlp_ack_factor *tlp_ack_factor;
+
+	/* Get TLP size */
+	falcon_pcie_core_read_reg(efx, PCIE_CORE_ADDR_PCIE_DEVICE_CTRL_STAT,
+				  &pcie_ctrl_stat_reg);
+	pcie_devicectrl = (u16) EFX_EXTRACT_DWORD(pcie_ctrl_stat_reg, 0, 15);
+	tlp_size = ((PCI_EXP_DEVCTL_PAYLOAD & pcie_devicectrl) >>
+		    ffs(PCI_EXP_DEVCTL_PAYLOAD));
+	EFX_WARN_ON_PARANOID(tlp_size > 3); /* => 1024 bytes */
+	tlp_ack_factor = &tlp_ack_factor_lut[tlp_size & 0x3];
+	tlp_size_decoded = tlp_ack_factor->tlp;
+
+	/* Get actual ack & actual and expected replay settings */
+	falcon_pcie_core_read_reg(efx, PCIE_CORE_ADDR_ACK_RPL_TIMER,
+				  &pcie_ack_rpl_reg);
+	current_replay = EFX_DWORD_FIELD(pcie_ack_rpl_reg, PCIE_CORE_RPL_TL);
+	current_ack_timer = EFX_DWORD_FIELD(pcie_ack_rpl_reg,
+					    PCIE_CORE_ACK_TL);
+
+	lut_index = ffs(num_lanes) - 1;
+	expected_replay = tlp_ack_factor->replay[lut_index & 0x3];
+
+	falcon_pcie_core_read_reg(efx, PCIE_CORE_ADDR_ACK_FREQ,
+				  &pcie_ack_freq_reg);
+	current_ack_freq = EFX_DWORD_FIELD(pcie_ack_freq_reg,
+					   PCIE_CORE_ACK_FREQ);
+
+	EFX_LOG(efx, "pcie x%d tlp=%d replay_reg=" EFX_DWORD_FMT " { ack=%d "
+		"current_replay=%d expected_replay=%d } ack_reg="
+		EFX_DWORD_FMT " { current_freq=%d expected_freq=%d }\n",
+		num_lanes, tlp_size_decoded,
+		EFX_DWORD_VAL(pcie_ack_rpl_reg), current_ack_timer,
+		current_replay, expected_replay,
+		EFX_DWORD_VAL(pcie_ack_rpl_reg), current_ack_freq, 0);
+
+	/* If expected replay setting needs to be bigger then set it */
+	if (expected_replay > current_replay) {
+		EFX_SET_DWORD_FIELD(pcie_ack_rpl_reg, PCIE_CORE_RPL_TL,
+				    expected_replay);
+
+		falcon_pcie_core_write_reg(efx, PCIE_CORE_ADDR_ACK_RPL_TIMER,
+					   pcie_ack_rpl_reg);
+	}
+}
+
+static int falcon_init_pcie_core(struct efx_nic *efx)
+{
+	int pciecap_offset;
+	unsigned num_lanes = 0;
+
+	/* Get num lanes */
+	pciecap_offset = pci_find_capability(efx->pci_dev, PCI_CAP_ID_EXP);
+	if (pciecap_offset) {
+		u16 pcie_linkstat;
+		int rc, link_sta;
+
+		link_sta = pciecap_offset + PCI_EXP_LNKSTA;
+		rc = pci_read_config_word(efx->pci_dev, link_sta,
+					  &pcie_linkstat);
+		if (rc)
+			return rc;
+
+		num_lanes = ((pcie_linkstat & PCI_EXP_LNKSTA_LNK_WID)
+			     >> PCI_EXP_LNKSTA_LNK_WID_LBN);
+		EFX_BUG_ON_PARANOID(num_lanes <= 0 || num_lanes > 8);
+
+		if (num_lanes < 8)
+			EFX_ERR(efx, "WARNING: the Solarflare Network Adapter "
+				"has been plugged into a PCI-Express slot with "
+				"less than 8 lanes (%d detected). This will "
+				"limit the maximum achievable bandwidth! "
+				"Consult your motherboard documentation to "
+				"find a slot that is 8 lanes electrically and "
+				"physically\n", num_lanes);
+	}
+
+	if (FALCON_REV(efx) <= FALCON_REV_A1)
+		return 0;
+
+	if (EFX_WORKAROUND_6943(efx) && num_lanes > 0)
+		falcon_init_ack_repl_timer(efx, num_lanes);
+
+	if (EFX_WORKAROUND_9096(efx)) {
+		efx_dword_t pcie_ack_freq_reg;
+
+		/* ensure ack freq timer is 0 = always ack after timeout */
+		falcon_pcie_core_read_reg(efx, PCIE_CORE_ADDR_ACK_FREQ,
+					  &pcie_ack_freq_reg);
+		EFX_SET_DWORD_FIELD(pcie_ack_freq_reg, PCIE_CORE_ACK_FREQ, 0);
+		falcon_pcie_core_write_reg(efx, PCIE_CORE_ADDR_ACK_FREQ,
+					   pcie_ack_freq_reg);
+	}
+
+	return 0;
+}
+
+static void falcon_fini_pcie_core(struct efx_nic *efx)
+{
+	efx_dword_t pcie_ack_freq_reg;
+
+	if (FALCON_REV(efx) <= FALCON_REV_A1)
+		return;
+
+	if (EFX_WORKAROUND_9096(efx)) {
+		/* Set the ACK frequency timer to 1, so TLP's are acked in
+		 * a timely fashion.
+		 */
+		falcon_pcie_core_read_reg(efx, PCIE_CORE_ADDR_ACK_FREQ,
+					  &pcie_ack_freq_reg);
+		EFX_SET_DWORD_FIELD(pcie_ack_freq_reg, PCIE_CORE_ACK_FREQ, 1);
+		falcon_pcie_core_write_reg(efx, PCIE_CORE_ADDR_ACK_FREQ,
+					   pcie_ack_freq_reg);
+	}
+}
+
+/* This call performs hardware-specific global initialisation, such as
+ * defining the descriptor cache sizes and number of RSS channels.
+ * It does not set up any buffers, descriptor rings or event queues.
+ */
+int falcon_init_nic(struct efx_nic *efx)
+{
+	struct falcon_nic_data *data;
+	efx_oword_t temp;
+	unsigned thresh;
+	int rc;
+
+	data = (struct falcon_nic_data *)efx->nic_data;
+
+	/* Set up the address region register. This is only needed
+	 * for the B0 FPGA, but since we are just pushing in the
+	 * reset defaults this may as well be unconditional. */
+	EFX_POPULATE_OWORD_4(temp, ADR_REGION0, 0,
+				   ADR_REGION1, (1 << 16),
+				   ADR_REGION2, (2 << 16),
+				   ADR_REGION3, (3 << 16));
+	falcon_write(efx, &temp, ADR_REGION_REG_KER);
+
+	/* Use on-chip SRAM if needed.
+	 */
+	falcon_read(efx, &temp, NIC_STAT_REG);
+	if (efx->external_sram_cfg == SRM_NB_BSZ_ONCHIP_ONLY)
+		EFX_SET_OWORD_FIELD(temp, ONCHIP_SRAM, 1);
+	else
+		EFX_SET_OWORD_FIELD(temp, ONCHIP_SRAM, 0);
+	falcon_write(efx, &temp, NIC_STAT_REG);
+
+	/* Check power requirements against PCIe power budgeting */
+	rc = falcon_check_power_limit(efx);
+	if (rc)
+		return rc;
+
+	/* Warn if <8 lanes of PCIe detected & set pcie timers */
+	rc = falcon_init_pcie_core(efx);
+	if (rc)
+		return rc;
+
+	/* Set buffer table mode */
+	EFX_POPULATE_OWORD_1(temp, BUF_TBL_MODE, BUF_TBL_MODE_FULL);
+	falcon_write(efx, &temp, BUF_TBL_CFG_REG_KER);
+
+	rc = falcon_reset_sram(efx);
+	if (rc)
+		return rc;
+
+	/* Set positions of descriptor caches in SRAM. */
+	EFX_POPULATE_OWORD_1(temp, SRM_TX_DC_BASE_ADR, data->tx_dc_base / 8);
+	falcon_write(efx, &temp, SRM_TX_DC_CFG_REG_KER);
+	EFX_POPULATE_OWORD_1(temp, SRM_RX_DC_BASE_ADR, data->rx_dc_base / 8);
+	falcon_write(efx, &temp, SRM_RX_DC_CFG_REG_KER);
+
+	/* Set TX descriptor cache size. */
+	EFX_POPULATE_OWORD_1(temp, TX_DC_SIZE, ffs(data->tx_dc_entries) - 4);
+	falcon_write(efx, &temp, TX_DC_CFG_REG_KER);
+
+	/* Set RX descriptor cache size.  Set low watermark to size-8, as
+	 * this allows most efficient prefetching.
+	 */
+	EFX_POPULATE_OWORD_1(temp, RX_DC_SIZE, ffs(data->rx_dc_entries) - 4);
+	falcon_write(efx, &temp, RX_DC_CFG_REG_KER);
+	EFX_POPULATE_OWORD_1(temp, RX_DC_PF_LWM, data->rx_dc_entries - 8);
+	falcon_write(efx, &temp, RX_DC_PF_WM_REG_KER);
+
+	/* Clear the parity enables on the TX data fifos as
+	 * they produce false parity errors because of timing issues
+	 */
+	if (EFX_WORKAROUND_5129(efx)) {
+		falcon_read(efx, &temp, SPARE_REG_KER);
+		EFX_SET_OWORD_FIELD(temp, MEM_PERR_EN_TX_DATA, 0);
+		falcon_write(efx, &temp, SPARE_REG_KER);
+	}
+
+	/* Enable all the genuinely fatal interrupts.  (They are still
+	 * masked by the overall interrupt mask, controlled by
+	 * falcon_interrupts()).
+	 *
+	 * Note: All other fatal interrupts are enabled
+	 */
+	EFX_POPULATE_OWORD_3(temp,
+			     ILL_ADR_INT_KER_EN, 1,
+			     RBUF_OWN_INT_KER_EN, 1,
+			     TBUF_OWN_INT_KER_EN, 1);
+	EFX_INVERT_OWORD(temp);
+	falcon_write(efx, &temp, FATAL_INTR_REG_KER);
+
+	/* Set number of RSS queues for receive path. */
+	falcon_read(efx, &temp, RX_FILTER_CTL_REG);
+	if (FALCON_REV(efx) >= FALCON_REV_B0)
+		EFX_SET_OWORD_FIELD(temp, NUM_KER, 0);
+	else
+		EFX_SET_OWORD_FIELD(temp, NUM_KER, efx->rss_queues - 1);
+	if (EFX_WORKAROUND_7244(efx)) {
+		EFX_SET_OWORD_FIELD(temp, UDP_FULL_SRCH_LIMIT, 8);
+		EFX_SET_OWORD_FIELD(temp, UDP_WILD_SRCH_LIMIT, 8);
+		EFX_SET_OWORD_FIELD(temp, TCP_FULL_SRCH_LIMIT, 8);
+		EFX_SET_OWORD_FIELD(temp, TCP_WILD_SRCH_LIMIT, 8);
+	}
+	falcon_write(efx, &temp, RX_FILTER_CTL_REG);
+
+	falcon_setup_rss_indir_table(efx);
+
+	/* Setup RX.  Wait for descriptor is broken and must
+	 * be disabled.  RXDP recovery shouldn't be needed, but is.
+	 */
+	falcon_read(efx, &temp, RX_SELF_RST_REG_KER);
+	EFX_SET_OWORD_FIELD(temp, RX_NODESC_WAIT_DIS, 1);
+	EFX_SET_OWORD_FIELD(temp, RX_RECOVERY_EN, 1);
+	if (EFX_WORKAROUND_5583(efx))
+		EFX_SET_OWORD_FIELD(temp, RX_ISCSI_DIS, 1);
+	falcon_write(efx, &temp, RX_SELF_RST_REG_KER);
+
+	/* Disable the ugly timer-based TX DMA backoff and allow TX DMA to be
+	 * controlled by the RX FIFO fill level. Set arbitration to one pkt/Q.
+	 */
+	falcon_read(efx, &temp, TX_CFG2_REG_KER);
+	EFX_SET_OWORD_FIELD(temp, TX_RX_SPACER, 0xfe);
+	EFX_SET_OWORD_FIELD(temp, TX_RX_SPACER_EN, 1);
+	EFX_SET_OWORD_FIELD(temp, TX_ONE_PKT_PER_Q, 1);
+	EFX_SET_OWORD_FIELD(temp, TX_CSR_PUSH_EN, 0);
+	EFX_SET_OWORD_FIELD(temp, TX_DIS_NON_IP_EV, 1);
+	/* Enable SW_EV to inherit in char driver - assume harmless here */
+	EFX_SET_OWORD_FIELD(temp, TX_SW_EV_EN, 1);
+	/* Prefetch threshold 2 => fetch when descriptor cache half empty */
+	EFX_SET_OWORD_FIELD(temp, TX_PREF_THRESHOLD, 2);
+	if (EFX_WORKAROUND_9008(efx))
+		EFX_SET_OWORD_FIELD(temp, TX_PREF_WD_TMR, (unsigned)0x3fffff);
+	/* Squash TX of packets of 16 bytes or less */
+	if (FALCON_REV(efx) >= FALCON_REV_B0 && EFX_WORKAROUND_9141(efx))
+		EFX_SET_OWORD_FIELD(temp, TX_FLUSH_MIN_LEN_EN_B0, 1);
+	falcon_write(efx, &temp, TX_CFG2_REG_KER);
+
+	/* Do not enable TX_NO_EOP_DISC_EN, since it limits packets to 16
+	 * descriptors (which is bad).
+	 */
+	falcon_read(efx, &temp, TX_CFG_REG_KER);
+	EFX_SET_OWORD_FIELD(temp, TX_NO_EOP_DISC_EN, 0);
+	falcon_write(efx, &temp, TX_CFG_REG_KER);
+
+	/* RX config */
+	falcon_read(efx, &temp, RX_CFG_REG_KER);
+	EFX_SET_OWORD_FIELD_VER(efx, temp, RX_DESC_PUSH_EN, 0);
+	if (EFX_WORKAROUND_7575(efx))
+		EFX_SET_OWORD_FIELD_VER(efx, temp, RX_USR_BUF_SIZE,
+					(3 * 4096) / 32);
+	if (FALCON_REV(efx) >= FALCON_REV_B0)
+		EFX_SET_OWORD_FIELD(temp, RX_INGR_EN_B0, 1);
+
+	/* RX FIFO flow control thresholds */
+	thresh = ((rx_xon_thresh_bytes >= 0) ?
+		  rx_xon_thresh_bytes : efx->type->rx_xon_thresh);
+	EFX_SET_OWORD_FIELD_VER(efx, temp, RX_XON_MAC_TH, thresh / 256);
+	thresh = ((rx_xoff_thresh_bytes >= 0) ?
+		  rx_xoff_thresh_bytes : efx->type->rx_xoff_thresh);
+	EFX_SET_OWORD_FIELD_VER(efx, temp, RX_XOFF_MAC_TH, thresh / 256);
+	/* RX control FIFO thresholds [32 entries] */
+	EFX_SET_OWORD_FIELD_VER(efx, temp, RX_XON_TX_TH, 25);
+	EFX_SET_OWORD_FIELD_VER(efx, temp, RX_XOFF_TX_TH, 20);
+	falcon_write(efx, &temp, RX_CFG_REG_KER);
+
+	/* Set destination of both TX and RX Flush events */
+	if (FALCON_REV(efx) >= FALCON_REV_B0) {
+		EFX_POPULATE_OWORD_1(temp, FLS_EVQ_ID, 0);
+		falcon_write(efx, &temp, DP_CTRL_REG);
+	}
+
+	return 0;
+}
+
+void falcon_fini_nic(struct efx_nic *efx)
+{
+	falcon_fini_pcie_core(efx);
+}
+
+void falcon_remove_nic(struct efx_nic *efx)
+{
+	/* Tear down the private nic state, and the driverlink nic params */
+	kfree(efx->nic_data);
+	efx->nic_data = efx->dl_info = NULL;
+
+	falcon_remove_spi_devices(efx);
+	falcon_free_buffer(efx, &efx->irq_status);
+
+	/* Reset the NIC finally */
+	(void) falcon_reset_hw(efx, RESET_TYPE_ALL);
+
+	/* Release the second function after the reset */
+	if (efx->pci_dev2) {
+		pci_dev_put(efx->pci_dev2);
+		efx->pci_dev2 = NULL;
+	}
+}
+
+void falcon_update_nic_stats(struct efx_nic *efx)
+{
+	efx_oword_t cnt;
+
+	/* Read the RX drop counter */
+	falcon_read(efx, &cnt, RX_NODESC_DROP_REG_KER);
+	efx->n_rx_nodesc_drop_cnt += EFX_OWORD_FIELD(cnt, RX_NODESC_DROP_CNT);
+}
+
+/**************************************************************************
+ *
+ * Revision-dependent attributes used by efx.c
+ *
+ **************************************************************************
+ */
+
+struct efx_nic_type falcon_a_nic_type = {
+	.is_dual_func = 1,
+	.mem_bar = 2,
+	.mem_map_size = 0x20000,
+	.txd_ptr_tbl_base = TX_DESC_PTR_TBL_KER_A1,
+	.rxd_ptr_tbl_base = RX_DESC_PTR_TBL_KER_A1,
+	.buf_tbl_base = BUF_TBL_KER_A1,
+	.evq_ptr_tbl_base = EVQ_PTR_TBL_KER_A1,
+	.evq_rptr_tbl_base = EVQ_RPTR_REG_KER_A1,
+	.txd_ring_mask = FALCON_TXD_RING_MASK,
+	.rxd_ring_mask = FALCON_RXD_RING_MASK,
+	.evq_size = FALCON_EVQ_SIZE,
+	.max_dma_mask = FALCON_DMA_MASK,
+	.tx_dma_mask = FALCON_TX_DMA_MASK,
+	.bug5391_mask = 0xf,
+	.rx_xoff_thresh = 2048,
+	.rx_xon_thresh = 512,
+	.rx_buffer_padding = 0x24,
+	.max_interrupt_mode = EFX_INT_MODE_MSI,
+	.phys_addr_channels = 4,
+};
+
+struct efx_nic_type falcon_b_nic_type = {
+	.is_dual_func = 0,
+	.mem_bar = 2,
+	/* Map everything up to and including the RSS indirection
+	 * table.  Don't map MSI-X table, MSI-X PBA since Linux
+	 * requires that they not be mapped.  */
+	.mem_map_size = RX_RSS_INDIR_TBL_B0 + 0x800,
+	.txd_ptr_tbl_base = TX_DESC_PTR_TBL_KER_B0,
+	.rxd_ptr_tbl_base = RX_DESC_PTR_TBL_KER_B0,
+	.buf_tbl_base = BUF_TBL_KER_B0,
+	.evq_ptr_tbl_base = EVQ_PTR_TBL_KER_B0,
+	.evq_rptr_tbl_base = EVQ_RPTR_REG_KER_B0,
+	.txd_ring_mask = FALCON_TXD_RING_MASK,
+	.rxd_ring_mask = FALCON_RXD_RING_MASK,
+	.evq_size = FALCON_EVQ_SIZE,
+	.max_dma_mask = FALCON_DMA_MASK,
+	.tx_dma_mask = FALCON_TX_DMA_MASK,
+	.bug5391_mask = 0,
+	.rx_xoff_thresh = 54272, /* ~80Kb - 3*max MTU */
+	.rx_xon_thresh = 27648,  /* ~3*max MTU */
+	.rx_buffer_padding = 0,
+	.max_interrupt_mode = EFX_INT_MODE_MSIX,
+	.phys_addr_channels = 32, /* Hardware limit is 64, but the legacy
+				   * interrupt handler only supports 32
+				   * channels */
+
+};
+
diff -rpuN linux-2.6.18.8/drivers/net/sfc/falcon_gmac.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/falcon_gmac.c
--- linux-2.6.18.8/drivers/net/sfc/falcon_gmac.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/falcon_gmac.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,320 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/delay.h>
+#include "net_driver.h"
+#include "efx.h"
+#include "falcon.h"
+#include "mac.h"
+#include "falcon_hwdefs.h"
+#include "falcon_io.h"
+#include "gmii.h"
+
+/**************************************************************************
+ *
+ * MAC register access
+ *
+ **************************************************************************/
+
+/* Offset of a GMAC register within Falcon */
+#define FALCON_GMAC_REG(mac_reg)					\
+	(FALCON_GMAC_REGBANK + ((mac_reg) * FALCON_GMAC_REG_SIZE))
+
+static void falcon_gmac_writel(struct efx_nic *efx,
+			       efx_dword_t *value, unsigned int mac_reg)
+{
+	efx_oword_t temp;
+
+	EFX_POPULATE_OWORD_1(temp, MAC_DATA, EFX_DWORD_FIELD(*value, MAC_DATA));
+	falcon_write(efx, &temp, FALCON_GMAC_REG(mac_reg));
+}
+
+static void falcon_gmac_readl(struct efx_nic *efx,
+			      efx_dword_t *value, unsigned int mac_reg)
+{
+	efx_oword_t temp;
+
+	falcon_read(efx, &temp, FALCON_GMAC_REG(mac_reg));
+	EFX_POPULATE_DWORD_1(*value, MAC_DATA, EFX_OWORD_FIELD(temp, MAC_DATA));
+}
+
+/**************************************************************************
+ *
+ * MAC operations
+ *
+ *************************************************************************/
+
+static int falcon_init_gmac(struct efx_nic *efx)
+{
+	int rc;
+
+	/* Reset the MAC */
+	mentormac_reset(efx);
+
+	/* Initialise PHY */
+	rc = efx->phy_op->init(efx);
+	if (rc)
+		return rc;
+
+	return 0;
+}
+
+static void falcon_reconfigure_gmac(struct efx_nic *efx)
+{
+	/* Reconfigure PHY and pick up PHY parameters.  This updates
+	 * the link status. */
+	efx->phy_op->reconfigure(efx);
+
+	/* Isolate the MAC. */
+	falcon_deconfigure_mac_wrapper(efx);
+
+	/* Reconfigure MAC */
+	mentormac_reconfigure(efx);
+
+	/* Reconfigure MAC wrapper */
+	falcon_reconfigure_mac_wrapper(efx);
+}
+
+static void falcon_fini_gmac(struct efx_nic *efx)
+{
+	/* Isolate the MAC - PHY */
+	falcon_deconfigure_mac_wrapper(efx);
+
+	/* Shut down PHY */
+	efx->phy_op->fini(efx);
+
+	/* Reset MAC */
+	mentormac_reset(efx);
+}
+
+static void falcon_update_stats_gmac(struct efx_nic *efx)
+{
+	struct efx_mac_stats *mac_stats = &efx->mac_stats;
+	unsigned long old_rx_pause, old_tx_pause;
+	unsigned long new_rx_pause, new_tx_pause;
+	int rc;
+
+	rc = falcon_dma_stats(efx, GDmaDone_offset);
+	if (rc)
+		return;
+
+	/* Pause frames are erroneously counted as errors (SFC bug 3269) */
+	old_rx_pause = mac_stats->rx_pause;
+	old_tx_pause = mac_stats->tx_pause;
+
+	/* Update MAC stats from DMAed values */
+	FALCON_STAT(efx, GRxGoodOct, rx_good_bytes);
+	FALCON_STAT(efx, GRxBadOct, rx_bad_bytes);
+	FALCON_STAT(efx, GRxMissPkt, rx_missed);
+	FALCON_STAT(efx, GRxFalseCRS, rx_false_carrier);
+	FALCON_STAT(efx, GRxPausePkt, rx_pause);
+	FALCON_STAT(efx, GRxBadPkt, rx_bad);
+	FALCON_STAT(efx, GRxUcastPkt, rx_unicast);
+	FALCON_STAT(efx, GRxMcastPkt, rx_multicast);
+	FALCON_STAT(efx, GRxBcastPkt, rx_broadcast);
+	FALCON_STAT(efx, GRxGoodLt64Pkt, rx_good_lt64);
+	FALCON_STAT(efx, GRxBadLt64Pkt, rx_bad_lt64);
+	FALCON_STAT(efx, GRx64Pkt, rx_64);
+	FALCON_STAT(efx, GRx65to127Pkt, rx_65_to_127);
+	FALCON_STAT(efx, GRx128to255Pkt, rx_128_to_255);
+	FALCON_STAT(efx, GRx256to511Pkt, rx_256_to_511);
+	FALCON_STAT(efx, GRx512to1023Pkt, rx_512_to_1023);
+	FALCON_STAT(efx, GRx1024to15xxPkt, rx_1024_to_15xx);
+	FALCON_STAT(efx, GRx15xxtoJumboPkt, rx_15xx_to_jumbo);
+	FALCON_STAT(efx, GRxGtJumboPkt, rx_gtjumbo);
+	FALCON_STAT(efx, GRxFcsErr64to15xxPkt, rx_bad_64_to_15xx);
+	FALCON_STAT(efx, GRxFcsErr15xxtoJumboPkt, rx_bad_15xx_to_jumbo);
+	FALCON_STAT(efx, GRxFcsErrGtJumboPkt, rx_bad_gtjumbo);
+	FALCON_STAT(efx, GTxGoodBadOct, tx_bytes);
+	FALCON_STAT(efx, GTxGoodOct, tx_good_bytes);
+	FALCON_STAT(efx, GTxSglColPkt, tx_single_collision);
+	FALCON_STAT(efx, GTxMultColPkt, tx_multiple_collision);
+	FALCON_STAT(efx, GTxExColPkt, tx_excessive_collision);
+	FALCON_STAT(efx, GTxDefPkt, tx_deferred);
+	FALCON_STAT(efx, GTxLateCol, tx_late_collision);
+	FALCON_STAT(efx, GTxExDefPkt, tx_excessive_deferred);
+	FALCON_STAT(efx, GTxPausePkt, tx_pause);
+	FALCON_STAT(efx, GTxBadPkt, tx_bad);
+	FALCON_STAT(efx, GTxUcastPkt, tx_unicast);
+	FALCON_STAT(efx, GTxMcastPkt, tx_multicast);
+	FALCON_STAT(efx, GTxBcastPkt, tx_broadcast);
+	FALCON_STAT(efx, GTxLt64Pkt, tx_lt64);
+	FALCON_STAT(efx, GTx64Pkt, tx_64);
+	FALCON_STAT(efx, GTx65to127Pkt, tx_65_to_127);
+	FALCON_STAT(efx, GTx128to255Pkt, tx_128_to_255);
+	FALCON_STAT(efx, GTx256to511Pkt, tx_256_to_511);
+	FALCON_STAT(efx, GTx512to1023Pkt, tx_512_to_1023);
+	FALCON_STAT(efx, GTx1024to15xxPkt, tx_1024_to_15xx);
+	FALCON_STAT(efx, GTx15xxtoJumboPkt, tx_15xx_to_jumbo);
+	FALCON_STAT(efx, GTxGtJumboPkt, tx_gtjumbo);
+	FALCON_STAT(efx, GTxNonTcpUdpPkt, tx_non_tcpudp);
+	FALCON_STAT(efx, GTxMacSrcErrPkt, tx_mac_src_error);
+	FALCON_STAT(efx, GTxIpSrcErrPkt, tx_ip_src_error);
+
+	/* Pause frames are erroneously counted as errors (SFC bug 3269) */
+	new_rx_pause = mac_stats->rx_pause;
+	new_tx_pause = mac_stats->tx_pause;
+	mac_stats->rx_bad -= (new_rx_pause - old_rx_pause);
+	mac_stats->tx_bad -= (new_tx_pause - old_tx_pause);
+
+	/* Derive stats that the MAC doesn't provide directly */
+	mac_stats->tx_bad_bytes =
+		mac_stats->tx_bytes - mac_stats->tx_good_bytes;
+	mac_stats->tx_packets =
+		mac_stats->tx_lt64 + mac_stats->tx_64 +
+		mac_stats->tx_65_to_127 + mac_stats->tx_128_to_255 +
+		mac_stats->tx_256_to_511 + mac_stats->tx_512_to_1023 +
+		mac_stats->tx_1024_to_15xx + mac_stats->tx_15xx_to_jumbo +
+		mac_stats->tx_gtjumbo;
+	mac_stats->tx_collision =
+		mac_stats->tx_single_collision +
+		mac_stats->tx_multiple_collision +
+		mac_stats->tx_excessive_collision +
+		mac_stats->tx_late_collision;
+	mac_stats->rx_bytes =
+		mac_stats->rx_good_bytes + mac_stats->rx_bad_bytes;
+	mac_stats->rx_packets =
+		mac_stats->rx_good_lt64 + mac_stats->rx_bad_lt64 +
+		mac_stats->rx_64 + mac_stats->rx_65_to_127 +
+		mac_stats->rx_128_to_255 + mac_stats->rx_256_to_511 +
+		mac_stats->rx_512_to_1023 + mac_stats->rx_1024_to_15xx +
+		mac_stats->rx_15xx_to_jumbo + mac_stats->rx_gtjumbo;
+	mac_stats->rx_good = mac_stats->rx_packets - mac_stats->rx_bad;
+	mac_stats->rx_lt64 = mac_stats->rx_good_lt64 + mac_stats->rx_bad_lt64;
+}
+
+static int falcon_check_gmac(struct efx_nic *efx)
+{
+	/* Nothing to do */
+	return 0;
+}
+
+static void falcon_gmac_sim_phy_event(struct efx_nic *efx)
+{
+	efx_qword_t phy_event;
+
+	EFX_POPULATE_QWORD_2(phy_event,
+			     EV_CODE, GLOBAL_EV_DECODE,
+			     G_PHY0_INTR, 1);
+	falcon_generate_event(&efx->channel[0], &phy_event);
+}
+
+static void falcon_gmac_reset_phy(struct efx_nic *efx)
+{
+	struct mii_if_info *gmii = &efx->mii;
+	int bmcr, i;
+
+	/* Perform software reset to make new settings take effect */
+	bmcr = gmii->mdio_read(gmii->dev, gmii->phy_id, MII_BMCR);
+	bmcr |= BMCR_RESET;
+	gmii->mdio_write(gmii->dev, gmii->phy_id, MII_BMCR, bmcr);
+
+	/* Wait for the reset to deassert */
+	for (i = 20; i; --i) {
+		udelay(10);
+		if ((gmii->mdio_read(gmii->dev, gmii->phy_id, MII_BMCR) &
+		    BMCR_RESET) == 0)
+			return;
+	}
+
+	EFX_ERR(efx, "wait for PHY reset timed out\n");
+}
+
+
+static int falcon_gmac_get_settings(struct efx_nic *efx,
+				    struct ethtool_cmd *ecmd)
+{
+	struct mii_if_info *gmii = &efx->mii;
+	int rc;
+
+	rc = mii_ethtool_gset(gmii, ecmd);
+	ecmd->supported &= ~(SUPPORTED_1000baseT_Half);
+	return rc;
+}
+
+static int falcon_gmac_set_settings(struct efx_nic *efx,
+				    struct ethtool_cmd *ecmd)
+{
+	struct mii_if_info *gmii = &efx->mii;
+	int rc;
+
+	/* 1000Mbps half-duplex is technically legal, but none of our
+	 * current hardware supports it, so just disallow it. */
+	if (ecmd->speed == SPEED_1000 && ecmd->duplex != DUPLEX_FULL) {
+		EFX_LOG(efx, "rejecting unsupported 1000Mbps HD"
+			" setting\n");
+		return -EINVAL;
+	}
+
+	/* Use MII to set all other settings */
+	rc = mii_ethtool_sset(gmii, ecmd);
+	if (rc)
+		return rc;
+
+	/* Reset the PHY */
+	falcon_gmac_reset_phy(efx);
+
+	return 0;
+}
+
+static int falcon_gmac_set_pause(struct efx_nic *efx,
+				 enum efx_fc_type flow_control)
+{
+	struct mii_if_info *gmii = &efx->mii;
+	int adv;
+
+	/* GMAC has tiny MAC FIFO, so TX flow control won't work */
+	if (flow_control & EFX_FC_TX)
+		return -EINVAL;
+
+	efx->flow_control = flow_control;
+
+	/* Push autonegotiation to PHY */
+	adv = gmii->mdio_read(gmii->dev, gmii->phy_id, MII_ADVERTISE);
+	adv &= ~GM_ADVERTISE_PAUSE_CAP;
+	adv |= (flow_control & EFX_FC_AUTO) ? GM_ADVERTISE_PAUSE_CAP : 0;
+	gmii->mdio_write(gmii->dev, gmii->phy_id, MII_ADVERTISE, adv);
+
+	falcon_gmac_reset_phy(efx);
+
+	return 0;
+}
+
+
+struct efx_mac_operations falcon_gmac_operations = {
+	.mac_writel	= falcon_gmac_writel,
+	.mac_readl	= falcon_gmac_readl,
+	.init		= falcon_init_gmac,
+	.reconfigure	= falcon_reconfigure_gmac,
+	.update_stats	= falcon_update_stats_gmac,
+	.fini		= falcon_fini_gmac,
+	.check_hw	= falcon_check_gmac,
+	.fake_phy_event	= falcon_gmac_sim_phy_event,
+	.get_settings   = falcon_gmac_get_settings,
+	.set_settings   = falcon_gmac_set_settings,
+	.set_pause      = falcon_gmac_set_pause,
+};
diff -rpuN linux-2.6.18.8/drivers/net/sfc/falcon.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/falcon.h
--- linux-2.6.18.8/drivers/net/sfc/falcon.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/falcon.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,177 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_FALCON_H
+#define EFX_FALCON_H
+
+#include <asm/io.h>
+#include <linux/spinlock.h>
+#include "net_driver.h"
+
+/*
+ * Falcon hardware control
+ */
+
+enum falcon_revision {
+	FALCON_REV_A0 = 0,
+	FALCON_REV_A1 = 1,
+	FALCON_REV_B0 = 2,
+};
+
+#if defined(EFX_USE_PCI_DEV_REVISION)
+#define FALCON_REV(efx) ((efx)->pci_dev->revision)
+#else
+#define FALCON_REV(efx) ((efx)->revision)
+#endif
+
+extern struct efx_nic_type falcon_a_nic_type;
+extern struct efx_nic_type falcon_b_nic_type;
+
+/**************************************************************************
+ *
+ * Externs
+ *
+ **************************************************************************
+ */
+
+/* TX data path */
+extern int falcon_probe_tx(struct efx_tx_queue *tx_queue);
+extern int falcon_init_tx(struct efx_tx_queue *tx_queue);
+extern void falcon_fini_tx(struct efx_tx_queue *tx_queue);
+extern void falcon_remove_tx(struct efx_tx_queue *tx_queue);
+#if defined(EFX_USE_FASTCALL)
+extern void fastcall falcon_push_buffers(struct efx_tx_queue *tx_queue);
+#else
+extern void falcon_push_buffers(struct efx_tx_queue *tx_queue);
+#endif
+
+/* RX data path */
+extern int falcon_probe_rx(struct efx_rx_queue *rx_queue);
+extern int falcon_init_rx(struct efx_rx_queue *rx_queue);
+extern void falcon_fini_rx(struct efx_rx_queue *rx_queue);
+extern void falcon_remove_rx(struct efx_rx_queue *rx_queue);
+#if defined(EFX_USE_FASTCALL)
+extern void fastcall falcon_notify_rx_desc(struct efx_rx_queue *rx_queue);
+#else
+extern void falcon_notify_rx_desc(struct efx_rx_queue *rx_queue);
+#endif
+
+/* Event data path */
+extern int falcon_probe_eventq(struct efx_channel *channel);
+extern int falcon_init_eventq(struct efx_channel *channel);
+extern void falcon_fini_eventq(struct efx_channel *channel);
+extern void falcon_remove_eventq(struct efx_channel *channel);
+#if defined(EFX_USE_FASTCALL)
+extern int fastcall falcon_process_eventq(struct efx_channel *channel,
+					  int *rx_quota);
+#else
+extern int falcon_process_eventq(struct efx_channel *channel, int *rx_quota);
+#endif
+#if defined(EFX_USE_FASTCALL)
+extern void fastcall falcon_eventq_read_ack(struct efx_channel *channel);
+#else
+extern void falcon_eventq_read_ack(struct efx_channel *channel);
+#endif
+
+/* Ports */
+extern int falcon_probe_port(struct efx_nic *efx);
+extern void falcon_remove_port(struct efx_nic *efx);
+
+/* MAC/PHY */
+extern void falcon_check_xaui_link_up(struct efx_nic *efx);
+extern int falcon_xaui_link_ok(struct efx_nic *efx);
+extern int falcon_dma_stats(struct efx_nic *efx,
+			    unsigned int done_offset);
+extern void falcon_drain_tx_fifo(struct efx_nic *efx);
+extern void falcon_deconfigure_mac_wrapper(struct efx_nic *efx);
+extern void falcon_reconfigure_mac_wrapper(struct efx_nic *efx);
+
+/* Interrupts and test events */
+extern int falcon_init_interrupt(struct efx_nic *efx);
+extern void falcon_enable_interrupts(struct efx_nic *efx);
+extern void falcon_generate_test_event(struct efx_channel *channel,
+				       unsigned int magic);
+extern void falcon_generate_interrupt(struct efx_nic *efx);
+extern void falcon_set_int_moderation(struct efx_channel *channel);
+extern void falcon_disable_interrupts(struct efx_nic *efx);
+extern void falcon_fini_interrupt(struct efx_nic *efx);
+
+/* Global Resources */
+extern int falcon_probe_nic(struct efx_nic *efx);
+extern int falcon_probe_resources(struct efx_nic *efx);
+extern int falcon_init_nic(struct efx_nic *efx);
+extern int falcon_reset_hw(struct efx_nic *efx, enum reset_type method);
+extern void falcon_fini_nic(struct efx_nic *efx);
+extern void falcon_remove_resources(struct efx_nic *efx);
+extern void falcon_remove_nic(struct efx_nic *efx);
+extern void falcon_update_nic_stats(struct efx_nic *efx);
+extern void falcon_set_multicast_hash(struct efx_nic *efx);
+extern int falcon_reset_xaui(struct efx_nic *efx);
+
+/**************************************************************************
+ *
+ * Falcon MAC stats
+ *
+ **************************************************************************
+ */
+
+#define FALCON_STAT_OFFSET(falcon_stat) EFX_VAL(falcon_stat, offset)
+#define FALCON_STAT_WIDTH(falcon_stat) EFX_VAL(falcon_stat, WIDTH)
+
+/* Retrieve statistic from statistics block */
+#define FALCON_STAT(efx, falcon_stat, efx_stat) do {		\
+	if (FALCON_STAT_WIDTH(falcon_stat) == 16)		\
+		(efx)->mac_stats.efx_stat += le16_to_cpu(	\
+			*((__force __le16 *)				\
+			  (efx->stats_buffer.addr +		\
+			   FALCON_STAT_OFFSET(falcon_stat))));	\
+	else if (FALCON_STAT_WIDTH(falcon_stat) == 32)		\
+		(efx)->mac_stats.efx_stat += le32_to_cpu(	\
+			*((__force __le32 *)				\
+			  (efx->stats_buffer.addr +		\
+			   FALCON_STAT_OFFSET(falcon_stat))));	\
+	else							\
+		(efx)->mac_stats.efx_stat += le64_to_cpu(	\
+			*((__force __le64 *)				\
+			  (efx->stats_buffer.addr +		\
+			   FALCON_STAT_OFFSET(falcon_stat))));	\
+	} while (0)
+
+#define FALCON_MAC_STATS_SIZE 0x100
+
+#define MAC_DATA_LBN 0
+#define MAC_DATA_WIDTH 32
+
+extern void falcon_generate_event(struct efx_channel *channel,
+				  efx_qword_t *event);
+
+#ifdef CONFIG_SFC_DEBUGFS
+struct seq_file;
+extern int falcon_debugfs_read_hardware_desc(struct seq_file *file, void *data);
+#endif
+
+#endif /* EFX_FALCON_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/falcon_hwdefs.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/falcon_hwdefs.h
--- linux-2.6.18.8/drivers/net/sfc/falcon_hwdefs.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/falcon_hwdefs.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1620 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_FALCON_HWDEFS_H
+#define EFX_FALCON_HWDEFS_H
+
+/*
+ * Falcon hardware value definitions.
+ * Falcon is the internal codename for the SFC4000 controller that is
+ * present in SFE400X evaluation boards
+ */
+
+/**************************************************************************
+ *
+ * Falcon registers
+ *
+ **************************************************************************
+ */
+
+/* Address region register */
+#define ADR_REGION_REG_KER	0x00
+#define ADR_REGION0_LBN	0
+#define ADR_REGION0_WIDTH	18
+#define ADR_REGION1_LBN	32
+#define ADR_REGION1_WIDTH	18
+#define ADR_REGION2_LBN	64
+#define ADR_REGION2_WIDTH	18
+#define ADR_REGION3_LBN	96
+#define ADR_REGION3_WIDTH	18
+
+/* Interrupt enable register */
+#define INT_EN_REG_KER 0x0010
+#define INT_LEVEL_SEL_LBN 8
+#define INT_LEVEL_SEL_WIDTH 6
+#define MEM_PERR_INT_EN_KER_LBN 5
+#define MEM_PERR_INT_EN_KER_WIDTH 1
+#define KER_INT_CHAR_LBN 4
+#define KER_INT_CHAR_WIDTH 1
+#define KER_INT_KER_LBN 3
+#define KER_INT_KER_WIDTH 1
+#define ILL_ADR_ERR_INT_EN_KER_LBN 2
+#define ILL_ADR_ERR_INT_EN_KER_WIDTH 1
+#define SRM_PERR_INT_EN_KER_LBN 1
+#define SRM_PERR_INT_EN_KER_WIDTH 1
+#define DRV_INT_EN_KER_LBN 0
+#define DRV_INT_EN_KER_WIDTH 1
+
+/* Interrupt status address register */
+#define INT_ADR_REG_KER	0x0030
+#define NORM_INT_VEC_DIS_KER_LBN 64
+#define NORM_INT_VEC_DIS_KER_WIDTH 1
+#define INT_ADR_KER_LBN 0
+#define INT_ADR_KER_WIDTH EFX_DMA_TYPE_WIDTH(64) /* not 46 for this one */
+
+/* Interrupt status register (B0 only) */
+#define INT_ISR0_B0 0x90
+#define INT_ISR1_B0 0xA0
+
+/* Interrupt acknowledge register (A0/A1 only) */
+#define INT_ACK_REG_KER_A1 0x0050
+#define INT_ACK_DUMMY_DATA_LBN 0
+#define INT_ACK_DUMMY_DATA_WIDTH 32
+
+/* Interrupt acknowledge work-around register (A0/A1 only )*/
+#define WORK_AROUND_BROKEN_PCI_READS_REG_KER_A1 0x0070
+
+/* Hardware initialisation register */
+#define HW_INIT_REG_KER 0x00c0
+#define BCSR_TARGET_MASK_LBN 101
+#define BCSR_TARGET_MASK_WIDTH 4
+#define PCIE_TIMEOUT_DIS_LBN 123
+#define PCIE_TIMEOUT_DIS_WIDTH 1
+#define B2B_REQ_EN_B0_LBN 45
+#define B2B_REQ_EN_B0_WIDTH 1
+#define FC_BLOCKING_EN_B0_LBN 44
+#define FC_BLOCKING_EN_B0_WIDTH 1
+
+/* SPI host command register */
+#define EE_SPI_HCMD_REG_KER 0x0100
+#define EE_SPI_HCMD_CMD_EN_LBN 31
+#define EE_SPI_HCMD_CMD_EN_WIDTH 1
+#define EE_WR_TIMER_ACTIVE_LBN 28
+#define EE_WR_TIMER_ACTIVE_WIDTH 1
+#define EE_SPI_HCMD_SF_SEL_LBN 24
+#define EE_SPI_HCMD_SF_SEL_WIDTH 1
+#define EE_SPI_EEPROM 0
+#define EE_SPI_FLASH 1
+#define EE_SPI_HCMD_DABCNT_LBN 16
+#define EE_SPI_HCMD_DABCNT_WIDTH 5
+#define EE_SPI_HCMD_READ_LBN 15
+#define EE_SPI_HCMD_READ_WIDTH 1
+#define EE_SPI_READ 1
+#define EE_SPI_WRITE 0
+#define EE_SPI_HCMD_DUBCNT_LBN 12
+#define EE_SPI_HCMD_DUBCNT_WIDTH 2
+#define EE_SPI_HCMD_ADBCNT_LBN 8
+#define EE_SPI_HCMD_ADBCNT_WIDTH 2
+#define EE_SPI_HCMD_ENC_LBN 0
+#define EE_SPI_HCMD_ENC_WIDTH 8
+
+/* SPI host address register */
+#define EE_SPI_HADR_REG_KER 0x0110
+#define EE_SPI_HADR_DUBYTE_LBN 24
+#define EE_SPI_HADR_DUBYTE_WIDTH 8
+#define EE_SPI_HADR_ADR_LBN 0
+#define EE_SPI_HADR_ADR_WIDTH 24
+
+/* SPI host data register */
+#define EE_SPI_HDATA_REG_KER 0x0120
+#define EE_SPI_HDATA3_LBN 96
+#define EE_SPI_HDATA3_WIDTH 32
+#define EE_SPI_HDATA2_LBN 64
+#define EE_SPI_HDATA2_WIDTH 32
+#define EE_SPI_HDATA1_LBN 32
+#define EE_SPI_HDATA1_WIDTH 32
+#define EE_SPI_HDATA0_LBN 0
+#define EE_SPI_HDATA0_WIDTH 32
+
+/* SPI/VPD config register */
+#define EE_VPD_CFG_REG_KER 0x0140
+#define EE_VPD_EN_LBN 0
+#define EE_VPD_EN_WIDTH 1
+#define EE_VPD_EN_AD9_MODE_LBN 1
+#define EE_VPD_EN_AD9_MODE_WIDTH 1
+#define EE_EE_CLOCK_DIV_LBN 112
+#define EE_EE_CLOCK_DIV_WIDTH 7
+#define EE_SF_CLOCK_DIV_LBN 120
+#define EE_SF_CLOCK_DIV_WIDTH 7
+
+/* PCIE CORE ACCESS REG */
+#define PCIE_CORE_INDIRECT_REG 0x01f0
+#define PCIE_CORE_ADDR_LBN 0
+#define PCIE_CORE_ADDR_WIDTH 12
+#define PCIE_CORE_RW_LBN 15
+#define PCIE_CORE_RW_WIDTH 1
+#define PCIE_CORE_VALUE_LBN 32
+#define PCIE_CORE_VALUE_WIDTH 32
+
+#define PCIE_CORE_ADDR_PCIE_DEVICE_CTRL_STAT 0x68
+#define PCIE_CORE_ADDR_PCIE_LINK_CTRL_STAT 0x70
+#define PCIE_CORE_ADDR_ACK_RPL_TIMER 0x700
+#define PCIE_CORE_ACK_TL_LBN 0
+#define PCIE_CORE_ACK_TL_WIDTH 16
+#define PCIE_CORE_RPL_TL_LBN 16
+#define PCIE_CORE_RPL_TL_WIDTH 16
+
+#define PCIE_CORE_ADDR_ACK_FREQ 0x70C
+#define PCIE_CORE_ACK_FREQ_LBN 0
+#define PCIE_CORE_ACK_FREQ_WIDTH 7
+
+
+/* NIC status register */
+#define NIC_STAT_REG 0x0200
+#define EE_STRAP_EN_LBN 31
+#define EE_STRAP_EN_WIDTH 1
+#define EE_STRAP_OVR_LBN 24
+#define EE_STRAP_OVR_WIDTH 4
+#define ONCHIP_SRAM_LBN 16
+#define ONCHIP_SRAM_WIDTH 1
+#define SF_PRST_LBN 9
+#define SF_PRST_WIDTH 1
+#define EE_PRST_LBN 8
+#define EE_PRST_WIDTH 1
+#define EE_STRAP_LBN 7
+#define EE_STRAP_WIDTH 1
+/* See pic_mode_t for decoding of this field */
+#define STRAP_ISCSI_EN_LBN 3
+#define STRAP_ISCSI_EN_WIDTH 1
+#define STRAP_PINS_LBN 0
+#define STRAP_PINS_WIDTH 3
+/* These bit definitions are extrapolated from the list of numerical
+ * values for STRAP_PINS.
+ */
+#define STRAP_10G_LBN 2
+#define STRAP_10G_WIDTH 1
+#define STRAP_DUAL_PORT_LBN 1
+#define STRAP_DUAL_PORT_WIDTH 1
+#define STRAP_PCIE_LBN 0
+#define STRAP_PCIE_WIDTH 1
+
+#define FLASH_PRESENT_LBN 7
+#define FLASH_PRESENT_WIDTH 1
+#define EEPROM_PRESENT_LBN 6
+#define EEPROM_PRESENT_WIDTH 1
+#define BOOTED_USING_NVDEVICE_LBN 3
+#define BOOTED_USING_NVDEVICE_WIDTH 1
+
+/* GPIO control register */
+
+#define GPIO_CTL_REG_KER 0x0210
+
+#define GPIO_USE_NIC_CLK_LBN (30)
+#define GPIO_USE_NIC_CLK_WIDTH (1)
+
+#define GPIO_OUTPUTS_LBN   (16)
+#define GPIO_OUTPUTS_WIDTH (4)
+
+#define GPIO_INPUTS_LBN (8)
+#define GPIO_INPUT_WIDTH (4)
+
+#define GPIO_DIRECTION_LBN (24)
+#define GPIO_DIRECTION_WIDTH (4)
+#define GPIO_DIRECTION_OUT (1)
+#define GPIO_SRAM_SLEEP (1 << 1)
+
+#define GPIO3_OEN_LBN (GPIO_DIRECTION_LBN + 3)
+#define	GPIO3_OEN_WIDTH 1
+#define	GPIO2_OEN_LBN (GPIO_DIRECTION_LBN + 2)
+#define	GPIO2_OEN_WIDTH 1
+#define	GPIO1_OEN_LBN (GPIO_DIRECTION_LBN + 1)
+#define	GPIO1_OEN_WIDTH 1
+#define GPIO0_OEN_LBN (GPIO_DIRECTION_LBN + 0)
+#define	GPIO0_OEN_WIDTH 1
+
+#define	GPIO3_OUT_LBN (GPIO_OUTPUTS_LBN + 3)
+#define	GPIO3_OUT_WIDTH 1
+#define	GPIO2_OUT_LBN (GPIO_OUTPUTS_LBN + 2)
+#define	GPIO2_OUT_WIDTH 1
+#define	GPIO1_OUT_LBN (GPIO_OUTPUTS_LBN + 1)
+#define	GPIO1_OUT_WIDTH 1
+#define	GPIO0_OUT_LBN (GPIO_OUTPUTS_LBN + 0)
+#define	GPIO0_OUT_WIDTH 1
+
+#define GPIO3_IN_LBN (GPIO_INPUTS_LBN + 3)
+#define	GPIO3_IN_WIDTH 1
+#define GPIO2_IN_LBN (GPIO_INPUTS_LBN + 2)
+#define	GPIO2_IN_WIDTH 1
+#define GPIO1_IN_LBN (GPIO_INPUTS_LBN + 1)
+#define	GPIO1_IN_WIDTH 1
+#define GPIO0_IN_LBN (GPIO_INPUTS_LBN + 0)
+#define	GPIO0_IN_WIDTH 1
+
+/* Global control register */
+#define GLB_CTL_REG_KER	0x0220
+#define EXT_PHY_RST_CTL_LBN 63
+#define EXT_PHY_RST_CTL_WIDTH 1
+#define PCIE_SD_RST_CTL_LBN 61
+#define PCIE_SD_RST_CTL_WIDTH 1
+#define PCIX_RST_CTL_LBN 60
+#define PCIX_RST_CTL_WIDTH 1
+#define PCIE_STCK_RST_CTL_LBN 59
+
+#define PCIE_STCK_RST_CTL_WIDTH 1
+#define PCIE_NSTCK_RST_CTL_LBN 58
+#define PCIE_NSTCK_RST_CTL_WIDTH 1
+#define PCIE_CORE_RST_CTL_LBN 57
+#define PCIE_CORE_RST_CTL_WIDTH 1
+#define EE_RST_CTL_LBN 49
+#define EE_RST_CTL_WIDTH 1
+#define CS_RST_CTL_LBN 48
+#define CS_RST_CTL_WIDTH 1
+#define RST_EXT_PHY_LBN 31
+#define RST_EXT_PHY_WIDTH 1
+#define RST_XGRX_LBN 24
+#define RST_XGRX_WIDTH 1
+#define RST_XGTX_LBN 23
+#define RST_XGTX_WIDTH 1
+#define RST_EM_LBN 22
+#define RST_EM_WIDTH 1
+#define INT_RST_DUR_LBN 4
+#define INT_RST_DUR_WIDTH 3
+#define EXT_PHY_RST_DUR_LBN 1
+#define EXT_PHY_RST_DUR_WIDTH 3
+#define SWRST_LBN 0
+#define SWRST_WIDTH 1
+#define INCLUDE_IN_RESET 0
+#define EXCLUDE_FROM_RESET 1
+
+/* Fatal interrupt register */
+#define FATAL_INTR_REG_KER 0x0230
+#define PCI_BUSERR_INT_KER_EN_LBN 43
+#define PCI_BUSERR_INT_KER_EN_WIDTH 1
+#define SRAM_OOB_INT_KER_EN_LBN 42
+#define SRAM_OOB_INT_KER_EN_WIDTH 1
+#define BUFID_OOB_INT_KER_EN_LBN 41
+#define BUFID_OOB_INT_KER_EN_WIDTH 1
+#define MEM_PERR_INT_KER_EN_LBN 40
+#define MEM_PERR_INT_KER_EN_WIDTH 1
+#define RBUF_OWN_INT_KER_EN_LBN 39
+#define RBUF_OWN_INT_KER_EN_WIDTH 1
+#define TBUF_OWN_INT_KER_EN_LBN 38
+#define TBUF_OWN_INT_KER_EN_WIDTH 1
+#define RDESCQ_OWN_INT_KER_EN_LBN 37
+#define RDESCQ_OWN_INT_KER_EN_WIDTH 1
+#define TDESCQ_OWN_INT_KER_EN_LBN 36
+#define TDESCQ_OWN_INT_KER_EN_WIDTH 1
+#define EVQ_OWN_INT_KER_EN_LBN 35
+#define EVQ_OWN_INT_KER_EN_WIDTH 1
+#define EVFF_OFLO_INT_KER_EN_LBN 34
+#define EVFF_OFLO_INT_KER_EN_WIDTH 1
+#define ILL_ADR_INT_KER_EN_LBN 33
+#define ILL_ADR_INT_KER_EN_WIDTH 1
+#define SRM_PERR_INT_KER_EN_LBN 32
+#define SRM_PERR_INT_KER_EN_WIDTH 1
+#define MEM_PERR_INT_KER_LBN 8
+#define MEM_PERR_INT_KER_WIDTH 1
+#define INT_KER_ERROR_LBN 0
+#define INT_KER_ERROR_WIDTH 12
+
+#define DP_CTRL_REG 0x250
+#define FLS_EVQ_ID_LBN 0
+#define FLS_EVQ_ID_WIDTH 11
+
+#define MEM_STAT_REG_KER 0x260
+
+/* Debug probe register */
+#define DEBUG_REG_KER 0x0270
+#define DEBUG_BLK_SEL2_LBN 47
+#define DEBUG_BLK_SEL2_WIDTH 3
+#define DEBUG_BLK_SEL1_LBN 44
+#define DEBUG_BLK_SEL1_WIDTH 3
+#define DEBUG_BLK_SEL0_LBN 41
+#define DEBUG_BLK_SEL0_WIDTH 3
+#define DEBUG_BLK_SEL_MISC 7
+#define DEBUG_BLK_SEL_SERDES 6
+#define DEBUG_BLK_SEL_EM 5
+#define DEBUG_BLK_SEL_SR 4
+#define DEBUG_BLK_SEL_EV 3
+#define DEBUG_BLK_SEL_RX 2
+#define DEBUG_BLK_SEL_TX 1
+#define DEBUG_BLK_SEL_BIU 0
+#define MISC_DEBUG_ADDR_LBN 36
+#define MISC_DEBUG_ADDR_WIDTH 5
+#define SERDES_DEBUG_ADDR_LBN 31
+#define SERDES_DEBUG_ADDR_WIDTH 5
+#define EM_DEBUG_ADDR_LBN 26
+#define EM_DEBUG_ADDR_WIDTH 5
+#define SR_DEBUG_ADDR_LBN 21
+#define SR_DEBUG_ADDR_WIDTH 5
+#define EV_DEBUG_ADDR_LBN 16
+#define EV_DEBUG_ADDR_WIDTH 5
+#define RX_DEBUG_ADDR_LBN 11
+#define RX_DEBUG_ADDR_WIDTH 5
+#define TX_DEBUG_ADDR_LBN 5
+#define TX_DEBUG_ADDR_WIDTH 5
+#define BIU_DEBUG_ADDR_LBN 1
+#define BIU_DEBUG_ADDR_WIDTH 5
+#define DEBUG_DIS_LBN 0
+#define DEBUG_DIS_WIDTH 1
+
+/* Scratch register 0 */
+#define DRIVER_REG0_KER 0x0280
+#define DRIVER_DW0_LBN 0
+#define DRIVER_DW0_WIDTH 32
+
+/* Scratch register 1 */
+#define DRIVER_REG1_KER 0x0290
+#define DRIVER_DW1_LBN 0
+#define DRIVER_DW1_WIDTH 32
+
+/* Scratch register 2 */
+#define DRIVER_REG2_KER 0x02A0
+#define DRIVER_DW2_LBN 0
+#define DRIVER_DW2_WIDTH 32
+
+/* Scratch register 3 */
+#define DRIVER_REG3_KER 0x02B0
+#define DRIVER_DW3_LBN 0
+#define DRIVER_DW3_WIDTH 32
+
+/* Scratch register 4 */
+#define DRIVER_REG4_KER 0x02C0
+#define DRIVER_DW3_LBN 0
+#define DRIVER_DW3_WIDTH 32
+
+/* Scratch register 5 */
+#define DRIVER_REG5_KER 0x02D0
+#define DRIVER_DW3_LBN 0
+#define DRIVER_DW3_WIDTH 32
+
+/* Scratch register 6 */
+#define DRIVER_REG6_KER 0x02E0
+#define DRIVER_DW3_LBN 0
+#define DRIVER_DW3_WIDTH 32
+
+/* Scratch register 7 */
+#define DRIVER_REG7_KER 0x02F0
+#define DRIVER_DW3_LBN 0
+#define DRIVER_DW3_WIDTH 32
+
+/* FPGA build version */
+#define ALTERA_BUILD_REG_KER 0x0300
+#define VER_MAJOR_LBN 24
+#define VER_MAJOR_WIDTH 8
+#define VER_MINOR_LBN 16
+#define VER_MINOR_WIDTH 8
+#define VER_BUILD_LBN 0
+#define VER_BUILD_WIDTH 16
+#define VER_ALL_LBN 0
+#define VER_ALL_WIDTH 32
+
+/* Spare EEPROM bits register (flash 0x390) */
+#define SPARE_REG_KER 0x310
+#define MEM_PERR_EN_LBN 64
+#define MEM_PERR_EN_WIDTH 38
+#define MEM_PERR_EN_TX_DATA_LBN 72
+#define MEM_PERR_EN_TX_DATA_WIDTH 2
+#define SPARE_EE_BITS_LBN 1
+#define SPARE_EE_BITS_WIDTH 31
+#define PCIE_LEGACY_ENDPOINT_LBN 0
+#define PCIE_LEGACY_ENDPOINT_WIDTH 1
+
+/* Page mapped view of bottom 1024 EVQ RPTRS */
+#define EVQ_RPTR_REG_P0	0x400
+/* Bit definitions are as for the densely mapped
+ * RPTR registers. */
+
+/* Timer table for kernel access */
+#define TIMER_CMD_REG_KER 0x420
+#define TIMER_MODE_LBN 12
+#define TIMER_MODE_WIDTH 2
+#define TIMER_MODE_DIS 0
+#define TIMER_MODE_INT_HLDOFF 2
+#define TIMER_VAL_LBN 0
+#define TIMER_VAL_WIDTH 12
+
+/* Driver generated event register */
+#define DRV_EV_REG_KER 0x440
+#define DRV_EV_QID_LBN 64
+#define DRV_EV_QID_WIDTH 12
+#define DRV_EV_DATA_LBN 0
+#define DRV_EV_DATA_WIDTH 64
+
+/* Event Queue control register */
+#define EVQ_CTL_REG_KER 0x450
+#define EVQ_FIFO_NOTAF_TH_LBN 0
+#define EVQ_FIFO_NOTAF_TH_WIDTH 6
+#define EVQ_FIFO_AF_TH_LBN 8
+#define EVQ_FIFO_AF_TH_WIDTH 6
+
+/* Buffer table configuration register */
+#define BUF_TBL_CFG_REG_KER 0x600
+#define BUF_TBL_MODE_LBN 3
+#define BUF_TBL_MODE_WIDTH 1
+#define BUF_TBL_MODE_HALF 0
+#define BUF_TBL_MODE_FULL 1
+
+/* SRAM receive descriptor cache configuration register */
+#define SRM_RX_DC_CFG_REG_KER 0x610
+#define SRM_RX_DC_BASE_ADR_LBN 0
+#define SRM_RX_DC_BASE_ADR_WIDTH 21
+
+/* SRAM transmit descriptor cache configuration register */
+#define SRM_TX_DC_CFG_REG_KER 0x620
+#define SRM_TX_DC_BASE_ADR_LBN 0
+#define SRM_TX_DC_BASE_ADR_WIDTH 21
+
+/* SRAM configuration register */
+#define SRM_CFG_REG_KER 0x630
+#define SRAM_OOB_ADR_INTEN_LBN 5
+#define SRAM_OOB_ADR_INTEN_WIDTH 1
+#define SRAM_OOB_BUF_INTEN_LBN 4
+#define SRAM_OOB_BUF_INTEN_WIDTH 1
+#define SRAM_OOB_BT_INIT_EN_LBN 3
+#define SRAM_OOB_BT_INIT_EN_WIDTH 1
+#define SRM_NUM_BANK_LBN 2
+#define SRM_NUM_BANK_WIDTH 1
+#define SRM_BANK_SIZE_LBN 0
+#define SRM_BANK_SIZE_WIDTH 2
+#define SRM_NUM_BANKS_AND_BANK_SIZE_LBN 0
+#define SRM_NUM_BANKS_AND_BANK_SIZE_WIDTH 3
+#define SRM_NB_BSZ_1BANKS_2M 0
+#define SRM_NB_BSZ_1BANKS_4M 1
+#define SRM_NB_BSZ_1BANKS_8M 2
+#define SRM_NB_BSZ_DEFAULT 3 /* char driver will set the default */
+#define SRM_NB_BSZ_2BANKS_4M 4
+#define SRM_NB_BSZ_2BANKS_8M 5
+#define SRM_NB_BSZ_2BANKS_16M 6
+#define SRM_NB_BSZ_RESERVED 7
+
+/* Special buffer table update register */
+#define BUF_TBL_UPD_REG_KER 0x0650
+#define BUF_UPD_CMD_LBN 63
+#define BUF_UPD_CMD_WIDTH 1
+#define BUF_CLR_CMD_LBN 62
+#define BUF_CLR_CMD_WIDTH 1
+#define BUF_CLR_END_ID_LBN 32
+#define BUF_CLR_END_ID_WIDTH 20
+#define BUF_CLR_START_ID_LBN 0
+#define BUF_CLR_START_ID_WIDTH 20
+
+/* Receive configuration register */
+#define RX_CFG_REG_KER 0x800
+
+/* B0 */
+#define RX_TOEP_TCP_SUPPRESS_B0_LBN 48
+#define RX_TOEP_TCP_SUPPRESS_B0_WIDTH 1
+#define RX_INGR_EN_B0_LBN 47
+#define RX_INGR_EN_B0_WIDTH 1
+#define RX_TOEP_IPV4_B0_LBN 46
+#define RX_TOEP_IPV4_B0_WIDTH 1
+#define RX_HASH_ALG_B0_LBN 45
+#define RX_HASH_ALG_B0_WIDTH 1
+#define RX_HASH_INSERT_HDR_B0_LBN 44
+#define RX_HASH_INSERT_HDR_B0_WIDTH 1
+#define RX_DESC_PUSH_EN_B0_LBN 43
+#define RX_DESC_PUSH_EN_B0_WIDTH 1
+#define RX_RDW_PATCH_EN_LBN 42 /* Non head of line blocking */
+#define RX_RDW_PATCH_EN_WIDTH 1
+#define RX_PCI_BURST_SIZE_B0_LBN 39
+#define RX_PCI_BURST_SIZE_B0_WIDTH 3
+#define RX_OWNERR_CTL_B0_LBN 38
+#define RX_OWNERR_CTL_B0_WIDTH 1
+#define RX_XON_TX_TH_B0_LBN 33
+#define RX_XON_TX_TH_B0_WIDTH 5
+#define RX_XOFF_TX_TH_B0_LBN 28
+#define RX_XOFF_TX_TH_B0_WIDTH 5
+#define RX_USR_BUF_SIZE_B0_LBN 19
+#define RX_USR_BUF_SIZE_B0_WIDTH 9
+#define RX_XON_MAC_TH_B0_LBN 10
+#define RX_XON_MAC_TH_B0_WIDTH 9
+#define RX_XOFF_MAC_TH_B0_LBN 1
+#define RX_XOFF_MAC_TH_B0_WIDTH 9
+#define RX_XOFF_MAC_EN_B0_LBN 0
+#define RX_XOFF_MAC_EN_B0_WIDTH 1
+
+/* A1 */
+#define RX_DESC_PUSH_EN_A1_LBN 35
+#define RX_DESC_PUSH_EN_A1_WIDTH 1
+#define RX_PCI_BURST_SIZE_A1_LBN 31
+#define RX_PCI_BURST_SIZE_A1_WIDTH 3
+#define RX_OWNERR_CTL_A1_LBN 30
+#define RX_OWNERR_CTL_A1_WIDTH 1
+#define RX_XON_TX_TH_A1_LBN 25
+#define RX_XON_TX_TH_A1_WIDTH 5
+#define RX_XOFF_TX_TH_A1_LBN 20
+#define RX_XOFF_TX_TH_A1_WIDTH 5
+#define RX_USR_BUF_SIZE_A1_LBN 11
+#define RX_USR_BUF_SIZE_A1_WIDTH 9
+#define RX_XON_MAC_TH_A1_LBN 6
+#define RX_XON_MAC_TH_A1_WIDTH 5
+#define RX_XOFF_MAC_TH_A1_LBN 1
+#define RX_XOFF_MAC_TH_A1_WIDTH 5
+#define RX_XOFF_MAC_EN_A1_LBN 0
+#define RX_XOFF_MAC_EN_A1_WIDTH 1
+
+/* Receive filter control register */
+#define RX_FILTER_CTL_REG 0x810
+#define SCATTER_ENBL_NO_MATCH_Q_B0_LBN 40
+#define SCATTER_ENBL_NO_MATCH_Q_B0_WIDTH 1
+#define UDP_FULL_SRCH_LIMIT_LBN 32
+#define UDP_FULL_SRCH_LIMIT_WIDTH 8
+#define NUM_KER_LBN 24
+#define NUM_KER_WIDTH 2
+#define UDP_WILD_SRCH_LIMIT_LBN 16
+#define UDP_WILD_SRCH_LIMIT_WIDTH 8
+#define TCP_WILD_SRCH_LIMIT_LBN 8
+#define TCP_WILD_SRCH_LIMIT_WIDTH 8
+#define TCP_FULL_SRCH_LIMIT_LBN 0
+#define TCP_FULL_SRCH_LIMIT_WIDTH 8
+
+/* RX queue flush register */
+#define RX_FLUSH_DESCQ_REG_KER 0x0820
+#define RX_FLUSH_DESCQ_CMD_LBN 24
+#define RX_FLUSH_DESCQ_CMD_WIDTH 1
+#define RX_FLUSH_DESCQ_LBN 0
+#define RX_FLUSH_DESCQ_WIDTH 12
+
+/* Receive descriptor update register */
+#define RX_DESC_UPD_REG_KER 0x0830
+#define RX_DESC_WPTR_LBN 96
+#define RX_DESC_WPTR_WIDTH 12
+#define RX_DESC_UPD_REG_KER_DWORD (RX_DESC_UPD_REG_KER + 12)
+#define RX_DESC_WPTR_DWORD_LBN 0
+#define RX_DESC_WPTR_DWORD_WIDTH 12
+
+/* Receive descriptor cache configuration register */
+#define RX_DC_CFG_REG_KER 0x840
+#define RX_DC_SIZE_LBN 0
+#define RX_DC_SIZE_WIDTH 2
+
+#define RX_DC_PF_WM_REG_KER 0x850
+#define RX_DC_PF_LWM_LBN 0
+#define RX_DC_PF_LWM_WIDTH 6
+
+/* RX no descriptor drop counter */
+#define RX_NODESC_DROP_REG_KER 0x880
+#define RX_NODESC_DROP_CNT_LBN 0
+#define RX_NODESC_DROP_CNT_WIDTH 16
+
+/* RX black magic register */
+#define RX_SELF_RST_REG_KER 0x890
+#define RX_ISCSI_DIS_LBN 17
+#define RX_ISCSI_DIS_WIDTH 1
+#define RX_PREFETCH_TIMEOUT_EN_LBN 10
+#define RX_PREFETCH_TIMEOUT_EN_WIDTH 1
+#define RX_NODESC_WAIT_DIS_LBN 9
+#define RX_NODESC_WAIT_DIS_WIDTH 1
+#define RX_RECOVERY_EN_LBN 8
+#define RX_RECOVERY_EN_WIDTH 1
+#define RX_SHUTDOWN_REASON_LBN	0
+#define RX_SHUTDOWN_REASON_WIDTH (3)
+
+/* TX queue flush register */
+#define TX_FLUSH_DESCQ_REG_KER 0x0a00
+#define TX_FLUSH_DESCQ_CMD_LBN 12
+#define TX_FLUSH_DESCQ_CMD_WIDTH 1
+#define TX_FLUSH_DESCQ_LBN 0
+#define TX_FLUSH_DESCQ_WIDTH 12
+
+/* Transmit descriptor update register */
+#define TX_DESC_UPD_REG_KER 0x0a10
+#define TX_DESC_WPTR_LBN 96
+#define TX_DESC_WPTR_WIDTH 12
+#define TX_DESC_UPD_REG_KER_DWORD (TX_DESC_UPD_REG_KER + 12)
+#define TX_DESC_WPTR_DWORD_LBN 0
+#define TX_DESC_WPTR_DWORD_WIDTH 12
+
+/* Transmit descriptor cache configuration register */
+#define TX_DC_CFG_REG_KER 0xa20
+#define TX_DC_SIZE_LBN 0
+#define TX_DC_SIZE_WIDTH 2
+
+/* Transmit checksum configuration register (A0/A1 only) */
+#define TX_CHKSM_CFG_REG_KER_A1 0xa30
+
+/* Transmit configuration register */
+#define TX_CFG_REG_KER 0xa50
+#define TX_NO_EOP_DISC_EN_LBN 5
+#define TX_NO_EOP_DISC_EN_WIDTH 1
+
+/* Transmit configuration register 2 */
+#define TX_CFG2_REG_KER 0xa80
+#define TX_CSR_PUSH_EN_LBN 89
+#define TX_CSR_PUSH_EN_WIDTH 1
+#define TX_RX_SPACER_LBN 64
+#define TX_RX_SPACER_WIDTH 8
+#define TX_SW_EV_EN_LBN 59
+#define TX_SW_EV_EN_WIDTH 1
+#define TX_RX_SPACER_EN_LBN 57
+#define TX_RX_SPACER_EN_WIDTH 1
+#define TX_PREF_WD_TMR_LBN 22
+#define TX_PREF_WD_TMR_WIDTH 22
+#define TX_PREF_THRESHOLD_LBN 19
+#define TX_PREF_THRESHOLD_WIDTH 2
+#define TX_ONE_PKT_PER_Q_LBN 18
+#define TX_ONE_PKT_PER_Q_WIDTH 1
+#define TX_DIS_NON_IP_EV_LBN 17
+#define TX_DIS_NON_IP_EV_WIDTH 1
+#define TX_DMA_SPACER_LBN 8
+#define TX_DMA_SPACER_WIDTH 8
+#define TX_FLUSH_MIN_LEN_EN_B0_LBN 7
+#define TX_FLUSH_MIN_LEN_EN_B0_WIDTH 1
+#define TX_TCP_DIS_A1_LBN 7
+#define TX_TCP_DIS_A1_WIDTH 1
+#define TX_IP_DIS_A1_LBN 6
+#define TX_IP_DIS_A1_WIDTH 1
+#define TX_MAX_CPL_LBN 2
+#define TX_MAX_CPL_WIDTH 2
+#define TX_MAX_PREF_LBN 0
+#define TX_MAX_PREF_WIDTH 2
+
+/* Transmit VLAN filter control register */
+#define TX_VLAN_REG 0xae0
+
+/* PHY management transmit data register */
+#define MD_TXD_REG_KER 0xc00
+#define MD_TXD_LBN 0
+#define MD_TXD_WIDTH 16
+
+/* PHY management receive data register */
+#define MD_RXD_REG_KER 0xc10
+#define MD_RXD_LBN 0
+#define MD_RXD_WIDTH 16
+
+/* PHY management configuration & status register */
+#define MD_CS_REG_KER 0xc20
+#define MD_PT_LBN 7
+#define MD_PT_WIDTH 3
+#define MD_PL_LBN 6
+#define MD_PL_WIDTH 1
+#define MD_INT_CLR_LBN 5
+#define MD_INT_CLR_WIDTH 1
+#define MD_GC_LBN 4
+#define MD_GC_WIDTH 1
+#define MD_PRSP_LBN 3
+#define MD_PRSP_WIDTH 1
+#define MD_RIC_LBN 2
+#define MD_RIC_WIDTH 1
+#define MD_RDC_LBN 1
+#define MD_RDC_WIDTH 1
+#define MD_WRC_LBN 0
+#define MD_WRC_WIDTH 1
+
+/* PHY management PHY address register */
+#define MD_PHY_ADR_REG_KER 0xc30
+#define MD_PHY_ADR_LBN 0
+#define MD_PHY_ADR_WIDTH 16
+
+/* PHY management ID register */
+#define MD_ID_REG_KER 0xc40
+#define MD_PRT_ADR_LBN 11
+#define MD_PRT_ADR_WIDTH 5
+#define MD_DEV_ADR_LBN 6
+#define MD_DEV_ADR_WIDTH 5
+/* Used for writing both at once */
+#define MD_PRT_DEV_ADR_LBN 6
+#define MD_PRT_DEV_ADR_WIDTH 10
+
+/* PHY management status & mask register (DWORD read only) */
+#define MD_STAT_REG_KER 0xc50
+#define MD_PINT_LBN 4
+#define MD_PINT_WIDTH 1
+#define MD_DONE_LBN 3
+#define MD_DONE_WIDTH 1
+#define MD_BSERR_LBN 2
+#define MD_BSERR_WIDTH 1
+#define MD_LNFL_LBN 1
+#define MD_LNFL_WIDTH 1
+#define MD_BSY_LBN 0
+#define MD_BSY_WIDTH 1
+
+/* Port 0 and 1 MAC stats registers */
+#define MAC0_STAT_DMA_REG_KER 0xc60
+#define MAC1_STAT_DMA_REG_KER 0xc70
+#define MAC_STAT_DMA_CMD_LBN 48
+#define MAC_STAT_DMA_CMD_WIDTH 1
+#define MAC_STAT_DMA_REGION_LBN 46
+#define MAC_STAT_DMA_REGION_WIDTH 2
+#define MAC_STAT_DMA_ADR_LBN 0
+#define MAC_STAT_DMA_ADR_WIDTH EFX_DMA_TYPE_WIDTH(46)
+
+/* Port 0 and 1 MAC control registers */
+#define MAC0_CTRL_REG_KER 0xc80
+#define MAC1_CTRL_REG_KER 0xc90
+#define MAC_XOFF_VAL_LBN 16
+#define MAC_XOFF_VAL_WIDTH 16
+#define TXFIFO_DRAIN_EN_B0_LBN 7
+#define TXFIFO_DRAIN_EN_B0_WIDTH 1
+#define MAC_XG_DISTXCRC_LBN 5
+#define MAC_XG_DISTXCRC_WIDTH 1
+#define MAC_BCAD_ACPT_LBN 4
+#define MAC_BCAD_ACPT_WIDTH 1
+#define MAC_UC_PROM_LBN 3
+#define MAC_UC_PROM_WIDTH 1
+#define MAC_LINK_STATUS_LBN 2
+#define MAC_LINK_STATUS_WIDTH 1
+#define MAC_SPEED_LBN 0
+#define MAC_SPEED_WIDTH 2
+
+/* External interrupt control (replaces MAC1_CTRL in B0) */
+#define GEN_MODE_REG_KER 0xc90
+#define XFP_PHY_INT_POL_SEL_LBN 3
+#define XFP_PHY_INT_POL_SEL_WIDTH 1
+#define XG_PHY_INT_POL_SEL_LBN 2
+#define XG_PHY_INT_POL_SEL_WIDTH 1
+#define XFP_PHY_INT_MASK_LBN 1
+#define XFP_PHY_INT_MASK_WIDTH 1
+#define XG_PHY_INT_MASK_LBN 0
+#define XG_PHY_INT_MASK_WIDTH 1
+
+
+/* 10G XAUI XGXS default values */
+#define XX_TXDRV_DEQ_DEFAULT 0xe /* deq=.6 */
+#define XX_TXDRV_DTX_DEFAULT 0x5 /* 1.25 */
+#define XX_SD_CTL_DRV_DEFAULT 0  /* 20mA */
+
+/* Multicast address hash table */
+#define MAC_MCAST_HASH_REG0_KER 0xca0
+#define MAC_MCAST_HASH_REG1_KER 0xcb0
+
+/* MAC test register. */
+#define MAC_TEST_REG_KER 0xcc0
+#define MAC_PTLOOP_EN_LBN 0
+#define MAC_PTLOOP_EN_WIDTH 1
+
+/* GMAC registers */
+#define FALCON_GMAC_REGBANK 0xe00
+#define FALCON_GMAC_REGBANK_SIZE 0x200
+#define FALCON_GMAC_REG_SIZE 0x10
+
+/* Source MAC filter control register */
+#define TX_SRC_MAC_CTRL_REG 0x1100
+
+/* XMAC registers */
+#define FALCON_XMAC_REGBANK 0x1200
+#define FALCON_XMAC_REGBANK_SIZE 0x200
+#define FALCON_XMAC_REG_SIZE 0x10
+
+/* XGMAC address register low */
+#define XM_ADR_LO_REG_MAC 0x00
+#define XM_ADR_3_LBN 24
+#define XM_ADR_3_WIDTH 8
+#define XM_ADR_2_LBN 16
+#define XM_ADR_2_WIDTH 8
+#define XM_ADR_1_LBN 8
+#define XM_ADR_1_WIDTH 8
+#define XM_ADR_0_LBN 0
+#define XM_ADR_0_WIDTH 8
+
+/* XGMAC address register high */
+#define XM_ADR_HI_REG_MAC 0x01
+#define XM_ADR_5_LBN 8
+#define XM_ADR_5_WIDTH 8
+#define XM_ADR_4_LBN 0
+#define XM_ADR_4_WIDTH 8
+
+/* XGMAC global configuration */
+#define XM_GLB_CFG_REG_MAC 0x02
+#define XM_LINE_LB_DEEP_RSVD_LBN 28
+#define XM_LINE_LB_DEEP_RSVD_WIDTH 1
+#define XM_RMTFLT_GEN_LBN 17
+#define XM_RMTFLT_GEN_WIDTH 1
+#define XM_DEBUG_MODE_LBN 16
+#define XM_DEBUG_MODE_WIDTH 1
+#define XM_RX_STAT_EN_LBN 11
+#define XM_RX_STAT_EN_WIDTH 1
+#define XM_TX_STAT_EN_LBN 10
+#define XM_TX_STAT_EN_WIDTH 1
+#define XM_RX_JUMBO_MODE_LBN 6
+#define XM_RX_JUMBO_MODE_WIDTH 1
+#define XM_WAN_MODE_LBN 5
+#define XM_WAN_MODE_WIDTH 1
+#define XM_AUTOCLR_MODE_LBN 4
+#define XM_AUTOCLR_MODE_WIDTH 1
+#define XM_INTCLR_MODE_LBN 3
+#define XM_INTCLR_MODE_WIDTH 1
+#define XM_CORE_RST_LBN 0
+#define XM_CORE_RST_WIDTH 1
+
+/* XGMAC transmit configuration */
+#define XM_TX_CFG_REG_MAC 0x03
+#define XM_TX_PROG_LBN 24
+#define XM_TX_PROG_WIDTH 1
+#define XM_IPG_LBN 16
+#define XM_IPG_WIDTH 4
+#define XM_FCNTL_LBN 10
+#define XM_FCNTL_WIDTH 1
+#define XM_TXCRC_LBN 8
+#define XM_TXCRC_WIDTH 1
+#define XM_EDRC_LBN 6
+#define XM_EDRC_WIDTH 1
+#define XM_AUTO_PAD_LBN 5
+#define XM_AUTO_PAD_WIDTH 1
+#define XM_TX_PRMBL_LBN 2
+#define XM_TX_PRMBL_WIDTH 1
+#define XM_TXEN_LBN 1
+#define XM_TXEN_WIDTH 1
+#define XM_TX_RST_LBN 0
+#define XM_TX_RST_WIDTH 1
+
+/* XGMAC receive configuration */
+#define XM_RX_CFG_REG_MAC 0x04
+#define XM_PASS_LENERR_LBN 26
+#define XM_PASS_LENERR_WIDTH 1
+#define XM_PASS_CRC_ERR_LBN 25
+#define XM_PASS_CRC_ERR_WIDTH 1
+#define XM_PASS_PRMBLE_ERR_LBN 24
+#define XM_PASS_PRMBLE_ERR_WIDTH 1
+#define XM_ACPT_ALL_MCAST_LBN 11
+#define XM_ACPT_ALL_MCAST_WIDTH 1
+#define XM_ACPT_ALL_UCAST_LBN 9
+#define XM_ACPT_ALL_UCAST_WIDTH 1
+#define XM_AUTO_DEPAD_LBN 8
+#define XM_AUTO_DEPAD_WIDTH 1
+#define XM_RXCRC_LBN 3
+#define XM_RXCRC_WIDTH 1
+#define XM_RX_PRMBL_LBN 2
+#define XM_RX_PRMBL_WIDTH 1
+#define XM_RXEN_LBN 1
+#define XM_RXEN_WIDTH 1
+#define XM_RX_RST_LBN 0
+#define XM_RX_RST_WIDTH 1
+
+/* XGMAC flow control register */
+#define XM_FC_REG_MAC 0x7
+#define XM_PAUSE_TIME_LBN 16
+#define XM_PAUSE_TIME_WIDTH 16
+#define XM_DIS_FCNTL_LBN 0
+#define XM_DIS_FCNTL_WIDTH 1
+
+/* XGMAC pause time count register */
+#define XM_PAUSE_TIME_REG_MAC 0x9
+#define XM_TX_PAUSE_CNT_LBN 16
+#define XM_TX_PAUSE_CNT_WIDTH 16
+#define XM_RX_PAUSE_CNT_LBN 0
+#define XM_RX_PAUSE_CNT_WIDTH 16
+
+/* XGMAC transmit parameter register */
+#define XM_TX_PARAM_REG_MAC 0x0d
+#define XM_TX_JUMBO_MODE_LBN 31
+#define XM_TX_JUMBO_MODE_WIDTH 1
+#define XM_MAX_TX_FRM_SIZE_LBN 16
+#define XM_MAX_TX_FRM_SIZE_WIDTH 14
+#define XM_PAD_CHAR_LBN 0
+#define XM_PAD_CHAR_WIDTH 8
+
+/* XGMAC receive parameter register */
+#define XM_RX_PARAM_REG_MAC 0x0e
+#define XM_MAX_RX_FRM_SIZE_LBN 0
+#define XM_MAX_RX_FRM_SIZE_WIDTH 14
+
+/* XGXS/XAUI powerdown/reset register */
+#define XX_PWR_RST_REG_MAC 0x10
+
+#define XX_PWRDND_EN_LBN 15
+#define XX_PWRDND_EN_WIDTH 1
+#define XX_PWRDNC_EN_LBN 14
+#define XX_PWRDNC_EN_WIDTH 1
+#define XX_PWRDNB_EN_LBN 13
+#define XX_PWRDNB_EN_WIDTH 1
+#define XX_PWRDNA_EN_LBN 12
+#define XX_PWRDNA_EN_WIDTH 1
+#define XX_RSTPLLCD_EN_LBN 9
+#define XX_RSTPLLCD_EN_WIDTH 1
+#define XX_RSTPLLAB_EN_LBN 8
+#define XX_RSTPLLAB_EN_WIDTH 1
+#define XX_RESETD_EN_LBN 7
+#define XX_RESETD_EN_WIDTH 1
+#define XX_RESETC_EN_LBN 6
+#define XX_RESETC_EN_WIDTH 1
+#define XX_RESETB_EN_LBN 5
+#define XX_RESETB_EN_WIDTH 1
+#define XX_RESETA_EN_LBN 4
+#define XX_RESETA_EN_WIDTH 1
+#define XX_RSTXGXSRX_EN_LBN 2
+#define XX_RSTXGXSRX_EN_WIDTH 1
+#define XX_RSTXGXSTX_EN_LBN 1
+#define XX_RSTXGXSTX_EN_WIDTH 1
+#define XX_RST_XX_EN_LBN 0
+#define XX_RST_XX_EN_WIDTH 1
+
+/* XGXS/XAUI powerdown/reset control register */
+#define XX_SD_CTL_REG_MAC 0x11
+#define XX_TERMADJ1_LBN 17
+#define XX_TERMADJ1_WIDTH 1
+#define XX_TERMADJ0_LBN 16
+#define XX_TERMADJ0_WIDTH 1
+#define XX_HIDRVD_LBN 15
+#define XX_HIDRVD_WIDTH 1
+#define XX_LODRVD_LBN 14
+#define XX_LODRVD_WIDTH 1
+#define XX_HIDRVC_LBN 13
+#define XX_HIDRVC_WIDTH 1
+#define XX_LODRVC_LBN 12
+#define XX_LODRVC_WIDTH 1
+#define XX_HIDRVB_LBN 11
+#define XX_HIDRVB_WIDTH 1
+#define XX_LODRVB_LBN 10
+#define XX_LODRVB_WIDTH 1
+#define XX_HIDRVA_LBN 9
+#define XX_HIDRVA_WIDTH 1
+#define XX_LODRVA_LBN 8
+#define XX_LODRVA_WIDTH 1
+#define XX_LPBKD_LBN 3
+#define XX_LPBKD_WIDTH 1
+#define XX_LPBKC_LBN 2
+#define XX_LPBKC_WIDTH 1
+#define XX_LPBKB_LBN 1
+#define XX_LPBKB_WIDTH 1
+#define XX_LPBKA_LBN 0
+#define XX_LPBKA_WIDTH 1
+
+#define XX_TXDRV_CTL_REG_MAC 0x12
+#define XX_DEQD_LBN 28
+#define XX_DEQD_WIDTH 4
+#define XX_DEQC_LBN 24
+#define XX_DEQC_WIDTH 4
+#define XX_DEQB_LBN 20
+#define XX_DEQB_WIDTH 4
+#define XX_DEQA_LBN 16
+#define XX_DEQA_WIDTH 4
+#define XX_DTXD_LBN 12
+#define XX_DTXD_WIDTH 4
+#define XX_DTXC_LBN 8
+#define XX_DTXC_WIDTH 4
+#define XX_DTXB_LBN 4
+#define XX_DTXB_WIDTH 4
+#define XX_DTXA_LBN 0
+#define XX_DTXA_WIDTH 4
+
+/* XAUI XGXS core status register */
+#define XX_FORCE_SIG_LBN 24
+#define XX_FORCE_SIG_WIDTH 8
+#define XX_FORCE_SIG_DECODE_FORCED 0xff
+#define XX_XGXS_LB_EN_LBN 23
+#define XX_XGXS_LB_EN_WIDTH 1
+#define XX_XGMII_LB_EN_LBN 22
+#define XX_XGMII_LB_EN_WIDTH 1
+#define XX_CORE_STAT_REG_MAC 0x16
+#define XX_ALIGN_DONE_LBN 20
+#define XX_ALIGN_DONE_WIDTH 1
+#define XX_SYNC_STAT_LBN 16
+#define XX_SYNC_STAT_WIDTH 4
+#define XX_SYNC_STAT_DECODE_SYNCED 0xf
+#define XX_COMMA_DET_LBN 12
+#define XX_COMMA_DET_WIDTH 4
+#define XX_COMMA_DET_DECODE_DETECTED 0xf
+#define XX_COMMA_DET_RESET 0xf
+#define XX_CHARERR_LBN 4
+#define XX_CHARERR_WIDTH 4
+#define XX_CHARERR_RESET 0xf
+#define XX_DISPERR_LBN 0
+#define XX_DISPERR_WIDTH 4
+#define XX_DISPERR_RESET 0xf
+
+/* Receive filter table */
+#define RX_FILTER_TBL0 0xF00000
+
+/* Receive descriptor pointer table */
+#define RX_DESC_PTR_TBL_KER_A1 0x11800
+#define RX_DESC_PTR_TBL_KER_B0 0xF40000
+#define RX_DESC_PTR_TBL_KER_P0 0x900
+#define RX_ISCSI_DDIG_EN_LBN 88
+#define RX_ISCSI_DDIG_EN_WIDTH 1
+#define RX_ISCSI_HDIG_EN_LBN 87
+#define RX_ISCSI_HDIG_EN_WIDTH 1
+#define RX_DESC_PREF_ACT_LBN 86
+#define RX_DESC_PREF_ACT_WIDTH 1
+#define RX_DC_HW_RPTR_LBN 80
+#define RX_DC_HW_RPTR_WIDTH 6
+#define RX_DESCQ_HW_RPTR_LBN 68
+#define RX_DESCQ_HW_RPTR_WIDTH 12
+#define RX_DESCQ_SW_WPTR_LBN 56
+#define RX_DESCQ_SW_WPTR_WIDTH 12
+#define RX_DESCQ_BUF_BASE_ID_LBN 36
+#define RX_DESCQ_BUF_BASE_ID_WIDTH 20
+#define RX_DESCQ_EVQ_ID_LBN 24
+#define RX_DESCQ_EVQ_ID_WIDTH 12
+#define RX_DESCQ_OWNER_ID_LBN 10
+#define RX_DESCQ_OWNER_ID_WIDTH 14
+#define RX_DESCQ_LABEL_LBN 5
+#define RX_DESCQ_LABEL_WIDTH 5
+#define RX_DESCQ_SIZE_LBN 3
+#define RX_DESCQ_SIZE_WIDTH 2
+#define RX_DESCQ_SIZE_4K 3
+#define RX_DESCQ_SIZE_2K 2
+#define RX_DESCQ_SIZE_1K 1
+#define RX_DESCQ_SIZE_512 0
+#define RX_DESCQ_TYPE_LBN 2
+#define RX_DESCQ_TYPE_WIDTH 1
+#define RX_DESCQ_JUMBO_LBN 1
+#define RX_DESCQ_JUMBO_WIDTH 1
+#define RX_DESCQ_EN_LBN 0
+#define RX_DESCQ_EN_WIDTH 1
+
+/* Transmit descriptor pointer table */
+#define TX_DESC_PTR_TBL_KER_A1 0x11900
+#define TX_DESC_PTR_TBL_KER_B0 0xF50000
+#define TX_DESC_PTR_TBL_KER_P0 0xa40
+#define TX_NON_IP_DROP_DIS_B0_LBN 91
+#define TX_NON_IP_DROP_DIS_B0_WIDTH 1
+#define TX_IP_CHKSM_DIS_B0_LBN 90
+#define TX_IP_CHKSM_DIS_B0_WIDTH 1
+#define TX_TCP_CHKSM_DIS_B0_LBN 89
+#define TX_TCP_CHKSM_DIS_B0_WIDTH 1
+#define TX_DESCQ_EN_LBN 88
+#define TX_DESCQ_EN_WIDTH 1
+#define TX_ISCSI_DDIG_EN_LBN 87
+#define TX_ISCSI_DDIG_EN_WIDTH 1
+#define TX_ISCSI_HDIG_EN_LBN 86
+#define TX_ISCSI_HDIG_EN_WIDTH 1
+#define TX_DC_HW_RPTR_LBN 80
+#define TX_DC_HW_RPTR_WIDTH 6
+#define TX_DESCQ_HW_RPTR_LBN 68
+#define TX_DESCQ_HW_RPTR_WIDTH 12
+#define TX_DESCQ_SW_WPTR_LBN 56
+#define TX_DESCQ_SW_WPTR_WIDTH 12
+#define TX_DESCQ_BUF_BASE_ID_LBN 36
+#define TX_DESCQ_BUF_BASE_ID_WIDTH 20
+#define TX_DESCQ_EVQ_ID_LBN 24
+#define TX_DESCQ_EVQ_ID_WIDTH 12
+#define TX_DESCQ_OWNER_ID_LBN 10
+#define TX_DESCQ_OWNER_ID_WIDTH 14
+#define TX_DESCQ_LABEL_LBN 5
+#define TX_DESCQ_LABEL_WIDTH 5
+#define TX_DESCQ_SIZE_LBN 3
+#define TX_DESCQ_SIZE_WIDTH 2
+#define TX_DESCQ_SIZE_4K 3
+#define TX_DESCQ_SIZE_2K 2
+#define TX_DESCQ_SIZE_1K 1
+#define TX_DESCQ_SIZE_512 0
+#define TX_DESCQ_TYPE_LBN 1
+#define TX_DESCQ_TYPE_WIDTH 2
+#define TX_DESCQ_FLUSH_LBN 0
+#define TX_DESCQ_FLUSH_WIDTH 1
+
+/* Event queue pointer */
+#define EVQ_PTR_TBL_KER_A1 0x11a00
+#define EVQ_PTR_TBL_KER_B0 0xf60000
+#define EVQ_PTR_TBL_KER_P0 0x500
+#define EVQ_WKUP_OR_INT_EN_LBN 39
+#define EVQ_WKUP_OR_INT_EN_WIDTH 1
+#define EVQ_NXT_WPTR_LBN 24
+#define EVQ_NXT_WPTR_WIDTH 15
+#define EVQ_EN_LBN 23
+#define EVQ_EN_WIDTH 1
+#define EVQ_SIZE_LBN 20
+#define EVQ_SIZE_WIDTH 3
+#define EVQ_SIZE_32K 6
+#define EVQ_SIZE_16K 5
+#define EVQ_SIZE_8K 4
+#define EVQ_SIZE_4K 3
+#define EVQ_SIZE_2K 2
+#define EVQ_SIZE_1K 1
+#define EVQ_SIZE_512 0
+#define EVQ_BUF_BASE_ID_LBN 0
+#define EVQ_BUF_BASE_ID_WIDTH 20
+
+/* Event queue read pointer */
+#define EVQ_RPTR_REG_KER_A1 0x11b00
+#define EVQ_RPTR_REG_KER_B0 0xfa0000
+#define EVQ_RPTR_LBN 0
+#define EVQ_RPTR_WIDTH 14
+#define EVQ_RPTR_REG_KER_DWORD (EVQ_RPTR_REG_KER + 0)
+#define EVQ_RPTR_DWORD_LBN 0
+#define EVQ_RPTR_DWORD_WIDTH 14
+
+/* RSS indirection table */
+#define RX_RSS_INDIR_TBL_B0 0xFB0000
+#define RX_RSS_INDIR_ENT_B0_LBN 0
+#define RX_RSS_INDIR_ENT_B0_WIDTH 6
+
+/* Special buffer descriptors (full-mode) */
+#define BUF_FULL_TBL_KER_A1 0x8000
+#define BUF_FULL_TBL_KER_B0 0x800000
+#define IP_DAT_BUF_SIZE_LBN 50
+#define IP_DAT_BUF_SIZE_WIDTH 1
+#define IP_DAT_BUF_SIZE_8K 1
+#define IP_DAT_BUF_SIZE_4K 0
+#define BUF_ADR_REGION_LBN 48
+#define BUF_ADR_REGION_WIDTH 2
+#define BUF_ADR_FBUF_LBN 14
+#define BUF_ADR_FBUF_WIDTH 34
+#define BUF_OWNER_ID_FBUF_LBN 0
+#define BUF_OWNER_ID_FBUF_WIDTH 14
+
+/* Special buffer descriptors (half-mode) */
+#define BUF_HALF_TBL_KER_A1 0x8000
+#define BUF_HALF_TBL_KER_B0 0x800000
+#define BUF_ADR_HBUF_ODD_LBN 44
+#define BUF_ADR_HBUF_ODD_WIDTH 20
+#define BUF_OWNER_ID_HBUF_ODD_LBN 32
+#define BUF_OWNER_ID_HBUF_ODD_WIDTH 12
+#define BUF_ADR_HBUF_EVEN_LBN 12
+#define BUF_ADR_HBUF_EVEN_WIDTH 20
+#define BUF_OWNER_ID_HBUF_EVEN_LBN 0
+#define BUF_OWNER_ID_HBUF_EVEN_WIDTH 12
+
+#define SRM_DBG_REG_B0 0x3000000
+
+/* Transmit descriptor */
+#define TX_KER_PORT_LBN 63
+#define TX_KER_PORT_WIDTH 1
+#define TX_KER_CONT_LBN 62
+#define TX_KER_CONT_WIDTH 1
+#define TX_KER_BYTE_CNT_LBN 48
+#define TX_KER_BYTE_CNT_WIDTH 14
+#define TX_KER_BUF_REGION_LBN 46
+#define TX_KER_BUF_REGION_WIDTH 2
+#define TX_KER_BUF_REGION0_DECODE 0
+#define TX_KER_BUF_REGION1_DECODE 1
+#define TX_KER_BUF_REGION2_DECODE 2
+#define TX_KER_BUF_REGION3_DECODE 3
+#define TX_KER_BUF_ADR_LBN 0
+#define TX_KER_BUF_ADR_WIDTH EFX_DMA_TYPE_WIDTH(46)
+
+/* Receive descriptor */
+#define RX_KER_BUF_SIZE_LBN 48
+#define RX_KER_BUF_SIZE_WIDTH 14
+#define RX_KER_BUF_REGION_LBN 46
+#define RX_KER_BUF_REGION_WIDTH 2
+#define RX_KER_BUF_REGION0_DECODE 0
+#define RX_KER_BUF_REGION1_DECODE 1
+#define RX_KER_BUF_REGION2_DECODE 2
+#define RX_KER_BUF_REGION3_DECODE 3
+#define RX_KER_BUF_ADR_LBN 0
+#define RX_KER_BUF_ADR_WIDTH EFX_DMA_TYPE_WIDTH(46)
+
+/**************************************************************************
+ *
+ * Falcon events
+ *
+ **************************************************************************
+ */
+
+/* Event queue entries */
+#define EV_CODE_LBN 60
+#define EV_CODE_WIDTH 4
+#define RX_IP_EV_DECODE 0
+#define TX_IP_EV_DECODE 2
+#define DRIVER_EV_DECODE 5
+#define GLOBAL_EV_DECODE 6
+#define DRV_GEN_EV_DECODE 7
+#define WHOLE_EVENT_LBN 0
+#define WHOLE_EVENT_WIDTH 64
+
+/* Receive events */
+#define RX_EV_PKT_OK_LBN 56
+#define RX_EV_PKT_OK_WIDTH 1
+#define RX_EV_PAUSE_FRM_ERR_LBN 55
+#define RX_EV_PAUSE_FRM_ERR_WIDTH 1
+#define RX_EV_BUF_OWNER_ID_ERR_LBN 54
+#define RX_EV_BUF_OWNER_ID_ERR_WIDTH 1
+#define RX_EV_IF_FRAG_ERR_LBN 53
+#define RX_EV_IF_FRAG_ERR_WIDTH 1
+#define RX_EV_IP_HDR_CHKSUM_ERR_LBN 52
+#define RX_EV_IP_HDR_CHKSUM_ERR_WIDTH 1
+#define RX_EV_TCP_UDP_CHKSUM_ERR_LBN 51
+#define RX_EV_TCP_UDP_CHKSUM_ERR_WIDTH 1
+#define RX_EV_ETH_CRC_ERR_LBN 50
+#define RX_EV_ETH_CRC_ERR_WIDTH 1
+#define RX_EV_FRM_TRUNC_LBN 49
+#define RX_EV_FRM_TRUNC_WIDTH 1
+#define RX_EV_DRIB_NIB_LBN 48
+#define RX_EV_DRIB_NIB_WIDTH 1
+#define RX_EV_TOBE_DISC_LBN 47
+#define RX_EV_TOBE_DISC_WIDTH 1
+#define RX_EV_PKT_TYPE_LBN 44
+#define RX_EV_PKT_TYPE_WIDTH 3
+#define RX_EV_PKT_TYPE_ETH_DECODE 0
+#define RX_EV_PKT_TYPE_LLC_DECODE 1
+#define RX_EV_PKT_TYPE_JUMBO_DECODE 2
+#define RX_EV_PKT_TYPE_VLAN_DECODE 3
+#define RX_EV_PKT_TYPE_VLAN_LLC_DECODE 4
+#define RX_EV_PKT_TYPE_VLAN_JUMBO_DECODE 5
+#define RX_EV_HDR_TYPE_LBN 42
+#define RX_EV_HDR_TYPE_WIDTH 2
+#define RX_EV_HDR_TYPE_TCP_IPV4_DECODE 0
+#define RX_EV_HDR_TYPE_UDP_IPV4_DECODE 1
+#define RX_EV_HDR_TYPE_OTHER_IP_DECODE 2
+#define RX_EV_HDR_TYPE_NON_IP_DECODE 3
+#define RX_EV_HDR_TYPE_HAS_CHECKSUMS(hdr_type) \
+	((hdr_type) <= RX_EV_HDR_TYPE_UDP_IPV4_DECODE)
+#define RX_EV_DESC_Q_EMPTY_LBN 41
+#define RX_EV_DESC_Q_EMPTY_WIDTH 1
+#define RX_EV_MCAST_HASH_MATCH_LBN 40
+#define RX_EV_MCAST_HASH_MATCH_WIDTH 1
+#define RX_EV_MCAST_PKT_LBN 39
+#define RX_EV_MCAST_PKT_WIDTH 1
+#define RX_EV_RECOVERY_FLAG_LBN 37
+#define RX_EV_RECOVERY_FLAG_WIDTH 1
+#define RX_EV_Q_LABEL_LBN 32
+#define RX_EV_Q_LABEL_WIDTH 5
+#define RX_EV_JUMBO_CONT_LBN 31
+#define RX_EV_JUMBO_CONT_WIDTH 1
+#define RX_EV_PORT_LBN 30
+#define RX_EV_PORT_WIDTH 1
+#define RX_EV_BYTE_CNT_LBN 16
+#define RX_EV_BYTE_CNT_WIDTH 14
+#define RX_EV_SOP_LBN 15
+#define RX_EV_SOP_WIDTH 1
+#define RX_ISCSI_DDIG_ERR_LBN 13
+#define RX_ISCSI_DDIG_ERR_WIDTH 1
+#define RX_ISCSI_HDIG_ERR_LBN 12
+#define RX_ISCSI_HDIG_ERR_WIDTH 1
+#define RX_EV_DESC_PTR_LBN 0
+#define RX_EV_DESC_PTR_WIDTH 12
+
+/* Transmit events */
+#define TX_EV_PKT_ERR_LBN 38
+#define TX_EV_PKT_ERR_WIDTH 1
+#define TX_EV_PKT_TOO_BIG_LBN 37
+#define TX_EV_PKT_TOO_BIG_WIDTH 1
+#define TX_EV_Q_LABEL_LBN 32
+#define TX_EV_Q_LABEL_WIDTH 5
+#define TX_EV_PORT_LBN 16
+#define TX_EV_PORT_WIDTH 1
+#define TX_EV_WQ_FF_FULL_LBN 15
+#define TX_EV_WQ_FF_FULL_WIDTH 1
+#define TX_EV_BUF_OWNER_ID_ERR_LBN 14
+#define TX_EV_BUF_OWNER_ID_ERR_WIDTH 1
+#define TX_EV_COMP_LBN 12
+#define TX_EV_COMP_WIDTH 1
+#define TX_EV_DESC_PTR_LBN 0
+#define TX_EV_DESC_PTR_WIDTH 12
+
+/* Driver events */
+#define DRIVER_EV_SUB_CODE_LBN 56
+#define DRIVER_EV_SUB_CODE_WIDTH 4
+#define DRIVER_EV_SUB_DATA_LBN 0
+#define DRIVER_EV_SUB_DATA_WIDTH 14
+#define TX_DESCQ_FLS_DONE_EV_DECODE 0
+#define RX_DESCQ_FLS_DONE_EV_DECODE 1
+#define EVQ_INIT_DONE_EV_DECODE 2
+#define EVQ_NOT_EN_EV_DECODE 3
+#define RX_DESCQ_FLSFF_OVFL_EV_DECODE 4
+#define SRM_UPD_DONE_EV_DECODE 5
+#define WAKE_UP_EV_DECODE 6
+#define TX_PKT_NON_TCP_UDP_DECODE 9
+#define TIMER_EV_DECODE 10
+#define RX_RECOVERY_EV_DECODE 11
+#define RX_DSC_ERROR_EV_DECODE 14
+#define TX_DSC_ERROR_EV_DECODE 15
+#define DRIVER_EV_TX_DESCQ_ID_LBN 0
+#define DRIVER_EV_TX_DESCQ_ID_WIDTH 12
+#define DRIVER_EV_RX_FLUSH_FAIL_LBN 12
+#define DRIVER_EV_RX_FLUSH_FAIL_WIDTH 1
+#define DRIVER_EV_RX_DESCQ_ID_LBN 0
+#define DRIVER_EV_RX_DESCQ_ID_WIDTH 12
+#define DRIVER_EV_EVQ_ID_LBN 0
+#define DRIVER_EV_EVQ_ID_WIDTH 12
+#define DRIVER_EV_SRM_UPD_LBN 0
+#define DRIVER_EV_SRM_UPD_WIDTH 2
+#define SRM_CLR_EV_DECODE 0
+#define SRM_UPD_EV_DECODE 1
+#define SRM_ILLCLR_EV_DECODE 2
+
+/* Global events */
+#define RX_RECOVERY_B0_LBN 12
+#define RX_RECOVERY_B0_WIDTH 1
+#define XG_MNT_INTR_B0_LBN 11
+#define XG_MNT_INTR_B0_WIDTH 1
+
+#define RX_RECOVERY_A1_LBN 11
+#define RX_RECOVERY_A1_WIDTH 1
+
+#define XFP_PHY_INTR_LBN 10
+#define XFP_PHY_INTR_WIDTH 1
+#define XG_PHY_INTR_LBN 9
+#define XG_PHY_INTR_WIDTH 1
+#define G_PHY1_INTR_LBN 8
+#define G_PHY1_INTR_WIDTH 1
+#define G_PHY0_INTR_LBN 7
+#define G_PHY0_INTR_WIDTH 1
+
+/* Driver-generated test events */
+#define DRV_GEN_EV_CODE_LBN 60
+#define DRV_GEN_EV_CODE_WIDTH 4
+#define DRV_GEN_EV_DATA_LBN 0
+#define DRV_GEN_EV_DATA_WIDTH 60
+#define EVQ_MAGIC_LBN 0
+#define EVQ_MAGIC_WIDTH 32
+
+/**************************************************************************
+ *
+ * Falcon MAC stats
+ *
+ **************************************************************************
+ *
+ */
+
+#define GRxGoodOct_offset 0x0
+#define GRxGoodOct_WIDTH 48
+#define GRxBadOct_offset 0x8
+#define GRxBadOct_WIDTH 48
+#define GRxMissPkt_offset 0x10
+#define GRxMissPkt_WIDTH 32
+#define GRxFalseCRS_offset 0x14
+#define GRxFalseCRS_WIDTH 32
+#define GRxPausePkt_offset 0x18
+#define GRxPausePkt_WIDTH 32
+#define GRxBadPkt_offset 0x1C
+#define GRxBadPkt_WIDTH 32
+#define GRxUcastPkt_offset 0x20
+#define GRxUcastPkt_WIDTH 32
+#define GRxMcastPkt_offset 0x24
+#define GRxMcastPkt_WIDTH 32
+#define GRxBcastPkt_offset 0x28
+#define GRxBcastPkt_WIDTH 32
+#define GRxGoodLt64Pkt_offset 0x2C
+#define GRxGoodLt64Pkt_WIDTH 32
+#define GRxBadLt64Pkt_offset 0x30
+#define GRxBadLt64Pkt_WIDTH 32
+#define GRx64Pkt_offset 0x34
+#define GRx64Pkt_WIDTH 32
+#define GRx65to127Pkt_offset 0x38
+#define GRx65to127Pkt_WIDTH 32
+#define GRx128to255Pkt_offset 0x3C
+#define GRx128to255Pkt_WIDTH 32
+#define GRx256to511Pkt_offset 0x40
+#define GRx256to511Pkt_WIDTH 32
+#define GRx512to1023Pkt_offset 0x44
+#define GRx512to1023Pkt_WIDTH 32
+#define GRx1024to15xxPkt_offset 0x48
+#define GRx1024to15xxPkt_WIDTH 32
+#define GRx15xxtoJumboPkt_offset 0x4C
+#define GRx15xxtoJumboPkt_WIDTH 32
+#define GRxGtJumboPkt_offset 0x50
+#define GRxGtJumboPkt_WIDTH 32
+#define GRxFcsErr64to15xxPkt_offset 0x54
+#define GRxFcsErr64to15xxPkt_WIDTH 32
+#define GRxFcsErr15xxtoJumboPkt_offset 0x58
+#define GRxFcsErr15xxtoJumboPkt_WIDTH 32
+#define GRxFcsErrGtJumboPkt_offset 0x5C
+#define GRxFcsErrGtJumboPkt_WIDTH 32
+#define GTxGoodBadOct_offset 0x80
+#define GTxGoodBadOct_WIDTH 48
+#define GTxGoodOct_offset 0x88
+#define GTxGoodOct_WIDTH 48
+#define GTxSglColPkt_offset 0x90
+#define GTxSglColPkt_WIDTH 32
+#define GTxMultColPkt_offset 0x94
+#define GTxMultColPkt_WIDTH 32
+#define GTxExColPkt_offset 0x98
+#define GTxExColPkt_WIDTH 32
+#define GTxDefPkt_offset 0x9C
+#define GTxDefPkt_WIDTH 32
+#define GTxLateCol_offset 0xA0
+#define GTxLateCol_WIDTH 32
+#define GTxExDefPkt_offset 0xA4
+#define GTxExDefPkt_WIDTH 32
+#define GTxPausePkt_offset 0xA8
+#define GTxPausePkt_WIDTH 32
+#define GTxBadPkt_offset 0xAC
+#define GTxBadPkt_WIDTH 32
+#define GTxUcastPkt_offset 0xB0
+#define GTxUcastPkt_WIDTH 32
+#define GTxMcastPkt_offset 0xB4
+#define GTxMcastPkt_WIDTH 32
+#define GTxBcastPkt_offset 0xB8
+#define GTxBcastPkt_WIDTH 32
+#define GTxLt64Pkt_offset 0xBC
+#define GTxLt64Pkt_WIDTH 32
+#define GTx64Pkt_offset 0xC0
+#define GTx64Pkt_WIDTH 32
+#define GTx65to127Pkt_offset 0xC4
+#define GTx65to127Pkt_WIDTH 32
+#define GTx128to255Pkt_offset 0xC8
+#define GTx128to255Pkt_WIDTH 32
+#define GTx256to511Pkt_offset 0xCC
+#define GTx256to511Pkt_WIDTH 32
+#define GTx512to1023Pkt_offset 0xD0
+#define GTx512to1023Pkt_WIDTH 32
+#define GTx1024to15xxPkt_offset 0xD4
+#define GTx1024to15xxPkt_WIDTH 32
+#define GTx15xxtoJumboPkt_offset 0xD8
+#define GTx15xxtoJumboPkt_WIDTH 32
+#define GTxGtJumboPkt_offset 0xDC
+#define GTxGtJumboPkt_WIDTH 32
+#define GTxNonTcpUdpPkt_offset 0xE0
+#define GTxNonTcpUdpPkt_WIDTH 16
+#define GTxMacSrcErrPkt_offset 0xE4
+#define GTxMacSrcErrPkt_WIDTH 16
+#define GTxIpSrcErrPkt_offset 0xE8
+#define GTxIpSrcErrPkt_WIDTH 16
+#define GDmaDone_offset 0xEC
+#define GDmaDone_WIDTH 32
+
+#define XgRxOctets_offset 0x0
+#define XgRxOctets_WIDTH 48
+#define XgRxOctetsOK_offset 0x8
+#define XgRxOctetsOK_WIDTH 48
+#define XgRxPkts_offset 0x10
+#define XgRxPkts_WIDTH 32
+#define XgRxPktsOK_offset 0x14
+#define XgRxPktsOK_WIDTH 32
+#define XgRxBroadcastPkts_offset 0x18
+#define XgRxBroadcastPkts_WIDTH 32
+#define XgRxMulticastPkts_offset 0x1C
+#define XgRxMulticastPkts_WIDTH 32
+#define XgRxUnicastPkts_offset 0x20
+#define XgRxUnicastPkts_WIDTH 32
+#define XgRxUndersizePkts_offset 0x24
+#define XgRxUndersizePkts_WIDTH 32
+#define XgRxOversizePkts_offset 0x28
+#define XgRxOversizePkts_WIDTH 32
+#define XgRxJabberPkts_offset 0x2C
+#define XgRxJabberPkts_WIDTH 32
+#define XgRxUndersizeFCSerrorPkts_offset 0x30
+#define XgRxUndersizeFCSerrorPkts_WIDTH 32
+#define XgRxDropEvents_offset 0x34
+#define XgRxDropEvents_WIDTH 32
+#define XgRxFCSerrorPkts_offset 0x38
+#define XgRxFCSerrorPkts_WIDTH 32
+#define XgRxAlignError_offset 0x3C
+#define XgRxAlignError_WIDTH 32
+#define XgRxSymbolError_offset 0x40
+#define XgRxSymbolError_WIDTH 32
+#define XgRxInternalMACError_offset 0x44
+#define XgRxInternalMACError_WIDTH 32
+#define XgRxControlPkts_offset 0x48
+#define XgRxControlPkts_WIDTH 32
+#define XgRxPausePkts_offset 0x4C
+#define XgRxPausePkts_WIDTH 32
+#define XgRxPkts64Octets_offset 0x50
+#define XgRxPkts64Octets_WIDTH 32
+#define XgRxPkts65to127Octets_offset 0x54
+#define XgRxPkts65to127Octets_WIDTH 32
+#define XgRxPkts128to255Octets_offset 0x58
+#define XgRxPkts128to255Octets_WIDTH 32
+#define XgRxPkts256to511Octets_offset 0x5C
+#define XgRxPkts256to511Octets_WIDTH 32
+#define XgRxPkts512to1023Octets_offset 0x60
+#define XgRxPkts512to1023Octets_WIDTH 32
+#define XgRxPkts1024to15xxOctets_offset 0x64
+#define XgRxPkts1024to15xxOctets_WIDTH 32
+#define XgRxPkts15xxtoMaxOctets_offset 0x68
+#define XgRxPkts15xxtoMaxOctets_WIDTH 32
+#define XgRxLengthError_offset 0x6C
+#define XgRxLengthError_WIDTH 32
+#define XgTxPkts_offset 0x80
+#define XgTxPkts_WIDTH 32
+#define XgTxOctets_offset 0x88
+#define XgTxOctets_WIDTH 48
+#define XgTxMulticastPkts_offset 0x90
+#define XgTxMulticastPkts_WIDTH 32
+#define XgTxBroadcastPkts_offset 0x94
+#define XgTxBroadcastPkts_WIDTH 32
+#define XgTxUnicastPkts_offset 0x98
+#define XgTxUnicastPkts_WIDTH 32
+#define XgTxControlPkts_offset 0x9C
+#define XgTxControlPkts_WIDTH 32
+#define XgTxPausePkts_offset 0xA0
+#define XgTxPausePkts_WIDTH 32
+#define XgTxPkts64Octets_offset 0xA4
+#define XgTxPkts64Octets_WIDTH 32
+#define XgTxPkts65to127Octets_offset 0xA8
+#define XgTxPkts65to127Octets_WIDTH 32
+#define XgTxPkts128to255Octets_offset 0xAC
+#define XgTxPkts128to255Octets_WIDTH 32
+#define XgTxPkts256to511Octets_offset 0xB0
+#define XgTxPkts256to511Octets_WIDTH 32
+#define XgTxPkts512to1023Octets_offset 0xB4
+#define XgTxPkts512to1023Octets_WIDTH 32
+#define XgTxPkts1024to15xxOctets_offset 0xB8
+#define XgTxPkts1024to15xxOctets_WIDTH 32
+#define XgTxPkts1519toMaxOctets_offset 0xBC
+#define XgTxPkts1519toMaxOctets_WIDTH 32
+#define XgTxUndersizePkts_offset 0xC0
+#define XgTxUndersizePkts_WIDTH 32
+#define XgTxOversizePkts_offset 0xC4
+#define XgTxOversizePkts_WIDTH 32
+#define XgTxNonTcpUdpPkt_offset 0xC8
+#define XgTxNonTcpUdpPkt_WIDTH 16
+#define XgTxMacSrcErrPkt_offset 0xCC
+#define XgTxMacSrcErrPkt_WIDTH 16
+#define XgTxIpSrcErrPkt_offset 0xD0
+#define XgTxIpSrcErrPkt_WIDTH 16
+#define XgDmaDone_offset 0xD4
+#define XgDmaDone_WIDTH 32
+
+#define FALCON_STATS_NOT_DONE 0x00000000
+#define FALCON_STATS_DONE 0xffffffff
+
+/* Interrupt status register bits */
+#define FATAL_INT_LBN 64
+#define FATAL_INT_WIDTH 1
+#define INT_EVQS_LBN 40
+#define INT_EVQS_WIDTH 4
+#define INT_FLAG_LBN 32
+#define INT_FLAG_WIDTH 1
+#define EVQ_FIFO_HF_LBN 1
+#define EVQ_FIFO_HF_WIDTH 1
+#define EVQ_FIFO_AF_LBN 0
+#define EVQ_FIFO_AF_WIDTH 1
+
+/**************************************************************************
+ *
+ * Falcon non-volatile configuration
+ *
+ **************************************************************************
+ */
+
+/* Board configuration v2 (v1 is obsolete; later versions are compatible) */
+struct falcon_nvconfig_board_v2 {
+	__le16 nports;
+	u8 port0_phy_addr;
+	u8 port0_phy_type;
+	u8 port1_phy_addr;
+	u8 port1_phy_type;
+	__le16 asic_sub_revision;
+	__le16 board_revision;
+} __attribute__ ((packed));
+
+/* Board configuration v3 extra information */
+struct falcon_nvconfig_board_v3 {
+	__le32 spi_device_type[2];
+} __attribute__ ((packed));
+
+/* Bit numbers for spi_device_type */
+#define SPI_DEV_TYPE_SIZE_LBN 0
+#define SPI_DEV_TYPE_SIZE_WIDTH 5
+#define SPI_DEV_TYPE_ADDR_LEN_LBN 6
+#define SPI_DEV_TYPE_ADDR_LEN_WIDTH 2
+#define SPI_DEV_TYPE_ERASE_CMD_LBN 8
+#define SPI_DEV_TYPE_ERASE_CMD_WIDTH 8
+#define SPI_DEV_TYPE_ERASE_SIZE_LBN 16
+#define SPI_DEV_TYPE_ERASE_SIZE_WIDTH 5
+#define SPI_DEV_TYPE_BLOCK_SIZE_LBN 24
+#define SPI_DEV_TYPE_BLOCK_SIZE_WIDTH 5
+#define SPI_DEV_TYPE_FIELD(type, field) \
+	(((type) >> EFX_LOW_BIT(field)) & EFX_MASK32(field))
+
+#define NVCONFIG_BASE 0x300
+#define NVCONFIG_BOARD_MAGIC_NUM 0xFA1C
+struct falcon_nvconfig {
+	efx_oword_t ee_vpd_cfg_reg;			/* 0x300 */
+	u8 mac_address[2][8];			/* 0x310 */
+	efx_oword_t pcie_sd_ctl0123_reg;		/* 0x320 */
+	efx_oword_t pcie_sd_ctl45_reg;			/* 0x330 */
+	efx_oword_t pcie_pcs_ctl_stat_reg;		/* 0x340 */
+	efx_oword_t hw_init_reg;			/* 0x350 */
+	efx_oword_t nic_stat_reg;			/* 0x360 */
+	efx_oword_t glb_ctl_reg;			/* 0x370 */
+	efx_oword_t srm_cfg_reg;			/* 0x380 */
+	efx_oword_t spare_reg;				/* 0x390 */
+	__le16 board_magic_num;			/* 0x3A0 */
+	__le16 board_struct_ver;
+	__le16 board_checksum;
+	struct falcon_nvconfig_board_v2 board_v2;
+	efx_oword_t ee_base_page_reg;			/* 0x3B0 */
+	struct falcon_nvconfig_board_v3 board_v3;	/* 0x3C0 */
+} __attribute__ ((packed));
+
+#endif /* EFX_FALCON_HWDEFS_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/falcon_io.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/falcon_io.h
--- linux-2.6.18.8/drivers/net/sfc/falcon_io.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/falcon_io.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,259 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_FALCON_IO_H
+#define EFX_FALCON_IO_H
+
+#include "net_driver.h"
+#include "falcon.h"
+
+/**************************************************************************
+ *
+ * Falcon hardware access
+ *
+ **************************************************************************
+ *
+ * Notes on locking strategy:
+ *
+ * Most Falcon registers require 16-byte (or 8-byte, for SRAM
+ * registers) atomic writes which necessitates locking.
+ * Under normal operation few writes to the Falcon BAR are made and these
+ * registers (EVQ_RPTR_REG, RX_DESC_UPD_REG and TX_DESC_UPD_REG) are special
+ * cased to allow 4-byte (hence lockless) accesses.
+ *
+ * It *is* safe to write to these 4-byte registers in the middle of an
+ * access to an 8-byte or 16-byte register.  We therefore use a
+ * spinlock to protect accesses to the larger registers, but no locks
+ * for the 4-byte registers.
+ *
+ * A write barrier is needed to ensure that DW3 is written after DW0/1/2
+ * due to the way the 16byte registers are "collected" in the Falcon BIU
+ *
+ * We also lock when carrying out reads, to ensure consistency of the
+ * data (made possible since the BIU reads all 128 bits into a cache).
+ * Reads are very rare, so this isn't a significant performance
+ * impact.  (Most data transferred from NIC to host is DMAed directly
+ * into host memory).
+ *
+ * I/O BAR access uses locks for both reads and writes (but is only provided
+ * for testing purposes).
+ */
+
+/* Special buffer descriptors (Falcon SRAM) */
+#define BUF_TBL_KER_A1 0x18000
+#define BUF_TBL_KER_B0 0x800000
+
+
+#if BITS_PER_LONG == 64
+#define FALCON_USE_QWORD_IO 1
+#endif
+
+#define _falcon_writeq(efx, value, reg) \
+	__raw_writeq((__force u64) (value), (efx)->membase + (reg))
+#define _falcon_writel(efx, value, reg) \
+	__raw_writel((__force u32) (value), (efx)->membase + (reg))
+#define _falcon_readq(efx, reg) \
+	((__force __le64) __raw_readq((efx)->membase + (reg)))
+#define _falcon_readl(efx, reg) \
+	((__force __le32) __raw_readl((efx)->membase + (reg)))
+
+/* Writes to a normal 16-byte Falcon register, locking as appropriate. */
+static inline void falcon_write(struct efx_nic *efx, efx_oword_t *value,
+				unsigned int reg)
+{
+	unsigned long flags __attribute__ ((unused));
+
+	EFX_REGDUMP(efx, "writing register %x with " EFX_OWORD_FMT "\n", reg,
+		    EFX_OWORD_VAL(*value));
+
+	spin_lock_irqsave(&efx->biu_lock, flags);
+#ifdef FALCON_USE_QWORD_IO
+	_falcon_writeq(efx, value->u64[0], reg + 0);
+	wmb();
+	_falcon_writeq(efx, value->u64[1], reg + 8);
+#else
+	_falcon_writel(efx, value->u32[0], reg + 0);
+	_falcon_writel(efx, value->u32[1], reg + 4);
+	_falcon_writel(efx, value->u32[2], reg + 8);
+	wmb();
+	_falcon_writel(efx, value->u32[3], reg + 12);
+#endif
+	mmiowb();
+	spin_unlock_irqrestore(&efx->biu_lock, flags);
+}
+
+/* Writes to an 8-byte Falcon SRAM register, locking as appropriate. */
+static inline void falcon_write_sram(struct efx_nic *efx, efx_qword_t *value,
+				     unsigned int index)
+{
+	unsigned int reg = efx->type->buf_tbl_base + (index * sizeof(*value));
+	unsigned long flags __attribute__ ((unused));
+
+	EFX_REGDUMP(efx, "writing SRAM register %x with " EFX_QWORD_FMT "\n",
+		    reg, EFX_QWORD_VAL(*value));
+
+	spin_lock_irqsave(&efx->biu_lock, flags);
+#ifdef FALCON_USE_QWORD_IO
+	_falcon_writeq(efx, value->u64[0], reg + 0);
+#else
+	_falcon_writel(efx, value->u32[0], reg + 0);
+	wmb();
+	_falcon_writel(efx, value->u32[1], reg + 4);
+#endif
+	mmiowb();
+	spin_unlock_irqrestore(&efx->biu_lock, flags);
+}
+
+/* Write dword to Falcon register that allows partial writes
+ *
+ * Some Falcon registers (EVQ_RPTR_REG, RX_DESC_UPD_REG and
+ * TX_DESC_UPD_REG) can be written to as a single dword.  This allows
+ * for lockless writes.
+ */
+static inline void falcon_writel(struct efx_nic *efx, efx_dword_t *value,
+				 unsigned int reg)
+{
+	EFX_REGDUMP(efx, "writing partial register %x with "EFX_DWORD_FMT"\n",
+		    reg, EFX_DWORD_VAL(*value));
+
+	/* No lock required */
+	_falcon_writel(efx, value->u32[0], reg);
+}
+
+/* Read from a Falcon register
+ *
+ * This reads an entire 16-byte Falcon register in one go, locking as
+ * appropriate.  It is essential to read the first dword first, as this
+ * prompts Falcon to load the current value into the shadow register.
+ */
+static inline void falcon_read(struct efx_nic *efx, efx_oword_t *value,
+			       unsigned int reg)
+{
+	unsigned long flags __attribute__ ((unused));
+
+	spin_lock_irqsave(&efx->biu_lock, flags);
+	value->u32[0] = _falcon_readl(efx, reg + 0);
+	rmb();
+	value->u32[1] = _falcon_readl(efx, reg + 4);
+	value->u32[2] = _falcon_readl(efx, reg + 8);
+	value->u32[3] = _falcon_readl(efx, reg + 12);
+	spin_unlock_irqrestore(&efx->biu_lock, flags);
+
+	EFX_REGDUMP(efx, "read from register %x, got " EFX_OWORD_FMT "\n", reg,
+		    EFX_OWORD_VAL(*value));
+}
+
+/* This reads an 8-byte Falcon SRAM entry in one go. */
+static inline void falcon_read_sram(struct efx_nic *efx, efx_qword_t *value,
+				    unsigned int index)
+{
+	unsigned int reg = efx->type->buf_tbl_base + (index * sizeof(*value));
+	unsigned long flags __attribute__ ((unused));
+
+	spin_lock_irqsave(&efx->biu_lock, flags);
+#ifdef FALCON_USE_QWORD_IO
+	value->u64[0] = _falcon_readq(efx, reg + 0);
+#else
+	value->u32[0] = _falcon_readl(efx, reg + 0);
+	rmb();
+	value->u32[1] = _falcon_readl(efx, reg + 4);
+#endif
+	spin_unlock_irqrestore(&efx->biu_lock, flags);
+
+	EFX_REGDUMP(efx, "read from SRAM register %x, got "EFX_QWORD_FMT"\n",
+		    reg, EFX_QWORD_VAL(*value));
+}
+
+/* Read dword from Falcon register that allows partial writes (sic) */
+static inline void falcon_readl(struct efx_nic *efx, efx_dword_t *value,
+				unsigned int reg)
+{
+	value->u32[0] = _falcon_readl(efx, reg);
+	EFX_REGDUMP(efx, "read from register %x, got "EFX_DWORD_FMT"\n",
+		    reg, EFX_DWORD_VAL(*value));
+}
+
+/* Write to a register forming part of a table */
+static inline void falcon_write_table(struct efx_nic *efx, efx_oword_t *value,
+				      unsigned int reg, unsigned int index)
+{
+	falcon_write(efx, value, reg + index * sizeof(efx_oword_t));
+}
+
+/* Read to a register forming part of a table */
+static inline void falcon_read_table(struct efx_nic *efx, efx_oword_t *value,
+				     unsigned int reg, unsigned int index)
+{
+	falcon_read(efx, value, reg + index * sizeof(efx_oword_t));
+}
+
+/* Write to a dword register forming part of a table */
+static inline void falcon_writel_table(struct efx_nic *efx, efx_dword_t *value,
+				       unsigned int reg, unsigned int index)
+{
+	falcon_writel(efx, value, reg + index * sizeof(efx_oword_t));
+}
+
+/* Page-mapped register block size */
+#define FALCON_PAGE_BLOCK_SIZE 0x2000
+
+/* Calculate offset to page-mapped register block */
+#define FALCON_PAGED_REG(page, reg) \
+	((page) * FALCON_PAGE_BLOCK_SIZE + (reg))
+
+/* As for falcon_write(), but for a page-mapped register. */
+static inline void falcon_write_page(struct efx_nic *efx, efx_oword_t *value,
+				     unsigned int reg, unsigned int page)
+{
+	falcon_write(efx, value, FALCON_PAGED_REG(page, reg));
+}
+
+/* As for falcon_writel(), but for a page-mapped register. */
+static inline void falcon_writel_page(struct efx_nic *efx, efx_dword_t *value,
+				      unsigned int reg, unsigned int page)
+{
+	falcon_writel(efx, value, FALCON_PAGED_REG(page, reg));
+}
+
+/* Write dword to Falcon page-mapped register with an extra lock.
+ *
+ * As for falcon_writel_page(), but for a register that suffers from
+ * SFC bug 3181. Take out a lock so the BIU collector cannot be
+ * confused. */
+static inline void falcon_writel_page_locked(struct efx_nic *efx,
+					     efx_dword_t *value,
+					     unsigned int reg,
+					     unsigned int page)
+{
+	unsigned long flags __attribute__ ((unused));
+
+	spin_lock_irqsave(&efx->biu_lock, flags);
+	falcon_writel(efx, value, FALCON_PAGED_REG(page, reg));
+	spin_unlock_irqrestore(&efx->biu_lock, flags);
+}
+
+#endif /* EFX_FALCON_IO_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/falcon_xmac.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/falcon_xmac.c
--- linux-2.6.18.8/drivers/net/sfc/falcon_xmac.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/falcon_xmac.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,691 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/delay.h>
+#include "net_driver.h"
+#include "efx.h"
+#include "falcon.h"
+#include "falcon_hwdefs.h"
+#include "falcon_io.h"
+#include "mac.h"
+#include "gmii.h"
+#include "mdio_10g.h"
+#include "phy.h"
+#include "boards.h"
+#include "workarounds.h"
+
+/**************************************************************************
+ *
+ * MAC register access
+ *
+ **************************************************************************/
+
+/* Offset of an XMAC register within Falcon */
+#define FALCON_XMAC_REG(mac_reg)					\
+	(FALCON_XMAC_REGBANK + ((mac_reg) * FALCON_XMAC_REG_SIZE))
+
+static void falcon_xmac_writel(struct efx_nic *efx,
+			       efx_dword_t *value, unsigned int mac_reg)
+{
+	efx_oword_t temp;
+
+	EFX_POPULATE_OWORD_1(temp, MAC_DATA, EFX_DWORD_FIELD(*value, MAC_DATA));
+	falcon_write(efx, &temp, FALCON_XMAC_REG(mac_reg));
+}
+
+static void falcon_xmac_readl(struct efx_nic *efx,
+			      efx_dword_t *value, unsigned int mac_reg)
+{
+	efx_oword_t temp;
+
+	falcon_read(efx, &temp, FALCON_XMAC_REG(mac_reg));
+	EFX_POPULATE_DWORD_1(*value, MAC_DATA, EFX_OWORD_FIELD(temp, MAC_DATA));
+}
+
+/**************************************************************************
+ *
+ * MAC operations
+ *
+ *************************************************************************/
+static int falcon_reset_xmac(struct efx_nic *efx)
+{
+	efx_dword_t reg;
+	int count;
+
+	EFX_POPULATE_DWORD_1(reg, XM_CORE_RST, 1);
+	efx->mac_op->mac_writel(efx, &reg, XM_GLB_CFG_REG_MAC);
+
+	for (count = 0; count < 10000; count++) {	/* wait upto 100ms */
+		efx->mac_op->mac_readl(efx, &reg, XM_GLB_CFG_REG_MAC);
+		if (EFX_DWORD_FIELD(reg, XM_CORE_RST) == 0)
+			return 0;
+		udelay(10);
+	}
+
+	EFX_ERR(efx, "timed out waiting for XMAC core reset\n");
+	return -ETIMEDOUT;
+}
+
+/* Configure the XAUI driver that is an output from Falcon */
+static void falcon_setup_xaui(struct efx_nic *efx)
+{
+	efx_dword_t sdctl, txdrv;
+
+	/* Move the XAUI into low power, unless there is no PHY, in
+	 * which case the XAUI will have to drive a cable. */
+	if (efx->phy_type == PHY_TYPE_NONE)
+		return;
+
+	efx->mac_op->mac_readl(efx, &sdctl, XX_SD_CTL_REG_MAC);
+	EFX_SET_DWORD_FIELD(sdctl, XX_HIDRVD, XX_SD_CTL_DRV_DEFAULT);
+	EFX_SET_DWORD_FIELD(sdctl, XX_LODRVD, XX_SD_CTL_DRV_DEFAULT);
+	EFX_SET_DWORD_FIELD(sdctl, XX_HIDRVC, XX_SD_CTL_DRV_DEFAULT);
+	EFX_SET_DWORD_FIELD(sdctl, XX_LODRVC, XX_SD_CTL_DRV_DEFAULT);
+	EFX_SET_DWORD_FIELD(sdctl, XX_HIDRVB, XX_SD_CTL_DRV_DEFAULT);
+	EFX_SET_DWORD_FIELD(sdctl, XX_LODRVB, XX_SD_CTL_DRV_DEFAULT);
+	EFX_SET_DWORD_FIELD(sdctl, XX_HIDRVA, XX_SD_CTL_DRV_DEFAULT);
+	EFX_SET_DWORD_FIELD(sdctl, XX_LODRVA, XX_SD_CTL_DRV_DEFAULT);
+	efx->mac_op->mac_writel(efx, &sdctl, XX_SD_CTL_REG_MAC);
+
+	EFX_POPULATE_DWORD_8(txdrv,
+			     XX_DEQD, XX_TXDRV_DEQ_DEFAULT,
+			     XX_DEQC, XX_TXDRV_DEQ_DEFAULT,
+			     XX_DEQB, XX_TXDRV_DEQ_DEFAULT,
+			     XX_DEQA, XX_TXDRV_DEQ_DEFAULT,
+			     XX_DTXD, XX_TXDRV_DTX_DEFAULT,
+			     XX_DTXC, XX_TXDRV_DTX_DEFAULT,
+			     XX_DTXB, XX_TXDRV_DTX_DEFAULT,
+			     XX_DTXA, XX_TXDRV_DTX_DEFAULT);
+	efx->mac_op->mac_writel(efx, &txdrv, XX_TXDRV_CTL_REG_MAC);
+}
+
+static void falcon_hold_xaui_in_rst(struct efx_nic *efx)
+{
+	efx_dword_t reg;
+
+	EFX_ZERO_DWORD(reg);
+	EFX_SET_DWORD_FIELD(reg, XX_PWRDNA_EN, 1);
+	EFX_SET_DWORD_FIELD(reg, XX_PWRDNB_EN, 1);
+	EFX_SET_DWORD_FIELD(reg, XX_PWRDNC_EN, 1);
+	EFX_SET_DWORD_FIELD(reg, XX_PWRDND_EN, 1);
+	EFX_SET_DWORD_FIELD(reg, XX_RSTPLLAB_EN, 1);
+	EFX_SET_DWORD_FIELD(reg, XX_RSTPLLCD_EN, 1);
+	EFX_SET_DWORD_FIELD(reg, XX_RESETA_EN, 1);
+	EFX_SET_DWORD_FIELD(reg, XX_RESETB_EN, 1);
+	EFX_SET_DWORD_FIELD(reg, XX_RESETC_EN, 1);
+	EFX_SET_DWORD_FIELD(reg, XX_RESETD_EN, 1);
+	EFX_SET_DWORD_FIELD(reg, XX_RSTXGXSRX_EN, 1);
+	EFX_SET_DWORD_FIELD(reg, XX_RSTXGXSTX_EN, 1);
+	efx->mac_op->mac_writel(efx, &reg, XX_PWR_RST_REG_MAC);
+	udelay(10);
+}
+
+static int _falcon_reset_xaui_a(struct efx_nic *efx)
+{
+	efx_dword_t reg;
+
+	if (!efx->is_asic)
+		return 0;
+
+	falcon_hold_xaui_in_rst(efx);
+	efx->mac_op->mac_readl(efx, &reg, XX_PWR_RST_REG_MAC);
+
+	/* Follow the RAMBUS XAUI data reset sequencing
+	 * Channels A and B first: power down, reset PLL, reset, clear
+	 */
+	EFX_SET_DWORD_FIELD(reg, XX_PWRDNA_EN, 0);
+	EFX_SET_DWORD_FIELD(reg, XX_PWRDNB_EN, 0);
+	efx->mac_op->mac_writel(efx, &reg, XX_PWR_RST_REG_MAC);
+	udelay(10);
+
+	EFX_SET_DWORD_FIELD(reg, XX_RSTPLLAB_EN, 0);
+	efx->mac_op->mac_writel(efx, &reg, XX_PWR_RST_REG_MAC);
+	udelay(10);
+
+	EFX_SET_DWORD_FIELD(reg, XX_RESETA_EN, 0);
+	EFX_SET_DWORD_FIELD(reg, XX_RESETB_EN, 0);
+	efx->mac_op->mac_writel(efx, &reg, XX_PWR_RST_REG_MAC);
+	udelay(10);
+
+	/* Channels C and D: power down, reset PLL, reset, clear */
+	EFX_SET_DWORD_FIELD(reg, XX_PWRDNC_EN, 0);
+	EFX_SET_DWORD_FIELD(reg, XX_PWRDND_EN, 0);
+	efx->mac_op->mac_writel(efx, &reg, XX_PWR_RST_REG_MAC);
+	udelay(10);
+
+	EFX_SET_DWORD_FIELD(reg, XX_RSTPLLCD_EN, 0);
+	efx->mac_op->mac_writel(efx, &reg, XX_PWR_RST_REG_MAC);
+	udelay(10);
+
+	EFX_SET_DWORD_FIELD(reg, XX_RESETC_EN, 0);
+	EFX_SET_DWORD_FIELD(reg, XX_RESETD_EN, 0);
+	efx->mac_op->mac_writel(efx, &reg, XX_PWR_RST_REG_MAC);
+	udelay(10);
+
+	/* Setup XAUI */
+	falcon_setup_xaui(efx);
+	udelay(10);
+
+	/* Take XGXS out of reset */
+	EFX_ZERO_DWORD(reg);
+	efx->mac_op->mac_writel(efx, &reg, XX_PWR_RST_REG_MAC);
+	udelay(10);
+
+	return 0;
+}
+
+static int _falcon_reset_xaui_b(struct efx_nic *efx)
+{
+	efx_dword_t reg;
+	int count;
+
+	if (!efx->is_asic)
+		return 0;
+
+	EFX_POPULATE_DWORD_1(reg, XX_RST_XX_EN, 1);
+	efx->mac_op->mac_writel(efx, &reg, XX_PWR_RST_REG_MAC);
+
+	/* Give some time for the link to establish */
+	for (count = 0; count < 1000; count++) { /* wait upto 10ms */
+		efx->mac_op->mac_readl(efx, &reg, XX_PWR_RST_REG_MAC);
+		if (EFX_DWORD_FIELD(reg, XX_RST_XX_EN) == 0) {
+			falcon_setup_xaui(efx);
+			return 0;
+		}
+		udelay(10);
+	}
+	EFX_ERR(efx, "timed out waiting for XAUI/XGXS reset\n");
+	return -ETIMEDOUT;
+}
+
+int falcon_reset_xaui(struct efx_nic *efx)
+{
+	int rc;
+
+	if (EFX_WORKAROUND_9388(efx)) {
+		falcon_hold_xaui_in_rst(efx);
+		efx->phy_op->reset_xaui(efx);
+		rc = _falcon_reset_xaui_a(efx);
+	} else {
+		rc = _falcon_reset_xaui_b(efx);
+	}
+	return rc;
+}
+
+static int falcon_init_xmac(struct efx_nic *efx)
+{
+	int rc;
+
+	/* Initialize the PHY first so the clock is around */
+	rc = efx->phy_op->init(efx);
+	if (rc)
+		goto fail1;
+
+	rc = falcon_reset_xaui(efx);
+	if (rc)
+		goto fail2;
+
+	/* Wait again. Give the PHY and MAC time to faff */
+	schedule_timeout_uninterruptible(HZ / 10);
+
+	/* Reset the MAC */
+	rc = falcon_reset_xmac(efx);
+	if (rc)
+		goto fail2;
+
+	return 0;
+
+ fail2:
+	efx->phy_op->fini(efx);
+ fail1:
+	return rc;
+}
+
+/* Get status of XAUI link */
+int falcon_xaui_link_ok(struct efx_nic *efx)
+{
+	efx_dword_t reg;
+	int align_done;
+	int sync_status;
+	int link_ok = 0;
+
+	/* If we're in internal loopback, then the link is up.
+	 * The A1 FPGA/4G has RX and TX XAUI wired together, so the link is up.
+	 * The B0 FPGA has XAUI offchip, so it is always up.
+	 */
+	if (!efx->is_asic || LOOPBACK_INTERNAL(efx))
+		return 1;
+
+	/* Read link status */
+	efx->mac_op->mac_readl(efx, &reg, XX_CORE_STAT_REG_MAC);
+
+	align_done = EFX_DWORD_FIELD(reg, XX_ALIGN_DONE);
+	sync_status = EFX_DWORD_FIELD(reg, XX_SYNC_STAT);
+	if (align_done && (sync_status == XX_SYNC_STAT_DECODE_SYNCED))
+		link_ok = 1;
+
+	/* Clear link status ready for next read */
+	EFX_SET_DWORD_FIELD(reg, XX_COMMA_DET, XX_COMMA_DET_RESET);
+	EFX_SET_DWORD_FIELD(reg, XX_CHARERR, XX_CHARERR_RESET);
+	EFX_SET_DWORD_FIELD(reg, XX_DISPERR, XX_DISPERR_RESET);
+
+	efx->mac_op->mac_writel(efx, &reg, XX_CORE_STAT_REG_MAC);
+
+	return link_ok;
+}
+
+/* Do most of the heavy lifting of falcon_reconfigure_xmac */
+static void falcon_reconfigure_xmac_core(struct efx_nic *efx)
+{
+	unsigned int max_frame_len;
+	efx_dword_t reg;
+	efx_oword_t mac_test_reg;
+	int rx_fc = (efx->flow_control & EFX_FC_RX) ? 1 : 0;
+
+	if (FALCON_REV(efx) <= FALCON_REV_A1 && !efx->is_asic) {
+		/* 10G FPGA's have the XAUI TX and RX wired together. Fake
+		 * the link status and configure the link options before
+		 * the MAC wrapper is configured */
+		efx->link_options = GM_LPA_10000FULL;
+		efx->link_up = 1;
+	}
+
+	/* Configure MAC  - cut-thru mode is hard wired on */
+	EFX_POPULATE_DWORD_3(reg,
+			     XM_RX_JUMBO_MODE, 1,
+			     XM_TX_STAT_EN, 1,
+			     XM_RX_STAT_EN, 1);
+	efx->mac_op->mac_writel(efx, &reg, XM_GLB_CFG_REG_MAC);
+
+	/* Configure TX */
+	EFX_POPULATE_DWORD_6(reg,
+			     XM_TXEN, 1,
+			     XM_TX_PRMBL, 1,
+			     XM_AUTO_PAD, 1,
+			     XM_TXCRC, 1,
+			     XM_FCNTL, 1,
+			     XM_IPG, 0x3);
+	efx->mac_op->mac_writel(efx, &reg, XM_TX_CFG_REG_MAC);
+
+	/* Configure RX */
+	EFX_POPULATE_DWORD_5(reg,
+			     XM_RXEN, 1,
+			     XM_AUTO_DEPAD, 0,
+			     XM_ACPT_ALL_MCAST, 1,
+			     XM_ACPT_ALL_UCAST, efx->promiscuous,
+			     XM_PASS_CRC_ERR, 1);
+	efx->mac_op->mac_writel(efx, &reg, XM_RX_CFG_REG_MAC);
+
+	/* Set frame length */
+	max_frame_len = EFX_MAX_FRAME_LEN(efx->net_dev->mtu);
+	EFX_POPULATE_DWORD_1(reg, XM_MAX_RX_FRM_SIZE, max_frame_len);
+	efx->mac_op->mac_writel(efx, &reg, XM_RX_PARAM_REG_MAC);
+	EFX_POPULATE_DWORD_2(reg,
+			     XM_MAX_TX_FRM_SIZE, max_frame_len,
+			     XM_TX_JUMBO_MODE, 1);
+	efx->mac_op->mac_writel(efx, &reg, XM_TX_PARAM_REG_MAC);
+
+	EFX_POPULATE_DWORD_2(reg,
+			     XM_PAUSE_TIME, 0xfffe, /* MAX PAUSE TIME */
+			     XM_DIS_FCNTL, rx_fc ? 0 : 1);
+	efx->mac_op->mac_writel(efx, &reg, XM_FC_REG_MAC);
+
+	/* Set MAC address */
+	EFX_POPULATE_DWORD_4(reg,
+			     XM_ADR_0, efx->net_dev->dev_addr[0],
+			     XM_ADR_1, efx->net_dev->dev_addr[1],
+			     XM_ADR_2, efx->net_dev->dev_addr[2],
+			     XM_ADR_3, efx->net_dev->dev_addr[3]);
+	efx->mac_op->mac_writel(efx, &reg, XM_ADR_LO_REG_MAC);
+	EFX_POPULATE_DWORD_2(reg,
+			     XM_ADR_4, efx->net_dev->dev_addr[4],
+			     XM_ADR_5, efx->net_dev->dev_addr[5]);
+	efx->mac_op->mac_writel(efx, &reg, XM_ADR_HI_REG_MAC);
+
+	/* Handle B0 FPGA loopback where RAMBUS XGXS block not present */
+	if (FALCON_REV(efx) == FALCON_REV_B0 && !efx->is_asic) {
+		int xgmii_loopback =
+			(efx->loopback_mode == LOOPBACK_XGMII) ? 1 : 0;
+
+		/* Set the MAC loopback bit. */
+		EFX_POPULATE_OWORD_1(mac_test_reg,
+				     MAC_PTLOOP_EN, xgmii_loopback);
+		falcon_write(efx, &mac_test_reg, MAC_TEST_REG_KER);
+	}
+}
+
+/* Do most of the heavy lifting of falcon_reconfigure_xmac */
+static void falcon_reconfigure_xgxs_core(struct efx_nic *efx)
+{
+	efx_dword_t reg;
+	int xgxs_loopback = (efx->loopback_mode == LOOPBACK_XGXS) ? 1 : 0;
+	int xaui_loopback = (efx->loopback_mode == LOOPBACK_XAUI) ? 1 : 0;
+	int xgmii_loopback =
+		(efx->loopback_mode == LOOPBACK_XGMII) ? 1 : 0;
+
+	if (FALCON_REV(efx) == FALCON_REV_B0 && !efx->is_asic)
+		/* RAMBUS XGXS block is not present */
+		return;
+
+	/* XGXS block is flaky and will need to be reset if moving
+	 * into our out of XGMII, XGXS or XAUI loopbacks. */
+	if (EFX_WORKAROUND_5147(efx)) {
+		int old_xgmii_loopback, old_xgxs_loopback, old_xaui_loopback;
+		int reset_xgxs;
+
+		efx->mac_op->mac_readl(efx, &reg,
+					    XX_CORE_STAT_REG_MAC);
+		old_xgxs_loopback = EFX_DWORD_FIELD(reg, XX_XGXS_LB_EN);
+		old_xgmii_loopback = EFX_DWORD_FIELD(reg, XX_XGMII_LB_EN);
+
+		efx->mac_op->mac_readl(efx, &reg, XX_SD_CTL_REG_MAC);
+		old_xaui_loopback = EFX_DWORD_FIELD(reg, XX_LPBKA);
+
+		/* The PHY driver may have turned XAUI off */
+		reset_xgxs = ((xgxs_loopback != old_xgxs_loopback) ||
+			      (xaui_loopback != old_xaui_loopback) ||
+			      (xgmii_loopback != old_xgmii_loopback));
+		if (reset_xgxs) {
+			efx->mac_op->mac_readl(efx, &reg,
+						    XX_PWR_RST_REG_MAC);
+			EFX_SET_DWORD_FIELD(reg, XX_RSTXGXSTX_EN, 1);
+			EFX_SET_DWORD_FIELD(reg, XX_RSTXGXSRX_EN, 1);
+			efx->mac_op->mac_writel(efx, &reg,
+						     XX_PWR_RST_REG_MAC);
+			udelay(1);
+			EFX_SET_DWORD_FIELD(reg, XX_RSTXGXSTX_EN, 0);
+			EFX_SET_DWORD_FIELD(reg, XX_RSTXGXSRX_EN, 0);
+			efx->mac_op->mac_writel(efx, &reg,
+						     XX_PWR_RST_REG_MAC);
+			udelay(1);
+		}
+	}
+
+	efx->mac_op->mac_readl(efx, &reg, XX_CORE_STAT_REG_MAC);
+	EFX_SET_DWORD_FIELD(reg, XX_FORCE_SIG,
+			    (xgxs_loopback || xaui_loopback) ?
+			    XX_FORCE_SIG_DECODE_FORCED : 0);
+	EFX_SET_DWORD_FIELD(reg, XX_XGXS_LB_EN, xgxs_loopback);
+	EFX_SET_DWORD_FIELD(reg, XX_XGMII_LB_EN, xgmii_loopback);
+	efx->mac_op->mac_writel(efx, &reg, XX_CORE_STAT_REG_MAC);
+
+	efx->mac_op->mac_readl(efx, &reg, XX_SD_CTL_REG_MAC);
+	EFX_SET_DWORD_FIELD(reg, XX_LPBKD, xaui_loopback);
+	EFX_SET_DWORD_FIELD(reg, XX_LPBKC, xaui_loopback);
+	EFX_SET_DWORD_FIELD(reg, XX_LPBKB, xaui_loopback);
+	EFX_SET_DWORD_FIELD(reg, XX_LPBKA, xaui_loopback);
+	efx->mac_op->mac_writel(efx, &reg, XX_SD_CTL_REG_MAC);
+}
+
+
+/* Sometimes the XAUI link between Falcon and XFP fails to come up. The state
+ * of the link is checked during phy_reconfigure(). After XAIU is reset then
+ * the MAC must be reconfigured.
+ */
+#define MAX_XAUI_TRIES (5)	/* It's never been seen to take more than 2 */
+
+void falcon_check_xaui_link_up(struct efx_nic *efx)
+{
+	int max_tries, tries;
+	tries = EFX_WORKAROUND_5147(efx) ? MAX_XAUI_TRIES : 1;
+	max_tries = tries;
+
+	if ((efx->loopback_mode == LOOPBACK_NETWORK) ||
+	    (efx->phy_type == PHY_TYPE_NONE) ||
+	    !efx->phy_powered)
+		return;
+
+	while (tries) {
+		if (falcon_xaui_link_ok(efx))
+			return;
+
+		EFX_LOG(efx, "%s Clobbering XAUI (%d tries left).\n",
+			__func__, tries);
+		(void) falcon_reset_xaui(efx);
+		/* Cannot use full reconfigure. Need to avoid recursion */
+
+		/* Give the poor thing time to sort itself out: if we retry
+		 * too fast it will never train. */
+		udelay(200);
+
+		falcon_reconfigure_xgxs_core(efx);
+
+		tries--;
+	}
+
+	EFX_ERR(efx, "Failed to bring XAUI link back up in %d tries!\n",
+		max_tries);
+}
+
+static void falcon_reconfigure_xmac(struct efx_nic *efx)
+{
+	falcon_deconfigure_mac_wrapper(efx);
+
+	/* In internal loopback modes disable transmit */
+	efx->tx_disabled = LOOPBACK_INTERNAL(efx);
+
+	efx->phy_op->reconfigure(efx);
+
+	falcon_reconfigure_xgxs_core(efx);
+	falcon_reconfigure_xmac_core(efx);
+
+	/* Reconfigure MAC wrapper */
+	falcon_reconfigure_mac_wrapper(efx);
+
+	/* Ensure XAUI link is up - might repeat reconfigure_xmac_core */
+	falcon_check_xaui_link_up(efx);
+}
+
+static void falcon_fini_xmac(struct efx_nic *efx)
+{
+	/* Isolate the MAC - PHY */
+	falcon_deconfigure_mac_wrapper(efx);
+
+	/* Potentially power down the PHY */
+	efx->phy_op->fini(efx);
+}
+
+static void falcon_update_stats_xmac(struct efx_nic *efx)
+{
+	struct efx_mac_stats *mac_stats = &efx->mac_stats;
+	int rc;
+
+	rc = falcon_dma_stats(efx, XgDmaDone_offset);
+	if (rc)
+		return;
+
+	/* Update MAC stats from DMAed values */
+	FALCON_STAT(efx, XgRxOctets, rx_bytes);
+	FALCON_STAT(efx, XgRxOctetsOK, rx_good_bytes);
+	FALCON_STAT(efx, XgRxPkts, rx_packets);
+	FALCON_STAT(efx, XgRxPktsOK, rx_good);
+	FALCON_STAT(efx, XgRxBroadcastPkts, rx_broadcast);
+	FALCON_STAT(efx, XgRxMulticastPkts, rx_multicast);
+	FALCON_STAT(efx, XgRxUnicastPkts, rx_unicast);
+	FALCON_STAT(efx, XgRxUndersizePkts, rx_lt64);
+	FALCON_STAT(efx, XgRxOversizePkts, rx_gtjumbo);
+	FALCON_STAT(efx, XgRxJabberPkts, rx_bad_gtjumbo);
+	FALCON_STAT(efx, XgRxUndersizeFCSerrorPkts, rx_bad_lt64);
+	FALCON_STAT(efx, XgRxDropEvents, rx_overflow);
+	FALCON_STAT(efx, XgRxFCSerrorPkts, rx_bad);
+	FALCON_STAT(efx, XgRxAlignError, rx_align_error);
+	FALCON_STAT(efx, XgRxSymbolError, rx_symbol_error);
+	FALCON_STAT(efx, XgRxInternalMACError, rx_internal_error);
+	FALCON_STAT(efx, XgRxControlPkts, rx_control);
+	FALCON_STAT(efx, XgRxPausePkts, rx_pause);
+	FALCON_STAT(efx, XgRxPkts64Octets, rx_64);
+	FALCON_STAT(efx, XgRxPkts65to127Octets, rx_65_to_127);
+	FALCON_STAT(efx, XgRxPkts128to255Octets, rx_128_to_255);
+	FALCON_STAT(efx, XgRxPkts256to511Octets, rx_256_to_511);
+	FALCON_STAT(efx, XgRxPkts512to1023Octets, rx_512_to_1023);
+	FALCON_STAT(efx, XgRxPkts1024to15xxOctets, rx_1024_to_15xx);
+	FALCON_STAT(efx, XgRxPkts15xxtoMaxOctets, rx_15xx_to_jumbo);
+	FALCON_STAT(efx, XgRxLengthError, rx_length_error);
+	FALCON_STAT(efx, XgTxPkts, tx_packets);
+	FALCON_STAT(efx, XgTxOctets, tx_bytes);
+	FALCON_STAT(efx, XgTxMulticastPkts, tx_multicast);
+	FALCON_STAT(efx, XgTxBroadcastPkts, tx_broadcast);
+	FALCON_STAT(efx, XgTxUnicastPkts, tx_unicast);
+	FALCON_STAT(efx, XgTxControlPkts, tx_control);
+	FALCON_STAT(efx, XgTxPausePkts, tx_pause);
+	FALCON_STAT(efx, XgTxPkts64Octets, tx_64);
+	FALCON_STAT(efx, XgTxPkts65to127Octets, tx_65_to_127);
+	FALCON_STAT(efx, XgTxPkts128to255Octets, tx_128_to_255);
+	FALCON_STAT(efx, XgTxPkts256to511Octets, tx_256_to_511);
+	FALCON_STAT(efx, XgTxPkts512to1023Octets, tx_512_to_1023);
+	FALCON_STAT(efx, XgTxPkts1024to15xxOctets, tx_1024_to_15xx);
+	FALCON_STAT(efx, XgTxPkts1519toMaxOctets, tx_15xx_to_jumbo);
+	FALCON_STAT(efx, XgTxUndersizePkts, tx_lt64);
+	FALCON_STAT(efx, XgTxOversizePkts, tx_gtjumbo);
+	FALCON_STAT(efx, XgTxNonTcpUdpPkt, tx_non_tcpudp);
+	FALCON_STAT(efx, XgTxMacSrcErrPkt, tx_mac_src_error);
+	FALCON_STAT(efx, XgTxIpSrcErrPkt, tx_ip_src_error);
+
+	/* Update derived statistics */
+	mac_stats->tx_good_bytes =
+		(mac_stats->tx_bytes - mac_stats->tx_bad_bytes);
+	mac_stats->rx_bad_bytes =
+		(mac_stats->rx_bytes - mac_stats->rx_good_bytes);
+}
+
+#define EFX_XAUI_RETRAIN_MAX 8
+
+static int falcon_check_xmac(struct efx_nic *efx)
+{
+	unsigned link_ok, phyxs_ok = 1;
+	unsigned has_phyxs = efx->phy_op->mmds & (1 << MDIO_MMD_PHYXS);
+
+	/* Check the remote XAUI link status */
+	link_ok = falcon_xaui_link_ok(efx);
+
+	if ((efx->loopback_mode == LOOPBACK_NETWORK) ||
+	    !efx->phy_powered)
+		return 0;
+
+	if (link_ok && has_phyxs && !LOOPBACK_INTERNAL(efx)) {
+		/* Does the PHYXS think we have lane sync? */
+		phyxs_ok = mdio_clause45_phyxgxs_lane_sync(efx);
+	}
+
+	if (EFX_WORKAROUND_5147(efx) && (!link_ok || !phyxs_ok)) {
+		(void) falcon_reset_xaui(efx);
+		falcon_reconfigure_xgxs_core(efx);
+	}
+
+	/* Call the PHY check_hw routine */
+	efx->phy_op->check_hw(efx);
+	return 0;
+}
+
+/* Simulate a PHY event */
+static void falcon_xmac_sim_phy_event(struct efx_nic *efx)
+{
+	efx_qword_t phy_event;
+
+	EFX_POPULATE_QWORD_2(phy_event,
+			     EV_CODE, GLOBAL_EV_DECODE,
+			     XG_PHY_INTR, 1);
+	falcon_generate_event(&efx->channel[0], &phy_event);
+}
+
+static int falcon_xmac_get_settings(struct efx_nic *efx,
+				    struct ethtool_cmd *ecmd)
+{
+	mdio_clause45_get_settings(efx, ecmd);
+	ecmd->transceiver = XCVR_INTERNAL;
+	ecmd->phy_address = efx->mii.phy_id;
+	ecmd->autoneg = AUTONEG_DISABLE;
+	ecmd->duplex = DUPLEX_FULL;
+	return 0;
+}
+
+static int falcon_xmac_set_settings(struct efx_nic *efx,
+				    struct ethtool_cmd *ecmd)
+{
+	if (ecmd->transceiver != XCVR_INTERNAL)
+		return -EINVAL;
+	if (ecmd->autoneg != AUTONEG_DISABLE)
+		return -EINVAL;
+	if (ecmd->duplex != DUPLEX_FULL)
+		return -EINVAL;
+
+	return mdio_clause45_set_settings(efx, ecmd);
+}
+
+
+static int falcon_xmac_set_pause(struct efx_nic *efx,
+				 enum efx_fc_type flow_control)
+{
+	int reset;
+
+	if (flow_control & EFX_FC_AUTO) {
+		EFX_LOG(efx, "10G does not support flow control "
+			"autonegotiation\n");
+		return -EINVAL;
+	}
+
+	if ((flow_control & EFX_FC_TX) && !(flow_control & EFX_FC_RX))
+		return -EINVAL;
+
+	/* TX flow control may automatically turn itself off if the
+	 * link partner (intermittently) stops responding to pause
+	 * frames. There isn't any indication that this has happened,
+	 * so the best we do is leave it up to the user to spot this
+	 * and fix it be cycling transmit flow control on this end. */
+	reset = ((flow_control & EFX_FC_TX) &&
+		 !(efx->flow_control & EFX_FC_TX));
+	if (EFX_WORKAROUND_11482(efx) && reset) {
+		if (FALCON_REV(efx) == FALCON_REV_B0) {
+			/* Recover by resetting the EM block */
+			mutex_lock(&efx->mac_lock);
+			if (efx->link_up)
+				falcon_drain_tx_fifo(efx);
+			mutex_unlock(&efx->mac_lock);
+		} else {
+			/* Schedule a reset to recover */
+			efx_schedule_reset(efx, RESET_TYPE_INVISIBLE);
+		}
+	}
+
+	efx->flow_control = flow_control;
+
+	return 0;
+}
+
+struct efx_mac_operations falcon_xmac_operations = {
+	.mac_writel	= falcon_xmac_writel,
+	.mac_readl	= falcon_xmac_readl,
+	.init		= falcon_init_xmac,
+	.reconfigure	= falcon_reconfigure_xmac,
+	.update_stats	= falcon_update_stats_xmac,
+	.fini		= falcon_fini_xmac,
+	.check_hw	= falcon_check_xmac,
+	.fake_phy_event = falcon_xmac_sim_phy_event,
+	.get_settings   = falcon_xmac_get_settings,
+	.set_settings   = falcon_xmac_set_settings,
+	.set_pause      = falcon_xmac_set_pause,
+};
diff -rpuN linux-2.6.18.8/drivers/net/sfc/gmii.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/gmii.h
--- linux-2.6.18.8/drivers/net/sfc/gmii.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/gmii.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,212 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_GMII_H
+#define EFX_GMII_H
+
+/*
+ * GMII interface
+ */
+
+#include <linux/mii.h>
+
+/* GMII registers, excluding registers already defined as MII
+ * registers in mii.h
+ */
+#define GMII_IER		0x12	/* Interrupt enable register */
+#define GMII_ISR		0x13	/* Interrupt status register */
+
+/* Interrupt enable register */
+#define IER_ANEG_ERR		0x8000	/* Bit 15 - autonegotiation error */
+#define IER_SPEED_CHG		0x4000	/* Bit 14 - speed changed */
+#define IER_DUPLEX_CHG		0x2000	/* Bit 13 - duplex changed */
+#define IER_PAGE_RCVD		0x1000	/* Bit 12 - page received */
+#define IER_ANEG_DONE		0x0800	/* Bit 11 - autonegotiation complete */
+#define IER_LINK_CHG		0x0400	/* Bit 10 - link status changed */
+#define IER_SYM_ERR		0x0200	/* Bit 9 - symbol error */
+#define IER_FALSE_CARRIER	0x0100	/* Bit 8 - false carrier */
+#define IER_FIFO_ERR		0x0080	/* Bit 7 - FIFO over/underflow */
+#define IER_MDIX_CHG		0x0040	/* Bit 6 - MDI crossover changed */
+#define IER_DOWNSHIFT		0x0020	/* Bit 5 - downshift */
+#define IER_ENERGY		0x0010	/* Bit 4 - energy detect */
+#define IER_DTE_POWER		0x0004	/* Bit 2 - DTE power detect */
+#define IER_POLARITY_CHG	0x0002	/* Bit 1 - polarity changed */
+#define IER_JABBER		0x0001	/* Bit 0 - jabber */
+
+/* Interrupt status register */
+#define ISR_ANEG_ERR		0x8000	/* Bit 15 - autonegotiation error */
+#define ISR_SPEED_CHG		0x4000	/* Bit 14 - speed changed */
+#define ISR_DUPLEX_CHG		0x2000	/* Bit 13 - duplex changed */
+#define ISR_PAGE_RCVD		0x1000	/* Bit 12 - page received */
+#define ISR_ANEG_DONE		0x0800	/* Bit 11 - autonegotiation complete */
+#define ISR_LINK_CHG		0x0400	/* Bit 10 - link status changed */
+#define ISR_SYM_ERR		0x0200	/* Bit 9 - symbol error */
+#define ISR_FALSE_CARRIER	0x0100	/* Bit 8 - false carrier */
+#define ISR_FIFO_ERR		0x0080	/* Bit 7 - FIFO over/underflow */
+#define ISR_MDIX_CHG		0x0040	/* Bit 6 - MDI crossover changed */
+#define ISR_DOWNSHIFT		0x0020	/* Bit 5 - downshift */
+#define ISR_ENERGY		0x0010	/* Bit 4 - energy detect */
+#define ISR_DTE_POWER		0x0004	/* Bit 2 - DTE power detect */
+#define ISR_POLARITY_CHG	0x0002	/* Bit 1 - polarity changed */
+#define ISR_JABBER		0x0001	/* Bit 0 - jabber */
+
+/* Logically extended advertisement register */
+#define GM_ADVERTISE_SLCT		ADVERTISE_SLCT
+#define GM_ADVERTISE_CSMA		ADVERTISE_CSMA
+#define GM_ADVERTISE_10HALF		ADVERTISE_10HALF
+#define GM_ADVERTISE_1000XFULL		ADVERTISE_1000XFULL
+#define GM_ADVERTISE_10FULL		ADVERTISE_10FULL
+#define GM_ADVERTISE_1000XHALF		ADVERTISE_1000XHALF
+#define GM_ADVERTISE_100HALF		ADVERTISE_100HALF
+#define GM_ADVERTISE_1000XPAUSE		ADVERTISE_1000XPAUSE
+#define GM_ADVERTISE_100FULL		ADVERTISE_100FULL
+#define GM_ADVERTISE_1000XPSE_ASYM	ADVERTISE_1000XPSE_ASYM
+#define GM_ADVERTISE_100BASE4		ADVERTISE_100BASE4
+#define GM_ADVERTISE_PAUSE_CAP		ADVERTISE_PAUSE_CAP
+#define GM_ADVERTISE_PAUSE_ASYM		ADVERTISE_PAUSE_ASYM
+#define GM_ADVERTISE_RESV		ADVERTISE_RESV
+#define GM_ADVERTISE_RFAULT		ADVERTISE_RFAULT
+#define GM_ADVERTISE_LPACK		ADVERTISE_LPACK
+#define GM_ADVERTISE_NPAGE		ADVERTISE_NPAGE
+#define GM_ADVERTISE_1000FULL		(ADVERTISE_1000FULL << 8)
+#define GM_ADVERTISE_1000HALF		(ADVERTISE_1000HALF << 8)
+#define GM_ADVERTISE_1000		(GM_ADVERTISE_1000FULL | \
+					 GM_ADVERTISE_1000HALF)
+#define GM_ADVERTISE_FULL		(GM_ADVERTISE_1000FULL | \
+					 ADVERTISE_FULL)
+#define GM_ADVERTISE_ALL		(GM_ADVERTISE_1000FULL | \
+					 GM_ADVERTISE_1000HALF | \
+					 ADVERTISE_ALL)
+
+/* Logically extended link partner ability register */
+#define GM_LPA_SLCT			LPA_SLCT
+#define GM_LPA_10HALF			LPA_10HALF
+#define GM_LPA_1000XFULL		LPA_1000XFULL
+#define GM_LPA_10FULL			LPA_10FULL
+#define GM_LPA_1000XHALF		LPA_1000XHALF
+#define GM_LPA_100HALF			LPA_100HALF
+#define GM_LPA_1000XPAUSE		LPA_1000XPAUSE
+#define GM_LPA_100FULL			LPA_100FULL
+#define GM_LPA_1000XPAUSE_ASYM		LPA_1000XPAUSE_ASYM
+#define GM_LPA_100BASE4			LPA_100BASE4
+#define GM_LPA_PAUSE_CAP		LPA_PAUSE_CAP
+#define GM_LPA_PAUSE_ASYM		LPA_PAUSE_ASYM
+#define GM_LPA_RESV			LPA_RESV
+#define GM_LPA_RFAULT			LPA_RFAULT
+#define GM_LPA_LPACK			LPA_LPACK
+#define GM_LPA_NPAGE			LPA_NPAGE
+#define GM_LPA_1000FULL			(LPA_1000FULL << 6)
+#define GM_LPA_1000HALF			(LPA_1000HALF << 6)
+#define GM_LPA_10000FULL		0x00040000
+#define GM_LPA_10000HALF		0x00080000
+#define GM_LPA_DUPLEX			(GM_LPA_1000FULL | GM_LPA_10000FULL \
+					 | LPA_DUPLEX)
+#define GM_LPA_10			(LPA_10FULL | LPA_10HALF)
+#define GM_LPA_100			LPA_100
+#define GM_LPA_1000			(GM_LPA_1000FULL | GM_LPA_1000HALF)
+#define GM_LPA_10000			(GM_LPA_10000FULL | GM_LPA_10000HALF)
+
+/* Retrieve GMII autonegotiation advertised abilities
+ *
+ * The MII advertisment register (MII_ADVERTISE) is logically extended
+ * to include advertisement bits ADVERTISE_1000FULL and
+ * ADVERTISE_1000HALF from MII_CTRL1000.  The result can be tested
+ * against the GM_ADVERTISE_xxx constants.
+ */
+static inline unsigned int gmii_advertised(struct mii_if_info *gmii)
+{
+	unsigned int advertise;
+	unsigned int ctrl1000;
+
+	advertise = gmii->mdio_read(gmii->dev, gmii->phy_id, MII_ADVERTISE);
+	ctrl1000 = gmii->mdio_read(gmii->dev, gmii->phy_id, MII_CTRL1000);
+	return (((ctrl1000 << 8) & GM_ADVERTISE_1000) | advertise);
+}
+
+/* Retrieve GMII autonegotiation link partner abilities
+ *
+ * The MII link partner ability register (MII_LPA) is logically
+ * extended by adding bits LPA_1000HALF and LPA_1000FULL from
+ * MII_STAT1000.  The result can be tested against the GM_LPA_xxx
+ * constants.
+ */
+static inline unsigned int gmii_lpa(struct mii_if_info *gmii)
+{
+	unsigned int lpa;
+	unsigned int stat1000;
+
+	lpa = gmii->mdio_read(gmii->dev, gmii->phy_id, MII_LPA);
+	stat1000 = gmii->mdio_read(gmii->dev, gmii->phy_id, MII_STAT1000);
+	return (((stat1000 << 6) & GM_LPA_1000) | lpa);
+}
+
+/* Calculate GMII autonegotiated link technology
+ *
+ * "negotiated" should be the result of gmii_advertised() logically
+ * ANDed with the result of gmii_lpa().
+ *
+ * "tech" will be negotiated with the unused bits masked out.  For
+ * example, if both ends of the link are capable of both
+ * GM_LPA_1000FULL and GM_LPA_100FULL, GM_LPA_100FULL will be masked
+ * out.
+ */
+static inline unsigned int gmii_nway_result(unsigned int negotiated)
+{
+	unsigned int other_bits;
+
+	/* Mask out the speed and duplexity bits */
+	other_bits = negotiated & ~(GM_LPA_10 | GM_LPA_100 | GM_LPA_1000);
+
+	if (negotiated & GM_LPA_1000FULL)
+		return (other_bits | GM_LPA_1000FULL);
+	else if (negotiated & GM_LPA_1000HALF)
+		return (other_bits | GM_LPA_1000HALF);
+	else
+		return (other_bits | mii_nway_result(negotiated));
+}
+
+/* Calculate GMII non-autonegotiated link technology
+ *
+ * This provides an equivalent to gmii_nway_result for the case when
+ * autonegotiation is disabled.
+ */
+static inline unsigned int gmii_forced_result(unsigned int bmcr)
+{
+	unsigned int result;
+	int full_duplex;
+
+	full_duplex = bmcr & BMCR_FULLDPLX;
+	if (bmcr & BMCR_SPEED1000)
+		result = full_duplex ? GM_LPA_1000FULL : GM_LPA_1000HALF;
+	else if (bmcr & BMCR_SPEED100)
+		result = full_duplex ? GM_LPA_100FULL : GM_LPA_100HALF;
+	else
+		result = full_duplex ? GM_LPA_10FULL : GM_LPA_10HALF;
+	return result;
+}
+
+#endif /* EFX_GMII_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/i2c-direct.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/i2c-direct.c
--- linux-2.6.18.8/drivers/net/sfc/i2c-direct.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/i2c-direct.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,398 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005:      Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <asm/io.h>
+#include <linux/delay.h>
+#include "net_driver.h"
+#include "i2c-direct.h"
+
+/* EEPROM access via I2C
+ * data (SDA) and clock (SCL) line read/writes
+ */
+
+static inline void setsda(struct efx_i2c_interface *i2c, int state)
+{
+	udelay(i2c->op->udelay);
+	i2c->sda = state;
+	i2c->op->setsda(i2c);
+	udelay(i2c->op->udelay);
+}
+
+static inline void setscl(struct efx_i2c_interface *i2c, int state)
+{
+	udelay(i2c->op->udelay);
+	i2c->scl = state;
+	i2c->op->setscl(i2c);
+	udelay(i2c->op->udelay);
+}
+
+static inline int getsda(struct efx_i2c_interface *i2c)
+{
+	int sda;
+
+	udelay(i2c->op->udelay);
+	sda = i2c->op->getsda(i2c);
+	udelay(i2c->op->udelay);
+	return sda;
+}
+
+static inline int getscl(struct efx_i2c_interface *i2c)
+{
+	int scl;
+
+	udelay(i2c->op->udelay);
+	scl = i2c->op->getscl(i2c);
+	udelay(i2c->op->udelay);
+	return scl;
+}
+
+/*
+ * I2C low-level protocol operations
+ *
+ */
+
+static inline void i2c_release(struct efx_i2c_interface *i2c)
+{
+	EFX_WARN_ON_PARANOID(!i2c->scl);
+	EFX_WARN_ON_PARANOID(!i2c->sda);
+	/* Just in case */
+	setscl(i2c, 1);
+	setsda(i2c, 1);
+	EFX_BUG_ON_PARANOID(getsda(i2c) != 1);
+	EFX_BUG_ON_PARANOID(getscl(i2c) != 1);
+}
+
+static inline void i2c_start(struct efx_i2c_interface *i2c)
+{
+	/* We may be restarting immediately after a {send,recv}_bit,
+	 * so SCL will not necessarily already be high.
+	 */
+	EFX_WARN_ON_PARANOID(!i2c->sda);
+	setscl(i2c, 1);
+	setsda(i2c, 0);
+	setscl(i2c, 0);
+	setsda(i2c, 1);
+}
+
+static inline void i2c_send_bit(struct efx_i2c_interface *i2c, int bit)
+{
+	EFX_WARN_ON_PARANOID(i2c->scl != 0);
+	setsda(i2c, bit);
+	setscl(i2c, 1);
+	setscl(i2c, 0);
+	setsda(i2c, 1);
+}
+
+static inline int i2c_recv_bit(struct efx_i2c_interface *i2c)
+{
+	int bit;
+
+	EFX_WARN_ON_PARANOID(i2c->scl != 0);
+	EFX_WARN_ON_PARANOID(!i2c->sda);
+	setscl(i2c, 1);
+	bit = getsda(i2c);
+	setscl(i2c, 0);
+	return bit;
+}
+
+static inline void i2c_stop(struct efx_i2c_interface *i2c)
+{
+	EFX_WARN_ON_PARANOID(i2c->scl != 0);
+	setsda(i2c, 0);
+	setscl(i2c, 1);
+	setsda(i2c, 1);
+}
+
+/*
+ * I2C mid-level protocol operations
+ *
+ */
+
+/* Sends a byte via the I2C bus and checks for an acknowledgement from
+ * the slave device.
+ */
+static int i2c_send_byte(struct efx_i2c_interface *i2c, u8 byte)
+{
+	int i;
+
+	/* Send byte */
+	for (i = 0; i < 8; i++) {
+		i2c_send_bit(i2c, !!(byte & 0x80));
+		byte <<= 1;
+	}
+
+	/* Check for acknowledgement from slave */
+	return (i2c_recv_bit(i2c) == 0 ? 0 : -EIO);
+}
+
+/* Receives a byte via the I2C bus and sends ACK/NACK to the slave device. */
+static u8 i2c_recv_byte(struct efx_i2c_interface *i2c, int ack)
+{
+	u8 value = 0;
+	int i;
+
+	/* Receive byte */
+	for (i = 0; i < 8; i++)
+		value = (value << 1) | i2c_recv_bit(i2c);
+
+	/* Send ACK/NACK */
+	i2c_send_bit(i2c, (ack ? 0 : 1));
+
+	return value;
+}
+
+/* Calculate command byte for a read operation */
+static inline u8 i2c_read_cmd(u8 device_id)
+{
+	return ((device_id << 1) | 1);
+}
+
+/* Calculate command byte for a write operation */
+static inline u8 i2c_write_cmd(u8 device_id)
+{
+	return ((device_id << 1) | 0);
+}
+
+int efx_i2c_check_presence(struct efx_i2c_interface *i2c, u8 device_id)
+{
+	int rc;
+
+	/* If someone is driving the bus low we just give up. */
+	if (getsda(i2c) == 0 || getscl(i2c) == 0) {
+		EFX_ERR(i2c->efx, "%s someone is holding the I2C bus low."
+			" Giving up.\n", __func__);
+		return -EFAULT;
+	}
+
+	/* Pretend to initiate a device write */
+	i2c_start(i2c);
+	rc = i2c_send_byte(i2c, i2c_write_cmd(device_id));
+	if (rc)
+		goto out;
+
+ out:
+	i2c_stop(i2c);
+	i2c_release(i2c);
+
+	return rc;
+}
+
+/* This performs a fast read of one or more consecutive bytes from an
+ * I2C device.  Not all devices support consecutive reads of more than
+ * one byte; for these devices use efx_i2c_read() instead.
+ */
+int efx_i2c_fast_read(struct efx_i2c_interface *i2c,
+		      u8 device_id, u8 offset, u8 *data, unsigned int len)
+{
+	int i;
+	int rc;
+
+	EFX_WARN_ON_PARANOID(getsda(i2c) != 1);
+	EFX_WARN_ON_PARANOID(getscl(i2c) != 1);
+	EFX_WARN_ON_PARANOID(data == NULL);
+	EFX_WARN_ON_PARANOID(len < 1);
+
+	/* Select device and starting offset */
+	i2c_start(i2c);
+	rc = i2c_send_byte(i2c, i2c_write_cmd(device_id));
+	if (rc)
+		goto out;
+	rc = i2c_send_byte(i2c, offset);
+	if (rc)
+		goto out;
+
+	/* Read data from device */
+	i2c_start(i2c);
+	rc = i2c_send_byte(i2c, i2c_read_cmd(device_id));
+	if (rc)
+		goto out;
+	for (i = 0; i < (len - 1); i++)
+		/* Read and acknowledge all but the last byte */
+		data[i] = i2c_recv_byte(i2c, 1);
+	/* Read last byte with no acknowledgement */
+	data[i] = i2c_recv_byte(i2c, 0);
+
+ out:
+	i2c_stop(i2c);
+	i2c_release(i2c);
+
+	return rc;
+}
+
+/* This performs a fast write of one or more consecutive bytes to an
+ * I2C device.  Not all devices support consecutive writes of more
+ * than one byte; for these devices use efx_i2c_write() instead.
+ */
+int efx_i2c_fast_write(struct efx_i2c_interface *i2c,
+		       u8 device_id, u8 offset,
+		       const u8 *data, unsigned int len)
+{
+	int i;
+	int rc;
+
+	EFX_WARN_ON_PARANOID(getsda(i2c) != 1);
+	EFX_WARN_ON_PARANOID(getscl(i2c) != 1);
+	EFX_WARN_ON_PARANOID(len < 1);
+
+	/* Select device and starting offset */
+	i2c_start(i2c);
+	rc = i2c_send_byte(i2c, i2c_write_cmd(device_id));
+	if (rc)
+		goto out;
+	rc = i2c_send_byte(i2c, offset);
+	if (rc)
+		goto out;
+
+	/* Write data to device */
+	for (i = 0; i < len; i++) {
+		rc = i2c_send_byte(i2c, data[i]);
+		if (rc)
+			goto out;
+	}
+
+ out:
+	i2c_stop(i2c);
+	i2c_release(i2c);
+
+	return rc;
+}
+
+/* I2C byte-by-byte read */
+int efx_i2c_read(struct efx_i2c_interface *i2c,
+		 u8 device_id, u8 offset, u8 *data, unsigned int len)
+{
+	int rc;
+
+	/* i2c_fast_read with length 1 is a single byte read */
+	for (; len > 0; offset++, data++, len--) {
+		rc = efx_i2c_fast_read(i2c, device_id, offset, data, 1);
+		if (rc)
+			return rc;
+	}
+
+	return 0;
+}
+
+/* I2C byte-by-byte write */
+int efx_i2c_write(struct efx_i2c_interface *i2c,
+		  u8 device_id, u8 offset, const u8 *data, unsigned int len)
+{
+	int rc;
+
+	/* i2c_fast_write with length 1 is a single byte write */
+	for (; len > 0; offset++, data++, len--) {
+		rc = efx_i2c_fast_write(i2c, device_id, offset, data, 1);
+		if (rc)
+			return rc;
+		mdelay(i2c->op->mdelay);
+	}
+
+	return 0;
+}
+
+
+/* This is just a slightly neater wrapper round efx_i2c_fast_write
+ * in the case where the target doesn't take an offset
+ */
+int efx_i2c_send_bytes(struct efx_i2c_interface *i2c,
+		       u8 device_id, const u8 *data, unsigned int len)
+{
+	return efx_i2c_fast_write(i2c, device_id, data[0], data + 1, len - 1);
+}
+
+/* I2C receiving of bytes - does not send an offset byte */
+int efx_i2c_recv_bytes(struct efx_i2c_interface *i2c, u8 device_id,
+		       u8 *bytes, unsigned int len)
+{
+	int i;
+	int rc;
+
+	EFX_WARN_ON_PARANOID(getsda(i2c) != 1);
+	EFX_WARN_ON_PARANOID(getscl(i2c) != 1);
+	EFX_WARN_ON_PARANOID(len < 1);
+
+	/* Select device */
+	i2c_start(i2c);
+
+	/* Read data from device */
+	rc = i2c_send_byte(i2c, i2c_read_cmd(device_id));
+	if (rc)
+		goto out;
+
+	for (i = 0; i < (len - 1); i++)
+		/* Read and acknowledge all but the last byte */
+		bytes[i] = i2c_recv_byte(i2c, 1);
+	/* Read last byte with no acknowledgement */
+	bytes[i] = i2c_recv_byte(i2c, 0);
+
+ out:
+	i2c_stop(i2c);
+	i2c_release(i2c);
+
+	return rc;
+}
+
+/* SMBus and some I2C devices will time out if the I2C clock is
+ * held low for too long. This is most likely to happen in virtualised
+ * systems (when the entire domain is descheduled) but could in
+ * principle happen due to preemption on any busy system (and given the
+ * potential length of an I2C operation turning preemption off is not
+ * a sensible option). The following functions deal with the failure by
+ * retrying up to a fixed number of times.
+  */
+
+#define I2C_MAX_RETRIES	(10)
+
+/* The timeout problem will result in -EIO. If the wrapped function
+ * returns any other error, pass this up and do not retry. */
+#define RETRY_WRAPPER(_f) \
+	int retries = I2C_MAX_RETRIES; \
+	int rc; \
+	while (retries) { \
+		rc = _f; \
+		if (rc != -EIO) \
+			return rc; \
+		retries--; \
+	} \
+	return rc; \
+
+int efx_i2c_check_presence_retry(struct efx_i2c_interface *i2c, u8 device_id)
+{
+	RETRY_WRAPPER(efx_i2c_check_presence(i2c, device_id))
+}
+
+int efx_i2c_read_retry(struct efx_i2c_interface *i2c,
+		 u8 device_id, u8 offset, u8 *data, unsigned int len)
+{
+	RETRY_WRAPPER(efx_i2c_read(i2c, device_id, offset, data, len))
+}
+
+int efx_i2c_write_retry(struct efx_i2c_interface *i2c,
+		  u8 device_id, u8 offset, const u8 *data, unsigned int len)
+{
+	RETRY_WRAPPER(efx_i2c_write(i2c, device_id, offset, data, len))
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/i2c-direct.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/i2c-direct.h
--- linux-2.6.18.8/drivers/net/sfc/i2c-direct.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/i2c-direct.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,108 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005:      Fen Systems Ltd.
+ * Copyright 2006:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_I2C_DIRECT_H
+#define EFX_I2C_DIRECT_H
+
+#include "net_driver.h"
+
+/*
+ * Direct control of an I2C bus
+ */
+
+struct efx_i2c_interface;
+
+/**
+ * struct efx_i2c_bit_operations - I2C bus direct control methods
+ *
+ * I2C bus direct control methods.
+ *
+ * @setsda: Set state of SDA line
+ * @setscl: Set state of SCL line
+ * @getsda: Get state of SDA line
+ * @getscl: Get state of SCL line
+ * @udelay: Delay between each bit operation
+ * @mdelay: Delay between each byte write
+ */
+struct efx_i2c_bit_operations {
+	void (*setsda) (struct efx_i2c_interface *i2c);
+	void (*setscl) (struct efx_i2c_interface *i2c);
+	int (*getsda) (struct efx_i2c_interface *i2c);
+	int (*getscl) (struct efx_i2c_interface *i2c);
+	unsigned int udelay;
+	unsigned int mdelay;
+};
+
+/**
+ * struct efx_i2c_interface - an I2C interface
+ *
+ * An I2C interface.
+ *
+ * @efx: Attached Efx NIC
+ * @op: I2C bus control methods
+ * @sda: Current output state of SDA line
+ * @scl: Current output state of SCL line
+ */
+struct efx_i2c_interface {
+	struct efx_nic *efx;
+	struct efx_i2c_bit_operations *op;
+	unsigned int sda:1;
+	unsigned int scl:1;
+};
+
+extern int efx_i2c_check_presence(struct efx_i2c_interface *i2c, u8 device_id);
+extern int efx_i2c_fast_read(struct efx_i2c_interface *i2c,
+			     u8 device_id, u8 offset,
+			     u8 *data, unsigned int len);
+extern int efx_i2c_fast_write(struct efx_i2c_interface *i2c,
+			      u8 device_id, u8 offset,
+			      const u8 *data, unsigned int len);
+extern int efx_i2c_read(struct efx_i2c_interface *i2c,
+			u8 device_id, u8 offset, u8 *data, unsigned int len);
+extern int efx_i2c_write(struct efx_i2c_interface *i2c,
+			 u8 device_id, u8 offset,
+			 const u8 *data, unsigned int len);
+
+extern int efx_i2c_send_bytes(struct efx_i2c_interface *i2c, u8 device_id,
+			      const u8 *bytes, unsigned int len);
+
+extern int efx_i2c_recv_bytes(struct efx_i2c_interface *i2c, u8 device_id,
+			      u8 *bytes, unsigned int len);
+
+
+/* Versions of the API that retry on failure. */
+extern int efx_i2c_check_presence_retry(struct efx_i2c_interface *i2c,
+					u8 device_id);
+
+extern int efx_i2c_read_retry(struct efx_i2c_interface *i2c,
+			u8 device_id, u8 offset, u8 *data, unsigned int len);
+
+extern int efx_i2c_write_retry(struct efx_i2c_interface *i2c,
+			 u8 device_id, u8 offset,
+			 const u8 *data, unsigned int len);
+
+#endif /* EFX_I2C_DIRECT_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/Kconfig linux-2.6.18-xen-3.3.0/drivers/net/sfc/Kconfig
--- linux-2.6.18.8/drivers/net/sfc/Kconfig	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/Kconfig	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,35 @@
+config SFC
+	tristate "Solarflare Solarstorm SFC4000 support"
+	depends on PCI && INET
+	select MII
+	help
+	  This driver supports 10-gigabit Ethernet cards based on
+	  the Solarflare Communications Solarstorm SFC4000 controller.
+
+	  To compile this driver as a module, choose M here.  The module
+	  will be called sfc.
+
+config SFC_DEBUGFS
+	bool "Solarflare Solarstorm SFC4000 debugging support"
+	depends on SFC && DEBUG_FS
+	default N
+	help
+	  This option creates an "sfc" subdirectory of debugfs with
+	  debugging information for the SFC4000 driver.
+
+	  If unsure, say N.
+
+config SFC_MTD
+	depends on SFC && MTD && MTD_PARTITIONS
+	tristate "Solarflare Solarstorm SFC4000 flash/EEPROM support"
+	help
+	  This module exposes the on-board flash and/or EEPROM memory as
+	  MTD devices (e.g. /dev/mtd1).  This makes it possible to upload a
+	  new boot ROM to the NIC.
+
+config SFC_RESOURCE
+	depends on SFC && X86
+	tristate "Solarflare Solarstorm SFC4000 resource driver"
+	help
+	  This module provides the SFC resource manager driver.
+
diff -rpuN linux-2.6.18.8/drivers/net/sfc/kernel_compat.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/kernel_compat.c
--- linux-2.6.18.8/drivers/net/sfc/kernel_compat.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/kernel_compat.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,654 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#define EFX_IN_KCOMPAT_C 1
+
+#include "net_driver.h"
+#include <linux/mii.h>
+#include <linux/ethtool.h>
+#include <linux/random.h>
+#include <linux/pci.h>
+#include <linux/spinlock.h>
+#include <linux/rtnetlink.h>
+#include <linux/bootmem.h>
+#include <asm/uaccess.h>
+#include "gmii.h"
+#include "ethtool.h"
+
+/*
+ * Kernel backwards compatibility
+ *
+ * This file provides functionality missing from earlier kernels.
+ */
+
+/**************************************************************************
+ *
+ * GMII-friendly versions of mii_ethtool_[gs]set
+ *
+ **************************************************************************
+ *
+ * Kernels prior to 2.6.12 don't support GMII PHYs via
+ * mii_ethtool_gset and mii_ethtool_sset.  These are those functions
+ * taken from a 2.6.12 kernel tree, with the tests for
+ * mii->supports_gmii removed (since that field doesn't exist in older
+ * kernels).
+ *
+ */
+
+#ifdef EFX_NEED_MII_ETHTOOL_FIX
+int efx_mii_ethtool_gset(struct mii_if_info *mii, struct ethtool_cmd *ecmd)
+{
+	struct net_device *dev = mii->dev;
+	u32 advert, bmcr, lpa, nego;
+	u32 advert2 = 0, bmcr2 = 0, lpa2 = 0;
+
+	ecmd->supported =
+	    (SUPPORTED_10baseT_Half | SUPPORTED_10baseT_Full |
+	     SUPPORTED_100baseT_Half | SUPPORTED_100baseT_Full |
+	     SUPPORTED_Autoneg | SUPPORTED_TP | SUPPORTED_MII);
+	ecmd->supported |= SUPPORTED_1000baseT_Half | SUPPORTED_1000baseT_Full;
+
+	/* only supports twisted-pair */
+	ecmd->port = PORT_MII;
+
+	/* only supports internal transceiver */
+	ecmd->transceiver = XCVR_INTERNAL;
+
+	/* this isn't fully supported at higher layers */
+	ecmd->phy_address = mii->phy_id;
+
+	ecmd->advertising = ADVERTISED_TP | ADVERTISED_MII;
+	advert = mii->mdio_read(dev, mii->phy_id, MII_ADVERTISE);
+	advert2 = mii->mdio_read(dev, mii->phy_id, MII_CTRL1000);
+
+	if (advert & ADVERTISE_10HALF)
+		ecmd->advertising |= ADVERTISED_10baseT_Half;
+	if (advert & ADVERTISE_10FULL)
+		ecmd->advertising |= ADVERTISED_10baseT_Full;
+	if (advert & ADVERTISE_100HALF)
+		ecmd->advertising |= ADVERTISED_100baseT_Half;
+	if (advert & ADVERTISE_100FULL)
+		ecmd->advertising |= ADVERTISED_100baseT_Full;
+	if (advert2 & ADVERTISE_1000HALF)
+		ecmd->advertising |= ADVERTISED_1000baseT_Half;
+	if (advert2 & ADVERTISE_1000FULL)
+		ecmd->advertising |= ADVERTISED_1000baseT_Full;
+
+	bmcr = mii->mdio_read(dev, mii->phy_id, MII_BMCR);
+	lpa = mii->mdio_read(dev, mii->phy_id, MII_LPA);
+	bmcr2 = mii->mdio_read(dev, mii->phy_id, MII_CTRL1000);
+	lpa2 = mii->mdio_read(dev, mii->phy_id, MII_STAT1000);
+	if (bmcr & BMCR_ANENABLE) {
+		ecmd->advertising |= ADVERTISED_Autoneg;
+		ecmd->autoneg = AUTONEG_ENABLE;
+
+		nego = mii_nway_result(advert & lpa);
+		if ((bmcr2 & (ADVERTISE_1000HALF | ADVERTISE_1000FULL)) &
+		    (lpa2 >> 2))
+			ecmd->speed = SPEED_1000;
+		else if (nego == LPA_100FULL || nego == LPA_100HALF)
+			ecmd->speed = SPEED_100;
+		else
+			ecmd->speed = SPEED_10;
+		if ((lpa2 & LPA_1000FULL) || nego == LPA_100FULL ||
+		    nego == LPA_10FULL) {
+			ecmd->duplex = DUPLEX_FULL;
+			mii->full_duplex = 1;
+		} else {
+			ecmd->duplex = DUPLEX_HALF;
+			mii->full_duplex = 0;
+		}
+	} else {
+		ecmd->autoneg = AUTONEG_DISABLE;
+
+		ecmd->speed = ((bmcr & BMCR_SPEED1000 &&
+				(bmcr & BMCR_SPEED100) == 0) ? SPEED_1000 :
+			       (bmcr & BMCR_SPEED100) ? SPEED_100 : SPEED_10);
+		ecmd->duplex =
+			(bmcr & BMCR_FULLDPLX) ? DUPLEX_FULL : DUPLEX_HALF;
+	}
+
+	/* ignore maxtxpkt, maxrxpkt for now */
+
+	return 0;
+}
+
+int efx_mii_ethtool_sset(struct mii_if_info *mii, struct ethtool_cmd *ecmd)
+{
+	struct net_device *dev = mii->dev;
+
+	if (ecmd->speed != SPEED_10 &&
+	    ecmd->speed != SPEED_100 &&
+	    ecmd->speed != SPEED_1000)
+		return -EINVAL;
+	if (ecmd->duplex != DUPLEX_HALF && ecmd->duplex != DUPLEX_FULL)
+		return -EINVAL;
+	if (ecmd->port != PORT_MII)
+		return -EINVAL;
+	if (ecmd->transceiver != XCVR_INTERNAL)
+		return -EINVAL;
+	if (ecmd->phy_address != mii->phy_id)
+		return -EINVAL;
+	if (ecmd->autoneg != AUTONEG_DISABLE && ecmd->autoneg != AUTONEG_ENABLE)
+		return -EINVAL;
+
+	/* ignore supported, maxtxpkt, maxrxpkt */
+
+	if (ecmd->autoneg == AUTONEG_ENABLE) {
+		u32 bmcr, advert, tmp;
+		u32 advert2 = 0, tmp2 = 0;
+
+		if ((ecmd->advertising & (ADVERTISED_10baseT_Half |
+					  ADVERTISED_10baseT_Full |
+					  ADVERTISED_100baseT_Half |
+					  ADVERTISED_100baseT_Full |
+					  ADVERTISED_1000baseT_Half |
+					  ADVERTISED_1000baseT_Full)) == 0)
+			return -EINVAL;
+
+		/* advertise only what has been requested */
+		advert = mii->mdio_read(dev, mii->phy_id, MII_ADVERTISE);
+		tmp = advert & ~(ADVERTISE_ALL | ADVERTISE_100BASE4);
+		advert2 = mii->mdio_read(dev, mii->phy_id, MII_CTRL1000);
+		tmp2 = advert2 & ~(ADVERTISE_1000HALF | ADVERTISE_1000FULL);
+		if (ecmd->advertising & ADVERTISED_10baseT_Half)
+			tmp |= ADVERTISE_10HALF;
+		if (ecmd->advertising & ADVERTISED_10baseT_Full)
+			tmp |= ADVERTISE_10FULL;
+		if (ecmd->advertising & ADVERTISED_100baseT_Half)
+			tmp |= ADVERTISE_100HALF;
+		if (ecmd->advertising & ADVERTISED_100baseT_Full)
+			tmp |= ADVERTISE_100FULL;
+		if (ecmd->advertising & ADVERTISED_1000baseT_Half)
+			tmp2 |= ADVERTISE_1000HALF;
+		if (ecmd->advertising & ADVERTISED_1000baseT_Full)
+			tmp2 |= ADVERTISE_1000FULL;
+		if (advert != tmp) {
+			mii->mdio_write(dev, mii->phy_id, MII_ADVERTISE, tmp);
+			mii->advertising = tmp;
+		}
+		if (advert2 != tmp2)
+			mii->mdio_write(dev, mii->phy_id, MII_CTRL1000, tmp2);
+
+		/* turn on autonegotiation, and force a renegotiate */
+		bmcr = mii->mdio_read(dev, mii->phy_id, MII_BMCR);
+		bmcr |= (BMCR_ANENABLE | BMCR_ANRESTART);
+		mii->mdio_write(dev, mii->phy_id, MII_BMCR, bmcr);
+
+		mii->force_media = 0;
+	} else {
+		u32 bmcr, tmp;
+
+		/* turn off auto negotiation, set speed and duplexity */
+		bmcr = mii->mdio_read(dev, mii->phy_id, MII_BMCR);
+		tmp = bmcr & ~(BMCR_ANENABLE | BMCR_SPEED100 |
+			       BMCR_SPEED1000 | BMCR_FULLDPLX);
+		if (ecmd->speed == SPEED_1000)
+			tmp |= BMCR_SPEED1000;
+		else if (ecmd->speed == SPEED_100)
+			tmp |= BMCR_SPEED100;
+		if (ecmd->duplex == DUPLEX_FULL) {
+			tmp |= BMCR_FULLDPLX;
+			mii->full_duplex = 1;
+		} else {
+			mii->full_duplex = 0;
+		}
+		if (bmcr != tmp)
+			mii->mdio_write(dev, mii->phy_id, MII_BMCR, tmp);
+
+		mii->force_media = 1;
+	}
+	return 0;
+}
+#endif /* NEED_EFX_MII_ETHTOOL_GSET */
+
+/**************************************************************************
+ *
+ * unregister_netdevice_notifier : Has a race before 2.6.17
+ *
+ **************************************************************************
+ *
+ */
+
+#ifdef EFX_NEED_UNREGISTER_NETDEVICE_NOTIFIER_FIX
+/**
+ * efx_unregister_netdevice_notifier - fixed unregister_netdevice_notifier
+ * @nb:		notifier to unregister
+ *
+ * unregister_netdevice_notifier() does not wait for the notifier
+ * to be unused before 2.6.17.  This wrapper fixes that.
+ */
+int efx_unregister_netdevice_notifier(struct notifier_block *nb)
+{
+	int res;
+
+	res = unregister_netdevice_notifier(nb);
+	/* Ensure any outstanding calls complete. */
+	rtnl_lock();
+	rtnl_unlock();
+	return res;
+}
+#endif /* NEED_EFX_UNREGISTER_NETDEVICE_NOTIFIER */
+
+/**************************************************************************
+ *
+ * IOMMU-locking versions of pci_[un]map_single and
+ * pci_{alloc,free}_consistent.  See SFC bug 4560.
+ *
+ **************************************************************************
+ *
+ */
+#ifdef EFX_NEED_IOMMU_LOCK
+
+/*
+ * efx_use_iommu_lock - IOMMU lock use control
+ *
+ * If set to 1, the driver will attempt to mitigate the race condition
+ * bug around IOMMU accesses in some 2.6 kernels.  If set to 2, the
+ * driver will use the lock even if it thinks it doesn't need to.
+ * Note that this is only a best-effort attempt; in particular, we
+ * cannot do anything about other drivers touching the IOMMU.
+ */
+static unsigned int efx_use_iommu_lock = 1;
+EXPORT_SYMBOL(efx_use_iommu_lock);
+
+/*
+ * efx_iommu_lock - lock around IOMMU accesses
+ *
+ * This spinlock should be held while calling functions that access
+ * the IOMMU if efx_use_iommu_lock is >= 2.  The efx_pci_*()
+ * functions do this where possible.
+ */
+static spinlock_t efx_iommu_lock = SPIN_LOCK_UNLOCKED;
+EXPORT_SYMBOL(efx_iommu_lock);
+
+/* Don't use the IOMMU lock if the device can access the whole of memory */
+#define EFX_DMA_CONSISTENT(_efx)			\
+	(((_efx)->dma_mask >> PAGE_SHIFT) >= max_pfn)
+/**
+ * efx_pci_map_single - map buffer for DMA, under IOMMU lock
+ * @pci:		PCI device
+ * @ptr:		Buffer
+ * @size:		Buffer length
+ * @direction:		DMA direction
+ *
+ * Wrapper for pci_map_single that uses efx_iommu_lock if necessary.
+ */
+dma_addr_t efx_pci_map_single(struct pci_dev *pci, void *ptr, size_t size,
+			      int direction)
+{
+	struct efx_nic *efx = pci_get_drvdata(pci);
+	unsigned long flags __attribute__ ((unused));
+	dma_addr_t dma_addr;
+
+	if (unlikely((efx_use_iommu_lock &&
+		      (!EFX_NO_IOMMU) &&
+		      (!EFX_DMA_CONSISTENT(efx))) ||
+		     efx_use_iommu_lock >= 2)) {
+		spin_lock_irqsave(&efx_iommu_lock, flags);
+		dma_addr = pci_map_single(pci, ptr, size, direction);
+		spin_unlock_irqrestore(&efx_iommu_lock, flags);
+	} else {
+		dma_addr = pci_map_single(pci, ptr, size, direction);
+	}
+	return dma_addr;
+}
+
+/**
+ * efx_pci_unmap_single - unmap buffer for DMA, under IOMMU lock
+ * @pci:		PCI device
+ * @dma_addr:		DMA address
+ * @size:		Buffer length
+ * @direction:		DMA direction
+ *
+ * Wrapper for pci_unmap_single that uses efx_iommu_lock if necessary.
+ */
+void efx_pci_unmap_single(struct pci_dev *pci, dma_addr_t dma_addr,
+			  size_t size, int direction)
+{
+	struct efx_nic *efx = pci_get_drvdata(pci);
+	unsigned long flags __attribute__ ((unused));
+
+	if (unlikely((efx_use_iommu_lock &&
+		      (!EFX_NO_IOMMU) &&
+		      (!EFX_DMA_CONSISTENT(efx))) ||
+		     efx_use_iommu_lock >= 2)) {
+		spin_lock_irqsave(&efx_iommu_lock, flags);
+		pci_unmap_single(pci, dma_addr, size, direction);
+		spin_unlock_irqrestore(&efx_iommu_lock, flags);
+	} else {
+		pci_unmap_single(pci, dma_addr, size, direction);
+	}
+}
+
+/**
+ * efx_pci_alloc_consistent - allocate DMA-consistent buffer, under IOMMU lock
+ * @pci:		PCI device
+ * @size:		Buffer length
+ * @dma_addr:		DMA address
+ *
+ * Wrapper for pci_alloc_consistent that uses efx_iommu_lock if necessary.
+ *
+ * Bugs: Currently this can't use the spinlock because
+ *	pci_alloc_consistent may block.
+ */
+void *efx_pci_alloc_consistent(struct pci_dev *pci, size_t size,
+			       dma_addr_t *dma_addr)
+{
+	return pci_alloc_consistent(pci, size, dma_addr);
+}
+
+/**
+ * efx_pci_free_consistent - free DMA-consistent buffer, under IOMMU lock
+ * @pci:		PCI device
+ * @size:		Buffer length
+ * @ptr:		Buffer
+ * @dma_addr:		DMA address
+ *
+ * Wrapper for pci_free_consistent that uses efx_iommu_lock if necessary.
+ */
+void efx_pci_free_consistent(struct pci_dev *pci, size_t size, void *ptr,
+			     dma_addr_t dma_addr)
+{
+	struct efx_nic *efx = pci_get_drvdata(pci);
+	unsigned long flags __attribute__ ((unused));
+
+	if (unlikely((efx_use_iommu_lock &&
+		      (!EFX_NO_IOMMU) &&
+		      (!EFX_DMA_CONSISTENT(efx))) ||
+		     efx_use_iommu_lock >= 2)) {
+		spin_lock_irqsave(&efx_iommu_lock, flags);
+		pci_free_consistent(pci, size, ptr, dma_addr);
+		spin_unlock_irqrestore(&efx_iommu_lock, flags);
+	} else {
+		pci_free_consistent(pci, size, ptr, dma_addr);
+	}
+}
+
+module_param(efx_use_iommu_lock, uint, 0644);
+MODULE_PARM_DESC(efx_use_iommu_lock, "Enable lock for bug in free_iommu");
+
+#endif
+
+#ifdef EFX_NEED_COMPOUND_PAGE_FIX
+
+void efx_compound_page_destructor(struct page *page)
+{
+	/* Fake up page state to keep __free_pages happy */
+	set_page_count(page, 1);
+	page[1].mapping = NULL;
+
+	__free_pages(page, (unsigned long)page[1].index);
+}
+
+#endif /* NEED_COMPOUND_PAGE_FIX */
+
+/**************************************************************************
+ *
+ * print_hex_dump, taken from lib/hexdump.c.
+ *
+ **************************************************************************
+ *
+ */
+#ifdef EFX_NEED_HEX_DUMP
+
+#define hex_asc(x)	"0123456789abcdef"[x]
+#define isascii(c) (((unsigned char)(c))<=0x7f)
+
+static void hex_dump_to_buffer(const void *buf, size_t len, int rowsize,
+			       int groupsize, char *linebuf, size_t linebuflen,
+			       int ascii)
+{
+        const u8 *ptr = buf;
+        u8 ch;
+        int j, lx = 0;
+        int ascii_column;
+
+        if (rowsize != 16 && rowsize != 32)
+                rowsize = 16;
+
+        if (!len)
+                goto nil;
+        if (len > rowsize)              /* limit to one line at a time */
+                len = rowsize;
+        if ((len % groupsize) != 0)     /* no mixed size output */
+                groupsize = 1;
+
+        switch (groupsize) {
+        case 8: {
+                const u64 *ptr8 = buf;
+                int ngroups = len / groupsize;
+
+                for (j = 0; j < ngroups; j++)
+                        lx += scnprintf(linebuf + lx, linebuflen - lx,
+				"%16.16llx ", (unsigned long long)*(ptr8 + j));
+                ascii_column = 17 * ngroups + 2;
+                break;
+        }
+
+        case 4: {
+                const u32 *ptr4 = buf;
+                int ngroups = len / groupsize;
+
+                for (j = 0; j < ngroups; j++)
+                        lx += scnprintf(linebuf + lx, linebuflen - lx,
+				"%8.8x ", *(ptr4 + j));
+                ascii_column = 9 * ngroups + 2;
+                break;
+        }
+
+        case 2: {
+                const u16 *ptr2 = buf;
+                int ngroups = len / groupsize;
+
+                for (j = 0; j < ngroups; j++)
+                        lx += scnprintf(linebuf + lx, linebuflen - lx,
+				"%4.4x ", *(ptr2 + j));
+                ascii_column = 5 * ngroups + 2;
+                break;
+        }
+
+        default:
+                for (j = 0; (j < rowsize) && (j < len) && (lx + 4) < linebuflen;
+                     j++) {
+                        ch = ptr[j];
+                        linebuf[lx++] = hex_asc(ch >> 4);
+                        linebuf[lx++] = hex_asc(ch & 0x0f);
+                        linebuf[lx++] = ' ';
+                }
+                ascii_column = 3 * rowsize + 2;
+                break;
+        }
+        if (!ascii)
+                goto nil;
+
+        while (lx < (linebuflen - 1) && lx < (ascii_column - 1))
+                linebuf[lx++] = ' ';
+	/* Removed is_print() check */
+        for (j = 0; (j < rowsize) && (j < len) && (lx + 2) < linebuflen; j++)
+                linebuf[lx++] = isascii(ptr[j]) ? ptr[j] : '.';
+nil:
+        linebuf[lx++] = '\0';
+}
+
+void print_hex_dump(const char *level, const char *prefix_str, int prefix_type,
+		    int rowsize, int groupsize,
+		    const void *buf, size_t len, int ascii)
+{
+        const u8 *ptr = buf;
+        int i, linelen, remaining = len;
+        char linebuf[200];
+
+        if (rowsize != 16 && rowsize != 32)
+                rowsize = 16;
+
+        for (i = 0; i < len; i += rowsize) {
+                linelen = min(remaining, rowsize);
+                remaining -= rowsize;
+                hex_dump_to_buffer(ptr + i, linelen, rowsize, groupsize,
+				   linebuf, sizeof(linebuf), ascii);
+
+                switch (prefix_type) {
+                case DUMP_PREFIX_ADDRESS:
+                        printk("%s%s%*p: %s\n", level, prefix_str,
+			       (int)(2 * sizeof(void *)), ptr + i, linebuf);
+                        break;
+                case DUMP_PREFIX_OFFSET:
+                        printk("%s%s%.8x: %s\n", level, prefix_str, i, linebuf);
+                        break;
+                default:
+                        printk("%s%s%s\n", level, prefix_str, linebuf);
+                        break;
+                }
+        }
+}
+
+#endif /* EFX_NEED_HEX_DUMP */
+
+/**************************************************************************
+ *
+ * print_mac, from net/ethernet/eth.c in v2.6.24
+ *
+ **************************************************************************
+ *
+ */
+#ifdef EFX_NEED_PRINT_MAC
+char *print_mac(char *buf, const u8 *addr)
+{
+        sprintf(buf, "%02x:%02x:%02x:%02x:%02x:%02x",
+                addr[0], addr[1], addr[2], addr[3], addr[4], addr[5]);
+        return buf;
+}
+#endif /* EFX_NEED_PRINT_MAC */
+
+#ifdef EFX_NEED_CSUM_TCPUDP_NOFOLD
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20)
+__wsum
+csum_tcpudp_nofold (__be32 saddr, __be32 daddr, unsigned short len,
+		    unsigned short proto, __wsum sum)
+#else
+__wsum
+csum_tcpudp_nofold (unsigned long saddr, unsigned long daddr,
+		    unsigned short len, unsigned short proto, __wsum sum)
+#endif
+{
+	unsigned long result;
+
+	result = (__force u64)saddr + (__force u64)daddr +
+		(__force u64)sum + ((len + proto) << 8);
+
+	/* Fold down to 32-bits so we don't lose in the typedef-less network stack.  */
+	/* 64 to 33 */
+	result = (result & 0xffffffff) + (result >> 32);
+	/* 33 to 32 */
+	result = (result & 0xffffffff) + (result >> 32);
+	return (__force __wsum)result;
+
+}
+#endif /* EFX_NEED_CSUM_TCPUDP_NOFOLD */
+
+#ifdef EFX_NEED_RANDOM_ETHER_ADDR
+/* Generate random MAC address */
+void efx_random_ether_addr(uint8_t *addr) {
+        get_random_bytes (addr, ETH_ALEN);
+	addr [0] &= 0xfe;       /* clear multicast bit */
+	addr [0] |= 0x02;       /* set local assignment bit (IEEE802) */
+}
+#endif /* EFX_NEED_RANDOM_ETHER_ADDR */
+
+#ifdef EFX_NEED_MSECS_TO_JIFFIES
+/*
+ * When we convert to jiffies then we interpret incoming values
+ * the following way:
+ *
+ * - negative values mean 'infinite timeout' (MAX_JIFFY_OFFSET)
+ *
+ * - 'too large' values [that would result in larger than
+ *   MAX_JIFFY_OFFSET values] mean 'infinite timeout' too.
+ *
+ * - all other values are converted to jiffies by either multiplying
+ *   the input value by a factor or dividing it with a factor
+ *
+ * We must also be careful about 32-bit overflows.
+ */
+#ifndef MSEC_PER_SEC
+#define MSEC_PER_SEC	1000L
+#endif
+unsigned long msecs_to_jiffies(const unsigned int m)
+{
+	/*
+	 * Negative value, means infinite timeout:
+	 */
+	if ((int)m < 0)
+		return MAX_JIFFY_OFFSET;
+
+#if HZ <= MSEC_PER_SEC && !(MSEC_PER_SEC % HZ)
+	/*
+	 * HZ is equal to or smaller than 1000, and 1000 is a nice
+	 * round multiple of HZ, divide with the factor between them,
+	 * but round upwards:
+	 */
+	return (m + (MSEC_PER_SEC / HZ) - 1) / (MSEC_PER_SEC / HZ);
+#elif HZ > MSEC_PER_SEC && !(HZ % MSEC_PER_SEC)
+	/*
+	 * HZ is larger than 1000, and HZ is a nice round multiple of
+	 * 1000 - simply multiply with the factor between them.
+	 *
+	 * But first make sure the multiplication result cannot
+	 * overflow:
+	 */
+	if (m > jiffies_to_msecs(MAX_JIFFY_OFFSET))
+		return MAX_JIFFY_OFFSET;
+
+	return m * (HZ / MSEC_PER_SEC);
+#else
+	/*
+	 * Generic case - multiply, round and divide. But first
+	 * check that if we are doing a net multiplication, that
+	 * we wouldnt overflow:
+	 */
+	if (HZ > MSEC_PER_SEC && m > jiffies_to_msecs(MAX_JIFFY_OFFSET))
+		return MAX_JIFFY_OFFSET;
+
+	return (m * HZ + MSEC_PER_SEC - 1) / MSEC_PER_SEC;
+#endif
+}
+#endif /* EFX_NEED_MSECS_TO_JIFFIES */
+
+#ifdef EFX_NEED_MSLEEP
+/**
+ * msleep - sleep safely even with waitqueue interruptions
+ * @msecs: Time in milliseconds to sleep for
+ */
+void msleep(unsigned int msecs)
+{
+	unsigned long timeout = msecs_to_jiffies(msecs) + 1;
+
+	while (timeout)
+		timeout = schedule_timeout_uninterruptible(timeout);
+}
+#endif
diff -rpuN linux-2.6.18.8/drivers/net/sfc/kernel_compat.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/kernel_compat.h
--- linux-2.6.18.8/drivers/net/sfc/kernel_compat.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/kernel_compat.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,925 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_KERNEL_COMPAT_H
+#define EFX_KERNEL_COMPAT_H
+
+#include <linux/version.h>
+#include <linux/sched.h>
+#include <linux/errno.h>
+#include <linux/pci.h>
+#include <linux/workqueue.h>
+#include <linux/moduleparam.h>
+#include <linux/interrupt.h>
+#include <linux/skbuff.h>
+#include <linux/netdevice.h>
+
+#include "extraversion.h"
+
+/*
+ * Kernel backwards compatibility
+ *
+ * This file provides macros that enable the driver to be compiled on
+ * any kernel from 2.6.9 onward (plus SLES 9 2.6.5), without requiring
+ * explicit version tests scattered throughout the code.
+ */
+
+/**************************************************************************
+ *
+ * Version/config/architecture tests to set feature flags
+ *
+ **************************************************************************
+ *
+ * NOTE: For simplicity, these initial version tests cover kernel.org
+ * releases only.  Backported features in "enterprise" kernels are
+ * handled further down.
+ */
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,9) &&	 \
+	!(defined(EFX_DIST_SUSE) &&			 \
+	  LINUX_VERSION_CODE == KERNEL_VERSION(2,6,5) && \
+	  EFX_DIST_KVER_LEVEL_1 == 7)
+	#error "This kernel version is now unsupported"
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,6)
+	#define EFX_NEED_RANDOM_ETHER_ADDR yes
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,7)
+	#define EFX_NEED_I2C_CLASS_HWMON yes
+	#define EFX_NEED_IF_MII yes
+	#define EFX_NEED_MSLEEP yes
+	#define EFX_NEED_MSECS_TO_JIFFIES yes
+#endif
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,8)
+	#define EFX_USE_MTD_ERASE_FAIL_ADDR yes
+#else
+	#define EFX_NEED_MTD_ERASE_CALLBACK yes
+	#define EFX_NEED_DUMMY_PCI_DISABLE_MSI yes
+	#define EFX_NEED_DUMMY_MSIX yes
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,9)
+	#define EFX_NEED_BYTEORDER_TYPES yes
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,10)
+	#define EFX_NEED_MMIOWB yes
+	#define EFX_NEED_PCI_SAVE_RESTORE_WRAPPERS yes
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,12)
+	#define EFX_NEED_DUMMY_SUPPORTS_GMII yes
+	#define EFX_NEED_MII_CONSTANTS yes
+	#define EFX_NEED_MII_ETHTOOL_FIX yes
+	#define EFX_HAVE_MSIX_TABLE_RESERVED yes
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,14)
+	#define EFX_NEED_SCHEDULE_TIMEOUT_INTERRUPTIBLE yes
+	#define EFX_NEED_SCHEDULE_TIMEOUT_UNINTERRUPTIBLE yes
+	#define EFX_NEED_GFP_T yes
+	#define EFX_NEED_KZALLOC yes
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,15)
+	#define EFX_NEED_SETUP_TIMER yes
+	#ifdef CONFIG_HUGETLB_PAGE
+		#define EFX_USE_COMPOUND_PAGES yes
+	#endif
+#else
+	#define EFX_USE_COMPOUND_PAGES yes
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,16)
+	#define EFX_NEED_MUTEX yes
+	#define EFX_NEED_SAFE_LISTS yes
+	#ifdef EFX_USE_COMPOUND_PAGES
+		#define EFX_NEED_COMPOUND_PAGE_FIX yes
+	#endif
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,17)
+	#define EFX_NEED_UNREGISTER_NETDEVICE_NOTIFIER_FIX yes
+	#define EFX_NEED_DEV_NOTICE yes
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,18)
+	#define EFX_NEED_IRQF_FLAGS yes
+	#define EFX_NEED_NETDEV_ALLOC_SKB yes
+	/* Fedora backported 2.6.18 netdevice.h changes */
+	#ifndef NETIF_F_GSO
+		#define EFX_NEED_NETIF_TX_LOCK yes
+	#endif
+#else
+	#define EFX_USE_MTD_WRITESIZE yes
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,19)
+	#define EFX_NEED_IRQ_HANDLER_T yes
+	#define EFX_HAVE_IRQ_HANDLER_REGS yes
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,20)
+	#define EFX_NEED_WORK_API_WRAPPERS yes
+	#define EFX_USE_FASTCALL yes
+	#define EFX_NEED_CSUM_UNFOLDED yes
+#endif
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,21)
+	/*
+	 * debugfs was introduced earlier, but only supports sym-links
+	 * from 2.6.21
+	 */
+	#ifdef CONFIG_DEBUG_FS
+		#define EFX_USE_DEBUGFS yes
+	#endif
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,22)
+	#define EFX_NEED_SKB_HEADER_MACROS yes
+	#define EFX_NEED_HEX_DUMP yes
+#else
+	#define EFX_USE_CANCEL_WORK_SYNC yes
+#endif
+
+#if LINUX_VERSION_CODE == KERNEL_VERSION(2,6,22)
+	#define EFX_NEED_HEX_DUMP_CONST_FIX yes
+#endif
+
+#if (LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,14)) && \
+    (LINUX_VERSION_CODE <  KERNEL_VERSION(2,6,23))
+	#define EFX_USE_ETHTOOL_GET_PERM_ADDR yes
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,23)
+	#ifdef __ia64__
+		/* csum_tcpudp_nofold() is extern but not exported */
+		#define EFX_NEED_CSUM_TCPUDP_NOFOLD yes
+	#endif
+#else
+	#define EFX_USE_PCI_DEV_REVISION yes
+	#define EFX_USE_CANCEL_DELAYED_WORK_SYNC yes
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,24)
+	#define EFX_HAVE_OLD_NAPI yes
+	#define EFX_NEED_GENERIC_LRO yes
+	#define EFX_NEED_PRINT_MAC yes
+#else
+	#define EFX_USE_ETHTOOL_FLAGS yes
+#endif
+
+/*
+ * SFC Bug 4560: Some kernels leak IOMMU entries under heavy load.  Use a
+ * spinlock to serialise access where possible to alleviate the
+ * problem.
+ *
+ * NB. The following definition is duplicated in
+ * the char driver.  Please keep in sync.
+ */
+#if (LINUX_VERSION_CODE < KERNEL_VERSION(2,6,17) && \
+     defined(__x86_64__) && defined(CONFIG_SMP))
+	#define EFX_NEED_IOMMU_LOCK yes
+	#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,15)
+		#if defined(CONFIG_GART_IOMMU)
+			#define EFX_NO_IOMMU no_iommu
+		#else
+			#define EFX_NO_IOMMU 1
+		#endif
+	#else
+		#define EFX_NO_IOMMU 0
+	#endif
+#endif
+
+#ifdef CONFIG_PPC64
+	/* __raw_writel and friends are broken on ppc64 */
+	#define EFX_NEED_RAW_READ_AND_WRITE_FIX yes
+#endif
+
+/**************************************************************************
+ *
+ * Exceptions for backported features
+ *
+ **************************************************************************
+ */
+
+/* RHEL4 */
+#if defined(EFX_DIST_RHEL) && LINUX_VERSION_CODE == KERNEL_VERSION(2,6,9)
+	#if EFX_DIST_KVER_LEVEL_1 >= 22
+		/* linux-2.6.9-mmiowb.patch */
+		#undef EFX_NEED_MMIOWB
+	#endif
+	#if EFX_DIST_KVER_LEVEL_1 >= 34
+		/* linux-2.6.9-net-mii-update.patch */
+		#undef EFX_NEED_DUMMY_SUPPORTS_GMII
+		#undef EFX_NEED_MII_CONSTANTS
+		#undef EFX_NEED_MII_ETHTOOL_FIX
+		/* linux-2.6.9-gfp_t-typedef.patch */
+		#undef EFX_NEED_GFP_T
+		/* linux-2.6.9-slab-update.patch */
+		#undef EFX_NEED_KZALLOC
+	#endif
+	#if EFX_DIST_KVER_LEVEL_1 >= 55
+		/* linux-2.6.18-sata-update.patch (ported from 2.6.18->2.6.9) */
+		#undef EFX_NEED_SCHEDULE_TIMEOUT_UNINTERRUPTIBLE
+		#undef EFX_NEED_IRQ_HANDLER_T
+	#endif
+#endif
+
+/* RHEL5 */
+#if defined(EFX_DIST_RHEL) && LINUX_VERSION_CODE == KERNEL_VERSION(2,6,18)
+	#if EFX_DIST_KVER_LEVEL_1 >= 53
+		/* linux-2.6.18-sata-update.patch */
+		#undef EFX_NEED_SCHEDULE_TIMEOUT_UNINTERRUPTIBLE
+		#undef EFX_NEED_IRQ_HANDLER_T
+	#endif
+#endif
+
+#if defined(EFX_DIST_RHEL)
+	#if (LINUX_VERSION_CODE != KERNEL_VERSION(2,6,9)) && \
+	     (LINUX_VERSION_CODE != KERNEL_VERSION(2,6,18))
+		#error "Unknown Red Hat Enterprise kernel version"
+	#endif
+#endif
+
+/* SLES9 */
+#if defined(EFX_DIST_SUSE) && LINUX_VERSION_CODE == KERNEL_VERSION(2,6,5) && \
+	EFX_DIST_KVER_LEVEL_1 == 7
+	#if EFX_DIST_KVER_LEVEL_2 >= 139
+		#undef EFX_NEED_MMIOWB
+	#endif
+	#if EFX_DIST_KVER_LEVEL_2 >= 191
+		#undef EFX_NEED_MSLEEP
+		#undef EFX_NEED_MSECS_TO_JIFFIES
+	#endif
+	#if EFX_DIST_KVER_LEVEL_2 >= 244
+		#undef EFX_NEED_BYTEORDER_TYPES
+	#endif
+	#if EFX_DIST_KVER_LEVEL_2 >= 252
+		#undef EFX_NEED_KZALLOC
+	#endif
+#endif
+
+/**************************************************************************
+ *
+ * Definitions of missing constants, types, functions and macros
+ *
+ **************************************************************************
+ *
+ */
+
+#ifndef DMA_40BIT_MASK
+	#define DMA_40BIT_MASK	0x000000ffffffffffULL
+#endif
+
+#ifndef spin_trylock_irqsave
+	#define spin_trylock_irqsave(lock, flags)	\
+	({						\
+		local_irq_save(flags);			\
+		spin_trylock(lock) ?			\
+		1 : ({local_irq_restore(flags); 0;});	\
+	})
+#endif
+
+#ifndef raw_smp_processor_id
+	#define raw_smp_processor_id() (current_thread_info()->cpu)
+#endif
+
+#ifndef NETIF_F_LRO
+	#define NETIF_F_LRO 0
+#endif
+
+/* Cope with small changes in PCI constants between minor kernel revisions */
+#if PCI_X_STATUS != 4
+	#undef PCI_X_STATUS
+	#define PCI_X_STATUS 4
+	#undef PCI_X_STATUS_MAX_SPLIT
+	#define PCI_X_STATUS_MAX_SPLIT 0x03800000
+#endif
+
+#ifndef PCI_EXP_LNKSTA
+	#define PCI_EXP_LNKSTA		18	    /* Link Status */
+#endif
+
+/* Used for struct pt_regs */
+#ifndef regs_return_value
+	#if defined(__x86_64__)
+		#define regs_return_value(regs) ((regs)->rax)
+	#elif defined(__i386__)
+		#define regs_return_value(regs) ((regs)->eax)
+	#elif defined(__ia64__)
+		#define regs_return_value(regs) ((regs)->r8)
+	#else
+		#error "Need definition for regs_return_value()"
+	#endif
+#endif
+
+#ifndef __GFP_COMP
+	#define __GFP_COMP 0
+#endif
+
+#ifndef __iomem
+	#define __iomem
+#endif
+
+#ifndef NET_IP_ALIGN
+	#define NET_IP_ALIGN 2
+#endif
+
+#ifndef PCI_CAP_ID_EXP
+#define PCI_CAP_ID_EXP		0x10    /* PCI Express */
+#endif
+
+#ifndef PCI_EXP_FLAGS
+#define PCI_EXP_FLAGS           2           /* Capabilities register */
+#define PCI_EXP_FLAGS_TYPE      0x00f0      /* Capability version */
+#define PCI_EXP_TYPE_ENDPOINT   0x0         /* Express Endpoint */
+#define PCI_EXP_TYPE_LEG_END    0x1         /* Legacy Endpoint */
+#define PCI_EXP_TYPE_ROOT_PORT  0x4         /* Root Port */
+#endif
+
+#ifndef PCI_EXP_DEVCAP
+#define PCI_EXP_DEVCAP          4           /* Device capabilities */
+#define PCI_EXP_DEVCAP_PAYLOAD  0x07        /* Max_Payload_Size */
+#define PCI_EXP_DEVCAP_PWR_VAL  0x3fc0000   /* Slot Power Limit Value */
+#define PCI_EXP_DEVCAP_PWR_SCL  0xc000000   /* Slot Power Limit Scale */
+#endif
+
+#ifndef PCI_EXP_DEVCTL
+#define PCI_EXP_DEVCTL          8           /* Device Control */
+#define PCI_EXP_DEVCTL_PAYLOAD  0x00e0      /* Max_Payload_Size */
+#define PCI_EXP_DEVCTL_READRQ   0x7000      /* Max_Read_Request_Size */
+#endif
+
+#ifndef PCI_EXP_LNKSTA
+#define PCI_EXP_LNKSTA		18	    /* Link Status */
+#endif
+
+#ifndef NETDEV_TX_OK
+	#define NETDEV_TX_OK 0
+#endif
+
+#ifndef NETDEV_TX_BUSY
+	#define NETDEV_TX_BUSY 1
+#endif
+
+#ifndef __force
+	#define __force
+#endif
+
+#if ! defined(for_each_cpu_mask) && ! defined(CONFIG_SMP)
+	#define for_each_cpu_mask(cpu, mask)            \
+		for ((cpu) = 0; (cpu) < 1; (cpu)++, (void)mask)
+#endif
+
+/**************************************************************************/
+
+#ifdef EFX_NEED_IRQ_HANDLER_T
+	typedef irqreturn_t (*irq_handler_t)(int, void *, struct pt_regs *);
+#endif
+
+#ifdef EFX_NEED_I2C_CLASS_HWMON
+	#define I2C_CLASS_HWMON (1<<0)
+#endif
+
+#ifdef EFX_NEED_MII_CONSTANTS
+	#define MII_CTRL1000		0x09
+	#define MII_STAT1000		0x0a
+	#define BMCR_SPEED1000		0x0040
+	#define ADVERTISE_PAUSE_ASYM	0x0800
+	#define ADVERTISE_PAUSE_CAP	0x0400
+	#define LPA_PAUSE_ASYM		0x0800
+	#define LPA_PAUSE_CAP		0x0400
+	#define ADVERTISE_1000FULL	0x0200
+	#define ADVERTISE_1000HALF	0x0100
+	#define LPA_1000FULL		0x0800
+	#define LPA_1000HALF		0x0400
+#endif
+
+#ifdef EFX_NEED_DUMMY_SUPPORTS_GMII
+	#include <linux/mii.h>
+	/* Ugly; redirect nonexistent new field to an old unused field */
+	#undef supports_gmii
+	#define supports_gmii full_duplex
+#endif
+
+#ifdef EFX_NEED_SKB_HEADER_MACROS
+	#define skb_mac_header(skb)	(skb)->mac.raw
+	#define skb_network_header(skb) (skb)->nh.raw
+	#define tcp_hdr(skb)		(skb)->h.th
+	#define ip_hdr(skb)		(skb)->nh.iph
+	#define skb_tail_pointer(skb)   (skb)->tail
+#endif
+
+#ifdef EFX_NEED_RAW_READ_AND_WRITE_FIX
+	#include <asm/io.h>
+	static inline void
+	efx_raw_writeb(u8 value, volatile void __iomem *addr)
+	{
+		writeb(value, addr);
+	}
+	static inline void
+	efx_raw_writew(u16 value, volatile void __iomem *addr)
+	{
+		writew(le16_to_cpu(value), addr);
+	}
+	static inline void
+	efx_raw_writel(u32 value, volatile void __iomem *addr)
+	{
+		writel(le32_to_cpu(value), addr);
+	}
+	static inline void
+	efx_raw_writeq(u64 value, volatile void __iomem *addr)
+	{
+		writeq(le64_to_cpu(value), addr);
+	}
+	static inline u8
+	efx_raw_readb(const volatile void __iomem *addr)
+	{
+		return readb(addr);
+	}
+	static inline u16
+	efx_raw_readw(const volatile void __iomem *addr)
+	{
+		return cpu_to_le16(readw(addr));
+	}
+	static inline u32
+	efx_raw_readl(const volatile void __iomem *addr)
+	{
+		return cpu_to_le32(readl(addr));
+	}
+	static inline u64
+	efx_raw_readq(const volatile void __iomem *addr)
+	{
+		return cpu_to_le64(readq(addr));
+	}
+
+	#undef __raw_writeb
+	#undef __raw_writew
+	#undef __raw_writel
+	#undef __raw_writeq
+	#undef __raw_readb
+	#undef __raw_readw
+	#undef __raw_readl
+	#undef __raw_readq
+	#define __raw_writeb efx_raw_writeb
+	#define __raw_writew efx_raw_writew
+	#define __raw_writel efx_raw_writel
+	#define __raw_writeq efx_raw_writeq
+	#define __raw_readb efx_raw_readb
+	#define __raw_readw efx_raw_readw
+	#define __raw_readl efx_raw_readl
+	#define __raw_readq efx_raw_readq
+#endif
+
+#ifdef EFX_NEED_SCHEDULE_TIMEOUT_INTERRUPTIBLE
+	static inline signed long
+	schedule_timeout_interruptible(signed long timeout)
+	{
+		set_current_state(TASK_INTERRUPTIBLE);
+		return schedule_timeout(timeout);
+	}
+#endif
+
+#ifdef EFX_NEED_SCHEDULE_TIMEOUT_UNINTERRUPTIBLE
+	static inline signed long
+	schedule_timeout_uninterruptible(signed long timeout)
+	{
+		set_current_state(TASK_UNINTERRUPTIBLE);
+		return schedule_timeout(timeout);
+	}
+#endif
+
+#ifdef EFX_NEED_MMIOWB
+	#if defined(__i386__) || defined(__x86_64__)
+		#define mmiowb()
+	#elif defined(__ia64__)
+		#ifndef ia64_mfa
+			#define ia64_mfa() asm volatile ("mf.a" ::: "memory")
+		#endif
+		#define mmiowb ia64_mfa
+	#else
+		#error "Need definition for mmiowb()"
+	#endif
+#endif
+
+#ifdef EFX_NEED_KZALLOC
+	static inline void *kzalloc(size_t size, int flags)
+	{
+		void *buf = kmalloc(size, flags);
+		if (buf)
+			memset(buf, 0,size);
+		return buf;
+	}
+#endif
+
+#ifdef EFX_NEED_SETUP_TIMER
+	static inline void setup_timer(struct timer_list * timer,
+				       void (*function)(unsigned long),
+				       unsigned long data)
+	{
+		timer->function = function;
+		timer->data = data;
+		init_timer(timer);
+	}
+#endif
+
+#ifdef EFX_NEED_MUTEX
+	#define EFX_DEFINE_MUTEX(x) DECLARE_MUTEX(x)
+	#undef DEFINE_MUTEX
+	#define DEFINE_MUTEX EFX_DEFINE_MUTEX
+
+	#define efx_mutex semaphore
+	#undef mutex
+	#define mutex efx_mutex
+
+	#define efx_mutex_init(x) init_MUTEX(x)
+	#undef mutex_init
+	#define mutex_init efx_mutex_init
+
+	#define efx_mutex_destroy(x) do { } while(0)
+	#undef mutex_destroy
+	#define mutex_destroy efx_mutex_destroy
+
+	#define efx_mutex_lock(x) down(x)
+	#undef mutex_lock
+	#define mutex_lock efx_mutex_lock
+
+	#define efx_mutex_lock_interruptible(x) down_interruptible(x)
+	#undef mutex_lock_interruptible
+	#define mutex_lock_interruptible efx_mutex_lock_interruptible
+
+	#define efx_mutex_unlock(x) up(x)
+	#undef mutex_unlock
+	#define mutex_unlock efx_mutex_unlock
+
+	#define efx_mutex_trylock(x) (!down_trylock(x))
+	#undef mutex_trylock
+	#define mutex_trylock efx_mutex_trylock
+
+	static inline int efx_mutex_is_locked(struct efx_mutex *m)
+	{
+		/* NB. This is quite inefficient, but it's the best we
+		 * can do with the semaphore API. */
+		if ( down_trylock(m) )
+			return 1;
+		/* Undo the effect of down_trylock. */
+		up(m);
+		return 0;
+	}
+	#undef mutex_is_locked
+	#define mutex_is_locked efx_mutex_is_locked
+#endif
+
+#ifndef NETIF_F_GSO
+	#define efx_gso_size tso_size
+	#undef gso_size
+	#define gso_size efx_gso_size
+	#define efx_gso_segs tso_segs
+	#undef gso_segs
+	#define gso_segs efx_gso_segs
+#endif
+
+#ifdef EFX_NEED_IRQF_FLAGS
+	#ifdef SA_PROBEIRQ
+		#define IRQF_PROBE_SHARED  SA_PROBEIRQ
+	#else
+		#define IRQF_PROBE_SHARED  0
+	#endif
+	#define IRQF_SHARED	   SA_SHIRQ
+#endif
+
+#ifdef EFX_NEED_NETDEV_ALLOC_SKB
+	#ifndef NET_SKB_PAD
+		#define NET_SKB_PAD 16
+	#endif
+
+	static inline
+	struct sk_buff *netdev_alloc_skb(struct net_device *dev,
+					 unsigned int length)
+	{
+		struct sk_buff *skb = alloc_skb(length + NET_SKB_PAD,
+						GFP_ATOMIC | __GFP_COLD);
+		if (likely(skb)) {
+			skb_reserve(skb, NET_SKB_PAD);
+			skb->dev = dev;
+		}
+		return skb;
+	}
+#endif
+
+#ifdef EFX_NEED_NETIF_TX_LOCK
+	static inline void netif_tx_lock(struct net_device *dev)
+	{
+		spin_lock(&dev->xmit_lock);
+		dev->xmit_lock_owner = smp_processor_id();
+	}
+	static inline void netif_tx_lock_bh(struct net_device *dev)
+	{
+		spin_lock_bh(&dev->xmit_lock);
+		dev->xmit_lock_owner = smp_processor_id();
+	}
+	static inline void netif_tx_unlock_bh(struct net_device *dev)
+	{
+		dev->xmit_lock_owner = -1;
+		spin_unlock_bh(&dev->xmit_lock);
+	}
+	static inline void netif_tx_unlock(struct net_device *dev)
+	{
+		dev->xmit_lock_owner = -1;
+		spin_unlock(&dev->xmit_lock);
+	}
+#endif
+
+#ifdef EFX_NEED_CSUM_UNFOLDED
+	typedef u32 __wsum;
+	#define csum_unfold(x) ((__force __wsum) x)
+#endif
+
+#ifdef EFX_NEED_HEX_DUMP
+	enum {
+		DUMP_PREFIX_NONE,
+		DUMP_PREFIX_ADDRESS,
+		DUMP_PREFIX_OFFSET
+	};
+#endif
+
+#ifdef EFX_NEED_PRINT_MAC
+	#define DECLARE_MAC_BUF(var) char var[18] __attribute__((unused))
+#endif
+
+#ifdef EFX_NEED_GFP_T
+	typedef unsigned int gfp_t;
+#endif
+
+#ifdef EFX_NEED_SAFE_LISTS
+	#define list_for_each_entry_safe_reverse(pos, n, head, member)	     \
+		for (pos = list_entry((head)->prev, typeof(*pos), member),   \
+		     n = list_entry(pos->member.prev, typeof(*pos), member); \
+		     &pos->member != (head);				     \
+		     pos = n,						     \
+		     n = list_entry(n->member.prev, typeof(*n), member))
+#endif
+
+#ifdef EFX_NEED_DEV_NOTICE
+	#define dev_notice dev_warn
+#endif
+
+#ifdef EFX_NEED_IF_MII
+	#include <linux/mii.h>
+	static inline struct mii_ioctl_data *efx_if_mii ( struct ifreq *rq ) {
+		return ( struct mii_ioctl_data * ) &rq->ifr_ifru;
+	}
+	#undef if_mii
+	#define if_mii efx_if_mii
+#endif
+
+#ifdef EFX_NEED_MTD_ERASE_CALLBACK
+	#include <linux/mtd/mtd.h>
+	static inline void efx_mtd_erase_callback(struct erase_info *instr) {
+		if ( instr->callback )
+			instr->callback ( instr );
+	}
+	#undef mtd_erase_callback
+	#define mtd_erase_callback efx_mtd_erase_callback
+#endif
+
+#ifdef EFX_NEED_DUMMY_PCI_DISABLE_MSI
+	#include <linux/pci.h>
+	static inline void dummy_pci_disable_msi ( struct pci_dev *dev ) {
+		/* Do nothing */
+	}
+	#undef pci_disable_msi
+	#define pci_disable_msi dummy_pci_disable_msi
+#endif
+
+#ifdef EFX_NEED_DUMMY_MSIX
+	struct msix_entry {
+		u16 	vector;	/* kernel uses to write allocated vector */
+		u16	entry;	/* driver uses to specify entry, OS writes */
+	};
+	static inline int pci_enable_msix(struct pci_dev* dev,
+					  struct msix_entry *entries, int nvec)
+		{return -1;}
+	static inline void pci_disable_msix(struct pci_dev *dev) { /* Do nothing */}
+#endif
+
+#ifdef EFX_NEED_BYTEORDER_TYPES
+	typedef __u16 __be16;
+	typedef __u32 __be32;
+	typedef __u64 __be64;
+	typedef __u16 __le16;
+	typedef __u32 __le32;
+	typedef __u64 __le64;
+#endif
+
+/**************************************************************************
+ *
+ * Missing functions provided by kernel_compat.c
+ *
+ **************************************************************************
+ *
+ */
+#ifdef EFX_NEED_RANDOM_ETHER_ADDR
+	extern void efx_random_ether_addr(uint8_t *addr);
+	#ifndef EFX_IN_KCOMPAT_C
+		#undef random_ether_addr
+		#define random_ether_addr efx_random_ether_addr
+	#endif
+#endif
+
+#ifdef EFX_NEED_MII_ETHTOOL_FIX
+	extern int efx_mii_ethtool_gset(struct mii_if_info *mii,
+					struct ethtool_cmd *ecmd);
+	extern int efx_mii_ethtool_sset(struct mii_if_info *mii,
+					struct ethtool_cmd *ecmd);
+	#ifndef EFX_IN_KCOMPAT_C
+		#undef mii_ethtool_gset
+		#define mii_ethtool_gset efx_mii_ethtool_gset
+		#undef mii_ethtool_sset
+		#define mii_ethtool_sset efx_mii_ethtool_sset
+	#endif
+#endif
+
+#ifdef EFX_NEED_UNREGISTER_NETDEVICE_NOTIFIER_FIX
+	extern int efx_unregister_netdevice_notifier(struct notifier_block *nb);
+	#ifndef EFX_IN_KCOMPAT_C
+		#undef unregister_netdevice_notifier
+		#define unregister_netdevice_notifier \
+				efx_unregister_netdevice_notifier
+	#endif
+#endif
+
+#ifdef EFX_NEED_IOMMU_LOCK
+	extern dma_addr_t efx_pci_map_single(struct pci_dev *pci, void *ptr,
+					     size_t size, int direction);
+	extern void efx_pci_unmap_single(struct pci_dev *pci,
+					 dma_addr_t dma_addr, size_t size,
+					 int direction);
+	extern void * efx_pci_alloc_consistent(struct pci_dev *pci,
+					       size_t size,
+					       dma_addr_t *dma_addr);
+	extern void efx_pci_free_consistent(struct pci_dev *pci,
+					    size_t size, void *ptr,
+					    dma_addr_t dma_addr);
+	#ifndef EFX_IN_KCOMPAT_C
+		#undef pci_map_single
+		#undef pci_unmap_single
+		#undef pci_alloc_consistent
+		#undef pci_free_consistent
+		#define pci_map_single efx_pci_map_single
+		#define pci_unmap_single efx_pci_unmap_single
+		#define pci_alloc_consistent efx_pci_alloc_consistent
+		#define pci_free_consistent efx_pci_free_consistent
+	#endif
+#endif
+
+#ifdef EFX_NEED_PRINT_MAC
+	extern char *print_mac(char *buf, const u8 *addr);
+#endif
+
+#ifdef EFX_NEED_COMPOUND_PAGE_FIX
+	extern void efx_compound_page_destructor(struct page *page);
+#endif
+
+#ifdef EFX_NEED_HEX_DUMP
+	extern void
+	print_hex_dump(const char *level, const char *prefix_str,
+		       int prefix_type, int rowsize, int groupsize,
+		       const void *buf, size_t len, int ascii);
+#endif
+
+#ifdef EFX_NEED_MSECS_TO_JIFFIES
+	extern unsigned long msecs_to_jiffies(const unsigned int m);
+#endif
+
+#ifdef EFX_NEED_MSLEEP
+	extern void msleep(unsigned int msecs);
+#endif
+
+/**************************************************************************
+ *
+ * Wrappers to fix bugs and parameter changes
+ *
+ **************************************************************************
+ *
+ */
+
+#ifdef EFX_NEED_PCI_SAVE_RESTORE_WRAPPERS
+	#define pci_save_state(_dev)					\
+		pci_save_state(_dev, (_dev)->saved_config_space)
+
+	#define pci_restore_state(_dev)					\
+		pci_restore_state(_dev, (_dev)->saved_config_space)
+#endif
+
+#ifdef EFX_NEED_WORK_API_WRAPPERS
+	/**
+	 * queue_delayed_work in pre 2.6.20 can't rearm from inside
+	 * the work member. So instead do a rather hacky sleep
+	 */
+	#define delayed_work work_struct
+	#define INIT_DELAYED_WORK INIT_WORK
+
+	static int inline efx_queue_delayed_work(struct workqueue_struct *wq,
+						 struct work_struct *work,
+						 unsigned long delay)
+	{
+		if (unlikely(delay > 0))
+			schedule_timeout_uninterruptible(delay);
+		return queue_work(wq, work);
+	}
+	#define queue_delayed_work efx_queue_delayed_work
+
+	/**
+	 * The old and new work-function prototypes just differ
+	 * in the type of the pointer returned, so it's safe
+	 * to cast between the prototypes.
+	 */
+	typedef void (*efx_old_work_func_t)(void *p);
+
+	#undef INIT_WORK
+	#define INIT_WORK(_work, _func)					\
+		do {							\
+			INIT_LIST_HEAD(&(_work)->entry);		\
+			(_work)->pending = 0;				\
+			PREPARE_WORK((_work),				\
+				     (efx_old_work_func_t) (_func),	\
+				     (_work));				\
+		} while (0)
+#endif
+
+#ifdef EFX_HAVE_OLD_NAPI
+	#define napi_str napi_dev[0]
+
+	static inline void netif_napi_add(struct net_device *dev,
+					  struct net_device *dummy,
+					  int (*poll) (struct net_device *,
+						       int *),
+					  int weight)
+	{
+		dev->weight = weight;
+		dev->poll = poll;
+	}
+
+	#define napi_enable netif_poll_enable
+	#define napi_disable netif_poll_disable
+
+	#define netif_rx_complete(dev, dummy) netif_rx_complete(dev)
+#endif
+
+#ifdef EFX_NEED_COMPOUND_PAGE_FIX
+	static inline
+	struct page *efx_alloc_pages(gfp_t flags, unsigned int order)
+	{
+		struct page *p = alloc_pages(flags, order);
+		if ((flags & __GFP_COMP) && (p != NULL) && (order > 0))
+			p[1].mapping = (void *)efx_compound_page_destructor;
+		return p;
+	}
+	#undef alloc_pages
+	#define alloc_pages efx_alloc_pages
+
+	static inline
+	void efx_free_pages(struct page *p, unsigned int order)
+	{
+		if ((order > 0) && (page_count(p) == 1))
+			p[1].mapping = NULL;
+		__free_pages(p, order);
+	}
+	#define __free_pages efx_free_pages
+#endif
+
+#ifdef EFX_NEED_HEX_DUMP_CONST_FIX
+	#define print_hex_dump(v,s,t,r,g,b,l,a) \
+		print_hex_dump((v),(s),(t),(r),(g),(void*)(b),(l),(a))
+#endif
+
+#endif /* EFX_KERNEL_COMPAT_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/lm87_support.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/lm87_support.c
--- linux-2.6.18.8/drivers/net/sfc/lm87_support.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/lm87_support.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,295 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2007:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************/
+
+#include "net_driver.h"
+#include "lm87_support.h"
+#include "workarounds.h"
+
+/* Setting this to 1 will cause efx_check_lm87 to dump the status when it
+ * detects an alarm. This will result in the canonical name (i.e. that in
+ * the LM87 data book) being printed for each set status bit, along with
+ * the reading for that sensor value, if applicable. If set to 0 only the
+ * raw status1 and status2 register values are printed. */
+#define LM87_VERBOSE_ALARMS	1
+
+/**************************************************************************
+ *
+ * Onboard LM87 temperature and voltage monitor
+ *
+ **************************************************************************
+ */
+
+/* LM87 channel mode: all current boards either do not use AIN1/FAN1 and 2
+ * or use them as AIN. */
+#define LM87_CHANNEL_MODE 0x16
+#define LM87_CHANNEL_AIN1 1
+#define LM87_CHANNEL_AIN2 2
+#define LM87_CHANNEL_INIT (LM87_CHANNEL_AIN2 | LM87_CHANNEL_AIN1)
+
+/* LM87 configuration register 1 */
+#define LM87_CONFIG_1 0x40
+#define LM87_START 0x01
+#define LM87_INTEN 0x02
+#define LM87_INITIALIZATION 0x80
+
+/* LM87 interrupt status register 1 */
+#define LM87_INT_STATUS_1 0x41
+
+/* LM87 interrupt status register 2 */
+#define LM87_INT_STATUS_2 0x42
+
+/* LM87 interrupt mask register 1 */
+#define LM87_INT_MASK_1 0x43
+
+/* LM87 interrupt mask register 2 */
+#define LM87_INT_MASK_2 0x44
+
+/* LM87 monitoring limits */
+#define LM87_LIMITS 0x2b
+
+
+int efx_probe_lm87(struct efx_nic *efx, int addr,
+		   const u8 *limits, int nlimits, const u16 irqmask)
+{
+	struct efx_i2c_interface *i2c = &efx->i2c;
+	u8 byte;
+	int rc;
+
+	/* Check for onboard LM87 */
+	rc = efx_i2c_check_presence_retry(i2c, addr);
+	if (rc) {
+		/* Not an error to lack an LM87, but failure to probe the
+		 * bus is worrying. */
+		if (rc == -EFAULT) {
+			EFX_ERR(efx, "Failed to probe I2C bus for LM87!\n");
+			return rc;
+		} else {
+			EFX_LOG(efx, "has no onboard LM87 chip\n");
+			return 0;
+		}
+	}
+	efx->board_info.lm87_addr = addr;
+	EFX_LOG(efx, "detected onboard LM87 chip at 0x%2x\n", addr);
+
+	/* Reset chip */
+	byte = LM87_INITIALIZATION;
+	rc = efx_i2c_write_retry(i2c, addr, LM87_CONFIG_1, &byte, 1);
+	if (rc) {
+		EFX_ERR(efx, "could not reset LM87\n");
+		return rc;
+	}
+
+	/* Configure channel mode: currently hardwire to make pins 5 and 6
+	 * AIN1 and AIN2 rather than FAN1, FAN2. */
+	byte = LM87_CHANNEL_INIT;
+	rc = efx_i2c_write_retry(i2c, addr, LM87_CHANNEL_MODE, &byte, 1);
+	if (rc) {
+		EFX_ERR(efx, "could not program LM87 chan. mode\n");
+		return rc;
+	}
+
+	/* Configure limits */
+	rc = efx_i2c_write_retry(i2c, addr, LM87_LIMITS, limits, nlimits);
+	if (rc) {
+		EFX_ERR(efx, "could not program LM87 limits\n");
+		return rc;
+	}
+
+	/* Mask off unwanted interrupts */
+	byte = (irqmask & 0xff);
+	rc = efx_i2c_write_retry(i2c, addr, LM87_INT_MASK_1, &byte, 1);
+	if (rc) {
+		EFX_ERR(efx, "could not mask LM87 interrupts\n");
+		return rc;
+	}
+
+	byte = (irqmask >> 8);
+	rc = efx_i2c_write_retry(i2c, addr, LM87_INT_MASK_2, &byte, 1);
+	if (rc) {
+		EFX_ERR(efx, "could not mask LM87 interrupts\n");
+		return rc;
+	}
+
+	/* Start monitoring */
+	byte = LM87_START;
+	if (irqmask != EFX_LM87_NO_INTS)
+		byte |= LM87_INTEN;
+
+	rc = efx_i2c_write_retry(i2c, addr, LM87_CONFIG_1, &byte, 1);
+	if (rc) {
+		EFX_ERR(efx, "could not start LM87\n");
+		return rc;
+	}
+
+	return rc;
+}
+
+void efx_remove_lm87(struct efx_nic *efx)
+{
+	struct efx_i2c_interface *i2c = &efx->i2c;
+	u8 byte;
+
+	if (!efx->board_info.lm87_addr)
+		return;
+
+	/* Reset chip */
+	byte = LM87_INITIALIZATION;
+	if (efx_i2c_write_retry(i2c, efx->board_info.lm87_addr,
+				LM87_CONFIG_1, &byte, 1) != 0)
+		EFX_ERR(efx, "could not reset LM87 on exit\n");
+}
+
+#if LM87_VERBOSE_ALARMS
+/* Bit number to name mapping for status1 */
+static const char *lm_stat_names[] = {
+/* Status 1 contents */
+	"+2.5Vin",
+	"Vccp1",
+	"Vcc",
+	"+5Vin",
+	"Int. Temp.",
+	"Ext. Temp.",
+	"FAN1/AIN1",
+	"FAN2/AIN2",
+/* Status 2 contents */
+	"+12Vin",
+	"Vccp2",
+	"Reserved",
+	"Reserved",
+	"CI",
+	"THERM#",
+	"D1 Fault",
+	"D2 Fault"
+};
+
+/* Where to read the value corresponding to an alarm bit. */
+static const int lm_stat_regs[] = {
+	0x20, 0x21, 0x22, 0x23, 0x27, 0x26, 0x28, 0x29,
+	0x24, 0x25, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff
+};
+
+/* The positions of the alarm bits do not correspond exactly to the
+ * order of the limit values. Convert so the user only needs to maintain
+ * one array */
+static int lm_bit_to_lim[] = {
+	0, /* 2.5V */
+	1, /* Vccp1 */
+	2, /* Vcc */
+	3, /* 5V */
+	7, /* Int temp. */
+	6, /* Ext temp. */
+	8, /* AIN1 */
+	9, /* AIN2 */
+	4, /* 12V */
+	5  /* Vccp2 */
+};
+
+/* These are bit numbers. I feel justified in hardwiring the max. */
+static const int lm_stat_max = 16;
+
+static void lm87_show_alarm(struct efx_nic *efx, int bit)
+{
+	char valbuf[8];
+	u8 val;
+
+	if (lm_stat_regs[bit] != 0xff) {
+		efx_i2c_read_retry(&efx->i2c, efx->board_info.lm87_addr,
+				   lm_stat_regs[bit], &val, 1);
+		sprintf(valbuf, "0x%02x  ", val);
+	} else {
+		strcpy(valbuf, "----  ");
+	}
+	/* If the board code knows what this sensor is wired to, let it tell
+	 * us, else just print the LM87 datasheet name of the input, and the
+	 * value. */
+	if (efx->board_info.interpret_sensor == NULL ||
+	    (bit < ARRAY_SIZE(lm_bit_to_lim) &&
+	     efx->board_info.interpret_sensor(efx, lm_bit_to_lim[bit], val)
+	     == 0))
+		EFX_ERR(efx, ": %10s  %4s\n",
+			STRING_TABLE_LOOKUP(bit, lm_stat), valbuf);
+}
+
+static void lm87_dump_alarms(struct efx_nic *efx, int stat1, int stat2)
+{
+	int i;
+	EFX_ERR(efx, "   NAME    value\n");
+	for (i = 0; i < 8; i++) {
+		if (stat1 & (1 << i))
+			lm87_show_alarm(efx, i);
+		if (stat2 & (1 << i))
+			lm87_show_alarm(efx, i + 8);
+	}
+}
+
+#else
+#define lm87_dump_alarms(_name, _stat1, _stat2) do {} while (0)
+#endif
+
+/* Read onboard LM87 (if present)
+ * Return error code if lm87 could not be read (-EIO)
+ * _or_ is raising an alarm (-ERANGE). 0 if AOK.
+ */
+int efx_check_lm87(struct efx_nic *efx, unsigned mask)
+{
+	struct efx_i2c_interface *i2c = &efx->i2c;
+	u8 int_status_1, int_status_2;
+	unsigned ints;
+	int rc = 0;
+
+	/* If link is up then do not monitor temperature */
+	if (EFX_WORKAROUND_7884(efx) && efx->link_up)
+		return 0;
+
+	if (!efx->board_info.lm87_addr)
+		return 0;
+
+	/* Read interrupt status registers */
+	rc = efx_i2c_read_retry(i2c, efx->board_info.lm87_addr,
+				LM87_INT_STATUS_1, &int_status_1, 1);
+	if (rc) {
+		EFX_ERR(efx, "could not read LM87 INT status 1\n");
+		return rc;
+	}
+	rc = efx_i2c_read_retry(i2c, efx->board_info.lm87_addr,
+				LM87_INT_STATUS_2, &int_status_2, 1);
+	if (rc) {
+		EFX_ERR(efx, "could not read LM87 INT status 2\n");
+		return rc;
+	}
+
+	int_status_1 &= mask;
+	int_status_2 &= (mask >> 8);
+	ints = ((int_status_2 << 8) | int_status_1);
+
+	/* Check interrupt status */
+	if (ints == 0)
+		return 0;
+
+	EFX_ERR(efx, "LM87 detected a hardware failure (status %02x:%02x)\n",
+		int_status_1, int_status_2);
+	lm87_dump_alarms(efx, int_status_1, int_status_2);
+
+	return -ERANGE;
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/lm87_support.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/lm87_support.h
--- linux-2.6.18.8/drivers/net/sfc/lm87_support.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/lm87_support.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,58 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2007:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************/
+
+#ifndef EFX_LM87_SUPPORT_H
+#define EFX_LM87_SUPPORT_H
+
+/* The interrupt bit masks. These are the same in the interrupt status and
+ * interrupt mask registers. */
+/* Register 1 bits */
+#define EFX_LM87_2_5V_INT	(1)
+#define EFX_LM87_VCCP1_INT	(2)
+#define EFX_LM87_VCC_INT	(4)
+#define EFX_LM87_5_V_INT	(8)
+#define EFX_LM87_ITMP_INT	(0x10)
+#define EFX_LM87_ETMP_INT	(0x20)
+#define EFX_LM87_FAN1_INT	(0x40)
+#define EFX_LM87_FAN2_INT	(0x80)
+/* Register 2 bits */
+#define EFX_LM87_12V_INT	(0x100)
+#define EFX_LM87_VCCP2_INT	(0x200)
+/* Bits 2 and 3 are reserved. */
+#define EFX_LM87_CI_INT		(0x1000)
+#define EFX_LM87_THERM_INT	(0x2000)
+#define EFX_LM87_D1_INT		(0x4000)
+#define EFX_LM87_D2_INT		(0x8000)
+
+#define EFX_LM87_NO_INTS	((u16)-1)
+
+extern
+int efx_probe_lm87(struct efx_nic *efx, int addr, const u8 *limits,
+		   int nlimits, const u16 irqmask);
+
+extern void efx_remove_lm87(struct efx_nic *efx);
+
+extern int efx_check_lm87(struct efx_nic *efx, unsigned mask);
+
+#endif /* EFX_LM87_SUPPORT_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/mac.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/mac.h
--- linux-2.6.18.8/drivers/net/sfc/mac.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/mac.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,38 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_MAC_H
+#define EFX_MAC_H
+
+#include "net_driver.h"
+
+extern void mentormac_reset(struct efx_nic *efx);
+extern void mentormac_reconfigure(struct efx_nic *efx);
+extern struct efx_mac_operations falcon_gmac_operations;
+extern struct efx_mac_operations falcon_xmac_operations;
+
+#endif
diff -rpuN linux-2.6.18.8/drivers/net/sfc/Makefile linux-2.6.18-xen-3.3.0/drivers/net/sfc/Makefile
--- linux-2.6.18.8/drivers/net/sfc/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,43 @@
+
+# Final objects
+sfc_o = sfc.o
+sfc_mtd_o = sfc_mtd.o
+
+# Constituent objects
+sfc_elements_o :=
+sfc_elements_o += efx.o
+sfc_elements_o += falcon.o
+sfc_elements_o += tx.o
+sfc_elements_o += rx.o
+sfc_elements_o += mentormac.o
+sfc_elements_o += falcon_gmac.o
+sfc_elements_o += falcon_xmac.o
+sfc_elements_o += alaska.o
+sfc_elements_o += i2c-direct.o
+sfc_elements_o += selftest.o
+sfc_elements_o += driverlink.o
+ifeq ($(CONFIG_SFC_DEBUGFS),y)
+sfc_elements_o += debugfs.o
+endif
+sfc_elements_o += ethtool.o
+sfc_elements_o += xfp_phy.o
+sfc_elements_o += mdio_10g.o
+sfc_elements_o += txc43128_phy.o
+sfc_elements_o += tenxpress.o
+sfc_elements_o += lm87_support.o
+sfc_elements_o += boards.o
+sfc_elements_o += sfe4001.o
+sfc_elements_o += pm8358_phy.o
+sfc_elements_o += null_phy.o
+sfc_elements_o += phy.o
+sfc_elements_o += kernel_compat.o
+
+sfc_mtd_elements_o := mtd.o
+
+obj-$(CONFIG_SFC) += $(sfc_o)
+obj-$(CONFIG_SFC_MTD) += $(sfc_mtd_o)
+
+sfc-objs = $(sfc_elements_o)
+sfc_mtd-objs = $(sfc_mtd_elements_o)
+
+obj-$(CONFIG_SFC_RESOURCE) += sfc_resource/
diff -rpuN linux-2.6.18.8/drivers/net/sfc/mdio_10g.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/mdio_10g.c
--- linux-2.6.18.8/drivers/net/sfc/mdio_10g.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/mdio_10g.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,441 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+/*
+ * Useful functions for working with MDIO clause 45 PHYs
+ */
+#include <linux/types.h>
+#include <linux/ethtool.h>
+#include <linux/delay.h>
+#include "net_driver.h"
+#include "mdio_10g.h"
+#include "boards.h"
+
+static const char *mmd_block_names[] = {
+	[0]               = "(illegal)",
+	[MDIO_MMD_PMAPMD] = "PMA/PMD",
+	[MDIO_MMD_WIS]    = "WIS",
+	[MDIO_MMD_PCS]    = "PCS",
+	[MDIO_MMD_PHYXS]  = "PHY XS",
+	[MDIO_MMD_DTEXS]  = "DTE XS",
+	[MDIO_MMD_TC]     = "TC",
+	[MDIO_MMD_AN]     = "AN",
+};
+static const int mmd_block_max = ARRAY_SIZE(mmd_block_names);
+
+const char *mdio_clause45_mmd_name(int mmd)
+{
+	return STRING_TABLE_LOOKUP(mmd, mmd_block);
+}
+
+int mdio_clause45_reset_mmd(struct efx_nic *port, int mmd,
+			    int spins, int spintime)
+{
+	u32 ctrl;
+	int phy_id = port->mii.phy_id;
+
+	/* Catch callers passing values in the wrong units (or just silly) */
+	EFX_BUG_ON_PARANOID(spins * spintime >= 5000);
+
+	mdio_clause45_write(port, phy_id, mmd, MDIO_MMDREG_CTRL1,
+			    (1 << MDIO_MMDREG_CTRL1_RESET_LBN));
+	/* Wait for the reset bit to clear. */
+	do {
+		msleep(spintime);
+		ctrl = mdio_clause45_read(port, phy_id, mmd, MDIO_MMDREG_CTRL1);
+		spins--;
+
+	} while (spins && (ctrl & (1 << MDIO_MMDREG_CTRL1_RESET_LBN)));
+
+	return spins ? spins : -ETIMEDOUT;
+}
+
+static int mdio_clause45_check_mmd(struct efx_nic *efx, int mmd,
+				   int fault_fatal)
+{
+	int status;
+	int phy_id = efx->mii.phy_id;
+	const char *mmdname = STRING_TABLE_LOOKUP(mmd, mmd_block);
+
+	if (LOOPBACK_INTERNAL(efx))
+		return 0;
+
+	/* Read MMD STATUS2 to check it is responding. */
+	status = mdio_clause45_read(efx, phy_id, mmd, MDIO_MMDREG_STAT2);
+	if (((status >> MDIO_MMDREG_STAT2_PRESENT_LBN) &
+	     ((1 << MDIO_MMDREG_STAT2_PRESENT_WIDTH) - 1)) !=
+	    MDIO_MMDREG_STAT2_PRESENT_VAL) {
+		EFX_ERR(efx, "PHY MMD %s not responding.\n", mmdname);
+		return -EIO;
+	}
+
+	/* Read MMD STATUS 1 to check for fault. */
+	status = mdio_clause45_read(efx, phy_id, mmd, MDIO_MMDREG_STAT1);
+	if ((status & (1 << MDIO_MMDREG_STAT1_FAULT_LBN)) != 0) {
+		if (fault_fatal) {
+			EFX_ERR(efx, "PHY MMD %s reporting fatal"
+				" fault: status %x\n", mmdname, status);
+			return -EIO;
+		} else {
+			EFX_LOG(efx, "PHY MMD %s reporting status"
+				" %x (expected)\n", mmdname, status);
+		}
+	}
+	return 0;
+}
+
+/* This ought to be ridiculous overkill. We expect it to fail rarely */
+#define MDIO45_RESET_TIME	HZ
+#define MDIO45_RESET_ITERS	(100)
+
+int mdio_clause45_wait_reset_mmds(struct efx_nic *efx,
+				  unsigned int mmd_mask)
+{
+	const int spintime = MDIO45_RESET_TIME / MDIO45_RESET_ITERS;
+	int tries = MDIO45_RESET_ITERS;
+	int rc = 0;
+	int in_reset;
+
+	while (tries) {
+		int mask = mmd_mask;
+		int mmd = 0;
+		int stat;
+		in_reset = 0;
+		while (mask) {
+			if (mask & 1) {
+				stat = mdio_clause45_read(efx,
+							  efx->mii.phy_id,
+							  mmd,
+							  MDIO_MMDREG_CTRL1);
+				if (stat < 0) {
+					EFX_ERR(efx, "failed to read"
+						" status of MMD %s\n",
+						STRING_TABLE_LOOKUP(mmd,
+								    mmd_block));
+					return -EIO;
+				}
+				if (stat & (1 << MDIO_MMDREG_CTRL1_RESET_LBN))
+					in_reset |= (1 << mmd);
+			}
+			mask = mask >> 1;
+			mmd++;
+		}
+		if (!in_reset)
+			break;
+		tries--;
+		msleep(spintime);
+	}
+	if (in_reset != 0) {
+		EFX_ERR(efx, "not all MMDs came out of reset in time."
+			" MMDs still in reset: %x\n", in_reset);
+		rc = -ETIMEDOUT;
+	}
+	return rc;
+}
+
+int mdio_clause45_check_mmds(struct efx_nic *efx,
+			     unsigned int mmd_mask, unsigned int fatal_mask)
+{
+	int devices, mmd = 0;
+	int probe_mmd;
+
+	/* Historically we have probed the PHYXS to find out what devices are
+	 * present,but that doesn't work so well if the PHYXS isn't expected
+	 * to exist, if so just find the first item in the list supplied. */
+	probe_mmd = (mmd_mask & MDIO_MMDREG_DEVS0_PHYXS) ? MDIO_MMD_PHYXS :
+	    __ffs(mmd_mask);
+	devices = mdio_clause45_read(efx, efx->mii.phy_id,
+				     probe_mmd, MDIO_MMDREG_DEVS0);
+
+	/* Check all the expected MMDs are present */
+	if (devices < 0) {
+		EFX_ERR(efx, "failed to read devices present\n");
+		return -EIO;
+	}
+	if ((devices & mmd_mask) != mmd_mask) {
+		EFX_ERR(efx, "required MMDs not present: got %x, "
+			"wanted %x\n", devices, mmd_mask);
+		return -ENODEV;
+	}
+	EFX_TRACE(efx, "Devices present: %x\n", devices);
+
+	/* Check all required MMDs are responding and happy. */
+	while (mmd_mask) {
+		if (mmd_mask & 1) {
+			int fault_fatal = fatal_mask & 1;
+			if (mdio_clause45_check_mmd(efx, mmd, fault_fatal))
+				return -EIO;
+		}
+		mmd_mask = mmd_mask >> 1;
+		fatal_mask = fatal_mask >> 1;
+		mmd++;
+	}
+
+	return 0;
+}
+
+int mdio_clause45_links_ok(struct efx_nic *efx, unsigned int mmd_mask)
+{
+	int phy_id = efx->mii.phy_id;
+	int status;
+	int ok = 1;
+	int mmd = 0;
+	int good;
+
+	/* If the port is in loopback, then we should only consider a subset
+	 * of mmd's */
+	if (LOOPBACK_INTERNAL(efx))
+		return 1;
+	else if (efx->loopback_mode == LOOPBACK_NETWORK)
+		return 0;
+	else if (!efx->phy_powered)
+		return 0;
+	else if (efx->loopback_mode == LOOPBACK_PHYXS)
+		mmd_mask &= ~(MDIO_MMDREG_DEVS0_PHYXS |
+			      MDIO_MMDREG_DEVS0_PCS |
+			      MDIO_MMDREG_DEVS0_PMAPMD);
+	else if (efx->loopback_mode == LOOPBACK_PCS)
+		mmd_mask &= ~(MDIO_MMDREG_DEVS0_PCS |
+			      MDIO_MMDREG_DEVS0_PMAPMD);
+	else if (efx->loopback_mode == LOOPBACK_PMAPMD)
+		mmd_mask &= ~MDIO_MMDREG_DEVS0_PMAPMD;
+
+	while (mmd_mask) {
+		if (mmd_mask & 1) {
+			/* Double reads because link state is latched, and a
+			 * read moves the current state into the register */
+			status = mdio_clause45_read(efx, phy_id,
+						    mmd, MDIO_MMDREG_STAT1);
+			status = mdio_clause45_read(efx, phy_id,
+						    mmd, MDIO_MMDREG_STAT1);
+
+			good = status & (1 << MDIO_MMDREG_STAT1_LINK_LBN);
+			ok = ok && good;
+		}
+		mmd_mask = (mmd_mask >> 1);
+		mmd++;
+	}
+	return ok;
+}
+
+void mdio_clause45_transmit_disable(struct efx_nic *efx, int disable)
+{
+	int phy_id = efx->mii.phy_id;
+	int ctrl1, ctrl2;
+
+	ctrl1 = ctrl2 = mdio_clause45_read(efx, phy_id, MDIO_MMD_PMAPMD,
+					   MDIO_MMDREG_TXDIS);
+	if (disable)
+		ctrl2 |= (1 << MDIO_MMDREG_TXDIS_GLOBAL_LBN);
+	else
+		ctrl1 &= ~(1 << MDIO_MMDREG_TXDIS_GLOBAL_LBN);
+	if (ctrl1 != ctrl2)
+		mdio_clause45_write(efx, phy_id, MDIO_MMD_PMAPMD,
+				    MDIO_MMDREG_TXDIS, ctrl2);
+}
+
+void mdio_clause45_phy_reconfigure(struct efx_nic *efx)
+{
+	int phy_id = efx->mii.phy_id;
+	int ctrl1, ctrl2;
+
+	/* Handle (with debouncing) PMA/PMD loopback */
+	ctrl1 = ctrl2 = mdio_clause45_read(efx, phy_id, MDIO_MMD_PMAPMD,
+					   MDIO_MMDREG_CTRL1);
+
+	if (efx->loopback_mode == LOOPBACK_PMAPMD)
+		ctrl2 |= (1 << MDIO_PMAPMD_CTRL1_LBACK_LBN);
+	else
+		ctrl2 &= ~(1 << MDIO_PMAPMD_CTRL1_LBACK_LBN);
+
+	if (ctrl1 != ctrl2)
+		mdio_clause45_write(efx, phy_id, MDIO_MMD_PMAPMD,
+				    MDIO_MMDREG_CTRL1, ctrl2);
+
+	/* Handle (with debouncing) PCS loopback */
+	ctrl1 = ctrl2 = mdio_clause45_read(efx, phy_id, MDIO_MMD_PCS,
+					   MDIO_MMDREG_CTRL1);
+	if (efx->loopback_mode == LOOPBACK_PCS)
+		ctrl2 |= (1 << MDIO_MMDREG_CTRL1_LBACK_LBN);
+	else
+		ctrl2 &= ~(1 << MDIO_MMDREG_CTRL1_LBACK_LBN);
+
+	if (ctrl1 != ctrl2)
+		mdio_clause45_write(efx, phy_id, MDIO_MMD_PCS,
+				    MDIO_MMDREG_CTRL1, ctrl2);
+
+	/* Handle (with debouncing) PHYXS network loopback */
+	ctrl1 = ctrl2 = mdio_clause45_read(efx, phy_id, MDIO_MMD_PHYXS,
+					   MDIO_MMDREG_CTRL1);
+	if (efx->loopback_mode == LOOPBACK_NETWORK)
+		ctrl2 |= (1 << MDIO_MMDREG_CTRL1_LBACK_LBN);
+	else
+		ctrl2 &= ~(1 << MDIO_MMDREG_CTRL1_LBACK_LBN);
+
+	if (ctrl1 != ctrl2)
+		mdio_clause45_write(efx, phy_id, MDIO_MMD_PHYXS,
+				    MDIO_MMDREG_CTRL1, ctrl2);
+}
+
+static void mdio_clause45_set_mmd_lpower(struct efx_nic *efx,
+					 int lpower, int mmd)
+{
+	const char *mmdname = STRING_TABLE_LOOKUP(mmd, mmd_block);
+	int phy = efx->mii.phy_id;
+	int stat = mdio_clause45_read(efx, phy, mmd, MDIO_MMDREG_STAT1);
+	int ctrl1, ctrl2;
+
+	EFX_TRACE(efx, "Setting low power mode for MMD %s to %d\n",
+		  mmdname, lpower);
+
+	if (stat & (1 << MDIO_MMDREG_STAT1_LPABLE_LBN)) {
+		ctrl1 = ctrl2 = mdio_clause45_read(efx, phy,
+						   mmd, MDIO_MMDREG_CTRL1);
+		if (lpower)
+			ctrl2 |= (1 << MDIO_MMDREG_CTRL1_LPOWER_LBN);
+		else
+			ctrl2 &= ~(1 << MDIO_MMDREG_CTRL1_LPOWER_LBN);
+		if (ctrl1 != ctrl2)
+			mdio_clause45_write(efx, phy, mmd,
+					    MDIO_MMDREG_CTRL1, ctrl2);
+	} else {
+		/* If we ever want a completely generic PHY driver
+		 * that which just does clause 45, we may consider not
+		 * complaining, but for now expect the driver to know
+		 * which MMDs to apply this to. */
+		EFX_ERR(efx, "Attempt change power setting of MMD %s"
+			" which doesn't support it.\n",	mmdname);
+	}
+}
+
+void mdio_clause45_set_mmds_lpower(struct efx_nic *efx,
+				   int low_power, unsigned int mmd_mask)
+{
+	int mmd = 0;
+	while (mmd_mask) {
+		if (mmd_mask & 1)
+			mdio_clause45_set_mmd_lpower(efx, low_power, mmd);
+		mmd_mask = (mmd_mask >> 1);
+		mmd++;
+	}
+}
+
+/**
+ * mdio_clause45_get_settings - Read (some of) the PHY settings over MDIO.
+ * @efx:		Efx NIC
+ * @ecmd: 		Buffer for settings
+ *
+ * On return the 'port', 'speed', 'supported' and 'advertising' fields of
+ * ecmd have been filled out based on the PMA type.
+ */
+void mdio_clause45_get_settings(struct efx_nic *efx,
+				struct ethtool_cmd *ecmd)
+{
+	int pma_type;
+
+	/* If no PMA is present we are presumably talking something XAUI-ish
+	 * like CX4. Which we report as FIBRE (see below) */
+	if ((efx->phy_op->mmds & DEV_PRESENT_BIT(MDIO_MMD_PMAPMD)) == 0) {
+		ecmd->speed = SPEED_10000;
+		ecmd->port = PORT_FIBRE;
+		ecmd->supported = SUPPORTED_FIBRE;
+		ecmd->advertising = ADVERTISED_FIBRE;
+		return;
+	}
+
+	pma_type = mdio_clause45_read(efx, efx->mii.phy_id,
+				      MDIO_MMD_PMAPMD, MDIO_MMDREG_CTRL2);
+	pma_type &= MDIO_PMAPMD_CTRL2_TYPE_MASK;
+
+	switch (pma_type) {
+		/* We represent CX4 as fibre in the absence of anything
+		   better. */
+	case MDIO_PMAPMD_CTRL2_10G_CX4:
+		ecmd->speed = SPEED_10000;
+		ecmd->port = PORT_FIBRE;
+		ecmd->supported = SUPPORTED_FIBRE;
+		ecmd->advertising = ADVERTISED_FIBRE;
+		break;
+		/* 10G Base-T */
+	case MDIO_PMAPMD_CTRL2_10G_BT:
+		ecmd->speed = SPEED_10000;
+		ecmd->port = PORT_TP;
+		ecmd->supported = SUPPORTED_TP | SUPPORTED_10000baseT_Full;
+		ecmd->advertising = (ADVERTISED_FIBRE
+				     | ADVERTISED_10000baseT_Full);
+		break;
+	case MDIO_PMAPMD_CTRL2_1G_BT:
+		ecmd->speed = SPEED_1000;
+		ecmd->port = PORT_TP;
+		ecmd->supported = SUPPORTED_TP | SUPPORTED_1000baseT_Full;
+		ecmd->advertising = (ADVERTISED_FIBRE
+				     | ADVERTISED_1000baseT_Full);
+		break;
+	case MDIO_PMAPMD_CTRL2_100_BT:
+		ecmd->speed = SPEED_100;
+		ecmd->port = PORT_TP;
+		ecmd->supported = SUPPORTED_TP | SUPPORTED_100baseT_Full;
+		ecmd->advertising = (ADVERTISED_FIBRE
+				     | ADVERTISED_100baseT_Full);
+		break;
+	case MDIO_PMAPMD_CTRL2_10_BT:
+		ecmd->speed = SPEED_10;
+		ecmd->port = PORT_TP;
+		ecmd->supported = SUPPORTED_TP | SUPPORTED_10baseT_Full;
+		ecmd->advertising = ADVERTISED_FIBRE | ADVERTISED_10baseT_Full;
+		break;
+	/* All the other defined modes are flavours of
+	 * 10G optical */
+	default:
+		ecmd->speed = SPEED_10000;
+		ecmd->port = PORT_FIBRE;
+		ecmd->supported = SUPPORTED_FIBRE;
+		ecmd->advertising = ADVERTISED_FIBRE;
+		break;
+	}
+}
+
+/**
+ * mdio_clause45_set_settings - Set (some of) the PHY settings over MDIO.
+ * @efx:		Efx NIC
+ * @ecmd: 		New settings
+ *
+ * Currently this just enforces that we are _not_ changing the
+ * 'port', 'speed', 'supported' or 'advertising' settings as these
+ * cannot be changed on any currently supported PHY.
+ */
+int mdio_clause45_set_settings(struct efx_nic *efx,
+			       struct ethtool_cmd *ecmd)
+{
+	struct ethtool_cmd tmpcmd;
+	mdio_clause45_get_settings(efx, &tmpcmd);
+	/* None of the current PHYs support more than one mode
+	 * of operation (and only 10GBT ever will), so keep things
+	 * simple for now */
+	if ((ecmd->speed == tmpcmd.speed) && (ecmd->port == tmpcmd.port) &&
+	    (ecmd->supported == tmpcmd.supported) &&
+	    (ecmd->advertising == tmpcmd.advertising))
+		return 0;
+	return -EOPNOTSUPP;
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/mdio_10g.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/mdio_10g.h
--- linux-2.6.18.8/drivers/net/sfc/mdio_10g.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/mdio_10g.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,295 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_MDIO_10G_H
+#define EFX_MDIO_10G_H
+
+/*
+ * Definitions needed for doing 10G MDIO as specified in clause 45
+ * MDIO, which do not appear in Linux yet. Also some helper functions.
+ */
+
+#include "efx.h"
+#include "boards.h"
+
+/* Numbering of the MDIO Manageable Devices (MMDs) */
+/* Physical Medium Attachment/ Physical Medium Dependent sublayer */
+#define MDIO_MMD_PMAPMD	(1)
+/* WAN Interface Sublayer */
+#define MDIO_MMD_WIS	(2)
+/* Physical Coding Sublayer */
+#define MDIO_MMD_PCS	(3)
+/* PHY Extender Sublayer */
+#define MDIO_MMD_PHYXS	(4)
+/* Extender Sublayer */
+#define MDIO_MMD_DTEXS	(5)
+/* Transmission convergence */
+#define MDIO_MMD_TC	(6)
+/* Auto negotiation */
+#define MDIO_MMD_AN	(7)
+
+/* Generic register locations */
+#define MDIO_MMDREG_CTRL1	(0)
+#define MDIO_MMDREG_STAT1	(1)
+#define MDIO_MMDREG_IDHI	(2)
+#define MDIO_MMDREG_IDLOW	(3)
+#define MDIO_MMDREG_SPEED	(4)
+#define MDIO_MMDREG_DEVS0	(5)
+#define MDIO_MMDREG_DEVS1	(6)
+#define MDIO_MMDREG_CTRL2	(7)
+#define MDIO_MMDREG_STAT2	(8)
+#define MDIO_MMDREG_TXDIS	(9)
+
+/* Bits in MMDREG_CTRL1 */
+/* Reset */
+#define MDIO_MMDREG_CTRL1_RESET_LBN	(15)
+#define MDIO_MMDREG_CTRL1_RESET_WIDTH	(1)
+/* Loopback */
+/* Note that while WIS, PCS, PHYXS and DTEXS have the loopback bit here,
+ * the PMA/PMD does not. */
+#define MDIO_MMDREG_CTRL1_LBACK_LBN	(14)
+#define MDIO_MMDREG_CTRL1_LBACK_WIDTH	(1)
+/* Low power */
+#define MDIO_MMDREG_CTRL1_LPOWER_LBN	(11)
+#define MDIO_MMDREG_CTRL1_LPOWER_WIDTH	(1)
+
+/* Bits in MMDREG_STAT1 */
+#define MDIO_MMDREG_STAT1_FAULT_LBN	(7)
+#define MDIO_MMDREG_STAT1_FAULT_WIDTH	(1)
+/* Link state */
+#define MDIO_MMDREG_STAT1_LINK_LBN	(2)
+#define MDIO_MMDREG_STAT1_LINK_WIDTH	(1)
+/* Low power ability */
+#define MDIO_MMDREG_STAT1_LPABLE_LBN	(1)
+#define MDIO_MMDREG_STAT1_LPABLE_WIDTH	(1)
+
+/* Bits in ID reg */
+#define MDIO_ID_REV(_id32)	(_id32 & 0xf)
+#define MDIO_ID_MODEL(_id32)	((_id32 >> 4) & 0x3f)
+#define MDIO_ID_OUI(_id32)	(_id32 >> 10)
+
+/* Bits in MMDREG_DEVS0. Someone thoughtfully layed things out
+ * so the 'bit present' bit number of an MMD is the number of
+ * that MMD */
+#define DEV_PRESENT_BIT(_b) (1 << _b)
+
+#define MDIO_MMDREG_DEVS0_DTEXS	 DEV_PRESENT_BIT(MDIO_MMD_DTEXS)
+#define MDIO_MMDREG_DEVS0_PHYXS	 DEV_PRESENT_BIT(MDIO_MMD_PHYXS)
+#define MDIO_MMDREG_DEVS0_PCS	 DEV_PRESENT_BIT(MDIO_MMD_PCS)
+#define MDIO_MMDREG_DEVS0_WIS	 DEV_PRESENT_BIT(MDIO_MMD_WIS)
+#define MDIO_MMDREG_DEVS0_PMAPMD DEV_PRESENT_BIT(MDIO_MMD_PMAPMD)
+#define MDIO_MMDREG_DEVS0_AN     DEV_PRESENT_BIT(MDIO_MMD_AN)
+
+
+/* Bits in MMDREG_STAT2 */
+#define MDIO_MMDREG_STAT2_PRESENT_VAL	(2)
+#define MDIO_MMDREG_STAT2_PRESENT_LBN	(14)
+#define MDIO_MMDREG_STAT2_PRESENT_WIDTH (2)
+
+/* Bits in MMDREG_TXDIS */
+#define MDIO_MMDREG_TXDIS_GLOBAL_LBN    (0)
+#define MDIO_MMDREG_TXDIS_GLOBAL_WIDTH  (1)
+
+/* MMD-specific bits, ordered by MMD, then register */
+#define MDIO_PMAPMD_CTRL1_LBACK_LBN	(0)
+#define MDIO_PMAPMD_CTRL1_LBACK_WIDTH	(1)
+
+/* PMA type (4 bits) */
+#define MDIO_PMAPMD_CTRL2_10G_CX4	(0x0)
+#define MDIO_PMAPMD_CTRL2_10G_EW	(0x1)
+#define MDIO_PMAPMD_CTRL2_10G_LW	(0x2)
+#define MDIO_PMAPMD_CTRL2_10G_SW	(0x3)
+#define MDIO_PMAPMD_CTRL2_10G_LX4	(0x4)
+#define MDIO_PMAPMD_CTRL2_10G_ER	(0x5)
+#define MDIO_PMAPMD_CTRL2_10G_LR	(0x6)
+#define MDIO_PMAPMD_CTRL2_10G_SR	(0x7)
+/* Reserved */
+#define MDIO_PMAPMD_CTRL2_10G_BT	(0x9)
+/* Reserved */
+/* Reserved */
+#define MDIO_PMAPMD_CTRL2_1G_BT		(0xc)
+/* Reserved */
+#define MDIO_PMAPMD_CTRL2_100_BT	(0xe)
+#define MDIO_PMAPMD_CTRL2_10_BT		(0xf)
+#define MDIO_PMAPMD_CTRL2_TYPE_MASK	(0xf)
+
+/* PCS 10GBT registers */
+#define MDIO_PCS_10GBT_STATUS		(32)
+#define MDIO_PCS_10GBT_STATUS2		(33)
+#define MDIO_PCS_10GBT_STATUS2_BER_LBN	(8)
+#define MDIO_PCS_10GBT_STATUS2_BER_WIDTH	(6)
+#define MDIO_PCS_10GBT_STATUS2_ERR_LBN	(0)
+#define MDIO_PCS_10GBT_STATUS2_ERR_WIDTH	(8)
+
+/* PHY XGXS lane state */
+#define MDIO_PHYXS_LANE_STATE		(0x18)
+#define MDIO_PHYXS_LANE_ALIGNED_LBN	(12)
+#define MDIO_PHYXS_LANE_SYNC0_LBN	(0)
+#define MDIO_PHYXS_LANE_SYNC1_LBN	(1)
+#define MDIO_PHYXS_LANE_SYNC2_LBN	(2)
+#define MDIO_PHYXS_LANE_SYNC3_LBN	(3)
+
+/* AN registers */
+#define MDIO_AN_STATUS			(1)
+#define MDIO_AN_STATUS_XNP_LBN		(7)
+#define MDIO_AN_STATUS_PAGE_LBN		(6)
+#define MDIO_AN_STATUS_AN_DONE_LBN	(5)
+#define MDIO_AN_STATUS_LP_AN_CAP_LBN	(0)
+
+#define MDIO_AN_10GBT_STATUS		(33)
+#define MDIO_AN_10GBT_STATUS_MS_FLT_LBN (15) /* MASTER/SLAVE config fault */
+#define MDIO_AN_10GBT_STATUS_MS_LBN     (14) /* MASTER/SLAVE config */
+#define MDIO_AN_10GBT_STATUS_LOC_OK_LBN (13) /* Local OK */
+#define MDIO_AN_10GBT_STATUS_REM_OK_LBN (12) /* Remote OK */
+#define MDIO_AN_10GBT_STATUS_LP_10G_LBN (11) /* Link partner is 10GBT capable */
+#define MDIO_AN_10GBT_STATUS_LP_LTA_LBN (10) /* LP loop timing ability */
+#define MDIO_AN_10GBT_STATUS_LP_TRR_LBN (9)  /* LP Training Reset Request */
+
+
+/* Packing of the prt and dev arguments of clause 45 style MDIO into a
+ * single int so they can be passed into the mdio_read/write functions
+ * that currently exist. Note that as Falcon is the only current user,
+ * the packed form is chosen to match what Falcon needs to write into
+ * a register. This is checked at compile-time so do not change it. If
+ * your target chip needs things layed out differently you will need
+ * to unpack the arguments in your chip-specific mdio functions.
+ */
+ /* These are defined by the standard. */
+#define MDIO45_PRT_ID_WIDTH  (5)
+#define MDIO45_DEV_ID_WIDTH  (5)
+
+/* The prt ID is just packed in immediately to the left of the dev ID */
+#define MDIO45_PRT_DEV_WIDTH (MDIO45_PRT_ID_WIDTH + MDIO45_DEV_ID_WIDTH)
+
+#define MDIO45_PRT_ID_MASK   ((1 << MDIO45_PRT_DEV_WIDTH) - 1)
+/* This is the prt + dev extended by 1 bit to hold the 'is clause 45' flag. */
+#define MDIO45_XPRT_ID_WIDTH   (MDIO45_PRT_DEV_WIDTH + 1)
+#define MDIO45_XPRT_ID_MASK   ((1 << MDIO45_XPRT_ID_WIDTH) - 1)
+#define MDIO45_XPRT_ID_IS10G   (1 << (MDIO45_XPRT_ID_WIDTH - 1))
+
+
+#define MDIO45_PRT_ID_COMP_LBN   MDIO45_DEV_ID_WIDTH
+#define MDIO45_PRT_ID_COMP_WIDTH  MDIO45_PRT_ID_WIDTH
+#define MDIO45_DEV_ID_COMP_LBN    0
+#define MDIO45_DEV_ID_COMP_WIDTH  MDIO45_DEV_ID_WIDTH
+
+/* Compose port and device into a phy_id */
+static inline int mdio_clause45_pack(u8 prt, u8 dev)
+{
+	efx_dword_t phy_id;
+	EFX_POPULATE_DWORD_2(phy_id, MDIO45_PRT_ID_COMP, prt,
+			     MDIO45_DEV_ID_COMP, dev);
+	return MDIO45_XPRT_ID_IS10G | EFX_DWORD_VAL(phy_id);
+}
+
+static inline void mdio_clause45_unpack(u32 val, u8 *prt, u8 *dev)
+{
+	efx_dword_t phy_id;
+	EFX_POPULATE_DWORD_1(phy_id, EFX_DWORD_0, val);
+	*prt = EFX_DWORD_FIELD(phy_id, MDIO45_PRT_ID_COMP);
+	*dev = EFX_DWORD_FIELD(phy_id, MDIO45_DEV_ID_COMP);
+}
+
+static inline int mdio_clause45_read(struct efx_nic *efx,
+				     u8 prt, u8 dev, u16 addr)
+{
+	return efx->mii.mdio_read(efx->net_dev,
+				  mdio_clause45_pack(prt, dev), addr);
+}
+
+static inline void mdio_clause45_write(struct efx_nic *efx,
+				       u8 prt, u8 dev, u16 addr, int value)
+{
+	efx->mii.mdio_write(efx->net_dev,
+			    mdio_clause45_pack(prt, dev), addr, value);
+}
+
+
+static inline u32 mdio_clause45_read_id(struct efx_nic *efx, int mmd)
+{
+	int phy_id = efx->mii.phy_id;
+	u16 id_low = mdio_clause45_read(efx, phy_id, mmd, MDIO_MMDREG_IDLOW);
+	u16 id_hi = mdio_clause45_read(efx, phy_id, mmd, MDIO_MMDREG_IDHI);
+	return (id_hi << 16) | (id_low);
+}
+
+static inline int mdio_clause45_phyxgxs_lane_sync(struct efx_nic *efx)
+{
+	int i, sync, lane_status;
+
+	for (i = 0; i < 2; ++i)
+		lane_status = mdio_clause45_read(efx, efx->mii.phy_id,
+						 MDIO_MMD_PHYXS,
+						 MDIO_PHYXS_LANE_STATE);
+
+	sync = (lane_status & (1 << MDIO_PHYXS_LANE_ALIGNED_LBN)) != 0;
+	if (!sync)
+		EFX_INFO(efx, "XGXS lane status: %x\n", lane_status);
+	return sync;
+}
+
+extern const char *mdio_clause45_mmd_name(int mmd);
+
+/*
+ * Reset a specific MMD and wait for reset to clear.
+ * Return number of spins left (>0) on success, -%ETIMEDOUT on failure.
+ *
+ * This function will sleep
+ */
+extern int mdio_clause45_reset_mmd(struct efx_nic *efx, int mmd,
+				   int spins, int spintime);
+
+/* As mdio_clause45_check_mmd but for multiple MMDs */
+int mdio_clause45_check_mmds(struct efx_nic *efx,
+			     unsigned int mmd_mask, unsigned int fatal_mask);
+
+/* Check the link status of specified mmds in bit mask */
+extern int mdio_clause45_links_ok(struct efx_nic *efx,
+				  unsigned int mmd_mask);
+
+/* Generic transmit disable support though PMAPMD */
+extern void mdio_clause45_transmit_disable(struct efx_nic *efx,
+					   int disable);
+
+/* Generic part of reconfigure: set/clear loopback bits */
+extern void mdio_clause45_phy_reconfigure(struct efx_nic *efx);
+
+/* Set the power state of the specified MMDs */
+extern void mdio_clause45_set_mmds_lpower(struct efx_nic *efx,
+					  int low_power, unsigned int mmd_mask);
+
+/* Read (some of) the PHY settings over MDIO */
+extern void mdio_clause45_get_settings(struct efx_nic *efx,
+				       struct ethtool_cmd *ecmd);
+
+/* Set (some of) the PHY settings over MDIO */
+extern int mdio_clause45_set_settings(struct efx_nic *efx,
+				      struct ethtool_cmd *ecmd);
+
+/* Wait for specified MMDs to exit reset within a timeout */
+extern int mdio_clause45_wait_reset_mmds(struct efx_nic *efx,
+					 unsigned int mmd_mask);
+
+#endif /* EFX_MDIO_10G_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/mentormac.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/mentormac.c
--- linux-2.6.18.8/drivers/net/sfc/mentormac.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/mentormac.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,506 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/delay.h>
+#include "net_driver.h"
+#include "gmii.h"
+#include "mac.h"
+
+/*
+ * Mentor MAC control
+ */
+
+/**************************************************************************
+ *
+ * Mentor MAC registers
+ *
+ **************************************************************************
+ *
+ * Register addresses are Mentor MAC register numbers.  Falcon maps these
+ * registers in at 16-byte intervals.  The mac_writel() and mac_readl()
+ * methods take care of abstracting away this difference.
+ */
+
+/* GMAC configuration register 1 */
+#define GM_CFG1_REG_MAC 0x00
+#define GM_SW_RST_LBN 31
+#define GM_SW_RST_WIDTH 1
+#define GM_SIM_RST_LBN 30
+#define GM_SIM_RST_WIDTH 1
+#define GM_RST_RX_MAC_CTL_LBN 19
+#define GM_RST_RX_MAC_CTL_WIDTH 1
+#define GM_RST_TX_MAC_CTL_LBN 18
+#define GM_RST_TX_MAC_CTL_WIDTH 1
+#define GM_RST_RX_FUNC_LBN 17
+#define GM_RST_RX_FUNC_WIDTH 1
+#define GM_RST_TX_FUNC_LBN 16
+#define GM_RST_TX_FUNC_WIDTH 1
+#define GM_LOOP_LBN 8
+#define GM_LOOP_WIDTH 1
+#define GM_RX_FC_EN_LBN 5
+#define GM_RX_FC_EN_WIDTH 1
+#define GM_TX_FC_EN_LBN 4
+#define GM_TX_FC_EN_WIDTH 1
+#define GM_SYNC_RXEN_LBN 3
+#define GM_SYNC_RXEN_WIDTH 1
+#define GM_RX_EN_LBN 2
+#define GM_RX_EN_WIDTH 1
+#define GM_SYNC_TXEN_LBN 1
+#define GM_SYNC_TXEN_WIDTH 1
+#define GM_TX_EN_LBN 0
+#define GM_TX_EN_WIDTH 1
+
+/* GMAC configuration register 2 */
+#define GM_CFG2_REG_MAC 0x01
+#define GM_PAMBL_LEN_LBN 12
+#define GM_PAMBL_LEN_WIDTH 4
+#define GM_IF_MODE_LBN 8
+#define GM_IF_MODE_WIDTH 2
+#define GM_HUGE_FRM_EN_LBN 5
+#define GM_HUGE_FRM_EN_WIDTH 1
+#define GM_LEN_CHK_LBN 4
+#define GM_LEN_CHK_WIDTH 1
+#define GM_PAD_CRC_EN_LBN 2
+#define GM_PAD_CRC_EN_WIDTH 1
+#define GM_CRC_EN_LBN 1
+#define GM_CRC_EN_WIDTH 1
+#define GM_FD_LBN 0
+#define GM_FD_WIDTH 1
+
+/* GMAC maximum frame length register */
+#define GM_MAX_FLEN_REG_MAC 0x04
+#define GM_MAX_FLEN_LBN 0
+#define GM_MAX_FLEN_WIDTH 16
+
+/* GMAC MII management configuration register */
+#define GM_MII_MGMT_CFG_REG_MAC 0x08
+#define GM_RST_MII_MGMT_LBN 31
+#define GM_RST_MII_MGMT_WIDTH 1
+#define GM_MGMT_SCAN_AUTO_INC_LBN 5
+#define GM_MGMT_SCAN_AUTO_INC_WIDTH 1
+#define GM_MGMT_PREM_SUPR_LBN 4
+#define GM_MGMT_PREM_SUPR_WIDTH 1
+#define GM_MGMT_CLK_SEL_LBN 0
+#define GM_MGMT_CLK_SEL_WIDTH 3
+
+/* GMAC MII management command register */
+#define GM_MII_MGMT_CMD_REG_MAC 0x09
+#define GM_MGMT_SCAN_CYC_LBN 1
+#define GM_MGMT_SCAN_CYC_WIDTH 1
+#define GM_MGMT_RD_CYC_LBN 0
+#define GM_MGMT_RD_CYC_WIDTH 1
+
+/* GMAC MII management address register */
+#define GM_MII_MGMT_ADR_REG_MAC 0x0a
+#define GM_MGMT_PHY_ADDR_LBN 8
+#define GM_MGMT_PHY_ADDR_WIDTH 5
+#define GM_MGMT_REG_ADDR_LBN 0
+#define GM_MGMT_REG_ADDR_WIDTH 5
+
+/* GMAC MII management control register */
+#define GM_MII_MGMT_CTL_REG_MAC 0x0b
+#define GM_MGMT_CTL_LBN 0
+#define GM_MGMT_CTL_WIDTH 16
+
+/* GMAC MII management status register */
+#define GM_MII_MGMT_STAT_REG_MAC 0x0c
+#define GM_MGMT_STAT_LBN 0
+#define GM_MGMT_STAT_WIDTH 16
+
+/* GMAC MII management indicators register */
+#define GM_MII_MGMT_IND_REG_MAC 0x0d
+#define GM_MGMT_NOT_VLD_LBN 2
+#define GM_MGMT_NOT_VLD_WIDTH 1
+#define GM_MGMT_SCANNING_LBN 1
+#define GM_MGMT_SCANNING_WIDTH 1
+#define GM_MGMT_BUSY_LBN 0
+#define GM_MGMT_BUSY_WIDTH 1
+
+/* GMAC station address register 1 */
+#define GM_ADR1_REG_MAC 0x10
+#define GM_HWADDR_5_LBN 24
+#define GM_HWADDR_5_WIDTH 8
+#define GM_HWADDR_4_LBN 16
+#define GM_HWADDR_4_WIDTH 8
+#define GM_HWADDR_3_LBN 8
+#define GM_HWADDR_3_WIDTH 8
+#define GM_HWADDR_2_LBN 0
+#define GM_HWADDR_2_WIDTH 8
+
+/* GMAC station address register 2 */
+#define GM_ADR2_REG_MAC 0x11
+#define GM_HWADDR_1_LBN 24
+#define GM_HWADDR_1_WIDTH 8
+#define GM_HWADDR_0_LBN 16
+#define GM_HWADDR_0_WIDTH 8
+
+/* GMAC FIFO configuration register 0 */
+#define GMF_CFG0_REG_MAC 0x12
+#define GMF_FTFENRPLY_LBN 20
+#define GMF_FTFENRPLY_WIDTH 1
+#define GMF_STFENRPLY_LBN 19
+#define GMF_STFENRPLY_WIDTH 1
+#define GMF_FRFENRPLY_LBN 18
+#define GMF_FRFENRPLY_WIDTH 1
+#define GMF_SRFENRPLY_LBN 17
+#define GMF_SRFENRPLY_WIDTH 1
+#define GMF_WTMENRPLY_LBN 16
+#define GMF_WTMENRPLY_WIDTH 1
+#define GMF_FTFENREQ_LBN 12
+#define GMF_FTFENREQ_WIDTH 1
+#define GMF_STFENREQ_LBN 11
+#define GMF_STFENREQ_WIDTH 1
+#define GMF_FRFENREQ_LBN 10
+#define GMF_FRFENREQ_WIDTH 1
+#define GMF_SRFENREQ_LBN 9
+#define GMF_SRFENREQ_WIDTH 1
+#define GMF_WTMENREQ_LBN 8
+#define GMF_WTMENREQ_WIDTH 1
+#define GMF_HSTRSTFT_LBN 4
+#define GMF_HSTRSTFT_WIDTH 1
+#define GMF_HSTRSTST_LBN 3
+#define GMF_HSTRSTST_WIDTH 1
+#define GMF_HSTRSTFR_LBN 2
+#define GMF_HSTRSTFR_WIDTH 1
+#define GMF_HSTRSTSR_LBN 1
+#define GMF_HSTRSTSR_WIDTH 1
+#define GMF_HSTRSTWT_LBN 0
+#define GMF_HSTRSTWT_WIDTH 1
+
+/* GMAC FIFO configuration register 1 */
+#define GMF_CFG1_REG_MAC 0x13
+#define GMF_CFGFRTH_LBN 16
+#define GMF_CFGFRTH_WIDTH 5
+#define GMF_CFGXOFFRTX_LBN 0
+#define GMF_CFGXOFFRTX_WIDTH 16
+
+/* GMAC FIFO configuration register 2 */
+#define GMF_CFG2_REG_MAC 0x14
+#define GMF_CFGHWM_LBN 16
+#define GMF_CFGHWM_WIDTH 6
+#define GMF_CFGLWM_LBN 0
+#define GMF_CFGLWM_WIDTH 6
+
+/* GMAC FIFO configuration register 3 */
+#define GMF_CFG3_REG_MAC 0x15
+#define GMF_CFGHWMFT_LBN 16
+#define GMF_CFGHWMFT_WIDTH 6
+#define GMF_CFGFTTH_LBN 0
+#define GMF_CFGFTTH_WIDTH 6
+
+/* GMAC FIFO configuration register 4 */
+#define GMF_CFG4_REG_MAC 0x16
+#define GMF_HSTFLTRFRM_LBN 0
+#define GMF_HSTFLTRFRM_WIDTH 18
+#define GMF_HSTFLTRFRM_PAUSE_LBN 12
+#define GMF_HSTFLTRFRM_PAUSE_WIDTH 12
+
+/* GMAC FIFO configuration register 5 */
+#define GMF_CFG5_REG_MAC 0x17
+#define GMF_CFGHDPLX_LBN 22
+#define GMF_CFGHDPLX_WIDTH 1
+#define GMF_SRFULL_LBN 21
+#define GMF_SRFULL_WIDTH 1
+#define GMF_HSTSRFULLCLR_LBN 20
+#define GMF_HSTSRFULLCLR_WIDTH 1
+#define GMF_CFGBYTMODE_LBN 19
+#define GMF_CFGBYTMODE_WIDTH 1
+#define GMF_HSTDRPLT64_LBN 18
+#define GMF_HSTDRPLT64_WIDTH 1
+#define GMF_HSTFLTRFRMDC_LBN 0
+#define GMF_HSTFLTRFRMDC_WIDTH 18
+#define GMF_HSTFLTRFRMDC_PAUSE_LBN 12
+#define GMF_HSTFLTRFRMDC_PAUSE_WIDTH 1
+
+/* TX total octet count */
+#define GM_TX_OCT_CNT_REG_MAC 0x40
+#define GM_STAT_LBN 0
+#define GM_STAT_WIDTH 32
+
+/* TX good octet count */
+#define GM_TX_GOOD_OCT_CNT_REG_MAC 0x41
+
+/* TX single collision packet count */
+#define GM_TX_SGLCOL_PKT_CNT_REG_MAC 0x42
+
+/* TX multiple collision packet count */
+#define GM_TX_MULTCOL_PKT_CNT_REG_MAC 0x43
+
+/* TX excessive collision packet count */
+#define GM_TX_EXCOL_PKT_CNT_REG_MAC 0x44
+
+/* TX deferred packet count */
+#define GM_TX_DEF_PKT_CNT_REG_MAC 0x45
+
+/* TX late packet count */
+#define GM_TX_LATECOL_PKT_CNT_REG_MAC 0x46
+
+/* TX excessive deferral packet count */
+#define GM_TX_EXDEF_PKT_CNT_REG_MAC 0x47
+
+/* TX pause packet count */
+#define GM_TX_PAUSE_PKT_CNT_REG_MAC 0x48
+
+/* TX bad packet count */
+#define GM_TX_BAD_PKT_CNT_REG_MAC 0x49
+
+/* TX unicast packet count */
+#define GM_TX_UCAST_PKT_CNT_REG_MAC 0x4a
+
+/* TX multicast packet count */
+#define GM_TX_MCAST_PKT_CNT_REG_MAC 0x4b
+
+/* TX broadcast packet count */
+#define GM_TX_BCAST_PKT_CNT_REG_MAC 0x4c
+
+/* TX <64-byte packet count */
+#define GM_TX_LT64_PKT_CNT_REG_MAC 0x4d
+
+/* TX 64-byte packet count */
+#define GM_TX_64_PKT_CNT_REG_MAC 0x4e
+
+/* TX 65-byte to 127-byte packet count */
+#define GM_TX_65_TO_127_PKT_CNT_REG_MAC 0x4f
+
+/* TX 128-byte to 255-byte packet count */
+#define GM_TX_128_TO_255_PKT_CNT_REG_MAC 0x50
+
+/* TX 256-byte to 511-byte packet count */
+#define GM_TX_256_TO_511_PKT_CNT_REG_MAC 0x51
+
+/* TX 512-byte to 1023-byte packet count */
+#define GM_TX_512_TO_1023_PKT_CNT_REG_MAC 0x52
+
+/* TX 1024-byte to 15xx-byte packet count */
+#define GM_TX_1024_TO_15XX_PKT_CNT_REG_MAC 0x53
+
+/* TX 15xx-byte to jumbo packet count */
+#define GM_TX_15XX_TO_JUMBO_PKT_CNT_REG_MAC 0x54
+
+/* TX >jumbo packet count */
+#define GM_TX_GTJUMBO_PKT_CNT_REG_MAC 0x55
+
+/* RX good octet count */
+#define GM_RX_GOOD_OCT_CNT_REG_MAC 0x60
+
+/* RX bad octet count */
+#define GM_RX_BAD_OCT_CNT_REG_MAC 0x61
+
+/* RX missed packet count */
+#define GM_RX_MISS_PKT_CNT_REG_MAC 0x62
+
+/* RX false carrier count */
+#define GM_RX_FALSE_CRS_CNT_REG_MAC 0x63
+
+/* RX pause packet count */
+#define GM_RX_PAUSE_PKT_CNT_REG_MAC 0x64
+
+/* RX bad packet count */
+#define GM_RX_BAD_PKT_CNT_REG_MAC 0x65
+
+/* RX unicast packet count */
+#define GM_RX_UCAST_PKT_CNT_REG_MAC 0x66
+
+/* RX multicast packet count */
+#define GM_RX_MCAST_PKT_CNT_REG_MAC 0x67
+
+/* RX broadcast packet count */
+#define GM_RX_BCAST_PKT_CNT_REG_MAC 0x68
+
+/* RX <64-byte good packet count */
+#define GM_RX_GOOD_LT64_PKT_CNT_REG_MAC 0x69
+
+/* RX <64-byte bad packet count */
+#define GM_RX_BAD_LT64_PKT_CNT_REG_MAC 0x6a
+
+/* RX 64-byte packet count */
+#define GM_RX_64_PKT_CNT_REG_MAC 0x6b
+
+/* RX 65-byte to 127-byte packet count */
+#define GM_RX_65_TO_127_PKT_CNT_REG_MAC 0x6c
+
+/* RX 128-byte to 255-byte packet count*/
+#define GM_RX_128_TO_255_PKT_CNT_REG_MAC 0x6d
+
+/* RX 256-byte to 511-byte packet count */
+#define GM_RX_256_TO_511_PKT_CNT_REG_MAC 0x6e
+
+/* RX 512-byte to 1023-byte packet count */
+#define GM_RX_512_TO_1023_PKT_CNT_REG_MAC 0x6f
+
+/* RX 1024-byte to 15xx-byte packet count */
+#define GM_RX_1024_TO_15XX_PKT_CNT_REG_MAC 0x70
+
+/* RX 15xx-byte to jumbo packet count */
+#define GM_RX_15XX_TO_JUMBO_PKT_CNT_REG_MAC 0x71
+
+/* RX >jumbo packet count */
+#define GM_RX_GTJUMBO_PKT_CNT_REG_MAC 0x72
+
+/* RX 64-byte to 15xx-byte bad crc packet count */
+#define GM_RX_BAD_64_TO_15XX_PKT_CNT_REG_MAC 0x73
+
+/* RX 15xx-byte to jumbo bad crc packet count */
+#define GM_RX_BAD_15XX_TO_JUMBO_PKT_CNT_REG_MAC 0x74
+
+/* RX >jumbo bad crc packet count */
+#define GM_RX_BAD_GTJUMBO_PKT_CNT_REG_MAC 0x75
+
+/**************************************************************************
+ *
+ * GMII access to PHY
+ *
+ **************************************************************************
+ */
+
+/* This does not reset the PHY, only the MAC.  However, TX and RX will
+ * both be disabled on the MAC after this, so the state of the PHY is
+ * somewhat irrelevant until the MAC is reinitialised.
+ */
+void mentormac_reset(struct efx_nic *efx)
+{
+	efx_dword_t reg;
+
+	EFX_POPULATE_DWORD_1(reg, GM_SW_RST, 1);
+	efx->mac_op->mac_writel(efx, &reg, GM_CFG1_REG_MAC);
+	udelay(1000);
+
+	EFX_POPULATE_DWORD_1(reg, GM_SW_RST, 0);
+	efx->mac_op->mac_writel(efx, &reg, GM_CFG1_REG_MAC);
+	udelay(1000);
+
+	/* Configure GMII interface so PHY is accessible */
+	EFX_POPULATE_DWORD_1(reg, GM_MGMT_CLK_SEL, 0x4);
+	efx->mac_op->mac_writel(efx, &reg,
+				     GM_MII_MGMT_CFG_REG_MAC);
+	udelay(10);
+}
+
+void mentormac_reconfigure(struct efx_nic *efx)
+{
+	int loopback, tx_fc, rx_fc, if_mode, full_duplex, bytemode, half_duplex;
+	unsigned int max_frame_len;
+	efx_dword_t reg;
+
+	/* Configuration register 1 */
+	tx_fc = (efx->flow_control & EFX_FC_TX) ? 1 : 0;
+	rx_fc = (efx->flow_control & EFX_FC_RX) ? 1 : 0;
+	loopback = (efx->loopback_mode == LOOPBACK_MAC) ? 1 : 0;
+	bytemode = (efx->link_options & GM_LPA_1000) ? 1 : 0;
+
+	if (efx->loopback_mode != LOOPBACK_NONE)
+		bytemode = 1;
+	if (!(efx->link_options & GM_LPA_DUPLEX))
+		/* Half-duplex operation requires TX flow control */
+		tx_fc = 1;
+	EFX_POPULATE_DWORD_5(reg,
+			     GM_LOOP, loopback,
+			     GM_TX_EN, 1,
+			     GM_TX_FC_EN, tx_fc,
+			     GM_RX_EN, 1,
+			     GM_RX_FC_EN, rx_fc);
+	efx->mac_op->mac_writel(efx, &reg, GM_CFG1_REG_MAC);
+	udelay(10);
+
+	/* Configuration register 2 */
+	if_mode = (bytemode) ? 2 : 1;
+	full_duplex = (efx->link_options & GM_LPA_DUPLEX) ? 1 : 0;
+	EFX_POPULATE_DWORD_4(reg,
+			     GM_IF_MODE, if_mode,
+			     GM_PAD_CRC_EN, 1,
+			     GM_FD, full_duplex,
+			     GM_PAMBL_LEN, 0x7/*datasheet recommended */);
+
+	efx->mac_op->mac_writel(efx, &reg, GM_CFG2_REG_MAC);
+	udelay(10);
+
+	/* Max frame len register */
+	max_frame_len = EFX_MAX_FRAME_LEN(efx->net_dev->mtu);
+	EFX_POPULATE_DWORD_1(reg, GM_MAX_FLEN, max_frame_len);
+	efx->mac_op->mac_writel(efx, &reg, GM_MAX_FLEN_REG_MAC);
+	udelay(10);
+
+	/* FIFO configuration register 0 */
+	EFX_POPULATE_DWORD_5(reg,
+			     GMF_FTFENREQ, 1,
+			     GMF_STFENREQ, 1,
+			     GMF_FRFENREQ, 1,
+			     GMF_SRFENREQ, 1,
+			     GMF_WTMENREQ, 1);
+	efx->mac_op->mac_writel(efx, &reg, GMF_CFG0_REG_MAC);
+	udelay(10);
+
+	/* FIFO configuration register 1 */
+	EFX_POPULATE_DWORD_2(reg,
+			     GMF_CFGFRTH, 0x12,
+			     GMF_CFGXOFFRTX, 0xffff);
+	efx->mac_op->mac_writel(efx, &reg, GMF_CFG1_REG_MAC);
+	udelay(10);
+
+	/* FIFO configuration register 2 */
+	EFX_POPULATE_DWORD_2(reg,
+			     GMF_CFGHWM, 0x3f,
+			     GMF_CFGLWM, 0xa);
+	efx->mac_op->mac_writel(efx, &reg, GMF_CFG2_REG_MAC);
+	udelay(10);
+
+	/* FIFO configuration register 3 */
+	EFX_POPULATE_DWORD_2(reg,
+			     GMF_CFGHWMFT, 0x1c,
+			     GMF_CFGFTTH, 0x08);
+	efx->mac_op->mac_writel(efx, &reg, GMF_CFG3_REG_MAC);
+	udelay(10);
+
+	/* FIFO configuration register 4 */
+	EFX_POPULATE_DWORD_1(reg, GMF_HSTFLTRFRM_PAUSE, 1);
+	efx->mac_op->mac_writel(efx, &reg, GMF_CFG4_REG_MAC);
+	udelay(10);
+
+	/* FIFO configuration register 5 */
+	half_duplex = (efx->link_options & GM_LPA_DUPLEX) ? 0 : 1;
+	efx->mac_op->mac_readl(efx, &reg, GMF_CFG5_REG_MAC);
+	EFX_SET_DWORD_FIELD(reg, GMF_CFGBYTMODE, bytemode);
+	EFX_SET_DWORD_FIELD(reg, GMF_CFGHDPLX, half_duplex);
+	EFX_SET_DWORD_FIELD(reg, GMF_HSTDRPLT64, half_duplex);
+	EFX_SET_DWORD_FIELD(reg, GMF_HSTFLTRFRMDC_PAUSE, 0);
+	efx->mac_op->mac_writel(efx, &reg, GMF_CFG5_REG_MAC);
+	udelay(10);
+
+	/* MAC address */
+	EFX_POPULATE_DWORD_4(reg,
+			     GM_HWADDR_5, efx->net_dev->dev_addr[5],
+			     GM_HWADDR_4, efx->net_dev->dev_addr[4],
+			     GM_HWADDR_3, efx->net_dev->dev_addr[3],
+			     GM_HWADDR_2, efx->net_dev->dev_addr[2]);
+	efx->mac_op->mac_writel(efx, &reg, GM_ADR1_REG_MAC);
+	udelay(10);
+	EFX_POPULATE_DWORD_2(reg,
+			     GM_HWADDR_1, efx->net_dev->dev_addr[1],
+			     GM_HWADDR_0, efx->net_dev->dev_addr[0]);
+	efx->mac_op->mac_writel(efx, &reg, GM_ADR2_REG_MAC);
+	udelay(10);
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/mtd.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/mtd.c
--- linux-2.6.18.8/drivers/net/sfc/mtd.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/mtd.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,602 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/module.h>
+#include <linux/mtd/mtd.h>
+#include <linux/mtd/partitions.h>
+#include <linux/delay.h>
+
+#define EFX_DRIVER_NAME "sfc_mtd"
+#include "driverlink_api.h"
+#include "net_driver.h"
+#include "spi.h"
+
+/*
+ * Flash and EEPROM (MTD) device driver
+ *
+ * This file provides a separate kernel module (sfc_mtd) which
+ * exposes the flash and EEPROM devices present on Solarflare NICs as
+ * MTD devices, enabling you to reflash the boot ROM code (or use the
+ * remaining space on the flash as a jffs2 filesystem, should you want
+ * to do so).
+ */
+
+#define EFX_MTD_VERIFY_BUF_LEN 16
+#define EFX_MAX_PARTITIONS 2
+#define EFX_FLASH_BOOTROM_OFFSET 0x8000U
+
+/* Write enable for EEPROM/flash configuration area
+ *
+ * Normally, writes to parts of non-volatile storage which contain
+ * critical configuration are disabled to prevent accidents.  This
+ * parameter allows enabling of such writes.
+ */
+static unsigned int efx_allow_nvconfig_writes;
+
+struct efx_mtd {
+	struct mtd_info mtd;
+	struct mtd_partition part[EFX_MAX_PARTITIONS];
+	char part_name[EFX_MAX_PARTITIONS][32];
+	char name[32];
+	struct efx_dl_device *efx_dev;
+	struct efx_nic *efx;
+	/* This must be held when using *spi; it guards against races
+	 * with device reset and between sequences of dependent
+	 * commands. */
+	struct semaphore access_lock;
+	struct efx_spi_device *spi;
+};
+
+/* SPI utilities */
+
+static int efx_spi_fast_wait(struct efx_mtd *efx_mtd)
+{
+	struct efx_spi_device *spi = efx_mtd->spi;
+	u8 status;
+	int i, rc;
+
+	/* Wait up to 1000us for flash/EEPROM to finish a fast operation. */
+	for (i = 0; i < 50; i++) {
+		udelay(20);
+
+		rc = spi->read(spi, efx_mtd->efx, SPI_RDSR, -1,
+			       &status, sizeof(status));
+		if (rc)
+			return rc;
+		if (!(status & SPI_STATUS_NRDY))
+			return 0;
+	}
+	EFX_ERR(efx_mtd->efx, "timed out waiting for %s last status=0x%02x\n",
+		efx_mtd->name, status);
+	return -ETIMEDOUT;
+}
+
+static int efx_spi_slow_wait(struct efx_mtd *efx_mtd, int uninterruptible)
+{
+	struct efx_spi_device *spi = efx_mtd->spi;
+	u8 status;
+	int rc, i;
+
+	/* Wait up to 4s for flash/EEPROM to finish a slow operation. */
+	for (i = 0; i < 40; i++) {
+		__set_current_state(uninterruptible ?
+				    TASK_UNINTERRUPTIBLE : TASK_INTERRUPTIBLE);
+		schedule_timeout(HZ / 10);
+		rc = spi->read(spi, efx_mtd->efx, SPI_RDSR, -1,
+			       &status, sizeof(status));
+		if (rc)
+			return rc;
+		if (!(status & SPI_STATUS_NRDY))
+			return 0;
+		if (signal_pending(current))
+			return -EINTR;
+	}
+	EFX_ERR(efx_mtd->efx, "timed out waiting for %s\n", efx_mtd->name);
+	return -ETIMEDOUT;
+}
+
+static int
+efx_spi_write_enable(struct efx_mtd *efx_mtd)
+{
+	struct efx_spi_device *spi = efx_mtd->spi;
+
+	return spi->write(spi, efx_mtd->efx, SPI_WREN, -1, NULL, 0);
+}
+
+static int efx_spi_unlock(struct efx_mtd *efx_mtd)
+{
+	struct efx_spi_device *spi = efx_mtd->spi;
+	const u8 unlock_mask = (SPI_STATUS_BP2 | SPI_STATUS_BP1 |
+				     SPI_STATUS_BP0);
+	u8 status;
+	int rc;
+
+	rc = spi->read(spi, efx_mtd->efx, SPI_RDSR, -1, &status,
+		       sizeof(status));
+	if (rc)
+		return rc;
+
+	if (!(status & unlock_mask))
+		return 0; /* already unlocked */
+
+	rc = efx_spi_write_enable(efx_mtd);
+	if (rc)
+		return rc;
+	rc = spi->write(spi, efx_mtd->efx, SPI_SST_EWSR, -1, NULL, 0);
+	if (rc)
+		return rc;
+
+	status &= ~unlock_mask;
+	rc = spi->write(spi, efx_mtd->efx, SPI_WRSR, -1, &status,
+			sizeof(status));
+	if (rc)
+		return rc;
+	rc = efx_spi_fast_wait(efx_mtd);
+	if (rc)
+		return rc;
+
+	return 0;
+}
+
+/* Dummy device used in case of a failed reset */
+
+static int efx_spi_dummy_read(const struct efx_spi_device *spi,
+			      struct efx_nic *efx, unsigned int command,
+			      int address, void *data, unsigned int len)
+{
+	return -EIO;
+}
+
+static int efx_spi_dummy_write(const struct efx_spi_device *spi,
+			       struct efx_nic *efx, unsigned int command,
+			       int address, const void *data, unsigned int len)
+{
+	return -EIO;
+}
+
+static struct efx_spi_device efx_spi_dummy_device = {
+	.block_size	= 1,
+	.erase_command	= 0xff,
+	.read		= efx_spi_dummy_read,
+	.write		= efx_spi_dummy_write,
+};
+
+/* MTD interface */
+
+static int efx_mtd_read(struct mtd_info *mtd, loff_t start, size_t len,
+			size_t *retlen, u8 *buffer)
+{
+	struct efx_mtd *efx_mtd = mtd->priv;
+	struct efx_spi_device *spi;
+	unsigned int command;
+	unsigned int block_len;
+	unsigned int pos = 0;
+	int rc;
+
+	rc = down_interruptible(&efx_mtd->access_lock);
+	if (rc)
+		goto out;
+	spi = efx_mtd->spi;
+
+	while (pos < len) {
+		block_len = min((unsigned int)len - pos,
+				efx_spi_read_limit(spi, start + pos));
+		command = efx_spi_munge_command(spi, SPI_READ, start + pos);
+		rc = spi->read(spi, efx_mtd->efx, command, start + pos,
+			       buffer + pos, block_len);
+		if (rc)
+			break;
+		pos += block_len;
+
+		/* Avoid locking up the system */
+		cond_resched();
+		if (signal_pending(current)) {
+			rc = -EINTR;
+			break;
+		}
+	}
+
+	up(&efx_mtd->access_lock);
+out:
+	*retlen = pos;
+	return rc;
+}
+
+/* Check that device contents match buffer.  If repeat is true, buffer
+ * contains a pattern of length EFX_MTD_VERIFY_BUF_LEN which the
+ * device contents should match repeatedly.
+ */
+static int efx_mtd_verify(struct mtd_info *mtd, loff_t start,
+			  size_t len, const u8 *buffer, int repeat)
+{
+	u8 verify_buffer[EFX_MTD_VERIFY_BUF_LEN];
+	unsigned int block_len;
+	size_t read_len;
+	unsigned int pos = 0;
+	int rc = 0;
+
+	while (pos < len) {
+		block_len = min(len - pos, sizeof(verify_buffer));
+		rc = efx_mtd_read(mtd, start + pos, block_len, &read_len,
+				  verify_buffer);
+		if (rc)
+			return rc;
+		if (memcmp(repeat ? buffer : buffer + pos, verify_buffer,
+			   block_len))
+			return -EIO;
+		pos += block_len;
+	}
+
+	return 0;
+}
+
+static int efx_mtd_erase(struct mtd_info *mtd, struct erase_info *erase)
+{
+	struct efx_mtd *efx_mtd = mtd->priv;
+	struct efx_spi_device *spi;
+	u8 empty[EFX_MTD_VERIFY_BUF_LEN];
+	int rc;
+
+	if (erase->len != mtd->erasesize) {
+		rc = -EINVAL;
+		goto out;
+	}
+
+	rc = down_interruptible(&efx_mtd->access_lock);
+	if (rc)
+		goto out;
+	spi = efx_mtd->spi;
+	if (spi->erase_command == 0) {
+		rc = -EOPNOTSUPP;
+		goto out_up;
+	}
+
+	rc = efx_spi_unlock(efx_mtd);
+	if (rc)
+		goto out_up;
+	rc = efx_spi_write_enable(efx_mtd);
+	if (rc)
+		goto out_up;
+	rc = spi->write(spi, efx_mtd->efx, spi->erase_command, erase->addr,
+			NULL, 0);
+	if (rc)
+		goto out_up;
+	rc = efx_spi_slow_wait(efx_mtd, 0);
+
+out_up:
+	up(&efx_mtd->access_lock);
+	if (rc)
+		goto out;
+
+	memset(empty, 0xff, sizeof(empty));
+	rc = efx_mtd_verify(mtd, erase->addr, erase->len, empty, 1);
+
+out:
+	if (rc == 0) {
+		erase->state = MTD_ERASE_DONE;
+	} else {
+		erase->state = MTD_ERASE_FAILED;
+#if defined(EFX_USE_MTD_ERASE_FAIL_ADDR)
+		erase->fail_addr = 0xffffffff;
+#endif
+	}
+	mtd_erase_callback(erase);
+	return rc;
+}
+
+static int efx_mtd_write(struct mtd_info *mtd, loff_t start,
+			 size_t len, size_t *retlen, const u8 *buffer)
+{
+	struct efx_mtd *efx_mtd = mtd->priv;
+	struct efx_spi_device *spi;
+	unsigned int command;
+	unsigned int block_len;
+	unsigned int pos = 0;
+	int rc;
+
+	rc = down_interruptible(&efx_mtd->access_lock);
+	if (rc)
+		goto out;
+	spi = efx_mtd->spi;
+
+	rc = efx_spi_unlock(efx_mtd);
+	if (rc)
+		goto out_up;
+
+	while (pos < len) {
+		rc = efx_spi_write_enable(efx_mtd);
+		if (rc)
+			break;
+
+		block_len = min((unsigned int)len - pos,
+				efx_spi_write_limit(spi, start + pos));
+		command = efx_spi_munge_command(spi, SPI_WRITE, start + pos);
+		rc = spi->write(spi, efx_mtd->efx, command, start + pos,
+				buffer + pos, block_len);
+		if (rc)
+			break;
+		pos += block_len;
+
+		rc = efx_spi_fast_wait(efx_mtd);
+		if (rc)
+			break;
+
+		/* Avoid locking up the system */
+		cond_resched();
+		if (signal_pending(current)) {
+			rc = -EINTR;
+			break;
+		}
+	}
+
+out_up:
+	up(&efx_mtd->access_lock);
+	if (rc == 0)
+		rc = efx_mtd_verify(mtd, start, len, buffer, 0);
+out:
+	*retlen = pos;
+	return rc;
+}
+
+static void efx_mtd_sync(struct mtd_info *mtd)
+{
+	struct efx_mtd *efx_mtd = mtd->priv;
+	int rc;
+
+	down(&efx_mtd->access_lock);
+	rc = efx_spi_slow_wait(efx_mtd, 1);
+	if (rc)
+		EFX_ERR(efx_mtd->efx, "%s sync failed (%d)\n",
+			efx_mtd->name, rc);
+	up(&efx_mtd->access_lock);
+}
+
+/* Driverlink interface */
+
+static void efx_mtd_reset_suspend(struct efx_dl_device *efx_dev)
+{
+	struct efx_mtd *efx_mtd = efx_dev->priv;
+
+	if (!efx_mtd)
+		return;
+
+	/* Acquire lock to ensure that any in-progress operations have
+	 * completed, and no new ones can start.
+	 */
+	down(&efx_mtd->access_lock);
+}
+
+static void efx_mtd_reset_resume(struct efx_dl_device *efx_dev, int ok)
+{
+	struct efx_mtd *efx_mtd = efx_dev->priv;
+
+	if (!efx_mtd)
+		return;
+
+	/* If device reset failed already, or SPI device doesn't
+	 * become ready, disable device.
+	 */
+	if (!ok || efx_spi_slow_wait(efx_mtd, 1) != 0) {
+		efx_mtd->spi = &efx_spi_dummy_device;
+		EFX_ERR(efx_mtd->efx, "%s disabled after failed reset\n",
+			efx_mtd->name);
+	}
+
+	up(&efx_mtd->access_lock);
+}
+
+static void efx_mtd_remove(struct efx_dl_device *efx_dev)
+{
+	struct efx_mtd *efx_mtd = efx_dev->priv;
+
+	del_mtd_partitions(&efx_mtd->mtd);
+	kfree(efx_mtd);
+	efx_dev->priv = NULL;
+}
+
+static __devinit int efx_mtd_register(struct efx_mtd *efx_mtd,
+				      struct efx_dl_device *efx_dev,
+				      struct efx_nic *efx,
+				      struct efx_spi_device *spi,
+				      const char *type_name,
+				      const char **part_type_name,
+				      unsigned int num_parts)
+{
+	int i;
+
+	efx_dev->priv = efx_mtd;
+
+	efx_mtd->efx_dev = efx_dev;
+	efx_mtd->efx = efx;
+	efx_mtd->spi = spi;
+	sema_init(&efx_mtd->access_lock, 1);
+
+	efx_mtd->mtd.size = spi->size;
+	efx_mtd->mtd.erasesize = spi->erase_size;
+#if defined(EFX_USE_MTD_WRITESIZE)
+	efx_mtd->mtd.writesize = 1;
+#endif
+	if (snprintf(efx_mtd->name, sizeof(efx_mtd->name),
+		     "%s %s", efx->name, type_name) >=
+	    sizeof(efx_mtd->name))
+		return -ENAMETOOLONG;
+
+	efx_mtd->mtd.priv = efx_mtd;
+	efx_mtd->mtd.name = efx_mtd->name;
+	efx_mtd->mtd.erase = efx_mtd_erase;
+	efx_mtd->mtd.read = efx_mtd_read;
+	efx_mtd->mtd.write = efx_mtd_write;
+	efx_mtd->mtd.sync = efx_mtd_sync;
+
+	for (i = 0; i < num_parts; i++) {
+		efx_mtd->part[i].name = efx_mtd->part_name[i];
+		if (snprintf(efx_mtd->part_name[i],
+			     sizeof(efx_mtd->part_name[i]),
+			     "%s %s", efx->name, part_type_name[i]) >=
+		    sizeof(efx_mtd->part_name[i]))
+			return -ENAMETOOLONG;
+
+		if (efx_allow_nvconfig_writes)
+			efx_mtd->part[i].mask_flags &= ~MTD_WRITEABLE;
+	}
+
+	return add_mtd_partitions(&efx_mtd->mtd, efx_mtd->part, num_parts);
+}
+
+static int __devinit
+efx_flash_probe(struct efx_dl_device *efx_dev,
+		const struct net_device *net_dev,
+		const struct efx_dl_device_info *dev_info,
+		const char *silicon_rev)
+{
+	struct efx_nic *efx = efx_dl_get_nic(efx_dev);
+	struct efx_mtd *efx_mtd;
+	const char *part_type_name[2];
+	unsigned int num_parts;
+	int rc;
+
+	if (!efx->spi_flash)
+		return -ENODEV;
+
+	efx_mtd = kzalloc(sizeof(*efx_mtd), GFP_KERNEL);
+	if (!efx_mtd)
+		return -ENOMEM;
+
+	efx_mtd->mtd.type = MTD_NORFLASH;
+	efx_mtd->mtd.flags = MTD_CAP_NORFLASH;
+
+	part_type_name[0] = "sfc_flash_config";
+	efx_mtd->part[0].offset = 0;
+	efx_mtd->part[0].size = min(efx->spi_flash->size,
+				    EFX_FLASH_BOOTROM_OFFSET);
+	efx_mtd->part[0].mask_flags = MTD_WRITEABLE;
+
+	if (efx->spi_flash->size <= EFX_FLASH_BOOTROM_OFFSET) {
+		num_parts = 1;
+	} else {
+		part_type_name[1] = "sfc_flash_bootrom";
+		efx_mtd->part[1].offset = EFX_FLASH_BOOTROM_OFFSET;
+		efx_mtd->part[1].size = (efx->spi_flash->size
+					 - EFX_FLASH_BOOTROM_OFFSET);
+		num_parts = 2;
+	}
+
+	rc = efx_mtd_register(efx_mtd, efx_dev, efx, efx->spi_flash,
+			      "sfc_flash", part_type_name, num_parts);
+	if (rc)
+		kfree(efx_mtd);
+	return rc;
+}
+
+static struct efx_dl_driver efx_flash_driver = {
+	.name		= "sfc_flash",
+	.probe		= efx_flash_probe,
+	.remove		= efx_mtd_remove,
+	.reset_suspend	= efx_mtd_reset_suspend,
+	.reset_resume	= efx_mtd_reset_resume,
+};
+
+static int __devinit
+efx_eeprom_probe(struct efx_dl_device *efx_dev,
+		 const struct net_device *net_dev,
+		 const struct efx_dl_device_info *dev_info,
+		 const char *silicon_rev)
+{
+	struct efx_nic *efx = efx_dl_get_nic(efx_dev);
+	struct efx_mtd *efx_mtd;
+	const char *type_name;
+	const char *part_type_name[1];
+	int rc;
+
+	if (!efx->spi_eeprom)
+		return -ENODEV;
+
+	efx_mtd = kzalloc(sizeof(*efx_mtd), GFP_KERNEL);
+	if (!efx_mtd)
+		return -ENOMEM;
+
+	efx_mtd->mtd.type = MTD_RAM;
+	efx_mtd->mtd.flags = MTD_CAP_RAM;
+
+	efx_mtd->part[0].offset = 0;
+	efx_mtd->part[0].size = efx->spi_eeprom->size;
+	efx_mtd->part[0].mask_flags = MTD_WRITEABLE;
+
+	if (efx->spi_eeprom->size <= 0x200) {
+		type_name = "sfc_small_eeprom";
+		part_type_name[0] = "sfc_small_config";
+	} else {
+		type_name = "sfc_large_eeprom";
+		part_type_name[0] = "sfc_large_config";
+	}
+
+	rc = efx_mtd_register(efx_mtd, efx_dev, efx, efx->spi_eeprom,
+			      type_name, part_type_name, 1);
+	if (rc)
+		kfree(efx_mtd);
+	return rc;
+}
+
+static struct efx_dl_driver efx_eeprom_driver = {
+	.name		= "sfc_eeprom",
+	.probe		= efx_eeprom_probe,
+	.remove		= efx_mtd_remove,
+	.reset_suspend	= efx_mtd_reset_suspend,
+	.reset_resume	= efx_mtd_reset_resume,
+};
+
+/* Kernel module interface */
+
+static int __init efx_mtd_init_module(void)
+{
+	int rc;
+
+	rc = efx_dl_register_driver(&efx_flash_driver);
+	if (rc)
+		return rc;
+	rc = efx_dl_register_driver(&efx_eeprom_driver);
+	if (rc) {
+		efx_dl_unregister_driver(&efx_flash_driver);
+		return rc;
+	}
+
+	return 0;
+}
+
+static void __exit efx_mtd_exit_module(void)
+{
+	efx_dl_unregister_driver(&efx_eeprom_driver);
+	efx_dl_unregister_driver(&efx_flash_driver);
+}
+
+module_init(efx_mtd_init_module);
+module_exit(efx_mtd_exit_module);
+
+MODULE_AUTHOR("Michael Brown <mbrown@fensystems.co.uk> and "
+	      "Solarflare Communications");
+MODULE_DESCRIPTION("SFC MTD driver");
+MODULE_LICENSE("GPL");
diff -rpuN linux-2.6.18.8/drivers/net/sfc/net_driver.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/net_driver.h
--- linux-2.6.18.8/drivers/net/sfc/net_driver.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/net_driver.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1096 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/* Common definitions for all Efx net driver code */
+
+#ifndef EFX_NET_DRIVER_H
+#define EFX_NET_DRIVER_H
+
+#include <linux/version.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/ethtool.h>
+#include <linux/if_vlan.h>
+#include <linux/timer.h>
+#include <linux/mii.h>
+#include <linux/list.h>
+#include <linux/pci.h>
+#include <linux/device.h>
+
+/* Must come before other headers */
+#include "kernel_compat.h"
+
+#include "enum.h"
+#include "bitfield.h"
+#include "driverlink_api.h"
+#include "driverlink.h"
+#include "i2c-direct.h"
+
+	#ifndef EFX_USE_DEBUGFS
+		/* Sick, but we have no other use for dentry */
+		#define dentry proc_dir_entry
+	#endif
+
+#define EFX_MAX_LRO_DESCRIPTORS 8
+#define EFX_MAX_LRO_AGGR MAX_SKB_FRAGS
+
+/**************************************************************************
+ *
+ * Build definitions
+ *
+ **************************************************************************/
+#ifndef EFX_DRIVER_NAME
+#define EFX_DRIVER_NAME	"sfc"
+#endif
+#define EFX_DRIVER_VERSION	"2.2.0101"
+
+#ifdef EFX_ENABLE_DEBUG
+#define EFX_BUG_ON_PARANOID(x) BUG_ON(x)
+#define EFX_WARN_ON_PARANOID(x) WARN_ON(x)
+#else
+#define EFX_BUG_ON_PARANOID(x) do {} while (0)
+#define EFX_WARN_ON_PARANOID(x) do {} while (0)
+#endif
+
+/* Include net device name in log messages if it has been registered.
+ * Use efx->name not efx->net_dev->name so that races with (un)registration
+ * are harmless.
+ */
+#define NET_DEV_NAME(efx) ((efx)->net_dev_registered ? (efx)->name : "")
+
+/* Un-rate-limited logging */
+#define EFX_ERR(efx, fmt, args...) \
+dev_err(&((efx)->pci_dev->dev), "ERR: %s " fmt, NET_DEV_NAME(efx), ##args)
+
+#define EFX_INFO(efx, fmt, args...) \
+dev_info(&((efx)->pci_dev->dev), "INFO: %s " fmt, NET_DEV_NAME(efx), ##args)
+
+#ifdef EFX_ENABLE_DEBUG
+#define EFX_LOG(efx, fmt, args...) \
+dev_info(&((efx)->pci_dev->dev), "DBG: %s " fmt, NET_DEV_NAME(efx), ##args)
+#else
+#define EFX_LOG(efx, fmt, args...) \
+dev_dbg(&((efx)->pci_dev->dev), "DBG: %s " fmt, NET_DEV_NAME(efx), ##args)
+#endif
+
+#define EFX_TRACE(efx, fmt, args...) do {} while (0)
+
+#define EFX_REGDUMP(efx, fmt, args...) do {} while (0)
+
+/* Rate-limited logging */
+#define EFX_ERR_RL(efx, fmt, args...) \
+do {if (net_ratelimit()) EFX_ERR(efx, fmt, ##args); } while (0)
+
+#define EFX_INFO_RL(efx, fmt, args...) \
+do {if (net_ratelimit()) EFX_INFO(efx, fmt, ##args); } while (0)
+
+#define EFX_LOG_RL(efx, fmt, args...) \
+do {if (net_ratelimit()) EFX_LOG(efx, fmt, ##args); } while (0)
+
+/* Kernel headers may redefine inline anyway */
+#ifndef inline
+#define inline inline __attribute__ ((always_inline))
+#endif
+
+/**************************************************************************
+ *
+ * Efx data structures
+ *
+ **************************************************************************/
+
+#define EFX_MAX_CHANNELS 32
+#define EFX_MAX_TX_QUEUES 1
+#define EFX_MAX_RX_QUEUES EFX_MAX_CHANNELS
+
+/**
+ * struct efx_special_buffer - An Efx special buffer
+ * @addr: CPU base address of the buffer
+ * @dma_addr: DMA base address of the buffer
+ * @len: Buffer length, in bytes
+ * @index: Buffer index within controller;s buffer table
+ * @entries: Number of buffer table entries
+ *
+ * Special buffers are used for the event queues and the TX and RX
+ * descriptor queues for each channel.  They are *not* used for the
+ * actual transmit and receive buffers.
+ *
+ * Note that for Falcon, TX and RX descriptor queues live in host memory.
+ * Allocation and freeing procedures must take this into account.
+ */
+struct efx_special_buffer {
+	void *addr;
+	dma_addr_t dma_addr;
+	unsigned int len;
+	int index;
+	int entries;
+};
+
+/**
+ * struct efx_tx_buffer - An Efx TX buffer
+ * @skb: The associated socket buffer.
+ *	Set only on the final fragment of a packet; %NULL for all other
+ *	fragments.  When this fragment completes, then we can free this
+ *	skb.
+ * @dma_addr: DMA address of the fragment.
+ * @len: Length of this fragment.
+ *	This field is zero when the queue slot is empty.
+ * @continuation: True if this fragment is not the end of a packet.
+ * @unmap_single: True if pci_unmap_single should be used.
+ * @unmap_addr: DMA address to unmap
+ * @unmap_len: Length of this fragment to unmap
+ */
+struct efx_tx_buffer {
+	const struct sk_buff *skb;
+	dma_addr_t dma_addr;
+	unsigned short len;
+	unsigned char continuation;
+	unsigned char unmap_single;
+	dma_addr_t unmap_addr;
+	unsigned short unmap_len;
+};
+
+/**
+ * struct efx_tx_queue - An Efx TX queue
+ *
+ * This is a ring buffer of TX fragments.
+ * Since the TX completion path always executes on the same
+ * CPU and the xmit path can operate on different CPUs,
+ * performance is increased by ensuring that the completion
+ * path and the xmit path operate on different cache lines.
+ * This is particularly important if the xmit path is always
+ * executing on one CPU which is different from the completion
+ * path.  There is also a cache line for members which are
+ * read but not written on the fast path.
+ *
+ * @efx: The associated Efx NIC
+ * @queue: DMA queue number
+ * @used: Queue is used by net driver
+ * @channel: The associated channel
+ * @buffer: The software buffer ring
+ * @txd: The hardware descriptor ring
+ * @debug_dir: debugfs directory
+ * @read_count: Current read pointer.
+ *	This is the number of buffers that have been removed from both rings.
+ * @stopped: Stopped flag.
+ *	Set if this TX queue is currently stopping its port.
+ * @insert_count: Current insert pointer
+ *	This is the number of buffers that have been added to the
+ *	software ring.
+ * @write_count: Current write pointer
+ *	This is the number of buffers that have been added to the
+ *	hardware ring.
+ * @old_read_count: The value of read_count when last checked.
+ *	This is here for performance reasons.  The xmit path will
+ *	only get the up-to-date value of read_count if this
+ *	variable indicates that the queue is full.  This is to
+ *	avoid cache-line ping-pong between the xmit path and the
+ *	completion path.
+ */
+struct efx_tx_queue {
+	/* Members which don't change on the fast path */
+	struct efx_nic *efx ____cacheline_aligned_in_smp;
+	int queue;
+	int used;
+	struct efx_channel *channel;
+	struct efx_nic *nic;
+	struct efx_tx_buffer *buffer;
+	struct efx_special_buffer txd;
+#ifdef CONFIG_SFC_DEBUGFS
+	struct dentry *debug_dir;
+#endif
+
+	/* Members used mainly on the completion path */
+	unsigned int read_count ____cacheline_aligned_in_smp;
+	int stopped;
+
+	/* Members used only on the xmit path */
+	unsigned int insert_count ____cacheline_aligned_in_smp;
+	unsigned int write_count;
+	unsigned int old_read_count;
+};
+
+/**
+ * struct efx_rx_buffer - An Efx RX data buffer
+ * @dma_addr: DMA base address of the buffer
+ * @skb: The associated socket buffer, if any.
+ *	If both this and page are %NULL, the buffer slot is currently free.
+ * @page: The associated page buffer, if any.
+ *	If both this and skb are %NULL, the buffer slot is currently free.
+ * @data: Pointer to ethernet header
+ * @len: Buffer length, in bytes.
+ * @unmap_addr: DMA address to unmap
+ */
+struct efx_rx_buffer {
+	dma_addr_t dma_addr;
+	struct sk_buff *skb;
+	struct page *page;
+	char *data;
+	unsigned int len;
+	dma_addr_t unmap_addr;
+};
+
+/**
+ * struct efx_rx_queue - An Efx RX queue
+ * @efx: The associated Efx NIC
+ * @queue: DMA queue number
+ * @used: Queue is used by net driver
+ * @channel: The associated channel
+ * @buffer: The software buffer ring
+ * @rxd: The hardware descriptor ring
+ * @added_count: Number of buffers added to the receive queue.
+ * @notified_count: Number of buffers given to NIC (<= @added_count).
+ * @removed_count: Number of buffers removed from the receive queue.
+ * @add_lock: Receive queue descriptor add spin lock.
+ *	This lock must be held in order to add buffers to the RX
+ *	descriptor ring (rxd and buffer) and to update added_count (but
+ *	not removed_count).
+ * @max_fill: RX descriptor maximum fill level (<= ring size)
+ * @fast_fill_trigger: RX descriptor fill level that will trigger a fast fill
+ *	(<= @max_fill)
+ * @fast_fill_limit: The level to which a fast fill will fill
+ *	(@fast_fill_trigger <= @fast_fill_limit <= @max_fill)
+ * @min_fill: RX descriptor minimum non-zero fill level.
+ *	This records the minimum fill level observed when a ring
+ *	refill was triggered.
+ * @min_overfill: RX descriptor minimum overflow fill level.
+ *	This records the minimum fill level at which RX queue
+ *	overflow was observed.  It should never be set.
+ * @alloc_page_count: RX allocation strategy counter.
+ * @alloc_skb_count: RX allocation strategy counter.
+ * @work: Descriptor push work thread
+ * @buf_page: Page for next RX buffer.
+ *	We can use a single page for multiple RX buffers. This tracks
+ *	the remaining space in the allocation.
+ * @buf_dma_addr: Page's DMA address.
+ * @buf_data: Page's host address.
+ * @debug_dir: debugfs directory
+ */
+struct efx_rx_queue {
+	struct efx_nic *efx;
+	int queue;
+	int used;
+	struct efx_channel *channel;
+	struct efx_rx_buffer *buffer;
+	struct efx_special_buffer rxd;
+
+	int added_count;
+	int notified_count;
+	int removed_count;
+	spinlock_t add_lock;
+	unsigned int max_fill;
+	unsigned int fast_fill_trigger;
+	unsigned int fast_fill_limit;
+	unsigned int min_fill;
+	unsigned int min_overfill;
+	unsigned int alloc_page_count;
+	unsigned int alloc_skb_count;
+	struct delayed_work work;
+	unsigned int slow_fill_count;
+
+	struct page *buf_page;
+	dma_addr_t buf_dma_addr;
+	char *buf_data;
+
+#ifdef CONFIG_SFC_DEBUGFS
+	struct dentry *debug_dir;
+#endif
+};
+
+/**
+ * struct efx_buffer - An Efx general-purpose buffer
+ * @addr: host base address of the buffer
+ * @dma_addr: DMA base address of the buffer
+ * @len: Buffer length, in bytes
+ *
+ * Falcon uses these buffers for its interrupt status registers and
+ * MAC stats dumps.
+ */
+struct efx_buffer {
+	void *addr;
+	dma_addr_t dma_addr;
+	unsigned int len;
+};
+
+
+/* Flags for channel->used_flags */
+#define EFX_USED_BY_RX 1
+#define EFX_USED_BY_TX 2
+#define EFX_USED_BY_RX_TX (EFX_USED_BY_RX | EFX_USED_BY_TX)
+
+enum efx_rx_alloc_method {
+	RX_ALLOC_METHOD_AUTO = 0,
+	RX_ALLOC_METHOD_SKB = 1,
+	RX_ALLOC_METHOD_PAGE = 2,
+};
+
+/**
+ * struct efx_channel - An Efx channel
+ *
+ * A channel comprises an event queue, at least one TX queue, at least
+ * one RX queue, and an associated tasklet for processing the event
+ * queue.
+ *
+ * @efx: Associated Efx NIC
+ * @evqnum: Event queue number
+ * @channel: Channel instance number
+ * @used_flags: Channel is used by net driver
+ * @enabled: Channel enabled indicator
+ * @irq: IRQ number (MSI and MSI-X only)
+ * @has_interrupt: Channel has an interrupt
+ * @irq_moderation: IRQ moderation value (in us)
+ * @napi_dev: Net device used with NAPI
+ * @napi_str: NAPI control structure
+ * @reset_work: Scheduled reset work thread
+ * @work_pending: Is work pending via NAPI?
+ * @eventq: Event queue buffer
+ * @eventq_read_ptr: Event queue read pointer
+ * @last_eventq_read_ptr: Last event queue read pointer value.
+ * @eventq_magic: Event queue magic value for driver-generated test events
+ * @debug_dir: debugfs directory
+ * @rx_alloc_level: Watermark based heuristic counter for pushing descriptors
+ *	and diagnostic counters
+ * @rx_alloc_push_pages: RX allocation method currently in use for pushing
+ *	descriptors
+ * @rx_alloc_pop_pages: RX allocation method currently in use for popping
+ *	descriptors
+ * @n_rx_tobe_disc: Count of RX_TOBE_DISC errors
+ * @n_rx_ip_frag_err: Count of RX IP fragment errors
+ * @n_rx_ip_hdr_chksum_err: Count of RX IP header checksum errors
+ * @n_rx_tcp_udp_chksum_err: Count of RX TCP and UDP checksum errors
+ * @n_rx_frm_trunc: Count of RX_FRM_TRUNC errors
+ * @n_rx_overlength: Count of RX_OVERLENGTH errors
+ * @n_skbuff_leaks: Count of skbuffs leaked due to RX overrun
+ */
+struct efx_channel {
+	struct efx_nic *efx;
+	int evqnum;
+	int channel;
+	int used_flags;
+	int enabled;
+	int irq;
+	unsigned int has_interrupt;
+	unsigned int irq_moderation;
+	struct net_device *napi_dev;
+#if !defined(EFX_HAVE_OLD_NAPI)
+	struct napi_struct napi_str;
+#endif
+	struct work_struct reset_work;
+	int work_pending;
+	struct efx_special_buffer eventq;
+	unsigned int eventq_read_ptr;
+	unsigned int last_eventq_read_ptr;
+	unsigned int eventq_magic;
+
+#ifdef CONFIG_SFC_DEBUGFS
+	struct dentry *debug_dir;
+#endif
+
+	int rx_alloc_level;
+	int rx_alloc_push_pages;
+	int rx_alloc_pop_pages;
+
+	unsigned n_rx_tobe_disc;
+	unsigned n_rx_ip_frag_err;
+	unsigned n_rx_ip_hdr_chksum_err;
+	unsigned n_rx_tcp_udp_chksum_err;
+	unsigned n_rx_frm_trunc;
+	unsigned n_rx_overlength;
+	unsigned n_skbuff_leaks;
+
+	/* Used to pipeline received packets in order to optimise memory
+	 * access with prefetches.
+	 */
+	struct efx_rx_buffer *rx_pkt;
+	int rx_pkt_csummed;
+
+};
+
+/**
+ * struct efx_blinker - S/W LED blinking context
+ * @led_num: LED ID (board-specific meaning)
+ * @state: Current state - on or off
+ * @resubmit: Timer resubmission flag
+ * @timer: Control timer for blinking
+ */
+struct efx_blinker {
+	int led_num;
+	int state;
+	int resubmit;
+	struct timer_list timer;
+};
+
+
+/**
+ * struct efx_board - board information
+ * @type: Board model type
+ * @major: Major rev. ('A', 'B' ...)
+ * @minor: Minor rev. (0, 1, ...)
+ * @init: Initialisation function
+ * @init_leds: Sets up board LEDs
+ * @set_fault_led: Turns the fault LED on or off
+ * @blink: Starts/stops blinking
+ * @monitor: Board-specific health check function
+ * @fini: Cleanup function
+ * @interpret_sensor: Function to interpret LM87 sensor meanings.
+ *	Returns %FALSE if no board-specific meaning.
+ * @mwatts: Power requirements (mW)
+ * @blinker: used to blink LEDs in software
+ * @lm87_addr: I2C address of LM87 monitoring chip (0 if absent)
+ */
+struct efx_board {
+	int type;
+	int major;
+	int minor;
+	int (*init) (struct efx_nic *nic);
+	/* As the LEDs are typically attached to the PHY, LEDs
+	 * have a separate init callback that happens later than
+	 * board init. */
+	int (*init_leds)(struct efx_nic *efx);
+	int (*monitor) (struct efx_nic *nic);
+	void (*set_fault_led) (struct efx_nic *efx, int state);
+	void (*blink) (struct efx_nic *efx, int start);
+	void (*fini) (struct efx_nic *nic);
+	int (*interpret_sensor) (struct efx_nic *nic, int num, unsigned val);
+	unsigned mwatts;
+	struct efx_blinker blinker;
+	unsigned int lm87_addr;
+};
+
+#define STRING_TABLE_LOOKUP(val, member) \
+	((val) < member ## _max) ? member ## _names[val] : "(invalid)"
+
+extern const char *efx_loopback_mode_names[];
+extern const unsigned int efx_loopback_mode_max;
+#define LOOPBACK_MODE(efx) \
+	STRING_TABLE_LOOKUP((efx)->loopback_mode, efx_loopback_mode)
+
+extern const char *efx_interrupt_mode_names[];
+extern const unsigned int efx_interrupt_mode_max;
+#define INT_MODE(efx) \
+	STRING_TABLE_LOOKUP(efx->interrupt_mode, efx_interrupt_mode)
+
+extern const char *efx_reset_type_names[];
+extern const unsigned int efx_reset_type_max;
+#define RESET_TYPE(type) \
+	STRING_TABLE_LOOKUP(type, efx_reset_type)
+
+extern const char *efx_phy_type_names[];
+extern const unsigned int efx_phy_type_max;
+#define PHY_TYPE(efx) \
+	STRING_TABLE_LOOKUP((efx)->phy_type, efx_phy_type)
+
+extern const char *efx_nic_state_names[];
+extern const unsigned int efx_nic_state_max;
+#define STATE_NAME(efx)	\
+	STRING_TABLE_LOOKUP((efx)->state, efx_nic_state)
+
+enum efx_int_mode {
+	/* Be careful if altering to correct macro below */
+	EFX_INT_MODE_MSIX = 0,
+	EFX_INT_MODE_MSI = 1,
+	EFX_INT_MODE_LEGACY = 2,
+	EFX_INT_MODE_MAX	/* Insert any new items before this */
+};
+#define EFX_INT_MODE_USE_MSI(x) (((x)->interrupt_mode) <= EFX_INT_MODE_MSI)
+
+enum phy_type {
+	PHY_TYPE_NONE = 0,
+	PHY_TYPE_CX4_RTMR = 1,
+	PHY_TYPE_1G_ALASKA = 2,
+	PHY_TYPE_10XPRESS = 3,
+	PHY_TYPE_XFP = 4,
+	PHY_TYPE_PM8358 = 6,
+	PHY_TYPE_MAX	/* Insert any new items before this */
+};
+
+#define PHY_ADDR_INVALID 0xff
+
+#define EFX_IS10G(efx) ((efx)->is_10g)
+#define EFX_ISCLAUSE45(efx) ((efx)->phy_type != PHY_TYPE_1G_ALASKA)
+
+enum nic_state {
+	STATE_INIT = 0,      /* suspend_lock always held */
+	STATE_RUNNING = 1,
+	STATE_FINI = 2,
+	STATE_RESETTING = 3, /* suspend_lock always held */
+	STATE_DISABLED = 4,
+	STATE_MAX,
+};
+
+/*
+ * Alignment of page-allocated RX buffers
+ *
+ * Controls the number of bytes inserted at the start of an RX buffer.
+ * This is the equivalent of NET_IP_ALIGN [which controls the alignment
+ * of the skb->head for hardware DMA].
+ */
+#ifdef __ia64__
+#define EFX_PAGE_IP_ALIGN 2
+#else
+#define EFX_PAGE_IP_ALIGN 0
+#endif
+
+/*
+ * Alignment of the skb->head which wraps a page-allocated RX buffer
+ *
+ * The skb allocated to wrap an rx_buffer can have this alignment. Since
+ * the data is memcpy'd from the rx_buf, it does not need to be equal to
+ * EFX_PAGE_IP_ALIGN.
+ */
+#define EFX_PAGE_SKB_ALIGN 2
+
+/* Forward declaration */
+struct efx_nic;
+
+/* Pseudo bit-mask flow control field */
+enum efx_fc_type {
+	EFX_FC_RX = 1,
+	EFX_FC_TX = 2,
+	EFX_FC_AUTO = 4,
+};
+
+/**
+ * struct efx_mac_operations - Efx MAC operations table
+ * @mac_writel: Write dword to MAC register
+ * @mac_readl: Read dword from a MAC register
+ * @init: Initialise MAC and PHY
+ * @reconfigure: Reconfigure MAC and PHY (e.g. for new link parameters)
+ * @update_stats: Update statistics
+ * @fini: Shut down MAC and PHY
+ * @check_hw: Check hardware
+ * @fake_phy_event: Simulate a PHY event on a port
+ * @get_settings: Get ethtool settings
+ * @set_settings: Set ethtool settings
+ * @set_pause: Set pause parameters
+ */
+struct efx_mac_operations {
+	void (*mac_writel) (struct efx_nic *efx,
+			    efx_dword_t *value, unsigned int mac_reg);
+	void (*mac_readl) (struct efx_nic *efx,
+			   efx_dword_t *value, unsigned int mac_reg);
+	int (*init) (struct efx_nic *efx);
+	void (*reconfigure) (struct efx_nic *efx);
+	void (*update_stats) (struct efx_nic *efx);
+	void (*fini) (struct efx_nic *efx);
+	int (*check_hw) (struct efx_nic *efx);
+	void (*fake_phy_event) (struct efx_nic *efx);
+
+	int (*get_settings) (struct efx_nic *efx,
+			     struct ethtool_cmd *ecmd);
+	int (*set_settings) (struct efx_nic *efx,
+			     struct ethtool_cmd *ecmd);
+	int (*set_pause) (struct efx_nic *efx,
+			  enum efx_fc_type pause_params);
+};
+
+/**
+ * struct efx_phy_operations - Efx PHY operations table
+ * @init: Initialise PHY
+ * @fini: Shut down PHY
+ * @reconfigure: Reconfigure PHY (e.g. for new link parameters)
+ * @clear_interrupt: Clear down interrupt
+ * @blink: Blink LEDs
+ * @check_hw: Check hardware
+ * @reset_xaui: Reset XAUI side of PHY for (software sequenced reset)
+ * @mmds: MMD presence mask
+ * @loopbacks: Supported loopback modes mask
+ * @startup_loopback: Loopback mode for start-of-day self-test
+ */
+struct efx_phy_operations {
+	int (*init) (struct efx_nic *efx);
+	void (*fini) (struct efx_nic *efx);
+	void (*reconfigure) (struct efx_nic *efx);
+	void (*clear_interrupt) (struct efx_nic *efx);
+	int (*check_hw) (struct efx_nic *efx);
+	void (*reset_xaui) (struct efx_nic *efx);
+	int mmds;
+	unsigned loopbacks;
+	unsigned startup_loopback;
+};
+
+/*
+ * Efx extended statistics
+ *
+ * Not all statistics are provided by all supported MACs.  The purpose
+ * is this structure is to contain the raw statistics provided by each
+ * MAC.
+ */
+struct efx_mac_stats {
+	u64 tx_bytes;
+	u64 tx_good_bytes;
+	u64 tx_bad_bytes;
+	unsigned long tx_packets;
+	unsigned long tx_bad;
+	unsigned long tx_pause;
+	unsigned long tx_control;
+	unsigned long tx_unicast;
+	unsigned long tx_multicast;
+	unsigned long tx_broadcast;
+	unsigned long tx_lt64;
+	unsigned long tx_64;
+	unsigned long tx_65_to_127;
+	unsigned long tx_128_to_255;
+	unsigned long tx_256_to_511;
+	unsigned long tx_512_to_1023;
+	unsigned long tx_1024_to_15xx;
+	unsigned long tx_15xx_to_jumbo;
+	unsigned long tx_gtjumbo;
+	unsigned long tx_collision;
+	unsigned long tx_single_collision;
+	unsigned long tx_multiple_collision;
+	unsigned long tx_excessive_collision;
+	unsigned long tx_deferred;
+	unsigned long tx_late_collision;
+	unsigned long tx_excessive_deferred;
+	unsigned long tx_non_tcpudp;
+	unsigned long tx_mac_src_error;
+	unsigned long tx_ip_src_error;
+	u64 rx_bytes;
+	u64 rx_good_bytes;
+	u64 rx_bad_bytes;
+	unsigned long rx_packets;
+	unsigned long rx_good;
+	unsigned long rx_bad;
+	unsigned long rx_pause;
+	unsigned long rx_control;
+	unsigned long rx_unicast;
+	unsigned long rx_multicast;
+	unsigned long rx_broadcast;
+	unsigned long rx_lt64;
+	unsigned long rx_64;
+	unsigned long rx_65_to_127;
+	unsigned long rx_128_to_255;
+	unsigned long rx_256_to_511;
+	unsigned long rx_512_to_1023;
+	unsigned long rx_1024_to_15xx;
+	unsigned long rx_15xx_to_jumbo;
+	unsigned long rx_gtjumbo;
+	unsigned long rx_bad_lt64;
+	unsigned long rx_bad_64_to_15xx;
+	unsigned long rx_bad_15xx_to_jumbo;
+	unsigned long rx_bad_gtjumbo;
+	unsigned long rx_overflow;
+	unsigned long rx_missed;
+	unsigned long rx_false_carrier;
+	unsigned long rx_symbol_error;
+	unsigned long rx_align_error;
+	unsigned long rx_length_error;
+	unsigned long rx_internal_error;
+	unsigned long rx_good_lt64;
+};
+
+/* Number of bits used in a multicast filter hash address */
+#define EFX_MCAST_HASH_BITS 8
+
+/* Number of (single-bit) entries in a multicast filter hash */
+#define EFX_MCAST_HASH_ENTRIES (1 << EFX_MCAST_HASH_BITS)
+
+/* An Efx multicast filter hash */
+union efx_multicast_hash {
+	u8 byte[EFX_MCAST_HASH_ENTRIES / sizeof(u8)];
+	efx_oword_t oword[EFX_MCAST_HASH_ENTRIES / sizeof(efx_oword_t)];
+};
+
+/* Efx Error condition statistics */
+struct efx_nic_errors {
+	atomic_t missing_event;
+	atomic_t rx_reset;
+	atomic_t rx_desc_fetch;
+	atomic_t tx_desc_fetch;
+	atomic_t spurious_tx;
+
+#ifdef CONFIG_SFC_DEBUGFS
+	struct dentry *debug_dir;
+#endif
+};
+
+/**
+ * struct efx_nic - an Efx NIC
+ * @name: Device name (net device name or bus id before net device registered)
+ * @pci_dev: The PCI device
+ * @pci_dev2: The secondary PCI device if present
+ * @type: Controller type attributes
+ * @dma_mask: DMA mask
+ * @legacy_irq: IRQ number
+ * @workqueue: Workqueue for resets, port reconfigures and the HW monitor
+ * @refill_workqueue: RX refill workqueue
+ * @reset_work: Scheduled reset workitem
+ * @monitor_work: Hardware monitor workitem
+ * @membase_phys: Memory BAR value as physical address
+ * @membase: Memory BAR value
+ * @biu_lock: BIU (bus interface unit) lock
+ * @interrupt_mode: Interrupt mode
+ * @is_asic: Is ASIC (else FPGA)
+ * @is_10g: Is set to 10G (else 1G)
+ * @external_sram_cfg: Size and number of banks of external SRAM
+ * @i2c: I2C interface
+ * @board_info: Board-level information
+ * @state: Device state flag. Can only be manipulated when both
+ *	suspend_lock and rtnl_lock are held.  Can be read when
+ *	either is held.
+ * @reset_pending: Pending reset method (normally RESET_TYPE_NONE)
+ * @suspend_lock: Device suspend lock.  This must not be acquired with
+ *	rtnl_lock held.
+ * @tx_queue: TX DMA queues
+ * @rx_queue: RX DMA queues
+ * @channel: Channels
+ * @rss_queues: Number of RSS queues
+ * @rx_buffer_len: RX buffer length
+ * @rx_buffer_order: Order (log2) of number of pages for each RX buffer
+ * @errors: Error condition stats
+ * @irq_status: Interrupt status buffer
+ * @last_irq_cpu: Last CPU to handle interrupt.
+ *	This register is written with the SMP processor ID whenever an
+ *	interrupt is handled.  It is used by falcon_test_interrupt()
+ *	to verify that an interrupt has occurred.
+ * @spi_flash: SPI flash device
+ *	This field will be %NULL if no flash device is present.
+ * @spi_eeprom: SPI EEPROM device
+ *	This field will be %NULL if no EEPROM device is present.
+ * @spi_lock: SPI bus lock
+ * @n_rx_nodesc_drop_cnt: RX no descriptor drop count
+ * @nic_data: Hardware dependant state
+ * @mac_lock: MAC access lock. Protects efx->port_enabled/net_dev_registered
+ *            and efx_reconfigure_port()
+ * @port_enabled: Port enabled indicator.
+ *	Serialises efx_stop_all and efx_start_all with kernel interfaces.
+ *	Safe to read under the rtnl_lock, mac_lock, or netif_tx_lock, but
+ *	all three must be held to modify it.
+ * @net_dev_registered: Port is registered with operating system.
+ * @port_initialized: Port initialized?
+ * @net_dev: Operating system network device. Consider holding the rtnl lock
+ * @rx_checksum_enabled: RX checksumming enabled
+ * @netif_stop_count: Port stop count
+ * @netif_stop_lock: Port stop lock
+ * @mac_stats: MAC statistics
+ * @stats: Net device statistics.
+ *	Hardware-specific code fills in @mac_stats, which provides a
+ *	detailed breakdown.  Generic code aggregates these statistics
+ *	into a standard &struct net_device_stats.
+ * @stats_buffer: DMA buffer for statistics
+ * @stats_lock: Statistics update lock
+ * @mac_op: MAC interface
+ * @mac_address: Permanent MAC address
+ * @phy_type: PHY type
+ * @phy_lock: PHY access lock
+ * @phy_op: PHY interface
+ * @phy_data: PHY private data (including PHY-specific stats)
+ * @mii: PHY interface
+ * @phy_powered: PHY power state
+ * @tx_disabled: PHY transmitter turned off
+ * @link_up: Link status
+ * @link_options: Link options (MII/GMII format)
+ * @n_link_state_changes: Number of times the link has changed state
+ * @promiscuous: Promiscuous flag. Protected by netif_tx_lock.
+ * @multicast_hash: Multicast hash table
+ * @flow_control: Flow control flags - separate RX/TX so can't use link_options
+ * @reconfigure_work: work item for dealing with PHY events
+ * @loopback_mode: Loopback status
+ * @loopback_modes: Supported loopback mode bitmask
+ * @loopback_selftest: Offline self-test private state
+ * @silicon_rev: Silicon revision description for driverlink
+ * @dl_info: Linked list of hardware parameters exposed through driverlink
+ * @dl_node: Driverlink port list
+ * @dl_device_list: Driverlink device list
+ * @dl_cb: Driverlink callbacks table
+ * @dl_cb_dev: Driverlink callback owner devices
+ * @debug_dir: NIC debugfs directory
+ * @debug_symlink: NIC debugfs sym-link (nic_eth\%d)
+ * @debug_port_dir: Port debugfs directory
+ * @debug_port_symlink: Port debugfs sym-link (if_eth\%d)
+ *
+ * The @priv field of the corresponding &struct net_device points to
+ * this.
+ */
+struct efx_nic {
+	char name[IFNAMSIZ];
+	struct pci_dev *pci_dev;
+	struct pci_dev *pci_dev2;
+#if !defined(EFX_USE_PCI_DEV_REVISION)
+	u8 revision;
+#endif
+	const struct efx_nic_type *type;
+	dma_addr_t dma_mask;
+	int legacy_irq;
+	struct workqueue_struct *workqueue;
+#if !defined(EFX_USE_CANCEL_DELAYED_WORK_SYNC)
+	/* Since we can't use cancel_delayed_work_sync efx_reset() has to
+	 * flush efx->workqueue to serialise against efx_reconfigure_port
+	 * and efx_monitor. So it can't also run on workqueue */
+	struct workqueue_struct *reset_workqueue;
+#endif
+	struct workqueue_struct *refill_workqueue;
+	struct work_struct reset_work;
+	struct delayed_work monitor_work;
+	unsigned long membase_phys;
+	void __iomem *membase;
+	spinlock_t biu_lock;
+	enum efx_int_mode interrupt_mode;
+	unsigned int is_asic:1;
+	unsigned int is_10g:1;
+	int external_sram_cfg;
+
+	struct efx_i2c_interface i2c;
+	struct efx_board board_info;
+
+	enum nic_state state;
+	enum reset_type reset_pending;
+
+	struct semaphore suspend_lock;
+
+	struct efx_tx_queue tx_queue[EFX_MAX_TX_QUEUES];
+	struct efx_rx_queue rx_queue[EFX_MAX_RX_QUEUES];
+	struct efx_channel channel[EFX_MAX_CHANNELS];
+
+	int rss_queues;
+	unsigned int rx_buffer_len;
+	unsigned int rx_buffer_order;
+
+	struct efx_nic_errors errors;
+
+	struct efx_buffer irq_status;
+	volatile signed int last_irq_cpu;
+
+	struct efx_spi_device *spi_flash;
+	struct efx_spi_device *spi_eeprom;
+	struct mutex spi_lock;
+
+	unsigned n_rx_nodesc_drop_cnt;
+
+	void *nic_data;
+
+	struct mutex mac_lock;
+	int port_enabled;
+	int net_dev_registered;
+
+	int port_initialized;
+	struct net_device *net_dev;
+	int rx_checksum_enabled;
+
+	atomic_t netif_stop_count;
+	spinlock_t netif_stop_lock;
+
+	struct efx_mac_stats mac_stats;
+	struct net_device_stats stats;
+	struct efx_buffer stats_buffer;
+	spinlock_t stats_lock;
+
+	struct efx_mac_operations *mac_op;
+	unsigned char mac_address[ETH_ALEN];
+
+	enum phy_type phy_type;
+	spinlock_t phy_lock;
+	struct efx_phy_operations *phy_op;
+	void *phy_data;
+	struct mii_if_info mii;
+	unsigned phy_powered;
+	unsigned tx_disabled;
+
+	int link_up;
+	unsigned int link_options;
+	unsigned int n_link_state_changes;
+
+	int promiscuous;
+	union efx_multicast_hash multicast_hash;
+	enum efx_fc_type flow_control;
+	struct work_struct reconfigure_work;
+
+	enum efx_loopback_mode loopback_mode;
+	unsigned int loopback_modes;
+	unsigned int startup_loopbacks;
+
+	void *loopback_selftest;
+
+	const char *silicon_rev;
+	struct efx_dl_device_info *dl_info;
+	struct list_head dl_node;
+	struct list_head dl_device_list;
+	struct efx_dl_callbacks dl_cb;
+	struct efx_dl_cb_devices dl_cb_dev;
+
+#ifdef CONFIG_SFC_DEBUGFS
+	struct dentry *debug_dir;
+	struct dentry *debug_symlink;
+	struct dentry *debug_port_dir;
+	struct dentry *debug_port_symlink;
+#endif
+};
+
+/**
+ * struct efx_nic_type - Efx device type definition
+ * @is_dual_func: Is dual-function (else single-function)
+ * @mem_bar: Memory BAR number
+ * @mem_map_size: Memory BAR mapped size
+ * @txd_ptr_tbl_base: TX descriptor ring base address
+ * @rxd_ptr_tbl_base: RX descriptor ring base address
+ * @buf_tbl_base: Buffer table base address
+ * @evq_ptr_tbl_base: Event queue pointer table base address
+ * @evq_rptr_tbl_base: Event queue read-pointer table base address
+ * @txd_ring_mask: TX descriptor ring size - 1 (must be a power of two - 1)
+ * @rxd_ring_mask: RX descriptor ring size - 1 (must be a power of two - 1)
+ * @evq_size: Event queue size (must be a power of two)
+ * @max_dma_mask: Maximum possible DMA mask
+ * @tx_dma_mask: TX DMA mask
+ * @bug5391_mask: Address mask for bug 5391 workaround
+ * @rx_xoff_thresh: RX FIFO XOFF watermark (bytes)
+ * @rx_xon_thresh: RX FIFO XON watermark (bytes)
+ * @rx_buffer_padding: Padding added to each RX buffer
+ * @max_interrupt_mode: Highest capability interrupt mode supported
+ *	from &enum efx_init_mode.
+ * @phys_addr_channels: Number of channels with physically addressed
+ *	descriptors
+ */
+struct efx_nic_type {
+	unsigned int is_dual_func;
+	unsigned int mem_bar;
+	unsigned int mem_map_size;
+	unsigned int txd_ptr_tbl_base;
+	unsigned int rxd_ptr_tbl_base;
+	unsigned int buf_tbl_base;
+	unsigned int evq_ptr_tbl_base;
+	unsigned int evq_rptr_tbl_base;
+
+	unsigned int txd_ring_mask;
+	unsigned int rxd_ring_mask;
+	unsigned int evq_size;
+	dma_addr_t max_dma_mask;
+	unsigned int tx_dma_mask;
+	unsigned bug5391_mask;
+
+	int rx_xoff_thresh;
+	int rx_xon_thresh;
+	unsigned int rx_buffer_padding;
+	unsigned int max_interrupt_mode;
+	unsigned int phys_addr_channels;
+};
+
+/**************************************************************************
+ *
+ * Prototypes and inline functions
+ *
+ *************************************************************************/
+
+/* Iterate over all used channels */
+#define efx_for_each_channel(_channel, _efx)				\
+	for (_channel = &_efx->channel[0];				\
+	     _channel < &_efx->channel[EFX_MAX_CHANNELS];		\
+	     _channel++)						\
+		if (!_channel->used_flags)				\
+			continue;					\
+		else
+
+/* Iterate over all used channels with interrupts */
+#define efx_for_each_channel_with_interrupt(_channel, _efx)		\
+	for (_channel = &_efx->channel[0];				\
+	     _channel < &_efx->channel[EFX_MAX_CHANNELS];		\
+	     _channel++)						\
+		if (!(_channel->used_flags && _channel->has_interrupt))	\
+			continue;					\
+		else
+
+/* Iterate over all used TX queues */
+#define efx_for_each_tx_queue(_tx_queue, _efx)				\
+	for (_tx_queue = &_efx->tx_queue[0];				\
+	     _tx_queue < &_efx->tx_queue[EFX_MAX_TX_QUEUES];		\
+	     _tx_queue++)						\
+		if (!_tx_queue->used)					\
+			continue;					\
+		else
+
+/* Iterate over all TX queues belonging to a channel */
+#define efx_for_each_channel_tx_queue(_tx_queue, _channel)		\
+	for (_tx_queue = &_channel->efx->tx_queue[0];			\
+	     _tx_queue < &_channel->efx->tx_queue[EFX_MAX_TX_QUEUES];	\
+	     _tx_queue++)						\
+		if ((!_tx_queue->used) ||				\
+		    (_tx_queue->channel != _channel))			\
+			continue;					\
+		else
+
+/* Iterate over all used RX queues */
+#define efx_for_each_rx_queue(_rx_queue, _efx)				\
+	for (_rx_queue = &_efx->rx_queue[0];				\
+	     _rx_queue < &_efx->rx_queue[EFX_MAX_RX_QUEUES];		\
+	     _rx_queue++)						\
+		if (!_rx_queue->used)					\
+			continue;					\
+		else
+
+/* Iterate over all RX queues belonging to a channel */
+#define efx_for_each_channel_rx_queue(_rx_queue, _channel)		\
+	for (_rx_queue = &_channel->efx->rx_queue[0];			\
+	     _rx_queue < &_channel->efx->rx_queue[EFX_MAX_RX_QUEUES];	\
+	     _rx_queue++)						\
+		if ((!_rx_queue->used) ||				\
+		    (_rx_queue->channel != _channel))			\
+			continue;					\
+		else
+
+/* Name formats */
+#define EFX_CHANNEL_NAME(_channel) "channel%d", _channel->channel
+#define EFX_TX_QUEUE_NAME(_tx_queue) "txq%d", _tx_queue->queue
+#define EFX_RX_QUEUE_NAME(_rx_queue) "rxq%d", _rx_queue->queue
+
+/* Returns a pointer to the specified receive buffer in the RX
+ * descriptor queue.
+ */
+static inline struct efx_rx_buffer *efx_rx_buffer(struct efx_rx_queue *rx_queue,
+						  unsigned int index)
+{
+	return (&rx_queue->buffer[index]);
+}
+
+/* Set bit in a little-endian bitfield */
+static inline void set_bit_le(int nr, unsigned char *addr)
+{
+	addr[nr / 8] |= (1 << (nr % 8));
+}
+
+/* Clear bit in a little-endian bitfield */
+static inline void clear_bit_le(int nr, unsigned char *addr)
+{
+	addr[nr / 8] &= ~(1 << (nr % 8));
+}
+
+
+/**
+ * EFX_MAX_FRAME_LEN - calculate maximum frame length
+ *
+ * This calculates the maximum frame length that will be used for a
+ * given MTU.  The frame length will be equal to the MTU plus a
+ * constant amount of header space and padding.  This is the quantity
+ * that the net driver will program into the MAC as the maximum frame
+ * length.
+ *
+ * The 10G MAC used in Falcon requires 8-byte alignment on the frame
+ * length, so we round up to the nearest 8.
+ */
+#define EFX_MAX_FRAME_LEN(mtu) \
+	((((mtu) + ETH_HLEN + VLAN_HLEN + 4/* FCS */) + 7) & ~7)
+
+
+#endif /* EFX_NET_DRIVER_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/null_phy.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/null_phy.c
--- linux-2.6.18.8/drivers/net/sfc/null_phy.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/null_phy.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,62 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include "efx.h"
+#include "falcon.h"
+#include "gmii.h"
+#include "phy.h"
+
+static int falcon_null_phy_check_hw(struct efx_nic *efx)
+{
+	int link_ok = falcon_xaui_link_ok(efx);
+
+	/* Generate PHY event that a PHY would have generated */
+	if (link_ok != efx->link_up) {
+		efx->link_up = link_ok;
+		efx->mac_op->fake_phy_event(efx);
+	}
+
+	return 0;
+}
+
+static void falcon_null_phy_reconfigure(struct efx_nic *efx)
+{
+	/* CX4 is always 10000FD only */
+	efx->link_options = GM_LPA_10000FULL;
+
+	falcon_null_phy_check_hw(efx);
+}
+
+struct efx_phy_operations falcon_null_phy_ops = {
+	.reconfigure     = falcon_null_phy_reconfigure,
+	.check_hw        = falcon_null_phy_check_hw,
+	.fini            = efx_port_dummy_op_void,
+	.clear_interrupt = efx_port_dummy_op_void,
+	.init            = efx_port_dummy_op_int,
+	.reset_xaui      = efx_port_dummy_op_void,
+	.mmds            = 0,
+	.loopbacks       = 0,
+	.startup_loopback = 0,
+};
diff -rpuN linux-2.6.18.8/drivers/net/sfc/phy.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/phy.c
--- linux-2.6.18.8/drivers/net/sfc/phy.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/phy.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,28 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2007:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include "net_driver.h"
+#include "phy.h"
+
diff -rpuN linux-2.6.18.8/drivers/net/sfc/phy.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/phy.h
--- linux-2.6.18.8/drivers/net/sfc/phy.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/phy.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,90 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2007:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_PHY_H
+#define EFX_PHY_H
+
+/****************************************************************************
+ * 10Xpress (SFX7101) PHY
+ */
+extern struct efx_phy_operations falcon_tenxpress_phy_ops;
+
+enum tenxpress_state {
+	TENXPRESS_STATUS_OFF = 0,
+	TENXPRESS_STATUS_OTEMP = 1,
+	TENXPRESS_STATUS_NORMAL = 2,
+};
+
+extern void tenxpress_set_state(struct efx_nic *efx,
+				enum tenxpress_state state);
+extern void tenxpress_phy_blink(struct efx_nic *efx, int blink);
+extern void tenxpress_crc_err(struct efx_nic *efx);
+
+/****************************************************************************
+ * Marvell 88E1111 "Alaska" PHY control
+ */
+extern struct efx_phy_operations alaska_phy_operations;
+
+/****************************************************************************
+* Exported functions from the driver for Transwitch CX4 retimer
+*/
+extern struct efx_phy_operations falcon_txc_phy_ops;
+
+#define TXC_GPIO_DIR_INPUT  (0)
+#define TXC_GPIO_DIR_OUTPUT (1)
+
+extern void txc_set_gpio_dir(struct efx_nic *p, int pin, int dir);
+extern void txc_set_gpio_val(struct efx_nic *p, int pin, int val);
+
+/****************************************************************************
+ * Exported functions from the driver for PMC PM8358 PHY
+ */
+extern struct efx_phy_operations falcon_pm8358_phy_ops;
+
+/****************************************************************************
+ * Exported functions from the driver for XFP optical PHYs
+ */
+extern struct efx_phy_operations falcon_xfp_phy_ops;
+
+/* The QUAKE XFP PHY provides various H/W control states for LEDs */
+#define QUAKE_LED_LINK_INVAL	(0)
+#define QUAKE_LED_LINK_STAT	(1)
+#define QUAKE_LED_LINK_ACT	(2)
+#define QUAKE_LED_LINK_ACTSTAT	(3)
+#define QUAKE_LED_OFF		(4)
+#define QUAKE_LED_ON		(5)
+#define QUAKE_LED_LINK_INPUT	(6)	/* Pin is an input. */
+/* What link the LED tracks */
+#define QUAKE_LED_TXLINK	(0)
+#define QUAKE_LED_RXLINK	(8)
+
+extern void xfp_set_led(struct efx_nic *p, int led, int state);
+
+/****************************************************************************
+ * NULL PHY ops
+ */
+extern struct efx_phy_operations falcon_null_phy_ops;
+
+#endif
diff -rpuN linux-2.6.18.8/drivers/net/sfc/pm8358_phy.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/pm8358_phy.c
--- linux-2.6.18.8/drivers/net/sfc/pm8358_phy.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/pm8358_phy.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,206 @@
+/* Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+/*
+ *  Driver for PMC-Sierra PM8358 XAUI PHY
+ */
+
+#include <linux/delay.h>
+#include "efx.h"
+#include "gmii.h"
+#include "mdio_10g.h"
+#include "phy.h"
+
+#define PM8358_REQUIRED_DEVS (MDIO_MMDREG_DEVS0_DTEXS)
+#define PM8358_LOOPBACKS (1 << LOOPBACK_PHY)
+
+/* PHY-specific definitions */
+/* Master ID and Global Performance Monitor Update */
+#define PMC_MASTER_REG (0xd000)
+/* Analog TX/RX settings under software control */
+#define PMC_MASTER_ANLG_CTRL (1 << 11)
+
+#define PMC_MCONF2_REG	(0xd002)
+/* Drive TX off centre of data eye (1) vs. clock edge (0) */
+#define	PMC_MCONF2_TEDGE (1 << 2)
+/* Drive RX off centre of data eye (1) vs. clock edge (0) */
+#define PMC_MCONF2_REDGE (1 << 3)
+
+/* Analog RX settings */
+#define PMC_ANALOG_RX_CFG0   (0xd025)
+#define PMC_ANALOG_RX_CFG1   (0xd02d)
+#define PMC_ANALOG_RX_CFG2   (0xd035)
+#define PMC_ANALOG_RX_CFG3   (0xd03d)
+#define PMC_ANALOG_RX_TERM     (1 << 15) /* Bit 15 of RX CFG: 0 for 100 ohms
+					    float, 1 for 50 to 1.2V */
+#define PMC_ANALOG_RX_EQ_MASK (3 << 8)
+#define PMC_ANALOG_RX_EQ_NONE (0 << 8)
+#define PMC_ANALOG_RX_EQ_HALF (1 << 8)
+#define PMC_ANALOG_RX_EQ_FULL (2 << 8)
+#define PMC_ANALOG_RX_EQ_RSVD (3 << 8)
+
+/* Reset the DTE XS MMD. */
+#define PMC_MAX_RESET_TIME 500
+#define PMC_RESET_WAIT 10
+
+static int pmc_reset_phy(struct efx_nic *efx)
+{
+	int rc = mdio_clause45_reset_mmd(efx, MDIO_MMD_DTEXS,
+					 PMC_MAX_RESET_TIME / PMC_RESET_WAIT,
+					 PMC_RESET_WAIT);
+	if (rc >= 0) {
+		EFX_TRACE(efx, "PMC8358: came out of reset with "
+			  "%d0 ms left\n", rc);
+		rc = 0;
+	} else {
+		EFX_ERR(efx, "PMC8358: reset timed out!\n");
+	}
+	return rc;
+}
+
+
+static void pmc_full_rx_eq(struct efx_nic *efx)
+{
+	int i, reg;
+
+	/* Enable software control of analog settings */
+	reg = mdio_clause45_read(efx, efx->mii.phy_id,
+				 MDIO_MMD_DTEXS, PMC_MASTER_REG);
+	reg |= PMC_MASTER_ANLG_CTRL;
+
+	mdio_clause45_write(efx, efx->mii.phy_id,
+			    MDIO_MMD_DTEXS, PMC_MASTER_REG, reg);
+
+	/* Turn RX eq on full for all channels. */
+	for (i = 0; i < 3; i++) {
+		/* The analog CFG registers are evenly spaced 8 apart */
+		u16 addr = PMC_ANALOG_RX_CFG0 + 8 * i;
+
+		reg = mdio_clause45_read(efx, efx->mii.phy_id,
+					 MDIO_MMD_DTEXS, addr);
+		reg = (reg & ~PMC_ANALOG_RX_EQ_MASK) | PMC_ANALOG_RX_EQ_FULL;
+		mdio_clause45_write(efx, efx->mii.phy_id,
+				    MDIO_MMD_DTEXS, addr, reg);
+	}
+}
+
+static void pmc_set_data_edges(struct efx_nic *efx)
+{
+	int reg;
+	/* Set TEDGE, clear REDGE */
+	reg = mdio_clause45_read(efx, efx->mii.phy_id,
+				 MDIO_MMD_DTEXS, PMC_MCONF2_REG);
+	reg &= ~PMC_MCONF2_REDGE;
+	reg |= PMC_MCONF2_TEDGE;
+
+	mdio_clause45_write(efx, efx->mii.phy_id,
+			    MDIO_MMD_DTEXS, PMC_MCONF2_REG, reg);
+}
+
+static int pm8358_phy_init(struct efx_nic *efx)
+{
+	u32 devid;
+	int rc;
+
+	/* The GLB_CTL reset line has been pulled before this is called,
+	 * and it may take up to 5ms for the PLL's to synchronise after
+	 * this is done. Best to wait 10ms here */
+	schedule_timeout_uninterruptible(HZ / 100);
+
+	rc = pmc_reset_phy(efx);
+	if (rc < 0)
+		return rc;
+
+	/* Check that all the MMDs we expect are present and responding. We
+	 * expect faults on some if the link is down, but not on the PHY XS */
+	rc = mdio_clause45_check_mmds(efx, PM8358_REQUIRED_DEVS, 0);
+	if (rc < 0)
+		return rc;
+
+	devid = mdio_clause45_read_id(efx, MDIO_MMD_DTEXS);
+	EFX_LOG(efx, "PM8358: PHY ID reg %x (OUI %x model %x revision"
+		" %x)\n", devid, MDIO_ID_OUI(devid), MDIO_ID_MODEL(devid),
+		MDIO_ID_REV(devid));
+
+	/* Turn on full RX equalisation */
+	pmc_full_rx_eq(efx);
+
+	/* Adjust RX and TX data edge position */
+	pmc_set_data_edges(efx);
+
+	EFX_LOG(efx, "PM8358: PHY init successful.\n");
+	return rc;
+}
+
+static int pm8358_link_ok(struct efx_nic *efx)
+{
+	return mdio_clause45_links_ok(efx, PM8358_REQUIRED_DEVS);
+}
+
+static int pm8358_phy_check_hw(struct efx_nic *efx)
+{
+	int rc = 0;
+	int link_up = pm8358_link_ok(efx);
+	/* Simulate a PHY event if link state has changed */
+	if (link_up != efx->link_up) {
+		efx->link_up = link_up;
+		efx->mac_op->fake_phy_event(efx);
+	}
+
+	return rc;
+}
+
+static void pm8358_phy_reconfigure(struct efx_nic *efx)
+{
+	int phy_id = efx->mii.phy_id;
+	int ctrl;
+	/* Handle DTE loopback */
+	ctrl = mdio_clause45_read(efx, phy_id, MDIO_MMD_DTEXS,
+				  MDIO_MMDREG_CTRL1);
+	if (efx->loopback_mode == LOOPBACK_PHY) {
+		EFX_TRACE(efx, "PM8358: setting DTE loopback\n");
+		ctrl |= (1 << MDIO_MMDREG_CTRL1_LBACK_LBN);
+	} else {
+		if (ctrl & (1 << MDIO_MMDREG_CTRL1_LBACK_LBN))
+			EFX_TRACE(efx,
+				  "PM8358: clearing DTE loopback\n");
+		ctrl &= ~(1 << MDIO_MMDREG_CTRL1_LBACK_LBN);
+	}
+	mdio_clause45_write(efx, phy_id, MDIO_MMD_DTEXS,
+			    MDIO_MMDREG_CTRL1, ctrl);
+
+	efx->link_up = pm8358_link_ok(efx);
+	efx->link_options = GM_LPA_10000FULL;
+}
+
+struct efx_phy_operations falcon_pm8358_phy_ops = {
+	.init            = pm8358_phy_init,
+	.reconfigure     = pm8358_phy_reconfigure,
+	.check_hw        = pm8358_phy_check_hw,
+	.fini            = efx_port_dummy_op_void,
+	.clear_interrupt = efx_port_dummy_op_void,
+	.reset_xaui      = efx_port_dummy_op_void,
+	.mmds            = PM8358_REQUIRED_DEVS,
+	.loopbacks       = PM8358_LOOPBACKS,
+	.startup_loopback = LOOPBACK_PHY,
+};
diff -rpuN linux-2.6.18.8/drivers/net/sfc/rx.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/rx.c
--- linux-2.6.18.8/drivers/net/sfc/rx.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/rx.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,798 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/socket.h>
+#include <linux/in.h>
+#include <linux/ip.h>
+#include <linux/tcp.h>
+#include <linux/udp.h>
+#include <net/ip.h>
+#include <net/checksum.h>
+#include "net_driver.h"
+#include "rx.h"
+#include "efx.h"
+#include "falcon.h"
+#include "selftest.h"
+#include "workarounds.h"
+
+
+/* Number of RX descriptors pushed at once. */
+#define EFX_RX_BATCH  8
+
+/* Size of buffer allocated for skb header area. */
+#define EFX_SKB_HEADERS  64u
+
+/*
+ * rx_alloc_method - RX buffer allocation method
+ *
+ * This driver supports two methods for allocating and using RX buffers:
+ * each RX buffer may be backed by an skb or by an order-n page.
+ *
+ * When LRO is in use then the second method has a lower overhead,
+ * since we don't have to allocate then free skbs on reassembled frames.
+ *
+ * Values:
+ *   - RX_ALLOC_METHOD_AUTO = 0
+ *   - RX_ALLOC_METHOD_SKB  = 1
+ *   - RX_ALLOC_METHOD_PAGE = 2
+ *
+ * The heuristic for %RX_ALLOC_METHOD_AUTO is a simple hysteresis count
+ * controlled by the parameters below.
+ *
+ *   - Since pushing and popping descriptors are separated by the rx_queue
+ *     size, so the watermarks should be ~rxd_size.
+ *   - The performance win by using page-based allocation for LRO is less
+ *     than the performance hit of using page-based allocation of non-LRO,
+ *     so the watermarks should reflect this.
+ *
+ * Per channel we maintain a single variable, updated by each channel:
+ *
+ *   rx_alloc_level += (lro_performed ? RX_ALLOC_FACTOR_LRO :
+ *                      RX_ALLOC_FACTOR_SKB)
+ * Per NAPI poll interval, we constrain rx_alloc_level to 0..MAX (which
+ * limits the hysteresis), and update the allocation strategy:
+ *
+ *   rx_alloc_method = (rx_alloc_level > RX_ALLOC_LEVEL_LRO ?
+ *                      RX_ALLOC_METHOD_PAGE : RX_ALLOC_METHOD_SKB)
+ */
+static int rx_alloc_method = RX_ALLOC_METHOD_PAGE;
+
+#define RX_ALLOC_LEVEL_LRO 0x2000
+#define RX_ALLOC_LEVEL_MAX 0x3000
+#define RX_ALLOC_FACTOR_LRO 1
+#define RX_ALLOC_FACTOR_SKB -2
+
+/* This is the percentage fill level below which new RX descriptors
+ * will be added to the RX descriptor ring.
+ */
+static unsigned int rx_refill_threshold = 90;
+
+/* This is the percentage fill level to which an RX queue will be refilled
+ * when the "RX refill threshold" is reached.
+ */
+static unsigned int rx_refill_limit = 95;
+
+/*
+ * RX maximum head room required.
+ *
+ * This must be at least 1 to prevent overflow and at least 2 to allow
+ * pipelined receives.
+ */
+#define EFX_RXD_HEAD_ROOM 2
+
+/* Macros for zero-order pages (potentially) containing multiple RX buffers */
+#define RX_DATA_OFFSET(_data)				\
+	(((unsigned long) (_data)) & (PAGE_SIZE-1))
+#define RX_BUF_OFFSET(_rx_buf)				\
+	RX_DATA_OFFSET((_rx_buf)->data)
+
+#define RX_PAGE_SIZE(_efx)				\
+	(PAGE_SIZE * (1u << (_efx)->rx_buffer_order))
+
+
+/**
+ * efx_init_rx_buffer_skb - create new RX buffer using skb-based allocation
+ *
+ * @rx_queue:		Efx RX queue
+ * @rx_buf:		RX buffer structure to populate
+ *
+ * This allocates memory for a new receive buffer, maps it for DMA,
+ * and populates a struct efx_rx_buffer with the relevant
+ * information.  Return a negative error code or 0 on success.
+ */
+static inline int efx_init_rx_buffer_skb(struct efx_rx_queue *rx_queue,
+					 struct efx_rx_buffer *rx_buf)
+{
+	struct efx_nic *efx = rx_queue->efx;
+	struct net_device *net_dev = efx->net_dev;
+	int skb_len = efx->rx_buffer_len;
+
+	rx_buf->skb = netdev_alloc_skb(net_dev, skb_len);
+	if (unlikely(!rx_buf->skb))
+		return -ENOMEM;
+
+	/* Adjust the SKB for padding and checksum */
+	skb_reserve(rx_buf->skb, NET_IP_ALIGN);
+	rx_buf->len = skb_len - NET_IP_ALIGN;
+	rx_buf->data = (char *)rx_buf->skb->data;
+
+	rx_buf->skb->ip_summed = CHECKSUM_UNNECESSARY;
+
+	/* Map for DMA */
+	rx_buf->dma_addr = pci_map_single(efx->pci_dev,
+					  rx_buf->data, rx_buf->len,
+					  PCI_DMA_FROMDEVICE);
+
+	if (unlikely(pci_dma_mapping_error(rx_buf->dma_addr))) {
+		/* Free the SKB */
+		dev_kfree_skb_any(rx_buf->skb);
+		rx_buf->skb = NULL;
+		return -EIO;
+	}
+
+	return 0;
+}
+
+/**
+ * efx_init_rx_buffer_page - create new RX buffer using page-based allocation
+ *
+ * @rx_queue:		Efx RX queue
+ * @rx_buf:		RX buffer structure to populate
+ *
+ * This allocates memory for a new receive buffer, maps it for DMA,
+ * and populates a struct efx_rx_buffer with the relevant
+ * information.  Return a negative error code or 0 on success.
+ */
+static inline int efx_init_rx_buffer_page(struct efx_rx_queue *rx_queue,
+					  struct efx_rx_buffer *rx_buf)
+{
+	struct efx_nic *efx = rx_queue->efx;
+	int bytes, space, offset;
+
+	bytes = efx->rx_buffer_len - EFX_PAGE_IP_ALIGN;
+
+	/* If there is space left in the previously allocated page,
+	 * then use it. Otherwise allocate a new one */
+	rx_buf->page = rx_queue->buf_page;
+	if (rx_buf->page == NULL) {
+		dma_addr_t dma_addr;
+
+		rx_buf->page = alloc_pages(__GFP_COLD | __GFP_COMP | GFP_ATOMIC,
+					   efx->rx_buffer_order);
+		if (unlikely(rx_buf->page == NULL))
+			return -ENOMEM;
+
+		/* Map the entire page for DMA */
+		dma_addr = pci_map_page(efx->pci_dev, rx_buf->page,
+					0, RX_PAGE_SIZE(efx),
+					PCI_DMA_FROMDEVICE);
+
+		if (unlikely(pci_dma_mapping_error(dma_addr))) {
+			__free_pages(rx_buf->page, efx->rx_buffer_order);
+			rx_buf->page = NULL;
+			return -EIO;
+		}
+
+		rx_queue->buf_page = rx_buf->page;
+		rx_queue->buf_dma_addr = dma_addr;
+		rx_queue->buf_data = ((char *) page_address(rx_buf->page) +
+				      EFX_PAGE_IP_ALIGN);
+	}
+
+	offset = RX_DATA_OFFSET(rx_queue->buf_data);
+	rx_buf->len = bytes;
+	rx_buf->dma_addr = rx_queue->buf_dma_addr + offset;
+	rx_buf->data = rx_queue->buf_data;
+
+	/* Try to pack multiple buffers per page */
+	if (efx->rx_buffer_order == 0) {
+		/* The next buffer starts on the next 512 byte boundary */
+		rx_queue->buf_data += ((bytes + 0x1ff) & ~0x1ff);
+		offset += ((bytes + 0x1ff) & ~0x1ff);
+
+		space = RX_PAGE_SIZE(efx) - offset;
+		if (space >= bytes) {
+			/* Refs dropped on kernel releasing each skb */
+			get_page(rx_queue->buf_page);
+			goto out;
+		}
+	}
+
+	/* This is the final RX buffer for this page, so mark it for
+	 * unmapping */
+	rx_queue->buf_page = NULL;
+	rx_buf->unmap_addr = rx_queue->buf_dma_addr;
+
+ out:
+	return 0;
+}
+
+/* This allocates memory for a new receive buffer, maps it for DMA,
+ * and populates a struct efx_rx_buffer with the relevant
+ * information.
+ */
+static inline int efx_init_rx_buffer(struct efx_rx_queue *rx_queue,
+				     struct efx_rx_buffer *new_rx_buf)
+{
+	int rc = 0;
+
+	if (rx_queue->channel->rx_alloc_push_pages) {
+		new_rx_buf->skb = NULL;
+		rc = efx_init_rx_buffer_page(rx_queue, new_rx_buf);
+		rx_queue->alloc_page_count++;
+	} else {
+		new_rx_buf->page = NULL;
+		rc = efx_init_rx_buffer_skb(rx_queue, new_rx_buf);
+		rx_queue->alloc_skb_count++;
+	}
+
+	if (unlikely(rc < 0))
+		EFX_LOG_RL(rx_queue->efx, "%s RXQ[%d] =%d\n", __func__,
+			   rx_queue->queue, rc);
+	return rc;
+}
+
+static inline void efx_unmap_rx_buffer(struct efx_nic *efx,
+				       struct efx_rx_buffer *rx_buf)
+{
+	if (rx_buf->page) {
+		EFX_BUG_ON_PARANOID(rx_buf->skb);
+		if (rx_buf->unmap_addr) {
+			pci_unmap_page(efx->pci_dev, rx_buf->unmap_addr,
+				       RX_PAGE_SIZE(efx), PCI_DMA_FROMDEVICE);
+			rx_buf->unmap_addr = 0;
+		}
+	} else if (likely(rx_buf->skb)) {
+		pci_unmap_single(efx->pci_dev, rx_buf->dma_addr,
+				 rx_buf->len, PCI_DMA_FROMDEVICE);
+	}
+}
+
+static inline void efx_free_rx_buffer(struct efx_nic *efx,
+				      struct efx_rx_buffer *rx_buf)
+{
+	if (rx_buf->page) {
+		__free_pages(rx_buf->page, efx->rx_buffer_order);
+		rx_buf->page = NULL;
+	} else if (likely(rx_buf->skb)) {
+		dev_kfree_skb_any(rx_buf->skb);
+		rx_buf->skb = NULL;
+	}
+}
+
+inline void efx_fini_rx_buffer(struct efx_rx_queue *rx_queue,
+			       struct efx_rx_buffer *rx_buf)
+{
+	/* Unmap for DMA */
+	efx_unmap_rx_buffer(rx_queue->efx, rx_buf);
+
+	/* Free the skb/page */
+	efx_free_rx_buffer(rx_queue->efx, rx_buf);
+}
+
+/**
+ * efx_fast_push_rx_descriptors - push new RX descriptors quickly
+ * @rx_queue:		RX descriptor queue
+ * @retry:              Recheck the fill level
+ * This will aim to fill the RX descriptor queue up to
+ * @rx_queue->@fast_fill_limit. If there is insufficient atomic
+ * memory to do so, the caller should retry.
+ */
+static int __efx_fast_push_rx_descriptors(struct efx_rx_queue *rx_queue,
+					  int retry)
+{
+	struct efx_rx_buffer *rx_buf;
+	unsigned fill_level, index;
+	int i, space, rc = 0;
+
+	/* Calculate current fill level.  Do this outside the lock,
+	 * because most of the time we'll end up not wanting to do the
+	 * fill anyway.
+	 */
+	fill_level = (rx_queue->added_count - rx_queue->removed_count);
+	EFX_BUG_ON_PARANOID(fill_level >
+			    rx_queue->efx->type->rxd_ring_mask + 1);
+
+	/* Don't fill if we don't need to */
+	if (fill_level >= rx_queue->fast_fill_trigger)
+		return 0;
+
+	/* Record minimum fill level */
+	if (unlikely(fill_level < rx_queue->min_fill))
+		if (fill_level)
+			rx_queue->min_fill = fill_level;
+
+	/* Acquire RX add lock.  If this lock is contended, then a fast
+	 * fill must already be in progress (e.g. in the refill
+	 * tasklet), so we don't need to do anything
+	 */
+	if (!spin_trylock_bh(&rx_queue->add_lock))
+		return -1;
+
+ retry:
+	/* Recalculate current fill level now that we have the lock */
+	fill_level = (rx_queue->added_count - rx_queue->removed_count);
+	EFX_BUG_ON_PARANOID(fill_level >
+			    rx_queue->efx->type->rxd_ring_mask + 1);
+	space = rx_queue->fast_fill_limit - fill_level;
+	if (space < EFX_RX_BATCH)
+		goto out_unlock;
+
+	EFX_TRACE(rx_queue->efx, "RX queue %d fast-filling descriptor ring from"
+		  " level %d to level %d using %s allocation\n",
+		  rx_queue->queue, fill_level, rx_queue->fast_fill_limit,
+		  rx_queue->channel->rx_alloc_push_pages ? "page" : "skb");
+
+	do {
+		for (i = 0; i < EFX_RX_BATCH; ++i) {
+			index = (rx_queue->added_count &
+				 rx_queue->efx->type->rxd_ring_mask);
+			rx_buf = efx_rx_buffer(rx_queue, index);
+			rc = efx_init_rx_buffer(rx_queue, rx_buf);
+			if (unlikely(rc))
+				goto out;
+			++rx_queue->added_count;
+		}
+	} while ((space -= EFX_RX_BATCH) >= EFX_RX_BATCH);
+
+	EFX_TRACE(rx_queue->efx, "RX queue %d fast-filled descriptor ring "
+		  "to level %d\n", rx_queue->queue,
+		  rx_queue->added_count - rx_queue->removed_count);
+
+ out:
+	/* Send write pointer to card.  */
+	falcon_notify_rx_desc(rx_queue);
+
+	/* If the fast fill is running inside from the refill tasklet, then
+	 * for SMP systems it may be running on a different CPU to
+	 * RX event processing, which means that the fill level may now be
+	 * out of date. */
+	if (unlikely(retry && (rc == 0)))
+		goto retry;
+
+ out_unlock:
+	/* Release RX add lock */
+	spin_unlock_bh(&rx_queue->add_lock);
+
+	return rc;
+}
+
+/**
+ * efx_fast_push_rx_descriptors - push new RX descriptors quickly
+ * @rx_queue:		RX descriptor queue
+ *
+ * This will aim to fill the RX descriptor queue up to
+ * @rx_queue->@fast_fill_limit.  If there is insufficient memory to do so,
+ * it will schedule a work item to immediately continue the fast fill
+ */
+void efx_fast_push_rx_descriptors(struct efx_rx_queue *rx_queue)
+{
+	int rc;
+
+	rc = __efx_fast_push_rx_descriptors(rx_queue, 0);
+	if (unlikely(rc)) {
+		/* Schedule the work item to run immediately. The hope is
+		 * that work is immediately pending to free some memory
+		 * (e.g. an RX event or TX completion)
+		 */
+		queue_delayed_work(rx_queue->efx->refill_workqueue,
+				   &rx_queue->work, 0);
+	}
+}
+
+void efx_rx_work(struct work_struct *data)
+{
+	struct efx_rx_queue *rx_queue;
+	int rc;
+
+#if !defined(EFX_NEED_WORK_API_WRAPPERS)
+	rx_queue = container_of(data, struct efx_rx_queue, work.work);
+#else
+	rx_queue = container_of(data, struct efx_rx_queue, work);
+#endif
+
+	if (unlikely(!rx_queue->channel->enabled))
+		return;
+
+	EFX_TRACE(rx_queue->efx, "RX queue %d worker thread executing on CPU "
+		  "%d\n", rx_queue->queue, raw_smp_processor_id());
+
+	++rx_queue->slow_fill_count;
+	/* Push new RX descriptors, allowing at least 1 jiffy for
+	 * the kernel to free some more memory. */
+	rc = __efx_fast_push_rx_descriptors(rx_queue, 1);
+	if (rc) {
+		queue_delayed_work(rx_queue->efx->refill_workqueue,
+				   &rx_queue->work, 1);
+	}
+}
+
+static inline void efx_rx_packet__check_len(struct efx_rx_queue *rx_queue,
+					    struct efx_rx_buffer *rx_buf,
+					    int len, int *discard,
+					    int *leak_packet)
+{
+	struct efx_nic *efx = rx_queue->efx;
+	unsigned max_len = rx_buf->len - efx->type->rx_buffer_padding;
+
+	if (likely(len <= max_len))
+		return;
+
+	/* The packet must be discarded, but this is only a fatal error
+	 * if the caller indicated it was
+	 */
+	*discard = 1;
+
+	if ((len > rx_buf->len) && EFX_WORKAROUND_8071(efx)) {
+		EFX_ERR_RL(efx, " RX queue %d seriously overlength "
+			   "RX event (0x%x > 0x%x+0x%x). Leaking\n",
+			   rx_queue->queue, len, max_len,
+			   efx->type->rx_buffer_padding);
+		/* If this buffer was skb-allocated, then the meta
+		 * data at the end of the skb will be trashed. So
+		 * we have no choice but to leak the fragment.
+		 */
+		*leak_packet = (rx_buf->skb != NULL);
+		efx_schedule_reset(efx, RESET_TYPE_RX_RECOVERY);
+	} else {
+		EFX_ERR_RL(efx, " RX queue %d overlength RX event "
+			   "(0x%x > 0x%x)\n", rx_queue->queue, len, max_len);
+	}
+
+	rx_queue->channel->n_rx_overlength++;
+}
+
+/* Allocate and construct an SKB around a struct page.*/
+static inline struct sk_buff *efx_rx_mk_skb(struct efx_rx_buffer *rx_buf,
+					    struct efx_nic *efx,
+					    int hdr_len)
+{
+	struct sk_buff *skb;
+
+	/* Allocate an SKB to store the headers */
+	skb = netdev_alloc_skb(efx->net_dev, hdr_len + EFX_PAGE_SKB_ALIGN);
+	if (unlikely(skb == NULL)) {
+		EFX_ERR_RL(efx, "RX out of memory for skb\n");
+		return NULL;
+	}
+
+	EFX_BUG_ON_PARANOID(skb_shinfo(skb)->nr_frags);
+	EFX_BUG_ON_PARANOID(rx_buf->len < hdr_len);
+
+	skb->ip_summed = CHECKSUM_UNNECESSARY;
+	skb_reserve(skb, EFX_PAGE_SKB_ALIGN);
+
+	skb->len = rx_buf->len;
+	skb->truesize = rx_buf->len + sizeof(struct sk_buff);
+	memcpy(skb->data, rx_buf->data, hdr_len);
+	skb->tail += hdr_len;
+
+	/* Append the remaining page onto the frag list */
+	if (unlikely(rx_buf->len > hdr_len)) {
+		struct skb_frag_struct *frag = skb_shinfo(skb)->frags;
+		frag->page = rx_buf->page;
+		frag->page_offset = RX_BUF_OFFSET(rx_buf) + hdr_len;
+		frag->size = skb->len - hdr_len;
+		skb_shinfo(skb)->nr_frags = 1;
+		skb->data_len = frag->size;
+	} else {
+		__free_pages(rx_buf->page, efx->rx_buffer_order);
+		skb->data_len = 0;
+	}
+
+	/* Ownership has transferred from the rx_buf to skb */
+	rx_buf->page = NULL;
+
+	/* Move past the ethernet header */
+	skb->protocol = eth_type_trans(skb, efx->net_dev);
+
+	return skb;
+}
+
+#if defined(EFX_USE_FASTCALL)
+void fastcall efx_rx_packet(struct efx_rx_queue *rx_queue,
+			    unsigned int index, unsigned int len,
+			    int checksummed, int discard)
+#else
+void efx_rx_packet(struct efx_rx_queue *rx_queue, unsigned int index,
+		   unsigned int len, int checksummed, int discard)
+#endif
+{
+	struct efx_nic *efx = rx_queue->efx;
+	struct efx_rx_buffer *rx_buf;
+	int leak_packet = 0;
+
+	rx_buf = efx_rx_buffer(rx_queue, index);
+	EFX_BUG_ON_PARANOID(!rx_buf->data);
+	EFX_BUG_ON_PARANOID(rx_buf->skb && rx_buf->page);
+	EFX_BUG_ON_PARANOID(!(rx_buf->skb || rx_buf->page));
+
+	/* This allows the refill path to post another buffer.
+	 * EFX_RXD_HEAD_ROOM ensures that the slot we are using
+	 * isn't overwritten yet.
+	 */
+	rx_queue->removed_count++;
+
+	/* Validate the length encoded in the event vs the descriptor pushed */
+	efx_rx_packet__check_len(rx_queue, rx_buf, len,
+				 &discard, &leak_packet);
+
+	EFX_TRACE(efx, "RX queue %d received id %x at %llx+%x %s%s\n",
+		  rx_queue->queue, index,
+		  (unsigned long long)rx_buf->dma_addr, len,
+		  (checksummed ? " [SUMMED]" : ""),
+		  (discard ? " [DISCARD]" : ""));
+
+	/* Discard packet, if instructed to do so */
+	if (unlikely(discard)) {
+		if (unlikely(leak_packet))
+			rx_queue->channel->n_skbuff_leaks++;
+		else
+			/* We haven't called efx_unmap_rx_buffer yet,
+			 * so fini the entire rx_buffer here */
+			efx_fini_rx_buffer(rx_queue, rx_buf);
+		return;
+	}
+
+	/* Release card resources - assumes all RX buffers consumed in-order
+	 * per RX queue
+	 */
+	efx_unmap_rx_buffer(efx, rx_buf);
+
+	/* Prefetch nice and early so data will (hopefully) be in cache by
+	 * the time we look at it.
+	 */
+	prefetch(rx_buf->data);
+
+	/* Pipeline receives so that we give time for packet headers to be
+	 * prefetched into cache.
+	 */
+	rx_buf->len = len;
+	if (rx_queue->channel->rx_pkt)
+		__efx_rx_packet(rx_queue->channel,
+				rx_queue->channel->rx_pkt,
+				rx_queue->channel->rx_pkt_csummed);
+	rx_queue->channel->rx_pkt = rx_buf;
+	rx_queue->channel->rx_pkt_csummed = checksummed;
+}
+
+/* Handle a received packet.  Second half: Touches packet payload. */
+void __efx_rx_packet(struct efx_channel *channel,
+		     struct efx_rx_buffer *rx_buf, int checksummed)
+{
+	struct efx_nic *efx = channel->efx;
+	enum efx_veto veto;
+	struct sk_buff *skb;
+
+	/* If we're in loopback test, then pass the packet directly to the
+	 * loopback layer, and free the rx_buf here
+	 */
+	if (unlikely(efx->loopback_selftest)) {
+		efx_loopback_rx_packet(efx, rx_buf->data, rx_buf->len);
+		efx_free_rx_buffer(efx, rx_buf);
+		goto done;
+	}
+
+	if (rx_buf->skb) {
+		/* Prefetch more information */
+		prefetch(skb_shinfo(rx_buf->skb));
+
+		/* Reserve space for the data */
+		skb_put(rx_buf->skb, rx_buf->len);
+
+		/* Move past the ethernet header. rx_buf->data still points
+		 * at the ethernet header */
+		rx_buf->skb->protocol = eth_type_trans(rx_buf->skb,
+						       efx->net_dev);
+	}
+
+	/* Both our generic-LRO and SFC-SSR support skb and page based
+	 * allocation, but neither support switching from one to the
+	 * other on the fly. If we spot that the allocation mode has
+	 * changed, then flush the LRO state.
+	 */
+	if (unlikely(channel->rx_alloc_pop_pages != (rx_buf->page != NULL))) {
+		channel->rx_alloc_pop_pages = (rx_buf->page != NULL);
+	}
+
+	/* Allow callback to veto the packet */
+	veto = EFX_DL_CALLBACK(efx, rx_packet, rx_buf->data, rx_buf->len);
+	if (unlikely(veto)) {
+		EFX_LOG(efx, "RX vetoed by driverlink %s driver\n",
+			efx->dl_cb_dev.rx_packet->driver->name);
+		/* Free the buffer now */
+		efx_free_rx_buffer(efx, rx_buf);
+		goto done;
+	}
+
+	/* Form an skb if required */
+	if (rx_buf->page) {
+		int hdr_len = min(rx_buf->len, EFX_SKB_HEADERS);
+		skb = efx_rx_mk_skb(rx_buf, efx, hdr_len);
+		if (unlikely(skb == NULL)) {
+			efx_free_rx_buffer(efx, rx_buf);
+			goto done;
+		}
+	} else {
+		/* We now own the SKB */
+		skb = rx_buf->skb;
+		rx_buf->skb = NULL;
+	}
+
+	EFX_BUG_ON_PARANOID(rx_buf->page);
+	EFX_BUG_ON_PARANOID(rx_buf->skb);
+	EFX_BUG_ON_PARANOID(!skb);
+
+	/* Set the SKB flags */
+	if (unlikely(!checksummed || !efx->rx_checksum_enabled))
+		skb->ip_summed = CHECKSUM_NONE;
+
+	/* Pass the packet up */
+	netif_receive_skb(skb);
+
+	/* Update allocation strategy method */
+	channel->rx_alloc_level += RX_ALLOC_FACTOR_SKB;
+
+	/* fall-thru */
+done:
+	/* Update statistics */
+	efx->net_dev->last_rx = jiffies;
+}
+
+void efx_rx_strategy(struct efx_channel *channel)
+{
+	enum efx_rx_alloc_method method = rx_alloc_method;
+
+	/* Only makes sense to use page based allocation if LRO is enabled */
+	if (!(channel->efx->net_dev->features & NETIF_F_LRO)) {
+		method = RX_ALLOC_METHOD_SKB;
+	} else if (method == RX_ALLOC_METHOD_AUTO) {
+		/* Constrain the rx_alloc_level */
+		if (channel->rx_alloc_level < 0)
+			channel->rx_alloc_level = 0;
+		else if (channel->rx_alloc_level > RX_ALLOC_LEVEL_MAX)
+			channel->rx_alloc_level = RX_ALLOC_LEVEL_MAX;
+
+		/* Decide on the allocation method */
+		method = ((channel->rx_alloc_level > RX_ALLOC_LEVEL_LRO) ?
+			  RX_ALLOC_METHOD_PAGE : RX_ALLOC_METHOD_SKB);
+	}
+
+	/* Push the option */
+	channel->rx_alloc_push_pages = (method == RX_ALLOC_METHOD_PAGE);
+}
+
+int efx_probe_rx_queue(struct efx_rx_queue *rx_queue)
+{
+	struct efx_nic *efx = rx_queue->efx;
+	unsigned int rxq_size;
+	int rc;
+
+	EFX_LOG(efx, "creating RX queue %d\n", rx_queue->queue);
+
+	/* Allocate RX buffers */
+	rxq_size = (efx->type->rxd_ring_mask + 1) * sizeof(*rx_queue->buffer);
+	rx_queue->buffer = kzalloc(rxq_size, GFP_KERNEL);
+	if (!rx_queue->buffer) {
+		rc = -ENOMEM;
+		goto fail1;
+	}
+
+	rc = falcon_probe_rx(rx_queue);
+	if (rc)
+		goto fail2;
+
+	return 0;
+
+ fail2:
+	kfree(rx_queue->buffer);
+	rx_queue->buffer = NULL;
+ fail1:
+	/* Mark queue as unused */
+	rx_queue->used = 0;
+
+	return rc;
+}
+
+int efx_init_rx_queue(struct efx_rx_queue *rx_queue)
+{
+	struct efx_nic *efx = rx_queue->efx;
+	unsigned int max_fill, trigger, limit;
+
+	EFX_LOG(rx_queue->efx, "initialising RX queue %d\n", rx_queue->queue);
+
+	ASSERT_RTNL();
+
+	/* Initialise ptr fields */
+	rx_queue->added_count = 0;
+	rx_queue->notified_count = 0;
+	rx_queue->removed_count = 0;
+	rx_queue->min_fill = -1U;
+	rx_queue->min_overfill = -1U;
+
+	/* Initialise limit fields */
+	max_fill = efx->type->rxd_ring_mask + 1 - EFX_RXD_HEAD_ROOM;
+	trigger = max_fill * min(rx_refill_threshold, 100U) / 100U;
+	limit = max_fill * min(rx_refill_limit, 100U) / 100U;
+
+	rx_queue->max_fill = max_fill;
+	rx_queue->fast_fill_trigger = trigger;
+	rx_queue->fast_fill_limit = limit;
+
+	/* Set up RX descriptor ring */
+	return falcon_init_rx(rx_queue);
+}
+
+void efx_fini_rx_queue(struct efx_rx_queue *rx_queue)
+{
+	int i;
+	struct efx_rx_buffer *rx_buf;
+
+	EFX_LOG(rx_queue->efx, "shutting down RX queue %d\n", rx_queue->queue);
+
+	ASSERT_RTNL();
+
+	/* Flush RX queue and remove descriptor ring */
+	falcon_fini_rx(rx_queue);
+
+	/* Release RX buffers NB start at index 0 not current HW ptr */
+	if (rx_queue->buffer) {
+		for (i = 0; i <= rx_queue->efx->type->rxd_ring_mask; i++) {
+			rx_buf = efx_rx_buffer(rx_queue, i);
+			efx_fini_rx_buffer(rx_queue, rx_buf);
+		}
+	}
+
+	/* For a page that is part-way through splitting into RX buffers */
+	if (rx_queue->buf_page != NULL) {
+		pci_unmap_page(rx_queue->efx->pci_dev, rx_queue->buf_dma_addr,
+			       RX_PAGE_SIZE(rx_queue->efx), PCI_DMA_FROMDEVICE);
+		__free_pages(rx_queue->buf_page,
+			     rx_queue->efx->rx_buffer_order);
+		rx_queue->buf_page = NULL;
+	}
+}
+
+void efx_remove_rx_queue(struct efx_rx_queue *rx_queue)
+{
+	EFX_LOG(rx_queue->efx, "destroying RX queue %d\n", rx_queue->queue);
+
+	falcon_remove_rx(rx_queue);
+
+	kfree(rx_queue->buffer);
+	rx_queue->buffer = NULL;
+	rx_queue->used = 0;
+}
+
+
+module_param(rx_alloc_method, int, 0644);
+MODULE_PARM_DESC(rx_alloc_method, "Allocation method used for RX buffers");
+
+module_param(rx_refill_threshold, uint, 0444);
+MODULE_PARM_DESC(rx_refill_threshold,
+		 "RX descriptor ring fast/slow fill threshold (%)");
+
diff -rpuN linux-2.6.18.8/drivers/net/sfc/rx.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/rx.h
--- linux-2.6.18.8/drivers/net/sfc/rx.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/rx.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,44 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2006:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_RX_H
+#define EFX_RX_H
+
+#include "net_driver.h"
+
+
+int efx_probe_rx_queue(struct efx_rx_queue *rx_queue);
+void efx_remove_rx_queue(struct efx_rx_queue *rx_queue);
+int efx_init_rx_queue(struct efx_rx_queue *rx_queue);
+void efx_fini_rx_queue(struct efx_rx_queue *rx_queue);
+
+void efx_rx_strategy(struct efx_channel *channel);
+void efx_fast_push_rx_descriptors(struct efx_rx_queue *rx_queue);
+void efx_rx_work(struct work_struct *data);
+void __efx_rx_packet(struct efx_channel *channel,
+		     struct efx_rx_buffer *rx_buf, int checksummed);
+
+
+#endif /* EFX_RX_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/selftest.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/selftest.c
--- linux-2.6.18.8/drivers/net/sfc/selftest.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/selftest.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,815 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/netdevice.h>
+#include <linux/module.h>
+#include <linux/delay.h>
+#include <linux/kernel_stat.h>
+#include <linux/pci.h>
+#include <linux/ethtool.h>
+#include <linux/ip.h>
+#include <linux/in.h>
+#include <linux/udp.h>
+#include <linux/rtnetlink.h>
+#include <asm/io.h>
+#include "net_driver.h"
+#include "ethtool.h"
+#include "efx.h"
+#include "falcon.h"
+#include "selftest.h"
+#include "boards.h"
+#include "workarounds.h"
+
+/* Self tests */
+
+/*
+ * Loopback test packet structure
+ *
+ * The self-test should stress every RSS vector, and unfortunately
+ * Falcon only performs RSS on TCP/UDP packets.
+ */
+struct efx_loopback_payload {
+	struct ethhdr header;
+	struct iphdr ip;
+	struct udphdr udp;
+	__be16 iteration;
+	const char msg[64];
+} __attribute__ ((packed));
+
+/* Loopback test source MAC address */
+static const unsigned char payload_source[ETH_ALEN] = {
+	0x00, 0x0f, 0x53, 0x1b, 0x1b, 0x1b,
+};
+
+static const char *payload_msg =
+	"Hello world! This is an Efx loopback test in progress!";
+
+struct efx_selftest_state {
+	/* Drop all packets in  efx_loopback_rx_packet */
+	int flush;
+
+	/* Number of packets being used in this test */
+	int packet_count;
+
+	/* RX good packet count */
+	atomic_t rx_good;
+
+	/* RX bad packet count */
+	atomic_t rx_bad;
+
+	/* Payload used in tests */
+	struct efx_loopback_payload payload;
+};
+
+/**************************************************************************
+ *
+ * Configurable values
+ *
+ **************************************************************************/
+
+/* Level of loopback testing
+ *
+ * The maximum packet burst length is 16**(n-1), i.e.
+ *
+ * - Level 0 : no packets
+ * - Level 1 : 1 packet
+ * - Level 2 : 17 packets (1 * 1 packet, 1 * 16 packets)
+ * - Level 3 : 273 packets (1 * 1 packet, 1 * 16 packet, 1 * 256 packets)
+ *
+ */
+static unsigned int loopback_test_level = 3;
+
+/**************************************************************************
+ *
+ * Interrupt and event queue testing
+ *
+ **************************************************************************/
+
+/*
+ * Test interrupts
+ *
+ * This generates a test interrupt and waits for it to be received by
+ * a CPU.  This routine must be called from process context and will
+ * sleep.
+ */
+static int efx_test_interrupts(struct efx_nic *efx,
+			       struct efx_self_tests *tests)
+{
+	struct efx_channel *channel;
+	unsigned long j_start;
+
+	EFX_LOG(efx, "testing interrupts\n");
+	tests->interrupt = -1;
+
+	/* Reset interrupt flag */
+	efx->last_irq_cpu = -1;
+	smp_wmb();
+
+	/* To guarantee that an interrupt is received make sure that the
+	 * channels are ack'd at least once, reenabling interrupts.
+	 * We will then receive an interrupt, either by explicitly
+	 * requesting one, or receiving traffic.
+	 */
+	efx_for_each_channel_with_interrupt(channel, efx) {
+		/* Process the eventq synchronously */
+		if (channel->work_pending)
+			efx_process_channel_now(channel);
+		/* Check if we haven't already received an interrupt */
+		if (efx->last_irq_cpu >= 0)
+			goto success;
+	}
+
+	j_start = jiffies;
+
+	/* Generate test interrupt. */
+	falcon_generate_interrupt(efx);
+
+	/* Wait for arrival of test interrupt. */
+	EFX_LOG(efx, "waiting for test interrupt\n");
+	schedule_timeout_uninterruptible(HZ / 10);
+	if (efx->last_irq_cpu >= 0)
+		goto success;
+
+	EFX_ERR(efx, "timed out in %ld jiffies waiting for interrupt\n",
+		jiffies - j_start);
+	return -ETIMEDOUT;
+
+ success:
+	EFX_LOG(efx, "%s test interrupt seen on CPU%d\n", INT_MODE(efx),
+		efx->last_irq_cpu);
+	tests->interrupt = 1;
+	return 0;
+}
+
+/*
+ * Test capability of events to receive events
+ * This generates a test event and waits for it to be processed
+ * This routine must be called from process context and will sleep.
+ */
+static int efx_test_eventq(struct efx_channel *channel,
+			   struct efx_self_tests *tests)
+{
+	unsigned int magic;
+
+	/* Channel specific code, limited to 20 bits */
+	magic = (0x00010150 + channel->channel);
+	EFX_LOG(channel->efx, "channel %d testing event queue with code %x\n",
+		channel->channel, magic);
+
+	tests->eventq_dma[channel->channel] = -1;
+	tests->eventq_int[channel->channel] = 1;	/* fake pass */
+	tests->eventq_poll[channel->channel] = 1;	/* fake pass */
+
+	/* Reset flag and zero magic word */
+	channel->efx->last_irq_cpu = -1;
+	channel->eventq_magic = 0;
+	smp_wmb();
+
+	/* Generate test event */
+	falcon_generate_test_event(channel, magic);
+	udelay(1);
+
+	efx_process_channel_now(channel);
+	if (channel->eventq_magic != magic) {
+		EFX_ERR(channel->efx, "channel %d  failed to see test event\n",
+			channel->channel);
+		return -ETIMEDOUT;
+	} else {
+		tests->eventq_dma[channel->channel] = 1;
+	}
+
+	return 0;
+}
+
+
+/*
+ * Test capability of events to generate interrupts
+ * This generates a test event and waits for it to be processed by an
+ * ISR.  This routine must be called from process context and will
+ * sleep.
+ */
+static int efx_test_eventq_irq(struct efx_channel *channel,
+			       struct efx_self_tests *tests)
+{
+	unsigned int magic, count;
+	unsigned long j_start = jiffies;
+
+	/* Channel specific code, limited to 20 bits */
+	magic = (0x00010150 + channel->channel);
+	EFX_LOG(channel->efx, "channel %d testing event queue with code %x\n",
+		channel->channel, magic);
+
+	tests->eventq_dma[channel->channel] = -1;
+	tests->eventq_int[channel->channel] = -1;
+	tests->eventq_poll[channel->channel] = -1;
+
+	/* Reset flag and zero magic word */
+	channel->efx->last_irq_cpu = -1;
+	channel->eventq_magic = 0;
+	smp_wmb();
+
+	/* Generate test event */
+	falcon_generate_test_event(channel, magic);
+
+	/* Wait for arrival of interrupt */
+	count = 0;
+	do {
+		schedule_timeout_uninterruptible(HZ / 100);
+
+		if (channel->work_pending)
+			efx_process_channel_now(channel);
+
+		if (channel->eventq_magic == magic)
+			goto eventq_ok;
+	}
+	while (++count < 2);
+
+	EFX_ERR(channel->efx, "channel %d timed out in %ld jiffies waiting for"
+		" event queue\n", channel->channel, jiffies - j_start);
+
+	/* See if interrupt arrived */
+	if (channel->efx->last_irq_cpu >= 0) {
+		EFX_ERR(channel->efx, "channel %d saw interrupt on CPU%d "
+			"during event queue test\n", channel->channel,
+			raw_smp_processor_id());
+		tests->eventq_int[channel->channel] = 1;
+	}
+
+	/* Check to see if event was received even if interrupt wasn't */
+	efx_process_channel_now(channel);
+	if (channel->eventq_magic == magic) {
+		EFX_ERR(channel->efx, "channel %d event was generated, but "
+			"failed to trigger an interrupt\n", channel->channel);
+		tests->eventq_dma[channel->channel] = 1;
+	}
+
+	return -ETIMEDOUT;
+ eventq_ok:
+	EFX_LOG(channel->efx, "channel %d event queue passed\n",
+		channel->channel);
+	tests->eventq_dma[channel->channel] = 1;
+	tests->eventq_int[channel->channel] = 1;
+	tests->eventq_poll[channel->channel] = 1;
+	return 0;
+}
+
+/**************************************************************************
+ *
+ * PHY testing
+ *
+ **************************************************************************/
+
+/*
+ * Check PHY presence
+ * This reads the PHY ID registers via GMII and checks that neither
+ * are all-zeroes or all-ones (indicating a nonexistent or
+ * uncommunicative PHY).
+ */
+static int efx_test_phy(struct efx_nic *efx,
+			struct efx_self_tests *tests)
+{
+	u16 physid1, physid2;
+	struct mii_if_info *mii = &efx->mii;
+	struct net_device *net_dev = efx->net_dev;
+
+	if (efx->phy_type == PHY_TYPE_NONE)
+		return 0;
+
+	EFX_LOG(efx, "testing PHY presence\n");
+	tests->phy_ok = -1;
+
+	physid1 = mii->mdio_read(net_dev, mii->phy_id, MII_PHYSID1);
+	physid2 = mii->mdio_read(net_dev, mii->phy_id, MII_PHYSID2);
+
+	if ((physid1 != 0x0000) && (physid1 != 0xffff) &&
+	    (physid2 != 0x0000) && (physid2 != 0xffff)) {
+		EFX_LOG(efx, "found MII PHY %d ID 0x%x:%x\n",
+			mii->phy_id, physid1, physid2);
+		tests->phy_ok = 1;
+		return 0;
+	}
+
+	EFX_ERR(efx, "no MII PHY present with ID %d\n", mii->phy_id);
+	return -ENODEV;
+}
+
+/**************************************************************************
+ *
+ * Loopback testing
+ * NB Only one loopback test can be executing concurrently.
+ *
+ **************************************************************************/
+
+/* Loopback test RX callback
+ * This is called for each received packet during loopback testing.
+ */
+void efx_loopback_rx_packet(struct efx_nic *efx,
+			    const char *buf_ptr, int pkt_len)
+{
+	struct efx_selftest_state *state = efx->loopback_selftest;
+	struct efx_loopback_payload *received;
+	struct efx_loopback_payload *payload;
+
+	BUG_ON(!buf_ptr);
+
+	/* If we are just flushing, then drop the packet */
+	if ((state == NULL) || state->flush)
+		return;
+
+	payload = &state->payload;
+
+	/* The packet should have been passed up to us before any LRO/SSR, so
+	 * we should be able to compare the data directly. Since the packet
+	 * is going to be thrown away by the caller, modify it in place.
+	 * efx_test_loopback guarantees to not touch state->payload during
+	 * the test */
+	received = (struct efx_loopback_payload *)(char *)buf_ptr;
+	received->ip.saddr = payload->ip.saddr;
+	received->ip.check = payload->ip.check;
+
+	/* Check that header exists */
+	if (pkt_len < sizeof(received->header)) {
+		EFX_ERR(efx, "saw runt RX packet (length %d) "
+			"in %s loopback test\n", pkt_len,
+			LOOPBACK_MODE(efx));
+		goto err;
+	}
+
+	/* Check that header matches */
+	if (memcmp(&received->header, &payload->header, ETH_HLEN) != 0) {
+		EFX_ERR(efx, "saw non-loopback RX packet in"
+			" %s loopback test\n",
+			LOOPBACK_MODE(efx));
+		goto err;
+	}
+
+	/* Check packet length */
+	if (pkt_len != sizeof(*payload)) {
+		EFX_ERR(efx, "saw incorrect RX packet length"
+			" %d (wanted %d) in %s loopback test\n",
+			pkt_len, (int)sizeof(*payload),
+			LOOPBACK_MODE(efx));
+		goto err;
+	}
+
+	/* Check that IP header matches */
+	if (memcmp(&received->ip, &payload->ip, sizeof(payload->ip)) != 0) {
+		EFX_ERR(efx, "saw corrupted IP header in %s "
+			"loopback test\n",
+			LOOPBACK_MODE(efx));
+		goto err;
+	}
+
+	/* Check that msg and padding matches */
+	if (memcmp(&received->msg, &payload->msg, sizeof(received->msg)) != 0) {
+		EFX_ERR(efx, "saw corrupted RX packet in %s "
+			"loopback test\n",
+			LOOPBACK_MODE(efx));
+		goto err;
+	}
+
+	/* Check that iteration matches */
+	if (received->iteration != payload->iteration) {
+		EFX_ERR(efx, "saw RX packet from iteration %d"
+			" (wanted %d) in %s loopback test\n",
+			ntohs(received->iteration), ntohs(payload->iteration),
+			LOOPBACK_MODE(efx));
+		goto err;
+	}
+
+	/* Increase correct RX count */
+	EFX_TRACE(efx, "got loopback RX in %s loopback test\n",
+		  LOOPBACK_MODE(efx));
+
+	atomic_inc(&state->rx_good);
+	return;
+
+ err:
+#ifdef EFX_ENABLE_DEBUG
+	if (atomic_read(&state->rx_bad) == 0) {
+		EFX_ERR(efx, "received packet:\n");
+		print_hex_dump(KERN_ERR, "", DUMP_PREFIX_OFFSET, 0x10, 1,
+			       buf_ptr, pkt_len, 0);
+		EFX_ERR(efx, "expected packet:\n");
+		print_hex_dump(KERN_ERR, "", DUMP_PREFIX_OFFSET, 0x10, 1,
+			       &state->payload, sizeof(state->payload), 0);
+	}
+#endif
+	atomic_inc(&state->rx_bad);
+}
+
+/* Initialise an efx_selftest_state for a new iteration */
+static void efx_iterate_state(struct efx_nic *efx)
+{
+	struct efx_selftest_state *state = efx->loopback_selftest;
+	struct net_device *net_dev = efx->net_dev;
+	struct efx_loopback_payload *payload = &state->payload;
+
+	/* Initialise the layerII header */
+	memcpy(&payload->header.h_dest, net_dev->dev_addr, ETH_ALEN);
+	memcpy(&payload->header.h_source, &payload_source, ETH_ALEN);
+	payload->header.h_proto = htons(ETH_P_IP);
+
+	/* saddr set later and used as incrementing count */
+	payload->ip.daddr = htonl(INADDR_LOOPBACK);
+	payload->ip.ihl = 5;
+	payload->ip.check = 0;		/* offloaded */
+	payload->ip.tot_len = htons(sizeof(*payload) - sizeof(struct ethhdr));
+	payload->ip.version = IPVERSION;
+	payload->ip.protocol = IPPROTO_UDP;
+
+	/* Initialise udp header */
+	payload->udp.source = 0;
+	payload->udp.len = htons(sizeof(*payload) - sizeof(struct ethhdr) -
+				 sizeof(struct iphdr));
+	payload->udp.check = 0;	/* checksum ignored */
+
+	/* Fill out payload */
+	payload->iteration = htons(ntohs(payload->iteration) + 1);
+	memcpy(&payload->msg, payload_msg, sizeof(payload_msg));
+
+	/* Fill out remaining state members */
+	atomic_set(&state->rx_good, 0);
+	atomic_set(&state->rx_bad, 0);
+	smp_wmb();
+}
+
+/*
+ * Perform loopback test with N packets
+ *
+ * This will transmit "num_packets" copies of a test packet, and check
+ * that they were both transmitted (i.e. a TX completion event was
+ * received) and received (i.e. the data arrived intact via loopback).
+ * The port must have already been placed into the desired loopback
+ * mode.
+ */
+static int efx_test_loopback(struct efx_nic *efx,
+			     struct efx_tx_queue *tx_queue,
+			     struct efx_loopback_self_tests *lb_tests)
+{
+#if !defined(EFX_HAVE_OLD_NAPI)
+	struct efx_channel *channel;
+#endif
+	struct efx_selftest_state *state = efx->loopback_selftest;
+	struct efx_loopback_payload *payload;
+	struct sk_buff *skb;
+	int rc = 0, i, tx_done, rx_good, rx_bad;
+
+	/* Fill out the packet contents */
+	efx_iterate_state(efx);
+
+	/* Create and fill skb */
+	skb = alloc_skb(sizeof(state->payload), GFP_KERNEL);
+	if (!skb) {
+		rc = -ENOMEM;
+		goto out1;
+	}
+	payload = ((struct efx_loopback_payload *)
+		   skb_put(skb, sizeof(state->payload)));
+	memcpy(payload, &state->payload, sizeof(state->payload));
+
+	/* Transmit N copies of buffer */
+	for (i = 0; i < state->packet_count; i++) {
+		/* Set the source address in the copy of the packet.
+		 * Incrementing the source address on a per-packet basis
+		 * should ensure that we stress all RSS vectors */
+		payload->ip.saddr = htonl(INADDR_LOOPBACK | (i << 2));
+		skb_get(skb);
+
+		if (efx_xmit(efx, tx_queue, skb) != NETDEV_TX_OK) {
+			EFX_ERR(efx, "TX queue %d could not transmit "
+				"packet %d of %d in %s loopback test\n",
+				tx_queue->queue, i + 1,
+				state->packet_count, LOOPBACK_MODE(efx));
+			rc = -EPIPE;
+			goto out2;
+		}
+
+		/* Avoid hogging the PCI bus */
+		udelay(10);
+	}
+
+#if !defined(EFX_HAVE_OLD_NAPI)
+	/* NAPI polling is not enabled, so process channels synchronously */
+	schedule_timeout_uninterruptible(HZ / 50);
+	efx_for_each_channel_with_interrupt(channel, efx) {
+		if (channel->work_pending)
+			efx_process_channel_now(channel);
+	}
+#else
+	/* Allow time for processing */
+	schedule_timeout_uninterruptible(HZ / 10);
+#endif
+
+	if (state->flush)
+		goto out3;
+
+	/* Check TX completion and received packet counts */
+	tx_done = state->packet_count - (atomic_read(&skb->users) - 1);
+	rx_good = atomic_read(&state->rx_good);
+	rx_bad = atomic_read(&state->rx_bad);
+	if (tx_done != state->packet_count) {
+		/* Don't free the skbs; they will be picked up on TX
+		 * overflow or channel teardown.
+		 */
+		EFX_ERR(efx, "TX queue %d saw only %d out of an "
+			"expected %d TX completion events in %s loopback "
+			"test\n", tx_queue->queue, tx_done,
+			state->packet_count, LOOPBACK_MODE(efx));
+		rc = -ETIMEDOUT;
+		/* Allow to fall through so we see the RX errors as well */
+	}
+
+	/* We may always be up to a flush away from our desired packet total */
+	if (rx_good != state->packet_count) {
+		EFX_LOG(efx, "TX queue %d saw only %d out of an "
+			"expected %d received packets in %s loopback "
+			"test\n", tx_queue->queue, rx_good,
+			state->packet_count, LOOPBACK_MODE(efx));
+		rc = -ETIMEDOUT;
+		/* Fall through */
+	}
+
+	/* Update loopback test structure */
+	lb_tests->tx_sent[tx_queue->queue] += state->packet_count;
+	lb_tests->tx_done[tx_queue->queue] += tx_done;
+	lb_tests->rx_good += rx_good;
+	lb_tests->rx_bad += rx_bad;
+
+ out3:
+ out2:
+	/* Free skb */
+	if (skb) {
+		/* If the selftest failed then the skb needs its reference
+		 * count decreasing */
+		while (skb_shared(skb))
+			kfree_skb(skb);
+		dev_kfree_skb_any(skb);
+	}
+
+ out1:
+	return rc;
+}
+
+/* Perform loopback test safely
+ *
+ * This performs a safe loopback test by starting with a single packet
+ * and only increasing the number of packets while the tests are
+ * passing.  This avoids flooding a network with garbage packets if
+ * e.g. setting the loopback mode fails.
+ */
+static int
+efx_test_loopback_safely(struct efx_nic *efx,
+			 struct efx_tx_queue *tx_queue,
+			 struct efx_loopback_self_tests *lb_tests)
+{
+	struct efx_selftest_state *state = efx->loopback_selftest;
+	int i, rc = 0;
+
+	for (i = 0; i < loopback_test_level; i++) {
+		/* Determine how many packets to send */
+		state->packet_count = (efx->type->txd_ring_mask + 1) / 3;
+		state->packet_count = min(1 << (i << 2), state->packet_count);
+		state->flush = 0;
+
+		EFX_LOG(efx, "TX queue %d testing %s loopback"
+			" with %d packets\n", tx_queue->queue,
+			LOOPBACK_MODE(efx), state->packet_count);
+
+		rc = efx_test_loopback(efx, tx_queue, lb_tests);
+		if (rc) {
+			/* Wait a while to ensure there are no packets
+			 * floating around after a failure.
+			 */
+			schedule_timeout_uninterruptible(HZ / 5);
+			return rc;
+		}
+	}
+
+	EFX_LOG(efx, "TX queue %d passed %s loopback test "
+		"with a burst length of %d packets\n",
+		tx_queue->queue, LOOPBACK_MODE(efx), state->packet_count);
+
+	return rc;
+}
+
+static int efx_test_loopbacks(struct efx_nic *efx,
+			      struct efx_self_tests *tests,
+			      unsigned int loopback_modes)
+{
+	struct efx_selftest_state *state = efx->loopback_selftest;
+	struct ethtool_cmd ecmd, ecmd_loopback;
+	struct efx_tx_queue *tx_queue;
+	enum efx_loopback_mode old_mode, mode;
+	int old_powered, count, rc = 0;
+	int retry = EFX_WORKAROUND_8909(efx);
+
+	/* Get current PHY settings */
+	rc = efx_ethtool_get_settings(efx->net_dev, &ecmd);
+	if (rc) {
+		EFX_ERR(efx, "could not get GMII settings\n");
+		return rc;
+	}
+	old_mode = efx->loopback_mode;
+	old_powered = efx->phy_powered;
+
+	/* Disable autonegotiation for the purposes of loopback */
+	memcpy(&ecmd_loopback, &ecmd, sizeof(ecmd_loopback));
+	if (ecmd_loopback.autoneg == AUTONEG_ENABLE) {
+		ecmd_loopback.autoneg = AUTONEG_DISABLE;
+		ecmd_loopback.duplex = DUPLEX_FULL;
+		ecmd_loopback.speed = EFX_IS10G(efx) ?
+			SPEED_10000 : SPEED_1000;
+	}
+
+	rc = efx_ethtool_set_settings(efx->net_dev, &ecmd_loopback);
+	if (rc) {
+		EFX_ERR(efx, "could not disable autonegotiation\n");
+		goto out;
+	}
+	tests->loopback_speed = ecmd_loopback.speed;
+	tests->loopback_full_duplex = ecmd_loopback.duplex;
+
+	/* Test all supported loopback modes */
+	for (mode = LOOPBACK_NONE; mode < LOOPBACK_TEST_MAX; mode++) {
+		if (!(loopback_modes & (1 << mode)))
+			continue;
+
+		/* Move the port into the specified loopback mode. */
+		state->flush = 1;
+		efx->phy_powered = 1;
+		efx->loopback_mode = mode;
+		efx_reconfigure_port(efx, 0);
+
+		/* Wait for the PHY to signal the link is up */
+		count = 0;
+		do {
+			struct efx_channel *channel = &efx->channel[0];
+
+			(void) efx->mac_op->check_hw(efx);
+			schedule_timeout_uninterruptible(HZ / 10);
+			if (channel->work_pending)
+				efx_process_channel_now(channel);
+			/* Wait for PHY events to be processed */
+			flush_workqueue(efx->workqueue);
+			rmb();
+		} while ((++count < 20) && !efx->link_up);
+
+		/* The link should now be up. If it isn't, there is no point
+		 * in attempting a loopback test */
+		if (!efx->link_up) {
+			EFX_ERR(efx, "loopback %s never came up\n",
+				LOOPBACK_MODE(efx));
+			rc = -EIO;
+			goto out;
+		}
+
+		EFX_LOG(efx, "link came up in %s loopback in %d iterations\n",
+			LOOPBACK_MODE(efx), count);
+
+		/* Test every TX queue */
+		efx_for_each_tx_queue(tx_queue, efx) {
+			rc |= efx_test_loopback_safely(efx, tx_queue,
+						       &tests->loopback[mode]);
+			if (rc)
+				goto fail;
+		}
+
+		continue;
+
+fail:
+		if (retry) {
+			/* Give the PHY a kick by pretending to move into
+			 * a Falcon internal loopback mode, then back out */
+			int first = ffs(efx->loopback_modes) - 1;
+
+			EFX_INFO(efx, "retrying %s loopback\n",
+				 LOOPBACK_MODE(efx));
+
+			state->flush = 1;
+			efx->loopback_mode = first;
+			efx_reconfigure_port(efx, 0);
+
+			retry = rc = 0;
+			--mode;
+			continue;
+		}
+		break;
+	}
+
+ out:
+	/* Take out of loopback and restore PHY settings */
+	state->flush = 1;
+	efx->loopback_mode = old_mode;
+	efx->phy_powered = old_powered;
+	/* Push the loopback change, and restore any other
+	 * settings we may have trodden on */
+	(void) efx_ethtool_set_settings(efx->net_dev, &ecmd);
+
+	return rc;
+}
+
+/**************************************************************************
+ *
+ * Entry points
+ *
+ *************************************************************************/
+
+/* Online (i.e. non-disruptive) testing
+ *
+ * This checks interrupt generation, event delivery and PHY presence.
+ * The caller should hold the suspend lock
+ *
+ */
+int efx_online_test(struct efx_nic *efx, struct efx_self_tests *tests)
+{
+	struct efx_channel *channel;
+	int rc = 0;
+
+	ASSERT_RTNL();
+
+	EFX_LOG(efx, "performing online self-tests\n");
+
+	rc |= efx_test_interrupts(efx, tests);
+	efx_for_each_channel(channel, efx) {
+		if (channel->has_interrupt)
+			rc |= efx_test_eventq_irq(channel, tests);
+		else
+			rc |= efx_test_eventq(channel, tests);
+	}
+	rc |= efx_test_phy(efx, tests);
+
+	if (rc)
+		EFX_ERR(efx, "failed online self-tests\n");
+
+	return rc;
+}
+
+/* Offline (i.e. disruptive) testing
+ * This checks MAC and PHY loopback on the specified port. The caller
+ * should hold the rtnl lock
+ */
+int efx_offline_test(struct efx_nic *efx,
+		     struct efx_self_tests *tests, unsigned int loopback_modes)
+{
+	struct efx_selftest_state *state;
+	int rc = 0;
+
+	ASSERT_RTNL();
+
+	EFX_LOG(efx, "performing offline self-tests\n");
+
+	/* Create a selftest_state structure to hold state for the test */
+	state = kzalloc(sizeof(*state), GFP_KERNEL);
+	if (state == NULL) {
+		rc = -ENOMEM;
+		goto out;
+	}
+
+	/* Set the port loopback_selftest member. From this point on
+	 * all received packets will be dropped. Mark the state as
+	 * "flushing" so all inflight packets are dropped */
+	BUG_ON(efx->loopback_selftest);
+	state->flush = 1;
+	efx->loopback_selftest = (void *)state;
+	wmb();
+
+	/* Test all loopback  modes */
+	rc = efx_test_loopbacks(efx, tests, loopback_modes);
+
+	/* Tidy up the port test state */
+	efx->loopback_selftest = NULL;
+	wmb();
+	kfree(state);
+
+ out:
+	if (rc)
+		EFX_ERR(efx, "failed offline self-tests\n");
+
+	return rc;
+}
+
diff -rpuN linux-2.6.18.8/drivers/net/sfc/selftest.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/selftest.h
--- linux-2.6.18.8/drivers/net/sfc/selftest.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/selftest.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,67 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_SELFTEST_H
+#define EFX_SELFTEST_H
+
+#include "net_driver.h"
+
+/*
+ * Self tests
+ */
+
+struct efx_loopback_self_tests {
+	int tx_sent[EFX_MAX_TX_QUEUES];
+	int tx_done[EFX_MAX_TX_QUEUES];
+	int rx_good;
+	int rx_bad;
+};
+
+/* Efx self test results
+ * For fields which are not counters, 1 indicates success and -1
+ * indicates failure.
+ */
+struct efx_self_tests {
+	int interrupt;
+	int eventq_dma[EFX_MAX_CHANNELS];
+	int eventq_int[EFX_MAX_CHANNELS];
+	int eventq_poll[EFX_MAX_CHANNELS];
+	int phy_ok;
+	int loopback_speed;
+	int loopback_full_duplex;
+	struct efx_loopback_self_tests loopback[LOOPBACK_TEST_MAX];
+};
+
+extern void efx_loopback_rx_packet(struct efx_nic *efx,
+				   const char *buf_ptr, int pkt_len);
+extern int efx_online_test(struct efx_nic *efx,
+			   struct efx_self_tests *tests);
+extern int efx_offline_test(struct efx_nic *efx,
+			    struct efx_self_tests *tests,
+			    unsigned int loopback_modes);
+
+#endif /* EFX_SELFTEST_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/assert_valid.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/assert_valid.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/assert_valid.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/assert_valid.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,95 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains functions to assert validness of resources and
+ * resource manager in DEBUG build of the resource driver.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/efrm/sysdep.h>
+
+#ifndef NDEBUG
+#include <ci/efrm/resource.h>
+#include <ci/efrm/driver_private.h>
+#include <ci/efrm/debug.h>
+
+void
+efrm_resource_manager_assert_valid(struct efrm_resource_manager *rm,
+				   const char *file, int line)
+{
+	_EFRM_ASSERT(rm, file, line);
+	_EFRM_ASSERT(rm->rm_name, file, line);
+	_EFRM_ASSERT(rm->rm_type < EFRM_RESOURCE_NUM, file, line);
+	_EFRM_ASSERT(rm->rm_table, file, line);
+	_EFRM_ASSERT(rm->rm_table_size > 0, file, line);
+	_EFRM_ASSERT(rm->rm_dtor, file, line);
+}
+EXPORT_SYMBOL(efrm_resource_manager_assert_valid);
+
+/*
+ * \param rs                    resource to validate
+ * \param ref_count_is_zero     One of 3 values
+ *                                > 0  - check ref count is zero
+ *                                = 0  - check ref count is non-zero
+ *                                < 0  - ref count could be any value
+ */
+void
+efrm_resource_assert_valid(struct efrm_resource *rs, int ref_count_is_zero,
+			   const char *file, int line)
+{
+	struct efrm_resource_manager *rm;
+
+	_EFRM_ASSERT(rs, file, line);
+
+	if (ref_count_is_zero >= 0) {
+		if (!(ref_count_is_zero || atomic_read(&rs->rs_ref_count) > 0)
+		    || !(!ref_count_is_zero
+			 || atomic_read(&rs->rs_ref_count) == 0))
+			EFRM_WARN("%s: check %szero ref=%d " EFRM_RESOURCE_FMT,
+				  __FUNCTION__,
+				  ref_count_is_zero == 0 ? "non-" : "",
+				  atomic_read(&rs->rs_ref_count),
+				  EFRM_RESOURCE_PRI_ARG(rs->rs_handle));
+
+		_EFRM_ASSERT(!(ref_count_is_zero == 0) ||
+			     atomic_read(&rs->rs_ref_count) != 0, file, line);
+		_EFRM_ASSERT(!(ref_count_is_zero > 0) ||
+			     atomic_read(&rs->rs_ref_count) == 0, file, line);
+	}
+
+	rm = efrm_rm_table[EFRM_RESOURCE_TYPE(rs->rs_handle)];
+	efrm_resource_manager_assert_valid(rm, file, line);
+}
+EXPORT_SYMBOL(efrm_resource_assert_valid);
+
+#endif
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/buddy.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/buddy.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/buddy.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/buddy.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,307 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains implementation of a buddy allocator.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/efhw/common.h> /* get uintXX types on win32 */
+#include <ci/efrm/sysdep.h>
+#include <ci/efrm/buddy.h>
+#include <ci/efrm/debug.h>
+
+#if 1
+#define DEBUG_ALLOC(x)
+#else
+#define DEBUG_ALLOC(x) x
+
+static inline void efrm_buddy_dump(struct efrm_buddy_allocator *b)
+{
+	unsigned o;
+
+	EFRM_NOTICE("%s: dump allocator with order %u",
+		    __FUNCTION__, b->order);
+	for (o = 0; o <= b->order; o++) {
+		struct list_head *l = &b->free_lists[o];
+		while (l->next != &b->free_lists[o]) {
+			l = l->next;
+			EFRM_NOTICE("%s: order %x: %zx", __FUNCTION__, o,
+				    l - b->links);
+		}
+	}
+}
+#endif
+
+/*
+ * The purpose of the following inline functions is to give the
+ * understandable names to the simple actions.
+ */
+static inline void
+efrm_buddy_free_list_add(struct efrm_buddy_allocator *b,
+			 unsigned order, unsigned addr)
+{
+	list_add(&b->links[addr], &b->free_lists[order]);
+	b->orders[addr] = (uint8_t) b->order;
+}
+static inline void
+efrm_buddy_free_list_del(struct efrm_buddy_allocator *b, unsigned addr)
+{
+	list_del(&b->links[addr]);
+	b->links[addr].next = NULL;
+}
+static inline int
+efrm_buddy_free_list_empty(struct efrm_buddy_allocator *b, unsigned order)
+{
+	return list_empty(&b->free_lists[order]);
+}
+static inline unsigned
+efrm_buddy_free_list_pop(struct efrm_buddy_allocator *b, unsigned order)
+{
+	struct list_head *l = list_pop(&b->free_lists[order]);
+	l->next = NULL;
+	return (unsigned)(l - b->links);
+}
+static inline int
+efrm_buddy_addr_in_free_list(struct efrm_buddy_allocator *b, unsigned addr)
+{
+	return b->links[addr].next != NULL;
+}
+static inline unsigned
+efrm_buddy_free_list_first(struct efrm_buddy_allocator *b, unsigned order)
+{
+	return (unsigned)(b->free_lists[order].next - b->links);
+}
+
+int efrm_buddy_ctor(struct efrm_buddy_allocator *b, unsigned order)
+{
+	unsigned o;
+	unsigned size = 1 << order;
+
+	DEBUG_ALLOC(EFRM_NOTICE("%s(%u)", __FUNCTION__, order));
+	EFRM_ASSERT(b);
+	EFRM_ASSERT(order <= sizeof(unsigned) * 8 - 1);
+
+	b->order = order;
+	b->free_lists = vmalloc((order + 1) * sizeof(struct list_head));
+	if (b->free_lists == NULL)
+		goto fail1;
+
+	b->links = vmalloc(size * sizeof(struct list_head));
+	if (b->links == NULL)
+		goto fail2;
+
+	b->orders = vmalloc(size);
+	if (b->orders == NULL)
+		goto fail3;
+
+	memset(b->links, 0, size * sizeof(struct list_head));
+
+	for (o = 0; o <= b->order; ++o)
+		INIT_LIST_HEAD(b->free_lists + o);
+
+	efrm_buddy_free_list_add(b, b->order, 0);
+
+	return 0;
+
+fail3:
+	vfree(b->links);
+fail2:
+	vfree(b->free_lists);
+fail1:
+	return -ENOMEM;
+}
+
+void efrm_buddy_dtor(struct efrm_buddy_allocator *b)
+{
+	EFRM_ASSERT(b);
+
+	vfree(b->free_lists);
+	vfree(b->links);
+	vfree(b->orders);
+}
+
+int efrm_buddy_alloc(struct efrm_buddy_allocator *b, unsigned order)
+{
+	unsigned smallest;
+	unsigned addr;
+
+	DEBUG_ALLOC(EFRM_NOTICE("%s(%u)", __FUNCTION__, order));
+	EFRM_ASSERT(b);
+
+	/* Find smallest chunk that is big enough.  ?? Can optimise this by
+	 ** keeping array of pointers to smallest chunk for each order.
+	 */
+	smallest = order;
+	while (smallest <= b->order &&
+	       efrm_buddy_free_list_empty(b, smallest))
+		++smallest;
+
+	if (smallest > b->order) {
+		DEBUG_ALLOC(EFRM_NOTICE
+			    ("buddy - alloc order %d failed - max order %d",
+			     order, b->order););
+		return -ENOMEM;
+	}
+
+	/* Split blocks until we get one of the correct size. */
+	addr = efrm_buddy_free_list_pop(b, smallest);
+
+	DEBUG_ALLOC(EFRM_NOTICE("buddy - alloc %x order %d cut from order %d",
+				addr, order, smallest););
+	while (smallest-- > order)
+		efrm_buddy_free_list_add(b, smallest, addr + (1 << smallest));
+
+	EFRM_DO_DEBUG(b->orders[addr] = (uint8_t) order);
+
+	EFRM_ASSERT(addr < 1u << b->order);
+	return addr;
+}
+
+void
+efrm_buddy_free(struct efrm_buddy_allocator *b, unsigned addr,
+		unsigned order)
+{
+	unsigned buddy_addr;
+
+	DEBUG_ALLOC(EFRM_NOTICE("%s(%u, %u)", __FUNCTION__, addr, order));
+	EFRM_ASSERT(b);
+	EFRM_ASSERT(order <= b->order);
+	EFRM_ASSERT((unsigned long)addr + ((unsigned long)1 << order) <=
+		    (unsigned long)1 << b->order);
+	EFRM_ASSERT(!efrm_buddy_addr_in_free_list(b, addr));
+	EFRM_ASSERT(b->orders[addr] == order);
+
+	/* merge free blocks */
+	while (order < b->order) {
+		buddy_addr = addr ^ (1 << order);
+		if (!efrm_buddy_addr_in_free_list(b, buddy_addr) ||
+		    b->orders[buddy_addr] != order)
+			break;
+		efrm_buddy_free_list_del(b, addr);
+		if (buddy_addr < addr)
+			addr = buddy_addr;
+		++order;
+	}
+
+	DEBUG_ALLOC(EFRM_NOTICE
+		    ("buddy - free %x merged into order %d", addr, order););
+	efrm_buddy_free_list_add(b, order, addr);
+}
+
+void efrm_buddy_reserve_at_start(struct efrm_buddy_allocator *b, unsigned n)
+{
+	int addr;
+	unsigned o;
+	EFRM_DO_DEBUG(int n_save = n);
+
+	DEBUG_ALLOC(EFRM_NOTICE("%s(%u)", __FUNCTION__, n));
+	EFRM_ASSERT(b);
+	EFRM_ASSERT(n <= 1u << b->order && n > 0);
+	/* Whole space must be free. */
+	EFRM_ASSERT(!efrm_buddy_free_list_empty(b, b->order));
+
+	o = fls(n);
+
+	while (n) {
+		while (((unsigned)1 << o) > n)
+			--o;
+		EFRM_ASSERT(((unsigned)1 << o) <= n);
+		addr = efrm_buddy_alloc(b, o);
+		EFRM_ASSERT(addr + (1 << o) <= n_save);
+		n -= 1 << o;
+	}
+}
+
+static int
+__efrm_buddy_reserve_at_end(struct efrm_buddy_allocator *b, unsigned order,
+			    int threshold)
+{
+	unsigned o, addr;
+
+	DEBUG_ALLOC(EFRM_NOTICE("%s(%u, %d)", __FUNCTION__, order, threshold));
+	EFRM_ASSERT(b);
+
+	/* Find largest block; there must be one big enough (or caller has
+	 ** goofed).
+	 */
+	for (o = b->order;; --o) {
+		if (efrm_buddy_free_list_empty(b, o))
+			continue;
+		addr = efrm_buddy_free_list_first(b, o);
+		if (addr + (1 << o) <= (unsigned)threshold)
+			continue;
+		break;
+	}
+	EFRM_ASSERT(o >= order);
+
+	/* Split down (keeping second half) until we reach
+	 * the requested size. */
+	addr = efrm_buddy_free_list_pop(b, o);
+
+	while (o-- > order) {
+		efrm_buddy_free_list_add(b, o, addr);
+		addr += 1 << o;
+	}
+
+	EFRM_DO_DEBUG(b->orders[addr] = (uint8_t) order);
+
+	return addr;
+}
+
+void efrm_buddy_reserve_at_end(struct efrm_buddy_allocator *b, unsigned n)
+{
+	int addr, threshold;
+	unsigned o;
+	EFRM_DO_DEBUG(int n_save = n);
+
+	DEBUG_ALLOC(EFRM_NOTICE("%s(%u)", __FUNCTION__, n));
+	DEBUG_ALLOC(efrm_buddy_dump(b));
+	EFRM_ASSERT(b);
+	EFRM_ASSERT(n <= 1u << b->order);
+
+	if (!n)
+	  return;
+
+	threshold = (1 << b->order) - n;
+	o = fls(n);
+
+	while (n) {
+		while (((unsigned)1 << o) > n)
+			--o;
+		EFRM_ASSERT(((unsigned)1 << o) <= n);
+		addr = __efrm_buddy_reserve_at_end(b, o, threshold);
+		EFRM_ASSERT(addr >= (1 << b->order) - n_save);
+		n -= 1 << o;
+	}
+	DEBUG_ALLOC(efrm_buddy_dump(b));
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/buffer_table.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/buffer_table.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/buffer_table.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/buffer_table.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,210 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains abstraction of the buffer table on the NIC.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*
+** Might be worth keeping a bitmap of which entries are clear.  Then we
+** wouldn't need to clear them all again when we free an allocation.
+*/
+
+#include <ci/efrm/debug.h>
+#include <ci/driver/efab/hardware.h>
+#include <ci/efrm/nic_table.h>
+#include <ci/efrm/buffer_table.h>
+#include <ci/efrm/buddy.h>
+
+/*! Comment? */
+struct efrm_buffer_table {
+	spinlock_t lock;
+	struct efrm_buddy_allocator buddy;
+};
+
+/* Efab buffer state. */
+static struct efrm_buffer_table efrm_buffers;
+
+int efrm_buffer_table_ctor(unsigned low, unsigned high)
+{
+	int log2_n_entries, rc;
+
+	EFRM_ASSERT(high > 0);
+	EFRM_ASSERT(low < high);
+
+	EFRM_TRACE("efrm_buffer_table_ctor: low=%u high=%u", low, high);
+	EFRM_NOTICE("efrm_buffer_table_ctor: low=%u high=%u", low, high);
+
+	log2_n_entries = fls(high - 1);
+
+	rc = efrm_buddy_ctor(&efrm_buffers.buddy, log2_n_entries);
+	if (rc < 0) {
+		EFRM_ERR("efrm_buffer_table_ctor: efrm_buddy_ctor(%d) "
+			 "failed (%d)", log2_n_entries, rc);
+		return rc;
+	}
+
+	spin_lock_init(&efrm_buffers.lock);
+
+	efrm_buddy_reserve_at_start(&efrm_buffers.buddy, low);
+	efrm_buddy_reserve_at_end(&efrm_buffers.buddy,
+				  (1 << log2_n_entries) - high);
+
+	EFRM_TRACE("efrm_buffer_table_ctor: done");
+
+	return 0;
+}
+
+void efrm_buffer_table_dtor(void)
+{
+	/* ?? debug check that all allocations have been freed? */
+
+	spin_lock_destroy(&efrm_buffers.lock);
+	efrm_buddy_dtor(&efrm_buffers.buddy);
+
+	EFRM_TRACE("efrm_buffer_table_dtor: done");
+}
+
+/**********************************************************************/
+
+int
+efrm_buffer_table_alloc(unsigned order,
+			struct efhw_buffer_table_allocation *a)
+{
+	irq_flags_t lock_flags;
+	int rc;
+
+	EFRM_ASSERT(&efrm_buffers.buddy);
+	EFRM_ASSERT(a);
+
+	/* Round up to multiple of two, as the buffer clear logic works in
+	 * pairs when not in "full" mode. */
+	order = max_t(unsigned, order, 1);
+
+	spin_lock_irqsave(&efrm_buffers.lock, lock_flags);
+	rc = efrm_buddy_alloc(&efrm_buffers.buddy, order);
+	spin_unlock_irqrestore(&efrm_buffers.lock, lock_flags);
+
+	if (rc < 0) {
+		EFRM_ERR("efrm_buffer_table_alloc: failed (n=%ld) rc %d",
+			 1ul << order, rc);
+		return rc;
+	}
+
+	EFRM_TRACE("efrm_buffer_table_alloc: base=%d n=%ld",
+		   rc, 1ul << order);
+	a->order = order;
+	a->base = (unsigned)rc;
+	return 0;
+}
+
+void efrm_buffer_table_free(struct efhw_buffer_table_allocation *a)
+{
+	irq_flags_t lock_flags;
+	struct efhw_nic *nic;
+	int nic_i;
+
+	EFRM_ASSERT(&efrm_buffers.buddy);
+	EFRM_ASSERT(a);
+	EFRM_ASSERT(a->base != -1);
+	EFRM_ASSERT((unsigned long)a->base + (1ul << a->order) <=
+		    efrm_buddy_size(&efrm_buffers.buddy));
+
+	EFRM_TRACE("efrm_buffer_table_free: base=%d n=%ld",
+		   a->base, (1ul << a->order));
+
+	EFRM_FOR_EACH_NIC(nic_i, nic)
+	    efhw_nic_buffer_table_clear(nic, a->base, 1ul << a->order);
+
+	spin_lock_irqsave(&efrm_buffers.lock, lock_flags);
+	efrm_buddy_free(&efrm_buffers.buddy, a->base, a->order);
+	spin_unlock_irqrestore(&efrm_buffers.lock, lock_flags);
+
+	EFRM_DO_DEBUG(a->base = a->order = -1);
+}
+
+/**********************************************************************/
+
+void
+efrm_buffer_table_set(struct efhw_buffer_table_allocation *a,
+		      unsigned i, dma_addr_t dma_addr, int owner)
+{
+	struct efhw_nic *nic;
+	int nic_i;
+
+	EFRM_ASSERT(a);
+	EFRM_ASSERT(i < (unsigned)1 << a->order);
+	EFRM_FOR_EACH_NIC(nic_i, nic)
+	    efhw_nic_buffer_table_set(nic, dma_addr, EFHW_NIC_PAGE_SIZE,
+				      0, owner, a->base + i);
+	/* NB. No commit Caller should call efrm_buffer_table_commit. There
+	   are underlying hardware constraints regarding the number of
+	   buffer table entries which can be pushed before commiting. */
+}
+
+unsigned long efrm_buffer_table_size(void)
+{
+	return efrm_buddy_size(&efrm_buffers.buddy);
+}
+
+/**********************************************************************/
+
+int
+efrm_page_register(dma_addr_t dma_addr, int owner,
+		   efhw_buffer_addr_t *buf_addr_out)
+{
+	struct efhw_buffer_table_allocation alloc;
+	int rc;
+
+	rc = efrm_buffer_table_alloc(0, &alloc);
+	if (rc == 0) {
+		efrm_buffer_table_set(&alloc, 0, dma_addr, owner);
+		efrm_buffer_table_commit();
+		*buf_addr_out = EFHW_BUFFER_ADDR(alloc.base, 0);
+	}
+	return rc;
+}
+EXPORT_SYMBOL(efrm_page_register);
+
+void efrm_page_unregister(efhw_buffer_addr_t buf_addr)
+{
+	struct efhw_buffer_table_allocation alloc;
+
+	alloc.order = 0;
+	alloc.base = EFHW_BUFFER_PAGE(buf_addr);
+	efrm_buffer_table_free(&alloc);
+}
+EXPORT_SYMBOL(efrm_page_unregister);
+
+void efrm_buffer_table_commit(void)
+{
+	struct efhw_nic *nic;
+	int nic_i;
+
+	EFRM_FOR_EACH_NIC(nic_i, nic)
+	    efhw_nic_buffer_table_commit(nic);
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/common.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/common.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/common.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/common.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,68 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides EtherFabric NIC hardware interface common
+ * definitions.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_DRIVER_EFAB_HARDWARE_COMMON_H__
+#define __CI_DRIVER_EFAB_HARDWARE_COMMON_H__
+
+/*----------------------------------------------------------------------------
+ *
+ * EtherFabric constants
+ *
+ *---------------------------------------------------------------------------*/
+
+#define EFHW_1K		0x00000400u
+#define EFHW_2K		0x00000800u
+#define EFHW_4K		0x00001000u
+#define EFHW_8K		0x00002000u
+#define EFHW_16K	0x00004000u
+#define EFHW_32K	0x00008000u
+#define EFHW_64K	0x00010000u
+#define EFHW_128K	0x00020000u
+#define EFHW_256K	0x00040000u
+#define EFHW_512K	0x00080000u
+#define EFHW_1M		0x00100000u
+#define EFHW_2M		0x00200000u
+#define EFHW_4M		0x00400000u
+#define EFHW_8M		0x00800000u
+#define EFHW_16M	0x01000000u
+#define EFHW_32M	0x02000000u
+#define EFHW_48M	0x03000000u
+#define EFHW_64M	0x04000000u
+#define EFHW_128M	0x08000000u
+#define EFHW_256M	0x10000000u
+#define EFHW_512M	0x20000000u
+#define EFHW_1G 	0x40000000u
+#define EFHW_2G		0x80000000u
+#define EFHW_4G		0x100000000ULL
+#define EFHW_8G		0x200000000ULL
+
+#endif /* __CI_DRIVER_EFAB_HARDWARE_COMMON_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_core.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_core.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_core.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_core.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1149 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides EtherFabric NIC - EFXXXX (aka Falcon) core register
+ * definitions.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#define  FALCON_EXTENDED_P_BAR 1
+
+/*************---- Bus Interface Unit Registers C Header ----*************/
+#define IOM_IND_ADR_REG_OFST 0x0 /* IO-mapped indirect access address
+				    register */
+  #define IOM_AUTO_ADR_INC_EN_LBN 16
+  #define IOM_AUTO_ADR_INC_EN_WIDTH 1
+  #define IOM_IND_ADR_LBN 0
+  #define IOM_IND_ADR_WIDTH 16
+#define IOM_IND_DAT_REG_OFST 0x4 /* IO-mapped indirect access data register */
+  #define IOM_IND_DAT_LBN 0
+  #define IOM_IND_DAT_WIDTH 32
+#define ADR_REGION_REG_KER_OFST 0x0 /* Address region register */
+#define ADR_REGION_REG_OFST 0x0 /* Address region register */
+  #define ADR_REGION3_LBN 96
+  #define ADR_REGION3_WIDTH 18
+  #define ADR_REGION2_LBN 64
+  #define ADR_REGION2_WIDTH 18
+  #define ADR_REGION1_LBN 32
+  #define ADR_REGION1_WIDTH 18
+  #define ADR_REGION0_LBN 0
+  #define ADR_REGION0_WIDTH 18
+#define INT_EN_REG_KER_OFST 0x10 /* Kernel driver Interrupt enable register */
+  #define KER_INT_CHAR_LBN 4
+  #define KER_INT_CHAR_WIDTH 1
+  #define KER_INT_KER_LBN 3
+  #define KER_INT_KER_WIDTH 1
+  #define ILL_ADR_ERR_INT_EN_KER_LBN 2
+  #define ILL_ADR_ERR_INT_EN_KER_WIDTH 1
+  #define SRM_PERR_INT_EN_KER_LBN 1
+  #define SRM_PERR_INT_EN_KER_WIDTH 1
+  #define DRV_INT_EN_KER_LBN 0
+  #define DRV_INT_EN_KER_WIDTH 1
+#define INT_EN_REG_CHAR_OFST 0x20 /* Char Driver interrupt enable register */
+  #define CHAR_INT_CHAR_LBN 4
+  #define CHAR_INT_CHAR_WIDTH 1
+  #define CHAR_INT_KER_LBN 3
+  #define CHAR_INT_KER_WIDTH 1
+  #define ILL_ADR_ERR_INT_EN_CHAR_LBN 2
+  #define ILL_ADR_ERR_INT_EN_CHAR_WIDTH 1
+  #define SRM_PERR_INT_EN_CHAR_LBN 1
+  #define SRM_PERR_INT_EN_CHAR_WIDTH 1
+  #define DRV_INT_EN_CHAR_LBN 0
+  #define DRV_INT_EN_CHAR_WIDTH 1
+#define INT_ADR_REG_KER_OFST 0x30 /* Interrupt host address for Kernel driver */
+  #define INT_ADR_KER_LBN 0
+  #define INT_ADR_KER_WIDTH 64
+  #define DRV_INT_KER_LBN 32
+  #define DRV_INT_KER_WIDTH 1
+  #define EV_FF_HALF_INT_KER_LBN 3
+  #define EV_FF_HALF_INT_KER_WIDTH 1
+  #define EV_FF_FULL_INT_KER_LBN 2
+  #define EV_FF_FULL_INT_KER_WIDTH 1
+  #define ILL_ADR_ERR_INT_KER_LBN 1
+  #define ILL_ADR_ERR_INT_KER_WIDTH 1
+  #define SRAM_PERR_INT_KER_LBN 0
+  #define SRAM_PERR_INT_KER_WIDTH 1
+#define INT_ADR_REG_CHAR_OFST 0x40 /* Interrupt host address for Char driver */
+  #define INT_ADR_CHAR_LBN 0
+  #define INT_ADR_CHAR_WIDTH 64
+  #define DRV_INT_CHAR_LBN 32
+  #define DRV_INT_CHAR_WIDTH 1
+  #define EV_FF_HALF_INT_CHAR_LBN 3
+  #define EV_FF_HALF_INT_CHAR_WIDTH 1
+  #define EV_FF_FULL_INT_CHAR_LBN 2
+  #define EV_FF_FULL_INT_CHAR_WIDTH 1
+  #define ILL_ADR_ERR_INT_CHAR_LBN 1
+  #define ILL_ADR_ERR_INT_CHAR_WIDTH 1
+  #define SRAM_PERR_INT_CHAR_LBN 0
+  #define SRAM_PERR_INT_CHAR_WIDTH 1
+#define INT_ISR0_B0_OFST 0x90 /* B0 only */
+#define INT_ISR1_B0_OFST 0xA0
+#define INT_ACK_REG_KER_A1_OFST 0x50 /* Kernel interrupt acknowledge register */
+  #define RESERVED_LBN 0
+  #define RESERVED_WIDTH 32
+#define INT_ACK_REG_CHAR_A1_OFST 0x60 /* CHAR interrupt acknowledge register */
+  #define RESERVED_LBN 0
+  #define RESERVED_WIDTH 32
+/*************---- Global CSR Registers C Header ----*************/
+#define STRAP_REG_KER_OFST 0x200 /* ASIC strap status register */
+#define STRAP_REG_OFST 0x200 /* ASIC strap status register */
+  #define ONCHIP_SRAM_LBN 16
+  #define ONCHIP_SRAM_WIDTH 0
+  #define STRAP_ISCSI_EN_LBN 3
+  #define STRAP_ISCSI_EN_WIDTH 1
+  #define STRAP_PINS_LBN 0
+  #define STRAP_PINS_WIDTH 3
+#define GPIO_CTL_REG_KER_OFST 0x210 /* GPIO control register */
+#define GPIO_CTL_REG_OFST 0x210 /* GPIO control register */
+  #define GPIO_OEN_LBN 24
+  #define GPIO_OEN_WIDTH 4
+  #define GPIO_OUT_LBN 16
+  #define GPIO_OUT_WIDTH 4
+  #define GPIO_IN_LBN 8
+  #define GPIO_IN_WIDTH 4
+  #define GPIO_PWRUP_VALUE_LBN 0
+  #define GPIO_PWRUP_VALUE_WIDTH 4
+#define GLB_CTL_REG_KER_OFST 0x220 /* Global control register */
+#define GLB_CTL_REG_OFST 0x220 /* Global control register */
+  #define SWRST_LBN 0
+  #define SWRST_WIDTH 1
+#define FATAL_INTR_REG_KER_OFST 0x230 /* Fatal interrupt register for Kernel */
+  #define PCI_BUSERR_INT_KER_EN_LBN 43
+  #define PCI_BUSERR_INT_KER_EN_WIDTH 1
+  #define SRAM_OOB_INT_KER_EN_LBN 42
+  #define SRAM_OOB_INT_KER_EN_WIDTH 1
+  #define BUFID_OOB_INT_KER_EN_LBN 41
+  #define BUFID_OOB_INT_KER_EN_WIDTH 1
+  #define MEM_PERR_INT_KER_EN_LBN 40
+  #define MEM_PERR_INT_KER_EN_WIDTH 1
+  #define RBUF_OWN_INT_KER_EN_LBN 39
+  #define RBUF_OWN_INT_KER_EN_WIDTH 1
+  #define TBUF_OWN_INT_KER_EN_LBN 38
+  #define TBUF_OWN_INT_KER_EN_WIDTH 1
+  #define RDESCQ_OWN_INT_KER_EN_LBN 37
+  #define RDESCQ_OWN_INT_KER_EN_WIDTH 1
+  #define TDESCQ_OWN_INT_KER_EN_LBN 36
+  #define TDESCQ_OWN_INT_KER_EN_WIDTH 1
+  #define EVQ_OWN_INT_KER_EN_LBN 35
+  #define EVQ_OWN_INT_KER_EN_WIDTH 1
+  #define EVFF_OFLO_INT_KER_EN_LBN 34
+  #define EVFF_OFLO_INT_KER_EN_WIDTH 1
+  #define ILL_ADR_INT_KER_EN_LBN 33
+  #define ILL_ADR_INT_KER_EN_WIDTH 1
+  #define SRM_PERR_INT_KER_EN_LBN 32
+  #define SRM_PERR_INT_KER_EN_WIDTH 1
+  #define PCI_BUSERR_INT_KER_LBN 11
+  #define PCI_BUSERR_INT_KER_WIDTH 1
+  #define SRAM_OOB_INT_KER_LBN 10
+  #define SRAM_OOB_INT_KER_WIDTH 1
+  #define BUFID_OOB_INT_KER_LBN 9
+  #define BUFID_OOB_INT_KER_WIDTH 1
+  #define MEM_PERR_INT_KER_LBN 8
+  #define MEM_PERR_INT_KER_WIDTH 1
+  #define RBUF_OWN_INT_KER_LBN 7
+  #define RBUF_OWN_INT_KER_WIDTH 1
+  #define TBUF_OWN_INT_KER_LBN 6
+  #define TBUF_OWN_INT_KER_WIDTH 1
+  #define RDESCQ_OWN_INT_KER_LBN 5
+  #define RDESCQ_OWN_INT_KER_WIDTH 1
+  #define TDESCQ_OWN_INT_KER_LBN 4
+  #define TDESCQ_OWN_INT_KER_WIDTH 1
+  #define EVQ_OWN_INT_KER_LBN 3
+  #define EVQ_OWN_INT_KER_WIDTH 1
+  #define EVFF_OFLO_INT_KER_LBN 2
+  #define EVFF_OFLO_INT_KER_WIDTH 1
+  #define ILL_ADR_INT_KER_LBN 1
+  #define ILL_ADR_INT_KER_WIDTH 1
+  #define SRM_PERR_INT_KER_LBN 0
+  #define SRM_PERR_INT_KER_WIDTH 1
+#define FATAL_INTR_REG_OFST 0x240 /* Fatal interrupt register for Char */
+  #define PCI_BUSERR_INT_CHAR_EN_LBN 43
+  #define PCI_BUSERR_INT_CHAR_EN_WIDTH 1
+  #define SRAM_OOB_INT_CHAR_EN_LBN 42
+  #define SRAM_OOB_INT_CHAR_EN_WIDTH 1
+  #define BUFID_OOB_INT_CHAR_EN_LBN 41
+  #define BUFID_OOB_INT_CHAR_EN_WIDTH 1
+  #define MEM_PERR_INT_CHAR_EN_LBN 40
+  #define MEM_PERR_INT_CHAR_EN_WIDTH 1
+  #define RBUF_OWN_INT_CHAR_EN_LBN 39
+  #define RBUF_OWN_INT_CHAR_EN_WIDTH 1
+  #define TBUF_OWN_INT_CHAR_EN_LBN 38
+  #define TBUF_OWN_INT_CHAR_EN_WIDTH 1
+  #define RDESCQ_OWN_INT_CHAR_EN_LBN 37
+  #define RDESCQ_OWN_INT_CHAR_EN_WIDTH 1
+  #define TDESCQ_OWN_INT_CHAR_EN_LBN 36
+  #define TDESCQ_OWN_INT_CHAR_EN_WIDTH 1
+  #define EVQ_OWN_INT_CHAR_EN_LBN 35
+  #define EVQ_OWN_INT_CHAR_EN_WIDTH 1
+  #define EVFF_OFLO_INT_CHAR_EN_LBN 34
+  #define EVFF_OFLO_INT_CHAR_EN_WIDTH 1
+  #define ILL_ADR_INT_CHAR_EN_LBN 33
+  #define ILL_ADR_INT_CHAR_EN_WIDTH 1
+  #define SRM_PERR_INT_CHAR_EN_LBN 32
+  #define SRM_PERR_INT_CHAR_EN_WIDTH 1
+  #define FATAL_INTR_REG_EN_BITS    0xffffffffffffffffULL
+  #define PCI_BUSERR_INT_CHAR_LBN 11
+  #define PCI_BUSERR_INT_CHAR_WIDTH 1
+  #define SRAM_OOB_INT_CHAR_LBN 10
+  #define SRAM_OOB_INT_CHAR_WIDTH 1
+  #define BUFID_OOB_INT_CHAR_LBN 9
+  #define BUFID_OOB_INT_CHAR_WIDTH 1
+  #define MEM_PERR_INT_CHAR_LBN 8
+  #define MEM_PERR_INT_CHAR_WIDTH 1
+  #define RBUF_OWN_INT_CHAR_LBN 7
+  #define RBUF_OWN_INT_CHAR_WIDTH 1
+  #define TBUF_OWN_INT_CHAR_LBN 6
+  #define TBUF_OWN_INT_CHAR_WIDTH 1
+  #define RDESCQ_OWN_INT_CHAR_LBN 5
+  #define RDESCQ_OWN_INT_CHAR_WIDTH 1
+  #define TDESCQ_OWN_INT_CHAR_LBN 4
+  #define TDESCQ_OWN_INT_CHAR_WIDTH 1
+  #define EVQ_OWN_INT_CHAR_LBN 3
+  #define EVQ_OWN_INT_CHAR_WIDTH 1
+  #define EVFF_OFLO_INT_CHAR_LBN 2
+  #define EVFF_OFLO_INT_CHAR_WIDTH 1
+  #define ILL_ADR_INT_CHAR_LBN 1
+  #define ILL_ADR_INT_CHAR_WIDTH 1
+  #define SRM_PERR_INT_CHAR_LBN 0
+  #define SRM_PERR_INT_CHAR_WIDTH 1
+#define DP_CTRL_REG_OFST 0x250 /* Datapath control register */
+  #define FLS_EVQ_ID_LBN 0
+  #define FLS_EVQ_ID_WIDTH 12
+#define MEM_STAT_REG_KER_OFST 0x260 /* Memory status register */
+#define MEM_STAT_REG_OFST 0x260 /* Memory status register */
+  #define MEM_PERR_VEC_LBN 53
+  #define MEM_PERR_VEC_WIDTH 38
+  #define MBIST_CORR_LBN 38
+  #define MBIST_CORR_WIDTH 15
+  #define MBIST_ERR_LBN 0
+  #define MBIST_ERR_WIDTH 38
+#define DEBUG_REG_KER_OFST 0x270 /* Debug register */
+#define DEBUG_REG_OFST 0x270 /* Debug register */
+  #define DEBUG_BLK_SEL2_LBN 47
+  #define DEBUG_BLK_SEL2_WIDTH 3
+  #define DEBUG_BLK_SEL1_LBN 44
+  #define DEBUG_BLK_SEL1_WIDTH 3
+  #define DEBUG_BLK_SEL0_LBN 41
+  #define DEBUG_BLK_SEL0_WIDTH 3
+  #define MISC_DEBUG_ADDR_LBN 36
+  #define MISC_DEBUG_ADDR_WIDTH 5
+  #define SERDES_DEBUG_ADDR_LBN 31
+  #define SERDES_DEBUG_ADDR_WIDTH 5
+  #define EM_DEBUG_ADDR_LBN 26
+  #define EM_DEBUG_ADDR_WIDTH 5
+  #define SR_DEBUG_ADDR_LBN 21
+  #define SR_DEBUG_ADDR_WIDTH 5
+  #define EV_DEBUG_ADDR_LBN 16
+  #define EV_DEBUG_ADDR_WIDTH 5
+  #define RX_DEBUG_ADDR_LBN 11
+  #define RX_DEBUG_ADDR_WIDTH 5
+  #define TX_DEBUG_ADDR_LBN 6
+  #define TX_DEBUG_ADDR_WIDTH 5
+  #define BIU_DEBUG_ADDR_LBN 1
+  #define BIU_DEBUG_ADDR_WIDTH 5
+  #define DEBUG_EN_LBN 0
+  #define DEBUG_EN_WIDTH 1
+#define DRIVER_REG0_KER_OFST 0x280 /* Driver scratch register 0 */
+#define DRIVER_REG0_OFST 0x280 /* Driver scratch register 0 */
+  #define DRIVER_DW0_LBN 0
+  #define DRIVER_DW0_WIDTH 32
+#define DRIVER_REG1_KER_OFST 0x290 /* Driver scratch register 1 */
+#define DRIVER_REG1_OFST 0x290 /* Driver scratch register 1 */
+  #define DRIVER_DW1_LBN 0
+  #define DRIVER_DW1_WIDTH 32
+#define DRIVER_REG2_KER_OFST 0x2A0 /* Driver scratch register 2 */
+#define DRIVER_REG2_OFST 0x2A0 /* Driver scratch register 2 */
+  #define DRIVER_DW2_LBN 0
+  #define DRIVER_DW2_WIDTH 32
+#define DRIVER_REG3_KER_OFST 0x2B0 /* Driver scratch register 3 */
+#define DRIVER_REG3_OFST 0x2B0 /* Driver scratch register 3 */
+  #define DRIVER_DW3_LBN 0
+  #define DRIVER_DW3_WIDTH 32
+#define DRIVER_REG4_KER_OFST 0x2C0 /* Driver scratch register 4 */
+#define DRIVER_REG4_OFST 0x2C0 /* Driver scratch register 4 */
+  #define DRIVER_DW4_LBN 0
+  #define DRIVER_DW4_WIDTH 32
+#define DRIVER_REG5_KER_OFST 0x2D0 /* Driver scratch register 5 */
+#define DRIVER_REG5_OFST 0x2D0 /* Driver scratch register 5 */
+  #define DRIVER_DW5_LBN 0
+  #define DRIVER_DW5_WIDTH 32
+#define DRIVER_REG6_KER_OFST 0x2E0 /* Driver scratch register 6 */
+#define DRIVER_REG6_OFST 0x2E0 /* Driver scratch register 6 */
+  #define DRIVER_DW6_LBN 0
+  #define DRIVER_DW6_WIDTH 32
+#define DRIVER_REG7_KER_OFST 0x2F0 /* Driver scratch register 7 */
+#define DRIVER_REG7_OFST 0x2F0 /* Driver scratch register 7 */
+  #define DRIVER_DW7_LBN 0
+  #define DRIVER_DW7_WIDTH 32
+#define ALTERA_BUILD_REG_OFST 0x300 /* Altera build register */
+#define ALTERA_BUILD_REG_OFST 0x300 /* Altera build register */
+  #define ALTERA_BUILD_VER_LBN 0
+  #define ALTERA_BUILD_VER_WIDTH 32
+
+/* so called CSR spare register
+    - contains separate parity enable bits for the various internal memory
+    blocks */
+#define MEM_PARITY_ERR_EN_REG_KER 0x310
+#define MEM_PARITY_ALL_BLOCKS_EN_LBN 64
+#define MEM_PARITY_ALL_BLOCKS_EN_WIDTH 38
+#define MEM_PARITY_TX_DATA_EN_LBN   72
+#define MEM_PARITY_TX_DATA_EN_WIDTH 2
+
+/*************---- Event & Timer Module Registers C Header ----*************/
+
+#if FALCON_EXTENDED_P_BAR
+#define EVQ_RPTR_REG_KER_OFST 0x11B00 /* Event queue read pointer register */
+#else
+#define EVQ_RPTR_REG_KER_OFST 0x1B00 /* Event queue read pointer register */
+#endif
+
+#define EVQ_RPTR_REG_OFST 0xFA0000 /* Event queue read pointer register
+				      array. */
+  #define EVQ_RPTR_LBN 0
+  #define EVQ_RPTR_WIDTH 15
+
+#if FALCON_EXTENDED_P_BAR
+#define EVQ_PTR_TBL_KER_OFST 0x11A00 /* Event queue pointer table for kernel
+					access */
+#else
+#define EVQ_PTR_TBL_KER_OFST 0x1A00 /* Event queue pointer table for kernel
+				       access */
+#endif
+
+#define EVQ_PTR_TBL_CHAR_OFST 0xF60000 /* Event queue pointer table for char
+					  direct access */
+  #define EVQ_WKUP_OR_INT_EN_LBN 39
+  #define EVQ_WKUP_OR_INT_EN_WIDTH 1
+  #define EVQ_NXT_WPTR_LBN 24
+  #define EVQ_NXT_WPTR_WIDTH 15
+  #define EVQ_EN_LBN 23
+  #define EVQ_EN_WIDTH 1
+  #define EVQ_SIZE_LBN 20
+  #define EVQ_SIZE_WIDTH 3
+  #define EVQ_BUF_BASE_ID_LBN 0
+  #define EVQ_BUF_BASE_ID_WIDTH 20
+#define TIMER_CMD_REG_KER_OFST 0x420 /* Timer table for kernel access.
+					Page-mapped */
+#define TIMER_CMD_REG_PAGE4_OFST 0x8420 /* Timer table for user-level access.
+					   Page-mapped. For lowest 1K queues.
+					 */
+#define TIMER_CMD_REG_PAGE123K_OFST 0x1000420 /* Timer table for user-level
+						 access. Page-mapped.
+						 For upper 3K queues. */
+#define TIMER_TBL_OFST 0xF70000 /* Timer table for char driver direct access */
+  #define TIMER_MODE_LBN 12
+  #define TIMER_MODE_WIDTH 2
+  #define TIMER_VAL_LBN 0
+  #define TIMER_VAL_WIDTH 12
+  #define TIMER_MODE_INT_HLDOFF 2
+  #define EVQ_BUF_SIZE_LBN 0
+  #define EVQ_BUF_SIZE_WIDTH 1
+#define DRV_EV_REG_KER_OFST 0x440 /* Driver generated event register */
+#define DRV_EV_REG_OFST 0x440 /* Driver generated event register */
+  #define DRV_EV_QID_LBN 64
+  #define DRV_EV_QID_WIDTH 12
+  #define DRV_EV_DATA_LBN 0
+  #define DRV_EV_DATA_WIDTH 64
+#define EVQ_CTL_REG_KER_OFST 0x450 /* Event queue control register */
+#define EVQ_CTL_REG_OFST 0x450 /* Event queue control register */
+  #define RX_EVQ_WAKEUP_MASK_B0_LBN 15
+  #define RX_EVQ_WAKEUP_MASK_B0_WIDTH 6
+  #define EVQ_OWNERR_CTL_LBN 14
+  #define EVQ_OWNERR_CTL_WIDTH 1
+  #define EVQ_FIFO_AF_TH_LBN 8
+  #define EVQ_FIFO_AF_TH_WIDTH 6
+  #define EVQ_FIFO_NOTAF_TH_LBN 0
+  #define EVQ_FIFO_NOTAF_TH_WIDTH 6
+/*************---- SRAM Module Registers C Header ----*************/
+#define BUF_TBL_CFG_REG_KER_OFST 0x600 /* Buffer table configuration register */
+#define BUF_TBL_CFG_REG_OFST 0x600 /* Buffer table configuration register */
+  #define BUF_TBL_MODE_LBN 3
+  #define BUF_TBL_MODE_WIDTH 1
+#define SRM_RX_DC_CFG_REG_KER_OFST 0x610 /* SRAM receive descriptor cache
+					    configuration register */
+#define SRM_RX_DC_CFG_REG_OFST 0x610 /* SRAM receive descriptor cache
+					configuration register */
+  #define SRM_RX_DC_BASE_ADR_LBN 0
+  #define SRM_RX_DC_BASE_ADR_WIDTH 21
+#define SRM_TX_DC_CFG_REG_KER_OFST 0x620 /* SRAM transmit descriptor cache
+					    configuration register */
+#define SRM_TX_DC_CFG_REG_OFST 0x620 /* SRAM transmit descriptor cache
+					configuration register */
+  #define SRM_TX_DC_BASE_ADR_LBN 0
+  #define SRM_TX_DC_BASE_ADR_WIDTH 21
+#define SRM_CFG_REG_KER_OFST 0x630 /* SRAM configuration register */
+#define SRM_CFG_REG_OFST 0x630 /* SRAM configuration register */
+  #define SRAM_OOB_ADR_INTEN_LBN 5
+  #define SRAM_OOB_ADR_INTEN_WIDTH 1
+  #define SRAM_OOB_BUF_INTEN_LBN 4
+  #define SRAM_OOB_BUF_INTEN_WIDTH 1
+  #define SRAM_BT_INIT_EN_LBN 3
+  #define SRAM_BT_INIT_EN_WIDTH 1
+  #define SRM_NUM_BANK_LBN 2
+  #define SRM_NUM_BANK_WIDTH 1
+  #define SRM_BANK_SIZE_LBN 0
+  #define SRM_BANK_SIZE_WIDTH 2
+#define BUF_TBL_UPD_REG_KER_OFST 0x650 /* Buffer table update register */
+#define BUF_TBL_UPD_REG_OFST 0x650 /* Buffer table update register */
+  #define BUF_UPD_CMD_LBN 63
+  #define BUF_UPD_CMD_WIDTH 1
+  #define BUF_CLR_CMD_LBN 62
+  #define BUF_CLR_CMD_WIDTH 1
+  #define BUF_CLR_END_ID_LBN 32
+  #define BUF_CLR_END_ID_WIDTH 20
+  #define BUF_CLR_START_ID_LBN 0
+  #define BUF_CLR_START_ID_WIDTH 20
+#define SRM_UPD_EVQ_REG_KER_OFST 0x660 /* Buffer table update register */
+#define SRM_UPD_EVQ_REG_OFST 0x660 /* Buffer table update register */
+  #define SRM_UPD_EVQ_ID_LBN 0
+  #define SRM_UPD_EVQ_ID_WIDTH 12
+#define SRAM_PARITY_REG_KER_OFST 0x670 /* SRAM parity register. */
+#define SRAM_PARITY_REG_OFST 0x670 /* SRAM parity register. */
+  #define FORCE_SRAM_PERR_LBN 0
+  #define FORCE_SRAM_PERR_WIDTH 1
+
+#if FALCON_EXTENDED_P_BAR
+#define BUF_HALF_TBL_KER_OFST 0x18000 /* Buffer table in half buffer table
+					 mode direct access by kernel driver */
+#else
+#define BUF_HALF_TBL_KER_OFST 0x8000 /* Buffer table in half buffer table
+					mode direct access by kernel driver */
+#endif
+
+
+#define BUF_HALF_TBL_OFST 0x800000 /* Buffer table in half buffer table mode
+				      direct access by char driver */
+  #define BUF_ADR_HBUF_ODD_LBN 44
+  #define BUF_ADR_HBUF_ODD_WIDTH 20
+  #define BUF_OWNER_ID_HBUF_ODD_LBN 32
+  #define BUF_OWNER_ID_HBUF_ODD_WIDTH 12
+  #define BUF_ADR_HBUF_EVEN_LBN 12
+  #define BUF_ADR_HBUF_EVEN_WIDTH 20
+  #define BUF_OWNER_ID_HBUF_EVEN_LBN 0
+  #define BUF_OWNER_ID_HBUF_EVEN_WIDTH 12
+
+
+#if FALCON_EXTENDED_P_BAR
+#define BUF_FULL_TBL_KER_OFST 0x18000 /* Buffer table in full buffer table
+					 mode direct access by kernel driver */
+#else
+#define BUF_FULL_TBL_KER_OFST 0x8000 /* Buffer table in full buffer table mode
+					direct access by kernel driver */
+#endif
+
+
+
+
+#define BUF_FULL_TBL_OFST 0x800000 /* Buffer table in full buffer table mode
+				      direct access by char driver */
+  #define IP_DAT_BUF_SIZE_LBN 50
+  #define IP_DAT_BUF_SIZE_WIDTH 1
+  #define BUF_ADR_REGION_LBN 48
+  #define BUF_ADR_REGION_WIDTH 2
+  #define BUF_ADR_FBUF_LBN 14
+  #define BUF_ADR_FBUF_WIDTH 34
+  #define BUF_OWNER_ID_FBUF_LBN 0
+  #define BUF_OWNER_ID_FBUF_WIDTH 14
+#define SRM_DBG_REG_OFST 0x3000000 /* SRAM debug access */
+  #define SRM_DBG_LBN 0
+  #define SRM_DBG_WIDTH 64
+/*************---- RX Datapath Registers C Header ----*************/
+
+#define RX_CFG_REG_KER_OFST 0x800 /* Receive configuration register */
+#define RX_CFG_REG_OFST 0x800 /* Receive configuration register */
+
+#if !defined(FALCON_64K_RXFIFO) && !defined(FALCON_PRE_02020029)
+# if !defined(FALCON_128K_RXFIFO)
+#  define FALCON_128K_RXFIFO
+# endif
+#endif
+
+#if defined(FALCON_128K_RXFIFO)
+
+/* new for B0 */
+  #define RX_TOEP_TCP_SUPPRESS_B0_LBN 48
+  #define RX_TOEP_TCP_SUPPRESS_B0_WIDTH 1
+  #define RX_INGR_EN_B0_LBN 47
+  #define RX_INGR_EN_B0_WIDTH 1
+  #define RX_TOEP_IPV4_B0_LBN 46
+  #define RX_TOEP_IPV4_B0_WIDTH 1
+  #define RX_HASH_ALG_B0_LBN 45
+  #define RX_HASH_ALG_B0_WIDTH 1
+  #define RX_HASH_INSERT_HDR_B0_LBN 44
+  #define RX_HASH_INSERT_HDR_B0_WIDTH 1
+/* moved for B0 */
+  #define RX_DESC_PUSH_EN_B0_LBN 43
+  #define RX_DESC_PUSH_EN_B0_WIDTH 1
+  #define RX_RDW_PATCH_EN_LBN 42 /* Non head of line blocking */
+  #define RX_RDW_PATCH_EN_WIDTH 1
+  #define RX_PCI_BURST_SIZE_B0_LBN 39
+  #define RX_PCI_BURST_SIZE_B0_WIDTH 3
+  #define RX_OWNERR_CTL_B0_LBN 38
+  #define RX_OWNERR_CTL_B0_WIDTH 1
+  #define RX_XON_TX_TH_B0_LBN 33
+  #define RX_XON_TX_TH_B0_WIDTH 5
+  #define RX_XOFF_TX_TH_B0_LBN 28
+  #define RX_XOFF_TX_TH_B0_WIDTH 5
+  #define RX_USR_BUF_SIZE_B0_LBN 19
+  #define RX_USR_BUF_SIZE_B0_WIDTH 9
+  #define RX_XON_MAC_TH_B0_LBN 10
+  #define RX_XON_MAC_TH_B0_WIDTH 9
+  #define RX_XOFF_MAC_TH_B0_LBN 1
+  #define RX_XOFF_MAC_TH_B0_WIDTH 9
+  #define RX_XOFF_MAC_EN_B0_LBN 0
+  #define RX_XOFF_MAC_EN_B0_WIDTH 1
+
+#elif !defined(FALCON_PRE_02020029)
+/* new for B0 */
+  #define RX_TOEP_TCP_SUPPRESS_B0_LBN 46
+  #define RX_TOEP_TCP_SUPPRESS_B0_WIDTH 1
+  #define RX_INGR_EN_B0_LBN 45
+  #define RX_INGR_EN_B0_WIDTH 1
+  #define RX_TOEP_IPV4_B0_LBN 44
+  #define RX_TOEP_IPV4_B0_WIDTH 1
+  #define RX_HASH_ALG_B0_LBN 43
+  #define RX_HASH_ALG_B0_WIDTH 41
+  #define RX_HASH_INSERT_HDR_B0_LBN 42
+  #define RX_HASH_INSERT_HDR_B0_WIDTH 1
+/* moved for B0 */
+  #define RX_DESC_PUSH_EN_B0_LBN 41
+  #define RX_DESC_PUSH_EN_B0_WIDTH 1
+  #define RX_PCI_BURST_SIZE_B0_LBN 37
+  #define RX_PCI_BURST_SIZE_B0_WIDTH 3
+  #define RX_OWNERR_CTL_B0_LBN 36
+  #define RX_OWNERR_CTL_B0_WIDTH 1
+  #define RX_XON_TX_TH_B0_LBN 31
+  #define RX_XON_TX_TH_B0_WIDTH 5
+  #define RX_XOFF_TX_TH_B0_LBN 26
+  #define RX_XOFF_TX_TH_B0_WIDTH 5
+  #define RX_USR_BUF_SIZE_B0_LBN 17
+  #define RX_USR_BUF_SIZE_B0_WIDTH 9
+  #define RX_XON_MAC_TH_B0_LBN 9
+  #define RX_XON_MAC_TH_B0_WIDTH 8
+  #define RX_XOFF_MAC_TH_B0_LBN 1
+  #define RX_XOFF_MAC_TH_B0_WIDTH 8
+  #define RX_XOFF_MAC_EN_B0_LBN 0
+  #define RX_XOFF_MAC_EN_B0_WIDTH 1
+
+#else
+/* new for B0 */
+  #define RX_TOEP_TCP_SUPPRESS_B0_LBN 44
+  #define RX_TOEP_TCP_SUPPRESS_B0_WIDTH 1
+  #define RX_INGR_EN_B0_LBN 43
+  #define RX_INGR_EN_B0_WIDTH 1
+  #define RX_TOEP_IPV4_B0_LBN 42
+  #define RX_TOEP_IPV4_B0_WIDTH 1
+  #define RX_HASH_ALG_B0_LBN 41
+  #define RX_HASH_ALG_B0_WIDTH 41
+  #define RX_HASH_INSERT_HDR_B0_LBN 40
+  #define RX_HASH_INSERT_HDR_B0_WIDTH 1
+/* moved for B0 */
+  #define RX_DESC_PUSH_EN_B0_LBN 35
+  #define RX_DESC_PUSH_EN_B0_WIDTH 1
+  #define RX_PCI_BURST_SIZE_B0_LBN 35
+  #define RX_PCI_BURST_SIZE_B0_WIDTH 2
+  #define RX_OWNERR_CTL_B0_LBN 34
+  #define RX_OWNERR_CTL_B0_WIDTH 1
+  #define RX_XON_TX_TH_B0_LBN 29
+  #define RX_XON_TX_TH_B0_WIDTH 5
+  #define RX_XOFF_TX_TH_B0_LBN 24
+  #define RX_XOFF_TX_TH_B0_WIDTH 5
+  #define RX_USR_BUF_SIZE_B0_LBN 15
+  #define RX_USR_BUF_SIZE_B0_WIDTH 9
+  #define RX_XON_MAC_TH_B0_LBN 8
+  #define RX_XON_MAC_TH_B0_WIDTH 7
+  #define RX_XOFF_MAC_TH_B0_LBN 1
+  #define RX_XOFF_MAC_TH_B0_WIDTH 7
+  #define RX_XOFF_MAC_EN_B0_LBN 0
+  #define RX_XOFF_MAC_EN_B0_WIDTH 1
+
+#endif
+
+/* A0/A1 */
+  #define RX_PUSH_EN_A1_LBN 35
+  #define RX_PUSH_EN_A1_WIDTH 1
+  #define RX_PCI_BURST_SIZE_A1_LBN 31
+  #define RX_PCI_BURST_SIZE_A1_WIDTH 3
+  #define RX_OWNERR_CTL_A1_LBN 30
+  #define RX_OWNERR_CTL_A1_WIDTH 1
+  #define RX_XON_TX_TH_A1_LBN 25
+  #define RX_XON_TX_TH_A1_WIDTH 5
+  #define RX_XOFF_TX_TH_A1_LBN 20
+  #define RX_XOFF_TX_TH_A1_WIDTH 5
+  #define RX_USR_BUF_SIZE_A1_LBN 11
+  #define RX_USR_BUF_SIZE_A1_WIDTH 9
+  #define RX_XON_MAC_TH_A1_LBN 6
+  #define RX_XON_MAC_TH_A1_WIDTH 5
+  #define RX_XOFF_MAC_TH_A1_LBN 1
+  #define RX_XOFF_MAC_TH_A1_WIDTH 5
+  #define RX_XOFF_MAC_EN_A1_LBN 0
+  #define RX_XOFF_MAC_EN_A1_WIDTH 1
+
+#define RX_FILTER_CTL_REG_OFST 0x810 /* Receive filter control registers */
+  #define SCATTER_ENBL_NO_MATCH_Q_B0_LBN 40
+  #define SCATTER_ENBL_NO_MATCH_Q_B0_WIDTH 1
+  #define UDP_FULL_SRCH_LIMIT_LBN 32
+  #define UDP_FULL_SRCH_LIMIT_WIDTH 8
+  #define NUM_KER_LBN 24
+  #define NUM_KER_WIDTH 2
+  #define UDP_WILD_SRCH_LIMIT_LBN 16
+  #define UDP_WILD_SRCH_LIMIT_WIDTH 8
+  #define TCP_WILD_SRCH_LIMIT_LBN 8
+  #define TCP_WILD_SRCH_LIMIT_WIDTH 8
+  #define TCP_FULL_SRCH_LIMIT_LBN 0
+  #define TCP_FULL_SRCH_LIMIT_WIDTH 8
+#define RX_FLUSH_DESCQ_REG_KER_OFST 0x820 /* Receive flush descriptor queue
+					     register */
+#define RX_FLUSH_DESCQ_REG_OFST 0x820 /* Receive flush descriptor queue
+					 register */
+  #define RX_FLUSH_DESCQ_CMD_LBN 24
+  #define RX_FLUSH_DESCQ_CMD_WIDTH 1
+  #define RX_FLUSH_EVQ_ID_LBN 12
+  #define RX_FLUSH_EVQ_ID_WIDTH 12
+  #define RX_FLUSH_DESCQ_LBN 0
+  #define RX_FLUSH_DESCQ_WIDTH 12
+#define RX_DESC_UPD_REG_KER_OFST 0x830 /* Kernel  receive descriptor update
+					  register. Page-mapped */
+#define RX_DESC_UPD_REG_PAGE4_OFST 0x8830 /* Char & user receive descriptor
+					     update register. Page-mapped.
+					     For lowest 1K queues. */
+#define RX_DESC_UPD_REG_PAGE123K_OFST 0x1000830 /* Char & user receive
+						   descriptor update register.
+						   Page-mapped. For upper
+						   3K queues. */
+  #define RX_DESC_WPTR_LBN 96
+  #define RX_DESC_WPTR_WIDTH 12
+  #define RX_DESC_PUSH_CMD_LBN 95
+  #define RX_DESC_PUSH_CMD_WIDTH 1
+  #define RX_DESC_LBN 0
+  #define RX_DESC_WIDTH 64
+  #define RX_KER_DESC_LBN 0
+  #define RX_KER_DESC_WIDTH 64
+  #define RX_USR_DESC_LBN 0
+  #define RX_USR_DESC_WIDTH 32
+#define RX_DC_CFG_REG_KER_OFST 0x840 /* Receive descriptor cache
+					configuration register */
+#define RX_DC_CFG_REG_OFST 0x840 /* Receive descriptor cache
+				    configuration register */
+  #define RX_DC_SIZE_LBN 0
+  #define RX_DC_SIZE_WIDTH 2
+#define RX_DC_PF_WM_REG_KER_OFST 0x850 /* Receive descriptor cache pre-fetch
+					  watermark register */
+#define RX_DC_PF_WM_REG_OFST 0x850 /* Receive descriptor cache pre-fetch
+				      watermark register */
+  #define RX_DC_PF_LWM_LO_LBN 0
+  #define RX_DC_PF_LWM_LO_WIDTH 6
+
+#define RX_RSS_TKEY_B0_OFST 0x860 /* RSS Toeplitz hash key (B0 only) */
+
+#define RX_NODESC_DROP_REG 0x880
+  #define RX_NODESC_DROP_CNT_LBN 0
+  #define RX_NODESC_DROP_CNT_WIDTH 16
+
+#define XM_TX_CFG_REG_OFST 0x1230
+  #define XM_AUTO_PAD_LBN 5
+  #define XM_AUTO_PAD_WIDTH 1
+
+#define RX_FILTER_TBL0_OFST 0xF00000 /* Receive filter table - even entries */
+  #define RSS_EN_0_B0_LBN 110
+  #define RSS_EN_0_B0_WIDTH 1
+  #define SCATTER_EN_0_B0_LBN 109
+  #define SCATTER_EN_0_B0_WIDTH 1
+  #define TCP_UDP_0_LBN 108
+  #define TCP_UDP_0_WIDTH 1
+  #define RXQ_ID_0_LBN 96
+  #define RXQ_ID_0_WIDTH 12
+  #define DEST_IP_0_LBN 64
+  #define DEST_IP_0_WIDTH 32
+  #define DEST_PORT_TCP_0_LBN 48
+  #define DEST_PORT_TCP_0_WIDTH 16
+  #define SRC_IP_0_LBN 16
+  #define SRC_IP_0_WIDTH 32
+  #define SRC_TCP_DEST_UDP_0_LBN 0
+  #define SRC_TCP_DEST_UDP_0_WIDTH 16
+#define RX_FILTER_TBL1_OFST 0xF00010 /* Receive filter table - odd entries */
+  #define RSS_EN_1_B0_LBN 110
+  #define RSS_EN_1_B0_WIDTH 1
+  #define SCATTER_EN_1_B0_LBN 109
+  #define SCATTER_EN_1_B0_WIDTH 1
+  #define TCP_UDP_1_LBN 108
+  #define TCP_UDP_1_WIDTH 1
+  #define RXQ_ID_1_LBN 96
+  #define RXQ_ID_1_WIDTH 12
+  #define DEST_IP_1_LBN 64
+  #define DEST_IP_1_WIDTH 32
+  #define DEST_PORT_TCP_1_LBN 48
+  #define DEST_PORT_TCP_1_WIDTH 16
+  #define SRC_IP_1_LBN 16
+  #define SRC_IP_1_WIDTH 32
+  #define SRC_TCP_DEST_UDP_1_LBN 0
+  #define SRC_TCP_DEST_UDP_1_WIDTH 16
+
+#if FALCON_EXTENDED_P_BAR
+#define RX_DESC_PTR_TBL_KER_OFST 0x11800 /* Receive descriptor pointer
+					    kernel access */
+#else
+#define RX_DESC_PTR_TBL_KER_OFST 0x1800 /* Receive descriptor pointer
+					   kernel access */
+#endif
+
+
+#define RX_DESC_PTR_TBL_OFST 0xF40000 /* Receive descriptor pointer table */
+  #define RX_ISCSI_DDIG_EN_LBN 88
+  #define RX_ISCSI_DDIG_EN_WIDTH 1
+  #define RX_ISCSI_HDIG_EN_LBN 87
+  #define RX_ISCSI_HDIG_EN_WIDTH 1
+  #define RX_DESC_PREF_ACT_LBN 86
+  #define RX_DESC_PREF_ACT_WIDTH 1
+  #define RX_DC_HW_RPTR_LBN 80
+  #define RX_DC_HW_RPTR_WIDTH 6
+  #define RX_DESCQ_HW_RPTR_LBN 68
+  #define RX_DESCQ_HW_RPTR_WIDTH 12
+  #define RX_DESCQ_SW_WPTR_LBN 56
+  #define RX_DESCQ_SW_WPTR_WIDTH 12
+  #define RX_DESCQ_BUF_BASE_ID_LBN 36
+  #define RX_DESCQ_BUF_BASE_ID_WIDTH 20
+  #define RX_DESCQ_EVQ_ID_LBN 24
+  #define RX_DESCQ_EVQ_ID_WIDTH 12
+  #define RX_DESCQ_OWNER_ID_LBN 10
+  #define RX_DESCQ_OWNER_ID_WIDTH 14
+  #define RX_DESCQ_LABEL_LBN 5
+  #define RX_DESCQ_LABEL_WIDTH 5
+  #define RX_DESCQ_SIZE_LBN 3
+  #define RX_DESCQ_SIZE_WIDTH 2
+  #define RX_DESCQ_TYPE_LBN 2
+  #define RX_DESCQ_TYPE_WIDTH 1
+  #define RX_DESCQ_JUMBO_LBN 1
+  #define RX_DESCQ_JUMBO_WIDTH 1
+  #define RX_DESCQ_EN_LBN 0
+  #define RX_DESCQ_EN_WIDTH 1
+
+
+#define RX_RSS_INDIR_TBL_B0_OFST 0xFB0000 /* RSS indirection table (B0 only) */
+  #define RX_RSS_INDIR_ENT_B0_LBN 0
+  #define RX_RSS_INDIR_ENT_B0_WIDTH 6
+
+/*************---- TX Datapath Registers C Header ----*************/
+#define TX_FLUSH_DESCQ_REG_KER_OFST 0xA00 /* Transmit flush descriptor
+					     queue register */
+#define TX_FLUSH_DESCQ_REG_OFST 0xA00 /* Transmit flush descriptor queue
+					 register */
+  #define TX_FLUSH_DESCQ_CMD_LBN 12
+  #define TX_FLUSH_DESCQ_CMD_WIDTH 1
+  #define TX_FLUSH_DESCQ_LBN 0
+  #define TX_FLUSH_DESCQ_WIDTH 12
+#define TX_DESC_UPD_REG_KER_OFST 0xA10 /* Kernel transmit descriptor update
+					  register. Page-mapped */
+#define TX_DESC_UPD_REG_PAGE4_OFST 0x8A10 /* Char & user transmit descriptor
+					     update register. Page-mapped */
+#define TX_DESC_UPD_REG_PAGE123K_OFST 0x1000A10 /* Char & user transmit
+						   descriptor update register.
+						   Page-mapped */
+  #define TX_DESC_WPTR_LBN 96
+  #define TX_DESC_WPTR_WIDTH 12
+  #define TX_DESC_PUSH_CMD_LBN 95
+  #define TX_DESC_PUSH_CMD_WIDTH 1
+  #define TX_DESC_LBN 0
+  #define TX_DESC_WIDTH 95
+  #define TX_KER_DESC_LBN 0
+  #define TX_KER_DESC_WIDTH 64
+  #define TX_USR_DESC_LBN 0
+  #define TX_USR_DESC_WIDTH 64
+#define TX_DC_CFG_REG_KER_OFST 0xA20 /* Transmit descriptor cache
+					configuration register */
+#define TX_DC_CFG_REG_OFST 0xA20 /* Transmit descriptor cache configuration
+				    register */
+  #define TX_DC_SIZE_LBN 0
+  #define TX_DC_SIZE_WIDTH 2
+
+#if FALCON_EXTENDED_P_BAR
+#define TX_DESC_PTR_TBL_KER_OFST 0x11900 /* Transmit descriptor pointer. */
+#else
+#define TX_DESC_PTR_TBL_KER_OFST 0x1900 /* Transmit descriptor pointer. */
+#endif
+
+
+#define TX_DESC_PTR_TBL_OFST 0xF50000 /* Transmit descriptor pointer */
+  #define TX_NON_IP_DROP_DIS_B0_LBN 91
+  #define TX_NON_IP_DROP_DIS_B0_WIDTH 1
+  #define TX_IP_CHKSM_DIS_B0_LBN 90
+  #define TX_IP_CHKSM_DIS_B0_WIDTH 1
+  #define TX_TCP_CHKSM_DIS_B0_LBN 89
+  #define TX_TCP_CHKSM_DIS_B0_WIDTH 1
+  #define TX_DESCQ_EN_LBN 88
+  #define TX_DESCQ_EN_WIDTH 1
+  #define TX_ISCSI_DDIG_EN_LBN 87
+  #define TX_ISCSI_DDIG_EN_WIDTH 1
+  #define TX_ISCSI_HDIG_EN_LBN 86
+  #define TX_ISCSI_HDIG_EN_WIDTH 1
+  #define TX_DC_HW_RPTR_LBN 80
+  #define TX_DC_HW_RPTR_WIDTH 6
+  #define TX_DESCQ_HW_RPTR_LBN 68
+  #define TX_DESCQ_HW_RPTR_WIDTH 12
+  #define TX_DESCQ_SW_WPTR_LBN 56
+  #define TX_DESCQ_SW_WPTR_WIDTH 12
+  #define TX_DESCQ_BUF_BASE_ID_LBN 36
+  #define TX_DESCQ_BUF_BASE_ID_WIDTH 20
+  #define TX_DESCQ_EVQ_ID_LBN 24
+  #define TX_DESCQ_EVQ_ID_WIDTH 12
+  #define TX_DESCQ_OWNER_ID_LBN 10
+  #define TX_DESCQ_OWNER_ID_WIDTH 14
+  #define TX_DESCQ_LABEL_LBN 5
+  #define TX_DESCQ_LABEL_WIDTH 5
+  #define TX_DESCQ_SIZE_LBN 3
+  #define TX_DESCQ_SIZE_WIDTH 2
+  #define TX_DESCQ_TYPE_LBN 1
+  #define TX_DESCQ_TYPE_WIDTH 2
+  #define TX_DESCQ_FLUSH_LBN 0
+  #define TX_DESCQ_FLUSH_WIDTH 1
+#define TX_CFG_REG_KER_OFST 0xA50 /* Transmit configuration register */
+#define TX_CFG_REG_OFST 0xA50 /* Transmit configuration register */
+  #define TX_IP_ID_P1_OFS_LBN 32
+  #define TX_IP_ID_P1_OFS_WIDTH 15
+  #define TX_IP_ID_P0_OFS_LBN 16
+  #define TX_IP_ID_P0_OFS_WIDTH 15
+  #define TX_TURBO_EN_LBN 3
+  #define TX_TURBO_EN_WIDTH 1
+  #define TX_OWNERR_CTL_LBN 2
+  #define TX_OWNERR_CTL_WIDTH 2
+  #define TX_NON_IP_DROP_DIS_LBN 1
+  #define TX_NON_IP_DROP_DIS_WIDTH 1
+  #define TX_IP_ID_REP_EN_LBN 0
+  #define TX_IP_ID_REP_EN_WIDTH 1
+#define TX_RESERVED_REG_KER_OFST 0xA80 /* Transmit configuration register */
+#define TX_RESERVED_REG_OFST 0xA80 /* Transmit configuration register */
+  #define TX_CSR_PUSH_EN_LBN 89
+  #define TX_CSR_PUSH_EN_WIDTH 1
+  #define TX_RX_SPACER_LBN 64
+  #define TX_RX_SPACER_WIDTH 8
+  #define TX_SW_EV_EN_LBN 59
+  #define TX_SW_EV_EN_WIDTH 1
+  #define TX_RX_SPACER_EN_LBN 57
+  #define TX_RX_SPACER_EN_WIDTH 1
+  #define TX_CSR_PREF_WD_TMR_LBN 24
+  #define TX_CSR_PREF_WD_TMR_WIDTH 16
+  #define TX_CSR_ONLY1TAG_LBN 21
+  #define TX_CSR_ONLY1TAG_WIDTH 1
+  #define TX_PREF_THRESHOLD_LBN 19
+  #define TX_PREF_THRESHOLD_WIDTH 2
+  #define TX_ONE_PKT_PER_Q_LBN 18
+  #define TX_ONE_PKT_PER_Q_WIDTH 1
+  #define TX_DIS_NON_IP_EV_LBN 17
+  #define TX_DIS_NON_IP_EV_WIDTH 1
+  #define TX_DMA_SPACER_LBN 8
+  #define TX_DMA_SPACER_WIDTH 8
+  #define TX_FLUSH_MIN_LEN_EN_B0_LBN 7
+  #define TX_FLUSH_MIN_LEN_EN_B0_WIDTH 1
+  #define TX_TCP_DIS_A1_LBN 7
+  #define TX_TCP_DIS_A1_WIDTH 1
+  #define TX_IP_DIS_A1_LBN 6
+  #define TX_IP_DIS_A1_WIDTH 1
+  #define TX_MAX_CPL_LBN 2
+  #define TX_MAX_CPL_WIDTH 2
+  #define TX_MAX_PREF_LBN 0
+  #define TX_MAX_PREF_WIDTH 2
+#define TX_VLAN_REG_OFST 0xAE0 /* Transmit VLAN tag register */
+  #define TX_VLAN_EN_LBN 127
+  #define TX_VLAN_EN_WIDTH 1
+  #define TX_VLAN7_PORT1_EN_LBN 125
+  #define TX_VLAN7_PORT1_EN_WIDTH 1
+  #define TX_VLAN7_PORT0_EN_LBN 124
+  #define TX_VLAN7_PORT0_EN_WIDTH 1
+  #define TX_VLAN7_LBN 112
+  #define TX_VLAN7_WIDTH 12
+  #define TX_VLAN6_PORT1_EN_LBN 109
+  #define TX_VLAN6_PORT1_EN_WIDTH 1
+  #define TX_VLAN6_PORT0_EN_LBN 108
+  #define TX_VLAN6_PORT0_EN_WIDTH 1
+  #define TX_VLAN6_LBN 96
+  #define TX_VLAN6_WIDTH 12
+  #define TX_VLAN5_PORT1_EN_LBN 93
+  #define TX_VLAN5_PORT1_EN_WIDTH 1
+  #define TX_VLAN5_PORT0_EN_LBN 92
+  #define TX_VLAN5_PORT0_EN_WIDTH 1
+  #define TX_VLAN5_LBN 80
+  #define TX_VLAN5_WIDTH 12
+  #define TX_VLAN4_PORT1_EN_LBN 77
+  #define TX_VLAN4_PORT1_EN_WIDTH 1
+  #define TX_VLAN4_PORT0_EN_LBN 76
+  #define TX_VLAN4_PORT0_EN_WIDTH 1
+  #define TX_VLAN4_LBN 64
+  #define TX_VLAN4_WIDTH 12
+  #define TX_VLAN3_PORT1_EN_LBN 61
+  #define TX_VLAN3_PORT1_EN_WIDTH 1
+  #define TX_VLAN3_PORT0_EN_LBN 60
+  #define TX_VLAN3_PORT0_EN_WIDTH 1
+  #define TX_VLAN3_LBN 48
+  #define TX_VLAN3_WIDTH 12
+  #define TX_VLAN2_PORT1_EN_LBN 45
+  #define TX_VLAN2_PORT1_EN_WIDTH 1
+  #define TX_VLAN2_PORT0_EN_LBN 44
+  #define TX_VLAN2_PORT0_EN_WIDTH 1
+  #define TX_VLAN2_LBN 32
+  #define TX_VLAN2_WIDTH 12
+  #define TX_VLAN1_PORT1_EN_LBN 29
+  #define TX_VLAN1_PORT1_EN_WIDTH 1
+  #define TX_VLAN1_PORT0_EN_LBN 28
+  #define TX_VLAN1_PORT0_EN_WIDTH 1
+  #define TX_VLAN1_LBN 16
+  #define TX_VLAN1_WIDTH 12
+  #define TX_VLAN0_PORT1_EN_LBN 13
+  #define TX_VLAN0_PORT1_EN_WIDTH 1
+  #define TX_VLAN0_PORT0_EN_LBN 12
+  #define TX_VLAN0_PORT0_EN_WIDTH 1
+  #define TX_VLAN0_LBN 0
+  #define TX_VLAN0_WIDTH 12
+#define TX_FIL_CTL_REG_OFST 0xAF0 /* Transmit filter control register */
+  #define TX_MADR1_FIL_EN_LBN 65
+  #define TX_MADR1_FIL_EN_WIDTH 1
+  #define TX_MADR0_FIL_EN_LBN 64
+  #define TX_MADR0_FIL_EN_WIDTH 1
+  #define TX_IPFIL31_PORT1_EN_LBN 63
+  #define TX_IPFIL31_PORT1_EN_WIDTH 1
+  #define TX_IPFIL31_PORT0_EN_LBN 62
+  #define TX_IPFIL31_PORT0_EN_WIDTH 1
+  #define TX_IPFIL30_PORT1_EN_LBN 61
+  #define TX_IPFIL30_PORT1_EN_WIDTH 1
+  #define TX_IPFIL30_PORT0_EN_LBN 60
+  #define TX_IPFIL30_PORT0_EN_WIDTH 1
+  #define TX_IPFIL29_PORT1_EN_LBN 59
+  #define TX_IPFIL29_PORT1_EN_WIDTH 1
+  #define TX_IPFIL29_PORT0_EN_LBN 58
+  #define TX_IPFIL29_PORT0_EN_WIDTH 1
+  #define TX_IPFIL28_PORT1_EN_LBN 57
+  #define TX_IPFIL28_PORT1_EN_WIDTH 1
+  #define TX_IPFIL28_PORT0_EN_LBN 56
+  #define TX_IPFIL28_PORT0_EN_WIDTH 1
+  #define TX_IPFIL27_PORT1_EN_LBN 55
+  #define TX_IPFIL27_PORT1_EN_WIDTH 1
+  #define TX_IPFIL27_PORT0_EN_LBN 54
+  #define TX_IPFIL27_PORT0_EN_WIDTH 1
+  #define TX_IPFIL26_PORT1_EN_LBN 53
+  #define TX_IPFIL26_PORT1_EN_WIDTH 1
+  #define TX_IPFIL26_PORT0_EN_LBN 52
+  #define TX_IPFIL26_PORT0_EN_WIDTH 1
+  #define TX_IPFIL25_PORT1_EN_LBN 51
+  #define TX_IPFIL25_PORT1_EN_WIDTH 1
+  #define TX_IPFIL25_PORT0_EN_LBN 50
+  #define TX_IPFIL25_PORT0_EN_WIDTH 1
+  #define TX_IPFIL24_PORT1_EN_LBN 49
+  #define TX_IPFIL24_PORT1_EN_WIDTH 1
+  #define TX_IPFIL24_PORT0_EN_LBN 48
+  #define TX_IPFIL24_PORT0_EN_WIDTH 1
+  #define TX_IPFIL23_PORT1_EN_LBN 47
+  #define TX_IPFIL23_PORT1_EN_WIDTH 1
+  #define TX_IPFIL23_PORT0_EN_LBN 46
+  #define TX_IPFIL23_PORT0_EN_WIDTH 1
+  #define TX_IPFIL22_PORT1_EN_LBN 45
+  #define TX_IPFIL22_PORT1_EN_WIDTH 1
+  #define TX_IPFIL22_PORT0_EN_LBN 44
+  #define TX_IPFIL22_PORT0_EN_WIDTH 1
+  #define TX_IPFIL21_PORT1_EN_LBN 43
+  #define TX_IPFIL21_PORT1_EN_WIDTH 1
+  #define TX_IPFIL21_PORT0_EN_LBN 42
+  #define TX_IPFIL21_PORT0_EN_WIDTH 1
+  #define TX_IPFIL20_PORT1_EN_LBN 41
+  #define TX_IPFIL20_PORT1_EN_WIDTH 1
+  #define TX_IPFIL20_PORT0_EN_LBN 40
+  #define TX_IPFIL20_PORT0_EN_WIDTH 1
+  #define TX_IPFIL19_PORT1_EN_LBN 39
+  #define TX_IPFIL19_PORT1_EN_WIDTH 1
+  #define TX_IPFIL19_PORT0_EN_LBN 38
+  #define TX_IPFIL19_PORT0_EN_WIDTH 1
+  #define TX_IPFIL18_PORT1_EN_LBN 37
+  #define TX_IPFIL18_PORT1_EN_WIDTH 1
+  #define TX_IPFIL18_PORT0_EN_LBN 36
+  #define TX_IPFIL18_PORT0_EN_WIDTH 1
+  #define TX_IPFIL17_PORT1_EN_LBN 35
+  #define TX_IPFIL17_PORT1_EN_WIDTH 1
+  #define TX_IPFIL17_PORT0_EN_LBN 34
+  #define TX_IPFIL17_PORT0_EN_WIDTH 1
+  #define TX_IPFIL16_PORT1_EN_LBN 33
+  #define TX_IPFIL16_PORT1_EN_WIDTH 1
+  #define TX_IPFIL16_PORT0_EN_LBN 32
+  #define TX_IPFIL16_PORT0_EN_WIDTH 1
+  #define TX_IPFIL15_PORT1_EN_LBN 31
+  #define TX_IPFIL15_PORT1_EN_WIDTH 1
+  #define TX_IPFIL15_PORT0_EN_LBN 30
+  #define TX_IPFIL15_PORT0_EN_WIDTH 1
+  #define TX_IPFIL14_PORT1_EN_LBN 29
+  #define TX_IPFIL14_PORT1_EN_WIDTH 1
+  #define TX_IPFIL14_PORT0_EN_LBN 28
+  #define TX_IPFIL14_PORT0_EN_WIDTH 1
+  #define TX_IPFIL13_PORT1_EN_LBN 27
+  #define TX_IPFIL13_PORT1_EN_WIDTH 1
+  #define TX_IPFIL13_PORT0_EN_LBN 26
+  #define TX_IPFIL13_PORT0_EN_WIDTH 1
+  #define TX_IPFIL12_PORT1_EN_LBN 25
+  #define TX_IPFIL12_PORT1_EN_WIDTH 1
+  #define TX_IPFIL12_PORT0_EN_LBN 24
+  #define TX_IPFIL12_PORT0_EN_WIDTH 1
+  #define TX_IPFIL11_PORT1_EN_LBN 23
+  #define TX_IPFIL11_PORT1_EN_WIDTH 1
+  #define TX_IPFIL11_PORT0_EN_LBN 22
+  #define TX_IPFIL11_PORT0_EN_WIDTH 1
+  #define TX_IPFIL10_PORT1_EN_LBN 21
+  #define TX_IPFIL10_PORT1_EN_WIDTH 1
+  #define TX_IPFIL10_PORT0_EN_LBN 20
+  #define TX_IPFIL10_PORT0_EN_WIDTH 1
+  #define TX_IPFIL9_PORT1_EN_LBN 19
+  #define TX_IPFIL9_PORT1_EN_WIDTH 1
+  #define TX_IPFIL9_PORT0_EN_LBN 18
+  #define TX_IPFIL9_PORT0_EN_WIDTH 1
+  #define TX_IPFIL8_PORT1_EN_LBN 17
+  #define TX_IPFIL8_PORT1_EN_WIDTH 1
+  #define TX_IPFIL8_PORT0_EN_LBN 16
+  #define TX_IPFIL8_PORT0_EN_WIDTH 1
+  #define TX_IPFIL7_PORT1_EN_LBN 15
+  #define TX_IPFIL7_PORT1_EN_WIDTH 1
+  #define TX_IPFIL7_PORT0_EN_LBN 14
+  #define TX_IPFIL7_PORT0_EN_WIDTH 1
+  #define TX_IPFIL6_PORT1_EN_LBN 13
+  #define TX_IPFIL6_PORT1_EN_WIDTH 1
+  #define TX_IPFIL6_PORT0_EN_LBN 12
+  #define TX_IPFIL6_PORT0_EN_WIDTH 1
+  #define TX_IPFIL5_PORT1_EN_LBN 11
+  #define TX_IPFIL5_PORT1_EN_WIDTH 1
+  #define TX_IPFIL5_PORT0_EN_LBN 10
+  #define TX_IPFIL5_PORT0_EN_WIDTH 1
+  #define TX_IPFIL4_PORT1_EN_LBN 9
+  #define TX_IPFIL4_PORT1_EN_WIDTH 1
+  #define TX_IPFIL4_PORT0_EN_LBN 8
+  #define TX_IPFIL4_PORT0_EN_WIDTH 1
+  #define TX_IPFIL3_PORT1_EN_LBN 7
+  #define TX_IPFIL3_PORT1_EN_WIDTH 1
+  #define TX_IPFIL3_PORT0_EN_LBN 6
+  #define TX_IPFIL3_PORT0_EN_WIDTH 1
+  #define TX_IPFIL2_PORT1_EN_LBN 5
+  #define TX_IPFIL2_PORT1_EN_WIDTH 1
+  #define TX_IPFIL2_PORT0_EN_LBN 4
+  #define TX_IPFIL2_PORT0_EN_WIDTH 1
+  #define TX_IPFIL1_PORT1_EN_LBN 3
+  #define TX_IPFIL1_PORT1_EN_WIDTH 1
+  #define TX_IPFIL1_PORT0_EN_LBN 2
+  #define TX_IPFIL1_PORT0_EN_WIDTH 1
+  #define TX_IPFIL0_PORT1_EN_LBN 1
+  #define TX_IPFIL0_PORT1_EN_WIDTH 1
+  #define TX_IPFIL0_PORT0_EN_LBN 0
+  #define TX_IPFIL0_PORT0_EN_WIDTH 1
+#define TX_IPFIL_TBL_OFST 0xB00 /* Transmit IP source address filter table */
+  #define TX_IPFIL_MASK_LBN 32
+  #define TX_IPFIL_MASK_WIDTH 32
+  #define TX_IP_SRC_ADR_LBN 0
+  #define TX_IP_SRC_ADR_WIDTH 32
+#define TX_PACE_REG_A1_OFST 0xF80000 /* Transmit pace control register */
+#define TX_PACE_REG_B0_OFST 0xA90    /* Transmit pace control register */
+  #define TX_PACE_SB_AF_LBN 19
+  #define TX_PACE_SB_AF_WIDTH 10
+  #define TX_PACE_SB_NOTAF_LBN 9
+  #define TX_PACE_SB_NOTAF_WIDTH 10
+  #define TX_PACE_FB_BASE_LBN 5
+  #define TX_PACE_FB_BASE_WIDTH 4
+  #define TX_PACE_BIN_TH_LBN 0
+  #define TX_PACE_BIN_TH_WIDTH 5
+#define TX_PACE_TBL_A1_OFST 0xF80040 /* Transmit pacing table */
+#define TX_PACE_TBL_FIRST_QUEUE_A1 4
+#define TX_PACE_TBL_B0_OFST 0xF80000 /* Transmit pacing table */
+#define TX_PACE_TBL_FIRST_QUEUE_B0 0
+  #define TX_PACE_LBN 0
+  #define TX_PACE_WIDTH 5
+
+/*************---- EE/Flash Registers C Header ----*************/
+#define EE_SPI_HCMD_REG_KER_OFST 0x100 /* SPI host command register */
+#define EE_SPI_HCMD_REG_OFST 0x100 /* SPI host command register */
+  #define EE_SPI_HCMD_CMD_EN_LBN 31
+  #define EE_SPI_HCMD_CMD_EN_WIDTH 1
+  #define EE_WR_TIMER_ACTIVE_LBN 28
+  #define EE_WR_TIMER_ACTIVE_WIDTH 1
+  #define EE_SPI_HCMD_SF_SEL_LBN 24
+  #define EE_SPI_HCMD_SF_SEL_WIDTH 1
+  #define EE_SPI_HCMD_DABCNT_LBN 16
+  #define EE_SPI_HCMD_DABCNT_WIDTH 5
+  #define EE_SPI_HCMD_READ_LBN 15
+  #define EE_SPI_HCMD_READ_WIDTH 1
+  #define EE_SPI_HCMD_DUBCNT_LBN 12
+  #define EE_SPI_HCMD_DUBCNT_WIDTH 2
+  #define EE_SPI_HCMD_ADBCNT_LBN 8
+  #define EE_SPI_HCMD_ADBCNT_WIDTH 2
+  #define EE_SPI_HCMD_ENC_LBN 0
+  #define EE_SPI_HCMD_ENC_WIDTH 8
+#define EE_SPI_HADR_REG_KER_OFST 0X110 /* SPI host address register */
+#define EE_SPI_HADR_REG_OFST 0X110 /* SPI host address register */
+  #define EE_SPI_HADR_DUBYTE_LBN 24
+  #define EE_SPI_HADR_DUBYTE_WIDTH 8
+  #define EE_SPI_HADR_ADR_LBN 0
+  #define EE_SPI_HADR_ADR_WIDTH 24
+#define EE_SPI_HDATA_REG_KER_OFST 0x120 /* SPI host data register */
+#define EE_SPI_HDATA_REG_OFST 0x120 /* SPI host data register */
+  #define EE_SPI_HDATA3_LBN 96
+  #define EE_SPI_HDATA3_WIDTH 32
+  #define EE_SPI_HDATA2_LBN 64
+  #define EE_SPI_HDATA2_WIDTH 32
+  #define EE_SPI_HDATA1_LBN 32
+  #define EE_SPI_HDATA1_WIDTH 32
+  #define EE_SPI_HDATA0_LBN 0
+  #define EE_SPI_HDATA0_WIDTH 32
+#define EE_BASE_PAGE_REG_KER_OFST 0x130 /* Expansion ROM base mirror register */
+#define EE_BASE_PAGE_REG_OFST 0x130 /* Expansion ROM base mirror register */
+  #define EE_EXP_ROM_WINDOW_BASE_LBN 16
+  #define EE_EXP_ROM_WINDOW_BASE_WIDTH 13
+  #define EE_EXPROM_MASK_LBN 0
+  #define EE_EXPROM_MASK_WIDTH 13
+#define EE_VPD_CFG0_REG_KER_OFST 0X140 /* SPI/VPD configuration register */
+#define EE_VPD_CFG0_REG_OFST 0X140 /* SPI/VPD configuration register */
+  #define EE_SF_FASTRD_EN_LBN 127
+  #define EE_SF_FASTRD_EN_WIDTH 1
+  #define EE_SF_CLOCK_DIV_LBN 120
+  #define EE_SF_CLOCK_DIV_WIDTH 7
+  #define EE_VPD_WIP_POLL_LBN 119
+  #define EE_VPD_WIP_POLL_WIDTH 1
+  #define EE_VPDW_LENGTH_LBN 80
+  #define EE_VPDW_LENGTH_WIDTH 15
+  #define EE_VPDW_BASE_LBN 64
+  #define EE_VPDW_BASE_WIDTH 15
+  #define EE_VPD_WR_CMD_EN_LBN 56
+  #define EE_VPD_WR_CMD_EN_WIDTH 8
+  #define EE_VPD_BASE_LBN 32
+  #define EE_VPD_BASE_WIDTH 24
+  #define EE_VPD_LENGTH_LBN 16
+  #define EE_VPD_LENGTH_WIDTH 13
+  #define EE_VPD_AD_SIZE_LBN 8
+  #define EE_VPD_AD_SIZE_WIDTH 5
+  #define EE_VPD_ACCESS_ON_LBN 5
+  #define EE_VPD_ACCESS_ON_WIDTH 1
+#define EE_VPD_SW_CNTL_REG_KER_OFST 0X150 /* VPD access SW control register */
+#define EE_VPD_SW_CNTL_REG_OFST 0X150 /* VPD access SW control register */
+  #define EE_VPD_CYCLE_PENDING_LBN 31
+  #define EE_VPD_CYCLE_PENDING_WIDTH 1
+  #define EE_VPD_CYC_WRITE_LBN 28
+  #define EE_VPD_CYC_WRITE_WIDTH 1
+  #define EE_VPD_CYC_ADR_LBN 0
+  #define EE_VPD_CYC_ADR_WIDTH 15
+#define EE_VPD_SW_DATA_REG_KER_OFST 0x160 /* VPD access SW data register */
+#define EE_VPD_SW_DATA_REG_OFST 0x160 /* VPD access SW data register */
+  #define EE_VPD_CYC_DAT_LBN 0
+  #define EE_VPD_CYC_DAT_WIDTH 32
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_desc.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_desc.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_desc.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_desc.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,75 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides EtherFabric NIC - EFXXXX (aka Falcon) descriptor
+ * definitions.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*************---- Descriptors C Headers ----*************/
+/* Receive Kernel IP Descriptor */
+  #define RX_KER_BUF_SIZE_LBN 48
+  #define RX_KER_BUF_SIZE_WIDTH 14
+  #define RX_KER_BUF_REGION_LBN 46
+  #define RX_KER_BUF_REGION_WIDTH 2
+      #define RX_KER_BUF_REGION0_DECODE 0
+      #define RX_KER_BUF_REGION1_DECODE 1
+      #define RX_KER_BUF_REGION2_DECODE 2
+      #define RX_KER_BUF_REGION3_DECODE 3
+  #define RX_KER_BUF_ADR_LBN 0
+  #define RX_KER_BUF_ADR_WIDTH 46
+/* Receive User IP Descriptor */
+  #define RX_USR_2BYTE_OFS_LBN 20
+  #define RX_USR_2BYTE_OFS_WIDTH 12
+  #define RX_USR_BUF_ID_LBN 0
+  #define RX_USR_BUF_ID_WIDTH 20
+/* Transmit Kernel IP Descriptor */
+  #define TX_KER_PORT_LBN 63
+  #define TX_KER_PORT_WIDTH 1
+  #define TX_KER_CONT_LBN 62
+  #define TX_KER_CONT_WIDTH 1
+  #define TX_KER_BYTE_CNT_LBN 48
+  #define TX_KER_BYTE_CNT_WIDTH 14
+  #define TX_KER_BUF_REGION_LBN 46
+  #define TX_KER_BUF_REGION_WIDTH 2
+      #define TX_KER_BUF_REGION0_DECODE 0
+      #define TX_KER_BUF_REGION1_DECODE 1
+      #define TX_KER_BUF_REGION2_DECODE 2
+      #define TX_KER_BUF_REGION3_DECODE 3
+  #define TX_KER_BUF_ADR_LBN 0
+  #define TX_KER_BUF_ADR_WIDTH 46
+/* Transmit User IP Descriptor */
+  #define TX_USR_PORT_LBN 47
+  #define TX_USR_PORT_WIDTH 1
+  #define TX_USR_CONT_LBN 46
+  #define TX_USR_CONT_WIDTH 1
+  #define TX_USR_BYTE_CNT_LBN 33
+  #define TX_USR_BYTE_CNT_WIDTH 13
+  #define TX_USR_BUF_ID_LBN 13
+  #define TX_USR_BUF_ID_WIDTH 20
+  #define TX_USR_BYTE_OFS_LBN 0
+  #define TX_USR_BYTE_OFS_WIDTH 13
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_event.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_event.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_event.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_event.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,155 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides EtherFabric NIC - EFXXXX (aka Falcon) event
+ * definitions.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*************---- Events Format C Header ----*************/
+/*************---- Event entry ----*************/
+  #define EV_CODE_LBN 60
+  #define EV_CODE_WIDTH 4
+      #define RX_IP_EV_DECODE 0
+      #define TX_IP_EV_DECODE 2
+      #define DRIVER_EV_DECODE 5
+      #define GLOBAL_EV_DECODE 6
+      #define DRV_GEN_EV_DECODE 7
+  #define EV_DATA_LBN 0
+  #define EV_DATA_WIDTH 60
+/******---- Receive IP events for both Kernel & User event queues ----******/
+  #define RX_EV_PKT_OK_LBN 56
+  #define RX_EV_PKT_OK_WIDTH 1
+  #define RX_EV_BUF_OWNER_ID_ERR_LBN 54
+  #define RX_EV_BUF_OWNER_ID_ERR_WIDTH 1
+  #define RX_EV_IP_HDR_CHKSUM_ERR_LBN 52
+  #define RX_EV_IP_HDR_CHKSUM_ERR_WIDTH 1
+  #define RX_EV_TCP_UDP_CHKSUM_ERR_LBN 51
+  #define RX_EV_TCP_UDP_CHKSUM_ERR_WIDTH 1
+  #define RX_EV_ETH_CRC_ERR_LBN 50
+  #define RX_EV_ETH_CRC_ERR_WIDTH 1
+  #define RX_EV_FRM_TRUNC_LBN 49
+  #define RX_EV_FRM_TRUNC_WIDTH 1
+  #define RX_EV_DRIB_NIB_LBN 48
+  #define RX_EV_DRIB_NIB_WIDTH 1
+  #define RX_EV_TOBE_DISC_LBN 47
+  #define RX_EV_TOBE_DISC_WIDTH 1
+  #define RX_EV_PKT_TYPE_LBN 44
+  #define RX_EV_PKT_TYPE_WIDTH 3
+      #define RX_EV_PKT_TYPE_ETH_DECODE 0
+      #define RX_EV_PKT_TYPE_LLC_DECODE 1
+      #define RX_EV_PKT_TYPE_JUMBO_DECODE 2
+      #define RX_EV_PKT_TYPE_VLAN_DECODE 3
+      #define RX_EV_PKT_TYPE_VLAN_LLC_DECODE 4
+      #define RX_EV_PKT_TYPE_VLAN_JUMBO_DECODE 5
+  #define RX_EV_HDR_TYPE_LBN 42
+  #define RX_EV_HDR_TYPE_WIDTH 2
+      #define RX_EV_HDR_TYPE_TCP_IPV4_DECODE 0
+      #define RX_EV_HDR_TYPE_UDP_IPV4_DECODE 1
+      #define RX_EV_HDR_TYPE_OTHER_IP_DECODE 2
+      #define RX_EV_HDR_TYPE_NON_IP_DECODE 3
+  #define RX_EV_DESC_Q_EMPTY_LBN 41
+  #define RX_EV_DESC_Q_EMPTY_WIDTH 1
+  #define RX_EV_MCAST_HASH_MATCH_LBN 40
+  #define RX_EV_MCAST_HASH_MATCH_WIDTH 1
+  #define RX_EV_MCAST_PKT_LBN 39
+  #define RX_EV_MCAST_PKT_WIDTH 1
+  #define RX_EV_Q_LABEL_LBN 32
+  #define RX_EV_Q_LABEL_WIDTH 5
+  #define RX_JUMBO_CONT_LBN 31
+  #define RX_JUMBO_CONT_WIDTH 1
+  #define RX_SOP_LBN 15
+  #define RX_SOP_WIDTH 1
+  #define RX_PORT_LBN 30
+  #define RX_PORT_WIDTH 1
+  #define RX_EV_BYTE_CNT_LBN 16
+  #define RX_EV_BYTE_CNT_WIDTH 14
+  #define RX_iSCSI_PKT_OK_LBN 14
+  #define RX_iSCSI_PKT_OK_WIDTH 1
+  #define RX_ISCSI_DDIG_ERR_LBN 13
+  #define RX_ISCSI_DDIG_ERR_WIDTH 1
+  #define RX_ISCSI_HDIG_ERR_LBN 12
+  #define RX_ISCSI_HDIG_ERR_WIDTH 1
+  #define RX_EV_DESC_PTR_LBN 0
+  #define RX_EV_DESC_PTR_WIDTH 12
+/******---- Transmit IP events for both Kernel & User event queues ----******/
+  #define TX_EV_PKT_ERR_LBN 38
+  #define TX_EV_PKT_ERR_WIDTH 1
+  #define TX_EV_PKT_TOO_BIG_LBN 37
+  #define TX_EV_PKT_TOO_BIG_WIDTH 1
+  #define TX_EV_Q_LABEL_LBN 32
+  #define TX_EV_Q_LABEL_WIDTH 5
+  #define TX_EV_PORT_LBN 16
+  #define TX_EV_PORT_WIDTH 1
+  #define TX_EV_WQ_FF_FULL_LBN 15
+  #define TX_EV_WQ_FF_FULL_WIDTH 1
+  #define TX_EV_BUF_OWNER_ID_ERR_LBN 14
+  #define TX_EV_BUF_OWNER_ID_ERR_WIDTH 1
+  #define TX_EV_COMP_LBN 12
+  #define TX_EV_COMP_WIDTH 1
+  #define TX_EV_DESC_PTR_LBN 0
+  #define TX_EV_DESC_PTR_WIDTH 12
+/*************---- Char or Kernel driver events ----*************/
+  #define DRIVER_EV_SUB_CODE_LBN 56
+  #define DRIVER_EV_SUB_CODE_WIDTH 4
+      #define TX_DESCQ_FLS_DONE_EV_DECODE 0x0
+      #define RX_DESCQ_FLS_DONE_EV_DECODE 0x1
+      #define EVQ_INIT_DONE_EV_DECODE 0x2
+      #define EVQ_NOT_EN_EV_DECODE 0x3
+      #define RX_DESCQ_FLSFF_OVFL_EV_DECODE 0x4
+      #define SRM_UPD_DONE_EV_DECODE 0x5
+      #define WAKE_UP_EV_DECODE 0x6
+      #define TX_PKT_NON_TCP_UDP_DECODE 0x9
+      #define TIMER_EV_DECODE 0xA
+      #define RX_DSC_ERROR_EV_DECODE 0xE
+  #define DRIVER_EV_TX_DESCQ_ID_LBN 0
+  #define DRIVER_EV_TX_DESCQ_ID_WIDTH 12
+  #define DRIVER_EV_RX_DESCQ_ID_LBN 0
+  #define DRIVER_EV_RX_DESCQ_ID_WIDTH 12
+  #define DRIVER_EV_EVQ_ID_LBN 0
+  #define DRIVER_EV_EVQ_ID_WIDTH 12
+  #define DRIVER_TMR_ID_LBN 0
+  #define DRIVER_TMR_ID_WIDTH 12
+  #define DRIVER_EV_SRM_UPD_LBN 0
+  #define DRIVER_EV_SRM_UPD_WIDTH 2
+      #define SRM_CLR_EV_DECODE 0
+      #define SRM_UPD_EV_DECODE 1
+      #define SRM_ILLCLR_EV_DECODE 2
+/********---- Global events. Sent to both event queue 0 and 4. ----********/
+  #define XFP_PHY_INTR_LBN 10
+  #define XFP_PHY_INTR_WIDTH 1
+  #define XG_PHY_INTR_LBN 9
+  #define XG_PHY_INTR_WIDTH 1
+  #define G_PHY1_INTR_LBN 8
+  #define G_PHY1_INTR_WIDTH 1
+  #define G_PHY0_INTR_LBN 7
+  #define G_PHY0_INTR_WIDTH 1
+/*************---- Driver generated events ----*************/
+  #define DRV_GEN_EV_CODE_LBN 60
+  #define DRV_GEN_EV_CODE_WIDTH 4
+  #define DRV_GEN_EV_DATA_LBN 0
+  #define DRV_GEN_EV_DATA_WIDTH 60
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_grmon.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_grmon.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_grmon.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_grmon.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,129 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides EtherFabric NIC - EFXXXX (aka Falcon) 1G MAC
+ * counters.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*************---- 1G MAC Statistical Counters C Header ----*************/
+#define GRxGoodOct_offset 0x0
+    #define GRxGoodOct_WIDTH 48
+#define GRxBadOct_offset 0x8
+    #define GRxBadOct_WIDTH 48
+#define GRxMissPkt_offset 0x10
+    #define GRxMissPkt_WIDTH 32
+#define GRxFalseCRS_offset 0x14
+    #define GRxFalseCRS_WIDTH 32
+#define GRxPausePkt_offset 0x18
+    #define GRxPausePkt_WIDTH 32
+#define GRxBadPkt_offset 0x1C
+    #define GRxBadPkt_WIDTH 32
+#define GRxUcastPkt_offset 0x20
+    #define GRxUcastPkt_WIDTH 32
+#define GRxMcastPkt_offset 0x24
+    #define GRxMcastPkt_WIDTH 32
+#define GRxBcastPkt_offset 0x28
+    #define GRxBcastPkt_WIDTH 32
+#define GRxGoodLt64Pkt_offset 0x2C
+    #define GRxGoodLt64Pkt_WIDTH 32
+#define GRxBadLt64Pkt_offset 0x30
+    #define GRxBadLt64Pkt_WIDTH 32
+#define GRx64Pkt_offset 0x34
+    #define GRx64Pkt_WIDTH 32
+#define GRx65to127Pkt_offset 0x38
+    #define GRx65to127Pkt_WIDTH 32
+#define GRx128to255Pkt_offset 0x3C
+    #define GRx128to255Pkt_WIDTH 32
+#define GRx256to511Pkt_offset 0x40
+    #define GRx256to511Pkt_WIDTH 32
+#define GRx512to1023Pkt_offset 0x44
+    #define GRx512to1023Pkt_WIDTH 32
+#define GRx1024to15xxPkt_offset 0x48
+    #define GRx1024to15xxPkt_WIDTH 32
+#define GRx15xxtoJumboPkt_offset 0x4C
+    #define GRx15xxtoJumboPkt_WIDTH 32
+#define GRxGtJumboPkt_offset 0x50
+    #define GRxGtJumboPkt_WIDTH 32
+#define GRxFcsErr64to15xxPkt_offset 0x54
+    #define GRxFcsErr64to15xxPkt_WIDTH 32
+#define GRxFcsErr15xxtoJumboPkt_offset 0x58
+    #define GRxFcsErr15xxtoJumboPkt_WIDTH 32
+#define GRxFcsErrGtJumboPkt_offset 0x5C
+    #define GRxFcsErrGtJumboPkt_WIDTH 32
+#define GTxGoodBadOct_offset 0x80
+    #define GTxGoodBadOct_WIDTH 48
+#define GTxGoodOct_offset 0x88
+    #define GTxGoodOct_WIDTH 48
+#define GTxSglColPkt_offset 0x90
+    #define GTxSglColPkt_WIDTH 32
+#define GTxMultColPkt_offset 0x94
+    #define GTxMultColPkt_WIDTH 32
+#define GTxExColPkt_offset 0x98
+    #define GTxExColPkt_WIDTH 32
+#define GTxDefPkt_offset 0x9C
+    #define GTxDefPkt_WIDTH 32
+#define GTxLateCol_offset 0xA0
+    #define GTxLateCol_WIDTH 32
+#define GTxExDefPkt_offset 0xA4
+    #define GTxExDefPkt_WIDTH 32
+#define GTxPausePkt_offset 0xA8
+    #define GTxPausePkt_WIDTH 32
+#define GTxBadPkt_offset 0xAC
+    #define GTxBadPkt_WIDTH 32
+#define GTxUcastPkt_offset 0xB0
+    #define GTxUcastPkt_WIDTH 32
+#define GTxMcastPkt_offset 0xB4
+    #define GTxMcastPkt_WIDTH 32
+#define GTxBcastPkt_offset 0xB8
+    #define GTxBcastPkt_WIDTH 32
+#define GTxLt64Pkt_offset 0xBC
+    #define GTxLt64Pkt_WIDTH 32
+#define GTx64Pkt_offset 0xC0
+    #define GTx64Pkt_WIDTH 32
+#define GTx65to127Pkt_offset 0xC4
+    #define GTx65to127Pkt_WIDTH 32
+#define GTx128to255Pkt_offset 0xC8
+    #define GTx128to255Pkt_WIDTH 32
+#define GTx256to511Pkt_offset 0xCC
+    #define GTx256to511Pkt_WIDTH 32
+#define GTx512to1023Pkt_offset 0xD0
+    #define GTx512to1023Pkt_WIDTH 32
+#define GTx1024to15xxPkt_offset 0xD4
+    #define GTx1024to15xxPkt_WIDTH 32
+#define GTx15xxtoJumboPkt_offset 0xD8
+    #define GTx15xxtoJumboPkt_WIDTH 32
+#define GTxGtJumboPkt_offset 0xDC
+    #define GTxGtJumboPkt_WIDTH 32
+#define GTxNonTcpUdpPkt_offset 0xE0
+    #define GTxNonTcpUdpPkt_WIDTH 16
+#define GTxMacSrcErrPkt_offset 0xE4
+    #define GTxMacSrcErrPkt_WIDTH 16
+#define GTxIpSrcErrPkt_offset 0xE8
+    #define GTxIpSrcErrPkt_WIDTH 16
+#define GDmaDone_offset 0xEC
+    #define GDmaDone_WIDTH 32
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_intr_vec.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_intr_vec.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_intr_vec.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_intr_vec.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,44 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides EtherFabric NIC - EFXXXX (aka Falcon) interrupt
+ * vector definitions.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*************---- Interrupt Vector Format C Header ----*************/
+#define DW0_OFST 0x0 /* Double-word 0: Event queue FIFO interrupts */
+  #define EVQ_FIFO_HF_LBN 1
+  #define EVQ_FIFO_HF_WIDTH 1
+  #define EVQ_FIFO_AF_LBN 0
+  #define EVQ_FIFO_AF_WIDTH 1
+#define DW1_OFST 0x4 /* Double-word 1: Interrupt indicator */
+  #define INT_FLAG_LBN 0
+  #define INT_FLAG_WIDTH 1
+#define DW2_OFST 0x8 /* Double-word 2: Fatal interrupts */
+  #define FATAL_INT_LBN 0
+  #define FATAL_INT_WIDTH 1
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_mac.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_mac.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_mac.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_mac.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,711 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides EtherFabric NIC - EFXXXX (aka Falcon) MAC register
+ * definitions.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*********---- 1G/10G Ethernet MAC Wrapper Registers C Header ----*********/
+#define MD_TXD_REG_KER_OFST 0xC00 /* PHY management transmit data register */
+#define MD_TXD_REG_OFST 0xC00 /* PHY management transmit data register */
+  #define MD_TXD_LBN 0
+  #define MD_TXD_WIDTH 16
+#define MD_RXD_REG_KER_OFST 0xC10 /* PHY management receive data register */
+#define MD_RXD_REG_OFST 0xC10 /* PHY management receive data register */
+  #define MD_RXD_LBN 0
+  #define MD_RXD_WIDTH 16
+#define MD_CS_REG_KER_OFST 0xC20 /* PHY management configuration &
+				    status register */
+#define MD_CS_REG_OFST 0xC20 /* PHY management configuration &
+				status register */
+  #define MD_PT_LBN 7
+  #define MD_PT_WIDTH 3
+  #define MD_PL_LBN 6
+  #define MD_PL_WIDTH 1
+  #define MD_INT_CLR_LBN 5
+  #define MD_INT_CLR_WIDTH 1
+  #define MD_GC_LBN 4
+  #define MD_GC_WIDTH 1
+  #define MD_PRSP_LBN 3
+  #define MD_PRSP_WIDTH 1
+  #define MD_RIC_LBN 2
+  #define MD_RIC_WIDTH 1
+  #define MD_RDC_LBN 1
+  #define MD_RDC_WIDTH 1
+  #define MD_WRC_LBN 0
+  #define MD_WRC_WIDTH 1
+#define MD_PHY_ADR_REG_KER_OFST 0xC30 /* PHY management PHY address register */
+#define MD_PHY_ADR_REG_OFST 0xC30 /* PHY management PHY address register */
+  #define MD_PHY_ADR_LBN 0
+  #define MD_PHY_ADR_WIDTH 16
+#define MD_ID_REG_KER_OFST 0xC40 /* PHY management ID register */
+#define MD_ID_REG_OFST 0xC40 /* PHY management ID register */
+  #define MD_PRT_ADR_LBN 11
+  #define MD_PRT_ADR_WIDTH 5
+  #define MD_DEV_ADR_LBN 6
+  #define MD_DEV_ADR_WIDTH 5
+#define MD_STAT_REG_KER_OFST 0xC50 /* PHY management status & mask register */
+#define MD_STAT_REG_OFST 0xC50 /* PHY management status & mask register */
+  #define MD_PINT_LBN 4
+  #define MD_PINT_WIDTH 1
+  #define MD_DONE_LBN 3
+  #define MD_DONE_WIDTH 1
+  #define MD_BSERR_LBN 2
+  #define MD_BSERR_WIDTH 1
+  #define MD_LNFL_LBN 1
+  #define MD_LNFL_WIDTH 1
+  #define MD_BSY_LBN 0
+  #define MD_BSY_WIDTH 1
+#define MAC0_STAT_DMA_REG_KER_OFST 0xC60 /* Port 0 MAC statistical counter
+					    DMA register */
+#define MAC0_STAT_DMA_REG_OFST 0xC60 /* Port 0 MAC statistical counter
+					DMA register */
+  #define MAC0_STAT_DMA_CMD_LBN 48
+  #define MAC0_STAT_DMA_CMD_WIDTH 1
+  #define MAC0_STAT_DMA_ADR_LBN 0
+  #define MAC0_STAT_DMA_ADR_WIDTH 48
+#define MAC1_STAT_DMA_REG_KER_OFST 0xC70 /* Port 1 MAC statistical counter
+					    DMA register */
+#define MAC1_STAT_DMA_REG_OFST 0xC70 /* Port 1 MAC statistical counter
+					DMA register */
+  #define MAC1_STAT_DMA_CMD_LBN 48
+  #define MAC1_STAT_DMA_CMD_WIDTH 1
+  #define MAC1_STAT_DMA_ADR_LBN 0
+  #define MAC1_STAT_DMA_ADR_WIDTH 48
+#define MAC0_CTRL_REG_KER_OFST 0xC80 /* Port 0 MAC control register */
+#define MAC0_CTRL_REG_OFST 0xC80 /* Port 0 MAC control register */
+  #define MAC0_XOFF_VAL_LBN 16
+  #define MAC0_XOFF_VAL_WIDTH 16
+  #define MAC0_BCAD_ACPT_LBN 4
+  #define MAC0_BCAD_ACPT_WIDTH 1
+  #define MAC0_UC_PROM_LBN 3
+  #define MAC0_UC_PROM_WIDTH 1
+  #define MAC0_LINK_STATUS_LBN 2
+  #define MAC0_LINK_STATUS_WIDTH 1
+  #define MAC0_SPEED_LBN 0
+  #define MAC0_SPEED_WIDTH 2
+#define MAC1_CTRL_REG_KER_OFST 0xC90 /* Port 1 MAC control register */
+#define MAC1_CTRL_REG_OFST 0xC90 /* Port 1 MAC control register */
+  #define MAC1_XOFF_VAL_LBN 16
+  #define MAC1_XOFF_VAL_WIDTH 16
+  #define MAC1_BCAD_ACPT_LBN 4
+  #define MAC1_BCAD_ACPT_WIDTH 1
+  #define MAC1_UC_PROM_LBN 3
+  #define MAC1_UC_PROM_WIDTH 1
+  #define MAC1_LINK_STATUS_LBN 2
+  #define MAC1_LINK_STATUS_WIDTH 1
+  #define MAC1_SPEED_LBN 0
+  #define MAC1_SPEED_WIDTH 2
+#define MAC_MC_HASH_REG0_KER_OFST 0xCA0 /* Multicast address hash table */
+#define MAC_MC_HASH_REG0_OFST 0xCA0 /* Multicast address hash table */
+  #define MAC_MCAST_HASH0_LBN 0
+  #define MAC_MCAST_HASH0_WIDTH 128
+#define MAC_MC_HASH_REG1_KER_OFST 0xCB0 /* Multicast address hash table */
+#define MAC_MC_HASH_REG1_OFST 0xCB0 /* Multicast address hash table */
+  #define MAC_MCAST_HASH1_LBN 0
+  #define MAC_MCAST_HASH1_WIDTH 128
+/*************---- 1G MAC Port 0 Registers C Header ----*************/
+#define GM_P0_BASE 0xE00
+#define GM_P1_BASE 0x1000
+#define GM_CFG1_REG_KER_OFST 0x00 /* GMAC configuration register 1 */
+#define GM_CFG1_REG_OFST 0x00 /* GMAC configuration register 1 */
+  #define GM_SW_RST_LBN 31
+  #define GM_SW_RST_WIDTH 1
+  #define GM_SIM_RST_LBN 30
+  #define GM_SIM_RST_WIDTH 1
+  #define GM_RST_RX_MAC_CTL_LBN 19
+  #define GM_RST_RX_MAC_CTL_WIDTH 1
+  #define GM_RST_TX_MAC_CTL_LBN 18
+  #define GM_RST_TX_MAC_CTL_WIDTH 1
+  #define GM_RST_RX_FUNC_LBN 17
+  #define GM_RST_RX_FUNC_WIDTH 1
+  #define GM_RST_TX_FUNC_LBN 16
+  #define GM_RST_TX_FUNC_WIDTH 1
+  #define GM_LOOP_LBN 8
+  #define GM_LOOP_WIDTH 1
+  #define GM_RX_FC_EN_LBN 5
+  #define GM_RX_FC_EN_WIDTH 1
+  #define GM_TX_FC_EN_LBN 4
+  #define GM_TX_FC_EN_WIDTH 1
+  #define GM_SYNC_RXEN_LBN 3
+  #define GM_SYNC_RXEN_WIDTH 1
+  #define GM_RX_EN_LBN 2
+  #define GM_RX_EN_WIDTH 1
+  #define GM_SYNC_TXEN_LBN 1
+  #define GM_SYNC_TXEN_WIDTH 1
+  #define GM_TX_EN_LBN 0
+  #define GM_TX_EN_WIDTH 1
+#define GM_CFG2_REG_KER_OFST 0x10 /* GMAC configuration register 2 */
+#define GM_CFG2_REG_OFST 0x10 /* GMAC configuration register 2 */
+  #define GM_PAMBL_LEN_LBN 12
+  #define GM_PAMBL_LEN_WIDTH 4
+  #define GM_IF_MODE_LBN 8
+  #define GM_IF_MODE_WIDTH 2
+  #define GM_HUGE_FRM_EN_LBN 5
+  #define GM_HUGE_FRM_EN_WIDTH 1
+  #define GM_LEN_CHK_LBN 4
+  #define GM_LEN_CHK_WIDTH 1
+  #define GM_PAD_CRC_EN_LBN 2
+  #define GM_PAD_CRC_EN_WIDTH 1
+  #define GM_CRC_EN_LBN 1
+  #define GM_CRC_EN_WIDTH 1
+  #define GM_FD_LBN 0
+  #define GM_FD_WIDTH 1
+#define GM_IPG_REG_KER_OFST 0x20 /* GMAC IPG register */
+#define GM_IPG_REG_OFST 0x20 /* GMAC IPG register */
+  #define GM_NONB2B_IPG1_LBN 24
+  #define GM_NONB2B_IPG1_WIDTH 7
+  #define GM_NONB2B_IPG2_LBN 16
+  #define GM_NONB2B_IPG2_WIDTH 7
+  #define GM_MIN_IPG_ENF_LBN 8
+  #define GM_MIN_IPG_ENF_WIDTH 8
+  #define GM_B2B_IPG_LBN 0
+  #define GM_B2B_IPG_WIDTH 7
+#define GM_HD_REG_KER_OFST 0x30 /* GMAC half duplex register */
+#define GM_HD_REG_OFST 0x30 /* GMAC half duplex register */
+  #define GM_ALT_BOFF_VAL_LBN 20
+  #define GM_ALT_BOFF_VAL_WIDTH 4
+  #define GM_ALT_BOFF_EN_LBN 19
+  #define GM_ALT_BOFF_EN_WIDTH 1
+  #define GM_BP_NO_BOFF_LBN 18
+  #define GM_BP_NO_BOFF_WIDTH 1
+  #define GM_DIS_BOFF_LBN 17
+  #define GM_DIS_BOFF_WIDTH 1
+  #define GM_EXDEF_TX_EN_LBN 16
+  #define GM_EXDEF_TX_EN_WIDTH 1
+  #define GM_RTRY_LIMIT_LBN 12
+  #define GM_RTRY_LIMIT_WIDTH 4
+  #define GM_COL_WIN_LBN 0
+  #define GM_COL_WIN_WIDTH 10
+#define GM_MAX_FLEN_REG_KER_OFST 0x40 /* GMAC maximum frame length register */
+#define GM_MAX_FLEN_REG_OFST 0x40 /* GMAC maximum frame length register */
+  #define GM_MAX_FLEN_LBN 0
+  #define GM_MAX_FLEN_WIDTH 16
+#define GM_TEST_REG_KER_OFST 0x70 /* GMAC test register */
+#define GM_TEST_REG_OFST 0x70 /* GMAC test register */
+  #define GM_MAX_BOFF_LBN 3
+  #define GM_MAX_BOFF_WIDTH 1
+  #define GM_REG_TX_FLOW_EN_LBN 2
+  #define GM_REG_TX_FLOW_EN_WIDTH 1
+  #define GM_TEST_PAUSE_LBN 1
+  #define GM_TEST_PAUSE_WIDTH 1
+  #define GM_SHORT_SLOT_LBN 0
+  #define GM_SHORT_SLOT_WIDTH 1
+#define GM_ADR1_REG_KER_OFST 0x100 /* GMAC station address register 1 */
+#define GM_ADR1_REG_OFST 0x100 /* GMAC station address register 1 */
+  #define GM_ADR1_LBN 0
+  #define GM_ADR1_WIDTH 32
+#define GM_ADR2_REG_KER_OFST 0x110 /* GMAC station address register 2 */
+#define GM_ADR2_REG_OFST 0x110 /* GMAC station address register 2 */
+  #define GM_ADR2_LBN 16
+  #define GM_ADR2_WIDTH 16
+#define GMF_CFG0_REG_KER_OFST 0x120 /* GMAC FIFO configuration register 0 */
+#define GMF_CFG0_REG_OFST 0x120 /* GMAC FIFO configuration register 0 */
+  #define GMF_FTFENRPLY_LBN 20
+  #define GMF_FTFENRPLY_WIDTH 1
+  #define GMF_STFENRPLY_LBN 19
+  #define GMF_STFENRPLY_WIDTH 1
+  #define GMF_FRFENRPLY_LBN 18
+  #define GMF_FRFENRPLY_WIDTH 1
+  #define GMF_SRFENRPLY_LBN 17
+  #define GMF_SRFENRPLY_WIDTH 1
+  #define GMF_WTMENRPLY_LBN 16
+  #define GMF_WTMENRPLY_WIDTH 1
+  #define GMF_FTFENREQ_LBN 12
+  #define GMF_FTFENREQ_WIDTH 1
+  #define GMF_STFENREQ_LBN 11
+  #define GMF_STFENREQ_WIDTH 1
+  #define GMF_FRFENREQ_LBN 10
+  #define GMF_FRFENREQ_WIDTH 1
+  #define GMF_SRFENREQ_LBN 9
+  #define GMF_SRFENREQ_WIDTH 1
+  #define GMF_WTMENREQ_LBN 8
+  #define GMF_WTMENREQ_WIDTH 1
+  #define GMF_HSTRSTFT_LBN 4
+  #define GMF_HSTRSTFT_WIDTH 1
+  #define GMF_HSTRSTST_LBN 3
+  #define GMF_HSTRSTST_WIDTH 1
+  #define GMF_HSTRSTFR_LBN 2
+  #define GMF_HSTRSTFR_WIDTH 1
+  #define GMF_HSTRSTSR_LBN 1
+  #define GMF_HSTRSTSR_WIDTH 1
+  #define GMF_HSTRSTWT_LBN 0
+  #define GMF_HSTRSTWT_WIDTH 1
+#define GMF_CFG1_REG_KER_OFST 0x130 /* GMAC FIFO configuration register 1 */
+#define GMF_CFG1_REG_OFST 0x130 /* GMAC FIFO configuration register 1 */
+  #define GMF_CFGFRTH_LBN 16
+  #define GMF_CFGFRTH_WIDTH 5
+  #define GMF_CFGXOFFRTX_LBN 0
+  #define GMF_CFGXOFFRTX_WIDTH 16
+#define GMF_CFG2_REG_KER_OFST 0x140 /* GMAC FIFO configuration register 2 */
+#define GMF_CFG2_REG_OFST 0x140 /* GMAC FIFO configuration register 2 */
+  #define GMF_CFGHWM_LBN 16
+  #define GMF_CFGHWM_WIDTH 6
+  #define GMF_CFGLWM_LBN 0
+  #define GMF_CFGLWM_WIDTH 6
+#define GMF_CFG3_REG_KER_OFST 0x150 /* GMAC FIFO configuration register 3 */
+#define GMF_CFG3_REG_OFST 0x150 /* GMAC FIFO configuration register 3 */
+  #define GMF_CFGHWMFT_LBN 16
+  #define GMF_CFGHWMFT_WIDTH 6
+  #define GMF_CFGFTTH_LBN 0
+  #define GMF_CFGFTTH_WIDTH 6
+#define GMF_CFG4_REG_KER_OFST 0x160 /* GMAC FIFO configuration register 4 */
+#define GMF_CFG4_REG_OFST 0x160 /* GMAC FIFO configuration register 4 */
+  #define GMF_HSTFLTRFRM_LBN 0
+  #define GMF_HSTFLTRFRM_WIDTH 18
+#define GMF_CFG5_REG_KER_OFST 0x170 /* GMAC FIFO configuration register 5 */
+#define GMF_CFG5_REG_OFST 0x170 /* GMAC FIFO configuration register 5 */
+  #define GMF_CFGHDPLX_LBN 22
+  #define GMF_CFGHDPLX_WIDTH 1
+  #define GMF_SRFULL_LBN 21
+  #define GMF_SRFULL_WIDTH 1
+  #define GMF_HSTSRFULLCLR_LBN 20
+  #define GMF_HSTSRFULLCLR_WIDTH 1
+  #define GMF_CFGBYTMODE_LBN 19
+  #define GMF_CFGBYTMODE_WIDTH 1
+  #define GMF_HSTDRPLT64_LBN 18
+  #define GMF_HSTDRPLT64_WIDTH 1
+  #define GMF_HSTFLTRFRMDC_LBN 0
+  #define GMF_HSTFLTRFRMDC_WIDTH 18
+/*************---- 10G MAC Registers C Header ----*************/
+#define XM_ADR_LO_REG_KER_P0_OFST 0x1200 /* XGMAC address register low -
+					    port 0 */
+#define XM_ADR_LO_REG_P0_OFST 0x1200 /* XGMAC address register low -
+					port 0 */
+  #define XM_ADR_LO_LBN 0
+  #define XM_ADR_LO_WIDTH 32
+#define XM_ADR_HI_REG_KER_P0_OFST 0x1210 /* XGMAC address register high -
+					    port 0 */
+#define XM_ADR_HI_REG_P0_OFST 0x1210 /* XGMAC address register high -
+					port 0 */
+  #define XM_ADR_HI_LBN 0
+  #define XM_ADR_HI_WIDTH 16
+#define XM_GLB_CFG_REG_KER_P0_OFST 0x1220 /* XGMAC global configuration -
+					     port 0 */
+#define XM_GLB_CFG_REG_P0_OFST 0x1220 /* XGMAC global configuration -
+					 port 0 */
+  #define XM_LINE_LB_DEEP_RSVD_LBN 28
+  #define XM_LINE_LB_DEEP_RSVD_WIDTH 1
+  #define XM_RMTFLT_GEN_LBN 17
+  #define XM_RMTFLT_GEN_WIDTH 1
+  #define XM_DEBUG_MODE_LBN 16
+  #define XM_DEBUG_MODE_WIDTH 1
+  #define XM_RX_STAT_EN_LBN 11
+  #define XM_RX_STAT_EN_WIDTH 1
+  #define XM_TX_STAT_EN_LBN 10
+  #define XM_TX_STAT_EN_WIDTH 1
+  #define XM_CUT_THRU_MODE_LBN 7
+  #define XM_CUT_THRU_MODE_WIDTH 1
+  #define XM_RX_JUMBO_MODE_LBN 6
+  #define XM_RX_JUMBO_MODE_WIDTH 1
+  #define XM_WAN_MODE_LBN 5
+  #define XM_WAN_MODE_WIDTH 1
+  #define XM_AUTOCLR_MODE_LBN 4
+  #define XM_AUTOCLR_MODE_WIDTH 1
+  #define XM_INTCLR_MODE_LBN 3
+  #define XM_INTCLR_MODE_WIDTH 1
+  #define XM_CORE_RST_LBN 0
+  #define XM_CORE_RST_WIDTH 1
+#define XM_TX_CFG_REG_KER_P0_OFST 0x1230 /* XGMAC transmit configuration -
+					    port 0 */
+#define XM_TX_CFG_REG_P0_OFST 0x1230 /* XGMAC transmit configuration -
+					port 0 */
+  #define XM_TX_PROG_LBN 24
+  #define XM_TX_PROG_WIDTH 1
+  #define XM_IPG_LBN 16
+  #define XM_IPG_WIDTH 4
+  #define XM_FCNTL_LBN 10
+  #define XM_FCNTL_WIDTH 1
+  #define XM_TXCRC_LBN 8
+  #define XM_TXCRC_WIDTH 1
+  #define XM_EDRC_LBN 6
+  #define XM_EDRC_WIDTH 1
+  #define XM_AUTO_PAD_LBN 5
+  #define XM_AUTO_PAD_WIDTH 1
+  #define XM_TX_PRMBL_LBN 2
+  #define XM_TX_PRMBL_WIDTH 1
+  #define XM_TXEN_LBN 1
+  #define XM_TXEN_WIDTH 1
+  #define XM_TX_RST_LBN 0
+  #define XM_TX_RST_WIDTH 1
+#define XM_RX_CFG_REG_KER_P0_OFST 0x1240 /* XGMAC receive configuration -
+					    port 0 */
+#define XM_RX_CFG_REG_P0_OFST 0x1240 /* XGMAC receive configuration -
+					port 0 */
+  #define XM_PASS_LENERR_LBN 26
+  #define XM_PASS_LENERR_WIDTH 1
+  #define XM_PASS_CRC_ERR_LBN 25
+  #define XM_PASS_CRC_ERR_WIDTH 1
+  #define XM_PASS_PRMBLE_ERR_LBN 24
+  #define XM_PASS_PRMBLE_ERR_WIDTH 1
+  #define XM_REJ_UCAST_LBN 18
+  #define XM_REJ_UCAST_WIDTH 1
+  #define XM_BSC_EN_LBN 17
+  #define XM_BSC_EN_WIDTH 1
+  #define XM_ACPT_ALL_MCAST_LBN 11
+  #define XM_ACPT_ALL_MCAST_WIDTH 1
+  #define XM_PASS_SAP_LBN 10
+  #define XM_PASS_SAP_WIDTH 1
+  #define XM_ACPT_ALL_UCAST_LBN 9
+  #define XM_ACPT_ALL_UCAST_WIDTH 1
+  #define XM_AUTO_DEPAD_LBN 8
+  #define XM_AUTO_DEPAD_WIDTH 1
+  #define XM_RXCRC_LBN 3
+  #define XM_RXCRC_WIDTH 1
+  #define XM_RX_PRMBL_LBN 2
+  #define XM_RX_PRMBL_WIDTH 1
+  #define XM_RXEN_LBN 1
+  #define XM_RXEN_WIDTH 1
+  #define XM_RX_RST_LBN 0
+  #define XM_RX_RST_WIDTH 1
+#define XM_FC_REG_KER_P0_OFST 0x1270 /* XGMAC flow control register -
+					port 0 */
+#define XM_FC_REG_P0_OFST 0x1270 /* XGMAC flow control register -
+				    port 0 */
+  #define XM_PAUSE_TIME_LBN 16
+  #define XM_PAUSE_TIME_WIDTH 16
+  #define XM_RX_MAC_STAT_LBN 11
+  #define XM_RX_MAC_STAT_WIDTH 1
+  #define XM_TX_MAC_STAT_LBN 10
+  #define XM_TX_MAC_STAT_WIDTH 1
+  #define XM_MCNTL_PASS_LBN 8
+  #define XM_MCNTL_PASS_WIDTH 2
+  #define XM_REJ_CNTL_UCAST_LBN 6
+  #define XM_REJ_CNTL_UCAST_WIDTH 1
+  #define XM_REJ_CNTL_MCAST_LBN 5
+  #define XM_REJ_CNTL_MCAST_WIDTH 1
+  #define XM_AUTO_XMIT_ZPAUSE_LBN 4
+  #define XM_AUTO_XMIT_ZPAUSE_WIDTH 1
+  #define XM_AUTO_XMIT_PAUSE_LBN 3
+  #define XM_AUTO_XMIT_PAUSE_WIDTH 1
+  #define XM_ZPAUSE_LBN 2
+  #define XM_ZPAUSE_WIDTH 1
+  #define XM_XMIT_PAUSE_LBN 1
+  #define XM_XMIT_PAUSE_WIDTH 1
+  #define XM_DIS_FCNTL_LBN 0
+  #define XM_DIS_FCNTL_WIDTH 1
+#define XM_PAUSE_TIME_REG_KER_P0_OFST 0x1290 /* XGMAC pause time register -
+						port 0 */
+#define XM_PAUSE_TIME_REG_P0_OFST 0x1290 /* XGMAC pause time register -
+					    port 0 */
+  #define XM_TX_PAUSE_CNT_LBN 16
+  #define XM_TX_PAUSE_CNT_WIDTH 16
+  #define XM_RX_PAUSE_CNT_LBN 0
+  #define XM_RX_PAUSE_CNT_WIDTH 16
+#define XM_TX_PARAM_REG_KER_P0_OFST 0x12D0 /* XGMAC transmit parameter
+					      register - port 0 */
+#define XM_TX_PARAM_REG_P0_OFST 0x12D0 /* XGMAC transmit parameter register -
+					  port 0 */
+  #define XM_TX_JUMBO_MODE_LBN 31
+  #define XM_TX_JUMBO_MODE_WIDTH 1
+  #define XM_MAX_TX_FRM_SIZE_LBN 16
+  #define XM_MAX_TX_FRM_SIZE_WIDTH 14
+  #define XM_PAD_CHAR_LBN 0
+  #define XM_PAD_CHAR_WIDTH 8
+#define XM_RX_PARAM_REG_KER_P0_OFST 0x12E0 /* XGMAC receive parameter
+					      register - port 0 */
+#define XM_RX_PARAM_REG_P0_OFST 0x12E0 /* XGMAC receive parameter register -
+					  port 0 */
+  #define XM_MAX_RX_FRM_SIZE_LBN 0
+  #define XM_MAX_RX_FRM_SIZE_WIDTH 14
+#define XX_PWR_RST_REG_KER_P0_OFST 0x1300 /* XGXS/XAUI powerdown/reset
+					     register */
+#define XX_PWR_RST_REG_P0_OFST 0x1300 /* XGXS/XAUI powerdown/reset register */
+  #define XX_PWRDND_SIG_LBN 31
+  #define XX_PWRDND_SIG_WIDTH 1
+  #define XX_PWRDNC_SIG_LBN 30
+  #define XX_PWRDNC_SIG_WIDTH 1
+  #define XX_PWRDNB_SIG_LBN 29
+  #define XX_PWRDNB_SIG_WIDTH 1
+  #define XX_PWRDNA_SIG_LBN 28
+  #define XX_PWRDNA_SIG_WIDTH 1
+  #define XX_SIM_MODE_LBN 27
+  #define XX_SIM_MODE_WIDTH 1
+  #define XX_RSTPLLCD_SIG_LBN 25
+  #define XX_RSTPLLCD_SIG_WIDTH 1
+  #define XX_RSTPLLAB_SIG_LBN 24
+  #define XX_RSTPLLAB_SIG_WIDTH 1
+  #define XX_RESETD_SIG_LBN 23
+  #define XX_RESETD_SIG_WIDTH 1
+  #define XX_RESETC_SIG_LBN 22
+  #define XX_RESETC_SIG_WIDTH 1
+  #define XX_RESETB_SIG_LBN 21
+  #define XX_RESETB_SIG_WIDTH 1
+  #define XX_RESETA_SIG_LBN 20
+  #define XX_RESETA_SIG_WIDTH 1
+  #define XX_RSTXGXSTX_SIG_LBN 18
+  #define XX_RSTXGXSTX_SIG_WIDTH 1
+  #define XX_RSTXGXSRX_SIG_LBN 17
+  #define XX_RSTXGXSRX_SIG_WIDTH 1
+  #define XX_SD_RST_ACT_LBN 16
+  #define XX_SD_RST_ACT_WIDTH 1
+  #define XX_PWRDND_EN_LBN 15
+  #define XX_PWRDND_EN_WIDTH 1
+  #define XX_PWRDNC_EN_LBN 14
+  #define XX_PWRDNC_EN_WIDTH 1
+  #define XX_PWRDNB_EN_LBN 13
+  #define XX_PWRDNB_EN_WIDTH 1
+  #define XX_PWRDNA_EN_LBN 12
+  #define XX_PWRDNA_EN_WIDTH 1
+  #define XX_RSTPLLCD_EN_LBN 9
+  #define XX_RSTPLLCD_EN_WIDTH 1
+  #define XX_RSTPLLAB_EN_LBN 8
+  #define XX_RSTPLLAB_EN_WIDTH 1
+  #define XX_RESETD_EN_LBN 7
+  #define XX_RESETD_EN_WIDTH 1
+  #define XX_RESETC_EN_LBN 6
+  #define XX_RESETC_EN_WIDTH 1
+  #define XX_RESETB_EN_LBN 5
+  #define XX_RESETB_EN_WIDTH 1
+  #define XX_RESETA_EN_LBN 4
+  #define XX_RESETA_EN_WIDTH 1
+  #define XX_RSTXGXSTX_EN_LBN 2
+  #define XX_RSTXGXSTX_EN_WIDTH 1
+  #define XX_RSTXGXSRX_EN_LBN 1
+  #define XX_RSTXGXSRX_EN_WIDTH 1
+  #define XX_RST_XX_EN_LBN 0
+  #define XX_RST_XX_EN_WIDTH 1
+#define XX_SD_CTL_REG_KER_P0_OFST 0x1310 /* XGXS/XAUI powerdown/reset control
+					    register */
+#define XX_SD_CTL_REG_P0_OFST 0x1310 /* XGXS/XAUI powerdown/reset control
+					register */
+  #define XX_TERMADJ1_LBN 17
+  #define XX_TERMADJ1_WIDTH 1
+  #define XX_TERMADJ0_LBN 16
+  #define XX_TERMADJ0_WIDTH 1
+  #define XX_HIDRVD_LBN 15
+  #define XX_HIDRVD_WIDTH 1
+  #define XX_LODRVD_LBN 14
+  #define XX_LODRVD_WIDTH 1
+  #define XX_HIDRVC_LBN 13
+  #define XX_HIDRVC_WIDTH 1
+  #define XX_LODRVC_LBN 12
+  #define XX_LODRVC_WIDTH 1
+  #define XX_HIDRVB_LBN 11
+  #define XX_HIDRVB_WIDTH 1
+  #define XX_LODRVB_LBN 10
+  #define XX_LODRVB_WIDTH 1
+  #define XX_HIDRVA_LBN 9
+  #define XX_HIDRVA_WIDTH 1
+  #define XX_LODRVA_LBN 8
+  #define XX_LODRVA_WIDTH 1
+  #define XX_LPBKD_LBN 3
+  #define XX_LPBKD_WIDTH 1
+  #define XX_LPBKC_LBN 2
+  #define XX_LPBKC_WIDTH 1
+  #define XX_LPBKB_LBN 1
+  #define XX_LPBKB_WIDTH 1
+  #define XX_LPBKA_LBN 0
+  #define XX_LPBKA_WIDTH 1
+#define XX_TXDRV_CTL_REG_KER_P0_OFST 0x1320 /* XAUI SerDes transmit drive
+					       control register */
+#define XX_TXDRV_CTL_REG_P0_OFST 0x1320 /* XAUI SerDes transmit drive
+					   control register */
+  #define XX_DEQD_LBN 28
+  #define XX_DEQD_WIDTH 4
+  #define XX_DEQC_LBN 24
+  #define XX_DEQC_WIDTH 4
+  #define XX_DEQB_LBN 20
+  #define XX_DEQB_WIDTH 4
+  #define XX_DEQA_LBN 16
+  #define XX_DEQA_WIDTH 4
+  #define XX_DTXD_LBN 12
+  #define XX_DTXD_WIDTH 4
+  #define XX_DTXC_LBN 8
+  #define XX_DTXC_WIDTH 4
+  #define XX_DTXB_LBN 4
+  #define XX_DTXB_WIDTH 4
+  #define XX_DTXA_LBN 0
+  #define XX_DTXA_WIDTH 4
+#define XX_PRBS_CTL_REG_KER_P0_OFST 0x1330 /* XAUI PRBS control register */
+#define XX_PRBS_CTL_REG_P0_OFST 0x1330 /* XAUI PRBS control register */
+  #define XX_CH3_RX_PRBS_SEL_LBN 30
+  #define XX_CH3_RX_PRBS_SEL_WIDTH 2
+  #define XX_CH3_RX_PRBS_INV_LBN 29
+  #define XX_CH3_RX_PRBS_INV_WIDTH 1
+  #define XX_CH3_RX_PRBS_CHKEN_LBN 28
+  #define XX_CH3_RX_PRBS_CHKEN_WIDTH 1
+  #define XX_CH2_RX_PRBS_SEL_LBN 26
+  #define XX_CH2_RX_PRBS_SEL_WIDTH 2
+  #define XX_CH2_RX_PRBS_INV_LBN 25
+  #define XX_CH2_RX_PRBS_INV_WIDTH 1
+  #define XX_CH2_RX_PRBS_CHKEN_LBN 24
+  #define XX_CH2_RX_PRBS_CHKEN_WIDTH 1
+  #define XX_CH1_RX_PRBS_SEL_LBN 22
+  #define XX_CH1_RX_PRBS_SEL_WIDTH 2
+  #define XX_CH1_RX_PRBS_INV_LBN 21
+  #define XX_CH1_RX_PRBS_INV_WIDTH 1
+  #define XX_CH1_RX_PRBS_CHKEN_LBN 20
+  #define XX_CH1_RX_PRBS_CHKEN_WIDTH 1
+  #define XX_CH0_RX_PRBS_SEL_LBN 18
+  #define XX_CH0_RX_PRBS_SEL_WIDTH 2
+  #define XX_CH0_RX_PRBS_INV_LBN 17
+  #define XX_CH0_RX_PRBS_INV_WIDTH 1
+  #define XX_CH0_RX_PRBS_CHKEN_LBN 16
+  #define XX_CH0_RX_PRBS_CHKEN_WIDTH 1
+  #define XX_CH3_TX_PRBS_SEL_LBN 14
+  #define XX_CH3_TX_PRBS_SEL_WIDTH 2
+  #define XX_CH3_TX_PRBS_INV_LBN 13
+  #define XX_CH3_TX_PRBS_INV_WIDTH 1
+  #define XX_CH3_TX_PRBS_CHKEN_LBN 12
+  #define XX_CH3_TX_PRBS_CHKEN_WIDTH 1
+  #define XX_CH2_TX_PRBS_SEL_LBN 10
+  #define XX_CH2_TX_PRBS_SEL_WIDTH 2
+  #define XX_CH2_TX_PRBS_INV_LBN 9
+  #define XX_CH2_TX_PRBS_INV_WIDTH 1
+  #define XX_CH2_TX_PRBS_CHKEN_LBN 8
+  #define XX_CH2_TX_PRBS_CHKEN_WIDTH 1
+  #define XX_CH1_TX_PRBS_SEL_LBN 6
+  #define XX_CH1_TX_PRBS_SEL_WIDTH 2
+  #define XX_CH1_TX_PRBS_INV_LBN 5
+  #define XX_CH1_TX_PRBS_INV_WIDTH 1
+  #define XX_CH1_TX_PRBS_CHKEN_LBN 4
+  #define XX_CH1_TX_PRBS_CHKEN_WIDTH 1
+  #define XX_CH0_TX_PRBS_SEL_LBN 2
+  #define XX_CH0_TX_PRBS_SEL_WIDTH 2
+  #define XX_CH0_TX_PRBS_INV_LBN 1
+  #define XX_CH0_TX_PRBS_INV_WIDTH 1
+  #define XX_CH0_TX_PRBS_CHKEN_LBN 0
+  #define XX_CH0_TX_PRBS_CHKEN_WIDTH 1
+#define XX_PRBS_CHK_REG_KER_P0_OFST 0x1340 /* XAUI PRBS checker control
+					      register */
+#define XX_PRBS_CHK_REG_P0_OFST 0x1340 /* XAUI PRBS checker control
+					  register */
+  #define XX_REV_LB_EN_LBN 16
+  #define XX_REV_LB_EN_WIDTH 1
+  #define XX_CH3_DEG_DET_LBN 15
+  #define XX_CH3_DEG_DET_WIDTH 1
+  #define XX_CH3_LFSR_LOCK_IND_LBN 14
+  #define XX_CH3_LFSR_LOCK_IND_WIDTH 1
+  #define XX_CH3_PRBS_FRUN_LBN 13
+  #define XX_CH3_PRBS_FRUN_WIDTH 1
+  #define XX_CH3_ERR_CHK_LBN 12
+  #define XX_CH3_ERR_CHK_WIDTH 1
+  #define XX_CH2_DEG_DET_LBN 11
+  #define XX_CH2_DEG_DET_WIDTH 1
+  #define XX_CH2_LFSR_LOCK_IND_LBN 10
+  #define XX_CH2_LFSR_LOCK_IND_WIDTH 1
+  #define XX_CH2_PRBS_FRUN_LBN 9
+  #define XX_CH2_PRBS_FRUN_WIDTH 1
+  #define XX_CH2_ERR_CHK_LBN 8
+  #define XX_CH2_ERR_CHK_WIDTH 1
+  #define XX_CH1_DEG_DET_LBN 7
+  #define XX_CH1_DEG_DET_WIDTH 1
+  #define XX_CH1_LFSR_LOCK_IND_LBN 6
+  #define XX_CH1_LFSR_LOCK_IND_WIDTH 1
+  #define XX_CH1_PRBS_FRUN_LBN 5
+  #define XX_CH1_PRBS_FRUN_WIDTH 1
+  #define XX_CH1_ERR_CHK_LBN 4
+  #define XX_CH1_ERR_CHK_WIDTH 1
+  #define XX_CH0_DEG_DET_LBN 3
+  #define XX_CH0_DEG_DET_WIDTH 1
+  #define XX_CH0_LFSR_LOCK_IND_LBN 2
+  #define XX_CH0_LFSR_LOCK_IND_WIDTH 1
+  #define XX_CH0_PRBS_FRUN_LBN 1
+  #define XX_CH0_PRBS_FRUN_WIDTH 1
+  #define XX_CH0_ERR_CHK_LBN 0
+  #define XX_CH0_ERR_CHK_WIDTH 1
+#define XX_PRBS_ERR_REG_KER_P0_OFST 0x1350 /* XAUI PRBS checker error
+					      count register */
+#define XX_PRBS_ERR_REG_P0_OFST 0x1350 /* XAUI PRBS checker error count
+					  register */
+  #define XX_CH3_PRBS_ERR_CNT_LBN 24
+  #define XX_CH3_PRBS_ERR_CNT_WIDTH 8
+  #define XX_CH2_PRBS_ERR_CNT_LBN 16
+  #define XX_CH2_PRBS_ERR_CNT_WIDTH 8
+  #define XX_CH1_PRBS_ERR_CNT_LBN 8
+  #define XX_CH1_PRBS_ERR_CNT_WIDTH 8
+  #define XX_CH0_PRBS_ERR_CNT_LBN 0
+  #define XX_CH0_PRBS_ERR_CNT_WIDTH 8
+#define XX_CORE_STAT_REG_KER_P0_OFST 0x1360 /* XAUI XGXS core status
+					       register */
+#define XX_CORE_STAT_REG_P0_OFST 0x1360 /* XAUI XGXS core status register */
+  #define XX_FORCE_SIG3_LBN 31
+  #define XX_FORCE_SIG3_WIDTH 1
+  #define XX_FORCE_SIG3_VAL_LBN 30
+  #define XX_FORCE_SIG3_VAL_WIDTH 1
+  #define XX_FORCE_SIG2_LBN 29
+  #define XX_FORCE_SIG2_WIDTH 1
+  #define XX_FORCE_SIG2_VAL_LBN 28
+  #define XX_FORCE_SIG2_VAL_WIDTH 1
+  #define XX_FORCE_SIG1_LBN 27
+  #define XX_FORCE_SIG1_WIDTH 1
+  #define XX_FORCE_SIG1_VAL_LBN 26
+  #define XX_FORCE_SIG1_VAL_WIDTH 1
+  #define XX_FORCE_SIG0_LBN 25
+  #define XX_FORCE_SIG0_WIDTH 1
+  #define XX_FORCE_SIG0_VAL_LBN 24
+  #define XX_FORCE_SIG0_VAL_WIDTH 1
+  #define XX_XGXS_LB_EN_LBN 23
+  #define XX_XGXS_LB_EN_WIDTH 1
+  #define XX_XGMII_LB_EN_LBN 22
+  #define XX_XGMII_LB_EN_WIDTH 1
+  #define XX_MATCH_FAULT_LBN 21
+  #define XX_MATCH_FAULT_WIDTH 1
+  #define XX_ALIGN_DONE_LBN 20
+  #define XX_ALIGN_DONE_WIDTH 1
+  #define XX_SYNC_STAT3_LBN 19
+  #define XX_SYNC_STAT3_WIDTH 1
+  #define XX_SYNC_STAT2_LBN 18
+  #define XX_SYNC_STAT2_WIDTH 1
+  #define XX_SYNC_STAT1_LBN 17
+  #define XX_SYNC_STAT1_WIDTH 1
+  #define XX_SYNC_STAT0_LBN 16
+  #define XX_SYNC_STAT0_WIDTH 1
+  #define XX_COMMA_DET_CH3_LBN 15
+  #define XX_COMMA_DET_CH3_WIDTH 1
+  #define XX_COMMA_DET_CH2_LBN 14
+  #define XX_COMMA_DET_CH2_WIDTH 1
+  #define XX_COMMA_DET_CH1_LBN 13
+  #define XX_COMMA_DET_CH1_WIDTH 1
+  #define XX_COMMA_DET_CH0_LBN 12
+  #define XX_COMMA_DET_CH0_WIDTH 1
+  #define XX_CGRP_ALIGN_CH3_LBN 11
+  #define XX_CGRP_ALIGN_CH3_WIDTH 1
+  #define XX_CGRP_ALIGN_CH2_LBN 10
+  #define XX_CGRP_ALIGN_CH2_WIDTH 1
+  #define XX_CGRP_ALIGN_CH1_LBN 9
+  #define XX_CGRP_ALIGN_CH1_WIDTH 1
+  #define XX_CGRP_ALIGN_CH0_LBN 8
+  #define XX_CGRP_ALIGN_CH0_WIDTH 1
+  #define XX_CHAR_ERR_CH3_LBN 7
+  #define XX_CHAR_ERR_CH3_WIDTH 1
+  #define XX_CHAR_ERR_CH2_LBN 6
+  #define XX_CHAR_ERR_CH2_WIDTH 1
+  #define XX_CHAR_ERR_CH1_LBN 5
+  #define XX_CHAR_ERR_CH1_WIDTH 1
+  #define XX_CHAR_ERR_CH0_LBN 4
+  #define XX_CHAR_ERR_CH0_WIDTH 1
+  #define XX_DISPERR_CH3_LBN 3
+  #define XX_DISPERR_CH3_WIDTH 1
+  #define XX_DISPERR_CH2_LBN 2
+  #define XX_DISPERR_CH2_WIDTH 1
+  #define XX_DISPERR_CH1_LBN 1
+  #define XX_DISPERR_CH1_WIDTH 1
+  #define XX_DISPERR_CH0_LBN 0
+  #define XX_DISPERR_CH0_WIDTH 1
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_xgrmon.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_xgrmon.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_xgrmon.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon/falcon_xgrmon.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,125 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides EtherFabric NIC - EFXXXX (aka Falcon) 10G MAC
+ * statistics register definitions.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*************---- 10G MAC Statistical Counters C Header ----*************/
+#define XgRxOctets_offset 0x0
+    #define XgRxOctets_WIDTH 48
+#define XgRxOctetsOK_offset 0x8
+    #define XgRxOctetsOK_WIDTH 48
+#define XgRxPkts_offset 0x10
+    #define XgRxPkts_WIDTH 32
+#define XgRxPktsOK_offset 0x14
+    #define XgRxPktsOK_WIDTH 32
+#define XgRxBroadcastPkts_offset 0x18
+    #define XgRxBroadcastPkts_WIDTH 32
+#define XgRxMulticastPkts_offset 0x1C
+    #define XgRxMulticastPkts_WIDTH 32
+#define XgRxUnicastPkts_offset 0x20
+    #define XgRxUnicastPkts_WIDTH 32
+#define XgRxUndersizePkts_offset 0x24
+    #define XgRxUndersizePkts_WIDTH 32
+#define XgRxOversizePkts_offset 0x28
+    #define XgRxOversizePkts_WIDTH 32
+#define XgRxJabberPkts_offset 0x2C
+    #define XgRxJabberPkts_WIDTH 32
+#define XgRxUndersizeFCSerrorPkts_offset 0x30
+    #define XgRxUndersizeFCSerrorPkts_WIDTH 32
+#define XgRxDropEvents_offset 0x34
+    #define XgRxDropEvents_WIDTH 32
+#define XgRxFCSerrorPkts_offset 0x38
+    #define XgRxFCSerrorPkts_WIDTH 32
+#define XgRxAlignError_offset 0x3C
+    #define XgRxAlignError_WIDTH 32
+#define XgRxSymbolError_offset 0x40
+    #define XgRxSymbolError_WIDTH 32
+#define XgRxInternalMACError_offset 0x44
+    #define XgRxInternalMACError_WIDTH 32
+#define XgRxControlPkts_offset 0x48
+    #define XgRxControlPkts_WIDTH 32
+#define XgRxPausePkts_offset 0x4C
+    #define XgRxPausePkts_WIDTH 32
+#define XgRxPkts64Octets_offset 0x50
+    #define XgRxPkts64Octets_WIDTH 32
+#define XgRxPkts65to127Octets_offset 0x54
+    #define XgRxPkts65to127Octets_WIDTH 32
+#define XgRxPkts128to255Octets_offset 0x58
+    #define XgRxPkts128to255Octets_WIDTH 32
+#define XgRxPkts256to511Octets_offset 0x5C
+    #define XgRxPkts256to511Octets_WIDTH 32
+#define XgRxPkts512to1023Octets_offset 0x60
+    #define XgRxPkts512to1023Octets_WIDTH 32
+#define XgRxPkts1024to15xxOctets_offset 0x64
+    #define XgRxPkts1024to15xxOctets_WIDTH 32
+#define XgRxPkts15xxtoMaxOctets_offset 0x68
+    #define XgRxPkts15xxtoMaxOctets_WIDTH 32
+#define XgRxLengthError_offset 0x6C
+    #define XgRxLengthError_WIDTH 32
+#define XgTxPkts_offset 0x80
+    #define XgTxPkts_WIDTH 32
+#define XgTxOctets_offset 0x88
+    #define XgTxOctets_WIDTH 48
+#define XgTxMulticastPkts_offset 0x90
+    #define XgTxMulticastPkts_WIDTH 32
+#define XgTxBroadcastPkts_offset 0x94
+    #define XgTxBroadcastPkts_WIDTH 32
+#define XgTxUnicastPkts_offset 0x98
+    #define XgTxUnicastPkts_WIDTH 32
+#define XgTxControlPkts_offset 0x9C
+    #define XgTxControlPkts_WIDTH 32
+#define XgTxPausePkts_offset 0xA0
+    #define XgTxPausePkts_WIDTH 32
+#define XgTxPkts64Octets_offset 0xA4
+    #define XgTxPkts64Octets_WIDTH 32
+#define XgTxPkts65to127Octets_offset 0xA8
+    #define XgTxPkts65to127Octets_WIDTH 32
+#define XgTxPkts128to255Octets_offset 0xAC
+    #define XgTxPkts128to255Octets_WIDTH 32
+#define XgTxPkts256to511Octets_offset 0xB0
+    #define XgTxPkts256to511Octets_WIDTH 32
+#define XgTxPkts512to1023Octets_offset 0xB4
+    #define XgTxPkts512to1023Octets_WIDTH 32
+#define XgTxPkts1024to15xxOctets_offset 0xB8
+    #define XgTxPkts1024to15xxOctets_WIDTH 32
+#define XgTxPkts1519toMaxOctets_offset 0xBC
+    #define XgTxPkts1519toMaxOctets_WIDTH 32
+#define XgTxUndersizePkts_offset 0xC0
+    #define XgTxUndersizePkts_WIDTH 32
+#define XgTxOversizePkts_offset 0xC4
+    #define XgTxOversizePkts_WIDTH 32
+#define xGTxNonTcpUdpPkt_offset 0xC8
+    #define xGTxNonTcpUdpPkt_WIDTH 16
+#define xGTxMacSrcErrPkt_offset 0xCC
+    #define xGTxMacSrcErrPkt_WIDTH 16
+#define xGTxIpSrcErrPkt_offset 0xD0
+    #define xGTxIpSrcErrPkt_WIDTH 16
+#define XgDmaDone_offset 0xD4
+    #define XgDmaDone_WIDTH 32
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/falcon.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,420 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides EtherFabric NIC - EFXXXX (aka Falcon) specific
+ * definitions.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_DRIVER_EFAB_HARDWARE_FALCON_H__
+#define __CI_DRIVER_EFAB_HARDWARE_FALCON_H__
+
+/*----------------------------------------------------------------------------
+ * Compile options
+ *---------------------------------------------------------------------------*/
+
+/* Falcon has an 8K maximum page size. */
+#define FALCON_MAX_PAGE_SIZE EFHW_8K
+
+/* include the register definitions */
+#include <ci/driver/efab/hardware/falcon/falcon_core.h>
+#include <ci/driver/efab/hardware/falcon/falcon_desc.h>
+#include <ci/driver/efab/hardware/falcon/falcon_event.h>
+#include <ci/driver/efab/hardware/falcon/falcon_mac.h>
+#include <ci/driver/efab/hardware/falcon/falcon_grmon.h>
+#include <ci/driver/efab/hardware/falcon/falcon_xgrmon.h>
+#include <ci/driver/efab/hardware/falcon/falcon_intr_vec.h>
+
+#define FALCON_DMA_TX_DESC_BYTES	8
+#define FALCON_DMA_RX_PHYS_DESC_BYTES	8
+#define FALCON_DMA_RX_BUF_DESC_BYTES	4
+
+
+/* ---- efhw_event_t helpers --- */
+
+#ifndef EFHW_IS_LITTLE_ENDIAN
+#error This needs lots of cpu_to_le64s() in
+#endif
+
+/*!\ TODO look at whether there is an efficiency gain to be had by
+  treating the event codes to 32bit masks as is done for EF1
+
+  These masks apply to the full 64 bits of the event to extract the
+  event code - followed by the common event codes to expect
+ */
+#define __FALCON_OPEN_MASK(WIDTH)  ((((uint64_t)1) << (WIDTH)) - 1)
+#define FALCON_EVENT_CODE_MASK \
+	(__FALCON_OPEN_MASK(EV_CODE_WIDTH) << EV_CODE_LBN)
+#define FALCON_EVENT_EV_Q_ID_MASK \
+	(__FALCON_OPEN_MASK(DRIVER_EV_EVQ_ID_WIDTH) << DRIVER_EV_EVQ_ID_LBN)
+#define FALCON_EVENT_TX_FLUSH_Q_ID_MASK \
+	(__FALCON_OPEN_MASK(DRIVER_EV_TX_DESCQ_ID_WIDTH) << \
+	 DRIVER_EV_TX_DESCQ_ID_LBN)
+#define FALCON_EVENT_RX_FLUSH_Q_ID_MASK \
+	(__FALCON_OPEN_MASK(DRIVER_EV_RX_DESCQ_ID_WIDTH) << \
+	 DRIVER_EV_RX_DESCQ_ID_LBN)
+#define FALCON_EVENT_DRV_SUBCODE_MASK \
+	(__FALCON_OPEN_MASK(DRIVER_EV_SUB_CODE_WIDTH) << \
+	 DRIVER_EV_SUB_CODE_LBN)
+
+#define FALCON_EVENT_FMT         "[ev:%x:%08x:%08x]"
+#define FALCON_EVENT_PRI_ARG(e) \
+	((unsigned)(((e).u64 & FALCON_EVENT_CODE_MASK) >> EV_CODE_LBN)), \
+	((unsigned)((e).u64 >> 32)), ((unsigned)((e).u64 & 0xFFFFFFFF))
+
+#define FALCON_EVENT_CODE(evp)		((evp)->u64 & FALCON_EVENT_CODE_MASK)
+#define FALCON_EVENT_WAKE_EVQ_ID(evp) \
+	(((evp)->u64 & FALCON_EVENT_EV_Q_ID_MASK) >> DRIVER_EV_EVQ_ID_LBN)
+#define FALCON_EVENT_TX_FLUSH_Q_ID(evp) \
+	(((evp)->u64 & FALCON_EVENT_TX_FLUSH_Q_ID_MASK) >> \
+	 DRIVER_EV_TX_DESCQ_ID_LBN)
+#define FALCON_EVENT_RX_FLUSH_Q_ID(evp) \
+	(((evp)->u64 & FALCON_EVENT_RX_FLUSH_Q_ID_MASK) >> \
+	 DRIVER_EV_RX_DESCQ_ID_LBN)
+#define FALCON_EVENT_DRIVER_SUBCODE(evp) \
+	(((evp)->u64 & FALCON_EVENT_DRV_SUBCODE_MASK) >> \
+	 DRIVER_EV_SUB_CODE_LBN)
+
+#define FALCON_EVENT_CODE_CHAR	((uint64_t)DRIVER_EV_DECODE << EV_CODE_LBN)
+#define FALCON_EVENT_CODE_SW	((uint64_t)DRV_GEN_EV_DECODE << EV_CODE_LBN)
+
+
+/* so this is the size in bytes of an awful lot of things */
+#define FALCON_REGISTER128          (16)
+
+/* we define some unique dummy values as a debug aid */
+#ifdef _WIN32
+#define FALCON_ATOMIC_BASE		0xdeadbeef00000000ui64
+#else
+#define FALCON_ATOMIC_BASE		0xdeadbeef00000000ULL
+#endif
+#define FALCON_ATOMIC_UPD_REG		(FALCON_ATOMIC_BASE | 0x1)
+#define FALCON_ATOMIC_PTR_TBL_REG	(FALCON_ATOMIC_BASE | 0x2)
+#define FALCON_ATOMIC_SRPM_UDP_EVQ_REG	(FALCON_ATOMIC_BASE | 0x3)
+#define FALCON_ATOMIC_RX_FLUSH_DESCQ	(FALCON_ATOMIC_BASE | 0x4)
+#define FALCON_ATOMIC_TX_FLUSH_DESCQ	(FALCON_ATOMIC_BASE | 0x5)
+#define FALCON_ATOMIC_INT_EN_REG	(FALCON_ATOMIC_BASE | 0x6)
+#define FALCON_ATOMIC_TIMER_CMD_REG	(FALCON_ATOMIC_BASE | 0x7)
+#define FALCON_ATOMIC_PACE_REG		(FALCON_ATOMIC_BASE | 0x8)
+#define FALCON_ATOMIC_INT_ACK_REG	(FALCON_ATOMIC_BASE | 0x9)
+/* XXX It crashed with odd value in FALCON_ATOMIC_INT_ADR_REG */
+#define FALCON_ATOMIC_INT_ADR_REG	(FALCON_ATOMIC_BASE | 0xa)
+
+/*----------------------------------------------------------------------------
+ *
+ * PCI control blocks for Falcon -
+ *          (P) primary is for NET
+ *          (S) secondary is for CHAR
+ *
+ *---------------------------------------------------------------------------*/
+
+#define FALCON_P_CTR_AP_BAR	2
+#define FALCON_S_CTR_AP_BAR	0
+#define FALCON_S_DEVID		0x6703
+
+
+/*----------------------------------------------------------------------------
+ *
+ * Falcon constants
+ *
+ *---------------------------------------------------------------------------*/
+
+#define FALCON_DMAQ_NUM		(EFHW_4K)
+#define FALCON_EVQ_TBL_NUM	(EFHW_4K)
+#define FALCON_TIMERS_NUM	(EFHW_4K)
+
+/* This value is an upper limit on the total number of filter table
+ * entries, including odd and even banks.  The actual size of filter table
+ * is determined at runtime, as it can vary.
+ */
+#define FALCON_FILTER_TBL_NUM		(EFHW_16K)
+
+/* max number of buffers which can be pushed before commiting */
+#define FALCON_BUFFER_UPD_MAX		(128)
+
+/* We can tell falcon to write its RX buffers in 32 byte quantums,
+   and since we pad packets 2 bytes to the right we can't use
+   a full page (not unless we use jumbo mode for all queues)
+
+   NOTE: tests/nic/dma.c assumes that the value here is the real NIC
+   value, so we explicitly round it down to the nearest 32 bytes */
+
+/* #define FALCON_RX_USR_BUF_SIZE    round_down(4096-2,32) */
+#define FALCON_RX_USR_BUF_SIZE		4064
+
+#define FALCON_EVQ_RPTR_REG_P0		0x400
+
+/*----------------------------------------------------------------------------
+ *
+ * Falcon requires user-space descriptor pushes to be:
+ *    dword[0-2]; wiob(); dword[3]
+ *
+ * Driver register access must be locked against other threads from
+ * the same driver but can be in any order: i.e dword[0-3]; wiob()
+ *
+ * The following helpers ensure that valid dword orderings are exercised
+ *
+ *---------------------------------------------------------------------------*/
+
+/* A union to allow writting 64bit values as 32bit values, without
+ * hitting the compilers aliasing rules. We hope the compiler optimises
+ * away the copy's anyway */
+union __u64to32 {
+	uint64_t u64;
+	struct {
+#ifdef EFHW_IS_LITTLE_ENDIAN
+		uint32_t a;
+		uint32_t b;
+#else
+		uint32_t b;
+		uint32_t a;
+#endif
+	} s;
+};
+
+static inline void
+falcon_write_ddd_d(efhw_ioaddr_t kva,
+		   uint32_t d0, uint32_t d1, uint32_t d2, uint32_t d3)
+{
+	writel(d0, kva + 0);
+	writel(d1, kva + 4);
+	writel(d2, kva + 8);
+	mmiowb();
+	writel(d3, kva + 12);
+}
+
+static inline void falcon_write_q(efhw_ioaddr_t kva, uint64_t q)
+{
+	union __u64to32 u;
+	u.u64 = q;
+
+	writel(u.s.a, kva);
+	mmiowb();
+	writel(u.s.b, kva + 4);
+}
+
+static inline void falcon_read_q(efhw_ioaddr_t addr, uint64_t *q0)
+{
+	/* It is essential that we read dword0 first, so that
+	 * the shadow register is updated with the latest value
+	 * and we get a self consistent value.
+	 */
+	union __u64to32 u;
+	u.s.a = readl(addr);
+	rmb();
+	u.s.b = readl(addr + 4);
+
+	*q0 = u.u64;
+}
+
+static inline void
+falcon_write_qq(efhw_ioaddr_t kva, uint64_t q0, uint64_t q1)
+{
+	writeq(q0, kva + 0);
+	falcon_write_q(kva + 8, q1);
+}
+
+static inline void
+falcon_read_qq(efhw_ioaddr_t addr, uint64_t *q0, uint64_t *q1)
+{
+	falcon_read_q(addr, q0);
+	*q1 = readq(addr + 8);
+}
+
+
+
+/*----------------------------------------------------------------------------
+ *
+ * Buffer virtual addresses (4K buffers)
+ *
+ *---------------------------------------------------------------------------*/
+
+/* Form a buffer virtual address from buffer ID and offset.  If the offset
+** is larger than the buffer size, then the buffer indexed will be
+** calculated appropriately.  It is the responsibility of the caller to
+** ensure that they have valid buffers programmed at that address.
+*/
+#define FALCON_VADDR_8K_S	(13)
+#define FALCON_VADDR_4K_S	(12)
+#define FALCON_VADDR_M		0xfffff	/* post shift mask  */
+
+#define FALCON_BUFFER_8K_ADDR(id, off)	(((id) << FALCON_VADDR_8K_S) + (off))
+#define FALCON_BUFFER_8K_PAGE(vaddr) \
+	(((vaddr) >> FALCON_VADDR_8K_S) & FALCON_VADDR_M)
+#define FALCON_BUFFER_8K_OFF(vaddr) \
+	((vaddr) & __FALCON_MASK32(FALCON_VADDR_8K_S))
+
+#define FALCON_BUFFER_4K_ADDR(id, off)	(((id) << FALCON_VADDR_4K_S) + (off))
+#define FALCON_BUFFER_4K_PAGE(vaddr) \
+	(((vaddr) >> FALCON_VADDR_4K_S) & FALCON_VADDR_M)
+#define FALCON_BUFFER_4K_OFF(vaddr) \
+	((vaddr) & __FALCON_MASK32(FALCON_VADDR_4K_S))
+
+/*----------------------------------------------------------------------------
+ *
+ * Timer helpers
+ *
+ *---------------------------------------------------------------------------*/
+
+static inline int falcon_timer_page_addr(uint idx)
+{
+
+	EFHW_ASSERT(TIMER_CMD_REG_KER_OFST ==
+		    (TIMER_CMD_REG_PAGE4_OFST - 4 * EFHW_8K));
+
+	EFHW_ASSERT(idx < FALCON_TIMERS_NUM);
+
+	if (idx < 4)
+		return TIMER_CMD_REG_KER_OFST + (idx * EFHW_8K);
+	else if (idx < 1024)
+		return TIMER_CMD_REG_PAGE4_OFST + ((idx - 4) * EFHW_8K);
+	else
+		return TIMER_CMD_REG_PAGE123K_OFST + ((idx - 1024) * EFHW_8K);
+}
+
+#define FALCON_TIMER_PAGE_MASK		(EFHW_8K-1)
+
+static inline int falcon_timer_page_offset(uint idx)
+{
+	return falcon_timer_page_addr(idx) & FALCON_TIMER_PAGE_MASK;
+}
+
+/*----------------------------------------------------------------------------
+ *
+ * DMA Queue helpers
+ *
+ *---------------------------------------------------------------------------*/
+
+/* iSCSI queue for A1; see bug 5427 for more details. */
+#define FALCON_A1_ISCSI_DMAQ 4
+
+/*! returns an address within a bar of the TX DMA doorbell */
+static inline uint falcon_tx_dma_page_addr(uint dmaq_idx)
+{
+	uint page;
+
+	EFHW_ASSERT((((TX_DESC_UPD_REG_PAGE123K_OFST) & (EFHW_8K - 1)) ==
+		     (((TX_DESC_UPD_REG_PAGE4_OFST) & (EFHW_8K - 1)))));
+
+	EFHW_ASSERT(dmaq_idx < FALCON_DMAQ_NUM);
+
+	if (dmaq_idx < 1024)
+		page = TX_DESC_UPD_REG_PAGE4_OFST + ((dmaq_idx - 4) * EFHW_8K);
+	else
+		page =
+		    TX_DESC_UPD_REG_PAGE123K_OFST +
+		    ((dmaq_idx - 1024) * EFHW_8K);
+
+	return page;
+}
+
+/*! returns an address within a bar of the RX DMA doorbell */
+static inline uint falcon_rx_dma_page_addr(uint dmaq_idx)
+{
+	uint page;
+
+	EFHW_ASSERT((((RX_DESC_UPD_REG_PAGE123K_OFST) & (EFHW_8K - 1)) ==
+		     ((RX_DESC_UPD_REG_PAGE4_OFST) & (EFHW_8K - 1))));
+
+	EFHW_ASSERT(dmaq_idx < FALCON_DMAQ_NUM);
+
+	if (dmaq_idx < 1024)
+		page = RX_DESC_UPD_REG_PAGE4_OFST + ((dmaq_idx - 4) * EFHW_8K);
+	else
+		page =
+		    RX_DESC_UPD_REG_PAGE123K_OFST +
+		    ((dmaq_idx - 1024) * EFHW_8K);
+
+	return page;
+}
+
+/*! "page"=NIC-dependent register set size */
+#define FALCON_DMA_PAGE_MASK  (EFHW_8K-1)
+
+/*! returns an address within a bar of the start of the "page"
+    containing the TX DMA doorbell */
+static inline int falcon_tx_dma_page_base(uint dma_idx)
+{
+	return falcon_tx_dma_page_addr(dma_idx) & ~FALCON_DMA_PAGE_MASK;
+}
+
+/*! returns an address within a bar of the start of the "page"
+    containing the RX DMA doorbell */
+static inline int falcon_rx_dma_page_base(uint dma_idx)
+{
+	return falcon_rx_dma_page_addr(dma_idx) & ~FALCON_DMA_PAGE_MASK;
+}
+
+/*! returns an offset within a "page" of the TX DMA doorbell */
+static inline int falcon_tx_dma_page_offset(uint dma_idx)
+{
+	return falcon_tx_dma_page_addr(dma_idx) & FALCON_DMA_PAGE_MASK;
+}
+
+/*! returns an offset within a "page" of the RX DMA doorbell */
+static inline int falcon_rx_dma_page_offset(uint dma_idx)
+{
+	return falcon_rx_dma_page_addr(dma_idx) & FALCON_DMA_PAGE_MASK;
+}
+
+/*----------------------------------------------------------------------------
+ *
+ * Events
+ *
+ *---------------------------------------------------------------------------*/
+
+/* Falcon nails down the event queue mappings */
+#define FALCON_EVQ_KERNEL0   (0)	/* hardwired for net driver */
+#define FALCON_EVQ_CHAR      (4)	/* char driver's event queue      */
+#define FALCON_EVQ_NONIRQ    (5)	/* char driver's non interrupting
+					   queue. Subsequent queues are
+					   available for user apps */
+
+/* reserved by the drivers */
+#define FALCON_EVQ_TBL_RESERVED	   (8)
+
+/* default DMA-Q sizes */
+#define FALCON_DMA_Q_DEFAULT_TX_SIZE  512
+
+#define FALCON_DMA_Q_DEFAULT_RX_SIZE  512
+
+#define FALCON_DMA_Q_DEFAULT_MMAP \
+	(FALCON_DMA_Q_DEFAULT_TX_SIZE * (FALCON_DMA_TX_DESC_BYTES * 2))
+
+/*----------------------------------------------------------------------------
+ *
+ * DEBUG - Analyser trigger
+ *
+ *---------------------------------------------------------------------------*/
+
+static inline void falcon_deadbeef(efhw_ioaddr_t efhw_kva, unsigned what)
+{
+	writel(what, efhw_kva + 0x300);
+	mmiowb();
+}
+#endif /* __CI_DRIVER_EFAB_HARDWARE_FALCON_H__ */
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/workarounds.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/workarounds.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/workarounds.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware/workarounds.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,75 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides workaround settings for EtherFabric NICs.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_DRIVER_EFAB_WORKAROUNDS_H__
+#define __CI_DRIVER_EFAB_WORKAROUNDS_H__
+
+/*----------------------------------------------------------------------------
+ *
+ * Hardware workarounds which have global scope
+ *
+ *---------------------------------------------------------------------------*/
+
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+
+#if defined(__CI_HARDWARE_CONFIG_FALCON_B0__)
+/*------------------------------- B0 ---------------------------------------*/
+
+#define BUG2175_WORKAROUND 0	/* TX event batching for dual port operation.
+				   This removes the effect (dup TX events)
+				   of the fix
+				   (TX event per packet + batch events) */
+#define BUG5302_WORKAROUND 0	/* unstick TX DMAQ after out-of-range wr ptr */
+#define BUG5475_WORKAROUND 1	/* 10G SNAP encapsulation broken */
+#define BUG5762_WORKAROUND 0	/* Set all queues to jumbo mode */
+#define BUG5391_WORKAROUND 0	/* Misaligned TX can't span 512-byte boundary */
+#define BUG7916_WORKAROUND 0	/* RX flush gets lost */
+
+#else
+/*------------------------------- A0/A1 ------------------------------------*/
+
+#define BUG2175_WORKAROUND 1	/* TX event batching for dual port operation.
+				   This removes the effect (dup TX events)
+				   of the fix
+				   (TX event per packet + batch events) */
+#define BUG5302_WORKAROUND 1	/* unstick TX DMAQ after out-of-range wr ptr */
+#define BUG5475_WORKAROUND 1	/* 10G SNAP encapsulation broken */
+#define BUG5762_WORKAROUND 1	/* Set all queues to jumbo mode */
+#define BUG5391_WORKAROUND 1	/* Misaligned TX can't span 512-byte boundary */
+#define BUG7916_WORKAROUND 1	/* RX flush gets lost */
+
+#endif /* B0/A01 */
+
+#else
+# error Need hw support.
+#endif
+
+#endif /* __CI_DRIVER_EFAB_WORKAROUNDS_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/efab/hardware.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,199 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides EtherFabric NIC hardware interface.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_DRIVER_EFAB_HARDWARE_H__
+#define __CI_DRIVER_EFAB_HARDWARE_H__
+
+#include "ci/driver/efab/hardware/workarounds.h"
+#include <ci/efhw/hardware_sysdep.h>
+
+
+/*----------------------------------------------------------------------------
+ *
+ * Common EtherFabric definitions
+ *
+ *---------------------------------------------------------------------------*/
+
+#include <ci/efhw/debug.h>
+#include <ci/efhw/common.h>
+#include <ci/driver/efab/hardware/common.h>
+
+/*----------------------------------------------------------------------------
+ *
+ * EtherFabric varients
+ *
+ *---------------------------------------------------------------------------*/
+
+#include <ci/driver/efab/hardware/falcon.h>
+
+/*----------------------------------------------------------------------------
+ *
+ * EtherFabric Portable Hardware Layer defines
+ *
+ *---------------------------------------------------------------------------*/
+
+  /*-------------- Initialisation ------------ */
+#define efhw_nic_close_hardware(nic) \
+	((nic)->efhw_func->close_hardware(nic))
+
+#define efhw_nic_init_hardware(nic, ev_handlers, mac_addr) \
+	((nic)->efhw_func->init_hardware((nic), (ev_handlers), (mac_addr)))
+
+/*-------------- Interrupt support  ------------ */
+/** Handle interrupt.  Return 0 if not handled, 1 if handled. */
+#define efhw_nic_interrupt(nic) \
+	((nic)->efhw_func->interrupt(nic))
+
+#define efhw_nic_interrupt_enable(nic, index) \
+	((nic)->efhw_func->interrupt_enable(nic, index))
+
+#define efhw_nic_interrupt_disable(nic, index) \
+	((nic)->efhw_func->interrupt_disable(nic, index))
+
+#define efhw_nic_set_interrupt_moderation(nic, index, val) \
+	((nic)->efhw_func->set_interrupt_moderation(nic, index, val))
+
+/*-------------- Event support  ------------ */
+
+#define efhw_nic_event_queue_enable(nic, evq, size, q_base, buf_base) \
+	((nic)->efhw_func->event_queue_enable(nic, evq, size, q_base, \
+					      buf_base))
+
+#define efhw_nic_event_queue_disable(nic, evq, timer_only) \
+	((nic)->efhw_func->event_queue_disable(nic, evq, timer_only))
+
+#define efhw_nic_wakeup_request(nic, q_base, index, evq) \
+	((nic)->efhw_func->wakeup_request(nic, q_base, index, evq))
+
+#define efhw_nic_sw_event(nic, data, ev) \
+	((nic)->efhw_func->sw_event(nic, data, ev))
+
+/*-------------- Filter support  ------------ */
+#define efhw_nic_ipfilter_set(nic, type, index, dmaq,		\
+			      saddr, sport, daddr, dport)	\
+	((nic)->efhw_func->ipfilter_set(nic, type, index, dmaq,	\
+					saddr, sport, daddr, dport))
+
+#define efhw_nic_ipfilter_attach(nic, index, dmaq) \
+	((nic)->efhw_func->ipfilter_attach(nic, index, dmaq))
+
+#define efhw_nic_ipfilter_detach(nic, index) \
+	((nic)->efhw_func->ipfilter_detach(nic, index))
+
+#define efhw_nic_ipfilter_clear(nic, index) \
+	((nic)->efhw_func->ipfilter_clear(nic, index))
+
+/*-------------- DMA support  ------------ */
+#define efhw_nic_dmaq_tx_q_init(nic, dmaq, evq, owner, tag,		\
+				dmaq_size, index, flags)		\
+	((nic)->efhw_func->dmaq_tx_q_init(nic, dmaq, evq, owner, tag,	\
+					  dmaq_size, index, flags))
+
+#define efhw_nic_dmaq_rx_q_init(nic, dmaq, evq, owner, tag,		\
+				dmaq_size, index, flags) \
+	((nic)->efhw_func->dmaq_rx_q_init(nic, dmaq, evq, owner, tag,	\
+					  dmaq_size, index, flags))
+
+#define efhw_nic_dmaq_tx_q_disable(nic, dmaq) \
+	((nic)->efhw_func->dmaq_tx_q_disable(nic, dmaq))
+
+#define efhw_nic_dmaq_rx_q_disable(nic, dmaq) \
+	((nic)->efhw_func->dmaq_rx_q_disable(nic, dmaq))
+
+#define efhw_nic_flush_tx_dma_channel(nic, dmaq) \
+	((nic)->efhw_func->flush_tx_dma_channel(nic, dmaq))
+
+#define efhw_nic_flush_rx_dma_channel(nic, dmaq) \
+	((nic)->efhw_func->flush_rx_dma_channel(nic, dmaq))
+
+/*-------------- MAC Low level interface ---- */
+#define efhw_gmac_get_mac_addr(nic) \
+	((nic)->gmac->get_mac_addr((nic)->gmac))
+
+/*-------------- Buffer table -------------- */
+#define efhw_nic_buffer_table_set(nic, addr, bufsz, region,		\
+				  own_id, buf_id)			\
+	((nic)->efhw_func->buffer_table_set(nic, addr, bufsz, region,	\
+					    own_id, buf_id))
+
+#define efhw_nic_buffer_table_set_n(nic, buf_id, addr, bufsz,		\
+				    region, n_pages, own_id) \
+	((nic)->efhw_func->buffer_table_set_n(nic, buf_id, addr, bufsz,	\
+					      region, n_pages, own_id))
+
+#define efhw_nic_buffer_table_clear(nic, id, num) \
+	((nic)->efhw_func->buffer_table_clear(nic, id, num))
+
+#define efhw_nic_buffer_table_commit(nic) \
+	((nic)->efhw_func->buffer_table_commit(nic))
+
+/*----------------------------------------------------------------------------
+ * Hardware specific portability macros for performance critical code.
+ *
+ * Warning: and driver code which is using these defines is not
+ * capable of supporting multiple NIC varients and should be built and
+ * marked appropriately
+ *
+ *---------------------------------------------------------------------------*/
+
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+
+/* --- DMA --- */
+#define EFHW_DMA_ADDRMASK		(0xffffffffffffffffULL)
+
+/* --- Buffers --- */
+#define EFHW_BUFFER_ADDR		FALCON_BUFFER_4K_ADDR
+#define EFHW_BUFFER_PAGE		FALCON_BUFFER_4K_PAGE
+#define EFHW_BUFFER_OFF			FALCON_BUFFER_4K_OFF
+
+/* --- Filters --- */
+#define EFHW_IP_FILTER_NUM		FALCON_FILTER_TBL_NUM
+
+#define EFHW_MAX_PAGE_SIZE		FALCON_MAX_PAGE_SIZE
+
+#else
+# error no hardware definition found
+#endif
+
+#if PAGE_SIZE <= EFHW_MAX_PAGE_SIZE
+#define EFHW_NIC_PAGE_SIZE PAGE_SIZE
+#else
+#define EFHW_NIC_PAGE_SIZE EFHW_MAX_PAGE_SIZE
+#endif
+#define EFHW_NIC_PAGE_MASK (~(EFHW_NIC_PAGE_SIZE-1))
+
+#endif /* __CI_DRIVER_EFAB_HARDWARE_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/resource/efx_vi.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/resource/efx_vi.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/resource/efx_vi.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/resource/efx_vi.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,276 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains public EFX VI API to Solarflare resource manager.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_DRIVER_RESOURCE_EFX_VI_H__
+#define __CI_DRIVER_RESOURCE_EFX_VI_H__
+
+/* Default size of event queue in the efx_vi resource.  Copied from
+ * CI_CFG_NETIF_EVENTQ_SIZE */
+#define EFX_VI_EVENTQ_SIZE_DEFAULT 1024
+
+extern int efx_vi_eventq_size;
+
+/**************************************************************************
+ * efx_vi_state types, allocation and free
+ **************************************************************************/
+
+/*! Handle for refering to a efx_vi */
+struct efx_vi_state;
+
+/*!
+ * Allocate an efx_vi, including event queue and pt_endpoint
+ *
+ * \param vih_out Pointer to a handle that is set on success
+ * \param nic_index Index of NIC to apply this resource to
+ * \return Zero on success (and vih_out set), non-zero on failure.
+ */
+extern int
+efx_vi_alloc(struct efx_vi_state **vih_out, int nic_index);
+
+/*!
+ * Free a previously allocated efx_vi
+ *
+ * \param vih The handle of the efx_vi to free
+ */
+extern void
+efx_vi_free(struct efx_vi_state *vih);
+
+/*!
+ * Reset a previously allocated efx_vi
+ *
+ * \param vih The handle of the efx_vi to reset
+ */
+extern void
+efx_vi_reset(struct efx_vi_state *vih);
+
+/**************************************************************************
+ * efx_vi_eventq types and functions
+ **************************************************************************/
+
+/*!
+ * Register a function to receive callbacks when event queue timeouts
+ * or wakeups occur.  Only one function per efx_vi can be registered
+ * at once.
+ *
+ * \param vih The handle to identify the efx_vi
+ * \param callback The function to callback
+ * \param context An argument to pass to the callback function
+ * \return Zero on success, non-zero on failure.
+ */
+extern int
+efx_vi_eventq_register_callback(struct efx_vi_state *vih,
+				void (*callback)(void *context, int is_timeout),
+				void *context);
+
+/*!
+ * Remove the current eventq timeout or wakeup callback function
+ *
+ * \param vih The handle to identify the efx_vi
+ * \return Zero on success, non-zero on failure
+ */
+extern int
+efx_vi_eventq_kill_callback(struct efx_vi_state *vih);
+
+/**************************************************************************
+ * efx_vi_dma_map types and functions
+ **************************************************************************/
+
+/*!
+ * Handle for refering to a efx_vi
+ */
+struct efx_vi_dma_map_state;
+
+/*!
+ * Map a list of buffer pages so they are registered with the hardware
+ *
+ * \param vih The handle to identify the efx_vi
+ * \param addrs An array of page pointers to map
+ * \param n_addrs Length of the page pointer array.  Must be a power of two.
+ * \param dmh_out Set on success to a handle used to refer to this mapping
+ * \return Zero on success, non-zero on failure.
+ */
+extern int
+efx_vi_dma_map_pages(struct efx_vi_state *vih, struct page **pages,
+			 int n_pages, struct efx_vi_dma_map_state **dmh_out);
+extern int
+efx_vi_dma_map_addrs(struct efx_vi_state *vih,
+		     unsigned long long *dev_bus_addrs, int n_pages,
+		     struct efx_vi_dma_map_state **dmh_out);
+
+/*!
+ * Unmap a previously mapped set of pages so they are no longer registered
+ * with the hardware.
+ *
+ * \param vih The handle to identify the efx_vi
+ * \param dmh The handle to identify the dma mapping
+ */
+extern void
+efx_vi_dma_unmap_pages(struct efx_vi_state *vih,
+		       struct efx_vi_dma_map_state *dmh);
+extern void
+efx_vi_dma_unmap_addrs(struct efx_vi_state *vih,
+		       struct efx_vi_dma_map_state *dmh);
+
+/*!
+ * Retrieve the buffer address of the mapping
+ *
+ * \param vih The handle to identify the efx_vi
+ * \param dmh The handle to identify the buffer mapping
+ * \return The buffer address on success, or zero on failure
+ */
+extern unsigned
+efx_vi_dma_get_map_addr(struct efx_vi_state *vih,
+			struct efx_vi_dma_map_state *dmh);
+
+/**************************************************************************
+ * efx_vi filter functions
+ **************************************************************************/
+
+#define EFX_VI_STATIC_FILTERS 32
+
+/*! Handle to refer to a filter instance */
+struct filter_resource_t;
+
+/*!
+ * Allocate and add a filter
+ *
+ * \param vih The handle to identify the efx_vi
+ * \param protocol The protocol of the new filter: UDP or TCP
+ * \param ip_addr_be32 The local ip address of the filter
+ * \param port_le16 The local port of the filter
+ * \param fh_out Set on success to be a handle to refer to this filter
+ * \return Zero on success, non-zero on failure.
+ */
+extern int
+efx_vi_filter(struct efx_vi_state *vih, int protocol, unsigned ip_addr_be32,
+	      int port_le16, struct filter_resource_t **fh_out);
+
+/*!
+ * Remove a filter and free resources associated with it
+ *
+ * \param vih The handle to identify the efx_vi
+ * \param fh The handle to identify the filter
+ * \return Zero on success, non-zero on failure
+ */
+extern int
+efx_vi_filter_stop(struct efx_vi_state *vih, struct filter_resource_t *fh);
+
+/**************************************************************************
+ * efx_vi hw resources types and functions
+ **************************************************************************/
+
+/*! Constants for the type field in efx_vi_hw_resource */
+#define EFX_VI_HW_RESOURCE_TXDMAQ    0x0	/* PFN of TX DMA Q */
+#define EFX_VI_HW_RESOURCE_RXDMAQ    0x1	/* PFN of RX DMA Q */
+#define EFX_VI_HW_RESOURCE_TXBELL    0x2	/* PFN of TX Doorbell (EF1) */
+#define EFX_VI_HW_RESOURCE_RXBELL    0x3	/* PFN of RX Doorbell (EF1) */
+#define EFX_VI_HW_RESOURCE_EVQTIMER  0x4	/* Address of event q timer */
+
+/* Address of event q pointer (EF1) */
+#define EFX_VI_HW_RESOURCE_EVQPTR    0x5
+/* Address of register pointer (Falcon A) */
+#define EFX_VI_HW_RESOURCE_EVQRPTR   0x6
+/* Offset of register pointer (Falcon B) */
+#define EFX_VI_HW_RESOURCE_EVQRPTR_OFFSET 0x7
+/* Address of mem KVA */
+#define EFX_VI_HW_RESOURCE_EVQMEMKVA 0x8
+/* PFN of doorbell page (Falcon) */
+#define EFX_VI_HW_RESOURCE_BELLPAGE  0x9
+
+/*! How large an array to allocate for the get_() functions - smaller
+  than the total number of constants as some are mutually exclusive */
+#define EFX_VI_HW_RESOURCE_MAXSIZE   0x7
+
+/*! Constants for the mem_type field in efx_vi_hw_resource */
+#define EFX_VI_HW_RESOURCE_IOBUFFER   0	/* Host memory */
+#define EFX_VI_HW_RESOURCE_PERIPHERAL 1	/* Card memory/registers */
+
+/*!
+ * Data structure providing information on a hardware resource mapping
+ */
+struct efx_vi_hw_resource {
+	u8 type;		/*!< What this resource represents */
+	u8 mem_type;		/*!< What type of memory is it in, eg,
+				 * host or iomem */
+	u8 more_to_follow;	/*!< Is this part of a multi-region resource */
+	u32 length;		/*!< Length of the resource in bytes */
+	unsigned long address;	/*!< Address of this resource */
+};
+
+/*!
+ * Metadata concerning the list of hardware resource mappings
+ */
+struct efx_vi_hw_resource_metadata {
+	int version;
+	int evq_order;
+	int evq_offs;
+	int evq_capacity;
+	int instance;
+	unsigned rx_capacity;
+	unsigned tx_capacity;
+	int nic_arch;
+	int nic_revision;
+	char nic_variant;
+};
+
+/*!
+ * Obtain a list of hardware resource mappings, using virtual addresses
+ *
+ * \param vih The handle to identify the efx_vi
+ * \param mdata Pointer to a structure to receive the metadata
+ * \param hw_res_array An array to receive the list of hardware resources
+ * \param length The length of hw_res_array.  Updated on success to contain
+ * the number of entries in the supplied array that were used.
+ * \return Zero on success, non-zero on failure
+ */
+extern int
+efx_vi_hw_resource_get_virt(struct efx_vi_state *vih,
+			    struct efx_vi_hw_resource_metadata *mdata,
+			    struct efx_vi_hw_resource *hw_res_array,
+			    int *length);
+
+/*!
+ * Obtain a list of hardware resource mappings, using physical addresses
+ *
+ * \param vih The handle to identify the efx_vi
+ * \param mdata Pointer to a structure to receive the metadata
+ * \param hw_res_array An array to receive the list of hardware resources
+ * \param length The length of hw_res_array.  Updated on success to contain
+ * the number of entries in the supplied array that were used.
+ * \return Zero on success, non-zero on failure
+ */
+extern int
+efx_vi_hw_resource_get_phys(struct efx_vi_state *vih,
+			    struct efx_vi_hw_resource_metadata *mdata,
+			    struct efx_vi_hw_resource *hw_res_array,
+			    int *length);
+
+#endif /* __CI_DRIVER_RESOURCE_EFX_VI_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/resource/linux_efhw_nic.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/resource/linux_efhw_nic.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/driver/resource/linux_efhw_nic.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/driver/resource/linux_efhw_nic.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,76 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains definition of the public type struct linux_efhw_nic.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_DRIVER_RESOURCE_LINUX_RESOURCE__
+#define __CI_DRIVER_RESOURCE_LINUX_RESOURCE__
+
+#ifndef __linux__
+# error Silly
+#endif
+#ifndef __KERNEL__
+# error Silly
+#endif
+
+#include <ci/efhw/efhw_types.h>
+#include <linux/interrupt.h>
+
+
+/************************************************************************
+ * Per-nic structure in the resource driver                             *
+ ************************************************************************/
+
+struct linux_efhw_nic {
+	struct efhw_nic nic;
+
+	struct pci_dev *pci_dev;	/*!< pci descriptor */
+	struct tasklet_struct tasklet;	/*!< for interrupt bottom half */
+
+	/* Physical addresses of the control aperture bar. */
+	unsigned long ctr_ap_pci_addr;
+
+	/*! Callbacks for driverlink, when needed. */
+	struct efx_dl_callbacks *dl_callbacks;
+
+	/*! Event handlers. */
+	struct efhw_ev_handler *ev_handlers;
+
+};
+
+#define linux_efhw_nic(efhw_nic)                \
+  container_of(efhw_nic, struct linux_efhw_nic, nic)
+
+#endif /* __CI_DRIVER_RESOURCE_LINUX_RESOURCE__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/checks.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/checks.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/checks.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/checks.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,118 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides helpers to turn bit shifts into dword shifts and
+ * check that the bit fields haven't overflown the dword etc.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_CHECK_H__
+#define __CI_EFHW_CHECK_H__
+
+/*----------------------------------------------------------------------------
+ *
+ * Helpers to turn bit shifts into dword shifts and check that the bit fields
+ * haven't overflown the dword etc. Aim is to preserve consistency with the
+ * autogenerated headers - once stable we could hard code.
+ *
+ *---------------------------------------------------------------------------*/
+
+/* mask constructors */
+#define __FALCON_MASK(WIDTH, T)	((((T)1) << (WIDTH)) - 1)
+#define __FALCON_MASK32(WIDTH)	__FALCON_MASK((WIDTH), uint32_t)
+#define __FALCON_MASK64(WIDTH)	__FALCON_MASK((WIDTH), uint64_t)
+
+#define __FALCON_MASKFIELD32(LBN, WIDTH) \
+	((uint32_t)(__FALCON_MASK32(WIDTH) << (LBN)))
+
+/* constructors for fields which span the first and second dwords */
+#define __LW(LBN)		(32 - LBN)
+#define __LOW(v, LBN, WIDTH) \
+	((uint32_t)(((v) & __FALCON_MASK64(__LW((LBN)))) << (LBN)))
+#define __HIGH(v, LBN, WIDTH) \
+	((uint32_t)(((v) >> __LW((LBN))) & \
+		    __FALCON_MASK64((WIDTH - __LW((LBN))))))
+/* constructors for fields within the second dword */
+#define __DW2(LBN)		((LBN) - 32)
+
+/* constructors for fields which span the second and third dwords */
+#define __LW2(LBN)		(64 - LBN)
+#define __LOW2(v, LBN, WIDTH) \
+	((uint32_t)(((v) & __FALCON_MASK64(__LW2((LBN)))) << ((LBN) - 32)))
+#define __HIGH2(v, LBN, WIDTH) \
+	((uint32_t)(((v) >> __LW2((LBN))) & \
+		    __FALCON_MASK64((WIDTH - __LW2((LBN))))))
+
+/* constructors for fields within the third dword */
+#define __DW3(LBN)		((LBN) - 64)
+
+/* constructors for fields which span the third and fourth dwords */
+#define __LW3(LBN)		(96 - LBN)
+#define __LOW3(v, LBN, WIDTH) \
+	((uint32_t)(((v) & __FALCON_MASK64(__LW3((LBN)))) << ((LBN) - 64)))
+#define __HIGH3(v, LBN, WIDTH) \
+	((ci_unit32)(((v) >> __LW3((LBN))) & \
+		     __FALCON_MASK64((WIDTH - __LW3((LBN))))))
+
+/* constructors for fields within the fourth dword */
+#define __DW4(LBN)		((LBN) - 96)
+
+/* checks that the autogenerated headers are consistent with our model */
+#define __WIDTHCHCK(a, b)	EFHW_ASSERT((a) == (b))
+#define __RANGECHCK(v, WIDTH) \
+	EFHW_ASSERT(((uint64_t)(v) & ~(__FALCON_MASK64((WIDTH)))) == 0)
+
+/* fields within the first dword */
+#define __DWCHCK(LBN, WIDTH) \
+	EFHW_ASSERT(((LBN) >= 0) && (((LBN)+(WIDTH)) <= 32))
+
+/* fields which span the first and second dwords */
+#define __LWCHK(LBN, WIDTH)	EFHW_ASSERT(WIDTH >= __LW(LBN))
+
+/* fields within the second dword */
+#define __DW2CHCK(LBN, WIDTH) \
+	EFHW_ASSERT(((LBN) >= 32) && (((LBN)+(WIDTH)) <= 64))
+
+/* fields which span the second and third dwords */
+#define __LW2CHK(LBN, WIDTH)	EFHW_ASSERT(WIDTH >= __LW2(LBN))
+
+/* fields within the third dword */
+#define __DW3CHCK(LBN, WIDTH) \
+	EFHW_ASSERT(((LBN) >= 64) && (((LBN)+(WIDTH)) <= 96))
+
+/* fields which span the third and fourth dwords */
+#define __LW3CHK(LBN, WIDTH)	EFHW_ASSERT(WIDTH >= __LW3(LBN))
+
+/* fields within the fourth dword */
+#define __DW4CHCK(LBN, WIDTH) \
+	EFHW_ASSERT(((LBN) >= 96) && (((LBN)+(WIDTH)) <= 128))
+
+/* fields in the first qword */
+#define __QWCHCK(LBN, WIDTH) \
+	EFHW_ASSERT(((LBN) >= 0) && (((LBN)+(WIDTH)) <= 64))
+
+#endif /* __CI_EFHW_CHECK_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/common.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/common.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/common.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/common.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,102 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides API of the efhw library which may be used both from
+ * the kernel and from the user-space code.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_COMMON_H__
+#define __CI_EFHW_COMMON_H__
+
+#include <ci/efhw/common_sysdep.h>
+
+enum efhw_arch {
+	EFHW_ARCH_FALCON,
+	EFHW_ARCH_SIENA,
+};
+
+typedef uint32_t efhw_buffer_addr_t;
+#define EFHW_BUFFER_ADDR_FMT	"[ba:%"PRIx32"]"
+
+/*! Comment? */
+typedef union {
+	uint64_t u64;
+	struct {
+		uint32_t a;
+		uint32_t b;
+	} opaque;
+	struct {
+		uint32_t code;
+		uint32_t status;
+	} ev1002;
+} efhw_event_t;
+
+/* Flags for TX/RX queues */
+#define EFHW_VI_JUMBO_EN           0x01  /*! scatter RX over multiple desc */
+#define EFHW_VI_ISCSI_RX_HDIG_EN   0x02  /*! iscsi rx header digest */
+#define EFHW_VI_ISCSI_TX_HDIG_EN   0x04  /*! iscsi tx header digest */
+#define EFHW_VI_ISCSI_RX_DDIG_EN   0x08  /*! iscsi rx data digest */
+#define EFHW_VI_ISCSI_TX_DDIG_EN   0x10  /*! iscsi tx data digest */
+#define EFHW_VI_TX_PHYS_ADDR_EN    0x20  /*! TX physical address mode */
+#define EFHW_VI_RX_PHYS_ADDR_EN    0x40  /*! RX physical address mode */
+#define EFHW_VI_RM_WITH_INTERRUPT  0x80  /*! VI with an interrupt */
+#define EFHW_VI_TX_IP_CSUM_DIS     0x100 /*! enable ip checksum generation */
+#define EFHW_VI_TX_TCPUDP_CSUM_DIS 0x200 /*! enable tcp/udp checksum
+					   generation */
+#define EFHW_VI_TX_TCPUDP_ONLY     0x400 /*! drop non-tcp/udp packets */
+
+/* Types of hardware filter */
+/* Each of these values implicitly selects scatter filters on B0 - or in
+   EFHW_IP_FILTER_TYPE_NOSCAT_B0_MASK if a non-scatter filter is required */
+#define EFHW_IP_FILTER_TYPE_UDP_WILDCARD  (0)	/* dest host only */
+#define EFHW_IP_FILTER_TYPE_UDP_FULL      (1)	/* dest host and port */
+#define EFHW_IP_FILTER_TYPE_TCP_WILDCARD  (2)	/* dest based filter */
+#define EFHW_IP_FILTER_TYPE_TCP_FULL      (3)	/* src  filter */
+/* Same again, but with RSS (for B0 only) */
+#define EFHW_IP_FILTER_TYPE_UDP_WILDCARD_RSS_B0  (4)
+#define EFHW_IP_FILTER_TYPE_UDP_FULL_RSS_B0      (5)
+#define EFHW_IP_FILTER_TYPE_TCP_WILDCARD_RSS_B0  (6)
+#define EFHW_IP_FILTER_TYPE_TCP_FULL_RSS_B0      (7)
+
+#define EFHW_IP_FILTER_TYPE_FULL_MASK      (0x1) /* Mask for full / wildcard */
+#define EFHW_IP_FILTER_TYPE_TCP_MASK       (0x2) /* Mask for TCP type */
+#define EFHW_IP_FILTER_TYPE_RSS_B0_MASK    (0x4) /* Mask for B0 RSS enable */
+#define EFHW_IP_FILTER_TYPE_NOSCAT_B0_MASK (0x8) /* Mask for B0 SCATTER dsbl */
+
+#define EFHW_IP_FILTER_TYPE_MASK	(0xffff) /* Mask of types above */
+
+#define EFHW_IP_FILTER_BROADCAST	(0x10000) /* driverlink filter
+						     support */
+
+#endif /* __CI_EFHW_COMMON_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/common_sysdep.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/common_sysdep.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/common_sysdep.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/common_sysdep.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,71 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides version-independent Linux kernel API for
+ * userland-to-kernel interfaces.
+ * Only kernels >=2.6.9 are supported.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_COMMON_LINUX_H__
+#define __CI_EFHW_COMMON_LINUX_H__
+
+#include <linux/types.h>
+#include <linux/version.h>
+
+/* Dirty hack, but Linux kernel does not provide DMA_ADDR_T_FMT */
+#if BITS_PER_LONG == 64 || defined(CONFIG_HIGHMEM64G)
+#define DMA_ADDR_T_FMT "%llx"
+#else
+#define DMA_ADDR_T_FMT "%x"
+#endif
+
+/* Linux kernel also does not provide PRIx32... Sigh. */
+#define PRIx32 "x"
+
+#ifdef __ia64__
+# define PRIx64 "lx"
+#else
+# define PRIx64 "llx"
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,19)
+enum {
+	false = 0,
+	true = 1
+};
+
+typedef _Bool bool;
+#endif /* LINUX_VERSION_CODE < 2.6.19 */
+
+#endif /* __CI_EFHW_COMMON_LINUX_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/debug.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/debug.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/debug.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/debug.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,84 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides debug-related API for efhw library using Linux kernel
+ * primitives.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_DEBUG_LINUX_H__
+#define __CI_EFHW_DEBUG_LINUX_H__
+
+#define EFHW_PRINTK_PREFIX "[sfc efhw] "
+
+#define EFHW_PRINTK(level, fmt, ...) \
+	printk(level EFHW_PRINTK_PREFIX fmt "\n", __VA_ARGS__)
+
+/* Following macros should be used with non-zero format parameters
+ * due to __VA_ARGS__ limitations.  Use "%s" with __FUNCTION__ if you can't
+ * find better parameters. */
+#define EFHW_ERR(fmt, ...)     EFHW_PRINTK(KERN_ERR, fmt, __VA_ARGS__)
+#define EFHW_WARN(fmt, ...)    EFHW_PRINTK(KERN_WARNING, fmt, __VA_ARGS__)
+#define EFHW_NOTICE(fmt, ...)  EFHW_PRINTK(KERN_NOTICE, fmt, __VA_ARGS__)
+#if 0 && !defined(NDEBUG)
+#define EFHW_TRACE(fmt, ...) EFHW_PRINTK(KERN_DEBUG, fmt, __VA_ARGS__)
+#else
+#define EFHW_TRACE(fmt, ...)
+#endif
+
+#ifndef NDEBUG
+#define EFHW_ASSERT(cond)  BUG_ON((cond) == 0)
+#define EFHW_DO_DEBUG(expr) expr
+#else
+#define EFHW_ASSERT(cond)
+#define EFHW_DO_DEBUG(expr)
+#endif
+
+#define EFHW_TEST(expr)			\
+	do {				\
+		if (unlikely(!(expr)))	\
+		BUG();			\
+	} while (0)
+
+/* Build time asserts. We paste the line number into the type name
+ * so that the macro can be used more than once per file even if the
+ * compiler objects to multiple identical typedefs. Collisions
+ * between use in different header files is still possible. */
+#ifndef EFHW_BUILD_ASSERT
+#define __EFHW_BUILD_ASSERT_NAME(_x) __EFHW_BUILD_ASSERT_ILOATHECPP(_x)
+#define __EFHW_BUILD_ASSERT_ILOATHECPP(_x)  __EFHW_BUILD_ASSERT__ ##_x
+#define EFHW_BUILD_ASSERT(e) \
+	typedef char __EFHW_BUILD_ASSERT_NAME(__LINE__)[(e) ? 1 : -1]
+#endif
+
+#endif /* __CI_EFHW_DEBUG_LINUX_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/efhw_config.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/efhw_config.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/efhw_config.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/efhw_config.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,43 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides some limits used in both kernel and userland code.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_EFAB_CONFIG_H__
+#define __CI_EFHW_EFAB_CONFIG_H__
+
+#define EFHW_MAX_NR_DEVS 5	/* max number of efhw devices supported */
+
+#endif /* __CI_EFHW_EFAB_CONFIG_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/efhw_types.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/efhw_types.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/efhw_types.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/efhw_types.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,342 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides struct efhw_nic and some related types.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_EFAB_TYPES_H__
+#define __CI_EFHW_EFAB_TYPES_H__
+
+#include <ci/efhw/efhw_config.h>
+#include <ci/efhw/hardware_sysdep.h>
+#include <ci/efhw/iopage_types.h>
+#include <ci/efhw/sysdep.h>
+
+/*--------------------------------------------------------------------
+ *
+ * hardware limits used in the types
+ *
+ *--------------------------------------------------------------------*/
+
+#define EFHW_KEVENTQ_MAX    8
+
+/*--------------------------------------------------------------------
+ *
+ * forward type declarations
+ *
+ *--------------------------------------------------------------------*/
+
+struct efhw_nic;
+
+/*--------------------------------------------------------------------
+ *
+ * Managed interface
+ *
+ *--------------------------------------------------------------------*/
+
+struct efhw_buffer_table_allocation{
+	unsigned base;
+	unsigned order;
+};
+
+struct eventq_resource_hardware {
+	/*!iobuffer allocated for eventq - can be larger than eventq */
+	efhw_iopages_t iobuff;
+	unsigned iobuff_off;
+	struct efhw_buffer_table_allocation buf_tbl_alloc;
+	int capacity;		/*!< capacity of event queue */
+};
+
+/*--------------------------------------------------------------------
+ *
+ * event queues and event driven callbacks
+ *
+ *--------------------------------------------------------------------*/
+
+struct efhw_keventq {
+	volatile int lock;
+	caddr_t evq_base;
+	int32_t evq_ptr;
+	uint32_t evq_mask;
+	unsigned instance;
+	struct eventq_resource_hardware hw;
+	struct efhw_ev_handler *ev_handlers;
+};
+
+/**********************************************************************
+ * Portable HW interface. ***************************************
+ **********************************************************************/
+
+/*--------------------------------------------------------------------
+ *
+ * EtherFabric Functional units - configuration and control
+ *
+ *--------------------------------------------------------------------*/
+
+struct efhw_func_ops {
+
+  /*-------------- Initialisation ------------ */
+
+	/*! close down all hardware functional units - leaves NIC in a safe
+	   state for driver unload */
+	void (*close_hardware) (struct efhw_nic *nic);
+
+	/*! initialise all hardware functional units */
+	int (*init_hardware) (struct efhw_nic *nic,
+			      struct efhw_ev_handler *,
+			      const uint8_t *mac_addr);
+
+  /*-------------- Interrupt support  ------------ */
+
+	/*! Main interrupt routine
+	 **        This function returns,
+	 **  - zero,       if the IRQ was not generated by EF1
+	 **  - non-zero,   if EF1 was the source of the IRQ
+	 **
+	 **
+	 ** opaque is an OS provided pointer for use by the OS callbacks
+	 ** e.g in Windows used to indicate DPC scheduled
+	 */
+	int (*interrupt) (struct efhw_nic *nic);
+
+	/*! Enable given interrupt mask for the given IRQ unit */
+	void (*interrupt_enable) (struct efhw_nic *nic, uint idx);
+
+	/*! Disable given interrupt mask for the given IRQ unit */
+	void (*interrupt_disable) (struct efhw_nic *nic, uint idx);
+
+	/*! Set interrupt moderation strategy for the given IRQ unit
+	 ** val is in usec
+	 */
+	void (*set_interrupt_moderation)(struct efhw_nic *nic,
+					 uint idx, uint val);
+
+  /*-------------- Event support  ------------ */
+
+	/*! Enable the given event queue
+	   depending on the underlying implementation (EF1 or Falcon) then
+	   either a q_base_addr in host memory, or a buffer base id should
+	   be proivded
+	 */
+	void (*event_queue_enable) (struct efhw_nic *nic,
+				    uint evq,	/* evnt queue index */
+				    uint evq_size,	/* units of #entries */
+				    dma_addr_t q_base_addr, uint buf_base_id);
+
+	/*! Disable the given event queue (and any associated timer) */
+	void (*event_queue_disable) (struct efhw_nic *nic, uint evq,
+				     int timer_only);
+
+	/*! request wakeup from the NIC on a given event Q */
+	void (*wakeup_request) (struct efhw_nic *nic, dma_addr_t q_base_addr,
+				int next_i, int evq);
+
+	/*! Push a SW event on a given eventQ */
+	void (*sw_event) (struct efhw_nic *nic, int data, int evq);
+
+  /*-------------- Filter support  ------------ */
+
+	/*! Setup a given filter - The software can request a filter_i,
+	 * but some EtherFabric implementations will override with
+	 * a more suitable index
+	 */
+	int (*ipfilter_set) (struct efhw_nic *nic, int type,
+			     int *filter_i, int dmaq,
+			     unsigned saddr_be32, unsigned sport_be16,
+			     unsigned daddr_be32, unsigned dport_be16);
+
+	/*! Attach a given filter to a DMAQ */
+	void (*ipfilter_attach) (struct efhw_nic *nic, int filter_idx,
+				 int dmaq_idx);
+
+	/*! Detach a filter from its DMAQ */
+	void (*ipfilter_detach) (struct efhw_nic *nic, int filter_idx);
+
+	/*! Clear down a given filter */
+	void (*ipfilter_clear) (struct efhw_nic *nic, int filter_idx);
+
+  /*-------------- DMA support  ------------ */
+
+	/*! Initialise NIC state for a given TX DMAQ */
+	void (*dmaq_tx_q_init) (struct efhw_nic *nic,
+				uint dmaq, uint evq, uint owner, uint tag,
+				uint dmaq_size, uint buf_idx, uint flags);
+
+	/*! Initialise NIC state for a given RX DMAQ */
+	void (*dmaq_rx_q_init) (struct efhw_nic *nic,
+				uint dmaq, uint evq, uint owner, uint tag,
+				uint dmaq_size, uint buf_idx, uint flags);
+
+	/*! Disable a given TX DMAQ */
+	void (*dmaq_tx_q_disable) (struct efhw_nic *nic, uint dmaq);
+
+	/*! Disable a given RX DMAQ */
+	void (*dmaq_rx_q_disable) (struct efhw_nic *nic, uint dmaq);
+
+	/*! Flush a given TX DMA channel */
+	int (*flush_tx_dma_channel) (struct efhw_nic *nic, uint dmaq);
+
+	/*! Flush a given RX DMA channel */
+	int (*flush_rx_dma_channel) (struct efhw_nic *nic, uint dmaq);
+
+  /*-------------- Buffer table Support ------------ */
+
+	/*! Initialise a buffer table page */
+	void (*buffer_table_set) (struct efhw_nic *nic,
+				  dma_addr_t dma_addr,
+				  uint bufsz, uint region,
+				  int own_id, int buffer_id);
+
+	/*! Initialise a block of buffer table pages */
+	void (*buffer_table_set_n) (struct efhw_nic *nic, int buffer_id,
+				    dma_addr_t dma_addr,
+				    uint bufsz, uint region,
+				    int n_pages, int own_id);
+
+	/*! Clear a block of buffer table pages */
+	void (*buffer_table_clear) (struct efhw_nic *nic, int buffer_id,
+				    int num);
+
+	/*! Commit a buffer table update  */
+	void (*buffer_table_commit) (struct efhw_nic *nic);
+
+};
+
+
+/*----------------------------------------------------------------------------
+ *
+ * NIC type
+ *
+ *---------------------------------------------------------------------------*/
+
+struct efhw_device_type {
+	int  arch;            /* enum efhw_arch */
+	char variant;         /* 'A', 'B', ... */
+	int  revision;        /* 0, 1, ... */
+};
+
+
+/*----------------------------------------------------------------------------
+ *
+ * EtherFabric NIC instance - nic.c for HW independent functions
+ *
+ *---------------------------------------------------------------------------*/
+
+/*! */
+struct efhw_nic {
+	/*! zero base index in efrm_nic_table.nic array */
+	volatile int index;
+	int ifindex;		/*!< OS level nic index */
+#ifdef HAS_NET_NAMESPACE
+	struct net *nd_net;
+#endif
+
+	struct efhw_device_type devtype;
+
+	/*! Options that can be set by user. */
+	unsigned options;
+# define NIC_OPT_EFTEST             0x1	/* owner is an eftest app */
+
+# define NIC_OPT_DEFAULT            0
+
+	/*! Internal flags that indicate hardware properties at runtime. */
+	unsigned flags;
+# define NIC_FLAG_NO_INTERRUPT          0x01 /* to be set at init time only */
+# define NIC_FLAG_TRY_MSI               0x02
+# define NIC_FLAG_MSI                   0x04
+# define NIC_FLAG_OS_IRQ_EN             0x08
+# define NIC_FLAG_10G                   0x10
+
+	unsigned mtu;		/*!< MAC MTU (includes MAC hdr) */
+
+	/* hardware resources */
+
+	/*! I/O address of the start of the bar */
+	efhw_ioaddr_t bar_ioaddr;
+
+	/*! Bar number of control aperture. */
+	unsigned ctr_ap_bar;
+	/*! Length of control aperture in bytes. */
+	unsigned ctr_ap_bytes;
+
+	uint8_t mac_addr[ETH_ALEN];	/*!< mac address  */
+
+	/*! EtherFabric Functional Units -- functions */
+	const struct efhw_func_ops *efhw_func;
+
+	/* Value read from FPGA version register.  Zero for asic. */
+	unsigned fpga_version;
+
+	/*! This lock protects a number of misc NIC resources.  It should
+	 * only be used for things that can be at the bottom of the lock
+	 * order.  ie. You mustn't attempt to grab any other lock while
+	 * holding this one.
+	 */
+	spinlock_t *reg_lock;
+	spinlock_t the_reg_lock;
+
+	int buf_commit_outstanding;	/*!< outstanding buffer commits */
+
+	/*! interrupt callbacks (hard-irq) */
+	void (*irq_handler) (struct efhw_nic *, int unit);
+
+	/*! event queues per driver */
+	struct efhw_keventq evq[EFHW_KEVENTQ_MAX];
+
+/* for marking when we are not using an IRQ unit
+      - 0 is a valid offset to an IRQ unit on EF1! */
+#define EFHW_IRQ_UNIT_UNUSED  0xffff
+	/*! interrupt unit in use  */
+	unsigned int irq_unit[EFHW_KEVENTQ_MAX];
+	efhw_iopage_t irq_iobuff;	/*!<  Falcon SYSERR interrupt */
+
+	/* The new driverlink infrastructure. */
+	struct efx_dl_device *net_driver_dev;
+	struct efx_dlfilt_cb_s *dlfilter_cb;
+
+	/*! Bit masks of the sizes of event queues and dma queues supported
+	 * by the nic. */
+	unsigned evq_sizes;
+	unsigned rxq_sizes;
+	unsigned txq_sizes;
+
+	/* Size of filter table (including odd and even banks). */
+	unsigned filter_tbl_size;
+};
+
+
+#define EFHW_KVA(nic)       ((nic)->bar_ioaddr)
+
+
+#endif /* __CI_EFHW_EFHW_TYPES_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/eventq.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/eventq.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/eventq.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/eventq.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,73 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains API provided by efhw/eventq.c file.  This file is not
+ * designed for use outside of the SFC resource driver.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_EVENTQ_H__
+#define __CI_EFHW_EVENTQ_H__
+
+#include <ci/efhw/efhw_types.h>
+#include <ci/efhw/eventq_macros.h>
+
+/*! Poll the event queue. */
+extern int efhw_keventq_poll(struct efhw_nic *, struct efhw_keventq *);
+
+/*! Callbacks for handling events. */
+struct efhw_ev_handler {
+	void (*wakeup_fn)(struct efhw_nic *nic, efhw_event_t *ev);
+	void (*timeout_fn)(struct efhw_nic *nic, efhw_event_t *ev);
+	void (*sw_fn)(struct efhw_nic *nic, efhw_event_t *ev);
+	void (*dmaq_flushed_fn) (struct efhw_nic *, int, int);
+};
+
+extern int efhw_keventq_ctor(struct efhw_nic *, int instance,
+			     struct efhw_keventq *, struct efhw_ev_handler *);
+extern void efhw_keventq_dtor(struct efhw_nic *, struct efhw_keventq *);
+
+extern void efhw_handle_txdmaq_flushed(struct efhw_nic *,
+				       struct efhw_ev_handler *,
+				       efhw_event_t *);
+extern void efhw_handle_rxdmaq_flushed(struct efhw_nic *,
+				       struct efhw_ev_handler *,
+				       efhw_event_t *);
+extern void efhw_handle_wakeup_event(struct efhw_nic *,
+				     struct efhw_ev_handler *,
+				     efhw_event_t *);
+extern void efhw_handle_timeout_event(struct efhw_nic *,
+				      struct efhw_ev_handler *,
+				      efhw_event_t *);
+
+#endif /* __CI_EFHW_EVENTQ_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/eventq_macros.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/eventq_macros.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/eventq_macros.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/eventq_macros.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,81 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides some event-related macros.  This file is designed for
+ * use from kernel and from the userland contexts.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_EVENTQ_MACROS_H__
+#define __CI_EFHW_EVENTQ_MACROS_H__
+
+#include <ci/efhw/common.h>
+
+/*--------------------------------------------------------------------
+ *
+ * Event Queue manipulation
+ *
+ *--------------------------------------------------------------------*/
+
+#define EFHW_EVENT_OFFSET(q, s, i)					\
+	(((s)->evq_ptr - (i) * (int32_t)sizeof(efhw_event_t))		\
+	 & (q)->evq_mask)
+
+#define EFHW_EVENT_PTR(q, s, i)						\
+	((efhw_event_t *)((q)->evq_base + EFHW_EVENT_OFFSET(q, s, i)))
+
+#define EFHW_EVENTQ_NEXT(s)						\
+	do { ((s)->evq_ptr += sizeof(efhw_event_t)); } while (0)
+
+#define EFHW_EVENTQ_PREV(s)						\
+	do { ((s)->evq_ptr -= sizeof(efhw_event_t)); } while (0)
+
+/* Be worried about this on byteswapped machines */
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+  /* Due to crazy chipsets, we see the event words being written in
+   ** arbitrary order (bug4539).  So test for presence of event must ensure
+   ** that both halves have changed from the null.
+   */
+	#define EFHW_IS_EVENT(evp)			\
+		(((evp)->opaque.a != (uint32_t)-1) &&	\
+		 ((evp)->opaque.b != (uint32_t)-1))
+	#define EFHW_CLEAR_EVENT(evp)       ((evp)->u64 = (uint64_t)-1)
+	#define EFHW_CLEAR_EVENT_VALUE      0xff
+#else
+	#error Fixme - unknown hardware configuration
+#endif
+
+#define EFHW_EVENT_OVERFLOW(evq, s)			\
+	(EFHW_IS_EVENT(EFHW_EVENT_PTR(evq, s, 1)))
+
+#endif /* __CI_EFHW_EVENTQ_MACROS_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/falcon.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/falcon.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/falcon.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/falcon.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,93 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains API provided by efhw/falcon.c file.  This file is not
+ * designed for use outside of the SFC resource driver.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_FALCON_H__
+#define __CI_EFHW_FALCON_H__
+
+#include <ci/efhw/efhw_types.h>
+#include <ci/efhw/common.h>
+
+/*----------------------------------------------------------------------------
+ *
+ * Locks - unfortunately required
+ *
+ *---------------------------------------------------------------------------*/
+
+#define FALCON_LOCK_DECL        irq_flags_t lock_state
+#define FALCON_LOCK_LOCK(nic) \
+	spin_lock_irqsave((nic)->reg_lock, lock_state)
+#define FALCON_LOCK_UNLOCK(nic) \
+	spin_unlock_irqrestore((nic)->reg_lock, lock_state)
+
+extern struct efhw_func_ops falcon_char_functional_units;
+
+/*! specify a pace value for a TX DMA Queue */
+extern void falcon_nic_pace(struct efhw_nic *nic, uint dmaq, uint pace);
+
+/*! confirm buffer table updates - should be used for items where
+   loss of data would be unacceptable. E.g for the buffers that back
+   an event or DMA queue */
+extern void falcon_nic_buffer_table_confirm(struct efhw_nic *nic);
+
+/*! Reset the all the TX DMA queue pointers. */
+extern void falcon_clobber_tx_dma_ptrs(struct efhw_nic *nic, uint dmaq);
+
+extern int
+falcon_handle_char_event(struct efhw_nic *nic,
+			 struct efhw_ev_handler *h, efhw_event_t *evp);
+
+/*! map event queue instance space (0,1,2,..) onto event queue
+  number. This function takes into account the allocation rules for
+  the underlying driver model */
+extern int falcon_idx_to_evq(struct efhw_nic *nic, uint idx);
+
+/*! Acknowledge to HW that processing is complete on a given event queue */
+extern void falcon_nic_evq_ack(struct efhw_nic *nic, uint evq,	/* evq id */
+			       uint rptr,	/* new read pointer update */
+			       bool wakeup	/* request a wakeup event if
+						   ptr's != */
+    );
+
+extern void
+falcon_nic_buffer_table_set_n(struct efhw_nic *nic, int buffer_id,
+			      dma_addr_t dma_addr, uint bufsz, uint region,
+			      int n_pages, int own_id);
+
+extern void falcon_nic_ipfilter_ctor(struct efhw_nic *nic);
+
+#endif /* __CI_EFHW_FALCON_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/falcon_hash.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/falcon_hash.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/falcon_hash.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/falcon_hash.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,58 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains API provided by efhw/falcon_hash.c file.
+ * Function declared in this file are not exported from the Linux
+ * sfc_resource driver.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_FALCON_HASH_H__
+#define __CI_EFHW_FALCON_HASH_H__
+
+/* All LE parameters */
+extern unsigned int
+falcon_hash_get_key(unsigned int src_ip, unsigned int src_port,
+		    unsigned int dest_ip, unsigned int dest_port,
+		    int tcp, int full);
+
+unsigned int falcon_hash_function1(unsigned int key, unsigned int nfilters);
+
+extern unsigned int
+falcon_hash_function2(unsigned int key, unsigned int nfitlers);
+
+extern unsigned int
+falcon_hash_iterator(unsigned int hash1, unsigned int hash2,
+		     unsigned int n_search, unsigned int nfilters);
+
+#endif /* __CI_EFHW_FALCON_HASH_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/hardware_sysdep.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/hardware_sysdep.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/hardware_sysdep.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/hardware_sysdep.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,84 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides version-independent Linux kernel API for header files
+ * with hardware-related definitions (in ci/driver/efab/hardware*).
+ * Only kernels >=2.6.9 are supported.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_HARDWARE_LINUX_H__
+#define __CI_EFHW_HARDWARE_LINUX_H__
+
+#include <asm/io.h>
+
+#ifdef __LITTLE_ENDIAN
+#define EFHW_IS_LITTLE_ENDIAN
+#elif __BIG_ENDIAN
+#define EFHW_IS_BIG_ENDIAN
+#else
+#error Unknown endianness
+#endif
+
+#ifndef mmiowb
+	#if defined(__i386__) || defined(__x86_64__)
+		#define mmiowb()
+	#elif defined(__ia64__)
+		#ifndef ia64_mfa
+			#define ia64_mfa() asm volatile ("mf.a" ::: "memory")
+		#endif
+	#define mmiowb ia64_mfa
+	#else
+	#error "Need definition for mmiowb()"
+	#endif
+#endif
+
+typedef char *efhw_ioaddr_t;
+
+#ifndef readq
+static inline uint64_t __readq(void __iomem *addr)
+{
+	return *(volatile uint64_t *)addr;
+}
+#define readq(x) __readq(x)
+#endif
+
+#ifndef writeq
+static inline void __writeq(uint64_t v, void __iomem *addr)
+{
+	*(volatile uint64_t *)addr = v;
+}
+#define writeq(val, addr) __writeq((val), (addr))
+#endif
+
+#endif /* __CI_EFHW_HARDWARE_LINUX_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/iopage.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/iopage.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/iopage.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/iopage.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,58 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains OS-independent API for allocating iopage types.
+ * The implementation of these functions is highly OS-dependent.
+ * This file is not designed for use outside of the SFC resource driver.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_DRIVER_RESOURCE_IOPAGE_H__
+#define __CI_DRIVER_RESOURCE_IOPAGE_H__
+
+#include <ci/efhw/efhw_types.h>
+
+/*--------------------------------------------------------------------
+ *
+ * memory allocation
+ *
+ *--------------------------------------------------------------------*/
+
+extern int efhw_iopage_alloc(struct efhw_nic *, efhw_iopage_t *p);
+extern void efhw_iopage_free(struct efhw_nic *, efhw_iopage_t *p);
+
+extern int efhw_iopages_alloc(struct efhw_nic *, efhw_iopages_t *p,
+			      unsigned order);
+extern void efhw_iopages_free(struct efhw_nic *, efhw_iopages_t *p);
+
+#endif /* __CI_DRIVER_RESOURCE_IOPAGE_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/iopage_types.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/iopage_types.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/iopage_types.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/iopage_types.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,188 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides efhw_page_t and efhw_iopage_t for Linux kernel.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_IOPAGE_LINUX_H__
+#define __CI_EFHW_IOPAGE_LINUX_H__
+
+#include <linux/gfp.h>
+#include <linux/hardirq.h>
+#include <ci/efhw/debug.h>
+
+/*--------------------------------------------------------------------
+ *
+ * efhw_page_t: A single page of memory.  Directly mapped in the driver,
+ * and can be mapped to userlevel.
+ *
+ *--------------------------------------------------------------------*/
+
+typedef struct {
+	unsigned long kva;
+} efhw_page_t;
+
+static inline int efhw_page_alloc(efhw_page_t *p)
+{
+	p->kva = __get_free_page(in_interrupt()? GFP_ATOMIC : GFP_KERNEL);
+	return p->kva ? 0 : -ENOMEM;
+}
+
+static inline int efhw_page_alloc_zeroed(efhw_page_t *p)
+{
+	p->kva = get_zeroed_page(in_interrupt()? GFP_ATOMIC : GFP_KERNEL);
+	return p->kva ? 0 : -ENOMEM;
+}
+
+static inline void efhw_page_free(efhw_page_t *p)
+{
+	free_page(p->kva);
+	EFHW_DO_DEBUG(memset(p, 0, sizeof(*p)));
+}
+
+static inline char *efhw_page_ptr(efhw_page_t *p)
+{
+	return (char *)p->kva;
+}
+
+static inline unsigned efhw_page_pfn(efhw_page_t *p)
+{
+	return (unsigned)(__pa(p->kva) >> PAGE_SHIFT);
+}
+
+static inline void efhw_page_mark_invalid(efhw_page_t *p)
+{
+	p->kva = 0;
+}
+
+static inline int efhw_page_is_valid(efhw_page_t *p)
+{
+	return p->kva != 0;
+}
+
+static inline void efhw_page_init_from_va(efhw_page_t *p, void *va)
+{
+	p->kva = (unsigned long)va;
+}
+
+/*--------------------------------------------------------------------
+ *
+ * efhw_iopage_t: A single page of memory.  Directly mapped in the driver,
+ * and can be mapped to userlevel.  Can also be accessed by the NIC.
+ *
+ *--------------------------------------------------------------------*/
+
+typedef struct {
+	efhw_page_t p;
+	dma_addr_t dma_addr;
+} efhw_iopage_t;
+
+static inline dma_addr_t efhw_iopage_dma_addr(efhw_iopage_t *p)
+{
+	return p->dma_addr;
+}
+
+#define efhw_iopage_ptr(iop)		efhw_page_ptr(&(iop)->p)
+#define efhw_iopage_pfn(iop)		efhw_page_pfn(&(iop)->p)
+#define efhw_iopage_mark_invalid(iop)	efhw_page_mark_invalid(&(iop)->p)
+#define efhw_iopage_is_valid(iop)	efhw_page_is_valid(&(iop)->p)
+
+/*--------------------------------------------------------------------
+ *
+ * efhw_iopages_t: A set of pages that are contiguous in physical memory.
+ * Directly mapped in the driver, and can be mapped to userlevel.  Can also
+ * be accessed by the NIC.
+ *
+ * NB. The O/S may be unwilling to allocate many, or even any of these.  So
+ * only use this type where the NIC really needs a physically contiguous
+ * buffer.
+ *
+ *--------------------------------------------------------------------*/
+
+typedef struct {
+	caddr_t kva;
+	unsigned order;
+	dma_addr_t dma_addr;
+} efhw_iopages_t;
+
+static inline caddr_t efhw_iopages_ptr(efhw_iopages_t *p)
+{
+	return p->kva;
+}
+
+static inline unsigned efhw_iopages_pfn(efhw_iopages_t *p)
+{
+	return (unsigned)(__pa(p->kva) >> PAGE_SHIFT);
+}
+
+static inline dma_addr_t efhw_iopages_dma_addr(efhw_iopages_t *p)
+{
+	return p->dma_addr;
+}
+
+static inline unsigned efhw_iopages_size(efhw_iopages_t *p)
+{
+	return 1u << (p->order + PAGE_SHIFT);
+}
+
+/* efhw_iopage_t <-> efhw_iopages_t conversions for handling physically
+ * contiguous allocations in iobufsets for iSCSI.  This allows the
+ * essential information about contiguous allocations from
+ * efhw_iopages_alloc() to be saved away in the efhw_iopage_t array in an
+ * iobufset.  (Changing the iobufset resource to use a union type would
+ * involve a lot of code changes, and make the iobufset's metadata larger
+ * which could be bad as it's supposed to fit into a single page on some
+ * platforms.)
+ */
+static inline void
+efhw_iopage_init_from_iopages(efhw_iopage_t *iopage,
+			    efhw_iopages_t *iopages, unsigned pageno)
+{
+	iopage->p.kva = ((unsigned long)efhw_iopages_ptr(iopages))
+	    + (pageno * PAGE_SIZE);
+	iopage->dma_addr = efhw_iopages_dma_addr(iopages) +
+	    (pageno * PAGE_SIZE);
+}
+
+static inline void
+efhw_iopages_init_from_iopage(efhw_iopages_t *iopages,
+			    efhw_iopage_t *iopage, unsigned order)
+{
+	iopages->kva = (caddr_t) efhw_iopage_ptr(iopage);
+	EFHW_ASSERT(iopages->kva);
+	iopages->order = order;
+	iopages->dma_addr = efhw_iopage_dma_addr(iopage);
+}
+
+#endif /* __CI_EFHW_IOPAGE_LINUX_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/nic.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/nic.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/nic.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/nic.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,62 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains API provided by efhw/nic.c file.  This file is not
+ * designed for use outside of the SFC resource driver.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_NIC_H__
+#define __CI_EFHW_NIC_H__
+
+#include <ci/efhw/efhw_types.h>
+#include <ci/efhw/public.h>
+
+
+/* Convert PCI info to device type.  Returns false when device is not
+ * recognised.
+ */
+extern int efhw_device_type_init(struct efhw_device_type *dt,
+				 int vendor_id, int device_id, int revision);
+
+/* Initialise fields that do not involve touching hardware. */
+extern void efhw_nic_init(struct efhw_nic *nic, unsigned flags,
+			  unsigned options, struct efhw_device_type dev_type);
+
+/*! Destruct NIC resources */
+extern void efhw_nic_dtor(struct efhw_nic *nic);
+
+/*! Shutdown interrupts */
+extern void efhw_nic_close_interrupts(struct efhw_nic *nic);
+
+#endif /* __CI_EFHW_NIC_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/public.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/public.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/public.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/public.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,83 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides public API of efhw library exported from the SFC
+ * resource driver.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_PUBLIC_H__
+#define __CI_EFHW_PUBLIC_H__
+
+#include <ci/efhw/common.h>
+#include <ci/efhw/efhw_types.h>
+
+/*! Returns true if we have some EtherFabric functional units -
+  whether configured or not */
+static inline int efhw_nic_have_functional_units(struct efhw_nic *nic)
+{
+	return nic->efhw_func != 0;
+}
+
+/*! Returns true if the EtherFabric functional units have been configured  */
+static inline int efhw_nic_have_hw(struct efhw_nic *nic)
+{
+	return efhw_nic_have_functional_units(nic) && (EFHW_KVA(nic) != 0);
+}
+
+/*! Helper function to allocate the iobuffer needed by an eventq
+ *   - it ensures the eventq has the correct alignment for the NIC
+ *
+ * \param rm        Event-queue resource manager
+ * \param instance  Event-queue instance (index)
+ * \param buf_bytes Requested size of eventq
+ * \return          < 0 if iobuffer allocation fails
+ */
+int efhw_nic_event_queue_alloc_iobuffer(struct efhw_nic *nic,
+					struct eventq_resource_hardware *h,
+					int evq_instance, unsigned buf_bytes);
+
+extern void falcon_nic_set_rx_usr_buf_size(struct efhw_nic *,
+					   int rx_usr_buf_size);
+
+extern void
+falcon_nic_rx_filter_ctl_set(struct efhw_nic *nic, uint32_t tcp_full,
+			     uint32_t tcp_wild,
+			     uint32_t udp_full, uint32_t udp_wild);
+
+extern void
+falcon_nic_rx_filter_ctl_get(struct efhw_nic *nic, uint32_t *tcp_full,
+			     uint32_t *tcp_wild,
+			     uint32_t *udp_full, uint32_t *udp_wild);
+
+#endif /* __CI_EFHW_PUBLIC_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/sysdep.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/sysdep.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efhw/sysdep.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efhw/sysdep.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,72 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides version-independent Linux kernel API for efhw library.
+ * Only kernels >=2.6.9 are supported.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_SYSDEP_LINUX_H__
+#define __CI_EFHW_SYSDEP_LINUX_H__
+
+#include <linux/version.h>
+#include <linux/module.h>
+#include <linux/spinlock.h>
+#include <linux/delay.h>
+#include <linux/if_ether.h>
+
+#include <linux/netdevice.h> /* necessary for etherdevice.h on some kernels */
+#include <linux/etherdevice.h>
+
+#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,21)
+static inline int is_local_ether_addr(const u8 *addr)
+{
+	return (0x02 & addr[0]);
+}
+#endif
+
+typedef unsigned long irq_flags_t;
+
+#define spin_lock_destroy(l_)  do {} while (0)
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,24)
+#define HAS_NET_NAMESPACE
+#endif
+
+/* Funny, but linux has round_up for x86 only, defined in
+ * x86-specific header */
+#ifndef round_up
+#define round_up(x, y) (((x) + (y) - 1) & ~((y)-1))
+#endif
+
+#endif /* __CI_EFHW_SYSDEP_LINUX_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/buddy.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/buddy.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/buddy.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/buddy.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,69 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides private API for buddy allocator.  This API is not
+ * designed for use outside of SFC resource driver.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_BUDDY_H__
+#define __CI_EFRM_BUDDY_H__
+
+#include <ci/efrm/sysdep.h>
+
+/*! Comment? */
+struct efrm_buddy_allocator {
+	struct list_head *free_lists;	/* array[order+1] */
+	struct list_head *links;	/* array[1<<order] */
+	uint8_t *orders;		/* array[1<<order] */
+	unsigned order;		/*!< total size == (1 << order) */
+	/* ?? Consider recording largest available order + for each order the
+	 ** smallest available order that is big enough.
+	 */
+};
+
+  /*! Returns total size of managed space. */
+static inline unsigned long efrm_buddy_size(struct efrm_buddy_allocator *b)
+{
+	return 1ul << b->order;
+}
+
+int efrm_buddy_ctor(struct efrm_buddy_allocator *b, unsigned order);
+void efrm_buddy_dtor(struct efrm_buddy_allocator *b);
+int efrm_buddy_alloc(struct efrm_buddy_allocator *b, unsigned order);
+void efrm_buddy_free(struct efrm_buddy_allocator *b, unsigned addr,
+		     unsigned order);
+void efrm_buddy_reserve_at_start(struct efrm_buddy_allocator *b, unsigned n);
+void efrm_buddy_reserve_at_end(struct efrm_buddy_allocator *b, unsigned n);
+
+#endif /* __CI_EFRM_BUDDY_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/buffer_table.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/buffer_table.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/buffer_table.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/buffer_table.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,86 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides private buffer table API.  This API is not designed
+ * for use outside of SFC resource driver.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_BUFFER_TABLE_H__
+#define __CI_EFRM_BUFFER_TABLE_H__
+
+#include <ci/efhw/efhw_types.h>
+
+/*--------------------------------------------------------------------
+ *
+ * NIC's buffer table.
+ *
+ *--------------------------------------------------------------------*/
+
+/*! Managed interface. */
+
+/*! construct a managed buffer table object, allocated over a region of
+ *  the NICs buffer table space
+ */
+extern int efrm_buffer_table_ctor(unsigned low, unsigned high);
+/*! destructor for above */
+extern void efrm_buffer_table_dtor(void);
+
+/*! allocate a contiguous region of buffer table space */
+extern int efrm_buffer_table_alloc(unsigned order,
+				   struct efhw_buffer_table_allocation *a);
+
+/*! current size of the buffer table.
+ * FIXME This function should be inline, but it is never used from
+ * the fast path, so let it as-is. */
+unsigned long efrm_buffer_table_size(void);
+
+/*--------------------------------------------------------------------
+ *
+ * buffer table operations through the HW independent API
+ *
+ *--------------------------------------------------------------------*/
+
+/*! free a previously allocated region of buffer table space */
+extern void efrm_buffer_table_free(struct efhw_buffer_table_allocation *a);
+
+/*! commit the update of a buffer table entry to every NIC */
+void efrm_buffer_table_commit(void);
+
+/*! set a given buffer table entry. [pa] should be the physical
+  address of pinned down memory. This function can only be called from
+  the char driver */
+void efrm_buffer_table_set(struct efhw_buffer_table_allocation *a,
+			   unsigned i, dma_addr_t dma_addr, int owner);
+
+#endif /* __CI_EFRM_BUFFER_TABLE_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/debug.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/debug.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/debug.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/debug.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,78 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides debug-related API for efrm library using Linux kernel
+ * primitives.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_DEBUG_LINUX_H__
+#define __CI_EFRM_DEBUG_LINUX_H__
+
+#define EFRM_PRINTK_PREFIX "[sfc efrm] "
+
+#define EFRM_PRINTK(level, fmt, ...) \
+	printk(level EFRM_PRINTK_PREFIX fmt "\n", __VA_ARGS__)
+
+/* Following macros should be used with non-zero format parameters
+ * due to __VA_ARGS__ limitations.  Use "%s" with __FUNCTION__ if you can't
+ * find better parameters. */
+#define EFRM_ERR(fmt, ...)     EFRM_PRINTK(KERN_ERR, fmt, __VA_ARGS__)
+#define EFRM_WARN(fmt, ...)    EFRM_PRINTK(KERN_WARNING, fmt, __VA_ARGS__)
+#define EFRM_NOTICE(fmt, ...)  EFRM_PRINTK(KERN_NOTICE, fmt, __VA_ARGS__)
+#if 0 && !defined(NDEBUG)
+#define EFRM_TRACE(fmt, ...) EFRM_PRINTK(KERN_DEBUG, fmt, __VA_ARGS__)
+#else
+#define EFRM_TRACE(fmt, ...)
+#endif
+
+#ifndef NDEBUG
+#define EFRM_ASSERT(cond)  BUG_ON((cond) == 0)
+#define _EFRM_ASSERT(cond, file, line) \
+	do {								\
+		if (unlikely(!(cond))) {				\
+			EFRM_ERR("assertion \"%s\" failed at %s %d",	\
+				 #cond, file, line);			\
+			BUG();						\
+		}							\
+	} while (0)
+
+#define EFRM_DO_DEBUG(expr) expr
+#define EFRM_VERIFY_EQ(expr, val) EFRM_ASSERT((expr) == (val))
+#else
+#define EFRM_ASSERT(cond)
+#define EFRM_DO_DEBUG(expr)
+#define EFRM_VERIFY_EQ(expr, val) expr
+#endif
+
+#endif /* __CI_EFRM_DEBUG_LINUX_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/driver_private.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/driver_private.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/driver_private.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/driver_private.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,86 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides private API of efrm library to be used from the SFC
+ * resource driver.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_DRIVER_PRIVATE_H__
+#define __CI_EFRM_DRIVER_PRIVATE_H__
+
+#include <ci/efrm/resource.h>
+#include <ci/efrm/sysdep.h>
+
+/*--------------------------------------------------------------------
+ *
+ * global variables
+ *
+ *--------------------------------------------------------------------*/
+
+/* Internal structure for resource driver */
+extern struct efrm_resource_manager *efrm_rm_table[];
+
+/*--------------------------------------------------------------------
+ *
+ * efrm_nic_table handling
+ *
+ *--------------------------------------------------------------------*/
+
+extern int efrm_driver_ctor(void);
+extern int efrm_driver_dtor(void);
+extern int efrm_driver_register_nic(struct efhw_nic *, int nic_index);
+extern int efrm_driver_unregister_nic(struct efhw_nic *);
+
+/*--------------------------------------------------------------------
+ *
+ * create/destroy resource managers
+ *
+ *--------------------------------------------------------------------*/
+
+struct vi_resource_dimensions {
+	unsigned evq_int_min, evq_int_max;
+	unsigned evq_timer_min, evq_timer_max;
+	unsigned rxq_min, rxq_max;
+	unsigned txq_min, txq_max;
+};
+
+/*! Initialise resources */
+extern int
+efrm_resources_init(const struct vi_resource_dimensions *,
+		    int buffer_table_min, int buffer_table_max);
+
+/*! Tear down resources */
+extern void efrm_resources_fini(void);
+
+#endif /* __CI_EFRM_DRIVER_PRIVATE_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/filter.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/filter.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/filter.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/filter.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,147 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides public API for filter resource.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_FILTER_H__
+#define __CI_EFRM_FILTER_H__
+
+#include <ci/efrm/resource.h>
+#include <ci/efrm/vi_resource.h>
+#include <ci/efrm/nic_set.h>
+#include <ci/efhw/common.h>
+
+/*! Comment? */
+struct filter_resource {
+	struct efrm_resource rs;
+	struct vi_resource *pt;
+	int filter_idx;
+	efrm_nic_set_t nic_set;
+};
+
+#define filter_resource(rs1)  container_of((rs1), struct filter_resource, rs)
+
+/*!
+ * Allocate filter resource.
+ *
+ * \param vi_parent VI resource to use as parent. The function takes
+ *                  reference to the VI resource on success.
+ * \param frs_out   pointer to return the new filter resource
+ *
+ * \return          status code; if non-zero, frs_out is unchanged
+ */
+extern int
+efrm_filter_resource_alloc(struct vi_resource *vi_parent,
+			   struct filter_resource **frs_out);
+
+/* efrm_filter_resource_free should be called only if
+ * __efrm_resource_ref_count_zero() returned true.
+ * The easiest way is to call efrm_filter_resource_release() */
+void efrm_filter_resource_free(struct filter_resource *frs);
+static inline void efrm_filter_resource_release(struct filter_resource *frs)
+{
+	unsigned id;
+
+	EFRM_RESOURCE_ASSERT_VALID(&frs->rs, 0);
+	id = EFRM_RESOURCE_INSTANCE(frs->rs.rs_handle);
+
+	if (atomic_dec_and_test(&frs->rs.rs_ref_count)) {
+		if (__efrm_resource_ref_count_zero(EFRM_RESOURCE_FILTER, id)) {
+			EFRM_ASSERT(EFRM_RESOURCE_INSTANCE(frs->rs.rs_handle) ==
+				    id);
+			efrm_filter_resource_free(frs);
+		}
+	}
+}
+
+/*--------------------------------------------------------------------
+ *!
+ * Called to set/change the PT endpoint of a filter
+ *
+ * Example of use is TCP helper when it finds a wildcard IP filter
+ * needs to change which application it delivers traffic to
+ *
+ * \param frs           filter resource
+ * \param pt_handle     handle of new PT endpoint
+ *
+ * \return              standard error codes
+ *
+ *--------------------------------------------------------------------*/
+extern int
+efrm_filter_resource_set_ptresource(struct filter_resource *frs,
+				    struct vi_resource *virs);
+
+extern int efrm_filter_resource_clear(struct filter_resource *frs);
+
+extern int __efrm_filter_resource_set(struct filter_resource *frs, int type,
+				      unsigned saddr_be32, uint16_t sport_be16,
+				      unsigned daddr_be32, uint16_t dport_be16);
+
+static inline int
+efrm_filter_resource_tcp_set(struct filter_resource *frs,
+			     unsigned saddr, uint16_t sport,
+			     unsigned daddr, uint16_t dport)
+{
+	int type;
+
+	EFRM_ASSERT((saddr && sport) || (!saddr && !sport));
+
+	type =
+	    saddr ? EFHW_IP_FILTER_TYPE_TCP_FULL :
+	    EFHW_IP_FILTER_TYPE_TCP_WILDCARD;
+
+	return __efrm_filter_resource_set(frs, type,
+					  saddr, sport, daddr, dport);
+}
+
+static inline int
+efrm_filter_resource_udp_set(struct filter_resource *frs,
+			     unsigned saddr, uint16_t sport,
+			     unsigned daddr, uint16_t dport)
+{
+	int type;
+
+	EFRM_ASSERT((saddr && sport) || (!saddr && !sport));
+
+	type =
+	    saddr ? EFHW_IP_FILTER_TYPE_UDP_FULL :
+	    EFHW_IP_FILTER_TYPE_UDP_WILDCARD;
+
+	return __efrm_filter_resource_set(frs,
+					  type, saddr, sport, daddr, dport);
+}
+
+#endif /* __CI_EFRM_FILTER_H__ */
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/iobufset.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/iobufset.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/iobufset.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/iobufset.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,123 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides public API for iobufset resource.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_IOBUFSET_H__
+#define __CI_EFRM_IOBUFSET_H__
+
+#include <ci/efrm/vi_resource.h>
+
+/*! Iobufset resource structture.
+ * Users should not access the structure fields directly, but use the API
+ * below.
+ * However, this structure should not be moved out of public headers,
+ * because part of API (ex. efrm_iobufset_dma_addr function) is inline and
+ * is used in the fast-path code.
+ */
+struct iobufset_resource {
+	struct efrm_resource rs;
+	struct vi_resource *evq;
+	struct efhw_buffer_table_allocation buf_tbl_alloc;
+	unsigned int faultonaccess;
+	unsigned int n_bufs;
+	unsigned int pages_per_contiguous_chunk;
+	unsigned order;
+	efhw_iopage_t bufs[1];
+	/*!< up to n_bufs can follow this, so this must be the last member */
+};
+
+#define iobufset_resource(rs1) \
+	container_of((rs1), struct iobufset_resource, rs)
+
+/*!
+ * Allocate iobufset resource.
+ *
+ * \param vi_evq    VI resource to use. The function takes
+ *                  reference to the VI resource on success.
+ * \param iobrs_out   pointer to return the new filter resource
+ *
+ * \return          status code; if non-zero, frs_out is unchanged
+ */
+extern int
+efrm_iobufset_resource_alloc(int32_t n_pages,
+			     int32_t pages_per_contiguous_chunk,
+			     struct vi_resource *vi_evq,
+			     bool phys_addr_mode,
+			     uint32_t faultonaccess,
+			     struct iobufset_resource **iobrs_out);
+
+/* efrm_iobufset_resource_free should be called only if
+ * __efrm_resource_ref_count_zero() returned true.
+ * The easiest way is to call efrm_iobufset_resource_release() */
+void efrm_iobufset_resource_free(struct iobufset_resource *rs);
+static inline void
+efrm_iobufset_resource_release(struct iobufset_resource *iobrs)
+{
+	unsigned id;
+
+	EFRM_RESOURCE_ASSERT_VALID(&iobrs->rs, 0);
+	id = EFRM_RESOURCE_INSTANCE(iobrs->rs.rs_handle);
+
+	if (atomic_dec_and_test(&iobrs->rs.rs_ref_count)) {
+		if (__efrm_resource_ref_count_zero(EFRM_RESOURCE_IOBUFSET, id))
+			efrm_iobufset_resource_free(iobrs);
+	}
+}
+
+static inline char *
+efrm_iobufset_ptr(struct iobufset_resource *rs, unsigned offs)
+{
+	EFRM_ASSERT(offs < (unsigned)(rs->n_bufs << PAGE_SHIFT));
+	return efhw_iopage_ptr(&rs->bufs[offs >> PAGE_SHIFT])
+	    + (offs & (PAGE_SIZE - 1));
+}
+
+static inline char *efrm_iobufset_page_ptr(struct iobufset_resource *rs,
+				       unsigned page_i)
+{
+	EFRM_ASSERT(page_i < (unsigned)rs->n_bufs);
+	return efhw_iopage_ptr(&rs->bufs[page_i]);
+}
+
+static inline dma_addr_t
+efrm_iobufset_dma_addr(struct iobufset_resource *rs, unsigned offs)
+{
+	EFRM_ASSERT(offs < (unsigned)(rs->n_bufs << PAGE_SHIFT));
+	return efhw_iopage_dma_addr(&rs->bufs[offs >> PAGE_SHIFT])
+	    + (offs & (PAGE_SIZE - 1));
+}
+
+#endif /* __CI_EFRM_IOBUFSET_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/nic_set.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/nic_set.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/nic_set.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/nic_set.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,104 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides public API for NIC sets.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_NIC_SET_H__
+#define __CI_EFRM_NIC_SET_H__
+
+#include <ci/efrm/debug.h>
+#include <ci/efhw/common_sysdep.h>
+#include <ci/efhw/efhw_config.h>
+
+/*--------------------------------------------------------------------
+ *
+ * efrm_nic_set_t - tracks which NICs something has been done on
+ *
+ *--------------------------------------------------------------------*/
+
+/* Internal suructure of efrm_nic_set_t should not be referenced outside of
+ * this file.  Add a new accessor if you should do it. */
+typedef struct {
+	uint32_t nics;
+} efrm_nic_set_t;
+
+#if EFHW_MAX_NR_DEVS > 32
+#error change efrm_nic_set to handle EFHW_MAX_NR_DEVS number of devices
+#endif
+
+static inline bool
+efrm_nic_set_read(const efrm_nic_set_t *nic_set, unsigned index)
+{
+	EFRM_ASSERT(nic_set);
+	EFRM_ASSERT(index < EFHW_MAX_NR_DEVS && index < 32);
+	return (nic_set->nics & (1 << index)) ? true : false;
+}
+
+static inline void
+efrm_nic_set_write(efrm_nic_set_t *nic_set, unsigned index, bool value)
+{
+	EFRM_ASSERT(nic_set);
+	EFRM_ASSERT(index < EFHW_MAX_NR_DEVS && index < 32);
+	EFRM_ASSERT(value == false || value == true);
+	nic_set->nics = (nic_set->nics & (~(1 << index))) + (value << index);
+}
+
+static inline void efrm_nic_set_clear(efrm_nic_set_t *nic_set)
+{
+	nic_set->nics = 0;
+}
+
+static inline void efrm_nic_set_all(efrm_nic_set_t *nic_set)
+{
+	nic_set->nics = 0xffffffff;
+}
+
+static inline bool efrm_nic_set_is_all_clear(efrm_nic_set_t *nic_set)
+{
+	return nic_set->nics == 0 ? true : false;
+}
+
+#define EFRM_NIC_SET_FMT "%x"
+
+static inline uint32_t efrm_nic_set_pri_arg(efrm_nic_set_t *nic_set)
+{
+	return nic_set->nics;
+}
+
+#define EFRM_FOR_EACH_NIC_INDEX_IN_SET(_set, _nic_i)			\
+	for ((_nic_i) = 0; (_nic_i) < EFHW_MAX_NR_DEVS; ++(_nic_i))	\
+		if (efrm_nic_set_read((_set), (_nic_i)))
+
+#endif /* __CI_EFRM_NIC_SET_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/nic_table.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/nic_table.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/nic_table.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/nic_table.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,98 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides public API for NIC table.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_NIC_TABLE_H__
+#define __CI_EFRM_NIC_TABLE_H__
+
+#include <ci/efhw/efhw_types.h>
+#include <ci/efrm/sysdep.h>
+
+/*--------------------------------------------------------------------
+ *
+ * struct efrm_nic_table - top level driver object keeping all NICs -
+ * implemented in driver_object.c
+ *
+ *--------------------------------------------------------------------*/
+
+/*! Comment? */
+struct efrm_nic_table {
+	/*! nics attached to this driver */
+	struct efhw_nic *nic[EFHW_MAX_NR_DEVS];
+	/*! pointer to an arbitrary struct efhw_nic if one exists;
+	 * for code which does not care which NIC it wants but
+	 * still needs one. Note you cannot assume nic[0] exists. */
+	struct efhw_nic *a_nic;
+	uint32_t nic_count;	/*!< number of nics attached to this driver */
+	spinlock_t lock;	/*!< lock for table modifications */
+	atomic_t ref_count;	/*!< refcount for users of nic table */
+};
+
+/* Resource driver structures used by other drivers as well */
+extern struct efrm_nic_table efrm_nic_table;
+
+static inline void efrm_nic_table_hold(void)
+{
+	atomic_inc(&efrm_nic_table.ref_count);
+}
+
+static inline void efrm_nic_table_rele(void)
+{
+	atomic_dec(&efrm_nic_table.ref_count);
+}
+
+static inline int efrm_nic_table_held(void)
+{
+	return (atomic_read(&efrm_nic_table.ref_count) != 0);
+}
+
+/* Run code block _x multiple times with variable nic set to each
+ * registered NIC in turn.
+ * DO NOT "break" out of this loop early. */
+#define EFRM_FOR_EACH_NIC(_nic_i, _nic)					\
+	for ((_nic_i) = (efrm_nic_table_hold(), 0);			\
+	     (_nic_i) < EFHW_MAX_NR_DEVS || (efrm_nic_table_rele(), 0);	\
+	     (_nic_i)++)						\
+		if (((_nic) = efrm_nic_table.nic[_nic_i]))
+
+#define EFRM_FOR_EACH_NIC_IN_SET(_set, _i, _nic)			\
+	for ((_i) = (efrm_nic_table_hold(), 0);				\
+	     (_i) < EFHW_MAX_NR_DEVS || (efrm_nic_table_rele(), 0);	\
+	     ++(_i))							\
+		if (((_nic) = efrm_nic_table.nic[_i]) &&		\
+		    efrm_nic_set_read((_set), (_i)))
+
+#endif /* __CI_EFRM_NIC_TABLE_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/private.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/private.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/private.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/private.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,141 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides private API of efrm library -- resource handling.
+ * This API is not designed for use outside of SFC resource driver.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_PRIVATE_H__
+#define __CI_EFRM_PRIVATE_H__
+
+#include <ci/efrm/resource.h>
+#include <ci/efrm/driver_private.h>
+#include <ci/efrm/sysdep.h>
+#include <ci/efrm/debug.h>
+
+/*--------------------------------------------------------------------
+ *
+ * create resource managers
+ *
+ *--------------------------------------------------------------------*/
+
+/*! Create a resource manager for various types of resources
+ */
+extern int
+efrm_create_iobufset_resource_manager(struct efrm_resource_manager **out);
+
+extern int
+efrm_create_filter_resource_manager(struct efrm_resource_manager **out);
+
+extern int
+efrm_create_vi_resource_manager(struct efrm_resource_manager **out,
+				const struct vi_resource_dimensions *);
+
+/*--------------------------------------------------------------------
+ *
+ * efrm_resource_handle_t handling
+ *
+ *--------------------------------------------------------------------*/
+
+/*! Initialize an area of memory to be used as a resource */
+static inline void efrm_resource_init(struct efrm_resource *rs,
+				      int type, int instance)
+{
+	EFRM_ASSERT(instance >= 0);
+	EFRM_ASSERT(type >= 0 && type < EFRM_RESOURCE_NUM);
+	atomic_set(&rs->rs_ref_count, 1);
+	rs->rs_handle.handle = (type << 28u) |
+		(((unsigned)jiffies & 0xfff) << 16) | instance;
+}
+
+/*--------------------------------------------------------------------
+ *
+ * Instance pool management
+ *
+ *--------------------------------------------------------------------*/
+
+/*! Allocate instance pool. Use kfifo_vfree to destroy it. */
+static inline int
+efrm_kfifo_id_ctor(struct kfifo **ids_out,
+		   unsigned int base, unsigned int limit, spinlock_t *lock)
+{
+	unsigned int i;
+	struct kfifo *ids;
+	unsigned char *buffer;
+	unsigned int size = roundup_pow_of_two((limit - base) * sizeof(int));
+
+	EFRM_ASSERT(base <= limit);
+	buffer = vmalloc(size);
+	ids = kfifo_init(buffer, size, GFP_KERNEL, lock);
+	if (IS_ERR(ids))
+		return PTR_ERR(ids);
+	for (i = base; i < limit; i++)
+		EFRM_VERIFY_EQ(__kfifo_put(ids, (unsigned char *)&i,
+					   sizeof(i)), sizeof(i));
+
+	*ids_out = ids;
+	return 0;
+}
+
+/*--------------------------------------------------------------------
+ *
+ * Various private functions
+ *
+ *--------------------------------------------------------------------*/
+
+/*! Initialize the fields in the provided resource manager memory area
+ *   \param rm         The area of memory to be initialized
+ *   \param dtor       A method to destroy the resource manager
+ *   \param name       A Textual name for the resource manager
+ *   \param type       The type of resource managed
+ *   \param initial_table_size Initial size of the ID table
+ *   \param auto_destroy Destroy resource manager on driver onload iff true
+ *
+ * A default table size is provided if the value 0 is provided.
+ */
+extern int
+efrm_resource_manager_ctor(struct efrm_resource_manager *rm,
+			   void (*dtor)(struct efrm_resource_manager *),
+			   const char *name, unsigned type,
+			   int initial_table_size);
+
+extern void efrm_resource_manager_dtor(struct efrm_resource_manager *rm);
+
+/*! Insert a resource into table in the resource manager.
+ *
+ * Caller should free the resource if this function returns non-zero.
+ */
+extern int efrm_resource_manager_insert(struct efrm_resource *rs);
+
+#endif /* __CI_EFRM_PRIVATE_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/resource.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/resource.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/resource.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/resource.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,122 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides public interface of efrm library -- resource handling.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_RESOURCE_H__
+#define __CI_EFRM_RESOURCE_H__
+
+/*--------------------------------------------------------------------
+ *
+ * headers for type dependencies
+ *
+ *--------------------------------------------------------------------*/
+
+#include <ci/efhw/efhw_types.h>
+#include <ci/efrm/resource_id.h>
+#include <ci/efrm/sysdep.h>
+#include <ci/efhw/common_sysdep.h>
+
+#ifndef __ci_driver__
+#error "Driver-only file"
+#endif
+
+/*--------------------------------------------------------------------
+ *
+ * struct efrm_resource - represents an allocated resource
+ *                   (eg. pinned pages of memory, or resource on a NIC)
+ *
+ *--------------------------------------------------------------------*/
+
+/*! Representation of an allocated resource */
+struct efrm_resource {
+	atomic_t rs_ref_count; /*!< users count; see
+				* __efrm_resource_ref_count_zero() */
+	efrm_resource_handle_t rs_handle;
+};
+
+/*--------------------------------------------------------------------
+ *
+ * managed resource abstraction
+ *
+ *--------------------------------------------------------------------*/
+
+/*! Factory for resources of a specific type */
+struct efrm_resource_manager {
+	const char *rm_name;	/*!< human readable only */
+	spinlock_t rm_lock;
+#ifndef NDEBUG
+	unsigned rm_type;
+#endif
+	int rm_resources;
+	int rm_resources_hiwat;
+	/*! table of allocated resources */
+	struct efrm_resource **rm_table;
+	unsigned rm_table_size;
+	/**
+	 * Destructor for the resource manager. Other resource managers
+	 * might be already dead, although the system guarantees that
+	 * managers are destructed in the order by which they were created
+	 */
+	void (*rm_dtor)(struct efrm_resource_manager *);
+};
+
+#ifdef NDEBUG
+# define EFRM_RESOURCE_ASSERT_VALID(rs, rc_mbz)
+# define EFRM_RESOURCE_MANAGER_ASSERT_VALID(rm)
+#else
+/*! Check validity of resource and report on failure */
+extern void efrm_resource_assert_valid(struct efrm_resource *,
+				       int rc_may_be_zero,
+				       const char *file, int line);
+# define EFRM_RESOURCE_ASSERT_VALID(rs, rc_mbz) \
+	efrm_resource_assert_valid((rs), (rc_mbz), __FILE__, __LINE__)
+
+/*! Check validity of resource manager and report on failure */
+extern void efrm_resource_manager_assert_valid(struct efrm_resource_manager *,
+					       const char *file, int line);
+# define EFRM_RESOURCE_MANAGER_ASSERT_VALID(rm) \
+	efrm_resource_manager_assert_valid((rm), __FILE__, __LINE__)
+#endif
+
+/*! Check the reference count on the resource provided and delete its
+ *  handle it in its owning resource manager if the
+ *  reference count has fallen to zero.
+ *
+ *  Returns TRUE if the caller should really free the resource.
+ */
+extern bool __efrm_resource_ref_count_zero(unsigned type, unsigned instance);
+
+#endif /* __CI_EFRM_RESOURCE_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/resource_id.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/resource_id.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/resource_id.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/resource_id.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,104 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides public type and definitions resource handle, and the
+ * definitions of resource types.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_DRIVER_EFRM_RESOURCE_ID_H__
+#define __CI_DRIVER_EFRM_RESOURCE_ID_H__
+
+/***********************************************************************
+ * Resource handles
+ *
+ * Resource handles are intended for identifying resources at kernel
+ * level, within the context of a particular NIC. particularly because
+ * for some resource types, the low 16 bites correspond to hardware
+ * IDs. They were historically also used at user level, with a nonce
+ * stored in the bits 16 to 27 (inclusive), but that approach is
+ * deprecated (but sill alive!).
+ *
+ * The handle value 0 is used to mean "no resource".
+ * Identify resources within the context of a file descriptor at user
+ * level.
+ ***********************************************************************/
+
+typedef struct efrm_resource_handle_s {
+	uint32_t handle;
+} efrm_resource_handle_t;
+
+/* You may think these following functions should all have
+ * _HANDLE_ in their names, but really we are providing an abstract set
+ * of methods on a (hypothetical) efrm_resource_t object, with
+ * efrm_resource_handle_t being just the reference one holds to access
+ * the object (aka "this" or "self").
+ */
+
+/* Below I use inline instead of macros where possible in order to get
+ * more type checking help from the compiler; hopefully we'll never
+ * have to rewrite these to use #define as we've found some horrible
+ * compiler on which we cannot make static inline do the Right Thing (tm).
+ *
+ * For consistency and to avoid pointless change I spell these
+ * routines as macro names (CAPTILIZE_UNDERSCORED), which also serves
+ * to remind people they are compact and inlined.
+ */
+
+#define EFRM_RESOURCE_FMT  "[rs:%08x]"
+
+static inline unsigned EFRM_RESOURCE_PRI_ARG(efrm_resource_handle_t h)
+{
+	return (h.handle);
+}
+
+static inline unsigned EFRM_RESOURCE_INSTANCE(efrm_resource_handle_t h)
+{
+	return (h.handle & 0x0000ffff);
+}
+
+static inline unsigned EFRM_RESOURCE_TYPE(efrm_resource_handle_t h)
+{
+	return (h.handle & 0xf0000000) >> 28;
+}
+
+/***********************************************************************
+ * Resource type codes
+ ***********************************************************************/
+
+#define EFRM_RESOURCE_IOBUFSET          0x0
+#define EFRM_RESOURCE_VI                0x1
+#define EFRM_RESOURCE_FILTER            0x2
+#define EFRM_RESOURCE_NUM               0x3	/* This isn't a resource! */
+
+#define	EFRM_RESOURCE_NAME(type) \
+	((type) == EFRM_RESOURCE_IOBUFSET?	"IOBUFSET"	: \
+	 (type) == EFRM_RESOURCE_VI?		"VI"		: \
+	 (type) == EFRM_RESOURCE_FILTER?	"FILTER"	: \
+						"<invalid>")
+
+#endif /* __CI_DRIVER_EFRM_RESOURCE_ID_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/sysdep.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/sysdep.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/sysdep.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/sysdep.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,54 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides Linux-like system-independent API for efrm library.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_SYSDEP_H__
+#define __CI_EFRM_SYSDEP_H__
+
+/* Spinlocks are defined in efhw/sysdep.h */
+#include <ci/efhw/sysdep.h>
+
+#if defined(__linux__) && defined(__KERNEL__)
+
+# include <ci/efrm/sysdep_linux.h>
+
+#else
+
+# include <ci/efrm/sysdep_ci2linux.h>
+
+#endif
+
+#endif /* __CI_EFRM_SYSDEP_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/sysdep_linux.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/sysdep_linux.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/sysdep_linux.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/sysdep_linux.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,248 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides version-independent Linux kernel API for efrm library.
+ * Only kernels >=2.6.9 are supported.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Kfifo API is partially stolen from linux-2.6.22/include/linux/list.h
+ * Copyright (C) 2004 Stelian Pop <stelian@popies.net>
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_SYSDEP_LINUX_H__
+#define __CI_EFRM_SYSDEP_LINUX_H__
+
+#include <linux/version.h>
+#include <linux/list.h>
+#include <linux/vmalloc.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/workqueue.h>
+#include <linux/gfp.h>
+#include <linux/slab.h>
+#include <linux/hardirq.h>
+#include <linux/kernel.h>
+#include <linux/if_ether.h>
+#include <linux/completion.h>
+#include <linux/in.h>
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20)
+/* get roundup_pow_of_two(), which was in kernel.h in early kernel versions */
+#include <linux/log2.h>
+#endif
+
+/********************************************************************
+ *
+ * List API
+ *
+ ********************************************************************/
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,18)
+static inline void
+list_replace_init(struct list_head *old, struct list_head *new)
+{
+	new->next = old->next;
+	new->next->prev = new;
+	new->prev = old->prev;
+	new->prev->next = new;
+	INIT_LIST_HEAD(old);
+}
+#endif
+
+static inline struct list_head *list_pop(struct list_head *list)
+{
+	struct list_head *link = list->next;
+	list_del(link);
+	return link;
+}
+
+static inline struct list_head *list_pop_tail(struct list_head *list)
+{
+	struct list_head *link = list->prev;
+	list_del(link);
+	return link;
+}
+
+/********************************************************************
+ *
+ * Workqueue API
+ *
+ ********************************************************************/
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,20)
+#define NEED_OLD_WORK_API
+
+/**
+ * The old and new work function prototypes just change
+ * the type of the pointer in the only argument, so it's
+ * safe to cast one function type to the other
+ */
+typedef void (*efrm_old_work_func_t) (void *p);
+
+#undef INIT_WORK
+#define INIT_WORK(_work, _func)					\
+	do {							\
+		INIT_LIST_HEAD(&(_work)->entry);		\
+		(_work)->pending = 0;				\
+		PREPARE_WORK((_work),				\
+			     (efrm_old_work_func_t) (_func),	\
+			     (_work));				\
+	} while (0)
+
+#endif
+
+/********************************************************************
+ *
+ * Kfifo API
+ *
+ ********************************************************************/
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,10)
+
+#if !defined(RHEL_RELEASE_CODE) || (RHEL_RELEASE_CODE < 1029)
+typedef unsigned gfp_t;
+#endif
+
+#define HAS_NO_KFIFO
+
+struct kfifo {
+	unsigned char *buffer;	/* the buffer holding the data */
+	unsigned int size;	/* the size of the allocated buffer */
+	unsigned int in;	/* data is added at offset (in % size) */
+	unsigned int out;	/* data is extracted from off. (out % size) */
+	spinlock_t *lock;	/* protects concurrent modifications */
+};
+
+extern struct kfifo *kfifo_init(unsigned char *buffer, unsigned int size,
+				gfp_t gfp_mask, spinlock_t *lock);
+extern struct kfifo *kfifo_alloc(unsigned int size, gfp_t gfp_mask,
+				 spinlock_t *lock);
+extern void kfifo_free(struct kfifo *fifo);
+extern unsigned int __kfifo_put(struct kfifo *fifo,
+				unsigned char *buffer, unsigned int len);
+extern unsigned int __kfifo_get(struct kfifo *fifo,
+				unsigned char *buffer, unsigned int len);
+
+/**
+ * kfifo_put - puts some data into the FIFO
+ * @fifo: the fifo to be used.
+ * @buffer: the data to be added.
+ * @len: the length of the data to be added.
+ *
+ * This function copies at most @len bytes from the @buffer into
+ * the FIFO depending on the free space, and returns the number of
+ * bytes copied.
+ */
+static inline unsigned int
+kfifo_put(struct kfifo *fifo, unsigned char *buffer, unsigned int len)
+{
+	unsigned long flags;
+	unsigned int ret;
+
+	spin_lock_irqsave(fifo->lock, flags);
+
+	ret = __kfifo_put(fifo, buffer, len);
+
+	spin_unlock_irqrestore(fifo->lock, flags);
+
+	return ret;
+}
+
+/**
+ * kfifo_get - gets some data from the FIFO
+ * @fifo: the fifo to be used.
+ * @buffer: where the data must be copied.
+ * @len: the size of the destination buffer.
+ *
+ * This function copies at most @len bytes from the FIFO into the
+ * @buffer and returns the number of copied bytes.
+ */
+static inline unsigned int
+kfifo_get(struct kfifo *fifo, unsigned char *buffer, unsigned int len)
+{
+	unsigned long flags;
+	unsigned int ret;
+
+	spin_lock_irqsave(fifo->lock, flags);
+
+	ret = __kfifo_get(fifo, buffer, len);
+
+	/*
+	 * optimization: if the FIFO is empty, set the indices to 0
+	 * so we don't wrap the next time
+	 */
+	if (fifo->in == fifo->out)
+		fifo->in = fifo->out = 0;
+
+	spin_unlock_irqrestore(fifo->lock, flags);
+
+	return ret;
+}
+
+/**
+ * __kfifo_len - returns the number of bytes available in the FIFO, no locking version
+ * @fifo: the fifo to be used.
+ */
+static inline unsigned int __kfifo_len(struct kfifo *fifo)
+{
+	return fifo->in - fifo->out;
+}
+
+/**
+ * kfifo_len - returns the number of bytes available in the FIFO
+ * @fifo: the fifo to be used.
+ */
+static inline unsigned int kfifo_len(struct kfifo *fifo)
+{
+	unsigned long flags;
+	unsigned int ret;
+
+	spin_lock_irqsave(fifo->lock, flags);
+
+	ret = __kfifo_len(fifo);
+
+	spin_unlock_irqrestore(fifo->lock, flags);
+
+	return ret;
+}
+
+#else
+#include <linux/kfifo.h>
+#endif
+
+static inline void kfifo_vfree(struct kfifo *fifo)
+{
+	vfree(fifo->buffer);
+	kfree(fifo);
+}
+
+#endif /* __CI_EFRM_SYSDEP_LINUX_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/vi_resource.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/vi_resource.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/vi_resource.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/vi_resource.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,171 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains public API for VI resource.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_VI_RESOURCE_H__
+#define __CI_EFRM_VI_RESOURCE_H__
+
+#include <ci/efhw/efhw_types.h>
+#include <ci/efrm/resource.h>
+#include <ci/efrm/debug.h>
+
+struct vi_resource;
+
+/* Make these inline instead of macros for type checking */
+static inline struct vi_resource *
+efrm_to_vi_resource(struct efrm_resource *rs)
+{
+	EFRM_ASSERT(EFRM_RESOURCE_TYPE(rs->rs_handle) == EFRM_RESOURCE_VI);
+	return (struct vi_resource *) rs;
+}
+static inline struct
+efrm_resource *efrm_from_vi_resource(struct vi_resource *rs)
+{
+	return (struct efrm_resource *)rs;
+}
+
+#define EFAB_VI_RESOURCE_INSTANCE(virs) \
+    EFRM_RESOURCE_INSTANCE(efrm_from_vi_resource(virs)->rs_handle)
+
+#define EFAB_VI_RESOURCE_PRI_ARG(virs) \
+    EFRM_RESOURCE_PRI_ARG(efrm_from_vi_resource(virs)->rs_handle)
+
+extern int
+efrm_vi_resource_alloc(struct vi_resource *evq_virs,
+		       uint16_t vi_flags, int32_t evq_capacity,
+		       int32_t txq_capacity, int32_t rxq_capacity,
+		       uint8_t tx_q_tag, uint8_t rx_q_tag,
+		       struct vi_resource **virs_in_out,
+		       uint32_t *out_io_mmap_bytes,
+		       uint32_t *out_mem_mmap_bytes,
+		       uint32_t *out_txq_capacity,
+		       uint32_t *out_rxq_capacity);
+
+static inline void efrm_vi_resource_ref(struct vi_resource *virs)
+{
+	atomic_inc(&efrm_from_vi_resource(virs)->rs_ref_count);
+}
+
+/* efrm_vi_resource_free should be called only if
+ * __efrm_resource_ref_count_zero() returned true.
+ * The easiest way is to call efrm_vi_resource_release() */
+extern void efrm_vi_resource_free(struct vi_resource *virs);
+static inline void efrm_vi_resource_release(struct vi_resource *virs)
+{
+	unsigned id;
+	struct efrm_resource *rs = efrm_from_vi_resource(virs);
+
+	id = EFRM_RESOURCE_INSTANCE(rs->rs_handle);
+
+	if (atomic_dec_and_test(&rs->rs_ref_count)) {
+		if (__efrm_resource_ref_count_zero(EFRM_RESOURCE_VI, id)) {
+			EFRM_ASSERT(EFRM_RESOURCE_INSTANCE(rs->rs_handle) ==
+				    id);
+			efrm_vi_resource_free(virs);
+		}
+	}
+}
+
+/*--------------------------------------------------------------------
+ *
+ * eventq handling
+ *
+ *--------------------------------------------------------------------*/
+
+/*! Reset an event queue and clear any associated timers */
+extern void efrm_eventq_reset(struct vi_resource *virs, int nic_index);
+
+/*! Register a kernel-level handler for the event queue.  This function is
+ * called whenever a timer expires, or whenever the event queue is woken
+ * but no thread is blocked on it.
+ *
+ * This function returns -EBUSY if a callback is already installed.
+ *
+ * \param rs      Event-queue resource
+ * \param handler Callback-handler
+ * \param arg     Argument to pass to callback-handler
+ * \return        Status code
+ */
+extern int
+efrm_eventq_register_callback(struct vi_resource *rs,
+			      void (*handler)(void *arg, int is_timeout,
+					      struct efhw_nic *nic),
+			      void *arg);
+
+/*! Kill the kernel-level callback.
+ *
+ * This function stops the timer from running and unregisters the callback
+ * function.  It waits for any running timeout handlers to complete before
+ * returning.
+ *
+ * \param rs      Event-queue resource
+ * \return        Nothing
+ */
+extern void efrm_eventq_kill_callback(struct vi_resource *rs);
+
+/*! Ask the NIC to generate a wakeup when an event is next delivered. */
+extern void efrm_eventq_request_wakeup(struct vi_resource *rs,
+				       unsigned current_ptr,
+				       unsigned nic_index);
+
+/*! Register a kernel-level handler for flush completions.
+ * \TODO Currently, it is unsafe to install a callback more than once.
+ *
+ * \param rs      VI resource being flushed.
+ * \param handler Callback handler function.
+ * \param arg     Argument to be passed to handler.
+ */
+extern void
+efrm_vi_register_flush_callback(struct vi_resource *rs,
+				void (*handler)(void *),
+				void *arg);
+
+int efrm_vi_resource_flush_retry(struct vi_resource *virs);
+
+/*! Comment? */
+extern int efrm_pt_flush(struct vi_resource *);
+
+/*! Comment? */
+extern int efrm_pt_pace(struct vi_resource *, unsigned int val);
+
+uint32_t efrm_vi_rm_txq_bytes(struct vi_resource *virs
+			      /*,struct efhw_nic *nic */ );
+uint32_t efrm_vi_rm_rxq_bytes(struct vi_resource *virs
+			      /*,struct efhw_nic *nic */ );
+uint32_t efrm_vi_rm_evq_bytes(struct vi_resource *virs
+			      /*,struct efhw_nic *nic */ );
+
+#endif /* __CI_EFRM_VI_RESOURCE_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/vi_resource_manager.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/vi_resource_manager.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/vi_resource_manager.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/vi_resource_manager.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,182 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains type definitions for VI resource.  These types
+ * may be used outside of the SFC resource driver, but such use is not
+ * recommended.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_DRIVER_EFAB_VI_RESOURCE_MANAGER_H__
+#define __CI_DRIVER_EFAB_VI_RESOURCE_MANAGER_H__
+
+#include <ci/efhw/common.h>
+#include <ci/efrm/vi_resource.h>
+#include <ci/efrm/nic_set.h>
+
+#define EFRM_VI_RM_DMA_QUEUE_COUNT 2
+#define EFRM_VI_RM_DMA_QUEUE_TX    0
+#define EFRM_VI_RM_DMA_QUEUE_RX    1
+
+/** Numbers of bits which can be set in the evq_state member of
+ * vi_resource_evq_info. */
+enum {
+  /** This bit is set if a wakeup has been requested on the NIC. */
+	VI_RESOURCE_EVQ_STATE_WAKEUP_PENDING,
+  /** This bit is set if the wakeup is valid for the sleeping
+   * process. */
+	VI_RESOURCE_EVQ_STATE_CALLBACK_REGISTERED,
+  /** This bit is set if a wakeup or timeout event is currently being
+   * processed. */
+	VI_RESOURCE_EVQ_STATE_BUSY,
+};
+#define VI_RESOURCE_EVQ_STATE(X) \
+	(((int32_t)1) << (VI_RESOURCE_EVQ_STATE_##X))
+
+/** Information about an event queue. */
+struct vi_resource_evq_info {
+  /** Flag bits indicating the state of wakeups. */
+	unsigned long evq_state;
+  /** A pointer to the resource instance for this queue.  This member
+   * is only valid if evq_state is non-zero or the resource is known
+   * to have a non-zero reference count. */
+	struct vi_resource *evq_virs;
+};
+
+#ifdef __ci_ul_driver__
+#define EFRM_VI_USE_WORKQUEUE 0
+#else
+#define EFRM_VI_USE_WORKQUEUE 1
+#endif
+
+/*! Global information for the VI resource manager. */
+struct vi_resource_manager {
+	struct efrm_resource_manager rm;
+
+	struct kfifo *instances_with_timer;
+	int with_timer_base;
+	int with_timer_limit;
+	struct kfifo *instances_with_interrupt;
+	int with_interrupt_base;
+	int with_interrupt_limit;
+
+	bool iscsi_dmaq_instance_is_free;
+	struct vi_resource_evq_info *evq_infos;
+
+	/* We keep VI resources which need flushing on these lists.  The VI
+	 * is put on the outstanding list when the flush request is issued
+	 * to the hardware and removed when the flush event arrives.  The
+	 * hardware can only handle a limited number of RX flush requests at
+	 * once, so VIs are placed in the waiting list until the flush can
+	 * be issued.  Flushes can be requested by the client or internally
+	 * by the VI resource manager.  In the former case, the reference
+	 * count must be non-zero for the duration of the flush and in the
+	 * later case, the reference count must be zero. */
+	struct list_head rx_flush_waiting_list;
+	struct list_head rx_flush_outstanding_list;
+	struct list_head tx_flush_outstanding_list;
+	int rx_flush_outstanding_count;
+
+	/* once the flush has happened we push the close into the work queue
+	 * so its OK on Windows to free the resources (Bug 3469).  Resources
+	 * on this list have zero reference count.
+	 */
+	struct list_head close_pending;
+	struct work_struct work_item;
+#if EFRM_VI_USE_WORKQUEUE
+	struct workqueue_struct *workqueue;
+#endif
+};
+
+struct vi_resource_nic_info {
+	struct eventq_resource_hardware evq_pages;
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+	efhw_iopages_t dmaq_pages[EFRM_VI_RM_DMA_QUEUE_COUNT];
+#endif
+};
+
+struct vi_resource {
+	/* Some macros make the assumption that the struct efrm_resource is
+	 * the first member of a struct vi_resource. */
+	struct efrm_resource rs;
+	atomic_t evq_refs;	/*!< Number of users of the event queue. */
+
+	efrm_nic_set_t nic_set;
+
+	uint32_t bar_mmap_bytes;
+	uint32_t mem_mmap_bytes;
+
+	int32_t evq_capacity;
+	int32_t dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_COUNT];
+
+	uint8_t dmaq_tag[EFRM_VI_RM_DMA_QUEUE_COUNT];
+	uint16_t flags;
+
+	/* we keep PT endpoints that have been destroyed on a list
+	 * until we have seen their TX and RX DMAQs flush complete
+	 * (see Bug 1217)
+	 */
+	struct list_head rx_flush_link;
+	struct list_head tx_flush_link;
+	efrm_nic_set_t rx_flush_nic_set;
+	efrm_nic_set_t rx_flush_outstanding_nic_set;
+	efrm_nic_set_t tx_flush_nic_set;
+	uint64_t flush_time;
+	int flush_count;
+
+	void (*flush_callback_fn)(void *);
+	void *flush_callback_arg;
+
+	void (*evq_callback_fn) (void *arg, int is_timeout,
+				 struct efhw_nic *nic);
+	void *evq_callback_arg;
+
+	struct vi_resource *evq_virs;	/*!< EVQ for DMA queues */
+
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+	 struct efhw_buffer_table_allocation
+	    dmaq_buf_tbl_alloc[EFRM_VI_RM_DMA_QUEUE_COUNT];
+#endif
+
+	struct vi_resource_nic_info nic_info[EFHW_MAX_NR_DEVS];
+};
+
+#undef vi_resource
+#define vi_resource(rs1)  container_of((rs1), struct vi_resource, rs)
+
+static inline dma_addr_t
+efrm_eventq_dma_addr(struct vi_resource *virs, uint32_t nic_index)
+{
+	struct eventq_resource_hardware *hw;
+	EFRM_ASSERT(efrm_nic_set_read(&virs->nic_set, nic_index));
+
+	hw = &(virs->nic_info[nic_index].evq_pages);
+
+	return efhw_iopages_dma_addr(&(hw->iobuff)) + hw->iobuff_off;
+}
+
+#endif /* __CI_DRIVER_EFAB_VI_RESOURCE_MANAGER_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/vi_resource_private.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/vi_resource_private.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/ci/efrm/vi_resource_private.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/ci/efrm/vi_resource_private.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,83 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains private API for VI resource.  The API is not designed
+ * to be used outside of the SFC resource driver.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_VI_RESOURCE_PRIVATE_H__
+#define __CI_EFRM_VI_RESOURCE_PRIVATE_H__
+
+#include <ci/efhw/common.h>
+#include <ci/efrm/vi_resource_manager.h>
+
+extern struct vi_resource_manager *efrm_vi_manager;
+
+/*************************************************************************/
+
+extern void efrm_vi_rm_delayed_free(struct work_struct *data);
+
+extern void efrm_vi_rm_salvage_flushed_vis(void);
+
+void efrm_vi_rm_free_flushed_resource(struct vi_resource *virs);
+
+void efrm_vi_rm_init_dmaq(struct vi_resource *virs, int queue_index,
+			  struct efhw_nic *nic);
+
+static inline int
+efrm_eventq_bytes(struct vi_resource *virs, uint32_t nic_index)
+{
+	EFRM_ASSERT(efrm_nic_set_read(&virs->nic_set, nic_index));
+
+	return efrm_vi_rm_evq_bytes(virs);
+}
+
+static inline efhw_event_t *
+efrm_eventq_base(struct vi_resource *virs, uint32_t nic_index)
+{
+	struct eventq_resource_hardware *hw;
+
+	EFRM_ASSERT(efrm_nic_set_read(&virs->nic_set, nic_index));
+
+	hw = &(virs->nic_info[nic_index].evq_pages);
+
+	return (efhw_event_t *) (efhw_iopages_ptr(&(hw->iobuff)) +
+				 hw->iobuff_off);
+}
+
+/*! Wakeup handler, see efhw_ev_handler_t for prototype */
+extern void efrm_handle_wakeup_event(struct efhw_nic *nic, efhw_event_t *ev);
+
+/*! Timeout handler, see efhw_ev_handler_t for prototype */
+extern void efrm_handle_timeout_event(struct efhw_nic *nic, efhw_event_t *ev);
+
+/*! DMA flush handler, see efhw_ev_handler_t for prototype */
+extern void efrm_handle_dmaq_flushed(struct efhw_nic *nic, int instance,
+				   int rx_flush);
+
+#endif /* __CI_EFRM_VI_RESOURCE_PRIVATE_H__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/driverlink_new.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/driverlink_new.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/driverlink_new.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/driverlink_new.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,290 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains driverlink code which interacts with the sfc network
+ * driver.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include "linux_resource_internal.h"
+#include "driverlink_api.h"
+#include "kernel_compat.h"
+#include <ci/efhw/falcon.h>
+
+#include <linux/rtnetlink.h>
+#include <linux/netdevice.h>
+
+/* The DL driver and associated calls */
+static int efrm_dl_probe(struct efx_dl_device *efrm_dev,
+			 const struct net_device *net_dev,
+			 const struct efx_dl_device_info *dev_info,
+			 const char *silicon_rev);
+
+static void efrm_dl_remove(struct efx_dl_device *efrm_dev);
+
+static void efrm_dl_reset_suspend(struct efx_dl_device *efrm_dev);
+
+static void efrm_dl_reset_resume(struct efx_dl_device *efrm_dev, int ok);
+
+static void efrm_dl_mtu_changed(struct efx_dl_device *, int);
+static void efrm_dl_event_falcon(struct efx_dl_device *efx_dev, void *p_event);
+
+static struct efx_dl_driver efrm_dl_driver = {
+	.name = "resource",
+	.probe = efrm_dl_probe,
+	.remove = efrm_dl_remove,
+	.reset_suspend = efrm_dl_reset_suspend,
+	.reset_resume = efrm_dl_reset_resume
+};
+
+static void
+init_vi_resource_dimensions(struct vi_resource_dimensions *rd,
+			    const struct efx_dl_falcon_resources *res)
+{
+	rd->evq_timer_min = res->evq_timer_min;
+	rd->evq_timer_max = res->evq_timer_max;
+	rd->evq_int_min = res->evq_int_min;
+	rd->evq_int_max = res->evq_int_max;
+	rd->rxq_min = res->rxq_min;
+	rd->rxq_max = res->rxq_max;
+	rd->txq_min = res->txq_min;
+	rd->txq_max = res->txq_max;
+	EFRM_TRACE
+	    ("Using evq_int(%d-%d) evq_timer(%d-%d) RXQ(%d-%d) TXQ(%d-%d)",
+	     res->evq_int_min, res->evq_int_max, res->evq_timer_min,
+	     res->evq_timer_max, res->rxq_min, res->rxq_max, res->txq_min,
+	     res->txq_max);
+}
+
+#if defined(EFX_NOT_UPSTREAM)
+/* We have a module parameter that can tell us to only load the char driver
+ * for 1 NIC (if there are multiple NICs in the system), and if so which one.
+ * This tells us the PCI bus and slot of the NIC to load for, or -1 to just
+ * load on all NICs (the default).
+ * Value is a hex number in the format
+ *   bbbbss
+ * where:
+ *   bbbb - PCI bus number
+ *   ss   - PCI slot number
+ */
+unsigned int only_NIC = -1;
+
+/** @ingroup module_params */
+module_param(only_NIC, uint, 0444);
+MODULE_PARM_DESC(only_NIC,
+		 "Initialise sfc_resource driver for one NIC only, "
+		 "with specified PCI bus and slot");
+#endif
+
+static int
+efrm_dl_probe(struct efx_dl_device *efrm_dev,
+	      const struct net_device *net_dev,
+	      const struct efx_dl_device_info *dev_info,
+	      const char *silicon_rev)
+{
+	struct vi_resource_dimensions res_dim;
+	struct efx_dl_falcon_resources *res;
+	struct linux_efhw_nic *lnic;
+	struct pci_dev *dev;
+	struct efhw_nic *nic;
+	unsigned probe_flags = 0;
+	int rc;
+
+	efrm_dev->priv = NULL;
+
+	efx_dl_for_each_device_info_matching(dev_info, EFX_DL_FALCON_RESOURCES,
+					     struct efx_dl_falcon_resources,
+					     hdr, res) {
+		/* break out, leaving res pointing at the falcon resources */
+		break;
+	}
+
+	if (res == NULL) {
+		EFRM_ERR("%s: Unable to find falcon driverlink resources",
+			 __FUNCTION__);
+		return -EINVAL;
+	}
+
+	if (res->flags & EFX_DL_FALCON_USE_MSI)
+		probe_flags |= NIC_FLAG_TRY_MSI;
+
+	dev = efrm_dev->pci_dev;
+	if (res->flags & EFX_DL_FALCON_DUAL_FUNC) {
+		unsigned vendor = dev->vendor;
+		EFRM_ASSERT(dev->bus != NULL);
+		dev = NULL;
+
+#if defined(EFX_NOT_UPSTREAM)
+		if (only_NIC != -1 &&
+		    (efrm_dev->pci_dev->bus->number !=
+		     ((only_NIC >> 8) & 0xFFFF)
+		     || PCI_SLOT(efrm_dev->pci_dev->devfn) !=
+		     (only_NIC & 0xFF))) {
+			EFRM_NOTICE("Hiding char device %x:%x",
+				    efrm_dev->pci_dev->bus->number,
+				    PCI_SLOT(efrm_dev->pci_dev->devfn));
+			return -ENODEV;
+		}
+#endif
+
+		while ((dev = pci_get_device(vendor, FALCON_S_DEVID, dev))
+		       != NULL) {
+			EFRM_ASSERT(dev->bus != NULL);
+			/* With PCIe (since it's point to point)
+			 * the slot ID is usually 0 and
+			 * the bus ID changes NIC to NIC, so we really
+			 * need to check both. */
+			if (PCI_SLOT(dev->devfn) ==
+			    PCI_SLOT(efrm_dev->pci_dev->devfn)
+			    && dev->bus->number ==
+			    efrm_dev->pci_dev->bus->number)
+				break;
+		}
+		if (dev == NULL) {
+			EFRM_ERR("%s: Unable to find falcon secondary "
+				 "PCI device.", __FUNCTION__);
+			return -ENODEV;
+		}
+		pci_dev_put(dev);
+	}
+
+	init_vi_resource_dimensions(&res_dim, res);
+
+	rc = efrm_nic_add(dev, probe_flags, net_dev->dev_addr, &lnic,
+			  res->biu_lock,
+			  res->buffer_table_min, res->buffer_table_max,
+			  &res_dim);
+	if (rc != 0)
+		return rc;
+
+	nic = &lnic->nic;
+	nic->mtu = net_dev->mtu + ETH_HLEN;
+	nic->net_driver_dev = efrm_dev;
+	nic->ifindex = net_dev->ifindex;
+#ifdef HAS_NET_NAMESPACE
+	nic->nd_net = net_dev->nd_net;
+#endif
+	efrm_dev->priv = nic;
+
+	/* Register a callback so we're told when MTU changes.
+	 * We dynamically allocate efx_dl_callbacks, because
+	 * the callbacks that we want depends on the NIC type.
+	 */
+	lnic->dl_callbacks =
+	    kmalloc(sizeof(struct efx_dl_callbacks), GFP_KERNEL);
+	if (!lnic->dl_callbacks) {
+		EFRM_ERR("Out of memory (%s)", __FUNCTION__);
+		efrm_nic_del(lnic);
+		return -ENOMEM;
+	}
+	memset(lnic->dl_callbacks, 0, sizeof(*lnic->dl_callbacks));
+	lnic->dl_callbacks->mtu_changed = efrm_dl_mtu_changed;
+
+	if ((res->flags & EFX_DL_FALCON_DUAL_FUNC) == 0) {
+		/* Net driver receives all management events.
+		 * Register a callback to receive the ones
+		 * we're interested in. */
+		lnic->dl_callbacks->event = efrm_dl_event_falcon;
+	}
+
+	rc = efx_dl_register_callbacks(efrm_dev, lnic->dl_callbacks);
+	if (rc < 0) {
+		EFRM_ERR("%s: efx_dl_register_callbacks failed (%d)",
+			 __FUNCTION__, rc);
+		kfree(lnic->dl_callbacks);
+		efrm_nic_del(lnic);
+		return rc;
+	}
+
+	return 0;
+}
+
+/* When we unregister ourselves on module removal, this function will be
+ * called for all the devices we claimed */
+static void efrm_dl_remove(struct efx_dl_device *efrm_dev)
+{
+	struct efhw_nic *nic = efrm_dev->priv;
+	struct linux_efhw_nic *lnic = linux_efhw_nic(nic);
+	EFRM_TRACE("%s called", __FUNCTION__);
+	if (lnic->dl_callbacks) {
+		efx_dl_unregister_callbacks(efrm_dev, lnic->dl_callbacks);
+		kfree(lnic->dl_callbacks);
+	}
+	if (efrm_dev->priv)
+		efrm_nic_del(lnic);
+	EFRM_TRACE("%s OK", __FUNCTION__);
+}
+
+static void efrm_dl_reset_suspend(struct efx_dl_device *efrm_dev)
+{
+	EFRM_NOTICE("%s:", __FUNCTION__);
+}
+
+static void efrm_dl_reset_resume(struct efx_dl_device *efrm_dev, int ok)
+{
+	EFRM_NOTICE("%s: ok=%d", __FUNCTION__, ok);
+}
+
+int efrm_driverlink_register(void)
+{
+	EFRM_TRACE("%s:", __FUNCTION__);
+	return efx_dl_register_driver(&efrm_dl_driver);
+}
+
+void efrm_driverlink_unregister(void)
+{
+	EFRM_TRACE("%s:", __FUNCTION__);
+	efx_dl_unregister_driver(&efrm_dl_driver);
+}
+
+static void efrm_dl_mtu_changed(struct efx_dl_device *efx_dev, int mtu)
+{
+	struct efhw_nic *nic = efx_dev->priv;
+
+	ASSERT_RTNL();	/* Since we're looking at efx_dl_device::port_net_dev */
+
+	EFRM_TRACE("%s: old=%d new=%d", __FUNCTION__, nic->mtu, mtu + ETH_HLEN);
+	/* If this happened we must have agreed to it above */
+	nic->mtu = mtu + ETH_HLEN;
+}
+
+static void efrm_dl_event_falcon(struct efx_dl_device *efx_dev, void *p_event)
+{
+	struct efhw_nic *nic = efx_dev->priv;
+	struct linux_efhw_nic *lnic = linux_efhw_nic(nic);
+	efhw_event_t *ev = p_event;
+
+	switch (FALCON_EVENT_CODE(ev)) {
+	case FALCON_EVENT_CODE_CHAR:
+		falcon_handle_char_event(nic, lnic->ev_handlers, ev);
+		break;
+	default:
+		EFRM_WARN("%s: unknown event type=%x", __FUNCTION__,
+			  (unsigned)FALCON_EVENT_CODE(ev));
+		break;
+	}
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/driver_object.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/driver_object.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/driver_object.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/driver_object.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,174 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains support for the global driver variables.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/efrm/nic_table.h>
+#include <ci/efrm/resource.h>
+#include <ci/efrm/debug.h>
+
+/* We use #define rather than static inline here so that the Windows
+ * "prefast" compiler can see its own locking primitive when these
+ * two function are used (and then perform extra checking where they
+ * are used)
+ *
+ * Both macros operate on an irq_flags_t
+*/
+
+#define efrm_driver_lock(irqlock_state) \
+	spin_lock_irqsave(&efrm_nic_table.lock, irqlock_state)
+
+#define efrm_driver_unlock(irqlock_state)		\
+	spin_unlock_irqrestore(&efrm_nic_table.lock,	\
+			       irqlock_state);
+
+/* These routines are all methods on the architecturally singleton
+   global variables: efrm_nic_table, efrm_rm_table.
+
+   I hope we never find a driver model that does not allow global
+   structure variables :) (but that would break almost every driver I've
+   ever seen).
+*/
+
+/*! Exported driver state */
+struct efrm_nic_table efrm_nic_table;
+EXPORT_SYMBOL(efrm_nic_table);
+
+/* Internal table with resource managers.
+ * We'd like to not export it, but we are still using efrm_rm_table
+ * in the char driver. So, it is declared in the private header with
+ * a purpose. */
+struct efrm_resource_manager *efrm_rm_table[EFRM_RESOURCE_NUM];
+EXPORT_SYMBOL(efrm_rm_table);
+
+int efrm_driver_ctor(void)
+{
+	memset(&efrm_nic_table, 0, sizeof(efrm_nic_table));
+	memset(&efrm_rm_table, 0, sizeof(efrm_rm_table));
+
+	spin_lock_init(&efrm_nic_table.lock);
+
+	EFRM_TRACE("%s: driver created", __FUNCTION__);
+	return 0;
+}
+
+int efrm_driver_dtor(void)
+{
+	EFRM_ASSERT(!efrm_nic_table_held());
+
+	spin_lock_destroy(&efrm_nic_table.lock);
+	EFRM_TRACE("%s: driver deleted", __FUNCTION__);
+	return 0;
+}
+
+int efrm_driver_register_nic(struct efhw_nic *nic, int nic_index)
+{
+	int rc = 0;
+	irq_flags_t lock_flags;
+
+	EFRM_ASSERT(nic_index >= 0);
+
+	efrm_driver_lock(lock_flags);
+
+	if (efrm_nic_table_held()) {
+		EFRM_WARN("%s: driver object is in use", __FUNCTION__);
+		rc = -EBUSY;
+		goto done;
+	}
+
+	if (efrm_nic_table.nic_count == EFHW_MAX_NR_DEVS) {
+		EFRM_WARN("%s: filled up NIC table size %d", __FUNCTION__,
+			  EFHW_MAX_NR_DEVS);
+		rc = -E2BIG;
+		goto done;
+	}
+
+	EFRM_ASSERT(efrm_nic_table.nic[nic_index] == NULL);
+	efrm_nic_table.nic[nic_index] = nic;
+	nic->index = nic_index;
+
+	if (efrm_nic_table.a_nic == NULL)
+		efrm_nic_table.a_nic = nic;
+
+	efrm_nic_table.nic_count++;
+	efrm_driver_unlock(lock_flags);
+	return rc;
+
+done:
+	efrm_driver_unlock(lock_flags);
+	return rc;
+}
+
+int efrm_driver_unregister_nic(struct efhw_nic *nic)
+{
+	int rc = 0;
+	int nic_index = nic->index;
+	irq_flags_t lock_flags;
+
+	EFRM_ASSERT(nic_index >= 0);
+
+	efrm_driver_lock(lock_flags);
+
+	if (efrm_nic_table_held()) {
+		EFRM_WARN("%s: driver object is in use", __FUNCTION__);
+		rc = -EBUSY;
+		goto done;
+	}
+
+	EFRM_ASSERT(efrm_nic_table.nic[nic_index] == nic);
+
+	nic->index = -1;
+	efrm_nic_table.nic[nic_index] = NULL;
+
+	--efrm_nic_table.nic_count;
+
+	if (efrm_nic_table.a_nic == nic) {
+		if (efrm_nic_table.nic_count == 0) {
+			efrm_nic_table.a_nic = NULL;
+		} else {
+			for (nic_index = 0; nic_index < EFHW_MAX_NR_DEVS;
+			     nic_index++) {
+				if (efrm_nic_table.nic[nic_index] != NULL)
+					efrm_nic_table.a_nic =
+					    efrm_nic_table.nic[nic_index];
+			}
+			EFRM_ASSERT(efrm_nic_table.a_nic);
+		}
+	}
+
+done:
+	efrm_driver_unlock(lock_flags);
+	return rc;
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/efx_vi_shm.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/efx_vi_shm.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/efx_vi_shm.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/efx_vi_shm.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,701 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides implementation of EFX VI API, used from Xen
+ * acceleration driver.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include "linux_resource_internal.h"
+#include <ci/efrm/nic_table.h>
+#include <ci/efrm/vi_resource_manager.h>
+#include <ci/driver/resource/efx_vi.h>
+#include <ci/efrm/filter.h>
+#include <ci/efrm/buffer_table.h>
+#include <linux/pci.h>
+#include "kernel_compat.h"
+
+#if EFX_VI_STATIC_FILTERS
+struct filter_list_t {
+	struct filter_list_t *next;
+	struct filter_resource *fres;
+};
+#endif
+
+struct efx_vi_state {
+	struct vi_resource *vi_res;
+
+	int nic_index;
+
+	void (*callback_fn)(void *arg, int is_timeout);
+	void *callback_arg;
+
+	struct completion flush_completion;
+
+#if EFX_VI_STATIC_FILTERS
+	struct filter_list_t fres[EFX_VI_STATIC_FILTERS];
+	struct filter_list_t *free_fres;
+	struct filter_list_t *used_fres;
+#endif
+};
+
+static void efx_vi_flush_complete(void *state_void)
+{
+	struct efx_vi_state *state = (struct efx_vi_state *)state_void;
+
+	complete(&state->flush_completion);
+}
+
+static inline int alloc_ep(struct efx_vi_state *state)
+{
+	int rc;
+
+	rc = efrm_vi_resource_alloc(NULL, EFHW_VI_JUMBO_EN,
+				    efx_vi_eventq_size,
+				    FALCON_DMA_Q_DEFAULT_TX_SIZE,
+				    FALCON_DMA_Q_DEFAULT_RX_SIZE,
+				    0, 0, &state->vi_res, NULL, NULL, NULL,
+				    NULL);
+	if (rc < 0) {
+		EFRM_ERR("%s: ERROR efrm_vi_resource_alloc error %d",
+			 __FUNCTION__, rc);
+		return rc;
+	}
+
+	efrm_vi_register_flush_callback(state->vi_res, &efx_vi_flush_complete,
+					(void *)state);
+
+	return 0;
+}
+
+static int free_ep(struct efx_vi_state *efx_state)
+{
+	efrm_vi_resource_release(efx_state->vi_res);
+
+	return 0;
+}
+
+#if EFX_VI_STATIC_FILTERS
+static int efx_vi_alloc_static_filters(struct efx_vi_state *efx_state)
+{
+	int i;
+	int rc;
+
+	efx_state->free_fres = efx_state->used_fres = NULL;
+
+	for (i = 0; i < EFX_VI_STATIC_FILTERS; i++) {
+		rc = efrm_filter_resource_alloc(efx_state->vi_res,
+						&efx_state->fres[i].fres);
+		if (rc < 0) {
+			EFRM_ERR("%s: efrm_filter_resource_alloc failed: %d",
+			     __FUNCTION__, rc);
+			while (i > 0) {
+				i--;
+				efrm_filter_resource_release(efx_state->
+							     fres[i].fres);
+			}
+			efx_state->free_fres = NULL;
+			return rc;
+		}
+		efx_state->fres[i].next = efx_state->free_fres;
+		efx_state->free_fres = &efx_state->fres[i];
+	}
+
+	return 0;
+}
+#endif
+
+int efx_vi_alloc(struct efx_vi_state **vih_out, int nic_index)
+{
+	struct efx_vi_state *efx_state;
+	int rc;
+
+	BUG_ON(nic_index < 0 || nic_index >= EFHW_MAX_NR_DEVS);
+
+	efx_state = kmalloc(sizeof(struct efx_vi_state), GFP_KERNEL);
+
+	if (!efx_state) {
+		EFRM_ERR("%s: failed to allocate memory for efx_vi_state",
+			 __FUNCTION__);
+		rc = -ENOMEM;
+		goto fail;
+	}
+
+	efx_state->nic_index = nic_index;
+	init_completion(&efx_state->flush_completion);
+
+	/* basically allocate_pt_endpoint() */
+	rc = alloc_ep(efx_state);
+	if (rc) {
+		EFRM_ERR("%s: alloc_ep failed: %d", __FUNCTION__, rc);
+		goto fail_no_pt;
+	}
+#if EFX_VI_STATIC_FILTERS
+	/* Statically allocate a set of filter resources - removes the
+	   restriction on not being able to use efx_vi_filter() from
+	   in_atomic() */
+	rc = efx_vi_alloc_static_filters(efx_state);
+	if (rc)
+		goto fail_no_filters;
+#endif
+
+	*vih_out = efx_state;
+
+	return 0;
+#if EFX_VI_STATIC_FILTERS
+fail_no_filters:
+	free_ep(efx_state);
+#endif
+fail_no_pt:
+	kfree(efx_state);
+fail:
+	return rc;
+}
+EXPORT_SYMBOL(efx_vi_alloc);
+
+void efx_vi_free(struct efx_vi_state *vih)
+{
+	struct efx_vi_state *efx_state = vih;
+
+	/* TODO flush dma channels, init dma queues?.  See ef_free_vnic() */
+#if EFX_VI_STATIC_FILTERS
+	int i;
+
+	for (i = 0; i < EFX_VI_STATIC_FILTERS; i++)
+		efrm_filter_resource_release(efx_state->fres[i].fres);
+#endif
+
+	if (efx_state->vi_res)
+		free_ep(efx_state);
+
+	kfree(efx_state);
+}
+EXPORT_SYMBOL(efx_vi_free);
+
+void efx_vi_reset(struct efx_vi_state *vih)
+{
+	struct efx_vi_state *efx_state = vih;
+
+	efrm_pt_flush(efx_state->vi_res);
+
+	while (wait_for_completion_timeout(&efx_state->flush_completion, HZ)
+	       == 0)
+		efrm_vi_resource_flush_retry(efx_state->vi_res);
+
+	/* Bosch the eventq */
+	efrm_eventq_reset(efx_state->vi_res, 0);
+	return;
+}
+EXPORT_SYMBOL(efx_vi_reset);
+
+static void
+efx_vi_eventq_callback(void *context, int is_timeout, struct efhw_nic *nic)
+{
+	struct efx_vi_state *efx_state = (struct efx_vi_state *)context;
+
+	EFRM_ASSERT(efx_state->callback_fn);
+
+	return efx_state->callback_fn(efx_state->callback_arg, is_timeout);
+}
+
+int
+efx_vi_eventq_register_callback(struct efx_vi_state *vih,
+			void (*callback)(void *context, int is_timeout),
+			void *context)
+{
+	struct efx_vi_state *efx_state = vih;
+
+	efx_state->callback_fn = callback;
+	efx_state->callback_arg = context;
+
+	/* Register the eventq timeout event callback */
+	efrm_eventq_register_callback(efx_state->vi_res,
+				      efx_vi_eventq_callback, efx_state);
+
+	return 0;
+}
+EXPORT_SYMBOL(efx_vi_eventq_register_callback);
+
+int efx_vi_eventq_kill_callback(struct efx_vi_state *vih)
+{
+	struct efx_vi_state *efx_state = vih;
+
+	if (efx_state->vi_res->evq_callback_fn)
+		efrm_eventq_kill_callback(efx_state->vi_res);
+
+	efx_state->callback_fn = NULL;
+	efx_state->callback_arg = NULL;
+
+	return 0;
+}
+EXPORT_SYMBOL(efx_vi_eventq_kill_callback);
+
+struct efx_vi_dma_map_state {
+	struct efhw_buffer_table_allocation bt_handle;
+	int n_pages;
+	dma_addr_t *dma_addrs;
+};
+
+int
+efx_vi_dma_map_pages(struct efx_vi_state *vih, struct page **pages,
+		     int n_pages, struct efx_vi_dma_map_state **dmh_out)
+{
+	struct efx_vi_state *efx_state = vih;
+	int order = fls(n_pages - 1), rc, i, evq_id;
+	dma_addr_t dma_addr;
+	struct efx_vi_dma_map_state *dm_state;
+
+	if (n_pages != (1 << order)) {
+		EFRM_WARN("%s: Can only allocate buffers in power of 2 "
+			  "sizes (not %d)", __FUNCTION__, n_pages);
+		return -EINVAL;
+	}
+
+	dm_state = kmalloc(sizeof(struct efx_vi_dma_map_state), GFP_KERNEL);
+	if (!dm_state)
+		return -ENOMEM;
+
+	dm_state->dma_addrs = kmalloc(sizeof(dma_addr_t) * n_pages,
+				      GFP_KERNEL);
+	if (!dm_state->dma_addrs) {
+		kfree(dm_state);
+		return -ENOMEM;
+	}
+
+	rc = efrm_buffer_table_alloc(order, &dm_state->bt_handle);
+	if (rc < 0) {
+		kfree(dm_state->dma_addrs);
+		kfree(dm_state);
+		return rc;
+	}
+
+	evq_id = EFRM_RESOURCE_INSTANCE(efx_state->vi_res->rs.rs_handle);
+	for (i = 0; i < n_pages; i++) {
+		/* TODO do we need to get_page() here ? */
+
+		dma_addr = pci_map_page
+		    (linux_efhw_nic(efrm_nic_table.nic[efx_state->nic_index])->
+		     pci_dev, pages[i], 0, PAGE_SIZE, PCI_DMA_TODEVICE);
+
+		efrm_buffer_table_set(&dm_state->bt_handle, i, dma_addr,
+				      evq_id);
+
+		dm_state->dma_addrs[i] = dma_addr;
+
+		/* Would be nice to not have to call commit each time, but
+		 * comment says there are hardware restrictions on how often
+		 * you can go without it, so do this to be safe */
+		efrm_buffer_table_commit();
+	}
+
+	dm_state->n_pages = n_pages;
+
+	*dmh_out = dm_state;
+
+	return 0;
+}
+EXPORT_SYMBOL(efx_vi_dma_map_pages);
+
+/* Function needed as Xen can't get pages for grants in dom0, but can
+   get dma address */
+int
+efx_vi_dma_map_addrs(struct efx_vi_state *vih,
+		     unsigned long long *bus_dev_addrs,
+		     int n_pages, struct efx_vi_dma_map_state **dmh_out)
+{
+	struct efx_vi_state *efx_state = vih;
+	int order = fls(n_pages - 1), rc, i, evq_id;
+	dma_addr_t dma_addr;
+	struct efx_vi_dma_map_state *dm_state;
+
+	if (n_pages != (1 << order)) {
+		EFRM_WARN("%s: Can only allocate buffers in power of 2 "
+			  "sizes (not %d)", __FUNCTION__, n_pages);
+		return -EINVAL;
+	}
+
+	dm_state = kmalloc(sizeof(struct efx_vi_dma_map_state), GFP_KERNEL);
+	if (!dm_state)
+		return -ENOMEM;
+
+	dm_state->dma_addrs = kmalloc(sizeof(dma_addr_t) * n_pages,
+				      GFP_KERNEL);
+	if (!dm_state->dma_addrs) {
+		kfree(dm_state);
+		return -ENOMEM;
+	}
+
+	rc = efrm_buffer_table_alloc(order, &dm_state->bt_handle);
+	if (rc < 0) {
+		kfree(dm_state->dma_addrs);
+		kfree(dm_state);
+		return rc;
+	}
+
+	evq_id = EFRM_RESOURCE_INSTANCE(efx_state->vi_res->rs.rs_handle);
+#if 0
+	EFRM_WARN("%s: mapping %d pages to evq %d, bt_ids %d-%d\n",
+		  __FUNCTION__, n_pages, evq_id,
+		  dm_state->bt_handle.base,
+		  dm_state->bt_handle.base + n_pages);
+#endif
+	for (i = 0; i < n_pages; i++) {
+
+		dma_addr = (dma_addr_t)bus_dev_addrs[i];
+
+		efrm_buffer_table_set(&dm_state->bt_handle, i, dma_addr,
+				      evq_id);
+
+		dm_state->dma_addrs[i] = dma_addr;
+
+		/* Would be nice to not have to call commit each time, but
+		 * comment says there are hardware restrictions on how often
+		 * you can go without it, so do this to be safe */
+		efrm_buffer_table_commit();
+	}
+
+	dm_state->n_pages = n_pages;
+
+	*dmh_out = dm_state;
+
+	return 0;
+}
+EXPORT_SYMBOL(efx_vi_dma_map_addrs);
+
+void
+efx_vi_dma_unmap_pages(struct efx_vi_state *vih,
+		       struct efx_vi_dma_map_state *dmh)
+{
+	struct efx_vi_state *efx_state = vih;
+	struct efx_vi_dma_map_state *dm_state =
+	    (struct efx_vi_dma_map_state *)dmh;
+	int i;
+
+	efrm_buffer_table_free(&dm_state->bt_handle);
+
+	for (i = 0; i < dm_state->n_pages; ++i)
+		pci_unmap_page(linux_efhw_nic
+			(efrm_nic_table.nic[efx_state->nic_index])->pci_dev,
+			dm_state->dma_addrs[i], PAGE_SIZE, PCI_DMA_TODEVICE);
+
+	kfree(dm_state->dma_addrs);
+	kfree(dm_state);
+
+	return;
+}
+EXPORT_SYMBOL(efx_vi_dma_unmap_pages);
+
+void
+efx_vi_dma_unmap_addrs(struct efx_vi_state *vih,
+		       struct efx_vi_dma_map_state *dmh)
+{
+	struct efx_vi_dma_map_state *dm_state =
+	    (struct efx_vi_dma_map_state *)dmh;
+
+	efrm_buffer_table_free(&dm_state->bt_handle);
+
+	kfree(dm_state->dma_addrs);
+	kfree(dm_state);
+
+	return;
+}
+EXPORT_SYMBOL(efx_vi_dma_unmap_addrs);
+
+unsigned
+efx_vi_dma_get_map_addr(struct efx_vi_state *vih,
+			struct efx_vi_dma_map_state *dmh)
+{
+	struct efx_vi_dma_map_state *dm_state =
+	    (struct efx_vi_dma_map_state *)dmh;
+
+	return EFHW_BUFFER_ADDR(dm_state->bt_handle.base, 0);
+}
+EXPORT_SYMBOL(efx_vi_dma_get_map_addr);
+
+#if EFX_VI_STATIC_FILTERS
+static int
+get_filter(struct efx_vi_state *efx_state,
+	   efrm_resource_handle_t pthandle, struct filter_resource **fres_out)
+{
+	struct filter_list_t *flist;
+	if (efx_state->free_fres == NULL)
+		return -ENOMEM;
+	else {
+		flist = efx_state->free_fres;
+		efx_state->free_fres = flist->next;
+		flist->next = efx_state->used_fres;
+		efx_state->used_fres = flist;
+		*fres_out = flist->fres;
+		return 0;
+	}
+}
+#endif
+
+static void
+release_filter(struct efx_vi_state *efx_state, struct filter_resource *fres)
+{
+#if EFX_VI_STATIC_FILTERS
+	struct filter_list_t *flist = efx_state->used_fres, *prev = NULL;
+	while (flist) {
+		if (flist->fres == fres) {
+			if (prev)
+				prev->next = flist->next;
+			else
+				efx_state->used_fres = flist->next;
+			flist->next = efx_state->free_fres;
+			efx_state->free_fres = flist;
+			return;
+		}
+		prev = flist;
+		flist = flist->next;
+	}
+	EFRM_ERR("%s: couldn't find filter", __FUNCTION__);
+#else
+	return efrm_filter_resource_release(fres);
+#endif
+}
+
+int
+efx_vi_filter(struct efx_vi_state *vih, int protocol,
+	      unsigned ip_addr_be32, int port_le16,
+	      struct filter_resource_t **fh_out)
+{
+	struct efx_vi_state *efx_state = vih;
+	struct filter_resource *frs;
+	int rc;
+
+#if EFX_VI_STATIC_FILTERS
+	rc = get_filter(efx_state, efx_state->vi_res->rs.rs_handle, &frs);
+#else
+	rc = efrm_filter_resource_alloc(efx_state->vi_res, &frs);
+#endif
+	if (rc < 0)
+		return rc;
+
+	/* Add the hardware filter. We pass in the source port and address
+	 * as 0 (wildcard) to minimise the number of filters needed. */
+	if (protocol == IPPROTO_TCP) {
+		rc = efrm_filter_resource_tcp_set(frs, 0, 0, ip_addr_be32,
+						  port_le16);
+	} else {
+		rc = efrm_filter_resource_udp_set(frs, 0, 0, ip_addr_be32,
+						  port_le16);
+	}
+
+	*fh_out = (struct filter_resource_t *)frs;
+
+	return rc;
+}
+EXPORT_SYMBOL(efx_vi_filter);
+
+int
+efx_vi_filter_stop(struct efx_vi_state *vih, struct filter_resource_t *fh)
+{
+	struct efx_vi_state *efx_state = vih;
+	struct filter_resource *frs = (struct filter_resource *)fh;
+	int rc;
+
+	rc = efrm_filter_resource_clear(frs);
+	release_filter(efx_state, frs);
+
+	return rc;
+}
+EXPORT_SYMBOL(efx_vi_filter_stop);
+
+int
+efx_vi_hw_resource_get_virt(struct efx_vi_state *vih,
+			    struct efx_vi_hw_resource_metadata *mdata,
+			    struct efx_vi_hw_resource *hw_res_array,
+			    int *length)
+{
+	EFRM_NOTICE("%s: TODO!", __FUNCTION__);
+
+	return 0;
+}
+EXPORT_SYMBOL(efx_vi_hw_resource_get_virt);
+
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+int
+efx_vi_hw_resource_get_phys(struct efx_vi_state *vih,
+			    struct efx_vi_hw_resource_metadata *mdata,
+			    struct efx_vi_hw_resource *hw_res_array,
+			    int *length)
+{
+	struct efx_vi_state *efx_state = vih;
+	int i, ni = efx_state->nic_index;
+	struct linux_efhw_nic *lnic = linux_efhw_nic(efrm_nic_table.nic[ni]);
+	unsigned long phys = lnic->ctr_ap_pci_addr;
+	struct efrm_resource *ep_res = &efx_state->vi_res->rs;
+	unsigned ep_mmap_bytes;
+
+	if (*length < EFX_VI_HW_RESOURCE_MAXSIZE)
+		return -EINVAL;
+
+	mdata->version = 0;
+
+	mdata->nic_arch = efrm_nic_table.nic[ni]->devtype.arch;
+	mdata->nic_variant = efrm_nic_table.nic[ni]->devtype.variant;
+	mdata->nic_revision = efrm_nic_table.nic[ni]->devtype.revision;
+
+	mdata->evq_order =
+	    efx_state->vi_res->nic_info[ni].evq_pages.iobuff.order;
+	mdata->evq_offs = efx_state->vi_res->nic_info[ni].evq_pages.iobuff_off;
+	mdata->evq_capacity = efx_vi_eventq_size;
+	mdata->instance = EFRM_RESOURCE_INSTANCE(ep_res->rs_handle);
+	mdata->rx_capacity = FALCON_DMA_Q_DEFAULT_RX_SIZE;
+	mdata->tx_capacity = FALCON_DMA_Q_DEFAULT_TX_SIZE;
+
+	ep_mmap_bytes = FALCON_DMA_Q_DEFAULT_MMAP;
+	EFRM_ASSERT(ep_mmap_bytes == PAGE_SIZE * 2);
+
+#ifndef NDEBUG
+	{
+		/* Sanity about doorbells */
+		unsigned long tx_dma_page_addr, rx_dma_page_addr;
+
+		/* get rx doorbell address */
+		rx_dma_page_addr =
+		    phys + falcon_rx_dma_page_addr(mdata->instance);
+		/* get tx doorbell address */
+		tx_dma_page_addr =
+		    phys + falcon_tx_dma_page_addr(mdata->instance);
+
+		/* Check the lower bits of the TX doorbell will be
+		 * consistent. */
+		EFRM_ASSERT((TX_DESC_UPD_REG_PAGE4_OFST &
+			     FALCON_DMA_PAGE_MASK) ==
+			    (TX_DESC_UPD_REG_PAGE123K_OFST &
+			     FALCON_DMA_PAGE_MASK));
+
+		/* Check the lower bits of the RX doorbell will be
+		 * consistent. */
+		EFRM_ASSERT((RX_DESC_UPD_REG_PAGE4_OFST &
+			     FALCON_DMA_PAGE_MASK) ==
+			    (RX_DESC_UPD_REG_PAGE123K_OFST &
+			     FALCON_DMA_PAGE_MASK));
+
+		/* Check that the doorbells will be in the same page. */
+		EFRM_ASSERT((TX_DESC_UPD_REG_PAGE4_OFST & PAGE_MASK) ==
+			    (RX_DESC_UPD_REG_PAGE4_OFST & PAGE_MASK));
+
+		/* Check that the doorbells are in the same page. */
+		EFRM_ASSERT((tx_dma_page_addr & PAGE_MASK) ==
+			    (rx_dma_page_addr & PAGE_MASK));
+
+		/* Check that the TX doorbell offset is correct. */
+		EFRM_ASSERT((TX_DESC_UPD_REG_PAGE4_OFST & ~PAGE_MASK) ==
+			    (tx_dma_page_addr & ~PAGE_MASK));
+
+		/* Check that the RX doorbell offset is correct. */
+		EFRM_ASSERT((RX_DESC_UPD_REG_PAGE4_OFST & ~PAGE_MASK) ==
+			    (rx_dma_page_addr & ~PAGE_MASK));
+	}
+#endif
+
+	i = 0;
+	hw_res_array[i].type = EFX_VI_HW_RESOURCE_TXDMAQ;
+	hw_res_array[i].mem_type = EFX_VI_HW_RESOURCE_PERIPHERAL;
+	hw_res_array[i].more_to_follow = 0;
+	hw_res_array[i].length = PAGE_SIZE;
+	hw_res_array[i].address =
+		(unsigned long)efx_state->vi_res->nic_info[ni].
+			dmaq_pages[EFRM_VI_RM_DMA_QUEUE_TX].kva;
+
+	i++;
+	hw_res_array[i].type = EFX_VI_HW_RESOURCE_RXDMAQ;
+	hw_res_array[i].mem_type = EFX_VI_HW_RESOURCE_PERIPHERAL;
+	hw_res_array[i].more_to_follow = 0;
+	hw_res_array[i].length = PAGE_SIZE;
+	hw_res_array[i].address =
+		(unsigned long)efx_state->vi_res->nic_info[ni].
+			dmaq_pages[EFRM_VI_RM_DMA_QUEUE_RX].kva;
+
+	/* NB EFX_VI_HW_RESOURCE_TXBELL not used on Falcon */
+	/* NB EFX_VI_HW_RESOURCE_RXBELL not used on Falcon */
+
+	i++;
+	hw_res_array[i].type = EFX_VI_HW_RESOURCE_EVQTIMER;
+	hw_res_array[i].mem_type = EFX_VI_HW_RESOURCE_PERIPHERAL;
+	hw_res_array[i].more_to_follow = 0;
+	hw_res_array[i].length = PAGE_SIZE;
+	hw_res_array[i].address =
+		(unsigned long)phys + falcon_timer_page_addr(mdata->instance);
+
+	/* NB EFX_VI_HW_RESOURCE_EVQPTR not used on Falcon */
+
+	i++;
+	switch (efrm_nic_table.nic[ni]->devtype.variant) {
+	case 'A':
+		hw_res_array[i].type = EFX_VI_HW_RESOURCE_EVQRPTR;
+		hw_res_array[i].mem_type = EFX_VI_HW_RESOURCE_PERIPHERAL;
+		hw_res_array[i].more_to_follow = 0;
+		hw_res_array[i].length = PAGE_SIZE;
+		hw_res_array[i].address = (unsigned long)phys +
+			EVQ_RPTR_REG_OFST +
+			(FALCON_REGISTER128 * mdata->instance);
+		break;
+	case 'B':
+		hw_res_array[i].type = EFX_VI_HW_RESOURCE_EVQRPTR_OFFSET;
+		hw_res_array[i].mem_type = EFX_VI_HW_RESOURCE_PERIPHERAL;
+		hw_res_array[i].more_to_follow = 0;
+		hw_res_array[i].length = PAGE_SIZE;
+		hw_res_array[i].address =
+			(unsigned long)FALCON_EVQ_RPTR_REG_P0;
+		break;
+	default:
+		EFRM_ASSERT(0);
+		break;
+	}
+
+	i++;
+	hw_res_array[i].type = EFX_VI_HW_RESOURCE_EVQMEMKVA;
+	hw_res_array[i].mem_type = EFX_VI_HW_RESOURCE_IOBUFFER;
+	hw_res_array[i].more_to_follow = 0;
+	hw_res_array[i].length = PAGE_SIZE;
+	hw_res_array[i].address = (unsigned long)efx_state->vi_res->
+		nic_info[ni].evq_pages.iobuff.kva;
+
+	i++;
+	hw_res_array[i].type = EFX_VI_HW_RESOURCE_BELLPAGE;
+	hw_res_array[i].mem_type = EFX_VI_HW_RESOURCE_PERIPHERAL;
+	hw_res_array[i].more_to_follow = 0;
+	hw_res_array[i].length = PAGE_SIZE;
+	hw_res_array[i].address =
+		(unsigned long)(phys +
+				falcon_tx_dma_page_addr(mdata->instance))
+		>> PAGE_SHIFT;
+
+	i++;
+
+	EFRM_ASSERT(i <= *length);
+
+	*length = i;
+
+	return 0;
+}
+EXPORT_SYMBOL(efx_vi_hw_resource_get_phys);
+#endif
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/eventq.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/eventq.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/eventq.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/eventq.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,320 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains event queue support.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/efhw/debug.h>
+#include <ci/efhw/iopage.h>
+#include <ci/driver/efab/hardware.h>
+#include <ci/efhw/eventq.h>
+#include <ci/efhw/falcon.h>
+#include <ci/efhw/nic.h>
+
+#define KEVENTQ_MAGIC 0x07111974
+
+/*! Helper function to allocate the iobuffer needed by an eventq
+ *   - it ensures the eventq has the correct alignment for the NIC
+ *
+ * \param rm        Event-queue resource manager
+ * \param instance  Event-queue instance (index)
+ * \param buf_bytes Requested size of eventq
+ * \return          < 0 if iobuffer allocation fails
+ */
+int
+efhw_nic_event_queue_alloc_iobuffer(struct efhw_nic *nic,
+				    struct eventq_resource_hardware *h,
+				    int evq_instance, unsigned buf_bytes)
+{
+	unsigned int page_order;
+	int rc;
+
+	/* Allocate an iobuffer. */
+	page_order = get_order(buf_bytes);
+
+	h->iobuff_off = 0;
+
+	EFHW_TRACE("allocating eventq size %x",
+		   1u << (page_order + PAGE_SHIFT));
+	rc = efhw_iopages_alloc(nic, &h->iobuff, page_order);
+	if (rc < 0) {
+		EFHW_WARN("%s: failed to allocate %u pages",
+			  __FUNCTION__, 1u << page_order);
+		return rc;
+	}
+
+	/* Set the eventq pages to match EFHW_CLEAR_EVENT() */
+	if (EFHW_CLEAR_EVENT_VALUE)
+		memset(efhw_iopages_ptr(&h->iobuff) + h->iobuff_off,
+		       EFHW_CLEAR_EVENT_VALUE, (1u << page_order) * PAGE_SIZE);
+
+	EFHW_TRACE("%s: allocated %u pages", __FUNCTION__, 1u << (page_order));
+
+	/* For Falcon the NIC is programmed with the base buffer address of a
+	 * contiguous region of buffer space. This means that larger than a
+	 * PAGE event queues can be expected to allocate even when the host's
+	 * physical memory is fragmented */
+	EFHW_ASSERT(efhw_nic_have_hw(nic));
+	EFHW_ASSERT(page_order <= h->buf_tbl_alloc.order);
+
+	/* Initialise the buffer table entries. */
+	falcon_nic_buffer_table_set_n(nic, h->buf_tbl_alloc.base,
+				      efhw_iopages_dma_addr(&h->iobuff) +
+				      h->iobuff_off, EFHW_NIC_PAGE_SIZE, 0,
+				      1 << page_order, 0);
+
+	if (evq_instance >= FALCON_EVQ_TBL_RESERVED)
+		falcon_nic_buffer_table_confirm(nic);
+	return 0;
+}
+
+/**********************************************************************
+ * Kernel event queue management.
+ */
+
+/* Values for [struct efhw_keventq::lock] field. */
+#define KEVQ_UNLOCKED      0
+#define KEVQ_LOCKED        1
+#define KEVQ_RECHECK       2
+
+int
+efhw_keventq_ctor(struct efhw_nic *nic, int instance,
+		  struct efhw_keventq *evq,
+		  struct efhw_ev_handler *ev_handlers)
+{
+	int rc;
+	unsigned buf_bytes = evq->hw.capacity * sizeof(efhw_event_t);
+
+	evq->instance = instance;
+	evq->ev_handlers = ev_handlers;
+
+	/* allocate an IObuffer for the eventq */
+	rc = efhw_nic_event_queue_alloc_iobuffer(nic, &evq->hw, evq->instance,
+						 buf_bytes);
+	if (rc < 0)
+		return rc;
+
+	/* Zero the timer-value for this queue.
+	   AND Tell the nic about the event queue. */
+	efhw_nic_event_queue_enable(nic, evq->instance, evq->hw.capacity,
+				    efhw_iopages_dma_addr(&evq->hw.iobuff) +
+				    evq->hw.iobuff_off,
+				    evq->hw.buf_tbl_alloc.base);
+
+	evq->lock = KEVQ_UNLOCKED;
+	evq->evq_base = efhw_iopages_ptr(&evq->hw.iobuff) + evq->hw.iobuff_off;
+	evq->evq_ptr = 0;
+	evq->evq_mask = (evq->hw.capacity * sizeof(efhw_event_t)) - 1u;
+
+	EFHW_TRACE("%s: [%d] base=%p end=%p", __FUNCTION__, evq->instance,
+		   evq->evq_base, evq->evq_base + buf_bytes);
+
+	return 0;
+}
+
+void efhw_keventq_dtor(struct efhw_nic *nic, struct efhw_keventq *evq)
+{
+	EFHW_ASSERT(evq);
+
+	EFHW_TRACE("%s: [%d]", __FUNCTION__, evq->instance);
+
+	/* Zero the timer-value for this queue.
+	   And Tell NIC to stop using this event queue. */
+	efhw_nic_event_queue_disable(nic, evq->instance, 0);
+
+	/* free the pages used by the eventq itself */
+	efhw_iopages_free(nic, &evq->hw.iobuff);
+}
+
+void
+efhw_handle_txdmaq_flushed(struct efhw_nic *nic, struct efhw_ev_handler *h,
+			   efhw_event_t *evp)
+{
+	int instance = (int)FALCON_EVENT_TX_FLUSH_Q_ID(evp);
+	EFHW_TRACE("%s: instance=%d", __FUNCTION__, instance);
+
+	if (!h->dmaq_flushed_fn) {
+		EFHW_WARN("%s: no handler registered", __FUNCTION__);
+		return;
+	}
+
+	h->dmaq_flushed_fn(nic, instance, false);
+}
+
+void
+efhw_handle_rxdmaq_flushed(struct efhw_nic *nic, struct efhw_ev_handler *h,
+			   efhw_event_t *evp)
+{
+	int instance = (int)FALCON_EVENT_RX_FLUSH_Q_ID(evp);
+	EFHW_TRACE("%s: instance=%d", __FUNCTION__, instance);
+
+	if (!h->dmaq_flushed_fn) {
+		EFHW_WARN("%s: no handler registered", __FUNCTION__);
+		return;
+	}
+
+	h->dmaq_flushed_fn(nic, instance, true);
+}
+
+void
+efhw_handle_wakeup_event(struct efhw_nic *nic, struct efhw_ev_handler *h,
+			 efhw_event_t *evp)
+{
+	if (!h->wakeup_fn) {
+		EFHW_WARN("%s: no handler registered", __FUNCTION__);
+		return;
+	}
+
+	h->wakeup_fn(nic, evp);
+}
+
+void
+efhw_handle_timeout_event(struct efhw_nic *nic, struct efhw_ev_handler *h,
+			  efhw_event_t *evp)
+{
+	if (!h->timeout_fn) {
+		EFHW_WARN("%s: no handler registered", __FUNCTION__);
+		return;
+	}
+
+	h->timeout_fn(nic, evp);
+}
+
+/**********************************************************************
+ * Kernel event queue event handling.
+ */
+
+int efhw_keventq_poll(struct efhw_nic *nic, struct efhw_keventq *q)
+{
+	efhw_event_t *ev;
+	int l, count = 0;
+
+	EFHW_ASSERT(nic);
+	EFHW_ASSERT(q);
+	EFHW_ASSERT(q->ev_handlers);
+
+	/* Acquire the lock, or mark the queue as needing re-checking. */
+	for (;;) {
+		l = q->lock;
+		if (l == KEVQ_UNLOCKED) {
+			if ((int)cmpxchg(&q->lock, l, KEVQ_LOCKED) == l)
+				break;
+		} else if (l == KEVQ_LOCKED) {
+			if ((int)cmpxchg(&q->lock, l, KEVQ_RECHECK) == l)
+				return 0;
+		} else {	/* already marked for re-checking */
+			EFHW_ASSERT(l == KEVQ_RECHECK);
+			return 0;
+		}
+	}
+
+	if (unlikely(EFHW_EVENT_OVERFLOW(q, q)))
+		goto overflow;
+
+	ev = EFHW_EVENT_PTR(q, q, 0);
+
+#ifndef NDEBUG
+	if (!EFHW_IS_EVENT(ev))
+		EFHW_TRACE("%s: %d NO EVENTS!", __FUNCTION__, q->instance);
+#endif
+
+	for (;;) {
+		/* Convention for return codes for handlers is:
+		 **   0   - no error, event consumed
+		 **   1   - no error, event not consumed
+		 **   -ve - error,    event not consumed
+		 */
+		if (likely(EFHW_IS_EVENT(ev))) {
+			count++;
+
+			switch (FALCON_EVENT_CODE(ev)) {
+
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+			case FALCON_EVENT_CODE_CHAR:
+				falcon_handle_char_event(nic, q->ev_handlers,
+							 ev);
+				break;
+#endif
+
+			default:
+				EFHW_ERR("efhw_keventq_poll: [%d] UNEXPECTED "
+					 "EVENT:"FALCON_EVENT_FMT,
+					 q->instance,
+					 FALCON_EVENT_PRI_ARG(*ev));
+			}
+
+			EFHW_CLEAR_EVENT(ev);
+			EFHW_EVENTQ_NEXT(q);
+
+			ev = EFHW_EVENT_PTR(q, q, 0);
+		} else {
+			/* No events left.  Release the lock (checking if we
+			 * need to re-poll to avoid race). */
+			l = q->lock;
+			if (l == KEVQ_LOCKED) {
+				if ((int)cmpxchg(&q->lock, l, KEVQ_UNLOCKED)
+				    == l) {
+					EFHW_TRACE
+					    ("efhw_keventq_poll: %d clean exit",
+					     q->instance);
+					goto clean_exit;
+				}
+			}
+
+			/* Potentially more work to do. */
+			l = q->lock;
+			EFHW_ASSERT(l == KEVQ_RECHECK);
+			EFHW_TEST((int)cmpxchg(&q->lock, l, KEVQ_LOCKED) == l);
+			EFHW_TRACE("efhw_keventq_poll: %d re-poll required",
+				   q->instance);
+		}
+	}
+
+	/* shouldn't get here */
+	EFHW_ASSERT(0);
+
+overflow:
+	/* ?? Oh dear.  Should we poll everything that could have possibly
+	 ** happened?  Or merely cry out in anguish...
+	 */
+	EFHW_WARN("efhw_keventq_poll: %d ***** OVERFLOW nic %d *****",
+		  q->instance, nic->index);
+
+	q->lock = KEVQ_UNLOCKED;
+	return count;
+
+clean_exit:
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+	/* Ack the processed events so that this event queue can potentially
+	   raise interrupts again */
+	falcon_nic_evq_ack(nic, q->instance,
+			   (EFHW_EVENT_OFFSET(q, q, 0) / sizeof(efhw_event_t)),
+			   false);
+#endif
+	return count;
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/falcon.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/falcon.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/falcon.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/falcon.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,2758 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains Falcon hardware support.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/driver/efab/hardware.h>
+#include <ci/efhw/debug.h>
+#include <ci/efhw/iopage.h>
+#include <ci/efhw/falcon.h>
+#include <ci/efhw/falcon_hash.h>
+#include <ci/efhw/nic.h>
+#include <ci/efhw/eventq.h>
+#include <ci/efhw/checks.h>
+
+
+/*----------------------------------------------------------------------------
+ *
+ * Workarounds and options
+ *
+ *---------------------------------------------------------------------------*/
+
+/* on for debug builds */
+#ifndef NDEBUG
+#  define FALCON_FULL_FILTER_CACHE 1	/* complete SW shadow of filter tbl */
+#  define FALCON_VERIFY_FILTERS    0
+#else /* Also adds duplicate filter check */
+#  define FALCON_FULL_FILTER_CACHE 1	/* keep this on for some security */
+#  define FALCON_VERIFY_FILTERS    0
+#endif
+
+/* options */
+#define RX_FILTER_CTL_SRCH_LIMIT_TCP_FULL 8	/* default search limit */
+#define RX_FILTER_CTL_SRCH_LIMIT_TCP_WILD 8	/* default search limit */
+#define RX_FILTER_CTL_SRCH_LIMIT_UDP_FULL 8	/* default search limit */
+#define RX_FILTER_CTL_SRCH_LIMIT_UDP_WILD 8	/* default search limit */
+#define RX_FILTER_CTL_SRCH_FUDGE_WILD 3	/* increase the search limit */
+#define RX_FILTER_CTL_SRCH_FUDGE_FULL 1	/* increase the search limit */
+
+#define FALCON_MAC_SET_TYPE_BY_SPEED           1
+
+/* FIXME: We should detect mode at runtime. */
+#define FALCON_BUFFER_TABLE_FULL_MODE          1
+
+/*----------------------------------------------------------------------------
+ *
+ * Debug Macros
+ *
+ *---------------------------------------------------------------------------*/
+
+#ifndef __KERNEL__
+#define _DEBUG_SYM_ extern
+#else
+#define _DEBUG_SYM_ static inline
+#endif
+
+ /*----------------------------------------------------------------------------
+  *
+  * Macros and forward declarations
+  *
+  *--------------------------------------------------------------------------*/
+
+#define FALCON_REGION_NUM 4	/* number of supported memory regions */
+
+#define FALCON_BUFFER_TBL_HALF_BYTES 4
+#define FALCON_BUFFER_TBL_FULL_BYTES 8
+
+/* Shadow buffer table - hack for testing only */
+#if FALCON_BUFFER_TABLE_FULL_MODE == 0
+# define FALCON_USE_SHADOW_BUFFER_TABLE 1
+#else
+# define FALCON_USE_SHADOW_BUFFER_TABLE 0
+#endif
+
+#if FALCON_USE_SHADOW_BUFFER_TABLE
+static uint64_t _falcon_buffer_table[FALCON_BUFFER_TBL_NUM];
+#endif
+
+/*----------------------------------------------------------------------------
+ *
+ * Header assertion checks
+ *
+ *---------------------------------------------------------------------------*/
+
+#define FALCON_ASSERT_VALID()	/* nothing yet */
+
+/* Falcon has a 128bit register model but most registers have useful
+   defaults or only implement a small number of bits. Some registers
+   can be programmed 32bits UNLOCKED all others should be interlocked
+   against other threads within the same protection domain.
+
+   Aim is for software to perform the minimum number of writes and
+   also to minimise the read-modify-write activity (which generally
+   indicates a lack of clarity in the use model).
+
+   Registers which are programmed in this module are listed below
+   together with the method of access. Care must be taken to ensure
+   remain adequate if the register spec changes.
+
+   All 128bits programmed
+    FALCON_BUFFER_TBL_HALF
+    RX_FILTER_TBL
+    TX_DESC_PTR_TBL
+    RX_DESC_PTR_TBL
+    DRV_EV_REG
+
+   All 64bits programmed
+    FALCON_BUFFER_TBL_FULL
+
+   32 bits are programmed (UNLOCKED)
+    EVQ_RPTR_REG
+
+   Low 64bits programmed remainder are written with a random number
+    RX_DC_CFG_REG
+    TX_DC_CFG_REG
+    SRM_RX_DC_CFG_REG
+    SRM_TX_DC_CFG_REG
+    BUF_TBL_CFG_REG
+    BUF_TBL_UPD_REG
+    SRM_UPD_EVQ_REG
+    EVQ_PTR_TBL
+    TIMER_CMD_REG
+    TX_PACE_TBL
+    FATAL_INTR_REG
+    INT_EN_REG (When enabling interrupts)
+    TX_FLUSH_DESCQ_REG
+    RX_FLUSH_DESCQ
+
+  Read Modify Write on low 32bits remainder are written with a random number
+    INT_EN_REG (When sending a driver interrupt)
+    DRIVER_REGX
+
+  Read Modify Write on low 64bits remainder are written with a random number
+   SRM_CFG_REG_OFST
+   RX_CFG_REG_OFST
+   RX_FILTER_CTL_REG
+
+  Read Modify Write on full 128bits
+   TXDP_RESERVED_REG  (aka TXDP_UNDOCUMENTED)
+   TX_CFG_REG
+
+*/
+
+/*----------------------------------------------------------------------------
+ *
+ * Filters static data
+ *
+ *---------------------------------------------------------------------------*/
+
+/* Defaults are set here to support dma.c */
+static unsigned tcp_full_srch_limit = RX_FILTER_CTL_SRCH_LIMIT_TCP_FULL;
+static unsigned tcp_wild_srch_limit = RX_FILTER_CTL_SRCH_LIMIT_TCP_WILD;
+static unsigned udp_full_srch_limit = RX_FILTER_CTL_SRCH_LIMIT_UDP_FULL;
+static unsigned udp_wild_srch_limit = RX_FILTER_CTL_SRCH_LIMIT_UDP_WILD;
+
+#if FALCON_VERIFY_FILTERS
+static void _falcon_nic_ipfilter_sanity(struct efhw_nic *nic);
+#endif
+
+/*----------------------------------------------------------------------------
+ *
+ * Filters low-level register interface
+ *
+ *---------------------------------------------------------------------------*/
+
+/* Build the filter entry */
+static void
+_falcon_nic_ipfilter_build(struct efhw_nic *nic,
+			   int tcp, int full, int rss_b0, int scat_b0,
+			   uint filter_i, uint dmaq_id,
+			   unsigned saddr_le32, unsigned sport_le16,
+			   unsigned daddr_le32, unsigned dport_le16,
+			   uint64_t *q0, uint64_t *q1)
+{
+	uint64_t v1, v2, v3, v4;
+	int type = tcp << 4 | full;
+
+	v4 = (((!tcp) << __DW4(TCP_UDP_1_LBN)) |
+	      (dmaq_id << __DW4(RXQ_ID_1_LBN)));
+
+	switch (nic->devtype.variant) {
+	case 'A':
+		EFHW_ASSERT(!rss_b0);
+		break;
+	case 'B':
+		v4 |= scat_b0 << __DW4(SCATTER_EN_1_B0_LBN);
+		v4 |= rss_b0 << __DW4(RSS_EN_1_B0_LBN);
+		break;
+	default:
+		EFHW_ASSERT(0);
+		break;
+	}
+
+	v3 = daddr_le32;
+
+	switch (type) {
+
+	case 0x11:		/* TCP_FULL */
+	case 0x01:		/* UDP_FULL */
+		v2 = ((dport_le16 << __DW2(DEST_PORT_TCP_1_LBN)) |
+		      (__HIGH(saddr_le32, SRC_IP_1_LBN, SRC_IP_1_WIDTH)));
+		v1 = ((__LOW(saddr_le32, SRC_IP_1_LBN, SRC_IP_1_WIDTH)) |
+		      (sport_le16 << SRC_TCP_DEST_UDP_1_LBN));
+		break;
+
+	case 0x10:		/* TCP_WILD */
+		v2 = ((uint64_t) dport_le16 << __DW2(DEST_PORT_TCP_1_LBN));
+		v1 = 0;
+		break;
+
+	case 0x00:		/* UDP_WILD */
+		v2 = 0;
+		v1 = ((uint64_t) dport_le16 << SRC_TCP_DEST_UDP_0_LBN);
+		break;
+
+	default:
+		EFHW_ASSERT(0);
+		v2 = 0;
+		v1 = 0;
+	}
+
+	*q0 = (v2 << 32) | v1;
+	*q1 = (v4 << 32) | v3;
+}
+
+static void
+_falcon_nic_ipfilter_set(struct efhw_nic *nic, int tcp,
+			 int full, int rss_b0, int scat_b0,
+			 uint filter_i, uint dmaq_id,
+			 unsigned saddr_le32, unsigned sport_le16,
+			 unsigned daddr_le32, unsigned dport_le16)
+{
+	uint64_t q0, q1;
+
+	/* wish you wouldn't do this */
+	EFHW_BUILD_ASSERT(RX_FILTER_TBL1_OFST ==
+			  RX_FILTER_TBL0_OFST + FALCON_REGISTER128);
+	EFHW_BUILD_ASSERT(TCP_UDP_1_LBN == TCP_UDP_0_LBN);
+	EFHW_BUILD_ASSERT(RXQ_ID_1_LBN == RXQ_ID_0_LBN);
+	EFHW_BUILD_ASSERT(DEST_IP_1_LBN == DEST_IP_0_LBN);
+	EFHW_BUILD_ASSERT(DEST_PORT_TCP_1_LBN == DEST_PORT_TCP_0_LBN);
+	EFHW_BUILD_ASSERT(SRC_IP_1_LBN == SRC_IP_0_LBN);
+	EFHW_BUILD_ASSERT(SRC_TCP_DEST_UDP_1_LBN == SRC_TCP_DEST_UDP_0_LBN);
+	EFHW_BUILD_ASSERT(SCATTER_EN_1_B0_LBN == SCATTER_EN_0_B0_LBN);
+	EFHW_BUILD_ASSERT(RSS_EN_1_B0_LBN == RSS_EN_0_B0_LBN);
+
+	EFHW_BUILD_ASSERT(TCP_UDP_1_WIDTH == TCP_UDP_0_WIDTH);
+	EFHW_BUILD_ASSERT(RXQ_ID_1_WIDTH == RXQ_ID_0_WIDTH);
+	EFHW_BUILD_ASSERT(DEST_IP_1_WIDTH == DEST_IP_0_WIDTH);
+	EFHW_BUILD_ASSERT(DEST_PORT_TCP_1_WIDTH == DEST_PORT_TCP_0_WIDTH);
+	EFHW_BUILD_ASSERT(SRC_IP_1_WIDTH == SRC_IP_0_WIDTH);
+	EFHW_BUILD_ASSERT(SRC_TCP_DEST_UDP_1_WIDTH == SRC_TCP_DEST_UDP_0_WIDTH);
+	EFHW_BUILD_ASSERT(SCATTER_EN_1_B0_WIDTH == SCATTER_EN_0_B0_WIDTH);
+	EFHW_BUILD_ASSERT(RSS_EN_1_B0_WIDTH == RSS_EN_0_B0_WIDTH);
+
+	/* TODO: Use filter table 1 as well */
+	ulong offset = RX_FILTER_TBL0_OFST + filter_i * 2 * FALCON_REGISTER128;
+
+	EFHW_TRACE("%s[%x]: offset=%lx", __FUNCTION__, filter_i, offset);
+
+	EFHW_TRACE("%s[%x]: filter %d tcp %d full %d src=%x:%x dest=%x:%x%s%s",
+		   __FUNCTION__, filter_i, tcp, full, dmaq_id,
+		   saddr_le32, sport_le16, daddr_le32, dport_le16,
+		   rss_b0 ? " RSS" : "", scat_b0 ? " SCAT" : "");
+
+	EFHW_ASSERT(filter_i < nic->filter_tbl_size);
+
+	/* dword 4 */
+	__DW4CHCK(TCP_UDP_1_LBN, TCP_UDP_1_WIDTH);
+	__DW4CHCK(RXQ_ID_1_LBN, RXQ_ID_1_WIDTH);
+
+	__RANGECHCK(tcp, TCP_UDP_1_WIDTH);
+	__RANGECHCK(dmaq_id, RXQ_ID_1_WIDTH);
+
+	/* dword 3 */
+	__DW3CHCK(DEST_IP_1_LBN, DEST_IP_1_WIDTH);
+	__RANGECHCK(daddr_le32, DEST_IP_1_WIDTH);
+
+	/* dword 2 */
+	__DW2CHCK(DEST_PORT_TCP_1_LBN, DEST_PORT_TCP_1_WIDTH);
+	__LWCHK(SRC_IP_1_LBN, SRC_IP_1_WIDTH);
+	__RANGECHCK(saddr_le32, SRC_IP_1_WIDTH);
+
+	/* dword 1 */
+	__DWCHCK(SRC_TCP_DEST_UDP_1_LBN, SRC_TCP_DEST_UDP_1_WIDTH);
+	__RANGECHCK(sport_le16, SRC_TCP_DEST_UDP_1_WIDTH);
+	__RANGECHCK(dport_le16, SRC_TCP_DEST_UDP_1_WIDTH);
+
+	/* Falcon requires 128 bit atomic access for this register */
+	_falcon_nic_ipfilter_build(nic, tcp, full, rss_b0, scat_b0,
+				   filter_i, dmaq_id, saddr_le32, sport_le16,
+				   daddr_le32, dport_le16, &q0, &q1);
+
+	EFHW_TRACE("%s[%x]@%p+%lx: %" PRIx64 " %" PRIx64, __FUNCTION__,
+		   filter_i, EFHW_KVA(nic), offset, q0, q1);
+
+	falcon_write_qq(EFHW_KVA(nic) + offset, q0, q1);
+	mmiowb();
+
+#if FALCON_VERIFY_FILTERS
+	{
+		uint64_t q0read, q1read;
+
+		/* Read a different entry first - entry BIU flushed shadow */
+		falcon_read_qq(EFHW_KVA(nic) + offset+0x10, &q0read, &q1read);
+		falcon_read_qq(EFHW_KVA(nic) + offset, &q0read, &q1read);
+		EFHW_ASSERT(q0read == q0);
+		EFHW_ASSERT(q1read == q1);
+
+		_falcon_nic_ipfilter_sanity(nic);
+	}
+#endif
+}
+
+static void _falcon_nic_ipfilter_clear(struct efhw_nic *nic, uint filter_i)
+{
+	/* TODO: Use filter table 1 as well */
+	ulong offset = RX_FILTER_TBL0_OFST + filter_i * 2 * FALCON_REGISTER128;
+
+	EFHW_ASSERT(filter_i < nic->filter_tbl_size);
+
+	EFHW_TRACE("%s[%x]", __FUNCTION__, filter_i);
+
+	/* Falcon requires 128 bit atomic access for this register */
+	falcon_write_qq(EFHW_KVA(nic) + offset, 0, 0);
+	mmiowb();
+#if FALCON_VERIFY_FILTERS
+	{
+		uint64_t q0read, q1read;
+
+		/* Read a different entry first - entry BIU flushed shadow */
+		falcon_read_qq(EFHW_KVA(nic) + offset+0x10, &q0read, &q1read);
+		falcon_read_qq(EFHW_KVA(nic) + offset, &q0read, &q1read);
+		EFHW_ASSERT(q0read == 0);
+		EFHW_ASSERT(q1read == 0);
+
+		_falcon_nic_ipfilter_sanity(nic);
+	}
+#endif
+}
+
+/*----------------------------------------------------------------------------
+ *
+ * DMAQ low-level register interface
+ *
+ *---------------------------------------------------------------------------*/
+
+static unsigned dmaq_sizes[] = {
+	512,
+	EFHW_1K,
+	EFHW_2K,
+	EFHW_4K,
+};
+
+#define N_DMAQ_SIZES  (sizeof(dmaq_sizes) / sizeof(dmaq_sizes[0]))
+
+static inline ulong falcon_dma_tx_q_offset(struct efhw_nic *nic, unsigned dmaq)
+{
+	EFHW_ASSERT(dmaq < FALCON_DMAQ_NUM);
+	return TX_DESC_PTR_TBL_OFST + dmaq * FALCON_REGISTER128;
+}
+
+static inline uint falcon_dma_tx_q_size_index(uint dmaq_size)
+{
+	uint i;
+
+	/* size must be one of the various options, otherwise we assert */
+	for (i = 0; i < N_DMAQ_SIZES; i++) {
+		if (dmaq_size == dmaq_sizes[i])
+			break;
+	}
+	EFHW_ASSERT(i < N_DMAQ_SIZES);
+	return i;
+}
+
+static void
+falcon_dmaq_tx_q_init(struct efhw_nic *nic,
+		      uint dmaq, uint evq_id, uint own_id,
+		      uint tag, uint dmaq_size, uint buf_idx, uint flags)
+{
+	FALCON_LOCK_DECL;
+	uint index, desc_type;
+	uint64_t val1, val2, val3;
+	ulong offset;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+
+	/* Q attributes */
+	int iscsi_hdig_en = ((flags & EFHW_VI_ISCSI_TX_HDIG_EN) != 0);
+	int iscsi_ddig_en = ((flags & EFHW_VI_ISCSI_TX_DDIG_EN) != 0);
+	int csum_ip_dis = ((flags & EFHW_VI_TX_IP_CSUM_DIS) != 0);
+	int csum_tcp_dis = ((flags & EFHW_VI_TX_TCPUDP_CSUM_DIS) != 0);
+	int non_ip_drop_dis = ((flags & EFHW_VI_TX_TCPUDP_ONLY) == 0);
+
+	/* initialise the TX descriptor queue pointer table */
+
+	/* NB physical vs buffer addressing is determined by the Queue ID. */
+
+	offset = falcon_dma_tx_q_offset(nic, dmaq);
+	index = falcon_dma_tx_q_size_index(dmaq_size);
+
+	/* allow VI flag to override this queue's descriptor type */
+	desc_type = (flags & EFHW_VI_TX_PHYS_ADDR_EN) ? 0 : 1;
+
+	/* bug9403: It is dangerous to allow buffer-addressed queues to
+	 * have owner_id=0. */
+	EFHW_ASSERT((own_id > 0) || desc_type == 0);
+
+	/* dword 1 */
+	__DWCHCK(TX_DESCQ_FLUSH_LBN, TX_DESCQ_FLUSH_WIDTH);
+	__DWCHCK(TX_DESCQ_TYPE_LBN, TX_DESCQ_TYPE_WIDTH);
+	__DWCHCK(TX_DESCQ_SIZE_LBN, TX_DESCQ_SIZE_WIDTH);
+	__DWCHCK(TX_DESCQ_LABEL_LBN, TX_DESCQ_LABEL_WIDTH);
+	__DWCHCK(TX_DESCQ_OWNER_ID_LBN, TX_DESCQ_OWNER_ID_WIDTH);
+
+	__LWCHK(TX_DESCQ_EVQ_ID_LBN, TX_DESCQ_EVQ_ID_WIDTH);
+
+	__RANGECHCK(1, TX_DESCQ_FLUSH_WIDTH);
+	__RANGECHCK(desc_type, TX_DESCQ_TYPE_WIDTH);
+	__RANGECHCK(index, TX_DESCQ_SIZE_WIDTH);
+	__RANGECHCK(tag, TX_DESCQ_LABEL_WIDTH);
+	__RANGECHCK(own_id, TX_DESCQ_OWNER_ID_WIDTH);
+	__RANGECHCK(evq_id, TX_DESCQ_EVQ_ID_WIDTH);
+
+	val1 = ((desc_type << TX_DESCQ_TYPE_LBN) |
+		(index << TX_DESCQ_SIZE_LBN) |
+		(tag << TX_DESCQ_LABEL_LBN) |
+		(own_id << TX_DESCQ_OWNER_ID_LBN) |
+		(__LOW(evq_id, TX_DESCQ_EVQ_ID_LBN, TX_DESCQ_EVQ_ID_WIDTH)));
+
+	/* dword 2 */
+	__DW2CHCK(TX_DESCQ_BUF_BASE_ID_LBN, TX_DESCQ_BUF_BASE_ID_WIDTH);
+	__RANGECHCK(buf_idx, TX_DESCQ_BUF_BASE_ID_WIDTH);
+
+	val2 = ((__HIGH(evq_id, TX_DESCQ_EVQ_ID_LBN, TX_DESCQ_EVQ_ID_WIDTH)) |
+		(buf_idx << __DW2(TX_DESCQ_BUF_BASE_ID_LBN)));
+
+	/* dword 3 */
+	__DW3CHCK(TX_ISCSI_HDIG_EN_LBN, TX_ISCSI_HDIG_EN_WIDTH);
+	__DW3CHCK(TX_ISCSI_DDIG_EN_LBN, TX_ISCSI_DDIG_EN_WIDTH);
+	__RANGECHCK(iscsi_hdig_en, TX_ISCSI_HDIG_EN_WIDTH);
+	__RANGECHCK(iscsi_ddig_en, TX_ISCSI_DDIG_EN_WIDTH);
+
+	val3 = ((iscsi_hdig_en << __DW3(TX_ISCSI_HDIG_EN_LBN)) |
+		(iscsi_ddig_en << __DW3(TX_ISCSI_DDIG_EN_LBN)) |
+		(1 << __DW3(TX_DESCQ_EN_LBN)));	/* queue enable bit */
+
+	switch (nic->devtype.variant) {
+	case 'B':
+		__DW3CHCK(TX_NON_IP_DROP_DIS_B0_LBN,
+			  TX_NON_IP_DROP_DIS_B0_WIDTH);
+		__DW3CHCK(TX_IP_CHKSM_DIS_B0_LBN, TX_IP_CHKSM_DIS_B0_WIDTH);
+		__DW3CHCK(TX_TCP_CHKSM_DIS_B0_LBN, TX_TCP_CHKSM_DIS_B0_WIDTH);
+
+		val3 |= ((non_ip_drop_dis << __DW3(TX_NON_IP_DROP_DIS_B0_LBN))|
+			 (csum_ip_dis << __DW3(TX_IP_CHKSM_DIS_B0_LBN)) |
+			 (csum_tcp_dis << __DW3(TX_TCP_CHKSM_DIS_B0_LBN)));
+		break;
+	case 'A':
+		if (csum_ip_dis || csum_tcp_dis || !non_ip_drop_dis)
+			EFHW_WARN
+				("%s: bad settings for A1 csum_ip_dis=%d "
+				 "csum_tcp_dis=%d non_ip_drop_dis=%d",
+				 __FUNCTION__, csum_ip_dis,
+				 csum_tcp_dis, non_ip_drop_dis);
+		break;
+	default:
+		EFHW_ASSERT(0);
+		break;
+	}
+
+	EFHW_TRACE("%s: txq %x evq %u tag %x id %x buf %x "
+		   "%x:%x:%x->%" PRIx64 ":%" PRIx64 ":%" PRIx64,
+		   __FUNCTION__,
+		   dmaq, evq_id, tag, own_id, buf_idx, dmaq_size,
+		   iscsi_hdig_en, iscsi_ddig_en, val1, val2, val3);
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_write_qq(efhw_kva + offset, ((val2 << 32) | val1), val3);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+	return;
+}
+
+static inline ulong
+falcon_dma_rx_q_offset(struct efhw_nic *nic, unsigned dmaq)
+{
+	EFHW_ASSERT(dmaq < FALCON_DMAQ_NUM);
+	return RX_DESC_PTR_TBL_OFST + dmaq * FALCON_REGISTER128;
+}
+
+static void
+falcon_dmaq_rx_q_init(struct efhw_nic *nic,
+		      uint dmaq, uint evq_id, uint own_id,
+		      uint tag, uint dmaq_size, uint buf_idx, uint flags)
+{
+	FALCON_LOCK_DECL;
+	uint i, desc_type = 1;
+	uint64_t val1, val2, val3;
+	ulong offset;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+
+	/* Q attributes */
+#if BUG5762_WORKAROUND
+	int jumbo = 1;		/* Queues must not have mixed types */
+#else
+	int jumbo = ((flags & EFHW_VI_JUMBO_EN) != 0);
+#endif
+	int iscsi_hdig_en = ((flags & EFHW_VI_ISCSI_RX_HDIG_EN) != 0);
+	int iscsi_ddig_en = ((flags & EFHW_VI_ISCSI_RX_DDIG_EN) != 0);
+
+	/* initialise the TX descriptor queue pointer table */
+	offset = falcon_dma_rx_q_offset(nic, dmaq);
+
+	/* size must be one of the various options, otherwise we assert */
+	for (i = 0; i < N_DMAQ_SIZES; i++) {
+		if (dmaq_size == dmaq_sizes[i])
+			break;
+	}
+	EFHW_ASSERT(i < N_DMAQ_SIZES);
+
+	/* allow VI flag to override this queue's descriptor type */
+	desc_type = (flags & EFHW_VI_RX_PHYS_ADDR_EN) ? 0 : 1;
+
+	/* bug9403: It is dangerous to allow buffer-addressed queues to have
+	 * owner_id=0 */
+	EFHW_ASSERT((own_id > 0) || desc_type == 0);
+
+	/* dword 1 */
+	__DWCHCK(RX_DESCQ_EN_LBN, RX_DESCQ_EN_WIDTH);
+	__DWCHCK(RX_DESCQ_JUMBO_LBN, RX_DESCQ_JUMBO_WIDTH);
+	__DWCHCK(RX_DESCQ_TYPE_LBN, RX_DESCQ_TYPE_WIDTH);
+	__DWCHCK(RX_DESCQ_SIZE_LBN, RX_DESCQ_SIZE_WIDTH);
+	__DWCHCK(RX_DESCQ_LABEL_LBN, RX_DESCQ_LABEL_WIDTH);
+	__DWCHCK(RX_DESCQ_OWNER_ID_LBN, RX_DESCQ_OWNER_ID_WIDTH);
+
+	__LWCHK(RX_DESCQ_EVQ_ID_LBN, RX_DESCQ_EVQ_ID_WIDTH);
+
+	__RANGECHCK(1, RX_DESCQ_EN_WIDTH);
+	__RANGECHCK(jumbo, RX_DESCQ_JUMBO_WIDTH);
+	__RANGECHCK(desc_type, RX_DESCQ_TYPE_WIDTH);
+	__RANGECHCK(i, RX_DESCQ_SIZE_WIDTH);
+	__RANGECHCK(tag, RX_DESCQ_LABEL_WIDTH);
+	__RANGECHCK(own_id, RX_DESCQ_OWNER_ID_WIDTH);
+	__RANGECHCK(evq_id, RX_DESCQ_EVQ_ID_WIDTH);
+
+	val1 = ((1 << RX_DESCQ_EN_LBN) |
+		(jumbo << RX_DESCQ_JUMBO_LBN) |
+		(desc_type << RX_DESCQ_TYPE_LBN) |
+		(i << RX_DESCQ_SIZE_LBN) |
+		(tag << RX_DESCQ_LABEL_LBN) |
+		(own_id << RX_DESCQ_OWNER_ID_LBN) |
+		(__LOW(evq_id, RX_DESCQ_EVQ_ID_LBN, RX_DESCQ_EVQ_ID_WIDTH)));
+
+	/* dword 2 */
+	__DW2CHCK(RX_DESCQ_BUF_BASE_ID_LBN, RX_DESCQ_BUF_BASE_ID_WIDTH);
+	__RANGECHCK(buf_idx, RX_DESCQ_BUF_BASE_ID_WIDTH);
+
+	val2 = ((__HIGH(evq_id, RX_DESCQ_EVQ_ID_LBN, RX_DESCQ_EVQ_ID_WIDTH)) |
+		(buf_idx << __DW2(RX_DESCQ_BUF_BASE_ID_LBN)));
+
+	/* dword 3 */
+	__DW3CHCK(RX_ISCSI_HDIG_EN_LBN, RX_ISCSI_HDIG_EN_WIDTH);
+	__DW3CHCK(RX_ISCSI_DDIG_EN_LBN, RX_ISCSI_DDIG_EN_WIDTH);
+	__RANGECHCK(iscsi_hdig_en, RX_ISCSI_HDIG_EN_WIDTH);
+	__RANGECHCK(iscsi_ddig_en, RX_ISCSI_DDIG_EN_WIDTH);
+
+	val3 = (iscsi_hdig_en << __DW3(RX_ISCSI_HDIG_EN_LBN)) |
+	    (iscsi_ddig_en << __DW3(RX_ISCSI_DDIG_EN_LBN));
+
+	EFHW_TRACE("%s: rxq %x evq %u tag %x id %x buf %x %s "
+		   "%x:%x:%x -> %" PRIx64 ":%" PRIx64 ":%" PRIx64,
+		   __FUNCTION__,
+		   dmaq, evq_id, tag, own_id, buf_idx,
+		   jumbo ? "jumbo" : "normal", dmaq_size,
+		   iscsi_hdig_en, iscsi_ddig_en, val1, val2, val3);
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_write_qq(efhw_kva + offset, ((val2 << 32) | val1), val3);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+	return;
+}
+
+static void falcon_dmaq_tx_q_disable(struct efhw_nic *nic, uint dmaq)
+{
+	FALCON_LOCK_DECL;
+	uint64_t val1, val2, val3;
+	ulong offset;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+
+	/* initialise the TX descriptor queue pointer table */
+
+	offset = falcon_dma_tx_q_offset(nic, dmaq);
+
+	/* dword 1 */
+	__DWCHCK(TX_DESCQ_TYPE_LBN, TX_DESCQ_TYPE_WIDTH);
+
+	val1 = ((uint64_t) 1 << TX_DESCQ_TYPE_LBN);
+
+	/* dword 2 */
+	val2 = 0;
+
+	/* dword 3 */
+	val3 = (0 << __DW3(TX_DESCQ_EN_LBN));	/* queue enable bit */
+
+	EFHW_TRACE("%s: %x->%" PRIx64 ":%" PRIx64 ":%" PRIx64,
+		   __FUNCTION__, dmaq, val1, val2, val3);
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_write_qq(efhw_kva + offset, ((val2 << 32) | val1), val3);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+	return;
+}
+
+static void falcon_dmaq_rx_q_disable(struct efhw_nic *nic, uint dmaq)
+{
+	FALCON_LOCK_DECL;
+	uint64_t val1, val2, val3;
+	ulong offset;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+
+	/* initialise the TX descriptor queue pointer table */
+	offset = falcon_dma_rx_q_offset(nic, dmaq);
+
+	/* dword 1 */
+	__DWCHCK(RX_DESCQ_EN_LBN, RX_DESCQ_EN_WIDTH);
+	__DWCHCK(RX_DESCQ_TYPE_LBN, RX_DESCQ_TYPE_WIDTH);
+
+	val1 = ((0 << RX_DESCQ_EN_LBN) | (1 << RX_DESCQ_TYPE_LBN));
+
+	/* dword 2 */
+	val2 = 0;
+
+	/* dword 3 */
+	val3 = 0;
+
+	EFHW_TRACE("falcon_dmaq_rx_q_disable: %x->%"
+		   PRIx64 ":%" PRIx64 ":%" PRIx64,
+		   dmaq, val1, val2, val3);
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_write_qq(efhw_kva + offset, ((val2 << 32) | val1), val3);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+	return;
+}
+
+
+/*----------------------------------------------------------------------------
+ *
+ * Buffer Table low-level register interface
+ *
+ *---------------------------------------------------------------------------*/
+
+/*! Convert a (potentially) 64-bit physical address to 32-bits.  Every use
+** of this function is a place where we're not 64-bit clean.
+*/
+static inline uint32_t dma_addr_to_u32(dma_addr_t addr)
+{
+	/* Top bits had better be zero! */
+	EFHW_ASSERT(addr == (addr & 0xffffffff));
+	return (uint32_t) addr;
+}
+
+static inline uint32_t
+falcon_nic_buffer_table_entry32_mk(dma_addr_t dma_addr, int own_id)
+{
+	uint32_t dma_addr32 = FALCON_BUFFER_4K_PAGE(dma_addr_to_u32(dma_addr));
+
+	/* don't do this to me */
+	EFHW_BUILD_ASSERT(BUF_ADR_HBUF_ODD_LBN == BUF_ADR_HBUF_EVEN_LBN + 32);
+	EFHW_BUILD_ASSERT(BUF_OWNER_ID_HBUF_ODD_LBN ==
+			  BUF_OWNER_ID_HBUF_EVEN_LBN + 32);
+
+	EFHW_BUILD_ASSERT(BUF_OWNER_ID_HBUF_ODD_WIDTH ==
+			  BUF_OWNER_ID_HBUF_EVEN_WIDTH);
+	EFHW_BUILD_ASSERT(BUF_ADR_HBUF_ODD_WIDTH == BUF_ADR_HBUF_EVEN_WIDTH);
+
+	__DWCHCK(BUF_ADR_HBUF_EVEN_LBN, BUF_ADR_HBUF_EVEN_WIDTH);
+	__DWCHCK(BUF_OWNER_ID_HBUF_EVEN_LBN, BUF_OWNER_ID_HBUF_EVEN_WIDTH);
+
+	__RANGECHCK(dma_addr32, BUF_ADR_HBUF_EVEN_WIDTH);
+	__RANGECHCK(own_id, BUF_OWNER_ID_HBUF_EVEN_WIDTH);
+
+	return ((dma_addr32 << BUF_ADR_HBUF_EVEN_LBN) |
+		(own_id << BUF_OWNER_ID_HBUF_EVEN_LBN));
+}
+
+static inline uint64_t
+falcon_nic_buffer_table_entry64_mk(dma_addr_t dma_addr,
+				   int bufsz,	/* bytes */
+				   int region, int own_id)
+{
+	__DW2CHCK(IP_DAT_BUF_SIZE_LBN, IP_DAT_BUF_SIZE_WIDTH);
+	__DW2CHCK(BUF_ADR_REGION_LBN, BUF_ADR_REGION_WIDTH);
+	__LWCHK(BUF_ADR_FBUF_LBN, BUF_ADR_FBUF_WIDTH);
+	__DWCHCK(BUF_OWNER_ID_FBUF_LBN, BUF_OWNER_ID_FBUF_WIDTH);
+
+	EFHW_ASSERT((bufsz == EFHW_4K) || (bufsz == EFHW_8K));
+
+	dma_addr = (dma_addr >> 12) & __FALCON_MASK64(BUF_ADR_FBUF_WIDTH);
+
+	__RANGECHCK(dma_addr, BUF_ADR_FBUF_WIDTH);
+	__RANGECHCK(1, IP_DAT_BUF_SIZE_WIDTH);
+	__RANGECHCK(region, BUF_ADR_REGION_WIDTH);
+	__RANGECHCK(own_id, BUF_OWNER_ID_FBUF_WIDTH);
+
+	return (((uint64_t) (bufsz == EFHW_8K) << IP_DAT_BUF_SIZE_LBN) |
+		((uint64_t) region << BUF_ADR_REGION_LBN) |
+		((uint64_t) dma_addr << BUF_ADR_FBUF_LBN) |
+		((uint64_t) own_id << BUF_OWNER_ID_FBUF_LBN));
+}
+
+static inline void
+_falcon_nic_buffer_table_set32(struct efhw_nic *nic,
+			       dma_addr_t dma_addr, uint bufsz,
+			       uint region, /* not used */
+			       int own_id, int buffer_id)
+{
+	/* programming the half table needs to be done in pairs. */
+	uint64_t entry, val, shift;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+	efhw_ioaddr_t offset;
+
+	EFHW_BUILD_ASSERT(BUF_ADR_HBUF_ODD_LBN == BUF_ADR_HBUF_EVEN_LBN + 32);
+	EFHW_BUILD_ASSERT(BUF_OWNER_ID_HBUF_ODD_LBN ==
+			  BUF_OWNER_ID_HBUF_EVEN_LBN + 32);
+
+	shift = (buffer_id & 1) ? 32 : 0;
+
+	offset = (efhw_kva + BUF_HALF_TBL_OFST +
+		  ((buffer_id & ~1) * FALCON_BUFFER_TBL_HALF_BYTES));
+
+	entry = falcon_nic_buffer_table_entry32_mk(dma_addr_to_u32(dma_addr),
+						   own_id);
+
+#if FALCON_USE_SHADOW_BUFFER_TABLE
+	val = _falcon_buffer_table[buffer_id & ~1];
+#else
+	/* This will not work unless we've completed
+	 * the buffer table updates */
+	falcon_read_q(offset, &val);
+#endif
+	val &= ~(((uint64_t) 0xffffffff) << shift);
+	val |= (entry << shift);
+
+	EFHW_TRACE("%s[%x]: " ci_dma_addr_fmt ":%x:%" PRIx64 "->%x = %"
+		   PRIx64, __FUNCTION__, buffer_id, dma_addr, own_id, entry,
+		   (unsigned)(offset - efhw_kva), val);
+
+	/* Falcon requires that access to this register is serialised */
+	falcon_write_q(offset, val);
+
+	/* NB. No mmiowb().  Caller should do that e.g by calling commit  */
+
+#if FALCON_USE_SHADOW_BUFFER_TABLE
+	_falcon_buffer_table[buffer_id & ~1] = val;
+#endif
+
+	/* Confirm the entry if the event queues haven't been set up. */
+	if (!nic->irq_handler) {
+		uint64_t new_val;
+		int count = 0;
+		while (1) {
+			mmiowb();
+			falcon_read_q(offset, &new_val);
+			if (new_val == val)
+				break;
+			count++;
+			if (count > 1000) {
+				EFHW_WARN("%s: poll Timeout", __FUNCTION__);
+				break;
+			}
+			udelay(1);
+		}
+	}
+}
+
+static inline void
+_falcon_nic_buffer_table_set64(struct efhw_nic *nic,
+			       dma_addr_t dma_addr, uint bufsz,
+			       uint region, int own_id, int buffer_id)
+{
+	efhw_ioaddr_t offset;
+	uint64_t entry;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+
+	EFHW_ASSERT(region < FALCON_REGION_NUM);
+
+	EFHW_ASSERT((bufsz == EFHW_4K) ||
+		    (bufsz == EFHW_8K && FALCON_BUFFER_TABLE_FULL_MODE));
+
+	offset = (efhw_kva + BUF_FULL_TBL_OFST +
+		  (buffer_id * FALCON_BUFFER_TBL_FULL_BYTES));
+
+	entry = falcon_nic_buffer_table_entry64_mk(dma_addr, bufsz, region,
+						   own_id);
+
+	EFHW_TRACE("%s[%x]: " ci_dma_addr_fmt
+		   ":bufsz=%x:region=%x:ownid=%x",
+		   __FUNCTION__, buffer_id, dma_addr, bufsz, region, own_id);
+
+	EFHW_TRACE("%s: BUF[%x]:NIC[%x]->%" PRIx64,
+		   __FUNCTION__, buffer_id,
+		   (unsigned int)(offset - efhw_kva), entry);
+
+	/* Falcon requires that access to this register is serialised */
+	falcon_write_q(offset, entry);
+
+	/* NB. No mmiowb().  Caller should do that e.g by calling commit */
+
+	/* Confirm the entry if the event queues haven't been set up. */
+	if (!nic->irq_handler) {
+		uint64_t new_entry;
+		int count = 0;
+		while (1) {
+			mmiowb();
+			falcon_read_q(offset, &new_entry);
+			if (new_entry == entry)
+				return;
+			count++;
+			if (count > 1000) {
+				EFHW_WARN("%s: poll Timeout waiting for "
+					  "value %"PRIx64
+					  " (last was %"PRIx64")",
+					  __FUNCTION__, entry, new_entry);
+				break;
+			}
+			udelay(1);
+		}
+	}
+}
+
+#if FALCON_BUFFER_TABLE_FULL_MODE
+#define _falcon_nic_buffer_table_set _falcon_nic_buffer_table_set64
+#else
+#define _falcon_nic_buffer_table_set _falcon_nic_buffer_table_set32
+#endif
+
+static inline void _falcon_nic_buffer_table_commit(struct efhw_nic *nic)
+{
+	/* MUST be called holding the FALCON_LOCK */
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+	uint64_t cmd;
+
+	EFHW_BUILD_ASSERT(BUF_TBL_UPD_REG_KER_OFST == BUF_TBL_UPD_REG_OFST);
+
+	__DW2CHCK(BUF_UPD_CMD_LBN, BUF_UPD_CMD_WIDTH);
+	__RANGECHCK(1, BUF_UPD_CMD_WIDTH);
+
+	cmd = ((uint64_t) 1 << BUF_UPD_CMD_LBN);
+
+	/* Falcon requires 128 bit atomic access for this register */
+	falcon_write_qq(efhw_kva + BUF_TBL_UPD_REG_OFST,
+			cmd, FALCON_ATOMIC_UPD_REG);
+	mmiowb();
+
+	nic->buf_commit_outstanding++;
+	EFHW_TRACE("COMMIT REQ out=%d", nic->buf_commit_outstanding);
+}
+
+static void falcon_nic_buffer_table_commit(struct efhw_nic *nic)
+{
+	/* nothing to do */
+}
+
+static inline void
+_falcon_nic_buffer_table_clear(struct efhw_nic *nic, int buffer_id, int num)
+{
+	uint64_t cmd;
+	uint64_t start_id = buffer_id;
+	uint64_t end_id = buffer_id + num - 1;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+
+	efhw_ioaddr_t offset = (efhw_kva + BUF_TBL_UPD_REG_OFST);
+
+	EFHW_BUILD_ASSERT(BUF_TBL_UPD_REG_KER_OFST == BUF_TBL_UPD_REG_OFST);
+
+#if !FALCON_BUFFER_TABLE_FULL_MODE
+	/* buffer_ids in half buffer mode reference pairs of buffers */
+	EFHW_ASSERT(buffer_id % 1 == 0);
+	EFHW_ASSERT(num % 1 == 0);
+	start_id = start_id >> 1;
+	end_id = end_id >> 1;
+#endif
+
+	EFHW_ASSERT(num >= 1);
+
+	__DWCHCK(BUF_CLR_START_ID_LBN, BUF_CLR_START_ID_WIDTH);
+	__DW2CHCK(BUF_CLR_END_ID_LBN, BUF_CLR_END_ID_WIDTH);
+
+	__DW2CHCK(BUF_CLR_CMD_LBN, BUF_CLR_CMD_WIDTH);
+	__RANGECHCK(1, BUF_CLR_CMD_WIDTH);
+
+	__RANGECHCK(start_id, BUF_CLR_START_ID_WIDTH);
+	__RANGECHCK(end_id, BUF_CLR_END_ID_WIDTH);
+
+	cmd = (((uint64_t) 1 << BUF_CLR_CMD_LBN) |
+	       (start_id << BUF_CLR_START_ID_LBN) |
+	       (end_id << BUF_CLR_END_ID_LBN));
+
+	/* Falcon requires 128 bit atomic access for this register */
+	falcon_write_qq(offset, cmd, FALCON_ATOMIC_UPD_REG);
+	mmiowb();
+
+	nic->buf_commit_outstanding++;
+	EFHW_TRACE("COMMIT CLEAR out=%d", nic->buf_commit_outstanding);
+}
+
+/*----------------------------------------------------------------------------
+ *
+ * Events low-level register interface
+ *
+ *---------------------------------------------------------------------------*/
+
+static unsigned eventq_sizes[] = {
+	512,
+	EFHW_1K,
+	EFHW_2K,
+	EFHW_4K,
+	EFHW_8K,
+	EFHW_16K,
+	EFHW_32K
+};
+
+#define N_EVENTQ_SIZES  (sizeof(eventq_sizes) / sizeof(eventq_sizes[0]))
+
+static inline void falcon_nic_srm_upd_evq(struct efhw_nic *nic, int evq)
+{
+	/* set up the eventq which will receive events from the SRAM module.
+	 * i.e buffer table updates and clears, TX and RX aperture table
+	 * updates */
+
+	FALCON_LOCK_DECL;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+
+	EFHW_BUILD_ASSERT(SRM_UPD_EVQ_REG_OFST == SRM_UPD_EVQ_REG_KER_OFST);
+
+	EFHW_ASSERT((evq == FALCON_EVQ_KERNEL0) || (evq == FALCON_EVQ_CHAR) ||
+		    (evq == FALCON_EVQ_NONIRQ));
+
+	__DWCHCK(SRM_UPD_EVQ_ID_LBN, SRM_UPD_EVQ_ID_WIDTH);
+	__RANGECHCK(evq, SRM_UPD_EVQ_ID_WIDTH);
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_write_qq(efhw_kva + SRM_UPD_EVQ_REG_OFST,
+			((uint64_t) evq << SRM_UPD_EVQ_ID_LBN),
+			FALCON_ATOMIC_SRPM_UDP_EVQ_REG);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+}
+
+static inline void
+falcon_nic_evq_ptr_tbl(struct efhw_nic *nic,
+		       uint evq,	/* evq id */
+		       uint enable,	/* 1 to enable, 0 to disable */
+		       uint buf_base_id,/* Buffer table base for EVQ */
+		       uint evq_size	/* Number of events */ )
+{
+	FALCON_LOCK_DECL;
+	uint i, val;
+	ulong offset;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+
+	/* size must be one of the various options, otherwise we assert */
+	for (i = 0; i < N_EVENTQ_SIZES; i++) {
+		if (evq_size <= eventq_sizes[i])
+			break;
+	}
+	EFHW_ASSERT(i < N_EVENTQ_SIZES);
+
+	__DWCHCK(EVQ_BUF_BASE_ID_LBN, EVQ_BUF_BASE_ID_WIDTH);
+	__DWCHCK(EVQ_SIZE_LBN, EVQ_SIZE_WIDTH);
+	__DWCHCK(EVQ_EN_LBN, EVQ_EN_WIDTH);
+
+	__RANGECHCK(i, EVQ_SIZE_WIDTH);
+	__RANGECHCK(buf_base_id, EVQ_BUF_BASE_ID_WIDTH);
+	__RANGECHCK(1, EVQ_EN_WIDTH);
+
+	/* if !enable then only evq needs to be correct, although valid
+	 * values need to be passed in for other arguments to prevent
+	 * assertions */
+
+	val = ((i << EVQ_SIZE_LBN) | (buf_base_id << EVQ_BUF_BASE_ID_LBN) |
+	       (enable ? (1 << EVQ_EN_LBN) : 0));
+
+	EFHW_ASSERT(evq < FALCON_EVQ_TBL_NUM);
+
+	offset = EVQ_PTR_TBL_CHAR_OFST;
+	offset += evq * FALCON_REGISTER128;
+
+	EFHW_TRACE("%s: evq %u en=%x:buf=%x:size=%x->%x at %lx",
+		   __FUNCTION__, evq, enable, buf_base_id, evq_size, val,
+		   offset);
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_write_qq(efhw_kva + offset, val, FALCON_ATOMIC_PTR_TBL_REG);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+
+	/* caller must wait for an update done event before writing any more
+	   table entries */
+
+	return;
+}
+
+void
+falcon_nic_evq_ack(struct efhw_nic *nic,
+		   uint evq,	/* evq id */
+		   uint rptr,	/* new read pointer update */
+		   bool wakeup	/* request a wakeup event if ptr's != */
+    )
+{
+	uint val;
+	ulong offset;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+
+	EFHW_BUILD_ASSERT(FALCON_EVQ_CHAR == 4);
+
+	__DWCHCK(EVQ_RPTR_LBN, EVQ_RPTR_WIDTH);
+	__RANGECHCK(rptr, EVQ_RPTR_WIDTH);
+
+	val = (rptr << EVQ_RPTR_LBN);
+
+	EFHW_ASSERT(evq < FALCON_EVQ_TBL_NUM);
+
+	if (evq < FALCON_EVQ_CHAR) {
+		offset = EVQ_RPTR_REG_KER_OFST;
+		offset += evq * FALCON_REGISTER128;
+
+		EFHW_ASSERT(!wakeup);	/* don't try this at home */
+	} else {
+		offset = EVQ_RPTR_REG_OFST + (FALCON_EVQ_CHAR *
+					      FALCON_REGISTER128);
+		offset += (evq - FALCON_EVQ_CHAR) * FALCON_REGISTER128;
+
+		/* nothing to do for interruptless event queues which do
+		 * not want a wakeup */
+		if (evq != FALCON_EVQ_CHAR && !wakeup)
+			return;
+	}
+
+	EFHW_TRACE("%s: %x %x %x->%x", __FUNCTION__, evq, rptr, wakeup, val);
+
+	writel(val, efhw_kva + offset);
+	mmiowb();
+}
+
+/*----------------------------------------------------------------------------
+ *
+ * Helper for evq mapping
+ *
+ * idx = 0 && char   => hw eventq[4]
+ * idx = 0 && net    => hw eventq[0]
+ *   0 < idx < 5     => hw eventq[idx]  (5 is non-interrupting)
+ *
+ *
+ *---------------------------------------------------------------------------*/
+
+int falcon_idx_to_evq(struct efhw_nic *nic, uint idx)
+{
+	EFHW_BUILD_ASSERT(FALCON_EVQ_CHAR == 4);
+	EFHW_ASSERT(idx <= FALCON_EVQ_NONIRQ);
+	return (idx > 0) ? idx : FALCON_EVQ_CHAR;
+}
+
+static inline int falcon_evq_is_interrupting(struct efhw_nic *nic, uint idx)
+{
+	EFHW_BUILD_ASSERT(FALCON_EVQ_CHAR == 4);
+	EFHW_ASSERT(idx <= FALCON_EVQ_NONIRQ);
+
+	/* only the first CHAR driver event queue is interrupting */
+	return (idx == FALCON_EVQ_CHAR);
+}
+
+static inline void
+falcon_drv_ev(struct efhw_nic *nic, uint64_t data, uint qid)
+{
+	FALCON_LOCK_DECL;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+
+	/* send an event from one driver to the other */
+	EFHW_BUILD_ASSERT(DRV_EV_REG_KER_OFST == DRV_EV_REG_OFST);
+	EFHW_BUILD_ASSERT(DRV_EV_DATA_LBN == 0);
+	EFHW_BUILD_ASSERT(DRV_EV_DATA_WIDTH == 64);
+	EFHW_BUILD_ASSERT(DRV_EV_QID_LBN == 64);
+	EFHW_BUILD_ASSERT(DRV_EV_QID_WIDTH == 12);
+
+	FALCON_LOCK_LOCK(nic);
+	falcon_write_qq(efhw_kva + DRV_EV_REG_OFST, data, qid);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+}
+
+_DEBUG_SYM_ void
+falcon_timer_cmd(struct efhw_nic *nic,
+		 uint evq,	/* timer id */
+		 uint mode,	/* mode bits */
+		 uint countdown /* counting value to set */ )
+{
+	FALCON_LOCK_DECL;
+	uint val;
+	ulong offset;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+
+	EFHW_BUILD_ASSERT(TIMER_VAL_LBN == 0);
+
+	__DWCHCK(TIMER_MODE_LBN, TIMER_MODE_WIDTH);
+	__DWCHCK(TIMER_VAL_LBN, TIMER_VAL_WIDTH);
+
+	__RANGECHCK(mode, TIMER_MODE_WIDTH);
+	__RANGECHCK(countdown, TIMER_VAL_WIDTH);
+
+	val = ((mode << TIMER_MODE_LBN) | (countdown << TIMER_VAL_LBN));
+
+	if (evq < FALCON_EVQ_CHAR) {
+		offset = TIMER_CMD_REG_KER_OFST;
+		offset += evq * EFHW_8K;	/* PAGE mapped register */
+	} else {
+		offset = TIMER_TBL_OFST;
+		offset += evq * FALCON_REGISTER128;
+	}
+	EFHW_ASSERT(evq < FALCON_EVQ_TBL_NUM);
+
+	EFHW_TRACE("%s: evq %u mode %x (%s) time %x -> %08x",
+		   __FUNCTION__, evq, mode,
+		   mode == 0 ? "DISABLE" :
+		   mode == 1 ? "IMMED" :
+		   mode == 2 ? (evq < 5 ? "HOLDOFF" : "RX_TRIG") :
+		   "<BAD>", countdown, val);
+
+	/* Falcon requires 128 bit atomic access for this register when
+	 * accessed from the driver. User access to timers is paged mapped
+	 */
+	FALCON_LOCK_LOCK(nic);
+	falcon_write_qq(efhw_kva + offset, val, FALCON_ATOMIC_TIMER_CMD_REG);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+	return;
+}
+
+/*--------------------------------------------------------------------
+ *
+ * Rate pacing - Low level interface
+ *
+ *--------------------------------------------------------------------*/
+void falcon_nic_pace(struct efhw_nic *nic, uint dmaq, uint pace)
+{
+	/* Pace specified in 2^(units of microseconds). This is the minimum
+	   additional delay imposed over and above the IPG.
+
+	   Pacing only available on the virtual interfaces
+	 */
+	FALCON_LOCK_DECL;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+	ulong offset;
+
+	if (pace > 20)
+		pace = 20;	/* maxm supported value */
+
+	__DWCHCK(TX_PACE_LBN, TX_PACE_WIDTH);
+	__RANGECHCK(pace, TX_PACE_WIDTH);
+
+	switch (nic->devtype.variant) {
+	case 'A':
+		EFHW_ASSERT(dmaq >= TX_PACE_TBL_FIRST_QUEUE_A1);
+		offset = TX_PACE_TBL_A1_OFST;
+		offset += (dmaq - TX_PACE_TBL_FIRST_QUEUE_A1) * 16;
+		break;
+	case 'B':
+		/* Would be nice to assert this, but as dmaq is unsigned and
+		 * TX_PACE_TBL_FIRST_QUEUE_B0 is 0, it makes no sense
+		 * EFHW_ASSERT(dmaq >= TX_PACE_TBL_FIRST_QUEUE_B0);
+		 */
+		offset = TX_PACE_TBL_B0_OFST;
+		offset += (dmaq - TX_PACE_TBL_FIRST_QUEUE_B0) * 16;
+		break;
+	default:
+		EFHW_ASSERT(0);
+		offset = 0;
+		break;
+	}
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_write_qq(efhw_kva + offset, pace, FALCON_ATOMIC_PACE_REG);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+
+	EFHW_TRACE("%s: txq %d offset=%lx pace=2^%x",
+		   __FUNCTION__, dmaq, offset, pace);
+}
+
+/*--------------------------------------------------------------------
+ *
+ * Interrupt - Low level interface
+ *
+ *--------------------------------------------------------------------*/
+
+static void falcon_nic_handle_fatal_int(struct efhw_nic *nic)
+{
+	FALCON_LOCK_DECL;
+	efhw_ioaddr_t offset;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+	uint64_t val;
+
+	offset = (efhw_kva + FATAL_INTR_REG_OFST);
+
+	/* Falcon requires 32 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	val = readl(offset);
+	FALCON_LOCK_UNLOCK(nic);
+
+	/* ?? BUG3249 - need to disable illegal address interrupt */
+	/* ?? BUG3114 - need to backport interrupt storm protection code */
+	EFHW_ERR("fatal interrupt: %s%s%s%s%s%s%s%s%s%s%s%s[%" PRIx64 "]",
+		 val & (1 << PCI_BUSERR_INT_CHAR_LBN) ? "PCI-bus-error " : "",
+		 val & (1 << SRAM_OOB_INT_CHAR_LBN) ? "SRAM-oob " : "",
+		 val & (1 << BUFID_OOB_INT_CHAR_LBN) ? "bufid-oob " : "",
+		 val & (1 << MEM_PERR_INT_CHAR_LBN) ? "int-parity " : "",
+		 val & (1 << RBUF_OWN_INT_CHAR_LBN) ? "rx-bufid-own " : "",
+		 val & (1 << TBUF_OWN_INT_CHAR_LBN) ? "tx-bufid-own " : "",
+		 val & (1 << RDESCQ_OWN_INT_CHAR_LBN) ? "rx-desc-own " : "",
+		 val & (1 << TDESCQ_OWN_INT_CHAR_LBN) ? "tx-desc-own " : "",
+		 val & (1 << EVQ_OWN_INT_CHAR_LBN) ? "evq-own " : "",
+		 val & (1 << EVFF_OFLO_INT_CHAR_LBN) ? "evq-fifo " : "",
+		 val & (1 << ILL_ADR_INT_CHAR_LBN) ? "ill-addr " : "",
+		 val & (1 << SRM_PERR_INT_CHAR_LBN) ? "sram-parity " : "", val);
+}
+
+static void falcon_nic_interrupt_hw_enable(struct efhw_nic *nic)
+{
+	FALCON_LOCK_DECL;
+	uint val;
+	efhw_ioaddr_t offset;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+
+	EFHW_BUILD_ASSERT(DRV_INT_EN_CHAR_WIDTH == 1);
+
+	if (nic->flags & NIC_FLAG_NO_INTERRUPT)
+		return;
+
+	offset = (efhw_kva + INT_EN_REG_CHAR_OFST);
+	val = 1 << DRV_INT_EN_CHAR_LBN;
+
+	EFHW_NOTICE("%s: %x -> %x", __FUNCTION__, (int)(offset - efhw_kva),
+		    val);
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_write_qq(offset, val, FALCON_ATOMIC_INT_EN_REG);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+}
+
+static void falcon_nic_interrupt_hw_disable(struct efhw_nic *nic)
+{
+	FALCON_LOCK_DECL;
+	efhw_ioaddr_t offset;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+
+	EFHW_BUILD_ASSERT(SRAM_PERR_INT_KER_WIDTH == 1);
+	EFHW_BUILD_ASSERT(DRV_INT_EN_KER_LBN == 0);
+	EFHW_BUILD_ASSERT(SRAM_PERR_INT_CHAR_WIDTH == 1);
+	EFHW_BUILD_ASSERT(DRV_INT_EN_CHAR_LBN == 0);
+	EFHW_BUILD_ASSERT(SRAM_PERR_INT_KER_LBN == SRAM_PERR_INT_CHAR_LBN);
+	EFHW_BUILD_ASSERT(DRV_INT_EN_KER_LBN == DRV_INT_EN_CHAR_LBN);
+
+	if (nic->flags & NIC_FLAG_NO_INTERRUPT)
+		return;
+
+	offset = (efhw_kva + INT_EN_REG_CHAR_OFST);
+
+	EFHW_NOTICE("%s: %x -> 0", __FUNCTION__, (int)(offset - efhw_kva));
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_write_qq(offset, 0, FALCON_ATOMIC_INT_EN_REG);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+}
+
+#ifndef __ci_ul_driver__
+
+static void falcon_nic_irq_addr_set(struct efhw_nic *nic, dma_addr_t dma_addr)
+{
+	FALCON_LOCK_DECL;
+	efhw_ioaddr_t offset;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+
+	offset = (efhw_kva + INT_ADR_REG_CHAR_OFST);
+
+	EFHW_NOTICE("%s: %x -> " DMA_ADDR_T_FMT, __FUNCTION__,
+		    (int)(offset - efhw_kva), dma_addr);
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_write_qq(offset, dma_addr, FALCON_ATOMIC_INT_ADR_REG);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+}
+
+#endif
+
+
+/*--------------------------------------------------------------------
+ *
+ * RXDP - low level interface
+ *
+ *--------------------------------------------------------------------*/
+
+void
+falcon_nic_set_rx_usr_buf_size(struct efhw_nic *nic, int usr_buf_bytes)
+{
+	FALCON_LOCK_DECL;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+	uint64_t val, val2, usr_buf_size = usr_buf_bytes / 32;
+	int rubs_lbn, rubs_width, roec_lbn;
+
+	EFHW_BUILD_ASSERT(RX_CFG_REG_OFST == RX_CFG_REG_KER_OFST);
+
+	switch (nic->devtype.variant) {
+	default:
+		EFHW_ASSERT(0);
+		/* Fall-through to avoid compiler warnings. */
+	case 'A':
+		rubs_lbn = RX_USR_BUF_SIZE_A1_LBN;
+		rubs_width = RX_USR_BUF_SIZE_A1_WIDTH;
+		roec_lbn = RX_OWNERR_CTL_A1_LBN;
+		break;
+	case 'B':
+		rubs_lbn = RX_USR_BUF_SIZE_B0_LBN;
+		rubs_width = RX_USR_BUF_SIZE_B0_WIDTH;
+		roec_lbn = RX_OWNERR_CTL_B0_LBN;
+		break;
+	}
+
+	__DWCHCK(rubs_lbn, rubs_width);
+	__QWCHCK(roec_lbn, 1);
+	__RANGECHCK(usr_buf_size, rubs_width);
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_read_qq(efhw_kva + RX_CFG_REG_OFST, &val, &val2);
+
+	val &= ~((__FALCON_MASK64(rubs_width)) << rubs_lbn);
+	val |= (usr_buf_size << rubs_lbn);
+
+	/* shouldn't be needed for a production driver */
+	val |= ((uint64_t) 1 << roec_lbn);
+
+	falcon_write_qq(efhw_kva + RX_CFG_REG_OFST, val, val2);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+}
+EXPORT_SYMBOL(falcon_nic_set_rx_usr_buf_size);
+
+void
+falcon_nic_rx_filter_ctl_get(struct efhw_nic *nic, uint32_t *tcp_full,
+			     uint32_t *tcp_wild,
+			     uint32_t *udp_full, uint32_t *udp_wild)
+{
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+	FALCON_LOCK_DECL;
+	uint64_t val;
+
+	FALCON_LOCK_LOCK(nic);
+	falcon_read_q(efhw_kva + RX_FILTER_CTL_REG_OFST, &val);
+	FALCON_LOCK_UNLOCK(nic);
+
+	*tcp_full = (uint32_t)((val >> TCP_FULL_SRCH_LIMIT_LBN) &
+			       (__FALCON_MASK64(TCP_FULL_SRCH_LIMIT_WIDTH)));
+
+	*tcp_wild = (uint32_t)((val >> TCP_WILD_SRCH_LIMIT_LBN) &
+			       (__FALCON_MASK64(TCP_WILD_SRCH_LIMIT_WIDTH)));
+
+	*udp_full = (uint32_t)((val >> UDP_FULL_SRCH_LIMIT_LBN) &
+			       (__FALCON_MASK64(UDP_FULL_SRCH_LIMIT_WIDTH)));
+
+	*udp_wild = (uint32_t)((val >> UDP_WILD_SRCH_LIMIT_LBN) &
+			       (__FALCON_MASK64(UDP_WILD_SRCH_LIMIT_WIDTH)));
+}
+EXPORT_SYMBOL(falcon_nic_rx_filter_ctl_get);
+
+void
+falcon_nic_rx_filter_ctl_set(struct efhw_nic *nic, uint32_t tcp_full,
+			     uint32_t tcp_wild,
+			     uint32_t udp_full, uint32_t udp_wild)
+{
+	uint64_t val, val2;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+	FALCON_LOCK_DECL;
+
+	EFHW_ASSERT(tcp_full < nic->filter_tbl_size);
+	EFHW_ASSERT(tcp_wild < nic->filter_tbl_size);
+	EFHW_ASSERT(udp_full < nic->filter_tbl_size);
+	EFHW_ASSERT(udp_wild < nic->filter_tbl_size);
+
+	/* until we implement a dynamic scaling of search limits we wish to
+	 * maintain the same limits set up by default in the net driver
+	 * when we initialize the char driver */
+	tcp_full_srch_limit = tcp_full;
+	tcp_wild_srch_limit = tcp_wild;
+	udp_full_srch_limit = udp_full;
+	udp_wild_srch_limit = udp_wild;
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_read_qq(efhw_kva + RX_FILTER_CTL_REG_OFST, &val, &val2);
+
+	/* Search limits */
+	val &= ~((__FALCON_MASK64(TCP_FULL_SRCH_LIMIT_WIDTH))
+		 << TCP_FULL_SRCH_LIMIT_LBN);
+
+	val |= ((uint64_t)tcp_full + RX_FILTER_CTL_SRCH_FUDGE_FULL)
+	    << TCP_FULL_SRCH_LIMIT_LBN;
+
+	val &= ~((__FALCON_MASK64(TCP_WILD_SRCH_LIMIT_WIDTH))
+		 << TCP_WILD_SRCH_LIMIT_LBN);
+
+	val |= ((uint64_t)tcp_wild + RX_FILTER_CTL_SRCH_FUDGE_WILD)
+	    << TCP_WILD_SRCH_LIMIT_LBN;
+
+	val &= ~((__FALCON_MASK64(UDP_FULL_SRCH_LIMIT_WIDTH))
+		 << UDP_FULL_SRCH_LIMIT_LBN);
+
+	val |= ((uint64_t)udp_full + RX_FILTER_CTL_SRCH_FUDGE_FULL)
+	    << UDP_FULL_SRCH_LIMIT_LBN;
+
+	val &= ~((__FALCON_MASK64(UDP_WILD_SRCH_LIMIT_WIDTH))
+		 << UDP_WILD_SRCH_LIMIT_LBN);
+
+	val |= ((uint64_t)udp_wild + RX_FILTER_CTL_SRCH_FUDGE_WILD)
+	    << UDP_WILD_SRCH_LIMIT_LBN;
+
+	falcon_write_qq(efhw_kva + RX_FILTER_CTL_REG_OFST, val, val2);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+}
+EXPORT_SYMBOL(falcon_nic_rx_filter_ctl_set);
+
+/*--------------------------------------------------------------------
+ *
+ * TXDP - low level interface
+ *
+ *--------------------------------------------------------------------*/
+
+_DEBUG_SYM_ void falcon_nic_tx_cfg(struct efhw_nic *nic, int unlocked)
+{
+	FALCON_LOCK_DECL;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+	uint64_t val1, val2;
+
+	EFHW_BUILD_ASSERT(TX_CFG_REG_OFST == TX_CFG_REG_KER_OFST);
+	__DWCHCK(TX_OWNERR_CTL_LBN, TX_OWNERR_CTL_WIDTH);
+	__DWCHCK(TX_NON_IP_DROP_DIS_LBN, TX_NON_IP_DROP_DIS_WIDTH);
+
+	FALCON_LOCK_LOCK(nic);
+	falcon_read_qq(efhw_kva + TX_CFG_REG_OFST, &val1, &val2);
+
+	/* Will flag fatal interrupts on owner id errors. This should not be
+	   on for production code because there is otherwise a denial of
+	   serivce attack possible */
+	val1 |= (1 << TX_OWNERR_CTL_LBN);
+
+	/* Setup user queue TCP/UDP only packet security */
+	if (unlocked)
+		val1 |= (1 << TX_NON_IP_DROP_DIS_LBN);
+	else
+		val1 &= ~(1 << TX_NON_IP_DROP_DIS_LBN);
+
+	falcon_write_qq(efhw_kva + TX_CFG_REG_OFST, val1, val2);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+}
+
+/*--------------------------------------------------------------------
+ *
+ * Random thresholds - Low level interface (Would like these to be op
+ * defaults wherever possible)
+ *
+ *--------------------------------------------------------------------*/
+
+static void falcon_nic_pace_cfg(struct efhw_nic *nic)
+{
+	FALCON_LOCK_DECL;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+	unsigned offset = 0;
+	uint64_t val;
+
+	val = 0xa81682;		/* !!!! */
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	switch (nic->devtype.variant) {
+	case 'A':  offset = TX_PACE_REG_A1_OFST;  break;
+	case 'B':  offset = TX_PACE_REG_B0_OFST;  break;
+	default:   EFHW_ASSERT(0);                break;
+	}
+	falcon_write_qq(efhw_kva + offset, val, 0);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+}
+
+/**********************************************************************
+ * Supporting modules. ************************************************
+ **********************************************************************/
+
+/*--------------------------------------------------------------------
+ *
+ * Filter support
+ *
+ *--------------------------------------------------------------------*/
+
+/*! \TODO this table should be per nic */
+struct falcon_cached_ipfilter {
+#if FALCON_FULL_FILTER_CACHE
+	unsigned dmaq;
+	unsigned saddr_le32;
+	unsigned daddr_le32;
+	unsigned sport_le16;
+	unsigned dport_le16;
+	unsigned tcp:1;
+	unsigned full:1;
+	unsigned rss_b0:1;
+	unsigned scat_b0:1;
+#endif
+	unsigned addr_valid:1;
+
+};
+
+
+/* TODO: Dynamically allocate this and store in struct efhw_nic. */
+static struct falcon_cached_ipfilter
+    host_ipfilter_cache[EFHW_MAX_NR_DEVS][FALCON_FILTER_TBL_NUM];
+
+
+static inline void host_ipfilter_cache_init(struct efhw_nic *nic)
+{
+	memset(host_ipfilter_cache[nic->index], 0,
+	       sizeof(host_ipfilter_cache[0][0]) * nic->filter_tbl_size);
+}
+
+static inline int host_ipfilter_cache_active(struct efhw_nic *nic, uint idx)
+{
+	EFHW_ASSERT(nic->index < EFHW_MAX_NR_DEVS);
+	EFHW_ASSERT(idx < nic->filter_tbl_size);
+
+	return (host_ipfilter_cache[nic->index][idx].addr_valid);
+
+}
+
+static inline void host_ipfilter_cache_flush(struct efhw_nic *nic, uint idx)
+{
+	EFHW_ASSERT(nic->index < EFHW_MAX_NR_DEVS);
+	EFHW_ASSERT(idx < nic->filter_tbl_size);
+
+	memset(&host_ipfilter_cache[nic->index][idx], 0,
+	       sizeof(struct falcon_cached_ipfilter));
+	mmiowb();
+}
+
+static inline void
+host_ipfilter_cache_set_addr(struct efhw_nic *nic, uint idx, uint dmaq,
+			     unsigned tcp, unsigned full,
+			     unsigned rss_b0, unsigned scat_b0,
+			     unsigned saddr_le32, unsigned sport_le16,
+			     unsigned daddr_le32, unsigned dport_le16)
+{
+	unsigned nic_i = nic->index;
+
+	EFHW_ASSERT(nic_i < EFHW_MAX_NR_DEVS);
+	EFHW_ASSERT(idx < nic->filter_tbl_size);
+	EFHW_ASSERT(!host_ipfilter_cache[nic_i][idx].addr_valid);
+
+	__RANGECHCK(sport_le16, SRC_TCP_DEST_UDP_1_WIDTH);
+	__RANGECHCK(dport_le16, SRC_TCP_DEST_UDP_1_WIDTH);
+
+#if FALCON_FULL_FILTER_CACHE
+	host_ipfilter_cache[nic_i][idx].dmaq = dmaq;
+	host_ipfilter_cache[nic_i][idx].saddr_le32 = saddr_le32;
+	host_ipfilter_cache[nic_i][idx].daddr_le32 = daddr_le32;
+	host_ipfilter_cache[nic_i][idx].sport_le16 = sport_le16;
+	host_ipfilter_cache[nic_i][idx].dport_le16 = dport_le16;
+	host_ipfilter_cache[nic_i][idx].tcp = tcp;
+	host_ipfilter_cache[nic_i][idx].full = full;
+	host_ipfilter_cache[nic_i][idx].rss_b0 = rss_b0;
+	host_ipfilter_cache[nic_i][idx].scat_b0 = scat_b0;
+#endif
+	host_ipfilter_cache[nic_i][idx].addr_valid = 1;
+	mmiowb();
+}
+
+#if FALCON_VERIFY_FILTERS
+/* Check that all active filters still exist by reading from H/W */
+static void _falcon_nic_ipfilter_sanity(struct efhw_nic *nic)
+{
+	unsigned i;
+	struct falcon_cached_ipfilter *f;
+	uint64_t q0_expect, q1_expect, q0_got, q1_got;
+
+	for (i = 0; i < nic->filter_tbl_size; i++) {
+		f = host_ipfilter_cache[nic->index] + i;
+		if (!f->addr_valid)
+			continue;
+
+		_falcon_nic_ipfilter_build(nic, f->tcp, f->full,
+					   f->rss_b0, f->scat_b0, i, f->dmaq,
+					   f->saddr_le32, f->sport_le16,
+					   f->daddr_le32, f->dport_le16,
+					   &q0_expect, &q1_expect);
+
+		falcon_read_qq(EFHW_KVA(nic) + RX_FILTER_TBL0_OFST +
+			       i * 2 * FALCON_REGISTER128,
+			       &q0_got, &q1_got);
+
+		if ((q0_got != q0_expect) || (q1_got != q1_expect)) {
+			falcon_write_qq(EFHW_KVA(nic) + 0x300,
+					q0_got, q1_got);
+			EFHW_ERR("ERROR: RX-filter[%d][%d] was "
+				 "%"PRIx64":%" PRIx64" expected "
+				 "%"PRIx64":%"PRIx64,
+				 nic->index, i, q0_got, q1_got,
+				 q0_expect, q1_expect);
+		}
+	}
+}
+#endif /* FALCON_VERIFY_FILTERS */
+
+#if FALCON_FULL_FILTER_CACHE
+static inline int
+host_ipfilter_cache_check_not(uint nic, uint idx, int tcp, int full,
+			      unsigned saddr_le32, unsigned sport_le16,
+			      unsigned daddr_le32, unsigned dport_le16)
+{
+	return ((host_ipfilter_cache[nic][idx].saddr_le32 != saddr_le32) ||
+		(host_ipfilter_cache[nic][idx].daddr_le32 != daddr_le32) ||
+		(host_ipfilter_cache[nic][idx].sport_le16 != sport_le16) ||
+		(host_ipfilter_cache[nic][idx].dport_le16 != dport_le16) ||
+		(host_ipfilter_cache[nic][idx].tcp != tcp) ||
+		(host_ipfilter_cache[nic][idx].full != full));
+}
+#endif
+
+#define host_ipfilter_cache_saddr_le32(nic, idx) \
+		host_ipfilter_cache[nic][idx].saddr_le32
+#define host_ipfilter_cache_daddr_le32(nic, idx) \
+		host_ipfilter_cache[nic][idx].daddr_le32
+#define host_ipfilter_cache_sport_le16(nic, idx) \
+		host_ipfilter_cache[nic][idx].sport_le16
+#define host_ipfilter_cache_dport_le16(nic, idx) \
+		host_ipfilter_cache[nic][idx].dport_le16
+#define host_ipfilter_cache_tcp(nic, idx) \
+		host_ipfilter_cache[nic][idx].tcp
+#define host_ipfilter_cache_full(nic, idx) \
+		host_ipfilter_cache[nic][idx].full
+
+/**********************************************************************
+ * Implementation of the HAL. ********************************************
+ **********************************************************************/
+
+/*----------------------------------------------------------------------------
+ *
+ * Initialisation and configuration discovery
+ *
+ *---------------------------------------------------------------------------*/
+
+#ifdef __ci_ul_driver__
+
+static int falcon_nic_init_irq_channel(struct efhw_nic *nic, int enable)
+{
+	EFHW_ERR("%s: not implemented for ul driver", __FUNCTION__);
+	return -EOPNOTSUPP;
+}
+
+#else
+
+static int falcon_nic_init_irq_channel(struct efhw_nic *nic, int enable)
+{
+	/* create a buffer for the irq channel */
+	int rc;
+
+	if (enable) {
+		rc = efhw_iopage_alloc(nic, &nic->irq_iobuff);
+		if (rc < 0)
+			return rc;
+
+		falcon_nic_irq_addr_set(nic,
+				efhw_iopage_dma_addr(&nic->irq_iobuff));
+	} else {
+		if (efhw_iopage_is_valid(&nic->irq_iobuff))
+			efhw_iopage_free(nic, &nic->irq_iobuff);
+
+		efhw_iopage_mark_invalid(&nic->irq_iobuff);
+		falcon_nic_irq_addr_set(nic, 0);
+	}
+
+	EFHW_TRACE("%s: " ci_dma_addr_fmt " %sable", __FUNCTION__,
+		   efhw_iopage_dma_addr(&nic->irq_iobuff), enable ?
+			"en" : "dis");
+
+	return 0;
+}
+
+#endif
+
+static void falcon_nic_close_hardware(struct efhw_nic *nic)
+{
+	/* check we are in possession of some hardware */
+	if (!efhw_nic_have_hw(nic))
+		return;
+
+	falcon_nic_init_irq_channel(nic, 0);
+
+	EFHW_NOTICE("%s:", __FUNCTION__);
+}
+
+#ifdef __ci_ul_driver__
+extern
+#else
+static
+#endif
+int falcon_nic_get_mac_config(struct efhw_nic *nic)
+{
+	efhw_ioaddr_t efhw_kva = nic->bar_ioaddr;
+	int is_mac_type_1g;
+	uint32_t strap, altera;
+	uint64_t rx_cfg, r;
+
+	altera = readl(efhw_kva + ALTERA_BUILD_REG_OFST);
+	strap = readl(efhw_kva + STRAP_REG_KER_OFST) & 0x7;
+
+	switch (nic->devtype.variant) {
+	case 'A':
+		if ((altera & 0x0fff0000) == 0x1130000) {
+			strap = 2;	/* FPGA - PCI-X 2G */
+		} else if ((altera & 0x00ff0000) == 0x140000) {
+			/* should be 114 */
+			strap = 4;	/* FPGA - PCI-X 4G */
+		} else if (strap < 2 || strap > 5) {
+			EFHW_ERR("Invalid strap option %d altera_buid_ver=%x",
+				 strap, altera);
+			return -EINVAL;
+		}
+		is_mac_type_1g = (0 != (strap & 2));
+		break;
+	case 'B':
+		/* Runtime check that the hardware and software agree about
+		 * the size of the RXFIFO. Write binary 11 across the left
+		 * most bit, and assert we get 1 back.
+		 */
+		r = 1LL << RX_TOEP_TCP_SUPPRESS_B0_LBN;
+		r |= (r << 1);
+
+		/* Save the original value */
+		falcon_read_q(efhw_kva + RX_CFG_REG_OFST, &rx_cfg);
+
+		/* Write and ready the dummy value */
+		falcon_write_qq(efhw_kva + RX_CFG_REG_OFST, r, 0);
+		falcon_read_q(efhw_kva + RX_CFG_REG_OFST, &r);
+
+		/* Restore the original value */
+		falcon_write_qq(efhw_kva + RX_CFG_REG_OFST, rx_cfg, 0);
+
+		if (r != (1LL << RX_TOEP_TCP_SUPPRESS_B0_LBN)) {
+			EFHW_ERR("The FPGA build (%x) RXFIFO size does not "
+				 "match the software", altera);
+			return -EINVAL;
+		}
+		is_mac_type_1g = (0 != (strap & 2));
+#if FALCON_MAC_SET_TYPE_BY_SPEED
+		/* Check the selected strap pins against the MAC speed -
+		 * and adjust if necessary.
+		 */
+		{
+			int speed;
+			speed = readl(efhw_kva + MAC0_CTRL_REG_OFST) & 0x3;
+			is_mac_type_1g = (speed <= 2);
+		}
+#endif
+		break;
+	default:
+		EFHW_ASSERT(0);
+		is_mac_type_1g = 0;
+		break;
+	}
+
+	nic->fpga_version = altera;
+
+	/* We can now set the MAC type correctly based on the strap pins. */
+	if (is_mac_type_1g) {
+		nic->flags &= ~NIC_FLAG_10G;
+	} else {
+		/* strap & 4 must be set according to checks above */
+		nic->flags |= NIC_FLAG_10G;
+	}
+	EFHW_NOTICE("Board has %s MAC: strap=%d",
+		    0 != (nic->flags & NIC_FLAG_10G) ? "10G" : "1G", strap);
+	return 0;
+}
+
+static int
+falcon_nic_init_hardware(struct efhw_nic *nic,
+			 struct efhw_ev_handler *ev_handlers,
+			 const uint8_t *mac_addr)
+{
+	int rc;
+
+	/* header sanity checks */
+	FALCON_ASSERT_VALID();
+
+	rc = falcon_nic_get_mac_config(nic);
+	if (rc < 0)
+		return rc;
+
+	/* Initialise supporting modules */
+	falcon_nic_ipfilter_ctor(nic);
+
+#if FALCON_USE_SHADOW_BUFFER_TABLE
+	CI_ZERO_ARRAY(_falcon_buffer_table, FALCON_BUFFER_TBL_NUM);
+#endif
+
+	/* Initialise the top level hardware blocks */
+	memcpy(nic->mac_addr, mac_addr, ETH_ALEN);
+
+	EFHW_TRACE("%s:", __FUNCTION__);
+
+	/* nic.c:efhw_nic_init marks all the interrupt units as unused.
+
+	   ?? TODO we should be able to request the non-interrupting event
+	   queue and the net driver's (for a net driver that is using libefhw)
+	   additional RSS queues here.
+
+	   Result would be that that net driver could call
+	   nic.c:efhw_nic_allocate_common_hardware_resources() and that the
+	   IFDEF FALCON's can be removed from
+	   nic.c:efhw_nic_allocate_common_hardware_resources()
+	 */
+	nic->irq_unit[0] = INT_EN_REG_CHAR_OFST;
+
+	/*****************************************************************
+	 * The rest of this function deals with initialization of the NICs
+	 * hardware (as opposed to the initialization of the
+	 * struct efhw_nic data structure */
+
+	/* char driver grabs SRM events onto the non interrupting
+	 * event queue */
+	falcon_nic_srm_upd_evq(nic, FALCON_EVQ_NONIRQ);
+
+	/* RXDP tweaks */
+
+	/* ?? bug2396 rx_cfg should be ok so long as the net driver
+	 * always pushes buffers big enough for the link MTU */
+
+	/* set the RX buffer cutoff size to be the same as PAGE_SIZE.
+	 * Use this value when we think that there will be a lot of
+	 * jumbo frames.
+	 *
+	 * The default value 1600 is useful when packets are small,
+	 * but would means that jumbo frame RX queues would need more
+	 * descriptors pushing */
+	falcon_nic_set_rx_usr_buf_size(nic, FALCON_RX_USR_BUF_SIZE);
+
+	/* TXDP tweaks */
+	/* ?? bug2396 looks ok */
+	falcon_nic_tx_cfg(nic, /*unlocked(for non-UDP/TCP)= */ 0);
+	falcon_nic_pace_cfg(nic);
+
+	/* ?? bug2396
+	 * netdriver must load first or else must RMW this register */
+	falcon_nic_rx_filter_ctl_set(nic, RX_FILTER_CTL_SRCH_LIMIT_TCP_FULL,
+				     RX_FILTER_CTL_SRCH_LIMIT_TCP_WILD,
+				     RX_FILTER_CTL_SRCH_LIMIT_UDP_FULL,
+				     RX_FILTER_CTL_SRCH_LIMIT_UDP_WILD);
+
+	if (!(nic->flags & NIC_FLAG_NO_INTERRUPT)) {
+		rc = efhw_keventq_ctor(nic, FALCON_EVQ_CHAR, &nic->evq[0],
+				       ev_handlers);
+		if (rc < 0) {
+			EFHW_ERR("%s: efhw_keventq_ctor() failed (%d) evq=%d",
+				 __FUNCTION__, rc, FALCON_EVQ_CHAR);
+			return rc;
+		}
+	}
+	rc = efhw_keventq_ctor(nic, FALCON_EVQ_NONIRQ,
+			       &nic->evq[FALCON_EVQ_NONIRQ], NULL);
+	if (rc < 0) {
+		EFHW_ERR("%s: efhw_keventq_ctor() failed (%d) evq=%d",
+			 __FUNCTION__, rc, FALCON_EVQ_NONIRQ);
+		return rc;
+	}
+
+	/* allocate IRQ channel */
+	rc = falcon_nic_init_irq_channel(nic, 1);
+	/* ignore failure at user-level for eftest */
+	if ((rc < 0) && !(nic->options & NIC_OPT_EFTEST))
+		return rc;
+
+	return 0;
+}
+
+/*--------------------------------------------------------------------
+ *
+ * Interrupt
+ *
+ *--------------------------------------------------------------------*/
+
+static void
+falcon_nic_interrupt_enable(struct efhw_nic *nic, unsigned idx)
+{
+	int evq;
+
+	if (idx || (nic->flags & NIC_FLAG_NO_INTERRUPT))
+		return;
+
+	/* Enable driver interrupts */
+	EFHW_NOTICE("%s: enable master interrupt", __FUNCTION__);
+	falcon_nic_interrupt_hw_enable(nic);
+
+	/* An interrupting eventq must start of day ack its read pointer */
+	evq = falcon_idx_to_evq(nic, idx);
+
+	if (falcon_evq_is_interrupting(nic, evq)) {
+		struct efhw_keventq *q = &nic->evq[idx];
+		unsigned rdptr =
+		    EFHW_EVENT_OFFSET(q, q, 1) / sizeof(efhw_event_t);
+		falcon_nic_evq_ack(nic, evq, rdptr, false);
+		EFHW_NOTICE("%s: ACK evq[%d]:%x", __FUNCTION__, evq, rdptr);
+	}
+}
+
+static void falcon_nic_interrupt_disable(struct efhw_nic *nic, uint idx)
+{
+	/* NB. No need to check for NIC_FLAG_NO_INTERRUPT, as
+	 ** falcon_nic_interrupt_hw_disable() will do it. */
+	if (idx)
+		return;
+	falcon_nic_interrupt_hw_disable(nic);
+}
+
+static void
+falcon_nic_set_interrupt_moderation(struct efhw_nic *nic, uint idx,
+				    uint32_t val)
+{
+	falcon_timer_cmd(nic, falcon_idx_to_evq(nic, idx),
+			 TIMER_MODE_INT_HLDOFF, val / 5);
+}
+
+static inline void legacy_irq_ack(struct efhw_nic *nic)
+{
+	EFHW_ASSERT(!(nic->flags & NIC_FLAG_NO_INTERRUPT));
+
+	if (!(nic->flags & NIC_FLAG_MSI)) {
+		writel(1, EFHW_KVA(nic) + INT_ACK_REG_CHAR_A1_OFST);
+		mmiowb();
+		/* ?? FIXME: We should be doing a read here to ensure IRQ is
+		 * thoroughly acked before we return from ISR. */
+	}
+}
+
+static int falcon_nic_interrupt(struct efhw_nic *nic)
+{
+	volatile uint32_t *syserr_ptr =
+	    (uint32_t *) efhw_iopage_ptr(&nic->irq_iobuff);
+	int handled = 0;
+	int done_ack = 0;
+
+	EFHW_ASSERT(!(nic->flags & NIC_FLAG_NO_INTERRUPT));
+	EFHW_ASSERT(syserr_ptr);
+
+	/* FIFO fill level interrupt - just log it. */
+	if (unlikely(*(syserr_ptr + (DW0_OFST / 4)))) {
+		EFHW_WARN("%s: *** FIFO *** %x", __FUNCTION__,
+			  *(syserr_ptr + (DW0_OFST / 4)));
+		*(syserr_ptr + (DW0_OFST / 4)) = 0;
+		handled++;
+	}
+
+	/* Fatal interrupts. */
+	if (unlikely(*(syserr_ptr + (DW2_OFST / 4)))) {
+		*(syserr_ptr + (DW2_OFST / 4)) = 0;
+		falcon_nic_handle_fatal_int(nic);
+		handled++;
+	}
+
+	/* Event queue interrupt.  For legacy interrupts we have to check
+	 * that the interrupt is for us, because it could be shared. */
+	if (*(syserr_ptr + (DW1_OFST / 4))) {
+		*(syserr_ptr + (DW1_OFST / 4)) = 0;
+		/* ACK must come before callback to handler fn. */
+		legacy_irq_ack(nic);
+		done_ack = 1;
+		handled++;
+		if (nic->irq_handler)
+			nic->irq_handler(nic, 0);
+	}
+
+	if (unlikely(!done_ack)) {
+		if (!handled)
+			/* Shared interrupt line (hopefully). */
+			return 0;
+		legacy_irq_ack(nic);
+	}
+
+	EFHW_TRACE("%s: handled %d", __FUNCTION__, handled);
+	return 1;
+}
+
+/*--------------------------------------------------------------------
+ *
+ * Event Management - and SW event posting
+ *
+ *--------------------------------------------------------------------*/
+
+static void
+falcon_nic_event_queue_enable(struct efhw_nic *nic, uint evq, uint evq_size,
+			      dma_addr_t q_base_addr,	/* not used */
+			      uint buf_base_id)
+{
+	EFHW_ASSERT(nic);
+
+	/*!\ TODO we can be more efficient if we know whether or not there
+	 * is a timer attached */
+	falcon_timer_cmd(nic, evq, 0 /* disable */ , 0);
+
+	falcon_nic_evq_ptr_tbl(nic, evq, 1, buf_base_id, evq_size);
+	EFHW_TRACE("%s: enable evq %u size %u", __FUNCTION__, evq, evq_size);
+}
+
+static void
+falcon_nic_event_queue_disable(struct efhw_nic *nic, uint evq, int timer_only)
+{
+	EFHW_ASSERT(nic);
+
+	/*!\ TODO we can be more efficient if we know whether or not there
+	 * is a timer attached */
+	falcon_timer_cmd(nic, evq, 0 /* disable */ , 0);
+
+	if (!timer_only)
+		falcon_nic_evq_ptr_tbl(nic, evq, 0, 0, 0);
+	EFHW_TRACE("%s: disenable evq %u", __FUNCTION__, evq);
+}
+
+static void
+falcon_nic_wakeup_request(struct efhw_nic *nic, dma_addr_t q_base_addr,
+			  int next_i, int evq)
+{
+	EFHW_ASSERT(evq > FALCON_EVQ_CHAR);
+	falcon_nic_evq_ack(nic, evq, next_i, true);
+	EFHW_TRACE("%s: evq %d next_i %d", __FUNCTION__, evq, next_i);
+}
+
+static void falcon_nic_sw_event(struct efhw_nic *nic, int data, int evq)
+{
+	uint64_t ev_data = data;
+
+	ev_data &= ~FALCON_EVENT_CODE_MASK;
+	ev_data |= FALCON_EVENT_CODE_SW;
+
+	falcon_drv_ev(nic, ev_data, evq);
+	EFHW_NOTICE("%s: evq[%d]->%x", __FUNCTION__, evq, data);
+}
+
+/*--------------------------------------------------------------------
+ *
+ * Filter support - TODO vary the depth of the search
+ *
+ *--------------------------------------------------------------------*/
+
+void
+falcon_nic_ipfilter_ctor(struct efhw_nic *nic)
+{
+	if (nic->devtype.variant == 'B' && nic->fpga_version)
+		nic->filter_tbl_size = 8 * 1024;
+	else
+		nic->filter_tbl_size = 16 * 1024;
+
+	host_ipfilter_cache_init(nic);
+}
+
+
+static int
+falcon_nic_ipfilter_set(struct efhw_nic *nic, int type, int *_filter_idx,
+			int dmaq,
+			unsigned saddr_be32, unsigned sport_be16,
+			unsigned daddr_be32, unsigned dport_be16)
+{
+	FALCON_LOCK_DECL;
+	int tcp;
+	int full;
+	int rss_b0;
+	int scat_b0;
+	int key, hash1, hash2, idx = -1;
+	int k;
+	int rc = 0;
+	unsigned max_srch = -1;
+
+	/* oh joy of joys .. maybe one day we'll optimise  */
+	unsigned int saddr = ntohl(saddr_be32);
+	unsigned int daddr = ntohl(daddr_be32);
+	unsigned int sport = ntohs(sport_be16);
+	unsigned int dport = ntohs(dport_be16);
+
+	__RANGECHCK(sport, SRC_TCP_DEST_UDP_1_WIDTH);
+	__RANGECHCK(dport, SRC_TCP_DEST_UDP_1_WIDTH);
+
+	tcp = ((type & EFHW_IP_FILTER_TYPE_TCP_MASK) != 0) ? 1 : 0;
+	full = ((type & EFHW_IP_FILTER_TYPE_FULL_MASK) != 0) ? 1 : 0;
+	rss_b0 = ((type & EFHW_IP_FILTER_TYPE_RSS_B0_MASK) != 0) ? 1 : 0;
+	scat_b0 = ((type & EFHW_IP_FILTER_TYPE_NOSCAT_B0_MASK) != 0) ? 0 : 1;
+	if (tcp && full)
+		max_srch = tcp_full_srch_limit;
+	else if (tcp && !full)
+		max_srch = tcp_wild_srch_limit;
+	else if (!tcp && full)
+		max_srch = udp_full_srch_limit;
+	else if (!tcp && !full)
+		max_srch = udp_wild_srch_limit;
+
+	EFHW_TRACE("%s: %x tcp %d full %d max_srch=%d",
+		   __FUNCTION__, type, tcp, full, max_srch);
+
+	/* The second hash function is simply
+	 * h2(key) = 13 LSB of (key * 2 -  1)
+	 * And the index(k), or the filter table address for kth search is
+	 * index(k) = 13 LSB of (h1(key) + k * h2(key))
+	 */
+	key = falcon_hash_get_key(saddr, sport, daddr, dport, tcp, full);
+	hash1 = falcon_hash_function1(key, nic->filter_tbl_size);
+	hash2 = falcon_hash_function2(key, nic->filter_tbl_size);
+
+	/* Avoid race to claim a filter entry */
+	FALCON_LOCK_LOCK(nic);
+
+	for (k = 0; (unsigned)k < max_srch; k++) {
+		idx = falcon_hash_iterator(hash1, hash2, k,
+					   nic->filter_tbl_size);
+
+		EFHW_TRACE("ipfilter_set[%d:%d:%d]: src=%x:%d dest=%x:%d %s",
+			   *_filter_idx, idx, k,
+			   saddr, sport, daddr, dport,
+			   host_ipfilter_cache_active(nic, idx) ?
+			   "Active" : "Clear");
+
+		if (!host_ipfilter_cache_active(nic, idx))
+			break;
+
+#if FALCON_FULL_FILTER_CACHE
+		/* Check that we are not duplicating the filter */
+		if (!host_ipfilter_cache_check_not(nic->index, idx, tcp, full,
+						   saddr, sport, daddr,
+						   dport)) {
+			EFHW_WARN("%s: ERROR: duplicate filter (disabling "
+				  "interrupts)", __FUNCTION__);
+			FALCON_LOCK_UNLOCK(nic);
+			falcon_nic_interrupt_hw_disable(nic);
+			return -EINVAL;
+		}
+#endif
+
+	}
+	if (k == max_srch) {
+		rc = -EADDRINUSE;
+		idx = -1;
+		goto fail1;
+	}
+
+	EFHW_ASSERT(idx < (int)nic->filter_tbl_size);
+
+	host_ipfilter_cache_set_addr(nic, idx, dmaq, tcp, full, rss_b0,
+				     scat_b0, saddr, sport, daddr, dport);
+
+	_falcon_nic_ipfilter_set(nic, tcp, full, rss_b0,
+				 scat_b0, idx, dmaq,
+				 saddr, sport, daddr, dport);
+
+	*_filter_idx = idx;
+
+	EFHW_TRACE("%s: filter %x rxq %d src " NIPQUAD_FMT
+		   ":%d dest " NIPQUAD_FMT ":%d set in %d",
+		   __FUNCTION__, idx, dmaq,
+		   NIPQUAD(&saddr), sport, NIPQUAD(&daddr), dport, k);
+
+fail1:
+	FALCON_LOCK_UNLOCK(nic);
+	return rc;
+}
+
+static void
+falcon_nic_ipfilter_attach(struct efhw_nic *nic, int filter_idx, int dmaq_idx)
+{
+	/* Intentionally empty - Falcon attaches and sets the filter
+	 * in filter_set */
+	EFHW_TRACE("%s: attach filter %x with rxq %d - ignored",
+		   __FUNCTION__, filter_idx, dmaq_idx);
+}
+
+static void falcon_nic_ipfilter_detach(struct efhw_nic *nic, int filter_idx)
+{
+	/* Intentionally empty - Falcon attaches and sets the filter
+	 * in filter_clear */
+	EFHW_TRACE("%s: detach filter %x from rxq - ignored",
+		   __FUNCTION__, filter_idx);
+}
+
+static void falcon_nic_ipfilter_clear(struct efhw_nic *nic, int filter_idx)
+{
+	FALCON_LOCK_DECL;
+
+	EFHW_TRACE("%s: filter %x", __FUNCTION__, filter_idx);
+
+	/* In case the filter has already been freed */
+	if (filter_idx == -1)
+		return;
+
+	FALCON_LOCK_LOCK(nic);
+
+	/* if we flush a chained hash then all we need to do is zero it out */
+	host_ipfilter_cache_flush(nic, filter_idx);
+	_falcon_nic_ipfilter_clear(nic, filter_idx);
+
+	FALCON_LOCK_UNLOCK(nic);
+	return;
+}
+
+/*--------------------------------------------------------------------
+ *
+ * Buffer table - helpers
+ *
+ *--------------------------------------------------------------------*/
+
+#define FALCON_LAZY_COMMIT_HWM (FALCON_BUFFER_UPD_MAX - 16)
+
+/* Note re.:
+ *  falcon_nic_buffer_table_lazy_commit(struct efhw_nic *nic)
+ *  falcon_nic_buffer_table_update_poll(struct efhw_nic *nic)
+ *  falcon_nic_buffer_table_confirm(struct efhw_nic *nic)
+ * -- these are no-ops in the user-level driver because it would need to
+ * coordinate with the real driver on the number of outstanding commits.
+ *
+ * An exception is made for eftest apps, which manage the hardware without
+ * using the char driver.
+ */
+
+static inline void falcon_nic_buffer_table_lazy_commit(struct efhw_nic *nic)
+{
+#if defined(__ci_ul_driver__)
+	if (!(nic->options & NIC_OPT_EFTEST))
+		return;
+#endif
+
+	/* Do nothing if operating in synchronous mode. */
+	if (!nic->irq_handler)
+		return;
+}
+
+static inline void falcon_nic_buffer_table_update_poll(struct efhw_nic *nic)
+{
+	FALCON_LOCK_DECL;
+	int count = 0, rc = 0;
+
+#if defined(__ci_ul_driver__)
+	if (!(nic->options & NIC_OPT_EFTEST))
+		return;
+#endif
+
+	/* We can be called here early days */
+	if (!nic->irq_handler)
+		return;
+
+	/* If we need to gather buffer update events then poll the
+	   non-interrupting event queue */
+
+	/* For each _buffer_table_commit there will be an update done
+	   event. We don't keep track of how many buffers each commit has
+	   committed, just make sure that all the expected events have been
+	   gathered */
+	FALCON_LOCK_LOCK(nic);
+
+	EFHW_TRACE("%s: %d", __FUNCTION__, nic->buf_commit_outstanding);
+
+	while (nic->buf_commit_outstanding > 0) {
+		/* we're not expecting to handle any events that require
+		 * upcalls into the core driver */
+		struct efhw_ev_handler handler;
+		memset(&handler, 0, sizeof(handler));
+		nic->evq[FALCON_EVQ_NONIRQ].ev_handlers = &handler;
+		rc = efhw_keventq_poll(nic, &nic->evq[FALCON_EVQ_NONIRQ]);
+		nic->evq[FALCON_EVQ_NONIRQ].ev_handlers = NULL;
+
+		if (rc < 0) {
+			EFHW_ERR("%s: poll ERROR (%d:%d) ***** ",
+				 __FUNCTION__, rc,
+				 nic->buf_commit_outstanding);
+			goto out;
+		}
+
+		FALCON_LOCK_UNLOCK(nic);
+
+		if (count++)
+			udelay(1);
+
+		if (count > 1000) {
+			EFHW_WARN("%s: poll Timeout ***** (%d)", __FUNCTION__,
+				  nic->buf_commit_outstanding);
+			nic->buf_commit_outstanding = 0;
+			return;
+		}
+		FALCON_LOCK_LOCK(nic);
+	}
+
+out:
+	FALCON_LOCK_UNLOCK(nic);
+	return;
+}
+
+void falcon_nic_buffer_table_confirm(struct efhw_nic *nic)
+{
+	/* confirm buffer table updates - should be used for items where
+	   loss of data would be unacceptable. E.g for the buffers that back
+	   an event or DMA queue */
+	FALCON_LOCK_DECL;
+
+#if defined(__ci_ul_driver__)
+	if (!(nic->options & NIC_OPT_EFTEST))
+		return;
+#endif
+
+	/* Do nothing if operating in synchronous mode. */
+	if (!nic->irq_handler)
+		return;
+
+	FALCON_LOCK_LOCK(nic);
+
+	_falcon_nic_buffer_table_commit(nic);
+
+	FALCON_LOCK_UNLOCK(nic);
+
+	falcon_nic_buffer_table_update_poll(nic);
+}
+
+/*--------------------------------------------------------------------
+ *
+ * Buffer table - API
+ *
+ *--------------------------------------------------------------------*/
+
+static void
+falcon_nic_buffer_table_clear(struct efhw_nic *nic, int buffer_id, int num)
+{
+	FALCON_LOCK_DECL;
+	FALCON_LOCK_LOCK(nic);
+	_falcon_nic_buffer_table_clear(nic, buffer_id, num);
+	FALCON_LOCK_UNLOCK(nic);
+}
+
+static void
+falcon_nic_buffer_table_set(struct efhw_nic *nic, dma_addr_t dma_addr,
+			    uint bufsz, uint region,
+			    int own_id, int buffer_id)
+{
+	FALCON_LOCK_DECL;
+
+	EFHW_ASSERT(region < FALCON_REGION_NUM);
+
+	EFHW_ASSERT((bufsz == EFHW_4K) ||
+		    (bufsz == EFHW_8K && FALCON_BUFFER_TABLE_FULL_MODE));
+
+	falcon_nic_buffer_table_update_poll(nic);
+
+	FALCON_LOCK_LOCK(nic);
+
+	_falcon_nic_buffer_table_set(nic, dma_addr, bufsz, region, own_id,
+				     buffer_id);
+
+	falcon_nic_buffer_table_lazy_commit(nic);
+
+	FALCON_LOCK_UNLOCK(nic);
+}
+
+void
+falcon_nic_buffer_table_set_n(struct efhw_nic *nic, int buffer_id,
+			      dma_addr_t dma_addr, uint bufsz, uint region,
+			      int n_pages, int own_id)
+{
+	/* used to set up a contiguous range of buffers */
+	FALCON_LOCK_DECL;
+
+	EFHW_ASSERT(region < FALCON_REGION_NUM);
+
+	EFHW_ASSERT((bufsz == EFHW_4K) ||
+		    (bufsz == EFHW_8K && FALCON_BUFFER_TABLE_FULL_MODE));
+
+	while (n_pages--) {
+
+		falcon_nic_buffer_table_update_poll(nic);
+
+		FALCON_LOCK_LOCK(nic);
+
+		_falcon_nic_buffer_table_set(nic, dma_addr, bufsz, region,
+					     own_id, buffer_id++);
+
+		falcon_nic_buffer_table_lazy_commit(nic);
+
+		FALCON_LOCK_UNLOCK(nic);
+
+		dma_addr += bufsz;
+	}
+}
+
+/*--------------------------------------------------------------------
+ *
+ * DMA Queues - mid level API
+ *
+ *--------------------------------------------------------------------*/
+
+#if BUG5302_WORKAROUND
+
+/* Tx queues can get stuck if the software write pointer is set to an index
+ * beyond the configured size of the queue, such that they will not flush.
+ * This code can be run before attempting a flush; it will detect the bogus
+ * value and reset it.  This fixes most instances of this problem, although
+ * sometimes it does not work, or we may not detect it in the first place,
+ * if the out-of-range value was replaced by an in-range value earlier.
+ * (In those cases we have to apply a bigger hammer later, if we see that
+ * the queue is still not flushing.)
+ */
+static void
+falcon_check_for_bogus_tx_dma_wptr(struct efhw_nic *nic, uint dmaq)
+{
+	FALCON_LOCK_DECL;
+	uint64_t val_low64, val_high64;
+	uint64_t size, hwptr, swptr, val;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+	ulong offset = falcon_dma_tx_q_offset(nic, dmaq);
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_read_qq(efhw_kva + offset, &val_low64, &val_high64);
+	FALCON_LOCK_UNLOCK(nic);
+
+	size = (val_low64 >> TX_DESCQ_SIZE_LBN)
+	    & __FALCON_MASK64(TX_DESCQ_SIZE_WIDTH);
+	size = (1 << size) * 512;
+	hwptr = (val_high64 >> __DW3(TX_DESCQ_HW_RPTR_LBN))
+	    & __FALCON_MASK64(TX_DESCQ_HW_RPTR_WIDTH);
+	swptr = (val_low64 >> TX_DESCQ_SW_WPTR_LBN)
+	    & __FALCON_MASK64(__LW2(TX_DESCQ_SW_WPTR_LBN));
+	val = (val_high64)
+	    &
+	    __FALCON_MASK64(__DW3
+			    (TX_DESCQ_SW_WPTR_LBN + TX_DESCQ_SW_WPTR_WIDTH));
+	val = val << __LW2(TX_DESCQ_SW_WPTR_LBN);
+	swptr = swptr | val;
+
+	if (swptr >= size) {
+		EFHW_WARN("Resetting bad write pointer for TXQ[%d]", dmaq);
+		writel((uint32_t) ((hwptr + 0) & (size - 1)),
+		       efhw_kva + falcon_tx_dma_page_addr(dmaq) + 12);
+		mmiowb();
+	}
+}
+
+/* Here's that "bigger hammer": we reset all the pointers (hardware read,
+ * hardware descriptor cache read, software write) to zero.
+ */
+void falcon_clobber_tx_dma_ptrs(struct efhw_nic *nic, uint dmaq)
+{
+	FALCON_LOCK_DECL;
+	uint64_t val_low64, val_high64;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+	ulong offset = falcon_dma_tx_q_offset(nic, dmaq);
+
+	EFHW_WARN("Recovering stuck TXQ[%d]", dmaq);
+	FALCON_LOCK_LOCK(nic);
+	falcon_read_qq(efhw_kva + offset, &val_low64, &val_high64);
+	val_high64 &= ~(__FALCON_MASK64(TX_DESCQ_HW_RPTR_WIDTH)
+			<< __DW3(TX_DESCQ_HW_RPTR_LBN));
+	val_high64 &= ~(__FALCON_MASK64(TX_DC_HW_RPTR_WIDTH)
+			<< __DW3(TX_DC_HW_RPTR_LBN));
+	falcon_write_qq(efhw_kva + offset, val_low64, val_high64);
+	mmiowb();
+	writel(0, efhw_kva + falcon_tx_dma_page_addr(dmaq) + 12);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+}
+
+#endif
+
+static inline int
+__falcon_really_flush_tx_dma_channel(struct efhw_nic *nic, uint dmaq)
+{
+	FALCON_LOCK_DECL;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+	uint val;
+
+	EFHW_BUILD_ASSERT(TX_FLUSH_DESCQ_REG_KER_OFST ==
+			  TX_FLUSH_DESCQ_REG_OFST);
+
+	__DWCHCK(TX_FLUSH_DESCQ_CMD_LBN, TX_FLUSH_DESCQ_CMD_WIDTH);
+	__DWCHCK(TX_FLUSH_DESCQ_LBN, TX_FLUSH_DESCQ_WIDTH);
+	__RANGECHCK(dmaq, TX_FLUSH_DESCQ_WIDTH);
+
+	val = ((1 << TX_FLUSH_DESCQ_CMD_LBN) | (dmaq << TX_FLUSH_DESCQ_LBN));
+
+	EFHW_TRACE("TX DMA flush[%d]", dmaq);
+
+#if BUG5302_WORKAROUND
+	falcon_check_for_bogus_tx_dma_wptr(nic, dmaq);
+#endif
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_write_qq(efhw_kva + TX_FLUSH_DESCQ_REG_OFST,
+			val, FALCON_ATOMIC_TX_FLUSH_DESCQ);
+
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+	return 0;
+}
+
+static inline int
+__falcon_is_tx_dma_channel_flushed(struct efhw_nic *nic, uint dmaq)
+{
+	FALCON_LOCK_DECL;
+	uint64_t val_low64, val_high64;
+	uint64_t enable, flush_pending;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+	ulong offset = falcon_dma_tx_q_offset(nic, dmaq);
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_read_qq(efhw_kva + offset, &val_low64, &val_high64);
+	FALCON_LOCK_UNLOCK(nic);
+
+	/* should see one of three values for these 2 bits
+	 *   1, queue enabled no flush pending
+	 *	- i.e. first flush request
+	 *   2, queue enabled, flush pending
+	 *	- i.e. request to reflush before flush finished
+	 *   3, queue disabled (no flush pending)
+	 *	- flush complete
+	 */
+	__DWCHCK(TX_DESCQ_FLUSH_LBN, TX_DESCQ_FLUSH_WIDTH);
+	__DW3CHCK(TX_DESCQ_EN_LBN, TX_DESCQ_EN_WIDTH);
+	enable = val_high64 & (1 << __DW3(TX_DESCQ_EN_LBN));
+	flush_pending = val_low64 & (1 << TX_DESCQ_FLUSH_LBN);
+
+	if (enable && !flush_pending)
+		return 0;
+
+	EFHW_TRACE("%d, %s: %s, %sflush pending", dmaq, __FUNCTION__,
+		   enable ? "enabled" : "disabled",
+		   flush_pending ? "" : "NO ");
+	/* still in progress */
+	if (enable && flush_pending)
+		return -EALREADY;
+
+	return -EAGAIN;
+}
+
+static int falcon_flush_tx_dma_channel(struct efhw_nic *nic, uint dmaq)
+{
+	int rc;
+	rc = __falcon_is_tx_dma_channel_flushed(nic, dmaq);
+	if (rc < 0) {
+		EFHW_WARN("%s: failed %d", __FUNCTION__, rc);
+		return rc;
+	}
+	return __falcon_really_flush_tx_dma_channel(nic, dmaq);
+}
+
+static int
+__falcon_really_flush_rx_dma_channel(struct efhw_nic *nic, uint dmaq)
+{
+	FALCON_LOCK_DECL;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+	uint val;
+
+	EFHW_BUILD_ASSERT(RX_FLUSH_DESCQ_REG_KER_OFST ==
+			  RX_FLUSH_DESCQ_REG_OFST);
+
+	__DWCHCK(RX_FLUSH_DESCQ_CMD_LBN, RX_FLUSH_DESCQ_CMD_WIDTH);
+	__DWCHCK(RX_FLUSH_DESCQ_LBN, RX_FLUSH_DESCQ_WIDTH);
+	__RANGECHCK(dmaq, RX_FLUSH_DESCQ_WIDTH);
+
+	val = ((1 << RX_FLUSH_DESCQ_CMD_LBN) | (dmaq << RX_FLUSH_DESCQ_LBN));
+
+	EFHW_TRACE("RX DMA flush[%d]", dmaq);
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_write_qq(efhw_kva + RX_FLUSH_DESCQ_REG_OFST, val,
+			FALCON_ATOMIC_RX_FLUSH_DESCQ);
+	mmiowb();
+	FALCON_LOCK_UNLOCK(nic);
+	return 0;
+}
+
+static inline int
+__falcon_is_rx_dma_channel_flushed(struct efhw_nic *nic, uint dmaq)
+{
+	FALCON_LOCK_DECL;
+	uint64_t val;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+	ulong offset = falcon_dma_rx_q_offset(nic, dmaq);
+
+	/* Falcon requires 128 bit atomic access for this register */
+	FALCON_LOCK_LOCK(nic);
+	falcon_read_q(efhw_kva + offset, &val);
+	FALCON_LOCK_UNLOCK(nic);
+
+	__DWCHCK(RX_DESCQ_EN_LBN, RX_DESCQ_EN_WIDTH);
+
+	/* is it enabled? */
+	return (val & (1 << RX_DESCQ_EN_LBN))
+	    ? 0 : -EAGAIN;
+}
+
+static int falcon_flush_rx_dma_channel(struct efhw_nic *nic, uint dmaq)
+{
+	int rc;
+	rc = __falcon_is_rx_dma_channel_flushed(nic, dmaq);
+	if (rc < 0) {
+		EFHW_ERR("%s: failed %d", __FUNCTION__, rc);
+		return rc;
+	}
+	return __falcon_really_flush_rx_dma_channel(nic, dmaq);
+}
+
+/*--------------------------------------------------------------------
+ *
+ * Falcon specific event callbacks
+ *
+ *--------------------------------------------------------------------*/
+
+int
+falcon_handle_char_event(struct efhw_nic *nic, struct efhw_ev_handler *h,
+			 efhw_event_t *ev)
+{
+	EFHW_TRACE("DRIVER EVENT: "FALCON_EVENT_FMT,
+		   FALCON_EVENT_PRI_ARG(*ev));
+
+	switch (FALCON_EVENT_DRIVER_SUBCODE(ev)) {
+
+	case TX_DESCQ_FLS_DONE_EV_DECODE:
+		EFHW_TRACE("TX[%d] flushed",
+			   (int)FALCON_EVENT_TX_FLUSH_Q_ID(ev));
+#if !defined(__ci_ul_driver__)
+		efhw_handle_txdmaq_flushed(nic, h, ev);
+#endif
+		break;
+
+	case RX_DESCQ_FLS_DONE_EV_DECODE:
+		EFHW_TRACE("RX[%d] flushed",
+			   (int)FALCON_EVENT_TX_FLUSH_Q_ID(ev));
+#if !defined(__ci_ul_driver__)
+		efhw_handle_rxdmaq_flushed(nic, h, ev);
+#endif
+		break;
+
+	case SRM_UPD_DONE_EV_DECODE:
+		nic->buf_commit_outstanding =
+		    max(0, nic->buf_commit_outstanding - 1);
+		EFHW_TRACE("COMMIT DONE %d", nic->buf_commit_outstanding);
+		break;
+
+	case EVQ_INIT_DONE_EV_DECODE:
+		EFHW_TRACE("EVQ INIT");
+		break;
+
+	case WAKE_UP_EV_DECODE:
+		EFHW_TRACE("WAKE UP");
+		efhw_handle_wakeup_event(nic, h, ev);
+		break;
+
+	case TIMER_EV_DECODE:
+		EFHW_TRACE("TIMER");
+		efhw_handle_timeout_event(nic, h, ev);
+		break;
+
+	case RX_DESCQ_FLSFF_OVFL_EV_DECODE:
+		/* This shouldn't happen. */
+		EFHW_ERR("%s: RX flush fifo overflowed", __FUNCTION__);
+		return -EINVAL;
+
+	default:
+		EFHW_TRACE("UNKOWN DRIVER EVENT: " FALCON_EVENT_FMT,
+			   FALCON_EVENT_PRI_ARG(*ev));
+		break;
+	}
+	return 0;
+}
+
+/*--------------------------------------------------------------------
+ *
+ * Abstraction Layer Hooks
+ *
+ *--------------------------------------------------------------------*/
+
+struct efhw_func_ops falcon_char_functional_units = {
+	falcon_nic_close_hardware,
+	falcon_nic_init_hardware,
+	falcon_nic_interrupt,
+	falcon_nic_interrupt_enable,
+	falcon_nic_interrupt_disable,
+	falcon_nic_set_interrupt_moderation,
+	falcon_nic_event_queue_enable,
+	falcon_nic_event_queue_disable,
+	falcon_nic_wakeup_request,
+	falcon_nic_sw_event,
+	falcon_nic_ipfilter_set,
+	falcon_nic_ipfilter_attach,
+	falcon_nic_ipfilter_detach,
+	falcon_nic_ipfilter_clear,
+	falcon_dmaq_tx_q_init,
+	falcon_dmaq_rx_q_init,
+	falcon_dmaq_tx_q_disable,
+	falcon_dmaq_rx_q_disable,
+	falcon_flush_tx_dma_channel,
+	falcon_flush_rx_dma_channel,
+	falcon_nic_buffer_table_set,
+	falcon_nic_buffer_table_set_n,
+	falcon_nic_buffer_table_clear,
+	falcon_nic_buffer_table_commit,
+};
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/falcon_hash.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/falcon_hash.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/falcon_hash.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/falcon_hash.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,178 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains EtherFabric NIC hash algorithms implementation.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/efhw/debug.h>
+#include <ci/driver/efab/hardware.h>
+
+
+/* this mask is per filter bank hence /2 */
+#define FILTER_MASK(n)  ((n) / 2u - 1u)
+
+/*
+ *  Main Functions related to the Hash Table Generation
+ *  Author: Srinivasaih, Nataraj
+ * Created: Thu May 13:32:41 PDT 2004
+ * $Id: falcon_hash.c,v 1.20 2008/01/29 08:28:56 ok_sasha Exp $
+ */
+/***************************************************************************
+Class	Maximum number of	Valid address ranges
+	hosts per network
+A	16777214		1.0.0.1 through 9.255.255.254
+				11.0.0.1 through 126.255.255.254
+B	65534			128.0.0.1 through 172.15.255.254
+				172.32.0.1 through 191.255.255.254
+C	254			192.0.0.1 through 192.167.255.254
+				192.169.0.1 through 223.255.255.254
+P	16777214		10.0.0.1 through 10.255.255.254 (10/8)
+	1048574			172.16.0.1 through 172.31.255.254 (172.16/12)
+	65534			192.168.0.1 through 192.168.255.254 (192.168/16)
+
+R	-			0.0.0.0 through 0.255.255.255
+				(used if host will be assigned a
+				valid address dynamically)
+				127.0.0.0 through 127.255.255.255
+				(loopback addresses)
+
+P : Private internets only
+R : Reserved
+****************************************************************************/
+
+/* All LE parameters */
+unsigned int
+falcon_hash_get_key(unsigned int src_ip, unsigned int src_port,
+		    unsigned int dest_ip, unsigned int dest_port,
+		    int tcp, int full)
+{
+
+	unsigned int result = 0;
+	int net_type;
+
+	EFHW_ASSERT(tcp == 0 || tcp == 1);
+	EFHW_ASSERT(full == 0 || full == 1);
+
+	net_type = tcp << 4 | full;
+
+	/* Note that src_ip and src_port ignored if a wildcard filter */
+	switch (net_type) {
+	case 0x01:		/* UDP Full */
+		result = ((dest_ip & 0xfffffffe) | (!(dest_ip & 1))) ^
+		    (((dest_port << 16) & 0xFFFF0000) |
+		     ((src_ip >> 16) & 0x0000FFFF)) ^
+		    (((src_ip << 16) & 0xFFFF0000) |
+		     ((src_port & 0x0000FFFF)));
+		EFHW_TRACE("falcon_hash_get_key: UDP Full %x", result);
+		break;
+	case 0x00:		/* UDP Wild Card */
+		result = ((dest_ip & 0xfffffffe) | (!(dest_ip & 1))) ^
+		    (((dest_port << 16) & 0x00000000) |
+		     ((src_ip >> 16) & 0x00000000)) ^
+		    (((src_ip << 16) & 0x00000000) |
+		     ((dest_port & 0x0000FFFF)));
+		EFHW_TRACE("falcon_hash_get_key: UDP Wildcard %x", result);
+		break;
+	case 0x10:		/* TCP Wild Card */
+		result = (dest_ip) ^
+		    (((dest_port << 16) & 0xFFFF0000) |
+		     ((src_ip >> 16) & 0x00000000)) ^
+		    (((src_ip << 16) & 0x00000000) |
+		     ((src_port & 0x00000000)));
+		EFHW_TRACE("falcon_hash_get_key: TCP Wildcard %x", result);
+		break;
+	case 0x11:		/* TCP Full */
+		result = (dest_ip) ^
+		    (((dest_port << 16) & 0xFFFF0000) |
+		     ((src_ip >> 16) & 0x0000FFFF)) ^
+		    (((src_ip << 16) & 0xFFFF0000) |
+		     ((src_port & 0x0000FFFF)));
+		EFHW_TRACE("falcon_hash_get_key: TCP Full %x", result);
+		break;
+	default:
+		EFHW_ASSERT(0);
+
+	}
+	return (result);
+}
+
+/* This function generates the First Hash key */
+unsigned int falcon_hash_function1(unsigned int key, unsigned int nfilters)
+{
+
+	unsigned short int lfsr_reg;
+	unsigned int tmp_key;
+	int index;
+
+	unsigned short int lfsr_input;
+	unsigned short int single_bit_key;
+	unsigned short int bit16_lfsr;
+	unsigned short int bit3_lfsr;
+
+	lfsr_reg = 0xFFFF;
+	tmp_key = key;
+
+	/* For Polynomial equation X^16+X^3+1 */
+	for (index = 0; index < 32; index++) {
+		/* Get the bit from key and shift the key */
+		single_bit_key = (tmp_key & 0x80000000) >> 31;
+		tmp_key = tmp_key << 1;
+
+		/* get the Tap bits to XOR operation */
+		bit16_lfsr = (lfsr_reg & 0x8000) >> 15;
+		bit3_lfsr = (lfsr_reg & 0x0004) >> 2;
+
+		/* Get the Input value to the LFSR */
+		lfsr_input = ((bit16_lfsr ^ bit3_lfsr) ^ single_bit_key);
+
+		/* Shift and store out of the two TAPs */
+		lfsr_reg = lfsr_reg << 1;
+		lfsr_reg = lfsr_reg | (lfsr_input & 0x0001);
+
+	}
+
+	lfsr_reg = lfsr_reg & FILTER_MASK(nfilters);
+
+	return lfsr_reg;
+}
+
+/* This function generates the Second Hash */
+unsigned int
+falcon_hash_function2(unsigned int key, unsigned int nfilters)
+{
+	return (unsigned int)(((unsigned long long)key * 2 - 1) &
+			      FILTER_MASK(nfilters));
+}
+
+/* This function iterates through the hash table */
+unsigned int
+falcon_hash_iterator(unsigned int hash1, unsigned int hash2,
+		     unsigned int n_search, unsigned int nfilters)
+{
+	return ((hash1 + (n_search * hash2)) & FILTER_MASK(nfilters));
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/falcon_mac.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/falcon_mac.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/falcon_mac.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/falcon_mac.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,171 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains MACs (Mentor MAC & GDACT1 ) support for Falcon.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/efhw/falcon.h>
+#include <ci/driver/efab/hardware.h>
+
+/********************************************************************
+ * Mentor MAC
+ */
+
+#define _PRE(x)       GM##x
+
+/*--------------------------------------------------------------------
+ *
+ * Debug Support
+ *
+ *--------------------------------------------------------------------*/
+
+#define MENTOR_MAC_ASSERT_VALID()					\
+    EFHW_ASSERT(nic);							\
+    EFHW_ASSERT(EFHW_KVA(nic));						\
+    EFHW_ASSERT(_PRE(_CFG1_REG_OFST) == _PRE(_CFG1_REG_KER_OFST));	\
+    EFHW_ASSERT(_PRE(_CFG2_REG_OFST) == _PRE(_CFG2_REG_KER_OFST));	\
+    EFHW_ASSERT(_PRE(_IPG_REG_OFST) == _PRE(_IPG_REG_KER_OFST));	\
+    EFHW_ASSERT(_PRE(_HD_REG_OFST) == _PRE(_HD_REG_KER_OFST));		\
+    EFHW_ASSERT(_PRE(_MAX_FLEN_REG_OFST) == _PRE(_MAX_FLEN_REG_KER_OFST)); \
+    EFHW_ASSERT(_PRE(_TEST_REG_OFST) == _PRE(_TEST_REG_KER_OFST));	\
+    EFHW_ASSERT(_PRE(_ADR1_REG_OFST) == _PRE(_ADR1_REG_KER_OFST));	\
+    EFHW_ASSERT(_PRE(_ADR2_REG_OFST) == _PRE(_ADR2_REG_KER_OFST));	\
+    EFHW_ASSERT(_PRE(F_CFG0_REG_OFST) == _PRE(F_CFG0_REG_KER_OFST));	\
+    EFHW_ASSERT(_PRE(F_CFG1_REG_OFST) == _PRE(F_CFG1_REG_KER_OFST));	\
+    EFHW_ASSERT(_PRE(F_CFG2_REG_OFST) == _PRE(F_CFG2_REG_KER_OFST));	\
+    EFHW_ASSERT(_PRE(F_CFG3_REG_OFST) == _PRE(F_CFG3_REG_KER_OFST));	\
+    EFHW_ASSERT(_PRE(F_CFG4_REG_OFST) == _PRE(F_CFG4_REG_KER_OFST));	\
+    EFHW_ASSERT(_PRE(F_CFG5_REG_OFST) == _PRE(F_CFG5_REG_KER_OFST));
+
+/*! Get MAC current address - i.e not necessarily the one in the EEPROM */
+static inline void mentormac_get_mac_addr(struct efhw_nic *nic)
+{
+	efhw_ioaddr_t mac_kva;
+	uint val1, val2;
+
+	MENTOR_MAC_ASSERT_VALID();
+
+	mac_kva = GM_P0_BASE + EFHW_KVA(nic);
+
+	val1 = readl(mac_kva + _PRE(_ADR1_REG_OFST));
+	val2 = readl(mac_kva + _PRE(_ADR2_REG_OFST));
+
+#if 0
+	nic->mac_addr[0] = (val1 & 0xff000000) >> 24;
+	nic->mac_addr[1] = (val1 & 0x00ff0000) >> 16;
+	nic->mac_addr[2] = (val1 & 0x0000ff00) >> 8;
+	nic->mac_addr[3] = (val1 & 0x000000ff) >> 0;
+	nic->mac_addr[4] = (val2 & 0xff000000) >> 24;
+	nic->mac_addr[5] = (val2 & 0x00ff0000) >> 16;
+#else
+	nic->mac_addr[5] = (val1 & 0xff000000) >> 24;
+	nic->mac_addr[4] = (val1 & 0x00ff0000) >> 16;
+	nic->mac_addr[3] = (val1 & 0x0000ff00) >> 8;
+	nic->mac_addr[2] = (val1 & 0x000000ff) >> 0;
+	nic->mac_addr[1] = (val2 & 0xff000000) >> 24;
+	nic->mac_addr[0] = (val2 & 0x00ff0000) >> 16;
+#endif
+}
+
+
+/********************************************************************
+ * GDACT10 MAC
+ */
+
+/*--------------------------------------------------------------------
+ *
+ * Debug Support
+ *
+ *--------------------------------------------------------------------*/
+
+#define GDACT10_MAC_ASSERT_VALID()					\
+    EFHW_ASSERT(nic);							\
+    EFHW_ASSERT(EFHW_KVA(nic));						\
+    EFHW_ASSERT(XM_GLB_CFG_REG_P0_OFST == XM_GLB_CFG_REG_KER_P0_OFST);	\
+    EFHW_ASSERT(XM_TX_CFG_REG_P0_OFST  == XM_TX_CFG_REG_KER_P0_OFST);	\
+    EFHW_ASSERT(XM_RX_CFG_REG_P0_OFST  == XM_RX_CFG_REG_KER_P0_OFST);	\
+    EFHW_ASSERT(MAC0_SPEED_LBN         == MAC1_SPEED_LBN);		\
+    EFHW_ASSERT(MAC0_SPEED_WIDTH       == MAC1_SPEED_WIDTH);		\
+    EFHW_ASSERT(MAC0_LINK_STATUS_LBN   == MAC1_LINK_STATUS_LBN);	\
+    EFHW_ASSERT(MAC0_LINK_STATUS_WIDTH == MAC1_LINK_STATUS_WIDTH);	\
+    EFHW_ASSERT(MAC1_BCAD_ACPT_LBN     == MAC0_BCAD_ACPT_LBN);		\
+    EFHW_ASSERT(MAC1_UC_PROM_LBN       == MAC0_UC_PROM_LBN);		\
+    EFHW_ASSERT(MAC0_CTRL_REG_KER_OFST == MAC0_CTRL_REG_OFST);		\
+    EFHW_ASSERT(MAC1_CTRL_REG_KER_OFST == MAC1_CTRL_REG_OFST);		\
+    EFHW_ASSERT(XM_ADR_LO_REG_KER_P0_OFST  == XM_ADR_LO_REG_P0_OFST);	\
+    EFHW_ASSERT(XM_ADR_HI_REG_KER_P0_OFST  == XM_ADR_HI_REG_P0_OFST);	\
+    EFHW_ASSERT(XM_RX_PARAM_REG_KER_P0_OFST == XM_RX_PARAM_REG_P0_OFST);
+
+/*--------------------------------------------------------------------
+ *
+ * Information gathering
+ *
+ *--------------------------------------------------------------------*/
+
+/*! Get MAC current address - i.e not necessarily the one in the EEPROM */
+static inline void GDACT10mac_get_mac_addr(struct efhw_nic *nic)
+{
+	uint val1, val2;
+	efhw_ioaddr_t efhw_kva = EFHW_KVA(nic);
+	FALCON_LOCK_DECL;
+
+	GDACT10_MAC_ASSERT_VALID();
+
+	EFHW_ASSERT(XM_ADR_LO_LBN == 0);
+	EFHW_ASSERT(XM_ADR_LO_WIDTH == 32);
+	EFHW_ASSERT(XM_ADR_HI_LBN == 0);
+	EFHW_ASSERT(XM_ADR_HI_WIDTH == 16);
+
+	FALCON_LOCK_LOCK(nic);
+
+	val1 = readl(efhw_kva + XM_ADR_LO_REG_P0_OFST);
+	val2 = readl(efhw_kva + XM_ADR_HI_REG_P0_OFST);
+
+	FALCON_LOCK_UNLOCK(nic);
+
+	/* The HW scores no points for consistency */
+	nic->mac_addr[5] = (val2 & 0x0000ff00) >> 8;
+	nic->mac_addr[4] = (val2 & 0x000000ff) >> 0;
+	nic->mac_addr[3] = (val1 & 0xff000000) >> 24;
+	nic->mac_addr[2] = (val1 & 0x00ff0000) >> 16;
+	nic->mac_addr[1] = (val1 & 0x0000ff00) >> 8;
+	nic->mac_addr[0] = (val1 & 0x000000ff) >> 0;
+}
+
+
+/********************************************************************
+ * Call one or another function
+ */
+
+void falcon_get_mac_addr(struct efhw_nic *nic)
+{
+	if (nic->flags & NIC_FLAG_10G)
+		GDACT10mac_get_mac_addr(nic);
+	else
+		mentormac_get_mac_addr(nic);
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/filter_resource.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/filter_resource.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/filter_resource.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/filter_resource.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,317 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains filters support.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/efrm/nic_table.h>
+#include <ci/driver/efab/hardware.h>
+#include <ci/efhw/falcon.h>
+#include <ci/efrm/vi_resource_manager.h>
+#include <ci/efrm/private.h>
+#include <ci/efrm/filter.h>
+#include <ci/efrm/buffer_table.h>
+
+struct filter_resource_manager {
+	struct efrm_resource_manager rm;
+	struct kfifo *free_ids;
+};
+
+static struct filter_resource_manager *efrm_filter_manager;
+
+void efrm_filter_resource_free(struct filter_resource *frs)
+{
+	struct efhw_nic *nic;
+	int nic_i;
+	int id;
+
+	EFRM_RESOURCE_ASSERT_VALID(&frs->rs, 1);
+
+	EFRM_TRACE("%s: " EFRM_RESOURCE_FMT, __FUNCTION__,
+		   EFRM_RESOURCE_PRI_ARG(frs->rs.rs_handle));
+
+	/* if we have a PT endpoint */
+	if (NULL != frs->pt) {
+		/* Detach the filter */
+		EFRM_FOR_EACH_NIC_IN_SET(&frs->nic_set, nic_i, nic)
+		    efhw_nic_ipfilter_detach(nic, frs->filter_idx);
+
+		/* Release our ref to the PT resource. */
+		EFRM_TRACE("%s: releasing PT resource reference",
+			   __FUNCTION__);
+		efrm_vi_resource_release(frs->pt);
+	}
+
+	/* Disable the filter. */
+	EFRM_FOR_EACH_NIC_IN_SET(&frs->nic_set, nic_i, nic)
+	    efhw_nic_ipfilter_clear(nic, frs->filter_idx);
+
+	/* Free this filter. */
+	id = EFRM_RESOURCE_INSTANCE(frs->rs.rs_handle);
+	EFRM_VERIFY_EQ(kfifo_put(efrm_filter_manager->free_ids,
+				 (unsigned char *)&id, sizeof(id)),
+		       sizeof(id));
+
+	EFRM_DO_DEBUG(memset(frs, 0, sizeof(*frs)));
+	kfree(frs);
+}
+EXPORT_SYMBOL(efrm_filter_resource_free);
+
+static void filter_rm_dtor(struct efrm_resource_manager *rm)
+{
+	EFRM_TRACE("filter_rm_dtor");
+
+	EFRM_RESOURCE_MANAGER_ASSERT_VALID(&efrm_filter_manager->rm);
+	EFRM_ASSERT(&efrm_filter_manager->rm == rm);
+
+	kfifo_vfree(efrm_filter_manager->free_ids);
+	EFRM_TRACE("filter_rm_dtor: done");
+}
+
+/**********************************************************************/
+/**********************************************************************/
+/**********************************************************************/
+
+int efrm_create_filter_resource_manager(struct efrm_resource_manager **rm_out)
+{
+	int rc;
+
+	EFRM_ASSERT(rm_out);
+
+	efrm_filter_manager =
+	    kmalloc(sizeof(struct filter_resource_manager), GFP_KERNEL);
+	if (efrm_filter_manager == 0)
+		return -ENOMEM;
+	memset(efrm_filter_manager, 0, sizeof(*efrm_filter_manager));
+
+	rc = efrm_resource_manager_ctor(&efrm_filter_manager->rm,
+					filter_rm_dtor, "FILTER",
+					EFRM_RESOURCE_FILTER, 0);
+	if (rc < 0)
+		goto fail1;
+
+	/* Create a pool of free instances */
+	rc = efrm_kfifo_id_ctor(&efrm_filter_manager->free_ids,
+				0, EFHW_IP_FILTER_NUM,
+				&efrm_filter_manager->rm.rm_lock);
+	if (rc != 0)
+		goto fail2;
+
+	*rm_out = &efrm_filter_manager->rm;
+	EFRM_TRACE("%s: filter resources created - %d IDs",
+		   __FUNCTION__, kfifo_len(efrm_filter_manager->free_ids));
+	return 0;
+
+fail2:
+	efrm_resource_manager_dtor(&efrm_filter_manager->rm);
+fail1:
+	memset(efrm_filter_manager, 0, sizeof(*efrm_filter_manager));
+	kfree(efrm_filter_manager);
+	return rc;
+
+}
+
+/*--------------------------------------------------------------------
+ *!
+ * Called to set/change the PT endpoint of a filter
+ *
+ * Example of use is TCP helper when it finds a wildcard IP filter
+ * needs to change which application it delivers traffic to
+ *
+ * \param frs           filter resource
+ * \param pt_handle     handle of new PT endpoint
+ *
+ * \return              standard error codes
+ *
+ *--------------------------------------------------------------------*/
+int
+efrm_filter_resource_set_ptresource(struct filter_resource *frs,
+				    struct vi_resource *ptrs)
+{
+	int rc, pti, nic_i;
+	struct efhw_nic *nic;
+
+	EFRM_ASSERT(frs);
+
+	/* if filter is attached to a valid PT endpoint */
+	if (NULL != frs->pt) {
+
+		EFRM_TRACE("%s: detaching PT resource " EFRM_RESOURCE_FMT
+			   " from filter ",
+			   __FUNCTION__,
+			   EFRM_RESOURCE_PRI_ARG(frs->rs.rs_handle));
+		/* Detach the filter */
+		EFRM_FOR_EACH_NIC_IN_SET(&frs->nic_set, nic_i, nic)
+		    efhw_nic_ipfilter_detach(nic, frs->filter_idx);
+
+		/* release reference */
+		efrm_vi_resource_release(frs->pt);
+		frs->pt = NULL;
+	}
+
+	if (ptrs != NULL) {
+		/* get PT endpoint index */
+		EFRM_RESOURCE_ASSERT_VALID(&ptrs->rs, 0);
+		EFRM_ASSERT(EFRM_RESOURCE_TYPE(ptrs->rs.rs_handle) ==
+			    EFRM_RESOURCE_VI);
+		pti = EFRM_RESOURCE_INSTANCE(ptrs->rs.rs_handle);
+		if (pti == 0) {
+			EFRM_ERR("%s: cannot filter for channel 0",
+				 __FUNCTION__);
+			rc = -EINVAL;
+			goto fail2;
+		}
+		frs->pt = ptrs;
+		EFRM_TRACE("%s: attaching PT resource " EFRM_RESOURCE_FMT
+			   " to filter",
+			   __FUNCTION__,
+			   EFRM_RESOURCE_PRI_ARG(frs->pt->rs.rs_handle));
+		EFRM_FOR_EACH_NIC_IN_SET(&frs->nic_set, nic_i, nic)
+		    efhw_nic_ipfilter_attach(nic, frs->filter_idx, pti);
+		efrm_vi_resource_ref(frs->pt);
+	}
+	return 0;
+
+fail2:
+	efrm_vi_resource_release(frs->pt);
+	return rc;
+}
+EXPORT_SYMBOL(efrm_filter_resource_set_ptresource);
+
+int efrm_filter_resource_clear(struct filter_resource *frs)
+{
+	struct efhw_nic *nic;
+	int nic_i;
+
+	EFRM_ASSERT(frs);
+	EFRM_FOR_EACH_NIC_IN_SET(&frs->nic_set, nic_i, nic)
+	    efhw_nic_ipfilter_clear(nic, frs->filter_idx);
+
+	return 0;
+}
+EXPORT_SYMBOL(efrm_filter_resource_clear);
+
+int
+__efrm_filter_resource_set(struct filter_resource *frs, int type,
+			   unsigned saddr, uint16_t sport,
+			   unsigned daddr, uint16_t dport)
+{
+	struct efhw_nic *nic;
+	int nic_i, rc = 0;
+	unsigned instance = EFRM_RESOURCE_INSTANCE(frs->pt->rs.rs_handle);
+
+	EFRM_ASSERT(frs);
+	EFRM_ASSERT(frs->pt);
+
+	if (efrm_nic_table.a_nic->devtype.variant >= 'B') {
+		/* Scatter setting must match the setting for
+		 * the corresponding RX queue */
+		if (!(frs->pt->flags & EFHW_VI_JUMBO_EN))
+			type |= EFHW_IP_FILTER_TYPE_NOSCAT_B0_MASK;
+	}
+
+	EFRM_FOR_EACH_NIC_IN_SET(&frs->nic_set, nic_i, nic)
+	    if (rc >= 0)
+		rc = efhw_nic_ipfilter_set(nic, type, &frs->filter_idx,
+					   instance,
+					   saddr, sport, daddr, dport);
+
+	return rc;
+}
+EXPORT_SYMBOL(__efrm_filter_resource_set);;
+
+int
+efrm_filter_resource_alloc(struct vi_resource *vi_parent,
+			   struct filter_resource **frs_out)
+{
+	struct efhw_nic *nic;
+	int nic_i, rc, instance;
+	struct filter_resource *frs;
+
+	EFRM_ASSERT(frs_out);
+	EFRM_ASSERT(efrm_filter_manager);
+	EFRM_RESOURCE_MANAGER_ASSERT_VALID(&efrm_filter_manager->rm);
+	EFRM_ASSERT(vi_parent == NULL ||
+		    EFRM_RESOURCE_TYPE(vi_parent->rs.rs_handle) ==
+		    EFRM_RESOURCE_VI);
+
+	/* Allocate resource data structure. */
+	frs = kmalloc(sizeof(struct filter_resource), GFP_KERNEL);
+	if (!frs)
+		return -ENOMEM;
+	efrm_nic_set_clear(&frs->nic_set);
+
+	/* Allocate an instance. */
+	rc = kfifo_get(efrm_filter_manager->free_ids,
+		       (unsigned char *)&instance, sizeof(instance));
+	if (rc != sizeof(instance)) {
+		EFRM_TRACE("%s: out of instances", __FUNCTION__);
+		EFRM_ASSERT(rc == 0);
+		rc = -EBUSY;
+		goto fail1;
+	}
+
+	/* Initialise the resource DS. */
+	efrm_resource_init(&frs->rs, EFRM_RESOURCE_FILTER, instance);
+	frs->pt = vi_parent;
+	if (frs->pt)
+		efrm_vi_resource_ref(frs->pt);
+	frs->filter_idx = -1;
+	EFRM_FOR_EACH_NIC(nic_i, nic)
+	    efrm_nic_set_write(&frs->nic_set, nic_i, true);
+
+	EFRM_TRACE("%s: " EFRM_RESOURCE_FMT " Q %d idx %x",
+		   __FUNCTION__,
+		   EFRM_RESOURCE_PRI_ARG(frs->rs.rs_handle),
+		   vi_parent == NULL ? -1 :
+		   EFRM_RESOURCE_INSTANCE(vi_parent->rs.rs_handle),
+		   frs->filter_idx);
+
+	/* Put it into the resource manager's table. */
+	rc = efrm_resource_manager_insert(&frs->rs);
+	if (rc != 0) {
+		if (atomic_dec_and_test(&frs->rs.rs_ref_count))
+			efrm_filter_resource_free(frs);
+		return rc;
+	}
+
+	*frs_out = frs;
+	return 0;
+
+fail1:
+	memset(frs, 0, sizeof(*frs));
+	kfree(frs);
+	return rc;
+}
+EXPORT_SYMBOL(efrm_filter_resource_alloc);
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/iobufset_resource.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/iobufset_resource.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/iobufset_resource.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/iobufset_resource.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,373 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains non-contiguous I/O buffers support.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/efrm/nic_table.h>
+#include <ci/efhw/iopage.h>
+#include <ci/driver/efab/hardware.h>
+#include <ci/efrm/private.h>
+#include <ci/efrm/iobufset.h>
+#include <ci/efrm/vi_resource_manager.h>
+#include <ci/efrm/buffer_table.h>
+
+#define EFRM_IOBUFSET_MAX_NUM_INSTANCES 0x00010000
+
+struct iobufset_resource_manager {
+	struct efrm_resource_manager rm;
+	struct kfifo *free_ids;
+};
+
+struct iobufset_resource_manager *efrm_iobufset_manager;
+
+#define iobsrs(rs1)  iobufset_resource(rs1)
+
+/* Returns size of iobufset resource data structure. */
+static inline size_t iobsrs_size(int no_pages)
+{
+	return offsetof(struct iobufset_resource, bufs) +
+	    no_pages * sizeof(efhw_iopage_t);
+}
+
+void efrm_iobufset_resource_free(struct iobufset_resource *rs)
+{
+	unsigned int no_pages;
+	unsigned int i;
+	int id;
+
+	EFRM_RESOURCE_ASSERT_VALID(&rs->rs, 1);
+	no_pages = rs->n_bufs;
+
+	if (rs->buf_tbl_alloc.base != (unsigned)-1)
+		efrm_buffer_table_free(&rs->buf_tbl_alloc);
+
+	/* see comment on call to efhw_iopage_alloc in the alloc routine above
+	   for discussion on use of efrm_nic_table.a_nic here */
+	EFRM_ASSERT(efrm_nic_table.a_nic);
+	if (rs->order == 0) {
+		for (i = 0; i < rs->n_bufs; ++i)
+			efhw_iopage_free(efrm_nic_table.a_nic, &rs->bufs[i]);
+	} else {
+		/* it is important that this is executed in increasing page
+		 * order because some implementations of
+		 * efhw_iopages_init_from_iopage() assume this */
+		for (i = 0; i < rs->n_bufs;
+		     i += rs->pages_per_contiguous_chunk) {
+			efhw_iopages_t iopages;
+			efhw_iopages_init_from_iopage(&iopages, &rs->bufs[i],
+						    rs->order);
+			efhw_iopages_free(efrm_nic_table.a_nic, &iopages);
+		}
+	}
+
+	/* free the instance number */
+	id = EFRM_RESOURCE_INSTANCE(rs->rs.rs_handle);
+	EFRM_VERIFY_EQ(kfifo_put(efrm_iobufset_manager->free_ids,
+				 (unsigned char *)&id, sizeof(id)), sizeof(id));
+
+	efrm_vi_resource_release(rs->evq);
+
+	EFRM_DO_DEBUG(memset(rs, 0, sizeof(*rs)));
+	if (iobsrs_size(no_pages) < PAGE_SIZE) {
+		kfree(rs);
+	} else {
+		vfree(rs);
+	}
+}
+EXPORT_SYMBOL(efrm_iobufset_resource_free);
+
+int
+efrm_iobufset_resource_alloc(int32_t n_pages,
+			     int32_t pages_per_contiguous_chunk,
+			     struct vi_resource *vi_evq,
+			     bool phys_addr_mode,
+			     uint32_t faultonaccess,
+			     struct iobufset_resource **iobrs_out)
+{
+	struct iobufset_resource *iobrs;
+	int rc, instance, object_size;
+	unsigned int i;
+
+	EFRM_ASSERT(iobrs_out);
+	EFRM_ASSERT(efrm_iobufset_manager);
+	EFRM_RESOURCE_MANAGER_ASSERT_VALID(&efrm_iobufset_manager->rm);
+	EFRM_RESOURCE_ASSERT_VALID(&vi_evq->rs, 0);
+	EFRM_ASSERT(EFRM_RESOURCE_TYPE(vi_evq->rs.rs_handle) ==
+		    EFRM_RESOURCE_VI);
+	EFRM_ASSERT(efrm_nic_table.a_nic);
+
+	/* allocate the resource data structure. */
+	object_size = iobsrs_size(n_pages);
+	if (object_size < PAGE_SIZE) {
+		/* this should be OK from a tasklet */
+		/* Necessary to do atomic alloc() as this
+		   can be called from a weird-ass iSCSI context that is
+		   !in_interrupt but is in_atomic - See BUG3163 */
+		iobrs = kmalloc(object_size, GFP_ATOMIC);
+	} else {		/* can't do this within a tasklet */
+#ifndef NDEBUG
+		if (in_interrupt() || in_atomic()) {
+			EFRM_ERR("%s(): alloc->u.iobufset.in_n_pages=%d",
+				 __FUNCTION__, n_pages);
+			EFRM_ASSERT(!in_interrupt());
+			EFRM_ASSERT(!in_atomic());
+		}
+#endif
+		iobrs = (struct iobufset_resource *) vmalloc(object_size);
+	}
+	if (iobrs == 0) {
+		rc = -ENOMEM;
+		goto fail1;
+	}
+
+	/* Allocate an instance number. */
+	rc = kfifo_get(efrm_iobufset_manager->free_ids,
+		       (unsigned char *)&instance, sizeof(instance));
+	if (rc != sizeof(instance)) {
+		EFRM_TRACE("%s: out of instances", __FUNCTION__);
+		EFRM_ASSERT(rc == 0);
+		rc = -EBUSY;
+		goto fail3;
+	}
+
+	efrm_resource_init(&iobrs->rs, EFRM_RESOURCE_IOBUFSET, instance);
+
+	iobrs->evq = vi_evq;
+	efrm_vi_resource_ref(iobrs->evq);
+
+	iobrs->n_bufs = n_pages;
+	iobrs->pages_per_contiguous_chunk = pages_per_contiguous_chunk;
+	iobrs->order = fls(iobrs->pages_per_contiguous_chunk - 1);
+	iobrs->faultonaccess = faultonaccess;
+
+	EFRM_TRACE("%s: " EFRM_RESOURCE_FMT " %u pages", __FUNCTION__,
+		   EFRM_RESOURCE_PRI_ARG(iobrs->rs.rs_handle), iobrs->n_bufs);
+
+	/* Allocate the iobuffers. */
+	if (iobrs->order == 0) {
+		/* make sure iobufs are in a known state in case we don't
+		 * finish our allocation */
+		for (i = 0; i < iobrs->n_bufs; ++i)
+			memset(&iobrs->bufs[i], 0, sizeof(iobrs->bufs[i]));
+
+		for (i = 0; i < iobrs->n_bufs; ++i) {
+			/* due to bug2426 we have to specifiy a NIC when
+			 * allocating a DMAable page, which is a bit messy.
+			 * For now we assume that if the page is suitable
+			 * (e.g. DMAable) by one nic (efrm_nic_table.a_nic),
+			 * it is suitable for all NICs.
+			 * XXX I bet that breaks in Solaris.
+			 */
+			rc = efhw_iopage_alloc(efrm_nic_table.a_nic,
+					     &iobrs->bufs[i]);
+			if (rc < 0) {
+				EFRM_ERR("%s: failed (rc %d) to allocate "
+					 "page (i=%u)", __FUNCTION__, rc, i);
+				goto fail4;
+			}
+		}
+	} else {
+		efhw_iopages_t iopages;
+		unsigned j;
+
+		/* make sure iobufs are in a known state in case we don't
+		 * finish our allocation */
+		for (i = 0; i < iobrs->n_bufs; ++i)
+			memset(&iobrs->bufs[i], 0, sizeof(iobrs->bufs[i]));
+
+		for (i = 0; i < iobrs->n_bufs;
+		     i += iobrs->pages_per_contiguous_chunk) {
+			rc = efhw_iopages_alloc(efrm_nic_table.a_nic,
+						&iopages, iobrs->order);
+			if (rc < 0) {
+				EFRM_ERR("%s: failed (rc %d) to allocate "
+					 "pages (i=%u order %d)",
+					 __FUNCTION__, rc, i, iobrs->order);
+				goto fail4;
+			}
+			for (j = 0; j < iobrs->pages_per_contiguous_chunk;
+			     j++) {
+				/* some implementation of
+				 * efhw_iopage_init_from_iopages() rely on
+				 * this function being called for
+				 * _all_ pages in the chunk */
+				efhw_iopage_init_from_iopages(
+							&iobrs->bufs[i + j],
+							&iopages, j);
+			}
+		}
+	}
+
+	iobrs->buf_tbl_alloc.base = (unsigned)-1;
+
+	if (!phys_addr_mode) {
+		unsigned instance = EFAB_VI_RESOURCE_INSTANCE(iobrs->evq);
+		/* Allocate space in the NIC's buffer table. */
+		rc = efrm_buffer_table_alloc(fls(iobrs->n_bufs - 1),
+					     &iobrs->buf_tbl_alloc);
+		if (rc < 0) {
+			EFRM_ERR("%s: failed (%d) to alloc %d buffer table "
+				 "entries", __FUNCTION__, rc, iobrs->n_bufs);
+			goto fail5;
+		}
+		EFRM_ASSERT(((unsigned)1 << iobrs->buf_tbl_alloc.order) >=
+			    (unsigned)iobrs->n_bufs);
+
+		/* Initialise the buffer table entries. */
+		for (i = 0; i < iobrs->n_bufs; ++i) {
+			/*\ ?? \TODO burst them! */
+			efrm_buffer_table_set(&iobrs->buf_tbl_alloc, i,
+					      efhw_iopage_dma_addr(&iobrs->
+								 bufs[i]),
+					      instance);
+		}
+		efrm_buffer_table_commit();
+	}
+
+	EFRM_TRACE("%s: " EFRM_RESOURCE_FMT " %d pages @ "
+		   EFHW_BUFFER_ADDR_FMT, __FUNCTION__,
+		   EFRM_RESOURCE_PRI_ARG(iobrs->rs.rs_handle),
+		   iobrs->n_bufs, EFHW_BUFFER_ADDR(iobrs->buf_tbl_alloc.base,
+						   0));
+
+	/* Put it into the resource manager's table. */
+	rc = efrm_resource_manager_insert(&iobrs->rs);
+	if (rc != 0) {
+		if (atomic_dec_and_test(&iobrs->rs.rs_ref_count))
+			efrm_iobufset_resource_free(iobrs);
+		return rc;
+	}
+
+	*iobrs_out = iobrs;
+	return 0;
+
+fail5:
+	i = iobrs->n_bufs;
+fail4:
+	/* see comment on call to efhw_iopage_alloc above for a discussion
+	 * on use of efrm_nic_table.a_nic here */
+	if (iobrs->order == 0) {
+		while (i--) {
+			efhw_iopage_t *page = &iobrs->bufs[i];
+			efhw_iopage_free(efrm_nic_table.a_nic, page);
+		}
+	} else {
+		unsigned int j;
+		for (j = 0; j < i; j += iobrs->pages_per_contiguous_chunk) {
+			efhw_iopages_t iopages;
+
+			EFRM_ASSERT(j % iobrs->pages_per_contiguous_chunk
+				    == 0);
+			/* it is important that this is executed in increasing
+			 * page order because some implementations of
+			 * efhw_iopages_init_from_iopage() assume this */
+			efhw_iopages_init_from_iopage(&iopages,
+						      &iobrs->bufs[j],
+						      iobrs->order);
+			efhw_iopages_free(efrm_nic_table.a_nic, &iopages);
+		}
+	}
+	efrm_vi_resource_release(iobrs->evq);
+fail3:
+	if (object_size < PAGE_SIZE) {
+		kfree(iobrs);
+	} else {
+		vfree(iobrs);
+	}
+fail1:
+	return rc;
+}
+EXPORT_SYMBOL(efrm_iobufset_resource_alloc);
+
+static void iobufset_rm_dtor(struct efrm_resource_manager *rm)
+{
+	EFRM_ASSERT(&efrm_iobufset_manager->rm == rm);
+	kfifo_vfree(efrm_iobufset_manager->free_ids);
+}
+
+int
+efrm_create_iobufset_resource_manager(struct efrm_resource_manager **rm_out)
+{
+	int rc, max;
+
+	EFRM_ASSERT(rm_out);
+
+	efrm_iobufset_manager =
+	    kmalloc(sizeof(*efrm_iobufset_manager), GFP_KERNEL);
+	if (efrm_iobufset_manager == 0)
+		return -ENOMEM;
+	memset(efrm_iobufset_manager, 0, sizeof(*efrm_iobufset_manager));
+
+	/*
+	 * Bug 1145, 1370: We need to set initial size of both the resource
+	 * table and instance id table so they never need to grow as we
+	 * want to be allocate new iobufset at tasklet time. Lets make
+	 * a pessimistic guess at maximum number of iobufsets possible.
+	 * Could be less because
+	 *   - jumbo frames have same no of packets per iobufset BUT more
+	 *     pages per buffer
+	 *   - buffer table entries used independently of iobufsets by
+	 *     sendfile
+	 *
+	 * Based on TCP/IP stack setting of PKTS_PER_SET_S=5 ...
+	 *  - can't use this define here as it breaks the layering.
+	 */
+#define MIN_PAGES_PER_IOBUFSET  (1 << 4)
+
+	max = efrm_buffer_table_size() / MIN_PAGES_PER_IOBUFSET;
+	max = min_t(int, max, EFRM_IOBUFSET_MAX_NUM_INSTANCES);
+
+	rc = efrm_kfifo_id_ctor(&efrm_iobufset_manager->free_ids,
+				0, max, &efrm_iobufset_manager->rm.rm_lock);
+	if (rc != 0)
+		goto fail1;
+
+	rc = efrm_resource_manager_ctor(&efrm_iobufset_manager->rm,
+					iobufset_rm_dtor, "IOBUFSET",
+					EFRM_RESOURCE_IOBUFSET, max);
+	if (rc < 0)
+		goto fail2;
+
+	*rm_out = &efrm_iobufset_manager->rm;
+	return 0;
+
+fail2:
+	kfifo_vfree(efrm_iobufset_manager->free_ids);
+fail1:
+	EFRM_DO_DEBUG(memset(efrm_iobufset_manager, 0,
+			     sizeof(*efrm_iobufset_manager)));
+	kfree(efrm_iobufset_manager);
+	return rc;
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/iopage.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/iopage.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/iopage.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/iopage.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,101 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides Linux-specific implementation for iopage API used
+ * from efhw library.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/driver/resource/linux_efhw_nic.h>
+#include "kernel_compat.h"
+#include <ci/efhw/common_sysdep.h> /* for dma_addr_t */
+
+int efhw_iopage_alloc(struct efhw_nic *nic, efhw_iopage_t *p)
+{
+	struct linux_efhw_nic *lnic = linux_efhw_nic(nic);
+	dma_addr_t handle;
+	void *kva;
+
+	kva = efrm_pci_alloc_consistent(lnic->pci_dev, PAGE_SIZE,
+					&handle);
+	if (kva == 0)
+		return -ENOMEM;
+
+	EFHW_ASSERT((handle & ~PAGE_MASK) == 0);
+
+	memset((void *)kva, 0, PAGE_SIZE);
+	efhw_page_init_from_va(&p->p, kva);
+
+	p->dma_addr = handle;
+
+	return 0;
+}
+
+void efhw_iopage_free(struct efhw_nic *nic, efhw_iopage_t *p)
+{
+	struct linux_efhw_nic *lnic = linux_efhw_nic(nic);
+	EFHW_ASSERT(efhw_page_is_valid(&p->p));
+
+	efrm_pci_free_consistent(lnic->pci_dev, PAGE_SIZE,
+				 efhw_iopage_ptr(p), p->dma_addr);
+}
+
+int efhw_iopages_alloc(struct efhw_nic *nic, efhw_iopages_t *p, unsigned order)
+{
+	unsigned bytes = 1u << (order + PAGE_SHIFT);
+	struct linux_efhw_nic *lnic = linux_efhw_nic(nic);
+	dma_addr_t handle;
+	caddr_t addr;
+	int gfp_flag;
+
+	/* Set __GFP_COMP if available to make reference counting work.
+	 * This is recommended here:
+	 *   http://www.forbiddenweb.org/viewtopic.php?id=83167&page=4#348331
+	 */
+	gfp_flag = ((in_atomic() ? GFP_ATOMIC : GFP_KERNEL) | __GFP_COMP);
+	addr = efrm_dma_alloc_coherent(&lnic->pci_dev->dev, bytes, &handle,
+				       gfp_flag);
+	if (addr == NULL)
+		return -ENOMEM;
+
+	EFHW_ASSERT((handle & ~PAGE_MASK) == 0);
+
+	p->order = order;
+	p->dma_addr = handle;
+	p->kva = addr;
+
+	return 0;
+}
+
+void efhw_iopages_free(struct efhw_nic *nic, efhw_iopages_t *p)
+{
+	unsigned bytes = 1u << (p->order + PAGE_SHIFT);
+	struct linux_efhw_nic *lnic = linux_efhw_nic(nic);
+
+	efrm_dma_free_coherent(&lnic->pci_dev->dev, bytes,
+			       (void *)p->kva, p->dma_addr);
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/kernel_compat.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/kernel_compat.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/kernel_compat.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/kernel_compat.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,584 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides compatibility layer for various Linux kernel versions
+ * (starting from 2.6.9 RHEL kernel).
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#define IN_KERNEL_COMPAT_C
+#include <linux/types.h>
+#include <ci/efrm/debug.h>
+#include "kernel_compat.h"
+
+/* Set this to 1 to enable very basic counting of iopage(s) allocations, then
+ * call dump_iopage_counts() to show the number of current allocations of
+ * orders 0-7.
+ */
+#define EFRM_IOPAGE_COUNTS_ENABLED 0
+
+
+
+/* I admit that it's a bit ugly going straight to the field, but it
+ * seems easiest given that get_page followed by put_page on a page
+ * with PG_reserved set will increment the ref count on 2.6.14 and
+ * below, but not 2.6.15.  Also, RedHat have hidden put_page_testzero
+ * in a header file which produces warnings when compiled.  This
+ * doesn't agree with our use of -Werror.
+ */
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,5)
+# define page_count_field(pg)  ((pg)->count)
+#else
+# define page_count_field(pg)  ((pg)->_count)
+#endif
+
+#define inc_page_count(page)   atomic_inc(&page_count_field(page))
+#define dec_page_count(page)   atomic_dec(&page_count_field(page))
+
+/* Bug 5531: set_page_count doesn't work if the new page count is an
+ * expression. */
+#define ci_set_page_count(page, n) set_page_count(page, (n))
+
+  /* Bug 3965: Gak!  Reference counts just don't work on memory
+   * allocated through pci_alloc_consistent.  Different versions and
+   * architectures do different things.  There are several interacting
+   * bugs/features which have been described below and then summarised
+   * in a table for each kernel version.  For each feature, there is a
+   * question, a short description, a hint at how to examine the
+   * kernel code for this feature and a description of the keys in the
+   * table.
+   *
+   * A. Is PG_compound set on multi-page allocations?
+   *
+   *    When a multi-page allocation succeeds, the kernel sets the
+   *    reference count of the first page to one and the count of the
+   *    remaining pages to zero.  This is an immediate problem because
+   *    if these pages are mapped into user space, the VM will do
+   *    get_page followed by put_page, at which point the reference
+   *    count will return to zero and the page will be freed.
+   *    PG_compound was introduced in 2.6.0 and back-ported to rhel3
+   *    kernels.  When it is set, all the pages have a pointer to the
+   *    first page so that they can share the reference count.  If
+   *    PG_compound is set, calling get_page(pg+1) can change
+   *    page_count(pg).  It was originally set on all multi-page
+   *    allocations, but later only set if the __GFP_COMP flag was
+   *    provided to the allocator.
+   *
+   *    See mm/page_alloc.c
+   *      Does prep_compound_page get called when __GFP_COMP not set?
+   *
+   * Keys:
+   *    NotDef - prep_compound_page and PG_compound are not defined.
+   *    Comp   - prep_compound_page is called for any multi-page allocation.
+   *    Opt    - prep_compound_page is only called if __GFP_COMP is set.
+   *    OptInv - prep_compound_page is only called if __GFP_NO_COMP is not set.
+   *
+   * B. Are bounce buffers ever used to satisfy pci_alloc_consistent?
+   *    (x86_64 only)
+   *
+   *    2.6 kernels introduced bounce buffers on x86_64 machines to access
+   *    memory above 4G when using the DMA mapping API.  At some point,
+   *    code was added to allow pci_alloc_consistent/dma_alloc_coherent to
+   *    allocate memory from the bounce buffers if the general purpose
+   *    allocator produced memory which wasn't suitable.  Such memory can
+   *    be recognised by the PG_reserved bit being set.  At a later point,
+   *    the __GFP_DMA32 flag was added and used to restrict the allocator
+   *    to below 4G.  The effect of this later change was that 4G capable
+   *    cards would no longer get memory from the bounce buffers, although
+   *    a card which can address less than 4G might get memory from the
+   *    bounce buffers.
+   *
+   *    See dma_alloc_coherent or pci_alloc_consistent in
+   *    arch/x86_64/kernel/pci-gart.c or arch/x86/kernel/pci-dma_64.c
+   *      Is (gfp |= GFP_DMA32) before dma_alloc_pages?
+   *      Is swiotlb_alloc_coherent called?
+   *
+   * Keys:
+   *    NU     - bounce buffers are Never Used
+   *    Used   - bounce buffers are sometimes used
+   *
+   * C. Does munmap decrement the reference count of a PG_reserved page?
+   *
+   *    Originally, the munmap code would not decrement the reference count
+   *    of a page which had PG_reserved set.  At some point in the 2.6
+   *    series, VM_PFNMAP was introduced and could be set on a vma to
+   *    indicate that no pages in that vma should have the reference count
+   *    decremented (unless they are copy-on-write copies).  At that point,
+   *    the check for PG_reserved pages in the munmap code path was
+   *    removed.  Some hackery in vm_normal_page means that a VM_PFNMAP vma
+   *    must map contiguous physical pages.  As a result, such pages should
+   *    be mapped during mmap using remap_pfn_range (for an example, see
+   *    drivers/char/mem.c).
+   *
+   *    In 2.6 kernels: See release_pages in mm/swap.c
+   *      Does PageReserved get tested?
+   *    In 2.6 kernels: See mm/memory.c
+   *      Is VM_PFNMAP used?
+   *    In 2.4 kernels: See __free_pte in mm/memory.c
+   *      Does PageReserved get tested?
+   *    In 2.4 kernels: See __free_pages in mm/page_alloc.c
+   *      Does PageReserved get tested?
+   *
+   * Keys:
+   *    resv   - The reference count is not touched for PG_reserved pages.
+   *    pfnmap - The VM_PFNMAP flag is checked instead of PG_reserved.
+   *
+   * D. Does munmap honour the PG_compound bit?
+   *
+   *    When PG_compound was originally introduced, the munmap code path
+   *    didn't check it before decrementing the reference count on the
+   *    page.  As a result, the wrong reference count would be updated if a
+   *    PG_compound page was ever mapped into user space.
+   *
+   *    In 2.6 kernels: See release_pages in mm/swap.c
+   *      Does PageCompound get tested?
+   *    In 2.4 kernels: See __free_pages in mm/page_alloc.c
+   *      Does PageCompound get tested?
+   *
+   * Keys:
+   *    NotHon - The PG_compound bit isn't honoured by munmap.
+   *    Hon    - The PG_compound bit is honoured by munmap.
+   *
+   *                 OS      A       B       C       D
+   * 2.4.18                  NotDef  NU      resv    NotHon
+   * 2.4.29                  NotDef  NU      resv    NotHon
+   * 2.4.20-31.9     rhl9    NotDef  NU      resv    NotHon
+   *
+   * 2.4.21-4.EL     rhel3   Comp    NU      resv    Hon
+   * 2.4.21-15.EL    rhel3   Comp    NU      resv    Hon
+   * 2.4.21-32.EL    rhel3   Comp    NU      resv    Hon
+   * 2.4.21-40.EL    rhel3   Comp    NU      resv    Hon
+   *
+   * 2.6.0                   Comp    NU      resv    NotHon
+   *
+   * 2.6.5-7.97      sles9   OptInv  NU      resv    NotHon
+   * 2.6.9           rhel4   Opt     NU      resv    NotHon
+   *
+   * 2.6.11          fc4     ?       ?       ?       ?
+   * 2.6.12          fc4     Opt     Used    resv    NotHon
+   * 2.6.13                  Opt     Used    resv    NotHon
+   *
+   * 2.6.15                  Opt     NU      pfnmap  NotHon
+   *
+   * 2.6.16                  Opt     NU      pfnmap  Hon
+   * 2.6.16.9                Opt     NU      pfnmap  Hon
+   * 2.6.17.2                Opt     NU      pfnmap  Hon
+   * 2.6.24-rc7      k.org   Opt     NU      pfnmap  Hon
+   *
+   * This LKML thread gives some low down on mapping pages into user
+   * space and using DMA.
+   *  http://www.forbiddenweb.org/viewtopic.php?id=83167&page=1
+   *
+   * There is no problem with single page allocations (until some
+   * kernel hands us a PG_reserved page and expects us to use
+   * VM_PFNMAP on the vma).
+   *
+   * Bug 5450: Most kernels set the reference count to one on the
+   * first sub-page of a high-order page allocation and zero on
+   * subsequent sub-pages.  Some kernels, however, set the page count
+   * to one on all the sub-pages.  The SLES 9 range are affected, as
+   * are kernels built without CONFIG_MMU defined.
+   *
+   * Possible strategies for multi-page allocations:
+   *
+   * EFRM_MMAP_USE_COMPOUND
+   * 1. Allocate a compound page.  Reference counting should then work
+   *    on the whole allocation.  This is a good theory, but is broken
+   *    by bug/feature D (above).
+   *
+   * EFRM_MMAP_USE_SPLIT
+   * 2. Convert the multi-page allocation to many single page
+   *    allocations.  This involves incrementing the reference counts
+   *    and clearing PG_compound on all the pages (including the
+   *    first).  The references should be released _after_ calling
+   *    pci_free_consistent so that that call doesn't release the
+   *    memory.
+   *
+   * EFRM_MMAP_USE_INCREMENT
+   * 3. Increment the reference count on all the pages after
+   *    allocating and decrement them again before freeing.  This gets
+   *    round the zero reference count problem.  It doesn't handle the
+   *    case where someone else is holding a reference to one of our
+   *    pages when we free the pages, but we think VM_IO stops this
+   *    from happening.
+   */
+
+/* Should we use strategy 1?  This can be forced on us by the OS. */
+#if defined(PG_compound)
+#define EFRM_MMAP_USE_COMPOUND 1
+#else
+#define EFRM_MMAP_USE_COMPOUND 0
+#endif
+
+/* Should we use strategy 2?  This can be used even if strategy 1 is
+ * used. */
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,16)
+#define EFRM_MMAP_USE_SPLIT    1
+#else
+#define EFRM_MMAP_USE_SPLIT    0
+#endif
+
+/* Should we use strategy 3?  There's no point doing this if either
+ * strategy 1 or strategy 2 is used. */
+#if !EFRM_MMAP_USE_COMPOUND && !EFRM_MMAP_USE_SPLIT
+#error "We shouldn't have to use this strategy."
+#define EFRM_MMAP_USE_INCREMENT 1
+#else
+#define EFRM_MMAP_USE_INCREMENT 0
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,17)
+#define EFRM_MMAP_RESET_REFCNT  1
+#else
+#define EFRM_MMAP_RESET_REFCNT  0
+#endif
+
+/* NB. 2.6.17 has renamed SetPageCompound to __SetPageCompound and
+ * ClearPageCompound to __ClearPageCompound. */
+#if ((defined(PageCompound)        !=  defined(PG_compound)) ||	\
+     (defined(SetPageCompound)     !=  defined(PG_compound) &&	\
+      defined(__SetPageCompound)   !=  defined(PG_compound)) ||	\
+     (defined(ClearPageCompound)   !=  defined(PG_compound) &&	\
+      defined(__ClearPageCompound) !=  defined(PG_compound)) ||	\
+     (defined(__GFP_COMP)          && !defined(PG_compound)))
+#error Mismatch of defined page-flags.
+#endif
+
+extern int use_pci_alloc;	/* Use pci_alloc_consistent to alloc iopages */
+
+/****************************************************************************
+ *
+ * allocate a buffer suitable for DMA to/from the NIC
+ *
+ ****************************************************************************/
+
+static inline void pci_mmap_pages_hack_after_alloc(caddr_t kva, unsigned order)
+{
+	unsigned pfn = __pa(kva) >> PAGE_SHIFT;
+	struct page *start_pg = pfn_to_page(pfn);
+#if !defined(NDEBUG) || EFRM_MMAP_USE_SPLIT
+	struct page *end_pg = start_pg + (1 << order);
+	struct page *pg;
+#endif
+
+	/* Compound pages don't get created for order 0 pages and there's no
+	 * fixing up needs to be done. */
+	if (order == 0)
+		return;
+
+	/* If we've been given a reserved page then it must have come from
+	 * the bounce buffer pool. */
+	if (PageReserved(start_pg)) {
+#if defined(VM_PFNMAP) || !defined(__x86_64__)
+		/* Kernel allocated reserved pages when not expected */
+		BUG();
+#endif
+		return;
+	}
+
+	/* Check the page count and PG_compound bit. */
+#ifndef NDEBUG
+#  if defined(PG_compound)
+	EFRM_ASSERT(PageCompound(start_pg) == EFRM_MMAP_USE_COMPOUND);
+#  endif
+	EFRM_ASSERT(page_count(start_pg) == 1);
+
+	{
+		/* Some kernels have the page count field hold (ref_count-1)
+		 * rather than (ref_count).  This is so that decrementing the
+		 * reference count to "zero" causes the internal value to change
+		 * from 0 to -1 which sets the carry flag.  Other kernels store
+		 * the real reference count value in the obvious way.  We handle
+		 * this here by reading the reference count field of the first
+		 * page, which is always 1. */
+		int pg_count_zero;
+		pg_count_zero = atomic_read(&page_count_field(start_pg)) - 1;
+		for (pg = start_pg + 1; pg < end_pg; pg++) {
+			int pg_count;
+#  if defined(PG_compound)
+			EFRM_ASSERT(PageCompound(pg) == EFRM_MMAP_USE_COMPOUND);
+#  endif
+
+			/* Bug 5450: Some kernels initialise the page count
+			 * to one for pages other than the first and some
+			 * leave it at zero.  We allow either behaviour
+			 * here, but disallow anything strange.  Newer
+			 * kernels only define set_page_count in an
+			 * internal header file, so we have to make do with
+			 * incrementing and decrementing the reference
+			 * count.  Fortunately, those kernels don't set the
+			 * reference count to one on all the pages. */
+			pg_count = atomic_read(&page_count_field(pg));
+#  if EFRM_MMAP_RESET_REFCNT
+			if (pg_count != pg_count_zero)
+				EFRM_ASSERT(pg_count == pg_count_zero + 1);
+#  else
+			EFRM_ASSERT(pg_count == pg_count_zero);
+#  endif
+		}
+	}
+#endif
+
+	/* Split the multi-page allocation if necessary. */
+#if EFRM_MMAP_USE_SPLIT
+	for (pg = start_pg; pg < end_pg; pg++) {
+
+		/* This is no longer a compound page. */
+#  if EFRM_MMAP_USE_COMPOUND
+		ClearPageCompound(pg);
+		EFRM_ASSERT(PageCompound(pg) == 0);
+#  endif
+
+#  ifndef NDEBUG
+		{
+			int pg_count = page_count(pg);
+			/* Bug 5450: The page count can be zero or one here. */
+			if (pg == start_pg) {
+				EFRM_ASSERT(pg_count == 1);
+			} else {
+#    if EFRM_MMAP_RESET_REFCNT
+				if (pg_count != 0)
+					EFRM_ASSERT(pg_count == 1);
+#    else
+				EFRM_ASSERT(pg_count == 0);
+#    endif
+			}
+		}
+#  endif
+
+		/* Get a reference which will be released after the pages have
+		 * been passed back to pci_free_consistent. */
+#  if EFRM_MMAP_RESET_REFCNT
+		/* Bug 5450: Reset the reference count since the count might
+		 * already be 1. */
+		ci_set_page_count(pg, (pg == start_pg) ? 2 : 1);
+#  else
+		get_page(pg);
+#  endif
+	}
+#endif
+
+	/* Fudge the reference count if necessary. */
+#if EFRM_MMAP_USE_INCREMENT
+	for (pg = start_pg; pg < end_pg; pg++)
+		inc_page_count(pg);
+#endif
+}
+
+static inline void pci_mmap_pages_hack_before_free(caddr_t kva, unsigned order)
+{
+#if EFRM_MMAP_USE_INCREMENT || !defined(NDEBUG)
+	/* Drop the references taken in pci_mmap_pages_hack_after_alloc */
+	unsigned pfn = __pa(kva) >> PAGE_SHIFT;
+	struct page *start_pg = pfn_to_page(pfn);
+	struct page *end_pg = start_pg + (1 << order);
+	struct page *pg;
+
+	/* Compound pages don't get created for order 0 pages and there's no
+	 * fixing up needs to be done. */
+	if (order == 0)
+		return;
+
+	if (PageReserved(start_pg))
+		return;
+
+#  if EFRM_MMAP_USE_INCREMENT
+	for (pg = start_pg; pg < end_pg; pg++)
+		dec_page_count(pg);
+#  endif
+
+#if !defined(NDEBUG)
+	EFRM_ASSERT(page_count(start_pg) == 1+EFRM_MMAP_USE_SPLIT);
+
+#  if EFRM_MMAP_USE_COMPOUND && !EFRM_MMAP_USE_SPLIT
+	for (pg = start_pg; pg < end_pg; pg++)
+		EFRM_ASSERT(PageCompound(pg));
+#  else
+	for (pg = start_pg+1; pg < end_pg; pg++) {
+		unsigned exp_pg_count = EFRM_MMAP_USE_SPLIT;
+		/* NB.  If this assertion fires, either we've messed up the
+		 * page counting or someone is holding on to a reference.
+		 */
+		EFRM_ASSERT(page_count(pg) == exp_pg_count);
+	}
+#  endif
+#endif
+
+#endif
+}
+
+static inline void pci_mmap_pages_hack_after_free(caddr_t kva, unsigned order)
+{
+#if EFRM_MMAP_USE_SPLIT
+	/* Drop the references taken in pci_mmap_pages_hack_after_alloc */
+	unsigned pfn = __pa(kva) >> PAGE_SHIFT;
+	struct page *start_pg = pfn_to_page(pfn);
+	struct page *end_pg = start_pg + (1 << order);
+	struct page *pg;
+
+	/* Compound pages don't get created for order 0 pages and there's no
+	 * fixing up needs to be done. */
+	if (order == 0)
+		return;
+
+	if (PageReserved(start_pg))
+		return;
+
+	for (pg = start_pg; pg < end_pg; pg++) {
+		EFRM_ASSERT(page_count(pg) == 1);
+		put_page(pg);
+	}
+#endif
+}
+
+
+#if EFRM_IOPAGE_COUNTS_ENABLED
+
+static int iopage_counts[8];
+
+void dump_iopage_counts(void)
+{
+	EFRM_NOTICE("iopage counts: %d %d %d %d %d %d %d %d", iopage_counts[0],
+		    iopage_counts[1], iopage_counts[2], iopage_counts[3],
+		    iopage_counts[4], iopage_counts[5], iopage_counts[6],
+		    iopage_counts[7]);
+}
+
+#endif
+
+
+
+/*********** pci_alloc_consistent / pci_free_consistent ***********/
+
+void *efrm_dma_alloc_coherent(struct device *dev, size_t size,
+			      dma_addr_t *dma_addr, int flag)
+{
+	struct pci_dev *pci_dev;
+	void *ptr;
+	unsigned order;
+	EFRM_IOMMU_DECL;
+
+	order = __ffs(size/PAGE_SIZE);
+	EFRM_ASSERT(size == (PAGE_SIZE<<order));
+
+	/* NB. The caller may well set __GFP_COMP.  However we can't
+	 * rely on this working on older kernels.  2.6.9 only acts on
+	 * __GFP_COMP if CONFIG_HUGETLB_PAGE is defined.  If the flag
+	 * did have an effect then PG_compound will be set on the
+	 * pages. */
+
+	if (use_pci_alloc) {
+		/* Can't take a spinlock here since the allocation can
+		 * block. */
+		ptr = dma_alloc_coherent(dev, size, dma_addr, flag);
+		if (ptr == NULL)
+			return ptr;
+	} else {
+#ifdef CONFIG_SWIOTLB		/* BUG1340 */
+		if (swiotlb) {
+			EFRM_ERR("%s: This kernel is using DMA bounce "
+				 "buffers.  Please upgrade kernel to "
+				 "linux2.6 or reduce the amount of RAM "
+				 "with mem=XXX.", __FUNCTION__);
+			return NULL;
+		}
+#endif
+		ptr = (void *)__get_free_pages(flag, order);
+
+		if (ptr == NULL)
+			return NULL;
+
+		EFRM_IOMMU_LOCK();
+		pci_dev = container_of(dev, struct pci_dev, dev);
+		*dma_addr = pci_map_single(pci_dev, ptr, size,
+					   PCI_DMA_BIDIRECTIONAL);
+		EFRM_IOMMU_UNLOCK();
+		if (pci_dma_mapping_error(*dma_addr)) {
+			free_pages((unsigned long)ptr, order);
+			return NULL;
+		}
+	}
+
+#ifndef CONFIG_IA64
+	pci_mmap_pages_hack_after_alloc(ptr, order);
+#endif
+
+#if EFRM_IOPAGE_COUNTS_ENABLED
+	if (order < 8)
+		iopage_counts[order]++;
+	else
+		EFRM_ERR("Huge iopages alloc (order=%d) ??? (not counted)",
+			 order);
+#endif
+
+	return ptr;
+}
+
+void efrm_dma_free_coherent(struct device *dev, size_t size,
+			    void *ptr, dma_addr_t dma_addr)
+{
+	struct pci_dev *pci_dev;
+	unsigned order;
+	EFRM_IOMMU_DECL;
+
+	order = __ffs(size/PAGE_SIZE);
+	EFRM_ASSERT(size == (PAGE_SIZE<<order));
+
+#if EFRM_IOPAGE_COUNTS_ENABLED
+	if (order < 8)
+		--iopage_counts[order];
+	else
+		EFRM_ERR("Huge iopages free (order=%d) ??? (not counted)",
+			 order);
+#endif
+#ifndef CONFIG_IA64
+	pci_mmap_pages_hack_before_free(ptr, order);
+#endif
+	if (use_pci_alloc) {
+		EFRM_IOMMU_LOCK();
+		dma_free_coherent(dev, size, ptr, dma_addr);
+		EFRM_IOMMU_UNLOCK();
+	} else {
+		pci_dev = container_of(dev, struct pci_dev, dev);
+		EFRM_IOMMU_LOCK();
+		efrm_pci_unmap_single(pci_dev, dma_addr, size,
+				      PCI_DMA_BIDIRECTIONAL);
+		EFRM_IOMMU_UNLOCK();
+
+		free_pages((unsigned long)ptr, order);
+	}
+
+#ifndef CONFIG_IA64
+	pci_mmap_pages_hack_after_free(ptr, order);
+#endif
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/kernel_compat.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/kernel_compat.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/kernel_compat.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/kernel_compat.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,239 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides compatibility layer for various Linux kernel versions
+ * (starting from 2.6.9 RHEL kernel).
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef DRIVER_LINUX_RESOURCE_KERNEL_COMPAT_H
+#define DRIVER_LINUX_RESOURCE_KERNEL_COMPAT_H
+
+#include <linux/version.h>
+
+/********* wait_for_completion_timeout() ********************/
+#include <linux/sched.h>
+
+/* RHEL_RELEASE_CODE from linux/version.h is only defined for 2.6.9-55EL
+ * UTS_RELEASE is unfortunately unusable
+ * Really only need this fix for <2.6.9-34EL
+ */
+#if (LINUX_VERSION_CODE < KERNEL_VERSION(2,6,11)) && \
+	!defined(RHEL_RELEASE_CODE)
+
+static inline unsigned long fastcall __sched
+efrm_wait_for_completion_timeout(struct completion *x, unsigned long timeout)
+{
+	might_sleep();
+
+	spin_lock_irq(&x->wait.lock);
+	if (!x->done) {
+		DECLARE_WAITQUEUE(wait, current);
+
+		wait.flags |= WQ_FLAG_EXCLUSIVE;
+		__add_wait_queue_tail(&x->wait, &wait);
+		do {
+			__set_current_state(TASK_UNINTERRUPTIBLE);
+			spin_unlock_irq(&x->wait.lock);
+			timeout = schedule_timeout(timeout);
+			spin_lock_irq(&x->wait.lock);
+			if (!timeout) {
+				__remove_wait_queue(&x->wait, &wait);
+				goto out;
+			}
+		} while (!x->done);
+		__remove_wait_queue(&x->wait, &wait);
+	}
+	x->done--;
+out:
+	spin_unlock_irq(&x->wait.lock);
+	return timeout;
+}
+
+#  ifdef wait_for_completion_timeout
+#    undef wait_for_completion_timeout
+#  endif
+#  define wait_for_completion_timeout efrm_wait_for_completion_timeout
+
+#endif
+
+/********* pci_map_*() ********************/
+
+#include <linux/pci.h>
+
+/* Bug 4560: Some kernels leak IOMMU entries under heavy load.  Use a
+ * spinlock to serialise access where possible to alleviate the
+ * problem.
+ *
+ * NB. This is duplicated in the net driver.  Please keep in sync. */
+#if ((LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,0)) && \
+     (LINUX_VERSION_CODE < KERNEL_VERSION(2,6,17)) && \
+      defined(__x86_64__) && defined(CONFIG_SMP))
+
+#define EFRM_HAVE_IOMMU_LOCK 1
+
+#if ((LINUX_VERSION_CODE == KERNEL_VERSION(2,6,5)) &&	\
+      defined(CONFIG_SUSE_KERNEL))
+#define EFRM_NEED_ALTERNATE_MAX_PFN 1
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,15)
+#if defined(CONFIG_GART_IOMMU)
+#define EFRM_NO_IOMMU no_iommu
+#else
+#define EFRM_NO_IOMMU 1
+#endif
+#else
+#define EFRM_NO_IOMMU 0
+#endif
+
+/* Set to 0 if we should never use the lock.  Set to 1 if we should
+ * automatically determine if we should use the lock.  Set to 2 if we
+ * should always use the lock. */
+extern unsigned int efx_use_iommu_lock;
+/* Defined in the net driver. */
+extern spinlock_t efx_iommu_lock;
+/* Non-zero if there is a card which needs the lock. */
+extern int efrm_need_iommu_lock;
+
+/* The IRQ state is needed if the lock is being used.  The flag is
+ * cached to ensure that every lock is followed by an unlock, even
+ * if the global flag changes in the middle of the operation. */
+
+#define EFRM_IOMMU_DECL				\
+	unsigned long efx_iommu_irq_state = 0;	\
+	int efx_iommu_using_lock;
+#define EFRM_IOMMU_LOCK()						\
+	do {								\
+		efx_iommu_using_lock = (efx_use_iommu_lock &&		\
+					(efrm_need_iommu_lock ||	\
+					 efx_use_iommu_lock >= 2));	\
+		if (efx_iommu_using_lock)				\
+		spin_lock_irqsave(&efx_iommu_lock, efx_iommu_irq_state);\
+	} while (0)
+#define EFRM_IOMMU_UNLOCK()						\
+	do {								\
+		if (efx_iommu_using_lock)				\
+		spin_unlock_irqrestore(&efx_iommu_lock,			\
+				       efx_iommu_irq_state);		\
+	} while (0)
+
+#else /* defined(__x86_64__) && defined(CONFIG_SMP) */
+
+#define EFRM_HAVE_IOMMU_LOCK 0
+#define EFRM_IOMMU_DECL
+#define EFRM_IOMMU_LOCK()    do {} while (0)
+#define EFRM_IOMMU_UNLOCK()  do {} while (0)
+
+#endif
+
+static inline dma_addr_t efrm_pci_map_single(struct pci_dev *hwdev, void *ptr,
+					     size_t size, int direction)
+{
+	dma_addr_t dma_addr;
+	EFRM_IOMMU_DECL;
+
+	EFRM_IOMMU_LOCK();
+	dma_addr = pci_map_single(hwdev, ptr, size, direction);
+	EFRM_IOMMU_UNLOCK();
+
+	return dma_addr;
+}
+
+static inline void efrm_pci_unmap_single(struct pci_dev *hwdev,
+					 dma_addr_t dma_addr, size_t size,
+					 int direction)
+{
+	EFRM_IOMMU_DECL;
+
+	EFRM_IOMMU_LOCK();
+	pci_unmap_single(hwdev, dma_addr, size, direction);
+	EFRM_IOMMU_UNLOCK();
+}
+
+static inline dma_addr_t efrm_pci_map_page(struct pci_dev *hwdev,
+					   struct page *page,
+					   unsigned long offset, size_t size,
+					   int direction)
+{
+	dma_addr_t dma_addr;
+	EFRM_IOMMU_DECL;
+
+	EFRM_IOMMU_LOCK();
+	dma_addr = pci_map_page(hwdev, page, offset, size, direction);
+	EFRM_IOMMU_UNLOCK();
+
+	return dma_addr;
+}
+
+static inline void efrm_pci_unmap_page(struct pci_dev *hwdev,
+				       dma_addr_t dma_addr, size_t size,
+				       int direction)
+{
+	EFRM_IOMMU_DECL;
+
+	EFRM_IOMMU_LOCK();
+	pci_unmap_page(hwdev, dma_addr, size, direction);
+	EFRM_IOMMU_UNLOCK();
+}
+
+#ifndef IN_KERNEL_COMPAT_C
+#  ifndef __GFP_COMP
+#    define __GFP_COMP 0
+#  endif
+#  ifndef __GFP_ZERO
+#    define __GFP_ZERO 0
+#  endif
+#endif
+
+extern void *efrm_dma_alloc_coherent(struct device *dev, size_t size,
+				     dma_addr_t *dma_addr, int flag);
+
+extern void efrm_dma_free_coherent(struct device *dev, size_t size,
+				   void *ptr, dma_addr_t dma_addr);
+
+static inline void *efrm_pci_alloc_consistent(struct pci_dev *hwdev,
+					      size_t size,
+					      dma_addr_t *dma_addr)
+{
+	return efrm_dma_alloc_coherent(&hwdev->dev, size, dma_addr,
+				       GFP_ATOMIC);
+}
+
+static inline void efrm_pci_free_consistent(struct pci_dev *hwdev, size_t size,
+					    void *ptr, dma_addr_t dma_addr)
+{
+	efrm_dma_free_coherent(&hwdev->dev, size, ptr, dma_addr);
+}
+
+#endif /* DRIVER_LINUX_RESOURCE_KERNEL_COMPAT_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/kernel_proc.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/kernel_proc.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/kernel_proc.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/kernel_proc.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,111 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains /proc/driver/sfc_resource/ implementation.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/efrm/debug.h>
+#include <ci/efrm/nic_table.h>
+#include <ci/efrm/driver_private.h>
+#include <linux/proc_fs.h>
+
+/** Top level directory for sfc specific stats **/
+static struct proc_dir_entry *efrm_proc_root; /* = NULL */
+
+static int
+efrm_resource_read_proc(char *buf, char **start, off_t offset, int count,
+			int *eof, void *data);
+
+int efrm_install_proc_entries(void)
+{
+	/* create the top-level directory for etherfabric specific stuff */
+	efrm_proc_root = proc_mkdir("sfc_resource", proc_root_driver);
+	if (!efrm_proc_root)
+		return -ENOMEM;
+	EFRM_ASSERT(efrm_proc_root);
+
+	if (create_proc_read_entry("resources", 0, efrm_proc_root,
+				   efrm_resource_read_proc, 0) == NULL) {
+		EFRM_WARN("%s: Unable to create /proc/drivers/sfc_resource/"
+			  "resources", __FUNCTION__);
+	}
+	return 0;
+}
+
+void efrm_uninstall_proc_entries(void)
+{
+	EFRM_ASSERT(efrm_proc_root);
+	remove_proc_entry("resources", efrm_proc_root);
+	remove_proc_entry("sfc_resource", proc_root_driver);
+	efrm_proc_root = NULL;
+}
+
+/****************************************************************************
+ *
+ * /proc/drivers/sfc/resources
+ *
+ ****************************************************************************/
+
+#define EFRM_PROC_PRINTF(buf, len, fmt, ...)				\
+	do {								\
+		if (count - len > 0)					\
+			len += snprintf(buf+len, count-len, (fmt),	\
+					__VA_ARGS__);			\
+	} while (0)
+
+static int
+efrm_resource_read_proc(char *buf, char **start, off_t offset, int count,
+			int *eof, void *data)
+{
+	irq_flags_t lock_flags;
+	int len = 0;
+	int type;
+	struct efrm_resource_manager *rm;
+
+	for (type = 0; type < EFRM_RESOURCE_NUM; type++) {
+		rm = efrm_rm_table[type];
+		if (rm == NULL)
+			continue;
+
+		EFRM_PROC_PRINTF(buf, len, "*** %s ***\n", rm->rm_name);
+
+		spin_lock_irqsave(&rm->rm_lock, lock_flags);
+		EFRM_PROC_PRINTF(buf, len, "current = %u\n", rm->rm_resources);
+		EFRM_PROC_PRINTF(buf, len, "    max = %u\n\n",
+				 rm->rm_resources_hiwat);
+		spin_unlock_irqrestore(&rm->rm_lock, lock_flags);
+	}
+
+	return count ? strlen(buf) : 0;
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/kfifo.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/kfifo.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/kfifo.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/kfifo.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,212 @@
+/*
+ * A simple kernel FIFO implementation.
+ *
+ * Copyright (C) 2004 Stelian Pop <stelian@popies.net>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ *
+ */
+
+/*
+ * This file is stolen from the Linux kernel sources
+ * (linux-2.6.22/kernel/kfifo.c) into sfc_resource driver.
+ * It should be used for old kernels without kfifo implementation.
+ * Most part of linux/kfifo.h is incorporated into
+ * ci/efrm/sysdep_linux.h.
+ */
+#include <ci/efrm/sysdep_linux.h>
+#ifdef HAS_NO_KFIFO
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/err.h>
+/*#include <linux/kfifo.h>*/
+
+/**
+ * kfifo_init - allocates a new FIFO using a preallocated buffer
+ * @buffer: the preallocated buffer to be used.
+ * @size: the size of the internal buffer, this have to be a power of 2.
+ * @gfp_mask: get_free_pages mask, passed to kmalloc()
+ * @lock: the lock to be used to protect the fifo buffer
+ *
+ * Do NOT pass the kfifo to kfifo_free() after use! Simply free the
+ * &struct kfifo with kfree().
+ */
+struct kfifo *kfifo_init(unsigned char *buffer, unsigned int size,
+			 gfp_t gfp_mask, spinlock_t * lock)
+{
+	struct kfifo *fifo;
+
+	/* size must be a power of 2 */
+	BUG_ON(size & (size - 1));
+
+	fifo = kmalloc(sizeof(struct kfifo), gfp_mask);
+	if (!fifo)
+		return ERR_PTR(-ENOMEM);
+
+	fifo->buffer = buffer;
+	fifo->size = size;
+	fifo->in = fifo->out = 0;
+	fifo->lock = lock;
+
+	return fifo;
+}
+
+EXPORT_SYMBOL(kfifo_init);
+
+/**
+ * kfifo_alloc - allocates a new FIFO and its internal buffer
+ * @size: the size of the internal buffer to be allocated.
+ * @gfp_mask: get_free_pages mask, passed to kmalloc()
+ * @lock: the lock to be used to protect the fifo buffer
+ *
+ * The size will be rounded-up to a power of 2.
+ */
+struct kfifo *kfifo_alloc(unsigned int size, gfp_t gfp_mask, spinlock_t * lock)
+{
+	unsigned char *buffer;
+	struct kfifo *ret;
+
+	/*
+	 * round up to the next power of 2, since our 'let the indices
+	 * wrap' tachnique works only in this case.
+	 */
+	if (size & (size - 1)) {
+		BUG_ON(size > 0x80000000);
+		size = roundup_pow_of_two(size);
+	}
+
+	buffer = kmalloc(size, gfp_mask);
+	if (!buffer)
+		return ERR_PTR(-ENOMEM);
+
+	ret = kfifo_init(buffer, size, gfp_mask, lock);
+
+	if (IS_ERR(ret))
+		kfree(buffer);
+
+	return ret;
+}
+
+EXPORT_SYMBOL(kfifo_alloc);
+
+/**
+ * kfifo_free - frees the FIFO
+ * @fifo: the fifo to be freed.
+ */
+void kfifo_free(struct kfifo *fifo)
+{
+	kfree(fifo->buffer);
+	kfree(fifo);
+}
+
+EXPORT_SYMBOL(kfifo_free);
+
+/**
+ * __kfifo_put - puts some data into the FIFO, no locking version
+ * @fifo: the fifo to be used.
+ * @buffer: the data to be added.
+ * @len: the length of the data to be added.
+ *
+ * This function copies at most @len bytes from the @buffer into
+ * the FIFO depending on the free space, and returns the number of
+ * bytes copied.
+ *
+ * Note that with only one concurrent reader and one concurrent
+ * writer, you don't need extra locking to use these functions.
+ */
+unsigned int
+__kfifo_put(struct kfifo *fifo, unsigned char *buffer, unsigned int len)
+{
+	unsigned int l;
+
+	len = min(len, fifo->size - fifo->in + fifo->out);
+
+	/*
+	 * Ensure that we sample the fifo->out index -before- we
+	 * start putting bytes into the kfifo.
+	 */
+
+	smp_mb();
+
+	/* first put the data starting from fifo->in to buffer end */
+	l = min(len, fifo->size - (fifo->in & (fifo->size - 1)));
+	memcpy(fifo->buffer + (fifo->in & (fifo->size - 1)), buffer, l);
+
+	/* then put the rest (if any) at the beginning of the buffer */
+	memcpy(fifo->buffer, buffer + l, len - l);
+
+	/*
+	 * Ensure that we add the bytes to the kfifo -before-
+	 * we update the fifo->in index.
+	 */
+
+	smp_wmb();
+
+	fifo->in += len;
+
+	return len;
+}
+
+EXPORT_SYMBOL(__kfifo_put);
+
+/**
+ * __kfifo_get - gets some data from the FIFO, no locking version
+ * @fifo: the fifo to be used.
+ * @buffer: where the data must be copied.
+ * @len: the size of the destination buffer.
+ *
+ * This function copies at most @len bytes from the FIFO into the
+ * @buffer and returns the number of copied bytes.
+ *
+ * Note that with only one concurrent reader and one concurrent
+ * writer, you don't need extra locking to use these functions.
+ */
+unsigned int
+__kfifo_get(struct kfifo *fifo, unsigned char *buffer, unsigned int len)
+{
+	unsigned int l;
+
+	len = min(len, fifo->in - fifo->out);
+
+	/*
+	 * Ensure that we sample the fifo->in index -before- we
+	 * start removing bytes from the kfifo.
+	 */
+
+	smp_rmb();
+
+	/* first get the data from fifo->out until the end of the buffer */
+	l = min(len, fifo->size - (fifo->out & (fifo->size - 1)));
+	memcpy(buffer, fifo->buffer + (fifo->out & (fifo->size - 1)), l);
+
+	/* then get the rest (if any) from the beginning of the buffer */
+	memcpy(buffer + l, fifo->buffer, len - l);
+
+	/*
+	 * Ensure that we remove the bytes from the kfifo -before-
+	 * we update the fifo->out index.
+	 */
+
+	smp_mb();
+
+	fifo->out += len;
+
+	return len;
+}
+
+EXPORT_SYMBOL(__kfifo_get);
+#endif
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/linux_resource_internal.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/linux_resource_internal.h
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/linux_resource_internal.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/linux_resource_internal.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,75 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains Linux-specific API internal for the resource driver.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __LINUX_RESOURCE_INTERNAL__
+#define __LINUX_RESOURCE_INTERNAL__
+
+#include <ci/driver/resource/linux_efhw_nic.h>
+#include <ci/efrm/debug.h>
+#include <ci/efrm/driver_private.h>
+#include <ci/driver/efab/hardware.h>
+
+
+/*! Linux specific EtherFabric initialisation */
+extern int
+linux_efrm_nic_ctor(struct linux_efhw_nic *, struct pci_dev *,
+		    spinlock_t *reg_lock,
+		    unsigned nic_flags, unsigned nic_options);
+
+/*! Linux specific EtherFabric initialisation */
+extern void linux_efrm_nic_dtor(struct linux_efhw_nic *);
+
+/*! Linux specific EtherFabric initialisation -- interrupt registration */
+extern int linux_efrm_irq_ctor(struct linux_efhw_nic *);
+
+/*! Linux specific  EtherFabric initialisation -- interrupt deregistration */
+extern void linux_efrm_irq_dtor(struct linux_efhw_nic *);
+
+extern int  efrm_driverlink_register(void);
+extern void efrm_driverlink_unregister(void);
+
+extern int
+efrm_nic_add(struct pci_dev *dev, unsigned int opts, const uint8_t *mac_addr,
+	     struct linux_efhw_nic **lnic_out, spinlock_t *reg_lock,
+	     int bt_min, int bt_max, const struct vi_resource_dimensions *);
+extern void efrm_nic_del(struct linux_efhw_nic *);
+
+
+extern int efrm_install_proc_entries(void);
+extern void efrm_uninstall_proc_entries(void);
+
+#endif  /* __LINUX_RESOURCE_INTERNAL__ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/Makefile linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/Makefile
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,15 @@
+obj-$(CONFIG_SFC_RESOURCE) := sfc_resource.o
+
+EXTRA_CFLAGS += -D__CI_HARDWARE_CONFIG_FALCON__
+EXTRA_CFLAGS += -D__ci_driver__ 
+EXTRA_CFLAGS += -Werror
+EXTRA_CFLAGS += -Idrivers/net/sfc -Idrivers/net/sfc/sfc_resource
+
+sfc_resource-objs := resource_driver.o iopage.o efx_vi_shm.o \
+	driverlink_new.o kernel_proc.o kfifo.o \
+	nic.o eventq.o falcon.o falcon_mac.o falcon_hash.o \
+	assert_valid.o buddy.o buffer_table.o filter_resource.o \
+	iobufset_resource.o resource_manager.o resources.o \
+	vi_resource_alloc.o vi_resource_event.o vi_resource_flush.o \
+	vi_resource_manager.o driver_object.o kernel_compat.o
+
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/nic.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/nic.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/nic.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/nic.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,190 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains EtherFabric Generic NIC instance (init, interrupts,
+ * etc)
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/efhw/debug.h>
+#include <ci/driver/efab/hardware.h>
+#include <ci/efhw/falcon.h>
+#include <ci/efhw/nic.h>
+#include <ci/efhw/eventq.h>
+
+
+int efhw_device_type_init(struct efhw_device_type *dt,
+			  int vendor_id, int device_id,
+			  int class_revision)
+{
+	if (vendor_id != 0x1924)
+		return 0;
+
+	switch (device_id) {
+	case 0x0703:
+	case 0x6703:
+		dt->arch = EFHW_ARCH_FALCON;
+		dt->variant = 'A';
+		switch (class_revision) {
+		case 0:
+			dt->revision = 0;
+			break;
+		case 1:
+			dt->revision = 1;
+			break;
+		default:
+			return 0;
+		}
+		break;
+	case 0x0710:
+		dt->arch = EFHW_ARCH_FALCON;
+		dt->variant = 'B';
+		switch (class_revision) {
+		case 2:
+			dt->revision = 0;
+			break;
+		default:
+			return 0;
+		}
+		break;
+	default:
+		return 0;
+	}
+
+	return 1;
+}
+
+
+/*--------------------------------------------------------------------
+ *
+ * NIC Initialisation
+ *
+ *--------------------------------------------------------------------*/
+
+/* make this separate from initialising data structure
+** to allow this to be called at a later time once we can access PCI
+** config space to find out what hardware we have
+*/
+void efhw_nic_init(struct efhw_nic *nic, unsigned flags, unsigned options,
+		   struct efhw_device_type dev_type)
+{
+	int i;
+
+	nic->devtype = dev_type;
+	nic->flags = flags;
+	nic->options = options;
+	nic->bar_ioaddr = 0;
+	spin_lock_init(&nic->the_reg_lock);
+	nic->reg_lock = &nic->the_reg_lock;
+	nic->mtu = 1500 + ETH_HLEN;
+
+	for (i = 0; i < EFHW_KEVENTQ_MAX; i++)
+		nic->irq_unit[i] = EFHW_IRQ_UNIT_UNUSED;
+
+	switch (nic->devtype.arch) {
+	case EFHW_ARCH_FALCON:
+		nic->evq_sizes = 512 | 1024 | 2048 | 4096 | 8192 |
+			16384 | 32768;
+		nic->txq_sizes = 512 | 1024 | 2048 | 4096;
+		nic->rxq_sizes = 512 | 1024 | 2048 | 4096;
+		nic->efhw_func = &falcon_char_functional_units;
+		nic->ctr_ap_bytes = EFHW_64M;
+		switch (nic->devtype.variant) {
+		case 'A':
+			nic->ctr_ap_bar = FALCON_S_CTR_AP_BAR;
+			break;
+		case 'B':
+			nic->flags |= NIC_FLAG_NO_INTERRUPT;
+			nic->ctr_ap_bar = FALCON_P_CTR_AP_BAR;
+			break;
+		default:
+			EFHW_ASSERT(0);
+			break;
+		}
+		break;
+	default:
+		EFHW_ASSERT(0);
+		break;
+	}
+}
+
+
+void efhw_nic_close_interrupts(struct efhw_nic *nic)
+{
+	int i;
+
+	EFHW_ASSERT(nic);
+	if (!efhw_nic_have_hw(nic))
+		return;
+
+	EFHW_ASSERT(efhw_nic_have_hw(nic));
+
+	for (i = 0; i < EFHW_KEVENTQ_MAX; i++) {
+		if (nic->irq_unit[i] != EFHW_IRQ_UNIT_UNUSED)
+			efhw_nic_interrupt_disable(nic, i);
+	}
+}
+
+void efhw_nic_dtor(struct efhw_nic *nic)
+{
+	EFHW_ASSERT(nic);
+
+	/* Check that we have functional units because the software only
+	 * driver doesn't initialise anything hardware related any more */
+
+#ifndef __ci_ul_driver__
+	/* close interrupts is called first because the act of deregistering
+	   the driver could cause this driver to change from master to slave
+	   and hence the implicit interrupt mappings would be wrong */
+
+	EFHW_TRACE("%s: functional units ... ", __FUNCTION__);
+
+	if (efhw_nic_have_functional_units(nic)) {
+		efhw_nic_close_interrupts(nic);
+		efhw_nic_close_hardware(nic);
+	}
+	EFHW_TRACE("%s: functional units ... done", __FUNCTION__);
+#endif
+
+	/* destroy event queues */
+	EFHW_TRACE("%s: event queues ... ", __FUNCTION__);
+
+#ifndef __ci_ul_driver__
+	{
+		int i;
+		for (i = 0; i < EFHW_KEVENTQ_MAX; ++i)
+			if (nic->evq[i].evq_mask)
+				efhw_keventq_dtor(nic, &nic->evq[i]);
+	}
+#endif
+
+	EFHW_TRACE("%s: event queues ... done", __FUNCTION__);
+
+	spin_lock_destroy(&nic->the_reg_lock);
+
+	EFHW_TRACE("%s: DONE", __FUNCTION__);
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/resource_driver.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/resource_driver.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/resource_driver.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/resource_driver.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,640 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains main driver entry points.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include "linux_resource_internal.h"
+#include "kernel_compat.h"
+#include <ci/efrm/nic_table.h>
+#include <ci/driver/resource/efx_vi.h>
+#include <ci/efhw/eventq.h>
+#include <ci/efhw/nic.h>
+#include <ci/efrm/buffer_table.h>
+#include <ci/efrm/vi_resource_private.h>
+#include <ci/efrm/driver_private.h>
+
+#if EFRM_HAVE_IOMMU_LOCK
+#ifdef EFRM_NEED_ALTERNATE_MAX_PFN
+extern unsigned long blk_max_pfn;
+#define max_pfn blk_max_pfn
+#else
+#include <linux/bootmem.h>
+#endif
+#endif
+
+MODULE_AUTHOR("Solarflare Communications");
+MODULE_LICENSE("GPL");
+
+static struct efhw_ev_handler ev_handler = {
+	.wakeup_fn = efrm_handle_wakeup_event,
+	.timeout_fn = efrm_handle_timeout_event,
+	.dmaq_flushed_fn = efrm_handle_dmaq_flushed,
+};
+
+#if EFRM_HAVE_IOMMU_LOCK
+int efrm_need_iommu_lock;
+EXPORT_SYMBOL(efrm_need_iommu_lock);
+#endif
+
+const int max_hardware_init_repeats = 10;
+
+/*--------------------------------------------------------------------
+ *
+ * Module load time variables
+ *
+ *--------------------------------------------------------------------*/
+/* See docs/notes/pci_alloc_consistent */
+int use_pci_alloc = 1;		/* Use pci_alloc_consistent to alloc iopages */
+static int do_irq = 1;		/* enable interrupts */
+
+#if defined(CONFIG_X86_XEN)
+static int irq_moderation = 60;	/* interrupt moderation (60 usec) */
+#else
+static int irq_moderation = 20;	/* interrupt moderation (20 usec) */
+#endif
+static int nic_options = NIC_OPT_DEFAULT;
+int efx_vi_eventq_size = EFX_VI_EVENTQ_SIZE_DEFAULT;
+
+module_param(do_irq, int, S_IRUGO);
+MODULE_PARM_DESC(do_irq, "Enable interrupts.  "
+		 "Do not turn it off unless you know what are you doing.");
+module_param(irq_moderation, int, S_IRUGO);
+MODULE_PARM_DESC(irq_moderation, "IRQ moderation in usec");
+module_param(nic_options, int, S_IRUGO);
+MODULE_PARM_DESC(nic_options, "Nic options -- see efhw_types.h");
+module_param(use_pci_alloc, int, S_IRUGO);
+MODULE_PARM_DESC(use_pci_alloc, "Use pci_alloc_consistent to alloc iopages "
+		 "(autodetected by kernel version)");
+module_param(efx_vi_eventq_size, int, S_IRUGO);
+MODULE_PARM_DESC(efx_vi_eventq_size,
+		 "Size of event queue allocated by efx_vi library");
+
+/*--------------------------------------------------------------------
+ *
+ * Linux specific NIC initialisation
+ *
+ *--------------------------------------------------------------------*/
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,19)
+# define IRQ_PT_REGS_ARG   , struct pt_regs *regs __attribute__ ((unused))
+#else
+# define IRQ_PT_REGS_ARG
+#endif
+
+#ifndef IRQF_SHARED
+# define IRQF_SHARED SA_SHIRQ
+#endif
+
+static inline irqreturn_t
+linux_efrm_interrupt(int irr, void *dev_id IRQ_PT_REGS_ARG)
+{
+	return efhw_nic_interrupt((struct efhw_nic *)dev_id);
+}
+
+int linux_efrm_irq_ctor(struct linux_efhw_nic *lnic)
+{
+	struct efhw_nic *nic = &lnic->nic;
+
+	nic->flags &= ~NIC_FLAG_MSI;
+	if (nic->flags & NIC_FLAG_TRY_MSI) {
+		int rc = pci_enable_msi(lnic->pci_dev);
+		if (rc < 0) {
+			EFRM_WARN("%s: Could not enable MSI (%d)",
+				  __FUNCTION__, rc);
+			EFRM_WARN("%s: Continuing with legacy interrupt mode",
+				  __FUNCTION__);
+		} else {
+			EFRM_NOTICE("%s: MSI enabled", __FUNCTION__);
+			nic->flags |= NIC_FLAG_MSI;
+		}
+	}
+
+	if (request_irq(lnic->pci_dev->irq, linux_efrm_interrupt,
+			IRQF_SHARED, "sfc_resource", nic)) {
+		EFRM_ERR("Request for interrupt #%d failed",
+			 lnic->pci_dev->irq);
+		nic->flags &= ~NIC_FLAG_OS_IRQ_EN;
+		return -EBUSY;
+	}
+	nic->flags |= NIC_FLAG_OS_IRQ_EN;
+
+	return 0;
+}
+
+void linux_efrm_irq_dtor(struct linux_efhw_nic *lnic)
+{
+	EFRM_TRACE("linux_efrm_irq_dtor: start");
+
+	if (lnic->nic.flags & NIC_FLAG_OS_IRQ_EN) {
+		free_irq(lnic->pci_dev->irq, &lnic->nic);
+		lnic->nic.flags &= ~NIC_FLAG_OS_IRQ_EN;
+	}
+
+	if (lnic->nic.flags & NIC_FLAG_MSI) {
+		pci_disable_msi(lnic->pci_dev);
+		lnic->nic.flags &= ~NIC_FLAG_MSI;
+	}
+
+	EFRM_TRACE("linux_efrm_irq_dtor: done");
+}
+
+/* Allocate buffer table entries for a particular NIC.
+ */
+static int efrm_nic_buffer_table_alloc(struct efhw_nic *nic)
+{
+	int capacity;
+	int page_order;
+	int i;
+	int rc;
+
+	/* Choose queue size. */
+	for (capacity = 8192; capacity <= nic->evq_sizes; capacity <<= 1) {
+		if (capacity > nic->evq_sizes) {
+			EFRM_ERR
+			    ("%s: Unable to choose EVQ size (supported=%x)",
+			     __FUNCTION__, nic->evq_sizes);
+			return -E2BIG;
+		} else if (capacity & nic->evq_sizes)
+			break;
+	}
+	for (i = 0; i < EFHW_KEVENTQ_MAX; ++i) {
+		nic->evq[i].hw.capacity = capacity;
+		nic->evq[i].hw.buf_tbl_alloc.base = (unsigned)-1;
+	}
+
+	/* allocate buffer table entries to map onto the iobuffer */
+	page_order = get_order(capacity * sizeof(efhw_event_t));
+	if (!(nic->flags & NIC_FLAG_NO_INTERRUPT)) {
+		rc = efrm_buffer_table_alloc(page_order,
+					     &nic->evq[0].hw.buf_tbl_alloc);
+		if (rc < 0) {
+			EFRM_WARN
+			    ("%s: failed (%d) to alloc %d buffer table entries",
+			     __FUNCTION__, rc, page_order);
+			return rc;
+		}
+	}
+	rc = efrm_buffer_table_alloc(page_order,
+				     &nic->evq[FALCON_EVQ_NONIRQ].hw.
+				     buf_tbl_alloc);
+	if (rc < 0) {
+		EFRM_WARN
+		    ("%s: failed (%d) to alloc %d buffer table entries",
+		     __FUNCTION__, rc, page_order);
+		return rc;
+	}
+
+	return 0;
+}
+
+/* Free buffer table entries allocated for a particular NIC.
+ */
+static void efrm_nic_buffer_table_free(struct efhw_nic *nic)
+{
+	int i;
+	for (i = 0; i <= FALCON_EVQ_NONIRQ; i++)
+		if (nic->evq[i].hw.buf_tbl_alloc.base != (unsigned)-1)
+			efrm_buffer_table_free(&nic->evq[i].hw.buf_tbl_alloc);
+
+}
+
+static int iomap_bar(struct linux_efhw_nic *lnic, size_t len)
+{
+	efhw_ioaddr_t ioaddr;
+
+	ioaddr = ioremap_nocache(lnic->ctr_ap_pci_addr, len);
+	if (ioaddr == 0)
+		return -ENOMEM;
+
+	lnic->nic.bar_ioaddr = ioaddr;
+	return 0;
+}
+
+static int linux_efhw_nic_map_ctr_ap(struct linux_efhw_nic *lnic)
+{
+	struct efhw_nic *nic = &lnic->nic;
+	int rc;
+
+	rc = iomap_bar(lnic, nic->ctr_ap_bytes);
+
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+	/* Bug 5195: workaround for now. */
+	if (rc != 0 && nic->ctr_ap_bytes > 16 * 1024 * 1024) {
+		/* Try half the size for now. */
+		nic->ctr_ap_bytes /= 2;
+		EFRM_WARN("Bug 5195 WORKAROUND: retrying iomap of %d bytes",
+			  nic->ctr_ap_bytes);
+		rc = iomap_bar(lnic, nic->ctr_ap_bytes);
+	}
+#endif
+
+	if (rc < 0) {
+		EFRM_ERR("Failed (%d) to map bar (%d bytes)",
+			 rc, nic->ctr_ap_bytes);
+		return rc;
+	}
+
+	return rc;
+}
+
+int
+linux_efrm_nic_ctor(struct linux_efhw_nic *lnic, struct pci_dev *dev,
+		    spinlock_t *reg_lock,
+		    unsigned nic_flags, unsigned nic_options)
+{
+	struct efhw_device_type dev_type;
+	struct efhw_nic *nic = &lnic->nic;
+	u8 class_revision;
+	int rc;
+
+	rc = pci_read_config_byte(dev, PCI_CLASS_REVISION, &class_revision);
+	if (rc != 0) {
+		EFRM_ERR("%s: pci_read_config_byte failed (%d)",
+			 __FUNCTION__, rc);
+		return rc;
+	}
+
+	if (!efhw_device_type_init(&dev_type, dev->vendor, dev->device,
+				   class_revision)) {
+		EFRM_ERR("%s: efhw_device_type_init failed %04x:%04x(%d)",
+			 __FUNCTION__, (unsigned) dev->vendor,
+			 (unsigned) dev->device, (int) class_revision);
+		return -ENODEV;
+	}
+
+	EFRM_NOTICE("attaching device type %04x:%04x %d:%c%d",
+		    (unsigned) dev->vendor, (unsigned) dev->device,
+		    dev_type.arch, dev_type.variant, dev_type.revision);
+
+	/* Initialise the adapter-structure. */
+	efhw_nic_init(nic, nic_flags, nic_options, dev_type);
+	lnic->pci_dev = dev;
+
+	rc = pci_enable_device(dev);
+	if (rc < 0) {
+		EFRM_ERR("%s: pci_enable_device failed (%d)",
+			 __FUNCTION__, rc);
+		return rc;
+	}
+
+	lnic->ctr_ap_pci_addr = pci_resource_start(dev, nic->ctr_ap_bar);
+
+	if (!pci_dma_supported(dev, (dma_addr_t)EFHW_DMA_ADDRMASK)) {
+		EFRM_ERR("%s: pci_dma_supported(%lx) failed", __FUNCTION__,
+			 (unsigned long)EFHW_DMA_ADDRMASK);
+		return -ENODEV;
+	}
+
+	if (pci_set_dma_mask(dev, (dma_addr_t)EFHW_DMA_ADDRMASK)) {
+		EFRM_ERR("%s: pci_set_dma_mask(%lx) failed", __FUNCTION__,
+			 (unsigned long)EFHW_DMA_ADDRMASK);
+		return -ENODEV;
+	}
+
+	if (pci_set_consistent_dma_mask(dev, (dma_addr_t)EFHW_DMA_ADDRMASK)) {
+		EFRM_ERR("%s: pci_set_consistent_dma_mask(%lx) failed",
+			 __FUNCTION__, (unsigned long)EFHW_DMA_ADDRMASK);
+		return -ENODEV;
+	}
+
+	rc = linux_efhw_nic_map_ctr_ap(lnic);
+	if (rc < 0)
+		return rc;
+
+	/* By default struct efhw_nic contains its own lock for protecting
+	 * access to nic registers.  We override it with a pointer to the
+	 * lock in the net driver.  This is needed when resource and net
+	 * drivers share a single PCI function (falcon B series).
+	 */
+	nic->reg_lock = reg_lock;
+	return 0;
+}
+
+void linux_efrm_nic_dtor(struct linux_efhw_nic *lnic)
+{
+	struct efhw_nic *nic = &lnic->nic;
+	efhw_ioaddr_t bar_ioaddr = nic->bar_ioaddr;
+
+	efhw_nic_dtor(nic);
+
+	efrm_nic_buffer_table_free(nic);
+
+	/* Unmap the bar. */
+	EFRM_ASSERT(bar_ioaddr);
+	iounmap(bar_ioaddr);
+	nic->bar_ioaddr = 0;
+}
+
+/****************************************************************************
+ *
+ * efrm_tasklet - used to poll the eventq which may result in further callbacks
+ *
+ ****************************************************************************/
+
+static void efrm_tasklet(unsigned long pdev)
+{
+	struct efhw_nic *nic = (struct efhw_nic *)pdev;
+
+	EFRM_ASSERT(!(nic->flags & NIC_FLAG_NO_INTERRUPT));
+
+	efhw_keventq_poll(nic, &nic->evq[0]);
+	EFRM_TRACE("tasklet complete");
+}
+
+/****************************************************************************
+ *
+ * char driver specific interrupt callbacks -- run at hard IRQL
+ *
+ ****************************************************************************/
+static void efrm_handle_eventq_irq(struct efhw_nic *nic, int evq)
+{
+	/* NB. The interrupt must have already been acked (for legacy mode). */
+
+	EFRM_TRACE("%s: starting tasklet", __FUNCTION__);
+	EFRM_ASSERT(!(nic->flags & NIC_FLAG_NO_INTERRUPT));
+
+	tasklet_schedule(&linux_efhw_nic(nic)->tasklet);
+}
+
+/* A count of how many NICs this driver knows about. */
+static int n_nics_probed;
+
+/****************************************************************************
+ *
+ * efrm_nic_add: add the NIC to the resource driver
+ *
+ * NOTE: the flow of control through this routine is quite subtle
+ * because of the number of operations that can fail. We therefore
+ * take the apporaching of keeping the return code (rc) variable
+ * accurate, and only do operations while it is non-negative. Tear down
+ * is done at the end if rc is negative, depending on what has been set up
+ * by that point.
+ *
+ * So basically just make sure that any code you add checks rc>=0 before
+ * doing any work and you'll be fine.
+ *
+ ****************************************************************************/
+int
+efrm_nic_add(struct pci_dev *dev, unsigned flags, const uint8_t *mac_addr,
+	     struct linux_efhw_nic **lnic_out, spinlock_t *reg_lock,
+	     int bt_min, int bt_max,
+	     const struct vi_resource_dimensions *res_dim)
+{
+	struct linux_efhw_nic *lnic = NULL;
+	struct efhw_nic *nic = NULL;
+	int count = 0, rc = 0, resources_init = 0;
+	int constructed = 0;
+	int registered_nic = 0;
+	int buffers_allocated = 0;
+	static unsigned nic_index; /* = 0; */
+
+	EFRM_TRACE("%s: device detected (Slot '%s', IRQ %d)", __FUNCTION__,
+		   pci_name(dev) ? pci_name(dev) : "?", dev->irq);
+
+	/* Ensure that we have room for the new adapter-structure. */
+	if (efrm_nic_table.nic_count == EFHW_MAX_NR_DEVS) {
+		EFRM_WARN("%s: WARNING: too many devices", __FUNCTION__);
+		rc = -ENOMEM;
+		goto failed;
+	}
+
+	if (n_nics_probed == 0) {
+		rc = efrm_resources_init(res_dim, bt_min, bt_max);
+		if (rc != 0)
+			goto failed;
+		resources_init = 1;
+	}
+
+	/* Allocate memory for the new adapter-structure. */
+	lnic = kmalloc(sizeof(*lnic), GFP_KERNEL);
+	if (lnic == NULL) {
+		EFRM_ERR("%s: ERROR: failed to allocate memory", __FUNCTION__);
+		rc = -ENOMEM;
+		goto failed;
+	}
+	memset(lnic, 0, sizeof(*lnic));
+	nic = &lnic->nic;
+
+	lnic->ev_handlers = &ev_handler;
+
+	/* OS specific hardware mappings */
+	rc = linux_efrm_nic_ctor(lnic, dev, reg_lock, flags, nic_options);
+	if (rc < 0) {
+		EFRM_ERR("%s: ERROR: initialisation failed", __FUNCTION__);
+		goto failed;
+	}
+
+	constructed = 1;
+
+	/* Tell the driver about the NIC - this needs to be done before the
+	   resources managers get created below. Note we haven't initialised
+	   the hardware yet, and I don't like doing this before the perhaps
+	   unreliable hardware initialisation. However, there's quite a lot
+	   of code to review if we wanted to hardware init before bringing
+	   up the resource managers. */
+	rc = efrm_driver_register_nic(nic, nic_index++);
+	if (rc < 0) {
+		EFRM_ERR("%s: cannot register nic %d with nic error code %d",
+			 __FUNCTION__, efrm_nic_table.nic_count, rc);
+		goto failed;
+	}
+	registered_nic = 1;
+
+	rc = efrm_nic_buffer_table_alloc(nic);
+	if (rc < 0)
+		goto failed;
+	buffers_allocated = 1;
+
+	/****************************************************/
+	/* hardware bringup                                 */
+	/****************************************************/
+	/* Detecting hardware can be a slightly unreliable process;
+	   we want to make sure that we maximise our chances, so we
+	   loop a few times until all is good. */
+	for (count = 0; count < max_hardware_init_repeats; count++) {
+		rc = efhw_nic_init_hardware(nic, &ev_handler, mac_addr);
+		if (rc >= 0)
+			break;
+
+		/* pain */
+		EFRM_ERR
+		    ("error - hardware initialisation failed code %d, "
+		     "attempt %d of %d", rc, count + 1,
+		     max_hardware_init_repeats);
+	}
+	if (rc < 0)
+		goto failed;
+
+	tasklet_init(&lnic->tasklet, efrm_tasklet, (ulong)nic);
+
+	/* set up interrupt handlers (hard-irq) */
+	nic->irq_handler = &efrm_handle_eventq_irq;
+
+	/* this device can now take management interrupts */
+	if (do_irq && !(nic->flags & NIC_FLAG_NO_INTERRUPT)) {
+		rc = linux_efrm_irq_ctor(lnic);
+		if (rc < 0) {
+			EFRM_ERR("Interrupt initialisation failed (%d)", rc);
+			goto failed;
+		}
+		efhw_nic_set_interrupt_moderation(nic, 0, irq_moderation);
+		efhw_nic_interrupt_enable(nic, 0);
+	}
+	EFRM_TRACE("interrupts are %sregistered", do_irq ? "" : "not ");
+
+#if EFRM_HAVE_IOMMU_LOCK
+	/* Bug 4560: We need the lock if there is memory which cannot be
+	 * accessed by the card and there is an IOMMU to access it.  In that
+	 * case, the kernel will use the IOMMU to access the high memory. */
+	if ((dev->dma_mask >> PAGE_SHIFT) < max_pfn && !EFRM_NO_IOMMU)
+		efrm_need_iommu_lock = 1;
+#endif
+
+	*lnic_out = lnic;
+	EFRM_ASSERT(rc == 0);
+	++n_nics_probed;
+	return 0;
+
+failed:
+	if (buffers_allocated)
+		efrm_nic_buffer_table_free(nic);
+	if (registered_nic)
+		efrm_driver_unregister_nic(nic);
+	if (constructed)
+		linux_efrm_nic_dtor(lnic);
+	kfree(lnic); /* safe in any case */
+	if (resources_init)
+		efrm_resources_fini();
+	return rc;
+}
+
+/****************************************************************************
+ *
+ * efrm_nic_del: Remove the nic from the resource driver structures
+ *
+ ****************************************************************************/
+void efrm_nic_del(struct linux_efhw_nic *lnic)
+{
+	struct efhw_nic *nic = &lnic->nic;
+
+	EFRM_TRACE("%s:", __FUNCTION__);
+	EFRM_ASSERT(nic);
+
+	efrm_driver_unregister_nic(nic);
+
+	/*
+	 * Synchronise here with any running ISR.
+	 * Remove the OS handler. There should be no IRQs being generated
+	 * by our NIC at this point.
+	 */
+	if (efhw_nic_have_functional_units(nic)) {
+		efhw_nic_close_interrupts(nic);
+		linux_efrm_irq_dtor(lnic);
+		tasklet_kill(&lnic->tasklet);
+	}
+
+	/* Close down hardware and free resources. */
+	linux_efrm_nic_dtor(lnic);
+	kfree(lnic);
+
+	if (--n_nics_probed == 0)
+		efrm_resources_fini();
+
+	EFRM_TRACE("NIC teardown: Done");
+}
+
+/****************************************************************************
+ *
+ * init_module: register as a PCI driver.
+ *
+ ****************************************************************************/
+static int init_sfc_resource(void)
+{
+	int rc = 0;
+
+	EFRM_TRACE("%s: RESOURCE driver starting", __FUNCTION__);
+
+	rc = efrm_driver_ctor();
+	if (rc < 0) {
+		EFRM_ERR("%s: efrm_driver_ctor: error %d", __FUNCTION__, rc);
+		goto fail_driver_ctor;
+	}
+
+	/* Register the driver so that our 'probe' function is called for
+	 * each EtherFabric device in the system.
+	 */
+	rc = efrm_driverlink_register();
+	if (rc == -ENODEV)
+		EFRM_ERR("%s: no devices found", __FUNCTION__);
+	if (rc < 0)
+		goto failed_driverlink;
+
+	if (efrm_install_proc_entries() != 0) {
+		/* Do not fail, but print a warning */
+		EFRM_WARN("%s: WARNING: failed to install /proc entries",
+			  __FUNCTION__);
+	}
+
+	return 0;
+
+failed_driverlink:
+	/* No need to release resource managers here since they register
+	 * destructors with the driver. */
+	efrm_driver_dtor();
+fail_driver_ctor:
+	EFRM_ASSERT(rc != 0);
+	return rc;
+}
+
+/****************************************************************************
+ *
+ * cleanup_module: module-removal entry-point
+ *
+ ****************************************************************************/
+static void cleanup_sfc_resource(void)
+{
+	efrm_uninstall_proc_entries();
+
+	efrm_driverlink_unregister();
+
+	/* Clean up char-driver specific initialisation.
+	   - driver dtor can use both work queue and buffer table entries */
+	efrm_driver_dtor();
+
+	EFRM_TRACE("unloaded");
+}
+
+module_init(init_sfc_resource);
+module_exit(cleanup_sfc_resource);
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/resource_manager.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/resource_manager.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/resource_manager.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/resource_manager.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,263 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains generic code for resources and resource managers.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/efrm/debug.h>
+#include <ci/efrm/nic_table.h>
+#include <ci/efhw/iopage.h>
+#include <ci/efrm/driver_private.h>
+
+/**********************************************************************
+ * Internal stuff.
+ */
+
+#define EFRM_RM_TABLE_SIZE_INIT 256
+
+static int grow_table(struct efrm_resource_manager *rm, unsigned min_size)
+{
+	irq_flags_t lock_flags;
+	struct efrm_resource **table, **old_table;
+	unsigned new_size;
+
+	EFRM_RESOURCE_MANAGER_ASSERT_VALID(rm);
+
+	spin_lock_irqsave(&rm->rm_lock, lock_flags);
+
+	/* Check whether the size of the table increased whilst the lock was
+	 * dropped. */
+	if (min_size <= rm->rm_table_size) {
+		spin_unlock_irqrestore(&rm->rm_lock, lock_flags);
+		return 0;
+	}
+
+	new_size = rm->rm_table_size << 1;
+	if (new_size < min_size)
+		new_size = min_size;
+
+	spin_unlock_irqrestore(&rm->rm_lock, lock_flags);
+	if (in_atomic()) {
+		EFRM_WARN("%s: in_atomic in grow_table()", __FUNCTION__);
+		EFRM_WARN("%s: allocating %u bytes", __FUNCTION__,
+			  (unsigned)(new_size *
+				     sizeof(struct efrm_resource *)));
+		return -ENOMEM;
+	}
+
+	table =
+	    (struct efrm_resource **)vmalloc(new_size *
+					     sizeof(struct efrm_resource *));
+	spin_lock_irqsave(&rm->rm_lock, lock_flags);
+
+	if (table == 0) {
+		EFRM_ERR("%s: out of memory in grow_table()", __FUNCTION__);
+		EFRM_ERR("%s: allocating %u bytes", __FUNCTION__,
+			 (unsigned)(new_size *
+				    sizeof(struct efrm_resource *)));
+		spin_unlock_irqrestore(&rm->rm_lock, lock_flags);
+		return -ENOMEM;
+	}
+
+	/* Could have got bigger while we dropped the lock... */
+	if (new_size <= rm->rm_table_size) {
+		spin_unlock_irqrestore(&rm->rm_lock, lock_flags);
+		vfree(table);
+		return 0;
+	}
+
+	memcpy(table, rm->rm_table, rm->rm_table_size * sizeof(*table));
+	memset(table + rm->rm_table_size, 0,
+	       sizeof(*table) * (new_size - rm->rm_table_size));
+	/* remember old table so we can free the
+	   memory after we drop the lock (bug 1040) */
+	old_table = rm->rm_table;
+	rm->rm_table = table;
+	rm->rm_table_size = new_size;
+	spin_unlock_irqrestore(&rm->rm_lock, lock_flags);
+	vfree(old_table);
+
+	return 0;
+}
+
+/**********************************************************************
+ * struct efrm_resource_manager
+ */
+
+void efrm_resource_manager_dtor(struct efrm_resource_manager *rm)
+{
+	EFRM_RESOURCE_MANAGER_ASSERT_VALID(rm);
+
+	/* call destructor */
+	EFRM_DO_DEBUG(if (rm->rm_resources)
+		      EFRM_ERR("%s: %s leaked %d resources",
+			       __FUNCTION__, rm->rm_name, rm->rm_resources));
+	EFRM_ASSERT(rm->rm_resources == 0);
+
+	rm->rm_dtor(rm);
+
+	/* clear out things built by efrm_resource_manager_ctor */
+	spin_lock_destroy(&rm->rm_lock);
+	vfree(rm->rm_table);
+
+	/* and the free the memory */
+	EFRM_DO_DEBUG(memset(rm, 0, sizeof(*rm)));
+	kfree(rm);
+}
+
+/* Construct a resource manager.  Resource managers are singletons. */
+int
+efrm_resource_manager_ctor(struct efrm_resource_manager *rm,
+			   void (*dtor)(struct efrm_resource_manager *),
+			   const char *name, unsigned type,
+			   int initial_table_size)
+{
+	EFRM_ASSERT(rm);
+	EFRM_ASSERT(dtor);
+
+	rm->rm_name = name;
+	EFRM_DO_DEBUG(rm->rm_type = type);
+	rm->rm_dtor = dtor;
+	spin_lock_init(&rm->rm_lock);
+	rm->rm_resources = 0;
+	rm->rm_resources_hiwat = 0;
+
+	/* if not set then pick a number */
+	rm->rm_table_size = (initial_table_size) ?
+		initial_table_size : EFRM_RM_TABLE_SIZE_INIT;
+
+	rm->rm_table = vmalloc(rm->rm_table_size *
+			       sizeof(struct efrm_resource *));
+
+	if (rm->rm_table == 0) {
+		spin_lock_destroy(&rm->rm_lock);
+		return -ENOMEM;
+	}
+	memset(rm->rm_table, 0, sizeof(*rm->rm_table) * rm->rm_table_size);
+
+	EFRM_RESOURCE_MANAGER_ASSERT_VALID(rm);
+	return 0;
+}
+
+int efrm_resource_manager_insert(struct efrm_resource *rs)
+{
+	irq_flags_t lock_flags;
+	struct efrm_resource_manager *rm;
+	int instance = EFRM_RESOURCE_INSTANCE(rs->rs_handle);
+
+	EFRM_ASSERT(EFRM_RESOURCE_TYPE(rs->rs_handle) < EFRM_RESOURCE_NUM);
+	rm = efrm_rm_table[EFRM_RESOURCE_TYPE(rs->rs_handle)];
+	EFRM_ASSERT(EFRM_RESOURCE_TYPE(rs->rs_handle) == rm->rm_type);
+	EFRM_RESOURCE_ASSERT_VALID(rs, 0);
+
+	/* Put an entry in the resource table. */
+	spin_lock_irqsave(&rm->rm_lock, lock_flags);
+	if ((unsigned)instance >= rm->rm_table_size) {
+		spin_unlock_irqrestore(&rm->rm_lock, lock_flags);
+		if (grow_table(rm, instance + 1) < 0)
+			return -ENOMEM;
+		spin_lock_irqsave(&rm->rm_lock, lock_flags);
+	}
+	EFRM_ASSERT(rm->rm_table_size > (unsigned)instance);
+	EFRM_ASSERT(rm->rm_table[instance] == NULL);
+	rm->rm_table[instance] = rs;
+	rm->rm_resources++;
+	if (rm->rm_resources > rm->rm_resources_hiwat)
+		rm->rm_resources_hiwat = rm->rm_resources;
+
+	/* Put the resource in the linked list. */
+	/* ?? broken list_add(&rm->rm_resources, &rs->rs_link); */
+	/* DJR wrote that it causes problem on driver unload, and DR tried
+	 * it and saw (probably) this cause an assertion failure due to a
+	 * bad link structure in
+	 * /runbench/results/2005/09/22/0_DupTester_15-16-46 */
+
+	spin_unlock_irqrestore(&rm->rm_lock, lock_flags);
+
+	return 0;
+}
+
+bool __efrm_resource_ref_count_zero(unsigned type, unsigned instance)
+{
+	/* This is rather nasty because when a resource's ref count goes to
+	 * zero there is still a pointer to it in the [rm_table].  Thus
+	 * arriving here does not guarantee that we have exclusive access
+	 * to the resource and can free it.  In fact the resource may
+	 * already have been freed by another thread (after we dropped our
+	 * ref, but before arriving here).
+	 *
+	 * At this point the only pointers to this resource should be [rs]
+	 * and the one in [rm_table].  EXCEPT: Someone could have got in
+	 * and looked-up the resource in the table before we got the lock.
+	 * In this case the ref will have been hiked again.
+	 *
+	 * Therefore, if ref count is non-zero here, we shouldn't do
+	 * anything, as someone else holds a ref to the resource, and will
+	 * eventually release it.
+	 *
+	 * Otherwise, we zero-out the table entry.  Therefore we have the
+	 * only pointer to the resource, and can kill it safely.
+	 */
+	struct efrm_resource_manager *rm = efrm_rm_table[type];
+	irq_flags_t lock_flags;
+	struct efrm_resource *rs;
+	bool do_free = false;
+
+	EFRM_TRACE("efrm_resource_ref_count_zero: type=%d instance=%d",
+		   rm->rm_type, instance);
+
+	EFRM_RESOURCE_MANAGER_ASSERT_VALID(rm);
+	EFRM_ASSERT(rm->rm_table_size > instance);
+
+	spin_lock_irqsave(&rm->rm_lock, lock_flags);
+
+	rs = rm->rm_table[instance];
+	if (rs != NULL) {
+		do_free = atomic_read(&rs->rs_ref_count) == 0;
+		if (do_free) {
+			EFRM_ASSERT(rm->rm_resources > 0);
+			--rm->rm_resources;
+			rm->rm_table[instance] = 0;
+		}
+	}
+
+	spin_unlock_irqrestore(&rm->rm_lock, lock_flags);
+
+	return do_free;
+}
+EXPORT_SYMBOL(__efrm_resource_ref_count_zero);
+
+/*
+ * vi: sw=8:ai:aw
+ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/resources.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/resources.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/resources.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/resources.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,94 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains resource managers initialisation functions.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/efrm/private.h>
+#include <ci/efrm/buffer_table.h>
+
+int
+efrm_resources_init(const struct vi_resource_dimensions *vi_res_dim,
+		    int buffer_table_min, int buffer_table_max)
+{
+	int i, rc;
+
+	rc = efrm_buffer_table_ctor(buffer_table_min, buffer_table_max);
+	if (rc != 0)
+		return rc;
+
+	/* Create resources in the correct order */
+	for (i = 0; i < EFRM_RESOURCE_NUM; ++i) {
+		struct efrm_resource_manager **rmp = &efrm_rm_table[i];
+
+		EFRM_ASSERT(*rmp == NULL);
+		switch (i) {
+		case EFRM_RESOURCE_VI:
+			rc = efrm_create_vi_resource_manager(rmp,
+							     vi_res_dim);
+			break;
+		case EFRM_RESOURCE_FILTER:
+			rc = efrm_create_filter_resource_manager(rmp);
+			break;
+		case EFRM_RESOURCE_IOBUFSET:
+			rc = efrm_create_iobufset_resource_manager(rmp);
+			break;
+		default:
+			rc = 0;
+			break;
+		}
+
+		if (rc < 0) {
+			EFRM_ERR("%s: failed type=%d (%d)",
+				 __FUNCTION__, i, rc);
+			efrm_buffer_table_dtor();
+			return rc;
+		}
+	}
+
+	return 0;
+}
+
+void efrm_resources_fini(void)
+{
+	int i;
+
+	for (i = EFRM_RESOURCE_NUM - 1; i >= 0; --i)
+		if (efrm_rm_table[i]) {
+			efrm_resource_manager_dtor(efrm_rm_table[i]);
+			efrm_rm_table[i] = NULL;
+		}
+
+	efrm_buffer_table_dtor();
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/vi_resource_alloc.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/vi_resource_alloc.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/vi_resource_alloc.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/vi_resource_alloc.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,876 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains allocation of VI resources.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/efrm/nic_table.h>
+#include <ci/efhw/iopage.h>
+#include <ci/driver/efab/hardware.h>
+#include <ci/efhw/public.h>
+#include <ci/efhw/falcon.h>
+#include <ci/efrm/private.h>
+#include <ci/efrm/buffer_table.h>
+#include <ci/efrm/vi_resource_private.h>
+
+/*** Data definitions ****************************************************/
+
+static const char *dmaq_names[] = { "TX", "RX" };
+
+struct vi_resource_manager *efrm_vi_manager;
+
+/*** Forward references **************************************************/
+
+static int
+efrm_vi_resource_alloc_or_free(int alloc, struct vi_resource *evq_virs,
+			       uint16_t vi_flags, int32_t evq_capacity,
+			       int32_t txq_capacity, int32_t rxq_capacity,
+			       uint8_t tx_q_tag, uint8_t rx_q_tag,
+			       struct vi_resource **virs_in_out);
+
+/*** Reference count handling ********************************************/
+
+static inline void efrm_vi_rm_get_ref(struct vi_resource *virs)
+{
+	atomic_inc(&virs->evq_refs);
+}
+
+static inline void efrm_vi_rm_drop_ref(struct vi_resource *virs)
+{
+	EFRM_ASSERT(atomic_read(&virs->evq_refs) != 0);
+	if (atomic_dec_and_test(&virs->evq_refs))
+		efrm_vi_resource_alloc_or_free(false, NULL, 0, 0, 0, 0, 0, 0,
+					       &virs);
+}
+
+/*** Instance numbers ****************************************************/
+
+static inline int efrm_vi_rm_alloc_id(uint16_t vi_flags, int32_t evq_capacity)
+{
+	irq_flags_t lock_flags;
+	int instance;
+	int rc;
+
+	if (efrm_nic_table.a_nic == NULL)	/* ?? FIXME: surely not right */
+		return -ENODEV;
+
+	spin_lock_irqsave(&efrm_vi_manager->rm.rm_lock, lock_flags);
+
+	/* Falcon A1 RX phys addr wierdness. */
+	if (efrm_nic_table.a_nic->devtype.variant == 'A' &&
+	    (vi_flags & EFHW_VI_RX_PHYS_ADDR_EN)) {
+		if (vi_flags & EFHW_VI_JUMBO_EN) {
+			/* Falcon-A cannot do phys + scatter. */
+			EFRM_WARN
+			    ("%s: falcon-A does not support phys+scatter mode",
+			     __FUNCTION__);
+			instance = -1;
+		} else if (efrm_vi_manager->iscsi_dmaq_instance_is_free
+			   && evq_capacity == 0) {
+			/* Falcon-A has a single RXQ that gives the correct
+			 * semantics for physical addressing.  However, it
+			 * happens to have the same instance number as the
+			 * 'char' event queue, so we cannot also hand out
+			 * the event queue. */
+			efrm_vi_manager->iscsi_dmaq_instance_is_free = false;
+			instance = FALCON_A1_ISCSI_DMAQ;
+		} else {
+			EFRM_WARN("%s: iSCSI receive queue not free",
+				  __FUNCTION__);
+			instance = -1;
+		}
+		goto unlock_out;
+	}
+
+	if (vi_flags & EFHW_VI_RM_WITH_INTERRUPT) {
+		rc = __kfifo_get(efrm_vi_manager->instances_with_interrupt,
+				 (unsigned char *)&instance, sizeof(instance));
+		if (rc != sizeof(instance)) {
+			EFRM_ASSERT(rc == 0);
+			instance = -1;
+		}
+		goto unlock_out;
+	}
+
+	/* Otherwise a normal run-of-the-mill VI. */
+	rc = __kfifo_get(efrm_vi_manager->instances_with_timer,
+			 (unsigned char *)&instance, sizeof(instance));
+	if (rc != sizeof(instance)) {
+		EFRM_ASSERT(rc == 0);
+		instance = -1;
+	}
+
+unlock_out:
+	spin_unlock_irqrestore(&efrm_vi_manager->rm.rm_lock, lock_flags);
+	return instance;
+}
+
+static void efrm_vi_rm_free_id(int instance)
+{
+	irq_flags_t lock_flags;
+	struct kfifo *instances;
+
+	if (efrm_nic_table.a_nic == NULL)	/* ?? FIXME: surely not right */
+		return;
+
+	if (efrm_nic_table.a_nic->devtype.variant == 'A' &&
+	    instance == FALCON_A1_ISCSI_DMAQ) {
+		EFRM_ASSERT(efrm_vi_manager->iscsi_dmaq_instance_is_free ==
+			    false);
+		spin_lock_irqsave(&efrm_vi_manager->rm.rm_lock, lock_flags);
+		efrm_vi_manager->iscsi_dmaq_instance_is_free = true;
+		spin_unlock_irqrestore(&efrm_vi_manager->rm.rm_lock,
+				       lock_flags);
+	} else {
+		if (instance >= efrm_vi_manager->with_timer_base &&
+		    instance < efrm_vi_manager->with_timer_limit) {
+			instances = efrm_vi_manager->instances_with_timer;
+		} else {
+			EFRM_ASSERT(instance >=
+				    efrm_vi_manager->with_interrupt_base);
+			EFRM_ASSERT(instance <
+				    efrm_vi_manager->with_interrupt_limit);
+			instances = efrm_vi_manager->instances_with_interrupt;
+		}
+
+		EFRM_VERIFY_EQ(kfifo_put(instances, (unsigned char *)&instance,
+					 sizeof(instance)), sizeof(instance));
+	}
+}
+
+/*** Queue sizes *********************************************************/
+
+/* NB. This should really take a nic as an argument, but that makes
+ * the buffer table allocation difficult. */
+uint32_t efrm_vi_rm_evq_bytes(struct vi_resource *virs
+			      /*,struct efhw_nic *nic */ )
+{
+	return virs->evq_capacity * sizeof(efhw_event_t);
+}
+EXPORT_SYMBOL(efrm_vi_rm_evq_bytes);
+
+/* NB. This should really take a nic as an argument, but that makes
+ * the buffer table allocation difficult. */
+uint32_t efrm_vi_rm_txq_bytes(struct vi_resource *virs
+			      /*,struct efhw_nic *nic */ )
+{
+	return virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_TX] *
+	    FALCON_DMA_TX_DESC_BYTES;
+}
+EXPORT_SYMBOL(efrm_vi_rm_txq_bytes);
+
+/* NB. This should really take a nic as an argument, but that makes
+ * the buffer table allocation difficult. */
+uint32_t efrm_vi_rm_rxq_bytes(struct vi_resource *virs
+			      /*,struct efhw_nic *nic */ )
+{
+	uint32_t bytes_per_desc = ((virs->flags & EFHW_VI_RX_PHYS_ADDR_EN)
+				   ? FALCON_DMA_RX_PHYS_DESC_BYTES
+				   : FALCON_DMA_RX_BUF_DESC_BYTES);
+	return virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_RX] * bytes_per_desc;
+}
+EXPORT_SYMBOL(efrm_vi_rm_rxq_bytes);
+
+static int choose_size(int size_rq, unsigned sizes)
+{
+	int size;
+
+	/* size_rq < 0 means default, but we interpret this as 'minimum'. */
+
+	for (size = 256;; size <<= 1)
+		if ((sizes & ~((size - 1) | size)) == 0)
+			return -1;
+		else if ((size & sizes) && size >= size_rq)
+			return size;
+}
+
+static int
+efrm_vi_rm_adjust_alloc_request(struct vi_resource *virs, struct efhw_nic *nic)
+{
+	int capacity;
+
+	EFRM_ASSERT(nic->efhw_func);
+
+	if (virs->evq_capacity) {
+		capacity = choose_size(virs->evq_capacity, nic->evq_sizes);
+		if (capacity < 0) {
+			EFRM_ERR("vi_resource: bad evq size %d (supported=%x)",
+				 virs->evq_capacity, nic->evq_sizes);
+			return -E2BIG;
+		}
+		virs->evq_capacity = capacity;
+	}
+	if (virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_TX]) {
+		capacity =
+		    choose_size(virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_TX],
+				nic->txq_sizes);
+		if (capacity < 0) {
+			EFRM_ERR("vi_resource: bad txq size %d (supported=%x)",
+				 virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_TX],
+				 nic->txq_sizes);
+			return -E2BIG;
+		}
+		virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_TX] = capacity;
+	}
+	if (virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_RX]) {
+		capacity =
+		    choose_size(virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_RX],
+				nic->rxq_sizes);
+		if (capacity < 0) {
+			EFRM_ERR("vi_resource: bad rxq size %d (supported=%x)",
+				 virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_RX],
+				 nic->rxq_sizes);
+			return -E2BIG;
+		}
+		virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_RX] = capacity;
+	}
+
+	return 0;
+}
+
+/* remove the reference to the event queue in this VI resource and decrement
+   the event queue's use count */
+static inline void efrm_vi_rm_detach_evq(struct vi_resource *virs)
+{
+	struct vi_resource *evq_virs;
+
+	EFRM_ASSERT(virs != NULL);
+
+	evq_virs = virs->evq_virs;
+
+	if (evq_virs != NULL) {
+		virs->evq_virs = NULL;
+		if (evq_virs == virs) {
+			EFRM_TRACE("%s: " EFRM_RESOURCE_FMT
+				   " had internal event queue ", __FUNCTION__,
+				   EFRM_RESOURCE_PRI_ARG(virs->rs.rs_handle));
+		} else {
+			efrm_vi_rm_drop_ref(evq_virs);
+			EFRM_TRACE("%s: " EFRM_RESOURCE_FMT " had event queue "
+				   EFRM_RESOURCE_FMT, __FUNCTION__,
+				   EFRM_RESOURCE_PRI_ARG(virs->rs.rs_handle),
+				   EFRM_RESOURCE_PRI_ARG(evq_virs->rs.
+							 rs_handle));
+		}
+	} else {
+		EFRM_TRACE("%s: " EFRM_RESOURCE_FMT
+			   " had no event queue (nothing to do)",
+			   __FUNCTION__,
+			   EFRM_RESOURCE_PRI_ARG(virs->rs.rs_handle));
+	}
+}
+
+/*** Buffer Table allocations ********************************************/
+
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+static int
+efrm_vi_rm_alloc_or_free_buffer_table(struct vi_resource *virs, bool is_alloc)
+{
+	uint32_t bytes;
+	int page_order;
+	int rc;
+
+	if (!is_alloc)
+		goto destroy;
+
+	if (virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_TX]) {
+		bytes = efrm_vi_rm_txq_bytes(virs);
+		page_order = get_order(bytes);
+		rc = efrm_buffer_table_alloc(page_order,
+					     (virs->dmaq_buf_tbl_alloc +
+					      EFRM_VI_RM_DMA_QUEUE_TX));
+		if (rc != 0) {
+			EFRM_TRACE
+			    ("%s: Error %d allocating TX buffer table entry",
+			     __FUNCTION__, rc);
+			goto fail_txq_alloc;
+		}
+	}
+
+	if (virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_RX]) {
+		bytes = efrm_vi_rm_rxq_bytes(virs);
+		page_order = get_order(bytes);
+		rc = efrm_buffer_table_alloc(page_order,
+					     (virs->dmaq_buf_tbl_alloc +
+					      EFRM_VI_RM_DMA_QUEUE_RX));
+		if (rc != 0) {
+			EFRM_TRACE
+			    ("%s: Error %d allocating RX buffer table entry",
+			     __FUNCTION__, rc);
+			goto fail_rxq_alloc;
+		}
+	}
+	return 0;
+
+destroy:
+	rc = 0;
+
+	if (virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_RX]) {
+		efrm_buffer_table_free(&virs->
+				       dmaq_buf_tbl_alloc
+				       [EFRM_VI_RM_DMA_QUEUE_RX]);
+	}
+fail_rxq_alloc:
+
+	if (virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_TX]) {
+		efrm_buffer_table_free(&virs->
+				       dmaq_buf_tbl_alloc
+				       [EFRM_VI_RM_DMA_QUEUE_TX]);
+	}
+fail_txq_alloc:
+
+	return rc;
+}
+
+#endif /* defined(__CI_HARDWARE_CONFIG_FALCON__) */
+
+/*** Per-NIC allocations *************************************************/
+
+static inline int
+efrm_vi_rm_init_evq(struct vi_resource *virs, int nic_index)
+{
+	int rc;
+	struct efhw_nic *nic = efrm_nic_table.nic[nic_index];
+	int instance = EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle);
+	struct eventq_resource_hardware *evq_hw =
+	    &virs->nic_info[nic_index].evq_pages;
+	uint32_t buf_bytes = efrm_vi_rm_evq_bytes(virs);
+
+	if (virs->evq_capacity == 0)
+		return 0;
+	evq_hw->capacity = virs->evq_capacity;
+
+	/* Allocate buffer table entries to map onto the iobuffer.  This
+	 * currently allocates its own buffer table entries on Falcon which is
+	 * a bit wasteful on a multi-NIC system. */
+	evq_hw->buf_tbl_alloc.base = (unsigned)-1;
+	rc = efrm_buffer_table_alloc(get_order(buf_bytes),
+				     &evq_hw->buf_tbl_alloc);
+	if (rc < 0) {
+		EFHW_WARN("%s: failed (%d) to alloc %d buffer table entries",
+			  __FUNCTION__, rc, get_order(buf_bytes));
+		return rc;
+	}
+
+	/* Allocate the event queue memory. */
+	rc = efhw_nic_event_queue_alloc_iobuffer(nic, evq_hw, instance,
+						 buf_bytes);
+	if (rc != 0) {
+		EFRM_ERR("%s: Error allocating iobuffer: %d", __FUNCTION__, rc);
+		efrm_buffer_table_free(&evq_hw->buf_tbl_alloc);
+		return rc;
+	}
+
+	/* Initialise the event queue hardware */
+	efhw_nic_event_queue_enable(nic, instance, virs->evq_capacity,
+				    efhw_iopages_dma_addr(&evq_hw->iobuff) +
+				    evq_hw->iobuff_off,
+				    evq_hw->buf_tbl_alloc.base);
+
+	EFRM_TRACE("%s: " EFRM_RESOURCE_FMT " capacity=%u", __FUNCTION__,
+		   EFRM_RESOURCE_PRI_ARG(virs->rs.rs_handle),
+		   virs->evq_capacity);
+
+#if defined(__ia64__)
+	/* Page size may be large, so for now just increase the
+	 * size of the requested evq up to a round number of
+	 * pages
+	 */
+	buf_bytes = CI_ROUND_UP(buf_bytes, PAGE_SIZE);
+#endif
+	EFRM_ASSERT(buf_bytes % PAGE_SIZE == 0);
+
+	virs->mem_mmap_bytes += buf_bytes;
+
+	return 0;
+}
+
+static inline void
+efrm_vi_rm_fini_evq(struct vi_resource *virs, int nic_index)
+{
+	struct efhw_nic *nic = efrm_nic_table.nic[nic_index];
+	int instance = EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle);
+	struct vi_resource_nic_info *nic_info = &virs->nic_info[nic_index];
+
+	if (virs->evq_capacity == 0)
+		return;
+
+	/* Zero the timer-value for this queue.
+	   And Tell NIC to stop using this event queue. */
+	efhw_nic_event_queue_disable(nic, instance, 0);
+
+	if (nic_info->evq_pages.buf_tbl_alloc.base != (unsigned)-1)
+		efrm_buffer_table_free(&nic_info->evq_pages.buf_tbl_alloc);
+
+	efhw_iopages_free(nic, &nic_info->evq_pages.iobuff);
+}
+
+/*! FIXME: we should make sure this number is never zero (=> unprotected) */
+/*! FIXME: put this definition in a relevant header (e.g. as (evqid)+1) */
+#define EFAB_EVQ_OWNER_ID(evqid) ((evqid))
+
+void
+efrm_vi_rm_init_dmaq(struct vi_resource *virs, int queue_type,
+		     struct efhw_nic *nic)
+{
+	int instance;
+	struct vi_resource *evq_virs;
+	int evq_instance;
+	efhw_buffer_addr_t buf_addr;
+
+	instance = EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle);
+	evq_virs = virs->evq_virs;
+	evq_instance = EFRM_RESOURCE_INSTANCE(evq_virs->rs.rs_handle);
+
+	buf_addr = virs->dmaq_buf_tbl_alloc[queue_type].base;
+
+	if (queue_type == EFRM_VI_RM_DMA_QUEUE_TX) {
+		efhw_nic_dmaq_tx_q_init(nic,
+			instance,	/* dmaq */
+			evq_instance,	/* evq */
+			EFAB_EVQ_OWNER_ID(evq_instance),	/* owner */
+			virs->dmaq_tag[queue_type],	/* tag */
+			virs->dmaq_capacity[queue_type], /* size of queue */
+			buf_addr,	/* buffer index */
+			virs->flags);	/* user specified Q attrs */
+	} else {
+		efhw_nic_dmaq_rx_q_init(nic,
+			instance,	/* dmaq */
+			evq_instance,	/* evq */
+			EFAB_EVQ_OWNER_ID(evq_instance),	/* owner */
+			virs->dmaq_tag[queue_type],	/* tag */
+			virs->dmaq_capacity[queue_type], /* size of queue */
+			buf_addr,	/* buffer index */
+			virs->flags);	/* user specified Q attrs */
+	}
+}
+
+static int
+efrm_vi_rm_init_or_fini_dmaq(struct vi_resource *virs,
+			     int queue_type, int init, int nic_index)
+{
+	int rc;
+	struct efhw_nic *nic = efrm_nic_table.nic[nic_index];
+	int instance = EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle);
+	uint32_t buf_bytes;
+	struct vi_resource *evq_virs;
+
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+	struct vi_resource_nic_info *nic_info = &virs->nic_info[nic_index];
+	int page_order;
+	uint32_t num_pages;
+	efhw_iopages_t *iobuff;
+#endif
+
+	if (!init)
+		goto destroy;
+
+	/* Ignore disabled queues. */
+	if (virs->dmaq_capacity[queue_type] == 0) {
+		if (queue_type == EFRM_VI_RM_DMA_QUEUE_TX)
+			efhw_nic_dmaq_tx_q_disable(nic, instance);
+		else
+			efhw_nic_dmaq_rx_q_disable(nic, instance);
+		return 0;
+	}
+
+	buf_bytes = (queue_type == EFRM_VI_RM_DMA_QUEUE_TX
+		     ? efrm_vi_rm_txq_bytes(virs)
+		     : efrm_vi_rm_rxq_bytes(virs));
+
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+	page_order = get_order(buf_bytes);
+
+	rc = efhw_iopages_alloc(nic, &nic_info->dmaq_pages[queue_type],
+			      page_order);
+	if (rc != 0) {
+		EFRM_ERR("%s: Failed to allocate %s DMA buffer.", __FUNCTION__,
+			 dmaq_names[queue_type]);
+		goto fail_iopages;
+	}
+
+	num_pages = 1 << page_order;
+	iobuff = &nic_info->dmaq_pages[queue_type];
+	efhw_nic_buffer_table_set_n(nic,
+				    virs->dmaq_buf_tbl_alloc[queue_type].base,
+				    efhw_iopages_dma_addr(iobuff),
+				    EFHW_NIC_PAGE_SIZE, 0, num_pages, 0);
+
+	falcon_nic_buffer_table_confirm(nic);
+
+	virs->mem_mmap_bytes += round_up(buf_bytes, PAGE_SIZE);
+#endif /* __CI_HARDWARE_CONFIG_FALCON__ */
+
+	evq_virs = virs->evq_virs;
+	EFRM_ASSERT(evq_virs);
+
+	/* Make sure there is an event queue. */
+	if (evq_virs->evq_capacity <= 0) {
+		EFRM_ERR("%s: Cannot use empty event queue for %s DMA",
+			 __FUNCTION__, dmaq_names[queue_type]);
+		rc = -EINVAL;
+		goto fail_evq;
+	}
+
+	efrm_vi_rm_init_dmaq(virs, queue_type, nic);
+
+	return 0;
+
+destroy:
+	rc = 0;
+
+	/* Ignore disabled queues. */
+	if (virs->dmaq_capacity[queue_type] == 0)
+		return 0;
+
+	/* No need to disable the queue here.  Nobody is using it anyway. */
+
+fail_evq:
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+	efhw_iopages_free(nic, &nic_info->dmaq_pages[queue_type]);
+fail_iopages:
+#endif
+
+	return rc;
+}
+
+static int
+efrm_vi_rm_init_or_fini_nic(struct vi_resource *virs, int init, int nic_index)
+{
+	struct vi_resource *evq_virs;
+	int rc;
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+#ifndef NDEBUG
+	int instance = EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle);
+#endif
+#endif
+
+	if (!init)
+		goto destroy;
+
+	evq_virs = virs->evq_virs;
+	if (evq_virs != virs) {
+		if (!efrm_nic_set_read(&evq_virs->nic_set, nic_index)) {
+			/* Ignore this NIC.  It's not supported by the event
+			 * queue. */
+			return 0;
+		}
+	}
+
+	rc = efrm_vi_rm_init_evq(virs, nic_index);
+	if (rc != 0)
+		goto fail_evq;
+
+	rc = efrm_vi_rm_init_or_fini_dmaq(virs, EFRM_VI_RM_DMA_QUEUE_TX,
+					  init, nic_index);
+	if (rc != 0)
+		goto fail_txq;
+
+	rc = efrm_vi_rm_init_or_fini_dmaq(virs, EFRM_VI_RM_DMA_QUEUE_RX,
+					  init, nic_index);
+	if (rc != 0)
+		goto fail_rxq;
+
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+	/* Allocate space for the control page. */
+	EFRM_ASSERT(falcon_tx_dma_page_offset(instance) < PAGE_SIZE);
+	EFRM_ASSERT(falcon_rx_dma_page_offset(instance) < PAGE_SIZE);
+	EFRM_ASSERT(falcon_timer_page_offset(instance) < PAGE_SIZE);
+	virs->bar_mmap_bytes += PAGE_SIZE;
+#endif
+
+	/* Mark the NIC as having been initialised. */
+	efrm_nic_set_write(&virs->nic_set, nic_index, true);
+
+	return 0;
+
+destroy:
+	rc = 0;
+
+	efrm_vi_rm_init_or_fini_dmaq(virs, EFRM_VI_RM_DMA_QUEUE_RX,
+				     false, nic_index);
+fail_rxq:
+
+	efrm_vi_rm_init_or_fini_dmaq(virs, EFRM_VI_RM_DMA_QUEUE_TX,
+				     false, nic_index);
+fail_txq:
+
+	efrm_vi_rm_fini_evq(virs, nic_index);
+fail_evq:
+
+	/* Mark the NIC as having been finalised. */
+	efrm_nic_set_write(&virs->nic_set, nic_index, false);
+	EFRM_ASSERT(rc != 0 || !init);
+
+	return rc;
+}
+
+static int
+efrm_vi_resource_alloc_or_free(int alloc, struct vi_resource *evq_virs,
+			       uint16_t vi_flags, int32_t evq_capacity,
+			       int32_t txq_capacity, int32_t rxq_capacity,
+			       uint8_t tx_q_tag, uint8_t rx_q_tag,
+			       struct vi_resource **virs_in_out)
+{
+	struct vi_resource *virs;
+	int rc;
+	int instance;
+	struct efhw_nic *nic;
+	int nic_i;
+
+	EFRM_ASSERT(virs_in_out);
+	EFRM_ASSERT(efrm_vi_manager);
+	EFRM_RESOURCE_MANAGER_ASSERT_VALID(&efrm_vi_manager->rm);
+
+	if (!alloc)
+		goto destroy;
+
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+	rx_q_tag &= (1 << TX_DESCQ_LABEL_WIDTH) - 1;
+	tx_q_tag &= (1 << RX_DESCQ_LABEL_WIDTH) - 1;
+#endif
+
+	virs = kmalloc(sizeof(*virs), GFP_KERNEL);
+	if (virs == NULL) {
+		EFRM_ERR("%s: Error allocating VI resource object",
+			 __FUNCTION__);
+		rc = -ENOMEM;
+		goto fail_alloc;
+	}
+	memset(virs, 0, sizeof(*virs));
+
+	/* Some macros make the assumption that the struct efrm_resource is
+	 * the first member of a struct vi_resource. */
+	EFRM_ASSERT(&virs->rs == (struct efrm_resource *) (virs));
+
+	instance = efrm_vi_rm_alloc_id(vi_flags, evq_capacity);
+	if (instance < 0) {
+		/* Clear out the close list... */
+		efrm_vi_rm_salvage_flushed_vis();
+		instance = efrm_vi_rm_alloc_id(vi_flags, evq_capacity);
+		if (instance >= 0)
+			EFRM_TRACE("%s: Salvaged a closed VI.", __FUNCTION__);
+	}
+
+	if (instance < 0) {
+		/* Could flush resources and try again here. */
+		EFRM_ERR("%s: Out of appropriate VI resources", __FUNCTION__);
+		rc = -EBUSY;
+		goto fail_alloc_id;
+	}
+
+	EFRM_TRACE("%s: new VI ID %d", __FUNCTION__, instance);
+	efrm_resource_init(&virs->rs, EFRM_RESOURCE_VI, instance);
+
+	/* Start with one reference.  Any external VIs using the EVQ of this
+	 * resource will increment this reference rather than the resource
+	 * reference to avoid DMAQ flushes from waiting for other DMAQ
+	 * flushes to complete.  When the resource reference goes to zero,
+	 * the DMAQ flush happens.  When the flush completes, this reference
+	 * is decremented.  When this reference reaches zero, the instance
+	 * is freed. */
+	atomic_set(&virs->evq_refs, 1);
+
+	efrm_nic_set_clear(&virs->nic_set);
+
+	virs->bar_mmap_bytes = 0;
+	virs->mem_mmap_bytes = 0;
+	virs->evq_capacity = evq_capacity;
+	virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_TX] = txq_capacity;
+	virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_RX] = rxq_capacity;
+	virs->dmaq_tag[EFRM_VI_RM_DMA_QUEUE_TX] = tx_q_tag;
+	virs->dmaq_tag[EFRM_VI_RM_DMA_QUEUE_RX] = rx_q_tag;
+	virs->flags = vi_flags;
+
+	INIT_LIST_HEAD(&virs->tx_flush_link);
+	INIT_LIST_HEAD(&virs->rx_flush_link);
+	efrm_nic_set_clear(&virs->tx_flush_nic_set);
+	efrm_nic_set_clear(&virs->rx_flush_nic_set);
+
+	memset(&efrm_vi_manager->evq_infos[instance], 0,
+	       sizeof(struct vi_resource_evq_info));
+	efrm_vi_manager->evq_infos[instance].evq_virs = virs;
+
+	/* Adjust the queue sizes. */
+	rc = 0;
+	EFRM_FOR_EACH_NIC(nic_i, nic)
+	    if (rc == 0)
+		rc = efrm_vi_rm_adjust_alloc_request(virs, nic);
+	if (rc != 0)
+		goto fail_adjust_request;
+
+	/* Attach the EVQ early so that we can ensure that the NIC sets
+	 * match. */
+	if (evq_virs == NULL) {
+		evq_virs = virs;
+		EFRM_TRACE("%s: " EFRM_RESOURCE_FMT
+			   " has no external event queue", __FUNCTION__,
+			   EFRM_RESOURCE_PRI_ARG(virs->rs.rs_handle));
+	} else {
+		/* Make sure the resource managers are the same. */
+		if (EFRM_RESOURCE_TYPE(evq_virs->rs.rs_handle) !=
+		    EFRM_RESOURCE_VI) {
+			EFRM_ERR("%s: Mismatched owner for event queue VI "
+				 EFRM_RESOURCE_FMT, __FUNCTION__,
+				 EFRM_RESOURCE_PRI_ARG(evq_virs->rs.rs_handle));
+			return -EINVAL;
+		}
+		EFRM_ASSERT(atomic_read(&evq_virs->evq_refs) != 0);
+		efrm_vi_rm_get_ref(evq_virs);
+		EFRM_TRACE("%s: " EFRM_RESOURCE_FMT " uses event queue "
+			   EFRM_RESOURCE_FMT,
+			   __FUNCTION__,
+			   EFRM_RESOURCE_PRI_ARG(virs->rs.rs_handle),
+			   EFRM_RESOURCE_PRI_ARG(evq_virs->rs.rs_handle));
+	}
+	virs->evq_virs = evq_virs;
+
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+	rc = efrm_vi_rm_alloc_or_free_buffer_table(virs, true);
+	if (rc != 0)
+		goto fail_buffer_table;
+#endif
+
+	rc = 0;
+	EFRM_FOR_EACH_NIC(nic_i, nic)
+	    if (rc == 0)
+		/* This updates virs->nic_set for the NICs which need
+		 * finalising. */
+		rc = efrm_vi_rm_init_or_fini_nic(virs, true, nic_i);
+	if (rc != 0)
+		goto fail_init_nic;
+
+	/* Put it into the resource manager's table. */
+	rc = efrm_resource_manager_insert(&virs->rs);
+	if (rc != 0) {
+		if (atomic_dec_and_test(&virs->rs.rs_ref_count))
+			efrm_vi_resource_free(virs);
+		return rc;
+	}
+
+	*virs_in_out = virs;
+	EFRM_TRACE("%s: Allocated " EFRM_RESOURCE_FMT, __FUNCTION__,
+		   EFRM_RESOURCE_PRI_ARG(virs->rs.rs_handle));
+	return 0;
+
+destroy:
+	virs = *virs_in_out;
+	EFRM_RESOURCE_ASSERT_VALID(&virs->rs, 1);
+	instance = EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle);
+
+	EFRM_TRACE("%s: Freeing %d", __FUNCTION__,
+		   EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle));
+
+	/* Destroying the VI.  The reference count must be zero. */
+	EFRM_ASSERT(atomic_read(&virs->evq_refs) == 0);
+
+	/* The EVQ should have gone (and DMA disabled) so that this
+	 * function can't be re-entered to destroy the EVQ VI. */
+	EFRM_ASSERT(virs->evq_virs == NULL);
+	rc = 0;
+
+fail_init_nic:
+	EFRM_FOR_EACH_NIC_IN_SET(&virs->nic_set, nic_i, nic)
+	    efrm_vi_rm_init_or_fini_nic(virs, false, nic_i);
+
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+	efrm_vi_rm_alloc_or_free_buffer_table(virs, false);
+fail_buffer_table:
+#endif
+
+	efrm_vi_rm_detach_evq(virs);
+
+fail_adjust_request:
+
+	EFRM_ASSERT(virs->evq_callback_fn == NULL);
+	memset(&efrm_vi_manager->evq_infos[instance], 0,
+	       sizeof(struct vi_resource_evq_info));
+	EFRM_TRACE("%s: delete VI ID %d", __FUNCTION__, instance);
+	efrm_vi_rm_free_id(instance);
+fail_alloc_id:
+
+	EFRM_DO_DEBUG(memset(virs, 0, sizeof(*virs)));
+	kfree(virs);
+fail_alloc:
+	*virs_in_out = NULL;
+
+	return rc;
+}
+
+/*** Resource object  ****************************************************/
+
+int
+efrm_vi_resource_alloc(struct vi_resource *evq_virs,
+		       uint16_t vi_flags, int32_t evq_capacity,
+		       int32_t txq_capacity, int32_t rxq_capacity,
+		       uint8_t tx_q_tag, uint8_t rx_q_tag,
+		       struct vi_resource **virs_out,
+		       uint32_t *out_io_mmap_bytes,
+		       uint32_t *out_mem_mmap_bytes,
+		       uint32_t *out_txq_capacity, uint32_t *out_rxq_capacity)
+{
+	int rc;
+	rc = efrm_vi_resource_alloc_or_free(true, evq_virs, vi_flags,
+					    evq_capacity, txq_capacity,
+					    rxq_capacity, tx_q_tag, rx_q_tag,
+					    virs_out);
+	if (rc == 0) {
+		if (out_io_mmap_bytes != NULL)
+			*out_io_mmap_bytes = (*virs_out)->bar_mmap_bytes;
+		if (out_mem_mmap_bytes != NULL)
+			*out_mem_mmap_bytes = (*virs_out)->mem_mmap_bytes;
+		if (out_txq_capacity != NULL)
+			*out_txq_capacity =
+			    (*virs_out)->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_TX];
+		if (out_rxq_capacity != NULL)
+			*out_rxq_capacity =
+			    (*virs_out)->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_RX];
+	}
+
+	return rc;
+}
+EXPORT_SYMBOL(efrm_vi_resource_alloc);
+
+void efrm_vi_rm_free_flushed_resource(struct vi_resource *virs)
+{
+	EFRM_ASSERT(virs != NULL);
+	EFRM_ASSERT(atomic_read(&virs->rs.rs_ref_count) == 0);
+
+	EFRM_TRACE("%s: " EFRM_RESOURCE_FMT, __FUNCTION__,
+		   EFRM_RESOURCE_PRI_ARG(virs->rs.rs_handle));
+	/* release the associated event queue then drop our own reference
+	 * count */
+	efrm_vi_rm_detach_evq(virs);
+	efrm_vi_rm_drop_ref(virs);
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/vi_resource_event.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/vi_resource_event.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/vi_resource_event.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/vi_resource_event.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,232 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains event handling for VI resource.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/efrm/nic_table.h>
+#include <ci/driver/efab/hardware.h>
+#include <ci/efhw/eventq.h>
+#include <ci/efrm/private.h>
+#include <ci/efrm/vi_resource_private.h>
+
+void
+efrm_eventq_request_wakeup(struct vi_resource *virs, unsigned current_ptr,
+			   unsigned nic_index)
+{
+	struct efhw_nic *nic;
+	int next_i;
+	EFRM_ASSERT(efrm_nic_set_read(&virs->nic_set, nic_index));
+	nic = efrm_nic_table.nic[nic_index];
+	EFRM_ASSERT(nic);
+	next_i = ((current_ptr / sizeof(efhw_event_t)) &
+		  (virs->evq_capacity - 1));
+
+	efhw_nic_wakeup_request(nic, efrm_eventq_dma_addr(virs, nic_index),
+				next_i,
+				EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle));
+}
+EXPORT_SYMBOL(efrm_eventq_request_wakeup);
+
+void efrm_eventq_reset(struct vi_resource *virs, int nic_index)
+{
+	struct efhw_nic *nic = efrm_nic_table.nic[nic_index];
+	int instance = EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle);
+
+	EFRM_ASSERT(virs->evq_capacity != 0);
+	EFRM_ASSERT(efrm_nic_set_read(&virs->nic_set, nic_index));
+
+	/* FIXME: Protect against concurrent resets. */
+
+	efhw_nic_event_queue_disable(nic, instance, 0);
+
+	memset(efrm_eventq_base(virs, nic_index), EFHW_CLEAR_EVENT_VALUE,
+	       efrm_eventq_bytes(virs, nic_index));
+	efhw_nic_event_queue_enable(nic, instance, virs->evq_capacity,
+				    efrm_eventq_dma_addr(virs, nic_index),
+				    virs->nic_info[nic_index].evq_pages.
+				    buf_tbl_alloc.base);
+	EFRM_TRACE("%s: " EFRM_RESOURCE_FMT, __FUNCTION__,
+		   EFRM_RESOURCE_PRI_ARG(virs->rs.rs_handle));
+}
+EXPORT_SYMBOL(efrm_eventq_reset);
+
+int
+efrm_eventq_register_callback(struct vi_resource *virs,
+			      void (*handler) (void *, int,
+					       struct efhw_nic *nic),
+			      void *arg)
+{
+	int instance;
+	int bit;
+
+	EFRM_RESOURCE_ASSERT_VALID(&virs->rs, 0);
+	EFRM_ASSERT(virs->evq_capacity != 0);
+
+	instance = EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle);
+
+	/* The handler can be set only once. */
+	bit = test_and_set_bit(VI_RESOURCE_EVQ_STATE_CALLBACK_REGISTERED,
+			       &efrm_vi_manager->evq_infos[instance].evq_state);
+	if (bit)
+		return -EBUSY;
+
+	/* Store the details. The order is important here. */
+	virs->evq_callback_arg = arg;
+	virs->evq_callback_fn = handler;
+
+	return 0;
+}
+EXPORT_SYMBOL(efrm_eventq_register_callback);
+
+void efrm_eventq_kill_callback(struct vi_resource *virs)
+{
+	int nic_i, instance;
+	struct efhw_nic *nic;
+	struct vi_resource_evq_info *evq_info;
+	int32_t evq_state;
+	int bit;
+
+	EFRM_RESOURCE_ASSERT_VALID(&virs->rs, 0);
+	EFRM_ASSERT(virs->evq_capacity != 0);
+
+	/* Clean out the callback so a new one can be installed. */
+	virs->evq_callback_fn = NULL;
+
+	instance = EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle);
+	evq_info = &efrm_vi_manager->evq_infos[instance];
+
+	/* Disable the event queue. */
+	EFRM_FOR_EACH_NIC_IN_SET(&virs->nic_set, nic_i, nic)
+	    efhw_nic_event_queue_disable(nic, instance, /*timer_only */ 1);
+
+	/* Disable the callback. */
+	bit = test_and_clear_bit(VI_RESOURCE_EVQ_STATE_CALLBACK_REGISTERED,
+				 &evq_info->evq_state);
+	EFRM_ASSERT(bit);	/* do not call me twice! */
+
+	/* Spin until the callback is complete. */
+	do {
+		rmb();
+
+		udelay(1);
+		evq_state = evq_info->evq_state;
+	} while ((evq_state & VI_RESOURCE_EVQ_STATE(BUSY)));
+}
+EXPORT_SYMBOL(efrm_eventq_kill_callback);
+
+static void
+efrm_eventq_do_callback(struct efhw_nic *nic, unsigned instance,
+			bool is_timeout)
+{
+	void (*handler) (void *, int is_timeout, struct efhw_nic *nic);
+	void *arg;
+	struct vi_resource_evq_info *evq_info;
+	int32_t evq_state;
+	int32_t new_evq_state;
+	struct vi_resource *virs;
+	int bit;
+
+	EFRM_TRACE("%s: q=%d %s", __FUNCTION__, instance,
+		   is_timeout ? "timeout" : "wakeup");
+	EFRM_ASSERT(efrm_vi_manager);
+
+	evq_info = &efrm_vi_manager->evq_infos[instance];
+
+	/* Set the BUSY bit and clear WAKEUP_PENDING.  Do this
+	 * before waking up the sleeper to avoid races. */
+	while (1) {
+		evq_state = evq_info->evq_state;
+		new_evq_state = evq_state;
+
+		if ((evq_state & VI_RESOURCE_EVQ_STATE(BUSY)) != 0) {
+			EFRM_ERR("%s:%d: evq_state[%d] corrupted!",
+				 __FUNCTION__, __LINE__, instance);
+			return;
+		}
+
+		if (!is_timeout)
+			new_evq_state &= ~VI_RESOURCE_EVQ_STATE(WAKEUP_PENDING);
+
+		if (evq_state & VI_RESOURCE_EVQ_STATE(CALLBACK_REGISTERED)) {
+			new_evq_state |= VI_RESOURCE_EVQ_STATE(BUSY);
+			if (cmpxchg(&evq_info->evq_state, evq_state,
+				    new_evq_state) == evq_state) {
+				virs = evq_info->evq_virs;
+				break;
+			}
+
+		} else {
+			/* Just update the state if necessary. */
+			if (new_evq_state == evq_state ||
+			    cmpxchg(&evq_info->evq_state, evq_state,
+				    new_evq_state) == evq_state)
+				return;
+		}
+
+		udelay(1);
+	}
+
+	/* Call the callback if any. */
+	if (evq_state & VI_RESOURCE_EVQ_STATE(CALLBACK_REGISTERED)) {
+		/* Retrieve the callback fn. */
+		handler = virs->evq_callback_fn;
+		arg = virs->evq_callback_arg;
+		if (handler != NULL)	/* avoid races */
+			handler(arg, is_timeout, nic);
+	}
+
+	/* Clear the BUSY bit. */
+	bit =
+	    test_and_clear_bit(VI_RESOURCE_EVQ_STATE_BUSY,
+			       &evq_info->evq_state);
+	if (!bit) {
+		EFRM_ERR("%s:%d: evq_state corrupted!",
+			 __FUNCTION__, __LINE__);
+	}
+}
+
+void efrm_handle_wakeup_event(struct efhw_nic *nic, efhw_event_t *ev)
+{
+	efrm_eventq_do_callback(nic,
+				(unsigned int)FALCON_EVENT_WAKE_EVQ_ID(ev),
+				false);
+}
+
+void efrm_handle_timeout_event(struct efhw_nic *nic, efhw_event_t *ev)
+{
+	efrm_eventq_do_callback(nic,
+				(unsigned int)FALCON_EVENT_WAKE_EVQ_ID(ev),
+				true);
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/vi_resource_flush.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/vi_resource_flush.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/vi_resource_flush.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/vi_resource_flush.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,506 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains DMA queue flushing of VI resources.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/efrm/nic_table.h>
+#include <ci/driver/efab/hardware.h>
+#include <ci/efhw/falcon.h>
+#include <ci/efrm/private.h>
+#include <ci/efrm/sysdep.h>
+#include <ci/efrm/buffer_table.h>
+#include <ci/efrm/vi_resource_private.h>
+
+#if EFRM_VI_USE_WORKQUEUE
+	/* can fail as workitem can already be scheuled -- ignore failure */
+#define EFRM_VI_RM_DELAYED_FREE(manager) \
+	queue_work(manager->workqueue, &manager->work_item)
+#else
+#define EFRM_VI_RM_DELAYED_FREE(manager) \
+	efrm_vi_rm_delayed_free(&manager->work_item)
+#endif
+
+static const int flush_fifo_hwm = 8 /* TODO should be a HW specific const */ ;
+
+static void
+efrm_vi_resource_rx_flush_done(struct vi_resource *virs, int nic_i,
+			       bool *completed)
+{
+	/* We should only get a flush event if there is a flush
+	 * outstanding. */
+	EFRM_ASSERT(efrm_nic_set_read
+		    (&virs->rx_flush_outstanding_nic_set, nic_i));
+
+	efrm_nic_set_write(&virs->rx_flush_outstanding_nic_set, nic_i, false);
+	efrm_nic_set_write(&virs->rx_flush_nic_set, nic_i, false);
+
+	if (efrm_nic_set_is_all_clear(&virs->rx_flush_outstanding_nic_set)) {
+		list_del(&virs->rx_flush_link);
+		efrm_vi_manager->rx_flush_outstanding_count--;
+
+		if (efrm_nic_set_is_all_clear(&virs->tx_flush_nic_set)) {
+			list_add_tail(&virs->rx_flush_link,
+				      &efrm_vi_manager->close_pending);
+			*completed = 1;
+		}
+	}
+}
+
+static void
+efrm_vi_resource_tx_flush_done(struct vi_resource *virs, int nic_i,
+			       bool *completed)
+{
+	/* We should only get a flush event if there is a flush
+	 * outstanding. */
+	EFRM_ASSERT(efrm_nic_set_read(&virs->tx_flush_nic_set, nic_i));
+
+	efrm_nic_set_write(&virs->tx_flush_nic_set, nic_i, false);
+
+	if (efrm_nic_set_is_all_clear(&virs->tx_flush_nic_set)) {
+		list_del(&virs->tx_flush_link);
+
+		if (efrm_nic_set_is_all_clear(&virs->rx_flush_nic_set)) {
+			list_add_tail(&virs->rx_flush_link,
+				      &efrm_vi_manager->close_pending);
+			*completed = 1;
+		}
+	}
+}
+
+static void
+efrm_vi_resource_issue_rx_flush(struct vi_resource *virs, bool *completed)
+{
+	struct efhw_nic *nic;
+	int nic_i;
+	int instance;
+	int rc;
+
+	instance = EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle);
+
+	list_add_tail(&virs->rx_flush_link,
+		      &efrm_vi_manager->rx_flush_outstanding_list);
+	virs->rx_flush_outstanding_nic_set = virs->rx_flush_nic_set;
+	efrm_vi_manager->rx_flush_outstanding_count++;
+
+	EFRM_FOR_EACH_NIC_IN_SET(&virs->nic_set, nic_i, nic) {
+		EFRM_TRACE("%s: rx queue %d flush requested for nic %d",
+			   __FUNCTION__, instance, nic->index);
+		rc = efhw_nic_flush_rx_dma_channel(nic, instance);
+		if (rc == -EAGAIN)
+			efrm_vi_resource_rx_flush_done(virs, nic_i, completed);
+	}
+}
+
+static void
+efrm_vi_resource_issue_tx_flush(struct vi_resource *virs, bool *completed)
+{
+	struct efhw_nic *nic;
+	int nic_i;
+	int instance;
+	int rc;
+
+	instance = EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle);
+
+	list_add_tail(&virs->tx_flush_link,
+		      &efrm_vi_manager->tx_flush_outstanding_list);
+
+	EFRM_FOR_EACH_NIC_IN_SET(&virs->nic_set, nic_i, nic) {
+		EFRM_TRACE("%s: tx queue %d flush requested for nic %d",
+			   __FUNCTION__, instance, nic->index);
+		rc = efhw_nic_flush_tx_dma_channel(nic, instance);
+		if (rc == -EAGAIN)
+			efrm_vi_resource_tx_flush_done(virs, nic_i, completed);
+	}
+}
+
+static void efrm_vi_resource_process_waiting_flushes(bool *completed)
+{
+	struct vi_resource *virs;
+
+	while (efrm_vi_manager->rx_flush_outstanding_count < flush_fifo_hwm &&
+	       !list_empty(&efrm_vi_manager->rx_flush_waiting_list)) {
+		virs =
+		    list_entry(list_pop
+			       (&efrm_vi_manager->rx_flush_waiting_list),
+			       struct vi_resource, rx_flush_link);
+		efrm_vi_resource_issue_rx_flush(virs, completed);
+	}
+}
+
+#if BUG7916_WORKAROUND || BUG5302_WORKAROUND
+static void
+efrm_vi_resource_flush_retry_vi(struct vi_resource *virs,
+				int64_t time_now, bool *completed)
+{
+	struct efhw_nic *nic;
+	int nic_i;
+	int instance;
+
+	instance = EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle);
+
+	virs->flush_count++;
+	virs->flush_time = time_now;
+
+#if BUG7916_WORKAROUND
+	if (!efrm_nic_set_is_all_clear(&virs->rx_flush_outstanding_nic_set)) {
+		EFRM_TRACE("%s: Retrying RX flush on instance %d",
+			   __FUNCTION__, instance);
+
+		list_del(&virs->rx_flush_link);
+		efrm_vi_manager->rx_flush_outstanding_count--;
+		efrm_vi_resource_issue_rx_flush(virs, completed);
+		efrm_vi_resource_process_waiting_flushes(completed);
+	}
+#endif
+
+#if BUG5302_WORKAROUND
+	if (!efrm_nic_set_is_all_clear(&virs->tx_flush_nic_set)) {
+		if (virs->flush_count > 5) {
+			EFRM_TRACE("%s: VI resource stuck flush pending "
+				   "(instance=%d, count=%d)",
+				   __FUNCTION__, instance, virs->flush_count);
+			EFRM_FOR_EACH_NIC_IN_SET(&virs->tx_flush_nic_set,
+						nic_i, nic) {
+				falcon_clobber_tx_dma_ptrs(nic, instance);
+			}
+		} else {
+			EFRM_TRACE("%s: Retrying TX flush on instance %d",
+				   __FUNCTION__, instance);
+		}
+
+		list_del(&virs->tx_flush_link);
+		efrm_vi_resource_issue_tx_flush(virs, completed);
+	}
+#endif
+}
+#endif
+
+int efrm_vi_resource_flush_retry(struct vi_resource *virs)
+{
+#if BUG7916_WORKAROUND || BUG5302_WORKAROUND
+	irq_flags_t lock_flags;
+	bool completed = false;
+
+	if (efrm_nic_set_is_all_clear(&virs->rx_flush_nic_set) &&
+	    efrm_nic_set_is_all_clear(&virs->tx_flush_nic_set))
+		return -EALREADY;
+
+	spin_lock_irqsave(&efrm_vi_manager->rm.rm_lock, lock_flags);
+	efrm_vi_resource_flush_retry_vi(virs, get_jiffies_64(), &completed);
+	spin_unlock_irqrestore(&efrm_vi_manager->rm.rm_lock, lock_flags);
+
+	if (completed)
+		EFRM_VI_RM_DELAYED_FREE(efrm_vi_manager);
+#endif
+
+	return 0;
+}
+EXPORT_SYMBOL(efrm_vi_resource_flush_retry);
+
+#if BUG7916_WORKAROUND || BUG5302_WORKAROUND
+/* resource manager lock should be taken before this call */
+static void efrm_vi_handle_flush_loss(bool *completed)
+{
+	struct list_head *pos, *temp;
+	struct vi_resource *virs;
+	int64_t time_now, time_pending;
+
+	/* It's possible we miss flushes - the list is sorted in order we
+	 * generate flushes, see if any are very old. It's also possible
+	 * that we decide an endpoint is flushed even though we've not
+	 * received all the flush events. We *should * mark as
+	 * completed, reclaim and loop again. ??
+	 * THIS NEEDS BACKPORTING FROM THE FALCON branch
+	 */
+	time_now = get_jiffies_64();
+
+#if BUG7916_WORKAROUND
+	list_for_each_safe(pos, temp,
+			   &efrm_vi_manager->rx_flush_outstanding_list) {
+		virs = container_of(pos, struct vi_resource, rx_flush_link);
+
+		time_pending = time_now - virs->flush_time;
+
+		/* List entries are held in reverse chronological order.  Only
+		 * process the old ones. */
+		if (time_pending <= 0x100000000LL)
+			break;
+
+		efrm_vi_resource_flush_retry_vi(virs, time_now, completed);
+	}
+#endif
+
+#if BUG5302_WORKAROUND
+	list_for_each_safe(pos, temp,
+			   &efrm_vi_manager->tx_flush_outstanding_list) {
+		virs = container_of(pos, struct vi_resource, tx_flush_link);
+
+		time_pending = time_now - virs->flush_time;
+
+		/* List entries are held in reverse chronological order.
+		 * Only process the old ones. */
+		if (time_pending <= 0x100000000LL)
+			break;
+
+		efrm_vi_resource_flush_retry_vi(virs, time_now, completed);
+	}
+#endif
+}
+#endif
+
+void
+efrm_vi_register_flush_callback(struct vi_resource *virs,
+				void (*handler)(void *), void *arg)
+{
+	if (handler == NULL) {
+		virs->flush_callback_fn = handler;
+		wmb();
+		virs->flush_callback_arg = arg;
+	} else {
+		virs->flush_callback_arg = arg;
+		wmb();
+		virs->flush_callback_fn = handler;
+	}
+}
+EXPORT_SYMBOL(efrm_vi_register_flush_callback);
+
+int efrm_pt_flush(struct vi_resource *virs)
+{
+	int instance;
+	irq_flags_t lock_flags;
+	bool completed = false;
+
+	instance = EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle);
+
+	EFRM_ASSERT(efrm_nic_set_is_all_clear(&virs->rx_flush_nic_set));
+	EFRM_ASSERT(efrm_nic_set_is_all_clear
+		    (&virs->rx_flush_outstanding_nic_set));
+	EFRM_ASSERT(efrm_nic_set_is_all_clear(&virs->tx_flush_nic_set));
+
+	EFRM_TRACE("%s: " EFRM_RESOURCE_FMT " EVQ=%d TXQ=%d RXQ=%d",
+		   __FUNCTION__, EFRM_RESOURCE_PRI_ARG(virs->rs.rs_handle),
+		   virs->evq_capacity,
+		   virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_TX],
+		   virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_RX]);
+
+	spin_lock_irqsave(&efrm_vi_manager->rm.rm_lock, lock_flags);
+
+	if (virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_RX] != 0)
+		virs->rx_flush_nic_set = virs->nic_set;
+
+	if (virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_TX] != 0)
+		virs->tx_flush_nic_set = virs->nic_set;
+
+	/* Clean up immediately if there are no flushes. */
+	if (efrm_nic_set_is_all_clear(&virs->rx_flush_nic_set) &&
+	    efrm_nic_set_is_all_clear(&virs->tx_flush_nic_set)) {
+		list_add_tail(&virs->rx_flush_link,
+			      &efrm_vi_manager->close_pending);
+		completed = true;
+	}
+
+	/* Issue the RX flush if possible or queue it for later. */
+	if (!efrm_nic_set_is_all_clear(&virs->rx_flush_nic_set)) {
+#if BUG7916_WORKAROUND || BUG5302_WORKAROUND
+		if (efrm_vi_manager->rx_flush_outstanding_count >=
+		    flush_fifo_hwm)
+			efrm_vi_handle_flush_loss(&completed);
+#endif
+		if (efrm_vi_manager->rx_flush_outstanding_count >=
+		    flush_fifo_hwm) {
+			list_add_tail(&virs->rx_flush_link,
+				      &efrm_vi_manager->rx_flush_waiting_list);
+		} else {
+			efrm_vi_resource_issue_rx_flush(virs, &completed);
+		}
+	}
+
+	/* Issue the TX flush.  There's no limit to the number of
+	 * outstanding TX flushes. */
+	if (!efrm_nic_set_is_all_clear(&virs->tx_flush_nic_set))
+		efrm_vi_resource_issue_tx_flush(virs, &completed);
+
+	virs->flush_time = get_jiffies_64();
+
+	spin_unlock_irqrestore(&efrm_vi_manager->rm.rm_lock, lock_flags);
+
+	if (completed)
+		EFRM_VI_RM_DELAYED_FREE(efrm_vi_manager);
+
+	return 0;
+}
+EXPORT_SYMBOL(efrm_pt_flush);
+
+static void
+efrm_handle_rx_dmaq_flushed(struct efhw_nic *flush_nic, int instance,
+			    bool *completed)
+{
+	struct list_head *pos, *temp;
+	struct vi_resource *virs;
+
+	list_for_each_safe(pos, temp,
+			   &efrm_vi_manager->rx_flush_outstanding_list) {
+		virs = container_of(pos, struct vi_resource, rx_flush_link);
+
+		if (instance == EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle)) {
+			efrm_vi_resource_rx_flush_done(virs,
+						       flush_nic->index,
+						       completed);
+			efrm_vi_resource_process_waiting_flushes(completed);
+			return;
+		}
+	}
+	EFRM_TRACE("%s: Unhandled rx flush event, nic %d, instance %d",
+		   __FUNCTION__, flush_nic->index, instance);
+}
+
+static void
+efrm_handle_tx_dmaq_flushed(struct efhw_nic *flush_nic, int instance,
+			    bool *completed)
+{
+	struct list_head *pos, *temp;
+	struct vi_resource *virs;
+
+	list_for_each_safe(pos, temp,
+			   &efrm_vi_manager->tx_flush_outstanding_list) {
+		virs = container_of(pos, struct vi_resource, tx_flush_link);
+
+		if (instance == EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle)) {
+			efrm_vi_resource_tx_flush_done(virs,
+						       flush_nic->index,
+						       completed);
+			return;
+		}
+	}
+	EFRM_TRACE("%s: Unhandled tx flush event, nic %d, instance %d",
+		   __FUNCTION__, flush_nic->index, instance);
+}
+
+void
+efrm_handle_dmaq_flushed(struct efhw_nic *flush_nic, int instance,
+			 int rx_flush)
+{
+	irq_flags_t lock_flags;
+	bool completed = false;
+
+	EFRM_TRACE("%s: nic_i=%d  instance=%d  rx_flush=%d", __FUNCTION__,
+		   flush_nic->index, instance, rx_flush);
+
+	spin_lock_irqsave(&efrm_vi_manager->rm.rm_lock, lock_flags);
+
+	if (rx_flush)
+		efrm_handle_rx_dmaq_flushed(flush_nic, instance, &completed);
+	else
+		efrm_handle_tx_dmaq_flushed(flush_nic, instance, &completed);
+
+#if BUG7916_WORKAROUND || BUG5302_WORKAROUND
+	efrm_vi_handle_flush_loss(&completed);
+#endif
+
+	spin_unlock_irqrestore(&efrm_vi_manager->rm.rm_lock, lock_flags);
+
+	if (completed)
+		EFRM_VI_RM_DELAYED_FREE(efrm_vi_manager);
+}
+
+static void
+efrm_vi_rm_reinit_dmaqs(struct vi_resource *virs)
+{
+	struct efhw_nic *nic;
+	int nic_i;
+
+	EFRM_FOR_EACH_NIC_IN_SET(&virs->nic_set, nic_i, nic) {
+		if (virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_TX] != 0)
+			efrm_vi_rm_init_dmaq(virs, EFRM_VI_RM_DMA_QUEUE_TX,
+					     nic);
+		if (virs->dmaq_capacity[EFRM_VI_RM_DMA_QUEUE_RX])
+			efrm_vi_rm_init_dmaq(virs, EFRM_VI_RM_DMA_QUEUE_RX,
+					     nic);
+	}
+}
+
+/* free any PT endpoints whose flush has now complete */
+void efrm_vi_rm_delayed_free(struct work_struct *data)
+{
+	irq_flags_t lock_flags;
+	struct list_head close_pending;
+	struct vi_resource *virs;
+
+	EFRM_RESOURCE_MANAGER_ASSERT_VALID(&efrm_vi_manager->rm);
+
+	spin_lock_irqsave(&efrm_vi_manager->rm.rm_lock, lock_flags);
+	list_replace_init(&efrm_vi_manager->close_pending, &close_pending);
+	spin_unlock_irqrestore(&efrm_vi_manager->rm.rm_lock, lock_flags);
+
+	EFRM_TRACE("%s: %p", __FUNCTION__, efrm_vi_manager);
+	while (!list_empty(&close_pending)) {
+		virs =
+		    list_entry(list_pop(&close_pending), struct vi_resource,
+			       rx_flush_link);
+		EFRM_TRACE("%s: flushed VI instance=%d", __FUNCTION__,
+			   EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle));
+
+		if (virs->flush_callback_fn != NULL) {
+			efrm_vi_rm_reinit_dmaqs(virs);
+			virs->flush_callback_fn(virs->flush_callback_arg);
+		} else
+			efrm_vi_rm_free_flushed_resource(virs);
+	}
+}
+
+void efrm_vi_rm_salvage_flushed_vis(void)
+{
+#if BUG7916_WORKAROUND || BUG5302_WORKAROUND
+	irq_flags_t lock_flags;
+	bool completed;
+
+	spin_lock_irqsave(&efrm_vi_manager->rm.rm_lock, lock_flags);
+	efrm_vi_handle_flush_loss(&completed);
+	spin_unlock_irqrestore(&efrm_vi_manager->rm.rm_lock, lock_flags);
+#endif
+
+	efrm_vi_rm_delayed_free(&efrm_vi_manager->work_item);
+}
+
+void efrm_vi_resource_free(struct vi_resource *virs)
+{
+	efrm_vi_register_flush_callback(virs, NULL, NULL);
+	efrm_pt_flush(virs);
+}
+EXPORT_SYMBOL(efrm_vi_resource_free);
+
+/*
+ * vi: sw=8:ai:aw
+ */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfc_resource/vi_resource_manager.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/vi_resource_manager.c
--- linux-2.6.18.8/drivers/net/sfc/sfc_resource/vi_resource_manager.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfc_resource/vi_resource_manager.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,259 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains the VI resource manager.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <ci/efrm/nic_table.h>
+#include <ci/driver/efab/hardware.h>
+#include <ci/efhw/falcon.h>
+#include <ci/efrm/private.h>
+#include <ci/efrm/vi_resource_private.h>
+
+int efrm_pt_pace(struct vi_resource *virs, unsigned int val)
+{
+#if defined(__CI_HARDWARE_CONFIG_FALCON__)
+	int instance, nic_i;
+	struct efhw_nic *nic;
+
+	EFRM_RESOURCE_ASSERT_VALID(&virs->rs, 0);
+	instance = EFRM_RESOURCE_INSTANCE(virs->rs.rs_handle);
+
+	EFRM_FOR_EACH_NIC_IN_SET(&virs->nic_set, nic_i, nic)
+		falcon_nic_pace(nic, instance, val);
+
+	EFRM_TRACE("%s[%d]=%d DONE", __FUNCTION__, instance, val);
+	return 0;
+#else
+	return -EOPNOTSUPP;
+#endif
+}
+EXPORT_SYMBOL(efrm_pt_pace);
+
+/*** Resource manager creation/destruction *******************************/
+
+static void efrm_vi_rm_dtor(struct efrm_resource_manager *rm);
+
+static int
+efrm_create_or_destroy_vi_resource_manager(
+				struct efrm_resource_manager **rm_in_out,
+				const struct vi_resource_dimensions *dims,
+				bool destroy)
+{
+	struct vi_resource *virs;
+	struct list_head *pos, *temp;
+	struct list_head flush_pending;
+	irq_flags_t lock_flags;
+	int rc, i, n_evqs;
+	unsigned dmaq_min, dmaq_max;
+
+	EFRM_ASSERT(rm_in_out);
+
+	if (destroy)
+		goto destroy;
+
+	EFRM_ASSERT(dims);
+	EFRM_NOTICE("vi_resource_manager: evq_int=%u-%u evq_timer=%u-%u",
+		    dims->evq_int_min, dims->evq_int_max,
+		    dims->evq_timer_min, dims->evq_timer_max);
+	EFRM_NOTICE("vi_resource_manager: rxq=%u-%u txq=%u-%u",
+		    dims->rxq_min, dims->rxq_max,
+		    dims->txq_min, dims->txq_max);
+
+	efrm_vi_manager = kmalloc(sizeof(*efrm_vi_manager), GFP_KERNEL);
+	if (efrm_vi_manager == NULL) {
+		rc = -ENOMEM;
+		goto fail_alloc;
+	}
+
+	memset(efrm_vi_manager, 0, sizeof(*efrm_vi_manager));
+
+	efrm_vi_manager->iscsi_dmaq_instance_is_free = true;
+
+	dmaq_min = max(dims->rxq_min, dims->txq_min);
+	dmaq_max = min(dims->rxq_max, dims->txq_max);
+
+	efrm_vi_manager->with_timer_base =
+	    max(dmaq_min, dims->evq_timer_min);
+	efrm_vi_manager->with_timer_limit =
+	    min(dmaq_max, dims->evq_timer_max);
+	rc = efrm_kfifo_id_ctor(&efrm_vi_manager->instances_with_timer,
+				efrm_vi_manager->with_timer_base,
+				efrm_vi_manager->with_timer_limit,
+				&efrm_vi_manager->rm.rm_lock);
+	if (rc < 0)
+		goto fail_with_timer_id_pool;
+
+	efrm_vi_manager->with_interrupt_base =
+	    max(dmaq_min, dims->evq_int_min);
+	efrm_vi_manager->with_interrupt_limit =
+	    min(dmaq_max, dims->evq_int_max);
+	efrm_vi_manager->with_interrupt_limit =
+		max(efrm_vi_manager->with_interrupt_limit,
+		    efrm_vi_manager->with_interrupt_base);
+	rc = efrm_kfifo_id_ctor(&efrm_vi_manager->instances_with_interrupt,
+				efrm_vi_manager->with_interrupt_base,
+				efrm_vi_manager->with_interrupt_limit,
+				&efrm_vi_manager->rm.rm_lock);
+	if (rc < 0)
+		goto fail_with_int_id_pool;
+
+	n_evqs = max(efrm_vi_manager->with_timer_limit,
+		     efrm_vi_manager->with_interrupt_limit);
+	rc = -ENOMEM;
+	efrm_vi_manager->evq_infos =
+		vmalloc(n_evqs * sizeof(struct vi_resource_evq_info));
+	if (efrm_vi_manager->evq_infos == NULL)
+		goto fail_alloc_evq_infos;
+
+	for (i = 0; i < n_evqs; ++i) {
+		efrm_vi_manager->evq_infos[i].evq_state = 0;
+		efrm_vi_manager->evq_infos[i].evq_virs = NULL;
+	}
+
+	INIT_LIST_HEAD(&efrm_vi_manager->rx_flush_waiting_list);
+	INIT_LIST_HEAD(&efrm_vi_manager->rx_flush_outstanding_list);
+	INIT_LIST_HEAD(&efrm_vi_manager->tx_flush_outstanding_list);
+	efrm_vi_manager->rx_flush_outstanding_count = 0;
+
+	INIT_LIST_HEAD(&efrm_vi_manager->close_pending);
+#if EFRM_VI_USE_WORKQUEUE
+	efrm_vi_manager->workqueue = create_workqueue("sfc_vi");
+	if (efrm_vi_manager->workqueue == NULL)
+		goto fail_create_workqueue;
+#endif
+	INIT_WORK(&efrm_vi_manager->work_item, efrm_vi_rm_delayed_free);
+
+	/* NB.  This must be the last step to avoid things getting tangled.
+	 * efrm_resource_manager_dtor calls the vi_rm_dtor which ends up in
+	 * this function. */
+	rc = efrm_resource_manager_ctor(&efrm_vi_manager->rm, efrm_vi_rm_dtor,
+					"VI", EFRM_RESOURCE_VI, 0);
+	if (rc < 0)
+		goto fail_rm_ctor;
+
+	*rm_in_out = &efrm_vi_manager->rm;
+	return 0;
+
+destroy:
+	rc = 0;
+	EFRM_RESOURCE_MANAGER_ASSERT_VALID(*rm_in_out);
+
+	/* Abort outstanding flushes.  Note, a VI resource can be on more
+	 * than one of these lists.  We handle this by starting with the TX
+	 * list and then append VIs to this list if they aren't on the TX
+	 * list already.  A VI is on the TX flush list if tx_flush_nic_set
+	 * is not empty. */
+	spin_lock_irqsave(&efrm_vi_manager->rm.rm_lock, lock_flags);
+
+	list_replace_init(&efrm_vi_manager->tx_flush_outstanding_list,
+			  &flush_pending);
+
+	list_for_each_safe(pos, temp,
+			   &efrm_vi_manager->rx_flush_waiting_list) {
+		virs = container_of(pos, struct vi_resource, rx_flush_link);
+
+		list_del(&virs->rx_flush_link);
+		if (efrm_nic_set_is_all_clear(&virs->tx_flush_nic_set))
+			list_add_tail(&virs->tx_flush_link, &flush_pending);
+	}
+
+	list_for_each_safe(pos, temp,
+			   &efrm_vi_manager->rx_flush_outstanding_list) {
+		virs = container_of(pos, struct vi_resource, rx_flush_link);
+
+		list_del(&virs->rx_flush_link);
+		if (efrm_nic_set_is_all_clear(&virs->tx_flush_nic_set))
+			list_add_tail(&virs->tx_flush_link, &flush_pending);
+	}
+
+	spin_unlock_irqrestore(&efrm_vi_manager->rm.rm_lock, lock_flags);
+
+	while (!list_empty(&flush_pending)) {
+		virs =
+		    list_entry(list_pop(&flush_pending), struct vi_resource,
+			       tx_flush_link);
+		EFRM_TRACE("%s: found PT endpoint " EFRM_RESOURCE_FMT
+			   " with flush pending [Tx=0x%x, Rx=0x%x, RxO=0x%x]",
+			   __FUNCTION__,
+			   EFRM_RESOURCE_PRI_ARG(virs->rs.rs_handle),
+			   virs->tx_flush_nic_set.nics,
+			   virs->rx_flush_nic_set.nics,
+			   virs->rx_flush_outstanding_nic_set.nics);
+		efrm_vi_rm_free_flushed_resource(virs);
+	}
+
+fail_rm_ctor:
+
+	/* Complete outstanding closes. */
+#if EFRM_VI_USE_WORKQUEUE
+	destroy_workqueue(efrm_vi_manager->workqueue);
+fail_create_workqueue:
+#endif
+	EFRM_ASSERT(list_empty(&efrm_vi_manager->close_pending));
+
+	n_evqs = max(efrm_vi_manager->with_timer_limit,
+		     efrm_vi_manager->with_interrupt_limit);
+	vfree(efrm_vi_manager->evq_infos);
+fail_alloc_evq_infos:
+
+	kfifo_vfree(efrm_vi_manager->instances_with_interrupt);
+fail_with_int_id_pool:
+
+	kfifo_vfree(efrm_vi_manager->instances_with_timer);
+fail_with_timer_id_pool:
+
+	if (destroy)
+		return 0;
+
+	EFRM_DO_DEBUG(memset(efrm_vi_manager, 0, sizeof(*efrm_vi_manager)));
+	kfree(efrm_vi_manager);
+fail_alloc:
+
+	*rm_in_out = NULL;
+	EFRM_ERR("%s: failed rc=%d", __FUNCTION__, rc);
+	return rc;
+}
+
+int
+efrm_create_vi_resource_manager(struct efrm_resource_manager **rm_out,
+				const struct vi_resource_dimensions *dims)
+{
+	return efrm_create_or_destroy_vi_resource_manager(rm_out, dims, false);
+}
+
+static void efrm_vi_rm_dtor(struct efrm_resource_manager *rm)
+{
+	efrm_create_or_destroy_vi_resource_manager(&rm, NULL, true);
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/sfe4001.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfe4001.c
--- linux-2.6.18.8/drivers/net/sfc/sfe4001.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/sfe4001.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,315 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2007:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************/
+
+/*****************************************************************************
+ * Support for the SFE4001 NIC: driver code for the PCA9539 I/O expander that
+ * controls the PHY power rails, and for the MAX6647 temp. sensor used to check
+ * the PHY
+ */
+#include <linux/delay.h>
+#include "efx.h"
+#include "phy.h"
+#include "boards.h"
+#include "falcon.h"
+#include "falcon_hwdefs.h"
+
+/**************************************************************************
+ *
+ * I2C IO Expander device
+ *
+ **************************************************************************/
+#define	PCA9539 0x74
+
+#define	P0_IN 0x00
+#define	P0_OUT 0x02
+#define	P0_INVERT 0x04
+#define	P0_CONFIG 0x06
+
+#define	P0_EN_1V0X_LBN 0
+#define	P0_EN_1V0X_WIDTH 1
+#define	P0_EN_1V2_LBN 1
+#define	P0_EN_1V2_WIDTH 1
+#define	P0_EN_2V5_LBN 2
+#define	P0_EN_2V5_WIDTH 1
+#define	P0_EN_3V3X_LBN 3
+#define	P0_EN_3V3X_WIDTH 1
+#define	P0_EN_5V_LBN 4
+#define	P0_EN_5V_WIDTH 1
+#define	P0_SHORTEN_JTAG_LBN 5
+#define	P0_SHORTEN_JTAG_WIDTH 1
+#define	P0_X_TRST_LBN 6
+#define	P0_X_TRST_WIDTH 1
+#define	P0_DSP_RESET_LBN 7
+#define	P0_DSP_RESET_WIDTH 1
+
+#define	P1_IN 0x01
+#define	P1_OUT 0x03
+#define	P1_INVERT 0x05
+#define	P1_CONFIG 0x07
+
+#define	P1_AFE_PWD_LBN 0
+#define	P1_AFE_PWD_WIDTH 1
+#define	P1_DSP_PWD25_LBN 1
+#define	P1_DSP_PWD25_WIDTH 1
+#define	P1_RESERVED_LBN 2
+#define	P1_RESERVED_WIDTH 2
+#define	P1_SPARE_LBN 4
+#define	P1_SPARE_WIDTH 4
+
+
+/**************************************************************************
+ *
+ * Temperature Sensor
+ *
+ **************************************************************************/
+#define	MAX6647	0x4e
+
+#define	RLTS	0x00
+#define	RLTE	0x01
+#define	RSL	0x02
+#define	RCL	0x03
+#define	RCRA	0x04
+#define	RLHN	0x05
+#define	RLLI	0x06
+#define	RRHI	0x07
+#define	RRLS	0x08
+#define	WCRW	0x0a
+#define	WLHO	0x0b
+#define	WRHA	0x0c
+#define	WRLN	0x0e
+#define	OSHT	0x0f
+#define	REET	0x10
+#define	RIET	0x11
+#define	RWOE	0x19
+#define	RWOI	0x20
+#define	HYS	0x21
+#define	QUEUE	0x22
+#define	MFID	0xfe
+#define	REVID	0xff
+
+/* Status bits */
+#define MAX6647_BUSY	(1 << 7)	/* ADC is converting */
+#define MAX6647_LHIGH	(1 << 6)	/* Local high temp. alarm */
+#define MAX6647_LLOW	(1 << 5)	/* Local low temp. alarm */
+#define MAX6647_RHIGH	(1 << 4)	/* Remote high temp. alarm */
+#define MAX6647_RLOW	(1 << 3)	/* Remote low temp. alarm */
+#define MAX6647_FAULT	(1 << 2)	/* DXN/DXP short/open circuit */
+#define MAX6647_EOT	(1 << 1)	/* Remote junction overtemp. */
+#define MAX6647_IOT	(1 << 0)	/* Local junction overtemp. */
+
+static const u8 xgphy_max_temperature = 90;
+
+void sfe4001_poweroff(struct efx_nic *efx)
+{
+	struct efx_i2c_interface *i2c = &efx->i2c;
+
+	u8 cfg, out, in;
+
+	EFX_INFO(efx, "%s\n", __func__);
+
+	/* Turn off all power rails */
+	out = 0xff;
+	(void) efx_i2c_write(i2c, PCA9539, P0_OUT, &out, EFX_BYTE);
+
+	/* Disable port 1 outputs on IO expander */
+	cfg = 0xff;
+	(void) efx_i2c_write(i2c, PCA9539, P1_CONFIG, &cfg, EFX_BYTE);
+
+	/* Disable port 0 outputs on IO expander */
+	cfg = 0xff;
+	(void) efx_i2c_write(i2c, PCA9539, P0_CONFIG, &cfg, EFX_BYTE);
+
+	/* Clear any over-temperature alert */
+	(void) efx_i2c_read(i2c, MAX6647, RSL, &in, EFX_BYTE);
+}
+
+static int sfe4001_check_hw(struct efx_nic *efx)
+{
+	struct efx_i2c_interface *i2c = &efx->i2c;
+	int rc;
+	u8 status;
+
+	/* Check the powered status of the PHY. Lack of power implies that
+	 * the MAX6647 has shut down power to it, probably due to a temp.
+	 * alarm. Reading the power status rather than the MAX6647 status
+	 * directly because the later is read-to-clear and would thus
+	 * start to power up the PHY again when polled, causing us to blip
+	 * the power undesirably */
+
+	/* If XAUI link is down, check power status. Reading
+	 * power requires a I2C byte read, which is too slow
+	 * to poll (see SFC bug 7884). */
+	if (falcon_xaui_link_ok(efx))
+		return 0;
+
+	rc = efx_i2c_read(i2c, PCA9539, P1_IN, &status, EFX_BYTE);
+	status &= ((1 << P1_AFE_PWD_LBN) | (1 << P1_DSP_PWD25_LBN));
+
+	/* We know we can read from the IO expander because we did
+	 * it during power-on. Assume failure now is bad news. */
+	if (rc != 0 || status == 0) {
+		sfe4001_poweroff(efx);
+
+		/* Note that the PHY is pining for the cooling fans */
+		tenxpress_set_state(efx, TENXPRESS_STATUS_OTEMP);
+
+		/* Log the info */
+		if (status == 0) {
+			EFX_ERR(efx, "%s: Temperature sensor reports "
+				"alarm! (0x%x) Shutting down PHY.\n",
+				__func__, status);
+			rc = -EIO;
+		} else {
+			EFX_ERR(efx, "%s: Failed to read PHY status!"
+				" Shutting down PHY.\n",
+				__func__);
+		}
+	}
+
+	return rc;
+}
+
+/* This board uses an I2C expander to provider power to the PHY, which needs to
+ * be turned on before the PHY can be used.
+ * Context: Process context, rtnl lock held
+ */
+int sfe4001_poweron(struct efx_nic *efx)
+{
+	struct efx_i2c_interface *i2c = &efx->i2c;
+	unsigned int count;
+	int rc;
+	u8 out, in, cfg;
+	efx_dword_t reg;
+
+	/* 10Xpress has fixed-function LED pins, so there is no board-specific
+	 * blink code. */
+	efx->board_info.blink = tenxpress_phy_blink;
+
+	/* Ensure that XGXS and XAUI SerDes are held in reset */
+	EFX_POPULATE_DWORD_7(reg, XX_PWRDNA_EN, 1,
+			     XX_PWRDNB_EN, 1,
+			     XX_RSTPLLAB_EN, 1,
+			     XX_RESETA_EN, 1,
+			     XX_RESETB_EN, 1,
+			     XX_RSTXGXSRX_EN, 1,
+			     XX_RSTXGXSTX_EN, 1);
+	efx->mac_op->mac_writel(efx, &reg, XX_PWR_RST_REG_MAC);
+	udelay(10);
+
+	efx->board_info.monitor = sfe4001_check_hw;
+	efx->board_info.fini = sfe4001_poweroff;
+
+	/* Set DSP over-temperature alert threshold */
+	EFX_INFO(efx, "DSP cut-out at %dC\n", xgphy_max_temperature);
+	rc = efx_i2c_write(i2c, MAX6647, WLHO,
+			   &xgphy_max_temperature, EFX_BYTE);
+	if (rc)
+		goto fail1;
+
+	/* Read it back and verify */
+	rc = efx_i2c_read(i2c, MAX6647, RLHN, &in, EFX_BYTE);
+	if (rc)
+		goto fail1;
+	if (in != xgphy_max_temperature) {
+		rc = -EFAULT;
+		goto fail1;
+	}
+
+	/* Clear any previous over-temperature alert */
+	rc = efx_i2c_read(i2c, MAX6647, RSL, &in, EFX_BYTE);
+	if (rc)
+		goto fail1;
+
+	/* Enable port 0 and port 1 outputs on IO expander */
+	cfg = 0x00;
+	rc = efx_i2c_write(i2c, PCA9539, P0_CONFIG, &cfg, EFX_BYTE);
+	if (rc)
+		goto fail1;
+	cfg = 0xff & ~(1 << P1_SPARE_LBN);
+	rc = efx_i2c_write(i2c, PCA9539, P1_CONFIG, &cfg, EFX_BYTE);
+	if (rc)
+		goto fail2;
+
+	/* Turn all power off then wait 1 sec. This ensures PHY is reset */
+	out = 0xff & ~((0 << P0_EN_1V2_LBN) | (0 << P0_EN_2V5_LBN) |
+		       (0 << P0_EN_3V3X_LBN) | (0 << P0_EN_5V_LBN) |
+		       (0 << P0_EN_1V0X_LBN));
+	rc = efx_i2c_write(i2c, PCA9539, P0_OUT, &out, EFX_BYTE);
+	if (rc)
+		goto fail3;
+
+	schedule_timeout_uninterruptible(HZ);
+	count = 0;
+	do {
+		/* Turn on 1.2V, 2.5V, 3.3V and 5V power rails */
+		out = 0xff & ~((1 << P0_EN_1V2_LBN) | (1 << P0_EN_2V5_LBN) |
+			       (1 << P0_EN_3V3X_LBN) | (1 << P0_EN_5V_LBN) |
+			       (1 << P0_X_TRST_LBN));
+
+		rc = efx_i2c_write(i2c, PCA9539, P0_OUT, &out, EFX_BYTE);
+		if (rc)
+			goto fail3;
+		msleep(10);
+
+		/* Turn on 1V power rail */
+		out &= ~(1 << P0_EN_1V0X_LBN);
+		rc = efx_i2c_write(i2c, PCA9539, P0_OUT, &out, EFX_BYTE);
+		if (rc)
+			goto fail3;
+
+		EFX_INFO(efx, "waiting for power (attempt %d)...\n", count);
+
+		schedule_timeout_uninterruptible(HZ);
+
+		/* Check DSP is powered */
+		rc = efx_i2c_read(i2c, PCA9539, P1_IN, &in, EFX_BYTE);
+		if (rc)
+			goto fail3;
+		if (in & (1 << P1_AFE_PWD_LBN))
+			goto done;
+
+	} while (++count < 20);
+
+	EFX_INFO(efx, "timed out waiting for power\n");
+	rc = -ETIMEDOUT;
+	goto fail3;
+
+done:
+	EFX_INFO(efx, "PHY is powered on\n");
+	return 0;
+
+fail3:
+	/* Turn off all power rails */
+	out = 0xff;
+	(void) efx_i2c_write(i2c, PCA9539, P0_OUT, &out, EFX_BYTE);
+	/* Disable port 1 outputs on IO expander */
+	out = 0xff;
+	(void) efx_i2c_write(i2c, PCA9539, P1_CONFIG, &out, EFX_BYTE);
+fail2:
+	/* Disable port 0 outputs on IO expander */
+	out = 0xff;
+	(void) efx_i2c_write(i2c, PCA9539, P1_CONFIG, &out, EFX_BYTE);
+fail1:
+	return rc;
+}
diff -rpuN linux-2.6.18.8/drivers/net/sfc/spi.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/spi.h
--- linux-2.6.18.8/drivers/net/sfc/spi.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/spi.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,186 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005:      Fen Systems Ltd.
+ * Copyright 2006:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_SPI_H
+#define EFX_SPI_H
+
+#include "net_driver.h"
+
+/**************************************************************************
+ *
+ * Basic SPI command set and bit definitions
+ *
+ *************************************************************************/
+
+/*
+ * Commands common to all known devices.
+ *
+ */
+
+/* Write status register */
+#define SPI_WRSR 0x01
+
+/* Write data to memory array */
+#define SPI_WRITE 0x02
+
+/* Read data from memory array */
+#define SPI_READ 0x03
+
+/* Reset write enable latch */
+#define SPI_WRDI 0x04
+
+/* Read status register */
+#define SPI_RDSR 0x05
+
+/* Set write enable latch */
+#define SPI_WREN 0x06
+
+/* SST: Enable write to status register */
+#define SPI_SST_EWSR 0x50
+
+/*
+ * Status register bits.  Not all bits are supported on all devices.
+ *
+ */
+
+/* Write-protect pin enabled */
+#define SPI_STATUS_WPEN 0x80
+
+/* Block protection bit 2 */
+#define SPI_STATUS_BP2 0x10
+
+/* Block protection bit 1 */
+#define SPI_STATUS_BP1 0x08
+
+/* Block protection bit 0 */
+#define SPI_STATUS_BP0 0x04
+
+/* State of the write enable latch */
+#define SPI_STATUS_WEN 0x02
+
+/* Device busy flag */
+#define SPI_STATUS_NRDY 0x01
+
+/**************************************************************************
+ *
+ * Efx SPI devices
+ *
+ **************************************************************************
+ */
+
+/**
+ * struct efx_spi_device - an Efx SPI (Serial Peripheral Interface) device
+ * @device_id:		Controller's id for the device
+ * @size:		Size (in bytes)
+ * @addr_len:		Number of address bytes in read/write commands
+ * @munge_address:	Flag whether addresses should be munged.
+ *	Some devices with 9-bit addresses (e.g. AT25040A EEPROM)
+ *	use bit 3 of the command byte as address bit A8, rather
+ *	than having a two-byte address.  If this flag is set, then
+ *	commands should be munged in this way.
+ * @erase_command:	Erase command (or 0 if sector erase not needed).
+ * @erase_size:		Erase sector size (in bytes)
+ *	Erase commands affect sectors with this size and alignment.
+ *	This must be a power of two.
+ * @block_size:		Write block size (in bytes).
+ *	Write commands are limited to blocks with this size and alignment.
+ * @read:		Read function for the device
+ * @write:		Write function for the device
+ */
+struct efx_spi_device {
+	int device_id;
+	unsigned int size;
+	unsigned int addr_len;
+	unsigned int munge_address:1;
+	u8 erase_command;
+	unsigned int erase_size;
+	unsigned int block_size;
+	int (*read) (const struct efx_spi_device *spi,
+		     struct efx_nic *efx, unsigned int command,
+		     int address, void *data, unsigned int len);
+	int (*write) (const struct efx_spi_device *spi,
+		      struct efx_nic *efx, unsigned int command,
+		      int address, const void *data, unsigned int len);
+};
+
+/* Maximum length for SPI read or write through Falcon */
+#define FALCON_SPI_MAX_LEN 16U
+
+/**
+ * efx_spi_write_limit - calculate maximum permitted length for write
+ * @spi:		SPI device description
+ * @start:		Starting address
+ *
+ * Return the maximum length for a write starting at the given address
+ * in the device.
+ *
+ * SPI writes must not cross block boundaries.  Devices tend
+ * to wrap addresses at block boundaries; e.g. trying to write 5 bytes
+ * starting at offset 14 with a block size of 16 might write
+ * {14,15,0,1,2} rather than {14,15,16,17,18}.
+ */
+static inline unsigned int
+efx_spi_write_limit(const struct efx_spi_device *spi, unsigned int start)
+{
+	return min(FALCON_SPI_MAX_LEN,
+		   (spi->block_size - (start & (spi->block_size - 1))));
+}
+
+/**
+ * efx_spi_read_limit - calculate maximum permitted length for read
+ * @spi:		SPI device description
+ * @start:		Starting address
+ *
+ * Return the maximum length for a read starting at the given address
+ * in the device.
+ */
+static inline unsigned int
+efx_spi_read_limit(const struct efx_spi_device *spi __attribute__ ((unused)),
+		   unsigned int start __attribute__ ((unused)))
+{
+	return FALCON_SPI_MAX_LEN;
+}
+
+/**
+ * efx_spi_munge_command - adjust command as necessary for given address
+ * @spi:		SPI device description
+ * @command:		Normal SPI command
+ * @address:		Address for command
+ *
+ * Some devices with 9-bit addresses (e.g. AT25040A EEPROM) use bit 3
+ * of the command byte as address bit A8, rather than having a
+ * two-byte address.  This function calculates the appropriate command
+ * byte for the device, taking this munging into account.
+ */
+static inline u8 efx_spi_munge_command(const struct efx_spi_device *spi,
+					    const u8 command,
+					    const unsigned int address)
+{
+	return (command | (((address >> 8) & spi->munge_address) << 3));
+}
+
+#endif /* EFX_SPI_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/tenxpress.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/tenxpress.c
--- linux-2.6.18.8/drivers/net/sfc/tenxpress.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/tenxpress.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,697 @@
+/****************************************************************************
+ * Driver for Solarflare 802.3an compliant PHY
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2007:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************/
+
+#include <linux/delay.h>
+#include <linux/seq_file.h>
+#include "efx.h"
+#include "debugfs.h"
+#include "gmii.h"
+#include "mdio_10g.h"
+#include "falcon.h"
+#include "phy.h"
+#include "falcon_hwdefs.h"
+#include "boards.h"
+
+/* We expect these MMDs to be in the package */
+/* AN not here as mdio_check_mmds() requires STAT2 support */
+#define TENXPRESS_REQUIRED_DEVS (MDIO_MMDREG_DEVS0_PMAPMD | \
+				 MDIO_MMDREG_DEVS0_PCS    | \
+				 MDIO_MMDREG_DEVS0_PHYXS)
+
+#define TENXPRESS_LOOPBACKS ((1 << LOOPBACK_PHYXS) |	\
+			     (1 << LOOPBACK_PCS) |	\
+			     (1 << LOOPBACK_PMAPMD) |	\
+			     (1 << LOOPBACK_NETWORK))
+
+/* We complain if we fail to see the link partner as 10G capable this many
+ * times in a row (must be > 1 as sampling the autoneg. registers is racy)
+ */
+#define MAX_BAD_LP_TRIES	(5)
+
+/* SNR operating margin register */
+#define PMA_PMD_SNR_MARGIN_0	(133)
+#define PMA_PMD_SNR_MARGIN_1	(134)
+#define PMA_PMD_SNR_MARGIN_2	(135)
+#define PMA_PMD_SNR_MARGIN_3	(136)
+
+/* Extended control register */
+#define	PMA_PMD_XCONTROL_REG 0xc000
+#define	PMA_PMD_LNPGA_POWERDOWN_LBN 8
+#define	PMA_PMD_LNPGA_POWERDOWN_WIDTH 1
+#define	PMA_PMD_AFE_POWERDOWN_LBN 9
+#define	PMA_PMD_AFE_POWERDOWN_WIDTH 1
+#define	PMA_PMD_DSP_POWERDOWN_LBN 10
+#define	PMA_PMD_DSP_POWERDOWN_WIDTH 1
+#define	PMA_PMD_PHY_POWERDOWN_LBN 11
+#define	PMA_PMD_PHY_POWERDOWN_WI
+
+/* extended status register */
+#define PMA_PMD_XSTATUS_REG 0xc001
+#define PMA_PMD_XSTAT_FLP_LBN   (12)
+
+
+/* LED control register */
+#define PMA_PMD_LED_CTRL_REG	(0xc007)
+#define PMA_PMA_LED_ACTIVITY_LBN	(3)
+
+/* LED function override register */
+#define PMA_PMD_LED_OVERR_REG	(0xc009)
+/* Bit positions for different LEDs (there are more but not wired on SFE4001)*/
+#define PMA_PMD_LED_LINK_LBN	(0)
+#define PMA_PMD_LED_SPEED_LBN	(2)
+#define PMA_PMD_LED_TX_LBN	(4)
+#define PMA_PMD_LED_RX_LBN	(6)
+/* Override settings */
+#define	PMA_PMD_LED_AUTO	(0)	/* H/W control */
+#define	PMA_PMD_LED_ON		(1)
+#define	PMA_PMD_LED_OFF		(2)
+#define PMA_PMD_LED_FLASH	(3)
+/* All LEDs under hardware control */
+#define PMA_PMD_LED_FULL_AUTO	(0)
+/* Green and Amber under hardware control, Red off */
+#define PMA_PMD_LED_DEFAULT	(PMA_PMD_LED_OFF << PMA_PMD_LED_RX_LBN)
+
+
+/* Self test (BIST) control register */
+#define PMA_PMD_BIST_CTRL_REG	(0xc014)
+#define PMA_PMD_BIST_BER_LBN	(2)	/* Run BER test */
+#define PMA_PMD_BIST_CONT_LBN	(1)	/* Run continuous BIST until cleared */
+#define PMA_PMD_BIST_SINGLE_LBN	(0)	/* Run 1 BIST iteration (self clears) */
+/* Self test status register */
+#define PMA_PMD_BIST_STAT_REG	(0xc015)
+#define PMA_PMD_BIST_ENX_LBN	(3)
+#define PMA_PMD_BIST_PMA_LBN	(2)
+#define PMA_PMD_BIST_RXD_LBN	(1)
+#define PMA_PMD_BIST_AFE_LBN	(0)
+
+/* Special Software reset register */
+#define PMA_PMD_EXT_CTRL_REG 49152
+#define PMA_PMD_EXT_SSR_LBN 15
+
+#define BIST_MAX_DELAY	(1000)
+#define BIST_POLL_DELAY	(10)
+
+static const char *bist_names[] = {
+	[PMA_PMD_BIST_AFE_LBN] = "AFE communication",
+	[PMA_PMD_BIST_RXD_LBN] = "RX data path",
+	[PMA_PMD_BIST_PMA_LBN] = "PMA loopback",
+	[PMA_PMD_BIST_ENX_LBN] = "ENX"
+};
+
+/* Identifier registers: each identifier has 4 part number and 2 revision
+ * registers starting at one of these addresses */
+#define PMA_PMD_AFE_ID_REG      49174
+#define PMA_PMD_DSP_ID_REG      49180
+#define PMA_PMD_FIRMWARE_ID_REG 49186
+
+/* Misc register defines */
+#define PCS_CLOCK_CTRL_REG 0xd801
+#define PLL312_RST_N_LBN 2
+
+#define PCS_SOFT_RST2_REG 0xd806
+#define SERDES_RST_N_LBN 13
+#define XGXS_RST_N_LBN 12
+
+#define	PCS_TEST_SELECT_REG 0xd807	/* PRM 10.5.8 */
+#define	CLK312_EN_LBN 3
+
+/* PHYXS registers */
+#define PHYXS_TEST1         (49162)
+#define LOOPBACK_NEAR_LBN   (8)
+#define LOOPBACK_NEAR_WIDTH (1)
+
+/* Boot status register */
+#define PCS_BOOT_STATUS_REG	(0xd000)
+#define PCS_BOOT_FATAL_ERR_LBN	(0)
+#define PCS_BOOT_PROGRESS_LBN	(1)
+#define PCS_BOOT_PROGRESS_WIDTH	(2)
+#define PCS_BOOT_COMPLETE_LBN	(3)
+
+#define PCS_BOOT_MAX_DELAY	(100)
+#define PCS_BOOT_POLL_DELAY	(10)
+
+#define TENXPRESS_ID_PN_LEN     (8)
+#define TENXPRESS_ID_REV_LEN    (4)
+#define TENXPRESS_ID_LEN        (TENXPRESS_ID_PN_LEN+1+TENXPRESS_ID_REV_LEN)
+
+static const int bist_max = ARRAY_SIZE(bist_names);
+
+/* Time to wait between powering down the LNPGA and turning off the power
+ * rails */
+#define LNPGA_PDOWN_WAIT	(HZ / 5)
+
+
+static int crc_error_reset_threshold = 100;
+module_param(crc_error_reset_threshold, int, 0644);
+MODULE_PARM_DESC(crc_error_reset_threshold,
+		 "Max number of CRC errors before XAUI reset");
+
+struct tenxpress_phy_data {
+#ifdef CONFIG_SFC_DEBUGFS
+	char phy_snr[4];
+	char phy_afe_id[TENXPRESS_ID_LEN + 1];
+	char phy_dsp_id[TENXPRESS_ID_LEN + 1];
+	char phy_firmware_id[TENXPRESS_ID_LEN + 1];
+	struct efx_nic *efx;
+#endif
+	enum tenxpress_state state;
+	enum efx_loopback_mode loopback_mode;
+	atomic_t bad_crc_count;
+	int phy_powered;
+	int tx_disabled;
+	int bad_lp_tries;
+};
+
+static int tenxpress_state_is(struct efx_nic *efx, int state)
+{
+	struct tenxpress_phy_data *phy_data = efx->phy_data;
+	return (phy_data != NULL) && (state == phy_data->state);
+}
+
+void tenxpress_set_state(struct efx_nic *efx,
+				enum tenxpress_state state)
+{
+	struct tenxpress_phy_data *phy_data = efx->phy_data;
+	if (phy_data != NULL)
+		phy_data->state = state;
+}
+
+void tenxpress_crc_err(struct efx_nic *efx)
+{
+	struct tenxpress_phy_data *phy_data = efx->phy_data;
+	if (phy_data != NULL)
+		atomic_inc(&phy_data->bad_crc_count);
+}
+
+#ifdef CONFIG_SFC_DEBUGFS
+
+/* debugfs entries for this PHY */
+static int tenxpress_ber_read(struct seq_file *file, void *data)
+{
+	struct efx_nic *efx = *(struct efx_nic **)data;
+	int reg, ber;
+
+	reg = mdio_clause45_read(efx, efx->mii.phy_id, MDIO_MMD_PCS,
+				 MDIO_PCS_10GBT_STATUS2);
+
+	/* Extract the BER */
+	ber = (reg >> MDIO_PCS_10GBT_STATUS2_BER_LBN) &
+		((1 << MDIO_PCS_10GBT_STATUS2_BER_WIDTH) - 1);
+
+	return seq_printf(file, "%d", ber);
+}
+
+
+static int tenxpress_snr_read(struct seq_file *file, void *data)
+{
+	struct tenxpress_phy_data *phy_data = NULL;
+	struct efx_nic *efx;
+	int lane = *(char *) data;
+	int reg, snr;
+
+	EFX_BUG_ON_PARANOID(lane < 0 || lane >= 4);
+	phy_data = container_of(data, struct tenxpress_phy_data, phy_snr[lane]);
+	efx = phy_data->efx;
+
+	reg = mdio_clause45_read(efx, efx->mii.phy_id,
+				 MDIO_MMD_PMAPMD, PMA_PMD_SNR_MARGIN_0 + lane);
+
+	/* Convert from SNR margin to SNR to match phychk output */
+	snr = (reg - 0x8000 + 238);
+
+	return seq_printf(file, "%d.%d", snr / 10, snr % 10);
+}
+
+
+static struct efx_debugfs_parameter debug_entries[] = {
+	EFX_PER_LANE_PARAMETER("phy_lane", "_snr",
+			       struct tenxpress_phy_data, phy_snr, char,
+			       tenxpress_snr_read),
+	EFX_NAMED_PARAMETER(phy_ber, struct tenxpress_phy_data, efx,
+			    struct efx_nic *, tenxpress_ber_read),
+	EFX_STRING_PARAMETER(struct tenxpress_phy_data, phy_afe_id),
+	EFX_STRING_PARAMETER(struct tenxpress_phy_data, phy_dsp_id),
+	EFX_STRING_PARAMETER(struct tenxpress_phy_data, phy_firmware_id),
+	{NULL}
+};
+
+static void tenxpress_phy_get_id(struct efx_nic *efx,
+				 char *id_buf, int id_addr)
+{
+	int i, reg;
+	char ch;
+
+	for (i = TENXPRESS_ID_PN_LEN / 2 - 1; i >= 0; --i) {
+		reg = mdio_clause45_read(efx, efx->mii.phy_id,
+					 MDIO_MMD_PMAPMD, id_addr + i);
+		ch = reg & 0xFF;
+		*id_buf++ = ch ? ch : ' ';
+		ch = (reg & 0xFF00) >> 8;
+		*id_buf++ = ch ? ch : ' ';
+	}
+	*id_buf++ = ' ';
+	for (i = TENXPRESS_ID_REV_LEN / 2 - 1; i >= 0; --i) {
+		reg = mdio_clause45_read(efx, efx->mii.phy_id,
+					 MDIO_MMD_PMAPMD,
+					 id_addr + TENXPRESS_ID_PN_LEN / 2 + i);
+		ch = reg & 0xFF;
+		*id_buf++ = ch ? ch : ' ';
+		ch = (reg & 0xFF00) >> 8;
+		*id_buf++ = ch ? ch : ' ';
+	}
+}
+
+static int tenxpress_debugfs_init(struct efx_nic *efx)
+{
+	struct tenxpress_phy_data *phy_data = efx->phy_data;
+	int lane, rc;
+
+	for (lane = 0; lane < 4; lane++)
+		phy_data->phy_snr[lane] = lane;
+
+	phy_data->efx = efx;
+	rc = efx_extend_debugfs_port(efx, efx->phy_data,
+				     debug_entries);
+	if (rc < 0)
+		return rc;
+
+	tenxpress_phy_get_id(efx, phy_data->phy_afe_id,
+			     PMA_PMD_AFE_ID_REG);
+	tenxpress_phy_get_id(efx, phy_data->phy_dsp_id,
+			     PMA_PMD_DSP_ID_REG);
+	tenxpress_phy_get_id(efx, phy_data->phy_firmware_id,
+			     PMA_PMD_FIRMWARE_ID_REG);
+
+	return 0;
+}
+
+#endif /* CONFIG_SFC_DEBUGFS */
+
+/* Check that the C166 has booted successfully */
+static int tenxpress_phy_check(struct efx_nic *efx)
+{
+	int phy_id = efx->mii.phy_id;
+	int count = PCS_BOOT_MAX_DELAY / PCS_BOOT_POLL_DELAY;
+	int boot_stat;
+
+	/* Wait for the boot to complete (or not) */
+	while (count) {
+		boot_stat = mdio_clause45_read(efx, phy_id,
+					       MDIO_MMD_PCS,
+					       PCS_BOOT_STATUS_REG);
+		if (boot_stat & (1 << PCS_BOOT_COMPLETE_LBN))
+			break;
+		count--;
+		udelay(PCS_BOOT_POLL_DELAY);
+	}
+
+	if (!count) {
+		EFX_ERR(efx, "%s: PHY boot timed out. Last status "
+			"%x\n", __func__,
+			(boot_stat >> PCS_BOOT_PROGRESS_LBN) &
+			((1 << PCS_BOOT_PROGRESS_WIDTH) - 1));
+		return -ETIMEDOUT;
+	}
+
+	return 0;
+}
+
+static void tenxpress_reset_xaui(struct efx_nic *efx);
+
+/* Initialise the part post power on reset or software special reset */
+static int tenxpress_init(struct efx_nic *efx)
+{
+	int rc, reg;
+
+	/* Turn on the clock  */
+	reg = (1 << CLK312_EN_LBN);
+	mdio_clause45_write(efx, efx->mii.phy_id,
+			    MDIO_MMD_PCS, PCS_TEST_SELECT_REG, reg);
+
+	rc = tenxpress_phy_check(efx);
+	if (rc < 0)
+		return rc;
+
+	/* Set the LEDs up as: Green = Link, Amber = Link/Act, Red = Off */
+	reg = mdio_clause45_read(efx, efx->mii.phy_id,
+				 MDIO_MMD_PMAPMD, PMA_PMD_LED_CTRL_REG);
+	reg |= (1 << PMA_PMA_LED_ACTIVITY_LBN);
+	mdio_clause45_write(efx, efx->mii.phy_id, MDIO_MMD_PMAPMD,
+			    PMA_PMD_LED_CTRL_REG, reg);
+
+	reg = PMA_PMD_LED_DEFAULT;
+	mdio_clause45_write(efx, efx->mii.phy_id, MDIO_MMD_PMAPMD,
+			    PMA_PMD_LED_OVERR_REG, reg);
+
+	return rc;
+}
+
+static int tenxpress_phy_init(struct efx_nic *efx)
+{
+	struct tenxpress_phy_data *phy_data;
+	int rc = 0;
+
+	phy_data = kzalloc(sizeof(*phy_data), GFP_KERNEL);
+	efx->phy_data = phy_data;
+	phy_data->phy_powered = efx->phy_powered;
+
+	tenxpress_set_state(efx, TENXPRESS_STATUS_NORMAL);
+
+	rc = mdio_clause45_wait_reset_mmds(efx,
+					   TENXPRESS_REQUIRED_DEVS);
+	if (rc < 0)
+		goto fail;
+
+	rc = mdio_clause45_check_mmds(efx, TENXPRESS_REQUIRED_DEVS, 0);
+	if (rc < 0)
+		goto fail;
+
+	rc = tenxpress_init(efx);
+	if (rc < 0)
+		goto fail;
+
+#ifdef CONFIG_SFC_DEBUGFS
+	rc = tenxpress_debugfs_init(efx);
+	if (rc < 0)
+		goto fail;
+#endif
+
+	schedule_timeout_uninterruptible(HZ / 5); /* 200ms */
+
+	/* Let XGXS and SerDes out of reset and resets 10XPress */
+	falcon_reset_xaui(efx);
+
+	return 0;
+
+ fail:
+	kfree(efx->phy_data);
+	efx->phy_data = NULL;
+	return rc;
+}
+
+static int tenxpress_special_reset(struct efx_nic *efx)
+{
+	int rc, reg;
+
+	EFX_TRACE(efx, "%s\n", __func__);
+
+	/* Initiate reset */
+	reg = mdio_clause45_read(efx, efx->mii.phy_id,
+				 MDIO_MMD_PMAPMD, PMA_PMD_EXT_CTRL_REG);
+	reg |= (1 << PMA_PMD_EXT_SSR_LBN);
+	mdio_clause45_write(efx, efx->mii.phy_id, MDIO_MMD_PMAPMD,
+			    PMA_PMD_EXT_CTRL_REG, reg);
+
+	msleep(200);
+
+	/* Wait for the blocks to come out of reset */
+	rc = mdio_clause45_wait_reset_mmds(efx,
+					   TENXPRESS_REQUIRED_DEVS);
+	if (rc < 0)
+		return rc;
+
+	/* Try and reconfigure the device */
+	rc = tenxpress_init(efx);
+	if (rc < 0)
+		return rc;
+
+	return 0;
+}
+
+static void tenxpress_set_bad_lp(struct efx_nic *efx, int bad_lp)
+{
+	struct tenxpress_phy_data *pd = efx->phy_data;
+	int reg;
+
+	/* Nothing to do if all is well and was previously so. */
+	if (!(bad_lp || pd->bad_lp_tries))
+		return;
+
+	reg = mdio_clause45_read(efx, efx->mii.phy_id,
+				 MDIO_MMD_PMAPMD, PMA_PMD_LED_OVERR_REG);
+
+	if (bad_lp)
+		pd->bad_lp_tries++;
+	else
+		pd->bad_lp_tries = 0;
+
+	if (pd->bad_lp_tries == MAX_BAD_LP_TRIES) {
+		pd->bad_lp_tries = 0;	/* Restart count */
+		reg &= ~(PMA_PMD_LED_FLASH << PMA_PMD_LED_RX_LBN);
+		reg |= (PMA_PMD_LED_FLASH << PMA_PMD_LED_RX_LBN);
+		EFX_ERR(efx, "This NIC appears to be plugged into"
+			" a port that is not 10GBASE-T capable.\n"
+			" This PHY is 10GBASE-T ONLY, so no link can"
+			" be established.\n");
+	} else {
+		reg |= (PMA_PMD_LED_OFF << PMA_PMD_LED_RX_LBN);
+	}
+	mdio_clause45_write(efx, efx->mii.phy_id, MDIO_MMD_PMAPMD,
+			    PMA_PMD_LED_OVERR_REG, reg);
+}
+
+/* Check link status and return a boolean OK value. If the link is NOT
+ * OK we have a quick rummage round to see if we appear to be plugged
+ * into a non-10GBT port and if so warn the user that they won't get
+ * link any time soon as we are 10GBT only, unless caller specified
+ * not to do this check (it isn't useful in loopback) */
+static int tenxpress_link_ok(struct efx_nic *efx, int check_lp)
+{
+	int ok = mdio_clause45_links_ok(efx, TENXPRESS_REQUIRED_DEVS);
+
+	if (ok) {
+		tenxpress_set_bad_lp(efx, 0);
+	} else if (check_lp) {
+		/* Are we plugged into the wrong sort of link? */
+		int bad_lp = 0;
+		int phy_id = efx->mii.phy_id;
+		int an_stat = mdio_clause45_read(efx, phy_id, MDIO_MMD_AN,
+						 MDIO_AN_STATUS);
+		int xphy_stat = mdio_clause45_read(efx, phy_id,
+						   MDIO_MMD_PMAPMD,
+						   PMA_PMD_XSTATUS_REG);
+		/* Are we plugged into anything that sends FLPs? If
+		 * not we can't distinguish between not being plugged
+		 * in and being plugged into a non-AN antique. The FLP
+		 * bit has the advantage of not clearing when autoneg
+		 * restarts. */
+		if (!(xphy_stat & (1 << PMA_PMD_XSTAT_FLP_LBN))) {
+			tenxpress_set_bad_lp(efx, 0);
+			return ok;
+		}
+
+		/* If it can do 10GBT it must be XNP capable */
+		bad_lp = !(an_stat & (1 << MDIO_AN_STATUS_XNP_LBN));
+		if (!bad_lp && (an_stat & (1 << MDIO_AN_STATUS_PAGE_LBN))) {
+			bad_lp = !(mdio_clause45_read(efx, phy_id,
+					MDIO_MMD_AN, MDIO_AN_10GBT_STATUS) &
+					(1 << MDIO_AN_10GBT_STATUS_LP_10G_LBN));
+		}
+		tenxpress_set_bad_lp(efx, bad_lp);
+	}
+	return ok;
+}
+
+static void tenxpress_phyxs_loopback(struct efx_nic *efx)
+{
+	int phy_id = efx->mii.phy_id;
+	int ctrl1, ctrl2;
+
+	ctrl1 = ctrl2 = mdio_clause45_read(efx, phy_id, MDIO_MMD_PHYXS,
+					   PHYXS_TEST1);
+	if (efx->loopback_mode == LOOPBACK_PHYXS)
+		ctrl2 |= (1 << LOOPBACK_NEAR_LBN);
+	else
+		ctrl2 &= ~(1 << LOOPBACK_NEAR_LBN);
+	if (ctrl1 != ctrl2)
+		mdio_clause45_write(efx, phy_id, MDIO_MMD_PHYXS,
+				    PHYXS_TEST1, ctrl2);
+}
+
+static void tenxpress_phy_reconfigure(struct efx_nic *efx)
+{
+	struct tenxpress_phy_data *phy_data = efx->phy_data;
+	int loop_change = LOOPBACK_OUT_OF(phy_data, efx,
+					  TENXPRESS_LOOPBACKS);
+
+	if (!tenxpress_state_is(efx, TENXPRESS_STATUS_NORMAL))
+		return;
+
+	/* When coming out of transmit disable, coming out of low power
+	 * mode, or moving out of any PHY internal loopback mode,
+	 * perform a special software reset */
+	if (((efx->phy_powered && !efx->tx_disabled) &&
+	     (!phy_data->phy_powered || phy_data->tx_disabled)) ||
+	    loop_change) {
+		(void) tenxpress_special_reset(efx);
+		falcon_reset_xaui(efx);
+	}
+
+	mdio_clause45_transmit_disable(efx, efx->tx_disabled);
+	mdio_clause45_phy_reconfigure(efx);
+	tenxpress_phyxs_loopback(efx);
+
+	phy_data->tx_disabled = efx->tx_disabled;
+	phy_data->loopback_mode = efx->loopback_mode;
+	phy_data->phy_powered = efx->phy_powered;
+	efx->link_up = tenxpress_link_ok(efx, 0);
+	efx->link_options = GM_LPA_10000FULL;
+}
+
+static void tenxpress_phy_clear_interrupt(struct efx_nic *efx)
+{
+	/* Nothing done here - LASI interrupts aren't reliable so poll  */
+}
+
+
+/* Poll PHY for interrupt */
+static int tenxpress_phy_check_hw(struct efx_nic *efx)
+{
+	struct tenxpress_phy_data *phy_data = efx->phy_data;
+	int phy_up = tenxpress_state_is(efx, TENXPRESS_STATUS_NORMAL);
+	int link_ok, rc = 0;
+
+	link_ok = phy_up && tenxpress_link_ok(efx, 1);
+
+	if (link_ok != efx->link_up) {
+		efx->link_up = link_ok;
+		efx->mac_op->fake_phy_event(efx);
+	}
+
+	/* Nothing to check if we've already shut down the PHY */
+	if (!phy_up)
+		return 0;
+
+	if (atomic_read(&phy_data->bad_crc_count) > crc_error_reset_threshold) {
+		EFX_ERR(efx, "Resetting XAUI due to too many CRC errors\n");
+		falcon_reset_xaui(efx);
+		atomic_set(&phy_data->bad_crc_count, 0);
+	}
+
+	rc = efx->board_info.monitor(efx);
+	if (rc)
+		efx->link_up = 0;
+
+	return rc;
+}
+
+static void tenxpress_phy_fini(struct efx_nic *efx)
+{
+	int reg;
+
+#ifdef CONFIG_SFC_DEBUGFS
+	efx_trim_debugfs_port(efx, debug_entries);
+#endif
+	/* Power down the LNPGA */
+	reg = (1 << PMA_PMD_LNPGA_POWERDOWN_LBN);
+	mdio_clause45_write(efx, efx->mii.phy_id, MDIO_MMD_PMAPMD,
+			    PMA_PMD_XCONTROL_REG, reg);
+
+	/* Waiting here ensures that the board fini, which can turn off the
+	 * power to the PHY, won't get run until the LNPGA powerdown has been
+	 * given long enough to complete. */
+	schedule_timeout_uninterruptible(LNPGA_PDOWN_WAIT); /* 200 ms */
+
+	kfree(efx->phy_data);
+	efx->phy_data = NULL;
+}
+
+
+/* Set the RX and TX LEDs and Link LED flashing. The other LEDs
+ * (which probably aren't wired anyway) are left in AUTO mode */
+void tenxpress_phy_blink(struct efx_nic *efx, int blink)
+{
+	int reg;
+
+	if (blink)
+		reg = (PMA_PMD_LED_FLASH << PMA_PMD_LED_TX_LBN) |
+			(PMA_PMD_LED_FLASH << PMA_PMD_LED_RX_LBN) |
+			(PMA_PMD_LED_FLASH << PMA_PMD_LED_LINK_LBN);
+	else
+		reg = PMA_PMD_LED_DEFAULT;
+
+	mdio_clause45_write(efx, efx->mii.phy_id, MDIO_MMD_PMAPMD,
+			    PMA_PMD_LED_OVERR_REG, reg);
+}
+
+static void tenxpress_reset_xaui(struct efx_nic *efx)
+{
+	int phy = efx->mii.phy_id;
+	int clk_ctrl, test_select, soft_rst2;
+
+	/* Real work is done on clock_ctrl other resets are thought to be
+	 * optional but make the reset more reliable
+	 */
+
+	/* Read */
+	clk_ctrl = mdio_clause45_read(efx, phy, MDIO_MMD_PCS,
+				      PCS_CLOCK_CTRL_REG);
+	test_select = mdio_clause45_read(efx, phy, MDIO_MMD_PCS,
+					 PCS_TEST_SELECT_REG);
+	soft_rst2 = mdio_clause45_read(efx, phy, MDIO_MMD_PCS,
+				       PCS_SOFT_RST2_REG);
+
+	/* Modify => put in reset */
+	test_select &= ~(1 << CLK312_EN_LBN);
+	mdio_clause45_write(efx, phy, MDIO_MMD_PCS,
+			    PCS_TEST_SELECT_REG, test_select);
+
+	soft_rst2 &= ~((1 << XGXS_RST_N_LBN) | (1 << SERDES_RST_N_LBN));
+	mdio_clause45_write(efx, phy, MDIO_MMD_PCS,
+			    PCS_SOFT_RST2_REG, soft_rst2);
+
+	clk_ctrl &= ~(1 << PLL312_RST_N_LBN);
+	mdio_clause45_write(efx, phy, MDIO_MMD_PCS,
+			    PCS_CLOCK_CTRL_REG, clk_ctrl);
+	udelay(10);
+
+	/* Modify => remove reset */
+	clk_ctrl |= (1 << PLL312_RST_N_LBN);
+	mdio_clause45_write(efx, phy, MDIO_MMD_PCS,
+			    PCS_CLOCK_CTRL_REG, clk_ctrl);
+	udelay(10);
+
+	soft_rst2 |= ((1 << XGXS_RST_N_LBN) | (1 << SERDES_RST_N_LBN));
+	mdio_clause45_write(efx, phy, MDIO_MMD_PCS,
+			    PCS_SOFT_RST2_REG, soft_rst2);
+	udelay(10);
+
+	test_select |= (1 << CLK312_EN_LBN);
+	mdio_clause45_write(efx, phy, MDIO_MMD_PCS,
+			    PCS_TEST_SELECT_REG, test_select);
+	udelay(10);
+}
+
+
+struct efx_phy_operations falcon_tenxpress_phy_ops = {
+	.init             = tenxpress_phy_init,
+	.reconfigure      = tenxpress_phy_reconfigure,
+	.check_hw         = tenxpress_phy_check_hw,
+	.fini             = tenxpress_phy_fini,
+	.clear_interrupt  = tenxpress_phy_clear_interrupt,
+	.reset_xaui       = tenxpress_reset_xaui,
+	.mmds             = TENXPRESS_REQUIRED_DEVS,
+	.loopbacks        = TENXPRESS_LOOPBACKS,
+	.startup_loopback = LOOPBACK_PCS,
+};
diff -rpuN linux-2.6.18.8/drivers/net/sfc/tx.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/tx.c
--- linux-2.6.18.8/drivers/net/sfc/tx.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/tx.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,522 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2005-2006: Fen Systems Ltd.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/pci.h>
+#include <linux/tcp.h>
+#include <linux/ip.h>
+#include <linux/in.h>
+#include <linux/if_ether.h>
+#include <linux/version.h>
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20)
+#include <linux/highmem.h>
+#endif
+#include "net_driver.h"
+#include "tx.h"
+#include "efx.h"
+#include "falcon.h"
+#include "workarounds.h"
+
+
+/*
+ * TX descriptor ring full threshold
+ *
+ * The tx_queue descriptor ring fill-level must fall below this value
+ * before we restart the netif queue
+ */
+#define EFX_NETDEV_TX_THRESHOLD(_tx_queue)	\
+	(_tx_queue->efx->type->txd_ring_mask / 2u)
+
+
+
+/* We want to be able to nest calls to netif_stop_queue(), since each
+ * channel can have an individual stop on the queue.
+ */
+void efx_stop_queue(struct efx_nic *efx)
+{
+	spin_lock_bh(&efx->netif_stop_lock);
+	EFX_TRACE(efx, "stop TX queue\n");
+
+	atomic_inc(&efx->netif_stop_count);
+	if (likely(efx->net_dev_registered))
+		netif_stop_queue(efx->net_dev);
+
+	spin_unlock_bh(&efx->netif_stop_lock);
+}
+
+/* Wake netif's TX queue
+ * We want to be able to nest calls to netif_stop_queue(), since each
+ * channel can have an individual stop on the queue.
+ */
+inline void efx_wake_queue(struct efx_nic *efx)
+{
+	local_bh_disable();
+	if (atomic_dec_and_lock(&efx->netif_stop_count,
+				&efx->netif_stop_lock)) {
+		EFX_TRACE(efx, "waking TX queue\n");
+		if (likely(efx->net_dev_registered))
+			netif_wake_queue(efx->net_dev);
+		spin_unlock(&efx->netif_stop_lock);
+	}
+	local_bh_enable();
+}
+
+/*
+ * Add a socket buffer to a TX queue
+ *
+ * This maps all fragments of a socket buffer for DMA and adds them to
+ * the TX queue.  The queue's insert pointer will be incremented by
+ * the number of fragments in the socket buffer.
+ *
+ * If any DMA mapping fails, any mapped fragments will be unmapped,
+ * the queue's insert pointer will be restored to its original value.
+ *
+ * Returns NETDEV_TX_OK or NETDEV_TX_BUSY
+ * You must hold netif_tx_lock() to call this function.
+ */
+static inline int efx_enqueue_skb(struct efx_tx_queue *tx_queue,
+				  const struct sk_buff *skb)
+{
+	struct efx_nic *efx = tx_queue->efx;
+	struct pci_dev *pci_dev = efx->pci_dev;
+	struct efx_tx_buffer *buffer;
+	skb_frag_t *fragment;
+	struct page *page;
+	int page_offset;
+	unsigned int len, unmap_len = 0, fill_level, insert_ptr, misalign;
+	dma_addr_t dma_addr, unmap_addr = 0;
+	unsigned int dma_len;
+	unsigned unmap_single;
+	int q_space, i = 0;
+	int rc = NETDEV_TX_OK;
+
+	EFX_BUG_ON_PARANOID(tx_queue->write_count != tx_queue->insert_count);
+
+	/* Get size of the initial fragment */
+	len = skb_headlen(skb);
+
+	fill_level = tx_queue->insert_count - tx_queue->old_read_count;
+	q_space = efx->type->txd_ring_mask - 1 - fill_level;
+
+	/* Map for DMA.  Use pci_map_single rather than pci_map_page
+	 * since this is more efficient on machines with sparse
+	 * memory.
+	 */
+	unmap_single = 1;
+	dma_addr = pci_map_single(pci_dev, skb->data, len, PCI_DMA_TODEVICE);
+
+	/* Process all fragments */
+	while (1) {
+		if (unlikely(pci_dma_mapping_error(dma_addr)))
+			goto pci_err;
+
+		/* Store fields for marking in the per-fragment final
+		 * descriptor */
+		unmap_len = len;
+		unmap_addr = dma_addr;
+
+		/* Add to TX queue, splitting across DMA boundaries */
+		do {
+			if (unlikely(q_space-- <= 0)) {
+				/* It might be that completions have
+				 * happened since the xmit path last
+				 * checked.  Update the xmit path's
+				 * copy of read_count.
+				 */
+				++tx_queue->stopped;
+				/* This memory barrier protects the
+				 * change of stopped from the access
+				 * of read_count. */
+				smp_mb();
+				tx_queue->old_read_count =
+					*(volatile unsigned *)
+					&tx_queue->read_count;
+				fill_level = (tx_queue->insert_count
+					      - tx_queue->old_read_count);
+				q_space = (efx->type->txd_ring_mask - 1 -
+					   fill_level);
+				if (unlikely(q_space-- <= 0))
+					goto stop;
+				smp_mb();
+				--tx_queue->stopped;
+			}
+
+			insert_ptr = (tx_queue->insert_count &
+				      efx->type->txd_ring_mask);
+			buffer = &tx_queue->buffer[insert_ptr];
+			EFX_BUG_ON_PARANOID(buffer->skb);
+			EFX_BUG_ON_PARANOID(buffer->len);
+			EFX_BUG_ON_PARANOID(buffer->continuation != 1);
+			EFX_BUG_ON_PARANOID(buffer->unmap_len);
+
+			dma_len = (((~dma_addr) & efx->type->tx_dma_mask) + 1);
+			if (likely(dma_len > len))
+				dma_len = len;
+
+			misalign = (unsigned)dma_addr & efx->type->bug5391_mask;
+			if (misalign && dma_len + misalign > 512)
+				dma_len = 512 - misalign;
+
+			/* Fill out per descriptor fields */
+			buffer->len = dma_len;
+			buffer->dma_addr = dma_addr;
+			len -= dma_len;
+			dma_addr += dma_len;
+			++tx_queue->insert_count;
+		} while (len);
+
+		/* Transfer ownership of the unmapping to the final buffer */
+		buffer->unmap_addr = unmap_addr;
+		buffer->unmap_single = unmap_single;
+		buffer->unmap_len = unmap_len;
+		unmap_len = 0;
+
+		/* Get address and size of next fragment */
+		if (i >= skb_shinfo(skb)->nr_frags)
+			break;
+		fragment = &skb_shinfo(skb)->frags[i];
+		len = fragment->size;
+		page = fragment->page;
+		page_offset = fragment->page_offset;
+		i++;
+		/* Map for DMA */
+		unmap_single = 0;
+		dma_addr = pci_map_page(pci_dev, page, page_offset, len,
+					PCI_DMA_TODEVICE);
+	}
+
+	/* Transfer ownership of the skb to the final buffer */
+	buffer->skb = skb;
+	buffer->continuation = 0;
+
+	/* Pass off to hardware */
+	falcon_push_buffers(tx_queue);
+
+	return NETDEV_TX_OK;
+
+ pci_err:
+	EFX_ERR_RL(efx, " TX queue %d could not map skb with %d bytes %d "
+		   "fragments for DMA\n", tx_queue->queue, skb->len,
+		   skb_shinfo(skb)->nr_frags + 1);
+
+	/* Mark the packet as transmitted, and free the SKB ourselves */
+	dev_kfree_skb_any((struct sk_buff *)skb);
+	goto unwind;
+
+ stop:
+	rc = NETDEV_TX_BUSY;
+
+	/* Stop the queue if it wasn't stopped before. */
+	if (tx_queue->stopped == 1)
+		efx_stop_queue(efx);
+
+ unwind:
+	/* Work backwards until we hit the original insert pointer value */
+	while (tx_queue->insert_count != tx_queue->write_count) {
+		--tx_queue->insert_count;
+		insert_ptr = tx_queue->insert_count & efx->type->txd_ring_mask;
+		buffer = &tx_queue->buffer[insert_ptr];
+		if (buffer->unmap_len) {
+			if (buffer->unmap_single)
+				pci_unmap_single(pci_dev, buffer->unmap_addr,
+						 buffer->unmap_len,
+						 PCI_DMA_TODEVICE);
+			else
+				pci_unmap_page(pci_dev, buffer->unmap_addr,
+					       buffer->unmap_len,
+					       PCI_DMA_TODEVICE);
+		}
+		buffer->unmap_len = 0;
+		buffer->len = 0;
+	}
+
+	/* Free the fragment we were mid-way through pushing */
+	if (unmap_len)
+		pci_unmap_page(pci_dev, unmap_addr, unmap_len,
+			       PCI_DMA_TODEVICE);
+
+	return rc;
+}
+
+/* Remove packets from the TX queue
+ *
+ * This removes packets from the TX queue, up to and including the
+ * specified index.
+ */
+static inline void efx_dequeue_buffers(struct efx_tx_queue *tx_queue,
+				       unsigned int index)
+{
+	struct pci_dev *pci_dev = tx_queue->efx->pci_dev;
+	struct efx_tx_buffer *buffer;
+	unsigned int stop_index, read_ptr;
+
+	/* Calculate the stopping point.  Doing the check this way
+	 * avoids wrongly completing every buffer in the ring if we
+	 * get called twice with the same index.  (Hardware should
+	 * never do this, since it can't complete that many buffers in
+	 * one go.)
+	 */
+	stop_index = (index + 1) & tx_queue->efx->type->txd_ring_mask;
+	read_ptr = tx_queue->read_count & tx_queue->efx->type->txd_ring_mask;
+
+	while (read_ptr != stop_index) {
+		buffer = &tx_queue->buffer[read_ptr];
+		if (unlikely(buffer->len == 0)) {
+			EFX_ERR(tx_queue->efx, "TX queue %d spurious TX "
+				"completion id %x\n", tx_queue->queue,
+				read_ptr);
+			atomic_inc(&tx_queue->efx->errors.spurious_tx);
+			/* Don't reset */
+		} else {
+			if (buffer->unmap_len) {
+				if (buffer->unmap_single)
+					pci_unmap_single(pci_dev,
+							 buffer->unmap_addr,
+							 buffer->unmap_len,
+							 PCI_DMA_TODEVICE);
+				else
+					pci_unmap_page(pci_dev,
+						       buffer->unmap_addr,
+						       buffer->unmap_len,
+						       PCI_DMA_TODEVICE);
+				buffer->unmap_single = 0;
+				buffer->unmap_len = 0;
+			}
+			if (buffer->skb) {
+				dev_kfree_skb_any((struct sk_buff *)
+						  buffer->skb);
+				buffer->skb = NULL;
+				EFX_TRACE(tx_queue->efx, "TX queue %d "
+					  "transmission id %x complete\n",
+					  tx_queue->queue, read_ptr);
+			}
+			buffer->continuation = 1;
+			buffer->len = 0;
+		}
+		++tx_queue->read_count;
+		read_ptr = (tx_queue->read_count &
+			    tx_queue->efx->type->txd_ring_mask);
+	}
+}
+
+/* Initiate a packet transmission on the specified TX queue.
+ * Note that returning anything other than NETDEV_TX_OK will cause the
+ * OS to free the skb.
+ *
+ * This function is split out from efx_hard_start_xmit to allow the
+ * loopback test to direct packets via specific TX queues.  It is
+ * therefore a non-static inline, so as not to penalise performance
+ * for non-loopback transmissions.
+ *
+ * Context: netif_tx_lock held
+ */
+inline int efx_xmit(struct efx_nic *efx,
+		    struct efx_tx_queue *tx_queue, struct sk_buff *skb)
+{
+	int rc;
+
+	/* Map fragments for DMA and add to TX queue */
+	rc = efx_enqueue_skb(tx_queue, skb);
+	if (unlikely(rc != NETDEV_TX_OK))
+		goto out;
+
+	/* Update last TX timer */
+	efx->net_dev->trans_start = jiffies;
+
+ out:
+	return rc;
+}
+
+/* Initiate a packet transmission.  We use one channel per CPU
+ * (sharing when we have more CPUs than channels).  On Falcon, the TX
+ * completion events will be directed back to the CPU that transmitted
+ * the packet, which should be cache-efficient.
+ *
+ * Context: non-blocking.
+ * Note that returning anything other than NETDEV_TX_OK will cause the
+ * OS to free the skb.
+ */
+int efx_hard_start_xmit(struct sk_buff *skb, struct net_device *net_dev)
+{
+	struct efx_nic *efx = net_dev->priv;
+	struct efx_tx_queue *tx_queue;
+	enum efx_veto veto;
+	int rc = NETDEV_TX_OK;
+
+	/* We have one TX queue. */
+	tx_queue = &efx->tx_queue[0];
+
+	/* See if driverlink wants to veto the packet. */
+	veto = EFX_DL_CALLBACK(efx, tx_packet, skb);
+	if (unlikely(veto)) {
+		EFX_LOG(efx, "TX queue %d packet vetoed by "
+			"driverlink %s driver\n", tx_queue->queue,
+			efx->dl_cb_dev.tx_packet->driver->name);
+		/* Free the skb; nothing else will do it */
+		dev_kfree_skb_any((struct sk_buff *)skb);
+		goto out;
+	}
+
+	rc = efx_xmit(efx, tx_queue, skb);
+out:
+	return rc;
+}
+
+#if defined(EFX_USE_FASTCALL)
+void fastcall efx_xmit_done(struct efx_tx_queue *tx_queue, unsigned int index)
+#else
+void efx_xmit_done(struct efx_tx_queue *tx_queue, unsigned int index)
+#endif
+{
+	unsigned long flags __attribute__ ((unused));
+	unsigned fill_level;
+	struct efx_nic *efx = tx_queue->efx;
+
+	EFX_BUG_ON_PARANOID(index > efx->type->txd_ring_mask);
+
+	/* Remove buffers from TX queue */
+	efx_dequeue_buffers(tx_queue, index);
+
+	/* See if we need to restart the netif queue.  This barrier
+	 * separates the update of read_count from the test of
+	 * stopped. */
+	smp_mb();
+	if (unlikely(tx_queue->stopped)) {
+		fill_level = tx_queue->insert_count - tx_queue->read_count;
+		if (fill_level < EFX_NETDEV_TX_THRESHOLD(tx_queue)) {
+			/* If the port is stopped and the net_dev isn't
+			 * registered, then the caller must be performing
+			 * flow control manually */
+			if (unlikely(!efx->net_dev_registered))
+				return;
+
+			/* Do this under netif_tx_lock(), to avoid racing
+			 * with efx_xmit(). */
+			netif_tx_lock(efx->net_dev);
+			if (tx_queue->stopped) {
+				tx_queue->stopped = 0;
+				efx_wake_queue(efx);
+			}
+			netif_tx_unlock(efx->net_dev);
+		}
+	}
+}
+
+int efx_probe_tx_queue(struct efx_tx_queue *tx_queue)
+{
+	struct efx_nic *efx = tx_queue->efx;
+	unsigned int txq_size;
+	int i, rc;
+
+	EFX_LOG(efx, "creating TX queue %d\n", tx_queue->queue);
+
+	/* Allocate software ring */
+	txq_size = (efx->type->txd_ring_mask + 1) * sizeof(*tx_queue->buffer);
+	tx_queue->buffer = kzalloc(txq_size, GFP_KERNEL);
+	if (!tx_queue->buffer) {
+		rc = -ENOMEM;
+		goto fail1;
+	}
+	for (i = 0; i <= efx->type->txd_ring_mask; ++i)
+		tx_queue->buffer[i].continuation = 1;
+
+	/* Allocate hardware ring */
+	rc = falcon_probe_tx(tx_queue);
+	if (rc)
+		goto fail2;
+
+	return 0;
+
+ fail2:
+	kfree(tx_queue->buffer);
+	tx_queue->buffer = NULL;
+ fail1:
+	/* Mark queue as unused */
+	tx_queue->used = 0;
+
+	return rc;
+}
+
+int efx_init_tx_queue(struct efx_tx_queue *tx_queue)
+{
+	EFX_LOG(tx_queue->efx, "initialising TX queue %d\n", tx_queue->queue);
+
+	ASSERT_RTNL();
+
+	/* Initialise fields */
+	tx_queue->insert_count = 0;
+	tx_queue->write_count = 0;
+	tx_queue->read_count = 0;
+	tx_queue->old_read_count = 0;
+	BUG_ON(tx_queue->stopped);
+
+	/* Set up TX descriptor ring */
+	return falcon_init_tx(tx_queue);
+}
+
+void efx_release_tx_buffers(struct efx_tx_queue *tx_queue)
+{
+	unsigned int last_index, mask;
+	if (tx_queue->buffer) {
+		/* Free any buffers left in the ring */
+		mask = tx_queue->efx->type->txd_ring_mask;
+		last_index = (tx_queue->insert_count - 1) & mask;
+		EFX_LOG(tx_queue->efx, "Will dequeue up to 0x%x from 0x%x\n",
+			last_index, tx_queue->read_count & mask);
+		efx_dequeue_buffers(tx_queue, last_index);
+	}
+}
+
+void efx_fini_tx_queue(struct efx_tx_queue *tx_queue)
+{
+	EFX_LOG(tx_queue->efx, "shutting down TX queue %d\n", tx_queue->queue);
+
+	ASSERT_RTNL();
+
+	/* Flush TX queue, remove descriptor ring */
+	falcon_fini_tx(tx_queue);
+
+	/* Release TX buffers */
+	efx_release_tx_buffers(tx_queue);
+
+	/* Release queue's stop on port, if any */
+	if (tx_queue->stopped) {
+		tx_queue->stopped = 0;
+		efx_wake_queue(tx_queue->efx);
+	}
+}
+
+void efx_remove_tx_queue(struct efx_tx_queue *tx_queue)
+{
+	EFX_LOG(tx_queue->efx, "destroying TX queue %d\n", tx_queue->queue);
+	falcon_remove_tx(tx_queue);
+
+	kfree(tx_queue->buffer);
+	tx_queue->buffer = NULL;
+	tx_queue->used = 0;
+}
+
+
diff -rpuN linux-2.6.18.8/drivers/net/sfc/txc43128_phy.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/txc43128_phy.c
--- linux-2.6.18.8/drivers/net/sfc/txc43128_phy.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/txc43128_phy.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,725 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+/*
+ * Driver for Transwitch/Mysticom CX4 retimer
+ * see www.transwitch.com, part is TXC-43128
+ */
+
+#include <linux/delay.h>
+#include <linux/seq_file.h>
+#include "efx.h"
+#include "debugfs.h"
+#include "gmii.h"
+#include "mdio_10g.h"
+#include "xenpack.h"
+#include "phy.h"
+#include "lm87_support.h"
+#include "falcon.h"
+#include "workarounds.h"
+
+/* We expect these MMDs to be in the package */
+#define TXC_REQUIRED_DEVS (MDIO_MMDREG_DEVS0_PCS |	\
+			   MDIO_MMDREG_DEVS0_PMAPMD |	\
+			   MDIO_MMDREG_DEVS0_PHYXS)
+
+#define TXC_LOOPBACKS ((1 << LOOPBACK_PCS) |		\
+		       (1 << LOOPBACK_PMAPMD) |		\
+		       (1 << LOOPBACK_NETWORK))
+
+/**************************************************************************
+ *
+ * Compile-time config
+ *
+ **************************************************************************
+ */
+#define TXCNAME "TXC43128"
+/* Total length of time we'll wait for the PHY to come out of reset */
+#define TXC_MAX_RESET_TIME 500
+/* Interval between checks */
+#define TXC_RESET_WAIT 10
+/* How long to run BIST: At 10Gbps 50 microseconds should be plenty to get
+ * some stats */
+#define TXC_BIST_DURATION (50)
+
+#define BER_INTERVAL (10 * efx_monitor_interval)
+
+/**************************************************************************
+ *
+ * Register definitions
+ *
+ **************************************************************************
+ */
+#define XAUI_NUM_LANES (4)
+
+/*** Global register bank */
+/* Silicon ID register */
+#define TXC_GLRGS_SLID		(0xc000)
+#define TXC_GLRGS_SLID_MASK	(0x1f)
+
+/* Command register */
+#define TXC_GLRGS_GLCMD		(0xc004)
+/* Useful bits in command register */
+/* Lane power-down */
+#define TXC_GLCMD_L01PD_LBN	(5)
+#define TXC_GLCMD_L23PD_LBN	(6)
+/* Limited SW reset: preserves configuration but
+ * initiates a logic reset. Self-clearing */
+#define TXC_GLCMD_LMTSWRST_LBN	(14)
+
+/* Signal Quality Control */
+#define TXC_GLRGS_GSGQLCTL	(0xc01a)
+/* Enable bit */
+#define TXC_GSGQLCT_SGQLEN_LBN	(15)
+/* Lane selection */
+#define TXC_GSGQLCT_LNSL_LBN	(13)
+#define TXC_GSGQLCT_LNSL_WIDTH	(2)
+
+/* Signal Quality Input */
+#define TXC_GLRGS_GSGQLIN	(0xc01b)
+/*  Signal Quality Grade */
+#define TXC_GLRGS_GSGQLGRD	(0xc01c)
+/* Drift sign */
+#define TXC_GSGQLGRD_DRFTSGN_LBN (15)
+/* Grade valid flag */
+#define TXC_GSGQLGRD_GRDVAL_LBN	(14)
+/* Remaining bits are the actual grade */
+#define TXC_GSGQLGRD_GRADE_LBN	(0)
+#define TXC_GSGQLGRD_GRADE_WIDTH (14)
+
+/*  Signal Quality Drift: 16-bit drift value */
+#define TXC_GLRGS_GSGQLDRFT	(0xc01d)
+
+/**** Analog register bank */
+#define TXC_ALRGS_ATXCTL	(0xc040)
+/* Lane power-down */
+#define TXC_ATXCTL_TXPD3_LBN	(15)
+#define TXC_ATXCTL_TXPD2_LBN	(14)
+#define TXC_ATXCTL_TXPD1_LBN	(13)
+#define TXC_ATXCTL_TXPD0_LBN	(12)
+
+/* Amplitude on lanes 0, 1 */
+#define  TXC_ALRGS_ATXAMP0	(0xc041)
+/* Amplitude on lanes 2, 3 */
+#define  TXC_ALRGS_ATXAMP1	(0xc042)
+/* Bit position of value for lane 0 (or 2) */
+#define TXC_ATXAMP_LANE02_LBN	(3)
+/* Bit position of value for lane 1 (or 3) */
+#define TXC_ATXAMP_LANE13_LBN	(11)
+
+#define TXC_ATXAMP_1280_mV	(0)
+#define TXC_ATXAMP_1200_mV	(8)
+#define TXC_ATXAMP_1120_mV	(12)
+#define TXC_ATXAMP_1060_mV	(14)
+#define TXC_ATXAMP_0820_mV	(25)
+#define TXC_ATXAMP_0720_mV	(26)
+#define TXC_ATXAMP_0580_mV	(27)
+#define TXC_ATXAMP_0440_mV	(28)
+
+#define TXC_ATXAMP_0820_BOTH					\
+	((TXC_ATXAMP_0820_mV << TXC_ATXAMP_LANE02_LBN)		\
+	 | (TXC_ATXAMP_0820_mV << TXC_ATXAMP_LANE13_LBN))
+
+#define TXC_ATXAMP_DEFAULT	(0x6060) /* From databook */
+
+/* Preemphasis on lanes 0, 1 */
+#define  TXC_ALRGS_ATXPRE0	(0xc043)
+/* Preemphasis on lanes 2, 3 */
+#define  TXC_ALRGS_ATXPRE1	(0xc044)
+
+#define TXC_ATXPRE_NONE (0)
+#define TXC_ATXPRE_DEFAULT	(0x1010) /* From databook */
+
+#define TXC_ALRGS_ARXCTL	(0xc045)
+/* Lane power-down */
+#define TXC_ARXCTL_RXPD3_LBN	(15)
+#define TXC_ARXCTL_RXPD2_LBN	(14)
+#define TXC_ARXCTL_RXPD1_LBN	(13)
+#define TXC_ARXCTL_RXPD0_LBN	(12)
+
+/*** receiver control registers: Bit Error Rate measurement */
+/* Per lane BER timers */
+#define TXC_RXCTL_BERTMR0	(0xc0d4)
+#define TXC_RXCTL_BERTMR1	(0xc154)
+#define TXC_RXCTL_BERTMR2	(0xc1d4)
+#define TXC_RXCTL_BERTMR3	(0xc254)
+/* Per lane BER counters */
+#define TXC_RXCTL_BERCNT0	(0xc0d5)
+#define TXC_RXCTL_BERCNT1	(0xc155)
+#define TXC_RXCTL_BERCNT2	(0xc1d5)
+#define TXC_RXCTL_BERCNT3	(0xc255)
+
+#define BER_REG_SPACING	(TXC_RXCTL_BERTMR1 - TXC_RXCTL_BERTMR0)
+
+/*** Main user-defined register set */
+/* Main control */
+#define TXC_MRGS_CTL		(0xc340)
+/* Bits in main control */
+#define TXC_MCTL_RESET_LBN	(15)	/* Self clear */
+#define TXC_MCTL_TXLED_LBN	(14)	/* 1 to show align status */
+#define TXC_MCTL_RXLED_LBN	(13)	/* 1 to show align status */
+
+/* GPIO output */
+#define TXC_GPIO_OUTPUT		(0xc346)
+#define TXC_GPIO_DIR		(0xc348)
+
+/*** Vendor-specific BIST registers */
+#define TXC_BIST_CTL		(0xc280)
+#define TXC_BIST_TXFRMCNT	(0xc281)
+#define TXC_BIST_RX0FRMCNT	(0xc282)
+#define TXC_BIST_RX1FRMCNT	(0xc283)
+#define TXC_BIST_RX2FRMCNT	(0xc284)
+#define TXC_BIST_RX3FRMCNT	(0xc285)
+#define TXC_BIST_RX0ERRCNT	(0xc286)
+#define TXC_BIST_RX1ERRCNT	(0xc287)
+#define TXC_BIST_RX2ERRCNT	(0xc288)
+#define TXC_BIST_RX3ERRCNT	(0xc289)
+
+/*** BIST control bits */
+/* BIST type (controls bit patter in test) */
+#define TXC_BIST_CTRL_TYPE_LBN	(10)
+#define	TXC_BIST_CTRL_TYPE_TSD	(0)	/* TranSwitch Deterministic */
+#define TXC_BIST_CTRL_TYPE_CRP	(1)	/* CRPAT standard */
+#define TXC_BIST_CTRL_TYPE_CJP	(2)	/* CJPAT standard */
+#define TXC_BIST_CTRL_TYPE_TSR	(3)	/* TranSwitch pseudo-random */
+/* Set this to 1 for 10 bit and 0 for 8 bit */
+#define TXC_BIST_CTRL_B10EN_LBN	(12)
+/* Enable BIST (write 0 to disable) */
+#define	TXC_BIST_CTRL_ENAB_LBN	(13)
+/*Stop BIST (self-clears when stop complete) */
+#define  TXC_BIST_CTRL_STOP_LBN	(14)
+/* Start BIST (cleared by writing 1 to STOP) */
+#define  TXC_BIST_CTRL_STRT_LBN	(15)
+
+/* Mt. Diablo test configuration */
+#define TXC_MTDIABLO_CTRL	(0xc34f)
+#define TXC_MTDIABLO_CTRL_PMA_LOOP_LBN	(10)
+
+struct txc43128_data {
+#ifdef CONFIG_SFC_DEBUGFS
+	/* BER stats update from check_hw. Note that this is in errors/second,
+	 * converting it to errors/bit is left as an exercise for user-space.
+	 */
+	unsigned phy_ber_pcs[4];
+	unsigned phy_ber_phyxs[4];
+#endif
+	unsigned bug10934_timer;
+	int phy_powered;
+	int tx_disabled;
+	enum efx_loopback_mode loopback_mode;
+};
+
+/* Perform the bug 10934 workaround every 5s */
+#define BUG10934_RESET_INTERVAL (5 * HZ)
+
+
+/* Perform a reset that doesn't clear configuration changes */
+static void txc_reset_logic(struct efx_nic *efx);
+
+/* Set the output value of a gpio */
+void txc_set_gpio_val(struct efx_nic *efx, int pin, int on)
+{
+	int outputs;
+
+	outputs = mdio_clause45_read(efx, efx->mii.phy_id,
+					MDIO_MMD_PHYXS, TXC_GPIO_OUTPUT);
+
+	outputs = (outputs & ~(1 << pin)) | (on << pin);
+
+	mdio_clause45_write(efx, efx->mii.phy_id,
+					MDIO_MMD_PHYXS, TXC_GPIO_OUTPUT,
+					outputs);
+}
+
+/* Set up the GPIO direction register */
+void txc_set_gpio_dir(struct efx_nic *efx, int pin, int dir)
+{
+	int dirs;
+
+	if (efx->board_info.minor < 3 &&
+		   efx->board_info.major == 0)
+		return;
+
+	dirs = mdio_clause45_read(efx, efx->mii.phy_id,
+			    MDIO_MMD_PHYXS, TXC_GPIO_DIR);
+	dirs = (dir & ~(1 << pin)) | (dir << pin);
+	mdio_clause45_write(efx, efx->mii.phy_id,
+			    MDIO_MMD_PHYXS, TXC_GPIO_DIR, dirs);
+
+}
+
+/* Reset the PMA/PMD MMD. The documentation is explicit that this does a
+ * global reset (it's less clear what reset of other MMDs does).*/
+static int txc_reset_phy(struct efx_nic *efx)
+{
+	int rc = mdio_clause45_reset_mmd(efx, MDIO_MMD_PMAPMD,
+					 TXC_MAX_RESET_TIME / TXC_RESET_WAIT,
+					 TXC_RESET_WAIT);
+	if (rc < 0)
+		goto fail;
+
+	/* Check that all the MMDs we expect are present and responding. We
+	 * expect faults on some if the link is down, but not on the PHY XS */
+	rc = mdio_clause45_check_mmds(efx, TXC_REQUIRED_DEVS, 0);
+	if (rc < 0)
+		goto fail;
+
+	return 0;
+
+ fail:
+	EFX_ERR(efx, TXCNAME ": reset timed out!\n");
+	return rc;
+}
+
+/* Run a single BIST on one MMD*/
+static int txc_bist_one(struct efx_nic *efx, int mmd, int test)
+{
+	int phy = efx->mii.phy_id;
+	int ctrl, bctl;
+	int lane;
+	int rc = 0;
+
+	EFX_INFO(efx, "" TXCNAME ": running BIST on %s MMD\n",
+		 mdio_clause45_mmd_name(mmd));
+
+	/* Set PMA to test into loopback using Mt Diablo reg as per app note */
+	ctrl = mdio_clause45_read(efx, phy, MDIO_MMD_PCS,
+				  TXC_MTDIABLO_CTRL);
+	ctrl |= (1 << TXC_MTDIABLO_CTRL_PMA_LOOP_LBN);
+	mdio_clause45_write(efx, phy, MDIO_MMD_PCS,
+			    TXC_MTDIABLO_CTRL, ctrl);
+
+
+	/* The BIST app. note lists these  as 3 distinct steps. */
+	/* Set the BIST type */
+	bctl = (test << TXC_BIST_CTRL_TYPE_LBN);
+	mdio_clause45_write(efx, phy, mmd, TXC_BIST_CTL, bctl);
+
+	/* Set the BSTEN bit in the BIST Control register to enable */
+	bctl |= (1 << TXC_BIST_CTRL_ENAB_LBN);
+	mdio_clause45_write(efx, phy, mmd, TXC_BIST_CTL, bctl);
+
+	/* Set the BSTRT bit in the BIST Control register */
+	mdio_clause45_write(efx, phy, mmd, TXC_BIST_CTL, bctl |
+			    (1 << TXC_BIST_CTRL_STRT_LBN));
+
+	/* Wait. */
+	udelay(TXC_BIST_DURATION);
+
+	/* Set the BSTOP bit in the BIST Control register */
+	bctl |= (1 << TXC_BIST_CTRL_STOP_LBN);
+	mdio_clause45_write(efx, phy, mmd, TXC_BIST_CTL, bctl);
+
+	/* The STOP bit should go off when things have stopped */
+	while (bctl & (1 << TXC_BIST_CTRL_STOP_LBN))
+		bctl = mdio_clause45_read(efx, phy, mmd, TXC_BIST_CTL);
+
+	/* Check all the error counts are 0 and all the frame counts are
+	   non-zero */
+	for (lane = 0; lane < 4; lane++) {
+		int count = mdio_clause45_read(efx, phy, mmd,
+					       TXC_BIST_RX0ERRCNT + lane);
+		if (count != 0) {
+			EFX_ERR(efx, ""TXCNAME": BIST error. "
+				"Lane %d had %d errs\n", lane, count);
+			rc = -EIO;
+		}
+		count = mdio_clause45_read(efx, phy, mmd,
+					   TXC_BIST_RX0FRMCNT + lane);
+		if (count == 0) {
+			EFX_ERR(efx, ""TXCNAME": BIST error. "
+				"Lane %d got 0 frames\n", lane);
+			rc = -EIO;
+		}
+	}
+
+	if (rc == 0)
+		EFX_INFO(efx, ""TXCNAME": BIST pass\n");
+
+	/* Disable BIST */
+	mdio_clause45_write(efx, phy, mmd, TXC_BIST_CTL, 0);
+
+	/* Turn off loopback */
+	ctrl &= ~(1 << TXC_MTDIABLO_CTRL_PMA_LOOP_LBN);
+	mdio_clause45_write(efx, phy, MDIO_MMD_PCS,
+			    TXC_MTDIABLO_CTRL, ctrl);
+
+	return rc;
+}
+
+/* Run all the desired BIST tests for the PHY */
+static int txc_bist(struct efx_nic *efx)
+{
+	int rc;
+	/*!\todo: experiment with running more of the BIST patterns to
+	 * see if it actually shows up more problems. */
+	rc = txc_bist_one(efx, MDIO_MMD_PCS, TXC_BIST_CTRL_TYPE_TSD);
+	return rc;
+}
+
+#ifdef CONFIG_SFC_DEBUGFS
+
+/* debugfs entries for this PHY */
+static struct efx_debugfs_parameter debug_entries[] = {
+	EFX_PER_LANE_PARAMETER("phy_ber_lane", "_pcs",
+			       struct txc43128_data, phy_ber_pcs,
+			       unsigned, efx_debugfs_read_uint),
+	EFX_PER_LANE_PARAMETER("phy_ber_lane", "_phyxs",
+			       struct txc43128_data, phy_ber_phyxs,
+			       unsigned, efx_debugfs_read_uint),
+	EFX_INT_PARAMETER(struct txc43128_data, phy_powered),
+	{NULL}
+};
+
+#endif /* CONFIG_SFC_DEBUGFS */
+
+/* Push the non-configurable defaults into the PHY. This must be
+ * done after every full reset */
+static void txc_apply_defaults(struct efx_nic *efx)
+{
+	int mctrl;
+
+	/* Turn amplitude down and preemphasis off on the host side
+	 * (PHY<->MAC) as this is believed less likely to upset Falcon
+	 * and no adverse effects have been noted. It probably also
+	 * saves a picowatt or two */
+
+	/* Turn off preemphasis */
+	mdio_clause45_write(efx, efx->mii.phy_id, MDIO_MMD_PHYXS,
+			    TXC_ALRGS_ATXPRE0, TXC_ATXPRE_NONE);
+	mdio_clause45_write(efx, efx->mii.phy_id, MDIO_MMD_PHYXS,
+			    TXC_ALRGS_ATXPRE1, TXC_ATXPRE_NONE);
+
+	/* Turn down the amplitude */
+	mdio_clause45_write(efx, efx->mii.phy_id, MDIO_MMD_PHYXS,
+			    TXC_ALRGS_ATXAMP0, TXC_ATXAMP_0820_BOTH);
+	mdio_clause45_write(efx, efx->mii.phy_id, MDIO_MMD_PHYXS,
+			    TXC_ALRGS_ATXAMP1, TXC_ATXAMP_0820_BOTH);
+
+	/* Set the line side amplitude and preemphasis to the databook
+	 * defaults as an erratum causes them to be 0 on at least some
+	 * PHY rev.s */
+	mdio_clause45_write(efx, efx->mii.phy_id, MDIO_MMD_PMAPMD,
+			    TXC_ALRGS_ATXPRE0, TXC_ATXPRE_DEFAULT);
+	mdio_clause45_write(efx, efx->mii.phy_id, MDIO_MMD_PMAPMD,
+			    TXC_ALRGS_ATXPRE1, TXC_ATXPRE_DEFAULT);
+	mdio_clause45_write(efx, efx->mii.phy_id, MDIO_MMD_PMAPMD,
+			    TXC_ALRGS_ATXAMP0, TXC_ATXAMP_DEFAULT);
+	mdio_clause45_write(efx, efx->mii.phy_id, MDIO_MMD_PMAPMD,
+			    TXC_ALRGS_ATXAMP1, TXC_ATXAMP_DEFAULT);
+
+	/* Set up the LEDs  */
+	mctrl = mdio_clause45_read(efx, efx->mii.phy_id,
+				   MDIO_MMD_PHYXS, TXC_MRGS_CTL);
+
+	/* Set the Green and Red LEDs to their default modes */
+	mctrl &= ~((1 << TXC_MCTL_TXLED_LBN) | (1 << TXC_MCTL_RXLED_LBN));
+	mdio_clause45_write(efx, efx->mii.phy_id,
+			    MDIO_MMD_PHYXS, TXC_MRGS_CTL, mctrl);
+
+	/* Databook recommends doing this after configuration changes */
+	txc_reset_logic(efx);
+
+	efx->board_info.init_leds(efx);
+}
+
+/* Initialisation entry point for this PHY driver */
+static int txc43128_phy_init(struct efx_nic *efx)
+{
+	u32 devid;
+	int rc = 0;
+	struct txc43128_data *phy_data;
+
+	devid = mdio_clause45_read_id(efx, MDIO_MMD_PHYXS);
+
+	phy_data = kzalloc(sizeof(*phy_data), GFP_KERNEL);
+	efx->phy_data = phy_data;
+
+	/* This is the default after reset */
+	phy_data->phy_powered = efx->phy_powered;
+	phy_data->tx_disabled = efx->tx_disabled;
+
+#ifdef CONFIG_SFC_DEBUGFS
+	rc = efx_extend_debugfs_port(efx, phy_data, debug_entries);
+	if (rc < 0)
+		goto fail1;
+#endif
+	EFX_INFO(efx, ""TXCNAME ": PHY ID reg %x (OUI %x model %x "
+		 "revision %x)\n", devid, MDIO_ID_OUI(devid),
+		 MDIO_ID_MODEL(devid), MDIO_ID_REV(devid));
+
+	EFX_INFO(efx, ""TXCNAME ": Silicon ID %x\n",
+		 mdio_clause45_read(efx, efx->mii.phy_id,
+				    MDIO_MMD_PHYXS, TXC_GLRGS_SLID) &
+					TXC_GLRGS_SLID_MASK);
+
+	rc = txc_reset_phy(efx);
+	if (rc < 0)
+		goto fail2;
+
+	rc = txc_bist(efx);
+	if (rc < 0)
+		goto fail2;
+
+	txc_apply_defaults(efx);
+
+	return 0;
+
+ fail2:
+#ifdef CONFIG_SFC_DEBUGFS
+	efx_trim_debugfs_port(efx, debug_entries);
+	/* fall-thru */
+ fail1:
+#endif
+	kfree(efx->phy_data);
+	efx->phy_data = NULL;
+	return rc;
+}
+
+/* Set the lane power down state in the global registers */
+static void txc_glrgs_lane_power(struct efx_nic *efx, int mmd)
+{
+	int pd = (1 << TXC_GLCMD_L01PD_LBN) | (1 << TXC_GLCMD_L23PD_LBN);
+	int ctl = mdio_clause45_read(efx, efx->mii.phy_id,
+				     mmd, TXC_GLRGS_GLCMD);
+
+	if (efx->phy_powered)
+		ctl &= ~pd;
+	else
+		ctl |= pd;
+
+	mdio_clause45_write(efx, efx->mii.phy_id,
+			    mmd, TXC_GLRGS_GLCMD, ctl);
+}
+
+/* Set the lane power down state in the analog control registers */
+static void txc_analog_lane_power(struct efx_nic *efx, int mmd)
+{
+	int txpd = (1 << TXC_ATXCTL_TXPD3_LBN) | (1 << TXC_ATXCTL_TXPD2_LBN)
+		| (1 << TXC_ATXCTL_TXPD1_LBN) | (1 << TXC_ATXCTL_TXPD0_LBN);
+
+	int rxpd = (1 << TXC_ATXCTL_TXPD3_LBN) | (1 << TXC_ATXCTL_TXPD2_LBN)
+		| (1 << TXC_ATXCTL_TXPD1_LBN) | (1 << TXC_ATXCTL_TXPD0_LBN);
+
+	int txctl = mdio_clause45_read(efx, efx->mii.phy_id,
+				       mmd, TXC_ALRGS_ATXCTL);
+	int rxctl = mdio_clause45_read(efx, efx->mii.phy_id,
+				       mmd, TXC_ALRGS_ARXCTL);
+
+	if (efx->phy_powered) {
+		txctl &= ~txpd;
+		rxctl &= ~rxpd;
+	} else {
+		txctl |= txpd;
+		rxctl |= rxpd;
+	}
+
+	mdio_clause45_write(efx, efx->mii.phy_id,
+			    mmd, TXC_ALRGS_ATXCTL, txctl);
+	mdio_clause45_write(efx, efx->mii.phy_id,
+			    mmd, TXC_ALRGS_ARXCTL, rxctl);
+}
+
+static void txc_set_power(struct efx_nic *efx)
+{
+	/* According to the data book, all the MMDs can do low power */
+	mdio_clause45_set_mmds_lpower(efx, !efx->phy_powered,
+				      TXC_REQUIRED_DEVS);
+
+	/* Global register bank is in PCS, PHY XS. These control the host
+	 * side and line side settings respectively. */
+	txc_glrgs_lane_power(efx, MDIO_MMD_PCS);
+	txc_glrgs_lane_power(efx, MDIO_MMD_PHYXS);
+
+	/* Analog register bank in PMA/PMD, PHY XS */
+	txc_analog_lane_power(efx, MDIO_MMD_PMAPMD);
+	txc_analog_lane_power(efx, MDIO_MMD_PHYXS);
+}
+
+
+static void txc_reset_logic_mmd(struct efx_nic *efx, int mmd)
+{
+	int portid = efx->mii.phy_id;
+	int val = mdio_clause45_read(efx, portid, mmd, TXC_GLRGS_GLCMD);
+	int tries = 50;
+	val |= (1 << TXC_GLCMD_LMTSWRST_LBN);
+	mdio_clause45_write(efx, portid, mmd, TXC_GLRGS_GLCMD, val);
+	while (tries--) {
+		val = mdio_clause45_read(efx, portid, mmd,
+					 TXC_GLRGS_GLCMD);
+		if (!(val & (1 << TXC_GLCMD_LMTSWRST_LBN)))
+			break;
+		udelay(1);
+	}
+	if (!tries)
+		EFX_INFO(efx, TXCNAME " Logic reset timed out!\n");
+}
+
+
+/* Perform a logic reset. This preserves the configuration registers
+ * and is needed for some configuration changes to take effect */
+static void txc_reset_logic(struct efx_nic *efx)
+{
+	/* The data sheet claims we can do the logic reset on either the
+	 * PCS or the PHYXS and the result is a reset of both host- and
+	 * line-side logic. */
+	txc_reset_logic_mmd(efx, MDIO_MMD_PCS);
+}
+
+static int txc43128_phy_read_link(struct efx_nic *efx)
+{
+	return mdio_clause45_links_ok(efx, TXC_REQUIRED_DEVS);
+}
+
+static void txc43128_phy_reconfigure(struct efx_nic *efx)
+{
+	struct txc43128_data *phy_data = efx->phy_data;
+	int power_change = (efx->phy_powered != phy_data->phy_powered);
+	int loop_change = LOOPBACK_CHANGED(phy_data, efx, TXC_LOOPBACKS);
+	int disable_change = (efx->tx_disabled != phy_data->tx_disabled);
+
+	if (!phy_data->tx_disabled && efx->tx_disabled) {
+		txc_reset_phy(efx);
+		txc_apply_defaults(efx);
+		falcon_reset_xaui(efx);
+		disable_change = 0;
+	}
+
+	mdio_clause45_transmit_disable(efx, efx->tx_disabled);
+	mdio_clause45_phy_reconfigure(efx);
+	if (power_change)
+		txc_set_power(efx);
+
+	/* The data sheet claims this is required after every reconfiguration
+	 * (note at end of 7.1), but we mustn't do it when nothing changes as
+	 * it glitches the link, and reconfigure gets called on link change,
+	 * so we get an IRQ storm on link up. */
+	if (loop_change || power_change || disable_change)
+		txc_reset_logic(efx);
+
+	phy_data->phy_powered = efx->phy_powered;
+	phy_data->loopback_mode = efx->loopback_mode;
+	phy_data->tx_disabled = efx->tx_disabled;
+	efx->link_up = txc43128_phy_read_link(efx);
+	efx->link_options = GM_LPA_10000FULL;
+}
+
+static void txc43128_phy_fini(struct efx_nic *efx)
+{
+	efx->board_info.blink(efx, 0);
+
+	/* Disable link events */
+	xenpack_disable_lasi_irqs(efx);
+
+#ifdef CONFIG_SFC_DEBUGFS
+	/* Remove the extra debug entries and free data */
+	efx_trim_debugfs_port(efx, debug_entries);
+#endif
+	kfree(efx->phy_data);
+	efx->phy_data = NULL;
+}
+
+/* Periodic callback: this exists mainly to poll link status as we currently
+ * don't use LASI interrupts. Also update the BER counters and poll the lm87 */
+static int txc43128_phy_check_hw(struct efx_nic *efx)
+{
+	struct txc43128_data *data = efx->phy_data;
+#ifdef CONFIG_SFC_DEBUGFS
+	int phy = efx->mii.phy_id;
+	int timer, count, i, mmd;
+#endif
+	int rc = 0;
+	int link_up = txc43128_phy_read_link(efx);
+
+	/* Simulate a PHY event if link state has changed */
+	if (link_up != efx->link_up) {
+		efx->link_up = link_up;
+		efx->mac_op->fake_phy_event(efx);
+	} else if (EFX_WORKAROUND_10934(efx)) {
+		if (link_up || (efx->loopback_mode != LOOPBACK_NONE))
+			data->bug10934_timer = jiffies;
+		else {
+			int delta = jiffies - data->bug10934_timer;
+			if (delta >= BUG10934_RESET_INTERVAL) {
+				data->bug10934_timer = jiffies;
+				txc_reset_logic(efx);
+			}
+		}
+	}
+
+	rc = efx->board_info.monitor(efx);
+	if (rc) {
+		EFX_ERR(efx, "" TXCNAME
+			": sensor alert! Putting PHY into low power.\n");
+		efx->phy_powered = 0;
+		txc_set_power(efx);
+	}
+
+#ifdef CONFIG_SFC_DEBUGFS
+	/* There are 2 MMDs with RX BER counters: PCS and PHY XS,
+	 * which happen to be consecutively numbered */
+	for (mmd = MDIO_MMD_PCS; mmd <= MDIO_MMD_PHYXS; mmd++) {
+		for (i = 0; i < XAUI_NUM_LANES; i++) {
+			timer = mdio_clause45_read(efx, phy, mmd,
+						   TXC_RXCTL_BERTMR0 +
+						   i * BER_REG_SPACING);
+			count = mdio_clause45_read(efx, phy, mmd,
+						   TXC_RXCTL_BERCNT0 +
+						   i * BER_REG_SPACING);
+			/* The BER timer counts down in seconds. If it would
+			 * expire before the next check_hw, update the stats &
+			 * restart the timer (clears the count) */
+			if (timer * HZ < efx_monitor_interval) {
+				/* Record count, allowing for the fact that the
+				 * timer may not have reached zero */
+				unsigned ber = (count * BER_INTERVAL) /
+					(BER_INTERVAL - timer * HZ);
+				if (mmd == MDIO_MMD_PCS)
+					data->phy_ber_pcs[i] = ber;
+				else
+					data->phy_ber_phyxs[i] = ber;
+				/* Reprogram the timer */
+				mdio_clause45_write(efx, phy, mmd,
+						    TXC_RXCTL_BERTMR0 +
+						    i * BER_REG_SPACING,
+						    BER_INTERVAL / HZ);
+			}
+		}
+		mmd = (mmd == MDIO_MMD_PCS) ? MDIO_MMD_PHYXS : 0;
+	}
+#endif /* CONFIG_SFC_DEBUGFS */
+	return rc;
+}
+
+struct efx_phy_operations falcon_txc_phy_ops = {
+	.init             = txc43128_phy_init,
+	.reconfigure      = txc43128_phy_reconfigure,
+	.check_hw         = txc43128_phy_check_hw,
+	.fini             = txc43128_phy_fini,
+	.clear_interrupt  = efx_port_dummy_op_void,
+	.reset_xaui       = efx_port_dummy_op_void,
+	.mmds             = TXC_REQUIRED_DEVS,
+	.loopbacks        = TXC_LOOPBACKS,
+	.startup_loopback = LOOPBACK_PMAPMD,
+};
diff -rpuN linux-2.6.18.8/drivers/net/sfc/tx.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/tx.h
--- linux-2.6.18.8/drivers/net/sfc/tx.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/tx.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,41 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2006:      Fen Systems Ltd.
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Initially developed by Michael Brown <mbrown@fensystems.co.uk>
+ * Maintained by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_TX_H
+#define EFX_TX_H
+
+#include "net_driver.h"
+
+int efx_probe_tx_queue(struct efx_tx_queue *tx_queue);
+void efx_remove_tx_queue(struct efx_tx_queue *tx_queue);
+int efx_init_tx_queue(struct efx_tx_queue *tx_queue);
+void efx_fini_tx_queue(struct efx_tx_queue *tx_queue);
+
+int efx_hard_start_xmit(struct sk_buff *skb, struct net_device *net_dev);
+void efx_release_tx_buffers(struct efx_tx_queue *tx_queue);
+
+#endif /* EFX_TX_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/workarounds.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/workarounds.h
--- linux-2.6.18.8/drivers/net/sfc/workarounds.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/workarounds.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,97 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_WORKAROUNDS_H
+#define EFX_WORKAROUNDS_H
+
+/*
+ * Hardware workarounds.
+ * Bug numbers are from Solarflare's Bugzilla.
+ */
+
+#define EFX_WORKAROUND_ALWAYS(efx) 1
+#define EFX_WORKAROUND_FALCON_A(efx) (FALCON_REV(efx) <= FALCON_REV_A1)
+#define EFX_WORKAROUND_FALCON_B0FPGA(efx) \
+	(FALCON_REV(efx) == FALCON_REV_B0 && !(efx)->is_asic)
+
+/* XAUI resets if link not detected */
+#define EFX_WORKAROUND_5147 EFX_WORKAROUND_ALWAYS
+/* SNAP frames have TOBE_DISC set */
+#define EFX_WORKAROUND_5475 EFX_WORKAROUND_ALWAYS
+/* PHY interrupts can go to the wrong port */
+#define EFX_WORKAROUND_6263 EFX_WORKAROUND_ALWAYS
+/* Reprog PCIe ACK timer to workaround issue in PCIe IP block */
+#define EFX_WORKAROUND_6943 EFX_WORKAROUND_ALWAYS
+/* RX PCIe double split performance issue */
+#define EFX_WORKAROUND_7575 EFX_WORKAROUND_ALWAYS
+/* Bit-bashed I2C reads cause performance drop */
+#define EFX_WORKAROUND_7884 EFX_WORKAROUND_ALWAYS
+/* Selftests need to be retried */
+#define EFX_WORKAROUND_8909 EFX_WORKAROUND_ALWAYS
+/* Queued ACKs aren't flushed before L1 entry */
+#define EFX_WORKAROUND_9096 EFX_WORKAROUND_ALWAYS
+/* TX pkt parser problem with <= 16 byte TXes */
+#define EFX_WORKAROUND_9141 EFX_WORKAROUND_ALWAYS
+/* XGXS and XAUI reset sequencing in SW */
+#define EFX_WORKAROUND_9388 EFX_WORKAROUND_ALWAYS
+/* Low rate CRC errors require XAUI reset */
+#define EFX_WORKAROUND_10750 EFX_WORKAROUND_ALWAYS
+/* TX_EV_PKT_ERR can be caused by a dangling TX descriptor
+ * or a PCIe error (bug 11028) */
+#define EFX_WORKAROUND_10727 EFX_WORKAROUND_ALWAYS
+/* CX4 retimer fails to bring link up after reset */
+#define EFX_WORKAROUND_10934 EFX_WORKAROUND_ALWAYS
+/* Transmit flow control may get disabled */
+#define EFX_WORKAROUND_11482 EFX_WORKAROUND_ALWAYS
+/* Flush events can take a very long time to appear */
+#define EFX_WORKAROUND_11557 EFX_WORKAROUND_ALWAYS
+
+/* Spurious parity errors in TSORT buffers */
+#define EFX_WORKAROUND_5129 EFX_WORKAROUND_FALCON_A
+/* No unaligned TX over 512 byte boundaries */
+#define EFX_WORKAROUND_5391 EFX_WORKAROUND_FALCON_A
+/* iSCSI parsing errors */
+#define EFX_WORKAROUND_5583 EFX_WORKAROUND_FALCON_A
+/* RX events go missing */
+#define EFX_WORKAROUND_5676 EFX_WORKAROUND_FALCON_A
+/* RX_RESET on A1 */
+#define EFX_WORKAROUND_6555 EFX_WORKAROUND_FALCON_A
+/* Spurious duplicate RX events */
+#define EFX_WORKAROUND_7062 EFX_WORKAROUND_FALCON_A
+/* Increase filter depth to avoid RX_RESET */
+#define EFX_WORKAROUND_7244 EFX_WORKAROUND_FALCON_A
+/* Flushes may never complete */
+#define EFX_WORKAROUND_7803 EFX_WORKAROUND_FALCON_A
+/* Leak overlength packets rather than free */
+#define EFX_WORKAROUND_8071 EFX_WORKAROUND_FALCON_A
+
+/* Memory needs clearing at start-of-day */
+#define EFX_WORKAROUND_8202 EFX_WORKAROUND_FALCON_B0FPGA
+/* MAC statistics are transient */
+#define EFX_WORKAROUND_8419 EFX_WORKAROUND_FALCON_B0FPGA
+/* Prefetch watchdog timer may trigger erroneously on busy systems */
+#define EFX_WORKAROUND_9008 EFX_WORKAROUND_FALCON_B0FPGA
+
+#endif /* EFX_WORKAROUNDS_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/xenpack.h linux-2.6.18-xen-3.3.0/drivers/net/sfc/xenpack.h
--- linux-2.6.18.8/drivers/net/sfc/xenpack.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/xenpack.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,80 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2006:      Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef EFX_XENPACK_H
+#define EFX_XENPACK_H
+
+/* Exported functions from Xenpack standard PHY control */
+
+#include "mdio_10g.h"
+
+/****************************************************************************/
+/* XENPACK MDIO register extensions */
+#define MDIO_XP_LASI_RX_CTRL	(0x9000)
+#define MDIO_XP_LASI_TX_CTRL	(0x9001)
+#define MDIO_XP_LASI_CTRL	(0x9002)
+#define MDIO_XP_LASI_RX_STAT	(0x9003)
+#define MDIO_XP_LASI_TX_STAT	(0x9004)
+#define MDIO_XP_LASI_STAT	(0x9005)
+
+/* Control/Status bits */
+#define XP_LASI_LS_ALARM	(1 << 0)
+#define XP_LASI_TX_ALARM	(1 << 1)
+#define XP_LASI_RX_ALARM	(1 << 2)
+/* These two are Quake vendor extensions to the standard XENPACK defines */
+#define XP_LASI_LS_INTB		(1 << 3)
+#define XP_LASI_TEST		(1 << 7)
+
+/* Enable LASI interrupts for PHY */
+static inline void xenpack_enable_lasi_irqs(struct efx_nic *efx)
+{
+	int reg;
+	int phy_id = efx->mii.phy_id;
+	/* Read to clear LASI status register */
+	reg = mdio_clause45_read(efx, phy_id, MDIO_MMD_PMAPMD,
+				 MDIO_XP_LASI_STAT);
+
+	/* Enable LASI interrupts from PMA/PMD */
+	mdio_clause45_write(efx, phy_id, MDIO_MMD_PMAPMD,
+			    MDIO_XP_LASI_CTRL, XP_LASI_LS_ALARM);
+}
+
+/* Read the LASI interrupt status to clear the interrupt. */
+static inline int xenpack_clear_lasi_irqs(struct efx_nic *efx)
+{
+	/* Read to clear link status alarm */
+	return mdio_clause45_read(efx, efx->mii.phy_id,
+				  MDIO_MMD_PMAPMD, MDIO_XP_LASI_STAT);
+}
+
+/* Turn off LASI interrupts */
+static inline void xenpack_disable_lasi_irqs(struct efx_nic *efx)
+{
+	/* Turn LASI interrupts off */
+	mdio_clause45_write(efx, efx->mii.phy_id, MDIO_MMD_PMAPMD,
+			    MDIO_XP_LASI_CTRL, 0);
+}
+
+#endif /* EFX_XENPACK_H */
diff -rpuN linux-2.6.18.8/drivers/net/sfc/xfp_phy.c linux-2.6.18-xen-3.3.0/drivers/net/sfc/xfp_phy.c
--- linux-2.6.18.8/drivers/net/sfc/xfp_phy.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/net/sfc/xfp_phy.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,206 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed by Solarflare Communications <linux-net-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+/*
+ * Driver for XFP optical PHYs (plus some support specific to the Quake 2032)
+ * See www.amcc.com for details (search for qt2032)
+ */
+
+#include <linux/timer.h>
+#include <linux/delay.h>
+#include "efx.h"
+#include "gmii.h"
+#include "mdio_10g.h"
+#include "xenpack.h"
+#include "phy.h"
+
+#define XFP_REQUIRED_DEVS (MDIO_MMDREG_DEVS0_PCS |	\
+			   MDIO_MMDREG_DEVS0_PMAPMD |	\
+			   MDIO_MMDREG_DEVS0_PHYXS)
+
+#define XFP_LOOPBACKS ((1 << LOOPBACK_PCS) |		\
+		       (1 << LOOPBACK_PMAPMD) |		\
+		       (1 << LOOPBACK_NETWORK))
+
+/****************************************************************************/
+/* Quake-specific MDIO registers */
+#define MDIO_QUAKE_LED0_REG	(0xD006)
+
+
+void xfp_set_led(struct efx_nic *p, int led, int mode)
+{
+	int addr = MDIO_QUAKE_LED0_REG + led;
+	mdio_clause45_write(p, p->mii.phy_id, MDIO_MMD_PMAPMD, addr,
+			    mode);
+}
+
+struct xfp_phy_data {
+	int phy_powered;
+	int tx_disabled;
+};
+
+
+#define XFP_MAX_RESET_TIME 500
+#define XFP_RESET_WAIT 10
+
+/* Reset the PHYXS MMD. This is documented (for the Quake PHY) as doing
+ * a complete soft reset.
+ */
+static int xfp_reset_phy(struct efx_nic *efx)
+{
+	int rc;
+
+	rc = mdio_clause45_reset_mmd(efx, MDIO_MMD_PHYXS,
+				     XFP_MAX_RESET_TIME / XFP_RESET_WAIT,
+				     XFP_RESET_WAIT);
+	if (rc < 0)
+		goto fail;
+
+	/* Wait 250ms for the PHY to complete bootup */
+	msleep(250);
+
+	/* Check that all the MMDs we expect are present and responding. We
+	 * expect faults on some if the link is down, but not on the PHY XS */
+	rc = mdio_clause45_check_mmds(efx, XFP_REQUIRED_DEVS,
+				      MDIO_MMDREG_DEVS0_PHYXS);
+	if (rc < 0)
+		goto fail;
+
+	efx->board_info.init_leds(efx);
+
+	return rc;
+
+ fail:
+	EFX_ERR(efx, "XFP: reset timed out!\n");
+	return rc;
+}
+
+
+static int xfp_phy_init(struct efx_nic *efx)
+{
+	struct xfp_phy_data *phy_data;
+	u32 devid = mdio_clause45_read_id(efx, MDIO_MMD_PHYXS);
+	int rc;
+
+	phy_data = kzalloc(sizeof(struct xfp_phy_data), GFP_KERNEL);
+	efx->phy_data = (void *) phy_data;
+
+	EFX_INFO(efx, "XFP: PHY ID reg %x (OUI %x model %x revision"
+		 " %x)\n", devid, MDIO_ID_OUI(devid), MDIO_ID_MODEL(devid),
+		 MDIO_ID_REV(devid));
+
+	phy_data->phy_powered = efx->phy_powered;
+	phy_data->tx_disabled = efx->tx_disabled;
+
+	rc = xfp_reset_phy(efx);
+	if (rc < 0)
+		goto fail;
+
+	EFX_INFO(efx, "XFP: PHY init %s.\n",
+		 rc ? "failed" : "successful");
+	return 0;
+
+ fail:
+	kfree(efx->phy_data);
+	efx->phy_data = NULL;
+	return rc;
+}
+
+static void xfp_phy_clear_interrupt(struct efx_nic *efx)
+{
+	xenpack_clear_lasi_irqs(efx);
+}
+
+static int xfp_link_ok(struct efx_nic *efx)
+{
+	return mdio_clause45_links_ok(efx, XFP_REQUIRED_DEVS);
+}
+
+static int xfp_phy_check_hw(struct efx_nic *efx)
+{
+	int rc = 0;
+	int link_up = xfp_link_ok(efx);
+	/* Simulate a PHY event if link state has changed */
+	if (link_up != efx->link_up) {
+		efx->link_up = link_up;
+		efx->mac_op->fake_phy_event(efx);
+	}
+
+	rc = efx->board_info.monitor(efx);
+	if (rc) {
+		EFX_ERR(efx, ": XFP sensor alert! Putting PHY into "
+			"low power.\n");
+		efx->phy_powered = 0;
+
+		mdio_clause45_set_mmds_lpower(efx, 1, XFP_REQUIRED_DEVS);
+	}
+
+	return rc;
+}
+
+static void xfp_phy_reconfigure(struct efx_nic *efx)
+{
+	struct xfp_phy_data *phy_data = efx->phy_data;
+
+	/* Reset the PHY when moving from transmitter off or powered off,
+	 * to transmitter on and powered on */
+	if ((efx->phy_powered && !efx->tx_disabled) &&
+	    (!phy_data->phy_powered || phy_data->tx_disabled))
+		xfp_reset_phy(efx);
+
+	mdio_clause45_transmit_disable(efx, efx->tx_disabled);
+	mdio_clause45_set_mmds_lpower(efx, !efx->phy_powered,
+				      XFP_REQUIRED_DEVS);
+	mdio_clause45_phy_reconfigure(efx);
+
+	phy_data->tx_disabled = efx->tx_disabled;
+	phy_data->phy_powered = efx->phy_powered;
+	efx->link_up = xfp_link_ok(efx);
+	efx->link_options = GM_LPA_10000FULL;
+}
+
+
+static void xfp_phy_fini(struct efx_nic *efx)
+{
+	/* Clobber the LED if it was blinking */
+	efx->board_info.blink(efx, 0);
+
+	/* Free the context block */
+	kfree(efx->phy_data);
+	efx->phy_data = NULL;
+}
+
+struct efx_phy_operations falcon_xfp_phy_ops = {
+	.init            = xfp_phy_init,
+	.reconfigure     = xfp_phy_reconfigure,
+	.check_hw        = xfp_phy_check_hw,
+	.fini            = xfp_phy_fini,
+	.clear_interrupt = xfp_phy_clear_interrupt,
+	.reset_xaui      = efx_port_dummy_op_void,
+	.mmds            = XFP_REQUIRED_DEVS,
+	.loopbacks       = XFP_LOOPBACKS,
+	/* No loopback appears to be reliable enough for self-test
+	 * operation. So don't do it. */
+	.startup_loopback = LOOPBACK_PCS,
+};
diff -rpuN linux-2.6.18.8/drivers/oprofile/buffer_sync.c linux-2.6.18-xen-3.3.0/drivers/oprofile/buffer_sync.c
--- linux-2.6.18.8/drivers/oprofile/buffer_sync.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/oprofile/buffer_sync.c	2008-08-21 11:36:07.000000000 +0200
@@ -6,6 +6,10 @@
  *
  * @author John Levon <levon@movementarian.org>
  *
+ * Modified by Aravind Menon for Xen
+ * These modifications are:
+ * Copyright (C) 2005 Hewlett-Packard Co.
+ *
  * This is the core of the buffer management. Each
  * CPU buffer is processed and entered into the
  * global event buffer. Such processing is necessary
@@ -38,6 +42,7 @@ static cpumask_t marked_cpus = CPU_MASK_
 static DEFINE_SPINLOCK(task_mortuary);
 static void process_task_mortuary(void);
 
+static int cpu_current_domain[NR_CPUS];
 
 /* Take ownership of the task struct and place it on the
  * list for processing. Only after two full buffer syncs
@@ -146,6 +151,11 @@ static void end_sync(void)
 int sync_start(void)
 {
 	int err;
+	int i;
+
+	for (i = 0; i < NR_CPUS; i++) {
+		cpu_current_domain[i] = COORDINATOR_DOMAIN;
+	}
 
 	start_cpu_work();
 
@@ -275,15 +285,31 @@ static void add_cpu_switch(int i)
 	last_cookie = INVALID_COOKIE;
 }
 
-static void add_kernel_ctx_switch(unsigned int in_kernel)
+static void add_cpu_mode_switch(unsigned int cpu_mode)
 {
 	add_event_entry(ESCAPE_CODE);
-	if (in_kernel)
-		add_event_entry(KERNEL_ENTER_SWITCH_CODE); 
-	else
-		add_event_entry(KERNEL_EXIT_SWITCH_CODE); 
+	switch (cpu_mode) {
+	case CPU_MODE_USER:
+		add_event_entry(USER_ENTER_SWITCH_CODE);
+		break;
+	case CPU_MODE_KERNEL:
+		add_event_entry(KERNEL_ENTER_SWITCH_CODE);
+		break;
+	case CPU_MODE_XEN:
+		add_event_entry(XEN_ENTER_SWITCH_CODE);
+	  	break;
+	default:
+		break;
+	}
 }
- 
+
+static void add_domain_switch(unsigned long domain_id)
+{
+	add_event_entry(ESCAPE_CODE);
+	add_event_entry(DOMAIN_SWITCH_CODE);
+	add_event_entry(domain_id);
+}
+
 static void
 add_user_ctx_switch(struct task_struct const * task, unsigned long cookie)
 {
@@ -348,9 +374,9 @@ static int add_us_sample(struct mm_struc
  * for later lookup from userspace.
  */
 static int
-add_sample(struct mm_struct * mm, struct op_sample * s, int in_kernel)
+add_sample(struct mm_struct * mm, struct op_sample * s, int cpu_mode)
 {
-	if (in_kernel) {
+	if (cpu_mode >= CPU_MODE_KERNEL) {
 		add_sample_entry(s->eip, s->event);
 		return 1;
 	} else if (mm) {
@@ -496,15 +522,21 @@ void sync_buffer(int cpu)
 	struct mm_struct *mm = NULL;
 	struct task_struct * new;
 	unsigned long cookie = 0;
-	int in_kernel = 1;
+	int cpu_mode = 1;
 	unsigned int i;
 	sync_buffer_state state = sb_buffer_start;
 	unsigned long available;
+	int domain_switch = 0;
 
 	mutex_lock(&buffer_mutex);
  
 	add_cpu_switch(cpu);
 
+	/* We need to assign the first samples in this CPU buffer to the
+	   same domain that we were processing at the last sync_buffer */
+	if (cpu_current_domain[cpu] != COORDINATOR_DOMAIN) {
+		add_domain_switch(cpu_current_domain[cpu]);
+	}
 	/* Remember, only we can modify tail_pos */
 
 	available = get_slots(cpu_buf);
@@ -512,16 +544,18 @@ void sync_buffer(int cpu)
 	for (i = 0; i < available; ++i) {
 		struct op_sample * s = &cpu_buf->buffer[cpu_buf->tail_pos];
  
-		if (is_code(s->eip)) {
-			if (s->event <= CPU_IS_KERNEL) {
-				/* kernel/userspace switch */
-				in_kernel = s->event;
+		if (is_code(s->eip) && !domain_switch) {
+			if (s->event <= CPU_MODE_XEN) {
+				/* xen/kernel/userspace switch */
+				cpu_mode = s->event;
 				if (state == sb_buffer_start)
 					state = sb_sample_start;
-				add_kernel_ctx_switch(s->event);
+				add_cpu_mode_switch(s->event);
 			} else if (s->event == CPU_TRACE_BEGIN) {
 				state = sb_bt_start;
 				add_trace_begin();
+			} else if (s->event == CPU_DOMAIN_SWITCH) {
+					domain_switch = 1;				
 			} else {
 				struct mm_struct * oldmm = mm;
 
@@ -535,11 +569,21 @@ void sync_buffer(int cpu)
 				add_user_ctx_switch(new, cookie);
 			}
 		} else {
-			if (state >= sb_bt_start &&
-			    !add_sample(mm, s, in_kernel)) {
-				if (state == sb_bt_start) {
-					state = sb_bt_ignore;
-					atomic_inc(&oprofile_stats.bt_lost_no_mapping);
+			if (domain_switch) {
+				cpu_current_domain[cpu] = s->eip;
+				add_domain_switch(s->eip);
+				domain_switch = 0;
+			} else {
+				if (cpu_current_domain[cpu] !=
+				    COORDINATOR_DOMAIN) {
+					add_sample_entry(s->eip, s->event);
+				}
+				else  if (state >= sb_bt_start &&
+				    !add_sample(mm, s, cpu_mode)) {
+					if (state == sb_bt_start) {
+						state = sb_bt_ignore;
+						atomic_inc(&oprofile_stats.bt_lost_no_mapping);
+					}
 				}
 			}
 		}
@@ -548,6 +592,11 @@ void sync_buffer(int cpu)
 	}
 	release_mm(mm);
 
+	/* We reset domain to COORDINATOR at each CPU switch */
+	if (cpu_current_domain[cpu] != COORDINATOR_DOMAIN) {
+		add_domain_switch(COORDINATOR_DOMAIN);
+	}
+
 	mark_done(cpu);
 
 	mutex_unlock(&buffer_mutex);
diff -rpuN linux-2.6.18.8/drivers/oprofile/cpu_buffer.c linux-2.6.18-xen-3.3.0/drivers/oprofile/cpu_buffer.c
--- linux-2.6.18.8/drivers/oprofile/cpu_buffer.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/oprofile/cpu_buffer.c	2008-08-21 11:36:07.000000000 +0200
@@ -6,6 +6,10 @@
  *
  * @author John Levon <levon@movementarian.org>
  *
+ * Modified by Aravind Menon for Xen
+ * These modifications are:
+ * Copyright (C) 2005 Hewlett-Packard Co.
+ *
  * Each CPU has a local buffer that stores PC value/event
  * pairs. We also log context switches when we notice them.
  * Eventually each CPU's buffer is processed into the global
@@ -34,6 +38,8 @@ static void wq_sync_buffer(void *);
 #define DEFAULT_TIMER_EXPIRE (HZ / 10)
 static int work_enabled;
 
+static int32_t current_domain = COORDINATOR_DOMAIN;
+
 void free_cpu_buffers(void)
 {
 	int i;
@@ -57,7 +63,7 @@ int alloc_cpu_buffers(void)
 			goto fail;
  
 		b->last_task = NULL;
-		b->last_is_kernel = -1;
+		b->last_cpu_mode = -1;
 		b->tracing = 0;
 		b->buffer_size = buffer_size;
 		b->tail_pos = 0;
@@ -113,7 +119,7 @@ void cpu_buffer_reset(struct oprofile_cp
 	 * collected will populate the buffer with proper
 	 * values to initialize the buffer
 	 */
-	cpu_buf->last_is_kernel = -1;
+	cpu_buf->last_cpu_mode = -1;
 	cpu_buf->last_task = NULL;
 }
 
@@ -163,13 +169,13 @@ add_code(struct oprofile_cpu_buffer * bu
  * because of the head/tail separation of the writer and reader
  * of the CPU buffer.
  *
- * is_kernel is needed because on some architectures you cannot
+ * cpu_mode is needed because on some architectures you cannot
  * tell if you are in kernel or user space simply by looking at
- * pc. We tag this in the buffer by generating kernel enter/exit
- * events whenever is_kernel changes
+ * pc. We tag this in the buffer by generating kernel/user (and xen)
+ *  enter events whenever cpu_mode changes
  */
 static int log_sample(struct oprofile_cpu_buffer * cpu_buf, unsigned long pc,
-		      int is_kernel, unsigned long event)
+		      int cpu_mode, unsigned long event)
 {
 	struct task_struct * task;
 
@@ -180,18 +186,18 @@ static int log_sample(struct oprofile_cp
 		return 0;
 	}
 
-	is_kernel = !!is_kernel;
-
 	task = current;
 
 	/* notice a switch from user->kernel or vice versa */
-	if (cpu_buf->last_is_kernel != is_kernel) {
-		cpu_buf->last_is_kernel = is_kernel;
-		add_code(cpu_buf, is_kernel);
+	if (cpu_buf->last_cpu_mode != cpu_mode) {
+		cpu_buf->last_cpu_mode = cpu_mode;
+		add_code(cpu_buf, cpu_mode);
 	}
-
+	
 	/* notice a task switch */
-	if (cpu_buf->last_task != task) {
+	/* if not processing other domain samples */
+	if ((cpu_buf->last_task != task) &&
+	    (current_domain == COORDINATOR_DOMAIN)) {
 		cpu_buf->last_task = task;
 		add_code(cpu_buf, (unsigned long)task);
 	}
@@ -275,6 +281,25 @@ void oprofile_add_trace(unsigned long pc
 	add_sample(cpu_buf, pc, 0);
 }
 
+int oprofile_add_domain_switch(int32_t domain_id)
+{
+	struct oprofile_cpu_buffer * cpu_buf = &cpu_buffer[smp_processor_id()];
+
+	/* should have space for switching into and out of domain 
+	   (2 slots each) plus one sample and one cpu mode switch */
+	if (((nr_available_slots(cpu_buf) < 6) && 
+	     (domain_id != COORDINATOR_DOMAIN)) ||
+	    (nr_available_slots(cpu_buf) < 2))
+		return 0;
+
+	add_code(cpu_buf, CPU_DOMAIN_SWITCH);
+	add_sample(cpu_buf, domain_id, 0);
+
+	current_domain = domain_id;
+
+	return 1;
+}
+
 /*
  * This serves to avoid cpu buffer overflow, and makes sure
  * the task mortuary progresses
diff -rpuN linux-2.6.18.8/drivers/oprofile/cpu_buffer.h linux-2.6.18-xen-3.3.0/drivers/oprofile/cpu_buffer.h
--- linux-2.6.18.8/drivers/oprofile/cpu_buffer.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/oprofile/cpu_buffer.h	2008-08-21 11:36:07.000000000 +0200
@@ -36,7 +36,7 @@ struct oprofile_cpu_buffer {
 	volatile unsigned long tail_pos;
 	unsigned long buffer_size;
 	struct task_struct * last_task;
-	int last_is_kernel;
+	int last_cpu_mode;
 	int tracing;
 	struct op_sample * buffer;
 	unsigned long sample_received;
@@ -51,7 +51,10 @@ extern struct oprofile_cpu_buffer cpu_bu
 void cpu_buffer_reset(struct oprofile_cpu_buffer * cpu_buf);
 
 /* transient events for the CPU buffer -> event buffer */
-#define CPU_IS_KERNEL 1
-#define CPU_TRACE_BEGIN 2
+#define CPU_MODE_USER           0
+#define CPU_MODE_KERNEL         1
+#define CPU_MODE_XEN            2
+#define CPU_TRACE_BEGIN         3
+#define CPU_DOMAIN_SWITCH       4
 
 #endif /* OPROFILE_CPU_BUFFER_H */
diff -rpuN linux-2.6.18.8/drivers/oprofile/event_buffer.h linux-2.6.18-xen-3.3.0/drivers/oprofile/event_buffer.h
--- linux-2.6.18.8/drivers/oprofile/event_buffer.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/oprofile/event_buffer.h	2008-08-21 11:36:07.000000000 +0200
@@ -29,15 +29,20 @@ void wake_up_buffer_waiter(void);
 #define CPU_SWITCH_CODE 		2
 #define COOKIE_SWITCH_CODE 		3
 #define KERNEL_ENTER_SWITCH_CODE	4
-#define KERNEL_EXIT_SWITCH_CODE		5
+#define USER_ENTER_SWITCH_CODE		5
 #define MODULE_LOADED_CODE		6
 #define CTX_TGID_CODE			7
 #define TRACE_BEGIN_CODE		8
 #define TRACE_END_CODE			9
+#define XEN_ENTER_SWITCH_CODE		10
+#define DOMAIN_SWITCH_CODE		11
  
 #define INVALID_COOKIE ~0UL
 #define NO_COOKIE 0UL
 
+/* Constant used to refer to coordinator domain (Xen) */
+#define COORDINATOR_DOMAIN -1
+
 /* add data to the event buffer */
 void add_event_entry(unsigned long data);
  
diff -rpuN linux-2.6.18.8/drivers/oprofile/oprof.c linux-2.6.18-xen-3.3.0/drivers/oprofile/oprof.c
--- linux-2.6.18.8/drivers/oprofile/oprof.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/oprofile/oprof.c	2008-08-21 11:36:07.000000000 +0200
@@ -5,6 +5,10 @@
  * @remark Read the file COPYING
  *
  * @author John Levon <levon@movementarian.org>
+ *
+ * Modified by Aravind Menon for Xen
+ * These modifications are:
+ * Copyright (C) 2005 Hewlett-Packard Co.
  */
 
 #include <linux/kernel.h>
@@ -19,7 +23,7 @@
 #include "cpu_buffer.h"
 #include "buffer_sync.h"
 #include "oprofile_stats.h"
- 
+
 struct oprofile_operations oprofile_ops;
 
 unsigned long oprofile_started;
@@ -33,6 +37,32 @@ static DEFINE_MUTEX(start_mutex);
  */
 static int timer = 0;
 
+int oprofile_set_active(int active_domains[], unsigned int adomains)
+{
+	int err;
+
+	if (!oprofile_ops.set_active)
+		return -EINVAL;
+
+	mutex_lock(&start_mutex);
+	err = oprofile_ops.set_active(active_domains, adomains);
+	mutex_unlock(&start_mutex);
+	return err;
+}
+
+int oprofile_set_passive(int passive_domains[], unsigned int pdomains)
+{
+	int err;
+
+	if (!oprofile_ops.set_passive)
+		return -EINVAL;
+
+	mutex_lock(&start_mutex);
+	err = oprofile_ops.set_passive(passive_domains, pdomains);
+	mutex_unlock(&start_mutex);
+	return err;
+}
+
 int oprofile_setup(void)
 {
 	int err;
diff -rpuN linux-2.6.18.8/drivers/oprofile/oprof.h linux-2.6.18-xen-3.3.0/drivers/oprofile/oprof.h
--- linux-2.6.18.8/drivers/oprofile/oprof.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/oprofile/oprof.h	2008-08-21 11:36:07.000000000 +0200
@@ -35,5 +35,8 @@ void oprofile_create_files(struct super_
 void oprofile_timer_init(struct oprofile_operations * ops);
 
 int oprofile_set_backtrace(unsigned long depth);
+
+int oprofile_set_active(int active_domains[], unsigned int adomains);
+int oprofile_set_passive(int passive_domains[], unsigned int pdomains);
  
 #endif /* OPROF_H */
diff -rpuN linux-2.6.18.8/drivers/oprofile/oprofile_files.c linux-2.6.18-xen-3.3.0/drivers/oprofile/oprofile_files.c
--- linux-2.6.18.8/drivers/oprofile/oprofile_files.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/oprofile/oprofile_files.c	2008-08-21 11:36:07.000000000 +0200
@@ -5,15 +5,21 @@
  * @remark Read the file COPYING
  *
  * @author John Levon <levon@movementarian.org>
+ *
+ * Modified by Aravind Menon for Xen
+ * These modifications are:
+ * Copyright (C) 2005 Hewlett-Packard Co.	
  */
 
 #include <linux/fs.h>
 #include <linux/oprofile.h>
+#include <asm/uaccess.h>
+#include <linux/ctype.h>
 
 #include "event_buffer.h"
 #include "oprofile_stats.h"
 #include "oprof.h"
- 
+
 unsigned long fs_buffer_size = 131072;
 unsigned long fs_cpu_buffer_size = 8192;
 unsigned long fs_buffer_watershed = 32768; /* FIXME: tune */
@@ -117,11 +123,202 @@ static ssize_t dump_write(struct file * 
 static struct file_operations dump_fops = {
 	.write		= dump_write,
 };
- 
+
+#define TMPBUFSIZE 512
+
+static unsigned int adomains = 0;
+static int active_domains[MAX_OPROF_DOMAINS + 1];
+static DEFINE_MUTEX(adom_mutex);
+
+static ssize_t adomain_write(struct file * file, char const __user * buf, 
+			     size_t count, loff_t * offset)
+{
+	char *tmpbuf;
+	char *startp, *endp;
+	int i;
+	unsigned long val;
+	ssize_t retval = count;
+	
+	if (*offset)
+		return -EINVAL;	
+	if (count > TMPBUFSIZE - 1)
+		return -EINVAL;
+
+	if (!(tmpbuf = kmalloc(TMPBUFSIZE, GFP_KERNEL)))
+		return -ENOMEM;
+
+	if (copy_from_user(tmpbuf, buf, count)) {
+		kfree(tmpbuf);
+		return -EFAULT;
+	}
+	tmpbuf[count] = 0;
+
+	mutex_lock(&adom_mutex);
+
+	startp = tmpbuf;
+	/* Parse one more than MAX_OPROF_DOMAINS, for easy error checking */
+	for (i = 0; i <= MAX_OPROF_DOMAINS; i++) {
+		val = simple_strtoul(startp, &endp, 0);
+		if (endp == startp)
+			break;
+		while (ispunct(*endp) || isspace(*endp))
+			endp++;
+		active_domains[i] = val;
+		if (active_domains[i] != val)
+			/* Overflow, force error below */
+			i = MAX_OPROF_DOMAINS + 1;
+		startp = endp;
+	}
+	/* Force error on trailing junk */
+	adomains = *startp ? MAX_OPROF_DOMAINS + 1 : i;
+
+	kfree(tmpbuf);
+
+	if (adomains > MAX_OPROF_DOMAINS
+	    || oprofile_set_active(active_domains, adomains)) {
+		adomains = 0;
+		retval = -EINVAL;
+	}
+
+	mutex_unlock(&adom_mutex);
+	return retval;
+}
+
+static ssize_t adomain_read(struct file * file, char __user * buf, 
+			    size_t count, loff_t * offset)
+{
+	char * tmpbuf;
+	size_t len;
+	int i;
+	ssize_t retval;
+
+	if (!(tmpbuf = kmalloc(TMPBUFSIZE, GFP_KERNEL)))
+		return -ENOMEM;
+
+	mutex_lock(&adom_mutex);
+
+	len = 0;
+	for (i = 0; i < adomains; i++)
+		len += snprintf(tmpbuf + len,
+				len < TMPBUFSIZE ? TMPBUFSIZE - len : 0,
+				"%u ", active_domains[i]);
+	WARN_ON(len > TMPBUFSIZE);
+	if (len != 0 && len <= TMPBUFSIZE)
+		tmpbuf[len-1] = '\n';
+
+	mutex_unlock(&adom_mutex);
+
+	retval = simple_read_from_buffer(buf, count, offset, tmpbuf, len);
+
+	kfree(tmpbuf);
+	return retval;
+}
+
+
+static struct file_operations active_domain_ops = {
+	.read		= adomain_read,
+	.write		= adomain_write,
+};
+
+static unsigned int pdomains = 0;
+static int passive_domains[MAX_OPROF_DOMAINS];
+static DEFINE_MUTEX(pdom_mutex);
+
+static ssize_t pdomain_write(struct file * file, char const __user * buf, 
+			     size_t count, loff_t * offset)
+{
+	char *tmpbuf;
+	char *startp, *endp;
+	int i;
+	unsigned long val;
+	ssize_t retval = count;
+	
+	if (*offset)
+		return -EINVAL;	
+	if (count > TMPBUFSIZE - 1)
+		return -EINVAL;
+
+	if (!(tmpbuf = kmalloc(TMPBUFSIZE, GFP_KERNEL)))
+		return -ENOMEM;
+
+	if (copy_from_user(tmpbuf, buf, count)) {
+		kfree(tmpbuf);
+		return -EFAULT;
+	}
+	tmpbuf[count] = 0;
+
+	mutex_lock(&pdom_mutex);
+
+	startp = tmpbuf;
+	/* Parse one more than MAX_OPROF_DOMAINS, for easy error checking */
+	for (i = 0; i <= MAX_OPROF_DOMAINS; i++) {
+		val = simple_strtoul(startp, &endp, 0);
+		if (endp == startp)
+			break;
+		while (ispunct(*endp) || isspace(*endp))
+			endp++;
+		passive_domains[i] = val;
+		if (passive_domains[i] != val)
+			/* Overflow, force error below */
+			i = MAX_OPROF_DOMAINS + 1;
+		startp = endp;
+	}
+	/* Force error on trailing junk */
+	pdomains = *startp ? MAX_OPROF_DOMAINS + 1 : i;
+
+	kfree(tmpbuf);
+
+	if (pdomains > MAX_OPROF_DOMAINS
+	    || oprofile_set_passive(passive_domains, pdomains)) {
+		pdomains = 0;
+		retval = -EINVAL;
+	}
+
+	mutex_unlock(&pdom_mutex);
+	return retval;
+}
+
+static ssize_t pdomain_read(struct file * file, char __user * buf, 
+			    size_t count, loff_t * offset)
+{
+	char * tmpbuf;
+	size_t len;
+	int i;
+	ssize_t retval;
+
+	if (!(tmpbuf = kmalloc(TMPBUFSIZE, GFP_KERNEL)))
+		return -ENOMEM;
+
+	mutex_lock(&pdom_mutex);
+
+	len = 0;
+	for (i = 0; i < pdomains; i++)
+		len += snprintf(tmpbuf + len,
+				len < TMPBUFSIZE ? TMPBUFSIZE - len : 0,
+				"%u ", passive_domains[i]);
+	WARN_ON(len > TMPBUFSIZE);
+	if (len != 0 && len <= TMPBUFSIZE)
+		tmpbuf[len-1] = '\n';
+
+	mutex_unlock(&pdom_mutex);
+
+	retval = simple_read_from_buffer(buf, count, offset, tmpbuf, len);
+
+	kfree(tmpbuf);
+	return retval;
+}
+
+static struct file_operations passive_domain_ops = {
+	.read		= pdomain_read,
+	.write		= pdomain_write,
+};
+
 void oprofile_create_files(struct super_block * sb, struct dentry * root)
 {
 	oprofilefs_create_file(sb, root, "enable", &enable_fops);
 	oprofilefs_create_file_perm(sb, root, "dump", &dump_fops, 0666);
+	oprofilefs_create_file(sb, root, "active_domains", &active_domain_ops);
+	oprofilefs_create_file(sb, root, "passive_domains", &passive_domain_ops);
 	oprofilefs_create_file(sb, root, "buffer", &event_buffer_fops);
 	oprofilefs_create_ulong(sb, root, "buffer_size", &fs_buffer_size);
 	oprofilefs_create_ulong(sb, root, "buffer_watershed", &fs_buffer_watershed);
diff -rpuN linux-2.6.18.8/drivers/pci/bus.c linux-2.6.18-xen-3.3.0/drivers/pci/bus.c
--- linux-2.6.18.8/drivers/pci/bus.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/pci/bus.c	2008-08-21 11:36:07.000000000 +0200
@@ -17,6 +17,8 @@
 
 #include "pci.h"
 
+extern int pci_mem_align;
+
 /**
  * pci_bus_alloc_resource - allocate a resource from a parent bus
  * @bus: PCI bus
@@ -44,6 +46,11 @@ pci_bus_alloc_resource(struct pci_bus *b
 
 	type_mask |= IORESOURCE_IO | IORESOURCE_MEM;
 
+	/* If the boot parameter 'pci-mem-align' was specified then we need to 
+	   align the memory addresses, at page size alignment. */
+	if (pci_mem_align && (align < (PAGE_SIZE-1)))
+		align = PAGE_SIZE - 1;
+
 	for (i = 0; i < PCI_BUS_NUM_RESOURCES; i++) {
 		struct resource *r = bus->resource[i];
 		if (!r)
diff -rpuN linux-2.6.18.8/drivers/pci/msi-xen.c linux-2.6.18-xen-3.3.0/drivers/pci/msi-xen.c
--- linux-2.6.18.8/drivers/pci/msi-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/pci/msi-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,746 @@
+/*
+ * File:	msi.c
+ * Purpose:	PCI Message Signaled Interrupt (MSI)
+ *
+ * Copyright (C) 2003-2004 Intel
+ * Copyright (C) Tom Long Nguyen (tom.l.nguyen@intel.com)
+ */
+
+#include <linux/mm.h>
+#include <linux/irq.h>
+#include <linux/interrupt.h>
+#include <linux/init.h>
+#include <linux/ioport.h>
+#include <linux/smp_lock.h>
+#include <linux/pci.h>
+#include <linux/proc_fs.h>
+
+#include <asm/errno.h>
+#include <asm/io.h>
+#include <asm/smp.h>
+
+#include "pci.h"
+#include "msi.h"
+
+static int pci_msi_enable = 1;
+
+static struct msi_ops *msi_ops;
+
+int msi_register(struct msi_ops *ops)
+{
+	msi_ops = ops;
+	return 0;
+}
+
+static LIST_HEAD(msi_dev_head);
+DEFINE_SPINLOCK(msi_dev_lock);
+
+struct msi_dev_list {
+	struct pci_dev *dev;
+	struct list_head list;
+	spinlock_t pirq_list_lock;
+	struct list_head pirq_list_head;
+};
+
+struct msi_pirq_entry {
+	struct list_head list;
+	int pirq;
+	int entry_nr;
+};
+
+static struct msi_dev_list *get_msi_dev_pirq_list(struct pci_dev *dev)
+{
+	struct msi_dev_list *msi_dev_list, *ret = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&msi_dev_lock, flags);
+
+	list_for_each_entry(msi_dev_list, &msi_dev_head, list)
+		if ( msi_dev_list->dev == dev )
+			ret = msi_dev_list;
+
+	if ( ret ) {
+		spin_unlock_irqrestore(&msi_dev_lock, flags);
+		return ret;
+	}
+
+	/* Has not allocate msi_dev until now. */
+	ret = kmalloc(sizeof(struct msi_dev_list), GFP_ATOMIC);
+
+	/* Failed to allocate msi_dev structure */
+	if ( !ret ) {
+		spin_unlock_irqrestore(&msi_dev_lock, flags);
+		return NULL;
+	}
+
+	spin_lock_init(&ret->pirq_list_lock);
+	INIT_LIST_HEAD(&ret->pirq_list_head);
+	list_add_tail(&ret->list, &msi_dev_head);
+	spin_unlock_irqrestore(&msi_dev_lock, flags);
+	return ret;
+}
+
+static int attach_pirq_entry(int pirq, int entry_nr,
+                             struct msi_dev_list *msi_dev_entry)
+{
+	struct msi_pirq_entry *entry = kmalloc(sizeof(*entry), GFP_ATOMIC);
+	unsigned long flags;
+
+	if (!entry)
+		return -ENOMEM;
+	entry->pirq = pirq;
+	entry->entry_nr = entry_nr;
+	spin_lock_irqsave(&msi_dev_entry->pirq_list_lock, flags);
+	list_add_tail(&entry->list, &msi_dev_entry->pirq_list_head);
+	spin_unlock_irqrestore(&msi_dev_entry->pirq_list_lock, flags);
+	return 0;
+}
+
+static void detach_pirq_entry(int entry_nr,
+ 							struct msi_dev_list *msi_dev_entry)
+{
+	unsigned long flags;
+	struct msi_pirq_entry *pirq_entry;
+
+	list_for_each_entry(pirq_entry, &msi_dev_entry->pirq_list_head, list) {
+		if (pirq_entry->entry_nr == entry_nr) {
+			spin_lock_irqsave(&msi_dev_entry->pirq_list_lock, flags);
+			list_del(&pirq_entry->list);
+			spin_unlock_irqrestore(&msi_dev_entry->pirq_list_lock, flags);
+			kfree(pirq_entry);
+			return;
+		}
+	}
+}
+
+/*
+ * pciback will provide device's owner
+ */
+static int (*get_owner)(struct pci_dev *dev);
+
+int register_msi_get_owner(int (*func)(struct pci_dev *dev))
+{
+	if (get_owner) {
+		printk(KERN_WARNING "register msi_get_owner again\n");
+		return -EEXIST;
+	}
+	get_owner = func;
+	return 0;
+}
+
+int unregister_msi_get_owner(int (*func)(struct pci_dev *dev))
+{
+	if (get_owner != func)
+		return -EINVAL;
+	get_owner = NULL;
+	return 0;
+}
+
+static int msi_get_dev_owner(struct pci_dev *dev)
+{
+	int owner;
+
+	BUG_ON(!is_initial_xendomain());
+	if (get_owner && (owner = get_owner(dev)) >= 0) {
+		printk(KERN_INFO "get owner for dev %x get %x \n",
+		       dev->devfn, owner);
+		return owner;
+	}
+
+	return DOMID_SELF;
+}
+
+static int msi_unmap_pirq(struct pci_dev *dev, int pirq)
+{
+	struct physdev_unmap_pirq unmap;
+	int rc;
+
+	unmap.domid = msi_get_dev_owner(dev);
+	unmap.pirq = pirq;
+
+	if ((rc = HYPERVISOR_physdev_op(PHYSDEVOP_unmap_pirq, &unmap)))
+		printk(KERN_WARNING "unmap irq %x failed\n", pirq);
+
+	if (rc < 0)
+		return rc;
+	return 0;
+}
+
+static u64 find_table_base(struct pci_dev *dev, int pos)
+{
+	u8 bar;
+	u32 reg;
+	unsigned long flags;
+
+ 	pci_read_config_dword(dev, msix_table_offset_reg(pos), &reg);
+	bar = reg & PCI_MSIX_FLAGS_BIRMASK;
+
+	flags = pci_resource_flags(dev, bar);
+	if (flags & (IORESOURCE_DISABLED | IORESOURCE_UNSET | IORESOURCE_BUSY))
+		return 0;
+
+	return pci_resource_start(dev, bar);
+}
+
+/*
+ * Protected by msi_lock
+ */
+static int msi_map_pirq_to_vector(struct pci_dev *dev, int pirq,
+				  int entry_nr, u64 table_base)
+{
+	struct physdev_map_pirq map_irq;
+	int rc;
+	domid_t domid = DOMID_SELF;
+
+	domid = msi_get_dev_owner(dev);
+
+	map_irq.domid = domid;
+	map_irq.type = MAP_PIRQ_TYPE_MSI;
+	map_irq.index = -1;
+	map_irq.pirq = pirq;
+	map_irq.bus = dev->bus->number;
+	map_irq.devfn = dev->devfn;
+	map_irq.entry_nr = entry_nr;
+	map_irq.table_base = table_base;
+
+	if ((rc = HYPERVISOR_physdev_op(PHYSDEVOP_map_pirq, &map_irq)))
+		printk(KERN_WARNING "map irq failed\n");
+
+	if (rc < 0)
+		return rc;
+
+	return map_irq.pirq;
+}
+
+static int msi_map_vector(struct pci_dev *dev, int entry_nr, u64 table_base)
+{
+	return msi_map_pirq_to_vector(dev, -1, entry_nr, table_base);
+}
+
+static int msi_init(void)
+{
+	static int status = 0;
+
+	if (pci_msi_quirk) {
+		pci_msi_enable = 0;
+		printk(KERN_WARNING "PCI: MSI quirk detected. MSI disabled.\n");
+		status = -EINVAL;
+	}
+
+	return status;
+}
+
+void pci_scan_msi_device(struct pci_dev *dev) { }
+
+void disable_msi_mode(struct pci_dev *dev, int pos, int type)
+{
+	u16 control;
+
+	pci_read_config_word(dev, msi_control_reg(pos), &control);
+	if (type == PCI_CAP_ID_MSI) {
+		/* Set enabled bits to single MSI & enable MSI_enable bit */
+		msi_disable(control);
+		pci_write_config_word(dev, msi_control_reg(pos), control);
+		dev->msi_enabled = 0;
+	} else {
+		msix_disable(control);
+		pci_write_config_word(dev, msi_control_reg(pos), control);
+		dev->msix_enabled = 0;
+	}
+    	if (pci_find_capability(dev, PCI_CAP_ID_EXP)) {
+		/* PCI Express Endpoint device detected */
+		pci_intx(dev, 1);  /* enable intx */
+	}
+}
+
+static void enable_msi_mode(struct pci_dev *dev, int pos, int type)
+{
+	u16 control;
+
+	pci_read_config_word(dev, msi_control_reg(pos), &control);
+	if (type == PCI_CAP_ID_MSI) {
+		/* Set enabled bits to single MSI & enable MSI_enable bit */
+		msi_enable(control, 1);
+		pci_write_config_word(dev, msi_control_reg(pos), control);
+		dev->msi_enabled = 1;
+	} else {
+		msix_enable(control);
+		pci_write_config_word(dev, msi_control_reg(pos), control);
+		dev->msix_enabled = 1;
+	}
+    	if (pci_find_capability(dev, PCI_CAP_ID_EXP)) {
+		/* PCI Express Endpoint device detected */
+		pci_intx(dev, 0);  /* disable intx */
+	}
+}
+
+#ifdef CONFIG_PM
+int pci_save_msi_state(struct pci_dev *dev)
+{
+	int pos;
+
+	pos = pci_find_capability(dev, PCI_CAP_ID_MSI);
+	if (pos <= 0 || dev->no_msi)
+		return 0;
+
+	if (!dev->msi_enabled)
+		return 0;
+
+	/* Restore dev->irq to its default pin-assertion vector */
+	msi_unmap_pirq(dev, dev->irq);
+	/* Disable MSI mode */
+	disable_msi_mode(dev, pos, PCI_CAP_ID_MSI);
+	/* Set the flags for use of restore */
+	dev->msi_enabled = 1;
+	return 0;
+}
+
+void pci_restore_msi_state(struct pci_dev *dev)
+{
+	int pos, pirq;
+
+	pos = pci_find_capability(dev, PCI_CAP_ID_MSI);
+	if (pos <= 0)
+		return;
+
+	if (!dev->msi_enabled)
+		return;
+
+	pirq = msi_map_pirq_to_vector(dev, dev->irq, 0, 0);
+	if (pirq < 0)
+		return;
+	enable_msi_mode(dev, pos, PCI_CAP_ID_MSI);
+}
+
+int pci_save_msix_state(struct pci_dev *dev)
+{
+	int pos;
+	unsigned long flags;
+	struct msi_dev_list *msi_dev_entry;
+	struct msi_pirq_entry *pirq_entry, *tmp;
+
+	pos = pci_find_capability(dev, PCI_CAP_ID_MSIX);
+	if (pos <= 0 || dev->no_msi)
+		return 0;
+
+	/* save the capability */
+	if (!dev->msix_enabled)
+		return 0;
+
+	msi_dev_entry = get_msi_dev_pirq_list(dev);
+
+	spin_lock_irqsave(&msi_dev_entry->pirq_list_lock, flags);
+        list_for_each_entry_safe(pirq_entry, tmp,
+                                 &msi_dev_entry->pirq_list_head, list)
+		msi_unmap_pirq(dev, pirq_entry->pirq);
+	spin_unlock_irqrestore(&msi_dev_entry->pirq_list_lock, flags);
+
+	disable_msi_mode(dev, pos, PCI_CAP_ID_MSIX);
+	/* Set the flags for use of restore */
+	dev->msix_enabled = 1;
+
+	return 0;
+}
+
+void pci_restore_msix_state(struct pci_dev *dev)
+{
+	int pos;
+	unsigned long flags;
+	u64 table_base;
+	struct msi_dev_list *msi_dev_entry;
+	struct msi_pirq_entry *pirq_entry, *tmp;
+
+	pos = pci_find_capability(dev, PCI_CAP_ID_MSIX);
+	if (pos <= 0)
+		return;
+
+	if (!dev->msix_enabled)
+		return;
+
+	msi_dev_entry = get_msi_dev_pirq_list(dev);
+	table_base = find_table_base(dev, pos);
+	if (!table_base)
+		return;
+
+	spin_lock_irqsave(&msi_dev_entry->pirq_list_lock, flags);
+	list_for_each_entry_safe(pirq_entry, tmp,
+				 &msi_dev_entry->pirq_list_head, list)
+		msi_map_pirq_to_vector(dev, pirq_entry->pirq,
+				       pirq_entry->entry_nr, table_base);
+	spin_unlock_irqrestore(&msi_dev_entry->pirq_list_lock, flags);
+
+	enable_msi_mode(dev, pos, PCI_CAP_ID_MSIX);
+}
+#endif
+
+/**
+ * msi_capability_init - configure device's MSI capability structure
+ * @dev: pointer to the pci_dev data structure of MSI device function
+ *
+ * Setup the MSI capability structure of device function with a single
+ * MSI vector, regardless of device function is capable of handling
+ * multiple messages. A return of zero indicates the successful setup
+ * of an entry zero with the new MSI vector or non-zero for otherwise.
+ **/
+static int msi_capability_init(struct pci_dev *dev)
+{
+	int pos, pirq;
+	u16 control;
+
+	pos = pci_find_capability(dev, PCI_CAP_ID_MSI);
+	pci_read_config_word(dev, msi_control_reg(pos), &control);
+
+	pirq = msi_map_vector(dev, 0, 0);
+	if (pirq < 0)
+		return -EBUSY;
+
+	dev->irq = pirq;
+	/* Set MSI enabled bits	 */
+	enable_msi_mode(dev, pos, PCI_CAP_ID_MSI);
+	dev->msi_enabled = 1;
+
+	return 0;
+}
+
+/**
+ * msix_capability_init - configure device's MSI-X capability
+ * @dev: pointer to the pci_dev data structure of MSI-X device function
+ * @entries: pointer to an array of struct msix_entry entries
+ * @nvec: number of @entries
+ *
+ * Setup the MSI-X capability structure of device function with a
+ * single MSI-X vector. A return of zero indicates the successful setup of
+ * requested MSI-X entries with allocated vectors or non-zero for otherwise.
+ **/
+static int msix_capability_init(struct pci_dev *dev,
+				struct msix_entry *entries, int nvec)
+{
+	u64 table_base;
+	int pirq, i, j, mapped, pos;
+	struct msi_dev_list *msi_dev_entry = get_msi_dev_pirq_list(dev);
+	struct msi_pirq_entry *pirq_entry;
+
+	if (!msi_dev_entry)
+		return -ENOMEM;
+
+	pos = pci_find_capability(dev, PCI_CAP_ID_MSIX);
+	table_base = find_table_base(dev, pos);
+	if (!table_base)
+		return -ENODEV;
+
+	/* MSI-X Table Initialization */
+	for (i = 0; i < nvec; i++) {
+		mapped = 0;
+		list_for_each_entry(pirq_entry, &msi_dev_entry->pirq_list_head, list) {
+			if (pirq_entry->entry_nr == entries[i].entry) {
+				printk(KERN_WARNING "msix entry %d for dev %02x:%02x:%01x are \
+				       not freed before acquire again.\n", entries[i].entry,
+					   dev->bus->number, PCI_SLOT(dev->devfn),
+					   PCI_FUNC(dev->devfn));
+				(entries + i)->vector = pirq_entry->pirq;
+				mapped = 1;
+				break;
+			}
+		}
+		if (mapped)
+			continue;
+		pirq = msi_map_vector(dev, entries[i].entry, table_base);
+		if (pirq < 0)
+			break;
+		attach_pirq_entry(pirq, entries[i].entry, msi_dev_entry);
+		(entries + i)->vector = pirq;
+	}
+
+	if (i != nvec) {
+		for (j = --i; j >= 0; j--) {
+			msi_unmap_pirq(dev, entries[j].vector);
+			detach_pirq_entry(entries[j].entry, msi_dev_entry);
+			entries[j].vector = 0;
+		}
+		return -EBUSY;
+	}
+
+	enable_msi_mode(dev, pos, PCI_CAP_ID_MSIX);
+	dev->msix_enabled = 1;
+
+	return 0;
+}
+
+/**
+ * pci_enable_msi - configure device's MSI capability structure
+ * @dev: pointer to the pci_dev data structure of MSI device function
+ *
+ * Setup the MSI capability structure of device function with
+ * a single MSI vector upon its software driver call to request for
+ * MSI mode enabled on its hardware device function. A return of zero
+ * indicates the successful setup of an entry zero with the new MSI
+ * vector or non-zero for otherwise.
+ **/
+extern int pci_frontend_enable_msi(struct pci_dev *dev);
+int pci_enable_msi(struct pci_dev* dev)
+{
+	struct pci_bus *bus;
+	int pos, temp, status = -EINVAL;
+
+	if (!pci_msi_enable || !dev)
+ 		return status;
+
+	if (dev->no_msi)
+		return status;
+
+	for (bus = dev->bus; bus; bus = bus->parent)
+		if (bus->bus_flags & PCI_BUS_FLAGS_NO_MSI)
+			return -EINVAL;
+
+	status = msi_init();
+	if (status < 0)
+		return status;
+
+#ifdef CONFIG_XEN_PCIDEV_FRONTEND
+	if (!is_initial_xendomain())
+	{
+		int ret;
+
+		temp = dev->irq;
+		ret = pci_frontend_enable_msi(dev);
+		if (ret)
+			return ret;
+
+		dev->irq_old = temp;
+
+		return ret;
+	}
+#endif
+
+	temp = dev->irq;
+
+	pos = pci_find_capability(dev, PCI_CAP_ID_MSI);
+	if (!pos)
+		return -EINVAL;
+
+	/* Check whether driver already requested for MSI-X vectors */
+	if (dev->msix_enabled) {
+		printk(KERN_INFO "PCI: %s: Can't enable MSI.  "
+			   "Device already has MSI-X vectors assigned\n",
+			   pci_name(dev));
+		dev->irq = temp;
+		return -EINVAL;
+	}
+
+	status = msi_capability_init(dev);
+	if ( !status )
+		dev->irq_old = temp;
+    else
+		dev->irq = temp;
+
+	return status;
+}
+
+extern void pci_frontend_disable_msi(struct pci_dev* dev);
+void pci_disable_msi(struct pci_dev* dev)
+{
+	int pos;
+	int pirq;
+
+	if (!pci_msi_enable)
+		return;
+	if (!dev)
+		return;
+
+#ifdef CONFIG_XEN_PCIDEV_FRONTEND
+	if (!is_initial_xendomain()) {
+		pci_frontend_disable_msi(dev);
+		dev->irq = dev->irq_old;
+		return;
+	}
+#endif
+
+	pos = pci_find_capability(dev, PCI_CAP_ID_MSI);
+	if (!pos)
+		return;
+
+	pirq = dev->irq;
+	/* Restore dev->irq to its default pin-assertion vector */
+	dev->irq = dev->irq_old;
+	msi_unmap_pirq(dev, pirq);
+
+	/* Disable MSI mode */
+	disable_msi_mode(dev, pos, PCI_CAP_ID_MSI);
+}
+
+/**
+ * pci_enable_msix - configure device's MSI-X capability structure
+ * @dev: pointer to the pci_dev data structure of MSI-X device function
+ * @entries: pointer to an array of MSI-X entries
+ * @nvec: number of MSI-X vectors requested for allocation by device driver
+ *
+ * Setup the MSI-X capability structure of device function with the number
+ * of requested vectors upon its software driver call to request for
+ * MSI-X mode enabled on its hardware device function. A return of zero
+ * indicates the successful configuration of MSI-X capability structure
+ * with new allocated MSI-X vectors. A return of < 0 indicates a failure.
+ * Or a return of > 0 indicates that driver request is exceeding the number
+ * of vectors available. Driver should use the returned value to re-send
+ * its request.
+ **/
+extern int pci_frontend_enable_msix(struct pci_dev *dev,
+		struct msix_entry *entries, int nvec);
+int pci_enable_msix(struct pci_dev* dev, struct msix_entry *entries, int nvec)
+{
+	struct pci_bus *bus;
+	int status, pos, nr_entries;
+	int i, j, temp;
+	u16 control;
+
+	if (!pci_msi_enable || !dev || !entries)
+ 		return -EINVAL;
+
+	if (dev->no_msi)
+		return -EINVAL;
+
+	for (bus = dev->bus; bus; bus = bus->parent)
+		if (bus->bus_flags & PCI_BUS_FLAGS_NO_MSI)
+			return -EINVAL;
+
+#ifdef CONFIG_XEN_PCIDEV_FRONTEND
+	if (!is_initial_xendomain()) {
+		int ret;
+
+		ret = pci_frontend_enable_msix(dev, entries, nvec);
+		if (ret) {
+			printk("get %x from pci_frontend_enable_msix\n", ret);
+			return ret;
+		}
+
+        return 0;
+	}
+#endif
+
+	status = msi_init();
+	if (status < 0)
+		return status;
+
+	pos = pci_find_capability(dev, PCI_CAP_ID_MSIX);
+	if (!pos)
+ 		return -EINVAL;
+
+	pci_read_config_word(dev, msi_control_reg(pos), &control);
+	nr_entries = multi_msix_capable(control);
+	if (nvec > nr_entries)
+		return -EINVAL;
+
+	/* Check for any invalid entries */
+	for (i = 0; i < nvec; i++) {
+		if (entries[i].entry >= nr_entries)
+			return -EINVAL;		/* invalid entry */
+		for (j = i + 1; j < nvec; j++) {
+			if (entries[i].entry == entries[j].entry)
+				return -EINVAL;	/* duplicate entry */
+		}
+	}
+
+	temp = dev->irq;
+	/* Check whether driver already requested for MSI vector */
+	if (dev->msi_enabled) {
+		printk(KERN_INFO "PCI: %s: Can't enable MSI-X.  "
+		       "Device already has an MSI vector assigned\n",
+		       pci_name(dev));
+		dev->irq = temp;
+		return -EINVAL;
+	}
+
+	status = msix_capability_init(dev, entries, nvec);
+
+	if ( !status )
+		dev->irq_old = temp;
+	else
+		dev->irq = temp;
+
+	return status;
+}
+
+extern void pci_frontend_disable_msix(struct pci_dev* dev);
+void pci_disable_msix(struct pci_dev* dev)
+{
+	int pos;
+	u16 control;
+
+
+	if (!pci_msi_enable)
+		return;
+	if (!dev)
+		return;
+
+#ifdef CONFIG_XEN_PCIDEV_FRONTEND
+	if (!is_initial_xendomain()) {
+		pci_frontend_disable_msix(dev);
+		dev->irq = dev->irq_old;
+		return;
+	}
+#endif
+
+	pos = pci_find_capability(dev, PCI_CAP_ID_MSIX);
+	if (!pos)
+		return;
+
+	pci_read_config_word(dev, msi_control_reg(pos), &control);
+	if (!(control & PCI_MSIX_FLAGS_ENABLE))
+		return;
+
+	msi_remove_pci_irq_vectors(dev);
+
+	/* Disable MSI mode */
+	disable_msi_mode(dev, pos, PCI_CAP_ID_MSIX);
+}
+
+/**
+ * msi_remove_pci_irq_vectors - reclaim MSI(X) vectors to unused state
+ * @dev: pointer to the pci_dev data structure of MSI(X) device function
+ *
+ * Being called during hotplug remove, from which the device function
+ * is hot-removed. All previous assigned MSI/MSI-X vectors, if
+ * allocated for this device function, are reclaimed to unused state,
+ * which may be used later on.
+ **/
+void msi_remove_pci_irq_vectors(struct pci_dev* dev)
+{
+	unsigned long flags;
+	struct msi_dev_list *msi_dev_entry;
+	struct msi_pirq_entry *pirq_entry, *tmp;
+
+	if (!pci_msi_enable || !dev)
+ 		return;
+
+	msi_dev_entry = get_msi_dev_pirq_list(dev);
+
+	spin_lock_irqsave(&msi_dev_entry->pirq_list_lock, flags);
+	if (!list_empty(&msi_dev_entry->pirq_list_head))
+	{
+		printk(KERN_WARNING "msix pirqs for dev %02x:%02x:%01x are not freed \
+		       before acquire again.\n", dev->bus->number, PCI_SLOT(dev->devfn),
+			   PCI_FUNC(dev->devfn));
+		list_for_each_entry_safe(pirq_entry, tmp,
+		                         &msi_dev_entry->pirq_list_head, list) {
+			msi_unmap_pirq(dev, pirq_entry->pirq);
+			list_del(&pirq_entry->list);
+			kfree(pirq_entry);
+		}
+	}
+	spin_unlock_irqrestore(&msi_dev_entry->pirq_list_lock, flags);
+	dev->irq = dev->irq_old;
+}
+
+void pci_no_msi(void)
+{
+	pci_msi_enable = 0;
+}
+
+EXPORT_SYMBOL(pci_enable_msi);
+EXPORT_SYMBOL(pci_disable_msi);
+EXPORT_SYMBOL(pci_enable_msix);
+EXPORT_SYMBOL(pci_disable_msix);
+#ifdef CONFIG_XEN
+EXPORT_SYMBOL(register_msi_get_owner);
+EXPORT_SYMBOL(unregister_msi_get_owner);
+#endif
+
diff -rpuN linux-2.6.18.8/drivers/pci/quirks.c linux-2.6.18-xen-3.3.0/drivers/pci/quirks.c
--- linux-2.6.18.8/drivers/pci/quirks.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/pci/quirks.c	2008-08-21 11:36:07.000000000 +0200
@@ -23,6 +23,40 @@
 #include <linux/acpi.h>
 #include "pci.h"
 
+/* A global flag which signals if we should page-align PCI mem windows. */
+int pci_mem_align = 0;
+
+static int __init set_pci_mem_align(char *str)
+{
+	pci_mem_align = 1;
+	return 1;
+}
+__setup("pci-mem-align", set_pci_mem_align);
+
+/* This quirk function enables us to force all memory resources which are 
+ * assigned to PCI devices, to be page-aligned.
+ */
+static void __devinit quirk_align_mem_resources(struct pci_dev *dev)
+{
+	int i;
+	struct resource *r;
+	resource_size_t old_start;
+
+	if (!pci_mem_align)
+		return;
+
+	for (i=0; i < DEVICE_COUNT_RESOURCE; i++) {
+		r = &dev->resource[i];
+		if ((r == NULL) || !(r->flags & IORESOURCE_MEM))
+			continue;
+
+		old_start = r->start;
+		r->start = (r->start + PAGE_SIZE - 1) & ~(PAGE_SIZE - 1);
+		r->end = r->end - (old_start - r->start);
+	}
+}
+DECLARE_PCI_FIXUP_HEADER(PCI_ANY_ID, PCI_ANY_ID, quirk_align_mem_resources);
+
 /* The Mellanox Tavor device gives false positive parity errors
  * Mark this device with a broken_parity_status, to allow
  * PCI scanning code to "skip" this now blacklisted device.
@@ -839,6 +873,25 @@ static void __init quirk_disable_pxb(str
 }
 DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_INTEL,	PCI_DEVICE_ID_INTEL_82454NX,	quirk_disable_pxb );
 
+static void __devinit quirk_sb600_sata(struct pci_dev *pdev)
+{
+	/* set sb600/sb700/sb800 sata to ahci mode */
+	u8 tmp;
+
+	pci_read_config_byte(pdev, PCI_CLASS_DEVICE, &tmp);
+	if (tmp == 0x01) {
+		pci_read_config_byte(pdev, 0x40, &tmp);
+		pci_write_config_byte(pdev, 0x40, tmp|1);
+		pci_write_config_byte(pdev, 0x9, 1);
+		pci_write_config_byte(pdev, 0xa, 6);
+		pci_write_config_byte(pdev, 0x40, tmp);
+
+		pdev->class = PCI_CLASS_STORAGE_SATA_AHCI;
+		dev_info(&pdev->dev, "set SATA to AHCI mode\n");
+	}
+}
+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_IXP700_SATA, quirk_sb600_sata);
+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_IXP600_SATA, quirk_sb600_sata);
 
 /*
  *	Serverworks CSB5 IDE does not fully support native mode
@@ -1494,10 +1547,11 @@ DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_N
 
 static void __devinit quirk_e100_interrupt(struct pci_dev *dev)
 {
-	u16 command;
+	u16 command, pmcsr;
 	u32 bar;
 	u8 __iomem *csr;
 	u8 cmd_hi;
+	int pm;
 
 	switch (dev->device) {
 	/* PCI IDs taken from drivers/net/e100.c */
@@ -1532,6 +1586,17 @@ static void __devinit quirk_e100_interru
 	if (!(command & PCI_COMMAND_MEMORY) || !bar)
 		return;
 
+        /*
+         * Check that the device is in the D0 power state. If it's not,
+         * there is no point to look any further.
+         */
+        pm = pci_find_capability(dev, PCI_CAP_ID_PM);
+        if (pm) {
+                pci_read_config_word(dev, pm + PCI_PM_CTRL, &pmcsr);
+                if ((pmcsr & PCI_PM_CTRL_STATE_MASK) != PCI_D0)
+                        return;
+        }
+
 	csr = ioremap(bar, 8);
 	if (!csr) {
 		printk(KERN_WARNING "PCI: Can't map %s e100 registers\n",
diff -rpuN linux-2.6.18.8/drivers/pnp/manager.c linux-2.6.18-xen-3.3.0/drivers/pnp/manager.c
--- linux-2.6.18.8/drivers/pnp/manager.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/pnp/manager.c	2008-08-21 11:36:07.000000000 +0200
@@ -168,7 +168,7 @@ static int pnp_assign_irq(struct pnp_dev
 	return 0;
 }
 
-static int pnp_assign_dma(struct pnp_dev *dev, struct pnp_dma *rule, int idx)
+static void pnp_assign_dma(struct pnp_dev *dev, struct pnp_dma *rule, int idx)
 {
 	resource_size_t *start, *end;
 	unsigned long *flags;
@@ -179,18 +179,14 @@ static int pnp_assign_dma(struct pnp_dev
 		1, 3, 5, 6, 7, 0, 2, 4
 	};
 
-	if (!dev || !rule)
-		return -EINVAL;
-
 	if (idx >= PNP_MAX_DMA) {
 		pnp_err("More than 2 dmas is incompatible with pnp specifications.");
-		/* pretend we were successful so at least the manager won't try again */
-		return 1;
+		return;
 	}
 
 	/* check if this resource has been manually set, if so skip */
 	if (!(dev->res.dma_resource[idx].flags & IORESOURCE_AUTO))
-		return 1;
+		return;
 
 	start = &dev->res.dma_resource[idx].start;
 	end = &dev->res.dma_resource[idx].end;
@@ -200,19 +196,17 @@ static int pnp_assign_dma(struct pnp_dev
 	*flags |= rule->flags | IORESOURCE_DMA;
 	*flags &=  ~IORESOURCE_UNSET;
 
-	if (!rule->map) {
-		*flags |= IORESOURCE_DISABLED;
-		return 1; /* skip disabled resource requests */
-	}
-
 	for (i = 0; i < 8; i++) {
 		if(rule->map & (1<<xtab[i])) {
 			*start = *end = xtab[i];
 			if(pnp_check_dma(dev, idx))
-				return 1;
+				return;
 		}
 	}
-	return 0;
+#ifdef MAX_DMA_CHANNELS
+	*start = *end = MAX_DMA_CHANNELS;
+#endif
+	*flags |= IORESOURCE_UNSET | IORESOURCE_DISABLED;
 }
 
 /**
@@ -331,8 +325,7 @@ static int pnp_assign_resources(struct p
 			irq = irq->next;
 		}
 		while (dma) {
-			if (!pnp_assign_dma(dev, dma, ndma))
-				goto fail;
+			pnp_assign_dma(dev, dma, ndma);
 			ndma++;
 			dma = dma->next;
 		}
@@ -367,8 +360,7 @@ static int pnp_assign_resources(struct p
 			irq = irq->next;
 		}
 		while (dma) {
-			if (!pnp_assign_dma(dev, dma, ndma))
-				goto fail;
+			pnp_assign_dma(dev, dma, ndma);
 			ndma++;
 			dma = dma->next;
 		}
diff -rpuN linux-2.6.18.8/drivers/scsi/ahci.c linux-2.6.18-xen-3.3.0/drivers/scsi/ahci.c
--- linux-2.6.18.8/drivers/scsi/ahci.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/scsi/ahci.c	2008-08-21 11:36:07.000000000 +0200
@@ -78,6 +78,7 @@ enum {
 
 	board_ahci		= 0,
 	board_ahci_vt8251	= 1,
+	board_ahci_sb700	= 2,
 
 	/* global controller registers */
 	HOST_CAP		= 0x00, /* host capabilities */
@@ -283,6 +284,16 @@ static const struct ata_port_info ahci_p
 		.udma_mask	= 0x7f, /* udma0-6 ; FIXME */
 		.port_ops	= &ahci_ops,
 	},
+	{
+		.sht            = &ahci_sht,
+		.host_flags     = ATA_FLAG_SATA | ATA_FLAG_NO_LEGACY |
+				  ATA_FLAG_MMIO | ATA_FLAG_PIO_DMA |
+				  ATA_FLAG_ACPI_SATA | ATA_FLAG_AN |
+				  ATA_FLAG_IPM,
+		.pio_mask       = 0x1f, /* pio0-4 */
+		.udma_mask      = 0x7f,
+		.port_ops       = &ahci_ops,
+	},
 };
 
 static const struct pci_device_id ahci_pci_tbl[] = {
@@ -317,6 +328,36 @@ static const struct pci_device_id ahci_p
 	  board_ahci }, /* ICH8M */
 	{ PCI_VENDOR_ID_INTEL, 0x282a, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
 	  board_ahci }, /* ICH8M */
+	{ PCI_VENDOR_ID_INTEL, 0x2922, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci }, /* ICH9 */
+	{ PCI_VENDOR_ID_INTEL, 0x2923, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci }, /* ICH9 */
+	{ PCI_VENDOR_ID_INTEL, 0x2924, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci }, /* ICH9 */
+	{ PCI_VENDOR_ID_INTEL, 0x2925, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci }, /* ICH9 */
+	{ PCI_VENDOR_ID_INTEL, 0x2927, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci }, /* ICH9 */
+	{ PCI_VENDOR_ID_INTEL, 0x2929, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci }, /* ICH9M */
+	{ PCI_VENDOR_ID_INTEL, 0x292a, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci }, /* ICH9M */
+	{ PCI_VENDOR_ID_INTEL, 0x292b, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci }, /* ICH9M */
+	{ PCI_VENDOR_ID_INTEL, 0x292f, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci }, /* ICH9M */
+	{ PCI_VENDOR_ID_INTEL, 0x294d, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci }, /* ICH9 */
+	{ PCI_VENDOR_ID_INTEL, 0x294e, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci }, /* ICH9M */
+	{ PCI_VENDOR_ID_INTEL, 0x3a02, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci }, /* ICH10 */
+	{ PCI_VENDOR_ID_INTEL, 0x3a05, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci }, /* ICH10 */
+	{ PCI_VENDOR_ID_INTEL, 0x3a22, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci }, /* ICH10 */
+	{ PCI_VENDOR_ID_INTEL, 0x3a25, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci }, /* ICH10 */
 
 	/* JMicron */
 	{ 0x197b, 0x2360, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
@@ -335,6 +376,18 @@ static const struct pci_device_id ahci_p
 	  board_ahci }, /* ATI SB600 non-raid */
 	{ PCI_VENDOR_ID_ATI, 0x4381, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
 	  board_ahci }, /* ATI SB600 raid */
+	{ PCI_VENDOR_ID_ATI, 0x4390, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci_sb700 },
+	{ PCI_VENDOR_ID_ATI, 0x4391, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci_sb700 },
+	{ PCI_VENDOR_ID_ATI, 0x4392, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci_sb700 },
+	{ PCI_VENDOR_ID_ATI, 0x4393, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci_sb700 },
+	{ PCI_VENDOR_ID_ATI, 0x4394, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci_sb700 },
+	{ PCI_VENDOR_ID_ATI, 0x4395, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
+	  board_ahci_sb700 },
 
 	/* VIA */
 	{ PCI_VENDOR_ID_VIA, 0x3349, PCI_ANY_ID, PCI_ANY_ID, 0, 0,
diff -rpuN linux-2.6.18.8/drivers/scsi/ata_piix.c linux-2.6.18-xen-3.3.0/drivers/scsi/ata_piix.c
--- linux-2.6.18.8/drivers/scsi/ata_piix.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/scsi/ata_piix.c	2008-08-21 11:36:07.000000000 +0200
@@ -125,6 +125,8 @@ enum {
 	ich6m_sata_ahci		= 6,
 	ich7m_sata_ahci		= 7,
 	ich8_sata_ahci		= 8,
+	ich9_sata_ahci		= 9,
+	ich8_2port_sata		= 10,
 
 	/* constants for mapping table */
 	P0			= 0,  /* port 0 */
@@ -192,12 +194,32 @@ static const struct pci_device_id piix_p
 	{ 0x8086, 0x27c4, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich7m_sata_ahci },
 	/* Enterprise Southbridge 2 (where's the datasheet?) */
 	{ 0x8086, 0x2680, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich6_sata_ahci },
-	/* SATA Controller 1 IDE (ICH8, no datasheet yet) */
+	/* SATA Controller 1 IDE (ICH8) */
 	{ 0x8086, 0x2820, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich8_sata_ahci },
-	/* SATA Controller 2 IDE (ICH8, ditto) */
-	{ 0x8086, 0x2825, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich8_sata_ahci },
+	/* SATA Controller 2 IDE (ICH8) */
+	{ 0x8086, 0x2825, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich8_2port_sata },
 	/* Mobile SATA Controller IDE (ICH8M, ditto) */
 	{ 0x8086, 0x2828, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich8_sata_ahci },
+	/* SATA Controller 1 IDE (ICH9) */
+	{ 0x8086, 0x2920, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich9_sata_ahci },
+	/* SATA Controller 1 IDE (ICH9) */
+	{ 0x8086, 0x2921, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich9_sata_ahci },
+	/* SATA Controller 2 IDE (ICH9) */
+	{ 0x8086, 0x2926, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich9_sata_ahci },
+	/* Mobile SATA Controller 1 IDE (ICH9M) */
+	{ 0x8086, 0x2928, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich9_sata_ahci },
+	/* Mobile SATA Controller 2 IDE (ICH9M) */
+	{ 0x8086, 0x292d, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich9_sata_ahci },
+	/* Mobile SATA Controller 2 IDE (ICH9M) */
+	{ 0x8086, 0x292e, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich9_sata_ahci },
+	/* SATA Controller IDE (ICH10) */
+	{ 0x8086, 0x3a00, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich8_sata_ahci },
+	/* SATA Controller IDE (ICH10) */
+	{ 0x8086, 0x3a06, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich8_2port_sata },
+	/* SATA Controller IDE (ICH10) */
+	{ 0x8086, 0x3a20, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich8_sata_ahci },
+	/* SATA Controller IDE (ICH10) */
+	{ 0x8086, 0x3a26, PCI_ANY_ID, PCI_ANY_ID, 0, 0, ich8_2port_sata },
 
 	{ }	/* terminate list */
 };
@@ -361,9 +383,34 @@ static const struct piix_map_db ich8_map
 	.present_shift = 8,
 	.map = {
 		/* PM   PS   SM   SS       MAP */
-		{  P0,  NA,  P1,  NA }, /* 00b (hardwired) */
+		{  P0,  P2,  P1,  P3 }, /* 00b (hardwired when in AHCI) */
 		{  RV,  RV,  RV,  RV },
-		{  RV,  RV,  RV,  RV }, /* 10b (never) */
+		{  IDE,  IDE,  NA,  NA }, /* 10b (IDE mode) */
+		{  RV,  RV,  RV,  RV },
+	},
+};
+
+static const struct piix_map_db ich9_map_db = {
+	.mask = 0x3,
+	.port_enable = 0x3,
+	.present_shift = 8,
+	.map = {
+		/* PM   PS   SM   SS       MAP */
+		{  P0,  P2,  P1,  P3 }, /* 00b (hardwired when in AHCI) */
+		{  RV,  RV,  RV,  RV },
+		{  IDE,  IDE,  NA,  NA }, /* 10b (IDE mode) */
+		{  RV,  RV,  RV,  RV },
+	},
+};
+
+static const struct piix_map_db ich8_2port_map_db = {
+	.mask = 0x3,
+	.port_enable = 0x3,
+	.map = {
+		/* PM   PS   SM   SS       MAP */
+		{  P0,  NA,  P1,  NA }, /* 00b */
+		{  RV,  RV,  RV,  RV }, /* 01b */
+		{  RV,  RV,  RV,  RV }, /* 10b */
 		{  RV,  RV,  RV,  RV },
 	},
 };
@@ -376,6 +423,8 @@ static const struct piix_map_db *piix_ma
 	[ich6m_sata_ahci]	= &ich6m_map_db,
 	[ich7m_sata_ahci]	= &ich7m_map_db,
 	[ich8_sata_ahci]	= &ich8_map_db,
+	[ich9_sata_ahci]	= &ich9_map_db,
+	[ich8_2port_sata]	= &ich8_2port_map_db,
 };
 
 static struct ata_port_info piix_port_info[] = {
@@ -487,6 +536,30 @@ static struct ata_port_info piix_port_in
 		.udma_mask	= 0x7f,	/* udma0-6 */
 		.port_ops	= &piix_sata_ops,
 	},
+
+	/* ich9_sata_ahci */
+	{
+		.sht		= &piix_sht,
+		.host_flags	= ATA_FLAG_SATA |
+				  PIIX_FLAG_CHECKINTR | PIIX_FLAG_SCR |
+				  PIIX_FLAG_AHCI,
+		.pio_mask	= 0x1f, /* pio0-4 */
+		.mwdma_mask	= 0x07, /* mwdma0-2 */
+		.udma_mask	= 0x7f, /* udma0-6 */
+		.port_ops	= &piix_sata_ops,
+	},
+	
+	/* ich8_2port_sata: 11: */
+	{
+		.sht		= &piix_sht,
+		.host_flags	= ATA_FLAG_SATA |
+				  PIIX_FLAG_CHECKINTR | PIIX_FLAG_SCR |
+				  PIIX_FLAG_AHCI,
+		.pio_mask	= 0x1f,	/* pio0-4 */
+		.mwdma_mask	= 0x07, /* mwdma0-2 */
+		.udma_mask	= ATA_UDMA6,
+		.port_ops	= &piix_pata_ops,
+	},
 };
 
 static struct pci_bits piix_enable_bits[] = {
diff -rpuN linux-2.6.18.8/drivers/serial/Kconfig linux-2.6.18-xen-3.3.0/drivers/serial/Kconfig
--- linux-2.6.18.8/drivers/serial/Kconfig	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/serial/Kconfig	2008-08-21 11:36:07.000000000 +0200
@@ -11,6 +11,7 @@ menu "Serial drivers"
 config SERIAL_8250
 	tristate "8250/16550 and compatible serial support"
 	depends on (BROKEN || !SPARC)
+	depends on !XEN_DISABLE_SERIAL
 	select SERIAL_CORE
 	---help---
 	  This selects whether you want to include the driver for the standard
diff -rpuN linux-2.6.18.8/drivers/video/console/Kconfig linux-2.6.18-xen-3.3.0/drivers/video/console/Kconfig
--- linux-2.6.18.8/drivers/video/console/Kconfig	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/video/console/Kconfig	2008-08-21 11:36:07.000000000 +0200
@@ -53,6 +53,7 @@ config VGACON_SOFT_SCROLLBACK_SIZE
 config VIDEO_SELECT
 	bool "Video mode selection support"
 	depends on  X86 && VGA_CONSOLE
+	depends on !XEN
 	---help---
 	  This enables support for text mode selection on kernel startup. If
 	  you want to take advantage of some high-resolution text mode your
diff -rpuN linux-2.6.18.8/drivers/video/Kconfig linux-2.6.18-xen-3.3.0/drivers/video/Kconfig
--- linux-2.6.18.8/drivers/video/Kconfig	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/video/Kconfig	2008-08-21 11:36:07.000000000 +0200
@@ -1254,7 +1254,7 @@ config FB_CYBLA
 	tristate "Cyberblade/i1 support"
 	depends on FB && PCI && X86_32 && !64BIT
 	select FB_CFB_IMAGEBLIT
-	select VIDEO_SELECT
+	select VIDEO_SELECT if !XEN
 	---help---
 	  This driver is supposed to support the Trident Cyberblade/i1
 	  graphics core integrated in the VIA VT8601A North Bridge,
diff -rpuN linux-2.6.18.8/drivers/xen/balloon/balloon.c linux-2.6.18-xen-3.3.0/drivers/xen/balloon/balloon.c
--- linux-2.6.18.8/drivers/xen/balloon/balloon.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/balloon/balloon.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,724 @@
+/******************************************************************************
+ * balloon.c
+ *
+ * Xen balloon driver - enables returning/claiming memory to/from Xen.
+ *
+ * Copyright (c) 2003, B Dragovic
+ * Copyright (c) 2003-2004, M Williamson, K Fraser
+ * Copyright (c) 2005 Dan M. Smith, IBM Corporation
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/sched.h>
+#include <linux/errno.h>
+#include <linux/mm.h>
+#include <linux/mman.h>
+#include <linux/smp_lock.h>
+#include <linux/pagemap.h>
+#include <linux/bootmem.h>
+#include <linux/highmem.h>
+#include <linux/vmalloc.h>
+#include <linux/mutex.h>
+#include <xen/xen_proc.h>
+#include <asm/hypervisor.h>
+#include <xen/balloon.h>
+#include <xen/interface/memory.h>
+#include <asm/maddr.h>
+#include <asm/page.h>
+#include <asm/pgalloc.h>
+#include <asm/pgtable.h>
+#include <asm/uaccess.h>
+#include <asm/tlb.h>
+#include <linux/highmem.h>
+#include <linux/list.h>
+#include <xen/xenbus.h>
+#include "common.h"
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+#ifdef CONFIG_PROC_FS
+static struct proc_dir_entry *balloon_pde;
+#endif
+
+static DEFINE_MUTEX(balloon_mutex);
+
+/*
+ * Protects atomic reservation decrease/increase against concurrent increases.
+ * Also protects non-atomic updates of current_pages and driver_pages, and
+ * balloon lists.
+ */
+DEFINE_SPINLOCK(balloon_lock);
+
+struct balloon_stats balloon_stats;
+
+/* We increase/decrease in batches which fit in a page */
+static unsigned long frame_list[PAGE_SIZE / sizeof(unsigned long)];
+
+/* VM /proc information for memory */
+extern unsigned long totalram_pages;
+
+#ifndef MODULE
+extern unsigned long totalhigh_pages;
+#define inc_totalhigh_pages() (totalhigh_pages++)
+#define dec_totalhigh_pages() (totalhigh_pages--)
+#else
+#define inc_totalhigh_pages() ((void)0)
+#define dec_totalhigh_pages() ((void)0)
+#endif
+
+/* List of ballooned pages, threaded through the mem_map array. */
+static LIST_HEAD(ballooned_pages);
+
+/* Main work function, always executed in process context. */
+static void balloon_process(void *unused);
+static DECLARE_WORK(balloon_worker, balloon_process, NULL);
+static struct timer_list balloon_timer;
+
+/* When ballooning out (allocating memory to return to Xen) we don't really 
+   want the kernel to try too hard since that can trigger the oom killer. */
+#define GFP_BALLOON \
+	(GFP_HIGHUSER|__GFP_NOWARN|__GFP_NORETRY|__GFP_NOMEMALLOC|__GFP_COLD)
+
+#define PAGE_TO_LIST(p) (&(p)->lru)
+#define LIST_TO_PAGE(l) list_entry((l), struct page, lru)
+#define UNLIST_PAGE(p)				\
+	do {					\
+		list_del(PAGE_TO_LIST(p));	\
+		PAGE_TO_LIST(p)->next = NULL;	\
+		PAGE_TO_LIST(p)->prev = NULL;	\
+	} while(0)
+
+#define IPRINTK(fmt, args...) \
+	printk(KERN_INFO "xen_mem: " fmt, ##args)
+#define WPRINTK(fmt, args...) \
+	printk(KERN_WARNING "xen_mem: " fmt, ##args)
+
+/* balloon_append: add the given page to the balloon. */
+static void balloon_append(struct page *page)
+{
+	/* Lowmem is re-populated first, so highmem pages go at list tail. */
+	if (PageHighMem(page)) {
+		list_add_tail(PAGE_TO_LIST(page), &ballooned_pages);
+		bs.balloon_high++;
+		dec_totalhigh_pages();
+	} else {
+		list_add(PAGE_TO_LIST(page), &ballooned_pages);
+		bs.balloon_low++;
+	}
+}
+
+/* balloon_retrieve: rescue a page from the balloon, if it is not empty. */
+static struct page *balloon_retrieve(void)
+{
+	struct page *page;
+
+	if (list_empty(&ballooned_pages))
+		return NULL;
+
+	page = LIST_TO_PAGE(ballooned_pages.next);
+	UNLIST_PAGE(page);
+
+	if (PageHighMem(page)) {
+		bs.balloon_high--;
+		inc_totalhigh_pages();
+	}
+	else
+		bs.balloon_low--;
+
+	return page;
+}
+
+static struct page *balloon_first_page(void)
+{
+	if (list_empty(&ballooned_pages))
+		return NULL;
+	return LIST_TO_PAGE(ballooned_pages.next);
+}
+
+static struct page *balloon_next_page(struct page *page)
+{
+	struct list_head *next = PAGE_TO_LIST(page)->next;
+	if (next == &ballooned_pages)
+		return NULL;
+	return LIST_TO_PAGE(next);
+}
+
+static inline void balloon_free_page(struct page *page)
+{
+#ifndef MODULE
+	if (put_page_testzero(page))
+		free_cold_page(page);
+#else
+	/* free_cold_page() is not being exported. */
+	__free_page(page);
+#endif
+}
+
+static void balloon_alarm(unsigned long unused)
+{
+	schedule_work(&balloon_worker);
+}
+
+static unsigned long current_target(void)
+{
+	unsigned long target = min(bs.target_pages, bs.hard_limit);
+	if (target > (bs.current_pages + bs.balloon_low + bs.balloon_high))
+		target = bs.current_pages + bs.balloon_low + bs.balloon_high;
+	return target;
+}
+
+static unsigned long minimum_target(void)
+{
+#ifndef CONFIG_XEN
+#define max_pfn num_physpages
+#endif
+	unsigned long min_pages, curr_pages = current_target();
+
+#define MB2PAGES(mb) ((mb) << (20 - PAGE_SHIFT))
+	/* Simple continuous piecewiese linear function:
+	 *  max MiB -> min MiB	gradient
+	 *       0	   0
+	 *      16	  16
+	 *      32	  24
+	 *     128	  72	(1/2)
+	 *     512 	 168	(1/4)
+	 *    2048	 360	(1/8)
+	 *    8192	 552	(1/32)
+	 *   32768	1320
+	 *  131072	4392
+	 */
+	if (max_pfn < MB2PAGES(128))
+		min_pages = MB2PAGES(8) + (max_pfn >> 1);
+	else if (max_pfn < MB2PAGES(512))
+		min_pages = MB2PAGES(40) + (max_pfn >> 2);
+	else if (max_pfn < MB2PAGES(2048))
+		min_pages = MB2PAGES(104) + (max_pfn >> 3);
+	else
+		min_pages = MB2PAGES(296) + (max_pfn >> 5);
+#undef MB2PAGES
+
+	/* Don't enforce growth */
+	return min(min_pages, curr_pages);
+#ifndef CONFIG_XEN
+#undef max_pfn
+#endif
+}
+
+static int increase_reservation(unsigned long nr_pages)
+{
+	unsigned long  pfn, i, flags;
+	struct page   *page;
+	long           rc;
+	struct xen_memory_reservation reservation = {
+		.address_bits = 0,
+		.extent_order = 0,
+		.domid        = DOMID_SELF
+	};
+
+	if (nr_pages > ARRAY_SIZE(frame_list))
+		nr_pages = ARRAY_SIZE(frame_list);
+
+	balloon_lock(flags);
+
+	page = balloon_first_page();
+	for (i = 0; i < nr_pages; i++) {
+		BUG_ON(page == NULL);
+		frame_list[i] = page_to_pfn(page);;
+		page = balloon_next_page(page);
+	}
+
+	set_xen_guest_handle(reservation.extent_start, frame_list);
+	reservation.nr_extents   = nr_pages;
+	rc = HYPERVISOR_memory_op(
+		XENMEM_populate_physmap, &reservation);
+	if (rc < nr_pages) {
+		if (rc > 0) {
+			int ret;
+
+			/* We hit the Xen hard limit: reprobe. */
+			reservation.nr_extents = rc;
+			ret = HYPERVISOR_memory_op(XENMEM_decrease_reservation,
+					&reservation);
+			BUG_ON(ret != rc);
+		}
+		if (rc >= 0)
+			bs.hard_limit = (bs.current_pages + rc -
+					 bs.driver_pages);
+		goto out;
+	}
+
+	for (i = 0; i < nr_pages; i++) {
+		page = balloon_retrieve();
+		BUG_ON(page == NULL);
+
+		pfn = page_to_pfn(page);
+		BUG_ON(!xen_feature(XENFEAT_auto_translated_physmap) &&
+		       phys_to_machine_mapping_valid(pfn));
+
+		set_phys_to_machine(pfn, frame_list[i]);
+
+#ifdef CONFIG_XEN
+		/* Link back into the page tables if not highmem. */
+		if (pfn < max_low_pfn) {
+			int ret;
+			ret = HYPERVISOR_update_va_mapping(
+				(unsigned long)__va(pfn << PAGE_SHIFT),
+				pfn_pte_ma(frame_list[i], PAGE_KERNEL),
+				0);
+			BUG_ON(ret);
+		}
+#endif
+
+		/* Relinquish the page back to the allocator. */
+		ClearPageReserved(page);
+		init_page_count(page);
+		balloon_free_page(page);
+	}
+
+	bs.current_pages += nr_pages;
+	totalram_pages = bs.current_pages;
+
+ out:
+	balloon_unlock(flags);
+
+	return 0;
+}
+
+static int decrease_reservation(unsigned long nr_pages)
+{
+	unsigned long  pfn, i, flags;
+	struct page   *page;
+	void          *v;
+	int            need_sleep = 0;
+	int ret;
+	struct xen_memory_reservation reservation = {
+		.address_bits = 0,
+		.extent_order = 0,
+		.domid        = DOMID_SELF
+	};
+
+	if (nr_pages > ARRAY_SIZE(frame_list))
+		nr_pages = ARRAY_SIZE(frame_list);
+
+	for (i = 0; i < nr_pages; i++) {
+		if ((page = alloc_page(GFP_BALLOON)) == NULL) {
+			nr_pages = i;
+			need_sleep = 1;
+			break;
+		}
+
+		pfn = page_to_pfn(page);
+		frame_list[i] = pfn_to_mfn(pfn);
+
+		if (!PageHighMem(page)) {
+			v = phys_to_virt(pfn << PAGE_SHIFT);
+			scrub_pages(v, 1);
+#ifdef CONFIG_XEN
+			ret = HYPERVISOR_update_va_mapping(
+				(unsigned long)v, __pte_ma(0), 0);
+			BUG_ON(ret);
+#endif
+		}
+#ifdef CONFIG_XEN_SCRUB_PAGES
+		else {
+			v = kmap(page);
+			scrub_pages(v, 1);
+			kunmap(page);
+		}
+#endif
+	}
+
+#ifdef CONFIG_XEN
+	/* Ensure that ballooned highmem pages don't have kmaps. */
+	kmap_flush_unused();
+	flush_tlb_all();
+#endif
+
+	balloon_lock(flags);
+
+	/* No more mappings: invalidate P2M and add to balloon. */
+	for (i = 0; i < nr_pages; i++) {
+		pfn = mfn_to_pfn(frame_list[i]);
+		set_phys_to_machine(pfn, INVALID_P2M_ENTRY);
+		balloon_append(pfn_to_page(pfn));
+	}
+
+	set_xen_guest_handle(reservation.extent_start, frame_list);
+	reservation.nr_extents   = nr_pages;
+	ret = HYPERVISOR_memory_op(XENMEM_decrease_reservation, &reservation);
+	BUG_ON(ret != nr_pages);
+
+	bs.current_pages -= nr_pages;
+	totalram_pages = bs.current_pages;
+
+	balloon_unlock(flags);
+
+	return need_sleep;
+}
+
+/*
+ * We avoid multiple worker processes conflicting via the balloon mutex.
+ * We may of course race updates of the target counts (which are protected
+ * by the balloon lock), or with changes to the Xen hard limit, but we will
+ * recover from these in time.
+ */
+static void balloon_process(void *unused)
+{
+	int need_sleep = 0;
+	long credit;
+
+	mutex_lock(&balloon_mutex);
+
+	do {
+		credit = current_target() - bs.current_pages;
+		if (credit > 0)
+			need_sleep = (increase_reservation(credit) != 0);
+		if (credit < 0)
+			need_sleep = (decrease_reservation(-credit) != 0);
+
+#ifndef CONFIG_PREEMPT
+		if (need_resched())
+			schedule();
+#endif
+	} while ((credit != 0) && !need_sleep);
+
+	/* Schedule more work if there is some still to be done. */
+	if (current_target() != bs.current_pages)
+		mod_timer(&balloon_timer, jiffies + HZ);
+
+	mutex_unlock(&balloon_mutex);
+}
+
+/* Resets the Xen limit, sets new target, and kicks off processing. */
+void balloon_set_new_target(unsigned long target)
+{
+	/* No need for lock. Not read-modify-write updates. */
+	bs.hard_limit   = ~0UL;
+	bs.target_pages = max(target, minimum_target());
+	schedule_work(&balloon_worker);
+}
+
+static struct xenbus_watch target_watch =
+{
+	.node = "memory/target"
+};
+
+/* React to a change in the target key */
+static void watch_target(struct xenbus_watch *watch,
+			 const char **vec, unsigned int len)
+{
+	unsigned long long new_target;
+	int err;
+
+	err = xenbus_scanf(XBT_NIL, "memory", "target", "%llu", &new_target);
+	if (err != 1) {
+		/* This is ok (for domain0 at least) - so just return */
+		return;
+	}
+
+	/* The given memory/target value is in KiB, so it needs converting to
+	 * pages. PAGE_SHIFT converts bytes to pages, hence PAGE_SHIFT - 10.
+	 */
+	balloon_set_new_target(new_target >> (PAGE_SHIFT - 10));
+}
+
+static int balloon_init_watcher(struct notifier_block *notifier,
+				unsigned long event,
+				void *data)
+{
+	int err;
+
+	err = register_xenbus_watch(&target_watch);
+	if (err)
+		printk(KERN_ERR "Failed to set balloon watcher\n");
+
+	return NOTIFY_DONE;
+}
+
+#ifdef CONFIG_PROC_FS
+static int balloon_write(struct file *file, const char __user *buffer,
+			 unsigned long count, void *data)
+{
+	char memstring[64], *endchar;
+	unsigned long long target_bytes;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	if (count <= 1)
+		return -EBADMSG; /* runt */
+	if (count > sizeof(memstring))
+		return -EFBIG;   /* too long */
+
+	if (copy_from_user(memstring, buffer, count))
+		return -EFAULT;
+	memstring[sizeof(memstring)-1] = '\0';
+
+	target_bytes = memparse(memstring, &endchar);
+	balloon_set_new_target(target_bytes >> PAGE_SHIFT);
+
+	return count;
+}
+
+static int balloon_read(char *page, char **start, off_t off,
+			int count, int *eof, void *data)
+{
+	int len;
+
+	len = sprintf(
+		page,
+		"Current allocation: %8lu kB\n"
+		"Requested target:   %8lu kB\n"
+		"Low-mem balloon:    %8lu kB\n"
+		"High-mem balloon:   %8lu kB\n"
+		"Driver pages:       %8lu kB\n"
+		"Xen hard limit:     ",
+		PAGES2KB(bs.current_pages), PAGES2KB(bs.target_pages), 
+		PAGES2KB(bs.balloon_low), PAGES2KB(bs.balloon_high),
+		PAGES2KB(bs.driver_pages));
+
+	if (bs.hard_limit != ~0UL)
+		len += sprintf(page + len, "%8lu kB\n",
+			       PAGES2KB(bs.hard_limit));
+	else
+		len += sprintf(page + len, "     ??? kB\n");
+
+	*eof = 1;
+	return len;
+}
+#endif
+
+static struct notifier_block xenstore_notifier;
+
+static int __init balloon_init(void)
+{
+#if defined(CONFIG_X86) && defined(CONFIG_XEN) 
+	unsigned long pfn;
+	struct page *page;
+#endif
+
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	IPRINTK("Initialising balloon driver.\n");
+
+#ifdef CONFIG_XEN
+	bs.current_pages = min(xen_start_info->nr_pages, max_pfn);
+	totalram_pages   = bs.current_pages;
+#else 
+	bs.current_pages = totalram_pages; 
+#endif
+	bs.target_pages  = bs.current_pages;
+	bs.balloon_low   = 0;
+	bs.balloon_high  = 0;
+	bs.driver_pages  = 0UL;
+	bs.hard_limit    = ~0UL;
+
+	init_timer(&balloon_timer);
+	balloon_timer.data = 0;
+	balloon_timer.function = balloon_alarm;
+    
+#ifdef CONFIG_PROC_FS
+	if ((balloon_pde = create_xen_proc_entry("balloon", 0644)) == NULL) {
+		WPRINTK("Unable to create /proc/xen/balloon.\n");
+		return -1;
+	}
+
+	balloon_pde->read_proc  = balloon_read;
+	balloon_pde->write_proc = balloon_write;
+#endif
+	balloon_sysfs_init();
+
+#if defined(CONFIG_X86) && defined(CONFIG_XEN) 
+	/* Initialise the balloon with excess memory space. */
+	for (pfn = xen_start_info->nr_pages; pfn < max_pfn; pfn++) {
+		page = pfn_to_page(pfn);
+		if (!PageReserved(page))
+			balloon_append(page);
+	}
+#endif
+
+	target_watch.callback = watch_target;
+	xenstore_notifier.notifier_call = balloon_init_watcher;
+
+	register_xenstore_notifier(&xenstore_notifier);
+    
+	return 0;
+}
+
+subsys_initcall(balloon_init);
+
+static void __exit balloon_exit(void)
+{
+    /* XXX - release balloon here */
+    return; 
+}
+
+module_exit(balloon_exit); 
+
+void balloon_update_driver_allowance(long delta)
+{
+	unsigned long flags;
+
+	balloon_lock(flags);
+	bs.driver_pages += delta;
+	balloon_unlock(flags);
+}
+
+#ifdef CONFIG_XEN
+static int dealloc_pte_fn(
+	pte_t *pte, struct page *pmd_page, unsigned long addr, void *data)
+{
+	unsigned long mfn = pte_mfn(*pte);
+	int ret;
+	struct xen_memory_reservation reservation = {
+		.nr_extents   = 1,
+		.extent_order = 0,
+		.domid        = DOMID_SELF
+	};
+	set_xen_guest_handle(reservation.extent_start, &mfn);
+	set_pte_at(&init_mm, addr, pte, __pte_ma(0));
+	set_phys_to_machine(__pa(addr) >> PAGE_SHIFT, INVALID_P2M_ENTRY);
+	ret = HYPERVISOR_memory_op(XENMEM_decrease_reservation, &reservation);
+	BUG_ON(ret != 1);
+	return 0;
+}
+#endif
+
+struct page **alloc_empty_pages_and_pagevec(int nr_pages)
+{
+	unsigned long flags;
+	void *v;
+	struct page *page, **pagevec;
+	int i, ret;
+
+	pagevec = kmalloc(sizeof(page) * nr_pages, GFP_KERNEL);
+	if (pagevec == NULL)
+		return NULL;
+
+	for (i = 0; i < nr_pages; i++) {
+		page = pagevec[i] = alloc_page(GFP_KERNEL|__GFP_COLD);
+		if (page == NULL)
+			goto err;
+
+		v = page_address(page);
+		scrub_pages(v, 1);
+
+		balloon_lock(flags);
+
+		if (xen_feature(XENFEAT_auto_translated_physmap)) {
+			unsigned long gmfn = page_to_pfn(page);
+			struct xen_memory_reservation reservation = {
+				.nr_extents   = 1,
+				.extent_order = 0,
+				.domid        = DOMID_SELF
+			};
+			set_xen_guest_handle(reservation.extent_start, &gmfn);
+			ret = HYPERVISOR_memory_op(XENMEM_decrease_reservation,
+						   &reservation);
+			if (ret == 1)
+				ret = 0; /* success */
+		} else {
+#ifdef CONFIG_XEN
+			ret = apply_to_page_range(&init_mm, (unsigned long)v,
+						  PAGE_SIZE, dealloc_pte_fn,
+						  NULL);
+#else
+			/* Cannot handle non-auto translate mode. */
+			ret = 1;
+#endif
+		}
+
+		if (ret != 0) {
+			balloon_unlock(flags);
+			balloon_free_page(page);
+			goto err;
+		}
+
+		totalram_pages = --bs.current_pages;
+
+		balloon_unlock(flags);
+	}
+
+ out:
+	schedule_work(&balloon_worker);
+#ifdef CONFIG_XEN
+	flush_tlb_all();
+#endif
+	return pagevec;
+
+ err:
+	balloon_lock(flags);
+	while (--i >= 0)
+		balloon_append(pagevec[i]);
+	balloon_unlock(flags);
+	kfree(pagevec);
+	pagevec = NULL;
+	goto out;
+}
+
+void free_empty_pages_and_pagevec(struct page **pagevec, int nr_pages)
+{
+	unsigned long flags;
+	int i;
+
+	if (pagevec == NULL)
+		return;
+
+	balloon_lock(flags);
+	for (i = 0; i < nr_pages; i++) {
+		BUG_ON(page_count(pagevec[i]) != 1);
+		balloon_append(pagevec[i]);
+	}
+	balloon_unlock(flags);
+
+	kfree(pagevec);
+
+	schedule_work(&balloon_worker);
+}
+
+void balloon_release_driver_page(struct page *page)
+{
+	unsigned long flags;
+
+	balloon_lock(flags);
+	balloon_append(page);
+	bs.driver_pages--;
+	balloon_unlock(flags);
+
+	schedule_work(&balloon_worker);
+}
+
+EXPORT_SYMBOL_GPL(balloon_update_driver_allowance);
+EXPORT_SYMBOL_GPL(alloc_empty_pages_and_pagevec);
+EXPORT_SYMBOL_GPL(free_empty_pages_and_pagevec);
+EXPORT_SYMBOL_GPL(balloon_release_driver_page);
+
+MODULE_LICENSE("Dual BSD/GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/balloon/common.h linux-2.6.18-xen-3.3.0/drivers/xen/balloon/common.h
--- linux-2.6.18.8/drivers/xen/balloon/common.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/balloon/common.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,58 @@
+/******************************************************************************
+ * balloon/common.h
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_BALLOON_COMMON_H__
+#define __XEN_BALLOON_COMMON_H__
+
+#define PAGES2KB(_p) ((_p)<<(PAGE_SHIFT-10))
+
+struct balloon_stats {
+	/* We aim for 'current allocation' == 'target allocation'. */
+	unsigned long current_pages;
+	unsigned long target_pages;
+	/* We may hit the hard limit in Xen. If we do then we remember it. */
+	unsigned long hard_limit;
+	/*
+	 * Drivers may alter the memory reservation independently, but they
+	 * must inform the balloon driver so we avoid hitting the hard limit.
+	 */
+	unsigned long driver_pages;
+	/* Number of pages in high- and low-memory balloons. */
+	unsigned long balloon_low;
+	unsigned long balloon_high;
+};
+
+extern struct balloon_stats balloon_stats;
+#define bs balloon_stats
+
+int balloon_sysfs_init(void);
+void balloon_sysfs_exit(void);
+
+void balloon_set_new_target(unsigned long target);
+
+#endif /* __XEN_BALLOON_COMMON_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/balloon/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/balloon/Makefile
--- linux-2.6.18.8/drivers/xen/balloon/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/balloon/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,2 @@
+
+obj-y := balloon.o sysfs.o
diff -rpuN linux-2.6.18.8/drivers/xen/balloon/sysfs.c linux-2.6.18-xen-3.3.0/drivers/xen/balloon/sysfs.c
--- linux-2.6.18.8/drivers/xen/balloon/sysfs.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/balloon/sysfs.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,170 @@
+/******************************************************************************
+ * balloon/sysfs.c
+ *
+ * Xen balloon driver - sysfs interfaces.
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/capability.h>
+#include <linux/errno.h>
+#include <linux/stat.h>
+#include <linux/string.h>
+#include <linux/sysdev.h>
+#include "common.h"
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+#define BALLOON_CLASS_NAME "xen_memory"
+
+#define BALLOON_SHOW(name, format, args...)			\
+	static ssize_t show_##name(struct sys_device *dev,	\
+				   char *buf)			\
+	{							\
+		return sprintf(buf, format, ##args);		\
+	}							\
+	static SYSDEV_ATTR(name, S_IRUGO, show_##name, NULL)
+
+BALLOON_SHOW(current_kb, "%lu\n", PAGES2KB(bs.current_pages));
+BALLOON_SHOW(low_kb, "%lu\n", PAGES2KB(bs.balloon_low));
+BALLOON_SHOW(high_kb, "%lu\n", PAGES2KB(bs.balloon_high));
+BALLOON_SHOW(hard_limit_kb,
+	     (bs.hard_limit!=~0UL) ? "%lu\n" : "???\n",
+	     (bs.hard_limit!=~0UL) ? PAGES2KB(bs.hard_limit) : 0);
+BALLOON_SHOW(driver_kb, "%lu\n", PAGES2KB(bs.driver_pages));
+
+static ssize_t show_target_kb(struct sys_device *dev, char *buf)
+{
+	return sprintf(buf, "%lu\n", PAGES2KB(bs.target_pages));
+}
+
+static ssize_t store_target_kb(struct sys_device *dev,
+			       const char *buf,
+			       size_t count)
+{
+	char memstring[64], *endchar;
+	unsigned long long target_bytes;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+	
+	if (count <= 1)
+		return -EBADMSG; /* runt */
+	if (count > sizeof(memstring))
+		return -EFBIG;   /* too long */
+	strcpy(memstring, buf);
+	
+	target_bytes = memparse(memstring, &endchar);
+	balloon_set_new_target(target_bytes >> PAGE_SHIFT);
+	
+	return count;
+}
+
+static SYSDEV_ATTR(target_kb, S_IRUGO | S_IWUSR,
+		   show_target_kb, store_target_kb);
+
+static struct sysdev_attribute *balloon_attrs[] = {
+	&attr_target_kb,
+};
+
+static struct attribute *balloon_info_attrs[] = {
+	&attr_current_kb.attr,
+	&attr_low_kb.attr,
+	&attr_high_kb.attr,
+	&attr_hard_limit_kb.attr,
+	&attr_driver_kb.attr,
+	NULL
+};
+
+static struct attribute_group balloon_info_group = {
+	.name = "info",
+	.attrs = balloon_info_attrs,
+};
+
+static struct sysdev_class balloon_sysdev_class = {
+	set_kset_name(BALLOON_CLASS_NAME),
+};
+
+static struct sys_device balloon_sysdev;
+
+static int register_balloon(struct sys_device *sysdev)
+{
+	int i, error;
+
+	error = sysdev_class_register(&balloon_sysdev_class);
+	if (error)
+		return error;
+
+	sysdev->id = 0;
+	sysdev->cls = &balloon_sysdev_class;
+
+	error = sysdev_register(sysdev);
+	if (error) {
+		sysdev_class_unregister(&balloon_sysdev_class);
+		return error;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(balloon_attrs); i++) {
+		error = sysdev_create_file(sysdev, balloon_attrs[i]);
+		if (error)
+			goto fail;
+	}
+
+	error = sysfs_create_group(&sysdev->kobj, &balloon_info_group);
+	if (error)
+		goto fail;
+	
+	return 0;
+
+ fail:
+	while (--i >= 0)
+		sysdev_remove_file(sysdev, balloon_attrs[i]);
+	sysdev_unregister(sysdev);
+	sysdev_class_unregister(&balloon_sysdev_class);
+	return error;
+}
+
+static void unregister_balloon(struct sys_device *sysdev)
+{
+	int i;
+
+	sysfs_remove_group(&sysdev->kobj, &balloon_info_group);
+	for (i = 0; i < ARRAY_SIZE(balloon_attrs); i++)
+		sysdev_remove_file(sysdev, balloon_attrs[i]);
+	sysdev_unregister(sysdev);
+	sysdev_class_unregister(&balloon_sysdev_class);
+}
+
+int balloon_sysfs_init(void)
+{
+	return register_balloon(&balloon_sysdev);
+}
+
+void balloon_sysfs_exit(void)
+{
+	unregister_balloon(&balloon_sysdev);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/blkback/blkback.c linux-2.6.18-xen-3.3.0/drivers/xen/blkback/blkback.c
--- linux-2.6.18.8/drivers/xen/blkback/blkback.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/blkback/blkback.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,646 @@
+/******************************************************************************
+ * arch/xen/drivers/blkif/backend/main.c
+ * 
+ * Back-end of the driver for virtual block devices. This portion of the
+ * driver exports a 'unified' block-device interface that can be accessed
+ * by any operating system that implements a compatible front end. A 
+ * reference front-end implementation can be found in:
+ *  arch/xen/drivers/blkif/frontend
+ * 
+ * Copyright (c) 2003-2004, Keir Fraser & Steve Hand
+ * Copyright (c) 2005, Christopher Clark
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/spinlock.h>
+#include <linux/kthread.h>
+#include <linux/list.h>
+#include <linux/delay.h>
+#include <xen/balloon.h>
+#include <asm/hypervisor.h>
+#include "common.h"
+
+/*
+ * These are rather arbitrary. They are fairly large because adjacent requests
+ * pulled from a communication ring are quite likely to end up being part of
+ * the same scatter/gather request at the disc.
+ * 
+ * ** TRY INCREASING 'blkif_reqs' IF WRITE SPEEDS SEEM TOO LOW **
+ * 
+ * This will increase the chances of being able to write whole tracks.
+ * 64 should be enough to keep us competitive with Linux.
+ */
+static int blkif_reqs = 64;
+module_param_named(reqs, blkif_reqs, int, 0);
+MODULE_PARM_DESC(reqs, "Number of blkback requests to allocate");
+
+/* Run-time switchable: /sys/module/blkback/parameters/ */
+static unsigned int log_stats = 0;
+static unsigned int debug_lvl = 0;
+module_param(log_stats, int, 0644);
+module_param(debug_lvl, int, 0644);
+
+/*
+ * Each outstanding request that we've passed to the lower device layers has a 
+ * 'pending_req' allocated to it. Each buffer_head that completes decrements 
+ * the pendcnt towards zero. When it hits zero, the specified domain has a 
+ * response queued for it, with the saved 'id' passed back.
+ */
+typedef struct {
+	blkif_t       *blkif;
+	u64            id;
+	int            nr_pages;
+	atomic_t       pendcnt;
+	unsigned short operation;
+	int            status;
+	struct list_head free_list;
+} pending_req_t;
+
+static pending_req_t *pending_reqs;
+static struct list_head pending_free;
+static DEFINE_SPINLOCK(pending_free_lock);
+static DECLARE_WAIT_QUEUE_HEAD(pending_free_wq);
+
+#define BLKBACK_INVALID_HANDLE (~0)
+
+static struct page **pending_pages;
+static grant_handle_t *pending_grant_handles;
+
+static inline int vaddr_pagenr(pending_req_t *req, int seg)
+{
+	return (req - pending_reqs) * BLKIF_MAX_SEGMENTS_PER_REQUEST + seg;
+}
+
+static inline unsigned long vaddr(pending_req_t *req, int seg)
+{
+	unsigned long pfn = page_to_pfn(pending_pages[vaddr_pagenr(req, seg)]);
+	return (unsigned long)pfn_to_kaddr(pfn);
+}
+
+#define pending_handle(_req, _seg) \
+	(pending_grant_handles[vaddr_pagenr(_req, _seg)])
+
+
+static int do_block_io_op(blkif_t *blkif);
+static void dispatch_rw_block_io(blkif_t *blkif,
+				 blkif_request_t *req,
+				 pending_req_t *pending_req);
+static void make_response(blkif_t *blkif, u64 id,
+			  unsigned short op, int st);
+
+/******************************************************************
+ * misc small helpers
+ */
+static pending_req_t* alloc_req(void)
+{
+	pending_req_t *req = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&pending_free_lock, flags);
+	if (!list_empty(&pending_free)) {
+		req = list_entry(pending_free.next, pending_req_t, free_list);
+		list_del(&req->free_list);
+	}
+	spin_unlock_irqrestore(&pending_free_lock, flags);
+	return req;
+}
+
+static void free_req(pending_req_t *req)
+{
+	unsigned long flags;
+	int was_empty;
+
+	spin_lock_irqsave(&pending_free_lock, flags);
+	was_empty = list_empty(&pending_free);
+	list_add(&req->free_list, &pending_free);
+	spin_unlock_irqrestore(&pending_free_lock, flags);
+	if (was_empty)
+		wake_up(&pending_free_wq);
+}
+
+static void unplug_queue(blkif_t *blkif)
+{
+	if (blkif->plug == NULL)
+		return;
+	if (blkif->plug->unplug_fn)
+		blkif->plug->unplug_fn(blkif->plug);
+	blk_put_queue(blkif->plug);
+	blkif->plug = NULL;
+}
+
+static void plug_queue(blkif_t *blkif, struct bio *bio)
+{
+	request_queue_t *q = bdev_get_queue(bio->bi_bdev);
+
+	if (q == blkif->plug)
+		return;
+	unplug_queue(blkif);
+	blk_get_queue(q);
+	blkif->plug = q;
+}
+
+static void fast_flush_area(pending_req_t *req)
+{
+	struct gnttab_unmap_grant_ref unmap[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+	unsigned int i, invcount = 0;
+	grant_handle_t handle;
+	int ret;
+
+	for (i = 0; i < req->nr_pages; i++) {
+		handle = pending_handle(req, i);
+		if (handle == BLKBACK_INVALID_HANDLE)
+			continue;
+		gnttab_set_unmap_op(&unmap[invcount], vaddr(req, i),
+				    GNTMAP_host_map, handle);
+		pending_handle(req, i) = BLKBACK_INVALID_HANDLE;
+		invcount++;
+	}
+
+	ret = HYPERVISOR_grant_table_op(
+		GNTTABOP_unmap_grant_ref, unmap, invcount);
+	BUG_ON(ret);
+}
+
+/******************************************************************
+ * SCHEDULER FUNCTIONS
+ */
+
+static void print_stats(blkif_t *blkif)
+{
+	printk(KERN_DEBUG "%s: oo %3d  |  rd %4d  |  wr %4d  |  br %4d\n",
+	       current->comm, blkif->st_oo_req,
+	       blkif->st_rd_req, blkif->st_wr_req, blkif->st_br_req);
+	blkif->st_print = jiffies + msecs_to_jiffies(10 * 1000);
+	blkif->st_rd_req = 0;
+	blkif->st_wr_req = 0;
+	blkif->st_oo_req = 0;
+}
+
+int blkif_schedule(void *arg)
+{
+	blkif_t *blkif = arg;
+
+	blkif_get(blkif);
+
+	if (debug_lvl)
+		printk(KERN_DEBUG "%s: started\n", current->comm);
+
+	while (!kthread_should_stop()) {
+		if (try_to_freeze())
+			continue;
+
+		wait_event_interruptible(
+			blkif->wq,
+			blkif->waiting_reqs || kthread_should_stop());
+		wait_event_interruptible(
+			pending_free_wq,
+			!list_empty(&pending_free) || kthread_should_stop());
+
+		blkif->waiting_reqs = 0;
+		smp_mb(); /* clear flag *before* checking for work */
+
+		if (do_block_io_op(blkif))
+			blkif->waiting_reqs = 1;
+		unplug_queue(blkif);
+
+		if (log_stats && time_after(jiffies, blkif->st_print))
+			print_stats(blkif);
+	}
+
+	if (log_stats)
+		print_stats(blkif);
+	if (debug_lvl)
+		printk(KERN_DEBUG "%s: exiting\n", current->comm);
+
+	blkif->xenblkd = NULL;
+	blkif_put(blkif);
+
+	return 0;
+}
+
+/******************************************************************
+ * COMPLETION CALLBACK -- Called as bh->b_end_io()
+ */
+
+static void __end_block_io_op(pending_req_t *pending_req, int error)
+{
+	/* An error fails the entire request. */
+	if ((pending_req->operation == BLKIF_OP_WRITE_BARRIER) &&
+	    (error == -EOPNOTSUPP)) {
+		DPRINTK("blkback: write barrier op failed, not supported\n");
+		blkback_barrier(XBT_NIL, pending_req->blkif->be, 0);
+		pending_req->status = BLKIF_RSP_EOPNOTSUPP;
+	} else if (error) {
+		DPRINTK("Buffer not up-to-date at end of operation, "
+			"error=%d\n", error);
+		pending_req->status = BLKIF_RSP_ERROR;
+	}
+
+	if (atomic_dec_and_test(&pending_req->pendcnt)) {
+		fast_flush_area(pending_req);
+		make_response(pending_req->blkif, pending_req->id,
+			      pending_req->operation, pending_req->status);
+		blkif_put(pending_req->blkif);
+		free_req(pending_req);
+	}
+}
+
+static int end_block_io_op(struct bio *bio, unsigned int done, int error)
+{
+	if (bio->bi_size != 0)
+		return 1;
+	__end_block_io_op(bio->bi_private, error);
+	bio_put(bio);
+	return error;
+}
+
+
+/******************************************************************************
+ * NOTIFICATION FROM GUEST OS.
+ */
+
+static void blkif_notify_work(blkif_t *blkif)
+{
+	blkif->waiting_reqs = 1;
+	wake_up(&blkif->wq);
+}
+
+irqreturn_t blkif_be_int(int irq, void *dev_id, struct pt_regs *regs)
+{
+	blkif_notify_work(dev_id);
+	return IRQ_HANDLED;
+}
+
+
+
+/******************************************************************
+ * DOWNWARD CALLS -- These interface with the block-device layer proper.
+ */
+
+static int do_block_io_op(blkif_t *blkif)
+{
+	blkif_back_rings_t *blk_rings = &blkif->blk_rings;
+	blkif_request_t req;
+	pending_req_t *pending_req;
+	RING_IDX rc, rp;
+	int more_to_do = 0;
+
+	rc = blk_rings->common.req_cons;
+	rp = blk_rings->common.sring->req_prod;
+	rmb(); /* Ensure we see queued requests up to 'rp'. */
+
+	while (rc != rp) {
+
+		if (RING_REQUEST_CONS_OVERFLOW(&blk_rings->common, rc))
+			break;
+
+		pending_req = alloc_req();
+		if (NULL == pending_req) {
+			blkif->st_oo_req++;
+			more_to_do = 1;
+			break;
+		}
+
+		if (kthread_should_stop()) {
+			more_to_do = 1;
+			break;
+		}
+
+		switch (blkif->blk_protocol) {
+		case BLKIF_PROTOCOL_NATIVE:
+			memcpy(&req, RING_GET_REQUEST(&blk_rings->native, rc), sizeof(req));
+			break;
+		case BLKIF_PROTOCOL_X86_32:
+			blkif_get_x86_32_req(&req, RING_GET_REQUEST(&blk_rings->x86_32, rc));
+			break;
+		case BLKIF_PROTOCOL_X86_64:
+			blkif_get_x86_64_req(&req, RING_GET_REQUEST(&blk_rings->x86_64, rc));
+			break;
+		default:
+			BUG();
+		}
+		blk_rings->common.req_cons = ++rc; /* before make_response() */
+
+		/* Apply all sanity checks to /private copy/ of request. */
+		barrier();
+
+		switch (req.operation) {
+		case BLKIF_OP_READ:
+			blkif->st_rd_req++;
+			dispatch_rw_block_io(blkif, &req, pending_req);
+			break;
+		case BLKIF_OP_WRITE_BARRIER:
+			blkif->st_br_req++;
+			/* fall through */
+		case BLKIF_OP_WRITE:
+			blkif->st_wr_req++;
+			dispatch_rw_block_io(blkif, &req, pending_req);
+			break;
+		default:
+			/* A good sign something is wrong: sleep for a while to
+			 * avoid excessive CPU consumption by a bad guest. */
+			msleep(1);
+			DPRINTK("error: unknown block io operation [%d]\n",
+				req.operation);
+			make_response(blkif, req.id, req.operation,
+				      BLKIF_RSP_ERROR);
+			free_req(pending_req);
+			break;
+		}
+
+		/* Yield point for this unbounded loop. */
+		cond_resched();
+	}
+
+	return more_to_do;
+}
+
+static void dispatch_rw_block_io(blkif_t *blkif,
+				 blkif_request_t *req,
+				 pending_req_t *pending_req)
+{
+	extern void ll_rw_block(int rw, int nr, struct buffer_head * bhs[]);
+	struct gnttab_map_grant_ref map[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+	struct phys_req preq;
+	struct { 
+		unsigned long buf; unsigned int nsec;
+	} seg[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+	unsigned int nseg;
+	struct bio *bio = NULL, *biolist[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+	int ret, i, nbio = 0;
+	int operation;
+
+	switch (req->operation) {
+	case BLKIF_OP_READ:
+		operation = READ;
+		break;
+	case BLKIF_OP_WRITE:
+		operation = WRITE;
+		break;
+	case BLKIF_OP_WRITE_BARRIER:
+		operation = WRITE_BARRIER;
+		break;
+	default:
+		operation = 0; /* make gcc happy */
+		BUG();
+	}
+
+	/* Check that number of segments is sane. */
+	nseg = req->nr_segments;
+	if (unlikely(nseg == 0 && operation != WRITE_BARRIER) || 
+	    unlikely(nseg > BLKIF_MAX_SEGMENTS_PER_REQUEST)) {
+		DPRINTK("Bad number of segments in request (%d)\n", nseg);
+		goto fail_response;
+	}
+
+	preq.dev           = req->handle;
+	preq.sector_number = req->sector_number;
+	preq.nr_sects      = 0;
+
+	pending_req->blkif     = blkif;
+	pending_req->id        = req->id;
+	pending_req->operation = req->operation;
+	pending_req->status    = BLKIF_RSP_OKAY;
+	pending_req->nr_pages  = nseg;
+
+	for (i = 0; i < nseg; i++) {
+		uint32_t flags;
+
+		seg[i].nsec = req->seg[i].last_sect -
+			req->seg[i].first_sect + 1;
+
+		if ((req->seg[i].last_sect >= (PAGE_SIZE >> 9)) ||
+		    (req->seg[i].last_sect < req->seg[i].first_sect))
+			goto fail_response;
+		preq.nr_sects += seg[i].nsec;
+
+		flags = GNTMAP_host_map;
+		if (operation != READ)
+			flags |= GNTMAP_readonly;
+		gnttab_set_map_op(&map[i], vaddr(pending_req, i), flags,
+				  req->seg[i].gref, blkif->domid);
+	}
+
+	ret = HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, map, nseg);
+	BUG_ON(ret);
+
+	for (i = 0; i < nseg; i++) {
+		if (unlikely(map[i].status != 0)) {
+			DPRINTK("invalid buffer -- could not remap it\n");
+			map[i].handle = BLKBACK_INVALID_HANDLE;
+			ret |= 1;
+		}
+
+		pending_handle(pending_req, i) = map[i].handle;
+
+		if (ret)
+			continue;
+
+		set_phys_to_machine(__pa(vaddr(
+			pending_req, i)) >> PAGE_SHIFT,
+			FOREIGN_FRAME(map[i].dev_bus_addr >> PAGE_SHIFT));
+		seg[i].buf  = map[i].dev_bus_addr | 
+			(req->seg[i].first_sect << 9);
+	}
+
+	if (ret)
+		goto fail_flush;
+
+	if (vbd_translate(&preq, blkif, operation) != 0) {
+		DPRINTK("access denied: %s of [%llu,%llu] on dev=%04x\n", 
+			operation == READ ? "read" : "write",
+			preq.sector_number,
+			preq.sector_number + preq.nr_sects, preq.dev);
+		goto fail_flush;
+	}
+
+	for (i = 0; i < nseg; i++) {
+		if (((int)preq.sector_number|(int)seg[i].nsec) &
+		    ((bdev_hardsect_size(preq.bdev) >> 9) - 1)) {
+			DPRINTK("Misaligned I/O request from domain %d",
+				blkif->domid);
+			goto fail_put_bio;
+		}
+
+		while ((bio == NULL) ||
+		       (bio_add_page(bio,
+				     virt_to_page(vaddr(pending_req, i)),
+				     seg[i].nsec << 9,
+				     seg[i].buf & ~PAGE_MASK) == 0)) {
+			bio = biolist[nbio++] = bio_alloc(GFP_KERNEL, nseg-i);
+			if (unlikely(bio == NULL))
+				goto fail_put_bio;
+
+			bio->bi_bdev    = preq.bdev;
+			bio->bi_private = pending_req;
+			bio->bi_end_io  = end_block_io_op;
+			bio->bi_sector  = preq.sector_number;
+		}
+
+		preq.sector_number += seg[i].nsec;
+	}
+
+	if (!bio) {
+		BUG_ON(operation != WRITE_BARRIER);
+		bio = biolist[nbio++] = bio_alloc(GFP_KERNEL, 0);
+		if (unlikely(bio == NULL))
+			goto fail_put_bio;
+
+		bio->bi_bdev    = preq.bdev;
+		bio->bi_private = pending_req;
+		bio->bi_end_io  = end_block_io_op;
+		bio->bi_sector  = -1;
+	}
+
+	plug_queue(blkif, bio);
+	atomic_set(&pending_req->pendcnt, nbio);
+	blkif_get(blkif);
+
+	for (i = 0; i < nbio; i++)
+		submit_bio(operation, biolist[i]);
+
+	if (operation == READ)
+		blkif->st_rd_sect += preq.nr_sects;
+	else if (operation == WRITE || operation == WRITE_BARRIER)
+		blkif->st_wr_sect += preq.nr_sects;
+
+	return;
+
+ fail_put_bio:
+	for (i = 0; i < (nbio-1); i++)
+		bio_put(biolist[i]);
+ fail_flush:
+	fast_flush_area(pending_req);
+ fail_response:
+	make_response(blkif, req->id, req->operation, BLKIF_RSP_ERROR);
+	free_req(pending_req);
+	msleep(1); /* back off a bit */
+} 
+
+
+
+/******************************************************************
+ * MISCELLANEOUS SETUP / TEARDOWN / DEBUGGING
+ */
+
+
+static void make_response(blkif_t *blkif, u64 id,
+			  unsigned short op, int st)
+{
+	blkif_response_t  resp;
+	unsigned long     flags;
+	blkif_back_rings_t *blk_rings = &blkif->blk_rings;
+	int more_to_do = 0;
+	int notify;
+
+	resp.id        = id;
+	resp.operation = op;
+	resp.status    = st;
+
+	spin_lock_irqsave(&blkif->blk_ring_lock, flags);
+	/* Place on the response ring for the relevant domain. */
+	switch (blkif->blk_protocol) {
+	case BLKIF_PROTOCOL_NATIVE:
+		memcpy(RING_GET_RESPONSE(&blk_rings->native, blk_rings->native.rsp_prod_pvt),
+		       &resp, sizeof(resp));
+		break;
+	case BLKIF_PROTOCOL_X86_32:
+		memcpy(RING_GET_RESPONSE(&blk_rings->x86_32, blk_rings->x86_32.rsp_prod_pvt),
+		       &resp, sizeof(resp));
+		break;
+	case BLKIF_PROTOCOL_X86_64:
+		memcpy(RING_GET_RESPONSE(&blk_rings->x86_64, blk_rings->x86_64.rsp_prod_pvt),
+		       &resp, sizeof(resp));
+		break;
+	default:
+		BUG();
+	}
+	blk_rings->common.rsp_prod_pvt++;
+	RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&blk_rings->common, notify);
+	if (blk_rings->common.rsp_prod_pvt == blk_rings->common.req_cons) {
+		/*
+		 * Tail check for pending requests. Allows frontend to avoid
+		 * notifications if requests are already in flight (lower
+		 * overheads and promotes batching).
+		 */
+		RING_FINAL_CHECK_FOR_REQUESTS(&blk_rings->common, more_to_do);
+
+	} else if (RING_HAS_UNCONSUMED_REQUESTS(&blk_rings->common)) {
+		more_to_do = 1;
+	}
+
+	spin_unlock_irqrestore(&blkif->blk_ring_lock, flags);
+
+	if (more_to_do)
+		blkif_notify_work(blkif);
+	if (notify)
+		notify_remote_via_irq(blkif->irq);
+}
+
+static int __init blkif_init(void)
+{
+	int i, mmap_pages;
+
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	mmap_pages = blkif_reqs * BLKIF_MAX_SEGMENTS_PER_REQUEST;
+
+	pending_reqs          = kmalloc(sizeof(pending_reqs[0]) *
+					blkif_reqs, GFP_KERNEL);
+	pending_grant_handles = kmalloc(sizeof(pending_grant_handles[0]) *
+					mmap_pages, GFP_KERNEL);
+	pending_pages         = alloc_empty_pages_and_pagevec(mmap_pages);
+
+	if (!pending_reqs || !pending_grant_handles || !pending_pages)
+		goto out_of_memory;
+
+	for (i = 0; i < mmap_pages; i++)
+		pending_grant_handles[i] = BLKBACK_INVALID_HANDLE;
+
+	blkif_interface_init();
+
+	memset(pending_reqs, 0, sizeof(pending_reqs));
+	INIT_LIST_HEAD(&pending_free);
+
+	for (i = 0; i < blkif_reqs; i++)
+		list_add_tail(&pending_reqs[i].free_list, &pending_free);
+
+	blkif_xenbus_init();
+
+	return 0;
+
+ out_of_memory:
+	kfree(pending_reqs);
+	kfree(pending_grant_handles);
+	free_empty_pages_and_pagevec(pending_pages, mmap_pages);
+	printk("%s: out of memory\n", __FUNCTION__);
+	return -ENOMEM;
+}
+
+module_init(blkif_init);
+
+MODULE_LICENSE("Dual BSD/GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/blkback/common.h linux-2.6.18-xen-3.3.0/drivers/xen/blkback/common.h
--- linux-2.6.18.8/drivers/xen/blkback/common.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/blkback/common.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,139 @@
+/* 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __BLKIF__BACKEND__COMMON_H__
+#define __BLKIF__BACKEND__COMMON_H__
+
+#include <linux/version.h>
+#include <linux/module.h>
+#include <linux/interrupt.h>
+#include <linux/slab.h>
+#include <linux/blkdev.h>
+#include <linux/vmalloc.h>
+#include <linux/wait.h>
+#include <asm/io.h>
+#include <asm/setup.h>
+#include <asm/pgalloc.h>
+#include <xen/evtchn.h>
+#include <asm/hypervisor.h>
+#include <xen/blkif.h>
+#include <xen/gnttab.h>
+#include <xen/driver_util.h>
+#include <xen/xenbus.h>
+
+#define DPRINTK(_f, _a...)			\
+	pr_debug("(file=%s, line=%d) " _f,	\
+		 __FILE__ , __LINE__ , ## _a )
+
+struct vbd {
+	blkif_vdev_t   handle;      /* what the domain refers to this vbd as */
+	unsigned char  readonly;    /* Non-zero -> read-only */
+	unsigned char  type;        /* VDISK_xxx */
+	u32            pdevice;     /* phys device that this vbd maps to */
+	struct block_device *bdev;
+};
+
+struct backend_info;
+
+typedef struct blkif_st {
+	/* Unique identifier for this interface. */
+	domid_t           domid;
+	unsigned int      handle;
+	/* Physical parameters of the comms window. */
+	unsigned int      irq;
+	/* Comms information. */
+	enum blkif_protocol blk_protocol;
+	blkif_back_rings_t blk_rings;
+	struct vm_struct *blk_ring_area;
+	/* The VBD attached to this interface. */
+	struct vbd        vbd;
+	/* Back pointer to the backend_info. */
+	struct backend_info *be;
+	/* Private fields. */
+	spinlock_t       blk_ring_lock;
+	atomic_t         refcnt;
+
+	wait_queue_head_t   wq;
+	struct task_struct  *xenblkd;
+	unsigned int        waiting_reqs;
+	request_queue_t     *plug;
+
+	/* statistics */
+	unsigned long       st_print;
+	int                 st_rd_req;
+	int                 st_wr_req;
+	int                 st_oo_req;
+	int                 st_br_req;
+	int                 st_rd_sect;
+	int                 st_wr_sect;
+
+	wait_queue_head_t waiting_to_free;
+
+	grant_handle_t shmem_handle;
+	grant_ref_t    shmem_ref;
+} blkif_t;
+
+blkif_t *blkif_alloc(domid_t domid);
+void blkif_disconnect(blkif_t *blkif);
+void blkif_free(blkif_t *blkif);
+int blkif_map(blkif_t *blkif, unsigned long shared_page, unsigned int evtchn);
+
+#define blkif_get(_b) (atomic_inc(&(_b)->refcnt))
+#define blkif_put(_b)					\
+	do {						\
+		if (atomic_dec_and_test(&(_b)->refcnt))	\
+			wake_up(&(_b)->waiting_to_free);\
+	} while (0)
+
+/* Create a vbd. */
+int vbd_create(blkif_t *blkif, blkif_vdev_t vdevice, unsigned major,
+	       unsigned minor, int readonly, int cdrom);
+void vbd_free(struct vbd *vbd);
+
+unsigned long long vbd_size(struct vbd *vbd);
+unsigned int vbd_info(struct vbd *vbd);
+unsigned long vbd_secsize(struct vbd *vbd);
+
+struct phys_req {
+	unsigned short       dev;
+	unsigned short       nr_sects;
+	struct block_device *bdev;
+	blkif_sector_t       sector_number;
+};
+
+int vbd_translate(struct phys_req *req, blkif_t *blkif, int operation);
+
+void blkif_interface_init(void);
+
+void blkif_xenbus_init(void);
+
+irqreturn_t blkif_be_int(int irq, void *dev_id, struct pt_regs *regs);
+int blkif_schedule(void *arg);
+
+int blkback_barrier(struct xenbus_transaction xbt,
+		    struct backend_info *be, int state);
+
+#endif /* __BLKIF__BACKEND__COMMON_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/blkback/interface.c linux-2.6.18-xen-3.3.0/drivers/xen/blkback/interface.c
--- linux-2.6.18.8/drivers/xen/blkback/interface.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/blkback/interface.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,181 @@
+/******************************************************************************
+ * arch/xen/drivers/blkif/backend/interface.c
+ * 
+ * Block-device interface management.
+ * 
+ * Copyright (c) 2004, Keir Fraser
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include "common.h"
+#include <xen/evtchn.h>
+#include <linux/kthread.h>
+
+static kmem_cache_t *blkif_cachep;
+
+blkif_t *blkif_alloc(domid_t domid)
+{
+	blkif_t *blkif;
+
+	blkif = kmem_cache_alloc(blkif_cachep, GFP_KERNEL);
+	if (!blkif)
+		return ERR_PTR(-ENOMEM);
+
+	memset(blkif, 0, sizeof(*blkif));
+	blkif->domid = domid;
+	spin_lock_init(&blkif->blk_ring_lock);
+	atomic_set(&blkif->refcnt, 1);
+	init_waitqueue_head(&blkif->wq);
+	blkif->st_print = jiffies;
+	init_waitqueue_head(&blkif->waiting_to_free);
+
+	return blkif;
+}
+
+static int map_frontend_page(blkif_t *blkif, unsigned long shared_page)
+{
+	struct gnttab_map_grant_ref op;
+
+	gnttab_set_map_op(&op, (unsigned long)blkif->blk_ring_area->addr,
+			  GNTMAP_host_map, shared_page, blkif->domid);
+
+	if (HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, &op, 1))
+		BUG();
+
+	if (op.status) {
+		DPRINTK(" Grant table operation failure !\n");
+		return op.status;
+	}
+
+	blkif->shmem_ref = shared_page;
+	blkif->shmem_handle = op.handle;
+
+	return 0;
+}
+
+static void unmap_frontend_page(blkif_t *blkif)
+{
+	struct gnttab_unmap_grant_ref op;
+
+	gnttab_set_unmap_op(&op, (unsigned long)blkif->blk_ring_area->addr,
+			    GNTMAP_host_map, blkif->shmem_handle);
+
+	if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1))
+		BUG();
+}
+
+int blkif_map(blkif_t *blkif, unsigned long shared_page, unsigned int evtchn)
+{
+	int err;
+
+	/* Already connected through? */
+	if (blkif->irq)
+		return 0;
+
+	if ( (blkif->blk_ring_area = alloc_vm_area(PAGE_SIZE)) == NULL )
+		return -ENOMEM;
+
+	err = map_frontend_page(blkif, shared_page);
+	if (err) {
+		free_vm_area(blkif->blk_ring_area);
+		return err;
+	}
+
+	switch (blkif->blk_protocol) {
+	case BLKIF_PROTOCOL_NATIVE:
+	{
+		blkif_sring_t *sring;
+		sring = (blkif_sring_t *)blkif->blk_ring_area->addr;
+		BACK_RING_INIT(&blkif->blk_rings.native, sring, PAGE_SIZE);
+		break;
+	}
+	case BLKIF_PROTOCOL_X86_32:
+	{
+		blkif_x86_32_sring_t *sring_x86_32;
+		sring_x86_32 = (blkif_x86_32_sring_t *)blkif->blk_ring_area->addr;
+		BACK_RING_INIT(&blkif->blk_rings.x86_32, sring_x86_32, PAGE_SIZE);
+		break;
+	}
+	case BLKIF_PROTOCOL_X86_64:
+	{
+		blkif_x86_64_sring_t *sring_x86_64;
+		sring_x86_64 = (blkif_x86_64_sring_t *)blkif->blk_ring_area->addr;
+		BACK_RING_INIT(&blkif->blk_rings.x86_64, sring_x86_64, PAGE_SIZE);
+		break;
+	}
+	default:
+		BUG();
+	}
+
+	err = bind_interdomain_evtchn_to_irqhandler(
+		blkif->domid, evtchn, blkif_be_int, 0, "blkif-backend", blkif);
+	if (err < 0)
+	{
+		unmap_frontend_page(blkif);
+		free_vm_area(blkif->blk_ring_area);
+		blkif->blk_rings.common.sring = NULL;
+		return err;
+	}
+	blkif->irq = err;
+
+	return 0;
+}
+
+void blkif_disconnect(blkif_t *blkif)
+{
+	if (blkif->xenblkd) {
+		kthread_stop(blkif->xenblkd);
+		blkif->xenblkd = NULL;
+	}
+
+	atomic_dec(&blkif->refcnt);
+	wait_event(blkif->waiting_to_free, atomic_read(&blkif->refcnt) == 0);
+	atomic_inc(&blkif->refcnt);
+
+	if (blkif->irq) {
+		unbind_from_irqhandler(blkif->irq, blkif);
+		blkif->irq = 0;
+	}
+
+	if (blkif->blk_rings.common.sring) {
+		unmap_frontend_page(blkif);
+		free_vm_area(blkif->blk_ring_area);
+		blkif->blk_rings.common.sring = NULL;
+	}
+}
+
+void blkif_free(blkif_t *blkif)
+{
+	if (!atomic_dec_and_test(&blkif->refcnt))
+		BUG();
+	kmem_cache_free(blkif_cachep, blkif);
+}
+
+void __init blkif_interface_init(void)
+{
+	blkif_cachep = kmem_cache_create("blkif_cache", sizeof(blkif_t), 
+					 0, 0, NULL, NULL);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/blkback/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/blkback/Makefile
--- linux-2.6.18.8/drivers/xen/blkback/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/blkback/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,3 @@
+obj-$(CONFIG_XEN_BLKDEV_BACKEND) := blkbk.o
+
+blkbk-y	:= blkback.o xenbus.o interface.o vbd.o
diff -rpuN linux-2.6.18.8/drivers/xen/blkback/vbd.c linux-2.6.18-xen-3.3.0/drivers/xen/blkback/vbd.c
--- linux-2.6.18.8/drivers/xen/blkback/vbd.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/blkback/vbd.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,118 @@
+/******************************************************************************
+ * blkback/vbd.c
+ * 
+ * Routines for managing virtual block devices (VBDs).
+ * 
+ * Copyright (c) 2003-2005, Keir Fraser & Steve Hand
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include "common.h"
+
+#define vbd_sz(_v)   ((_v)->bdev->bd_part ?				\
+	(_v)->bdev->bd_part->nr_sects : (_v)->bdev->bd_disk->capacity)
+
+unsigned long long vbd_size(struct vbd *vbd)
+{
+	return vbd_sz(vbd);
+}
+
+unsigned int vbd_info(struct vbd *vbd)
+{
+	return vbd->type | (vbd->readonly?VDISK_READONLY:0);
+}
+
+unsigned long vbd_secsize(struct vbd *vbd)
+{
+	return bdev_hardsect_size(vbd->bdev);
+}
+
+int vbd_create(blkif_t *blkif, blkif_vdev_t handle, unsigned major,
+	       unsigned minor, int readonly, int cdrom)
+{
+	struct vbd *vbd;
+	struct block_device *bdev;
+
+	vbd = &blkif->vbd;
+	vbd->handle   = handle; 
+	vbd->readonly = readonly;
+	vbd->type     = 0;
+
+	vbd->pdevice  = MKDEV(major, minor);
+
+	bdev = open_by_devnum(vbd->pdevice,
+			      vbd->readonly ? FMODE_READ : FMODE_WRITE);
+
+	if (IS_ERR(bdev)) {
+		DPRINTK("vbd_creat: device %08x could not be opened.\n",
+			vbd->pdevice);
+		return -ENOENT;
+	}
+
+	vbd->bdev = bdev;
+
+	if (vbd->bdev->bd_disk == NULL) {
+		DPRINTK("vbd_creat: device %08x doesn't exist.\n",
+			vbd->pdevice);
+		vbd_free(vbd);
+		return -ENOENT;
+	}
+
+	if (vbd->bdev->bd_disk->flags & GENHD_FL_CD || cdrom)
+		vbd->type |= VDISK_CDROM;
+	if (vbd->bdev->bd_disk->flags & GENHD_FL_REMOVABLE)
+		vbd->type |= VDISK_REMOVABLE;
+
+	DPRINTK("Successful creation of handle=%04x (dom=%u)\n",
+		handle, blkif->domid);
+	return 0;
+}
+
+void vbd_free(struct vbd *vbd)
+{
+	if (vbd->bdev)
+		blkdev_put(vbd->bdev);
+	vbd->bdev = NULL;
+}
+
+int vbd_translate(struct phys_req *req, blkif_t *blkif, int operation)
+{
+	struct vbd *vbd = &blkif->vbd;
+	int rc = -EACCES;
+
+	if ((operation != READ) && vbd->readonly)
+		goto out;
+
+	if (unlikely((req->sector_number + req->nr_sects) > vbd_sz(vbd)))
+		goto out;
+
+	req->dev  = vbd->pdevice;
+	req->bdev = vbd->bdev;
+	rc = 0;
+
+ out:
+	return rc;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/blkback/xenbus.c linux-2.6.18-xen-3.3.0/drivers/xen/blkback/xenbus.c
--- linux-2.6.18.8/drivers/xen/blkback/xenbus.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/blkback/xenbus.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,541 @@
+/*  Xenbus code for blkif backend
+    Copyright (C) 2005 Rusty Russell <rusty@rustcorp.com.au>
+    Copyright (C) 2005 XenSource Ltd
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program; if not, write to the Free Software
+    Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+*/
+
+#include <stdarg.h>
+#include <linux/module.h>
+#include <linux/kthread.h>
+#include "common.h"
+
+#undef DPRINTK
+#define DPRINTK(fmt, args...)				\
+	pr_debug("blkback/xenbus (%s:%d) " fmt ".\n",	\
+		 __FUNCTION__, __LINE__, ##args)
+
+struct backend_info
+{
+	struct xenbus_device *dev;
+	blkif_t *blkif;
+	struct xenbus_watch backend_watch;
+	unsigned major;
+	unsigned minor;
+	char *mode;
+};
+
+static void connect(struct backend_info *);
+static int connect_ring(struct backend_info *);
+static void backend_changed(struct xenbus_watch *, const char **,
+			    unsigned int);
+
+static int blkback_name(blkif_t *blkif, char *buf)
+{
+	char *devpath, *devname;
+	struct xenbus_device *dev = blkif->be->dev;
+
+	devpath = xenbus_read(XBT_NIL, dev->nodename, "dev", NULL);
+	if (IS_ERR(devpath)) 
+		return PTR_ERR(devpath);
+	
+	if ((devname = strstr(devpath, "/dev/")) != NULL)
+		devname += strlen("/dev/");
+	else
+		devname  = devpath;
+
+	snprintf(buf, TASK_COMM_LEN, "blkback.%d.%s", blkif->domid, devname);
+	kfree(devpath);
+	
+	return 0;
+}
+
+static void update_blkif_status(blkif_t *blkif)
+{ 
+	int err;
+	char name[TASK_COMM_LEN];
+
+	/* Not ready to connect? */
+	if (!blkif->irq || !blkif->vbd.bdev)
+		return;
+
+	/* Already connected? */
+	if (blkif->be->dev->state == XenbusStateConnected)
+		return;
+
+	/* Attempt to connect: exit if we fail to. */
+	connect(blkif->be);
+	if (blkif->be->dev->state != XenbusStateConnected)
+		return;
+
+	err = blkback_name(blkif, name);
+	if (err) {
+		xenbus_dev_error(blkif->be->dev, err, "get blkback dev name");
+		return;
+	}
+
+	blkif->xenblkd = kthread_run(blkif_schedule, blkif, name);
+	if (IS_ERR(blkif->xenblkd)) {
+		err = PTR_ERR(blkif->xenblkd);
+		blkif->xenblkd = NULL;
+		xenbus_dev_error(blkif->be->dev, err, "start xenblkd");
+	}
+}
+
+
+/****************************************************************
+ *  sysfs interface for VBD I/O requests
+ */
+
+#define VBD_SHOW(name, format, args...)					\
+	static ssize_t show_##name(struct device *_dev,			\
+				   struct device_attribute *attr,	\
+				   char *buf)				\
+	{								\
+		struct xenbus_device *dev = to_xenbus_device(_dev);	\
+		struct backend_info *be = dev->dev.driver_data;		\
+									\
+		return sprintf(buf, format, ##args);			\
+	}								\
+	static DEVICE_ATTR(name, S_IRUGO, show_##name, NULL)
+
+VBD_SHOW(oo_req,  "%d\n", be->blkif->st_oo_req);
+VBD_SHOW(rd_req,  "%d\n", be->blkif->st_rd_req);
+VBD_SHOW(wr_req,  "%d\n", be->blkif->st_wr_req);
+VBD_SHOW(br_req,  "%d\n", be->blkif->st_br_req);
+VBD_SHOW(rd_sect, "%d\n", be->blkif->st_rd_sect);
+VBD_SHOW(wr_sect, "%d\n", be->blkif->st_wr_sect);
+
+static struct attribute *vbdstat_attrs[] = {
+	&dev_attr_oo_req.attr,
+	&dev_attr_rd_req.attr,
+	&dev_attr_wr_req.attr,
+	&dev_attr_br_req.attr,
+	&dev_attr_rd_sect.attr,
+	&dev_attr_wr_sect.attr,
+	NULL
+};
+
+static struct attribute_group vbdstat_group = {
+	.name = "statistics",
+	.attrs = vbdstat_attrs,
+};
+
+VBD_SHOW(physical_device, "%x:%x\n", be->major, be->minor);
+VBD_SHOW(mode, "%s\n", be->mode);
+
+int xenvbd_sysfs_addif(struct xenbus_device *dev)
+{
+	int error;
+	
+	error = device_create_file(&dev->dev, &dev_attr_physical_device);
+ 	if (error)
+		goto fail1;
+
+	error = device_create_file(&dev->dev, &dev_attr_mode);
+	if (error)
+		goto fail2;
+
+	error = sysfs_create_group(&dev->dev.kobj, &vbdstat_group);
+	if (error)
+		goto fail3;
+
+	return 0;
+
+fail3:	sysfs_remove_group(&dev->dev.kobj, &vbdstat_group);
+fail2:	device_remove_file(&dev->dev, &dev_attr_mode);
+fail1:	device_remove_file(&dev->dev, &dev_attr_physical_device);
+	return error;
+}
+
+void xenvbd_sysfs_delif(struct xenbus_device *dev)
+{
+	sysfs_remove_group(&dev->dev.kobj, &vbdstat_group);
+	device_remove_file(&dev->dev, &dev_attr_mode);
+	device_remove_file(&dev->dev, &dev_attr_physical_device);
+}
+
+static int blkback_remove(struct xenbus_device *dev)
+{
+	struct backend_info *be = dev->dev.driver_data;
+
+	DPRINTK("");
+
+	if (be->major || be->minor)
+		xenvbd_sysfs_delif(dev);
+
+	if (be->backend_watch.node) {
+		unregister_xenbus_watch(&be->backend_watch);
+		kfree(be->backend_watch.node);
+		be->backend_watch.node = NULL;
+	}
+
+	if (be->blkif) {
+		blkif_disconnect(be->blkif);
+		vbd_free(&be->blkif->vbd);
+		blkif_free(be->blkif);
+		be->blkif = NULL;
+	}
+
+	kfree(be);
+	dev->dev.driver_data = NULL;
+	return 0;
+}
+
+int blkback_barrier(struct xenbus_transaction xbt,
+		    struct backend_info *be, int state)
+{
+	struct xenbus_device *dev = be->dev;
+	int err;
+
+	err = xenbus_printf(xbt, dev->nodename, "feature-barrier",
+			    "%d", state);
+	if (err)
+		xenbus_dev_fatal(dev, err, "writing feature-barrier");
+
+	return err;
+}
+
+/**
+ * Entry point to this code when a new device is created.  Allocate the basic
+ * structures, and watch the store waiting for the hotplug scripts to tell us
+ * the device's physical major and minor numbers.  Switch to InitWait.
+ */
+static int blkback_probe(struct xenbus_device *dev,
+			 const struct xenbus_device_id *id)
+{
+	int err;
+	struct backend_info *be = kzalloc(sizeof(struct backend_info),
+					  GFP_KERNEL);
+	if (!be) {
+		xenbus_dev_fatal(dev, -ENOMEM,
+				 "allocating backend structure");
+		return -ENOMEM;
+	}
+	be->dev = dev;
+	dev->dev.driver_data = be;
+
+	be->blkif = blkif_alloc(dev->otherend_id);
+	if (IS_ERR(be->blkif)) {
+		err = PTR_ERR(be->blkif);
+		be->blkif = NULL;
+		xenbus_dev_fatal(dev, err, "creating block interface");
+		goto fail;
+	}
+
+	/* setup back pointer */
+	be->blkif->be = be;
+
+	err = xenbus_watch_path2(dev, dev->nodename, "physical-device",
+				 &be->backend_watch, backend_changed);
+	if (err)
+		goto fail;
+
+	err = xenbus_switch_state(dev, XenbusStateInitWait);
+	if (err)
+		goto fail;
+
+	return 0;
+
+fail:
+	DPRINTK("failed");
+	blkback_remove(dev);
+	return err;
+}
+
+
+/**
+ * Callback received when the hotplug scripts have placed the physical-device
+ * node.  Read it and the mode node, and create a vbd.  If the frontend is
+ * ready, connect.
+ */
+static void backend_changed(struct xenbus_watch *watch,
+			    const char **vec, unsigned int len)
+{
+	int err;
+	unsigned major;
+	unsigned minor;
+	struct backend_info *be
+		= container_of(watch, struct backend_info, backend_watch);
+	struct xenbus_device *dev = be->dev;
+	int cdrom = 0;
+	char *device_type;
+
+	DPRINTK("");
+
+	err = xenbus_scanf(XBT_NIL, dev->nodename, "physical-device", "%x:%x",
+			   &major, &minor);
+	if (XENBUS_EXIST_ERR(err)) {
+		/* Since this watch will fire once immediately after it is
+		   registered, we expect this.  Ignore it, and wait for the
+		   hotplug scripts. */
+		return;
+	}
+	if (err != 2) {
+		xenbus_dev_fatal(dev, err, "reading physical-device");
+		return;
+	}
+
+	if ((be->major || be->minor) &&
+	    ((be->major != major) || (be->minor != minor))) {
+		printk(KERN_WARNING
+		       "blkback: changing physical device (from %x:%x to "
+		       "%x:%x) not supported.\n", be->major, be->minor,
+		       major, minor);
+		return;
+	}
+
+	be->mode = xenbus_read(XBT_NIL, dev->nodename, "mode", NULL);
+	if (IS_ERR(be->mode)) {
+		err = PTR_ERR(be->mode);
+		be->mode = NULL;
+		xenbus_dev_fatal(dev, err, "reading mode");
+		return;
+	}
+
+	device_type = xenbus_read(XBT_NIL, dev->otherend, "device-type", NULL);
+	if (!IS_ERR(device_type)) {
+		cdrom = strcmp(device_type, "cdrom") == 0;
+		kfree(device_type);
+	}
+
+	if (be->major == 0 && be->minor == 0) {
+		/* Front end dir is a number, which is used as the handle. */
+
+		char *p = strrchr(dev->otherend, '/') + 1;
+		long handle = simple_strtoul(p, NULL, 0);
+
+		be->major = major;
+		be->minor = minor;
+
+		err = vbd_create(be->blkif, handle, major, minor,
+				 (NULL == strchr(be->mode, 'w')), cdrom);
+		if (err) {
+			be->major = be->minor = 0;
+			xenbus_dev_fatal(dev, err, "creating vbd structure");
+			return;
+		}
+
+		err = xenvbd_sysfs_addif(dev);
+		if (err) {
+			vbd_free(&be->blkif->vbd);
+			be->major = be->minor = 0;
+			xenbus_dev_fatal(dev, err, "creating sysfs entries");
+			return;
+		}
+
+		/* We're potentially connected now */
+		update_blkif_status(be->blkif);
+	}
+}
+
+
+/**
+ * Callback received when the frontend's state changes.
+ */
+static void frontend_changed(struct xenbus_device *dev,
+			     enum xenbus_state frontend_state)
+{
+	struct backend_info *be = dev->dev.driver_data;
+	int err;
+
+	DPRINTK("%s", xenbus_strstate(frontend_state));
+
+	switch (frontend_state) {
+	case XenbusStateInitialising:
+		if (dev->state == XenbusStateClosed) {
+			printk(KERN_INFO "%s: %s: prepare for reconnect\n",
+			       __FUNCTION__, dev->nodename);
+			xenbus_switch_state(dev, XenbusStateInitWait);
+		}
+		break;
+
+	case XenbusStateInitialised:
+	case XenbusStateConnected:
+		/* Ensure we connect even when two watches fire in 
+		   close successsion and we miss the intermediate value 
+		   of frontend_state. */
+		if (dev->state == XenbusStateConnected)
+			break;
+
+		err = connect_ring(be);
+		if (err)
+			break;
+		update_blkif_status(be->blkif);
+		break;
+
+	case XenbusStateClosing:
+		blkif_disconnect(be->blkif);
+		xenbus_switch_state(dev, XenbusStateClosing);
+		break;
+
+	case XenbusStateClosed:
+		xenbus_switch_state(dev, XenbusStateClosed);
+		if (xenbus_dev_is_online(dev))
+			break;
+		/* fall through if not online */
+	case XenbusStateUnknown:
+		device_unregister(&dev->dev);
+		break;
+
+	default:
+		xenbus_dev_fatal(dev, -EINVAL, "saw state %d at frontend",
+				 frontend_state);
+		break;
+	}
+}
+
+
+/* ** Connection ** */
+
+
+/**
+ * Write the physical details regarding the block device to the store, and
+ * switch to Connected state.
+ */
+static void connect(struct backend_info *be)
+{
+	struct xenbus_transaction xbt;
+	int err;
+	struct xenbus_device *dev = be->dev;
+
+	DPRINTK("%s", dev->otherend);
+
+	/* Supply the information about the device the frontend needs */
+again:
+	err = xenbus_transaction_start(&xbt);
+	if (err) {
+		xenbus_dev_fatal(dev, err, "starting transaction");
+		return;
+	}
+
+	err = blkback_barrier(xbt, be, 1);
+	if (err)
+		goto abort;
+
+	err = xenbus_printf(xbt, dev->nodename, "sectors", "%llu",
+			    vbd_size(&be->blkif->vbd));
+	if (err) {
+		xenbus_dev_fatal(dev, err, "writing %s/sectors",
+				 dev->nodename);
+		goto abort;
+	}
+
+	/* FIXME: use a typename instead */
+	err = xenbus_printf(xbt, dev->nodename, "info", "%u",
+			    vbd_info(&be->blkif->vbd));
+	if (err) {
+		xenbus_dev_fatal(dev, err, "writing %s/info",
+				 dev->nodename);
+		goto abort;
+	}
+	err = xenbus_printf(xbt, dev->nodename, "sector-size", "%lu",
+			    vbd_secsize(&be->blkif->vbd));
+	if (err) {
+		xenbus_dev_fatal(dev, err, "writing %s/sector-size",
+				 dev->nodename);
+		goto abort;
+	}
+
+	err = xenbus_transaction_end(xbt, 0);
+	if (err == -EAGAIN)
+		goto again;
+	if (err)
+		xenbus_dev_fatal(dev, err, "ending transaction");
+
+	err = xenbus_switch_state(dev, XenbusStateConnected);
+	if (err)
+		xenbus_dev_fatal(dev, err, "switching to Connected state",
+				 dev->nodename);
+
+	return;
+ abort:
+	xenbus_transaction_end(xbt, 1);
+}
+
+
+static int connect_ring(struct backend_info *be)
+{
+	struct xenbus_device *dev = be->dev;
+	unsigned long ring_ref;
+	unsigned int evtchn;
+	char protocol[64] = "";
+	int err;
+
+	DPRINTK("%s", dev->otherend);
+
+	err = xenbus_gather(XBT_NIL, dev->otherend, "ring-ref", "%lu", &ring_ref,
+			    "event-channel", "%u", &evtchn, NULL);
+	if (err) {
+		xenbus_dev_fatal(dev, err,
+				 "reading %s/ring-ref and event-channel",
+				 dev->otherend);
+		return err;
+	}
+
+	be->blkif->blk_protocol = BLKIF_PROTOCOL_NATIVE;
+	err = xenbus_gather(XBT_NIL, dev->otherend, "protocol",
+			    "%63s", protocol, NULL);
+	if (err)
+		strcpy(protocol, "unspecified, assuming native");
+	else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_NATIVE))
+		be->blkif->blk_protocol = BLKIF_PROTOCOL_NATIVE;
+	else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_X86_32))
+		be->blkif->blk_protocol = BLKIF_PROTOCOL_X86_32;
+	else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_X86_64))
+		be->blkif->blk_protocol = BLKIF_PROTOCOL_X86_64;
+	else {
+		xenbus_dev_fatal(dev, err, "unknown fe protocol %s", protocol);
+		return -1;
+	}
+	printk(KERN_INFO
+	       "blkback: ring-ref %ld, event-channel %d, protocol %d (%s)\n",
+	       ring_ref, evtchn, be->blkif->blk_protocol, protocol);
+
+	/* Map the shared frame, irq etc. */
+	err = blkif_map(be->blkif, ring_ref, evtchn);
+	if (err) {
+		xenbus_dev_fatal(dev, err, "mapping ring-ref %lu port %u",
+				 ring_ref, evtchn);
+		return err;
+	}
+
+	return 0;
+}
+
+
+/* ** Driver Registration ** */
+
+
+static const struct xenbus_device_id blkback_ids[] = {
+	{ "vbd" },
+	{ "" }
+};
+
+
+static struct xenbus_driver blkback = {
+	.name = "vbd",
+	.owner = THIS_MODULE,
+	.ids = blkback_ids,
+	.probe = blkback_probe,
+	.remove = blkback_remove,
+	.otherend_changed = frontend_changed
+};
+
+
+void blkif_xenbus_init(void)
+{
+	xenbus_register_backend(&blkback);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/blkfront/blkfront.c linux-2.6.18-xen-3.3.0/drivers/xen/blkfront/blkfront.c
--- linux-2.6.18.8/drivers/xen/blkfront/blkfront.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/blkfront/blkfront.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,936 @@
+/******************************************************************************
+ * blkfront.c
+ * 
+ * XenLinux virtual block-device driver.
+ * 
+ * Copyright (c) 2003-2004, Keir Fraser & Steve Hand
+ * Modifications by Mark A. Williamson are (c) Intel Research Cambridge
+ * Copyright (c) 2004, Christian Limpach
+ * Copyright (c) 2004, Andrew Warfield
+ * Copyright (c) 2005, Christopher Clark
+ * Copyright (c) 2005, XenSource Ltd
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/version.h>
+#include "block.h"
+#include <linux/cdrom.h>
+#include <linux/sched.h>
+#include <linux/interrupt.h>
+#include <scsi/scsi.h>
+#include <xen/evtchn.h>
+#include <xen/xenbus.h>
+#include <xen/interface/grant_table.h>
+#include <xen/interface/io/protocols.h>
+#include <xen/gnttab.h>
+#include <asm/hypervisor.h>
+#include <asm/maddr.h>
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+#define BLKIF_STATE_DISCONNECTED 0
+#define BLKIF_STATE_CONNECTED    1
+#define BLKIF_STATE_SUSPENDED    2
+
+#define MAXIMUM_OUTSTANDING_BLOCK_REQS \
+    (BLKIF_MAX_SEGMENTS_PER_REQUEST * BLK_RING_SIZE)
+#define GRANT_INVALID_REF	0
+
+static void connect(struct blkfront_info *);
+static void blkfront_closing(struct xenbus_device *);
+static int blkfront_remove(struct xenbus_device *);
+static int talk_to_backend(struct xenbus_device *, struct blkfront_info *);
+static int setup_blkring(struct xenbus_device *, struct blkfront_info *);
+
+static void kick_pending_request_queues(struct blkfront_info *);
+
+static irqreturn_t blkif_int(int irq, void *dev_id, struct pt_regs *ptregs);
+static void blkif_restart_queue(void *arg);
+static void blkif_recover(struct blkfront_info *);
+static void blkif_completion(struct blk_shadow *);
+static void blkif_free(struct blkfront_info *, int);
+
+
+/**
+ * Entry point to this code when a new device is created.  Allocate the basic
+ * structures and the ring buffer for communication with the backend, and
+ * inform the backend of the appropriate details for those.  Switch to
+ * Initialised state.
+ */
+static int blkfront_probe(struct xenbus_device *dev,
+			  const struct xenbus_device_id *id)
+{
+	int err, vdevice, i;
+	struct blkfront_info *info;
+
+	/* FIXME: Use dynamic device id if this is not set. */
+	err = xenbus_scanf(XBT_NIL, dev->nodename,
+			   "virtual-device", "%i", &vdevice);
+	if (err != 1) {
+		/* go looking in the extended area instead */
+		err = xenbus_scanf(XBT_NIL, dev->nodename, "virtual-device-ext",
+			"%i", &vdevice);
+		if (err != 1) {
+			xenbus_dev_fatal(dev, err, "reading virtual-device");
+			return err;
+		}
+	}
+
+	info = kzalloc(sizeof(*info), GFP_KERNEL);
+	if (!info) {
+		xenbus_dev_fatal(dev, -ENOMEM, "allocating info structure");
+		return -ENOMEM;
+	}
+
+	info->xbdev = dev;
+	info->vdevice = vdevice;
+	info->connected = BLKIF_STATE_DISCONNECTED;
+	INIT_WORK(&info->work, blkif_restart_queue, (void *)info);
+
+	for (i = 0; i < BLK_RING_SIZE; i++)
+		info->shadow[i].req.id = i+1;
+	info->shadow[BLK_RING_SIZE-1].req.id = 0x0fffffff;
+
+	/* Front end dir is a number, which is used as the id. */
+	info->handle = simple_strtoul(strrchr(dev->nodename,'/')+1, NULL, 0);
+	dev->dev.driver_data = info;
+
+	err = talk_to_backend(dev, info);
+	if (err) {
+		kfree(info);
+		dev->dev.driver_data = NULL;
+		return err;
+	}
+
+	return 0;
+}
+
+
+/**
+ * We are reconnecting to the backend, due to a suspend/resume, or a backend
+ * driver restart.  We tear down our blkif structure and recreate it, but
+ * leave the device-layer structures intact so that this is transparent to the
+ * rest of the kernel.
+ */
+static int blkfront_resume(struct xenbus_device *dev)
+{
+	struct blkfront_info *info = dev->dev.driver_data;
+	int err;
+
+	DPRINTK("blkfront_resume: %s\n", dev->nodename);
+
+	blkif_free(info, info->connected == BLKIF_STATE_CONNECTED);
+
+	err = talk_to_backend(dev, info);
+	if (info->connected == BLKIF_STATE_SUSPENDED && !err)
+		blkif_recover(info);
+
+	return err;
+}
+
+
+/* Common code used when first setting up, and when resuming. */
+static int talk_to_backend(struct xenbus_device *dev,
+			   struct blkfront_info *info)
+{
+	const char *message = NULL;
+	struct xenbus_transaction xbt;
+	int err;
+
+	/* Create shared ring, alloc event channel. */
+	err = setup_blkring(dev, info);
+	if (err)
+		goto out;
+
+again:
+	err = xenbus_transaction_start(&xbt);
+	if (err) {
+		xenbus_dev_fatal(dev, err, "starting transaction");
+		goto destroy_blkring;
+	}
+
+	err = xenbus_printf(xbt, dev->nodename,
+			    "ring-ref","%u", info->ring_ref);
+	if (err) {
+		message = "writing ring-ref";
+		goto abort_transaction;
+	}
+	err = xenbus_printf(xbt, dev->nodename, "event-channel", "%u",
+			    irq_to_evtchn_port(info->irq));
+	if (err) {
+		message = "writing event-channel";
+		goto abort_transaction;
+	}
+	err = xenbus_printf(xbt, dev->nodename, "protocol", "%s",
+			    XEN_IO_PROTO_ABI_NATIVE);
+	if (err) {
+		message = "writing protocol";
+		goto abort_transaction;
+	}
+
+	err = xenbus_transaction_end(xbt, 0);
+	if (err) {
+		if (err == -EAGAIN)
+			goto again;
+		xenbus_dev_fatal(dev, err, "completing transaction");
+		goto destroy_blkring;
+	}
+
+	xenbus_switch_state(dev, XenbusStateInitialised);
+
+	return 0;
+
+ abort_transaction:
+	xenbus_transaction_end(xbt, 1);
+	if (message)
+		xenbus_dev_fatal(dev, err, "%s", message);
+ destroy_blkring:
+	blkif_free(info, 0);
+ out:
+	return err;
+}
+
+
+static int setup_blkring(struct xenbus_device *dev,
+			 struct blkfront_info *info)
+{
+	blkif_sring_t *sring;
+	int err;
+
+	info->ring_ref = GRANT_INVALID_REF;
+
+	sring = (blkif_sring_t *)__get_free_page(GFP_NOIO | __GFP_HIGH);
+	if (!sring) {
+		xenbus_dev_fatal(dev, -ENOMEM, "allocating shared ring");
+		return -ENOMEM;
+	}
+	SHARED_RING_INIT(sring);
+	FRONT_RING_INIT(&info->ring, sring, PAGE_SIZE);
+
+	err = xenbus_grant_ring(dev, virt_to_mfn(info->ring.sring));
+	if (err < 0) {
+		free_page((unsigned long)sring);
+		info->ring.sring = NULL;
+		goto fail;
+	}
+	info->ring_ref = err;
+
+	err = bind_listening_port_to_irqhandler(
+		dev->otherend_id, blkif_int, SA_SAMPLE_RANDOM, "blkif", info);
+	if (err <= 0) {
+		xenbus_dev_fatal(dev, err,
+				 "bind_listening_port_to_irqhandler");
+		goto fail;
+	}
+	info->irq = err;
+
+	return 0;
+fail:
+	blkif_free(info, 0);
+	return err;
+}
+
+
+/**
+ * Callback received when the backend's state changes.
+ */
+static void backend_changed(struct xenbus_device *dev,
+			    enum xenbus_state backend_state)
+{
+	struct blkfront_info *info = dev->dev.driver_data;
+	struct block_device *bd;
+
+	DPRINTK("blkfront:backend_changed.\n");
+
+	switch (backend_state) {
+	case XenbusStateInitialising:
+	case XenbusStateInitWait:
+	case XenbusStateInitialised:
+	case XenbusStateReconfiguring:
+	case XenbusStateReconfigured:
+	case XenbusStateUnknown:
+	case XenbusStateClosed:
+		break;
+
+	case XenbusStateConnected:
+		connect(info);
+		break;
+
+	case XenbusStateClosing:
+		bd = bdget(info->dev);
+		if (bd == NULL)
+			xenbus_dev_fatal(dev, -ENODEV, "bdget failed");
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,17)
+		down(&bd->bd_sem);
+#else
+		mutex_lock(&bd->bd_mutex);
+#endif
+		if (info->users > 0)
+			xenbus_dev_error(dev, -EBUSY,
+					 "Device in use; refusing to close");
+		else
+			blkfront_closing(dev);
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,17)
+		up(&bd->bd_sem);
+#else
+		mutex_unlock(&bd->bd_mutex);
+#endif
+		bdput(bd);
+		break;
+	}
+}
+
+
+/* ** Connection ** */
+
+
+/*
+ * Invoked when the backend is finally 'ready' (and has told produced
+ * the details about the physical device - #sectors, size, etc).
+ */
+static void connect(struct blkfront_info *info)
+{
+	unsigned long long sectors;
+	unsigned long sector_size;
+	unsigned int binfo;
+	int err;
+
+	if ((info->connected == BLKIF_STATE_CONNECTED) ||
+	    (info->connected == BLKIF_STATE_SUSPENDED) )
+		return;
+
+	DPRINTK("blkfront.c:connect:%s.\n", info->xbdev->otherend);
+
+	err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
+			    "sectors", "%Lu", &sectors,
+			    "info", "%u", &binfo,
+			    "sector-size", "%lu", &sector_size,
+			    NULL);
+	if (err) {
+		xenbus_dev_fatal(info->xbdev, err,
+				 "reading backend fields at %s",
+				 info->xbdev->otherend);
+		return;
+	}
+
+	err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
+			    "feature-barrier", "%lu", &info->feature_barrier,
+			    NULL);
+	if (err)
+		info->feature_barrier = 0;
+
+	err = xlvbd_add(sectors, info->vdevice, binfo, sector_size, info);
+	if (err) {
+		xenbus_dev_fatal(info->xbdev, err, "xlvbd_add at %s",
+				 info->xbdev->otherend);
+		return;
+	}
+
+	err = xlvbd_sysfs_addif(info);
+	if (err) {
+		xenbus_dev_fatal(info->xbdev, err, "xlvbd_sysfs_addif at %s",
+				 info->xbdev->otherend);
+		return;
+	}
+
+	(void)xenbus_switch_state(info->xbdev, XenbusStateConnected);
+
+	/* Kick pending requests. */
+	spin_lock_irq(&blkif_io_lock);
+	info->connected = BLKIF_STATE_CONNECTED;
+	kick_pending_request_queues(info);
+	spin_unlock_irq(&blkif_io_lock);
+
+	add_disk(info->gd);
+
+	info->is_ready = 1;
+}
+
+/**
+ * Handle the change of state of the backend to Closing.  We must delete our
+ * device-layer structures now, to ensure that writes are flushed through to
+ * the backend.  Once is this done, we can switch to Closed in
+ * acknowledgement.
+ */
+static void blkfront_closing(struct xenbus_device *dev)
+{
+	struct blkfront_info *info = dev->dev.driver_data;
+	unsigned long flags;
+
+	DPRINTK("blkfront_closing: %s removed\n", dev->nodename);
+
+	if (info->rq == NULL)
+		goto out;
+
+	spin_lock_irqsave(&blkif_io_lock, flags);
+	/* No more blkif_request(). */
+	blk_stop_queue(info->rq);
+	/* No more gnttab callback work. */
+	gnttab_cancel_free_callback(&info->callback);
+	spin_unlock_irqrestore(&blkif_io_lock, flags);
+
+	/* Flush gnttab callback work. Must be done with no locks held. */
+	flush_scheduled_work();
+
+	xlvbd_sysfs_delif(info);
+
+	xlvbd_del(info);
+
+ out:
+	xenbus_frontend_closed(dev);
+}
+
+
+static int blkfront_remove(struct xenbus_device *dev)
+{
+	struct blkfront_info *info = dev->dev.driver_data;
+
+	DPRINTK("blkfront_remove: %s removed\n", dev->nodename);
+
+	blkif_free(info, 0);
+
+	kfree(info);
+
+	return 0;
+}
+
+
+static inline int GET_ID_FROM_FREELIST(
+	struct blkfront_info *info)
+{
+	unsigned long free = info->shadow_free;
+	BUG_ON(free > BLK_RING_SIZE);
+	info->shadow_free = info->shadow[free].req.id;
+	info->shadow[free].req.id = 0x0fffffee; /* debug */
+	return free;
+}
+
+static inline void ADD_ID_TO_FREELIST(
+	struct blkfront_info *info, unsigned long id)
+{
+	info->shadow[id].req.id  = info->shadow_free;
+	info->shadow[id].request = 0;
+	info->shadow_free = id;
+}
+
+static inline void flush_requests(struct blkfront_info *info)
+{
+	int notify;
+
+	RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&info->ring, notify);
+
+	if (notify)
+		notify_remote_via_irq(info->irq);
+}
+
+static void kick_pending_request_queues(struct blkfront_info *info)
+{
+	if (!RING_FULL(&info->ring)) {
+		/* Re-enable calldowns. */
+		blk_start_queue(info->rq);
+		/* Kick things off immediately. */
+		do_blkif_request(info->rq);
+	}
+}
+
+static void blkif_restart_queue(void *arg)
+{
+	struct blkfront_info *info = (struct blkfront_info *)arg;
+	spin_lock_irq(&blkif_io_lock);
+	if (info->connected == BLKIF_STATE_CONNECTED)
+		kick_pending_request_queues(info);
+	spin_unlock_irq(&blkif_io_lock);
+}
+
+static void blkif_restart_queue_callback(void *arg)
+{
+	struct blkfront_info *info = (struct blkfront_info *)arg;
+	schedule_work(&info->work);
+}
+
+int blkif_open(struct inode *inode, struct file *filep)
+{
+	struct blkfront_info *info = inode->i_bdev->bd_disk->private_data;
+	info->users++;
+	return 0;
+}
+
+
+int blkif_release(struct inode *inode, struct file *filep)
+{
+	struct blkfront_info *info = inode->i_bdev->bd_disk->private_data;
+	info->users--;
+	if (info->users == 0) {
+		/* Check whether we have been instructed to close.  We will
+		   have ignored this request initially, as the device was
+		   still mounted. */
+		struct xenbus_device * dev = info->xbdev;
+		enum xenbus_state state = xenbus_read_driver_state(dev->otherend);
+
+		if (state == XenbusStateClosing && info->is_ready)
+			blkfront_closing(dev);
+	}
+	return 0;
+}
+
+
+int blkif_ioctl(struct inode *inode, struct file *filep,
+		unsigned command, unsigned long argument)
+{
+	int i;
+
+	DPRINTK_IOCTL("command: 0x%x, argument: 0x%lx, dev: 0x%04x\n",
+		      command, (long)argument, inode->i_rdev);
+
+	switch (command) {
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,16)
+	case HDIO_GETGEO: {
+		struct block_device *bd = inode->i_bdev;
+		struct hd_geometry geo;
+		int ret;
+
+                if (!argument)
+                        return -EINVAL;
+
+		geo.start = get_start_sect(bd);
+		ret = blkif_getgeo(bd, &geo);
+		if (ret)
+			return ret;
+
+		if (copy_to_user((struct hd_geometry __user *)argument, &geo,
+				 sizeof(geo)))
+                        return -EFAULT;
+
+                return 0;
+	}
+#endif
+	case CDROMMULTISESSION:
+		DPRINTK("FIXME: support multisession CDs later\n");
+		for (i = 0; i < sizeof(struct cdrom_multisession); i++)
+			if (put_user(0, (char __user *)(argument + i)))
+				return -EFAULT;
+		return 0;
+
+	case CDROM_GET_CAPABILITY: {
+		struct blkfront_info *info =
+			inode->i_bdev->bd_disk->private_data;
+		struct gendisk *gd = info->gd;
+		if (gd->flags & GENHD_FL_CD)
+			return 0;
+		return -EINVAL;
+	}
+	default:
+		/*printk(KERN_ALERT "ioctl %08x not supported by Xen blkdev\n",
+		  command);*/
+		return -EINVAL; /* same return as native Linux */
+	}
+
+	return 0;
+}
+
+
+int blkif_getgeo(struct block_device *bd, struct hd_geometry *hg)
+{
+	/* We don't have real geometry info, but let's at least return
+	   values consistent with the size of the device */
+	sector_t nsect = get_capacity(bd->bd_disk);
+	sector_t cylinders = nsect;
+
+	hg->heads = 0xff;
+	hg->sectors = 0x3f;
+	sector_div(cylinders, hg->heads * hg->sectors);
+	hg->cylinders = cylinders;
+	if ((sector_t)(hg->cylinders + 1) * hg->heads * hg->sectors < nsect)
+		hg->cylinders = 0xffff;
+	return 0;
+}
+
+
+/*
+ * blkif_queue_request
+ *
+ * request block io
+ *
+ * id: for guest use only.
+ * operation: BLKIF_OP_{READ,WRITE,PROBE}
+ * buffer: buffer to read/write into. this should be a
+ *   virtual address in the guest os.
+ */
+static int blkif_queue_request(struct request *req)
+{
+	struct blkfront_info *info = req->rq_disk->private_data;
+	unsigned long buffer_mfn;
+	blkif_request_t *ring_req;
+	struct bio *bio;
+	struct bio_vec *bvec;
+	int idx;
+	unsigned long id;
+	unsigned int fsect, lsect;
+	int ref;
+	grant_ref_t gref_head;
+
+	if (unlikely(info->connected != BLKIF_STATE_CONNECTED))
+		return 1;
+
+	if (gnttab_alloc_grant_references(
+		BLKIF_MAX_SEGMENTS_PER_REQUEST, &gref_head) < 0) {
+		gnttab_request_free_callback(
+			&info->callback,
+			blkif_restart_queue_callback,
+			info,
+			BLKIF_MAX_SEGMENTS_PER_REQUEST);
+		return 1;
+	}
+
+	/* Fill out a communications ring structure. */
+	ring_req = RING_GET_REQUEST(&info->ring, info->ring.req_prod_pvt);
+	id = GET_ID_FROM_FREELIST(info);
+	info->shadow[id].request = (unsigned long)req;
+
+	ring_req->id = id;
+	ring_req->sector_number = (blkif_sector_t)req->sector;
+	ring_req->handle = info->handle;
+
+	ring_req->operation = rq_data_dir(req) ?
+		BLKIF_OP_WRITE : BLKIF_OP_READ;
+	if (blk_barrier_rq(req))
+		ring_req->operation = BLKIF_OP_WRITE_BARRIER;
+
+	ring_req->nr_segments = 0;
+	rq_for_each_bio (bio, req) {
+		bio_for_each_segment (bvec, bio, idx) {
+			BUG_ON(ring_req->nr_segments
+			       == BLKIF_MAX_SEGMENTS_PER_REQUEST);
+			buffer_mfn = page_to_phys(bvec->bv_page) >> PAGE_SHIFT;
+			fsect = bvec->bv_offset >> 9;
+			lsect = fsect + (bvec->bv_len >> 9) - 1;
+			/* install a grant reference. */
+			ref = gnttab_claim_grant_reference(&gref_head);
+			BUG_ON(ref == -ENOSPC);
+
+			gnttab_grant_foreign_access_ref(
+				ref,
+				info->xbdev->otherend_id,
+				buffer_mfn,
+				rq_data_dir(req) ? GTF_readonly : 0 );
+
+			info->shadow[id].frame[ring_req->nr_segments] =
+				mfn_to_pfn(buffer_mfn);
+
+			ring_req->seg[ring_req->nr_segments] =
+				(struct blkif_request_segment) {
+					.gref       = ref,
+					.first_sect = fsect,
+					.last_sect  = lsect };
+
+			ring_req->nr_segments++;
+		}
+	}
+
+	info->ring.req_prod_pvt++;
+
+	/* Keep a private copy so we can reissue requests when recovering. */
+	info->shadow[id].req = *ring_req;
+
+	gnttab_free_grant_references(gref_head);
+
+	return 0;
+}
+
+/*
+ * do_blkif_request
+ *  read a block; request is in a request queue
+ */
+void do_blkif_request(request_queue_t *rq)
+{
+	struct blkfront_info *info = NULL;
+	struct request *req;
+	int queued;
+
+	DPRINTK("Entered do_blkif_request\n");
+
+	queued = 0;
+
+	while ((req = elv_next_request(rq)) != NULL) {
+		info = req->rq_disk->private_data;
+		if (!blk_fs_request(req)) {
+			end_request(req, 0);
+			continue;
+		}
+
+		if (RING_FULL(&info->ring))
+			goto wait;
+
+		DPRINTK("do_blk_req %p: cmd %p, sec %llx, "
+			"(%u/%li) buffer:%p [%s]\n",
+			req, req->cmd, (long long)req->sector,
+			req->current_nr_sectors,
+			req->nr_sectors, req->buffer,
+			rq_data_dir(req) ? "write" : "read");
+
+
+		blkdev_dequeue_request(req);
+		if (blkif_queue_request(req)) {
+			blk_requeue_request(rq, req);
+		wait:
+			/* Avoid pointless unplugs. */
+			blk_stop_queue(rq);
+			break;
+		}
+
+		queued++;
+	}
+
+	if (queued != 0)
+		flush_requests(info);
+}
+
+
+static irqreturn_t blkif_int(int irq, void *dev_id, struct pt_regs *ptregs)
+{
+	struct request *req;
+	blkif_response_t *bret;
+	RING_IDX i, rp;
+	unsigned long flags;
+	struct blkfront_info *info = (struct blkfront_info *)dev_id;
+	int uptodate;
+
+	spin_lock_irqsave(&blkif_io_lock, flags);
+
+	if (unlikely(info->connected != BLKIF_STATE_CONNECTED)) {
+		spin_unlock_irqrestore(&blkif_io_lock, flags);
+		return IRQ_HANDLED;
+	}
+
+ again:
+	rp = info->ring.sring->rsp_prod;
+	rmb(); /* Ensure we see queued responses up to 'rp'. */
+
+	for (i = info->ring.rsp_cons; i != rp; i++) {
+		unsigned long id;
+		int ret;
+
+		bret = RING_GET_RESPONSE(&info->ring, i);
+		id   = bret->id;
+		req  = (struct request *)info->shadow[id].request;
+
+		blkif_completion(&info->shadow[id]);
+
+		ADD_ID_TO_FREELIST(info, id);
+
+		uptodate = (bret->status == BLKIF_RSP_OKAY);
+		switch (bret->operation) {
+		case BLKIF_OP_WRITE_BARRIER:
+			if (unlikely(bret->status == BLKIF_RSP_EOPNOTSUPP)) {
+				printk("blkfront: %s: write barrier op failed\n",
+				       info->gd->disk_name);
+				uptodate = -EOPNOTSUPP;
+				info->feature_barrier = 0;
+			        xlvbd_barrier(info);
+			}
+			/* fall through */
+		case BLKIF_OP_READ:
+		case BLKIF_OP_WRITE:
+			if (unlikely(bret->status != BLKIF_RSP_OKAY))
+				DPRINTK("Bad return from blkdev data "
+					"request: %x\n", bret->status);
+
+			ret = end_that_request_first(req, uptodate,
+				req->hard_nr_sectors);
+			BUG_ON(ret);
+			end_that_request_last(req, uptodate);
+			break;
+		default:
+			BUG();
+		}
+	}
+
+	info->ring.rsp_cons = i;
+
+	if (i != info->ring.req_prod_pvt) {
+		int more_to_do;
+		RING_FINAL_CHECK_FOR_RESPONSES(&info->ring, more_to_do);
+		if (more_to_do)
+			goto again;
+	} else
+		info->ring.sring->rsp_event = i + 1;
+
+	kick_pending_request_queues(info);
+
+	spin_unlock_irqrestore(&blkif_io_lock, flags);
+
+	return IRQ_HANDLED;
+}
+
+static void blkif_free(struct blkfront_info *info, int suspend)
+{
+	/* Prevent new requests being issued until we fix things up. */
+	spin_lock_irq(&blkif_io_lock);
+	info->connected = suspend ?
+		BLKIF_STATE_SUSPENDED : BLKIF_STATE_DISCONNECTED;
+	/* No more blkif_request(). */
+	if (info->rq)
+		blk_stop_queue(info->rq);
+	/* No more gnttab callback work. */
+	gnttab_cancel_free_callback(&info->callback);
+	spin_unlock_irq(&blkif_io_lock);
+
+	/* Flush gnttab callback work. Must be done with no locks held. */
+	flush_scheduled_work();
+
+	/* Free resources associated with old device channel. */
+	if (info->ring_ref != GRANT_INVALID_REF) {
+		gnttab_end_foreign_access(info->ring_ref, 
+					  (unsigned long)info->ring.sring);
+		info->ring_ref = GRANT_INVALID_REF;
+		info->ring.sring = NULL;
+	}
+	if (info->irq)
+		unbind_from_irqhandler(info->irq, info);
+	info->irq = 0;
+}
+
+static void blkif_completion(struct blk_shadow *s)
+{
+	int i;
+	for (i = 0; i < s->req.nr_segments; i++)
+		gnttab_end_foreign_access(s->req.seg[i].gref, 0UL);
+}
+
+static void blkif_recover(struct blkfront_info *info)
+{
+	int i;
+	blkif_request_t *req;
+	struct blk_shadow *copy;
+	int j;
+
+	/* Stage 1: Make a safe copy of the shadow state. */
+	copy = kmalloc(sizeof(info->shadow), GFP_NOIO | __GFP_NOFAIL | __GFP_HIGH);
+	memcpy(copy, info->shadow, sizeof(info->shadow));
+
+	/* Stage 2: Set up free list. */
+	memset(&info->shadow, 0, sizeof(info->shadow));
+	for (i = 0; i < BLK_RING_SIZE; i++)
+		info->shadow[i].req.id = i+1;
+	info->shadow_free = info->ring.req_prod_pvt;
+	info->shadow[BLK_RING_SIZE-1].req.id = 0x0fffffff;
+
+	/* Stage 3: Find pending requests and requeue them. */
+	for (i = 0; i < BLK_RING_SIZE; i++) {
+		/* Not in use? */
+		if (copy[i].request == 0)
+			continue;
+
+		/* Grab a request slot and copy shadow state into it. */
+		req = RING_GET_REQUEST(
+			&info->ring, info->ring.req_prod_pvt);
+		*req = copy[i].req;
+
+		/* We get a new request id, and must reset the shadow state. */
+		req->id = GET_ID_FROM_FREELIST(info);
+		memcpy(&info->shadow[req->id], &copy[i], sizeof(copy[i]));
+
+		/* Rewrite any grant references invalidated by susp/resume. */
+		for (j = 0; j < req->nr_segments; j++)
+			gnttab_grant_foreign_access_ref(
+				req->seg[j].gref,
+				info->xbdev->otherend_id,
+				pfn_to_mfn(info->shadow[req->id].frame[j]),
+				rq_data_dir((struct request *)
+					    info->shadow[req->id].request) ?
+				GTF_readonly : 0);
+		info->shadow[req->id].req = *req;
+
+		info->ring.req_prod_pvt++;
+	}
+
+	kfree(copy);
+
+	(void)xenbus_switch_state(info->xbdev, XenbusStateConnected);
+
+	spin_lock_irq(&blkif_io_lock);
+
+	/* Now safe for us to use the shared ring */
+	info->connected = BLKIF_STATE_CONNECTED;
+
+	/* Send off requeued requests */
+	flush_requests(info);
+
+	/* Kick any other new requests queued since we resumed */
+	kick_pending_request_queues(info);
+
+	spin_unlock_irq(&blkif_io_lock);
+}
+
+int blkfront_is_ready(struct xenbus_device *dev)
+{
+	struct blkfront_info *info = dev->dev.driver_data;
+
+	return info->is_ready;
+}
+
+
+/* ** Driver Registration ** */
+
+
+static const struct xenbus_device_id blkfront_ids[] = {
+	{ "vbd" },
+	{ "" }
+};
+MODULE_ALIAS("xen:vbd");
+
+static struct xenbus_driver blkfront = {
+	.name = "vbd",
+	.owner = THIS_MODULE,
+	.ids = blkfront_ids,
+	.probe = blkfront_probe,
+	.remove = blkfront_remove,
+	.resume = blkfront_resume,
+	.otherend_changed = backend_changed,
+	.is_ready = blkfront_is_ready,
+};
+
+
+static int __init xlblk_init(void)
+{
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	return xenbus_register_frontend(&blkfront);
+}
+module_init(xlblk_init);
+
+
+static void __exit xlblk_exit(void)
+{
+	return xenbus_unregister_driver(&blkfront);
+}
+module_exit(xlblk_exit);
+
+MODULE_LICENSE("Dual BSD/GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/blkfront/block.h linux-2.6.18-xen-3.3.0/drivers/xen/blkfront/block.h
--- linux-2.6.18.8/drivers/xen/blkfront/block.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/blkfront/block.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,158 @@
+/******************************************************************************
+ * block.h
+ * 
+ * Shared definitions between all levels of XenLinux Virtual block devices.
+ * 
+ * Copyright (c) 2003-2004, Keir Fraser & Steve Hand
+ * Modifications by Mark A. Williamson are (c) Intel Research Cambridge
+ * Copyright (c) 2004-2005, Christian Limpach
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_DRIVERS_BLOCK_H__
+#define __XEN_DRIVERS_BLOCK_H__
+
+#include <linux/version.h>
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/string.h>
+#include <linux/errno.h>
+#include <linux/fs.h>
+#include <linux/hdreg.h>
+#include <linux/blkdev.h>
+#include <linux/major.h>
+#include <asm/hypervisor.h>
+#include <xen/xenbus.h>
+#include <xen/gnttab.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/io/blkif.h>
+#include <xen/interface/io/ring.h>
+#include <asm/io.h>
+#include <asm/atomic.h>
+#include <asm/uaccess.h>
+
+#define DPRINTK(_f, _a...) pr_debug(_f, ## _a)
+
+#if 0
+#define DPRINTK_IOCTL(_f, _a...) printk(KERN_ALERT _f, ## _a)
+#else
+#define DPRINTK_IOCTL(_f, _a...) ((void)0)
+#endif
+
+struct xlbd_type_info
+{
+	int partn_shift;
+	int disks_per_major;
+	char *devname;
+	char *diskname;
+};
+
+struct xlbd_major_info
+{
+	int major;
+	int index;
+	int usage;
+	struct xlbd_type_info *type;
+};
+
+struct blk_shadow {
+	blkif_request_t req;
+	unsigned long request;
+	unsigned long frame[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+};
+
+#define BLK_RING_SIZE __RING_SIZE((blkif_sring_t *)0, PAGE_SIZE)
+
+/*
+ * We have one of these per vbd, whether ide, scsi or 'other'.  They
+ * hang in private_data off the gendisk structure. We may end up
+ * putting all kinds of interesting stuff here :-)
+ */
+struct blkfront_info
+{
+	struct xenbus_device *xbdev;
+	dev_t dev;
+ 	struct gendisk *gd;
+	int vdevice;
+	blkif_vdev_t handle;
+	int connected;
+	int ring_ref;
+	blkif_front_ring_t ring;
+	unsigned int irq;
+	struct xlbd_major_info *mi;
+	request_queue_t *rq;
+	struct work_struct work;
+	struct gnttab_free_callback callback;
+	struct blk_shadow shadow[BLK_RING_SIZE];
+	unsigned long shadow_free;
+	int feature_barrier;
+	int is_ready;
+
+	/**
+	 * The number of people holding this device open.  We won't allow a
+	 * hot-unplug unless this is 0.
+	 */
+	int users;
+};
+
+extern spinlock_t blkif_io_lock;
+
+extern int blkif_open(struct inode *inode, struct file *filep);
+extern int blkif_release(struct inode *inode, struct file *filep);
+extern int blkif_ioctl(struct inode *inode, struct file *filep,
+		       unsigned command, unsigned long argument);
+extern int blkif_getgeo(struct block_device *, struct hd_geometry *);
+extern int blkif_check(dev_t dev);
+extern int blkif_revalidate(dev_t dev);
+extern void do_blkif_request (request_queue_t *rq);
+
+/* Virtual block-device subsystem. */
+/* Note that xlvbd_add doesn't call add_disk for you: you're expected
+   to call add_disk on info->gd once the disk is properly connected
+   up. */
+int xlvbd_add(blkif_sector_t capacity, int device,
+	      u16 vdisk_info, u16 sector_size, struct blkfront_info *info);
+void xlvbd_del(struct blkfront_info *info);
+int xlvbd_barrier(struct blkfront_info *info);
+
+#ifdef CONFIG_SYSFS
+int xlvbd_sysfs_addif(struct blkfront_info *info);
+void xlvbd_sysfs_delif(struct blkfront_info *info);
+#else
+static inline int xlvbd_sysfs_addif(struct blkfront_info *info)
+{
+	return 0;
+}
+
+static inline void xlvbd_sysfs_delif(struct blkfront_info *info)
+{
+	;
+}
+#endif
+
+#endif /* __XEN_DRIVERS_BLOCK_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/blkfront/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/blkfront/Makefile
--- linux-2.6.18.8/drivers/xen/blkfront/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/blkfront/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,5 @@
+
+obj-$(CONFIG_XEN_BLKDEV_FRONTEND)	:= xenblk.o
+
+xenblk-objs := blkfront.o vbd.o
+
diff -rpuN linux-2.6.18.8/drivers/xen/blkfront/vbd.c linux-2.6.18-xen-3.3.0/drivers/xen/blkfront/vbd.c
--- linux-2.6.18.8/drivers/xen/blkfront/vbd.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/blkfront/vbd.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,460 @@
+/******************************************************************************
+ * vbd.c
+ * 
+ * XenLinux virtual block-device driver (xvd).
+ * 
+ * Copyright (c) 2003-2004, Keir Fraser & Steve Hand
+ * Modifications by Mark A. Williamson are (c) Intel Research Cambridge
+ * Copyright (c) 2004-2005, Christian Limpach
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include "block.h"
+#include <linux/blkdev.h>
+#include <linux/list.h>
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+#define BLKIF_MAJOR(dev) ((dev)>>8)
+#define BLKIF_MINOR(dev) ((dev) & 0xff)
+
+#define EXT_SHIFT 28
+#define EXTENDED (1<<EXT_SHIFT)
+#define VDEV_IS_EXTENDED(dev) ((dev)&(EXTENDED))
+#define BLKIF_MINOR_EXT(dev) ((dev)&(~EXTENDED))
+
+/*
+ * For convenience we distinguish between ide, scsi and 'other' (i.e.,
+ * potentially combinations of the two) in the naming scheme and in a few other
+ * places.
+ */
+
+#define NUM_IDE_MAJORS 10
+#define NUM_SCSI_MAJORS 17
+#define NUM_VBD_MAJORS 2
+
+static struct xlbd_type_info xlbd_ide_type = {
+	.partn_shift = 6,
+	.disks_per_major = 2,
+	.devname = "ide",
+	.diskname = "hd",
+};
+
+static struct xlbd_type_info xlbd_scsi_type = {
+	.partn_shift = 4,
+	.disks_per_major = 16,
+	.devname = "sd",
+	.diskname = "sd",
+};
+
+static struct xlbd_type_info xlbd_vbd_type = {
+	.partn_shift = 4,
+	.disks_per_major = 16,
+	.devname = "xvd",
+	.diskname = "xvd",
+};
+
+static struct xlbd_type_info xlbd_vbd_type_ext = {
+	.partn_shift = 8,
+	.disks_per_major = 256,
+	.devname = "xvd",
+	.diskname = "xvd",
+};
+
+static struct xlbd_major_info *major_info[NUM_IDE_MAJORS + NUM_SCSI_MAJORS +
+					 NUM_VBD_MAJORS];
+
+#define XLBD_MAJOR_IDE_START	0
+#define XLBD_MAJOR_SCSI_START	(NUM_IDE_MAJORS)
+#define XLBD_MAJOR_VBD_START	(NUM_IDE_MAJORS + NUM_SCSI_MAJORS)
+
+#define XLBD_MAJOR_IDE_RANGE	XLBD_MAJOR_IDE_START ... XLBD_MAJOR_SCSI_START - 1
+#define XLBD_MAJOR_SCSI_RANGE	XLBD_MAJOR_SCSI_START ... XLBD_MAJOR_VBD_START - 1
+#define XLBD_MAJOR_VBD_RANGE	XLBD_MAJOR_VBD_START ... XLBD_MAJOR_VBD_START + NUM_VBD_MAJORS - 1
+
+static struct block_device_operations xlvbd_block_fops =
+{
+	.owner = THIS_MODULE,
+	.open = blkif_open,
+	.release = blkif_release,
+	.ioctl  = blkif_ioctl,
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,16)
+	.getgeo = blkif_getgeo
+#endif
+};
+
+DEFINE_SPINLOCK(blkif_io_lock);
+
+static struct xlbd_major_info *
+xlbd_alloc_major_info(int major, int minor, int index)
+{
+	struct xlbd_major_info *ptr;
+	int do_register;
+
+	ptr = kzalloc(sizeof(struct xlbd_major_info), GFP_KERNEL);
+	if (ptr == NULL)
+		return NULL;
+
+	ptr->major = major;
+	do_register = 1;
+
+	switch (index) {
+	case XLBD_MAJOR_IDE_RANGE:
+		ptr->type = &xlbd_ide_type;
+		ptr->index = index - XLBD_MAJOR_IDE_START;
+		break;
+	case XLBD_MAJOR_SCSI_RANGE:
+		ptr->type = &xlbd_scsi_type;
+		ptr->index = index - XLBD_MAJOR_SCSI_START;
+		break;
+	case XLBD_MAJOR_VBD_RANGE:
+		ptr->index = 0;
+		if ((index - XLBD_MAJOR_VBD_START) == 0)
+			ptr->type = &xlbd_vbd_type;
+		else
+			ptr->type = &xlbd_vbd_type_ext;
+
+		/* 
+		 * if someone already registered block major 202,
+		 * don't try to register it again
+		 */
+		if (major_info[XLBD_MAJOR_VBD_START] != NULL)
+			do_register = 0;
+		break;
+	}
+
+	if (do_register) {
+		if (register_blkdev(ptr->major, ptr->type->devname)) {
+			kfree(ptr);
+			return NULL;
+		}
+
+		printk("xen-vbd: registered block device major %i\n", ptr->major);
+	}
+
+	major_info[index] = ptr;
+	return ptr;
+}
+
+static struct xlbd_major_info *
+xlbd_get_major_info(int major, int minor, int vdevice)
+{
+	struct xlbd_major_info *mi;
+	int index;
+
+	switch (major) {
+	case IDE0_MAJOR: index = 0; break;
+	case IDE1_MAJOR: index = 1; break;
+	case IDE2_MAJOR: index = 2; break;
+	case IDE3_MAJOR: index = 3; break;
+	case IDE4_MAJOR: index = 4; break;
+	case IDE5_MAJOR: index = 5; break;
+	case IDE6_MAJOR: index = 6; break;
+	case IDE7_MAJOR: index = 7; break;
+	case IDE8_MAJOR: index = 8; break;
+	case IDE9_MAJOR: index = 9; break;
+	case SCSI_DISK0_MAJOR: index = 10; break;
+	case SCSI_DISK1_MAJOR ... SCSI_DISK7_MAJOR:
+		index = 11 + major - SCSI_DISK1_MAJOR;
+		break;
+        case SCSI_DISK8_MAJOR ... SCSI_DISK15_MAJOR:
+                index = 18 + major - SCSI_DISK8_MAJOR;
+                break;
+        case SCSI_CDROM_MAJOR: index = 26; break;
+        default:
+		if (!VDEV_IS_EXTENDED(vdevice))
+			index = 27;
+		else
+			index = 28;
+		break;
+	}
+
+	mi = ((major_info[index] != NULL) ? major_info[index] :
+	      xlbd_alloc_major_info(major, minor, index));
+	if (mi)
+		mi->usage++;
+	return mi;
+}
+
+static void
+xlbd_put_major_info(struct xlbd_major_info *mi)
+{
+	mi->usage--;
+	/* XXX: release major if 0 */
+}
+
+static int
+xlvbd_init_blk_queue(struct gendisk *gd, u16 sector_size)
+{
+	request_queue_t *rq;
+
+	rq = blk_init_queue(do_blkif_request, &blkif_io_lock);
+	if (rq == NULL)
+		return -1;
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,10)
+	elevator_init(rq, "noop");
+#else
+	elevator_init(rq, &elevator_noop);
+#endif
+
+	/* Hard sector size and max sectors impersonate the equiv. hardware. */
+	blk_queue_hardsect_size(rq, sector_size);
+	blk_queue_max_sectors(rq, 512);
+
+	/* Each segment in a request is up to an aligned page in size. */
+	blk_queue_segment_boundary(rq, PAGE_SIZE - 1);
+	blk_queue_max_segment_size(rq, PAGE_SIZE);
+
+	/* Ensure a merged request will fit in a single I/O ring slot. */
+	blk_queue_max_phys_segments(rq, BLKIF_MAX_SEGMENTS_PER_REQUEST);
+	blk_queue_max_hw_segments(rq, BLKIF_MAX_SEGMENTS_PER_REQUEST);
+
+	/* Make sure buffer addresses are sector-aligned. */
+	blk_queue_dma_alignment(rq, 511);
+
+	/* Make sure we don't use bounce buffers. */
+	blk_queue_bounce_limit(rq, BLK_BOUNCE_ANY);
+
+	gd->queue = rq;
+
+	return 0;
+}
+
+static int
+xlvbd_alloc_gendisk(int major, int minor, blkif_sector_t capacity, int vdevice,
+		    u16 vdisk_info, u16 sector_size,
+		    struct blkfront_info *info)
+{
+	struct gendisk *gd;
+	struct xlbd_major_info *mi;
+	int nr_minors = 1;
+	int err = -ENODEV;
+	unsigned int offset;
+
+	BUG_ON(info->gd != NULL);
+	BUG_ON(info->mi != NULL);
+	BUG_ON(info->rq != NULL);
+
+	mi = xlbd_get_major_info(major, minor, vdevice);
+	if (mi == NULL)
+		goto out;
+	info->mi = mi;
+
+	if ((minor & ((1 << mi->type->partn_shift) - 1)) == 0)
+		nr_minors = 1 << mi->type->partn_shift;
+
+	gd = alloc_disk(nr_minors);
+	if (gd == NULL)
+		goto out;
+
+	offset =  mi->index * mi->type->disks_per_major +
+			(minor >> mi->type->partn_shift);
+	if (nr_minors > 1) {
+		if (offset < 26) {
+			sprintf(gd->disk_name, "%s%c",
+				 mi->type->diskname, 'a' + offset );
+		}
+		else {
+			sprintf(gd->disk_name, "%s%c%c",
+				mi->type->diskname,
+				'a' + ((offset/26)-1), 'a' + (offset%26) );
+		}
+	}
+	else {
+		if (offset < 26) {
+			sprintf(gd->disk_name, "%s%c%d",
+				mi->type->diskname,
+				'a' + offset,
+				minor & ((1 << mi->type->partn_shift) - 1));
+		}
+		else {
+			sprintf(gd->disk_name, "%s%c%c%d",
+				mi->type->diskname,
+				'a' + ((offset/26)-1), 'a' + (offset%26),
+				minor & ((1 << mi->type->partn_shift) - 1));
+		}
+	}
+
+	gd->major = mi->major;
+	gd->first_minor = minor;
+	gd->fops = &xlvbd_block_fops;
+	gd->private_data = info;
+	gd->driverfs_dev = &(info->xbdev->dev);
+	set_capacity(gd, capacity);
+
+	if (xlvbd_init_blk_queue(gd, sector_size)) {
+		del_gendisk(gd);
+		goto out;
+	}
+
+	info->rq = gd->queue;
+	info->gd = gd;
+
+	if (info->feature_barrier)
+		xlvbd_barrier(info);
+
+	if (vdisk_info & VDISK_READONLY)
+		set_disk_ro(gd, 1);
+
+	if (vdisk_info & VDISK_REMOVABLE)
+		gd->flags |= GENHD_FL_REMOVABLE;
+
+	if (vdisk_info & VDISK_CDROM)
+		gd->flags |= GENHD_FL_CD;
+
+	return 0;
+
+ out:
+	if (mi)
+		xlbd_put_major_info(mi);
+	info->mi = NULL;
+	return err;
+}
+
+int
+xlvbd_add(blkif_sector_t capacity, int vdevice, u16 vdisk_info,
+	  u16 sector_size, struct blkfront_info *info)
+{
+	struct block_device *bd;
+	int err = 0;
+	int major, minor;
+
+	if ((vdevice>>EXT_SHIFT) > 1) {
+		/* this is above the extended range; something is wrong */
+		printk(KERN_WARNING "blkfront: vdevice 0x%x is above the extended range; ignoring\n", vdevice);
+		return -ENODEV;
+	}
+
+	if (!VDEV_IS_EXTENDED(vdevice)) {
+		major = BLKIF_MAJOR(vdevice);
+		minor = BLKIF_MINOR(vdevice);
+	}
+	else {
+		major = 202;
+		minor = BLKIF_MINOR_EXT(vdevice);
+	}
+
+	info->dev = MKDEV(major, minor);
+	bd = bdget(info->dev);
+	if (bd == NULL)
+		return -ENODEV;
+
+	err = xlvbd_alloc_gendisk(major, minor, capacity, vdevice, vdisk_info,
+				  sector_size, info);
+
+	bdput(bd);
+	return err;
+}
+
+void
+xlvbd_del(struct blkfront_info *info)
+{
+	if (info->mi == NULL)
+		return;
+
+	BUG_ON(info->gd == NULL);
+	del_gendisk(info->gd);
+	put_disk(info->gd);
+	info->gd = NULL;
+
+	xlbd_put_major_info(info->mi);
+	info->mi = NULL;
+
+	BUG_ON(info->rq == NULL);
+	blk_cleanup_queue(info->rq);
+	info->rq = NULL;
+}
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,16)
+int
+xlvbd_barrier(struct blkfront_info *info)
+{
+	int err;
+
+	err = blk_queue_ordered(info->rq,
+		info->feature_barrier ? QUEUE_ORDERED_DRAIN : QUEUE_ORDERED_NONE, NULL);
+	if (err)
+		return err;
+	printk(KERN_INFO "blkfront: %s: barriers %s\n",
+	       info->gd->disk_name, info->feature_barrier ? "enabled" : "disabled");
+	return 0;
+}
+#else
+int
+xlvbd_barrier(struct blkfront_info *info)
+{
+	printk(KERN_INFO "blkfront: %s: barriers disabled\n", info->gd->disk_name);
+	return -ENOSYS;
+}
+#endif
+
+#ifdef CONFIG_SYSFS
+static ssize_t show_media(struct device *dev,
+		                  struct device_attribute *attr, char *buf)
+{
+	struct xenbus_device *xendev = to_xenbus_device(dev);
+	struct blkfront_info *info = xendev->dev.driver_data;
+
+	if (info->gd->flags & GENHD_FL_CD)
+		return sprintf(buf, "cdrom\n");
+	return sprintf(buf, "disk\n");
+}
+
+static struct device_attribute xlvbd_attrs[] = {
+	__ATTR(media, S_IRUGO, show_media, NULL),
+};
+
+int xlvbd_sysfs_addif(struct blkfront_info *info)
+{
+	int i;
+	int error = 0;
+
+	for (i = 0; i < ARRAY_SIZE(xlvbd_attrs); i++) {
+		error = device_create_file(info->gd->driverfs_dev,
+				&xlvbd_attrs[i]);
+		if (error)
+			goto fail;
+	}
+	return 0;
+
+fail:
+	while (--i >= 0)
+		device_remove_file(info->gd->driverfs_dev, &xlvbd_attrs[i]);
+	return error;
+}
+
+void xlvbd_sysfs_delif(struct blkfront_info *info)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(xlvbd_attrs); i++)
+		device_remove_file(info->gd->driverfs_dev, &xlvbd_attrs[i]);
+}
+
+#endif /* CONFIG_SYSFS */
diff -rpuN linux-2.6.18.8/drivers/xen/blktap/blktap.c linux-2.6.18-xen-3.3.0/drivers/xen/blktap/blktap.c
--- linux-2.6.18.8/drivers/xen/blktap/blktap.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/blktap/blktap.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1681 @@
+/******************************************************************************
+ * drivers/xen/blktap/blktap.c
+ * 
+ * Back-end driver for user level virtual block devices. This portion of the
+ * driver exports a 'unified' block-device interface that can be accessed
+ * by any operating system that implements a compatible front end. Requests
+ * are remapped to a user-space memory region.
+ *
+ * Based on the blkback driver code.
+ * 
+ * Copyright (c) 2004-2005, Andrew Warfield and Julian Chesterfield
+ *
+ * Clean ups and fix ups:
+ *    Copyright (c) 2006, Steven Rostedt - Red Hat, Inc.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/spinlock.h>
+#include <linux/kthread.h>
+#include <linux/list.h>
+#include <asm/hypervisor.h>
+#include "common.h"
+#include <xen/balloon.h>
+#include <xen/driver_util.h>
+#include <linux/kernel.h>
+#include <linux/fs.h>
+#include <linux/mm.h>
+#include <linux/errno.h>
+#include <linux/major.h>
+#include <linux/gfp.h>
+#include <linux/poll.h>
+#include <linux/delay.h>
+#include <asm/tlbflush.h>
+
+#define MAX_TAP_DEV 256     /*the maximum number of tapdisk ring devices    */
+#define MAX_DEV_NAME 100    /*the max tapdisk ring device name e.g. blktap0 */
+
+/*
+ * The maximum number of requests that can be outstanding at any time
+ * is determined by 
+ *
+ *   [mmap_alloc * MAX_PENDING_REQS * BLKIF_MAX_SEGMENTS_PER_REQUEST] 
+ *
+ * where mmap_alloc < MAX_DYNAMIC_MEM.
+ *
+ * TODO:
+ * mmap_alloc is initialised to 2 and should be adjustable on the fly via
+ * sysfs.
+ */
+#define BLK_RING_SIZE		__RING_SIZE((blkif_sring_t *)0, PAGE_SIZE)
+#define MAX_DYNAMIC_MEM		BLK_RING_SIZE
+#define MAX_PENDING_REQS	BLK_RING_SIZE
+#define MMAP_PAGES (MAX_PENDING_REQS * BLKIF_MAX_SEGMENTS_PER_REQUEST)
+#define MMAP_VADDR(_start, _req,_seg)                                   \
+        (_start +                                                       \
+         ((_req) * BLKIF_MAX_SEGMENTS_PER_REQUEST * PAGE_SIZE) +        \
+         ((_seg) * PAGE_SIZE))
+static int blkif_reqs = MAX_PENDING_REQS;
+static int mmap_pages = MMAP_PAGES;
+
+#define RING_PAGES 1 /* BLKTAP - immediately before the mmap area, we
+		      * have a bunch of pages reserved for shared
+		      * memory rings.
+		      */
+
+/*Data struct handed back to userspace for tapdisk device to VBD mapping*/
+typedef struct domid_translate {
+	unsigned short domid;
+	unsigned short busid;
+} domid_translate_t ;
+
+typedef struct domid_translate_ext {
+	unsigned short domid;
+	u32 busid;
+} domid_translate_ext_t ;
+
+/*Data struct associated with each of the tapdisk devices*/
+typedef struct tap_blkif {
+	struct vm_area_struct *vma;   /*Shared memory area                   */
+	unsigned long rings_vstart;   /*Kernel memory mapping                */
+	unsigned long user_vstart;    /*User memory mapping                  */
+	unsigned long dev_inuse;      /*One process opens device at a time.  */
+	unsigned long dev_pending;    /*In process of being opened           */
+	unsigned long ring_ok;        /*make this ring->state                */
+	blkif_front_ring_t ufe_ring;  /*Rings up to user space.              */
+	wait_queue_head_t wait;       /*for poll                             */
+	unsigned long mode;           /*current switching mode               */
+	int minor;                    /*Minor number for tapdisk device      */
+	pid_t pid;                    /*tapdisk process id                   */
+	enum { RUNNING, CLEANSHUTDOWN } status; /*Detect a clean userspace 
+						  shutdown                   */
+	unsigned long *idx_map;       /*Record the user ring id to kern 
+					[req id, idx] tuple                  */
+	blkif_t *blkif;               /*Associate blkif with tapdev          */
+	struct domid_translate_ext trans; /*Translation from domid to bus.   */
+} tap_blkif_t;
+
+static struct tap_blkif *tapfds[MAX_TAP_DEV];
+static int blktap_next_minor;
+
+module_param(blkif_reqs, int, 0);
+/* Run-time switchable: /sys/module/blktap/parameters/ */
+static unsigned int log_stats = 0;
+static unsigned int debug_lvl = 0;
+module_param(log_stats, int, 0644);
+module_param(debug_lvl, int, 0644);
+
+/*
+ * Each outstanding request that we've passed to the lower device layers has a 
+ * 'pending_req' allocated to it. Each buffer_head that completes decrements 
+ * the pendcnt towards zero. When it hits zero, the specified domain has a 
+ * response queued for it, with the saved 'id' passed back.
+ */
+typedef struct {
+	blkif_t       *blkif;
+	u64            id;
+	unsigned short mem_idx;
+	int            nr_pages;
+	atomic_t       pendcnt;
+	unsigned short operation;
+	int            status;
+	struct list_head free_list;
+	int            inuse;
+} pending_req_t;
+
+static pending_req_t *pending_reqs[MAX_PENDING_REQS];
+static struct list_head pending_free;
+static DEFINE_SPINLOCK(pending_free_lock);
+static DECLARE_WAIT_QUEUE_HEAD (pending_free_wq);
+static int alloc_pending_reqs;
+
+typedef unsigned int PEND_RING_IDX;
+
+static inline int MASK_PEND_IDX(int i) { 
+	return (i & (MAX_PENDING_REQS-1));
+}
+
+static inline unsigned int RTN_PEND_IDX(pending_req_t *req, int idx) {
+	return (req - pending_reqs[idx]);
+}
+
+#define NR_PENDING_REQS (MAX_PENDING_REQS - pending_prod + pending_cons)
+
+#define BLKBACK_INVALID_HANDLE (~0)
+
+static struct page **foreign_pages[MAX_DYNAMIC_MEM];
+static inline unsigned long idx_to_kaddr(
+	unsigned int mmap_idx, unsigned int req_idx, unsigned int sg_idx)
+{
+	unsigned int arr_idx = req_idx*BLKIF_MAX_SEGMENTS_PER_REQUEST + sg_idx;
+	unsigned long pfn = page_to_pfn(foreign_pages[mmap_idx][arr_idx]);
+	return (unsigned long)pfn_to_kaddr(pfn);
+}
+
+static unsigned short mmap_alloc = 0;
+static unsigned short mmap_lock = 0;
+static unsigned short mmap_inuse = 0;
+
+/******************************************************************
+ * GRANT HANDLES
+ */
+
+/* When using grant tables to map a frame for device access then the
+ * handle returned must be used to unmap the frame. This is needed to
+ * drop the ref count on the frame.
+ */
+struct grant_handle_pair
+{
+        grant_handle_t kernel;
+        grant_handle_t user;
+};
+#define INVALID_GRANT_HANDLE	0xFFFF
+
+static struct grant_handle_pair 
+    pending_grant_handles[MAX_DYNAMIC_MEM][MMAP_PAGES];
+#define pending_handle(_id, _idx, _i) \
+    (pending_grant_handles[_id][((_idx) * BLKIF_MAX_SEGMENTS_PER_REQUEST) \
+    + (_i)])
+
+
+static int blktap_read_ufe_ring(tap_blkif_t *info); /*local prototypes*/
+
+#define BLKTAP_MINOR 0  /*/dev/xen/blktap has a dynamic major */
+#define BLKTAP_DEV_DIR  "/dev/xen"
+
+static int blktap_major;
+
+/* blktap IOCTLs: */
+#define BLKTAP_IOCTL_KICK_FE         1
+#define BLKTAP_IOCTL_KICK_BE         2 /* currently unused */
+#define BLKTAP_IOCTL_SETMODE         3
+#define BLKTAP_IOCTL_SENDPID	     4
+#define BLKTAP_IOCTL_NEWINTF	     5
+#define BLKTAP_IOCTL_MINOR	     6
+#define BLKTAP_IOCTL_MAJOR	     7
+#define BLKTAP_QUERY_ALLOC_REQS      8
+#define BLKTAP_IOCTL_FREEINTF        9
+#define BLKTAP_IOCTL_NEWINTF_EXT     50
+#define BLKTAP_IOCTL_PRINT_IDXS      100  
+
+/* blktap switching modes: (Set with BLKTAP_IOCTL_SETMODE)             */
+#define BLKTAP_MODE_PASSTHROUGH      0x00000000  /* default            */
+#define BLKTAP_MODE_INTERCEPT_FE     0x00000001
+#define BLKTAP_MODE_INTERCEPT_BE     0x00000002  /* unimp.             */
+
+#define BLKTAP_MODE_INTERPOSE \
+           (BLKTAP_MODE_INTERCEPT_FE | BLKTAP_MODE_INTERCEPT_BE)
+
+
+static inline int BLKTAP_MODE_VALID(unsigned long arg)
+{
+	return ((arg == BLKTAP_MODE_PASSTHROUGH ) ||
+		(arg == BLKTAP_MODE_INTERCEPT_FE) ||
+                (arg == BLKTAP_MODE_INTERPOSE   ));
+}
+
+/* Requests passing through the tap to userspace are re-assigned an ID.
+ * We must record a mapping between the BE [IDX,ID] tuple and the userspace
+ * ring ID. 
+ */
+
+static inline unsigned long MAKE_ID(domid_t fe_dom, PEND_RING_IDX idx)
+{
+        return ((fe_dom << 16) | MASK_PEND_IDX(idx));
+}
+
+extern inline PEND_RING_IDX ID_TO_IDX(unsigned long id)
+{
+        return (PEND_RING_IDX)(id & 0x0000ffff);
+}
+
+extern inline int ID_TO_MIDX(unsigned long id)
+{
+        return (int)(id >> 16);
+}
+
+#define INVALID_REQ 0xdead0000
+
+/*TODO: Convert to a free list*/
+static inline int GET_NEXT_REQ(unsigned long *idx_map)
+{
+	int i;
+	for (i = 0; i < MAX_PENDING_REQS; i++)
+		if (idx_map[i] == INVALID_REQ)
+			return i;
+
+	return INVALID_REQ;
+}
+
+static inline int OFFSET_TO_USR_IDX(int offset)
+{
+	return offset / BLKIF_MAX_SEGMENTS_PER_REQUEST;
+}
+
+static inline int OFFSET_TO_SEG(int offset)
+{
+	return offset % BLKIF_MAX_SEGMENTS_PER_REQUEST;
+}
+
+
+#define BLKTAP_INVALID_HANDLE(_g) \
+    (((_g->kernel) == INVALID_GRANT_HANDLE) &&  \
+     ((_g->user) == INVALID_GRANT_HANDLE))
+
+#define BLKTAP_INVALIDATE_HANDLE(_g) do {       \
+    (_g)->kernel = INVALID_GRANT_HANDLE; (_g)->user = INVALID_GRANT_HANDLE; \
+    } while(0)
+
+
+/******************************************************************
+ * BLKTAP VM OPS
+ */
+
+static struct page *blktap_nopage(struct vm_area_struct *vma,
+				  unsigned long address,
+				  int *type)
+{
+	/*
+	 * if the page has not been mapped in by the driver then return
+	 * NOPAGE_SIGBUS to the domain.
+	 */
+
+	return NOPAGE_SIGBUS;
+}
+
+static pte_t blktap_clear_pte(struct vm_area_struct *vma,
+			      unsigned long uvaddr,
+			      pte_t *ptep, int is_fullmm)
+{
+	pte_t copy;
+	tap_blkif_t *info;
+	int offset, seg, usr_idx, pending_idx, mmap_idx;
+	unsigned long uvstart = vma->vm_start + (RING_PAGES << PAGE_SHIFT);
+	unsigned long kvaddr;
+	struct page **map;
+	struct page *pg;
+	struct grant_handle_pair *khandle;
+	struct gnttab_unmap_grant_ref unmap[2];
+	int count = 0;
+
+	/*
+	 * If the address is before the start of the grant mapped region or
+	 * if vm_file is NULL (meaning mmap failed and we have nothing to do)
+	 */
+	if (uvaddr < uvstart || vma->vm_file == NULL)
+		return ptep_get_and_clear_full(vma->vm_mm, uvaddr, 
+					       ptep, is_fullmm);
+
+	info = vma->vm_file->private_data;
+	map = vma->vm_private_data;
+
+	/* TODO Should these be changed to if statements? */
+	BUG_ON(!info);
+	BUG_ON(!info->idx_map);
+	BUG_ON(!map);
+
+	offset = (int) ((uvaddr - uvstart) >> PAGE_SHIFT);
+	usr_idx = OFFSET_TO_USR_IDX(offset);
+	seg = OFFSET_TO_SEG(offset);
+
+	pending_idx = MASK_PEND_IDX(ID_TO_IDX(info->idx_map[usr_idx]));
+	mmap_idx = ID_TO_MIDX(info->idx_map[usr_idx]);
+
+	kvaddr = idx_to_kaddr(mmap_idx, pending_idx, seg);
+	pg = pfn_to_page(__pa(kvaddr) >> PAGE_SHIFT);
+	ClearPageReserved(pg);
+	map[offset + RING_PAGES] = NULL;
+
+	khandle = &pending_handle(mmap_idx, pending_idx, seg);
+
+	if (khandle->kernel != INVALID_GRANT_HANDLE) {
+		gnttab_set_unmap_op(&unmap[count], kvaddr, 
+				    GNTMAP_host_map, khandle->kernel);
+		count++;
+
+		set_phys_to_machine(__pa(kvaddr) >> PAGE_SHIFT, 
+				    INVALID_P2M_ENTRY);
+	}
+
+	if (khandle->user != INVALID_GRANT_HANDLE) {
+		BUG_ON(xen_feature(XENFEAT_auto_translated_physmap));
+
+		copy = *ptep;
+		gnttab_set_unmap_op(&unmap[count], virt_to_machine(ptep), 
+				    GNTMAP_host_map 
+				    | GNTMAP_application_map 
+				    | GNTMAP_contains_pte,
+				    khandle->user);
+		count++;
+	} else {
+		BUG_ON(!xen_feature(XENFEAT_auto_translated_physmap));
+
+		/* USING SHADOW PAGE TABLES. */
+		copy = ptep_get_and_clear_full(vma->vm_mm, uvaddr, ptep,
+					       is_fullmm);
+	}
+
+	if (count) {
+		BLKTAP_INVALIDATE_HANDLE(khandle);
+		if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref,
+					      unmap, count))
+			BUG();
+	}
+
+	return copy;
+}
+
+struct vm_operations_struct blktap_vm_ops = {
+	nopage:   blktap_nopage,
+	zap_pte:  blktap_clear_pte,
+};
+
+/******************************************************************
+ * BLKTAP FILE OPS
+ */
+ 
+/*Function Declarations*/
+static tap_blkif_t *get_next_free_dev(void);
+static int blktap_open(struct inode *inode, struct file *filp);
+static int blktap_release(struct inode *inode, struct file *filp);
+static int blktap_mmap(struct file *filp, struct vm_area_struct *vma);
+static int blktap_ioctl(struct inode *inode, struct file *filp,
+                        unsigned int cmd, unsigned long arg);
+static unsigned int blktap_poll(struct file *file, poll_table *wait);
+
+static const struct file_operations blktap_fops = {
+	.owner   = THIS_MODULE,
+	.poll    = blktap_poll,
+	.ioctl   = blktap_ioctl,
+	.open    = blktap_open,
+	.release = blktap_release,
+	.mmap    = blktap_mmap,
+};
+
+
+static tap_blkif_t *get_next_free_dev(void)
+{
+	struct class *class;
+	tap_blkif_t *info;
+	int minor;
+
+	/*
+	 * This is called only from the ioctl, which
+	 * means we should always have interrupts enabled.
+	 */
+	BUG_ON(irqs_disabled());
+
+	spin_lock_irq(&pending_free_lock);
+
+	/* tapfds[0] is always NULL */
+
+	for (minor = 1; minor < blktap_next_minor; minor++) {
+		info = tapfds[minor];
+		/* we could have failed a previous attempt. */
+		if (!info ||
+		    ((info->dev_inuse == 0) &&
+		     (info->dev_pending == 0)) ) {
+			info->dev_pending = 1;
+			goto found;
+		}
+	}
+	info = NULL;
+	minor = -1;
+
+	/*
+	 * We didn't find free device. If we can still allocate
+	 * more, then we grab the next device minor that is
+	 * available.  This is done while we are still under
+	 * the protection of the pending_free_lock.
+	 */
+	if (blktap_next_minor < MAX_TAP_DEV)
+		minor = blktap_next_minor++;
+found:
+	spin_unlock_irq(&pending_free_lock);
+
+	if (!info && minor > 0) {
+		info = kzalloc(sizeof(*info), GFP_KERNEL);
+		if (unlikely(!info)) {
+			/*
+			 * If we failed here, try to put back
+			 * the next minor number. But if one
+			 * was just taken, then we just lose this
+			 * minor.  We can try to allocate this
+			 * minor again later.
+			 */
+			spin_lock_irq(&pending_free_lock);
+			if (blktap_next_minor == minor+1)
+				blktap_next_minor--;
+			spin_unlock_irq(&pending_free_lock);
+			goto out;
+		}
+
+		info->minor = minor;
+		/*
+		 * Make sure that we have a minor before others can
+		 * see us.
+		 */
+		wmb();
+		tapfds[minor] = info;
+
+		if ((class = get_xen_class()) != NULL)
+			class_device_create(class, NULL,
+					    MKDEV(blktap_major, minor), NULL,
+					    "blktap%d", minor);
+	}
+
+out:
+	return info;
+}
+
+int dom_to_devid(domid_t domid, int xenbus_id, blkif_t *blkif) 
+{
+	tap_blkif_t *info;
+	int i;
+
+	for (i = 1; i < blktap_next_minor; i++) {
+		info = tapfds[i];
+		if ( info &&
+		     (info->trans.domid == domid) &&
+		     (info->trans.busid == xenbus_id) ) {
+			info->blkif = blkif;
+			info->status = RUNNING;
+			return i;
+		}
+	}
+	return -1;
+}
+
+void signal_tapdisk(int idx) 
+{
+	tap_blkif_t *info;
+	struct task_struct *ptask;
+
+	/*
+	 * if the userland tools set things up wrong, this could be negative;
+	 * just don't try to signal in this case
+	 */
+	if (idx < 0)
+		return;
+
+	info = tapfds[idx];
+	if ((idx < 0) || (idx > MAX_TAP_DEV) || !info)
+		return;
+
+	if (info->pid > 0) {
+		ptask = find_task_by_pid(info->pid);
+		if (ptask)
+			info->status = CLEANSHUTDOWN;
+	}
+	info->blkif = NULL;
+
+	return;
+}
+
+static int blktap_open(struct inode *inode, struct file *filp)
+{
+	blkif_sring_t *sring;
+	int idx = iminor(inode) - BLKTAP_MINOR;
+	tap_blkif_t *info;
+	int i;
+	
+	/* ctrl device, treat differently */
+	if (!idx)
+		return 0;
+
+	info = tapfds[idx];
+
+	if ((idx < 0) || (idx > MAX_TAP_DEV) || !info) {
+		WPRINTK("Unable to open device /dev/xen/blktap%d\n",
+			idx);
+		return -ENODEV;
+	}
+
+	DPRINTK("Opening device /dev/xen/blktap%d\n",idx);
+	
+	/*Only one process can access device at a time*/
+	if (test_and_set_bit(0, &info->dev_inuse))
+		return -EBUSY;
+
+	info->dev_pending = 0;
+	    
+	/* Allocate the fe ring. */
+	sring = (blkif_sring_t *)get_zeroed_page(GFP_KERNEL);
+	if (sring == NULL)
+		goto fail_nomem;
+
+	SetPageReserved(virt_to_page(sring));
+    
+	SHARED_RING_INIT(sring);
+	FRONT_RING_INIT(&info->ufe_ring, sring, PAGE_SIZE);
+	
+	filp->private_data = info;
+	info->vma = NULL;
+
+	info->idx_map = kmalloc(sizeof(unsigned long) * MAX_PENDING_REQS, 
+				GFP_KERNEL);
+	
+	if (info->idx_map == NULL)
+		goto fail_nomem;
+
+	if (idx > 0) {
+		init_waitqueue_head(&info->wait);
+		for (i = 0; i < MAX_PENDING_REQS; i++) 
+			info->idx_map[i] = INVALID_REQ;
+	}
+
+	DPRINTK("Tap open: device /dev/xen/blktap%d\n",idx);
+	return 0;
+
+ fail_nomem:
+	return -ENOMEM;
+}
+
+static int blktap_release(struct inode *inode, struct file *filp)
+{
+	tap_blkif_t *info = filp->private_data;
+	
+	/* check for control device */
+	if (!info)
+		return 0;
+
+	info->dev_inuse = 0;
+	DPRINTK("Freeing device [/dev/xen/blktap%d]\n",info->minor);
+
+	/* Free the ring page. */
+	ClearPageReserved(virt_to_page(info->ufe_ring.sring));
+	free_page((unsigned long) info->ufe_ring.sring);
+
+	/* Clear any active mappings and free foreign map table */
+	if (info->vma) {
+		zap_page_range(
+			info->vma, info->vma->vm_start, 
+			info->vma->vm_end - info->vma->vm_start, NULL);
+
+		kfree(info->vma->vm_private_data);
+
+		info->vma = NULL;
+	}
+
+	if (info->idx_map) {
+		kfree(info->idx_map);
+		info->idx_map = NULL;
+	}
+
+	if ( (info->status != CLEANSHUTDOWN) && (info->blkif != NULL) ) {
+		if (info->blkif->xenblkd != NULL) {
+			kthread_stop(info->blkif->xenblkd);
+			info->blkif->xenblkd = NULL;
+		}
+		info->status = CLEANSHUTDOWN;
+	}
+
+	return 0;
+}
+
+
+/* Note on mmap:
+ * We need to map pages to user space in a way that will allow the block
+ * subsystem set up direct IO to them.  This couldn't be done before, because
+ * there isn't really a sane way to translate a user virtual address down to a 
+ * physical address when the page belongs to another domain.
+ *
+ * My first approach was to map the page in to kernel memory, add an entry
+ * for it in the physical frame list (using alloc_lomem_region as in blkback)
+ * and then attempt to map that page up to user space.  This is disallowed
+ * by xen though, which realizes that we don't really own the machine frame
+ * underlying the physical page.
+ *
+ * The new approach is to provide explicit support for this in xen linux.
+ * The VMA now has a flag, VM_FOREIGN, to indicate that it contains pages
+ * mapped from other vms.  vma->vm_private_data is set up as a mapping 
+ * from pages to actual page structs.  There is a new clause in get_user_pages
+ * that does the right thing for this sort of mapping.
+ */
+static int blktap_mmap(struct file *filp, struct vm_area_struct *vma)
+{
+	int size;
+	struct page **map;
+	int i;
+	tap_blkif_t *info = filp->private_data;
+	int ret;
+
+	if (info == NULL) {
+		WPRINTK("blktap: mmap, retrieving idx failed\n");
+		return -ENOMEM;
+	}
+	
+	vma->vm_flags |= VM_RESERVED;
+	vma->vm_ops = &blktap_vm_ops;
+
+	size = vma->vm_end - vma->vm_start;
+	if (size != ((mmap_pages + RING_PAGES) << PAGE_SHIFT)) {
+		WPRINTK("you _must_ map exactly %d pages!\n",
+		       mmap_pages + RING_PAGES);
+		return -EAGAIN;
+	}
+
+	size >>= PAGE_SHIFT;
+	info->rings_vstart = vma->vm_start;
+	info->user_vstart  = info->rings_vstart + (RING_PAGES << PAGE_SHIFT);
+    
+	/* Map the ring pages to the start of the region and reserve it. */
+	if (xen_feature(XENFEAT_auto_translated_physmap))
+		ret = vm_insert_page(vma, vma->vm_start,
+				     virt_to_page(info->ufe_ring.sring));
+	else
+		ret = remap_pfn_range(vma, vma->vm_start,
+				      __pa(info->ufe_ring.sring) >> PAGE_SHIFT,
+				      PAGE_SIZE, vma->vm_page_prot);
+	if (ret) {
+		WPRINTK("Mapping user ring failed!\n");
+		goto fail;
+	}
+
+	/* Mark this VM as containing foreign pages, and set up mappings. */
+	map = kzalloc(((vma->vm_end - vma->vm_start) >> PAGE_SHIFT)
+		      * sizeof(struct page *),
+		      GFP_KERNEL);
+	if (map == NULL) {
+		WPRINTK("Couldn't alloc VM_FOREIGN map.\n");
+		goto fail;
+	}
+
+	for (i = 0; i < ((vma->vm_end - vma->vm_start) >> PAGE_SHIFT); i++)
+		map[i] = NULL;
+    
+	vma->vm_private_data = map;
+	vma->vm_flags |= VM_FOREIGN;
+	vma->vm_flags |= VM_DONTCOPY;
+
+#ifdef CONFIG_X86
+	vma->vm_mm->context.has_foreign_mappings = 1;
+#endif
+
+	info->vma = vma;
+	info->ring_ok = 1;
+	return 0;
+ fail:
+	/* Clear any active mappings. */
+	zap_page_range(vma, vma->vm_start, 
+		       vma->vm_end - vma->vm_start, NULL);
+
+	return -ENOMEM;
+}
+
+
+static int blktap_ioctl(struct inode *inode, struct file *filp,
+                        unsigned int cmd, unsigned long arg)
+{
+	tap_blkif_t *info = filp->private_data;
+
+	switch(cmd) {
+	case BLKTAP_IOCTL_KICK_FE: 
+	{
+		/* There are fe messages to process. */
+		return blktap_read_ufe_ring(info);
+	}
+	case BLKTAP_IOCTL_SETMODE:
+	{
+		if (info) {
+			if (BLKTAP_MODE_VALID(arg)) {
+				info->mode = arg;
+				/* XXX: may need to flush rings here. */
+				DPRINTK("blktap: set mode to %lx\n", 
+				       arg);
+				return 0;
+			}
+		}
+		return 0;
+	}
+	case BLKTAP_IOCTL_PRINT_IDXS:
+        {
+		if (info) {
+			printk("User Rings: \n-----------\n");
+			printk("UF: rsp_cons: %2d, req_prod_prv: %2d "
+				"| req_prod: %2d, rsp_prod: %2d\n",
+				info->ufe_ring.rsp_cons,
+				info->ufe_ring.req_prod_pvt,
+				info->ufe_ring.sring->req_prod,
+				info->ufe_ring.sring->rsp_prod);
+		}
+            	return 0;
+        }
+	case BLKTAP_IOCTL_SENDPID:
+	{
+		if (info) {
+			info->pid = (pid_t)arg;
+			DPRINTK("blktap: pid received %d\n", 
+			       info->pid);
+		}
+		return 0;
+	}
+	case BLKTAP_IOCTL_NEWINTF:
+	{		
+		uint64_t val = (uint64_t)arg;
+		domid_translate_t *tr = (domid_translate_t *)&val;
+
+		DPRINTK("NEWINTF Req for domid %d and bus id %d\n", 
+		       tr->domid, tr->busid);
+		info = get_next_free_dev();
+		if (!info) {
+			WPRINTK("Error initialising /dev/xen/blktap - "
+				"No more devices\n");
+			return -1;
+		}
+		info->trans.domid = tr->domid;
+		info->trans.busid = tr->busid;
+		return info->minor;
+	}
+	case BLKTAP_IOCTL_NEWINTF_EXT:
+	{
+		void __user *udata = (void __user *) arg;
+		domid_translate_ext_t tr;
+
+		if (copy_from_user(&tr, udata, sizeof(domid_translate_ext_t)))
+			return -EFAULT;
+
+		DPRINTK("NEWINTF_EXT Req for domid %d and bus id %d\n", 
+		       tr.domid, tr.busid);
+		info = get_next_free_dev();
+		if (!info) {
+			WPRINTK("Error initialising /dev/xen/blktap - "
+				"No more devices\n");
+			return -1;
+		}
+		info->trans.domid = tr.domid;
+		info->trans.busid = tr.busid;
+		return info->minor;
+	}
+	case BLKTAP_IOCTL_FREEINTF:
+	{
+		unsigned long dev = arg;
+		unsigned long flags;
+
+		info = tapfds[dev];
+
+		if ((dev > MAX_TAP_DEV) || !info)
+			return 0; /* should this be an error? */
+
+		spin_lock_irqsave(&pending_free_lock, flags);
+		if (info->dev_pending)
+			info->dev_pending = 0;
+		spin_unlock_irqrestore(&pending_free_lock, flags);
+
+		return 0;
+	}
+	case BLKTAP_IOCTL_MINOR:
+	{
+		unsigned long dev = arg;
+
+		info = tapfds[dev];
+
+		if ((dev > MAX_TAP_DEV) || !info)
+			return -EINVAL;
+
+		return info->minor;
+	}
+	case BLKTAP_IOCTL_MAJOR:
+		return blktap_major;
+
+	case BLKTAP_QUERY_ALLOC_REQS:
+	{
+		WPRINTK("BLKTAP_QUERY_ALLOC_REQS ioctl: %d/%d\n",
+		       alloc_pending_reqs, blkif_reqs);
+		return (alloc_pending_reqs/blkif_reqs) * 100;
+	}
+	}
+	return -ENOIOCTLCMD;
+}
+
+static unsigned int blktap_poll(struct file *filp, poll_table *wait)
+{
+	tap_blkif_t *info = filp->private_data;
+	
+	/* do not work on the control device */
+	if (!info)
+		return 0;
+
+	poll_wait(filp, &info->wait, wait);
+	if (info->ufe_ring.req_prod_pvt != info->ufe_ring.sring->req_prod) {
+		RING_PUSH_REQUESTS(&info->ufe_ring);
+		return POLLIN | POLLRDNORM;
+	}
+	return 0;
+}
+
+void blktap_kick_user(int idx)
+{
+	tap_blkif_t *info;
+
+	info = tapfds[idx];
+
+	if ((idx < 0) || (idx > MAX_TAP_DEV) || !info)
+		return;
+
+	wake_up_interruptible(&info->wait);
+
+	return;
+}
+
+static int do_block_io_op(blkif_t *blkif);
+static void dispatch_rw_block_io(blkif_t *blkif,
+				 blkif_request_t *req,
+				 pending_req_t *pending_req);
+static void make_response(blkif_t *blkif, u64 id,
+                          unsigned short op, int st);
+
+/******************************************************************
+ * misc small helpers
+ */
+static int req_increase(void)
+{
+	int i, j;
+
+	if (mmap_alloc >= MAX_PENDING_REQS || mmap_lock) 
+		return -EINVAL;
+
+	pending_reqs[mmap_alloc]  = kzalloc(sizeof(pending_req_t)
+					    * blkif_reqs, GFP_KERNEL);
+	foreign_pages[mmap_alloc] = alloc_empty_pages_and_pagevec(mmap_pages);
+
+	if (!pending_reqs[mmap_alloc] || !foreign_pages[mmap_alloc])
+		goto out_of_memory;
+
+	DPRINTK("%s: reqs=%d, pages=%d\n",
+		__FUNCTION__, blkif_reqs, mmap_pages);
+
+	for (i = 0; i < MAX_PENDING_REQS; i++) {
+		list_add_tail(&pending_reqs[mmap_alloc][i].free_list, 
+			      &pending_free);
+		pending_reqs[mmap_alloc][i].mem_idx = mmap_alloc;
+		for (j = 0; j < BLKIF_MAX_SEGMENTS_PER_REQUEST; j++)
+			BLKTAP_INVALIDATE_HANDLE(&pending_handle(mmap_alloc, 
+								 i, j));
+	}
+
+	mmap_alloc++;
+	DPRINTK("# MMAPs increased to %d\n",mmap_alloc);
+	return 0;
+
+ out_of_memory:
+	free_empty_pages_and_pagevec(foreign_pages[mmap_alloc], mmap_pages);
+	kfree(pending_reqs[mmap_alloc]);
+	WPRINTK("%s: out of memory\n", __FUNCTION__);
+	return -ENOMEM;
+}
+
+static void mmap_req_del(int mmap)
+{
+	BUG_ON(!spin_is_locked(&pending_free_lock));
+
+	kfree(pending_reqs[mmap]);
+	pending_reqs[mmap] = NULL;
+
+	free_empty_pages_and_pagevec(foreign_pages[mmap_alloc], mmap_pages);
+	foreign_pages[mmap] = NULL;
+
+	mmap_lock = 0;
+	DPRINTK("# MMAPs decreased to %d\n",mmap_alloc);
+	mmap_alloc--;
+}
+
+static pending_req_t* alloc_req(void)
+{
+	pending_req_t *req = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&pending_free_lock, flags);
+
+	if (!list_empty(&pending_free)) {
+		req = list_entry(pending_free.next, pending_req_t, free_list);
+		list_del(&req->free_list);
+	}
+
+	if (req) {
+		req->inuse = 1;
+		alloc_pending_reqs++;
+	}
+	spin_unlock_irqrestore(&pending_free_lock, flags);
+
+	return req;
+}
+
+static void free_req(pending_req_t *req)
+{
+	unsigned long flags;
+	int was_empty;
+
+	spin_lock_irqsave(&pending_free_lock, flags);
+
+	alloc_pending_reqs--;
+	req->inuse = 0;
+	if (mmap_lock && (req->mem_idx == mmap_alloc-1)) {
+		mmap_inuse--;
+		if (mmap_inuse == 0) mmap_req_del(mmap_alloc-1);
+		spin_unlock_irqrestore(&pending_free_lock, flags);
+		return;
+	}
+	was_empty = list_empty(&pending_free);
+	list_add(&req->free_list, &pending_free);
+
+	spin_unlock_irqrestore(&pending_free_lock, flags);
+
+	if (was_empty)
+		wake_up(&pending_free_wq);
+}
+
+static void fast_flush_area(pending_req_t *req, int k_idx, int u_idx,
+			    int tapidx)
+{
+	struct gnttab_unmap_grant_ref unmap[BLKIF_MAX_SEGMENTS_PER_REQUEST*2];
+	unsigned int i, invcount = 0;
+	struct grant_handle_pair *khandle;
+	uint64_t ptep;
+	int ret, mmap_idx;
+	unsigned long kvaddr, uvaddr;
+	tap_blkif_t *info;
+	
+
+	info = tapfds[tapidx];
+
+	if ((tapidx < 0) || (tapidx > MAX_TAP_DEV) || !info) {
+		WPRINTK("fast_flush: Couldn't get info!\n");
+		return;
+	}
+
+	if (info->vma != NULL &&
+	    xen_feature(XENFEAT_auto_translated_physmap)) {
+		down_write(&info->vma->vm_mm->mmap_sem);
+		zap_page_range(info->vma, 
+			       MMAP_VADDR(info->user_vstart, u_idx, 0), 
+			       req->nr_pages << PAGE_SHIFT, NULL);
+		up_write(&info->vma->vm_mm->mmap_sem);
+		return;
+	}
+
+	mmap_idx = req->mem_idx;
+
+	for (i = 0; i < req->nr_pages; i++) {
+		kvaddr = idx_to_kaddr(mmap_idx, k_idx, i);
+		uvaddr = MMAP_VADDR(info->user_vstart, u_idx, i);
+
+		khandle = &pending_handle(mmap_idx, k_idx, i);
+
+		if (khandle->kernel != INVALID_GRANT_HANDLE) {
+			gnttab_set_unmap_op(&unmap[invcount],
+					    idx_to_kaddr(mmap_idx, k_idx, i),
+					    GNTMAP_host_map, khandle->kernel);
+			invcount++;
+
+			set_phys_to_machine(
+				__pa(idx_to_kaddr(mmap_idx, k_idx, i))
+				>> PAGE_SHIFT, INVALID_P2M_ENTRY);
+		}
+
+		if (khandle->user != INVALID_GRANT_HANDLE) {
+			BUG_ON(xen_feature(XENFEAT_auto_translated_physmap));
+			if (create_lookup_pte_addr(
+				info->vma->vm_mm,
+				MMAP_VADDR(info->user_vstart, u_idx, i),
+				&ptep) !=0) {
+				WPRINTK("Couldn't get a pte addr!\n");
+				return;
+			}
+
+			gnttab_set_unmap_op(&unmap[invcount], ptep,
+					    GNTMAP_host_map
+					    | GNTMAP_application_map
+					    | GNTMAP_contains_pte,
+					    khandle->user);
+			invcount++;
+		}
+
+		BLKTAP_INVALIDATE_HANDLE(khandle);
+	}
+	ret = HYPERVISOR_grant_table_op(
+		GNTTABOP_unmap_grant_ref, unmap, invcount);
+	BUG_ON(ret);
+	
+	if (info->vma != NULL && !xen_feature(XENFEAT_auto_translated_physmap))
+		zap_page_range(info->vma, 
+			       MMAP_VADDR(info->user_vstart, u_idx, 0), 
+			       req->nr_pages << PAGE_SHIFT, NULL);
+}
+
+/******************************************************************
+ * SCHEDULER FUNCTIONS
+ */
+
+static void print_stats(blkif_t *blkif)
+{
+	printk(KERN_DEBUG "%s: oo %3d  |  rd %4d  |  wr %4d\n",
+	       current->comm, blkif->st_oo_req,
+	       blkif->st_rd_req, blkif->st_wr_req);
+	blkif->st_print = jiffies + msecs_to_jiffies(10 * 1000);
+	blkif->st_rd_req = 0;
+	blkif->st_wr_req = 0;
+	blkif->st_oo_req = 0;
+}
+
+int tap_blkif_schedule(void *arg)
+{
+	blkif_t *blkif = arg;
+
+	blkif_get(blkif);
+
+	if (debug_lvl)
+		printk(KERN_DEBUG "%s: started\n", current->comm);
+
+	while (!kthread_should_stop()) {
+		if (try_to_freeze())
+			continue;
+
+		wait_event_interruptible(
+			blkif->wq,
+			blkif->waiting_reqs || kthread_should_stop());
+		wait_event_interruptible(
+			pending_free_wq,
+			!list_empty(&pending_free) || kthread_should_stop());
+
+		blkif->waiting_reqs = 0;
+		smp_mb(); /* clear flag *before* checking for work */
+
+		if (do_block_io_op(blkif))
+			blkif->waiting_reqs = 1;
+
+		if (log_stats && time_after(jiffies, blkif->st_print))
+			print_stats(blkif);
+	}
+
+	if (log_stats)
+		print_stats(blkif);
+	if (debug_lvl)
+		printk(KERN_DEBUG "%s: exiting\n", current->comm);
+
+	blkif->xenblkd = NULL;
+	blkif_put(blkif);
+
+	return 0;
+}
+
+/******************************************************************
+ * COMPLETION CALLBACK -- Called by user level ioctl()
+ */
+
+static int blktap_read_ufe_ring(tap_blkif_t *info)
+{
+	/* This is called to read responses from the UFE ring. */
+	RING_IDX i, j, rp;
+	blkif_response_t *resp;
+	blkif_t *blkif=NULL;
+	int pending_idx, usr_idx, mmap_idx;
+	pending_req_t *pending_req;
+	
+	if (!info)
+		return 0;
+
+	/* We currently only forward packets in INTERCEPT_FE mode. */
+	if (!(info->mode & BLKTAP_MODE_INTERCEPT_FE))
+		return 0;
+
+	/* for each outstanding message on the UFEring  */
+	rp = info->ufe_ring.sring->rsp_prod;
+	rmb();
+        
+	for (i = info->ufe_ring.rsp_cons; i != rp; i++) {
+		blkif_response_t res;
+		resp = RING_GET_RESPONSE(&info->ufe_ring, i);
+		memcpy(&res, resp, sizeof(res));
+		mb(); /* rsp_cons read by RING_FULL() in do_block_io_op(). */
+		++info->ufe_ring.rsp_cons;
+
+		/*retrieve [usr_idx] to [mmap_idx,pending_idx] mapping*/
+		usr_idx = (int)res.id;
+		pending_idx = MASK_PEND_IDX(ID_TO_IDX(info->idx_map[usr_idx]));
+		mmap_idx = ID_TO_MIDX(info->idx_map[usr_idx]);
+
+		if ( (mmap_idx >= mmap_alloc) || 
+		   (ID_TO_IDX(info->idx_map[usr_idx]) >= MAX_PENDING_REQS) )
+			WPRINTK("Incorrect req map"
+			       "[%d], internal map [%d,%d (%d)]\n", 
+			       usr_idx, mmap_idx, 
+			       ID_TO_IDX(info->idx_map[usr_idx]),
+			       MASK_PEND_IDX(
+				       ID_TO_IDX(info->idx_map[usr_idx])));
+
+		pending_req = &pending_reqs[mmap_idx][pending_idx];
+		blkif = pending_req->blkif;
+
+		for (j = 0; j < pending_req->nr_pages; j++) {
+
+			unsigned long kvaddr, uvaddr;
+			struct page **map = info->vma->vm_private_data;
+			struct page *pg;
+			int offset;
+
+			uvaddr = MMAP_VADDR(info->user_vstart, usr_idx, j);
+			kvaddr = idx_to_kaddr(mmap_idx, pending_idx, j);
+
+			pg = pfn_to_page(__pa(kvaddr) >> PAGE_SHIFT);
+			ClearPageReserved(pg);
+			offset = (uvaddr - info->vma->vm_start) 
+				>> PAGE_SHIFT;
+			map[offset] = NULL;
+		}
+		fast_flush_area(pending_req, pending_idx, usr_idx, info->minor);
+		info->idx_map[usr_idx] = INVALID_REQ;
+		make_response(blkif, pending_req->id, res.operation,
+			      res.status);
+		blkif_put(pending_req->blkif);
+		free_req(pending_req);
+	}
+		
+	return 0;
+}
+
+
+/******************************************************************************
+ * NOTIFICATION FROM GUEST OS.
+ */
+
+static void blkif_notify_work(blkif_t *blkif)
+{
+	blkif->waiting_reqs = 1;
+	wake_up(&blkif->wq);
+}
+
+irqreturn_t tap_blkif_be_int(int irq, void *dev_id, struct pt_regs *regs)
+{
+	blkif_notify_work(dev_id);
+	return IRQ_HANDLED;
+}
+
+
+
+/******************************************************************
+ * DOWNWARD CALLS -- These interface with the block-device layer proper.
+ */
+static int print_dbug = 1;
+static int do_block_io_op(blkif_t *blkif)
+{
+	blkif_back_rings_t *blk_rings = &blkif->blk_rings;
+	blkif_request_t req;
+	pending_req_t *pending_req;
+	RING_IDX rc, rp;
+	int more_to_do = 0;
+	tap_blkif_t *info;
+
+	rc = blk_rings->common.req_cons;
+	rp = blk_rings->common.sring->req_prod;
+	rmb(); /* Ensure we see queued requests up to 'rp'. */
+
+	/*Check blkif has corresponding UE ring*/
+	if (blkif->dev_num < 0) {
+		/*oops*/
+		if (print_dbug) {
+			WPRINTK("Corresponding UE " 
+			       "ring does not exist!\n");
+			print_dbug = 0; /*We only print this message once*/
+		}
+		return 0;
+	}
+
+	info = tapfds[blkif->dev_num];
+
+	if (blkif->dev_num > MAX_TAP_DEV || !info || !info->dev_inuse) {
+		if (print_dbug) {
+			WPRINTK("Can't get UE info!\n");
+			print_dbug = 0;
+		}
+		return 0;
+	}
+
+	while (rc != rp) {
+		
+		if (RING_FULL(&info->ufe_ring)) {
+			WPRINTK("RING_FULL! More to do\n");
+			more_to_do = 1;
+			break;
+		}
+
+		if (RING_REQUEST_CONS_OVERFLOW(&blk_rings->common, rc)) {
+			WPRINTK("RING_REQUEST_CONS_OVERFLOW!"
+			       " More to do\n");
+			more_to_do = 1;
+			break;		
+		}
+
+		pending_req = alloc_req();
+		if (NULL == pending_req) {
+			blkif->st_oo_req++;
+			more_to_do = 1;
+			break;
+		}
+
+		if (kthread_should_stop()) {
+			more_to_do = 1;
+			break;
+		}
+
+		switch (blkif->blk_protocol) {
+		case BLKIF_PROTOCOL_NATIVE:
+			memcpy(&req, RING_GET_REQUEST(&blk_rings->native, rc),
+			       sizeof(req));
+			break;
+		case BLKIF_PROTOCOL_X86_32:
+			blkif_get_x86_32_req(&req, RING_GET_REQUEST(&blk_rings->x86_32, rc));
+			break;
+		case BLKIF_PROTOCOL_X86_64:
+			blkif_get_x86_64_req(&req, RING_GET_REQUEST(&blk_rings->x86_64, rc));
+			break;
+		default:
+			BUG();
+		}
+		blk_rings->common.req_cons = ++rc; /* before make_response() */
+
+		/* Apply all sanity checks to /private copy/ of request. */
+		barrier();
+
+		switch (req.operation) {
+		case BLKIF_OP_READ:
+			blkif->st_rd_req++;
+			dispatch_rw_block_io(blkif, &req, pending_req);
+			break;
+
+		case BLKIF_OP_WRITE:
+			blkif->st_wr_req++;
+			dispatch_rw_block_io(blkif, &req, pending_req);
+			break;
+
+		default:
+			/* A good sign something is wrong: sleep for a while to
+			 * avoid excessive CPU consumption by a bad guest. */
+			msleep(1);
+			WPRINTK("unknown operation [%d]\n",
+				req.operation);
+			make_response(blkif, req.id, req.operation,
+				      BLKIF_RSP_ERROR);
+			free_req(pending_req);
+			break;
+		}
+
+		/* Yield point for this unbounded loop. */
+		cond_resched();
+	}
+		
+	blktap_kick_user(blkif->dev_num);
+
+	return more_to_do;
+}
+
+static void dispatch_rw_block_io(blkif_t *blkif,
+				 blkif_request_t *req,
+				 pending_req_t *pending_req)
+{
+	extern void ll_rw_block(int rw, int nr, struct buffer_head * bhs[]);
+	int op, operation = (req->operation == BLKIF_OP_WRITE) ? WRITE : READ;
+	struct gnttab_map_grant_ref map[BLKIF_MAX_SEGMENTS_PER_REQUEST*2];
+	unsigned int nseg;
+	int ret, i, nr_sects = 0;
+	tap_blkif_t *info;
+	blkif_request_t *target;
+	int pending_idx = RTN_PEND_IDX(pending_req,pending_req->mem_idx);
+	int usr_idx;
+	uint16_t mmap_idx = pending_req->mem_idx;
+
+	if (blkif->dev_num < 0 || blkif->dev_num > MAX_TAP_DEV)
+		goto fail_response;
+
+	info = tapfds[blkif->dev_num];
+	if (info == NULL)
+		goto fail_response;
+
+	/* Check we have space on user ring - should never fail. */
+	usr_idx = GET_NEXT_REQ(info->idx_map);
+	if (usr_idx == INVALID_REQ) {
+		BUG();
+		goto fail_response;
+	}
+
+	/* Check that number of segments is sane. */
+	nseg = req->nr_segments;
+	if ( unlikely(nseg == 0) || 
+	    unlikely(nseg > BLKIF_MAX_SEGMENTS_PER_REQUEST) ) {
+		WPRINTK("Bad number of segments in request (%d)\n", nseg);
+		goto fail_response;
+	}
+	
+	/* Make sure userspace is ready. */
+	if (!info->ring_ok) {
+		WPRINTK("blktap: ring not ready for requests!\n");
+		goto fail_response;
+	}
+
+	if (RING_FULL(&info->ufe_ring)) {
+		WPRINTK("blktap: fe_ring is full, can't add "
+			"IO Request will be dropped. %d %d\n",
+			RING_SIZE(&info->ufe_ring),
+			RING_SIZE(&blkif->blk_rings.common));
+		goto fail_response;
+	}
+
+	pending_req->blkif     = blkif;
+	pending_req->id        = req->id;
+	pending_req->operation = operation;
+	pending_req->status    = BLKIF_RSP_OKAY;
+	pending_req->nr_pages  = nseg;
+	op = 0;
+	for (i = 0; i < nseg; i++) {
+		unsigned long uvaddr;
+		unsigned long kvaddr;
+		uint64_t ptep;
+		uint32_t flags;
+
+		uvaddr = MMAP_VADDR(info->user_vstart, usr_idx, i);
+		kvaddr = idx_to_kaddr(mmap_idx, pending_idx, i);
+
+		flags = GNTMAP_host_map;
+		if (operation == WRITE)
+			flags |= GNTMAP_readonly;
+		gnttab_set_map_op(&map[op], kvaddr, flags,
+				  req->seg[i].gref, blkif->domid);
+		op++;
+
+		if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+			/* Now map it to user. */
+			ret = create_lookup_pte_addr(info->vma->vm_mm, 
+						     uvaddr, &ptep);
+			if (ret) {
+				WPRINTK("Couldn't get a pte addr!\n");
+				goto fail_flush;
+			}
+
+			flags = GNTMAP_host_map | GNTMAP_application_map
+				| GNTMAP_contains_pte;
+			if (operation == WRITE)
+				flags |= GNTMAP_readonly;
+			gnttab_set_map_op(&map[op], ptep, flags,
+					  req->seg[i].gref, blkif->domid);
+			op++;
+		}
+
+		nr_sects += (req->seg[i].last_sect - 
+			     req->seg[i].first_sect + 1);
+	}
+
+	ret = HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, map, op);
+	BUG_ON(ret);
+
+	if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+		for (i = 0; i < (nseg*2); i+=2) {
+			unsigned long uvaddr;
+			unsigned long kvaddr;
+			unsigned long offset;
+			struct page *pg;
+
+			uvaddr = MMAP_VADDR(info->user_vstart, usr_idx, i/2);
+			kvaddr = idx_to_kaddr(mmap_idx, pending_idx, i/2);
+
+			if (unlikely(map[i].status != 0)) {
+				WPRINTK("invalid kernel buffer -- "
+					"could not remap it\n");
+				ret |= 1;
+				map[i].handle = INVALID_GRANT_HANDLE;
+			}
+
+			if (unlikely(map[i+1].status != 0)) {
+				WPRINTK("invalid user buffer -- "
+					"could not remap it\n");
+				ret |= 1;
+				map[i+1].handle = INVALID_GRANT_HANDLE;
+			}
+
+			pending_handle(mmap_idx, pending_idx, i/2).kernel 
+				= map[i].handle;
+			pending_handle(mmap_idx, pending_idx, i/2).user   
+				= map[i+1].handle;
+
+			if (ret)
+				continue;
+
+			set_phys_to_machine(__pa(kvaddr) >> PAGE_SHIFT,
+					    FOREIGN_FRAME(map[i].dev_bus_addr
+							  >> PAGE_SHIFT));
+			offset = (uvaddr - info->vma->vm_start) >> PAGE_SHIFT;
+			pg = pfn_to_page(__pa(kvaddr) >> PAGE_SHIFT);
+			((struct page **)info->vma->vm_private_data)[offset] =
+				pg;
+		}
+	} else {
+		for (i = 0; i < nseg; i++) {
+			unsigned long uvaddr;
+			unsigned long kvaddr;
+			unsigned long offset;
+			struct page *pg;
+
+			uvaddr = MMAP_VADDR(info->user_vstart, usr_idx, i);
+			kvaddr = idx_to_kaddr(mmap_idx, pending_idx, i);
+
+			if (unlikely(map[i].status != 0)) {
+				WPRINTK("invalid kernel buffer -- "
+					"could not remap it\n");
+				ret |= 1;
+				map[i].handle = INVALID_GRANT_HANDLE;
+			}
+
+			pending_handle(mmap_idx, pending_idx, i).kernel 
+				= map[i].handle;
+
+			if (ret)
+				continue;
+
+			offset = (uvaddr - info->vma->vm_start) >> PAGE_SHIFT;
+			pg = pfn_to_page(__pa(kvaddr) >> PAGE_SHIFT);
+			((struct page **)info->vma->vm_private_data)[offset] =
+				pg;
+		}
+	}
+
+	if (ret)
+		goto fail_flush;
+
+	if (xen_feature(XENFEAT_auto_translated_physmap))
+		down_write(&info->vma->vm_mm->mmap_sem);
+	/* Mark mapped pages as reserved: */
+	for (i = 0; i < req->nr_segments; i++) {
+		unsigned long kvaddr;
+		struct page *pg;
+
+		kvaddr = idx_to_kaddr(mmap_idx, pending_idx, i);
+		pg = pfn_to_page(__pa(kvaddr) >> PAGE_SHIFT);
+		SetPageReserved(pg);
+		if (xen_feature(XENFEAT_auto_translated_physmap)) {
+			ret = vm_insert_page(info->vma,
+					     MMAP_VADDR(info->user_vstart,
+							usr_idx, i), pg);
+			if (ret) {
+				up_write(&info->vma->vm_mm->mmap_sem);
+				goto fail_flush;
+			}
+		}
+	}
+	if (xen_feature(XENFEAT_auto_translated_physmap))
+		up_write(&info->vma->vm_mm->mmap_sem);
+	
+	/*record [mmap_idx,pending_idx] to [usr_idx] mapping*/
+	info->idx_map[usr_idx] = MAKE_ID(mmap_idx, pending_idx);
+
+	blkif_get(blkif);
+	/* Finally, write the request message to the user ring. */
+	target = RING_GET_REQUEST(&info->ufe_ring,
+				  info->ufe_ring.req_prod_pvt);
+	memcpy(target, req, sizeof(*req));
+	target->id = usr_idx;
+	wmb(); /* blktap_poll() reads req_prod_pvt asynchronously */
+	info->ufe_ring.req_prod_pvt++;
+
+	if (operation == READ)
+		blkif->st_rd_sect += nr_sects;
+	else if (operation == WRITE)
+		blkif->st_wr_sect += nr_sects;
+
+	return;
+
+ fail_flush:
+	WPRINTK("Reached Fail_flush\n");
+	fast_flush_area(pending_req, pending_idx, usr_idx, blkif->dev_num);
+ fail_response:
+	make_response(blkif, req->id, req->operation, BLKIF_RSP_ERROR);
+	free_req(pending_req);
+	msleep(1); /* back off a bit */
+}
+
+
+
+/******************************************************************
+ * MISCELLANEOUS SETUP / TEARDOWN / DEBUGGING
+ */
+
+
+static void make_response(blkif_t *blkif, u64 id,
+                          unsigned short op, int st)
+{
+	blkif_response_t  resp;
+	unsigned long     flags;
+	blkif_back_rings_t *blk_rings = &blkif->blk_rings;
+	int more_to_do = 0;
+	int notify;
+
+	resp.id        = id;
+	resp.operation = op;
+	resp.status    = st;
+
+	spin_lock_irqsave(&blkif->blk_ring_lock, flags);
+	/* Place on the response ring for the relevant domain. */
+	switch (blkif->blk_protocol) {
+	case BLKIF_PROTOCOL_NATIVE:
+		memcpy(RING_GET_RESPONSE(&blk_rings->native,
+					 blk_rings->native.rsp_prod_pvt),
+		       &resp, sizeof(resp));
+		break;
+	case BLKIF_PROTOCOL_X86_32:
+		memcpy(RING_GET_RESPONSE(&blk_rings->x86_32,
+					 blk_rings->x86_32.rsp_prod_pvt),
+		       &resp, sizeof(resp));
+		break;
+	case BLKIF_PROTOCOL_X86_64:
+		memcpy(RING_GET_RESPONSE(&blk_rings->x86_64,
+					 blk_rings->x86_64.rsp_prod_pvt),
+		       &resp, sizeof(resp));
+		break;
+	default:
+		BUG();
+	}
+	blk_rings->common.rsp_prod_pvt++;
+	RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&blk_rings->common, notify);
+
+	if (blk_rings->common.rsp_prod_pvt == blk_rings->common.req_cons) {
+		/*
+		 * Tail check for pending requests. Allows frontend to avoid
+		 * notifications if requests are already in flight (lower
+		 * overheads and promotes batching).
+		 */
+		RING_FINAL_CHECK_FOR_REQUESTS(&blk_rings->common, more_to_do);
+	} else if (RING_HAS_UNCONSUMED_REQUESTS(&blk_rings->common)) {
+		more_to_do = 1;
+	}
+
+	spin_unlock_irqrestore(&blkif->blk_ring_lock, flags);
+	if (more_to_do)
+		blkif_notify_work(blkif);
+	if (notify)
+		notify_remote_via_irq(blkif->irq);
+}
+
+static int __init blkif_init(void)
+{
+	int i, ret;
+	struct class *class;
+
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	INIT_LIST_HEAD(&pending_free);
+        for(i = 0; i < 2; i++) {
+		ret = req_increase();
+		if (ret)
+			break;
+	}
+	if (i == 0)
+		return ret;
+
+	tap_blkif_interface_init();
+
+	alloc_pending_reqs = 0;
+
+	tap_blkif_xenbus_init();
+
+	/* Dynamically allocate a major for this device */
+	ret = register_chrdev(0, "blktap", &blktap_fops);
+
+	if (ret < 0) {
+		WPRINTK("Couldn't register /dev/xen/blktap\n");
+		return -ENOMEM;
+	}	
+	
+	blktap_major = ret;
+
+	/* tapfds[0] is always NULL */
+	blktap_next_minor++;
+
+	DPRINTK("Created misc_dev [/dev/xen/blktap%d]\n",i);
+
+	/* Make sure the xen class exists */
+	if ((class = get_xen_class()) != NULL) {
+		/*
+		 * This will allow udev to create the blktap ctrl device.
+		 * We only want to create blktap0 first.  We don't want
+		 * to flood the sysfs system with needless blktap devices.
+		 * We only create the device when a request of a new device is
+		 * made.
+		 */
+		class_device_create(class, NULL,
+				    MKDEV(blktap_major, 0), NULL,
+				    "blktap0");
+	} else {
+		/* this is bad, but not fatal */
+		WPRINTK("blktap: sysfs xen_class not created\n");
+	}
+
+	DPRINTK("Blktap device successfully created\n");
+
+	return 0;
+}
+
+module_init(blkif_init);
+
+MODULE_LICENSE("Dual BSD/GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/blktap/common.h linux-2.6.18-xen-3.3.0/drivers/xen/blktap/common.h
--- linux-2.6.18.8/drivers/xen/blktap/common.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/blktap/common.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,121 @@
+/* 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __BLKIF__BACKEND__COMMON_H__
+#define __BLKIF__BACKEND__COMMON_H__
+
+#include <linux/version.h>
+#include <linux/module.h>
+#include <linux/interrupt.h>
+#include <linux/slab.h>
+#include <linux/blkdev.h>
+#include <linux/vmalloc.h>
+#include <asm/io.h>
+#include <asm/setup.h>
+#include <asm/pgalloc.h>
+#include <xen/evtchn.h>
+#include <asm/hypervisor.h>
+#include <xen/blkif.h>
+#include <xen/gnttab.h>
+#include <xen/driver_util.h>
+
+#define DPRINTK(_f, _a...) pr_debug("(file=%s, line=%d) " _f, \
+                                    __FILE__ , __LINE__ , ## _a )
+
+#define WPRINTK(fmt, args...) printk(KERN_WARNING "blk_tap: " fmt, ##args)
+
+struct backend_info;
+
+typedef struct blkif_st {
+	/* Unique identifier for this interface. */
+	domid_t           domid;
+	unsigned int      handle;
+	/* Physical parameters of the comms window. */
+	unsigned int      irq;
+	/* Comms information. */
+	enum blkif_protocol blk_protocol;
+	blkif_back_rings_t blk_rings;
+	struct vm_struct *blk_ring_area;
+	/* Back pointer to the backend_info. */
+	struct backend_info *be;
+	/* Private fields. */
+	spinlock_t       blk_ring_lock;
+	atomic_t         refcnt;
+
+	wait_queue_head_t   wq;
+	struct task_struct  *xenblkd;
+	unsigned int        waiting_reqs;
+	request_queue_t     *plug;
+
+	/* statistics */
+	unsigned long       st_print;
+	int                 st_rd_req;
+	int                 st_wr_req;
+	int                 st_oo_req;
+	int                 st_rd_sect;
+	int                 st_wr_sect;
+
+	wait_queue_head_t waiting_to_free;
+
+	grant_handle_t shmem_handle;
+	grant_ref_t    shmem_ref;
+	
+	int		dev_num;
+	uint64_t        sectors;
+} blkif_t;
+
+blkif_t *tap_alloc_blkif(domid_t domid);
+void tap_blkif_free(blkif_t *blkif);
+int tap_blkif_map(blkif_t *blkif, unsigned long shared_page, 
+		  unsigned int evtchn);
+void tap_blkif_unmap(blkif_t *blkif);
+
+#define blkif_get(_b) (atomic_inc(&(_b)->refcnt))
+#define blkif_put(_b)					\
+	do {						\
+		if (atomic_dec_and_test(&(_b)->refcnt))	\
+			wake_up(&(_b)->waiting_to_free);\
+	} while (0)
+
+
+struct phys_req {
+	unsigned short       dev;
+	unsigned short       nr_sects;
+	struct block_device *bdev;
+	blkif_sector_t       sector_number;
+};
+
+void tap_blkif_interface_init(void);
+
+void tap_blkif_xenbus_init(void);
+
+irqreturn_t tap_blkif_be_int(int irq, void *dev_id, struct pt_regs *regs);
+int tap_blkif_schedule(void *arg);
+
+int dom_to_devid(domid_t domid, int xenbus_id, blkif_t *blkif);
+void signal_tapdisk(int idx);
+
+#endif /* __BLKIF__BACKEND__COMMON_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/blktap/interface.c linux-2.6.18-xen-3.3.0/drivers/xen/blktap/interface.c
--- linux-2.6.18.8/drivers/xen/blktap/interface.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/blktap/interface.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,174 @@
+/******************************************************************************
+ * drivers/xen/blktap/interface.c
+ * 
+ * Block-device interface management.
+ * 
+ * Copyright (c) 2004, Keir Fraser
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+
+ */
+
+#include "common.h"
+#include <xen/evtchn.h>
+
+static kmem_cache_t *blkif_cachep;
+
+blkif_t *tap_alloc_blkif(domid_t domid)
+{
+	blkif_t *blkif;
+
+	blkif = kmem_cache_alloc(blkif_cachep, GFP_KERNEL);
+	if (!blkif)
+		return ERR_PTR(-ENOMEM);
+
+	memset(blkif, 0, sizeof(*blkif));
+	blkif->domid = domid;
+	spin_lock_init(&blkif->blk_ring_lock);
+	atomic_set(&blkif->refcnt, 1);
+	init_waitqueue_head(&blkif->wq);
+	blkif->st_print = jiffies;
+	init_waitqueue_head(&blkif->waiting_to_free);
+
+	return blkif;
+}
+
+static int map_frontend_page(blkif_t *blkif, unsigned long shared_page)
+{
+	struct gnttab_map_grant_ref op;
+
+	gnttab_set_map_op(&op, (unsigned long)blkif->blk_ring_area->addr,
+			  GNTMAP_host_map, shared_page, blkif->domid);
+
+	if (HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, &op, 1))
+		BUG();
+
+	if (op.status) {
+		DPRINTK(" Grant table operation failure !\n");
+		return op.status;
+	}
+
+	blkif->shmem_ref = shared_page;
+	blkif->shmem_handle = op.handle;
+
+	return 0;
+}
+
+static void unmap_frontend_page(blkif_t *blkif)
+{
+	struct gnttab_unmap_grant_ref op;
+
+	gnttab_set_unmap_op(&op, (unsigned long)blkif->blk_ring_area->addr,
+			    GNTMAP_host_map, blkif->shmem_handle);
+
+	if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1))
+		BUG();
+}
+
+int tap_blkif_map(blkif_t *blkif, unsigned long shared_page, 
+		  unsigned int evtchn)
+{
+	int err;
+
+	/* Already connected through? */
+	if (blkif->irq)
+		return 0;
+
+	if ( (blkif->blk_ring_area = alloc_vm_area(PAGE_SIZE)) == NULL )
+		return -ENOMEM;
+
+	err = map_frontend_page(blkif, shared_page);
+	if (err) {
+		free_vm_area(blkif->blk_ring_area);
+		return err;
+	}
+
+	switch (blkif->blk_protocol) {
+	case BLKIF_PROTOCOL_NATIVE:
+	{
+		blkif_sring_t *sring;
+		sring = (blkif_sring_t *)blkif->blk_ring_area->addr;
+		BACK_RING_INIT(&blkif->blk_rings.native, sring, PAGE_SIZE);
+		break;
+	}
+	case BLKIF_PROTOCOL_X86_32:
+	{
+		blkif_x86_32_sring_t *sring_x86_32;
+		sring_x86_32 = (blkif_x86_32_sring_t *)blkif->blk_ring_area->addr;
+		BACK_RING_INIT(&blkif->blk_rings.x86_32, sring_x86_32, PAGE_SIZE);
+		break;
+	}
+	case BLKIF_PROTOCOL_X86_64:
+	{
+		blkif_x86_64_sring_t *sring_x86_64;
+		sring_x86_64 = (blkif_x86_64_sring_t *)blkif->blk_ring_area->addr;
+		BACK_RING_INIT(&blkif->blk_rings.x86_64, sring_x86_64, PAGE_SIZE);
+		break;
+	}
+	default:
+		BUG();
+	}
+
+	err = bind_interdomain_evtchn_to_irqhandler(
+		blkif->domid, evtchn, tap_blkif_be_int,
+		0, "blkif-backend", blkif);
+	if (err < 0) {
+		unmap_frontend_page(blkif);
+		free_vm_area(blkif->blk_ring_area);
+		blkif->blk_rings.common.sring = NULL;
+		return err;
+	}
+	blkif->irq = err;
+
+	return 0;
+}
+
+void tap_blkif_unmap(blkif_t *blkif)
+{
+	if (blkif->irq) {
+		unbind_from_irqhandler(blkif->irq, blkif);
+		blkif->irq = 0;
+	}
+	if (blkif->blk_rings.common.sring) {
+		unmap_frontend_page(blkif);
+		free_vm_area(blkif->blk_ring_area);
+		blkif->blk_rings.common.sring = NULL;
+	}
+}
+
+void tap_blkif_free(blkif_t *blkif)
+{
+	atomic_dec(&blkif->refcnt);
+	wait_event(blkif->waiting_to_free, atomic_read(&blkif->refcnt) == 0);
+
+	tap_blkif_unmap(blkif);
+	kmem_cache_free(blkif_cachep, blkif);
+}
+
+void __init tap_blkif_interface_init(void)
+{
+	blkif_cachep = kmem_cache_create("blktapif_cache", sizeof(blkif_t), 
+					 0, 0, NULL, NULL);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/blktap/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/blktap/Makefile
--- linux-2.6.18.8/drivers/xen/blktap/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/blktap/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,5 @@
+LINUXINCLUDE += -I../xen/include/public/io
+
+obj-$(CONFIG_XEN_BLKDEV_TAP) := xenblktap.o
+
+xenblktap-y := xenbus.o interface.o blktap.o 
diff -rpuN linux-2.6.18.8/drivers/xen/blktap/xenbus.c linux-2.6.18-xen-3.3.0/drivers/xen/blktap/xenbus.c
--- linux-2.6.18.8/drivers/xen/blktap/xenbus.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/blktap/xenbus.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,477 @@
+/* drivers/xen/blktap/xenbus.c
+ *
+ * Xenbus code for blktap
+ *
+ * Copyright (c) 2004-2005, Andrew Warfield and Julian Chesterfield
+ *
+ * Based on the blkback xenbus code:
+ *
+ * Copyright (C) 2005 Rusty Russell <rusty@rustcorp.com.au>
+ * Copyright (C) 2005 XenSource Ltd
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <stdarg.h>
+#include <linux/module.h>
+#include <linux/kthread.h>
+#include <xen/xenbus.h>
+#include "common.h"
+
+
+struct backend_info
+{
+	struct xenbus_device *dev;
+	blkif_t *blkif;
+	struct xenbus_watch backend_watch;
+	int xenbus_id;
+	int group_added;
+};
+
+
+static void connect(struct backend_info *);
+static int connect_ring(struct backend_info *);
+static int blktap_remove(struct xenbus_device *dev);
+static int blktap_probe(struct xenbus_device *dev,
+			 const struct xenbus_device_id *id);
+static void tap_backend_changed(struct xenbus_watch *, const char **,
+			    unsigned int);
+static void tap_frontend_changed(struct xenbus_device *dev,
+			     enum xenbus_state frontend_state);
+
+static int strsep_len(const char *str, char c, unsigned int len)
+{
+        unsigned int i;
+
+        for (i = 0; str[i]; i++)
+                if (str[i] == c) {
+                        if (len == 0)
+                                return i;
+                        len--;
+                }
+        return (len == 0) ? i : -ERANGE;
+}
+
+static long get_id(const char *str)
+{
+        int len,end;
+        const char *ptr;
+        char *tptr, num[10];
+	
+        len = strsep_len(str, '/', 2);
+        end = strlen(str);
+        if ( (len < 0) || (end < 0) ) return -1;
+	
+        ptr = str + len + 1;
+        strncpy(num,ptr,end - len);
+        tptr = num + (end - (len + 1));
+        *tptr = '\0';
+	DPRINTK("Get_id called for %s (%s)\n",str,num);
+	
+        return simple_strtol(num, NULL, 10);
+}				
+
+static int blktap_name(blkif_t *blkif, char *buf)
+{
+	char *devpath, *devname;
+	struct xenbus_device *dev = blkif->be->dev;
+
+	devpath = xenbus_read(XBT_NIL, dev->nodename, "dev", NULL);
+	if (IS_ERR(devpath)) 
+		return PTR_ERR(devpath);
+	
+	if ((devname = strstr(devpath, "/dev/")) != NULL)
+		devname += strlen("/dev/");
+	else
+		devname  = devpath;
+
+	snprintf(buf, TASK_COMM_LEN, "blktap.%d.%s", blkif->domid, devname);
+	kfree(devpath);
+	
+	return 0;
+}
+
+/****************************************************************
+ *  sysfs interface for I/O requests of blktap device
+ */
+
+#define VBD_SHOW(name, format, args...)					\
+	static ssize_t show_##name(struct device *_dev,			\
+				   struct device_attribute *attr,	\
+				   char *buf)				\
+	{								\
+		struct xenbus_device *dev = to_xenbus_device(_dev);	\
+		struct backend_info *be = dev->dev.driver_data;		\
+									\
+		return sprintf(buf, format, ##args);			\
+	}								\
+	static DEVICE_ATTR(name, S_IRUGO, show_##name, NULL)
+
+VBD_SHOW(oo_req,  "%d\n", be->blkif->st_oo_req);
+VBD_SHOW(rd_req,  "%d\n", be->blkif->st_rd_req);
+VBD_SHOW(wr_req,  "%d\n", be->blkif->st_wr_req);
+VBD_SHOW(rd_sect, "%d\n", be->blkif->st_rd_sect);
+VBD_SHOW(wr_sect, "%d\n", be->blkif->st_wr_sect);
+
+static struct attribute *tapstat_attrs[] = {
+	&dev_attr_oo_req.attr,
+	&dev_attr_rd_req.attr,
+	&dev_attr_wr_req.attr,
+	&dev_attr_rd_sect.attr,
+	&dev_attr_wr_sect.attr,
+	NULL
+};
+
+static struct attribute_group tapstat_group = {
+	.name = "statistics",
+	.attrs = tapstat_attrs,
+};
+
+int xentap_sysfs_addif(struct xenbus_device *dev)
+{
+	int err;
+	struct backend_info *be = dev->dev.driver_data;
+	err = sysfs_create_group(&dev->dev.kobj, &tapstat_group);
+	if (!err)
+		be->group_added = 1;
+	return err;
+}
+
+void xentap_sysfs_delif(struct xenbus_device *dev)
+{
+	struct backend_info *be = dev->dev.driver_data;
+	sysfs_remove_group(&dev->dev.kobj, &tapstat_group);
+	be->group_added = 0;
+}
+
+static int blktap_remove(struct xenbus_device *dev)
+{
+	struct backend_info *be = dev->dev.driver_data;
+
+	if (be->group_added)
+		xentap_sysfs_delif(be->dev);
+	if (be->backend_watch.node) {
+		unregister_xenbus_watch(&be->backend_watch);
+		kfree(be->backend_watch.node);
+		be->backend_watch.node = NULL;
+	}
+	if (be->blkif) {
+		if (be->blkif->xenblkd)
+			kthread_stop(be->blkif->xenblkd);
+		signal_tapdisk(be->blkif->dev_num);
+		tap_blkif_free(be->blkif);
+		be->blkif = NULL;
+	}
+	kfree(be);
+	dev->dev.driver_data = NULL;
+	return 0;
+}
+
+static void tap_update_blkif_status(blkif_t *blkif)
+{ 
+	int err;
+	char name[TASK_COMM_LEN];
+
+	/* Not ready to connect? */
+	if(!blkif->irq || !blkif->sectors) {
+		return;
+	} 
+
+	/* Already connected? */
+	if (blkif->be->dev->state == XenbusStateConnected)
+		return;
+
+	/* Attempt to connect: exit if we fail to. */
+	connect(blkif->be);
+	if (blkif->be->dev->state != XenbusStateConnected)
+		return;
+
+	err = blktap_name(blkif, name);
+	if (err) {
+		xenbus_dev_error(blkif->be->dev, err, "get blktap dev name");
+		return;
+	}
+
+	if (!blkif->be->group_added) {
+		err = xentap_sysfs_addif(blkif->be->dev);
+		if (err) {
+			xenbus_dev_fatal(blkif->be->dev, err, 
+					 "creating sysfs entries");
+			return;
+		}
+	}
+
+	blkif->xenblkd = kthread_run(tap_blkif_schedule, blkif, name);
+	if (IS_ERR(blkif->xenblkd)) {
+		err = PTR_ERR(blkif->xenblkd);
+		blkif->xenblkd = NULL;
+		xenbus_dev_fatal(blkif->be->dev, err, "start xenblkd");
+		WPRINTK("Error starting thread\n");
+	}
+}
+
+/**
+ * Entry point to this code when a new device is created.  Allocate
+ * the basic structures, and watch the store waiting for the
+ * user-space program to tell us the physical device info.  Switch to
+ * InitWait.
+ */
+static int blktap_probe(struct xenbus_device *dev,
+			 const struct xenbus_device_id *id)
+{
+	int err;
+	struct backend_info *be = kzalloc(sizeof(struct backend_info),
+					  GFP_KERNEL);
+	if (!be) {
+		xenbus_dev_fatal(dev, -ENOMEM,
+				 "allocating backend structure");
+		return -ENOMEM;
+	}
+
+	be->dev = dev;
+	dev->dev.driver_data = be;
+	be->xenbus_id = get_id(dev->nodename);
+
+	be->blkif = tap_alloc_blkif(dev->otherend_id);
+	if (IS_ERR(be->blkif)) {
+		err = PTR_ERR(be->blkif);
+		be->blkif = NULL;
+		xenbus_dev_fatal(dev, err, "creating block interface");
+		goto fail;
+	}
+
+	/* setup back pointer */
+	be->blkif->be = be;
+	be->blkif->sectors = 0;
+
+	/* set a watch on disk info, waiting for userspace to update details*/
+	err = xenbus_watch_path2(dev, dev->nodename, "info",
+				 &be->backend_watch, tap_backend_changed);
+	if (err)
+		goto fail;
+	
+	err = xenbus_switch_state(dev, XenbusStateInitWait);
+	if (err)
+		goto fail;
+	return 0;
+
+fail:
+	DPRINTK("blktap probe failed\n");
+	blktap_remove(dev);
+	return err;
+}
+
+
+/**
+ * Callback received when the user space code has placed the device
+ * information in xenstore. 
+ */
+static void tap_backend_changed(struct xenbus_watch *watch,
+			    const char **vec, unsigned int len)
+{
+	int err;
+	unsigned long info;
+	struct backend_info *be
+		= container_of(watch, struct backend_info, backend_watch);
+	struct xenbus_device *dev = be->dev;
+	
+	/** 
+	 * Check to see whether userspace code has opened the image 
+	 * and written sector
+	 * and disk info to xenstore
+	 */
+	err = xenbus_gather(XBT_NIL, dev->nodename, "info", "%lu", &info, 
+			    NULL);
+	if (XENBUS_EXIST_ERR(err))
+		return;
+	if (err) {
+		xenbus_dev_error(dev, err, "getting info");
+		return;
+	}
+
+	DPRINTK("Userspace update on disk info, %lu\n",info);
+
+	err = xenbus_gather(XBT_NIL, dev->nodename, "sectors", "%llu", 
+			    &be->blkif->sectors, NULL);
+
+	/* Associate tap dev with domid*/
+	be->blkif->dev_num = dom_to_devid(be->blkif->domid, be->xenbus_id, 
+					  be->blkif);
+	DPRINTK("Thread started for domid [%d], connecting disk\n", 
+		be->blkif->dev_num);
+
+	tap_update_blkif_status(be->blkif);
+}
+
+/**
+ * Callback received when the frontend's state changes.
+ */
+static void tap_frontend_changed(struct xenbus_device *dev,
+			     enum xenbus_state frontend_state)
+{
+	struct backend_info *be = dev->dev.driver_data;
+	int err;
+
+	DPRINTK("\n");
+
+	switch (frontend_state) {
+	case XenbusStateInitialising:
+		if (dev->state == XenbusStateClosed) {
+			printk(KERN_INFO "%s: %s: prepare for reconnect\n",
+			       __FUNCTION__, dev->nodename);
+			xenbus_switch_state(dev, XenbusStateInitWait);
+		}
+		break;
+
+	case XenbusStateInitialised:
+	case XenbusStateConnected:
+		/* Ensure we connect even when two watches fire in 
+		   close successsion and we miss the intermediate value 
+		   of frontend_state. */
+		if (dev->state == XenbusStateConnected)
+			break;
+
+		err = connect_ring(be);
+		if (err)
+			break;
+		tap_update_blkif_status(be->blkif);
+		break;
+
+	case XenbusStateClosing:
+		if (be->blkif->xenblkd) {
+			kthread_stop(be->blkif->xenblkd);
+			be->blkif->xenblkd = NULL;
+		}
+		xenbus_switch_state(dev, XenbusStateClosing);
+		break;
+
+	case XenbusStateClosed:
+		xenbus_switch_state(dev, XenbusStateClosed);
+		if (xenbus_dev_is_online(dev))
+			break;
+		/* fall through if not online */
+	case XenbusStateUnknown:
+		device_unregister(&dev->dev);
+		break;
+
+	default:
+		xenbus_dev_fatal(dev, -EINVAL, "saw state %d at frontend",
+				 frontend_state);
+		break;
+	}
+}
+
+
+/**
+ * Switch to Connected state.
+ */
+static void connect(struct backend_info *be)
+{
+	int err;
+
+	struct xenbus_device *dev = be->dev;
+
+	err = xenbus_switch_state(dev, XenbusStateConnected);
+	if (err)
+		xenbus_dev_fatal(dev, err, "switching to Connected state",
+				 dev->nodename);
+
+	return;
+}
+
+
+static int connect_ring(struct backend_info *be)
+{
+	struct xenbus_device *dev = be->dev;
+	unsigned long ring_ref;
+	unsigned int evtchn;
+	char protocol[64];
+	int err;
+
+	DPRINTK("%s\n", dev->otherend);
+
+	err = xenbus_gather(XBT_NIL, dev->otherend, "ring-ref", "%lu", 
+			    &ring_ref, "event-channel", "%u", &evtchn, NULL);
+	if (err) {
+		xenbus_dev_fatal(dev, err,
+				 "reading %s/ring-ref and event-channel",
+				 dev->otherend);
+		return err;
+	}
+
+	be->blkif->blk_protocol = BLKIF_PROTOCOL_NATIVE;
+	err = xenbus_gather(XBT_NIL, dev->otherend, "protocol",
+			    "%63s", protocol, NULL);
+	if (err)
+		strcpy(protocol, "unspecified, assuming native");
+	else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_NATIVE))
+		be->blkif->blk_protocol = BLKIF_PROTOCOL_NATIVE;
+	else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_X86_32))
+		be->blkif->blk_protocol = BLKIF_PROTOCOL_X86_32;
+	else if (0 == strcmp(protocol, XEN_IO_PROTO_ABI_X86_64))
+		be->blkif->blk_protocol = BLKIF_PROTOCOL_X86_64;
+	else {
+		xenbus_dev_fatal(dev, err, "unknown fe protocol %s", protocol);
+		return -1;
+	}
+	printk(KERN_INFO
+	       "blktap: ring-ref %ld, event-channel %d, protocol %d (%s)\n",
+	       ring_ref, evtchn, be->blkif->blk_protocol, protocol);
+
+	/* Map the shared frame, irq etc. */
+	err = tap_blkif_map(be->blkif, ring_ref, evtchn);
+	if (err) {
+		xenbus_dev_fatal(dev, err, "mapping ring-ref %lu port %u",
+				 ring_ref, evtchn);
+		return err;
+	} 
+
+	return 0;
+}
+
+
+/* ** Driver Registration ** */
+
+
+static const struct xenbus_device_id blktap_ids[] = {
+	{ "tap" },
+	{ "" }
+};
+
+
+static struct xenbus_driver blktap = {
+	.name = "tap",
+	.owner = THIS_MODULE,
+	.ids = blktap_ids,
+	.probe = blktap_probe,
+	.remove = blktap_remove,
+	.otherend_changed = tap_frontend_changed
+};
+
+
+void tap_blkif_xenbus_init(void)
+{
+	xenbus_register_backend(&blktap);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/char/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/char/Makefile
--- linux-2.6.18.8/drivers/xen/char/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/char/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1 @@
+obj-$(CONFIG_XEN_DEVMEM)	:= mem.o
diff -rpuN linux-2.6.18.8/drivers/xen/char/mem.c linux-2.6.18-xen-3.3.0/drivers/xen/char/mem.c
--- linux-2.6.18.8/drivers/xen/char/mem.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/char/mem.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,190 @@
+/*
+ *  Originally from linux/drivers/char/mem.c
+ *
+ *  Copyright (C) 1991, 1992  Linus Torvalds
+ *
+ *  Added devfs support. 
+ *    Jan-11-1998, C. Scott Ananian <cananian@alumni.princeton.edu>
+ *  Shared /dev/zero mmaping support, Feb 2000, Kanoj Sarcar <kanoj@sgi.com>
+ */
+
+#include <linux/mm.h>
+#include <linux/miscdevice.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <linux/mman.h>
+#include <linux/random.h>
+#include <linux/init.h>
+#include <linux/raw.h>
+#include <linux/tty.h>
+#include <linux/capability.h>
+#include <linux/smp_lock.h>
+#include <linux/ptrace.h>
+#include <linux/device.h>
+#include <asm/pgalloc.h>
+#include <asm/uaccess.h>
+#include <asm/io.h>
+#include <asm/hypervisor.h>
+
+static inline int uncached_access(struct file *file)
+{
+	if (file->f_flags & O_SYNC)
+		return 1;
+	/* Xen sets correct MTRR type on non-RAM for us. */
+	return 0;
+}
+
+/*
+ * This funcion reads the *physical* memory. The f_pos points directly to the 
+ * memory location. 
+ */
+static ssize_t read_mem(struct file * file, char __user * buf,
+			size_t count, loff_t *ppos)
+{
+	unsigned long p = *ppos, ignored;
+	ssize_t read = 0, sz;
+	void __iomem *v;
+
+	while (count > 0) {
+		/*
+		 * Handle first page in case it's not aligned
+		 */
+		if (-p & (PAGE_SIZE - 1))
+			sz = -p & (PAGE_SIZE - 1);
+		else
+			sz = PAGE_SIZE;
+
+		sz = min_t(unsigned long, sz, count);
+
+		v = ioremap(p, sz);
+		if (IS_ERR(v) || v == NULL) {
+			/*
+			 * Some programs (e.g., dmidecode) groove off into
+			 * weird RAM areas where no tables can possibly exist
+			 * (because Xen will have stomped on them!). These
+			 * programs get rather upset if we let them know that
+			 * Xen failed their access, so we fake out a read of
+			 * all zeroes.
+			 */
+			if (clear_user(buf, count))
+				return -EFAULT;
+			read += count;
+			break;
+		}
+
+		ignored = copy_to_user(buf, v, sz);
+		iounmap(v);
+		if (ignored)
+			return -EFAULT;
+		buf += sz;
+		p += sz;
+		count -= sz;
+		read += sz;
+	}
+
+	*ppos += read;
+	return read;
+}
+
+static ssize_t write_mem(struct file * file, const char __user * buf, 
+			 size_t count, loff_t *ppos)
+{
+	unsigned long p = *ppos, ignored;
+	ssize_t written = 0, sz;
+	void __iomem *v;
+
+	while (count > 0) {
+		/*
+		 * Handle first page in case it's not aligned
+		 */
+		if (-p & (PAGE_SIZE - 1))
+			sz = -p & (PAGE_SIZE - 1);
+		else
+			sz = PAGE_SIZE;
+
+		sz = min_t(unsigned long, sz, count);
+
+		v = ioremap(p, sz);
+		if (v == NULL)
+			break;
+		if (IS_ERR(v)) {
+			if (written == 0)
+				return PTR_ERR(v);
+			break;
+		}
+
+		ignored = copy_from_user(v, buf, sz);
+		iounmap(v);
+		if (ignored) {
+			written += sz - ignored;
+			if (written)
+				break;
+			return -EFAULT;
+		}
+		buf += sz;
+		p += sz;
+		count -= sz;
+		written += sz;
+	}
+
+	*ppos += written;
+	return written;
+}
+
+#ifndef ARCH_HAS_DEV_MEM_MMAP_MEM
+static int xen_mmap_mem(struct file * file, struct vm_area_struct * vma)
+{
+	size_t size = vma->vm_end - vma->vm_start;
+
+	if (uncached_access(file))
+		vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
+
+	/* We want to return the real error code, not EAGAIN. */
+	return direct_remap_pfn_range(vma, vma->vm_start, vma->vm_pgoff,
+				      size, vma->vm_page_prot, DOMID_IO);
+}
+#endif
+
+/*
+ * The memory devices use the full 32/64 bits of the offset, and so we cannot
+ * check against negative addresses: they are ok. The return value is weird,
+ * though, in that case (0).
+ *
+ * also note that seeking relative to the "end of file" isn't supported:
+ * it has no meaning, so it returns -EINVAL.
+ */
+static loff_t memory_lseek(struct file * file, loff_t offset, int orig)
+{
+	loff_t ret;
+
+	mutex_lock(&file->f_dentry->d_inode->i_mutex);
+	switch (orig) {
+		case 0:
+			file->f_pos = offset;
+			ret = file->f_pos;
+			force_successful_syscall_return();
+			break;
+		case 1:
+			file->f_pos += offset;
+			ret = file->f_pos;
+			force_successful_syscall_return();
+			break;
+		default:
+			ret = -EINVAL;
+	}
+	mutex_unlock(&file->f_dentry->d_inode->i_mutex);
+	return ret;
+}
+
+static int open_mem(struct inode * inode, struct file * filp)
+{
+	return capable(CAP_SYS_RAWIO) ? 0 : -EPERM;
+}
+
+const struct file_operations mem_fops = {
+	.llseek		= memory_lseek,
+	.read		= read_mem,
+	.write		= write_mem,
+	.mmap		= xen_mmap_mem,
+	.open		= open_mem,
+};
diff -rpuN linux-2.6.18.8/drivers/xen/console/console.c linux-2.6.18-xen-3.3.0/drivers/xen/console/console.c
--- linux-2.6.18.8/drivers/xen/console/console.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/console/console.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,731 @@
+/******************************************************************************
+ * console.c
+ * 
+ * Virtual console driver.
+ * 
+ * Copyright (c) 2002-2004, K A Fraser.
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/version.h>
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/signal.h>
+#include <linux/sched.h>
+#include <linux/interrupt.h>
+#include <linux/tty.h>
+#include <linux/tty_flip.h>
+#include <linux/serial.h>
+#include <linux/major.h>
+#include <linux/ptrace.h>
+#include <linux/ioport.h>
+#include <linux/mm.h>
+#include <linux/slab.h>
+#include <linux/init.h>
+#include <linux/console.h>
+#include <linux/bootmem.h>
+#include <linux/sysrq.h>
+#include <linux/screen_info.h>
+#include <linux/vt.h>
+#include <asm/io.h>
+#include <asm/irq.h>
+#include <asm/uaccess.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/event_channel.h>
+#include <asm/hypervisor.h>
+#include <xen/evtchn.h>
+#include <xen/xenbus.h>
+#include <xen/xencons.h>
+
+/*
+ * Modes:
+ *  'xencons=off'  [XC_OFF]:     Console is disabled.
+ *  'xencons=tty'  [XC_TTY]:     Console attached to '/dev/tty[0-9]+'.
+ *  'xencons=ttyS' [XC_SERIAL]:  Console attached to '/dev/ttyS[0-9]+'.
+ *  'xencons=xvc'  [XC_XVC]:     Console attached to '/dev/xvc0'.
+ *  default:                     XC_XVC
+ * 
+ * NB. In mode XC_TTY, we create dummy consoles for tty2-63. This suppresses
+ * warnings from standard distro startup scripts.
+ */
+static enum {
+	XC_OFF, XC_TTY, XC_SERIAL, XC_XVC
+} xc_mode = XC_XVC;
+static int xc_num = -1;
+
+/* /dev/xvc0 device number allocated by lanana.org. */
+#define XEN_XVC_MAJOR 204
+#define XEN_XVC_MINOR 191
+
+#ifdef CONFIG_MAGIC_SYSRQ
+static unsigned long sysrq_requested;
+extern int sysrq_enabled;
+#endif
+
+static int __init xencons_setup(char *str)
+{
+	char *q;
+	int n;
+	extern int console_use_vt;
+
+	console_use_vt = 1;
+	if (!strncmp(str, "ttyS", 4)) {
+		xc_mode = XC_SERIAL;
+		str += 4;
+	} else if (!strncmp(str, "tty", 3)) {
+		xc_mode = XC_TTY;
+		str += 3;
+		console_use_vt = 0;
+	} else if (!strncmp(str, "xvc", 3)) {
+		xc_mode = XC_XVC;
+		str += 3;
+	} else if (!strncmp(str, "off", 3)) {
+		xc_mode = XC_OFF;
+		str += 3;
+	}
+
+	n = simple_strtol(str, &q, 10);
+	if (q != str)
+		xc_num = n;
+
+	return 1;
+}
+__setup("xencons=", xencons_setup);
+
+/* The kernel and user-land drivers share a common transmit buffer. */
+static unsigned int wbuf_size = 4096;
+#define WBUF_MASK(_i) ((_i)&(wbuf_size-1))
+static char *wbuf;
+static unsigned int wc, wp; /* write_cons, write_prod */
+
+static int __init xencons_bufsz_setup(char *str)
+{
+	unsigned int goal;
+	goal = simple_strtoul(str, NULL, 0);
+	if (goal) {
+		goal = roundup_pow_of_two(goal);
+		if (wbuf_size < goal)
+			wbuf_size = goal;
+	}
+	return 1;
+}
+__setup("xencons_bufsz=", xencons_bufsz_setup);
+
+/* This lock protects accesses to the common transmit buffer. */
+static DEFINE_SPINLOCK(xencons_lock);
+
+/* Common transmit-kick routine. */
+static void __xencons_tx_flush(void);
+
+static struct tty_driver *xencons_driver;
+
+/******************** Kernel console driver ********************************/
+
+static void kcons_write(struct console *c, const char *s, unsigned int count)
+{
+	int           i = 0;
+	unsigned long flags;
+
+	spin_lock_irqsave(&xencons_lock, flags);
+
+	while (i < count) {
+		for (; i < count; i++) {
+			if ((wp - wc) >= (wbuf_size - 1))
+				break;
+			if ((wbuf[WBUF_MASK(wp++)] = s[i]) == '\n')
+				wbuf[WBUF_MASK(wp++)] = '\r';
+		}
+
+		__xencons_tx_flush();
+	}
+
+	spin_unlock_irqrestore(&xencons_lock, flags);
+}
+
+static void kcons_write_dom0(struct console *c, const char *s, unsigned int count)
+{
+
+	while (count > 0) {
+		int rc;
+		rc = HYPERVISOR_console_io( CONSOLEIO_write, count, (char *)s);
+		if (rc <= 0)
+			break;
+		count -= rc;
+		s += rc;
+	}
+}
+
+static struct tty_driver *kcons_device(struct console *c, int *index)
+{
+	*index = 0;
+	return xencons_driver;
+}
+
+static struct console kcons_info = {
+	.device	= kcons_device,
+	.flags	= CON_PRINTBUFFER | CON_ENABLED,
+	.index	= -1,
+};
+
+static int __init xen_console_init(void)
+{
+	if (!is_running_on_xen())
+		goto out;
+
+	if (is_initial_xendomain()) {
+		kcons_info.write = kcons_write_dom0;
+	} else {
+		if (!xen_start_info->console.domU.evtchn)
+			goto out;
+		kcons_info.write = kcons_write;
+	}
+
+	switch (xc_mode) {
+	case XC_XVC:
+		strcpy(kcons_info.name, "xvc");
+		if (xc_num == -1)
+			xc_num = 0;
+		break;
+
+	case XC_SERIAL:
+		strcpy(kcons_info.name, "ttyS");
+		if (xc_num == -1)
+			xc_num = 0;
+		break;
+
+	case XC_TTY:
+		strcpy(kcons_info.name, "tty");
+		if (xc_num == -1)
+			xc_num = 1;
+		break;
+
+	default:
+		goto out;
+	}
+
+	wbuf = alloc_bootmem(wbuf_size);
+
+	register_console(&kcons_info);
+
+ out:
+	return 0;
+}
+console_initcall(xen_console_init);
+
+/*** Useful function for console debugging -- goes straight to Xen. ***/
+asmlinkage int xprintk(const char *fmt, ...)
+{
+	va_list args;
+	int printk_len;
+	static char printk_buf[1024];
+
+	/* Emit the output into the temporary buffer */
+	va_start(args, fmt);
+	printk_len = vsnprintf(printk_buf, sizeof(printk_buf), fmt, args);
+	va_end(args);
+
+	/* Send the processed output directly to Xen. */
+	kcons_write_dom0(NULL, printk_buf, printk_len);
+
+	return 0;
+}
+
+/*** Forcibly flush console data before dying. ***/
+void xencons_force_flush(void)
+{
+	int sz;
+
+	/* Emergency console is synchronous, so there's nothing to flush. */
+	if (!is_running_on_xen() ||
+	    is_initial_xendomain() ||
+	    !xen_start_info->console.domU.evtchn)
+		return;
+
+	/* Spin until console data is flushed through to the daemon. */
+	while (wc != wp) {
+		int sent = 0;
+		if ((sz = wp - wc) == 0)
+			continue;
+		sent = xencons_ring_send(&wbuf[WBUF_MASK(wc)], sz);
+		if (sent > 0)
+			wc += sent;
+	}
+}
+
+
+void __init dom0_init_screen_info(const struct dom0_vga_console_info *info, size_t size)
+{
+	/* This is drawn from a dump from vgacon:startup in
+	 * standard Linux. */
+	screen_info.orig_video_mode = 3;
+	screen_info.orig_video_isVGA = 1;
+	screen_info.orig_video_lines = 25;
+	screen_info.orig_video_cols = 80;
+	screen_info.orig_video_ega_bx = 3;
+	screen_info.orig_video_points = 16;
+	screen_info.orig_y = screen_info.orig_video_lines - 1;
+
+	switch (info->video_type) {
+	case XEN_VGATYPE_TEXT_MODE_3:
+		if (size < offsetof(struct dom0_vga_console_info, u.text_mode_3)
+		           + sizeof(info->u.text_mode_3))
+			break;
+		screen_info.orig_video_lines = info->u.text_mode_3.rows;
+		screen_info.orig_video_cols = info->u.text_mode_3.columns;
+		screen_info.orig_x = info->u.text_mode_3.cursor_x;
+		screen_info.orig_y = info->u.text_mode_3.cursor_y;
+		screen_info.orig_video_points =
+			info->u.text_mode_3.font_height;
+		break;
+
+	case XEN_VGATYPE_VESA_LFB:
+		if (size < offsetof(struct dom0_vga_console_info,
+		                    u.vesa_lfb.gbl_caps))
+			break;
+		screen_info.orig_video_isVGA = VIDEO_TYPE_VLFB;
+		screen_info.lfb_width = info->u.vesa_lfb.width;
+		screen_info.lfb_height = info->u.vesa_lfb.height;
+		screen_info.lfb_depth = info->u.vesa_lfb.bits_per_pixel;
+		screen_info.lfb_base = info->u.vesa_lfb.lfb_base;
+		screen_info.lfb_size = info->u.vesa_lfb.lfb_size;
+		screen_info.lfb_linelength = info->u.vesa_lfb.bytes_per_line;
+		screen_info.red_size = info->u.vesa_lfb.red_size;
+		screen_info.red_pos = info->u.vesa_lfb.red_pos;
+		screen_info.green_size = info->u.vesa_lfb.green_size;
+		screen_info.green_pos = info->u.vesa_lfb.green_pos;
+		screen_info.blue_size = info->u.vesa_lfb.blue_size;
+		screen_info.blue_pos = info->u.vesa_lfb.blue_pos;
+		screen_info.rsvd_size = info->u.vesa_lfb.rsvd_size;
+		screen_info.rsvd_pos = info->u.vesa_lfb.rsvd_pos;
+		if (size >= offsetof(struct dom0_vga_console_info,
+		                     u.vesa_lfb.gbl_caps)
+		            + sizeof(info->u.vesa_lfb.gbl_caps))
+			screen_info.capabilities = info->u.vesa_lfb.gbl_caps;
+		if (size >= offsetof(struct dom0_vga_console_info,
+		                     u.vesa_lfb.mode_attrs)
+		            + sizeof(info->u.vesa_lfb.mode_attrs))
+			screen_info.vesa_attributes = info->u.vesa_lfb.mode_attrs;
+		break;
+	}
+}
+
+
+/******************** User-space console driver (/dev/console) ************/
+
+#define DRV(_d)         (_d)
+#define DUMMY_TTY(_tty) ((xc_mode == XC_TTY) &&		\
+			 ((_tty)->index != (xc_num - 1)))
+
+static struct termios *xencons_termios[MAX_NR_CONSOLES];
+static struct termios *xencons_termios_locked[MAX_NR_CONSOLES];
+static struct tty_struct *xencons_tty;
+static int xencons_priv_irq;
+static char x_char;
+
+void xencons_rx(char *buf, unsigned len, struct pt_regs *regs)
+{
+	int           i;
+	unsigned long flags;
+
+	spin_lock_irqsave(&xencons_lock, flags);
+	if (xencons_tty == NULL)
+		goto out;
+
+	for (i = 0; i < len; i++) {
+#ifdef CONFIG_MAGIC_SYSRQ
+		if (sysrq_enabled) {
+			if (buf[i] == '\x0f') { /* ^O */
+				if (!sysrq_requested) {
+					sysrq_requested = jiffies;
+					continue; /* don't print sysrq key */
+				}
+				sysrq_requested = 0;
+			} else if (sysrq_requested) {
+				unsigned long sysrq_timeout =
+					sysrq_requested + HZ*2;
+				sysrq_requested = 0;
+				if (time_before(jiffies, sysrq_timeout)) {
+					spin_unlock_irqrestore(
+						&xencons_lock, flags);
+					handle_sysrq(
+						buf[i], regs, xencons_tty);
+					spin_lock_irqsave(
+						&xencons_lock, flags);
+					continue;
+				}
+			}
+		}
+#endif
+		tty_insert_flip_char(xencons_tty, buf[i], 0);
+	}
+	tty_flip_buffer_push(xencons_tty);
+
+ out:
+	spin_unlock_irqrestore(&xencons_lock, flags);
+}
+
+static void __xencons_tx_flush(void)
+{
+	int sent, sz, work_done = 0;
+
+	if (x_char) {
+		if (is_initial_xendomain())
+			kcons_write_dom0(NULL, &x_char, 1);
+		else
+			while (x_char)
+				if (xencons_ring_send(&x_char, 1) == 1)
+					break;
+		x_char = 0;
+		work_done = 1;
+	}
+
+	while (wc != wp) {
+		sz = wp - wc;
+		if (sz > (wbuf_size - WBUF_MASK(wc)))
+			sz = wbuf_size - WBUF_MASK(wc);
+		if (is_initial_xendomain()) {
+			kcons_write_dom0(NULL, &wbuf[WBUF_MASK(wc)], sz);
+			wc += sz;
+		} else {
+			sent = xencons_ring_send(&wbuf[WBUF_MASK(wc)], sz);
+			if (sent == 0)
+				break;
+			wc += sent;
+		}
+		work_done = 1;
+	}
+
+	if (work_done && (xencons_tty != NULL)) {
+		wake_up_interruptible(&xencons_tty->write_wait);
+		if ((xencons_tty->flags & (1 << TTY_DO_WRITE_WAKEUP)) &&
+		    (xencons_tty->ldisc.write_wakeup != NULL))
+			(xencons_tty->ldisc.write_wakeup)(xencons_tty);
+	}
+}
+
+void xencons_tx(void)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&xencons_lock, flags);
+	__xencons_tx_flush();
+	spin_unlock_irqrestore(&xencons_lock, flags);
+}
+
+/* Privileged receive callback and transmit kicker. */
+static irqreturn_t xencons_priv_interrupt(int irq, void *dev_id,
+					  struct pt_regs *regs)
+{
+	static char rbuf[16];
+	int         l;
+
+	while ((l = HYPERVISOR_console_io(CONSOLEIO_read, 16, rbuf)) > 0)
+		xencons_rx(rbuf, l, regs);
+
+	xencons_tx();
+
+	return IRQ_HANDLED;
+}
+
+static int xencons_write_room(struct tty_struct *tty)
+{
+	return wbuf_size - (wp - wc);
+}
+
+static int xencons_chars_in_buffer(struct tty_struct *tty)
+{
+	return wp - wc;
+}
+
+static void xencons_send_xchar(struct tty_struct *tty, char ch)
+{
+	unsigned long flags;
+
+	if (DUMMY_TTY(tty))
+		return;
+
+	spin_lock_irqsave(&xencons_lock, flags);
+	x_char = ch;
+	__xencons_tx_flush();
+	spin_unlock_irqrestore(&xencons_lock, flags);
+}
+
+static void xencons_throttle(struct tty_struct *tty)
+{
+	if (DUMMY_TTY(tty))
+		return;
+
+	if (I_IXOFF(tty))
+		xencons_send_xchar(tty, STOP_CHAR(tty));
+}
+
+static void xencons_unthrottle(struct tty_struct *tty)
+{
+	if (DUMMY_TTY(tty))
+		return;
+
+	if (I_IXOFF(tty)) {
+		if (x_char != 0)
+			x_char = 0;
+		else
+			xencons_send_xchar(tty, START_CHAR(tty));
+	}
+}
+
+static void xencons_flush_buffer(struct tty_struct *tty)
+{
+	unsigned long flags;
+
+	if (DUMMY_TTY(tty))
+		return;
+
+	spin_lock_irqsave(&xencons_lock, flags);
+	wc = wp = 0;
+	spin_unlock_irqrestore(&xencons_lock, flags);
+}
+
+static inline int __xencons_put_char(int ch)
+{
+	char _ch = (char)ch;
+	if ((wp - wc) == wbuf_size)
+		return 0;
+	wbuf[WBUF_MASK(wp++)] = _ch;
+	return 1;
+}
+
+static int xencons_write(
+	struct tty_struct *tty,
+	const unsigned char *buf,
+	int count)
+{
+	int i;
+	unsigned long flags;
+
+	if (DUMMY_TTY(tty))
+		return count;
+
+	spin_lock_irqsave(&xencons_lock, flags);
+
+	for (i = 0; i < count; i++)
+		if (!__xencons_put_char(buf[i]))
+			break;
+
+	if (i != 0)
+		__xencons_tx_flush();
+
+	spin_unlock_irqrestore(&xencons_lock, flags);
+
+	return i;
+}
+
+static void xencons_put_char(struct tty_struct *tty, u_char ch)
+{
+	unsigned long flags;
+
+	if (DUMMY_TTY(tty))
+		return;
+
+	spin_lock_irqsave(&xencons_lock, flags);
+	(void)__xencons_put_char(ch);
+	spin_unlock_irqrestore(&xencons_lock, flags);
+}
+
+static void xencons_flush_chars(struct tty_struct *tty)
+{
+	unsigned long flags;
+
+	if (DUMMY_TTY(tty))
+		return;
+
+	spin_lock_irqsave(&xencons_lock, flags);
+	__xencons_tx_flush();
+	spin_unlock_irqrestore(&xencons_lock, flags);
+}
+
+static void xencons_wait_until_sent(struct tty_struct *tty, int timeout)
+{
+	unsigned long orig_jiffies = jiffies;
+
+	if (DUMMY_TTY(tty))
+		return;
+
+	while (DRV(tty->driver)->chars_in_buffer(tty)) {
+		set_current_state(TASK_INTERRUPTIBLE);
+		schedule_timeout(1);
+		if (signal_pending(current))
+			break;
+		if (timeout && time_after(jiffies, orig_jiffies + timeout))
+			break;
+	}
+
+	set_current_state(TASK_RUNNING);
+}
+
+static int xencons_open(struct tty_struct *tty, struct file *filp)
+{
+	unsigned long flags;
+
+	if (DUMMY_TTY(tty))
+		return 0;
+
+	spin_lock_irqsave(&xencons_lock, flags);
+	tty->driver_data = NULL;
+	if (xencons_tty == NULL)
+		xencons_tty = tty;
+	__xencons_tx_flush();
+	spin_unlock_irqrestore(&xencons_lock, flags);
+
+	return 0;
+}
+
+static void xencons_close(struct tty_struct *tty, struct file *filp)
+{
+	unsigned long flags;
+
+	if (DUMMY_TTY(tty))
+		return;
+
+	mutex_lock(&tty_mutex);
+
+	if (tty->count != 1) {
+		mutex_unlock(&tty_mutex);
+		return;
+	}
+
+	/* Prevent other threads from re-opening this tty. */
+	set_bit(TTY_CLOSING, &tty->flags);
+	mutex_unlock(&tty_mutex);
+
+	tty->closing = 1;
+	tty_wait_until_sent(tty, 0);
+	if (DRV(tty->driver)->flush_buffer != NULL)
+		DRV(tty->driver)->flush_buffer(tty);
+	if (tty->ldisc.flush_buffer != NULL)
+		tty->ldisc.flush_buffer(tty);
+	tty->closing = 0;
+	spin_lock_irqsave(&xencons_lock, flags);
+	xencons_tty = NULL;
+	spin_unlock_irqrestore(&xencons_lock, flags);
+}
+
+static struct tty_operations xencons_ops = {
+	.open = xencons_open,
+	.close = xencons_close,
+	.write = xencons_write,
+	.write_room = xencons_write_room,
+	.put_char = xencons_put_char,
+	.flush_chars = xencons_flush_chars,
+	.chars_in_buffer = xencons_chars_in_buffer,
+	.send_xchar = xencons_send_xchar,
+	.flush_buffer = xencons_flush_buffer,
+	.throttle = xencons_throttle,
+	.unthrottle = xencons_unthrottle,
+	.wait_until_sent = xencons_wait_until_sent,
+};
+
+static int __init xencons_init(void)
+{
+	int rc;
+
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	if (xc_mode == XC_OFF)
+		return 0;
+
+	if (!is_initial_xendomain()) {
+		rc = xencons_ring_init();
+		if (rc)
+			return rc;
+	}
+
+	xencons_driver = alloc_tty_driver((xc_mode == XC_TTY) ?
+					  MAX_NR_CONSOLES : 1);
+	if (xencons_driver == NULL)
+		return -ENOMEM;
+
+	DRV(xencons_driver)->name            = "xencons";
+	DRV(xencons_driver)->major           = TTY_MAJOR;
+	DRV(xencons_driver)->type            = TTY_DRIVER_TYPE_SERIAL;
+	DRV(xencons_driver)->subtype         = SERIAL_TYPE_NORMAL;
+	DRV(xencons_driver)->init_termios    = tty_std_termios;
+	DRV(xencons_driver)->flags           =
+		TTY_DRIVER_REAL_RAW |
+		TTY_DRIVER_RESET_TERMIOS;
+	DRV(xencons_driver)->termios         = xencons_termios;
+	DRV(xencons_driver)->termios_locked  = xencons_termios_locked;
+
+	switch (xc_mode) {
+	case XC_XVC:
+		DRV(xencons_driver)->name        = "xvc";
+		DRV(xencons_driver)->major       = XEN_XVC_MAJOR;
+		DRV(xencons_driver)->minor_start = XEN_XVC_MINOR;
+		DRV(xencons_driver)->name_base   = xc_num;
+		break;
+	case XC_SERIAL:
+		DRV(xencons_driver)->name        = "ttyS";
+		DRV(xencons_driver)->minor_start = 64 + xc_num;
+		DRV(xencons_driver)->name_base   = xc_num;
+		break;
+	default:
+		DRV(xencons_driver)->name        = "tty";
+		DRV(xencons_driver)->minor_start = 1;
+		DRV(xencons_driver)->name_base   = 1;
+		break;
+	}
+
+	tty_set_operations(xencons_driver, &xencons_ops);
+
+	if ((rc = tty_register_driver(DRV(xencons_driver))) != 0) {
+		printk("WARNING: Failed to register Xen virtual "
+		       "console driver as '%s%d'\n",
+		       DRV(xencons_driver)->name,
+		       DRV(xencons_driver)->name_base);
+		put_tty_driver(xencons_driver);
+		xencons_driver = NULL;
+		return rc;
+	}
+
+	if (is_initial_xendomain()) {
+		xencons_priv_irq = bind_virq_to_irqhandler(
+			VIRQ_CONSOLE,
+			0,
+			xencons_priv_interrupt,
+			0,
+			"console",
+			NULL);
+		BUG_ON(xencons_priv_irq < 0);
+	}
+
+	printk("Xen virtual console successfully installed as %s%d\n",
+	       DRV(xencons_driver)->name, xc_num);
+
+	return 0;
+}
+
+module_init(xencons_init);
+
+MODULE_LICENSE("Dual BSD/GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/console/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/console/Makefile
--- linux-2.6.18.8/drivers/xen/console/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/console/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,2 @@
+
+obj-y	:= console.o xencons_ring.o
diff -rpuN linux-2.6.18.8/drivers/xen/console/xencons_ring.c linux-2.6.18-xen-3.3.0/drivers/xen/console/xencons_ring.c
--- linux-2.6.18.8/drivers/xen/console/xencons_ring.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/console/xencons_ring.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,143 @@
+/* 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/version.h>
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/signal.h>
+#include <linux/sched.h>
+#include <linux/interrupt.h>
+#include <linux/tty.h>
+#include <linux/tty_flip.h>
+#include <linux/serial.h>
+#include <linux/major.h>
+#include <linux/ptrace.h>
+#include <linux/ioport.h>
+#include <linux/mm.h>
+#include <linux/slab.h>
+
+#include <asm/hypervisor.h>
+#include <xen/evtchn.h>
+#include <xen/xencons.h>
+#include <linux/wait.h>
+#include <linux/interrupt.h>
+#include <linux/sched.h>
+#include <linux/err.h>
+#include <xen/interface/io/console.h>
+
+static int xencons_irq;
+
+static inline struct xencons_interface *xencons_interface(void)
+{
+	return mfn_to_virt(xen_start_info->console.domU.mfn);
+}
+
+static inline void notify_daemon(void)
+{
+	/* Use evtchn: this is called early, before irq is set up. */
+	notify_remote_via_evtchn(xen_start_info->console.domU.evtchn);
+}
+
+int xencons_ring_send(const char *data, unsigned len)
+{
+	int sent = 0;
+	struct xencons_interface *intf = xencons_interface();
+	XENCONS_RING_IDX cons, prod;
+
+	cons = intf->out_cons;
+	prod = intf->out_prod;
+	mb();
+	BUG_ON((prod - cons) > sizeof(intf->out));
+
+	while ((sent < len) && ((prod - cons) < sizeof(intf->out)))
+		intf->out[MASK_XENCONS_IDX(prod++, intf->out)] = data[sent++];
+
+	wmb();
+	intf->out_prod = prod;
+
+	notify_daemon();
+
+	return sent;
+}
+
+static irqreturn_t handle_input(int irq, void *unused, struct pt_regs *regs)
+{
+	struct xencons_interface *intf = xencons_interface();
+	XENCONS_RING_IDX cons, prod;
+
+	cons = intf->in_cons;
+	prod = intf->in_prod;
+	mb();
+	BUG_ON((prod - cons) > sizeof(intf->in));
+
+	while (cons != prod) {
+		xencons_rx(intf->in+MASK_XENCONS_IDX(cons,intf->in), 1, regs);
+		cons++;
+	}
+
+	mb();
+	intf->in_cons = cons;
+
+	notify_daemon();
+
+	xencons_tx();
+
+	return IRQ_HANDLED;
+}
+
+int xencons_ring_init(void)
+{
+	int irq;
+
+	if (xencons_irq)
+		unbind_from_irqhandler(xencons_irq, NULL);
+	xencons_irq = 0;
+
+	if (!is_running_on_xen() ||
+	    is_initial_xendomain() ||
+	    !xen_start_info->console.domU.evtchn)
+		return -ENODEV;
+
+	irq = bind_caller_port_to_irqhandler(
+		xen_start_info->console.domU.evtchn,
+		handle_input, 0, "xencons", NULL);
+	if (irq < 0) {
+		printk(KERN_ERR "XEN console request irq failed %i\n", irq);
+		return irq;
+	}
+
+	xencons_irq = irq;
+
+	/* In case we have in-flight data after save/restore... */
+	notify_daemon();
+
+	return 0;
+}
+
+void xencons_resume(void)
+{
+	(void)xencons_ring_init();
+}
diff -rpuN linux-2.6.18.8/drivers/xen/core/cpu_hotplug.c linux-2.6.18-xen-3.3.0/drivers/xen/core/cpu_hotplug.c
--- linux-2.6.18.8/drivers/xen/core/cpu_hotplug.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/core/cpu_hotplug.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,173 @@
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/notifier.h>
+#include <linux/cpu.h>
+#include <xen/cpu_hotplug.h>
+#include <xen/xenbus.h>
+
+/*
+ * Set of CPUs that remote admin software will allow us to bring online.
+ * Notified to us via xenbus.
+ */
+static cpumask_t xenbus_allowed_cpumask;
+
+/* Set of CPUs that local admin will allow us to bring online. */
+static cpumask_t local_allowed_cpumask = CPU_MASK_ALL;
+
+static int local_cpu_hotplug_request(void)
+{
+	/*
+	 * We assume a CPU hotplug request comes from local admin if it is made
+	 * via a userspace process (i.e., one with a real mm_struct).
+	 */
+	return (current->mm != NULL);
+}
+
+static void vcpu_hotplug(unsigned int cpu)
+{
+	int err;
+	char dir[32], state[32];
+
+	if ((cpu >= NR_CPUS) || !cpu_possible(cpu))
+		return;
+
+	sprintf(dir, "cpu/%u", cpu);
+	err = xenbus_scanf(XBT_NIL, dir, "availability", "%s", state);
+	if (err != 1) {
+		printk(KERN_ERR "XENBUS: Unable to read cpu state\n");
+		return;
+	}
+
+	if (strcmp(state, "online") == 0) {
+		cpu_set(cpu, xenbus_allowed_cpumask);
+		(void)cpu_up(cpu);
+	} else if (strcmp(state, "offline") == 0) {
+		cpu_clear(cpu, xenbus_allowed_cpumask);
+		(void)cpu_down(cpu);
+	} else {
+		printk(KERN_ERR "XENBUS: unknown state(%s) on CPU%d\n",
+		       state, cpu);
+	}
+}
+
+static void handle_vcpu_hotplug_event(
+	struct xenbus_watch *watch, const char **vec, unsigned int len)
+{
+	unsigned int cpu;
+	char *cpustr;
+	const char *node = vec[XS_WATCH_PATH];
+
+	if ((cpustr = strstr(node, "cpu/")) != NULL) {
+		sscanf(cpustr, "cpu/%u", &cpu);
+		vcpu_hotplug(cpu);
+	}
+}
+
+static int smpboot_cpu_notify(struct notifier_block *notifier,
+			      unsigned long action, void *hcpu)
+{
+	unsigned int cpu = (long)hcpu;
+
+	/*
+	 * We do this in a callback notifier rather than __cpu_disable()
+	 * because local_cpu_hotplug_request() does not work in the latter
+	 * as it's always executed from within a stopmachine kthread.
+	 */
+	if ((action == CPU_DOWN_PREPARE) && local_cpu_hotplug_request())
+		cpu_clear(cpu, local_allowed_cpumask);
+
+	return NOTIFY_OK;
+}
+
+static int setup_cpu_watcher(struct notifier_block *notifier,
+			      unsigned long event, void *data)
+{
+	unsigned int i;
+
+	static struct xenbus_watch cpu_watch = {
+		.node = "cpu",
+		.callback = handle_vcpu_hotplug_event,
+		.flags = XBWF_new_thread };
+	(void)register_xenbus_watch(&cpu_watch);
+
+	if (!is_initial_xendomain()) {
+		for_each_possible_cpu(i)
+			vcpu_hotplug(i);
+		printk(KERN_INFO "Brought up %ld CPUs\n",
+		       (long)num_online_cpus());
+	}
+
+	return NOTIFY_DONE;
+}
+
+static int __init setup_vcpu_hotplug_event(void)
+{
+	static struct notifier_block hotplug_cpu = {
+		.notifier_call = smpboot_cpu_notify };
+	static struct notifier_block xsn_cpu = {
+		.notifier_call = setup_cpu_watcher };
+
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	register_cpu_notifier(&hotplug_cpu);
+	register_xenstore_notifier(&xsn_cpu);
+
+	return 0;
+}
+
+arch_initcall(setup_vcpu_hotplug_event);
+
+int smp_suspend(void)
+{
+	unsigned int cpu;
+	int err;
+
+	for_each_online_cpu(cpu) {
+		if (cpu == 0)
+			continue;
+		err = cpu_down(cpu);
+		if (err) {
+			printk(KERN_CRIT "Failed to take all CPUs "
+			       "down: %d.\n", err);
+			for_each_possible_cpu(cpu)
+				vcpu_hotplug(cpu);
+			return err;
+		}
+	}
+
+	return 0;
+}
+
+void smp_resume(void)
+{
+	unsigned int cpu;
+
+	for_each_possible_cpu(cpu)
+		vcpu_hotplug(cpu);
+}
+
+int cpu_up_check(unsigned int cpu)
+{
+	int rc = 0;
+
+	if (local_cpu_hotplug_request()) {
+		cpu_set(cpu, local_allowed_cpumask);
+		if (!cpu_isset(cpu, xenbus_allowed_cpumask)) {
+			printk("%s: attempt to bring up CPU %u disallowed by "
+			       "remote admin.\n", __FUNCTION__, cpu);
+			rc = -EBUSY;
+		}
+	} else if (!cpu_isset(cpu, local_allowed_cpumask) ||
+		   !cpu_isset(cpu, xenbus_allowed_cpumask)) {
+		rc = -EBUSY;
+	}
+
+	return rc;
+}
+
+void init_xenbus_allowed_cpumask(void)
+{
+	xenbus_allowed_cpumask = cpu_present_map;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/core/evtchn.c linux-2.6.18-xen-3.3.0/drivers/xen/core/evtchn.c
--- linux-2.6.18.8/drivers/xen/core/evtchn.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/core/evtchn.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1061 @@
+/******************************************************************************
+ * evtchn.c
+ * 
+ * Communication via Xen event channels.
+ * 
+ * Copyright (c) 2002-2005, K A Fraser
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/module.h>
+#include <linux/irq.h>
+#include <linux/interrupt.h>
+#include <linux/sched.h>
+#include <linux/kernel_stat.h>
+#include <linux/version.h>
+#include <asm/atomic.h>
+#include <asm/system.h>
+#include <asm/ptrace.h>
+#include <asm/synch_bitops.h>
+#include <xen/evtchn.h>
+#include <xen/interface/event_channel.h>
+#include <xen/interface/physdev.h>
+#include <asm/hypervisor.h>
+#include <linux/mc146818rtc.h> /* RTC_IRQ */
+
+/*
+ * This lock protects updates to the following mapping and reference-count
+ * arrays. The lock does not need to be acquired to read the mapping tables.
+ */
+static DEFINE_SPINLOCK(irq_mapping_update_lock);
+
+/* IRQ <-> event-channel mappings. */
+static int evtchn_to_irq[NR_EVENT_CHANNELS] = {
+	[0 ...  NR_EVENT_CHANNELS-1] = -1 };
+
+/* Packed IRQ information: binding type, sub-type index, and event channel. */
+static u32 irq_info[NR_IRQS];
+
+/* Binding types. */
+enum {
+	IRQT_UNBOUND,
+	IRQT_PIRQ,
+	IRQT_VIRQ,
+	IRQT_IPI,
+	IRQT_LOCAL_PORT,
+	IRQT_CALLER_PORT
+};
+
+/* Constructor for packed IRQ information. */
+static inline u32 mk_irq_info(u32 type, u32 index, u32 evtchn)
+{
+	return ((type << 24) | (index << 16) | evtchn);
+}
+
+/* Convenient shorthand for packed representation of an unbound IRQ. */
+#define IRQ_UNBOUND	mk_irq_info(IRQT_UNBOUND, 0, 0)
+
+/*
+ * Accessors for packed IRQ information.
+ */
+
+static inline unsigned int evtchn_from_irq(int irq)
+{
+	return (u16)(irq_info[irq]);
+}
+
+static inline unsigned int index_from_irq(int irq)
+{
+	return (u8)(irq_info[irq] >> 16);
+}
+
+static inline unsigned int type_from_irq(int irq)
+{
+	return (u8)(irq_info[irq] >> 24);
+}
+
+/* IRQ <-> VIRQ mapping. */
+DEFINE_PER_CPU(int, virq_to_irq[NR_VIRQS]) = {[0 ... NR_VIRQS-1] = -1};
+
+/* IRQ <-> IPI mapping. */
+#ifndef NR_IPIS
+#define NR_IPIS 1
+#endif
+DEFINE_PER_CPU(int, ipi_to_irq[NR_IPIS]) = {[0 ... NR_IPIS-1] = -1};
+
+/* Reference counts for bindings to IRQs. */
+static int irq_bindcount[NR_IRQS];
+
+/* Bitmap indicating which PIRQs require Xen to be notified on unmask. */
+static DECLARE_BITMAP(pirq_needs_eoi, NR_PIRQS);
+
+#ifdef CONFIG_SMP
+
+static u8 cpu_evtchn[NR_EVENT_CHANNELS];
+static unsigned long cpu_evtchn_mask[NR_CPUS][NR_EVENT_CHANNELS/BITS_PER_LONG];
+
+static inline unsigned long active_evtchns(unsigned int cpu, shared_info_t *sh,
+					   unsigned int idx)
+{
+	return (sh->evtchn_pending[idx] &
+		cpu_evtchn_mask[cpu][idx] &
+		~sh->evtchn_mask[idx]);
+}
+
+static void bind_evtchn_to_cpu(unsigned int chn, unsigned int cpu)
+{
+	shared_info_t *s = HYPERVISOR_shared_info;
+	int irq = evtchn_to_irq[chn];
+
+	BUG_ON(!test_bit(chn, s->evtchn_mask));
+
+	if (irq != -1)
+		set_native_irq_info(irq, cpumask_of_cpu(cpu));
+
+	clear_bit(chn, (unsigned long *)cpu_evtchn_mask[cpu_evtchn[chn]]);
+	set_bit(chn, (unsigned long *)cpu_evtchn_mask[cpu]);
+	cpu_evtchn[chn] = cpu;
+}
+
+static void init_evtchn_cpu_bindings(void)
+{
+	int i;
+
+	/* By default all event channels notify CPU#0. */
+	for (i = 0; i < NR_IRQS; i++)
+		set_native_irq_info(i, cpumask_of_cpu(0));
+
+	memset(cpu_evtchn, 0, sizeof(cpu_evtchn));
+	memset(cpu_evtchn_mask[0], ~0, sizeof(cpu_evtchn_mask[0]));
+}
+
+static inline unsigned int cpu_from_evtchn(unsigned int evtchn)
+{
+	return cpu_evtchn[evtchn];
+}
+
+#else
+
+static inline unsigned long active_evtchns(unsigned int cpu, shared_info_t *sh,
+					   unsigned int idx)
+{
+	return (sh->evtchn_pending[idx] & ~sh->evtchn_mask[idx]);
+}
+
+static void bind_evtchn_to_cpu(unsigned int chn, unsigned int cpu)
+{
+}
+
+static void init_evtchn_cpu_bindings(void)
+{
+}
+
+static inline unsigned int cpu_from_evtchn(unsigned int evtchn)
+{
+	return 0;
+}
+
+#endif
+
+/* Upcall to generic IRQ layer. */
+#ifdef CONFIG_X86
+extern fastcall unsigned int do_IRQ(struct pt_regs *regs);
+void __init xen_init_IRQ(void);
+void __init init_IRQ(void)
+{
+	irq_ctx_init(0);
+	xen_init_IRQ();
+}
+#if defined (__i386__)
+static inline void exit_idle(void) {}
+#define IRQ_REG orig_eax
+#elif defined (__x86_64__)
+#include <asm/idle.h>
+#define IRQ_REG orig_rax
+#endif
+#define do_IRQ(irq, regs) do {		\
+	(regs)->IRQ_REG = ~(irq);	\
+	do_IRQ((regs));			\
+} while (0)
+#endif
+
+/* Xen will never allocate port zero for any purpose. */
+#define VALID_EVTCHN(chn)	((chn) != 0)
+
+/*
+ * Force a proper event-channel callback from Xen after clearing the
+ * callback mask. We do this in a very simple manner, by making a call
+ * down into Xen. The pending flag will be checked by Xen on return.
+ */
+void force_evtchn_callback(void)
+{
+	VOID(HYPERVISOR_xen_version(0, NULL));
+}
+/* Not a GPL symbol: used in ubiquitous macros, so too restrictive. */
+EXPORT_SYMBOL(force_evtchn_callback);
+
+static DEFINE_PER_CPU(unsigned int, upcall_count) = { 0 };
+static DEFINE_PER_CPU(unsigned int, last_processed_l1i) = { BITS_PER_LONG - 1 };
+static DEFINE_PER_CPU(unsigned int, last_processed_l2i) = { BITS_PER_LONG - 1 };
+
+/* NB. Interrupts are disabled on entry. */
+asmlinkage void evtchn_do_upcall(struct pt_regs *regs)
+{
+	unsigned long       l1, l2;
+	unsigned long       masked_l1, masked_l2;
+	unsigned int        l1i, l2i, port, count;
+	int                 irq;
+	unsigned int        cpu = smp_processor_id();
+	shared_info_t      *s = HYPERVISOR_shared_info;
+	vcpu_info_t        *vcpu_info = &s->vcpu_info[cpu];
+
+
+	do {
+		/* Avoid a callback storm when we reenable delivery. */
+		vcpu_info->evtchn_upcall_pending = 0;
+
+		/* Nested invocations bail immediately. */
+		if (unlikely(per_cpu(upcall_count, cpu)++))
+			return;
+
+#ifndef CONFIG_X86 /* No need for a barrier -- XCHG is a barrier on x86. */
+		/* Clear master flag /before/ clearing selector flag. */
+		wmb();
+#endif
+		l1 = xchg(&vcpu_info->evtchn_pending_sel, 0);
+
+		l1i = per_cpu(last_processed_l1i, cpu);
+		l2i = per_cpu(last_processed_l2i, cpu);
+
+		while (l1 != 0) {
+
+			l1i = (l1i + 1) % BITS_PER_LONG;
+			masked_l1 = l1 & ((~0UL) << l1i);
+
+			if (masked_l1 == 0) { /* if we masked out all events, wrap around to the beginning */
+				l1i = BITS_PER_LONG - 1;
+				l2i = BITS_PER_LONG - 1;
+				continue;
+			}
+			l1i = __ffs(masked_l1);
+
+			do {
+				l2 = active_evtchns(cpu, s, l1i);
+
+				l2i = (l2i + 1) % BITS_PER_LONG;
+				masked_l2 = l2 & ((~0UL) << l2i);
+
+				if (masked_l2 == 0) { /* if we masked out all events, move on */
+					l2i = BITS_PER_LONG - 1;
+					break;
+				}
+
+				l2i = __ffs(masked_l2);
+
+				/* process port */
+				port = (l1i * BITS_PER_LONG) + l2i;
+				if ((irq = evtchn_to_irq[port]) != -1)
+					do_IRQ(irq, regs);
+				else {
+					exit_idle();
+					evtchn_device_upcall(port);
+				}
+
+				/* if this is the final port processed, we'll pick up here+1 next time */
+				per_cpu(last_processed_l1i, cpu) = l1i;
+				per_cpu(last_processed_l2i, cpu) = l2i;
+
+			} while (l2i != BITS_PER_LONG - 1);
+
+			l2 = active_evtchns(cpu, s, l1i);
+			if (l2 == 0) /* we handled all ports, so we can clear the selector bit */
+				l1 &= ~(1UL << l1i);
+
+		}
+
+		/* If there were nested callbacks then we have more to do. */
+		count = per_cpu(upcall_count, cpu);
+		per_cpu(upcall_count, cpu) = 0;
+	} while (unlikely(count != 1));
+}
+
+static int find_unbound_irq(void)
+{
+	static int warned;
+	int dynirq, irq;
+
+	for (dynirq = 0; dynirq < NR_DYNIRQS; dynirq++) {
+		irq = dynirq_to_irq(dynirq);
+		if (irq_bindcount[irq] == 0)
+			return irq;
+	}
+
+	if (!warned) {
+		warned = 1;
+		printk(KERN_WARNING "No available IRQ to bind to: "
+		       "increase NR_DYNIRQS.\n");
+	}
+
+	return -ENOSPC;
+}
+
+static int bind_caller_port_to_irq(unsigned int caller_port)
+{
+	int irq;
+
+	spin_lock(&irq_mapping_update_lock);
+
+	if ((irq = evtchn_to_irq[caller_port]) == -1) {
+		if ((irq = find_unbound_irq()) < 0)
+			goto out;
+
+		evtchn_to_irq[caller_port] = irq;
+		irq_info[irq] = mk_irq_info(IRQT_CALLER_PORT, 0, caller_port);
+	}
+
+	irq_bindcount[irq]++;
+
+ out:
+	spin_unlock(&irq_mapping_update_lock);
+	return irq;
+}
+
+static int bind_local_port_to_irq(unsigned int local_port)
+{
+	int irq;
+
+	spin_lock(&irq_mapping_update_lock);
+
+	BUG_ON(evtchn_to_irq[local_port] != -1);
+
+	if ((irq = find_unbound_irq()) < 0) {
+		struct evtchn_close close = { .port = local_port };
+		if (HYPERVISOR_event_channel_op(EVTCHNOP_close, &close))
+			BUG();
+		goto out;
+	}
+
+	evtchn_to_irq[local_port] = irq;
+	irq_info[irq] = mk_irq_info(IRQT_LOCAL_PORT, 0, local_port);
+	irq_bindcount[irq]++;
+
+ out:
+	spin_unlock(&irq_mapping_update_lock);
+	return irq;
+}
+
+static int bind_listening_port_to_irq(unsigned int remote_domain)
+{
+	struct evtchn_alloc_unbound alloc_unbound;
+	int err;
+
+	alloc_unbound.dom        = DOMID_SELF;
+	alloc_unbound.remote_dom = remote_domain;
+
+	err = HYPERVISOR_event_channel_op(EVTCHNOP_alloc_unbound,
+					  &alloc_unbound);
+
+	return err ? : bind_local_port_to_irq(alloc_unbound.port);
+}
+
+static int bind_interdomain_evtchn_to_irq(unsigned int remote_domain,
+					  unsigned int remote_port)
+{
+	struct evtchn_bind_interdomain bind_interdomain;
+	int err;
+
+	bind_interdomain.remote_dom  = remote_domain;
+	bind_interdomain.remote_port = remote_port;
+
+	err = HYPERVISOR_event_channel_op(EVTCHNOP_bind_interdomain,
+					  &bind_interdomain);
+
+	return err ? : bind_local_port_to_irq(bind_interdomain.local_port);
+}
+
+static int bind_virq_to_irq(unsigned int virq, unsigned int cpu)
+{
+	struct evtchn_bind_virq bind_virq;
+	int evtchn, irq;
+
+	spin_lock(&irq_mapping_update_lock);
+
+	if ((irq = per_cpu(virq_to_irq, cpu)[virq]) == -1) {
+		if ((irq = find_unbound_irq()) < 0)
+			goto out;
+
+		bind_virq.virq = virq;
+		bind_virq.vcpu = cpu;
+		if (HYPERVISOR_event_channel_op(EVTCHNOP_bind_virq,
+						&bind_virq) != 0)
+			BUG();
+		evtchn = bind_virq.port;
+
+		evtchn_to_irq[evtchn] = irq;
+		irq_info[irq] = mk_irq_info(IRQT_VIRQ, virq, evtchn);
+
+		per_cpu(virq_to_irq, cpu)[virq] = irq;
+
+		bind_evtchn_to_cpu(evtchn, cpu);
+	}
+
+	irq_bindcount[irq]++;
+
+ out:
+	spin_unlock(&irq_mapping_update_lock);
+	return irq;
+}
+
+static int bind_ipi_to_irq(unsigned int ipi, unsigned int cpu)
+{
+	struct evtchn_bind_ipi bind_ipi;
+	int evtchn, irq;
+
+	spin_lock(&irq_mapping_update_lock);
+
+	if ((irq = per_cpu(ipi_to_irq, cpu)[ipi]) == -1) {
+		if ((irq = find_unbound_irq()) < 0)
+			goto out;
+
+		bind_ipi.vcpu = cpu;
+		if (HYPERVISOR_event_channel_op(EVTCHNOP_bind_ipi,
+						&bind_ipi) != 0)
+			BUG();
+		evtchn = bind_ipi.port;
+
+		evtchn_to_irq[evtchn] = irq;
+		irq_info[irq] = mk_irq_info(IRQT_IPI, ipi, evtchn);
+
+		per_cpu(ipi_to_irq, cpu)[ipi] = irq;
+
+		bind_evtchn_to_cpu(evtchn, cpu);
+	}
+
+	irq_bindcount[irq]++;
+
+ out:
+	spin_unlock(&irq_mapping_update_lock);
+	return irq;
+}
+
+static void unbind_from_irq(unsigned int irq)
+{
+	struct evtchn_close close;
+	unsigned int cpu;
+	int evtchn = evtchn_from_irq(irq);
+
+	spin_lock(&irq_mapping_update_lock);
+
+	if ((--irq_bindcount[irq] == 0) && VALID_EVTCHN(evtchn)) {
+		close.port = evtchn;
+		if ((type_from_irq(irq) != IRQT_CALLER_PORT) &&
+		    HYPERVISOR_event_channel_op(EVTCHNOP_close, &close))
+			BUG();
+
+		switch (type_from_irq(irq)) {
+		case IRQT_VIRQ:
+			per_cpu(virq_to_irq, cpu_from_evtchn(evtchn))
+				[index_from_irq(irq)] = -1;
+			break;
+		case IRQT_IPI:
+			per_cpu(ipi_to_irq, cpu_from_evtchn(evtchn))
+				[index_from_irq(irq)] = -1;
+			break;
+		default:
+			break;
+		}
+
+		/* Closed ports are implicitly re-bound to VCPU0. */
+		bind_evtchn_to_cpu(evtchn, 0);
+
+		evtchn_to_irq[evtchn] = -1;
+		irq_info[irq] = IRQ_UNBOUND;
+
+		/* Zap stats across IRQ changes of use. */
+		for_each_possible_cpu(cpu)
+			kstat_cpu(cpu).irqs[irq] = 0;
+	}
+
+	spin_unlock(&irq_mapping_update_lock);
+}
+
+int bind_caller_port_to_irqhandler(
+	unsigned int caller_port,
+	irqreturn_t (*handler)(int, void *, struct pt_regs *),
+	unsigned long irqflags,
+	const char *devname,
+	void *dev_id)
+{
+	int irq, retval;
+
+	irq = bind_caller_port_to_irq(caller_port);
+	if (irq < 0)
+		return irq;
+
+	retval = request_irq(irq, handler, irqflags, devname, dev_id);
+	if (retval != 0) {
+		unbind_from_irq(irq);
+		return retval;
+	}
+
+	return irq;
+}
+EXPORT_SYMBOL_GPL(bind_caller_port_to_irqhandler);
+
+int bind_listening_port_to_irqhandler(
+	unsigned int remote_domain,
+	irqreturn_t (*handler)(int, void *, struct pt_regs *),
+	unsigned long irqflags,
+	const char *devname,
+	void *dev_id)
+{
+	int irq, retval;
+
+	irq = bind_listening_port_to_irq(remote_domain);
+	if (irq < 0)
+		return irq;
+
+	retval = request_irq(irq, handler, irqflags, devname, dev_id);
+	if (retval != 0) {
+		unbind_from_irq(irq);
+		return retval;
+	}
+
+	return irq;
+}
+EXPORT_SYMBOL_GPL(bind_listening_port_to_irqhandler);
+
+int bind_interdomain_evtchn_to_irqhandler(
+	unsigned int remote_domain,
+	unsigned int remote_port,
+	irqreturn_t (*handler)(int, void *, struct pt_regs *),
+	unsigned long irqflags,
+	const char *devname,
+	void *dev_id)
+{
+	int irq, retval;
+
+	irq = bind_interdomain_evtchn_to_irq(remote_domain, remote_port);
+	if (irq < 0)
+		return irq;
+
+	retval = request_irq(irq, handler, irqflags, devname, dev_id);
+	if (retval != 0) {
+		unbind_from_irq(irq);
+		return retval;
+	}
+
+	return irq;
+}
+EXPORT_SYMBOL_GPL(bind_interdomain_evtchn_to_irqhandler);
+
+int bind_virq_to_irqhandler(
+	unsigned int virq,
+	unsigned int cpu,
+	irqreturn_t (*handler)(int, void *, struct pt_regs *),
+	unsigned long irqflags,
+	const char *devname,
+	void *dev_id)
+{
+	int irq, retval;
+
+	irq = bind_virq_to_irq(virq, cpu);
+	if (irq < 0)
+		return irq;
+
+	retval = request_irq(irq, handler, irqflags, devname, dev_id);
+	if (retval != 0) {
+		unbind_from_irq(irq);
+		return retval;
+	}
+
+	return irq;
+}
+EXPORT_SYMBOL_GPL(bind_virq_to_irqhandler);
+
+int bind_ipi_to_irqhandler(
+	unsigned int ipi,
+	unsigned int cpu,
+	irqreturn_t (*handler)(int, void *, struct pt_regs *),
+	unsigned long irqflags,
+	const char *devname,
+	void *dev_id)
+{
+	int irq, retval;
+
+	irq = bind_ipi_to_irq(ipi, cpu);
+	if (irq < 0)
+		return irq;
+
+	retval = request_irq(irq, handler, irqflags, devname, dev_id);
+	if (retval != 0) {
+		unbind_from_irq(irq);
+		return retval;
+	}
+
+	return irq;
+}
+EXPORT_SYMBOL_GPL(bind_ipi_to_irqhandler);
+
+void unbind_from_irqhandler(unsigned int irq, void *dev_id)
+{
+	free_irq(irq, dev_id);
+	unbind_from_irq(irq);
+}
+EXPORT_SYMBOL_GPL(unbind_from_irqhandler);
+
+#ifdef CONFIG_SMP
+void rebind_evtchn_to_cpu(int port, unsigned int cpu)
+{
+	struct evtchn_bind_vcpu ebv = { .port = port, .vcpu = cpu };
+	int masked;
+
+	masked = test_and_set_evtchn_mask(port);
+	if (HYPERVISOR_event_channel_op(EVTCHNOP_bind_vcpu, &ebv) == 0)
+		bind_evtchn_to_cpu(port, cpu);
+	if (!masked)
+		unmask_evtchn(port);
+}
+
+static void rebind_irq_to_cpu(unsigned int irq, unsigned int tcpu)
+{
+	int evtchn = evtchn_from_irq(irq);
+
+	if (VALID_EVTCHN(evtchn))
+		rebind_evtchn_to_cpu(evtchn, tcpu);
+}
+
+static void set_affinity_irq(unsigned int irq, cpumask_t dest)
+{
+	unsigned tcpu = first_cpu(dest);
+	rebind_irq_to_cpu(irq, tcpu);
+}
+#endif
+
+int resend_irq_on_evtchn(unsigned int irq)
+{
+	int masked, evtchn = evtchn_from_irq(irq);
+	shared_info_t *s = HYPERVISOR_shared_info;
+
+	if (!VALID_EVTCHN(evtchn))
+		return 1;
+
+	masked = test_and_set_evtchn_mask(evtchn);
+	synch_set_bit(evtchn, s->evtchn_pending);
+	if (!masked)
+		unmask_evtchn(evtchn);
+
+	return 1;
+}
+
+/*
+ * Interface to generic handling in irq.c
+ */
+
+static unsigned int startup_dynirq(unsigned int irq)
+{
+	int evtchn = evtchn_from_irq(irq);
+
+	if (VALID_EVTCHN(evtchn))
+		unmask_evtchn(evtchn);
+	return 0;
+}
+
+static void shutdown_dynirq(unsigned int irq)
+{
+	int evtchn = evtchn_from_irq(irq);
+
+	if (VALID_EVTCHN(evtchn))
+		mask_evtchn(evtchn);
+}
+
+static void enable_dynirq(unsigned int irq)
+{
+	int evtchn = evtchn_from_irq(irq);
+
+	if (VALID_EVTCHN(evtchn))
+		unmask_evtchn(evtchn);
+}
+
+static void disable_dynirq(unsigned int irq)
+{
+	int evtchn = evtchn_from_irq(irq);
+
+	if (VALID_EVTCHN(evtchn))
+		mask_evtchn(evtchn);
+}
+
+static void ack_dynirq(unsigned int irq)
+{
+	int evtchn = evtchn_from_irq(irq);
+
+	move_native_irq(irq);
+
+	if (VALID_EVTCHN(evtchn)) {
+		mask_evtchn(evtchn);
+		clear_evtchn(evtchn);
+	}
+}
+
+static void end_dynirq(unsigned int irq)
+{
+	int evtchn = evtchn_from_irq(irq);
+
+	if (VALID_EVTCHN(evtchn) && !(irq_desc[irq].status & IRQ_DISABLED))
+		unmask_evtchn(evtchn);
+}
+
+static struct hw_interrupt_type dynirq_type = {
+	.typename = "Dynamic-irq",
+	.startup  = startup_dynirq,
+	.shutdown = shutdown_dynirq,
+	.enable   = enable_dynirq,
+	.disable  = disable_dynirq,
+	.ack      = ack_dynirq,
+	.end      = end_dynirq,
+#ifdef CONFIG_SMP
+	.set_affinity = set_affinity_irq,
+#endif
+	.retrigger = resend_irq_on_evtchn,
+};
+
+static inline void pirq_unmask_notify(int pirq)
+{
+	struct physdev_eoi eoi = { .irq = pirq };
+	if (unlikely(test_bit(pirq, pirq_needs_eoi)))
+		VOID(HYPERVISOR_physdev_op(PHYSDEVOP_eoi, &eoi));
+}
+
+static inline void pirq_query_unmask(int pirq)
+{
+	struct physdev_irq_status_query irq_status;
+	irq_status.irq = pirq;
+	if (HYPERVISOR_physdev_op(PHYSDEVOP_irq_status_query, &irq_status))
+		irq_status.flags = 0;
+	clear_bit(pirq, pirq_needs_eoi);
+	if (irq_status.flags & XENIRQSTAT_needs_eoi)
+		set_bit(pirq, pirq_needs_eoi);
+}
+
+/*
+ * On startup, if there is no action associated with the IRQ then we are
+ * probing. In this case we should not share with others as it will confuse us.
+ */
+#define probing_irq(_irq) (irq_desc[(_irq)].action == NULL)
+
+static unsigned int startup_pirq(unsigned int irq)
+{
+	struct evtchn_bind_pirq bind_pirq;
+	int evtchn = evtchn_from_irq(irq);
+
+	if (VALID_EVTCHN(evtchn))
+		goto out;
+
+	bind_pirq.pirq  = irq;
+	/* NB. We are happy to share unless we are probing. */
+	bind_pirq.flags = probing_irq(irq) ? 0 : BIND_PIRQ__WILL_SHARE;
+	if (HYPERVISOR_event_channel_op(EVTCHNOP_bind_pirq, &bind_pirq) != 0) {
+		if (!probing_irq(irq))
+			printk(KERN_INFO "Failed to obtain physical IRQ %d\n",
+			       irq);
+		return 0;
+	}
+	evtchn = bind_pirq.port;
+
+	pirq_query_unmask(irq_to_pirq(irq));
+
+	evtchn_to_irq[evtchn] = irq;
+	bind_evtchn_to_cpu(evtchn, 0);
+	irq_info[irq] = mk_irq_info(IRQT_PIRQ, irq, evtchn);
+
+ out:
+	unmask_evtchn(evtchn);
+	pirq_unmask_notify(irq_to_pirq(irq));
+
+	return 0;
+}
+
+static void shutdown_pirq(unsigned int irq)
+{
+	struct evtchn_close close;
+	int evtchn = evtchn_from_irq(irq);
+
+	if (!VALID_EVTCHN(evtchn))
+		return;
+
+	mask_evtchn(evtchn);
+
+	close.port = evtchn;
+	if (HYPERVISOR_event_channel_op(EVTCHNOP_close, &close) != 0)
+		BUG();
+
+	bind_evtchn_to_cpu(evtchn, 0);
+	evtchn_to_irq[evtchn] = -1;
+	irq_info[irq] = IRQ_UNBOUND;
+}
+
+static void enable_pirq(unsigned int irq)
+{
+	startup_pirq(irq);
+}
+
+static void disable_pirq(unsigned int irq)
+{
+}
+
+static void ack_pirq(unsigned int irq)
+{
+	int evtchn = evtchn_from_irq(irq);
+
+	move_native_irq(irq);
+
+	if (VALID_EVTCHN(evtchn)) {
+		mask_evtchn(evtchn);
+		clear_evtchn(evtchn);
+	}
+}
+
+static void end_pirq(unsigned int irq)
+{
+	int evtchn = evtchn_from_irq(irq);
+
+	if ((irq_desc[irq].status & (IRQ_DISABLED|IRQ_PENDING)) ==
+	    (IRQ_DISABLED|IRQ_PENDING)) {
+		shutdown_pirq(irq);
+	} else if (VALID_EVTCHN(evtchn)) {
+		unmask_evtchn(evtchn);
+		pirq_unmask_notify(irq_to_pirq(irq));
+	}
+}
+
+static struct hw_interrupt_type pirq_type = {
+	.typename = "Phys-irq",
+	.startup  = startup_pirq,
+	.shutdown = shutdown_pirq,
+	.enable   = enable_pirq,
+	.disable  = disable_pirq,
+	.ack      = ack_pirq,
+	.end      = end_pirq,
+#ifdef CONFIG_SMP
+	.set_affinity = set_affinity_irq,
+#endif
+	.retrigger = resend_irq_on_evtchn,
+};
+
+int irq_ignore_unhandled(unsigned int irq)
+{
+	struct physdev_irq_status_query irq_status = { .irq = irq };
+
+	if (!is_running_on_xen())
+		return 0;
+
+	if (HYPERVISOR_physdev_op(PHYSDEVOP_irq_status_query, &irq_status))
+		return 0;
+	return !!(irq_status.flags & XENIRQSTAT_shared);
+}
+
+void notify_remote_via_irq(int irq)
+{
+	int evtchn = evtchn_from_irq(irq);
+
+	if (VALID_EVTCHN(evtchn))
+		notify_remote_via_evtchn(evtchn);
+}
+EXPORT_SYMBOL_GPL(notify_remote_via_irq);
+
+int irq_to_evtchn_port(int irq)
+{
+	return evtchn_from_irq(irq);
+}
+EXPORT_SYMBOL_GPL(irq_to_evtchn_port);
+
+void mask_evtchn(int port)
+{
+	shared_info_t *s = HYPERVISOR_shared_info;
+	synch_set_bit(port, s->evtchn_mask);
+}
+EXPORT_SYMBOL_GPL(mask_evtchn);
+
+void unmask_evtchn(int port)
+{
+	shared_info_t *s = HYPERVISOR_shared_info;
+	unsigned int cpu = smp_processor_id();
+	vcpu_info_t *vcpu_info = &s->vcpu_info[cpu];
+
+	BUG_ON(!irqs_disabled());
+
+	/* Slow path (hypercall) if this is a non-local port. */
+	if (unlikely(cpu != cpu_from_evtchn(port))) {
+		struct evtchn_unmask unmask = { .port = port };
+		VOID(HYPERVISOR_event_channel_op(EVTCHNOP_unmask, &unmask));
+		return;
+	}
+
+	synch_clear_bit(port, s->evtchn_mask);
+
+	/* Did we miss an interrupt 'edge'? Re-fire if so. */
+	if (synch_test_bit(port, s->evtchn_pending) &&
+	    !synch_test_and_set_bit(port / BITS_PER_LONG,
+				    &vcpu_info->evtchn_pending_sel))
+		vcpu_info->evtchn_upcall_pending = 1;
+}
+EXPORT_SYMBOL_GPL(unmask_evtchn);
+
+void disable_all_local_evtchn(void)
+{
+	unsigned i, cpu = smp_processor_id();
+	shared_info_t *s = HYPERVISOR_shared_info;
+
+	for (i = 0; i < NR_EVENT_CHANNELS; ++i)
+		if (cpu_from_evtchn(i) == cpu)
+			synch_set_bit(i, &s->evtchn_mask[0]);
+}
+
+static void restore_cpu_virqs(unsigned int cpu)
+{
+	struct evtchn_bind_virq bind_virq;
+	int virq, irq, evtchn;
+
+	for (virq = 0; virq < NR_VIRQS; virq++) {
+		if ((irq = per_cpu(virq_to_irq, cpu)[virq]) == -1)
+			continue;
+
+		BUG_ON(irq_info[irq] != mk_irq_info(IRQT_VIRQ, virq, 0));
+
+		/* Get a new binding from Xen. */
+		bind_virq.virq = virq;
+		bind_virq.vcpu = cpu;
+		if (HYPERVISOR_event_channel_op(EVTCHNOP_bind_virq,
+						&bind_virq) != 0)
+			BUG();
+		evtchn = bind_virq.port;
+
+		/* Record the new mapping. */
+		evtchn_to_irq[evtchn] = irq;
+		irq_info[irq] = mk_irq_info(IRQT_VIRQ, virq, evtchn);
+		bind_evtchn_to_cpu(evtchn, cpu);
+
+		/* Ready for use. */
+		unmask_evtchn(evtchn);
+	}
+}
+
+static void restore_cpu_ipis(unsigned int cpu)
+{
+	struct evtchn_bind_ipi bind_ipi;
+	int ipi, irq, evtchn;
+
+	for (ipi = 0; ipi < NR_IPIS; ipi++) {
+		if ((irq = per_cpu(ipi_to_irq, cpu)[ipi]) == -1)
+			continue;
+
+		BUG_ON(irq_info[irq] != mk_irq_info(IRQT_IPI, ipi, 0));
+
+		/* Get a new binding from Xen. */
+		bind_ipi.vcpu = cpu;
+		if (HYPERVISOR_event_channel_op(EVTCHNOP_bind_ipi,
+						&bind_ipi) != 0)
+			BUG();
+		evtchn = bind_ipi.port;
+
+		/* Record the new mapping. */
+		evtchn_to_irq[evtchn] = irq;
+		irq_info[irq] = mk_irq_info(IRQT_IPI, ipi, evtchn);
+		bind_evtchn_to_cpu(evtchn, cpu);
+
+		/* Ready for use. */
+		unmask_evtchn(evtchn);
+
+	}
+}
+
+void irq_resume(void)
+{
+	unsigned int cpu, pirq, irq, evtchn;
+
+	init_evtchn_cpu_bindings();
+
+	/* New event-channel space is not 'live' yet. */
+	for (evtchn = 0; evtchn < NR_EVENT_CHANNELS; evtchn++)
+		mask_evtchn(evtchn);
+
+	/* Check that no PIRQs are still bound. */
+	for (pirq = 0; pirq < NR_PIRQS; pirq++)
+		BUG_ON(irq_info[pirq_to_irq(pirq)] != IRQ_UNBOUND);
+
+	/* No IRQ <-> event-channel mappings. */
+	for (irq = 0; irq < NR_IRQS; irq++)
+		irq_info[irq] &= ~0xFFFF; /* zap event-channel binding */
+	for (evtchn = 0; evtchn < NR_EVENT_CHANNELS; evtchn++)
+		evtchn_to_irq[evtchn] = -1;
+
+	for_each_possible_cpu(cpu) {
+		restore_cpu_virqs(cpu);
+		restore_cpu_ipis(cpu);
+	}
+
+}
+
+void __init xen_init_IRQ(void)
+{
+	unsigned int i;
+
+	init_evtchn_cpu_bindings();
+
+	/* No event channels are 'live' right now. */
+	for (i = 0; i < NR_EVENT_CHANNELS; i++)
+		mask_evtchn(i);
+
+	/* No IRQ -> event-channel mappings. */
+	for (i = 0; i < NR_IRQS; i++)
+		irq_info[i] = IRQ_UNBOUND;
+
+	/* Dynamic IRQ space is currently unbound. Zero the refcnts. */
+	for (i = 0; i < NR_DYNIRQS; i++) {
+		irq_bindcount[dynirq_to_irq(i)] = 0;
+
+		irq_desc[dynirq_to_irq(i)].status = IRQ_DISABLED;
+		irq_desc[dynirq_to_irq(i)].action = NULL;
+		irq_desc[dynirq_to_irq(i)].depth = 1;
+		irq_desc[dynirq_to_irq(i)].chip = &dynirq_type;
+	}
+
+	/* Phys IRQ space is statically bound (1:1 mapping). Nail refcnts. */
+	for (i = 0; i < NR_PIRQS; i++) {
+		irq_bindcount[pirq_to_irq(i)] = 1;
+
+#ifdef RTC_IRQ
+		/* If not domain 0, force our RTC driver to fail its probe. */
+		if ((i == RTC_IRQ) && !is_initial_xendomain())
+			continue;
+#endif
+
+		irq_desc[pirq_to_irq(i)].status = IRQ_DISABLED;
+		irq_desc[pirq_to_irq(i)].action = NULL;
+		irq_desc[pirq_to_irq(i)].depth = 1;
+		irq_desc[pirq_to_irq(i)].chip = &pirq_type;
+	}
+}
diff -rpuN linux-2.6.18.8/drivers/xen/core/features.c linux-2.6.18-xen-3.3.0/drivers/xen/core/features.c
--- linux-2.6.18.8/drivers/xen/core/features.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/core/features.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,34 @@
+/******************************************************************************
+ * features.c
+ *
+ * Xen feature flags.
+ *
+ * Copyright (c) 2006, Ian Campbell, XenSource Inc.
+ */
+#include <linux/types.h>
+#include <linux/cache.h>
+#include <linux/module.h>
+#include <asm/hypervisor.h>
+#include <xen/features.h>
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+u8 xen_features[XENFEAT_NR_SUBMAPS * 32] __read_mostly;
+/* Not a GPL symbol: used in ubiquitous macros, so too restrictive. */
+EXPORT_SYMBOL(xen_features);
+
+void setup_xen_features(void)
+{
+	xen_feature_info_t fi;
+	int i, j;
+
+	for (i = 0; i < XENFEAT_NR_SUBMAPS; i++) {
+		fi.submap_idx = i;
+		if (HYPERVISOR_xen_version(XENVER_get_features, &fi) < 0)
+			break;
+		for (j=0; j<32; j++)
+			xen_features[i*32+j] = !!(fi.submap & 1<<j);
+	}
+}
diff -rpuN linux-2.6.18.8/drivers/xen/core/firmware.c linux-2.6.18-xen-3.3.0/drivers/xen/core/firmware.c
--- linux-2.6.18.8/drivers/xen/core/firmware.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/core/firmware.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,74 @@
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/init.h>
+#include <linux/edd.h>
+#include <video/edid.h>
+#include <xen/interface/platform.h>
+#include <asm/hypervisor.h>
+
+#if defined(CONFIG_EDD) || defined(CONFIG_EDD_MODULE)
+void __init copy_edd(void)
+{
+	int ret;
+	struct xen_platform_op op;
+
+	if (!is_initial_xendomain())
+		return;
+
+	op.cmd = XENPF_firmware_info;
+
+	op.u.firmware_info.type = XEN_FW_DISK_INFO;
+	for (op.u.firmware_info.index = 0;
+	     edd.edd_info_nr < EDDMAXNR;
+	     op.u.firmware_info.index++) {
+		struct edd_info *info = edd.edd_info + edd.edd_info_nr;
+
+		info->params.length = sizeof(info->params);
+		set_xen_guest_handle(op.u.firmware_info.u.disk_info.edd_params,
+				     &info->params);
+		ret = HYPERVISOR_platform_op(&op);
+		if (ret)
+			break;
+
+#define C(x) info->x = op.u.firmware_info.u.disk_info.x
+		C(device);
+		C(version);
+		C(interface_support);
+		C(legacy_max_cylinder);
+		C(legacy_max_head);
+		C(legacy_sectors_per_track);
+#undef C
+
+		edd.edd_info_nr++;
+	}
+
+	op.u.firmware_info.type = XEN_FW_DISK_MBR_SIGNATURE;
+	for (op.u.firmware_info.index = 0;
+	     edd.mbr_signature_nr < EDD_MBR_SIG_MAX;
+	     op.u.firmware_info.index++) {
+		ret = HYPERVISOR_platform_op(&op);
+		if (ret)
+			break;
+		edd.mbr_signature[edd.mbr_signature_nr++] =
+			op.u.firmware_info.u.disk_mbr_signature.mbr_signature;
+	}
+}
+#endif
+
+void __init copy_edid(void)
+{
+#if defined(CONFIG_FIRMWARE_EDID) && defined(CONFIG_X86)
+	struct xen_platform_op op;
+
+	if (!is_initial_xendomain())
+		return;
+
+	op.cmd = XENPF_firmware_info;
+	op.u.firmware_info.index = 0;
+	op.u.firmware_info.type = XEN_FW_VBEDDC_INFO;
+	set_xen_guest_handle(op.u.firmware_info.u.vbeddc_info.edid,
+			     edid_info.dummy);
+	if (HYPERVISOR_platform_op(&op) != 0)
+		memset(edid_info.dummy, 0x13, sizeof(edid_info.dummy));
+#endif
+}
diff -rpuN linux-2.6.18.8/drivers/xen/core/gnttab.c linux-2.6.18-xen-3.3.0/drivers/xen/core/gnttab.c
--- linux-2.6.18.8/drivers/xen/core/gnttab.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/core/gnttab.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,769 @@
+/******************************************************************************
+ * gnttab.c
+ *
+ * Granting foreign access to our memory reservation.
+ *
+ * Copyright (c) 2005-2006, Christopher Clark
+ * Copyright (c) 2004-2005, K A Fraser
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/module.h>
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/seqlock.h>
+#include <xen/interface/xen.h>
+#include <xen/gnttab.h>
+#include <asm/pgtable.h>
+#include <asm/uaccess.h>
+#include <asm/synch_bitops.h>
+#include <asm/io.h>
+#include <xen/interface/memory.h>
+#include <xen/driver_util.h>
+#include <asm/gnttab_dma.h>
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+/* External tools reserve first few grant table entries. */
+#define NR_RESERVED_ENTRIES 8
+#define GNTTAB_LIST_END 0xffffffff
+#define ENTRIES_PER_GRANT_FRAME (PAGE_SIZE / sizeof(grant_entry_t))
+
+static grant_ref_t **gnttab_list;
+static unsigned int nr_grant_frames;
+static unsigned int boot_max_nr_grant_frames;
+static int gnttab_free_count;
+static grant_ref_t gnttab_free_head;
+static DEFINE_SPINLOCK(gnttab_list_lock);
+
+static struct grant_entry *shared;
+
+static struct gnttab_free_callback *gnttab_free_callback_list;
+
+static int gnttab_expand(unsigned int req_entries);
+
+#define RPP (PAGE_SIZE / sizeof(grant_ref_t))
+#define gnttab_entry(entry) (gnttab_list[(entry) / RPP][(entry) % RPP])
+
+#define nr_freelist_frames(grant_frames)				\
+	(((grant_frames) * ENTRIES_PER_GRANT_FRAME + RPP - 1) / RPP)
+
+static int get_free_entries(int count)
+{
+	unsigned long flags;
+	int ref, rc;
+	grant_ref_t head;
+
+	spin_lock_irqsave(&gnttab_list_lock, flags);
+
+	if ((gnttab_free_count < count) &&
+	    ((rc = gnttab_expand(count - gnttab_free_count)) < 0)) {
+		spin_unlock_irqrestore(&gnttab_list_lock, flags);
+		return rc;
+	}
+
+	ref = head = gnttab_free_head;
+	gnttab_free_count -= count;
+	while (count-- > 1)
+		head = gnttab_entry(head);
+ 	gnttab_free_head = gnttab_entry(head);
+	gnttab_entry(head) = GNTTAB_LIST_END;
+
+	spin_unlock_irqrestore(&gnttab_list_lock, flags);
+
+	return ref;
+}
+
+#define get_free_entry() get_free_entries(1)
+
+static void do_free_callbacks(void)
+{
+	struct gnttab_free_callback *callback, *next;
+
+	callback = gnttab_free_callback_list;
+	gnttab_free_callback_list = NULL;
+
+	while (callback != NULL) {
+		next = callback->next;
+		if (gnttab_free_count >= callback->count) {
+			callback->next = NULL;
+			callback->fn(callback->arg);
+		} else {
+			callback->next = gnttab_free_callback_list;
+			gnttab_free_callback_list = callback;
+		}
+		callback = next;
+	}
+}
+
+static inline void check_free_callbacks(void)
+{
+	if (unlikely(gnttab_free_callback_list))
+		do_free_callbacks();
+}
+
+static void put_free_entry(grant_ref_t ref)
+{
+	unsigned long flags;
+	spin_lock_irqsave(&gnttab_list_lock, flags);
+	gnttab_entry(ref) = gnttab_free_head;
+	gnttab_free_head = ref;
+	gnttab_free_count++;
+	check_free_callbacks();
+	spin_unlock_irqrestore(&gnttab_list_lock, flags);
+}
+
+/*
+ * Public grant-issuing interface functions
+ */
+
+int gnttab_grant_foreign_access(domid_t domid, unsigned long frame,
+				int flags)
+{
+	int ref;
+
+	if (unlikely((ref = get_free_entry()) < 0))
+		return -ENOSPC;
+
+	shared[ref].frame = frame;
+	shared[ref].domid = domid;
+	wmb();
+	BUG_ON(flags & (GTF_accept_transfer | GTF_reading | GTF_writing));
+	shared[ref].flags = GTF_permit_access | flags;
+
+	return ref;
+}
+EXPORT_SYMBOL_GPL(gnttab_grant_foreign_access);
+
+void gnttab_grant_foreign_access_ref(grant_ref_t ref, domid_t domid,
+				     unsigned long frame, int flags)
+{
+	shared[ref].frame = frame;
+	shared[ref].domid = domid;
+	wmb();
+	BUG_ON(flags & (GTF_accept_transfer | GTF_reading | GTF_writing));
+	shared[ref].flags = GTF_permit_access | flags;
+}
+EXPORT_SYMBOL_GPL(gnttab_grant_foreign_access_ref);
+
+
+int gnttab_query_foreign_access(grant_ref_t ref)
+{
+	u16 nflags;
+
+	nflags = shared[ref].flags;
+
+	return (nflags & (GTF_reading|GTF_writing));
+}
+EXPORT_SYMBOL_GPL(gnttab_query_foreign_access);
+
+int gnttab_end_foreign_access_ref(grant_ref_t ref)
+{
+	u16 flags, nflags;
+
+	nflags = shared[ref].flags;
+	do {
+		if ((flags = nflags) & (GTF_reading|GTF_writing)) {
+			printk(KERN_DEBUG "WARNING: g.e. still in use!\n");
+			return 0;
+		}
+	} while ((nflags = synch_cmpxchg_subword(&shared[ref].flags, flags, 0)) !=
+		 flags);
+
+	return 1;
+}
+EXPORT_SYMBOL_GPL(gnttab_end_foreign_access_ref);
+
+void gnttab_end_foreign_access(grant_ref_t ref, unsigned long page)
+{
+	if (gnttab_end_foreign_access_ref(ref)) {
+		put_free_entry(ref);
+		if (page != 0)
+			free_page(page);
+	} else {
+		/* XXX This needs to be fixed so that the ref and page are
+		   placed on a list to be freed up later. */
+		printk(KERN_DEBUG
+		       "WARNING: leaking g.e. and page still in use!\n");
+	}
+}
+EXPORT_SYMBOL_GPL(gnttab_end_foreign_access);
+
+int gnttab_grant_foreign_transfer(domid_t domid, unsigned long pfn)
+{
+	int ref;
+
+	if (unlikely((ref = get_free_entry()) < 0))
+		return -ENOSPC;
+	gnttab_grant_foreign_transfer_ref(ref, domid, pfn);
+
+	return ref;
+}
+EXPORT_SYMBOL_GPL(gnttab_grant_foreign_transfer);
+
+void gnttab_grant_foreign_transfer_ref(grant_ref_t ref, domid_t domid,
+				       unsigned long pfn)
+{
+	shared[ref].frame = pfn;
+	shared[ref].domid = domid;
+	wmb();
+	shared[ref].flags = GTF_accept_transfer;
+}
+EXPORT_SYMBOL_GPL(gnttab_grant_foreign_transfer_ref);
+
+unsigned long gnttab_end_foreign_transfer_ref(grant_ref_t ref)
+{
+	unsigned long frame;
+	u16           flags;
+
+	/*
+	 * If a transfer is not even yet started, try to reclaim the grant
+	 * reference and return failure (== 0).
+	 */
+	while (!((flags = shared[ref].flags) & GTF_transfer_committed)) {
+		if (synch_cmpxchg_subword(&shared[ref].flags, flags, 0) == flags)
+			return 0;
+		cpu_relax();
+	}
+
+	/* If a transfer is in progress then wait until it is completed. */
+	while (!(flags & GTF_transfer_completed)) {
+		flags = shared[ref].flags;
+		cpu_relax();
+	}
+
+	/* Read the frame number /after/ reading completion status. */
+	rmb();
+	frame = shared[ref].frame;
+	BUG_ON(frame == 0);
+
+	return frame;
+}
+EXPORT_SYMBOL_GPL(gnttab_end_foreign_transfer_ref);
+
+unsigned long gnttab_end_foreign_transfer(grant_ref_t ref)
+{
+	unsigned long frame = gnttab_end_foreign_transfer_ref(ref);
+	put_free_entry(ref);
+	return frame;
+}
+EXPORT_SYMBOL_GPL(gnttab_end_foreign_transfer);
+
+void gnttab_free_grant_reference(grant_ref_t ref)
+{
+	put_free_entry(ref);
+}
+EXPORT_SYMBOL_GPL(gnttab_free_grant_reference);
+
+void gnttab_free_grant_references(grant_ref_t head)
+{
+	grant_ref_t ref;
+	unsigned long flags;
+	int count = 1;
+	if (head == GNTTAB_LIST_END)
+		return;
+	spin_lock_irqsave(&gnttab_list_lock, flags);
+	ref = head;
+	while (gnttab_entry(ref) != GNTTAB_LIST_END) {
+		ref = gnttab_entry(ref);
+		count++;
+	}
+	gnttab_entry(ref) = gnttab_free_head;
+	gnttab_free_head = head;
+	gnttab_free_count += count;
+	check_free_callbacks();
+	spin_unlock_irqrestore(&gnttab_list_lock, flags);
+}
+EXPORT_SYMBOL_GPL(gnttab_free_grant_references);
+
+int gnttab_alloc_grant_references(u16 count, grant_ref_t *head)
+{
+	int h = get_free_entries(count);
+
+	if (h < 0)
+		return -ENOSPC;
+
+	*head = h;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(gnttab_alloc_grant_references);
+
+int gnttab_empty_grant_references(const grant_ref_t *private_head)
+{
+	return (*private_head == GNTTAB_LIST_END);
+}
+EXPORT_SYMBOL_GPL(gnttab_empty_grant_references);
+
+int gnttab_claim_grant_reference(grant_ref_t *private_head)
+{
+	grant_ref_t g = *private_head;
+	if (unlikely(g == GNTTAB_LIST_END))
+		return -ENOSPC;
+	*private_head = gnttab_entry(g);
+	return g;
+}
+EXPORT_SYMBOL_GPL(gnttab_claim_grant_reference);
+
+void gnttab_release_grant_reference(grant_ref_t *private_head,
+				    grant_ref_t release)
+{
+	gnttab_entry(release) = *private_head;
+	*private_head = release;
+}
+EXPORT_SYMBOL_GPL(gnttab_release_grant_reference);
+
+void gnttab_request_free_callback(struct gnttab_free_callback *callback,
+				  void (*fn)(void *), void *arg, u16 count)
+{
+	unsigned long flags;
+	spin_lock_irqsave(&gnttab_list_lock, flags);
+	if (callback->next)
+		goto out;
+	callback->fn = fn;
+	callback->arg = arg;
+	callback->count = count;
+	callback->next = gnttab_free_callback_list;
+	gnttab_free_callback_list = callback;
+	check_free_callbacks();
+out:
+	spin_unlock_irqrestore(&gnttab_list_lock, flags);
+}
+EXPORT_SYMBOL_GPL(gnttab_request_free_callback);
+
+void gnttab_cancel_free_callback(struct gnttab_free_callback *callback)
+{
+	struct gnttab_free_callback **pcb;
+	unsigned long flags;
+
+	spin_lock_irqsave(&gnttab_list_lock, flags);
+	for (pcb = &gnttab_free_callback_list; *pcb; pcb = &(*pcb)->next) {
+		if (*pcb == callback) {
+			*pcb = callback->next;
+			break;
+		}
+	}
+	spin_unlock_irqrestore(&gnttab_list_lock, flags);
+}
+EXPORT_SYMBOL_GPL(gnttab_cancel_free_callback);
+
+static int grow_gnttab_list(unsigned int more_frames)
+{
+	unsigned int new_nr_grant_frames, extra_entries, i;
+	unsigned int nr_glist_frames, new_nr_glist_frames;
+
+	new_nr_grant_frames = nr_grant_frames + more_frames;
+	extra_entries       = more_frames * ENTRIES_PER_GRANT_FRAME;
+
+	nr_glist_frames = nr_freelist_frames(nr_grant_frames);
+	new_nr_glist_frames = nr_freelist_frames(new_nr_grant_frames);
+	for (i = nr_glist_frames; i < new_nr_glist_frames; i++) {
+		gnttab_list[i] = (grant_ref_t *)__get_free_page(GFP_ATOMIC);
+		if (!gnttab_list[i])
+			goto grow_nomem;
+	}
+
+	for (i = ENTRIES_PER_GRANT_FRAME * nr_grant_frames;
+	     i < ENTRIES_PER_GRANT_FRAME * new_nr_grant_frames - 1; i++)
+		gnttab_entry(i) = i + 1;
+
+	gnttab_entry(i) = gnttab_free_head;
+	gnttab_free_head = ENTRIES_PER_GRANT_FRAME * nr_grant_frames;
+	gnttab_free_count += extra_entries;
+
+	nr_grant_frames = new_nr_grant_frames;
+
+	check_free_callbacks();
+
+	return 0;
+	
+grow_nomem:
+	for ( ; i >= nr_glist_frames; i--)
+		free_page((unsigned long) gnttab_list[i]);
+	return -ENOMEM;
+}
+
+static unsigned int __max_nr_grant_frames(void)
+{
+	struct gnttab_query_size query;
+	int rc;
+
+	query.dom = DOMID_SELF;
+
+	rc = HYPERVISOR_grant_table_op(GNTTABOP_query_size, &query, 1);
+	if ((rc < 0) || (query.status != GNTST_okay))
+		return 4; /* Legacy max supported number of frames */
+
+	return query.max_nr_frames;
+}
+
+static inline unsigned int max_nr_grant_frames(void)
+{
+	unsigned int xen_max = __max_nr_grant_frames();
+
+	if (xen_max > boot_max_nr_grant_frames)
+		return boot_max_nr_grant_frames;
+	return xen_max;
+}
+
+#ifdef CONFIG_XEN
+
+static DEFINE_SEQLOCK(gnttab_dma_lock);
+
+#ifdef CONFIG_X86
+static int map_pte_fn(pte_t *pte, struct page *pmd_page,
+		      unsigned long addr, void *data)
+{
+	unsigned long **frames = (unsigned long **)data;
+
+	set_pte_at(&init_mm, addr, pte, pfn_pte_ma((*frames)[0], PAGE_KERNEL));
+	(*frames)++;
+	return 0;
+}
+
+static int unmap_pte_fn(pte_t *pte, struct page *pmd_page,
+			unsigned long addr, void *data)
+{
+
+	set_pte_at(&init_mm, addr, pte, __pte(0));
+	return 0;
+}
+
+void *arch_gnttab_alloc_shared(unsigned long *frames)
+{
+	struct vm_struct *area;
+	area = alloc_vm_area(PAGE_SIZE * max_nr_grant_frames());
+	BUG_ON(area == NULL);
+	return area->addr;
+}
+#endif /* CONFIG_X86 */
+
+static int gnttab_map(unsigned int start_idx, unsigned int end_idx)
+{
+	struct gnttab_setup_table setup;
+	unsigned long *frames;
+	unsigned int nr_gframes = end_idx + 1;
+	int rc;
+
+	frames = kmalloc(nr_gframes * sizeof(unsigned long), GFP_ATOMIC);
+	if (!frames)
+		return -ENOMEM;
+
+	setup.dom        = DOMID_SELF;
+	setup.nr_frames  = nr_gframes;
+	set_xen_guest_handle(setup.frame_list, frames);
+
+	rc = HYPERVISOR_grant_table_op(GNTTABOP_setup_table, &setup, 1);
+	if (rc == -ENOSYS) {
+		kfree(frames);
+		return -ENOSYS;
+	}
+
+	BUG_ON(rc || setup.status);
+
+	if (shared == NULL)
+		shared = arch_gnttab_alloc_shared(frames);
+
+#ifdef CONFIG_X86
+	rc = apply_to_page_range(&init_mm, (unsigned long)shared,
+				 PAGE_SIZE * nr_gframes,
+				 map_pte_fn, &frames);
+	BUG_ON(rc);
+	frames -= nr_gframes; /* adjust after map_pte_fn() */
+#endif /* CONFIG_X86 */
+
+	kfree(frames);
+
+	return 0;
+}
+
+static void gnttab_page_free(struct page *page)
+{
+	ClearPageForeign(page);
+	gnttab_reset_grant_page(page);
+	put_page(page);
+}
+
+/*
+ * Must not be called with IRQs off.  This should only be used on the
+ * slow path.
+ *
+ * Copy a foreign granted page to local memory.
+ */
+int gnttab_copy_grant_page(grant_ref_t ref, struct page **pagep)
+{
+	struct gnttab_unmap_and_replace unmap;
+	mmu_update_t mmu;
+	struct page *page;
+	struct page *new_page;
+	void *new_addr;
+	void *addr;
+	paddr_t pfn;
+	maddr_t mfn;
+	maddr_t new_mfn;
+	int err;
+
+	page = *pagep;
+	if (!get_page_unless_zero(page))
+		return -ENOENT;
+
+	err = -ENOMEM;
+	new_page = alloc_page(GFP_ATOMIC | __GFP_NOWARN);
+	if (!new_page)
+		goto out;
+
+	new_addr = page_address(new_page);
+	addr = page_address(page);
+	memcpy(new_addr, addr, PAGE_SIZE);
+
+	pfn = page_to_pfn(page);
+	mfn = pfn_to_mfn(pfn);
+	new_mfn = virt_to_mfn(new_addr);
+
+	write_seqlock(&gnttab_dma_lock);
+
+	/* Make seq visible before checking page_mapped. */
+	smp_mb();
+
+	/* Has the page been DMA-mapped? */
+	if (unlikely(page_mapped(page))) {
+		write_sequnlock(&gnttab_dma_lock);
+		put_page(new_page);
+		err = -EBUSY;
+		goto out;
+	}
+
+	if (!xen_feature(XENFEAT_auto_translated_physmap))
+		set_phys_to_machine(pfn, new_mfn);
+
+	gnttab_set_replace_op(&unmap, (unsigned long)addr,
+			      (unsigned long)new_addr, ref);
+
+	err = HYPERVISOR_grant_table_op(GNTTABOP_unmap_and_replace,
+					&unmap, 1);
+	BUG_ON(err);
+	BUG_ON(unmap.status);
+
+	write_sequnlock(&gnttab_dma_lock);
+
+	if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+		set_phys_to_machine(page_to_pfn(new_page), INVALID_P2M_ENTRY);
+
+		mmu.ptr = (new_mfn << PAGE_SHIFT) | MMU_MACHPHYS_UPDATE;
+		mmu.val = pfn;
+		err = HYPERVISOR_mmu_update(&mmu, 1, NULL, DOMID_SELF);
+		BUG_ON(err);
+	}
+
+	new_page->mapping = page->mapping;
+	new_page->index = page->index;
+	set_bit(PG_foreign, &new_page->flags);
+	*pagep = new_page;
+
+	SetPageForeign(page, gnttab_page_free);
+	page->mapping = NULL;
+
+out:
+	put_page(page);
+	return err;
+}
+EXPORT_SYMBOL_GPL(gnttab_copy_grant_page);
+
+void gnttab_reset_grant_page(struct page *page)
+{
+	init_page_count(page);
+	reset_page_mapcount(page);
+}
+EXPORT_SYMBOL_GPL(gnttab_reset_grant_page);
+
+/*
+ * Keep track of foreign pages marked as PageForeign so that we don't
+ * return them to the remote domain prematurely.
+ *
+ * PageForeign pages are pinned down by increasing their mapcount.
+ *
+ * All other pages are simply returned as is.
+ */
+void __gnttab_dma_map_page(struct page *page)
+{
+	unsigned int seq;
+
+	if (!is_running_on_xen() || !PageForeign(page))
+		return;
+
+	do {
+		seq = read_seqbegin(&gnttab_dma_lock);
+
+		if (gnttab_dma_local_pfn(page))
+			break;
+
+		atomic_set(&page->_mapcount, 0);
+
+		/* Make _mapcount visible before read_seqretry. */
+		smp_mb();
+	} while (unlikely(read_seqretry(&gnttab_dma_lock, seq)));
+}
+
+int gnttab_resume(void)
+{
+	if (max_nr_grant_frames() < nr_grant_frames)
+		return -ENOSYS;
+	return gnttab_map(0, nr_grant_frames - 1);
+}
+
+int gnttab_suspend(void)
+{
+#ifdef CONFIG_X86
+	apply_to_page_range(&init_mm, (unsigned long)shared,
+			    PAGE_SIZE * nr_grant_frames,
+			    unmap_pte_fn, NULL);
+#endif
+	return 0;
+}
+
+#else /* !CONFIG_XEN */
+
+#include <platform-pci.h>
+
+static unsigned long resume_frames;
+
+static int gnttab_map(unsigned int start_idx, unsigned int end_idx)
+{
+	struct xen_add_to_physmap xatp;
+	unsigned int i = end_idx;
+
+	/* Loop backwards, so that the first hypercall has the largest index,
+	 * ensuring that the table will grow only once.
+	 */
+	do {
+		xatp.domid = DOMID_SELF;
+		xatp.idx = i;
+		xatp.space = XENMAPSPACE_grant_table;
+		xatp.gpfn = (resume_frames >> PAGE_SHIFT) + i;
+		if (HYPERVISOR_memory_op(XENMEM_add_to_physmap, &xatp))
+			BUG();
+	} while (i-- > start_idx);
+
+	return 0;
+}
+
+int gnttab_resume(void)
+{
+	unsigned int max_nr_gframes, nr_gframes;
+
+	nr_gframes = nr_grant_frames;
+	max_nr_gframes = max_nr_grant_frames();
+	if (max_nr_gframes < nr_gframes)
+		return -ENOSYS;
+
+	if (!resume_frames) {
+		resume_frames = alloc_xen_mmio(PAGE_SIZE * max_nr_gframes);
+		shared = ioremap(resume_frames, PAGE_SIZE * max_nr_gframes);
+		if (shared == NULL) {
+			printk("error to ioremap gnttab share frames\n");
+			return -1;
+		}
+	}
+
+	gnttab_map(0, nr_gframes - 1);
+
+	return 0;
+}
+
+#endif /* !CONFIG_XEN */
+
+static int gnttab_expand(unsigned int req_entries)
+{
+	int rc;
+	unsigned int cur, extra;
+
+	cur = nr_grant_frames;
+	extra = ((req_entries + (ENTRIES_PER_GRANT_FRAME-1)) /
+		 ENTRIES_PER_GRANT_FRAME);
+	if (cur + extra > max_nr_grant_frames())
+		return -ENOSPC;
+
+	if ((rc = gnttab_map(cur, cur + extra - 1)) == 0)
+		rc = grow_gnttab_list(extra);
+
+	return rc;
+}
+
+int __devinit gnttab_init(void)
+{
+	int i;
+	unsigned int max_nr_glist_frames, nr_glist_frames;
+	unsigned int nr_init_grefs;
+
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	nr_grant_frames = 1;
+	boot_max_nr_grant_frames = __max_nr_grant_frames();
+
+	/* Determine the maximum number of frames required for the
+	 * grant reference free list on the current hypervisor.
+	 */
+	max_nr_glist_frames = nr_freelist_frames(boot_max_nr_grant_frames);
+
+	gnttab_list = kmalloc(max_nr_glist_frames * sizeof(grant_ref_t *),
+			      GFP_KERNEL);
+	if (gnttab_list == NULL)
+		return -ENOMEM;
+
+	nr_glist_frames = nr_freelist_frames(nr_grant_frames);
+	for (i = 0; i < nr_glist_frames; i++) {
+		gnttab_list[i] = (grant_ref_t *)__get_free_page(GFP_KERNEL);
+		if (gnttab_list[i] == NULL)
+			goto ini_nomem;
+	}
+
+	if (gnttab_resume() < 0)
+		return -ENODEV;
+
+	nr_init_grefs = nr_grant_frames * ENTRIES_PER_GRANT_FRAME;
+
+	for (i = NR_RESERVED_ENTRIES; i < nr_init_grefs - 1; i++)
+		gnttab_entry(i) = i + 1;
+
+	gnttab_entry(nr_init_grefs - 1) = GNTTAB_LIST_END;
+	gnttab_free_count = nr_init_grefs - NR_RESERVED_ENTRIES;
+	gnttab_free_head  = NR_RESERVED_ENTRIES;
+
+	return 0;
+
+ ini_nomem:
+	for (i--; i >= 0; i--)
+		free_page((unsigned long)gnttab_list[i]);
+	kfree(gnttab_list);
+	return -ENOMEM;
+}
+
+#ifdef CONFIG_XEN
+core_initcall(gnttab_init);
+#endif
diff -rpuN linux-2.6.18.8/drivers/xen/core/hypervisor_sysfs.c linux-2.6.18-xen-3.3.0/drivers/xen/core/hypervisor_sysfs.c
--- linux-2.6.18.8/drivers/xen/core/hypervisor_sysfs.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/core/hypervisor_sysfs.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,57 @@
+/*
+ *  copyright (c) 2006 IBM Corporation
+ *  Authored by: Mike D. Day <ncmike@us.ibm.com>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License version 2 as
+ *  published by the Free Software Foundation.
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/kobject.h>
+#include <xen/hypervisor_sysfs.h>
+#include <asm/hypervisor.h>
+
+static ssize_t hyp_sysfs_show(struct kobject *kobj,
+			      struct attribute *attr,
+			      char *buffer)
+{
+	struct hyp_sysfs_attr *hyp_attr;
+	hyp_attr = container_of(attr, struct hyp_sysfs_attr, attr);
+	if (hyp_attr->show)
+		return hyp_attr->show(hyp_attr, buffer);
+	return 0;
+}
+
+static ssize_t hyp_sysfs_store(struct kobject *kobj,
+			       struct attribute *attr,
+			       const char *buffer,
+			       size_t len)
+{
+	struct hyp_sysfs_attr *hyp_attr;
+	hyp_attr = container_of(attr, struct hyp_sysfs_attr, attr);
+	if (hyp_attr->store)
+		return hyp_attr->store(hyp_attr, buffer, len);
+	return 0;
+}
+
+static struct sysfs_ops hyp_sysfs_ops = {
+	.show = hyp_sysfs_show,
+	.store = hyp_sysfs_store,
+};
+
+static struct kobj_type hyp_sysfs_kobj_type = {
+	.sysfs_ops = &hyp_sysfs_ops,
+};
+
+static int __init hypervisor_subsys_init(void)
+{
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	hypervisor_subsys.kset.kobj.ktype = &hyp_sysfs_kobj_type;
+	return 0;
+}
+
+device_initcall(hypervisor_subsys_init);
diff -rpuN linux-2.6.18.8/drivers/xen/core/machine_kexec.c linux-2.6.18-xen-3.3.0/drivers/xen/core/machine_kexec.c
--- linux-2.6.18.8/drivers/xen/core/machine_kexec.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/core/machine_kexec.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,207 @@
+/*
+ * drivers/xen/core/machine_kexec.c 
+ * handle transition of Linux booting another kernel
+ */
+
+#include <linux/kexec.h>
+#include <xen/interface/kexec.h>
+#include <linux/mm.h>
+#include <linux/bootmem.h>
+
+extern void machine_kexec_setup_load_arg(xen_kexec_image_t *xki, 
+					 struct kimage *image);
+extern int machine_kexec_setup_resources(struct resource *hypervisor,
+					 struct resource *phys_cpus,
+					 int nr_phys_cpus);
+extern void machine_kexec_register_resources(struct resource *res);
+
+static int __initdata xen_max_nr_phys_cpus;
+static struct resource xen_hypervisor_res;
+static struct resource *xen_phys_cpus;
+
+size_t vmcoreinfo_size_xen;
+unsigned long paddr_vmcoreinfo_xen;
+
+void __init xen_machine_kexec_setup_resources(void)
+{
+	xen_kexec_range_t range;
+	struct resource *res;
+	int k = 0;
+
+	if (!is_initial_xendomain())
+		return;
+
+	/* determine maximum number of physical cpus */
+
+	while (1) {
+		memset(&range, 0, sizeof(range));
+		range.range = KEXEC_RANGE_MA_CPU;
+		range.nr = k;
+
+		if(HYPERVISOR_kexec_op(KEXEC_CMD_kexec_get_range, &range))
+			break;
+
+		k++;
+	}
+
+	if (k == 0)
+		return;
+
+	xen_max_nr_phys_cpus = k;
+
+	/* allocate xen_phys_cpus */
+
+	xen_phys_cpus = alloc_bootmem_low(k * sizeof(struct resource));
+	BUG_ON(xen_phys_cpus == NULL);
+
+	/* fill in xen_phys_cpus with per-cpu crash note information */
+
+	for (k = 0; k < xen_max_nr_phys_cpus; k++) {
+		memset(&range, 0, sizeof(range));
+		range.range = KEXEC_RANGE_MA_CPU;
+		range.nr = k;
+
+		if (HYPERVISOR_kexec_op(KEXEC_CMD_kexec_get_range, &range))
+			goto err;
+
+		res = xen_phys_cpus + k;
+
+		memset(res, 0, sizeof(*res));
+		res->name = "Crash note";
+		res->start = range.start;
+		res->end = range.start + range.size - 1;
+		res->flags = IORESOURCE_BUSY | IORESOURCE_MEM;
+	}
+
+	/* fill in xen_hypervisor_res with hypervisor machine address range */
+
+	memset(&range, 0, sizeof(range));
+	range.range = KEXEC_RANGE_MA_XEN;
+
+	if (HYPERVISOR_kexec_op(KEXEC_CMD_kexec_get_range, &range))
+		goto err;
+
+	xen_hypervisor_res.name = "Hypervisor code and data";
+	xen_hypervisor_res.start = range.start;
+	xen_hypervisor_res.end = range.start + range.size - 1;
+	xen_hypervisor_res.flags = IORESOURCE_BUSY | IORESOURCE_MEM;
+
+	/* fill in crashk_res if range is reserved by hypervisor */
+
+	memset(&range, 0, sizeof(range));
+	range.range = KEXEC_RANGE_MA_CRASH;
+
+	if (HYPERVISOR_kexec_op(KEXEC_CMD_kexec_get_range, &range))
+		goto err;
+
+	if (range.size) {
+		crashk_res.start = range.start;
+		crashk_res.end = range.start + range.size - 1;
+	}
+
+	/* get physical address of vmcoreinfo */
+	memset(&range, 0, sizeof(range));
+	range.range = KEXEC_RANGE_MA_VMCOREINFO;
+
+	if (HYPERVISOR_kexec_op(KEXEC_CMD_kexec_get_range, &range))
+		return;
+
+	if (range.size) {
+		paddr_vmcoreinfo_xen = range.start;
+		vmcoreinfo_size_xen = range.size;
+	}
+
+	if (machine_kexec_setup_resources(&xen_hypervisor_res, xen_phys_cpus,
+					  xen_max_nr_phys_cpus))
+		goto err;
+
+	return;
+
+ err:
+	/*
+	 * It isn't possible to free xen_phys_cpus this early in the
+	 * boot. Failure at this stage is unexpected and the amount of
+	 * memory is small therefore we tolerate the potential leak.
+         */
+	xen_max_nr_phys_cpus = 0;
+	return;
+}
+
+void __init xen_machine_kexec_register_resources(struct resource *res)
+{
+	request_resource(res, &xen_hypervisor_res);
+	machine_kexec_register_resources(res);
+}
+
+static void setup_load_arg(xen_kexec_image_t *xki, struct kimage *image)
+{
+	machine_kexec_setup_load_arg(xki, image);
+
+	xki->indirection_page = image->head;
+	xki->start_address = image->start;
+}
+
+/*
+ * Load the image into xen so xen can kdump itself
+ * This might have been done in prepare, but prepare
+ * is currently called too early. It might make sense
+ * to move prepare, but for now, just add an extra hook.
+ */
+int xen_machine_kexec_load(struct kimage *image)
+{
+	xen_kexec_load_t xkl;
+
+	memset(&xkl, 0, sizeof(xkl));
+	xkl.type = image->type;
+	setup_load_arg(&xkl.image, image);
+	return HYPERVISOR_kexec_op(KEXEC_CMD_kexec_load, &xkl);
+}
+
+/*
+ * Unload the image that was stored by machine_kexec_load()
+ * This might have been done in machine_kexec_cleanup() but it
+ * is called too late, and its possible xen could try and kdump
+ * using resources that have been freed.
+ */
+void xen_machine_kexec_unload(struct kimage *image)
+{
+	xen_kexec_load_t xkl;
+
+	memset(&xkl, 0, sizeof(xkl));
+	xkl.type = image->type;
+	WARN_ON(HYPERVISOR_kexec_op(KEXEC_CMD_kexec_unload, &xkl));
+}
+
+/*
+ * Do not allocate memory (or fail in any way) in machine_kexec().
+ * We are past the point of no return, committed to rebooting now.
+ *
+ * This has the hypervisor move to the prefered reboot CPU, 
+ * stop all CPUs and kexec. That is it combines machine_shutdown()
+ * and machine_kexec() in Linux kexec terms.
+ */
+NORET_TYPE void machine_kexec(struct kimage *image)
+{
+	xen_kexec_exec_t xke;
+
+	memset(&xke, 0, sizeof(xke));
+	xke.type = image->type;
+	VOID(HYPERVISOR_kexec_op(KEXEC_CMD_kexec, &xke));
+	panic("KEXEC_CMD_kexec hypercall should not return\n");
+}
+
+void machine_shutdown(void)
+{
+	/* do nothing */
+}
+
+
+/*
+ * Local variables:
+ *  c-file-style: "linux"
+ *  indent-tabs-mode: t
+ *  c-indent-level: 8
+ *  c-basic-offset: 8
+ *  tab-width: 8
+ * End:
+ */
diff -rpuN linux-2.6.18.8/drivers/xen/core/machine_reboot.c linux-2.6.18-xen-3.3.0/drivers/xen/core/machine_reboot.c
--- linux-2.6.18.8/drivers/xen/core/machine_reboot.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/core/machine_reboot.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,247 @@
+#include <linux/version.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/unistd.h>
+#include <linux/module.h>
+#include <linux/reboot.h>
+#include <linux/sysrq.h>
+#include <linux/stringify.h>
+#include <linux/stop_machine.h>
+#include <asm/irq.h>
+#include <asm/mmu_context.h>
+#include <xen/evtchn.h>
+#include <asm/hypervisor.h>
+#include <xen/xenbus.h>
+#include <linux/cpu.h>
+#include <xen/gnttab.h>
+#include <xen/xencons.h>
+#include <xen/cpu_hotplug.h>
+#include <xen/interface/vcpu.h>
+
+#if defined(__i386__) || defined(__x86_64__)
+
+/*
+ * Power off function, if any
+ */
+void (*pm_power_off)(void);
+EXPORT_SYMBOL(pm_power_off);
+
+void machine_emergency_restart(void)
+{
+	/* We really want to get pending console data out before we die. */
+	xencons_force_flush();
+	HYPERVISOR_shutdown(SHUTDOWN_reboot);
+}
+
+void machine_restart(char * __unused)
+{
+	machine_emergency_restart();
+}
+
+void machine_halt(void)
+{
+	machine_power_off();
+}
+
+void machine_power_off(void)
+{
+	/* We really want to get pending console data out before we die. */
+	xencons_force_flush();
+	if (pm_power_off)
+		pm_power_off();
+	HYPERVISOR_shutdown(SHUTDOWN_poweroff);
+}
+
+int reboot_thru_bios = 0;	/* for dmi_scan.c */
+EXPORT_SYMBOL(machine_restart);
+EXPORT_SYMBOL(machine_halt);
+EXPORT_SYMBOL(machine_power_off);
+
+static void pre_suspend(void)
+{
+	HYPERVISOR_shared_info = (shared_info_t *)empty_zero_page;
+	WARN_ON(HYPERVISOR_update_va_mapping(fix_to_virt(FIX_SHARED_INFO),
+					     __pte_ma(0), 0));
+
+	xen_start_info->store_mfn = mfn_to_pfn(xen_start_info->store_mfn);
+	xen_start_info->console.domU.mfn =
+		mfn_to_pfn(xen_start_info->console.domU.mfn);
+}
+
+static void post_suspend(int suspend_cancelled)
+{
+	int i, j, k, fpp;
+	unsigned long shinfo_mfn;
+	extern unsigned long max_pfn;
+	extern unsigned long *pfn_to_mfn_frame_list_list;
+	extern unsigned long *pfn_to_mfn_frame_list[];
+
+	if (suspend_cancelled) {
+		xen_start_info->store_mfn =
+			pfn_to_mfn(xen_start_info->store_mfn);
+		xen_start_info->console.domU.mfn =
+			pfn_to_mfn(xen_start_info->console.domU.mfn);
+	} else {
+#ifdef CONFIG_SMP
+		cpu_initialized_map = cpu_online_map;
+#endif
+	}
+
+	shinfo_mfn = xen_start_info->shared_info >> PAGE_SHIFT;
+	if (HYPERVISOR_update_va_mapping(fix_to_virt(FIX_SHARED_INFO),
+					 pfn_pte_ma(shinfo_mfn, PAGE_KERNEL),
+					 0))
+		BUG();
+	HYPERVISOR_shared_info = (shared_info_t *)fix_to_virt(FIX_SHARED_INFO);
+
+	memset(empty_zero_page, 0, PAGE_SIZE);
+
+	fpp = PAGE_SIZE/sizeof(unsigned long);
+	for (i = 0, j = 0, k = -1; i < max_pfn; i += fpp, j++) {
+		if ((j % fpp) == 0) {
+			k++;
+			pfn_to_mfn_frame_list_list[k] =
+				virt_to_mfn(pfn_to_mfn_frame_list[k]);
+			j = 0;
+		}
+		pfn_to_mfn_frame_list[k][j] =
+			virt_to_mfn(&phys_to_machine_mapping[i]);
+	}
+	HYPERVISOR_shared_info->arch.max_pfn = max_pfn;
+	HYPERVISOR_shared_info->arch.pfn_to_mfn_frame_list_list =
+		virt_to_mfn(pfn_to_mfn_frame_list_list);
+}
+
+#else /* !(defined(__i386__) || defined(__x86_64__)) */
+
+#ifndef HAVE_XEN_PRE_SUSPEND
+#define xen_pre_suspend()	((void)0)
+#endif
+
+#ifndef HAVE_XEN_POST_SUSPEND
+#define xen_post_suspend(x)	((void)0)
+#endif
+
+#define switch_idle_mm()	((void)0)
+#define mm_pin_all()		((void)0)
+#define pre_suspend()		xen_pre_suspend()
+#define post_suspend(x)		xen_post_suspend(x)
+
+#endif
+
+struct suspend {
+	int fast_suspend;
+	void (*resume_notifier)(int);
+};
+
+static int take_machine_down(void *_suspend)
+{
+	struct suspend *suspend = _suspend;
+	int suspend_cancelled, err;
+	extern void time_resume(void);
+
+	if (suspend->fast_suspend) {
+		BUG_ON(!irqs_disabled());
+	} else {
+		BUG_ON(irqs_disabled());
+
+		for (;;) {
+			err = smp_suspend();
+			if (err)
+				return err;
+
+			xenbus_suspend();
+			preempt_disable();
+
+			if (num_online_cpus() == 1)
+				break;
+
+			preempt_enable();
+			xenbus_suspend_cancel();
+		}
+
+		local_irq_disable();
+	}
+
+	mm_pin_all();
+	gnttab_suspend();
+	pre_suspend();
+
+	/*
+	 * This hypercall returns 1 if suspend was cancelled or the domain was
+	 * merely checkpointed, and 0 if it is resuming in a new domain.
+	 */
+	suspend_cancelled = HYPERVISOR_suspend(virt_to_mfn(xen_start_info));
+
+	suspend->resume_notifier(suspend_cancelled);
+	post_suspend(suspend_cancelled);
+	gnttab_resume();
+	if (!suspend_cancelled) {
+		irq_resume();
+#ifdef __x86_64__
+		/*
+		 * Older versions of Xen do not save/restore the user %cr3.
+		 * We do it here just in case, but there's no need if we are
+		 * in fast-suspend mode as that implies a new enough Xen.
+		 */
+		if (!suspend->fast_suspend)
+			xen_new_user_pt(__pa(__user_pgd(
+				current->active_mm->pgd)));
+#endif
+	}
+	time_resume();
+
+	if (!suspend->fast_suspend)
+		local_irq_enable();
+
+	return suspend_cancelled;
+}
+
+int __xen_suspend(int fast_suspend, void (*resume_notifier)(int))
+{
+	int err, suspend_cancelled;
+	struct suspend suspend;
+
+	BUG_ON(smp_processor_id() != 0);
+	BUG_ON(in_interrupt());
+
+#if defined(__i386__) || defined(__x86_64__)
+	if (xen_feature(XENFEAT_auto_translated_physmap)) {
+		printk(KERN_WARNING "Cannot suspend in "
+		       "auto_translated_physmap mode.\n");
+		return -EOPNOTSUPP;
+	}
+#endif
+
+	/* If we are definitely UP then 'slow mode' is actually faster. */
+	if (num_possible_cpus() == 1)
+		fast_suspend = 0;
+
+	suspend.fast_suspend = fast_suspend;
+	suspend.resume_notifier = resume_notifier;
+
+	if (fast_suspend) {
+		xenbus_suspend();
+		err = stop_machine_run(take_machine_down, &suspend, 0);
+		if (err < 0)
+			xenbus_suspend_cancel();
+	} else {
+		err = take_machine_down(&suspend);
+	}
+
+	if (err < 0)
+		return err;
+
+	suspend_cancelled = err;
+	if (!suspend_cancelled) {
+		xencons_resume();
+		xenbus_resume();
+	} else {
+		xenbus_suspend_cancel();
+	}
+
+	if (!fast_suspend)
+		smp_resume();
+
+	return 0;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/core/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/core/Makefile
--- linux-2.6.18.8/drivers/xen/core/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/core/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,14 @@
+#
+# Makefile for the linux kernel.
+#
+
+obj-y := evtchn.o gnttab.o features.o reboot.o machine_reboot.o firmware.o
+
+obj-$(CONFIG_PCI)		+= pci.o
+obj-$(CONFIG_PROC_FS)		+= xen_proc.o
+obj-$(CONFIG_SYS_HYPERVISOR)	+= hypervisor_sysfs.o
+obj-$(CONFIG_HOTPLUG_CPU)	+= cpu_hotplug.o
+obj-$(CONFIG_XEN_SYSFS)		+= xen_sysfs.o
+obj-$(CONFIG_XEN_SMPBOOT)	+= smpboot.o
+obj-$(CONFIG_KEXEC)		+= machine_kexec.o
+obj-$(CONFIG_XEN_XENCOMM)	+= xencomm.o
diff -rpuN linux-2.6.18.8/drivers/xen/core/pci.c linux-2.6.18-xen-3.3.0/drivers/xen/core/pci.c
--- linux-2.6.18.8/drivers/xen/core/pci.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/core/pci.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,67 @@
+/*
+ * vim:shiftwidth=8:noexpandtab
+ */
+
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/pci.h>
+#include <xen/interface/physdev.h>
+
+static int (*pci_bus_probe)(struct device *dev);
+static int (*pci_bus_remove)(struct device *dev);
+
+static int pci_bus_probe_wrapper(struct device *dev)
+{
+	int r;
+	struct pci_dev *pci_dev = to_pci_dev(dev);
+	struct physdev_manage_pci manage_pci;
+	manage_pci.bus = pci_dev->bus->number;
+	manage_pci.devfn = pci_dev->devfn;
+
+	r = HYPERVISOR_physdev_op(PHYSDEVOP_manage_pci_add, &manage_pci);
+	if (r && r != -ENOSYS)
+		return r;
+
+	r = pci_bus_probe(dev);
+	if (r) {
+		int ret;
+
+		ret = HYPERVISOR_physdev_op(PHYSDEVOP_manage_pci_remove,
+					    &manage_pci);
+		WARN_ON(ret && ret != -ENOSYS);
+	}
+
+	return r;
+}
+
+static int pci_bus_remove_wrapper(struct device *dev)
+{
+	int r;
+	struct pci_dev *pci_dev = to_pci_dev(dev);
+	struct physdev_manage_pci manage_pci;
+	manage_pci.bus = pci_dev->bus->number;
+	manage_pci.devfn = pci_dev->devfn;
+
+	r = pci_bus_remove(dev);
+	/* dev and pci_dev are no longer valid!! */
+
+	WARN_ON(HYPERVISOR_physdev_op(PHYSDEVOP_manage_pci_remove,
+		&manage_pci));
+	return r;
+}
+
+static int __init hook_pci_bus(void)
+{
+	if (!is_running_on_xen() || !is_initial_xendomain())
+		return 0;
+
+	pci_bus_probe = pci_bus_type.probe;
+	pci_bus_type.probe = pci_bus_probe_wrapper;
+
+	pci_bus_remove = pci_bus_type.remove;
+	pci_bus_type.remove = pci_bus_remove_wrapper;
+
+	return 0;
+}
+
+core_initcall(hook_pci_bus);
diff -rpuN linux-2.6.18.8/drivers/xen/core/reboot.c linux-2.6.18-xen-3.3.0/drivers/xen/core/reboot.c
--- linux-2.6.18.8/drivers/xen/core/reboot.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/core/reboot.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,335 @@
+#define __KERNEL_SYSCALLS__
+#include <linux/version.h>
+#include <linux/kernel.h>
+#include <linux/unistd.h>
+#include <linux/module.h>
+#include <linux/reboot.h>
+#include <linux/sysrq.h>
+#include <asm/hypervisor.h>
+#include <xen/xenbus.h>
+#include <xen/evtchn.h>
+#include <linux/kmod.h>
+#include <linux/slab.h>
+#include <linux/workqueue.h>
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+MODULE_LICENSE("Dual BSD/GPL");
+
+#define SHUTDOWN_INVALID  -1
+#define SHUTDOWN_POWEROFF  0
+#define SHUTDOWN_SUSPEND   2
+#define SHUTDOWN_RESUMING  3
+#define SHUTDOWN_HALT      4
+
+/* Ignore multiple shutdown requests. */
+static int shutting_down = SHUTDOWN_INVALID;
+
+/* Was last suspend request cancelled? */
+static int suspend_cancelled;
+
+/* Can we leave APs online when we suspend? */
+static int fast_suspend;
+
+static void __shutdown_handler(void *unused);
+static DECLARE_WORK(shutdown_work, __shutdown_handler, NULL);
+
+static int setup_suspend_evtchn(void);
+
+int __xen_suspend(int fast_suspend, void (*resume_notifier)(int));
+
+static int shutdown_process(void *__unused)
+{
+	static char *envp[] = { "HOME=/", "TERM=linux",
+				"PATH=/sbin:/usr/sbin:/bin:/usr/bin", NULL };
+	static char *poweroff_argv[] = { "/sbin/poweroff", NULL };
+
+	extern asmlinkage long sys_reboot(int magic1, int magic2,
+					  unsigned int cmd, void *arg);
+
+	if ((shutting_down == SHUTDOWN_POWEROFF) ||
+	    (shutting_down == SHUTDOWN_HALT)) {
+		if (call_usermodehelper("/sbin/poweroff", poweroff_argv,
+					envp, 0) < 0) {
+#ifdef CONFIG_XEN
+			sys_reboot(LINUX_REBOOT_MAGIC1,
+				   LINUX_REBOOT_MAGIC2,
+				   LINUX_REBOOT_CMD_POWER_OFF,
+				   NULL);
+#endif /* CONFIG_XEN */
+		}
+	}
+
+	shutting_down = SHUTDOWN_INVALID; /* could try again */
+
+	return 0;
+}
+
+static void xen_resume_notifier(int _suspend_cancelled)
+{
+	int old_state = xchg(&shutting_down, SHUTDOWN_RESUMING);
+	BUG_ON(old_state != SHUTDOWN_SUSPEND);
+	suspend_cancelled = _suspend_cancelled;
+}
+
+static int xen_suspend(void *__unused)
+{
+	int err, old_state;
+
+	daemonize("suspend");
+	err = set_cpus_allowed(current, cpumask_of_cpu(0));
+	if (err) {
+		printk(KERN_ERR "Xen suspend can't run on CPU0 (%d)\n", err);
+		goto fail;
+	}
+
+	do {
+		err = __xen_suspend(fast_suspend, xen_resume_notifier);
+		if (err) {
+			printk(KERN_ERR "Xen suspend failed (%d)\n", err);
+			goto fail;
+		}
+		if (!suspend_cancelled)
+			setup_suspend_evtchn();
+		old_state = cmpxchg(
+			&shutting_down, SHUTDOWN_RESUMING, SHUTDOWN_INVALID);
+	} while (old_state == SHUTDOWN_SUSPEND);
+
+	switch (old_state) {
+	case SHUTDOWN_INVALID:
+	case SHUTDOWN_SUSPEND:
+		BUG();
+	case SHUTDOWN_RESUMING:
+		break;
+	default:
+		schedule_work(&shutdown_work);
+		break;
+	}
+
+	return 0;
+
+ fail:
+	old_state = xchg(&shutting_down, SHUTDOWN_INVALID);
+	BUG_ON(old_state != SHUTDOWN_SUSPEND);
+	return 0;
+}
+
+static void switch_shutdown_state(int new_state)
+{
+	int prev_state, old_state = SHUTDOWN_INVALID;
+
+	/* We only drive shutdown_state into an active state. */
+	if (new_state == SHUTDOWN_INVALID)
+		return;
+
+	do {
+		/* We drop this transition if already in an active state. */
+		if ((old_state != SHUTDOWN_INVALID) &&
+		    (old_state != SHUTDOWN_RESUMING))
+			return;
+		/* Attempt to transition. */
+		prev_state = old_state;
+		old_state = cmpxchg(&shutting_down, old_state, new_state);
+	} while (old_state != prev_state);
+
+	/* Either we kick off the work, or we leave it to xen_suspend(). */
+	if (old_state == SHUTDOWN_INVALID)
+		schedule_work(&shutdown_work);
+	else
+		BUG_ON(old_state != SHUTDOWN_RESUMING);
+}
+
+static void __shutdown_handler(void *unused)
+{
+	int err;
+
+	err = kernel_thread((shutting_down == SHUTDOWN_SUSPEND) ?
+			    xen_suspend : shutdown_process,
+			    NULL, CLONE_FS | CLONE_FILES);
+
+	if (err < 0) {
+		printk(KERN_WARNING "Error creating shutdown process (%d): "
+		       "retrying...\n", -err);
+		schedule_delayed_work(&shutdown_work, HZ/2);
+	}
+}
+
+static void shutdown_handler(struct xenbus_watch *watch,
+			     const char **vec, unsigned int len)
+{
+	extern void ctrl_alt_del(void);
+	char *str;
+	struct xenbus_transaction xbt;
+	int err, new_state = SHUTDOWN_INVALID;
+
+	if ((shutting_down != SHUTDOWN_INVALID) &&
+	    (shutting_down != SHUTDOWN_RESUMING))
+		return;
+
+ again:
+	err = xenbus_transaction_start(&xbt);
+	if (err)
+		return;
+
+	str = (char *)xenbus_read(xbt, "control", "shutdown", NULL);
+	/* Ignore read errors and empty reads. */
+	if (XENBUS_IS_ERR_READ(str)) {
+		xenbus_transaction_end(xbt, 1);
+		return;
+	}
+
+	xenbus_write(xbt, "control", "shutdown", "");
+
+	err = xenbus_transaction_end(xbt, 0);
+	if (err == -EAGAIN) {
+		kfree(str);
+		goto again;
+	}
+
+	if (strcmp(str, "poweroff") == 0)
+		new_state = SHUTDOWN_POWEROFF;
+	else if (strcmp(str, "reboot") == 0)
+		ctrl_alt_del();
+	else if (strcmp(str, "suspend") == 0)
+		new_state = SHUTDOWN_SUSPEND;
+	else if (strcmp(str, "halt") == 0)
+		new_state = SHUTDOWN_HALT;
+	else
+		printk("Ignoring shutdown request: %s\n", str);
+
+	switch_shutdown_state(new_state);
+
+	kfree(str);
+}
+
+static void sysrq_handler(struct xenbus_watch *watch, const char **vec,
+			  unsigned int len)
+{
+	char sysrq_key = '\0';
+	struct xenbus_transaction xbt;
+	int err;
+
+ again:
+	err = xenbus_transaction_start(&xbt);
+	if (err)
+		return;
+	if (!xenbus_scanf(xbt, "control", "sysrq", "%c", &sysrq_key)) {
+		printk(KERN_ERR "Unable to read sysrq code in "
+		       "control/sysrq\n");
+		xenbus_transaction_end(xbt, 1);
+		return;
+	}
+
+	if (sysrq_key != '\0')
+		xenbus_printf(xbt, "control", "sysrq", "%c", '\0');
+
+	err = xenbus_transaction_end(xbt, 0);
+	if (err == -EAGAIN)
+		goto again;
+
+#ifdef CONFIG_MAGIC_SYSRQ
+	if (sysrq_key != '\0')
+		handle_sysrq(sysrq_key, NULL, NULL);
+#endif
+}
+
+static struct xenbus_watch shutdown_watch = {
+	.node = "control/shutdown",
+	.callback = shutdown_handler
+};
+
+static struct xenbus_watch sysrq_watch = {
+	.node = "control/sysrq",
+	.callback = sysrq_handler
+};
+
+static irqreturn_t suspend_int(int irq, void* dev_id, struct pt_regs *ptregs)
+{
+	switch_shutdown_state(SHUTDOWN_SUSPEND);
+	return IRQ_HANDLED;
+}
+
+static int setup_suspend_evtchn(void)
+{
+	static int irq;
+	int port;
+	char portstr[16];
+
+	if (irq > 0)
+		unbind_from_irqhandler(irq, NULL);
+
+	irq = bind_listening_port_to_irqhandler(0, suspend_int, 0, "suspend",
+						NULL);
+	if (irq <= 0)
+		return -1;
+
+	port = irq_to_evtchn_port(irq);
+	printk(KERN_INFO "suspend: event channel %d\n", port);
+	sprintf(portstr, "%d", port);
+	xenbus_write(XBT_NIL, "device/suspend", "event-channel", portstr);
+
+	return 0;
+}
+
+static int setup_shutdown_watcher(void)
+{
+	int err;
+
+	xenbus_scanf(XBT_NIL, "control",
+		     "platform-feature-multiprocessor-suspend",
+		     "%d", &fast_suspend);
+
+	err = register_xenbus_watch(&shutdown_watch);
+	if (err) {
+		printk(KERN_ERR "Failed to set shutdown watcher\n");
+		return err;
+	}
+
+	err = register_xenbus_watch(&sysrq_watch);
+	if (err) {
+		printk(KERN_ERR "Failed to set sysrq watcher\n");
+		return err;
+	}
+
+	/* suspend event channel */
+	err = setup_suspend_evtchn();
+	if (err) {
+		printk(KERN_ERR "Failed to register suspend event channel\n");
+		return err;
+	}
+
+	return 0;
+}
+
+#ifdef CONFIG_XEN
+
+static int shutdown_event(struct notifier_block *notifier,
+			  unsigned long event,
+			  void *data)
+{
+	setup_shutdown_watcher();
+	return NOTIFY_DONE;
+}
+
+static int __init setup_shutdown_event(void)
+{
+	static struct notifier_block xenstore_notifier = {
+		.notifier_call = shutdown_event
+	};
+	register_xenstore_notifier(&xenstore_notifier);
+
+	return 0;
+}
+
+subsys_initcall(setup_shutdown_event);
+
+#else /* !defined(CONFIG_XEN) */
+
+int xen_reboot_init(void)
+{
+	return setup_shutdown_watcher();
+}
+
+#endif /* !defined(CONFIG_XEN) */
diff -rpuN linux-2.6.18.8/drivers/xen/core/smpboot.c linux-2.6.18-xen-3.3.0/drivers/xen/core/smpboot.c
--- linux-2.6.18.8/drivers/xen/core/smpboot.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/core/smpboot.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,464 @@
+/*
+ *	Xen SMP booting functions
+ *
+ *	See arch/i386/kernel/smpboot.c for copyright and credits for derived
+ *	portions of this file.
+ */
+
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/mm.h>
+#include <linux/sched.h>
+#include <linux/kernel_stat.h>
+#include <linux/smp_lock.h>
+#include <linux/irq.h>
+#include <linux/bootmem.h>
+#include <linux/notifier.h>
+#include <linux/cpu.h>
+#include <linux/percpu.h>
+#include <asm/desc.h>
+#include <asm/arch_hooks.h>
+#include <asm/pgalloc.h>
+#include <xen/evtchn.h>
+#include <xen/interface/vcpu.h>
+#include <xen/cpu_hotplug.h>
+#include <xen/xenbus.h>
+
+extern irqreturn_t smp_reschedule_interrupt(int, void *, struct pt_regs *);
+extern irqreturn_t smp_call_function_interrupt(int, void *, struct pt_regs *);
+
+extern int local_setup_timer(unsigned int cpu);
+extern void local_teardown_timer(unsigned int cpu);
+
+extern void hypervisor_callback(void);
+extern void failsafe_callback(void);
+extern void system_call(void);
+extern void smp_trap_init(trap_info_t *);
+
+/* Number of siblings per CPU package */
+int smp_num_siblings = 1;
+
+cpumask_t cpu_online_map;
+EXPORT_SYMBOL(cpu_online_map);
+cpumask_t cpu_possible_map;
+EXPORT_SYMBOL(cpu_possible_map);
+cpumask_t cpu_initialized_map;
+
+struct cpuinfo_x86 cpu_data[NR_CPUS] __cacheline_aligned;
+EXPORT_SYMBOL(cpu_data);
+
+#ifdef CONFIG_HOTPLUG_CPU
+DEFINE_PER_CPU(int, cpu_state) = { 0 };
+#endif
+
+static DEFINE_PER_CPU(int, resched_irq);
+static DEFINE_PER_CPU(int, callfunc_irq);
+static char resched_name[NR_CPUS][15];
+static char callfunc_name[NR_CPUS][15];
+
+u8 cpu_2_logical_apicid[NR_CPUS] = { [0 ... NR_CPUS-1] = BAD_APICID };
+
+cpumask_t cpu_sibling_map[NR_CPUS] __cacheline_aligned;
+cpumask_t cpu_core_map[NR_CPUS] __cacheline_aligned;
+EXPORT_SYMBOL(cpu_core_map);
+
+#if defined(__i386__)
+u8 x86_cpu_to_apicid[NR_CPUS] = { [0 ... NR_CPUS-1] = 0xff };
+EXPORT_SYMBOL(x86_cpu_to_apicid);
+#elif !defined(CONFIG_X86_IO_APIC)
+unsigned int maxcpus = NR_CPUS;
+#endif
+
+void __init prefill_possible_map(void)
+{
+	int i, rc;
+
+	for_each_possible_cpu(i)
+	    if (i != smp_processor_id())
+		return;
+
+	for (i = 0; i < NR_CPUS; i++) {
+		rc = HYPERVISOR_vcpu_op(VCPUOP_is_up, i, NULL);
+		if (rc >= 0)
+			cpu_set(i, cpu_possible_map);
+	}
+}
+
+void __init smp_alloc_memory(void)
+{
+}
+
+static inline void
+set_cpu_sibling_map(unsigned int cpu)
+{
+	cpu_data[cpu].phys_proc_id = cpu;
+	cpu_data[cpu].cpu_core_id  = 0;
+
+	cpu_sibling_map[cpu] = cpumask_of_cpu(cpu);
+	cpu_core_map[cpu]    = cpumask_of_cpu(cpu);
+
+	cpu_data[cpu].booted_cores = 1;
+}
+
+static void
+remove_siblinginfo(unsigned int cpu)
+{
+	cpu_data[cpu].phys_proc_id = BAD_APICID;
+	cpu_data[cpu].cpu_core_id  = BAD_APICID;
+
+	cpus_clear(cpu_sibling_map[cpu]);
+	cpus_clear(cpu_core_map[cpu]);
+
+	cpu_data[cpu].booted_cores = 0;
+}
+
+static int __cpuinit xen_smp_intr_init(unsigned int cpu)
+{
+	int rc;
+
+	per_cpu(resched_irq, cpu) = per_cpu(callfunc_irq, cpu) = -1;
+
+	sprintf(resched_name[cpu], "resched%u", cpu);
+	rc = bind_ipi_to_irqhandler(RESCHEDULE_VECTOR,
+				    cpu,
+				    smp_reschedule_interrupt,
+				    SA_INTERRUPT,
+				    resched_name[cpu],
+				    NULL);
+	if (rc < 0)
+		goto fail;
+	per_cpu(resched_irq, cpu) = rc;
+
+	sprintf(callfunc_name[cpu], "callfunc%u", cpu);
+	rc = bind_ipi_to_irqhandler(CALL_FUNCTION_VECTOR,
+				    cpu,
+				    smp_call_function_interrupt,
+				    SA_INTERRUPT,
+				    callfunc_name[cpu],
+				    NULL);
+	if (rc < 0)
+		goto fail;
+	per_cpu(callfunc_irq, cpu) = rc;
+
+	if ((cpu != 0) && ((rc = local_setup_timer(cpu)) != 0))
+		goto fail;
+
+	return 0;
+
+ fail:
+	if (per_cpu(resched_irq, cpu) >= 0)
+		unbind_from_irqhandler(per_cpu(resched_irq, cpu), NULL);
+	if (per_cpu(callfunc_irq, cpu) >= 0)
+		unbind_from_irqhandler(per_cpu(callfunc_irq, cpu), NULL);
+	return rc;
+}
+
+#ifdef CONFIG_HOTPLUG_CPU
+static void xen_smp_intr_exit(unsigned int cpu)
+{
+	if (cpu != 0)
+		local_teardown_timer(cpu);
+
+	unbind_from_irqhandler(per_cpu(resched_irq, cpu), NULL);
+	unbind_from_irqhandler(per_cpu(callfunc_irq, cpu), NULL);
+}
+#endif
+
+void __cpuinit cpu_bringup(void)
+{
+	cpu_init();
+	identify_cpu(cpu_data + smp_processor_id());
+	touch_softlockup_watchdog();
+	preempt_disable();
+	local_irq_enable();
+}
+
+static void __cpuinit cpu_bringup_and_idle(void)
+{
+	cpu_bringup();
+	cpu_idle();
+}
+
+static void __cpuinit cpu_initialize_context(unsigned int cpu)
+{
+	/* vcpu_guest_context_t is too large to allocate on the stack.
+	 * Hence we allocate statically and protect it with a lock */
+	static vcpu_guest_context_t ctxt;
+	static DEFINE_SPINLOCK(ctxt_lock);
+
+	struct task_struct *idle = idle_task(cpu);
+#ifdef __x86_64__
+	struct desc_ptr *gdt_descr = &cpu_gdt_descr[cpu];
+#else
+	struct Xgt_desc_struct *gdt_descr = &per_cpu(cpu_gdt_descr, cpu);
+#endif
+
+	if (cpu_test_and_set(cpu, cpu_initialized_map))
+		return;
+
+	spin_lock(&ctxt_lock);
+
+	memset(&ctxt, 0, sizeof(ctxt));
+
+	ctxt.flags = VGCF_IN_KERNEL;
+	ctxt.user_regs.ds = __USER_DS;
+	ctxt.user_regs.es = __USER_DS;
+	ctxt.user_regs.fs = 0;
+	ctxt.user_regs.gs = 0;
+	ctxt.user_regs.ss = __KERNEL_DS;
+	ctxt.user_regs.eip = (unsigned long)cpu_bringup_and_idle;
+	ctxt.user_regs.eflags = X86_EFLAGS_IF | 0x1000; /* IOPL_RING1 */
+
+	memset(&ctxt.fpu_ctxt, 0, sizeof(ctxt.fpu_ctxt));
+
+	smp_trap_init(ctxt.trap_ctxt);
+
+	ctxt.ldt_ents = 0;
+
+	ctxt.gdt_frames[0] = virt_to_mfn(gdt_descr->address);
+	ctxt.gdt_ents      = gdt_descr->size / 8;
+
+#ifdef __i386__
+	ctxt.user_regs.cs = __KERNEL_CS;
+	ctxt.user_regs.esp = idle->thread.esp0 - sizeof(struct pt_regs);
+
+	ctxt.kernel_ss = __KERNEL_DS;
+	ctxt.kernel_sp = idle->thread.esp0;
+
+	ctxt.event_callback_cs     = __KERNEL_CS;
+	ctxt.event_callback_eip    = (unsigned long)hypervisor_callback;
+	ctxt.failsafe_callback_cs  = __KERNEL_CS;
+	ctxt.failsafe_callback_eip = (unsigned long)failsafe_callback;
+
+	ctxt.ctrlreg[3] = xen_pfn_to_cr3(virt_to_mfn(swapper_pg_dir));
+#else /* __x86_64__ */
+	ctxt.user_regs.cs = __KERNEL_CS;
+	ctxt.user_regs.esp = idle->thread.rsp0 - sizeof(struct pt_regs);
+
+	ctxt.kernel_ss = __KERNEL_DS;
+	ctxt.kernel_sp = idle->thread.rsp0;
+
+	ctxt.event_callback_eip    = (unsigned long)hypervisor_callback;
+	ctxt.failsafe_callback_eip = (unsigned long)failsafe_callback;
+	ctxt.syscall_callback_eip  = (unsigned long)system_call;
+
+	ctxt.ctrlreg[3] = xen_pfn_to_cr3(virt_to_mfn(init_level4_pgt));
+
+	ctxt.gs_base_kernel = (unsigned long)(cpu_pda(cpu));
+#endif
+
+	if (HYPERVISOR_vcpu_op(VCPUOP_initialise, cpu, &ctxt))
+		BUG();
+
+	spin_unlock(&ctxt_lock);
+}
+
+void __init smp_prepare_cpus(unsigned int max_cpus)
+{
+	unsigned int cpu;
+	struct task_struct *idle;
+	int apicid, acpiid;
+	struct vcpu_get_physid cpu_id;
+#ifdef __x86_64__
+	struct desc_ptr *gdt_descr;
+#else
+	struct Xgt_desc_struct *gdt_descr;
+#endif
+
+	apicid = 0;
+	if (HYPERVISOR_vcpu_op(VCPUOP_get_physid, 0, &cpu_id) == 0) {
+		apicid = xen_vcpu_physid_to_x86_apicid(cpu_id.phys_id);
+		acpiid = xen_vcpu_physid_to_x86_acpiid(cpu_id.phys_id);
+#ifdef CONFIG_ACPI
+		if (acpiid != 0xff)
+			x86_acpiid_to_apicid[acpiid] = apicid;
+#endif
+	}
+	boot_cpu_data.apicid = apicid;
+	cpu_data[0] = boot_cpu_data;
+
+	cpu_2_logical_apicid[0] = apicid;
+	x86_cpu_to_apicid[0] = apicid;
+
+	current_thread_info()->cpu = 0;
+
+	for (cpu = 0; cpu < NR_CPUS; cpu++) {
+		cpus_clear(cpu_sibling_map[cpu]);
+		cpus_clear(cpu_core_map[cpu]);
+	}
+
+	set_cpu_sibling_map(0);
+
+	if (xen_smp_intr_init(0))
+		BUG();
+
+	cpu_initialized_map = cpumask_of_cpu(0);
+
+	/* Restrict the possible_map according to max_cpus. */
+	while ((num_possible_cpus() > 1) && (num_possible_cpus() > max_cpus)) {
+		for (cpu = NR_CPUS-1; !cpu_isset(cpu, cpu_possible_map); cpu--)
+			continue;
+		cpu_clear(cpu, cpu_possible_map);
+	}
+
+	for_each_possible_cpu (cpu) {
+		if (cpu == 0)
+			continue;
+
+#ifdef __x86_64__
+		gdt_descr = &cpu_gdt_descr[cpu];
+#else
+		gdt_descr = &per_cpu(cpu_gdt_descr, cpu);
+#endif
+		gdt_descr->address = get_zeroed_page(GFP_KERNEL);
+		if (unlikely(!gdt_descr->address)) {
+			printk(KERN_CRIT "CPU%d failed to allocate GDT\n",
+			       cpu);
+			continue;
+		}
+		gdt_descr->size = GDT_SIZE;
+		memcpy((void *)gdt_descr->address, cpu_gdt_table, GDT_SIZE);
+		make_page_readonly(
+			(void *)gdt_descr->address,
+			XENFEAT_writable_descriptor_tables);
+
+		apicid = cpu;
+		if (HYPERVISOR_vcpu_op(VCPUOP_get_physid, cpu, &cpu_id) == 0) {
+			apicid = xen_vcpu_physid_to_x86_apicid(cpu_id.phys_id);
+			acpiid = xen_vcpu_physid_to_x86_acpiid(cpu_id.phys_id);
+#ifdef CONFIG_ACPI
+			if (acpiid != 0xff)
+				x86_acpiid_to_apicid[acpiid] = apicid;
+#endif
+		}
+		cpu_data[cpu] = boot_cpu_data;
+		cpu_data[cpu].apicid = apicid;
+
+		cpu_2_logical_apicid[cpu] = apicid;
+		x86_cpu_to_apicid[cpu] = apicid;
+
+		idle = fork_idle(cpu);
+		if (IS_ERR(idle))
+			panic("failed fork for CPU %d", cpu);
+
+#ifdef __x86_64__
+		cpu_pda(cpu)->pcurrent = idle;
+		cpu_pda(cpu)->cpunumber = cpu;
+		clear_ti_thread_flag(idle->thread_info, TIF_FORK);
+#endif
+
+		irq_ctx_init(cpu);
+
+#ifdef CONFIG_HOTPLUG_CPU
+		if (is_initial_xendomain())
+			cpu_set(cpu, cpu_present_map);
+#else
+		cpu_set(cpu, cpu_present_map);
+#endif
+	}
+
+	init_xenbus_allowed_cpumask();
+
+#ifdef CONFIG_X86_IO_APIC
+	/*
+	 * Here we can be sure that there is an IO-APIC in the system. Let's
+	 * go and set it up:
+	 */
+	if (!skip_ioapic_setup && nr_ioapics)
+		setup_IO_APIC();
+#endif
+}
+
+void __devinit smp_prepare_boot_cpu(void)
+{
+	prefill_possible_map();
+}
+
+#ifdef CONFIG_HOTPLUG_CPU
+
+/*
+ * Initialize cpu_present_map late to skip SMP boot code in init/main.c.
+ * But do it early enough to catch critical for_each_present_cpu() loops
+ * in i386-specific code.
+ */
+static int __init initialize_cpu_present_map(void)
+{
+	cpu_present_map = cpu_possible_map;
+	return 0;
+}
+core_initcall(initialize_cpu_present_map);
+
+int __cpu_disable(void)
+{
+	cpumask_t map = cpu_online_map;
+	unsigned int cpu = smp_processor_id();
+
+	if (cpu == 0)
+		return -EBUSY;
+
+	remove_siblinginfo(cpu);
+
+	cpu_clear(cpu, map);
+	fixup_irqs(map);
+	cpu_clear(cpu, cpu_online_map);
+
+	return 0;
+}
+
+void __cpu_die(unsigned int cpu)
+{
+	while (HYPERVISOR_vcpu_op(VCPUOP_is_up, cpu, NULL)) {
+		current->state = TASK_UNINTERRUPTIBLE;
+		schedule_timeout(HZ/10);
+	}
+
+	xen_smp_intr_exit(cpu);
+
+	if (num_online_cpus() == 1)
+		alternatives_smp_switch(0);
+}
+
+#endif /* CONFIG_HOTPLUG_CPU */
+
+int __cpuinit __cpu_up(unsigned int cpu)
+{
+	int rc;
+
+	rc = cpu_up_check(cpu);
+	if (rc)
+		return rc;
+
+	cpu_initialize_context(cpu);
+
+	if (num_online_cpus() == 1)
+		alternatives_smp_switch(1);
+
+	/* This must be done before setting cpu_online_map */
+	set_cpu_sibling_map(cpu);
+	wmb();
+
+	rc = xen_smp_intr_init(cpu);
+	if (rc) {
+		remove_siblinginfo(cpu);
+		return rc;
+	}
+
+	cpu_set(cpu, cpu_online_map);
+
+	rc = HYPERVISOR_vcpu_op(VCPUOP_up, cpu, NULL);
+	BUG_ON(rc);
+
+	return 0;
+}
+
+void __init smp_cpus_done(unsigned int max_cpus)
+{
+}
+
+#ifndef CONFIG_X86_LOCAL_APIC
+int setup_profiling_timer(unsigned int multiplier)
+{
+	return -EINVAL;
+}
+#endif
diff -rpuN linux-2.6.18.8/drivers/xen/core/xencomm.c linux-2.6.18-xen-3.3.0/drivers/xen/core/xencomm.c
--- linux-2.6.18.8/drivers/xen/core/xencomm.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/core/xencomm.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,229 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ * 
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ * 
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ *
+ * Copyright (C) IBM Corp. 2006
+ *
+ * Authors: Hollis Blanchard <hollisb@us.ibm.com>
+ */
+
+#include <linux/gfp.h>
+#include <linux/mm.h>
+#include <asm/page.h>
+#include <xen/xencomm.h>
+#include <xen/interface/xen.h>
+#ifdef __ia64__
+#include <asm/xen/xencomm.h>	/* for is_kern_addr() */
+#endif
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+static int xencomm_init(struct xencomm_desc *desc,
+			void *buffer, unsigned long bytes)
+{
+	unsigned long recorded = 0;
+	int i = 0;
+
+	while ((recorded < bytes) && (i < desc->nr_addrs)) {
+		unsigned long vaddr = (unsigned long)buffer + recorded;
+		unsigned long paddr;
+		int offset;
+		int chunksz;
+
+		offset = vaddr % PAGE_SIZE; /* handle partial pages */
+		chunksz = min(PAGE_SIZE - offset, bytes - recorded);
+
+		paddr = xencomm_vtop(vaddr);
+		if (paddr == ~0UL) {
+			printk("%s: couldn't translate vaddr %lx\n",
+			       __func__, vaddr);
+			return -EINVAL;
+		}
+
+		desc->address[i++] = paddr;
+		recorded += chunksz;
+	}
+
+	if (recorded < bytes) {
+		printk("%s: could only translate %ld of %ld bytes\n",
+		       __func__, recorded, bytes);
+		return -ENOSPC;
+	}
+
+	/* mark remaining addresses invalid (just for safety) */
+	while (i < desc->nr_addrs)
+		desc->address[i++] = XENCOMM_INVALID;
+
+	desc->magic = XENCOMM_MAGIC;
+
+	return 0;
+}
+
+static struct xencomm_desc *xencomm_alloc(gfp_t gfp_mask,
+					  void *buffer, unsigned long bytes)
+{
+	struct xencomm_desc *desc;
+	unsigned long buffer_ulong = (unsigned long)buffer;
+	unsigned long start = buffer_ulong & PAGE_MASK;
+	unsigned long end = (buffer_ulong + bytes) | ~PAGE_MASK;
+	unsigned long nr_addrs = (end - start + 1) >> PAGE_SHIFT;
+	unsigned long size = sizeof(*desc) +
+		sizeof(desc->address[0]) * nr_addrs;
+
+	/*
+	 * slab allocator returns at least sizeof(void*) aligned pointer.
+	 * When sizeof(*desc) > sizeof(void*), struct xencomm_desc might
+	 * cross page boundary.
+	 */
+	if (sizeof(*desc) > sizeof(void*)) {
+		unsigned long order = get_order(size);
+		desc = (struct xencomm_desc *)__get_free_pages(gfp_mask,
+							       order);
+		if (desc == NULL)
+			return NULL;
+
+		desc->nr_addrs =
+			((PAGE_SIZE << order) - sizeof(struct xencomm_desc)) /
+			sizeof(*desc->address);
+	} else {
+		desc = kmalloc(size, gfp_mask);
+		if (desc == NULL)
+			return NULL;
+
+		desc->nr_addrs = nr_addrs;
+	}
+	return desc;
+}
+
+void xencomm_free(struct xencomm_handle *desc)
+{
+	if (desc && !((ulong)desc & XENCOMM_INLINE_FLAG)) {
+		struct xencomm_desc *desc__ = (struct xencomm_desc*)desc;
+		if (sizeof(*desc__) > sizeof(void*)) {
+			unsigned long size = sizeof(*desc__) +
+				sizeof(desc__->address[0]) * desc__->nr_addrs;
+			unsigned long order = get_order(size);
+			free_pages((unsigned long)__va(desc), order);
+		} else
+			kfree(__va(desc));
+	}
+}
+
+static int xencomm_create(void *buffer, unsigned long bytes, struct xencomm_desc **ret, gfp_t gfp_mask)
+{
+	struct xencomm_desc *desc;
+	int rc;
+
+	pr_debug("%s: %p[%ld]\n", __func__, buffer, bytes);
+
+	if (bytes == 0) {
+		/* don't create a descriptor; Xen recognizes NULL. */
+		BUG_ON(buffer != NULL);
+		*ret = NULL;
+		return 0;
+	}
+
+	BUG_ON(buffer == NULL); /* 'bytes' is non-zero */
+
+	desc = xencomm_alloc(gfp_mask, buffer, bytes);
+	if (!desc) {
+		printk("%s failure\n", "xencomm_alloc");
+		return -ENOMEM;
+	}
+
+	rc = xencomm_init(desc, buffer, bytes);
+	if (rc) {
+		printk("%s failure: %d\n", "xencomm_init", rc);
+		xencomm_free((struct xencomm_handle *)__pa(desc));
+		return rc;
+	}
+
+	*ret = desc;
+	return 0;
+}
+
+/* check if memory address is within VMALLOC region  */
+static int is_phys_contiguous(unsigned long addr)
+{
+	if (!is_kernel_addr(addr))
+		return 0;
+
+	return (addr < VMALLOC_START) || (addr >= VMALLOC_END);
+}
+
+static struct xencomm_handle *xencomm_create_inline(void *ptr)
+{
+	unsigned long paddr;
+
+	BUG_ON(!is_phys_contiguous((unsigned long)ptr));
+
+	paddr = (unsigned long)xencomm_pa(ptr);
+	BUG_ON(paddr & XENCOMM_INLINE_FLAG);
+	return (struct xencomm_handle *)(paddr | XENCOMM_INLINE_FLAG);
+}
+
+/* "mini" routine, for stack-based communications: */
+static int xencomm_create_mini(void *buffer,
+	unsigned long bytes, struct xencomm_mini *xc_desc,
+	struct xencomm_desc **ret)
+{
+	int rc = 0;
+	struct xencomm_desc *desc;
+	BUG_ON(((unsigned long)xc_desc) % sizeof(*xc_desc) != 0);
+
+	desc = (void *)xc_desc; 
+
+	desc->nr_addrs = XENCOMM_MINI_ADDRS;
+
+	if (!(rc = xencomm_init(desc, buffer, bytes)))
+		*ret = desc;
+
+	return rc;
+}
+
+struct xencomm_handle *xencomm_map(void *ptr, unsigned long bytes)
+{
+	int rc;
+	struct xencomm_desc *desc;
+
+	if (is_phys_contiguous((unsigned long)ptr))
+		return xencomm_create_inline(ptr);
+
+	rc = xencomm_create(ptr, bytes, &desc, GFP_KERNEL);
+
+	if (rc || desc == NULL)
+		return NULL;
+
+	return xencomm_pa(desc);
+}
+
+struct xencomm_handle *__xencomm_map_no_alloc(void *ptr, unsigned long bytes, 
+			struct xencomm_mini *xc_desc)
+{
+	int rc;
+	struct xencomm_desc *desc = NULL;
+
+	if (is_phys_contiguous((unsigned long)ptr))
+		return xencomm_create_inline(ptr);
+
+	rc = xencomm_create_mini(ptr, bytes, xc_desc,
+				&desc);
+
+	if (rc)
+		return NULL;
+ 
+	return xencomm_pa(desc);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/core/xen_proc.c linux-2.6.18-xen-3.3.0/drivers/xen/core/xen_proc.c
--- linux-2.6.18.8/drivers/xen/core/xen_proc.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/core/xen_proc.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,23 @@
+
+#include <linux/module.h>
+#include <linux/proc_fs.h>
+#include <xen/xen_proc.h>
+
+static struct proc_dir_entry *xen_base;
+
+struct proc_dir_entry *create_xen_proc_entry(const char *name, mode_t mode)
+{
+	if ( xen_base == NULL )
+		if ( (xen_base = proc_mkdir("xen", &proc_root)) == NULL )
+			panic("Couldn't create /proc/xen");
+	return create_proc_entry(name, mode, xen_base);
+}
+
+EXPORT_SYMBOL_GPL(create_xen_proc_entry); 
+
+void remove_xen_proc_entry(const char *name)
+{
+	remove_proc_entry(name, xen_base);
+}
+
+EXPORT_SYMBOL_GPL(remove_xen_proc_entry); 
diff -rpuN linux-2.6.18.8/drivers/xen/core/xen_sysfs.c linux-2.6.18-xen-3.3.0/drivers/xen/core/xen_sysfs.c
--- linux-2.6.18.8/drivers/xen/core/xen_sysfs.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/core/xen_sysfs.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,420 @@
+/*
+ *  copyright (c) 2006 IBM Corporation
+ *  Authored by: Mike D. Day <ncmike@us.ibm.com>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License version 2 as
+ *  published by the Free Software Foundation.
+ */
+
+#include <linux/err.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <asm/hypervisor.h>
+#include <xen/features.h>
+#include <xen/hypervisor_sysfs.h>
+#include <xen/xenbus.h>
+#include <xen/interface/kexec.h>
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Mike D. Day <ncmike@us.ibm.com>");
+
+static ssize_t type_show(struct hyp_sysfs_attr *attr, char *buffer)
+{
+	return sprintf(buffer, "xen\n");
+}
+
+HYPERVISOR_ATTR_RO(type);
+
+static int __init xen_sysfs_type_init(void)
+{
+	return sysfs_create_file(&hypervisor_subsys.kset.kobj, &type_attr.attr);
+}
+
+static void xen_sysfs_type_destroy(void)
+{
+	sysfs_remove_file(&hypervisor_subsys.kset.kobj, &type_attr.attr);
+}
+
+/* xen version attributes */
+static ssize_t major_show(struct hyp_sysfs_attr *attr, char *buffer)
+{
+	int version = HYPERVISOR_xen_version(XENVER_version, NULL);
+	if (version)
+		return sprintf(buffer, "%d\n", version >> 16);
+	return -ENODEV;
+}
+
+HYPERVISOR_ATTR_RO(major);
+
+static ssize_t minor_show(struct hyp_sysfs_attr *attr, char *buffer)
+{
+	int version = HYPERVISOR_xen_version(XENVER_version, NULL);
+	if (version)
+		return sprintf(buffer, "%d\n", version & 0xff);
+	return -ENODEV;
+}
+
+HYPERVISOR_ATTR_RO(minor);
+
+static ssize_t extra_show(struct hyp_sysfs_attr *attr, char *buffer)
+{
+	int ret = -ENOMEM;
+	char *extra;
+
+	extra = kmalloc(XEN_EXTRAVERSION_LEN, GFP_KERNEL);
+	if (extra) {
+		ret = HYPERVISOR_xen_version(XENVER_extraversion, extra);
+		if (!ret)
+			ret = sprintf(buffer, "%s\n", extra);
+		kfree(extra);
+	}
+
+	return ret;
+}
+
+HYPERVISOR_ATTR_RO(extra);
+
+static struct attribute *version_attrs[] = {
+	&major_attr.attr,
+	&minor_attr.attr,
+	&extra_attr.attr,
+	NULL
+};
+
+static struct attribute_group version_group = {
+	.name = "version",
+	.attrs = version_attrs,
+};
+
+static int __init xen_sysfs_version_init(void)
+{
+	return sysfs_create_group(&hypervisor_subsys.kset.kobj,
+				  &version_group);
+}
+
+static void xen_sysfs_version_destroy(void)
+{
+	sysfs_remove_group(&hypervisor_subsys.kset.kobj, &version_group);
+}
+
+/* UUID */
+
+static ssize_t uuid_show(struct hyp_sysfs_attr *attr, char *buffer)
+{
+	char *vm, *val;
+	int ret;
+	extern int xenstored_ready;
+
+	if (!xenstored_ready)
+		return -EBUSY;
+
+	vm = xenbus_read(XBT_NIL, "vm", "", NULL);
+	if (IS_ERR(vm))
+		return PTR_ERR(vm);
+	val = xenbus_read(XBT_NIL, vm, "uuid", NULL);
+	kfree(vm);
+	if (IS_ERR(val))
+		return PTR_ERR(val);
+	ret = sprintf(buffer, "%s\n", val);
+	kfree(val);
+	return ret;
+}
+
+HYPERVISOR_ATTR_RO(uuid);
+
+static int __init xen_sysfs_uuid_init(void)
+{
+	return sysfs_create_file(&hypervisor_subsys.kset.kobj, &uuid_attr.attr);
+}
+
+static void xen_sysfs_uuid_destroy(void)
+{
+	sysfs_remove_file(&hypervisor_subsys.kset.kobj, &uuid_attr.attr);
+}
+
+/* xen compilation attributes */
+
+static ssize_t compiler_show(struct hyp_sysfs_attr *attr, char *buffer)
+{
+	int ret = -ENOMEM;
+	struct xen_compile_info *info;
+
+	info = kmalloc(sizeof(struct xen_compile_info), GFP_KERNEL);
+	if (info) {
+		ret = HYPERVISOR_xen_version(XENVER_compile_info, info);
+		if (!ret)
+			ret = sprintf(buffer, "%s\n", info->compiler);
+		kfree(info);
+	}
+
+	return ret;
+}
+
+HYPERVISOR_ATTR_RO(compiler);
+
+static ssize_t compiled_by_show(struct hyp_sysfs_attr *attr, char *buffer)
+{
+	int ret = -ENOMEM;
+	struct xen_compile_info *info;
+
+	info = kmalloc(sizeof(struct xen_compile_info), GFP_KERNEL);
+	if (info) {
+		ret = HYPERVISOR_xen_version(XENVER_compile_info, info);
+		if (!ret)
+			ret = sprintf(buffer, "%s\n", info->compile_by);
+		kfree(info);
+	}
+
+	return ret;
+}
+
+HYPERVISOR_ATTR_RO(compiled_by);
+
+static ssize_t compile_date_show(struct hyp_sysfs_attr *attr, char *buffer)
+{
+	int ret = -ENOMEM;
+	struct xen_compile_info *info;
+
+	info = kmalloc(sizeof(struct xen_compile_info), GFP_KERNEL);
+	if (info) {
+		ret = HYPERVISOR_xen_version(XENVER_compile_info, info);
+		if (!ret)
+			ret = sprintf(buffer, "%s\n", info->compile_date);
+		kfree(info);
+	}
+
+	return ret;
+}
+
+HYPERVISOR_ATTR_RO(compile_date);
+
+static struct attribute *xen_compile_attrs[] = {
+	&compiler_attr.attr,
+	&compiled_by_attr.attr,
+	&compile_date_attr.attr,
+	NULL
+};
+
+static struct attribute_group xen_compilation_group = {
+	.name = "compilation",
+	.attrs = xen_compile_attrs,
+};
+
+int __init static xen_compilation_init(void)
+{
+	return sysfs_create_group(&hypervisor_subsys.kset.kobj,
+				  &xen_compilation_group);
+}
+
+static void xen_compilation_destroy(void)
+{
+	sysfs_remove_group(&hypervisor_subsys.kset.kobj,
+			   &xen_compilation_group);
+}
+
+/* xen properties info */
+
+static ssize_t capabilities_show(struct hyp_sysfs_attr *attr, char *buffer)
+{
+	int ret = -ENOMEM;
+	char *caps;
+
+	caps = kmalloc(XEN_CAPABILITIES_INFO_LEN, GFP_KERNEL);
+	if (caps) {
+		ret = HYPERVISOR_xen_version(XENVER_capabilities, caps);
+		if (!ret)
+			ret = sprintf(buffer, "%s\n", caps);
+		kfree(caps);
+	}
+
+	return ret;
+}
+
+HYPERVISOR_ATTR_RO(capabilities);
+
+static ssize_t changeset_show(struct hyp_sysfs_attr *attr, char *buffer)
+{
+	int ret = -ENOMEM;
+	char *cset;
+
+	cset = kmalloc(XEN_CHANGESET_INFO_LEN, GFP_KERNEL);
+	if (cset) {
+		ret = HYPERVISOR_xen_version(XENVER_changeset, cset);
+		if (!ret)
+			ret = sprintf(buffer, "%s\n", cset);
+		kfree(cset);
+	}
+
+	return ret;
+}
+
+HYPERVISOR_ATTR_RO(changeset);
+
+static ssize_t virtual_start_show(struct hyp_sysfs_attr *attr, char *buffer)
+{
+	int ret = -ENOMEM;
+	struct xen_platform_parameters *parms;
+
+	parms = kmalloc(sizeof(struct xen_platform_parameters), GFP_KERNEL);
+	if (parms) {
+		ret = HYPERVISOR_xen_version(XENVER_platform_parameters,
+					     parms);
+		if (!ret)
+			ret = sprintf(buffer, "%lx\n", parms->virt_start);
+		kfree(parms);
+	}
+
+	return ret;
+}
+
+HYPERVISOR_ATTR_RO(virtual_start);
+
+static ssize_t pagesize_show(struct hyp_sysfs_attr *attr, char *buffer)
+{
+	int ret;
+
+	ret = HYPERVISOR_xen_version(XENVER_pagesize, NULL);
+	if (ret > 0)
+		ret = sprintf(buffer, "%x\n", ret);
+
+	return ret;
+}
+
+HYPERVISOR_ATTR_RO(pagesize);
+
+/* eventually there will be several more features to export */
+static ssize_t xen_feature_show(int index, char *buffer)
+{
+	int ret = -ENOMEM;
+	struct xen_feature_info *info;
+
+	info = kmalloc(sizeof(struct xen_feature_info), GFP_KERNEL);
+	if (info) {
+		info->submap_idx = index;
+		ret = HYPERVISOR_xen_version(XENVER_get_features, info);
+		if (!ret)
+			ret = sprintf(buffer, "%d\n", info->submap);
+		kfree(info);
+	}
+
+	return ret;
+}
+
+static ssize_t writable_pt_show(struct hyp_sysfs_attr *attr, char *buffer)
+{
+	return xen_feature_show(XENFEAT_writable_page_tables, buffer);
+}
+
+HYPERVISOR_ATTR_RO(writable_pt);
+
+static struct attribute *xen_properties_attrs[] = {
+	&capabilities_attr.attr,
+	&changeset_attr.attr,
+	&virtual_start_attr.attr,
+	&pagesize_attr.attr,
+	&writable_pt_attr.attr,
+	NULL
+};
+
+static struct attribute_group xen_properties_group = {
+	.name = "properties",
+	.attrs = xen_properties_attrs,
+};
+
+static int __init xen_properties_init(void)
+{
+	return sysfs_create_group(&hypervisor_subsys.kset.kobj,
+				  &xen_properties_group);
+}
+
+static void xen_properties_destroy(void)
+{
+	sysfs_remove_group(&hypervisor_subsys.kset.kobj,
+			   &xen_properties_group);
+}
+
+#ifdef CONFIG_KEXEC
+
+static ssize_t vmcoreinfo_show(struct hyp_sysfs_attr *attr, char *page)
+{
+	extern size_t vmcoreinfo_size_xen;
+	extern unsigned long paddr_vmcoreinfo_xen;
+
+	return sprintf(page, "%lx %zx\n",
+		paddr_vmcoreinfo_xen, vmcoreinfo_size_xen);
+}
+
+HYPERVISOR_ATTR_RO(vmcoreinfo);
+
+static int __init xen_sysfs_vmcoreinfo_init(void)
+{
+	return sysfs_create_file(&hypervisor_subsys.kset.kobj,
+				 &vmcoreinfo_attr.attr);
+}
+
+static void xen_sysfs_vmcoreinfo_destroy(void)
+{
+	sysfs_remove_file(&hypervisor_subsys.kset.kobj, &vmcoreinfo_attr.attr);
+}
+
+#else
+
+#define xen_sysfs_vmcoreinfo_init()	0
+#define xen_sysfs_vmcoreinfo_destroy()	((void)0)
+
+#endif
+
+static int __init hyper_sysfs_init(void)
+{
+	int ret;
+
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	ret = xen_sysfs_type_init();
+	if (ret)
+		goto out;
+	ret = xen_sysfs_version_init();
+	if (ret)
+		goto version_out;
+	ret = xen_compilation_init();
+	if (ret)
+		goto comp_out;
+	ret = xen_sysfs_uuid_init();
+	if (ret)
+		goto uuid_out;
+	ret = xen_properties_init();
+	if (ret)
+		goto prop_out;
+	ret = xen_sysfs_vmcoreinfo_init();
+	if (!ret)
+		goto out;
+
+	xen_properties_destroy();
+prop_out:
+	xen_sysfs_uuid_destroy();
+uuid_out:
+	xen_compilation_destroy();
+comp_out:
+	xen_sysfs_version_destroy();
+version_out:
+	xen_sysfs_type_destroy();
+out:
+	return ret;
+}
+
+static void __exit hyper_sysfs_exit(void)
+{
+	xen_sysfs_vmcoreinfo_destroy();
+	xen_properties_destroy();
+	xen_compilation_destroy();
+	xen_sysfs_uuid_destroy();
+	xen_sysfs_version_destroy();
+	xen_sysfs_type_destroy();
+
+}
+
+module_init(hyper_sysfs_init);
+module_exit(hyper_sysfs_exit);
diff -rpuN linux-2.6.18.8/drivers/xen/evtchn/evtchn.c linux-2.6.18-xen-3.3.0/drivers/xen/evtchn/evtchn.c
--- linux-2.6.18.8/drivers/xen/evtchn/evtchn.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/evtchn/evtchn.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,560 @@
+/******************************************************************************
+ * evtchn.c
+ * 
+ * Driver for receiving and demuxing event-channel signals.
+ * 
+ * Copyright (c) 2004-2005, K A Fraser
+ * Multi-process extensions Copyright (c) 2004, Steven Smith
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/string.h>
+#include <linux/errno.h>
+#include <linux/fs.h>
+#include <linux/errno.h>
+#include <linux/miscdevice.h>
+#include <linux/major.h>
+#include <linux/proc_fs.h>
+#include <linux/stat.h>
+#include <linux/poll.h>
+#include <linux/irq.h>
+#include <linux/init.h>
+#include <linux/gfp.h>
+#include <linux/mutex.h>
+#include <linux/cpu.h>
+#include <xen/evtchn.h>
+#include <xen/public/evtchn.h>
+
+struct per_user_data {
+	/* Notification ring, accessed via /dev/xen/evtchn. */
+#define EVTCHN_RING_SIZE     (PAGE_SIZE / sizeof(evtchn_port_t))
+#define EVTCHN_RING_MASK(_i) ((_i)&(EVTCHN_RING_SIZE-1))
+	evtchn_port_t *ring;
+	unsigned int ring_cons, ring_prod, ring_overflow;
+	struct mutex ring_cons_mutex; /* protect against concurrent readers */
+
+	/* Processes wait on this queue when ring is empty. */
+	wait_queue_head_t evtchn_wait;
+	struct fasync_struct *evtchn_async_queue;
+
+	int bind_cpu;
+	int nr_event_wrong_delivery;
+};
+
+/* Who's bound to each port? */
+static struct per_user_data *port_user[NR_EVENT_CHANNELS];
+static spinlock_t port_user_lock;
+
+void evtchn_device_upcall(int port)
+{
+	struct per_user_data *u;
+
+	spin_lock(&port_user_lock);
+
+	mask_evtchn(port);
+	clear_evtchn(port);
+
+	if ((u = port_user[port]) != NULL) {
+		if ((u->ring_prod - u->ring_cons) < EVTCHN_RING_SIZE) {
+			u->ring[EVTCHN_RING_MASK(u->ring_prod)] = port;
+			wmb(); /* Ensure ring contents visible */
+			if (u->ring_cons == u->ring_prod++) {
+				wake_up_interruptible(&u->evtchn_wait);
+				kill_fasync(&u->evtchn_async_queue,
+					    SIGIO, POLL_IN);
+			}
+		} else {
+			u->ring_overflow = 1;
+		}
+	}
+
+	spin_unlock(&port_user_lock);
+}
+
+static void evtchn_check_wrong_delivery(struct per_user_data *u)
+{
+	evtchn_port_t port;
+	unsigned int current_cpu = smp_processor_id();
+
+	/* Delivered to correct CPU? All is good. */
+	if (u->bind_cpu == current_cpu) {
+		u->nr_event_wrong_delivery = 0;
+		return;
+	}
+
+	/* Tolerate up to 100 consecutive misdeliveries. */
+	if (++u->nr_event_wrong_delivery < 100)
+		return;
+
+	spin_lock_irq(&port_user_lock);
+
+	for (port = 0; port < NR_EVENT_CHANNELS; port++)
+		if (port_user[port] == u)
+			rebind_evtchn_to_cpu(port, current_cpu);
+
+	u->bind_cpu = current_cpu;
+	u->nr_event_wrong_delivery = 0;
+
+	spin_unlock_irq(&port_user_lock);
+}
+
+static ssize_t evtchn_read(struct file *file, char __user *buf,
+			   size_t count, loff_t *ppos)
+{
+	int rc;
+	unsigned int c, p, bytes1 = 0, bytes2 = 0;
+	struct per_user_data *u = file->private_data;
+
+	/* Whole number of ports. */
+	count &= ~(sizeof(evtchn_port_t)-1);
+
+	if (count == 0)
+		return 0;
+
+	if (count > PAGE_SIZE)
+		count = PAGE_SIZE;
+
+	for (;;) {
+		mutex_lock(&u->ring_cons_mutex);
+
+		rc = -EFBIG;
+		if (u->ring_overflow)
+			goto unlock_out;
+
+		if ((c = u->ring_cons) != (p = u->ring_prod))
+			break;
+
+		mutex_unlock(&u->ring_cons_mutex);
+
+		if (file->f_flags & O_NONBLOCK)
+			return -EAGAIN;
+
+		rc = wait_event_interruptible(
+			u->evtchn_wait, u->ring_cons != u->ring_prod);
+		if (rc)
+			return rc;
+	}
+
+	/* Byte lengths of two chunks. Chunk split (if any) is at ring wrap. */
+	if (((c ^ p) & EVTCHN_RING_SIZE) != 0) {
+		bytes1 = (EVTCHN_RING_SIZE - EVTCHN_RING_MASK(c)) *
+			sizeof(evtchn_port_t);
+		bytes2 = EVTCHN_RING_MASK(p) * sizeof(evtchn_port_t);
+	} else {
+		bytes1 = (p - c) * sizeof(evtchn_port_t);
+		bytes2 = 0;
+	}
+
+	/* Truncate chunks according to caller's maximum byte count. */
+	if (bytes1 > count) {
+		bytes1 = count;
+		bytes2 = 0;
+	} else if ((bytes1 + bytes2) > count) {
+		bytes2 = count - bytes1;
+	}
+
+	rc = -EFAULT;
+	rmb(); /* Ensure that we see the port before we copy it. */
+	if (copy_to_user(buf, &u->ring[EVTCHN_RING_MASK(c)], bytes1) ||
+	    ((bytes2 != 0) &&
+	     copy_to_user(&buf[bytes1], &u->ring[0], bytes2)))
+		goto unlock_out;
+	
+	evtchn_check_wrong_delivery(u);
+
+	u->ring_cons += (bytes1 + bytes2) / sizeof(evtchn_port_t);
+	rc = bytes1 + bytes2;
+
+ unlock_out:
+	mutex_unlock(&u->ring_cons_mutex);
+	return rc;
+}
+
+static ssize_t evtchn_write(struct file *file, const char __user *buf,
+			    size_t count, loff_t *ppos)
+{
+	int rc, i;
+	evtchn_port_t *kbuf = (evtchn_port_t *)__get_free_page(GFP_KERNEL);
+	struct per_user_data *u = file->private_data;
+
+	if (kbuf == NULL)
+		return -ENOMEM;
+
+	/* Whole number of ports. */
+	count &= ~(sizeof(evtchn_port_t)-1);
+
+	rc = 0;
+	if (count == 0)
+		goto out;
+
+	if (count > PAGE_SIZE)
+		count = PAGE_SIZE;
+
+	rc = -EFAULT;
+	if (copy_from_user(kbuf, buf, count) != 0)
+		goto out;
+
+	spin_lock_irq(&port_user_lock);
+	for (i = 0; i < (count/sizeof(evtchn_port_t)); i++)
+		if ((kbuf[i] < NR_EVENT_CHANNELS) && (port_user[kbuf[i]] == u))
+			unmask_evtchn(kbuf[i]);
+	spin_unlock_irq(&port_user_lock);
+
+	rc = count;
+
+ out:
+	free_page((unsigned long)kbuf);
+	return rc;
+}
+
+static unsigned int next_bind_cpu(cpumask_t map)
+{
+	static unsigned int bind_cpu;
+	bind_cpu = next_cpu(bind_cpu, map);
+	if (bind_cpu >= NR_CPUS)
+		bind_cpu = first_cpu(map);
+	return bind_cpu;
+}
+
+static void evtchn_bind_to_user(struct per_user_data *u, int port)
+{
+	spin_lock_irq(&port_user_lock);
+
+	BUG_ON(port_user[port] != NULL);
+	port_user[port] = u;
+
+	if (u->bind_cpu == -1)
+		u->bind_cpu = next_bind_cpu(cpu_online_map);
+
+	rebind_evtchn_to_cpu(port, u->bind_cpu);
+
+	unmask_evtchn(port);
+
+	spin_unlock_irq(&port_user_lock);
+}
+
+static long evtchn_ioctl(struct file *file,
+			 unsigned int cmd, unsigned long arg)
+{
+	int rc;
+	struct per_user_data *u = file->private_data;
+	void __user *uarg = (void __user *) arg;
+
+	switch (cmd) {
+	case IOCTL_EVTCHN_BIND_VIRQ: {
+		struct ioctl_evtchn_bind_virq bind;
+		struct evtchn_bind_virq bind_virq;
+
+		rc = -EFAULT;
+		if (copy_from_user(&bind, uarg, sizeof(bind)))
+			break;
+
+		bind_virq.virq = bind.virq;
+		bind_virq.vcpu = 0;
+		rc = HYPERVISOR_event_channel_op(EVTCHNOP_bind_virq,
+						 &bind_virq);
+		if (rc != 0)
+			break;
+
+		rc = bind_virq.port;
+		evtchn_bind_to_user(u, rc);
+		break;
+	}
+
+	case IOCTL_EVTCHN_BIND_INTERDOMAIN: {
+		struct ioctl_evtchn_bind_interdomain bind;
+		struct evtchn_bind_interdomain bind_interdomain;
+
+		rc = -EFAULT;
+		if (copy_from_user(&bind, uarg, sizeof(bind)))
+			break;
+
+		bind_interdomain.remote_dom  = bind.remote_domain;
+		bind_interdomain.remote_port = bind.remote_port;
+		rc = HYPERVISOR_event_channel_op(EVTCHNOP_bind_interdomain,
+						 &bind_interdomain);
+		if (rc != 0)
+			break;
+
+		rc = bind_interdomain.local_port;
+		evtchn_bind_to_user(u, rc);
+		break;
+	}
+
+	case IOCTL_EVTCHN_BIND_UNBOUND_PORT: {
+		struct ioctl_evtchn_bind_unbound_port bind;
+		struct evtchn_alloc_unbound alloc_unbound;
+
+		rc = -EFAULT;
+		if (copy_from_user(&bind, uarg, sizeof(bind)))
+			break;
+
+		alloc_unbound.dom        = DOMID_SELF;
+		alloc_unbound.remote_dom = bind.remote_domain;
+		rc = HYPERVISOR_event_channel_op(EVTCHNOP_alloc_unbound,
+						 &alloc_unbound);
+		if (rc != 0)
+			break;
+
+		rc = alloc_unbound.port;
+		evtchn_bind_to_user(u, rc);
+		break;
+	}
+
+	case IOCTL_EVTCHN_UNBIND: {
+		struct ioctl_evtchn_unbind unbind;
+		struct evtchn_close close;
+		int ret;
+
+		rc = -EFAULT;
+		if (copy_from_user(&unbind, uarg, sizeof(unbind)))
+			break;
+
+		rc = -EINVAL;
+		if (unbind.port >= NR_EVENT_CHANNELS)
+			break;
+
+		spin_lock_irq(&port_user_lock);
+    
+		rc = -ENOTCONN;
+		if (port_user[unbind.port] != u) {
+			spin_unlock_irq(&port_user_lock);
+			break;
+		}
+
+		port_user[unbind.port] = NULL;
+		mask_evtchn(unbind.port);
+		rebind_evtchn_to_cpu(unbind.port, 0);
+
+		spin_unlock_irq(&port_user_lock);
+
+		close.port = unbind.port;
+		ret = HYPERVISOR_event_channel_op(EVTCHNOP_close, &close);
+		BUG_ON(ret);
+
+		rc = 0;
+		break;
+	}
+
+	case IOCTL_EVTCHN_NOTIFY: {
+		struct ioctl_evtchn_notify notify;
+
+		rc = -EFAULT;
+		if (copy_from_user(&notify, uarg, sizeof(notify)))
+			break;
+
+		if (notify.port >= NR_EVENT_CHANNELS) {
+			rc = -EINVAL;
+		} else if (port_user[notify.port] != u) {
+			rc = -ENOTCONN;
+		} else {
+			notify_remote_via_evtchn(notify.port);
+			rc = 0;
+		}
+		break;
+	}
+
+	case IOCTL_EVTCHN_RESET: {
+		/* Initialise the ring to empty. Clear errors. */
+		mutex_lock(&u->ring_cons_mutex);
+		spin_lock_irq(&port_user_lock);
+		u->ring_cons = u->ring_prod = u->ring_overflow = 0;
+		spin_unlock_irq(&port_user_lock);
+		mutex_unlock(&u->ring_cons_mutex);
+		rc = 0;
+		break;
+	}
+
+	default:
+		rc = -ENOSYS;
+		break;
+	}
+
+	return rc;
+}
+
+static unsigned int evtchn_poll(struct file *file, poll_table *wait)
+{
+	unsigned int mask = POLLOUT | POLLWRNORM;
+	struct per_user_data *u = file->private_data;
+
+	poll_wait(file, &u->evtchn_wait, wait);
+	if (u->ring_cons != u->ring_prod)
+		mask |= POLLIN | POLLRDNORM;
+	if (u->ring_overflow)
+		mask = POLLERR;
+	return mask;
+}
+
+static int evtchn_fasync(int fd, struct file *filp, int on)
+{
+	struct per_user_data *u = filp->private_data;
+	return fasync_helper(fd, filp, on, &u->evtchn_async_queue);
+}
+
+static int evtchn_open(struct inode *inode, struct file *filp)
+{
+	struct per_user_data *u;
+
+	if ((u = kmalloc(sizeof(*u), GFP_KERNEL)) == NULL)
+		return -ENOMEM;
+
+	memset(u, 0, sizeof(*u));
+	init_waitqueue_head(&u->evtchn_wait);
+
+	u->ring = (evtchn_port_t *)__get_free_page(GFP_KERNEL);
+	if (u->ring == NULL) {
+		kfree(u);
+		return -ENOMEM;
+	}
+
+	mutex_init(&u->ring_cons_mutex);
+
+	filp->private_data = u;
+
+	u->bind_cpu = -1;
+
+	return 0;
+}
+
+static int evtchn_release(struct inode *inode, struct file *filp)
+{
+	int i;
+	struct per_user_data *u = filp->private_data;
+	struct evtchn_close close;
+
+	spin_lock_irq(&port_user_lock);
+
+	free_page((unsigned long)u->ring);
+
+	for (i = 0; i < NR_EVENT_CHANNELS; i++) {
+		int ret;
+		if (port_user[i] != u)
+			continue;
+
+		port_user[i] = NULL;
+		mask_evtchn(i);
+		rebind_evtchn_to_cpu(i, 0);
+
+		close.port = i;
+		ret = HYPERVISOR_event_channel_op(EVTCHNOP_close, &close);
+		BUG_ON(ret);
+	}
+
+	spin_unlock_irq(&port_user_lock);
+
+	kfree(u);
+
+	return 0;
+}
+
+static const struct file_operations evtchn_fops = {
+	.owner   = THIS_MODULE,
+	.read    = evtchn_read,
+	.write   = evtchn_write,
+	.unlocked_ioctl = evtchn_ioctl,
+	.poll    = evtchn_poll,
+	.fasync  = evtchn_fasync,
+	.open    = evtchn_open,
+	.release = evtchn_release,
+};
+
+static struct miscdevice evtchn_miscdev = {
+	.minor        = MISC_DYNAMIC_MINOR,
+	.name         = "evtchn",
+	.fops         = &evtchn_fops,
+};
+
+static int __cpuinit evtchn_cpu_notify(struct notifier_block *nfb,
+			unsigned long action, void *hcpu)
+{
+	int hotcpu = (unsigned long)hcpu;
+	cpumask_t map = cpu_online_map;
+	int port, newcpu;
+	struct per_user_data *u;
+
+	switch (action) {
+	case CPU_DOWN_PREPARE:
+		cpu_clear(hotcpu, map);
+		spin_lock_irq(&port_user_lock);
+		for (port = 0; port < NR_EVENT_CHANNELS; port++) {
+			if ((u = port_user[port]) != NULL && 
+			    u->bind_cpu == hotcpu &&
+			    (newcpu = next_bind_cpu(map)) < NR_CPUS) {
+				rebind_evtchn_to_cpu(port, newcpu);
+				u->bind_cpu = newcpu;
+			}
+		}
+		spin_unlock_irq(&port_user_lock);
+		break;
+	default:
+		return NOTIFY_DONE;
+	}
+	return NOTIFY_OK;
+}
+
+static struct notifier_block __cpuinitdata evtchn_cpu_nfb = {
+	.notifier_call = evtchn_cpu_notify
+};
+
+static int __init evtchn_init(void)
+{
+	int err;
+
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	spin_lock_init(&port_user_lock);
+	memset(port_user, 0, sizeof(port_user));
+
+	/* Create '/dev/misc/evtchn'. */
+	err = misc_register(&evtchn_miscdev);
+	if (err != 0) {
+		printk(KERN_ALERT "Could not register /dev/misc/evtchn\n");
+		return err;
+	}
+
+	register_cpu_notifier(&evtchn_cpu_nfb);
+
+	printk("Event-channel device installed.\n");
+
+	return 0;
+}
+
+static void __exit evtchn_cleanup(void)
+{
+	misc_deregister(&evtchn_miscdev);
+	unregister_cpu_notifier(&evtchn_cpu_nfb);
+}
+
+module_init(evtchn_init);
+module_exit(evtchn_cleanup);
+
+MODULE_LICENSE("Dual BSD/GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/evtchn/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/evtchn/Makefile
--- linux-2.6.18.8/drivers/xen/evtchn/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/evtchn/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,2 @@
+
+obj-y	:= evtchn.o
diff -rpuN linux-2.6.18.8/drivers/xen/fbfront/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/fbfront/Makefile
--- linux-2.6.18.8/drivers/xen/fbfront/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/fbfront/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,2 @@
+obj-$(CONFIG_XEN_FRAMEBUFFER)	:= xenfb.o
+obj-$(CONFIG_XEN_KEYBOARD)	+= xenkbd.o
diff -rpuN linux-2.6.18.8/drivers/xen/fbfront/xenfb.c linux-2.6.18-xen-3.3.0/drivers/xen/fbfront/xenfb.c
--- linux-2.6.18.8/drivers/xen/fbfront/xenfb.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/fbfront/xenfb.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,887 @@
+/*
+ * linux/drivers/video/xenfb.c -- Xen para-virtual frame buffer device
+ *
+ * Copyright (C) 2005-2006 Anthony Liguori <aliguori@us.ibm.com>
+ * Copyright (C) 2006 Red Hat, Inc., Markus Armbruster <armbru@redhat.com>
+ *
+ *  Based on linux/drivers/video/q40fb.c
+ *
+ *  This file is subject to the terms and conditions of the GNU General Public
+ *  License. See the file COPYING in the main directory of this archive for
+ *  more details.
+ */
+
+/*
+ * TODO:
+ *
+ * Switch to grant tables when they become capable of dealing with the
+ * frame buffer.
+ */
+
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/fb.h>
+#include <linux/module.h>
+#include <linux/vmalloc.h>
+#include <linux/mm.h>
+#include <linux/mutex.h>
+#include <asm/hypervisor.h>
+#include <xen/evtchn.h>
+#include <xen/interface/io/fbif.h>
+#include <xen/interface/io/protocols.h>
+#include <xen/xenbus.h>
+#include <linux/kthread.h>
+
+struct xenfb_mapping
+{
+	struct list_head	link;
+	struct vm_area_struct	*vma;
+	atomic_t		map_refs;
+	int			faults;
+	struct xenfb_info	*info;
+};
+
+struct xenfb_info
+{
+	struct task_struct	*kthread;
+	wait_queue_head_t	wq;
+
+	unsigned char		*fb;
+	struct fb_info		*fb_info;
+	struct timer_list	refresh;
+	int			dirty;
+	int			x1, y1, x2, y2;	/* dirty rectangle,
+						   protected by dirty_lock */
+	spinlock_t		dirty_lock;
+	struct mutex		mm_lock;
+	int			nr_pages;
+	struct page		**pages;
+	struct list_head	mappings; /* protected by mm_lock */
+
+	int			irq;
+	struct xenfb_page	*page;
+	unsigned long 		*mfns;
+	int			update_wanted; /* XENFB_TYPE_UPDATE wanted */
+	int			feature_resize; /* Backend has resize feature */
+	struct xenfb_resize	resize;
+	int			resize_dpy;
+	spinlock_t		resize_lock;
+
+	struct xenbus_device	*xbdev;
+};
+
+/*
+ * There are three locks:
+ *    spinlock resize_lock protecting resize_dpy and resize
+ *    spinlock dirty_lock protecting the dirty rectangle
+ *    mutex mm_lock protecting mappings.
+ *
+ * How the dirty and mapping locks work together
+ *
+ * The problem is that dirty rectangle and mappings aren't
+ * independent: the dirty rectangle must cover all faulted pages in
+ * mappings.  We need to prove that our locking maintains this
+ * invariant.
+ *
+ * There are several kinds of critical regions:
+ *
+ * 1. Holding only dirty_lock: xenfb_refresh().  May run in
+ *    interrupts.  Extends the dirty rectangle.  Trivially preserves
+ *    invariant.
+ *
+ * 2. Holding only mm_lock: xenfb_mmap() and xenfb_vm_close().  Touch
+ *    only mappings.  The former creates unfaulted pages.  Preserves
+ *    invariant.  The latter removes pages.  Preserves invariant.
+ *
+ * 3. Holding both locks: xenfb_vm_nopage().  Extends the dirty
+ *    rectangle and updates mappings consistently.  Preserves
+ *    invariant.
+ *
+ * 4. The ugliest one: xenfb_update_screen().  Clear the dirty
+ *    rectangle and update mappings consistently.
+ *
+ *    We can't simply hold both locks, because zap_page_range() cannot
+ *    be called with a spinlock held.
+ *
+ *    Therefore, we first clear the dirty rectangle with both locks
+ *    held.  Then we unlock dirty_lock and update the mappings.
+ *    Critical regions that hold only dirty_lock may interfere with
+ *    that.  This can only be region 1: xenfb_refresh().  But that
+ *    just extends the dirty rectangle, which can't harm the
+ *    invariant.
+ *
+ * But FIXME: the invariant is too weak.  It misses that the fault
+ * record in mappings must be consistent with the mapping of pages in
+ * the associated address space!  do_no_page() updates the PTE after
+ * xenfb_vm_nopage() returns, i.e. outside the critical region.  This
+ * allows the following race:
+ *
+ * X writes to some address in the Xen frame buffer
+ * Fault - call do_no_page()
+ *     call xenfb_vm_nopage()
+ *         grab mm_lock
+ *         map->faults++;
+ *         release mm_lock
+ *     return back to do_no_page()
+ * (preempted, or SMP)
+ * Xen worker thread runs.
+ *      grab mm_lock
+ *      look at mappings
+ *          find this mapping, zaps its pages (but page not in pte yet)
+ *          clear map->faults
+ *      releases mm_lock
+ * (back to X process)
+ *     put page in X's pte
+ *
+ * Oh well, we wont be updating the writes to this page anytime soon.
+ */
+#define MB_ (1024*1024)
+#define XENFB_DEFAULT_FB_LEN (XENFB_WIDTH * XENFB_HEIGHT * XENFB_DEPTH / 8)
+
+enum {KPARAM_MEM, KPARAM_WIDTH, KPARAM_HEIGHT, KPARAM_CNT};
+static int video[KPARAM_CNT] = {2, XENFB_WIDTH, XENFB_HEIGHT};
+module_param_array(video, int, NULL, 0);
+MODULE_PARM_DESC(video,
+		"Size of video memory in MB and width,height in pixels, default = (2,800,600)");
+
+static int xenfb_fps = 20;
+
+static int xenfb_remove(struct xenbus_device *);
+static void xenfb_init_shared_page(struct xenfb_info *, struct fb_info *);
+static int xenfb_connect_backend(struct xenbus_device *, struct xenfb_info *);
+static void xenfb_disconnect_backend(struct xenfb_info *);
+
+static void xenfb_send_event(struct xenfb_info *info,
+		union xenfb_out_event *event)
+{
+	__u32 prod;
+
+	prod = info->page->out_prod;
+	/* caller ensures !xenfb_queue_full() */
+	mb();			/* ensure ring space available */
+	XENFB_OUT_RING_REF(info->page, prod) = *event;
+	wmb();			/* ensure ring contents visible */
+	info->page->out_prod = prod + 1;
+
+	notify_remote_via_irq(info->irq);
+}
+
+static void xenfb_do_update(struct xenfb_info *info,
+			    int x, int y, int w, int h)
+{
+	union xenfb_out_event event;
+
+	memset(&event, 0, sizeof(event));
+	event.type = XENFB_TYPE_UPDATE;
+	event.update.x = x;
+	event.update.y = y;
+	event.update.width = w;
+	event.update.height = h;
+
+	/* caller ensures !xenfb_queue_full() */
+	xenfb_send_event(info, &event);
+}
+
+static void xenfb_do_resize(struct xenfb_info *info)
+{
+	union xenfb_out_event event;
+
+	memset(&event, 0, sizeof(event));
+	event.resize = info->resize;
+
+	/* caller ensures !xenfb_queue_full() */
+	xenfb_send_event(info, &event);
+}
+
+static int xenfb_queue_full(struct xenfb_info *info)
+{
+	__u32 cons, prod;
+
+	prod = info->page->out_prod;
+	cons = info->page->out_cons;
+	return prod - cons == XENFB_OUT_RING_LEN;
+}
+
+static void xenfb_update_screen(struct xenfb_info *info)
+{
+	unsigned long flags;
+	int y1, y2, x1, x2;
+	struct xenfb_mapping *map;
+
+	if (!info->update_wanted)
+		return;
+	if (xenfb_queue_full(info))
+		return;
+
+	mutex_lock(&info->mm_lock);
+
+	spin_lock_irqsave(&info->dirty_lock, flags);
+	y1 = info->y1;
+	y2 = info->y2;
+	x1 = info->x1;
+	x2 = info->x2;
+	info->x1 = info->y1 = INT_MAX;
+	info->x2 = info->y2 = 0;
+	spin_unlock_irqrestore(&info->dirty_lock, flags);
+
+	list_for_each_entry(map, &info->mappings, link) {
+		if (!map->faults)
+			continue;
+		zap_page_range(map->vma, map->vma->vm_start,
+			       map->vma->vm_end - map->vma->vm_start, NULL);
+		map->faults = 0;
+	}
+
+	mutex_unlock(&info->mm_lock);
+
+	if (x2 < x1 || y2 < y1) {
+		printk("xenfb_update_screen bogus rect %d %d %d %d\n",
+		       x1, x2, y1, y2);
+		WARN_ON(1);
+	}
+	xenfb_do_update(info, x1, y1, x2 - x1, y2 - y1);
+}
+
+static void xenfb_handle_resize_dpy(struct xenfb_info *info)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&info->resize_lock, flags);
+	if (info->resize_dpy) {
+		if (!xenfb_queue_full(info)) {
+			info->resize_dpy = 0;
+			xenfb_do_resize(info);
+		}
+	}
+	spin_unlock_irqrestore(&info->resize_lock, flags);
+}
+
+static int xenfb_thread(void *data)
+{
+	struct xenfb_info *info = data;
+
+	while (!kthread_should_stop()) {
+		xenfb_handle_resize_dpy(info);
+		if (info->dirty) {
+			info->dirty = 0;
+			xenfb_update_screen(info);
+		}
+		wait_event_interruptible(info->wq,
+			kthread_should_stop() || info->dirty);
+		try_to_freeze();
+	}
+	return 0;
+}
+
+static int xenfb_setcolreg(unsigned regno, unsigned red, unsigned green,
+			   unsigned blue, unsigned transp,
+			   struct fb_info *info)
+{
+	u32 v;
+
+	if (regno > info->cmap.len)
+		return 1;
+
+	red   >>= (16 - info->var.red.length);
+	green >>= (16 - info->var.green.length);
+	blue  >>= (16 - info->var.blue.length);
+
+	v = (red << info->var.red.offset) |
+	    (green << info->var.green.offset) |
+	    (blue << info->var.blue.offset);
+
+	/* FIXME is this sane?  check against xxxfb_setcolreg()!  */
+	switch (info->var.bits_per_pixel) {
+	case 16:
+	case 24:
+	case 32:
+		((u32 *)info->pseudo_palette)[regno] = v;
+		break;
+	}
+	
+	return 0;
+}
+
+static void xenfb_timer(unsigned long data)
+{
+	struct xenfb_info *info = (struct xenfb_info *)data;
+	wake_up(&info->wq);
+}
+
+static void __xenfb_refresh(struct xenfb_info *info,
+			    int x1, int y1, int w, int h)
+{
+	int y2, x2;
+
+	y2 = y1 + h;
+	x2 = x1 + w;
+
+	if (info->y1 > y1)
+		info->y1 = y1;
+	if (info->y2 < y2)
+		info->y2 = y2;
+	if (info->x1 > x1)
+		info->x1 = x1;
+	if (info->x2 < x2)
+		info->x2 = x2;
+	info->dirty = 1;
+
+	if (timer_pending(&info->refresh))
+		return;
+
+	mod_timer(&info->refresh, jiffies + HZ/xenfb_fps);
+}
+
+static void xenfb_refresh(struct xenfb_info *info,
+			  int x1, int y1, int w, int h)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&info->dirty_lock, flags);
+	__xenfb_refresh(info, x1, y1, w, h);
+	spin_unlock_irqrestore(&info->dirty_lock, flags);
+}
+
+static void xenfb_fillrect(struct fb_info *p, const struct fb_fillrect *rect)
+{
+	struct xenfb_info *info = p->par;
+
+	cfb_fillrect(p, rect);
+	xenfb_refresh(info, rect->dx, rect->dy, rect->width, rect->height);
+}
+
+static void xenfb_imageblit(struct fb_info *p, const struct fb_image *image)
+{
+	struct xenfb_info *info = p->par;
+
+	cfb_imageblit(p, image);
+	xenfb_refresh(info, image->dx, image->dy, image->width, image->height);
+}
+
+static void xenfb_copyarea(struct fb_info *p, const struct fb_copyarea *area)
+{
+	struct xenfb_info *info = p->par;
+
+	cfb_copyarea(p, area);
+	xenfb_refresh(info, area->dx, area->dy, area->width, area->height);
+}
+
+static void xenfb_vm_open(struct vm_area_struct *vma)
+{
+	struct xenfb_mapping *map = vma->vm_private_data;
+	atomic_inc(&map->map_refs);
+}
+
+static void xenfb_vm_close(struct vm_area_struct *vma)
+{
+	struct xenfb_mapping *map = vma->vm_private_data;
+	struct xenfb_info *info = map->info;
+
+	mutex_lock(&info->mm_lock);
+	if (atomic_dec_and_test(&map->map_refs)) {
+		list_del(&map->link);
+		kfree(map);
+	}
+	mutex_unlock(&info->mm_lock);
+}
+
+static struct page *xenfb_vm_nopage(struct vm_area_struct *vma,
+				    unsigned long vaddr, int *type)
+{
+	struct xenfb_mapping *map = vma->vm_private_data;
+	struct xenfb_info *info = map->info;
+	int pgnr = (vaddr - vma->vm_start) >> PAGE_SHIFT;
+	unsigned long flags;
+	struct page *page;
+	int y1, y2;
+
+	if (pgnr >= info->nr_pages)
+		return NOPAGE_SIGBUS;
+
+	mutex_lock(&info->mm_lock);
+	spin_lock_irqsave(&info->dirty_lock, flags);
+	page = info->pages[pgnr];
+	get_page(page);
+	map->faults++;
+
+	y1 = pgnr * PAGE_SIZE / info->fb_info->fix.line_length;
+	y2 = (pgnr * PAGE_SIZE + PAGE_SIZE - 1) / info->fb_info->fix.line_length;
+	if (y2 > info->fb_info->var.yres)
+		y2 = info->fb_info->var.yres;
+	__xenfb_refresh(info, 0, y1, info->fb_info->var.xres, y2 - y1);
+	spin_unlock_irqrestore(&info->dirty_lock, flags);
+	mutex_unlock(&info->mm_lock);
+
+	if (type)
+		*type = VM_FAULT_MINOR;
+
+	return page;
+}
+
+static struct vm_operations_struct xenfb_vm_ops = {
+	.open	= xenfb_vm_open,
+	.close	= xenfb_vm_close,
+	.nopage	= xenfb_vm_nopage,
+};
+
+static int xenfb_mmap(struct fb_info *fb_info, struct vm_area_struct *vma)
+{
+	struct xenfb_info *info = fb_info->par;
+	struct xenfb_mapping *map;
+	int map_pages;
+
+	if (!(vma->vm_flags & VM_WRITE))
+		return -EINVAL;
+	if (!(vma->vm_flags & VM_SHARED))
+		return -EINVAL;
+	if (vma->vm_pgoff != 0)
+		return -EINVAL;
+
+	map_pages = (vma->vm_end - vma->vm_start + PAGE_SIZE-1) >> PAGE_SHIFT;
+	if (map_pages > info->nr_pages)
+		return -EINVAL;
+
+	map = kzalloc(sizeof(*map), GFP_KERNEL);
+	if (map == NULL)
+		return -ENOMEM;
+
+	map->vma = vma;
+	map->faults = 0;
+	map->info = info;
+	atomic_set(&map->map_refs, 1);
+
+	mutex_lock(&info->mm_lock);
+	list_add(&map->link, &info->mappings);
+	mutex_unlock(&info->mm_lock);
+
+	vma->vm_ops = &xenfb_vm_ops;
+	vma->vm_flags |= (VM_DONTEXPAND | VM_RESERVED);
+	vma->vm_private_data = map;
+
+	return 0;
+}
+
+static int
+xenfb_check_var(struct fb_var_screeninfo *var, struct fb_info *info)
+{
+	struct xenfb_info *xenfb_info;
+	int required_mem_len;
+
+	xenfb_info = info->par;
+
+	if (!xenfb_info->feature_resize) {
+		if (var->xres == video[KPARAM_WIDTH] &&
+			var->yres == video[KPARAM_HEIGHT] &&
+			var->bits_per_pixel == xenfb_info->page->depth) {
+			return 0;
+		}
+		return -EINVAL;
+	}
+
+	/* Can't resize past initial width and height */
+	if (var->xres > video[KPARAM_WIDTH] || var->yres > video[KPARAM_HEIGHT])
+		return -EINVAL;
+
+	required_mem_len = var->xres * var->yres * (xenfb_info->page->depth / 8);
+	if (var->bits_per_pixel == xenfb_info->page->depth &&
+		var->xres <= info->fix.line_length / (XENFB_DEPTH / 8) &&
+		required_mem_len <= info->fix.smem_len) {
+		var->xres_virtual = var->xres;
+		var->yres_virtual = var->yres;
+		return 0;
+	}
+	return -EINVAL;
+}
+
+static int xenfb_set_par(struct fb_info *info)
+{
+	struct xenfb_info *xenfb_info;
+	unsigned long flags;
+
+	xenfb_info = info->par;
+
+	spin_lock_irqsave(&xenfb_info->resize_lock, flags);
+	xenfb_info->resize.type = XENFB_TYPE_RESIZE;
+	xenfb_info->resize.width = info->var.xres;
+	xenfb_info->resize.height = info->var.yres;
+	xenfb_info->resize.stride = info->fix.line_length;
+	xenfb_info->resize.depth = info->var.bits_per_pixel;
+	xenfb_info->resize.offset = 0;
+	xenfb_info->resize_dpy = 1;
+	spin_unlock_irqrestore(&xenfb_info->resize_lock, flags);
+	return 0;
+}
+
+static struct fb_ops xenfb_fb_ops = {
+	.owner		= THIS_MODULE,
+	.fb_setcolreg	= xenfb_setcolreg,
+	.fb_fillrect	= xenfb_fillrect,
+	.fb_copyarea	= xenfb_copyarea,
+	.fb_imageblit	= xenfb_imageblit,
+	.fb_mmap	= xenfb_mmap,
+	.fb_check_var	= xenfb_check_var,
+	.fb_set_par     = xenfb_set_par,
+};
+
+static irqreturn_t xenfb_event_handler(int rq, void *dev_id,
+				       struct pt_regs *regs)
+{
+	/*
+	 * No in events recognized, simply ignore them all.
+	 * If you need to recognize some, see xenbkd's input_handler()
+	 * for how to do that.
+	 */
+	struct xenfb_info *info = dev_id;
+	struct xenfb_page *page = info->page;
+
+	if (page->in_cons != page->in_prod) {
+		info->page->in_cons = info->page->in_prod;
+		notify_remote_via_irq(info->irq);
+	}
+	return IRQ_HANDLED;
+}
+
+static unsigned long vmalloc_to_mfn(void *address)
+{
+	return pfn_to_mfn(vmalloc_to_pfn(address));
+}
+
+static int __devinit xenfb_probe(struct xenbus_device *dev,
+				 const struct xenbus_device_id *id)
+{
+	struct xenfb_info *info;
+	struct fb_info *fb_info;
+	int fb_size;
+	int val;
+	int ret;
+
+	info = kzalloc(sizeof(*info), GFP_KERNEL);
+	if (info == NULL) {
+		xenbus_dev_fatal(dev, -ENOMEM, "allocating info structure");
+		return -ENOMEM;
+	}
+
+	/* Limit kernel param videoram amount to what is in xenstore */
+	if (xenbus_scanf(XBT_NIL, dev->otherend, "videoram", "%d", &val) == 1) {
+		if (val < video[KPARAM_MEM])
+			video[KPARAM_MEM] = val;
+	}
+
+	/* If requested res does not fit in available memory, use default */
+	fb_size = video[KPARAM_MEM] * MB_;
+	if (video[KPARAM_WIDTH] * video[KPARAM_HEIGHT] * XENFB_DEPTH/8 > fb_size) {
+		video[KPARAM_WIDTH] = XENFB_WIDTH;
+		video[KPARAM_HEIGHT] = XENFB_HEIGHT;
+		fb_size = XENFB_DEFAULT_FB_LEN;
+	}
+
+	dev->dev.driver_data = info;
+	info->xbdev = dev;
+	info->irq = -1;
+	info->x1 = info->y1 = INT_MAX;
+	spin_lock_init(&info->dirty_lock);
+	spin_lock_init(&info->resize_lock);
+	mutex_init(&info->mm_lock);
+	init_waitqueue_head(&info->wq);
+	init_timer(&info->refresh);
+	info->refresh.function = xenfb_timer;
+	info->refresh.data = (unsigned long)info;
+	INIT_LIST_HEAD(&info->mappings);
+
+	info->fb = vmalloc(fb_size);
+	if (info->fb == NULL)
+		goto error_nomem;
+	memset(info->fb, 0, fb_size);
+
+	info->nr_pages = (fb_size + PAGE_SIZE - 1) >> PAGE_SHIFT;
+
+	info->pages = kmalloc(sizeof(struct page *) * info->nr_pages,
+			      GFP_KERNEL);
+	if (info->pages == NULL)
+		goto error_nomem;
+
+	info->mfns = vmalloc(sizeof(unsigned long) * info->nr_pages);
+	if (!info->mfns)
+		goto error_nomem;
+
+	/* set up shared page */
+	info->page = (void *)__get_free_page(GFP_KERNEL | __GFP_ZERO);
+	if (!info->page)
+		goto error_nomem;
+
+	fb_info = framebuffer_alloc(sizeof(u32) * 256, NULL);
+				/* see fishy hackery below */
+	if (fb_info == NULL)
+		goto error_nomem;
+
+	/* FIXME fishy hackery */
+	fb_info->pseudo_palette = fb_info->par;
+	fb_info->par = info;
+	/* /FIXME */
+	fb_info->screen_base = info->fb;
+
+	fb_info->fbops = &xenfb_fb_ops;
+	fb_info->var.xres_virtual = fb_info->var.xres = video[KPARAM_WIDTH];
+	fb_info->var.yres_virtual = fb_info->var.yres = video[KPARAM_HEIGHT];
+	fb_info->var.bits_per_pixel = XENFB_DEPTH;
+
+	fb_info->var.red = (struct fb_bitfield){16, 8, 0};
+	fb_info->var.green = (struct fb_bitfield){8, 8, 0};
+	fb_info->var.blue = (struct fb_bitfield){0, 8, 0};
+
+	fb_info->var.activate = FB_ACTIVATE_NOW;
+	fb_info->var.height = -1;
+	fb_info->var.width = -1;
+	fb_info->var.vmode = FB_VMODE_NONINTERLACED;
+
+	fb_info->fix.visual = FB_VISUAL_TRUECOLOR;
+	fb_info->fix.line_length = fb_info->var.xres * (XENFB_DEPTH / 8);
+	fb_info->fix.smem_start = 0;
+	fb_info->fix.smem_len = fb_size;
+	strcpy(fb_info->fix.id, "xen");
+	fb_info->fix.type = FB_TYPE_PACKED_PIXELS;
+	fb_info->fix.accel = FB_ACCEL_NONE;
+
+	fb_info->flags = FBINFO_FLAG_DEFAULT;
+
+	ret = fb_alloc_cmap(&fb_info->cmap, 256, 0);
+	if (ret < 0) {
+		framebuffer_release(fb_info);
+		xenbus_dev_fatal(dev, ret, "fb_alloc_cmap");
+		goto error;
+	}
+
+	xenfb_init_shared_page(info, fb_info);
+
+	ret = register_framebuffer(fb_info);
+	if (ret) {
+		fb_dealloc_cmap(&info->fb_info->cmap);
+		framebuffer_release(fb_info);
+		xenbus_dev_fatal(dev, ret, "register_framebuffer");
+		goto error;
+	}
+	info->fb_info = fb_info;
+
+	/* FIXME should this be delayed until backend XenbusStateConnected? */
+	info->kthread = kthread_run(xenfb_thread, info, "xenfb thread");
+	if (IS_ERR(info->kthread)) {
+		ret = PTR_ERR(info->kthread);
+		info->kthread = NULL;
+		xenbus_dev_fatal(dev, ret, "register_framebuffer");
+		goto error;
+	}
+
+	ret = xenfb_connect_backend(dev, info);
+	if (ret < 0)
+		goto error;
+
+	return 0;
+
+ error_nomem:
+	ret = -ENOMEM;
+	xenbus_dev_fatal(dev, ret, "allocating device memory");
+ error:
+	xenfb_remove(dev);
+	return ret;
+}
+
+static int xenfb_resume(struct xenbus_device *dev)
+{
+	struct xenfb_info *info = dev->dev.driver_data;
+
+	xenfb_disconnect_backend(info);
+	xenfb_init_shared_page(info, info->fb_info);
+	return xenfb_connect_backend(dev, info);
+}
+
+static int xenfb_remove(struct xenbus_device *dev)
+{
+	struct xenfb_info *info = dev->dev.driver_data;
+
+	del_timer(&info->refresh);
+	if (info->kthread)
+		kthread_stop(info->kthread);
+	xenfb_disconnect_backend(info);
+	if (info->fb_info) {
+		unregister_framebuffer(info->fb_info);
+		fb_dealloc_cmap(&info->fb_info->cmap);
+		framebuffer_release(info->fb_info);
+	}
+	free_page((unsigned long)info->page);
+	vfree(info->mfns);
+	kfree(info->pages);
+	vfree(info->fb);
+	kfree(info);
+
+	return 0;
+}
+
+static void xenfb_init_shared_page(struct xenfb_info *info,
+                                   struct fb_info * fb_info)
+{
+	int i;
+	int epd = PAGE_SIZE / sizeof(info->mfns[0]);
+
+	for (i = 0; i < info->nr_pages; i++)
+		info->pages[i] = vmalloc_to_page(info->fb + i * PAGE_SIZE);
+
+	for (i = 0; i < info->nr_pages; i++)
+		info->mfns[i] = vmalloc_to_mfn(info->fb + i * PAGE_SIZE);
+
+	for (i = 0; i * epd < info->nr_pages; i++)
+		info->page->pd[i] = vmalloc_to_mfn(&info->mfns[i * epd]);
+
+	info->page->width = fb_info->var.xres;
+	info->page->height = fb_info->var.yres;
+	info->page->depth = fb_info->var.bits_per_pixel;
+	info->page->line_length = fb_info->fix.line_length;
+	info->page->mem_length = fb_info->fix.smem_len;
+	info->page->in_cons = info->page->in_prod = 0;
+	info->page->out_cons = info->page->out_prod = 0;
+}
+
+static int xenfb_connect_backend(struct xenbus_device *dev,
+				 struct xenfb_info *info)
+{
+	int ret;
+	struct xenbus_transaction xbt;
+
+	ret = bind_listening_port_to_irqhandler(
+		dev->otherend_id, xenfb_event_handler, 0, "xenfb", info);
+	if (ret < 0) {
+		xenbus_dev_fatal(dev, ret,
+				 "bind_listening_port_to_irqhandler");
+		return ret;
+	}
+	info->irq = ret;
+
+ again:
+	ret = xenbus_transaction_start(&xbt);
+	if (ret) {
+		xenbus_dev_fatal(dev, ret, "starting transaction");
+		return ret;
+	}
+	ret = xenbus_printf(xbt, dev->nodename, "page-ref", "%lu",
+			    virt_to_mfn(info->page));
+	if (ret)
+		goto error_xenbus;
+	ret = xenbus_printf(xbt, dev->nodename, "event-channel", "%u",
+			    irq_to_evtchn_port(info->irq));
+	if (ret)
+		goto error_xenbus;
+	ret = xenbus_printf(xbt, dev->nodename, "protocol", "%s",
+			    XEN_IO_PROTO_ABI_NATIVE);
+	if (ret)
+		goto error_xenbus;
+	ret = xenbus_printf(xbt, dev->nodename, "feature-update", "1");
+	if (ret)
+		goto error_xenbus;
+	ret = xenbus_transaction_end(xbt, 0);
+	if (ret) {
+		if (ret == -EAGAIN)
+			goto again;
+		xenbus_dev_fatal(dev, ret, "completing transaction");
+		return ret;
+	}
+
+	xenbus_switch_state(dev, XenbusStateInitialised);
+	return 0;
+
+ error_xenbus:
+	xenbus_transaction_end(xbt, 1);
+	xenbus_dev_fatal(dev, ret, "writing xenstore");
+	return ret;
+}
+
+static void xenfb_disconnect_backend(struct xenfb_info *info)
+{
+	if (info->irq >= 0)
+		unbind_from_irqhandler(info->irq, info);
+	info->irq = -1;
+}
+
+static void xenfb_backend_changed(struct xenbus_device *dev,
+				  enum xenbus_state backend_state)
+{
+	struct xenfb_info *info = dev->dev.driver_data;
+	int val;
+
+	switch (backend_state) {
+	case XenbusStateInitialising:
+	case XenbusStateInitialised:
+	case XenbusStateReconfiguring:
+	case XenbusStateReconfigured:
+	case XenbusStateUnknown:
+	case XenbusStateClosed:
+		break;
+
+	case XenbusStateInitWait:
+	InitWait:
+		xenbus_switch_state(dev, XenbusStateConnected);
+		break;
+
+	case XenbusStateConnected:
+		/*
+		 * Work around xenbus race condition: If backend goes
+		 * through InitWait to Connected fast enough, we can
+		 * get Connected twice here.
+		 */
+		if (dev->state != XenbusStateConnected)
+			goto InitWait; /* no InitWait seen yet, fudge it */
+
+		if (xenbus_scanf(XBT_NIL, info->xbdev->otherend,
+				 "request-update", "%d", &val) < 0)
+			val = 0;
+		if (val)
+			info->update_wanted = 1;
+
+		if (xenbus_scanf(XBT_NIL, dev->otherend,
+					"feature-resize", "%d", &val) < 0)
+			val = 0;
+		info->feature_resize = val;
+		break;
+
+	case XenbusStateClosing:
+		// FIXME is this safe in any dev->state?
+		xenbus_frontend_closed(dev);
+		break;
+	}
+}
+
+static const struct xenbus_device_id xenfb_ids[] = {
+	{ "vfb" },
+	{ "" }
+};
+MODULE_ALIAS("xen:vfb");
+
+static struct xenbus_driver xenfb_driver = {
+	.name = "vfb",
+	.owner = THIS_MODULE,
+	.ids = xenfb_ids,
+	.probe = xenfb_probe,
+	.remove = xenfb_remove,
+	.resume = xenfb_resume,
+	.otherend_changed = xenfb_backend_changed,
+};
+
+static int __init xenfb_init(void)
+{
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	/* Nothing to do if running in dom0. */
+	if (is_initial_xendomain())
+		return -ENODEV;
+
+	return xenbus_register_frontend(&xenfb_driver);
+}
+
+static void __exit xenfb_cleanup(void)
+{
+	return xenbus_unregister_driver(&xenfb_driver);
+}
+
+module_init(xenfb_init);
+module_exit(xenfb_cleanup);
+
+MODULE_LICENSE("GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/fbfront/xenkbd.c linux-2.6.18-xen-3.3.0/drivers/xen/fbfront/xenkbd.c
--- linux-2.6.18.8/drivers/xen/fbfront/xenkbd.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/fbfront/xenkbd.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,354 @@
+/*
+ * linux/drivers/input/keyboard/xenkbd.c -- Xen para-virtual input device
+ *
+ * Copyright (C) 2005 Anthony Liguori <aliguori@us.ibm.com>
+ * Copyright (C) 2006 Red Hat, Inc., Markus Armbruster <armbru@redhat.com>
+ *
+ *  Based on linux/drivers/input/mouse/sermouse.c
+ *
+ *  This file is subject to the terms and conditions of the GNU General Public
+ *  License. See the file COPYING in the main directory of this archive for
+ *  more details.
+ */
+
+/*
+ * TODO:
+ *
+ * Switch to grant tables together with xenfb.c.
+ */
+
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/module.h>
+#include <linux/input.h>
+#include <asm/hypervisor.h>
+#include <xen/evtchn.h>
+#include <xen/interface/io/fbif.h>
+#include <xen/interface/io/kbdif.h>
+#include <xen/xenbus.h>
+
+struct xenkbd_info
+{
+	struct input_dev *kbd;
+	struct input_dev *ptr;
+	struct xenkbd_page *page;
+	int irq;
+	struct xenbus_device *xbdev;
+	char phys[32];
+};
+
+static int xenkbd_remove(struct xenbus_device *);
+static int xenkbd_connect_backend(struct xenbus_device *, struct xenkbd_info *);
+static void xenkbd_disconnect_backend(struct xenkbd_info *);
+
+/*
+ * Note: if you need to send out events, see xenfb_do_update() for how
+ * to do that.
+ */
+
+static irqreturn_t input_handler(int rq, void *dev_id, struct pt_regs *regs)
+{
+	struct xenkbd_info *info = dev_id;
+	struct xenkbd_page *page = info->page;
+	__u32 cons, prod;
+
+	prod = page->in_prod;
+	if (prod == page->in_cons)
+		return IRQ_HANDLED;
+	rmb();			/* ensure we see ring contents up to prod */
+	for (cons = page->in_cons; cons != prod; cons++) {
+		union xenkbd_in_event *event;
+		struct input_dev *dev;
+		event = &XENKBD_IN_RING_REF(page, cons);
+
+		dev = info->ptr;
+		switch (event->type) {
+		case XENKBD_TYPE_MOTION:
+			if (event->motion.rel_z)
+				input_report_rel(dev, REL_WHEEL,
+						 -event->motion.rel_z);
+			input_report_rel(dev, REL_X, event->motion.rel_x);
+			input_report_rel(dev, REL_Y, event->motion.rel_y);
+			break;
+		case XENKBD_TYPE_KEY:
+			dev = NULL;
+			if (test_bit(event->key.keycode, info->kbd->keybit))
+				dev = info->kbd;
+			if (test_bit(event->key.keycode, info->ptr->keybit))
+				dev = info->ptr;
+			if (dev)
+				input_report_key(dev, event->key.keycode,
+						 event->key.pressed);
+			else
+				printk("xenkbd: unhandled keycode 0x%x\n",
+				       event->key.keycode);
+			break;
+		case XENKBD_TYPE_POS:
+			if (event->pos.rel_z)
+				input_report_rel(dev, REL_WHEEL,
+						 -event->pos.rel_z);
+			input_report_abs(dev, ABS_X, event->pos.abs_x);
+			input_report_abs(dev, ABS_Y, event->pos.abs_y);
+			break;
+		}
+		if (dev)
+			input_sync(dev);
+	}
+	mb();			/* ensure we got ring contents */
+	page->in_cons = cons;
+	notify_remote_via_irq(info->irq);
+
+	return IRQ_HANDLED;
+}
+
+int __devinit xenkbd_probe(struct xenbus_device *dev,
+			   const struct xenbus_device_id *id)
+{
+	int ret, i;
+	struct xenkbd_info *info;
+	struct input_dev *kbd, *ptr;
+
+	info = kzalloc(sizeof(*info), GFP_KERNEL);
+	if (!info) {
+		xenbus_dev_fatal(dev, -ENOMEM, "allocating info structure");
+		return -ENOMEM;
+	}
+	dev->dev.driver_data = info;
+	info->xbdev = dev;
+	snprintf(info->phys, sizeof(info->phys), "xenbus/%s", dev->nodename);
+
+	info->page = (void *)__get_free_page(GFP_KERNEL);
+	if (!info->page)
+		goto error_nomem;
+	info->page->in_cons = info->page->in_prod = 0;
+	info->page->out_cons = info->page->out_prod = 0;
+
+	/* keyboard */
+	kbd = input_allocate_device();
+	if (!kbd)
+		goto error_nomem;
+	kbd->name = "Xen Virtual Keyboard";
+	kbd->phys = info->phys;
+	kbd->id.bustype = BUS_PCI;
+	kbd->id.vendor = 0x5853;
+	kbd->id.product = 0xffff;
+	kbd->evbit[0] = BIT(EV_KEY);
+	for (i = KEY_ESC; i < KEY_UNKNOWN; i++)
+		set_bit(i, kbd->keybit);
+	for (i = KEY_OK; i < KEY_MAX; i++)
+		set_bit(i, kbd->keybit);
+
+	ret = input_register_device(kbd);
+	if (ret) {
+		input_free_device(kbd);
+		xenbus_dev_fatal(dev, ret, "input_register_device(kbd)");
+		goto error;
+	}
+	info->kbd = kbd;
+
+	/* pointing device */
+	ptr = input_allocate_device();
+	if (!ptr)
+		goto error_nomem;
+	ptr->name = "Xen Virtual Pointer";
+	ptr->phys = info->phys;
+	ptr->id.bustype = BUS_PCI;
+	ptr->id.vendor = 0x5853;
+	ptr->id.product = 0xfffe;
+	ptr->evbit[0] = BIT(EV_KEY) | BIT(EV_REL) | BIT(EV_ABS);
+	for (i = BTN_LEFT; i <= BTN_TASK; i++)
+		set_bit(i, ptr->keybit);
+	ptr->relbit[0] = BIT(REL_X) | BIT(REL_Y) | BIT(REL_WHEEL);
+	input_set_abs_params(ptr, ABS_X, 0, XENFB_WIDTH, 0, 0);
+	input_set_abs_params(ptr, ABS_Y, 0, XENFB_HEIGHT, 0, 0);
+
+	ret = input_register_device(ptr);
+	if (ret) {
+		input_free_device(ptr);
+		xenbus_dev_fatal(dev, ret, "input_register_device(ptr)");
+		goto error;
+	}
+	info->ptr = ptr;
+
+	ret = xenkbd_connect_backend(dev, info);
+	if (ret < 0)
+		goto error;
+
+	return 0;
+
+ error_nomem:
+	ret = -ENOMEM;
+	xenbus_dev_fatal(dev, ret, "allocating device memory");
+ error:
+	xenkbd_remove(dev);
+	return ret;
+}
+
+static int xenkbd_resume(struct xenbus_device *dev)
+{
+	struct xenkbd_info *info = dev->dev.driver_data;
+
+	xenkbd_disconnect_backend(info);
+	info->page->in_cons = info->page->in_prod = 0;
+	info->page->out_cons = info->page->out_prod = 0;
+	return xenkbd_connect_backend(dev, info);
+}
+
+static int xenkbd_remove(struct xenbus_device *dev)
+{
+	struct xenkbd_info *info = dev->dev.driver_data;
+
+	xenkbd_disconnect_backend(info);
+	input_unregister_device(info->kbd);
+	input_unregister_device(info->ptr);
+	free_page((unsigned long)info->page);
+	kfree(info);
+	return 0;
+}
+
+static int xenkbd_connect_backend(struct xenbus_device *dev,
+				  struct xenkbd_info *info)
+{
+	int ret;
+	struct xenbus_transaction xbt;
+
+	ret = bind_listening_port_to_irqhandler(
+		dev->otherend_id, input_handler, 0, "xenkbd", info);
+	if (ret < 0) {
+		xenbus_dev_fatal(dev, ret,
+				 "bind_listening_port_to_irqhandler");
+		return ret;
+	}
+	info->irq = ret;
+
+ again:
+	ret = xenbus_transaction_start(&xbt);
+	if (ret) {
+		xenbus_dev_fatal(dev, ret, "starting transaction");
+		return ret;
+	}
+	ret = xenbus_printf(xbt, dev->nodename, "page-ref", "%lu",
+			    virt_to_mfn(info->page));
+	if (ret)
+		goto error_xenbus;
+	ret = xenbus_printf(xbt, dev->nodename, "event-channel", "%u",
+			    irq_to_evtchn_port(info->irq));
+	if (ret)
+		goto error_xenbus;
+	ret = xenbus_transaction_end(xbt, 0);
+	if (ret) {
+		if (ret == -EAGAIN)
+			goto again;
+		xenbus_dev_fatal(dev, ret, "completing transaction");
+		return ret;
+	}
+
+	xenbus_switch_state(dev, XenbusStateInitialised);
+	return 0;
+
+ error_xenbus:
+	xenbus_transaction_end(xbt, 1);
+	xenbus_dev_fatal(dev, ret, "writing xenstore");
+	return ret;
+}
+
+static void xenkbd_disconnect_backend(struct xenkbd_info *info)
+{
+	if (info->irq >= 0)
+		unbind_from_irqhandler(info->irq, info);
+	info->irq = -1;
+}
+
+static void xenkbd_backend_changed(struct xenbus_device *dev,
+				   enum xenbus_state backend_state)
+{
+	struct xenkbd_info *info = dev->dev.driver_data;
+	int ret, val;
+
+	switch (backend_state) {
+	case XenbusStateInitialising:
+	case XenbusStateInitialised:
+	case XenbusStateReconfiguring:
+	case XenbusStateReconfigured:
+	case XenbusStateUnknown:
+	case XenbusStateClosed:
+		break;
+
+	case XenbusStateInitWait:
+	InitWait:
+		ret = xenbus_scanf(XBT_NIL, info->xbdev->otherend,
+				   "feature-abs-pointer", "%d", &val);
+		if (ret < 0)
+			val = 0;
+		if (val) {
+			ret = xenbus_printf(XBT_NIL, info->xbdev->nodename,
+					    "request-abs-pointer", "1");
+			if (ret)
+				; /* FIXME */
+		}
+		xenbus_switch_state(dev, XenbusStateConnected);
+		break;
+
+	case XenbusStateConnected:
+		/*
+		 * Work around xenbus race condition: If backend goes
+		 * through InitWait to Connected fast enough, we can
+		 * get Connected twice here.
+		 */
+		if (dev->state != XenbusStateConnected)
+			goto InitWait; /* no InitWait seen yet, fudge it */
+
+		/* Set input abs params to match backend screen res */
+		if (xenbus_scanf(XBT_NIL, info->xbdev->otherend,
+				   "width", "%d", &val) > 0 )
+			input_set_abs_params(info->ptr, ABS_X, 0, val, 0, 0);
+
+		if (xenbus_scanf(XBT_NIL, info->xbdev->otherend,
+				   "height", "%d", &val) > 0 )
+			input_set_abs_params(info->ptr, ABS_Y, 0, val, 0, 0);
+
+		break;
+
+	case XenbusStateClosing:
+		xenbus_frontend_closed(dev);
+		break;
+	}
+}
+
+static const struct xenbus_device_id xenkbd_ids[] = {
+	{ "vkbd" },
+	{ "" }
+};
+MODULE_ALIAS("xen:vkbd");
+
+static struct xenbus_driver xenkbd_driver = {
+	.name = "vkbd",
+	.owner = THIS_MODULE,
+	.ids = xenkbd_ids,
+	.probe = xenkbd_probe,
+	.remove = xenkbd_remove,
+	.resume = xenkbd_resume,
+	.otherend_changed = xenkbd_backend_changed,
+};
+
+static int __init xenkbd_init(void)
+{
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	/* Nothing to do if running in dom0. */
+	if (is_initial_xendomain())
+		return -ENODEV;
+
+	return xenbus_register_frontend(&xenkbd_driver);
+}
+
+static void __exit xenkbd_cleanup(void)
+{
+	return xenbus_unregister_driver(&xenkbd_driver);
+}
+
+module_init(xenkbd_init);
+module_exit(xenkbd_cleanup);
+
+MODULE_LICENSE("GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/gntdev/gntdev.c linux-2.6.18-xen-3.3.0/drivers/xen/gntdev/gntdev.c
--- linux-2.6.18.8/drivers/xen/gntdev/gntdev.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/gntdev/gntdev.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1074 @@
+/******************************************************************************
+ * gntdev.c
+ * 
+ * Device for accessing (in user-space) pages that have been granted by other
+ * domains.
+ *
+ * Copyright (c) 2006-2007, D G Murray.
+ * 
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ * 
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#include <asm/atomic.h>
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/fs.h>
+#include <linux/device.h>
+#include <linux/mm.h>
+#include <linux/mman.h>
+#include <asm/uaccess.h>
+#include <asm/io.h>
+#include <xen/gnttab.h>
+#include <asm/hypervisor.h>
+#include <xen/balloon.h>
+#include <xen/evtchn.h>
+#include <xen/driver_util.h>
+
+#include <linux/types.h>
+#include <xen/public/gntdev.h>
+
+
+#define DRIVER_AUTHOR "Derek G. Murray <Derek.Murray@cl.cam.ac.uk>"
+#define DRIVER_DESC   "User-space granted page access driver"
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR(DRIVER_AUTHOR);
+MODULE_DESCRIPTION(DRIVER_DESC);
+
+#define MAX_GRANTS_LIMIT   1024
+#define DEFAULT_MAX_GRANTS 128
+
+/* A slot can be in one of three states:
+ *
+ * 0. GNTDEV_SLOT_INVALID:
+ *    This slot is not associated with a grant reference, and is therefore free
+ *    to be overwritten by a new grant reference.
+ *
+ * 1. GNTDEV_SLOT_NOT_YET_MAPPED:
+ *    This slot is associated with a grant reference (via the 
+ *    IOCTL_GNTDEV_MAP_GRANT_REF ioctl), but it has not yet been mmap()-ed.
+ *
+ * 2. GNTDEV_SLOT_MAPPED:
+ *    This slot is associated with a grant reference, and has been mmap()-ed.
+ */
+typedef enum gntdev_slot_state {
+	GNTDEV_SLOT_INVALID = 0,
+	GNTDEV_SLOT_NOT_YET_MAPPED,
+	GNTDEV_SLOT_MAPPED
+} gntdev_slot_state_t;
+
+#define GNTDEV_INVALID_HANDLE    -1
+#define GNTDEV_FREE_LIST_INVALID -1
+/* Each opened instance of gntdev is associated with a list of grants,
+ * represented by an array of elements of the following type,
+ * gntdev_grant_info_t.
+ */
+typedef struct gntdev_grant_info {
+	gntdev_slot_state_t state;
+	union {
+		uint32_t free_list_index;
+		struct {
+			domid_t domid;
+			grant_ref_t ref;
+			grant_handle_t kernel_handle;
+			grant_handle_t user_handle;
+			uint64_t dev_bus_addr;
+		} valid;
+	} u;
+} gntdev_grant_info_t;
+
+/* Private data structure, which is stored in the file pointer for files
+ * associated with this device.
+ */
+typedef struct gntdev_file_private_data {
+  
+	/* Array of grant information. */
+	gntdev_grant_info_t *grants;
+	uint32_t grants_size;
+
+	/* Read/write semaphore used to protect the grants array. */
+	struct rw_semaphore grants_sem;
+
+	/* An array of indices of free slots in the grants array.
+	 * N.B. An entry in this list may temporarily have the value
+	 * GNTDEV_FREE_LIST_INVALID if the corresponding slot has been removed
+	 * from the list by the contiguous allocator, but the list has not yet
+	 * been compressed. However, this is not visible across invocations of
+	 * the device.
+	 */
+	int32_t *free_list;
+	
+	/* The number of free slots in the grants array. */
+	uint32_t free_list_size;
+
+	/* Read/write semaphore used to protect the free list. */
+	struct rw_semaphore free_list_sem;
+	
+	/* Index of the next slot after the most recent contiguous allocation, 
+	 * for use in a next-fit allocator.
+	 */
+	uint32_t next_fit_index;
+
+	/* Used to map grants into the kernel, before mapping them into user
+	 * space.
+	 */
+	struct page **foreign_pages;
+
+} gntdev_file_private_data_t;
+
+/* Module lifecycle operations. */
+static int __init gntdev_init(void);
+static void __exit gntdev_exit(void);
+
+module_init(gntdev_init);
+module_exit(gntdev_exit);
+
+/* File operations. */
+static int gntdev_open(struct inode *inode, struct file *flip);
+static int gntdev_release(struct inode *inode, struct file *flip);
+static int gntdev_mmap(struct file *flip, struct vm_area_struct *vma);
+static long gntdev_ioctl(struct file *flip,
+			 unsigned int cmd, unsigned long arg);
+
+static const struct file_operations gntdev_fops = {
+	.owner = THIS_MODULE,
+	.open = gntdev_open,
+	.release = gntdev_release,
+	.mmap = gntdev_mmap,
+	.unlocked_ioctl = gntdev_ioctl
+};
+
+/* VM operations. */
+static void gntdev_vma_close(struct vm_area_struct *vma);
+static pte_t gntdev_clear_pte(struct vm_area_struct *vma, unsigned long addr,
+			      pte_t *ptep, int is_fullmm);
+
+static struct vm_operations_struct gntdev_vmops = {
+	.close = gntdev_vma_close,
+	.zap_pte = gntdev_clear_pte
+};
+
+/* Global variables. */
+
+/* The driver major number, for use when unregistering the driver. */
+static int gntdev_major;
+
+#define GNTDEV_NAME "gntdev"
+
+/* Memory mapping functions
+ * ------------------------
+ *
+ * Every granted page is mapped into both kernel and user space, and the two
+ * following functions return the respective virtual addresses of these pages.
+ *
+ * When shadow paging is disabled, the granted page is mapped directly into
+ * user space; when it is enabled, it is mapped into the kernel and remapped
+ * into user space using vm_insert_page() (see gntdev_mmap(), below).
+ */
+
+/* Returns the virtual address (in user space) of the @page_index'th page
+ * in the given VM area.
+ */
+static inline unsigned long get_user_vaddr (struct vm_area_struct *vma,
+					    int page_index)
+{
+	return (unsigned long) vma->vm_start + (page_index << PAGE_SHIFT);
+}
+
+/* Returns the virtual address (in kernel space) of the @slot_index'th page
+ * mapped by the gntdev instance that owns the given private data struct.
+ */
+static inline unsigned long get_kernel_vaddr (gntdev_file_private_data_t *priv,
+					      int slot_index)
+{
+	unsigned long pfn;
+	void *kaddr;
+	pfn = page_to_pfn(priv->foreign_pages[slot_index]);
+	kaddr = pfn_to_kaddr(pfn);
+	return (unsigned long) kaddr;
+}
+
+/* Helper functions. */
+
+/* Adds information about a grant reference to the list of grants in the file's
+ * private data structure. Returns non-zero on failure. On success, sets the
+ * value of *offset to the offset that should be mmap()-ed in order to map the
+ * grant reference.
+ */
+static int add_grant_reference(struct file *flip,
+			       struct ioctl_gntdev_grant_ref *op,
+			       uint64_t *offset)
+{
+	gntdev_file_private_data_t *private_data 
+		= (gntdev_file_private_data_t *) flip->private_data;
+
+	uint32_t slot_index;
+
+	if (unlikely(private_data->free_list_size == 0)) {
+		return -ENOMEM;
+	}
+
+	slot_index = private_data->free_list[--private_data->free_list_size];
+	private_data->free_list[private_data->free_list_size]
+		= GNTDEV_FREE_LIST_INVALID;
+
+	/* Copy the grant information into file's private data. */
+	private_data->grants[slot_index].state = GNTDEV_SLOT_NOT_YET_MAPPED;
+	private_data->grants[slot_index].u.valid.domid = op->domid;
+	private_data->grants[slot_index].u.valid.ref = op->ref;
+
+	/* The offset is calculated as the index of the chosen entry in the
+	 * file's private data's array of grant information. This is then
+	 * shifted to give an offset into the virtual "file address space".
+	 */
+	*offset = slot_index << PAGE_SHIFT;
+
+	return 0;
+}
+
+/* Adds the @count grant references to the contiguous range in the slot array
+ * beginning at @first_slot. It is assumed that @first_slot was returned by a
+ * previous invocation of find_contiguous_free_range(), during the same
+ * invocation of the driver.
+ */
+static int add_grant_references(struct file *flip,
+				int count,
+				struct ioctl_gntdev_grant_ref *ops,
+				uint32_t first_slot)
+{
+	gntdev_file_private_data_t *private_data 
+		= (gntdev_file_private_data_t *) flip->private_data;
+	int i;
+	
+	for (i = 0; i < count; ++i) {
+
+		/* First, mark the slot's entry in the free list as invalid. */
+		int free_list_index = 
+			private_data->grants[first_slot+i].u.free_list_index;
+		private_data->free_list[free_list_index] = 
+			GNTDEV_FREE_LIST_INVALID;
+
+		/* Now, update the slot. */
+		private_data->grants[first_slot+i].state = 
+			GNTDEV_SLOT_NOT_YET_MAPPED;
+		private_data->grants[first_slot+i].u.valid.domid =
+			ops[i].domid;
+		private_data->grants[first_slot+i].u.valid.ref = ops[i].ref;
+	}
+
+	return 0;	
+}
+
+/* Scans through the free list for @flip, removing entries that are marked as
+ * GNTDEV_SLOT_INVALID. This will reduce the recorded size of the free list to
+ * the number of valid entries.
+ */
+static void compress_free_list(struct file *flip) 
+{
+	gntdev_file_private_data_t *private_data 
+		= (gntdev_file_private_data_t *) flip->private_data;
+	int i, j = 0, old_size, slot_index;
+	
+	old_size = private_data->free_list_size;
+	for (i = 0; i < old_size; ++i) {
+		if (private_data->free_list[i] != GNTDEV_FREE_LIST_INVALID) {
+			if (i > j) {
+				slot_index = private_data->free_list[i];
+				private_data->free_list[j] = slot_index;
+				private_data->grants[slot_index].u
+					.free_list_index = j;
+				private_data->free_list[i] 
+					= GNTDEV_FREE_LIST_INVALID;
+			}
+			++j;
+		} else {
+			--private_data->free_list_size;
+		}
+	}
+}
+
+/* Searches the grant array in the private data of @flip for a range of
+ * @num_slots contiguous slots in the GNTDEV_SLOT_INVALID state.
+ *
+ * Returns the index of the first slot if a range is found, otherwise -ENOMEM.
+ */
+static int find_contiguous_free_range(struct file *flip,
+				      uint32_t num_slots) 
+{
+	gntdev_file_private_data_t *private_data 
+		= (gntdev_file_private_data_t *) flip->private_data;
+	
+	int i;
+	int start_index = private_data->next_fit_index;
+	int range_start = 0, range_length;
+
+	if (private_data->free_list_size < num_slots) {
+		return -ENOMEM;
+	}
+
+	/* First search from the start_index to the end of the array. */
+	range_length = 0;
+	for (i = start_index; i < private_data->grants_size; ++i) {
+		if (private_data->grants[i].state == GNTDEV_SLOT_INVALID) {
+			if (range_length == 0) {
+				range_start = i;
+			}
+			++range_length;
+			if (range_length == num_slots) {
+				return range_start;
+			}
+		}
+	}
+	
+	/* Now search from the start of the array to the start_index. */
+	range_length = 0;
+	for (i = 0; i < start_index; ++i) {
+		if (private_data->grants[i].state == GNTDEV_SLOT_INVALID) {
+			if (range_length == 0) {
+				range_start = i;
+			}
+			++range_length;
+			if (range_length == num_slots) {
+				return range_start;
+			}
+		}
+	}
+	
+	return -ENOMEM;
+}
+
+static int init_private_data(gntdev_file_private_data_t *priv,
+			     uint32_t max_grants)
+{
+	int i;
+
+	/* Allocate space for the kernel-mapping of granted pages. */
+	priv->foreign_pages = 
+		alloc_empty_pages_and_pagevec(max_grants);
+	if (!priv->foreign_pages)
+		goto nomem_out;
+
+	/* Allocate the grant list and free-list. */
+	priv->grants = kmalloc(max_grants * sizeof(gntdev_grant_info_t),
+			       GFP_KERNEL);
+	if (!priv->grants)
+		goto nomem_out2;
+	priv->free_list = kmalloc(max_grants * sizeof(int32_t), GFP_KERNEL);
+	if (!priv->free_list)
+		goto nomem_out3;
+
+	/* Initialise the free-list, which contains all slots at first. */
+	for (i = 0; i < max_grants; ++i) {
+		priv->free_list[max_grants - i - 1] = i;
+		priv->grants[i].state = GNTDEV_SLOT_INVALID;
+		priv->grants[i].u.free_list_index = max_grants - i - 1;
+	}
+	priv->grants_size = max_grants;
+	priv->free_list_size = max_grants;
+	priv->next_fit_index = 0;
+
+	return 0;
+
+nomem_out3:
+	kfree(priv->grants);
+nomem_out2:
+	free_empty_pages_and_pagevec(priv->foreign_pages, max_grants);
+nomem_out:
+	return -ENOMEM;
+
+}
+
+/* Interface functions. */
+
+/* Initialises the driver. Called when the module is loaded. */
+static int __init gntdev_init(void)
+{
+	struct class *class;
+	struct class_device *device;
+
+	if (!is_running_on_xen()) {
+		printk(KERN_ERR "You must be running Xen to use gntdev\n");
+		return -ENODEV;
+	}
+
+	gntdev_major = register_chrdev(0, GNTDEV_NAME, &gntdev_fops);
+	if (gntdev_major < 0)
+	{
+		printk(KERN_ERR "Could not register gntdev device\n");
+		return -ENOMEM;
+	}
+
+	/* Note that if the sysfs code fails, we will still initialise the
+	 * device, and output the major number so that the device can be
+	 * created manually using mknod.
+	 */
+	if ((class = get_xen_class()) == NULL) {
+		printk(KERN_ERR "Error setting up xen_class\n");
+		printk(KERN_ERR "gntdev created with major number = %d\n", 
+		       gntdev_major);
+		return 0;
+	}
+
+	device = class_device_create(class, NULL, MKDEV(gntdev_major, 0),
+				     NULL, GNTDEV_NAME);
+	if (IS_ERR(device)) {
+		printk(KERN_ERR "Error creating gntdev device in xen_class\n");
+		printk(KERN_ERR "gntdev created with major number = %d\n",
+		       gntdev_major);
+		return 0;
+	}
+
+	return 0;
+}
+
+/* Cleans up and unregisters the driver. Called when the driver is unloaded.
+ */
+static void __exit gntdev_exit(void)
+{
+	struct class *class;
+	if ((class = get_xen_class()) != NULL)
+		class_device_destroy(class, MKDEV(gntdev_major, 0));
+	unregister_chrdev(gntdev_major, GNTDEV_NAME);
+}
+
+/* Called when the device is opened. */
+static int gntdev_open(struct inode *inode, struct file *flip)
+{
+	gntdev_file_private_data_t *private_data;
+
+	try_module_get(THIS_MODULE);
+
+	/* Allocate space for the per-instance private data. */
+	private_data = kmalloc(sizeof(*private_data), GFP_KERNEL);
+	if (!private_data)
+		goto nomem_out;
+
+	/* These will be lazily initialised by init_private_data. */
+	private_data->grants = NULL;
+	private_data->free_list = NULL;
+	private_data->foreign_pages = NULL;
+
+	init_rwsem(&private_data->grants_sem);
+	init_rwsem(&private_data->free_list_sem);
+
+	flip->private_data = private_data;
+
+	return 0;
+
+nomem_out:
+	return -ENOMEM;
+}
+
+/* Called when the device is closed.
+ */
+static int gntdev_release(struct inode *inode, struct file *flip)
+{
+	if (flip->private_data) {
+		gntdev_file_private_data_t *private_data = 
+			(gntdev_file_private_data_t *) flip->private_data;
+		if (private_data->foreign_pages)
+			free_empty_pages_and_pagevec
+				(private_data->foreign_pages,
+				 private_data->grants_size);
+		if (private_data->grants) 
+			kfree(private_data->grants);
+		if (private_data->free_list)
+			kfree(private_data->free_list);
+		kfree(private_data);
+	}
+	module_put(THIS_MODULE);
+	return 0;
+}
+
+/* Called when an attempt is made to mmap() the device. The private data from
+ * @flip contains the list of grant references that can be mapped. The vm_pgoff
+ * field of @vma contains the index into that list that refers to the grant
+ * reference that will be mapped. Only mappings that are a multiple of
+ * PAGE_SIZE are handled.
+ */
+static int gntdev_mmap (struct file *flip, struct vm_area_struct *vma) 
+{
+	struct gnttab_map_grant_ref op;
+	unsigned long slot_index = vma->vm_pgoff;
+	unsigned long kernel_vaddr, user_vaddr;
+	uint32_t size = (vma->vm_end - vma->vm_start) >> PAGE_SHIFT;
+	uint64_t ptep;
+	int ret;
+	int flags;
+	int i;
+	struct page *page;
+	gntdev_file_private_data_t *private_data = flip->private_data;
+
+	if (unlikely(!private_data)) {
+		printk(KERN_ERR "File's private data is NULL.\n");
+		return -EINVAL;
+	}
+
+	/* Test to make sure that the grants array has been initialised. */
+	down_read(&private_data->grants_sem);
+	if (unlikely(!private_data->grants)) {
+		up_read(&private_data->grants_sem);
+		printk(KERN_ERR "Attempted to mmap before ioctl.\n");
+		return -EINVAL;
+	}
+	up_read(&private_data->grants_sem);
+
+	if (unlikely((size <= 0) || 
+		     (size + slot_index) > private_data->grants_size)) {
+		printk(KERN_ERR "Invalid number of pages or offset"
+		       "(num_pages = %d, first_slot = %ld).\n",
+		       size, slot_index);
+		return -ENXIO;
+	}
+
+	if ((vma->vm_flags & VM_WRITE) && !(vma->vm_flags & VM_SHARED)) {
+		printk(KERN_ERR "Writable mappings must be shared.\n");
+		return -EINVAL;
+	}
+
+	/* Slots must be in the NOT_YET_MAPPED state. */
+	down_write(&private_data->grants_sem);
+	for (i = 0; i < size; ++i) {
+		if (private_data->grants[slot_index + i].state != 
+		    GNTDEV_SLOT_NOT_YET_MAPPED) {
+			printk(KERN_ERR "Slot (index = %ld) is in the wrong "
+			       "state (%d).\n", slot_index + i, 
+			       private_data->grants[slot_index + i].state);
+			up_write(&private_data->grants_sem);
+			return -EINVAL;
+		}
+	}
+
+	/* Install the hook for unmapping. */
+	vma->vm_ops = &gntdev_vmops;
+    
+	/* The VM area contains pages from another VM. */
+	vma->vm_flags |= VM_FOREIGN;
+	vma->vm_private_data = kzalloc(size * sizeof(struct page *),
+				       GFP_KERNEL);
+	if (vma->vm_private_data == NULL) {
+		printk(KERN_ERR "Couldn't allocate mapping structure for VM "
+		       "area.\n");
+		return -ENOMEM;
+	}
+
+	/* This flag prevents Bad PTE errors when the memory is unmapped. */
+	vma->vm_flags |= VM_RESERVED;
+
+	/* This flag prevents this VM area being copied on a fork(). A better
+	 * behaviour might be to explicitly carry out the appropriate mappings
+	 * on fork(), but I don't know if there's a hook for this.
+	 */
+	vma->vm_flags |= VM_DONTCOPY;
+
+#ifdef CONFIG_X86
+	/* This flag ensures that the page tables are not unpinned before the
+	 * VM area is unmapped. Therefore Xen still recognises the PTE as
+	 * belonging to an L1 pagetable, and the grant unmap operation will
+	 * succeed, even if the process does not exit cleanly.
+	 */
+	vma->vm_mm->context.has_foreign_mappings = 1;
+#endif
+
+	for (i = 0; i < size; ++i) {
+
+		flags = GNTMAP_host_map;
+		if (!(vma->vm_flags & VM_WRITE))
+			flags |= GNTMAP_readonly;
+
+		kernel_vaddr = get_kernel_vaddr(private_data, slot_index + i);
+		user_vaddr = get_user_vaddr(vma, i);
+		page = pfn_to_page(__pa(kernel_vaddr) >> PAGE_SHIFT);
+
+		gnttab_set_map_op(&op, kernel_vaddr, flags,   
+				  private_data->grants[slot_index+i]
+				  .u.valid.ref, 
+				  private_data->grants[slot_index+i]
+				  .u.valid.domid);
+
+		/* Carry out the mapping of the grant reference. */
+		ret = HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, 
+						&op, 1);
+		BUG_ON(ret);
+		if (op.status) {
+			printk(KERN_ERR "Error mapping the grant reference "
+			       "into the kernel (%d). domid = %d; ref = %d\n",
+			       op.status,
+			       private_data->grants[slot_index+i]
+			       .u.valid.domid,
+			       private_data->grants[slot_index+i]
+			       .u.valid.ref);
+			goto undo_map_out;
+		}
+
+		/* Store a reference to the page that will be mapped into user
+		 * space.
+		 */
+		((struct page **) vma->vm_private_data)[i] = page;
+
+		/* Mark mapped page as reserved. */
+		SetPageReserved(page);
+
+		/* Record the grant handle, for use in the unmap operation. */
+		private_data->grants[slot_index+i].u.valid.kernel_handle = 
+			op.handle;
+		private_data->grants[slot_index+i].u.valid.dev_bus_addr = 
+			op.dev_bus_addr;
+		
+		private_data->grants[slot_index+i].state = GNTDEV_SLOT_MAPPED;
+		private_data->grants[slot_index+i].u.valid.user_handle =
+			GNTDEV_INVALID_HANDLE;
+
+		/* Now perform the mapping to user space. */
+		if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+
+			/* NOT USING SHADOW PAGE TABLES. */
+			/* In this case, we map the grant(s) straight into user
+			 * space.
+			 */
+
+			/* Get the machine address of the PTE for the user 
+			 *  page.
+			 */
+			if ((ret = create_lookup_pte_addr(vma->vm_mm, 
+							  vma->vm_start 
+							  + (i << PAGE_SHIFT), 
+							  &ptep)))
+			{
+				printk(KERN_ERR "Error obtaining PTE pointer "
+				       "(%d).\n", ret);
+				goto undo_map_out;
+			}
+			
+			/* Configure the map operation. */
+		
+			/* The reference is to be used by host CPUs. */
+			flags = GNTMAP_host_map;
+			
+			/* Specifies a user space mapping. */
+			flags |= GNTMAP_application_map;
+			
+			/* The map request contains the machine address of the
+			 * PTE to update.
+			 */
+			flags |= GNTMAP_contains_pte;
+			
+			if (!(vma->vm_flags & VM_WRITE))
+				flags |= GNTMAP_readonly;
+
+			gnttab_set_map_op(&op, ptep, flags, 
+					  private_data->grants[slot_index+i]
+					  .u.valid.ref, 
+					  private_data->grants[slot_index+i]
+					  .u.valid.domid);
+
+			/* Carry out the mapping of the grant reference. */
+			ret = HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref,
+							&op, 1);
+			BUG_ON(ret);
+			if (op.status) {
+				printk(KERN_ERR "Error mapping the grant "
+				       "reference into user space (%d). domid "
+				       "= %d; ref = %d\n", op.status,
+				       private_data->grants[slot_index+i].u
+				       .valid.domid,
+				       private_data->grants[slot_index+i].u
+				       .valid.ref);
+				goto undo_map_out;
+			}
+			
+			/* Record the grant handle, for use in the unmap 
+			 * operation. 
+			 */
+			private_data->grants[slot_index+i].u.
+				valid.user_handle = op.handle;
+
+			/* Update p2m structure with the new mapping. */
+			set_phys_to_machine(__pa(kernel_vaddr) >> PAGE_SHIFT,
+					    FOREIGN_FRAME(private_data->
+							  grants[slot_index+i]
+							  .u.valid.dev_bus_addr
+							  >> PAGE_SHIFT));
+		} else {
+			/* USING SHADOW PAGE TABLES. */
+			/* In this case, we simply insert the page into the VM
+			 * area. */
+			ret = vm_insert_page(vma, user_vaddr, page);
+		}
+
+	}
+
+	up_write(&private_data->grants_sem);
+	return 0;
+
+undo_map_out:
+	/* If we have a mapping failure, the unmapping will be taken care of
+	 * by do_mmap_pgoff(), which will eventually call gntdev_clear_pte().
+	 * All we need to do here is free the vma_private_data.
+	 */
+	kfree(vma->vm_private_data);
+
+	/* THIS IS VERY UNPLEASANT: do_mmap_pgoff() will set the vma->vm_file
+	 * to NULL on failure. However, we need this in gntdev_clear_pte() to
+	 * unmap the grants. Therefore, we smuggle a reference to the file's
+	 * private data in the VM area's private data pointer.
+	 */
+	vma->vm_private_data = private_data;
+	
+	up_write(&private_data->grants_sem);
+
+	return -ENOMEM;
+}
+
+static pte_t gntdev_clear_pte(struct vm_area_struct *vma, unsigned long addr,
+			      pte_t *ptep, int is_fullmm)
+{
+	int slot_index, ret;
+	pte_t copy;
+	struct gnttab_unmap_grant_ref op;
+	gntdev_file_private_data_t *private_data;
+
+	/* THIS IS VERY UNPLEASANT: do_mmap_pgoff() will set the vma->vm_file
+	 * to NULL on failure. However, we need this in gntdev_clear_pte() to
+	 * unmap the grants. Therefore, we smuggle a reference to the file's
+	 * private data in the VM area's private data pointer.
+	 */
+	if (vma->vm_file) {
+		private_data = (gntdev_file_private_data_t *)
+			vma->vm_file->private_data;
+	} else if (vma->vm_private_data) {
+		private_data = (gntdev_file_private_data_t *)
+			vma->vm_private_data;
+	} else {
+		private_data = NULL; /* gcc warning */
+		BUG();
+	}
+
+	/* Copy the existing value of the PTE for returning. */
+	copy = *ptep;
+
+	/* Calculate the grant relating to this PTE. */
+	slot_index = vma->vm_pgoff + ((addr - vma->vm_start) >> PAGE_SHIFT);
+
+	/* Only unmap grants if the slot has been mapped. This could be being
+	 * called from a failing mmap().
+	 */
+	if (private_data->grants[slot_index].state == GNTDEV_SLOT_MAPPED) {
+
+		/* First, we clear the user space mapping, if it has been made.
+		 */
+		if (private_data->grants[slot_index].u.valid.user_handle !=
+		    GNTDEV_INVALID_HANDLE && 
+		    !xen_feature(XENFEAT_auto_translated_physmap)) {
+			/* NOT USING SHADOW PAGE TABLES. */
+			gnttab_set_unmap_op(&op, virt_to_machine(ptep), 
+					    GNTMAP_contains_pte,
+					    private_data->grants[slot_index]
+					    .u.valid.user_handle);
+			ret = HYPERVISOR_grant_table_op(
+				GNTTABOP_unmap_grant_ref, &op, 1);
+			BUG_ON(ret);
+			if (op.status)
+				printk("User unmap grant status = %d\n", 
+				       op.status);
+		} else {
+			/* USING SHADOW PAGE TABLES. */
+			pte_clear_full(vma->vm_mm, addr, ptep, is_fullmm);
+		}
+
+		/* Finally, we unmap the grant from kernel space. */
+		gnttab_set_unmap_op(&op, 
+				    get_kernel_vaddr(private_data, slot_index),
+				    GNTMAP_host_map, 
+				    private_data->grants[slot_index].u.valid
+				    .kernel_handle);
+		ret = HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, 
+						&op, 1);
+		BUG_ON(ret);
+		if (op.status)
+			printk("Kernel unmap grant status = %d\n", op.status);
+
+
+		/* Return slot to the not-yet-mapped state, so that it may be
+		 * mapped again, or removed by a subsequent ioctl.
+		 */
+		private_data->grants[slot_index].state = 
+			GNTDEV_SLOT_NOT_YET_MAPPED;
+
+		/* Invalidate the physical to machine mapping for this page. */
+		set_phys_to_machine(__pa(get_kernel_vaddr(private_data, 
+							  slot_index)) 
+				    >> PAGE_SHIFT, INVALID_P2M_ENTRY);
+
+	} else {
+		pte_clear_full(vma->vm_mm, addr, ptep, is_fullmm);
+	}
+
+	return copy;
+}
+
+/* "Destructor" for a VM area.
+ */
+static void gntdev_vma_close(struct vm_area_struct *vma) {
+	if (vma->vm_private_data) {
+		kfree(vma->vm_private_data);
+	}
+}
+
+/* Called when an ioctl is made on the device.
+ */
+static long gntdev_ioctl(struct file *flip,
+			 unsigned int cmd, unsigned long arg)
+{
+	int rc = 0;
+	gntdev_file_private_data_t *private_data = 
+		(gntdev_file_private_data_t *) flip->private_data;
+
+	/* On the first invocation, we will lazily initialise the grant array
+	 * and free-list.
+	 */
+	if (unlikely(!private_data->grants) 
+	    && likely(cmd != IOCTL_GNTDEV_SET_MAX_GRANTS)) {
+		down_write(&private_data->grants_sem);
+		
+		if (unlikely(private_data->grants)) {
+			up_write(&private_data->grants_sem);
+			goto private_data_initialised;
+		}
+		
+		/* Just use the default. Setting to a non-default is handled
+		 * in the ioctl switch.
+		 */
+		rc = init_private_data(private_data, DEFAULT_MAX_GRANTS);
+		
+		up_write(&private_data->grants_sem);
+
+		if (rc) {
+			printk (KERN_ERR "Initialising gntdev private data "
+				"failed.\n");
+			return rc;
+		}
+	}
+	    
+private_data_initialised:
+	switch (cmd) {
+	case IOCTL_GNTDEV_MAP_GRANT_REF:
+	{
+		struct ioctl_gntdev_map_grant_ref op;
+		down_write(&private_data->grants_sem);
+		down_write(&private_data->free_list_sem);
+
+		if ((rc = copy_from_user(&op, (void __user *) arg, 
+					 sizeof(op)))) {
+			rc = -EFAULT;
+			goto map_out;
+		}
+		if (unlikely(op.count <= 0)) {
+			rc = -EINVAL;
+			goto map_out;
+		}
+
+		if (op.count == 1) {
+			if ((rc = add_grant_reference(flip, &op.refs[0],
+						      &op.index)) < 0) {
+				printk(KERN_ERR "Adding grant reference "
+				       "failed (%d).\n", rc);
+				goto map_out;
+			}
+		} else {
+			struct ioctl_gntdev_grant_ref *refs, *u;
+			refs = kmalloc(op.count * sizeof(*refs), GFP_KERNEL);
+			if (!refs) {
+				rc = -ENOMEM;
+				goto map_out;
+			}
+			u = ((struct ioctl_gntdev_map_grant_ref *)arg)->refs;
+			if ((rc = copy_from_user(refs,
+						 (void __user *)u,
+						 sizeof(*refs) * op.count))) {
+				printk(KERN_ERR "Copying refs from user failed"
+				       " (%d).\n", rc);
+				rc = -EINVAL;
+				goto map_out;
+			}
+			if ((rc = find_contiguous_free_range(flip, op.count))
+			    < 0) {
+				printk(KERN_ERR "Finding contiguous range "
+				       "failed (%d).\n", rc);
+				kfree(refs);
+				goto map_out;
+			}
+			op.index = rc << PAGE_SHIFT;
+			if ((rc = add_grant_references(flip, op.count,
+						       refs, rc))) {
+				printk(KERN_ERR "Adding grant references "
+				       "failed (%d).\n", rc);
+				kfree(refs);
+				goto map_out;
+			}
+			compress_free_list(flip);
+			kfree(refs);
+		}
+		if ((rc = copy_to_user((void __user *) arg, 
+				       &op, 
+				       sizeof(op)))) {
+			printk(KERN_ERR "Copying result back to user failed "
+			       "(%d)\n", rc);
+			rc = -EFAULT;
+			goto map_out;
+		}
+	map_out:
+		up_write(&private_data->grants_sem);
+		up_write(&private_data->free_list_sem);
+		return rc;
+	}
+	case IOCTL_GNTDEV_UNMAP_GRANT_REF:
+	{
+		struct ioctl_gntdev_unmap_grant_ref op;
+		int i, start_index;
+
+		down_write(&private_data->grants_sem);
+		down_write(&private_data->free_list_sem);
+
+		if ((rc = copy_from_user(&op, 
+					 (void __user *) arg, 
+					 sizeof(op)))) {
+			rc = -EFAULT;
+			goto unmap_out;
+		}
+
+		start_index = op.index >> PAGE_SHIFT;
+
+		/* First, check that all pages are in the NOT_YET_MAPPED
+		 * state.
+		 */
+		for (i = 0; i < op.count; ++i) {
+			if (unlikely
+			    (private_data->grants[start_index + i].state
+			     != GNTDEV_SLOT_NOT_YET_MAPPED)) {
+				if (private_data->grants[start_index + i].state
+				    == GNTDEV_SLOT_INVALID) {
+					printk(KERN_ERR
+					       "Tried to remove an invalid "
+					       "grant at offset 0x%x.",
+					       (start_index + i) 
+					       << PAGE_SHIFT);
+					rc = -EINVAL;
+				} else {
+					printk(KERN_ERR
+					       "Tried to remove a grant which "
+					       "is currently mmap()-ed at "
+					       "offset 0x%x.",
+					       (start_index + i) 
+					       << PAGE_SHIFT);
+					rc = -EBUSY;
+				}
+				goto unmap_out;
+			}
+		}
+
+		/* Unmap pages and add them to the free list.
+		 */
+		for (i = 0; i < op.count; ++i) {
+			private_data->grants[start_index+i].state = 
+				GNTDEV_SLOT_INVALID;
+			private_data->grants[start_index+i].u.free_list_index =
+				private_data->free_list_size;
+			private_data->free_list[private_data->free_list_size] =
+				start_index + i;
+			++private_data->free_list_size;
+		}
+
+	unmap_out:
+		up_write(&private_data->grants_sem);
+		up_write(&private_data->free_list_sem);
+		return rc;
+	}
+	case IOCTL_GNTDEV_GET_OFFSET_FOR_VADDR:
+	{
+		struct ioctl_gntdev_get_offset_for_vaddr op;
+		struct vm_area_struct *vma;
+		unsigned long vaddr;
+
+		if ((rc = copy_from_user(&op, 
+					 (void __user *) arg, 
+					 sizeof(op)))) {
+			rc = -EFAULT;
+			goto get_offset_out;
+		}
+		vaddr = (unsigned long)op.vaddr;
+
+		down_read(&current->mm->mmap_sem);		
+		vma = find_vma(current->mm, vaddr);
+		if (vma == NULL) {
+			rc = -EFAULT;
+			goto get_offset_unlock_out;
+		}
+		if ((!vma->vm_ops) || (vma->vm_ops != &gntdev_vmops)) {
+			printk(KERN_ERR "The vaddr specified does not belong "
+			       "to a gntdev instance: %#lx\n", vaddr);
+			rc = -EFAULT;
+			goto get_offset_unlock_out;
+		}
+		if (vma->vm_start != vaddr) {
+			printk(KERN_ERR "The vaddr specified in an "
+			       "IOCTL_GNTDEV_GET_OFFSET_FOR_VADDR must be at "
+			       "the start of the VM area. vma->vm_start = "
+			       "%#lx; vaddr = %#lx\n",
+			       vma->vm_start, vaddr);
+			rc = -EFAULT;
+			goto get_offset_unlock_out;
+		}
+		op.offset = vma->vm_pgoff << PAGE_SHIFT;
+		op.count = (vma->vm_end - vma->vm_start) >> PAGE_SHIFT;
+		up_read(&current->mm->mmap_sem);
+		if ((rc = copy_to_user((void __user *) arg, 
+				       &op, 
+				       sizeof(op)))) {
+			rc = -EFAULT;
+			goto get_offset_out;
+		}
+		goto get_offset_out;
+	get_offset_unlock_out:
+		up_read(&current->mm->mmap_sem);
+	get_offset_out:
+		return rc;
+	}
+	case IOCTL_GNTDEV_SET_MAX_GRANTS:
+	{
+		struct ioctl_gntdev_set_max_grants op;
+		if ((rc = copy_from_user(&op, 
+					 (void __user *) arg, 
+					 sizeof(op)))) {
+			rc = -EFAULT;
+			goto set_max_out;
+		}
+		down_write(&private_data->grants_sem);
+		if (private_data->grants) {
+			rc = -EBUSY;
+			goto set_max_unlock_out;
+		}
+		if (op.count > MAX_GRANTS_LIMIT) {
+			rc = -EINVAL;
+			goto set_max_unlock_out;
+		}						 
+		rc = init_private_data(private_data, op.count);
+	set_max_unlock_out:
+		up_write(&private_data->grants_sem);
+	set_max_out:
+		return rc;
+	}
+	default:
+		return -ENOIOCTLCMD;
+	}
+
+	return 0;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/gntdev/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/gntdev/Makefile
--- linux-2.6.18.8/drivers/xen/gntdev/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/gntdev/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1 @@
+obj-$(CONFIG_XEN_GRANT_DEV) := gntdev.o
diff -rpuN linux-2.6.18.8/drivers/xen/Kconfig linux-2.6.18-xen-3.3.0/drivers/xen/Kconfig
--- linux-2.6.18.8/drivers/xen/Kconfig	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/Kconfig	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,314 @@
+#
+# This Kconfig describe xen options
+#
+
+mainmenu "Xen Configuration"
+
+config XEN
+	bool
+
+if XEN
+config XEN_INTERFACE_VERSION
+	hex
+	default 0x00030207
+
+menu "XEN"
+
+config XEN_PRIVILEGED_GUEST
+	bool "Privileged Guest (domain 0)"
+	help
+	  Support for privileged operation (domain 0)
+
+config XEN_UNPRIVILEGED_GUEST
+	def_bool !XEN_PRIVILEGED_GUEST
+
+config XEN_PRIVCMD
+	def_bool y
+	depends on PROC_FS
+
+config XEN_XENBUS_DEV
+	def_bool y
+	depends on PROC_FS
+
+config XEN_NETDEV_ACCEL_SFC_UTIL
+    	depends on X86    
+	tristate
+
+config XEN_BACKEND
+        tristate "Backend driver support"
+        default XEN_PRIVILEGED_GUEST
+        help
+          Support for backend device drivers that provide I/O services
+          to other virtual machines.
+
+config XEN_BLKDEV_BACKEND
+	tristate "Block-device backend driver"
+        depends on XEN_BACKEND
+	default XEN_BACKEND
+	help
+	  The block-device backend driver allows the kernel to export its
+	  block devices to other guests via a high-performance shared-memory
+	  interface.
+
+config XEN_BLKDEV_TAP
+	tristate "Block-device tap backend driver"
+	depends on XEN_BACKEND
+	default XEN_BACKEND
+	help
+	  The block tap driver is an alternative to the block back driver 
+          and allows VM block requests to be redirected to userspace through
+          a device interface.  The tap allows user-space development of 
+          high-performance block backends, where disk images may be implemented
+          as files, in memory, or on other hosts across the network.  This 
+	  driver can safely coexist with the existing blockback driver.
+
+config XEN_NETDEV_BACKEND
+	tristate "Network-device backend driver"
+        depends on XEN_BACKEND && NET
+	default XEN_BACKEND
+	help
+	  The network-device backend driver allows the kernel to export its
+	  network devices to other guests via a high-performance shared-memory
+	  interface.
+
+config XEN_NETDEV_PIPELINED_TRANSMITTER
+	bool "Pipelined transmitter (DANGEROUS)"
+	depends on XEN_NETDEV_BACKEND
+	help
+	  If the net backend is a dumb domain, such as a transparent Ethernet
+	  bridge with no local IP interface, it is safe to say Y here to get
+	  slightly lower network overhead.
+	  If the backend has a local IP interface; or may be doing smart things
+	  like reassembling packets to perform firewall filtering; or if you
+	  are unsure; or if you experience network hangs when this option is
+	  enabled; then you must say N here.
+
+config XEN_NETDEV_ACCEL_SFC_BACKEND
+	tristate "Network-device backend driver acceleration for Solarflare NICs"
+	depends on XEN_NETDEV_BACKEND && SFC && SFC_RESOURCE && X86
+	select XEN_NETDEV_ACCEL_SFC_UTIL
+	default m
+
+config XEN_NETDEV_LOOPBACK
+	tristate "Network-device loopback driver"
+	depends on XEN_NETDEV_BACKEND
+	help
+	  A two-interface loopback device to emulate a local netfront-netback
+	  connection. If unsure, it is probably safe to say N here.
+
+config XEN_PCIDEV_BACKEND
+	tristate "PCI-device backend driver"
+	depends on PCI && XEN_BACKEND
+	default XEN_BACKEND
+	help
+	  The PCI device backend driver allows the kernel to export arbitrary
+	  PCI devices to other guests. If you select this to be a module, you
+	  will need to make sure no other driver has bound to the device(s)
+	  you want to make visible to other guests.
+
+choice
+	prompt "PCI Backend Mode"
+	depends on XEN_PCIDEV_BACKEND
+	default XEN_PCIDEV_BACKEND_VPCI if !IA64
+	default XEN_PCIDEV_BACKEND_CONTROLLER if IA64
+
+config XEN_PCIDEV_BACKEND_VPCI
+	bool "Virtual PCI"
+	---help---
+	  This PCI Backend hides the true PCI topology and makes the frontend
+	  think there is a single PCI bus with only the exported devices on it.
+	  For example, a device at 03:05.0 will be re-assigned to 00:00.0. A
+	  second device at 02:1a.1 will be re-assigned to 00:01.1.
+
+config XEN_PCIDEV_BACKEND_PASS
+	bool "Passthrough"
+	---help---
+	  This PCI Backend provides a real view of the PCI topology to the
+	  frontend (for example, a device at 06:01.b will still appear at
+	  06:01.b to the frontend). This is similar to how Xen 2.0.x exposed
+	  PCI devices to its driver domains. This may be required for drivers
+	  which depend on finding their hardward in certain bus/slot
+	  locations.
+
+config XEN_PCIDEV_BACKEND_SLOT
+	bool "Slot"
+	---help---
+	  This PCI Backend hides the true PCI topology and makes the frontend
+	  think there is a single PCI bus with only the exported devices on it.
+	  Contrary to the virtual PCI backend, a function becomes a new slot.
+	  For example, a device at 03:05.2 will be re-assigned to 00:00.0. A
+	  second device at 02:1a.1 will be re-assigned to 00:01.0.
+
+config XEN_PCIDEV_BACKEND_CONTROLLER
+	bool "Controller"
+	depends on IA64
+	---help---
+	  This PCI backend virtualizes the PCI bus topology by providing a
+	  virtual bus per PCI root device.  Devices which are physically under
+	  the same root bus will appear on the same virtual bus.  For systems
+	  with complex I/O addressing, this is the only backend which supports
+	  extended I/O port spaces and MMIO translation offsets.  This backend
+	  also supports slot virtualization.  For example, a device at
+	  0000:01:02.1 will be re-assigned to 0000:00:00.0.  A second device
+	  at 0000:02:05.0 (behind a P2P bridge on bus 0000:01) will be
+	  re-assigned to 0000:00:01.0.  A third device at 0000:16:05.0 (under
+	  a different PCI root bus) will be re-assigned to 0000:01:00.0.
+
+endchoice
+
+config XEN_PCIDEV_BE_DEBUG
+	bool "PCI Backend Debugging"
+	depends on XEN_PCIDEV_BACKEND
+
+config XEN_TPMDEV_BACKEND
+	tristate "TPM-device backend driver"
+        depends on XEN_BACKEND
+	help
+	  The TPM-device backend driver
+
+config XEN_SCSI_BACKEND
+	tristate "SCSI backend driver"
+	depends on SCSI && XEN_BACKEND
+	default m
+	help
+	  The SCSI backend driver allows the kernel to export its SCSI Devices
+	  to other guests via a high-performance shared-memory interface.
+
+config XEN_BLKDEV_FRONTEND
+	tristate "Block-device frontend driver"
+	default y
+	help
+	  The block-device frontend driver allows the kernel to access block
+	  devices mounted within another guest OS. Unless you are building a
+	  dedicated device-driver domain, or your master control domain
+	  (domain 0), then you almost certainly want to say Y here.
+
+config XEN_NETDEV_FRONTEND
+	tristate "Network-device frontend driver"
+	depends on NET
+	default y
+	help
+	  The network-device frontend driver allows the kernel to access
+	  network interfaces within another guest OS. Unless you are building a
+	  dedicated device-driver domain, or your master control domain
+	  (domain 0), then you almost certainly want to say Y here.
+
+config XEN_NETDEV_ACCEL_SFC_FRONTEND
+	tristate "Network-device frontend driver acceleration for Solarflare NICs"
+	depends on XEN_NETDEV_FRONTEND && X86
+	select XEN_NETDEV_ACCEL_SFC_UTIL
+	default m
+
+config XEN_SCSI_FRONTEND
+	tristate "SCSI frontend driver"
+	depends on SCSI
+	default m
+	help
+	  The SCSI frontend driver allows the kernel to access SCSI Devices
+	  within another guest OS.
+
+config XEN_GRANT_DEV
+	tristate "User-space granted page access driver"
+	default XEN_PRIVILEGED_GUEST
+	help
+	  Device for accessing (in user-space) pages that have been granted
+	  by other domains.
+
+config XEN_FRAMEBUFFER
+	tristate "Framebuffer-device frontend driver"
+	depends on FB
+	select FB_CFB_FILLRECT
+	select FB_CFB_COPYAREA
+	select FB_CFB_IMAGEBLIT
+	default y
+	help
+	  The framebuffer-device frontend drivers allows the kernel to create a
+	  virtual framebuffer.  This framebuffer can be viewed in another
+	  domain.  Unless this domain has access to a real video card, you
+	  probably want to say Y here.
+
+config XEN_KEYBOARD
+	tristate "Keyboard-device frontend driver"
+	depends on XEN_FRAMEBUFFER && INPUT
+	default y
+	help
+	  The keyboard-device frontend driver allows the kernel to create a
+	  virtual keyboard.  This keyboard can then be driven by another
+	  domain.  If you've said Y to CONFIG_XEN_FRAMEBUFFER, you probably
+	  want to say Y here.
+
+config XEN_SCRUB_PAGES
+	bool "Scrub memory before freeing it to Xen"
+	default y
+	help
+	  Erase memory contents before freeing it back to Xen's global
+	  pool. This ensures that any secrets contained within that
+	  memory (e.g., private keys) cannot be found by other guests that
+	  may be running on the machine. Most people will want to say Y here.
+	  If security is not a concern then you may increase performance by
+	  saying N.
+
+config XEN_DISABLE_SERIAL
+	bool "Disable serial port drivers"
+	default y
+	help
+	  Disable serial port drivers, allowing the Xen console driver
+	  to provide a serial console at ttyS0.
+
+config XEN_SYSFS
+	tristate "Export Xen attributes in sysfs"
+	depends on SYSFS
+	select SYS_HYPERVISOR
+	default y
+	help
+	  Xen hypervisor attributes will show up under /sys/hypervisor/.
+
+choice
+	prompt "Xen version compatibility"
+	default XEN_COMPAT_030002_AND_LATER
+
+	config XEN_COMPAT_030002_AND_LATER
+		bool "3.0.2 and later"
+
+	config XEN_COMPAT_030004_AND_LATER
+		bool "3.0.4 and later"
+
+	config XEN_COMPAT_030100_AND_LATER
+		bool "3.1.0 and later"
+
+	config XEN_COMPAT_LATEST_ONLY
+		bool "no compatibility code"
+
+endchoice
+
+config XEN_COMPAT
+	hex
+	default 0xffffff if XEN_COMPAT_LATEST_ONLY
+	default 0x030100 if XEN_COMPAT_030100_AND_LATER
+	default 0x030004 if XEN_COMPAT_030004_AND_LATER
+	default 0x030002 if XEN_COMPAT_030002_AND_LATER
+	default 0
+
+endmenu
+
+config HAVE_IRQ_IGNORE_UNHANDLED
+	def_bool y
+
+config NO_IDLE_HZ
+	def_bool y
+
+config XEN_SMPBOOT
+	def_bool y
+	depends on SMP && !PPC_XEN
+
+config XEN_BALLOON
+	def_bool y
+	depends on !PPC_XEN
+
+config XEN_XENCOMM
+	bool
+
+config XEN_DEVMEM
+	def_bool y
+
+endif
diff -rpuN linux-2.6.18.8/drivers/xen/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/Makefile
--- linux-2.6.18.8/drivers/xen/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,25 @@
+obj-y	+= core/
+obj-y	+= console/
+obj-y	+= evtchn/
+obj-y	+= xenbus/
+obj-y	+= char/
+
+obj-y	+= util.o
+obj-$(CONFIG_XEN_BALLOON)		+= balloon/
+obj-$(CONFIG_XEN_BLKDEV_BACKEND)	+= blkback/
+obj-$(CONFIG_XEN_BLKDEV_TAP)		+= blktap/
+obj-$(CONFIG_XEN_NETDEV_BACKEND)	+= netback/
+obj-$(CONFIG_XEN_TPMDEV_BACKEND)	+= tpmback/
+obj-$(CONFIG_XEN_BLKDEV_FRONTEND)	+= blkfront/
+obj-$(CONFIG_XEN_NETDEV_FRONTEND)	+= netfront/
+obj-$(CONFIG_XEN_PCIDEV_BACKEND)	+= pciback/
+obj-$(CONFIG_XEN_PCIDEV_FRONTEND)	+= pcifront/
+obj-$(CONFIG_XEN_FRAMEBUFFER)		+= fbfront/
+obj-$(CONFIG_XEN_KEYBOARD)		+= fbfront/
+obj-$(CONFIG_XEN_SCSI_BACKEND)		+= scsiback/
+obj-$(CONFIG_XEN_SCSI_FRONTEND)		+= scsifront/
+obj-$(CONFIG_XEN_PRIVCMD)	+= privcmd/
+obj-$(CONFIG_XEN_GRANT_DEV)	+= gntdev/
+obj-$(CONFIG_XEN_NETDEV_ACCEL_SFC_UTIL)		+= sfc_netutil/
+obj-$(CONFIG_XEN_NETDEV_ACCEL_SFC_FRONTEND)	+= sfc_netfront/
+obj-$(CONFIG_XEN_NETDEV_ACCEL_SFC_BACKEND)	+= sfc_netback/
diff -rpuN linux-2.6.18.8/drivers/xen/netback/accel.c linux-2.6.18-xen-3.3.0/drivers/xen/netback/accel.c
--- linux-2.6.18.8/drivers/xen/netback/accel.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/netback/accel.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,269 @@
+/******************************************************************************
+ * drivers/xen/netback/accel.c
+ *
+ * Interface between backend virtual network device and accelerated plugin. 
+ * 
+ * Copyright (C) 2007 Solarflare Communications, Inc
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/list.h>
+#include <asm/atomic.h>
+#include <xen/xenbus.h>
+#include <linux/mutex.h>
+
+#include "common.h"
+
+#if 0
+#undef DPRINTK
+#define DPRINTK(fmt, args...)						\
+	printk("netback/accel (%s:%d) " fmt ".\n", __FUNCTION__, __LINE__, ##args)
+#endif
+
+/* 
+ * A list of available netback accelerator plugin modules (each list
+ * entry is of type struct netback_accelerator) 
+ */ 
+static struct list_head accelerators_list;
+/* Lock used to protect access to accelerators_list */
+DEFINE_MUTEX(accelerators_mutex);
+
+/* 
+ * Compare a backend to an accelerator, and decide if they are
+ * compatible (i.e. if the accelerator should be used by the
+ * backend) 
+ */
+static int match_accelerator(struct xenbus_device *xendev,
+			     struct backend_info *be, 
+			     struct netback_accelerator *accelerator)
+{
+	int rc = 0;
+	char *eth_name = xenbus_read(XBT_NIL, xendev->nodename, "accel", NULL);
+	
+	if (IS_ERR(eth_name)) {
+		/* Probably means not present */
+		DPRINTK("%s: no match due to xenbus_read accel error %d\n", 
+			__FUNCTION__, PTR_ERR(eth_name));
+		return 0;
+	} else {
+		if (!strcmp(eth_name, accelerator->eth_name))
+			rc = 1;
+		kfree(eth_name);
+		return rc;
+	}
+}
+
+
+static void do_probe(struct backend_info *be, 
+		     struct netback_accelerator *accelerator,
+		     struct xenbus_device *xendev) 
+{
+	be->accelerator = accelerator;
+	atomic_inc(&be->accelerator->use_count);
+	if (be->accelerator->hooks->probe(xendev) != 0) {
+		atomic_dec(&be->accelerator->use_count);
+		module_put(be->accelerator->hooks->owner);
+		be->accelerator = NULL;
+	}
+}
+
+
+/*
+ * Notify suitable backends that a new accelerator is available and
+ * connected.  This will also notify the accelerator plugin module
+ * that it is being used for a device through the probe hook.
+ */
+static int netback_accelerator_probe_backend(struct device *dev, void *arg)
+{
+	struct netback_accelerator *accelerator = 
+		(struct netback_accelerator *)arg;
+	struct xenbus_device *xendev = to_xenbus_device(dev);
+
+	if (!strcmp("vif", xendev->devicetype)) {
+		struct backend_info *be = xendev->dev.driver_data;
+
+		if (match_accelerator(xendev, be, accelerator) &&
+		    try_module_get(accelerator->hooks->owner)) {
+			do_probe(be, accelerator, xendev);
+		}
+	}
+	return 0;
+}
+
+
+/*
+ * Notify suitable backends that an accelerator is unavailable.
+ */
+static int netback_accelerator_remove_backend(struct device *dev, void *arg)
+{
+	struct xenbus_device *xendev = to_xenbus_device(dev);
+	struct netback_accelerator *accelerator = 
+		(struct netback_accelerator *)arg;
+	
+	if (!strcmp("vif", xendev->devicetype)) {
+		struct backend_info *be = xendev->dev.driver_data;
+
+		if (be->accelerator == accelerator) {
+			be->accelerator->hooks->remove(xendev);
+			atomic_dec(&be->accelerator->use_count);
+			module_put(be->accelerator->hooks->owner);
+			be->accelerator = NULL;
+		}
+	}
+	return 0;
+}
+
+
+
+/*
+ * Entry point for an netback accelerator plugin module.  Called to
+ * advertise its presence, and connect to any suitable backends.
+ */
+int netback_connect_accelerator(unsigned version, int id, const char *eth_name, 
+				struct netback_accel_hooks *hooks)
+{
+	struct netback_accelerator *new_accelerator;
+	unsigned eth_name_len;
+
+	if (version != NETBACK_ACCEL_VERSION) {
+		if (version > NETBACK_ACCEL_VERSION) {
+			/* Caller has higher version number, leave it
+			   up to them to decide whether to continue.
+			   They can recall with a lower number if
+			   they're happy to be compatible with us */
+			return NETBACK_ACCEL_VERSION;
+		} else {
+			/* We have a more recent version than caller.
+			   Currently reject, but may in future be able
+			   to be backwardly compatible */
+			return -EPROTO;
+		}
+	}
+
+	new_accelerator = 
+		kmalloc(sizeof(struct netback_accelerator), GFP_KERNEL);
+	if (!new_accelerator) {
+		DPRINTK("%s: failed to allocate memory for accelerator\n",
+			__FUNCTION__);
+		return -ENOMEM;
+	}
+
+	new_accelerator->id = id;
+	
+	eth_name_len = strlen(eth_name)+1;
+	new_accelerator->eth_name = kmalloc(eth_name_len, GFP_KERNEL);
+	if (!new_accelerator->eth_name) {
+		DPRINTK("%s: failed to allocate memory for eth_name string\n",
+			__FUNCTION__);
+		kfree(new_accelerator);
+		return -ENOMEM;
+	}
+	strlcpy(new_accelerator->eth_name, eth_name, eth_name_len);
+	
+	new_accelerator->hooks = hooks;
+
+	atomic_set(&new_accelerator->use_count, 0);
+	
+	mutex_lock(&accelerators_mutex);
+	list_add(&new_accelerator->link, &accelerators_list);
+	
+	/* tell existing backends about new plugin */
+	xenbus_for_each_backend(new_accelerator, 
+				netback_accelerator_probe_backend);
+
+	mutex_unlock(&accelerators_mutex);
+
+	return 0;
+
+}
+EXPORT_SYMBOL_GPL(netback_connect_accelerator);
+
+
+/* 
+ * Disconnect an accelerator plugin module that has previously been
+ * connected.
+ */
+void netback_disconnect_accelerator(int id, const char *eth_name)
+{
+	struct netback_accelerator *accelerator, *next;
+
+	mutex_lock(&accelerators_mutex);
+	list_for_each_entry_safe(accelerator, next, &accelerators_list, link) {
+		if (!strcmp(eth_name, accelerator->eth_name)) {
+			xenbus_for_each_backend
+				(accelerator, netback_accelerator_remove_backend);
+			BUG_ON(atomic_read(&accelerator->use_count) != 0);
+			list_del(&accelerator->link);				
+			kfree(accelerator->eth_name);
+			kfree(accelerator);
+			break;
+		}
+	}
+	mutex_unlock(&accelerators_mutex);
+}
+EXPORT_SYMBOL_GPL(netback_disconnect_accelerator);
+
+
+void netback_probe_accelerators(struct backend_info *be,
+				struct xenbus_device *dev)
+{
+	struct netback_accelerator *accelerator;
+
+	/* 
+	 * Check list of accelerators to see if any is suitable, and
+	 * use it if it is.
+	 */
+	mutex_lock(&accelerators_mutex);
+	list_for_each_entry(accelerator, &accelerators_list, link) { 
+		if (match_accelerator(dev, be, accelerator) &&
+		    try_module_get(accelerator->hooks->owner)) {
+			do_probe(be, accelerator, dev);
+			break;
+		}
+	}
+	mutex_unlock(&accelerators_mutex);
+}
+
+
+void netback_remove_accelerators(struct backend_info *be,
+				 struct xenbus_device *dev)
+{
+	mutex_lock(&accelerators_mutex);
+	/* Notify the accelerator (if any) of this device's removal */
+	if (be->accelerator != NULL) {
+		be->accelerator->hooks->remove(dev);
+		atomic_dec(&be->accelerator->use_count);
+		module_put(be->accelerator->hooks->owner);
+		be->accelerator = NULL;
+	}
+	mutex_unlock(&accelerators_mutex);
+}
+
+
+void netif_accel_init(void)
+{
+	INIT_LIST_HEAD(&accelerators_list);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/netback/common.h linux-2.6.18-xen-3.3.0/drivers/xen/netback/common.h
--- linux-2.6.18.8/drivers/xen/netback/common.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/netback/common.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,217 @@
+/******************************************************************************
+ * arch/xen/drivers/netif/backend/common.h
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __NETIF__BACKEND__COMMON_H__
+#define __NETIF__BACKEND__COMMON_H__
+
+#include <linux/version.h>
+#include <linux/module.h>
+#include <linux/interrupt.h>
+#include <linux/slab.h>
+#include <linux/ip.h>
+#include <linux/in.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/wait.h>
+#include <xen/evtchn.h>
+#include <xen/interface/io/netif.h>
+#include <asm/io.h>
+#include <asm/pgalloc.h>
+#include <xen/interface/grant_table.h>
+#include <xen/gnttab.h>
+#include <xen/driver_util.h>
+#include <xen/xenbus.h>
+
+#define DPRINTK(_f, _a...)			\
+	pr_debug("(file=%s, line=%d) " _f,	\
+		 __FILE__ , __LINE__ , ## _a )
+#define IPRINTK(fmt, args...)				\
+	printk(KERN_INFO "xen_net: " fmt, ##args)
+#define WPRINTK(fmt, args...)				\
+	printk(KERN_WARNING "xen_net: " fmt, ##args)
+
+typedef struct netif_st {
+	/* Unique identifier for this interface. */
+	domid_t          domid;
+	unsigned int     handle;
+
+	u8               fe_dev_addr[6];
+
+	/* Physical parameters of the comms window. */
+	grant_handle_t   tx_shmem_handle;
+	grant_ref_t      tx_shmem_ref;
+	grant_handle_t   rx_shmem_handle;
+	grant_ref_t      rx_shmem_ref;
+	unsigned int     irq;
+
+	/* The shared rings and indexes. */
+	netif_tx_back_ring_t tx;
+	netif_rx_back_ring_t rx;
+	struct vm_struct *tx_comms_area;
+	struct vm_struct *rx_comms_area;
+
+	/* Set of features that can be turned on in dev->features. */
+	int features;
+
+	/* Internal feature information. */
+	u8 can_queue:1;	/* can queue packets for receiver? */
+	u8 copying_receiver:1;	/* copy packets to receiver?       */
+
+	/* Allow netif_be_start_xmit() to peek ahead in the rx request ring. */
+	RING_IDX rx_req_cons_peek;
+
+	/* Transmit shaping: allow 'credit_bytes' every 'credit_usec'. */
+	unsigned long   credit_bytes;
+	unsigned long   credit_usec;
+	unsigned long   remaining_credit;
+	struct timer_list credit_timeout;
+
+	/* Enforce draining of the transmit queue. */
+	struct timer_list tx_queue_timeout;
+
+	/* Miscellaneous private stuff. */
+	struct list_head list;  /* scheduling list */
+	atomic_t         refcnt;
+	struct net_device *dev;
+	struct net_device_stats stats;
+
+	unsigned int carrier;
+
+	wait_queue_head_t waiting_to_free;
+} netif_t;
+
+/*
+ * Implement our own carrier flag: the network stack's version causes delays
+ * when the carrier is re-enabled (in particular, dev_activate() may not
+ * immediately be called, which can cause packet loss; also the etherbridge
+ * can be rather lazy in activating its port).
+ */
+#define netback_carrier_on(netif)	((netif)->carrier = 1)
+#define netback_carrier_off(netif)	((netif)->carrier = 0)
+#define netback_carrier_ok(netif)	((netif)->carrier)
+
+enum {
+	NETBK_DONT_COPY_SKB,
+	NETBK_DELAYED_COPY_SKB,
+	NETBK_ALWAYS_COPY_SKB,
+};
+
+extern int netbk_copy_skb_mode;
+
+/* Function pointers into netback accelerator plugin modules */
+struct netback_accel_hooks {
+	struct module *owner;
+	int  (*probe)(struct xenbus_device *dev);
+	int (*remove)(struct xenbus_device *dev);
+};
+
+/* Structure to track the state of a netback accelerator plugin */
+struct netback_accelerator {
+	struct list_head link;
+	int id;
+	char *eth_name;
+	atomic_t use_count;
+	struct netback_accel_hooks *hooks;
+};
+
+struct backend_info {
+	struct xenbus_device *dev;
+	netif_t *netif;
+	enum xenbus_state frontend_state;
+
+	/* State relating to the netback accelerator */
+	void *netback_accel_priv;
+	/* The accelerator that this backend is currently using */
+	struct netback_accelerator *accelerator;
+};
+
+#define NETBACK_ACCEL_VERSION 0x00010001
+
+/* 
+ * Connect an accelerator plugin module to netback.  Returns zero on
+ * success, < 0 on error, > 0 (with highest version number supported)
+ * if version mismatch.
+ */
+extern int netback_connect_accelerator(unsigned version,
+				       int id, const char *eth_name, 
+				       struct netback_accel_hooks *hooks);
+/* Disconnect a previously connected accelerator plugin module */
+extern void netback_disconnect_accelerator(int id, const char *eth_name);
+
+
+extern
+void netback_probe_accelerators(struct backend_info *be,
+				struct xenbus_device *dev);
+extern
+void netback_remove_accelerators(struct backend_info *be,
+				 struct xenbus_device *dev);
+extern
+void netif_accel_init(void);
+
+
+#define NET_TX_RING_SIZE __RING_SIZE((netif_tx_sring_t *)0, PAGE_SIZE)
+#define NET_RX_RING_SIZE __RING_SIZE((netif_rx_sring_t *)0, PAGE_SIZE)
+
+void netif_disconnect(netif_t *netif);
+
+netif_t *netif_alloc(domid_t domid, unsigned int handle);
+int netif_map(netif_t *netif, unsigned long tx_ring_ref,
+	      unsigned long rx_ring_ref, unsigned int evtchn);
+
+#define netif_get(_b) (atomic_inc(&(_b)->refcnt))
+#define netif_put(_b)						\
+	do {							\
+		if ( atomic_dec_and_test(&(_b)->refcnt) )	\
+			wake_up(&(_b)->waiting_to_free);	\
+	} while (0)
+
+void netif_xenbus_init(void);
+
+#define netif_schedulable(netif)				\
+	(netif_running((netif)->dev) && netback_carrier_ok(netif))
+
+void netif_schedule_work(netif_t *netif);
+void netif_deschedule_work(netif_t *netif);
+
+int netif_be_start_xmit(struct sk_buff *skb, struct net_device *dev);
+struct net_device_stats *netif_be_get_stats(struct net_device *dev);
+irqreturn_t netif_be_int(int irq, void *dev_id, struct pt_regs *regs);
+
+static inline int netbk_can_queue(struct net_device *dev)
+{
+	netif_t *netif = netdev_priv(dev);
+	return netif->can_queue;
+}
+
+static inline int netbk_can_sg(struct net_device *dev)
+{
+	netif_t *netif = netdev_priv(dev);
+	return netif->features & NETIF_F_SG;
+}
+
+#endif /* __NETIF__BACKEND__COMMON_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/netback/interface.c linux-2.6.18-xen-3.3.0/drivers/xen/netback/interface.c
--- linux-2.6.18.8/drivers/xen/netback/interface.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/netback/interface.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,336 @@
+/******************************************************************************
+ * arch/xen/drivers/netif/backend/interface.c
+ * 
+ * Network-device interface management.
+ * 
+ * Copyright (c) 2004-2005, Keir Fraser
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include "common.h"
+#include <linux/ethtool.h>
+#include <linux/rtnetlink.h>
+
+/*
+ * Module parameter 'queue_length':
+ * 
+ * Enables queuing in the network stack when a client has run out of receive
+ * descriptors. Although this feature can improve receive bandwidth by avoiding
+ * packet loss, it can also result in packets sitting in the 'tx_queue' for
+ * unbounded time. This is bad if those packets hold onto foreign resources.
+ * For example, consider a packet that holds onto resources belonging to the
+ * guest for which it is queued (e.g., packet received on vif1.0, destined for
+ * vif1.1 which is not activated in the guest): in this situation the guest
+ * will never be destroyed, unless vif1.1 is taken down. To avoid this, we
+ * run a timer (tx_queue_timeout) to drain the queue when the interface is
+ * blocked.
+ */
+static unsigned long netbk_queue_length = 32;
+module_param_named(queue_length, netbk_queue_length, ulong, 0);
+
+static void __netif_up(netif_t *netif)
+{
+	enable_irq(netif->irq);
+	netif_schedule_work(netif);
+}
+
+static void __netif_down(netif_t *netif)
+{
+	disable_irq(netif->irq);
+	netif_deschedule_work(netif);
+}
+
+static int net_open(struct net_device *dev)
+{
+	netif_t *netif = netdev_priv(dev);
+	if (netback_carrier_ok(netif)) {
+		__netif_up(netif);
+		netif_start_queue(dev);
+	}
+	return 0;
+}
+
+static int net_close(struct net_device *dev)
+{
+	netif_t *netif = netdev_priv(dev);
+	if (netback_carrier_ok(netif))
+		__netif_down(netif);
+	netif_stop_queue(dev);
+	return 0;
+}
+
+static int netbk_change_mtu(struct net_device *dev, int mtu)
+{
+	int max = netbk_can_sg(dev) ? 65535 - ETH_HLEN : ETH_DATA_LEN;
+
+	if (mtu > max)
+		return -EINVAL;
+	dev->mtu = mtu;
+	return 0;
+}
+
+static int netbk_set_sg(struct net_device *dev, u32 data)
+{
+	if (data) {
+		netif_t *netif = netdev_priv(dev);
+
+		if (!(netif->features & NETIF_F_SG))
+			return -ENOSYS;
+	}
+
+	return ethtool_op_set_sg(dev, data);
+}
+
+static int netbk_set_tso(struct net_device *dev, u32 data)
+{
+	if (data) {
+		netif_t *netif = netdev_priv(dev);
+
+		if (!(netif->features & NETIF_F_TSO))
+			return -ENOSYS;
+	}
+
+	return ethtool_op_set_tso(dev, data);
+}
+
+static struct ethtool_ops network_ethtool_ops =
+{
+	.get_tx_csum = ethtool_op_get_tx_csum,
+	.set_tx_csum = ethtool_op_set_tx_csum,
+	.get_sg = ethtool_op_get_sg,
+	.set_sg = netbk_set_sg,
+	.get_tso = ethtool_op_get_tso,
+	.set_tso = netbk_set_tso,
+	.get_link = ethtool_op_get_link,
+};
+
+netif_t *netif_alloc(domid_t domid, unsigned int handle)
+{
+	int err = 0;
+	struct net_device *dev;
+	netif_t *netif;
+	char name[IFNAMSIZ] = {};
+
+	snprintf(name, IFNAMSIZ - 1, "vif%u.%u", domid, handle);
+	dev = alloc_netdev(sizeof(netif_t), name, ether_setup);
+	if (dev == NULL) {
+		DPRINTK("Could not create netif: out of memory\n");
+		return ERR_PTR(-ENOMEM);
+	}
+
+	netif = netdev_priv(dev);
+	memset(netif, 0, sizeof(*netif));
+	netif->domid  = domid;
+	netif->handle = handle;
+	atomic_set(&netif->refcnt, 1);
+	init_waitqueue_head(&netif->waiting_to_free);
+	netif->dev = dev;
+
+	netback_carrier_off(netif);
+
+	netif->credit_bytes = netif->remaining_credit = ~0UL;
+	netif->credit_usec  = 0UL;
+	init_timer(&netif->credit_timeout);
+	/* Initialize 'expires' now: it's used to track the credit window. */
+	netif->credit_timeout.expires = jiffies;
+
+	init_timer(&netif->tx_queue_timeout);
+
+	dev->hard_start_xmit = netif_be_start_xmit;
+	dev->get_stats       = netif_be_get_stats;
+	dev->open            = net_open;
+	dev->stop            = net_close;
+	dev->change_mtu	     = netbk_change_mtu;
+	dev->features        = NETIF_F_IP_CSUM;
+
+	SET_ETHTOOL_OPS(dev, &network_ethtool_ops);
+
+	dev->tx_queue_len = netbk_queue_length;
+
+	/*
+	 * Initialise a dummy MAC address. We choose the numerically
+	 * largest non-broadcast address to prevent the address getting
+	 * stolen by an Ethernet bridge for STP purposes.
+	 * (FE:FF:FF:FF:FF:FF)
+	 */ 
+	memset(dev->dev_addr, 0xFF, ETH_ALEN);
+	dev->dev_addr[0] &= ~0x01;
+
+	rtnl_lock();
+	err = register_netdevice(dev);
+	rtnl_unlock();
+	if (err) {
+		DPRINTK("Could not register new net device %s: err=%d\n",
+			dev->name, err);
+		free_netdev(dev);
+		return ERR_PTR(err);
+	}
+
+	DPRINTK("Successfully created netif\n");
+	return netif;
+}
+
+static int map_frontend_pages(
+	netif_t *netif, grant_ref_t tx_ring_ref, grant_ref_t rx_ring_ref)
+{
+	struct gnttab_map_grant_ref op;
+
+	gnttab_set_map_op(&op, (unsigned long)netif->tx_comms_area->addr,
+			  GNTMAP_host_map, tx_ring_ref, netif->domid);
+    
+	if (HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, &op, 1))
+		BUG();
+
+	if (op.status) { 
+		DPRINTK(" Gnttab failure mapping tx_ring_ref!\n");
+		return op.status;
+	}
+
+	netif->tx_shmem_ref    = tx_ring_ref;
+	netif->tx_shmem_handle = op.handle;
+
+	gnttab_set_map_op(&op, (unsigned long)netif->rx_comms_area->addr,
+			  GNTMAP_host_map, rx_ring_ref, netif->domid);
+
+	if (HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, &op, 1))
+		BUG();
+
+	if (op.status) {
+		DPRINTK(" Gnttab failure mapping rx_ring_ref!\n");
+		return op.status;
+	}
+
+	netif->rx_shmem_ref    = rx_ring_ref;
+	netif->rx_shmem_handle = op.handle;
+
+	return 0;
+}
+
+static void unmap_frontend_pages(netif_t *netif)
+{
+	struct gnttab_unmap_grant_ref op;
+
+	gnttab_set_unmap_op(&op, (unsigned long)netif->tx_comms_area->addr,
+			    GNTMAP_host_map, netif->tx_shmem_handle);
+
+	if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1))
+		BUG();
+
+	gnttab_set_unmap_op(&op, (unsigned long)netif->rx_comms_area->addr,
+			    GNTMAP_host_map, netif->rx_shmem_handle);
+
+	if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1))
+		BUG();
+}
+
+int netif_map(netif_t *netif, unsigned long tx_ring_ref,
+	      unsigned long rx_ring_ref, unsigned int evtchn)
+{
+	int err = -ENOMEM;
+	netif_tx_sring_t *txs;
+	netif_rx_sring_t *rxs;
+
+	/* Already connected through? */
+	if (netif->irq)
+		return 0;
+
+	netif->tx_comms_area = alloc_vm_area(PAGE_SIZE);
+	if (netif->tx_comms_area == NULL)
+		return -ENOMEM;
+	netif->rx_comms_area = alloc_vm_area(PAGE_SIZE);
+	if (netif->rx_comms_area == NULL)
+		goto err_rx;
+
+	err = map_frontend_pages(netif, tx_ring_ref, rx_ring_ref);
+	if (err)
+		goto err_map;
+
+	err = bind_interdomain_evtchn_to_irqhandler(
+		netif->domid, evtchn, netif_be_int, 0,
+		netif->dev->name, netif);
+	if (err < 0)
+		goto err_hypervisor;
+	netif->irq = err;
+	disable_irq(netif->irq);
+
+	txs = (netif_tx_sring_t *)netif->tx_comms_area->addr;
+	BACK_RING_INIT(&netif->tx, txs, PAGE_SIZE);
+
+	rxs = (netif_rx_sring_t *)
+		((char *)netif->rx_comms_area->addr);
+	BACK_RING_INIT(&netif->rx, rxs, PAGE_SIZE);
+
+	netif->rx_req_cons_peek = 0;
+
+	netif_get(netif);
+
+	rtnl_lock();
+	netback_carrier_on(netif);
+	if (netif_running(netif->dev))
+		__netif_up(netif);
+	rtnl_unlock();
+
+	return 0;
+err_hypervisor:
+	unmap_frontend_pages(netif);
+err_map:
+	free_vm_area(netif->rx_comms_area);
+err_rx:
+	free_vm_area(netif->tx_comms_area);
+	return err;
+}
+
+void netif_disconnect(netif_t *netif)
+{
+	if (netback_carrier_ok(netif)) {
+		rtnl_lock();
+		netback_carrier_off(netif);
+		netif_carrier_off(netif->dev); /* discard queued packets */
+		if (netif_running(netif->dev))
+			__netif_down(netif);
+		rtnl_unlock();
+		netif_put(netif);
+	}
+
+	atomic_dec(&netif->refcnt);
+	wait_event(netif->waiting_to_free, atomic_read(&netif->refcnt) == 0);
+
+	del_timer_sync(&netif->credit_timeout);
+	del_timer_sync(&netif->tx_queue_timeout);
+
+	if (netif->irq)
+		unbind_from_irqhandler(netif->irq, netif);
+	
+	unregister_netdev(netif->dev);
+
+	if (netif->tx.sring) {
+		unmap_frontend_pages(netif);
+		free_vm_area(netif->tx_comms_area);
+		free_vm_area(netif->rx_comms_area);
+	}
+
+	free_netdev(netif->dev);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/netback/loopback.c linux-2.6.18-xen-3.3.0/drivers/xen/netback/loopback.c
--- linux-2.6.18.8/drivers/xen/netback/loopback.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/netback/loopback.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,324 @@
+/******************************************************************************
+ * netback/loopback.c
+ * 
+ * A two-interface loopback device to emulate a local netfront-netback
+ * connection. This ensures that local packet delivery looks identical
+ * to inter-domain delivery. Most importantly, packets delivered locally
+ * originating from other domains will get *copied* when they traverse this
+ * driver. This prevents unbounded delays in socket-buffer queues from
+ * causing the netback driver to "seize up".
+ * 
+ * This driver creates a symmetric pair of loopback interfaces with names
+ * vif0.0 and veth0. The intention is that 'vif0.0' is bound to an Ethernet
+ * bridge, just like a proper netback interface, while a local IP interface
+ * is configured on 'veth0'.
+ * 
+ * As with a real netback interface, vif0.0 is configured with a suitable
+ * dummy MAC address. No default is provided for veth0: a reasonable strategy
+ * is to transfer eth0's MAC address to veth0, and give eth0 a dummy address
+ * (to avoid confusing the Etherbridge).
+ * 
+ * Copyright (c) 2005 K A Fraser
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/module.h>
+#include <linux/netdevice.h>
+#include <linux/inetdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/skbuff.h>
+#include <linux/ethtool.h>
+#include <net/dst.h>
+#include <net/xfrm.h>		/* secpath_reset() */
+#include <asm/hypervisor.h>	/* is_initial_xendomain() */
+
+static int nloopbacks = -1;
+module_param(nloopbacks, int, 0);
+MODULE_PARM_DESC(nloopbacks, "Number of netback-loopback devices to create");
+
+struct net_private {
+	struct net_device *loopback_dev;
+	struct net_device_stats stats;
+};
+
+static int loopback_open(struct net_device *dev)
+{
+	struct net_private *np = netdev_priv(dev);
+	memset(&np->stats, 0, sizeof(np->stats));
+	netif_start_queue(dev);
+	return 0;
+}
+
+static int loopback_close(struct net_device *dev)
+{
+	netif_stop_queue(dev);
+	return 0;
+}
+
+#ifdef CONFIG_X86
+static int is_foreign(unsigned long pfn)
+{
+	/* NB. Play it safe for auto-translation mode. */
+	return (xen_feature(XENFEAT_auto_translated_physmap) ||
+		(phys_to_machine_mapping[pfn] & FOREIGN_FRAME_BIT));
+}
+#else
+/* How to detect a foreign mapping? Play it safe. */
+#define is_foreign(pfn)	(1)
+#endif
+
+static int skb_remove_foreign_references(struct sk_buff *skb)
+{
+	struct page *page;
+	unsigned long pfn;
+	int i, off;
+	char *vaddr;
+
+	BUG_ON(skb_shinfo(skb)->frag_list);
+
+	if (skb_cloned(skb) &&
+	    unlikely(pskb_expand_head(skb, 0, 0, GFP_ATOMIC)))
+		return 0;
+
+	for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
+		pfn = page_to_pfn(skb_shinfo(skb)->frags[i].page);
+		if (!is_foreign(pfn))
+			continue;
+		
+		page = alloc_page(GFP_ATOMIC | __GFP_NOWARN);
+		if (unlikely(!page))
+			return 0;
+
+		vaddr = kmap_skb_frag(&skb_shinfo(skb)->frags[i]);
+		off = skb_shinfo(skb)->frags[i].page_offset;
+		memcpy(page_address(page) + off,
+		       vaddr + off,
+		       skb_shinfo(skb)->frags[i].size);
+		kunmap_skb_frag(vaddr);
+
+		put_page(skb_shinfo(skb)->frags[i].page);
+		skb_shinfo(skb)->frags[i].page = page;
+	}
+
+	return 1;
+}
+
+static int loopback_start_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+	struct net_private *np = netdev_priv(dev);
+
+	if (!skb_remove_foreign_references(skb)) {
+		np->stats.tx_dropped++;
+		dev_kfree_skb(skb);
+		return 0;
+	}
+
+	dst_release(skb->dst);
+	skb->dst = NULL;
+
+	skb_orphan(skb);
+
+	np->stats.tx_bytes += skb->len;
+	np->stats.tx_packets++;
+
+	/* Switch to loopback context. */
+	dev = np->loopback_dev;
+	np  = netdev_priv(dev);
+
+	np->stats.rx_bytes += skb->len;
+	np->stats.rx_packets++;
+
+	if (skb->ip_summed == CHECKSUM_HW) {
+		/* Defer checksum calculation. */
+		skb->proto_csum_blank = 1;
+		/* Must be a local packet: assert its integrity. */
+		skb->proto_data_valid = 1;
+	}
+
+	skb->ip_summed = skb->proto_data_valid ?
+		CHECKSUM_UNNECESSARY : CHECKSUM_NONE;
+
+	skb->pkt_type = PACKET_HOST; /* overridden by eth_type_trans() */
+	skb->protocol = eth_type_trans(skb, dev);
+	skb->dev      = dev;
+	dev->last_rx  = jiffies;
+
+	/* Flush netfilter context: rx'ed skbuffs not expected to have any. */
+	nf_reset(skb);
+	secpath_reset(skb);
+
+	netif_rx(skb);
+
+	return 0;
+}
+
+static struct net_device_stats *loopback_get_stats(struct net_device *dev)
+{
+	struct net_private *np = netdev_priv(dev);
+	return &np->stats;
+}
+
+static struct ethtool_ops network_ethtool_ops =
+{
+	.get_tx_csum = ethtool_op_get_tx_csum,
+	.set_tx_csum = ethtool_op_set_tx_csum,
+	.get_sg = ethtool_op_get_sg,
+	.set_sg = ethtool_op_set_sg,
+	.get_tso = ethtool_op_get_tso,
+	.set_tso = ethtool_op_set_tso,
+	.get_link = ethtool_op_get_link,
+};
+
+/*
+ * Nothing to do here. Virtual interface is point-to-point and the
+ * physical interface is probably promiscuous anyway.
+ */
+static void loopback_set_multicast_list(struct net_device *dev)
+{
+}
+
+static void loopback_construct(struct net_device *dev, struct net_device *lo)
+{
+	struct net_private *np = netdev_priv(dev);
+
+	np->loopback_dev     = lo;
+
+	dev->open            = loopback_open;
+	dev->stop            = loopback_close;
+	dev->hard_start_xmit = loopback_start_xmit;
+	dev->get_stats       = loopback_get_stats;
+	dev->set_multicast_list = loopback_set_multicast_list;
+	dev->change_mtu	     = NULL; /* allow arbitrary mtu */
+
+	dev->tx_queue_len    = 0;
+
+	dev->features        = (NETIF_F_HIGHDMA |
+				NETIF_F_LLTX |
+				NETIF_F_TSO |
+				NETIF_F_SG |
+				NETIF_F_IP_CSUM);
+
+	SET_ETHTOOL_OPS(dev, &network_ethtool_ops);
+
+	/*
+	 * We do not set a jumbo MTU on the interface. Otherwise the network
+	 * stack will try to send large packets that will get dropped by the
+	 * Ethernet bridge (unless the physical Ethernet interface is
+	 * configured to transfer jumbo packets). If a larger MTU is desired
+	 * then the system administrator can specify it using the 'ifconfig'
+	 * command.
+	 */
+	/*dev->mtu             = 16*1024;*/
+}
+
+static int __init make_loopback(int i)
+{
+	struct net_device *dev1, *dev2;
+	char dev_name[IFNAMSIZ];
+	int err = -ENOMEM;
+
+	sprintf(dev_name, "vif0.%d", i);
+	dev1 = alloc_netdev(sizeof(struct net_private), dev_name, ether_setup);
+	if (!dev1)
+		return err;
+
+	sprintf(dev_name, "veth%d", i);
+	dev2 = alloc_netdev(sizeof(struct net_private), dev_name, ether_setup);
+	if (!dev2)
+		goto fail_netdev2;
+
+	loopback_construct(dev1, dev2);
+	loopback_construct(dev2, dev1);
+
+	/*
+	 * Initialise a dummy MAC address for the 'dummy backend' interface. We
+	 * choose the numerically largest non-broadcast address to prevent the
+	 * address getting stolen by an Ethernet bridge for STP purposes.
+	 */
+	memset(dev1->dev_addr, 0xFF, ETH_ALEN);
+	dev1->dev_addr[0] &= ~0x01;
+
+	if ((err = register_netdev(dev1)) != 0)
+		goto fail;
+
+	if ((err = register_netdev(dev2)) != 0) {
+		unregister_netdev(dev1);
+		goto fail;
+	}
+
+	return 0;
+
+ fail:
+	free_netdev(dev2);
+ fail_netdev2:
+	free_netdev(dev1);
+	return err;
+}
+
+static void __exit clean_loopback(int i)
+{
+	struct net_device *dev1, *dev2;
+	char dev_name[IFNAMSIZ];
+
+	sprintf(dev_name, "vif0.%d", i);
+	dev1 = dev_get_by_name(dev_name);
+	sprintf(dev_name, "veth%d", i);
+	dev2 = dev_get_by_name(dev_name);
+	if (dev1 && dev2) {
+		unregister_netdev(dev2);
+		unregister_netdev(dev1);
+		free_netdev(dev2);
+		free_netdev(dev1);
+	}
+}
+
+static int __init loopback_init(void)
+{
+	int i, err = 0;
+
+	if (nloopbacks == -1)
+		nloopbacks = is_initial_xendomain() ? 4 : 0;
+
+	for (i = 0; i < nloopbacks; i++)
+		if ((err = make_loopback(i)) != 0)
+			break;
+
+	return err;
+}
+
+module_init(loopback_init);
+
+static void __exit loopback_exit(void)
+{
+	int i;
+
+	for (i = nloopbacks; i-- > 0; )
+		clean_loopback(i);
+}
+
+module_exit(loopback_exit);
+
+MODULE_LICENSE("Dual BSD/GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/netback/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/netback/Makefile
--- linux-2.6.18.8/drivers/xen/netback/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/netback/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,5 @@
+obj-$(CONFIG_XEN_NETDEV_BACKEND) := netbk.o
+obj-$(CONFIG_XEN_NETDEV_LOOPBACK) += netloop.o
+
+netbk-y   := netback.o xenbus.o interface.o accel.o
+netloop-y := loopback.o
diff -rpuN linux-2.6.18.8/drivers/xen/netback/netback.c linux-2.6.18-xen-3.3.0/drivers/xen/netback/netback.c
--- linux-2.6.18.8/drivers/xen/netback/netback.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/netback/netback.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1614 @@
+/******************************************************************************
+ * drivers/xen/netback/netback.c
+ * 
+ * Back-end of the driver for virtual network devices. This portion of the
+ * driver exports a 'unified' network-device interface that can be accessed
+ * by any operating system that implements a compatible front end. A 
+ * reference front-end implementation can be found in:
+ *  drivers/xen/netfront/netfront.c
+ * 
+ * Copyright (c) 2002-2005, K A Fraser
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include "common.h"
+#include <xen/balloon.h>
+#include <xen/interface/memory.h>
+
+/*define NETBE_DEBUG_INTERRUPT*/
+
+/* extra field used in struct page */
+#define netif_page_index(pg) (*(long *)&(pg)->mapping)
+
+struct netbk_rx_meta {
+	skb_frag_t frag;
+	int id;
+	u8 copy:1;
+};
+
+struct netbk_tx_pending_inuse {
+	struct list_head list;
+	unsigned long alloc_time;
+};
+
+static void netif_idx_release(u16 pending_idx);
+static void netif_page_release(struct page *page);
+static void make_tx_response(netif_t *netif, 
+			     netif_tx_request_t *txp,
+			     s8       st);
+static netif_rx_response_t *make_rx_response(netif_t *netif, 
+					     u16      id, 
+					     s8       st,
+					     u16      offset,
+					     u16      size,
+					     u16      flags);
+
+static void net_tx_action(unsigned long unused);
+static DECLARE_TASKLET(net_tx_tasklet, net_tx_action, 0);
+
+static void net_rx_action(unsigned long unused);
+static DECLARE_TASKLET(net_rx_tasklet, net_rx_action, 0);
+
+static struct timer_list net_timer;
+static struct timer_list netbk_tx_pending_timer;
+
+#define MAX_PENDING_REQS 256
+
+static struct sk_buff_head rx_queue;
+
+static struct page **mmap_pages;
+static inline unsigned long idx_to_pfn(unsigned int idx)
+{
+	return page_to_pfn(mmap_pages[idx]);
+}
+
+static inline unsigned long idx_to_kaddr(unsigned int idx)
+{
+	return (unsigned long)pfn_to_kaddr(idx_to_pfn(idx));
+}
+
+#define PKT_PROT_LEN 64
+
+static struct pending_tx_info {
+	netif_tx_request_t req;
+	netif_t *netif;
+} pending_tx_info[MAX_PENDING_REQS];
+static u16 pending_ring[MAX_PENDING_REQS];
+typedef unsigned int PEND_RING_IDX;
+#define MASK_PEND_IDX(_i) ((_i)&(MAX_PENDING_REQS-1))
+static PEND_RING_IDX pending_prod, pending_cons;
+#define NR_PENDING_REQS (MAX_PENDING_REQS - pending_prod + pending_cons)
+
+/* Freed TX SKBs get batched on this ring before return to pending_ring. */
+static u16 dealloc_ring[MAX_PENDING_REQS];
+static PEND_RING_IDX dealloc_prod, dealloc_cons;
+
+/* Doubly-linked list of in-use pending entries. */
+static struct netbk_tx_pending_inuse pending_inuse[MAX_PENDING_REQS];
+static LIST_HEAD(pending_inuse_head);
+
+static struct sk_buff_head tx_queue;
+
+static grant_handle_t grant_tx_handle[MAX_PENDING_REQS];
+static gnttab_unmap_grant_ref_t tx_unmap_ops[MAX_PENDING_REQS];
+static gnttab_map_grant_ref_t tx_map_ops[MAX_PENDING_REQS];
+
+static struct list_head net_schedule_list;
+static spinlock_t net_schedule_list_lock;
+
+#define MAX_MFN_ALLOC 64
+static unsigned long mfn_list[MAX_MFN_ALLOC];
+static unsigned int alloc_index = 0;
+
+/* Setting this allows the safe use of this driver without netloop. */
+static int MODPARM_copy_skb = 1;
+module_param_named(copy_skb, MODPARM_copy_skb, bool, 0);
+MODULE_PARM_DESC(copy_skb, "Copy data received from netfront without netloop");
+
+int netbk_copy_skb_mode;
+
+static inline unsigned long alloc_mfn(void)
+{
+	BUG_ON(alloc_index == 0);
+	return mfn_list[--alloc_index];
+}
+
+static int check_mfn(int nr)
+{
+	struct xen_memory_reservation reservation = {
+		.extent_order = 0,
+		.domid        = DOMID_SELF
+	};
+	int rc;
+
+	if (likely(alloc_index >= nr))
+		return 0;
+
+	set_xen_guest_handle(reservation.extent_start, mfn_list + alloc_index);
+	reservation.nr_extents = MAX_MFN_ALLOC - alloc_index;
+	rc = HYPERVISOR_memory_op(XENMEM_increase_reservation, &reservation);
+	if (likely(rc > 0))
+		alloc_index += rc;
+
+	return alloc_index >= nr ? 0 : -ENOMEM;
+}
+
+static inline void maybe_schedule_tx_action(void)
+{
+	smp_mb();
+	if ((NR_PENDING_REQS < (MAX_PENDING_REQS/2)) &&
+	    !list_empty(&net_schedule_list))
+		tasklet_schedule(&net_tx_tasklet);
+}
+
+static struct sk_buff *netbk_copy_skb(struct sk_buff *skb)
+{
+	struct skb_shared_info *ninfo;
+	struct sk_buff *nskb;
+	unsigned long offset;
+	int ret;
+	int len;
+	int headlen;
+
+	BUG_ON(skb_shinfo(skb)->frag_list != NULL);
+
+	nskb = alloc_skb(SKB_MAX_HEAD(0), GFP_ATOMIC | __GFP_NOWARN);
+	if (unlikely(!nskb))
+		goto err;
+
+	skb_reserve(nskb, 16 + NET_IP_ALIGN);
+	headlen = nskb->end - nskb->data;
+	if (headlen > skb_headlen(skb))
+		headlen = skb_headlen(skb);
+	ret = skb_copy_bits(skb, 0, __skb_put(nskb, headlen), headlen);
+	BUG_ON(ret);
+
+	ninfo = skb_shinfo(nskb);
+	ninfo->gso_size = skb_shinfo(skb)->gso_size;
+	ninfo->gso_type = skb_shinfo(skb)->gso_type;
+
+	offset = headlen;
+	len = skb->len - headlen;
+
+	nskb->len = skb->len;
+	nskb->data_len = len;
+	nskb->truesize += len;
+
+	while (len) {
+		struct page *page;
+		int copy;
+		int zero;
+
+		if (unlikely(ninfo->nr_frags >= MAX_SKB_FRAGS)) {
+			dump_stack();
+			goto err_free;
+		}
+
+		copy = len >= PAGE_SIZE ? PAGE_SIZE : len;
+		zero = len >= PAGE_SIZE ? 0 : __GFP_ZERO;
+
+		page = alloc_page(GFP_ATOMIC | __GFP_NOWARN | zero);
+		if (unlikely(!page))
+			goto err_free;
+
+		ret = skb_copy_bits(skb, offset, page_address(page), copy);
+		BUG_ON(ret);
+
+		ninfo->frags[ninfo->nr_frags].page = page;
+		ninfo->frags[ninfo->nr_frags].page_offset = 0;
+		ninfo->frags[ninfo->nr_frags].size = copy;
+		ninfo->nr_frags++;
+
+		offset += copy;
+		len -= copy;
+	}
+
+	offset = nskb->data - skb->data;
+
+	nskb->h.raw = skb->h.raw + offset;
+	nskb->nh.raw = skb->nh.raw + offset;
+	nskb->mac.raw = skb->mac.raw + offset;
+
+	return nskb;
+
+ err_free:
+	kfree_skb(nskb);
+ err:
+	return NULL;
+}
+
+static inline int netbk_max_required_rx_slots(netif_t *netif)
+{
+	if (netif->features & (NETIF_F_SG|NETIF_F_TSO))
+		return MAX_SKB_FRAGS + 2; /* header + extra_info + frags */
+	return 1; /* all in one */
+}
+
+static inline int netbk_queue_full(netif_t *netif)
+{
+	RING_IDX peek   = netif->rx_req_cons_peek;
+	RING_IDX needed = netbk_max_required_rx_slots(netif);
+
+	return ((netif->rx.sring->req_prod - peek) < needed) ||
+	       ((netif->rx.rsp_prod_pvt + NET_RX_RING_SIZE - peek) < needed);
+}
+
+static void tx_queue_callback(unsigned long data)
+{
+	netif_t *netif = (netif_t *)data;
+	if (netif_schedulable(netif))
+		netif_wake_queue(netif->dev);
+}
+
+int netif_be_start_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+	netif_t *netif = netdev_priv(dev);
+
+	BUG_ON(skb->dev != dev);
+
+	/* Drop the packet if the target domain has no receive buffers. */
+	if (unlikely(!netif_schedulable(netif) || netbk_queue_full(netif)))
+		goto drop;
+
+	/*
+	 * Copy the packet here if it's destined for a flipping interface
+	 * but isn't flippable (e.g. extra references to data).
+	 * XXX For now we also copy skbuffs whose head crosses a page
+	 * boundary, because netbk_gop_skb can't handle them.
+	 */
+	if (!netif->copying_receiver ||
+	    ((skb_headlen(skb) + offset_in_page(skb->data)) >= PAGE_SIZE)) {
+		struct sk_buff *nskb = netbk_copy_skb(skb);
+		if ( unlikely(nskb == NULL) )
+			goto drop;
+		/* Copy only the header fields we use in this driver. */
+		nskb->dev = skb->dev;
+		nskb->ip_summed = skb->ip_summed;
+		nskb->proto_data_valid = skb->proto_data_valid;
+		dev_kfree_skb(skb);
+		skb = nskb;
+	}
+
+	netif->rx_req_cons_peek += skb_shinfo(skb)->nr_frags + 1 +
+				   !!skb_shinfo(skb)->gso_size;
+	netif_get(netif);
+
+	if (netbk_can_queue(dev) && netbk_queue_full(netif)) {
+		netif->rx.sring->req_event = netif->rx_req_cons_peek +
+			netbk_max_required_rx_slots(netif);
+		mb(); /* request notification /then/ check & stop the queue */
+		if (netbk_queue_full(netif)) {
+			netif_stop_queue(dev);
+			/*
+			 * Schedule 500ms timeout to restart the queue, thus
+			 * ensuring that an inactive queue will be drained.
+			 * Packets will be immediately be dropped until more
+			 * receive buffers become available (see
+			 * netbk_queue_full() check above).
+			 */
+			netif->tx_queue_timeout.data = (unsigned long)netif;
+			netif->tx_queue_timeout.function = tx_queue_callback;
+			__mod_timer(&netif->tx_queue_timeout, jiffies + HZ/2);
+		}
+	}
+
+	skb_queue_tail(&rx_queue, skb);
+	tasklet_schedule(&net_rx_tasklet);
+
+	return 0;
+
+ drop:
+	netif->stats.tx_dropped++;
+	dev_kfree_skb(skb);
+	return 0;
+}
+
+#if 0
+static void xen_network_done_notify(void)
+{
+	static struct net_device *eth0_dev = NULL;
+	if (unlikely(eth0_dev == NULL))
+		eth0_dev = __dev_get_by_name("eth0");
+	netif_rx_schedule(eth0_dev);
+}
+/* 
+ * Add following to poll() function in NAPI driver (Tigon3 is example):
+ *  if ( xen_network_done() )
+ *      tg3_enable_ints(tp);
+ */
+int xen_network_done(void)
+{
+	return skb_queue_empty(&rx_queue);
+}
+#endif
+
+struct netrx_pending_operations {
+	unsigned trans_prod, trans_cons;
+	unsigned mmu_prod, mmu_mcl;
+	unsigned mcl_prod, mcl_cons;
+	unsigned copy_prod, copy_cons;
+	unsigned meta_prod, meta_cons;
+	mmu_update_t *mmu;
+	gnttab_transfer_t *trans;
+	gnttab_copy_t *copy;
+	multicall_entry_t *mcl;
+	struct netbk_rx_meta *meta;
+};
+
+/* Set up the grant operations for this fragment.  If it's a flipping
+   interface, we also set up the unmap request from here. */
+static u16 netbk_gop_frag(netif_t *netif, struct netbk_rx_meta *meta,
+			  int i, struct netrx_pending_operations *npo,
+			  struct page *page, unsigned long size,
+			  unsigned long offset)
+{
+	mmu_update_t *mmu;
+	gnttab_transfer_t *gop;
+	gnttab_copy_t *copy_gop;
+	multicall_entry_t *mcl;
+	netif_rx_request_t *req;
+	unsigned long old_mfn, new_mfn;
+
+	old_mfn = virt_to_mfn(page_address(page));
+
+	req = RING_GET_REQUEST(&netif->rx, netif->rx.req_cons + i);
+	if (netif->copying_receiver) {
+		/* The fragment needs to be copied rather than
+		   flipped. */
+		meta->copy = 1;
+		copy_gop = npo->copy + npo->copy_prod++;
+		copy_gop->flags = GNTCOPY_dest_gref;
+		if (PageForeign(page)) {
+			struct pending_tx_info *src_pend =
+				&pending_tx_info[netif_page_index(page)];
+			copy_gop->source.domid = src_pend->netif->domid;
+			copy_gop->source.u.ref = src_pend->req.gref;
+			copy_gop->flags |= GNTCOPY_source_gref;
+		} else {
+			copy_gop->source.domid = DOMID_SELF;
+			copy_gop->source.u.gmfn = old_mfn;
+		}
+		copy_gop->source.offset = offset;
+		copy_gop->dest.domid = netif->domid;
+		copy_gop->dest.offset = 0;
+		copy_gop->dest.u.ref = req->gref;
+		copy_gop->len = size;
+	} else {
+		meta->copy = 0;
+		if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+			new_mfn = alloc_mfn();
+
+			/*
+			 * Set the new P2M table entry before
+			 * reassigning the old data page. Heed the
+			 * comment in pgtable-2level.h:pte_page(). :-)
+			 */
+			set_phys_to_machine(page_to_pfn(page), new_mfn);
+
+			mcl = npo->mcl + npo->mcl_prod++;
+			MULTI_update_va_mapping(mcl,
+					     (unsigned long)page_address(page),
+					     pfn_pte_ma(new_mfn, PAGE_KERNEL),
+					     0);
+
+			mmu = npo->mmu + npo->mmu_prod++;
+			mmu->ptr = ((maddr_t)new_mfn << PAGE_SHIFT) |
+				MMU_MACHPHYS_UPDATE;
+			mmu->val = page_to_pfn(page);
+		}
+
+		gop = npo->trans + npo->trans_prod++;
+		gop->mfn = old_mfn;
+		gop->domid = netif->domid;
+		gop->ref = req->gref;
+	}
+	return req->id;
+}
+
+static void netbk_gop_skb(struct sk_buff *skb,
+			  struct netrx_pending_operations *npo)
+{
+	netif_t *netif = netdev_priv(skb->dev);
+	int nr_frags = skb_shinfo(skb)->nr_frags;
+	int i;
+	int extra;
+	struct netbk_rx_meta *head_meta, *meta;
+
+	head_meta = npo->meta + npo->meta_prod++;
+	head_meta->frag.page_offset = skb_shinfo(skb)->gso_type;
+	head_meta->frag.size = skb_shinfo(skb)->gso_size;
+	extra = !!head_meta->frag.size + 1;
+
+	for (i = 0; i < nr_frags; i++) {
+		meta = npo->meta + npo->meta_prod++;
+		meta->frag = skb_shinfo(skb)->frags[i];
+		meta->id = netbk_gop_frag(netif, meta, i + extra, npo,
+					  meta->frag.page,
+					  meta->frag.size,
+					  meta->frag.page_offset);
+	}
+
+	/*
+	 * This must occur at the end to ensure that we don't trash skb_shinfo
+	 * until we're done. We know that the head doesn't cross a page
+	 * boundary because such packets get copied in netif_be_start_xmit.
+	 */
+	head_meta->id = netbk_gop_frag(netif, head_meta, 0, npo,
+				       virt_to_page(skb->data),
+				       skb_headlen(skb),
+				       offset_in_page(skb->data));
+
+	netif->rx.req_cons += nr_frags + extra;
+}
+
+static inline void netbk_free_pages(int nr_frags, struct netbk_rx_meta *meta)
+{
+	int i;
+
+	for (i = 0; i < nr_frags; i++)
+		put_page(meta[i].frag.page);
+}
+
+/* This is a twin to netbk_gop_skb.  Assume that netbk_gop_skb was
+   used to set up the operations on the top of
+   netrx_pending_operations, which have since been done.  Check that
+   they didn't give any errors and advance over them. */
+static int netbk_check_gop(int nr_frags, domid_t domid,
+			   struct netrx_pending_operations *npo)
+{
+	multicall_entry_t *mcl;
+	gnttab_transfer_t *gop;
+	gnttab_copy_t     *copy_op;
+	int status = NETIF_RSP_OKAY;
+	int i;
+
+	for (i = 0; i <= nr_frags; i++) {
+		if (npo->meta[npo->meta_cons + i].copy) {
+			copy_op = npo->copy + npo->copy_cons++;
+			if (copy_op->status != GNTST_okay) {
+				DPRINTK("Bad status %d from copy to DOM%d.\n",
+					copy_op->status, domid);
+				status = NETIF_RSP_ERROR;
+			}
+		} else {
+			if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+				mcl = npo->mcl + npo->mcl_cons++;
+				/* The update_va_mapping() must not fail. */
+				BUG_ON(mcl->result != 0);
+			}
+
+			gop = npo->trans + npo->trans_cons++;
+			/* Check the reassignment error code. */
+			if (gop->status != 0) {
+				DPRINTK("Bad status %d from grant transfer to DOM%u\n",
+					gop->status, domid);
+				/*
+				 * Page no longer belongs to us unless
+				 * GNTST_bad_page, but that should be
+				 * a fatal error anyway.
+				 */
+				BUG_ON(gop->status == GNTST_bad_page);
+				status = NETIF_RSP_ERROR;
+			}
+		}
+	}
+
+	return status;
+}
+
+static void netbk_add_frag_responses(netif_t *netif, int status,
+				     struct netbk_rx_meta *meta, int nr_frags)
+{
+	int i;
+	unsigned long offset;
+
+	for (i = 0; i < nr_frags; i++) {
+		int id = meta[i].id;
+		int flags = (i == nr_frags - 1) ? 0 : NETRXF_more_data;
+
+		if (meta[i].copy)
+			offset = 0;
+		else
+			offset = meta[i].frag.page_offset;
+		make_rx_response(netif, id, status, offset,
+				 meta[i].frag.size, flags);
+	}
+}
+
+static void net_rx_action(unsigned long unused)
+{
+	netif_t *netif = NULL;
+	s8 status;
+	u16 id, irq, flags;
+	netif_rx_response_t *resp;
+	multicall_entry_t *mcl;
+	struct sk_buff_head rxq;
+	struct sk_buff *skb;
+	int notify_nr = 0;
+	int ret;
+	int nr_frags;
+	int count;
+	unsigned long offset;
+
+	/*
+	 * Putting hundreds of bytes on the stack is considered rude.
+	 * Static works because a tasklet can only be on one CPU at any time.
+	 */
+	static multicall_entry_t rx_mcl[NET_RX_RING_SIZE+3];
+	static mmu_update_t rx_mmu[NET_RX_RING_SIZE];
+	static gnttab_transfer_t grant_trans_op[NET_RX_RING_SIZE];
+	static gnttab_copy_t grant_copy_op[NET_RX_RING_SIZE];
+	static unsigned char rx_notify[NR_IRQS];
+	static u16 notify_list[NET_RX_RING_SIZE];
+	static struct netbk_rx_meta meta[NET_RX_RING_SIZE];
+
+	struct netrx_pending_operations npo = {
+		mmu: rx_mmu,
+		trans: grant_trans_op,
+		copy: grant_copy_op,
+		mcl: rx_mcl,
+		meta: meta};
+
+	skb_queue_head_init(&rxq);
+
+	count = 0;
+
+	while ((skb = skb_dequeue(&rx_queue)) != NULL) {
+		nr_frags = skb_shinfo(skb)->nr_frags;
+		*(int *)skb->cb = nr_frags;
+
+		if (!xen_feature(XENFEAT_auto_translated_physmap) &&
+		    !((netif_t *)netdev_priv(skb->dev))->copying_receiver &&
+		    check_mfn(nr_frags + 1)) {
+			/* Memory squeeze? Back off for an arbitrary while. */
+			if ( net_ratelimit() )
+				WPRINTK("Memory squeeze in netback "
+					"driver.\n");
+			mod_timer(&net_timer, jiffies + HZ);
+			skb_queue_head(&rx_queue, skb);
+			break;
+		}
+
+		netbk_gop_skb(skb, &npo);
+
+		count += nr_frags + 1;
+
+		__skb_queue_tail(&rxq, skb);
+
+		/* Filled the batch queue? */
+		if (count + MAX_SKB_FRAGS >= NET_RX_RING_SIZE)
+			break;
+	}
+
+	BUG_ON(npo.meta_prod > ARRAY_SIZE(meta));
+
+	npo.mmu_mcl = npo.mcl_prod;
+	if (npo.mcl_prod) {
+		BUG_ON(xen_feature(XENFEAT_auto_translated_physmap));
+		BUG_ON(npo.mmu_prod > ARRAY_SIZE(rx_mmu));
+		mcl = npo.mcl + npo.mcl_prod++;
+
+		BUG_ON(mcl[-1].op != __HYPERVISOR_update_va_mapping);
+		mcl[-1].args[MULTI_UVMFLAGS_INDEX] = UVMF_TLB_FLUSH|UVMF_ALL;
+
+		mcl->op = __HYPERVISOR_mmu_update;
+		mcl->args[0] = (unsigned long)rx_mmu;
+		mcl->args[1] = npo.mmu_prod;
+		mcl->args[2] = 0;
+		mcl->args[3] = DOMID_SELF;
+	}
+
+	if (npo.trans_prod) {
+		BUG_ON(npo.trans_prod > ARRAY_SIZE(grant_trans_op));
+		mcl = npo.mcl + npo.mcl_prod++;
+		mcl->op = __HYPERVISOR_grant_table_op;
+		mcl->args[0] = GNTTABOP_transfer;
+		mcl->args[1] = (unsigned long)grant_trans_op;
+		mcl->args[2] = npo.trans_prod;
+	}
+
+	if (npo.copy_prod) {
+		BUG_ON(npo.copy_prod > ARRAY_SIZE(grant_copy_op));
+		mcl = npo.mcl + npo.mcl_prod++;
+		mcl->op = __HYPERVISOR_grant_table_op;
+		mcl->args[0] = GNTTABOP_copy;
+		mcl->args[1] = (unsigned long)grant_copy_op;
+		mcl->args[2] = npo.copy_prod;
+	}
+
+	/* Nothing to do? */
+	if (!npo.mcl_prod)
+		return;
+
+	BUG_ON(npo.mcl_prod > ARRAY_SIZE(rx_mcl));
+
+	ret = HYPERVISOR_multicall(npo.mcl, npo.mcl_prod);
+	BUG_ON(ret != 0);
+	/* The mmu_machphys_update() must not fail. */
+	BUG_ON(npo.mmu_mcl && npo.mcl[npo.mmu_mcl].result != 0);
+
+	while ((skb = __skb_dequeue(&rxq)) != NULL) {
+		nr_frags = *(int *)skb->cb;
+
+		netif = netdev_priv(skb->dev);
+		/* We can't rely on skb_release_data to release the
+		   pages used by fragments for us, since it tries to
+		   touch the pages in the fraglist.  If we're in
+		   flipping mode, that doesn't work.  In copying mode,
+		   we still have access to all of the pages, and so
+		   it's safe to let release_data deal with it. */
+		/* (Freeing the fragments is safe since we copy
+		   non-linear skbs destined for flipping interfaces) */
+		if (!netif->copying_receiver) {
+			atomic_set(&(skb_shinfo(skb)->dataref), 1);
+			skb_shinfo(skb)->frag_list = NULL;
+			skb_shinfo(skb)->nr_frags = 0;
+			netbk_free_pages(nr_frags, meta + npo.meta_cons + 1);
+		}
+
+		netif->stats.tx_bytes += skb->len;
+		netif->stats.tx_packets++;
+
+		status = netbk_check_gop(nr_frags, netif->domid, &npo);
+
+		id = meta[npo.meta_cons].id;
+		flags = nr_frags ? NETRXF_more_data : 0;
+
+		if (skb->ip_summed == CHECKSUM_HW) /* local packet? */
+			flags |= NETRXF_csum_blank | NETRXF_data_validated;
+		else if (skb->proto_data_valid) /* remote but checksummed? */
+			flags |= NETRXF_data_validated;
+
+		if (meta[npo.meta_cons].copy)
+			offset = 0;
+		else
+			offset = offset_in_page(skb->data);
+		resp = make_rx_response(netif, id, status, offset,
+					skb_headlen(skb), flags);
+
+		if (meta[npo.meta_cons].frag.size) {
+			struct netif_extra_info *gso =
+				(struct netif_extra_info *)
+				RING_GET_RESPONSE(&netif->rx,
+						  netif->rx.rsp_prod_pvt++);
+
+			resp->flags |= NETRXF_extra_info;
+
+			gso->u.gso.size = meta[npo.meta_cons].frag.size;
+			gso->u.gso.type = XEN_NETIF_GSO_TYPE_TCPV4;
+			gso->u.gso.pad = 0;
+			gso->u.gso.features = 0;
+
+			gso->type = XEN_NETIF_EXTRA_TYPE_GSO;
+			gso->flags = 0;
+		}
+
+		netbk_add_frag_responses(netif, status,
+					 meta + npo.meta_cons + 1,
+					 nr_frags);
+
+		RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&netif->rx, ret);
+		irq = netif->irq;
+		if (ret && !rx_notify[irq]) {
+			rx_notify[irq] = 1;
+			notify_list[notify_nr++] = irq;
+		}
+
+		if (netif_queue_stopped(netif->dev) &&
+		    netif_schedulable(netif) &&
+		    !netbk_queue_full(netif))
+			netif_wake_queue(netif->dev);
+
+		netif_put(netif);
+		dev_kfree_skb(skb);
+		npo.meta_cons += nr_frags + 1;
+	}
+
+	while (notify_nr != 0) {
+		irq = notify_list[--notify_nr];
+		rx_notify[irq] = 0;
+		notify_remote_via_irq(irq);
+	}
+
+	/* More work to do? */
+	if (!skb_queue_empty(&rx_queue) && !timer_pending(&net_timer))
+		tasklet_schedule(&net_rx_tasklet);
+#if 0
+	else
+		xen_network_done_notify();
+#endif
+}
+
+static void net_alarm(unsigned long unused)
+{
+	tasklet_schedule(&net_rx_tasklet);
+}
+
+static void netbk_tx_pending_timeout(unsigned long unused)
+{
+	tasklet_schedule(&net_tx_tasklet);
+}
+
+struct net_device_stats *netif_be_get_stats(struct net_device *dev)
+{
+	netif_t *netif = netdev_priv(dev);
+	return &netif->stats;
+}
+
+static int __on_net_schedule_list(netif_t *netif)
+{
+	return netif->list.next != NULL;
+}
+
+static void remove_from_net_schedule_list(netif_t *netif)
+{
+	spin_lock_irq(&net_schedule_list_lock);
+	if (likely(__on_net_schedule_list(netif))) {
+		list_del(&netif->list);
+		netif->list.next = NULL;
+		netif_put(netif);
+	}
+	spin_unlock_irq(&net_schedule_list_lock);
+}
+
+static void add_to_net_schedule_list_tail(netif_t *netif)
+{
+	if (__on_net_schedule_list(netif))
+		return;
+
+	spin_lock_irq(&net_schedule_list_lock);
+	if (!__on_net_schedule_list(netif) &&
+	    likely(netif_schedulable(netif))) {
+		list_add_tail(&netif->list, &net_schedule_list);
+		netif_get(netif);
+	}
+	spin_unlock_irq(&net_schedule_list_lock);
+}
+
+/*
+ * Note on CONFIG_XEN_NETDEV_PIPELINED_TRANSMITTER:
+ * If this driver is pipelining transmit requests then we can be very
+ * aggressive in avoiding new-packet notifications -- frontend only needs to
+ * send a notification if there are no outstanding unreceived responses.
+ * If we may be buffer transmit buffers for any reason then we must be rather
+ * more conservative and treat this as the final check for pending work.
+ */
+void netif_schedule_work(netif_t *netif)
+{
+	int more_to_do;
+
+#ifdef CONFIG_XEN_NETDEV_PIPELINED_TRANSMITTER
+	more_to_do = RING_HAS_UNCONSUMED_REQUESTS(&netif->tx);
+#else
+	RING_FINAL_CHECK_FOR_REQUESTS(&netif->tx, more_to_do);
+#endif
+
+	if (more_to_do) {
+		add_to_net_schedule_list_tail(netif);
+		maybe_schedule_tx_action();
+	}
+}
+
+void netif_deschedule_work(netif_t *netif)
+{
+	remove_from_net_schedule_list(netif);
+}
+
+
+static void tx_add_credit(netif_t *netif)
+{
+	unsigned long max_burst, max_credit;
+
+	/*
+	 * Allow a burst big enough to transmit a jumbo packet of up to 128kB.
+	 * Otherwise the interface can seize up due to insufficient credit.
+	 */
+	max_burst = RING_GET_REQUEST(&netif->tx, netif->tx.req_cons)->size;
+	max_burst = min(max_burst, 131072UL);
+	max_burst = max(max_burst, netif->credit_bytes);
+
+	/* Take care that adding a new chunk of credit doesn't wrap to zero. */
+	max_credit = netif->remaining_credit + netif->credit_bytes;
+	if (max_credit < netif->remaining_credit)
+		max_credit = ULONG_MAX; /* wrapped: clamp to ULONG_MAX */
+
+	netif->remaining_credit = min(max_credit, max_burst);
+}
+
+static void tx_credit_callback(unsigned long data)
+{
+	netif_t *netif = (netif_t *)data;
+	tx_add_credit(netif);
+	netif_schedule_work(netif);
+}
+
+static inline int copy_pending_req(PEND_RING_IDX pending_idx)
+{
+	return gnttab_copy_grant_page(grant_tx_handle[pending_idx],
+				      &mmap_pages[pending_idx]);
+}
+
+inline static void net_tx_action_dealloc(void)
+{
+	struct netbk_tx_pending_inuse *inuse, *n;
+	gnttab_unmap_grant_ref_t *gop;
+	u16 pending_idx;
+	PEND_RING_IDX dc, dp;
+	netif_t *netif;
+	int ret;
+	LIST_HEAD(list);
+
+	dc = dealloc_cons;
+	gop = tx_unmap_ops;
+
+	/*
+	 * Free up any grants we have finished using
+	 */
+	do {
+		dp = dealloc_prod;
+
+		/* Ensure we see all indices enqueued by netif_idx_release(). */
+		smp_rmb();
+
+		while (dc != dp) {
+			unsigned long pfn;
+
+			pending_idx = dealloc_ring[MASK_PEND_IDX(dc++)];
+			list_move_tail(&pending_inuse[pending_idx].list, &list);
+
+			pfn = idx_to_pfn(pending_idx);
+			/* Already unmapped? */
+			if (!phys_to_machine_mapping_valid(pfn))
+				continue;
+
+			gnttab_set_unmap_op(gop, idx_to_kaddr(pending_idx),
+					    GNTMAP_host_map,
+					    grant_tx_handle[pending_idx]);
+			gop++;
+		}
+
+		if (netbk_copy_skb_mode != NETBK_DELAYED_COPY_SKB ||
+		    list_empty(&pending_inuse_head))
+			break;
+
+		/* Copy any entries that have been pending for too long. */
+		list_for_each_entry_safe(inuse, n, &pending_inuse_head, list) {
+			if (time_after(inuse->alloc_time + HZ / 2, jiffies))
+				break;
+
+			switch (copy_pending_req(inuse - pending_inuse)) {
+			case 0:
+				list_move_tail(&inuse->list, &list);
+				continue;
+			case -EBUSY:
+				list_del_init(&inuse->list);
+				continue;
+			case -ENOENT:
+				continue;
+			}
+
+			break;
+		}
+	} while (dp != dealloc_prod);
+
+	dealloc_cons = dc;
+
+	ret = HYPERVISOR_grant_table_op(
+		GNTTABOP_unmap_grant_ref, tx_unmap_ops, gop - tx_unmap_ops);
+	BUG_ON(ret);
+
+	list_for_each_entry_safe(inuse, n, &list, list) {
+		pending_idx = inuse - pending_inuse;
+
+		netif = pending_tx_info[pending_idx].netif;
+
+		make_tx_response(netif, &pending_tx_info[pending_idx].req, 
+				 NETIF_RSP_OKAY);
+
+		/* Ready for next use. */
+		gnttab_reset_grant_page(mmap_pages[pending_idx]);
+
+		pending_ring[MASK_PEND_IDX(pending_prod++)] = pending_idx;
+
+		netif_put(netif);
+
+		list_del_init(&inuse->list);
+	}
+}
+
+static void netbk_tx_err(netif_t *netif, netif_tx_request_t *txp, RING_IDX end)
+{
+	RING_IDX cons = netif->tx.req_cons;
+
+	do {
+		make_tx_response(netif, txp, NETIF_RSP_ERROR);
+		if (cons >= end)
+			break;
+		txp = RING_GET_REQUEST(&netif->tx, cons++);
+	} while (1);
+	netif->tx.req_cons = cons;
+	netif_schedule_work(netif);
+	netif_put(netif);
+}
+
+static int netbk_count_requests(netif_t *netif, netif_tx_request_t *first,
+				netif_tx_request_t *txp, int work_to_do)
+{
+	RING_IDX cons = netif->tx.req_cons;
+	int frags = 0;
+
+	if (!(first->flags & NETTXF_more_data))
+		return 0;
+
+	do {
+		if (frags >= work_to_do) {
+			DPRINTK("Need more frags\n");
+			return -frags;
+		}
+
+		if (unlikely(frags >= MAX_SKB_FRAGS)) {
+			DPRINTK("Too many frags\n");
+			return -frags;
+		}
+
+		memcpy(txp, RING_GET_REQUEST(&netif->tx, cons + frags),
+		       sizeof(*txp));
+		if (txp->size > first->size) {
+			DPRINTK("Frags galore\n");
+			return -frags;
+		}
+
+		first->size -= txp->size;
+		frags++;
+
+		if (unlikely((txp->offset + txp->size) > PAGE_SIZE)) {
+			DPRINTK("txp->offset: %x, size: %u\n",
+				txp->offset, txp->size);
+			return -frags;
+		}
+	} while ((txp++)->flags & NETTXF_more_data);
+
+	return frags;
+}
+
+static gnttab_map_grant_ref_t *netbk_get_requests(netif_t *netif,
+						  struct sk_buff *skb,
+						  netif_tx_request_t *txp,
+						  gnttab_map_grant_ref_t *mop)
+{
+	struct skb_shared_info *shinfo = skb_shinfo(skb);
+	skb_frag_t *frags = shinfo->frags;
+	unsigned long pending_idx = *((u16 *)skb->data);
+	int i, start;
+
+	/* Skip first skb fragment if it is on same page as header fragment. */
+	start = ((unsigned long)shinfo->frags[0].page == pending_idx);
+
+	for (i = start; i < shinfo->nr_frags; i++, txp++) {
+		pending_idx = pending_ring[MASK_PEND_IDX(pending_cons++)];
+
+		gnttab_set_map_op(mop++, idx_to_kaddr(pending_idx),
+				  GNTMAP_host_map | GNTMAP_readonly,
+				  txp->gref, netif->domid);
+
+		memcpy(&pending_tx_info[pending_idx].req, txp, sizeof(*txp));
+		netif_get(netif);
+		pending_tx_info[pending_idx].netif = netif;
+		frags[i].page = (void *)pending_idx;
+	}
+
+	return mop;
+}
+
+static int netbk_tx_check_mop(struct sk_buff *skb,
+			       gnttab_map_grant_ref_t **mopp)
+{
+	gnttab_map_grant_ref_t *mop = *mopp;
+	int pending_idx = *((u16 *)skb->data);
+	netif_t *netif = pending_tx_info[pending_idx].netif;
+	netif_tx_request_t *txp;
+	struct skb_shared_info *shinfo = skb_shinfo(skb);
+	int nr_frags = shinfo->nr_frags;
+	int i, err, start;
+
+	/* Check status of header. */
+	err = mop->status;
+	if (unlikely(err)) {
+		txp = &pending_tx_info[pending_idx].req;
+		make_tx_response(netif, txp, NETIF_RSP_ERROR);
+		pending_ring[MASK_PEND_IDX(pending_prod++)] = pending_idx;
+		netif_put(netif);
+	} else {
+		set_phys_to_machine(
+			__pa(idx_to_kaddr(pending_idx)) >> PAGE_SHIFT,
+			FOREIGN_FRAME(mop->dev_bus_addr >> PAGE_SHIFT));
+		grant_tx_handle[pending_idx] = mop->handle;
+	}
+
+	/* Skip first skb fragment if it is on same page as header fragment. */
+	start = ((unsigned long)shinfo->frags[0].page == pending_idx);
+
+	for (i = start; i < nr_frags; i++) {
+		int j, newerr;
+
+		pending_idx = (unsigned long)shinfo->frags[i].page;
+
+		/* Check error status: if okay then remember grant handle. */
+		newerr = (++mop)->status;
+		if (likely(!newerr)) {
+			set_phys_to_machine(
+				__pa(idx_to_kaddr(pending_idx))>>PAGE_SHIFT,
+				FOREIGN_FRAME(mop->dev_bus_addr>>PAGE_SHIFT));
+			grant_tx_handle[pending_idx] = mop->handle;
+			/* Had a previous error? Invalidate this fragment. */
+			if (unlikely(err))
+				netif_idx_release(pending_idx);
+			continue;
+		}
+
+		/* Error on this fragment: respond to client with an error. */
+		txp = &pending_tx_info[pending_idx].req;
+		make_tx_response(netif, txp, NETIF_RSP_ERROR);
+		pending_ring[MASK_PEND_IDX(pending_prod++)] = pending_idx;
+		netif_put(netif);
+
+		/* Not the first error? Preceding frags already invalidated. */
+		if (err)
+			continue;
+
+		/* First error: invalidate header and preceding fragments. */
+		pending_idx = *((u16 *)skb->data);
+		netif_idx_release(pending_idx);
+		for (j = start; j < i; j++) {
+			pending_idx = (unsigned long)shinfo->frags[i].page;
+			netif_idx_release(pending_idx);
+		}
+
+		/* Remember the error: invalidate all subsequent fragments. */
+		err = newerr;
+	}
+
+	*mopp = mop + 1;
+	return err;
+}
+
+static void netbk_fill_frags(struct sk_buff *skb)
+{
+	struct skb_shared_info *shinfo = skb_shinfo(skb);
+	int nr_frags = shinfo->nr_frags;
+	int i;
+
+	for (i = 0; i < nr_frags; i++) {
+		skb_frag_t *frag = shinfo->frags + i;
+		netif_tx_request_t *txp;
+		unsigned long pending_idx;
+
+		pending_idx = (unsigned long)frag->page;
+
+		pending_inuse[pending_idx].alloc_time = jiffies;
+		list_add_tail(&pending_inuse[pending_idx].list,
+			      &pending_inuse_head);
+
+		txp = &pending_tx_info[pending_idx].req;
+		frag->page = virt_to_page(idx_to_kaddr(pending_idx));
+		frag->size = txp->size;
+		frag->page_offset = txp->offset;
+
+		skb->len += txp->size;
+		skb->data_len += txp->size;
+		skb->truesize += txp->size;
+	}
+}
+
+int netbk_get_extras(netif_t *netif, struct netif_extra_info *extras,
+		     int work_to_do)
+{
+	struct netif_extra_info extra;
+	RING_IDX cons = netif->tx.req_cons;
+
+	do {
+		if (unlikely(work_to_do-- <= 0)) {
+			DPRINTK("Missing extra info\n");
+			return -EBADR;
+		}
+
+		memcpy(&extra, RING_GET_REQUEST(&netif->tx, cons),
+		       sizeof(extra));
+		if (unlikely(!extra.type ||
+			     extra.type >= XEN_NETIF_EXTRA_TYPE_MAX)) {
+			netif->tx.req_cons = ++cons;
+			DPRINTK("Invalid extra type: %d\n", extra.type);
+			return -EINVAL;
+		}
+
+		memcpy(&extras[extra.type - 1], &extra, sizeof(extra));
+		netif->tx.req_cons = ++cons;
+	} while (extra.flags & XEN_NETIF_EXTRA_FLAG_MORE);
+
+	return work_to_do;
+}
+
+static int netbk_set_skb_gso(struct sk_buff *skb, struct netif_extra_info *gso)
+{
+	if (!gso->u.gso.size) {
+		DPRINTK("GSO size must not be zero.\n");
+		return -EINVAL;
+	}
+
+	/* Currently only TCPv4 S.O. is supported. */
+	if (gso->u.gso.type != XEN_NETIF_GSO_TYPE_TCPV4) {
+		DPRINTK("Bad GSO type %d.\n", gso->u.gso.type);
+		return -EINVAL;
+	}
+
+	skb_shinfo(skb)->gso_size = gso->u.gso.size;
+	skb_shinfo(skb)->gso_type = SKB_GSO_TCPV4;
+
+	/* Header must be checked, and gso_segs computed. */
+	skb_shinfo(skb)->gso_type |= SKB_GSO_DODGY;
+	skb_shinfo(skb)->gso_segs = 0;
+
+	return 0;
+}
+
+/* Called after netfront has transmitted */
+static void net_tx_action(unsigned long unused)
+{
+	struct list_head *ent;
+	struct sk_buff *skb;
+	netif_t *netif;
+	netif_tx_request_t txreq;
+	netif_tx_request_t txfrags[MAX_SKB_FRAGS];
+	struct netif_extra_info extras[XEN_NETIF_EXTRA_TYPE_MAX - 1];
+	u16 pending_idx;
+	RING_IDX i;
+	gnttab_map_grant_ref_t *mop;
+	unsigned int data_len;
+	int ret, work_to_do;
+
+	if (dealloc_cons != dealloc_prod)
+		net_tx_action_dealloc();
+
+	mop = tx_map_ops;
+	while (((NR_PENDING_REQS + MAX_SKB_FRAGS) < MAX_PENDING_REQS) &&
+		!list_empty(&net_schedule_list)) {
+		/* Get a netif from the list with work to do. */
+		ent = net_schedule_list.next;
+		netif = list_entry(ent, netif_t, list);
+		netif_get(netif);
+		remove_from_net_schedule_list(netif);
+
+		RING_FINAL_CHECK_FOR_REQUESTS(&netif->tx, work_to_do);
+		if (!work_to_do) {
+			netif_put(netif);
+			continue;
+		}
+
+		i = netif->tx.req_cons;
+		rmb(); /* Ensure that we see the request before we copy it. */
+		memcpy(&txreq, RING_GET_REQUEST(&netif->tx, i), sizeof(txreq));
+
+		/* Credit-based scheduling. */
+		if (txreq.size > netif->remaining_credit) {
+			unsigned long now = jiffies;
+			unsigned long next_credit = 
+				netif->credit_timeout.expires +
+				msecs_to_jiffies(netif->credit_usec / 1000);
+
+			/* Timer could already be pending in rare cases. */
+			if (timer_pending(&netif->credit_timeout)) {
+				netif_put(netif);
+				continue;
+			}
+
+			/* Passed the point where we can replenish credit? */
+			if (time_after_eq(now, next_credit)) {
+				netif->credit_timeout.expires = now;
+				tx_add_credit(netif);
+			}
+
+			/* Still too big to send right now? Set a callback. */
+			if (txreq.size > netif->remaining_credit) {
+				netif->credit_timeout.data     =
+					(unsigned long)netif;
+				netif->credit_timeout.function =
+					tx_credit_callback;
+				__mod_timer(&netif->credit_timeout,
+					    next_credit);
+				netif_put(netif);
+				continue;
+			}
+		}
+		netif->remaining_credit -= txreq.size;
+
+		work_to_do--;
+		netif->tx.req_cons = ++i;
+
+		memset(extras, 0, sizeof(extras));
+		if (txreq.flags & NETTXF_extra_info) {
+			work_to_do = netbk_get_extras(netif, extras,
+						      work_to_do);
+			i = netif->tx.req_cons;
+			if (unlikely(work_to_do < 0)) {
+				netbk_tx_err(netif, &txreq, i);
+				continue;
+			}
+		}
+
+		ret = netbk_count_requests(netif, &txreq, txfrags, work_to_do);
+		if (unlikely(ret < 0)) {
+			netbk_tx_err(netif, &txreq, i - ret);
+			continue;
+		}
+		i += ret;
+
+		if (unlikely(txreq.size < ETH_HLEN)) {
+			DPRINTK("Bad packet size: %d\n", txreq.size);
+			netbk_tx_err(netif, &txreq, i);
+			continue;
+		}
+
+		/* No crossing a page as the payload mustn't fragment. */
+		if (unlikely((txreq.offset + txreq.size) > PAGE_SIZE)) {
+			DPRINTK("txreq.offset: %x, size: %u, end: %lu\n", 
+				txreq.offset, txreq.size, 
+				(txreq.offset &~PAGE_MASK) + txreq.size);
+			netbk_tx_err(netif, &txreq, i);
+			continue;
+		}
+
+		pending_idx = pending_ring[MASK_PEND_IDX(pending_cons)];
+
+		data_len = (txreq.size > PKT_PROT_LEN &&
+			    ret < MAX_SKB_FRAGS) ?
+			PKT_PROT_LEN : txreq.size;
+
+		skb = alloc_skb(data_len + 16 + NET_IP_ALIGN,
+				GFP_ATOMIC | __GFP_NOWARN);
+		if (unlikely(skb == NULL)) {
+			DPRINTK("Can't allocate a skb in start_xmit.\n");
+			netbk_tx_err(netif, &txreq, i);
+			break;
+		}
+
+		/* Packets passed to netif_rx() must have some headroom. */
+		skb_reserve(skb, 16 + NET_IP_ALIGN);
+
+		if (extras[XEN_NETIF_EXTRA_TYPE_GSO - 1].type) {
+			struct netif_extra_info *gso;
+			gso = &extras[XEN_NETIF_EXTRA_TYPE_GSO - 1];
+
+			if (netbk_set_skb_gso(skb, gso)) {
+				kfree_skb(skb);
+				netbk_tx_err(netif, &txreq, i);
+				continue;
+			}
+		}
+
+		gnttab_set_map_op(mop, idx_to_kaddr(pending_idx),
+				  GNTMAP_host_map | GNTMAP_readonly,
+				  txreq.gref, netif->domid);
+		mop++;
+
+		memcpy(&pending_tx_info[pending_idx].req,
+		       &txreq, sizeof(txreq));
+		pending_tx_info[pending_idx].netif = netif;
+		*((u16 *)skb->data) = pending_idx;
+
+		__skb_put(skb, data_len);
+
+		skb_shinfo(skb)->nr_frags = ret;
+		if (data_len < txreq.size) {
+			skb_shinfo(skb)->nr_frags++;
+			skb_shinfo(skb)->frags[0].page =
+				(void *)(unsigned long)pending_idx;
+		} else {
+			/* Discriminate from any valid pending_idx value. */
+			skb_shinfo(skb)->frags[0].page = (void *)~0UL;
+		}
+
+		__skb_queue_tail(&tx_queue, skb);
+
+		pending_cons++;
+
+		mop = netbk_get_requests(netif, skb, txfrags, mop);
+
+		netif->tx.req_cons = i;
+		netif_schedule_work(netif);
+
+		if ((mop - tx_map_ops) >= ARRAY_SIZE(tx_map_ops))
+			break;
+	}
+
+	if (mop == tx_map_ops)
+		return;
+
+	ret = HYPERVISOR_grant_table_op(
+		GNTTABOP_map_grant_ref, tx_map_ops, mop - tx_map_ops);
+	BUG_ON(ret);
+
+	mop = tx_map_ops;
+	while ((skb = __skb_dequeue(&tx_queue)) != NULL) {
+		netif_tx_request_t *txp;
+
+		pending_idx = *((u16 *)skb->data);
+		netif       = pending_tx_info[pending_idx].netif;
+		txp         = &pending_tx_info[pending_idx].req;
+
+		/* Check the remap error code. */
+		if (unlikely(netbk_tx_check_mop(skb, &mop))) {
+			DPRINTK("netback grant failed.\n");
+			skb_shinfo(skb)->nr_frags = 0;
+			kfree_skb(skb);
+			continue;
+		}
+
+		data_len = skb->len;
+		memcpy(skb->data,
+		       (void *)(idx_to_kaddr(pending_idx)|txp->offset),
+		       data_len);
+		if (data_len < txp->size) {
+			/* Append the packet payload as a fragment. */
+			txp->offset += data_len;
+			txp->size -= data_len;
+		} else {
+			/* Schedule a response immediately. */
+			netif_idx_release(pending_idx);
+		}
+
+		/*
+		 * Old frontends do not assert data_validated but we
+		 * can infer it from csum_blank so test both flags.
+		 */
+		if (txp->flags & (NETTXF_data_validated|NETTXF_csum_blank)) {
+			skb->ip_summed = CHECKSUM_UNNECESSARY;
+			skb->proto_data_valid = 1;
+		} else {
+			skb->ip_summed = CHECKSUM_NONE;
+			skb->proto_data_valid = 0;
+		}
+		skb->proto_csum_blank = !!(txp->flags & NETTXF_csum_blank);
+
+		netbk_fill_frags(skb);
+
+		skb->dev      = netif->dev;
+		skb->protocol = eth_type_trans(skb, skb->dev);
+
+		netif->stats.rx_bytes += skb->len;
+		netif->stats.rx_packets++;
+
+		if (unlikely(netbk_copy_skb_mode == NETBK_ALWAYS_COPY_SKB) &&
+		    unlikely(skb_linearize(skb))) {
+			DPRINTK("Can't linearize skb in net_tx_action.\n");
+			kfree_skb(skb);
+			continue;
+		}
+
+		netif_rx(skb);
+		netif->dev->last_rx = jiffies;
+	}
+
+	if (netbk_copy_skb_mode == NETBK_DELAYED_COPY_SKB &&
+	    !list_empty(&pending_inuse_head)) {
+		struct netbk_tx_pending_inuse *oldest;
+
+		oldest = list_entry(pending_inuse_head.next,
+				    struct netbk_tx_pending_inuse, list);
+		mod_timer(&netbk_tx_pending_timer, oldest->alloc_time + HZ);
+	}
+}
+
+static void netif_idx_release(u16 pending_idx)
+{
+	static DEFINE_SPINLOCK(_lock);
+	unsigned long flags;
+
+	spin_lock_irqsave(&_lock, flags);
+	dealloc_ring[MASK_PEND_IDX(dealloc_prod)] = pending_idx;
+	/* Sync with net_tx_action_dealloc: insert idx /then/ incr producer. */
+	smp_wmb();
+	dealloc_prod++;
+	spin_unlock_irqrestore(&_lock, flags);
+
+	tasklet_schedule(&net_tx_tasklet);
+}
+
+static void netif_page_release(struct page *page)
+{
+	netif_idx_release(netif_page_index(page));
+}
+
+irqreturn_t netif_be_int(int irq, void *dev_id, struct pt_regs *regs)
+{
+	netif_t *netif = dev_id;
+
+	add_to_net_schedule_list_tail(netif);
+	maybe_schedule_tx_action();
+
+	if (netif_schedulable(netif) && !netbk_queue_full(netif))
+		netif_wake_queue(netif->dev);
+
+	return IRQ_HANDLED;
+}
+
+static void make_tx_response(netif_t *netif, 
+			     netif_tx_request_t *txp,
+			     s8       st)
+{
+	RING_IDX i = netif->tx.rsp_prod_pvt;
+	netif_tx_response_t *resp;
+	int notify;
+
+	resp = RING_GET_RESPONSE(&netif->tx, i);
+	resp->id     = txp->id;
+	resp->status = st;
+
+	if (txp->flags & NETTXF_extra_info)
+		RING_GET_RESPONSE(&netif->tx, ++i)->status = NETIF_RSP_NULL;
+
+	netif->tx.rsp_prod_pvt = ++i;
+	RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&netif->tx, notify);
+	if (notify)
+		notify_remote_via_irq(netif->irq);
+
+#ifdef CONFIG_XEN_NETDEV_PIPELINED_TRANSMITTER
+	if (i == netif->tx.req_cons) {
+		int more_to_do;
+		RING_FINAL_CHECK_FOR_REQUESTS(&netif->tx, more_to_do);
+		if (more_to_do)
+			add_to_net_schedule_list_tail(netif);
+	}
+#endif
+}
+
+static netif_rx_response_t *make_rx_response(netif_t *netif, 
+					     u16      id, 
+					     s8       st,
+					     u16      offset,
+					     u16      size,
+					     u16      flags)
+{
+	RING_IDX i = netif->rx.rsp_prod_pvt;
+	netif_rx_response_t *resp;
+
+	resp = RING_GET_RESPONSE(&netif->rx, i);
+	resp->offset     = offset;
+	resp->flags      = flags;
+	resp->id         = id;
+	resp->status     = (s16)size;
+	if (st < 0)
+		resp->status = (s16)st;
+
+	netif->rx.rsp_prod_pvt = ++i;
+
+	return resp;
+}
+
+#ifdef NETBE_DEBUG_INTERRUPT
+static irqreturn_t netif_be_dbg(int irq, void *dev_id, struct pt_regs *regs)
+{
+	struct list_head *ent;
+	netif_t *netif;
+	int i = 0;
+
+	printk(KERN_ALERT "netif_schedule_list:\n");
+	spin_lock_irq(&net_schedule_list_lock);
+
+	list_for_each (ent, &net_schedule_list) {
+		netif = list_entry(ent, netif_t, list);
+		printk(KERN_ALERT " %d: private(rx_req_cons=%08x "
+		       "rx_resp_prod=%08x\n",
+		       i, netif->rx.req_cons, netif->rx.rsp_prod_pvt);
+		printk(KERN_ALERT "   tx_req_cons=%08x tx_resp_prod=%08x)\n",
+		       netif->tx.req_cons, netif->tx.rsp_prod_pvt);
+		printk(KERN_ALERT "   shared(rx_req_prod=%08x "
+		       "rx_resp_prod=%08x\n",
+		       netif->rx.sring->req_prod, netif->rx.sring->rsp_prod);
+		printk(KERN_ALERT "   rx_event=%08x tx_req_prod=%08x\n",
+		       netif->rx.sring->rsp_event, netif->tx.sring->req_prod);
+		printk(KERN_ALERT "   tx_resp_prod=%08x, tx_event=%08x)\n",
+		       netif->tx.sring->rsp_prod, netif->tx.sring->rsp_event);
+		i++;
+	}
+
+	spin_unlock_irq(&net_schedule_list_lock);
+	printk(KERN_ALERT " ** End of netif_schedule_list **\n");
+
+	return IRQ_HANDLED;
+}
+#endif
+
+static int __init netback_init(void)
+{
+	int i;
+	struct page *page;
+
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	/* We can increase reservation by this much in net_rx_action(). */
+	balloon_update_driver_allowance(NET_RX_RING_SIZE);
+
+	skb_queue_head_init(&rx_queue);
+	skb_queue_head_init(&tx_queue);
+
+	init_timer(&net_timer);
+	net_timer.data = 0;
+	net_timer.function = net_alarm;
+
+	init_timer(&netbk_tx_pending_timer);
+	netbk_tx_pending_timer.data = 0;
+	netbk_tx_pending_timer.function = netbk_tx_pending_timeout;
+
+	mmap_pages = alloc_empty_pages_and_pagevec(MAX_PENDING_REQS);
+	if (mmap_pages == NULL) {
+		printk("%s: out of memory\n", __FUNCTION__);
+		return -ENOMEM;
+	}
+
+	for (i = 0; i < MAX_PENDING_REQS; i++) {
+		page = mmap_pages[i];
+		SetPageForeign(page, netif_page_release);
+		netif_page_index(page) = i;
+		INIT_LIST_HEAD(&pending_inuse[i].list);
+	}
+
+	pending_cons = 0;
+	pending_prod = MAX_PENDING_REQS;
+	for (i = 0; i < MAX_PENDING_REQS; i++)
+		pending_ring[i] = i;
+
+	spin_lock_init(&net_schedule_list_lock);
+	INIT_LIST_HEAD(&net_schedule_list);
+
+	netbk_copy_skb_mode = NETBK_DONT_COPY_SKB;
+	if (MODPARM_copy_skb) {
+		if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_and_replace,
+					      NULL, 0))
+			netbk_copy_skb_mode = NETBK_ALWAYS_COPY_SKB;
+		else
+			netbk_copy_skb_mode = NETBK_DELAYED_COPY_SKB;
+	}
+
+	netif_accel_init();
+
+	netif_xenbus_init();
+
+#ifdef NETBE_DEBUG_INTERRUPT
+	(void)bind_virq_to_irqhandler(VIRQ_DEBUG,
+				      0,
+				      netif_be_dbg,
+				      SA_SHIRQ, 
+				      "net-be-dbg",
+				      &netif_be_dbg);
+#endif
+
+	return 0;
+}
+
+module_init(netback_init);
+
+MODULE_LICENSE("Dual BSD/GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/netback/xenbus.c linux-2.6.18-xen-3.3.0/drivers/xen/netback/xenbus.c
--- linux-2.6.18.8/drivers/xen/netback/xenbus.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/netback/xenbus.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,454 @@
+/*  Xenbus code for netif backend
+    Copyright (C) 2005 Rusty Russell <rusty@rustcorp.com.au>
+    Copyright (C) 2005 XenSource Ltd
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program; if not, write to the Free Software
+    Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+*/
+
+#include <stdarg.h>
+#include <linux/module.h>
+#include <xen/xenbus.h>
+#include "common.h"
+
+#if 0
+#undef DPRINTK
+#define DPRINTK(fmt, args...) \
+    printk("netback/xenbus (%s:%d) " fmt ".\n", __FUNCTION__, __LINE__, ##args)
+#endif
+
+
+static int connect_rings(struct backend_info *);
+static void connect(struct backend_info *);
+static void backend_create_netif(struct backend_info *be);
+
+static int netback_remove(struct xenbus_device *dev)
+{
+	struct backend_info *be = dev->dev.driver_data;
+
+	netback_remove_accelerators(be, dev);
+
+	if (be->netif) {
+		kobject_uevent(&dev->dev.kobj, KOBJ_OFFLINE);
+		netif_disconnect(be->netif);
+		be->netif = NULL;
+	}
+	kfree(be);
+	dev->dev.driver_data = NULL;
+	return 0;
+}
+
+
+/**
+ * Entry point to this code when a new device is created.  Allocate the basic
+ * structures and switch to InitWait.
+ */
+static int netback_probe(struct xenbus_device *dev,
+			 const struct xenbus_device_id *id)
+{
+	const char *message;
+	struct xenbus_transaction xbt;
+	int err;
+	int sg;
+	struct backend_info *be = kzalloc(sizeof(struct backend_info),
+					  GFP_KERNEL);
+	if (!be) {
+		xenbus_dev_fatal(dev, -ENOMEM,
+				 "allocating backend structure");
+		return -ENOMEM;
+	}
+
+	be->dev = dev;
+	dev->dev.driver_data = be;
+
+	sg = 1;
+	if (netbk_copy_skb_mode == NETBK_ALWAYS_COPY_SKB)
+		sg = 0;
+
+	do {
+		err = xenbus_transaction_start(&xbt);
+		if (err) {
+			xenbus_dev_fatal(dev, err, "starting transaction");
+			goto fail;
+		}
+
+		err = xenbus_printf(xbt, dev->nodename, "feature-sg", "%d", sg);
+		if (err) {
+			message = "writing feature-sg";
+			goto abort_transaction;
+		}
+
+		err = xenbus_printf(xbt, dev->nodename, "feature-gso-tcpv4",
+				    "%d", sg);
+		if (err) {
+			message = "writing feature-gso-tcpv4";
+			goto abort_transaction;
+		}
+
+		/* We support rx-copy path. */
+		err = xenbus_printf(xbt, dev->nodename,
+				    "feature-rx-copy", "%d", 1);
+		if (err) {
+			message = "writing feature-rx-copy";
+			goto abort_transaction;
+		}
+
+		/*
+		 * We don't support rx-flip path (except old guests who don't
+		 * grok this feature flag).
+		 */
+		err = xenbus_printf(xbt, dev->nodename,
+				    "feature-rx-flip", "%d", 0);
+		if (err) {
+			message = "writing feature-rx-flip";
+			goto abort_transaction;
+		}
+
+		err = xenbus_transaction_end(xbt, 0);
+	} while (err == -EAGAIN);
+
+	if (err) {
+		xenbus_dev_fatal(dev, err, "completing transaction");
+		goto fail;
+	}
+
+	netback_probe_accelerators(be, dev);
+
+	err = xenbus_switch_state(dev, XenbusStateInitWait);
+	if (err)
+		goto fail;
+
+	/* This kicks hotplug scripts, so do it immediately. */
+	backend_create_netif(be);
+
+	return 0;
+
+abort_transaction:
+	xenbus_transaction_end(xbt, 1);
+	xenbus_dev_fatal(dev, err, "%s", message);
+fail:
+	DPRINTK("failed");
+	netback_remove(dev);
+	return err;
+}
+
+
+/**
+ * Handle the creation of the hotplug script environment.  We add the script
+ * and vif variables to the environment, for the benefit of the vif-* hotplug
+ * scripts.
+ */
+static int netback_uevent(struct xenbus_device *xdev, char **envp,
+			  int num_envp, char *buffer, int buffer_size)
+{
+	struct backend_info *be = xdev->dev.driver_data;
+	netif_t *netif = be->netif;
+	int i = 0, length = 0;
+	char *val;
+
+	DPRINTK("netback_uevent");
+
+	val = xenbus_read(XBT_NIL, xdev->nodename, "script", NULL);
+	if (IS_ERR(val)) {
+		int err = PTR_ERR(val);
+		xenbus_dev_fatal(xdev, err, "reading script");
+		return err;
+	}
+	else {
+		add_uevent_var(envp, num_envp, &i, buffer, buffer_size,
+			       &length, "script=%s", val);
+		kfree(val);
+	}
+
+	add_uevent_var(envp, num_envp, &i, buffer, buffer_size, &length,
+		       "vif=%s", netif->dev->name);
+
+	envp[i] = NULL;
+
+	return 0;
+}
+
+
+static void backend_create_netif(struct backend_info *be)
+{
+	int err;
+	long handle;
+	struct xenbus_device *dev = be->dev;
+
+	if (be->netif != NULL)
+		return;
+
+	err = xenbus_scanf(XBT_NIL, dev->nodename, "handle", "%li", &handle);
+	if (err != 1) {
+		xenbus_dev_fatal(dev, err, "reading handle");
+		return;
+	}
+
+	be->netif = netif_alloc(dev->otherend_id, handle);
+	if (IS_ERR(be->netif)) {
+		err = PTR_ERR(be->netif);
+		be->netif = NULL;
+		xenbus_dev_fatal(dev, err, "creating interface");
+		return;
+	}
+
+	kobject_uevent(&dev->dev.kobj, KOBJ_ONLINE);
+}
+
+
+/**
+ * Callback received when the frontend's state changes.
+ */
+static void frontend_changed(struct xenbus_device *dev,
+			     enum xenbus_state frontend_state)
+{
+	struct backend_info *be = dev->dev.driver_data;
+
+	DPRINTK("%s", xenbus_strstate(frontend_state));
+
+	be->frontend_state = frontend_state;
+
+	switch (frontend_state) {
+	case XenbusStateInitialising:
+		if (dev->state == XenbusStateClosed) {
+			printk(KERN_INFO "%s: %s: prepare for reconnect\n",
+			       __FUNCTION__, dev->nodename);
+			xenbus_switch_state(dev, XenbusStateInitWait);
+		}
+		break;
+
+	case XenbusStateInitialised:
+		break;
+
+	case XenbusStateConnected:
+		if (dev->state == XenbusStateConnected)
+			break;
+		backend_create_netif(be);
+		if (be->netif)
+			connect(be);
+		break;
+
+	case XenbusStateClosing:
+		if (be->netif) {
+			kobject_uevent(&dev->dev.kobj, KOBJ_OFFLINE);
+			netif_disconnect(be->netif);
+			be->netif = NULL;
+		}
+		xenbus_switch_state(dev, XenbusStateClosing);
+		break;
+
+	case XenbusStateClosed:
+		xenbus_switch_state(dev, XenbusStateClosed);
+		if (xenbus_dev_is_online(dev))
+			break;
+		/* fall through if not online */
+	case XenbusStateUnknown:
+		device_unregister(&dev->dev);
+		break;
+
+	default:
+		xenbus_dev_fatal(dev, -EINVAL, "saw state %d at frontend",
+				 frontend_state);
+		break;
+	}
+}
+
+
+static void xen_net_read_rate(struct xenbus_device *dev,
+			      unsigned long *bytes, unsigned long *usec)
+{
+	char *s, *e;
+	unsigned long b, u;
+	char *ratestr;
+
+	/* Default to unlimited bandwidth. */
+	*bytes = ~0UL;
+	*usec = 0;
+
+	ratestr = xenbus_read(XBT_NIL, dev->nodename, "rate", NULL);
+	if (IS_ERR(ratestr))
+		return;
+
+	s = ratestr;
+	b = simple_strtoul(s, &e, 10);
+	if ((s == e) || (*e != ','))
+		goto fail;
+
+	s = e + 1;
+	u = simple_strtoul(s, &e, 10);
+	if ((s == e) || (*e != '\0'))
+		goto fail;
+
+	*bytes = b;
+	*usec = u;
+
+	kfree(ratestr);
+	return;
+
+ fail:
+	WPRINTK("Failed to parse network rate limit. Traffic unlimited.\n");
+	kfree(ratestr);
+}
+
+static int xen_net_read_mac(struct xenbus_device *dev, u8 mac[])
+{
+	char *s, *e, *macstr;
+	int i;
+
+	macstr = s = xenbus_read(XBT_NIL, dev->nodename, "mac", NULL);
+	if (IS_ERR(macstr))
+		return PTR_ERR(macstr);
+
+	for (i = 0; i < ETH_ALEN; i++) {
+		mac[i] = simple_strtoul(s, &e, 16);
+		if ((s == e) || (*e != ((i == ETH_ALEN-1) ? '\0' : ':'))) {
+			kfree(macstr);
+			return -ENOENT;
+		}
+		s = e+1;
+	}
+
+	kfree(macstr);
+	return 0;
+}
+
+static void connect(struct backend_info *be)
+{
+	int err;
+	struct xenbus_device *dev = be->dev;
+
+	err = connect_rings(be);
+	if (err)
+		return;
+
+	err = xen_net_read_mac(dev, be->netif->fe_dev_addr);
+	if (err) {
+		xenbus_dev_fatal(dev, err, "parsing %s/mac", dev->nodename);
+		return;
+	}
+
+	xen_net_read_rate(dev, &be->netif->credit_bytes,
+			  &be->netif->credit_usec);
+	be->netif->remaining_credit = be->netif->credit_bytes;
+
+	xenbus_switch_state(dev, XenbusStateConnected);
+
+	netif_wake_queue(be->netif->dev);
+}
+
+
+static int connect_rings(struct backend_info *be)
+{
+	struct xenbus_device *dev = be->dev;
+	unsigned long tx_ring_ref, rx_ring_ref;
+	unsigned int evtchn, rx_copy;
+	int err;
+	int val;
+
+	DPRINTK("");
+
+	err = xenbus_gather(XBT_NIL, dev->otherend,
+			    "tx-ring-ref", "%lu", &tx_ring_ref,
+			    "rx-ring-ref", "%lu", &rx_ring_ref,
+			    "event-channel", "%u", &evtchn, NULL);
+	if (err) {
+		xenbus_dev_fatal(dev, err,
+				 "reading %s/ring-ref and event-channel",
+				 dev->otherend);
+		return err;
+	}
+
+	err = xenbus_scanf(XBT_NIL, dev->otherend, "request-rx-copy", "%u",
+			   &rx_copy);
+	if (err == -ENOENT) {
+		err = 0;
+		rx_copy = 0;
+	}
+	if (err < 0) {
+		xenbus_dev_fatal(dev, err, "reading %s/request-rx-copy",
+				 dev->otherend);
+		return err;
+	}
+	be->netif->copying_receiver = !!rx_copy;
+
+	if (be->netif->dev->tx_queue_len != 0) {
+		if (xenbus_scanf(XBT_NIL, dev->otherend,
+				 "feature-rx-notify", "%d", &val) < 0)
+			val = 0;
+		if (val)
+			be->netif->can_queue = 1;
+		else
+			/* Must be non-zero for pfifo_fast to work. */
+			be->netif->dev->tx_queue_len = 1;
+	}
+
+	if (xenbus_scanf(XBT_NIL, dev->otherend, "feature-sg", "%d", &val) < 0)
+		val = 0;
+	if (val) {
+		be->netif->features |= NETIF_F_SG;
+		be->netif->dev->features |= NETIF_F_SG;
+	}
+
+	if (xenbus_scanf(XBT_NIL, dev->otherend, "feature-gso-tcpv4", "%d",
+			 &val) < 0)
+		val = 0;
+	if (val) {
+		be->netif->features |= NETIF_F_TSO;
+		be->netif->dev->features |= NETIF_F_TSO;
+	}
+
+	if (xenbus_scanf(XBT_NIL, dev->otherend, "feature-no-csum-offload",
+			 "%d", &val) < 0)
+		val = 0;
+	if (val) {
+		be->netif->features &= ~NETIF_F_IP_CSUM;
+		be->netif->dev->features &= ~NETIF_F_IP_CSUM;
+	}
+
+	/* Map the shared frame, irq etc. */
+	err = netif_map(be->netif, tx_ring_ref, rx_ring_ref, evtchn);
+	if (err) {
+		xenbus_dev_fatal(dev, err,
+				 "mapping shared-frames %lu/%lu port %u",
+				 tx_ring_ref, rx_ring_ref, evtchn);
+		return err;
+	}
+	return 0;
+}
+
+
+/* ** Driver Registration ** */
+
+
+static const struct xenbus_device_id netback_ids[] = {
+	{ "vif" },
+	{ "" }
+};
+
+
+static struct xenbus_driver netback = {
+	.name = "vif",
+	.owner = THIS_MODULE,
+	.ids = netback_ids,
+	.probe = netback_probe,
+	.remove = netback_remove,
+	.uevent = netback_uevent,
+	.otherend_changed = frontend_changed,
+};
+
+
+void netif_xenbus_init(void)
+{
+	xenbus_register_backend(&netback);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/netfront/accel.c linux-2.6.18-xen-3.3.0/drivers/xen/netfront/accel.c
--- linux-2.6.18.8/drivers/xen/netfront/accel.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/netfront/accel.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,824 @@
+/******************************************************************************
+ * Virtual network driver for conversing with remote driver backends.
+ *
+ * Copyright (C) 2007 Solarflare Communications, Inc.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/netdevice.h>
+#include <linux/skbuff.h>
+#include <linux/list.h>
+#include <linux/mutex.h>
+#include <asm/hypervisor.h>
+#include <xen/xenbus.h>
+
+#include "netfront.h"
+
+#define DPRINTK(fmt, args...)				\
+	pr_debug("netfront/accel (%s:%d) " fmt,		\
+	       __FUNCTION__, __LINE__, ##args)
+#define IPRINTK(fmt, args...)				\
+	printk(KERN_INFO "netfront/accel: " fmt, ##args)
+#define WPRINTK(fmt, args...)				\
+	printk(KERN_WARNING "netfront/accel: " fmt, ##args)
+
+static int netfront_remove_accelerator(struct netfront_info *np,
+				       struct xenbus_device *dev);
+static int netfront_load_accelerator(struct netfront_info *np, 
+				     struct xenbus_device *dev, 
+				     const char *frontend);
+
+/*
+ * List of all netfront accelerator plugin modules available.  Each
+ * list entry is of type struct netfront_accelerator.
+ */ 
+static struct list_head accelerators_list;
+
+/* Lock to protect access to accelerators_list */
+static spinlock_t accelerators_lock;
+
+/* Workqueue to process acceleration configuration changes */
+struct workqueue_struct *accel_watch_workqueue;
+
+/* Mutex to prevent concurrent loads and suspends, etc. */
+DEFINE_MUTEX(accelerator_mutex);
+
+void netif_init_accel(void)
+{
+	INIT_LIST_HEAD(&accelerators_list);
+	spin_lock_init(&accelerators_lock);
+
+	accel_watch_workqueue = create_workqueue("net_accel");
+}
+
+void netif_exit_accel(void)
+{
+	struct netfront_accelerator *accelerator, *tmp;
+	unsigned long flags;
+
+	flush_workqueue(accel_watch_workqueue);
+	destroy_workqueue(accel_watch_workqueue);
+
+	spin_lock_irqsave(&accelerators_lock, flags);
+
+	list_for_each_entry_safe(accelerator, tmp, &accelerators_list, link) {
+		BUG_ON(!list_empty(&accelerator->vif_states));
+
+		list_del(&accelerator->link);
+		kfree(accelerator->frontend);
+		kfree(accelerator);
+	}
+
+	spin_unlock_irqrestore(&accelerators_lock, flags);
+}
+
+
+/* 
+ * Watch the configured accelerator and change plugin if it's modified 
+ */
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20)
+static void accel_watch_work(struct work_struct *context)
+#else
+static void accel_watch_work(void *context)
+#endif
+{
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20)
+	struct netfront_accel_vif_state *vif_state = 
+		container_of(context, struct netfront_accel_vif_state, 
+			     accel_work);
+#else
+        struct netfront_accel_vif_state *vif_state = 
+		(struct netfront_accel_vif_state *)context;
+#endif
+	struct netfront_info *np = vif_state->np;
+	char *accel_frontend;
+	int accel_len, rc = -1;
+
+	mutex_lock(&accelerator_mutex);
+
+	accel_frontend = xenbus_read(XBT_NIL, np->xbdev->otherend, 
+				     "accel-frontend", &accel_len);
+	if (IS_ERR(accel_frontend)) {
+		accel_frontend = NULL;
+		netfront_remove_accelerator(np, np->xbdev);
+	} else {
+		/* If this is the first time, request the accelerator,
+		   otherwise only request one if it has changed */
+		if (vif_state->accel_frontend == NULL) {
+			rc = netfront_load_accelerator(np, np->xbdev, 
+						       accel_frontend);
+		} else {
+			if (strncmp(vif_state->accel_frontend, accel_frontend,
+				    accel_len)) {
+				netfront_remove_accelerator(np, np->xbdev);
+				rc = netfront_load_accelerator(np, np->xbdev, 
+							       accel_frontend);
+			}
+		}
+	}
+
+	/* Get rid of previous state and replace with the new name */
+	if (vif_state->accel_frontend != NULL)
+		kfree(vif_state->accel_frontend);
+	vif_state->accel_frontend = accel_frontend;
+
+	mutex_unlock(&accelerator_mutex);
+
+	if (rc == 0) {
+		DPRINTK("requesting module %s\n", accel_frontend);
+		request_module("%s", accel_frontend);
+		/*
+		 * Module should now call netfront_accelerator_loaded() once
+		 * it's up and running, and we can continue from there 
+		 */
+	}
+}
+
+
+static void accel_watch_changed(struct xenbus_watch *watch,
+				const char **vec, unsigned int len)
+{
+	struct netfront_accel_vif_state *vif_state = 
+		container_of(watch, struct netfront_accel_vif_state,
+			     accel_watch);
+	queue_work(accel_watch_workqueue, &vif_state->accel_work);
+}
+
+
+void netfront_accelerator_add_watch(struct netfront_info *np)
+{
+	int err;
+	
+	/* Check we're not trying to overwrite an existing watch */
+	BUG_ON(np->accel_vif_state.accel_watch.node != NULL);
+
+	/* Get a watch on the accelerator plugin */
+	err = xenbus_watch_path2(np->xbdev, np->xbdev->otherend, 
+				 "accel-frontend", 
+				 &np->accel_vif_state.accel_watch,
+				 accel_watch_changed);
+	if (err) {
+		DPRINTK("%s: Failed to register accel watch: %d\n",
+                        __FUNCTION__, err);
+		np->accel_vif_state.accel_watch.node = NULL;
+        }
+}
+
+
+static
+void netfront_accelerator_remove_watch(struct netfront_info *np)
+{
+	struct netfront_accel_vif_state *vif_state = &np->accel_vif_state;
+
+	/* Get rid of watch on accelerator plugin */
+	if (vif_state->accel_watch.node != NULL) {
+		unregister_xenbus_watch(&vif_state->accel_watch);
+		kfree(vif_state->accel_watch.node);
+		vif_state->accel_watch.node = NULL;
+
+		flush_workqueue(accel_watch_workqueue);
+
+		/* Clean up any state left from watch */
+		if (vif_state->accel_frontend != NULL) {
+			kfree(vif_state->accel_frontend);
+			vif_state->accel_frontend = NULL;
+		}
+	}	
+}
+
+
+/* 
+ * Initialise the accel_vif_state field in the netfront state
+ */ 
+void init_accelerator_vif(struct netfront_info *np,
+			  struct xenbus_device *dev)
+{
+	np->accelerator = NULL;
+
+	/* It's assumed that these things don't change */
+	np->accel_vif_state.np = np;
+	np->accel_vif_state.dev = dev;
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20)
+	INIT_WORK(&np->accel_vif_state.accel_work, accel_watch_work);
+#else
+	INIT_WORK(&np->accel_vif_state.accel_work, accel_watch_work, 
+		  &np->accel_vif_state);
+#endif
+}
+
+
+/*
+ * Compare a frontend description string against an accelerator to see
+ * if they match.  Would ultimately be nice to replace the string with
+ * a unique numeric identifier for each accelerator.
+ */
+static int match_accelerator(const char *frontend, 
+			     struct netfront_accelerator *accelerator)
+{
+	return strcmp(frontend, accelerator->frontend) == 0;
+}
+
+
+/* 
+ * Add a frontend vif to the list of vifs that is using a netfront
+ * accelerator plugin module.
+ */
+static void add_accelerator_vif(struct netfront_accelerator *accelerator,
+				struct netfront_info *np)
+{
+	unsigned long flags;
+
+	/* Need lock to write list */
+	spin_lock_irqsave(&accelerator->vif_states_lock, flags);
+
+	if (np->accelerator == NULL) {
+		np->accelerator = accelerator;
+		
+		list_add(&np->accel_vif_state.link, &accelerator->vif_states);
+	} else {
+		/* 
+		 * May get here legitimately if suspend_cancel is
+		 * called, but in that case configuration should not
+		 * have changed
+		 */
+		BUG_ON(np->accelerator != accelerator);
+	}
+
+	spin_unlock_irqrestore(&accelerator->vif_states_lock, flags);
+}
+
+
+/*
+ * Initialise the state to track an accelerator plugin module.
+ */ 
+static int init_accelerator(const char *frontend, 
+			    struct netfront_accelerator **result,
+			    struct netfront_accel_hooks *hooks)
+{
+	struct netfront_accelerator *accelerator = 
+		kmalloc(sizeof(struct netfront_accelerator), GFP_KERNEL);
+	unsigned long flags;
+	int frontend_len;
+
+	if (!accelerator) {
+		DPRINTK("no memory for accelerator\n");
+		return -ENOMEM;
+	}
+
+	frontend_len = strlen(frontend) + 1;
+	accelerator->frontend = kmalloc(frontend_len, GFP_KERNEL);
+	if (!accelerator->frontend) {
+		DPRINTK("no memory for accelerator\n");
+		kfree(accelerator);
+		return -ENOMEM;
+	}
+	strlcpy(accelerator->frontend, frontend, frontend_len);
+	
+	INIT_LIST_HEAD(&accelerator->vif_states);
+	spin_lock_init(&accelerator->vif_states_lock);
+
+	accelerator->hooks = hooks;
+
+	spin_lock_irqsave(&accelerators_lock, flags);
+	list_add(&accelerator->link, &accelerators_list);
+	spin_unlock_irqrestore(&accelerators_lock, flags);
+
+	*result = accelerator;
+
+	return 0;
+}					
+
+
+/* 
+ * Modify the hooks stored in the per-vif state to match that in the
+ * netfront accelerator's state.
+ */
+static void 
+accelerator_set_vif_state_hooks(struct netfront_accel_vif_state *vif_state)
+{
+	/* This function must be called with the vif_states_lock held */
+
+	DPRINTK("%p\n",vif_state);
+
+	/* Make sure there are no data path operations going on */
+	netif_poll_disable(vif_state->np->netdev);
+	netif_tx_lock_bh(vif_state->np->netdev);
+
+	vif_state->hooks = vif_state->np->accelerator->hooks;
+
+	netif_tx_unlock_bh(vif_state->np->netdev);
+	netif_poll_enable(vif_state->np->netdev);
+}
+
+
+static void accelerator_probe_new_vif(struct netfront_info *np,
+				      struct xenbus_device *dev, 
+				      struct netfront_accelerator *accelerator)
+{
+	struct netfront_accel_hooks *hooks;
+	unsigned long flags;
+
+	DPRINTK("\n");
+
+	/* Include this frontend device on the accelerator's list */
+	add_accelerator_vif(accelerator, np);
+	
+	hooks = accelerator->hooks;
+	
+	if (hooks) {
+		if (hooks->new_device(np->netdev, dev) == 0) {
+			spin_lock_irqsave
+				(&accelerator->vif_states_lock, flags);
+
+			accelerator_set_vif_state_hooks(&np->accel_vif_state);
+
+			spin_unlock_irqrestore
+				(&accelerator->vif_states_lock, flags);
+		}
+	}
+
+	return;
+}
+
+
+/*  
+ * Request that a particular netfront accelerator plugin is loaded.
+ * Usually called as a result of the vif configuration specifying
+ * which one to use. Must be called with accelerator_mutex held 
+ */
+static int netfront_load_accelerator(struct netfront_info *np, 
+				     struct xenbus_device *dev, 
+				     const char *frontend)
+{
+	struct netfront_accelerator *accelerator;
+	int rc = 0;
+
+	DPRINTK(" %s\n", frontend);
+
+	/* 
+	 * Look at list of loaded accelerators to see if the requested
+	 * one is already there 
+	 */
+	list_for_each_entry(accelerator, &accelerators_list, link) {
+		if (match_accelerator(frontend, accelerator)) {
+			accelerator_probe_new_vif(np, dev, accelerator);
+			return 0;
+		}
+	}
+
+	/* Couldn't find it, so create a new one and load the module */
+	if ((rc = init_accelerator(frontend, &accelerator, NULL)) < 0) {
+		return rc;
+	}
+
+	/* Include this frontend device on the accelerator's list */
+	add_accelerator_vif(accelerator, np);
+
+	return rc;
+}
+
+
+/*
+ * Go through all the netfront vifs and see if they have requested
+ * this accelerator.  Notify the accelerator plugin of the relevant
+ * device if so.  Called when an accelerator plugin module is first
+ * loaded and connects to netfront.
+ */
+static void 
+accelerator_probe_vifs(struct netfront_accelerator *accelerator,
+		       struct netfront_accel_hooks *hooks)
+{
+	struct netfront_accel_vif_state *vif_state, *tmp;
+	unsigned long flags;
+
+	DPRINTK("%p\n", accelerator);
+
+	/* 
+	 * Store the hooks for future calls to probe a new device, and
+	 * to wire into the vif_state once the accelerator plugin is
+	 * ready to accelerate each vif
+	 */
+	BUG_ON(hooks == NULL);
+	accelerator->hooks = hooks;
+
+	/* 
+	 *  currently hold accelerator_mutex, so don't need
+	 *  vif_states_lock to read the list
+	 */
+	list_for_each_entry_safe(vif_state, tmp, &accelerator->vif_states,
+				 link) {
+		struct netfront_info *np = vif_state->np;
+		
+		if (hooks->new_device(np->netdev, vif_state->dev) == 0) {
+			spin_lock_irqsave
+				(&accelerator->vif_states_lock, flags);
+
+			accelerator_set_vif_state_hooks(vif_state);
+
+			spin_unlock_irqrestore
+				(&accelerator->vif_states_lock, flags);
+		}
+	}
+}
+
+
+/* 
+ * Called by the netfront accelerator plugin module when it has loaded 
+ */
+int netfront_accelerator_loaded(int version, const char *frontend, 
+				struct netfront_accel_hooks *hooks)
+{
+	struct netfront_accelerator *accelerator;
+
+	if (is_initial_xendomain())
+		return -EINVAL;
+
+	if (version != NETFRONT_ACCEL_VERSION) {
+		if (version > NETFRONT_ACCEL_VERSION) {
+			/* Caller has higher version number, leave it
+			   up to them to decide whether to continue.
+			   They can re-call with a lower number if
+			   they're happy to be compatible with us */
+			return NETFRONT_ACCEL_VERSION;
+		} else {
+			/* We have a more recent version than caller.
+			   Currently reject, but may in future be able
+			   to be backwardly compatible */
+			return -EPROTO;
+		}
+	}
+
+	mutex_lock(&accelerator_mutex);
+
+	/* 
+	 * Look through list of accelerators to see if it has already
+	 * been requested
+	 */
+	list_for_each_entry(accelerator, &accelerators_list, link) {
+		if (match_accelerator(frontend, accelerator)) {
+			accelerator_probe_vifs(accelerator, hooks);
+			goto out;
+		}
+	}
+
+	/*
+	 * If it wasn't in the list, add it now so that when it is
+	 * requested the caller will find it
+	 */
+	DPRINTK("Couldn't find matching accelerator (%s)\n",
+		frontend);
+
+	init_accelerator(frontend, &accelerator, hooks);
+
+ out:
+	mutex_unlock(&accelerator_mutex);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(netfront_accelerator_loaded);
+
+
+/* 
+ * Remove the hooks from a single vif state.
+ */
+static void 
+accelerator_remove_single_hook(struct netfront_accelerator *accelerator,
+			       struct netfront_accel_vif_state *vif_state)
+{
+	/* Make sure there are no data path operations going on */
+	netif_poll_disable(vif_state->np->netdev);
+	netif_tx_lock_bh(vif_state->np->netdev);
+
+	/* 
+	 * Remove the hooks, but leave the vif_state on the
+	 * accelerator's list as that signifies this vif is
+	 * interested in using that accelerator if it becomes
+	 * available again
+	 */
+	vif_state->hooks = NULL;
+	
+	netif_tx_unlock_bh(vif_state->np->netdev);
+	netif_poll_enable(vif_state->np->netdev);		       
+}
+
+
+/* 
+ * Safely remove the accelerator function hooks from a netfront state.
+ */
+static void accelerator_remove_hooks(struct netfront_accelerator *accelerator)
+{
+	struct netfront_accel_hooks *hooks;
+	struct netfront_accel_vif_state *vif_state, *tmp;
+	unsigned long flags;
+
+	/* Mutex is held so don't need vif_states_lock to iterate list */
+	list_for_each_entry_safe(vif_state, tmp,
+				 &accelerator->vif_states,
+				 link) {
+		spin_lock_irqsave(&accelerator->vif_states_lock, flags);
+
+		if(vif_state->hooks) {
+			hooks = vif_state->hooks;
+			
+			/* Last chance to get statistics from the accelerator */
+			hooks->get_stats(vif_state->np->netdev,
+					 &vif_state->np->stats);
+
+			spin_unlock_irqrestore(&accelerator->vif_states_lock,
+					       flags);
+
+			accelerator_remove_single_hook(accelerator, vif_state);
+
+			accelerator->hooks->remove(vif_state->dev);
+		} else {
+			spin_unlock_irqrestore(&accelerator->vif_states_lock,
+					       flags);
+		}
+	}
+	
+	accelerator->hooks = NULL;
+}
+
+
+/* 
+ * Called by a netfront accelerator when it is unloaded.  This safely
+ * removes the hooks into the plugin and blocks until all devices have
+ * finished using it, so on return it is safe to unload.
+ */
+void netfront_accelerator_stop(const char *frontend)
+{
+	struct netfront_accelerator *accelerator;
+	unsigned long flags;
+
+	mutex_lock(&accelerator_mutex);
+	spin_lock_irqsave(&accelerators_lock, flags);
+
+	list_for_each_entry(accelerator, &accelerators_list, link) {
+		if (match_accelerator(frontend, accelerator)) {
+			spin_unlock_irqrestore(&accelerators_lock, flags);
+
+			accelerator_remove_hooks(accelerator);
+
+			goto out;
+		}
+	}
+	spin_unlock_irqrestore(&accelerators_lock, flags);
+ out:
+	mutex_unlock(&accelerator_mutex);
+}
+EXPORT_SYMBOL_GPL(netfront_accelerator_stop);
+
+
+/* Helper for call_remove and do_suspend */
+static int do_remove(struct netfront_info *np, struct xenbus_device *dev,
+		     unsigned long *lock_flags)
+{
+	struct netfront_accelerator *accelerator = np->accelerator;
+ 	struct netfront_accel_hooks *hooks;
+ 	int rc = 0;
+ 
+	if (np->accel_vif_state.hooks) {
+		hooks = np->accel_vif_state.hooks;
+
+		/* Last chance to get statistics from the accelerator */
+		hooks->get_stats(np->netdev, &np->stats);
+
+		spin_unlock_irqrestore(&accelerator->vif_states_lock, 
+				       *lock_flags);
+
+ 		/* 
+ 		 * Try and do the opposite of accelerator_probe_new_vif
+ 		 * to ensure there's no state pointing back at the 
+ 		 * netdev 
+ 		 */
+		accelerator_remove_single_hook(accelerator, 
+ 					       &np->accel_vif_state);
+
+		rc = accelerator->hooks->remove(dev);
+
+		spin_lock_irqsave(&accelerator->vif_states_lock, *lock_flags);
+	}
+ 
+ 	return rc;
+}
+
+
+static int netfront_remove_accelerator(struct netfront_info *np,
+				       struct xenbus_device *dev)
+{
+	struct netfront_accelerator *accelerator;
+ 	struct netfront_accel_vif_state *tmp_vif_state;
+  	unsigned long flags;
+	int rc = 0; 
+
+ 	/* Check that we've got a device that was accelerated */
+ 	if (np->accelerator == NULL)
+		return rc;
+
+	accelerator = np->accelerator;
+
+	spin_lock_irqsave(&accelerator->vif_states_lock, flags); 
+
+	list_for_each_entry(tmp_vif_state, &accelerator->vif_states,
+			    link) {
+		if (tmp_vif_state == &np->accel_vif_state) {
+			list_del(&np->accel_vif_state.link);
+			break;
+		}
+	}
+
+	rc = do_remove(np, dev, &flags);
+
+	np->accelerator = NULL;
+
+	spin_unlock_irqrestore(&accelerator->vif_states_lock, flags); 
+
+	return rc;
+}
+
+
+int netfront_accelerator_call_remove(struct netfront_info *np,
+				     struct xenbus_device *dev)
+{
+	int rc;
+	netfront_accelerator_remove_watch(np);
+	mutex_lock(&accelerator_mutex);
+	rc = netfront_remove_accelerator(np, dev);
+	mutex_unlock(&accelerator_mutex);
+	return rc;
+}
+
+  
+int netfront_accelerator_suspend(struct netfront_info *np,
+ 				 struct xenbus_device *dev)
+{
+	unsigned long flags;
+	int rc = 0;
+	
+	netfront_accelerator_remove_watch(np);
+
+	mutex_lock(&accelerator_mutex);
+
+ 	/* Check that we've got a device that was accelerated */
+ 	if (np->accelerator == NULL)
+		goto out;
+
+	/* 
+	 * Call the remove accelerator hook, but leave the vif_state
+	 * on the accelerator's list in case there is a suspend_cancel.
+	 */
+	spin_lock_irqsave(&np->accelerator->vif_states_lock, flags); 
+	
+	rc = do_remove(np, dev, &flags);
+
+	spin_unlock_irqrestore(&np->accelerator->vif_states_lock, flags); 
+ out:
+	mutex_unlock(&accelerator_mutex);
+	return rc;
+}
+  
+  
+int netfront_accelerator_suspend_cancel(struct netfront_info *np,
+ 					struct xenbus_device *dev)
+{
+	/* 
+	 * Setting the watch will cause it to fire and probe the
+	 * accelerator, so no need to call accelerator_probe_new_vif()
+	 * directly here
+	 */
+	if (dev->state == XenbusStateConnected)
+		netfront_accelerator_add_watch(np);
+	return 0;
+}
+ 
+ 
+void netfront_accelerator_resume(struct netfront_info *np,
+ 				 struct xenbus_device *dev)
+{
+ 	struct netfront_accel_vif_state *accel_vif_state = NULL;
+ 	spinlock_t *vif_states_lock;
+ 	unsigned long flags;
+ 
+ 	mutex_lock(&accelerator_mutex);
+
+	/* Check that we've got a device that was accelerated */
+ 	if(np->accelerator == NULL)
+		goto out;
+
+ 	/* Find the vif_state from the accelerator's list */
+ 	list_for_each_entry(accel_vif_state, &np->accelerator->vif_states, 
+ 			    link) {
+ 		if (accel_vif_state->dev == dev) {
+ 			BUG_ON(accel_vif_state != &np->accel_vif_state);
+ 
+ 			vif_states_lock = &np->accelerator->vif_states_lock;
+			spin_lock_irqsave(vif_states_lock, flags); 
+ 
+ 			/* 
+ 			 * Remove it from the accelerator's list so
+ 			 * state is consistent for probing new vifs
+ 			 * when they get connected
+ 			 */
+ 			list_del(&accel_vif_state->link);
+ 			np->accelerator = NULL;
+ 
+ 			spin_unlock_irqrestore(vif_states_lock, flags); 
+ 			
+			break;
+ 		}
+ 	}
+
+ out:
+	mutex_unlock(&accelerator_mutex);
+	return;
+}
+
+
+int netfront_check_accelerator_queue_ready(struct net_device *dev,
+					   struct netfront_info *np)
+{
+	struct netfront_accelerator *accelerator;
+	struct netfront_accel_hooks *hooks;
+	int rc = 1;
+	unsigned long flags;
+
+	accelerator = np->accelerator;
+
+	/* Call the check_ready accelerator hook. */ 
+	if (np->accel_vif_state.hooks && accelerator) {
+		spin_lock_irqsave(&accelerator->vif_states_lock, flags); 
+		hooks = np->accel_vif_state.hooks;
+		if (hooks && np->accelerator == accelerator)
+			rc = np->accel_vif_state.hooks->check_ready(dev);
+		spin_unlock_irqrestore(&accelerator->vif_states_lock, flags);
+	}
+
+	return rc;
+}
+
+
+void netfront_accelerator_call_stop_napi_irq(struct netfront_info *np,
+					     struct net_device *dev)
+{
+	struct netfront_accelerator *accelerator;
+	struct netfront_accel_hooks *hooks;
+	unsigned long flags;
+
+	accelerator = np->accelerator;
+
+	/* Call the stop_napi_interrupts accelerator hook. */
+	if (np->accel_vif_state.hooks && accelerator != NULL) {
+		spin_lock_irqsave(&accelerator->vif_states_lock, flags); 
+		hooks = np->accel_vif_state.hooks;
+		if (hooks && np->accelerator == accelerator)
+ 			np->accel_vif_state.hooks->stop_napi_irq(dev);
+		spin_unlock_irqrestore(&accelerator->vif_states_lock, flags);
+	}
+}
+
+
+int netfront_accelerator_call_get_stats(struct netfront_info *np,
+					struct net_device *dev)
+{
+	struct netfront_accelerator *accelerator;
+	struct netfront_accel_hooks *hooks;
+	unsigned long flags;
+	int rc = 0;
+
+	accelerator = np->accelerator;
+
+	/* Call the get_stats accelerator hook. */
+	if (np->accel_vif_state.hooks && accelerator != NULL) {
+		spin_lock_irqsave(&accelerator->vif_states_lock, flags); 
+		hooks = np->accel_vif_state.hooks;
+		if (hooks && np->accelerator == accelerator)
+ 			rc = np->accel_vif_state.hooks->get_stats(dev,
+								  &np->stats);
+		spin_unlock_irqrestore(&accelerator->vif_states_lock, flags);
+	}
+	return rc;
+}
+
diff -rpuN linux-2.6.18.8/drivers/xen/netfront/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/netfront/Makefile
--- linux-2.6.18.8/drivers/xen/netfront/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/netfront/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,4 @@
+
+obj-$(CONFIG_XEN_NETDEV_FRONTEND)	:= xennet.o
+
+xennet-objs := netfront.o accel.o
diff -rpuN linux-2.6.18.8/drivers/xen/netfront/netfront.c linux-2.6.18-xen-3.3.0/drivers/xen/netfront/netfront.c
--- linux-2.6.18.8/drivers/xen/netfront/netfront.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/netfront/netfront.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,2240 @@
+/******************************************************************************
+ * Virtual network driver for conversing with remote driver backends.
+ *
+ * Copyright (c) 2002-2005, K A Fraser
+ * Copyright (c) 2005, XenSource Ltd
+ * Copyright (C) 2007 Solarflare Communications, Inc.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/module.h>
+#include <linux/version.h>
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/string.h>
+#include <linux/errno.h>
+#include <linux/netdevice.h>
+#include <linux/inetdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/skbuff.h>
+#include <linux/init.h>
+#include <linux/bitops.h>
+#include <linux/ethtool.h>
+#include <linux/in.h>
+#include <linux/if_ether.h>
+#include <linux/io.h>
+#include <linux/moduleparam.h>
+#include <net/sock.h>
+#include <net/pkt_sched.h>
+#include <net/arp.h>
+#include <net/route.h>
+#include <asm/uaccess.h>
+#include <xen/evtchn.h>
+#include <xen/xenbus.h>
+#include <xen/interface/io/netif.h>
+#include <xen/interface/memory.h>
+#include <xen/balloon.h>
+#include <asm/page.h>
+#include <asm/maddr.h>
+#include <asm/uaccess.h>
+#include <xen/interface/grant_table.h>
+#include <xen/gnttab.h>
+
+struct netfront_cb {
+	struct page *page;
+	unsigned offset;
+};
+
+#define NETFRONT_SKB_CB(skb)	((struct netfront_cb *)((skb)->cb))
+
+#include "netfront.h"
+
+/*
+ * Mutually-exclusive module options to select receive data path:
+ *  rx_copy : Packets are copied by network backend into local memory
+ *  rx_flip : Page containing packet data is transferred to our ownership
+ * For fully-virtualised guests there is no option - copying must be used.
+ * For paravirtualised guests, flipping is the default.
+ */
+#ifdef CONFIG_XEN
+static int MODPARM_rx_copy = 0;
+module_param_named(rx_copy, MODPARM_rx_copy, bool, 0);
+MODULE_PARM_DESC(rx_copy, "Copy packets from network card (rather than flip)");
+static int MODPARM_rx_flip = 0;
+module_param_named(rx_flip, MODPARM_rx_flip, bool, 0);
+MODULE_PARM_DESC(rx_flip, "Flip packets from network card (rather than copy)");
+#else
+static const int MODPARM_rx_copy = 1;
+static const int MODPARM_rx_flip = 0;
+#endif
+
+#define RX_COPY_THRESHOLD 256
+
+/* If we don't have GSO, fake things up so that we never try to use it. */
+#if defined(NETIF_F_GSO)
+#define HAVE_GSO			1
+#define HAVE_TSO			1 /* TSO is a subset of GSO */
+#define HAVE_CSUM_OFFLOAD		1
+static inline void dev_disable_gso_features(struct net_device *dev)
+{
+	/* Turn off all GSO bits except ROBUST. */
+	dev->features &= (1 << NETIF_F_GSO_SHIFT) - 1;
+	dev->features |= NETIF_F_GSO_ROBUST;
+}
+#elif defined(NETIF_F_TSO)
+#define HAVE_GSO		       0
+#define HAVE_TSO                       1
+
+/* Some older kernels cannot cope with incorrect checksums,
+ * particularly in netfilter. I'm not sure there is 100% correlation
+ * with the presence of NETIF_F_TSO but it appears to be a good first
+ * approximiation.
+ */
+#define HAVE_CSUM_OFFLOAD              0
+
+#define gso_size tso_size
+#define gso_segs tso_segs
+static inline void dev_disable_gso_features(struct net_device *dev)
+{
+       /* Turn off all TSO bits. */
+       dev->features &= ~NETIF_F_TSO;
+}
+static inline int skb_is_gso(const struct sk_buff *skb)
+{
+        return skb_shinfo(skb)->tso_size;
+}
+static inline int skb_gso_ok(struct sk_buff *skb, int features)
+{
+        return (features & NETIF_F_TSO);
+}
+
+static inline int netif_needs_gso(struct net_device *dev, struct sk_buff *skb)
+{
+        return skb_is_gso(skb) &&
+               (!skb_gso_ok(skb, dev->features) ||
+                unlikely(skb->ip_summed != CHECKSUM_HW));
+}
+#else
+#define HAVE_GSO			0
+#define HAVE_TSO			0
+#define HAVE_CSUM_OFFLOAD		0
+#define netif_needs_gso(dev, skb)	0
+#define dev_disable_gso_features(dev)	((void)0)
+#define ethtool_op_set_tso(dev, data)	(-ENOSYS)
+#endif
+
+#define GRANT_INVALID_REF	0
+
+struct netfront_rx_info {
+	struct netif_rx_response rx;
+	struct netif_extra_info extras[XEN_NETIF_EXTRA_TYPE_MAX - 1];
+};
+
+/*
+ * Implement our own carrier flag: the network stack's version causes delays
+ * when the carrier is re-enabled (in particular, dev_activate() may not
+ * immediately be called, which can cause packet loss).
+ */
+#define netfront_carrier_on(netif)	((netif)->carrier = 1)
+#define netfront_carrier_off(netif)	((netif)->carrier = 0)
+#define netfront_carrier_ok(netif)	((netif)->carrier)
+
+/*
+ * Access macros for acquiring freeing slots in tx_skbs[].
+ */
+
+static inline void add_id_to_freelist(struct sk_buff **list, unsigned short id)
+{
+	list[id] = list[0];
+	list[0]  = (void *)(unsigned long)id;
+}
+
+static inline unsigned short get_id_from_freelist(struct sk_buff **list)
+{
+	unsigned int id = (unsigned int)(unsigned long)list[0];
+	list[0] = list[id];
+	return id;
+}
+
+static inline int xennet_rxidx(RING_IDX idx)
+{
+	return idx & (NET_RX_RING_SIZE - 1);
+}
+
+static inline struct sk_buff *xennet_get_rx_skb(struct netfront_info *np,
+						RING_IDX ri)
+{
+	int i = xennet_rxidx(ri);
+	struct sk_buff *skb = np->rx_skbs[i];
+	np->rx_skbs[i] = NULL;
+	return skb;
+}
+
+static inline grant_ref_t xennet_get_rx_ref(struct netfront_info *np,
+					    RING_IDX ri)
+{
+	int i = xennet_rxidx(ri);
+	grant_ref_t ref = np->grant_rx_ref[i];
+	np->grant_rx_ref[i] = GRANT_INVALID_REF;
+	return ref;
+}
+
+#define DPRINTK(fmt, args...)				\
+	pr_debug("netfront (%s:%d) " fmt,		\
+		 __FUNCTION__, __LINE__, ##args)
+#define IPRINTK(fmt, args...)				\
+	printk(KERN_INFO "netfront: " fmt, ##args)
+#define WPRINTK(fmt, args...)				\
+	printk(KERN_WARNING "netfront: " fmt, ##args)
+
+static int setup_device(struct xenbus_device *, struct netfront_info *);
+static struct net_device *create_netdev(struct xenbus_device *);
+
+static void end_access(int, void *);
+static void netif_disconnect_backend(struct netfront_info *);
+
+static int network_connect(struct net_device *);
+static void network_tx_buf_gc(struct net_device *);
+static void network_alloc_rx_buffers(struct net_device *);
+static void send_fake_arp(struct net_device *);
+
+static irqreturn_t netif_int(int irq, void *dev_id, struct pt_regs *ptregs);
+
+#ifdef CONFIG_SYSFS
+static int xennet_sysfs_addif(struct net_device *netdev);
+static void xennet_sysfs_delif(struct net_device *netdev);
+#else /* !CONFIG_SYSFS */
+#define xennet_sysfs_addif(dev) (0)
+#define xennet_sysfs_delif(dev) do { } while(0)
+#endif
+
+static inline int xennet_can_sg(struct net_device *dev)
+{
+	return dev->features & NETIF_F_SG;
+}
+
+/**
+ * Entry point to this code when a new device is created.  Allocate the basic
+ * structures and the ring buffers for communication with the backend, and
+ * inform the backend of the appropriate details for those.
+ */
+static int __devinit netfront_probe(struct xenbus_device *dev,
+				    const struct xenbus_device_id *id)
+{
+	int err;
+	struct net_device *netdev;
+	struct netfront_info *info;
+
+	netdev = create_netdev(dev);
+	if (IS_ERR(netdev)) {
+		err = PTR_ERR(netdev);
+		xenbus_dev_fatal(dev, err, "creating netdev");
+		return err;
+	}
+
+	info = netdev_priv(netdev);
+	dev->dev.driver_data = info;
+
+	err = register_netdev(info->netdev);
+	if (err) {
+		printk(KERN_WARNING "%s: register_netdev err=%d\n",
+		       __FUNCTION__, err);
+		goto fail;
+	}
+
+	err = xennet_sysfs_addif(info->netdev);
+	if (err) {
+		unregister_netdev(info->netdev);
+		printk(KERN_WARNING "%s: add sysfs failed err=%d\n",
+		       __FUNCTION__, err);
+		goto fail;
+	}
+
+	return 0;
+
+ fail:
+	free_netdev(netdev);
+	dev->dev.driver_data = NULL;
+	return err;
+}
+
+static int __devexit netfront_remove(struct xenbus_device *dev)
+{
+	struct netfront_info *info = dev->dev.driver_data;
+
+	DPRINTK("%s\n", dev->nodename);
+
+	netfront_accelerator_call_remove(info, dev);
+
+	netif_disconnect_backend(info);
+
+	del_timer_sync(&info->rx_refill_timer);
+
+	xennet_sysfs_delif(info->netdev);
+
+	unregister_netdev(info->netdev);
+
+	free_netdev(info->netdev);
+
+	return 0;
+}
+
+
+static int netfront_suspend(struct xenbus_device *dev)
+{
+	struct netfront_info *info = dev->dev.driver_data;
+	return netfront_accelerator_suspend(info, dev);
+}
+
+
+static int netfront_suspend_cancel(struct xenbus_device *dev)
+{
+	struct netfront_info *info = dev->dev.driver_data;
+	return netfront_accelerator_suspend_cancel(info, dev);
+}
+
+
+/**
+ * We are reconnecting to the backend, due to a suspend/resume, or a backend
+ * driver restart.  We tear down our netif structure and recreate it, but
+ * leave the device-layer structures intact so that this is transparent to the
+ * rest of the kernel.
+ */
+static int netfront_resume(struct xenbus_device *dev)
+{
+	struct netfront_info *info = dev->dev.driver_data;
+
+	DPRINTK("%s\n", dev->nodename);
+
+	netfront_accelerator_resume(info, dev);
+
+	netif_disconnect_backend(info);
+	return 0;
+}
+
+static int xen_net_read_mac(struct xenbus_device *dev, u8 mac[])
+{
+	char *s, *e, *macstr;
+	int i;
+
+	macstr = s = xenbus_read(XBT_NIL, dev->nodename, "mac", NULL);
+	if (IS_ERR(macstr))
+		return PTR_ERR(macstr);
+
+	for (i = 0; i < ETH_ALEN; i++) {
+		mac[i] = simple_strtoul(s, &e, 16);
+		if ((s == e) || (*e != ((i == ETH_ALEN-1) ? '\0' : ':'))) {
+			kfree(macstr);
+			return -ENOENT;
+		}
+		s = e+1;
+	}
+
+	kfree(macstr);
+	return 0;
+}
+
+/* Common code used when first setting up, and when resuming. */
+static int talk_to_backend(struct xenbus_device *dev,
+			   struct netfront_info *info)
+{
+	const char *message;
+	struct xenbus_transaction xbt;
+	int err;
+
+	/* Read mac only in the first setup. */
+	if (!is_valid_ether_addr(info->mac)) {
+		err = xen_net_read_mac(dev, info->mac);
+		if (err) {
+			xenbus_dev_fatal(dev, err, "parsing %s/mac",
+					 dev->nodename);
+			goto out;
+		}
+	}
+
+	/* Create shared ring, alloc event channel. */
+	err = setup_device(dev, info);
+	if (err)
+		goto out;
+
+	/* This will load an accelerator if one is configured when the
+	 * watch fires */
+	netfront_accelerator_add_watch(info);
+
+again:
+	err = xenbus_transaction_start(&xbt);
+	if (err) {
+		xenbus_dev_fatal(dev, err, "starting transaction");
+		goto destroy_ring;
+	}
+
+	err = xenbus_printf(xbt, dev->nodename, "tx-ring-ref","%u",
+			    info->tx_ring_ref);
+	if (err) {
+		message = "writing tx ring-ref";
+		goto abort_transaction;
+	}
+	err = xenbus_printf(xbt, dev->nodename, "rx-ring-ref","%u",
+			    info->rx_ring_ref);
+	if (err) {
+		message = "writing rx ring-ref";
+		goto abort_transaction;
+	}
+	err = xenbus_printf(xbt, dev->nodename,
+			    "event-channel", "%u",
+			    irq_to_evtchn_port(info->irq));
+	if (err) {
+		message = "writing event-channel";
+		goto abort_transaction;
+	}
+
+	err = xenbus_printf(xbt, dev->nodename, "request-rx-copy", "%u",
+			    info->copying_receiver);
+	if (err) {
+		message = "writing request-rx-copy";
+		goto abort_transaction;
+	}
+
+	err = xenbus_printf(xbt, dev->nodename, "feature-rx-notify", "%d", 1);
+	if (err) {
+		message = "writing feature-rx-notify";
+		goto abort_transaction;
+	}
+
+	err = xenbus_printf(xbt, dev->nodename, "feature-no-csum-offload",
+			    "%d", !HAVE_CSUM_OFFLOAD);
+	if (err) {
+		message = "writing feature-no-csum-offload";
+		goto abort_transaction;
+	}
+
+	err = xenbus_printf(xbt, dev->nodename, "feature-sg", "%d", 1);
+	if (err) {
+		message = "writing feature-sg";
+		goto abort_transaction;
+	}
+
+	err = xenbus_printf(xbt, dev->nodename, "feature-gso-tcpv4", "%d",
+			    HAVE_TSO);
+	if (err) {
+		message = "writing feature-gso-tcpv4";
+		goto abort_transaction;
+	}
+
+	err = xenbus_transaction_end(xbt, 0);
+	if (err) {
+		if (err == -EAGAIN)
+			goto again;
+		xenbus_dev_fatal(dev, err, "completing transaction");
+		goto destroy_ring;
+	}
+
+	return 0;
+
+ abort_transaction:
+	xenbus_transaction_end(xbt, 1);
+	xenbus_dev_fatal(dev, err, "%s", message);
+ destroy_ring:
+	netfront_accelerator_call_remove(info, dev);
+	netif_disconnect_backend(info);
+ out:
+	return err;
+}
+
+static int setup_device(struct xenbus_device *dev, struct netfront_info *info)
+{
+	struct netif_tx_sring *txs;
+	struct netif_rx_sring *rxs;
+	int err;
+	struct net_device *netdev = info->netdev;
+
+	info->tx_ring_ref = GRANT_INVALID_REF;
+	info->rx_ring_ref = GRANT_INVALID_REF;
+	info->rx.sring = NULL;
+	info->tx.sring = NULL;
+	info->irq = 0;
+
+	txs = (struct netif_tx_sring *)get_zeroed_page(GFP_NOIO | __GFP_HIGH);
+	if (!txs) {
+		err = -ENOMEM;
+		xenbus_dev_fatal(dev, err, "allocating tx ring page");
+		goto fail;
+	}
+	SHARED_RING_INIT(txs);
+	FRONT_RING_INIT(&info->tx, txs, PAGE_SIZE);
+
+	err = xenbus_grant_ring(dev, virt_to_mfn(txs));
+	if (err < 0) {
+		free_page((unsigned long)txs);
+		goto fail;
+	}
+	info->tx_ring_ref = err;
+
+	rxs = (struct netif_rx_sring *)get_zeroed_page(GFP_NOIO | __GFP_HIGH);
+	if (!rxs) {
+		err = -ENOMEM;
+		xenbus_dev_fatal(dev, err, "allocating rx ring page");
+		goto fail;
+	}
+	SHARED_RING_INIT(rxs);
+	FRONT_RING_INIT(&info->rx, rxs, PAGE_SIZE);
+
+	err = xenbus_grant_ring(dev, virt_to_mfn(rxs));
+	if (err < 0) {
+		free_page((unsigned long)rxs);
+		goto fail;
+	}
+	info->rx_ring_ref = err;
+
+	memcpy(netdev->dev_addr, info->mac, ETH_ALEN);
+
+	err = bind_listening_port_to_irqhandler(
+		dev->otherend_id, netif_int, SA_SAMPLE_RANDOM, netdev->name,
+		netdev);
+	if (err < 0)
+		goto fail;
+	info->irq = err;
+
+	return 0;
+
+ fail:
+	return err;
+}
+
+/**
+ * Callback received when the backend's state changes.
+ */
+static void backend_changed(struct xenbus_device *dev,
+			    enum xenbus_state backend_state)
+{
+	struct netfront_info *np = dev->dev.driver_data;
+	struct net_device *netdev = np->netdev;
+
+	DPRINTK("%s\n", xenbus_strstate(backend_state));
+
+	switch (backend_state) {
+	case XenbusStateInitialising:
+	case XenbusStateInitialised:
+	case XenbusStateConnected:
+	case XenbusStateReconfiguring:
+	case XenbusStateReconfigured:
+	case XenbusStateUnknown:
+	case XenbusStateClosed:
+		break;
+
+	case XenbusStateInitWait:
+		if (dev->state != XenbusStateInitialising)
+			break;
+		if (network_connect(netdev) != 0)
+			break;
+		xenbus_switch_state(dev, XenbusStateConnected);
+		send_fake_arp(netdev);
+		break;
+
+	case XenbusStateClosing:
+		xenbus_frontend_closed(dev);
+		break;
+	}
+}
+
+/** Send a packet on a net device to encourage switches to learn the
+ * MAC. We send a fake ARP request.
+ *
+ * @param dev device
+ * @return 0 on success, error code otherwise
+ */
+static void send_fake_arp(struct net_device *dev)
+{
+#ifdef CONFIG_INET
+	struct sk_buff *skb;
+	u32             src_ip, dst_ip;
+
+	dst_ip = INADDR_BROADCAST;
+	src_ip = inet_select_addr(dev, dst_ip, RT_SCOPE_LINK);
+
+	/* No IP? Then nothing to do. */
+	if (src_ip == 0)
+		return;
+
+	skb = arp_create(ARPOP_REPLY, ETH_P_ARP,
+			 dst_ip, dev, src_ip,
+			 /*dst_hw*/ NULL, /*src_hw*/ NULL,
+			 /*target_hw*/ dev->dev_addr);
+	if (skb == NULL)
+		return;
+
+	dev_queue_xmit(skb);
+#endif
+}
+
+static inline int netfront_tx_slot_available(struct netfront_info *np)
+{
+	return ((np->tx.req_prod_pvt - np->tx.rsp_cons) <
+		(TX_MAX_TARGET - MAX_SKB_FRAGS - 2));
+}
+
+
+static inline void network_maybe_wake_tx(struct net_device *dev)
+{
+	struct netfront_info *np = netdev_priv(dev);
+
+	if (unlikely(netif_queue_stopped(dev)) &&
+	    netfront_tx_slot_available(np) &&
+	    likely(netif_running(dev)) &&
+	    netfront_check_accelerator_queue_ready(dev, np))
+		netif_wake_queue(dev);
+}
+
+
+int netfront_check_queue_ready(struct net_device *dev)
+{
+	struct netfront_info *np = netdev_priv(dev);
+
+	return unlikely(netif_queue_stopped(dev)) &&
+		netfront_tx_slot_available(np) &&
+		likely(netif_running(dev));
+}
+EXPORT_SYMBOL(netfront_check_queue_ready);
+
+
+static int network_open(struct net_device *dev)
+{
+	struct netfront_info *np = netdev_priv(dev);
+
+	memset(&np->stats, 0, sizeof(np->stats));
+
+	spin_lock_bh(&np->rx_lock);
+	if (netfront_carrier_ok(np)) {
+		network_alloc_rx_buffers(dev);
+		np->rx.sring->rsp_event = np->rx.rsp_cons + 1;
+		if (RING_HAS_UNCONSUMED_RESPONSES(&np->rx)){
+			netfront_accelerator_call_stop_napi_irq(np, dev);
+
+			netif_rx_schedule(dev);
+		}
+	}
+	spin_unlock_bh(&np->rx_lock);
+
+	network_maybe_wake_tx(dev);
+
+	return 0;
+}
+
+static void network_tx_buf_gc(struct net_device *dev)
+{
+	RING_IDX cons, prod;
+	unsigned short id;
+	struct netfront_info *np = netdev_priv(dev);
+	struct sk_buff *skb;
+
+	BUG_ON(!netfront_carrier_ok(np));
+
+	do {
+		prod = np->tx.sring->rsp_prod;
+		rmb(); /* Ensure we see responses up to 'rp'. */
+
+		for (cons = np->tx.rsp_cons; cons != prod; cons++) {
+			struct netif_tx_response *txrsp;
+
+			txrsp = RING_GET_RESPONSE(&np->tx, cons);
+			if (txrsp->status == NETIF_RSP_NULL)
+				continue;
+
+			id  = txrsp->id;
+			skb = np->tx_skbs[id];
+			if (unlikely(gnttab_query_foreign_access(
+				np->grant_tx_ref[id]) != 0)) {
+				printk(KERN_ALERT "network_tx_buf_gc: warning "
+				       "-- grant still in use by backend "
+				       "domain.\n");
+				BUG();
+			}
+			gnttab_end_foreign_access_ref(np->grant_tx_ref[id]);
+			gnttab_release_grant_reference(
+				&np->gref_tx_head, np->grant_tx_ref[id]);
+			np->grant_tx_ref[id] = GRANT_INVALID_REF;
+			add_id_to_freelist(np->tx_skbs, id);
+			dev_kfree_skb_irq(skb);
+		}
+
+		np->tx.rsp_cons = prod;
+
+		/*
+		 * Set a new event, then check for race with update of tx_cons.
+		 * Note that it is essential to schedule a callback, no matter
+		 * how few buffers are pending. Even if there is space in the
+		 * transmit ring, higher layers may be blocked because too much
+		 * data is outstanding: in such cases notification from Xen is
+		 * likely to be the only kick that we'll get.
+		 */
+		np->tx.sring->rsp_event =
+			prod + ((np->tx.sring->req_prod - prod) >> 1) + 1;
+		mb();
+	} while ((cons == prod) && (prod != np->tx.sring->rsp_prod));
+
+	network_maybe_wake_tx(dev);
+}
+
+static void rx_refill_timeout(unsigned long data)
+{
+	struct net_device *dev = (struct net_device *)data;
+	struct netfront_info *np = netdev_priv(dev);
+
+	netfront_accelerator_call_stop_napi_irq(np, dev);
+
+	netif_rx_schedule(dev);
+}
+
+static void network_alloc_rx_buffers(struct net_device *dev)
+{
+	unsigned short id;
+	struct netfront_info *np = netdev_priv(dev);
+	struct sk_buff *skb;
+	struct page *page;
+	int i, batch_target, notify;
+	RING_IDX req_prod = np->rx.req_prod_pvt;
+	struct xen_memory_reservation reservation;
+	grant_ref_t ref;
+ 	unsigned long pfn;
+ 	void *vaddr;
+	int nr_flips;
+	netif_rx_request_t *req;
+
+	if (unlikely(!netfront_carrier_ok(np)))
+		return;
+
+	/*
+	 * Allocate skbuffs greedily, even though we batch updates to the
+	 * receive ring. This creates a less bursty demand on the memory
+	 * allocator, so should reduce the chance of failed allocation requests
+	 * both for ourself and for other kernel subsystems.
+	 */
+	batch_target = np->rx_target - (req_prod - np->rx.rsp_cons);
+	for (i = skb_queue_len(&np->rx_batch); i < batch_target; i++) {
+		/*
+		 * Allocate an skb and a page. Do not use __dev_alloc_skb as
+		 * that will allocate page-sized buffers which is not
+		 * necessary here.
+		 * 16 bytes added as necessary headroom for netif_receive_skb.
+		 */
+		skb = alloc_skb(RX_COPY_THRESHOLD + 16 + NET_IP_ALIGN,
+				GFP_ATOMIC | __GFP_NOWARN);
+		if (unlikely(!skb))
+			goto no_skb;
+
+		page = alloc_page(GFP_ATOMIC | __GFP_NOWARN);
+		if (!page) {
+			kfree_skb(skb);
+no_skb:
+			/* Any skbuffs queued for refill? Force them out. */
+			if (i != 0)
+				goto refill;
+			/* Could not allocate any skbuffs. Try again later. */
+			mod_timer(&np->rx_refill_timer,
+				  jiffies + (HZ/10));
+			break;
+		}
+
+		skb_reserve(skb, 16 + NET_IP_ALIGN); /* mimic dev_alloc_skb() */
+		skb_shinfo(skb)->frags[0].page = page;
+		skb_shinfo(skb)->nr_frags = 1;
+		__skb_queue_tail(&np->rx_batch, skb);
+	}
+
+	/* Is the batch large enough to be worthwhile? */
+	if (i < (np->rx_target/2)) {
+		if (req_prod > np->rx.sring->req_prod)
+			goto push;
+		return;
+	}
+
+	/* Adjust our fill target if we risked running out of buffers. */
+	if (((req_prod - np->rx.sring->rsp_prod) < (np->rx_target / 4)) &&
+	    ((np->rx_target *= 2) > np->rx_max_target))
+		np->rx_target = np->rx_max_target;
+
+ refill:
+	for (nr_flips = i = 0; ; i++) {
+		if ((skb = __skb_dequeue(&np->rx_batch)) == NULL)
+			break;
+
+		skb->dev = dev;
+
+		id = xennet_rxidx(req_prod + i);
+
+		BUG_ON(np->rx_skbs[id]);
+		np->rx_skbs[id] = skb;
+
+		ref = gnttab_claim_grant_reference(&np->gref_rx_head);
+		BUG_ON((signed short)ref < 0);
+		np->grant_rx_ref[id] = ref;
+
+		pfn = page_to_pfn(skb_shinfo(skb)->frags[0].page);
+		vaddr = page_address(skb_shinfo(skb)->frags[0].page);
+
+		req = RING_GET_REQUEST(&np->rx, req_prod + i);
+		if (!np->copying_receiver) {
+			gnttab_grant_foreign_transfer_ref(ref,
+							  np->xbdev->otherend_id,
+							  pfn);
+			np->rx_pfn_array[nr_flips] = pfn_to_mfn(pfn);
+			if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+				/* Remove this page before passing
+				 * back to Xen. */
+				set_phys_to_machine(pfn, INVALID_P2M_ENTRY);
+				MULTI_update_va_mapping(np->rx_mcl+i,
+							(unsigned long)vaddr,
+							__pte(0), 0);
+			}
+			nr_flips++;
+		} else {
+			gnttab_grant_foreign_access_ref(ref,
+							np->xbdev->otherend_id,
+							pfn_to_mfn(pfn),
+							0);
+		}
+
+		req->id = id;
+		req->gref = ref;
+	}
+
+	if ( nr_flips != 0 ) {
+		/* Tell the ballon driver what is going on. */
+		balloon_update_driver_allowance(i);
+
+		set_xen_guest_handle(reservation.extent_start,
+				     np->rx_pfn_array);
+		reservation.nr_extents   = nr_flips;
+		reservation.extent_order = 0;
+		reservation.address_bits = 0;
+		reservation.domid        = DOMID_SELF;
+
+		if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+			/* After all PTEs have been zapped, flush the TLB. */
+			np->rx_mcl[i-1].args[MULTI_UVMFLAGS_INDEX] =
+				UVMF_TLB_FLUSH|UVMF_ALL;
+
+			/* Give away a batch of pages. */
+			np->rx_mcl[i].op = __HYPERVISOR_memory_op;
+			np->rx_mcl[i].args[0] = XENMEM_decrease_reservation;
+			np->rx_mcl[i].args[1] = (unsigned long)&reservation;
+
+			/* Zap PTEs and give away pages in one big
+			 * multicall. */
+			if (unlikely(HYPERVISOR_multicall(np->rx_mcl, i+1)))
+				BUG();
+
+			/* Check return status of HYPERVISOR_memory_op(). */
+			if (unlikely(np->rx_mcl[i].result != i))
+				panic("Unable to reduce memory reservation\n");
+			while (nr_flips--)
+				BUG_ON(np->rx_mcl[nr_flips].result);
+		} else {
+			if (HYPERVISOR_memory_op(XENMEM_decrease_reservation,
+						 &reservation) != i)
+				panic("Unable to reduce memory reservation\n");
+		}
+	} else {
+		wmb();
+	}
+
+	/* Above is a suitable barrier to ensure backend will see requests. */
+	np->rx.req_prod_pvt = req_prod + i;
+ push:
+	RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&np->rx, notify);
+	if (notify)
+		notify_remote_via_irq(np->irq);
+}
+
+static void xennet_make_frags(struct sk_buff *skb, struct net_device *dev,
+			      struct netif_tx_request *tx)
+{
+	struct netfront_info *np = netdev_priv(dev);
+	char *data = skb->data;
+	unsigned long mfn;
+	RING_IDX prod = np->tx.req_prod_pvt;
+	int frags = skb_shinfo(skb)->nr_frags;
+	unsigned int offset = offset_in_page(data);
+	unsigned int len = skb_headlen(skb);
+	unsigned int id;
+	grant_ref_t ref;
+	int i;
+
+	while (len > PAGE_SIZE - offset) {
+		tx->size = PAGE_SIZE - offset;
+		tx->flags |= NETTXF_more_data;
+		len -= tx->size;
+		data += tx->size;
+		offset = 0;
+
+		id = get_id_from_freelist(np->tx_skbs);
+		np->tx_skbs[id] = skb_get(skb);
+		tx = RING_GET_REQUEST(&np->tx, prod++);
+		tx->id = id;
+		ref = gnttab_claim_grant_reference(&np->gref_tx_head);
+		BUG_ON((signed short)ref < 0);
+
+		mfn = virt_to_mfn(data);
+		gnttab_grant_foreign_access_ref(ref, np->xbdev->otherend_id,
+						mfn, GTF_readonly);
+
+		tx->gref = np->grant_tx_ref[id] = ref;
+		tx->offset = offset;
+		tx->size = len;
+		tx->flags = 0;
+	}
+
+	for (i = 0; i < frags; i++) {
+		skb_frag_t *frag = skb_shinfo(skb)->frags + i;
+
+		tx->flags |= NETTXF_more_data;
+
+		id = get_id_from_freelist(np->tx_skbs);
+		np->tx_skbs[id] = skb_get(skb);
+		tx = RING_GET_REQUEST(&np->tx, prod++);
+		tx->id = id;
+		ref = gnttab_claim_grant_reference(&np->gref_tx_head);
+		BUG_ON((signed short)ref < 0);
+
+		mfn = pfn_to_mfn(page_to_pfn(frag->page));
+		gnttab_grant_foreign_access_ref(ref, np->xbdev->otherend_id,
+						mfn, GTF_readonly);
+
+		tx->gref = np->grant_tx_ref[id] = ref;
+		tx->offset = frag->page_offset;
+		tx->size = frag->size;
+		tx->flags = 0;
+	}
+
+	np->tx.req_prod_pvt = prod;
+}
+
+static int network_start_xmit(struct sk_buff *skb, struct net_device *dev)
+{
+	unsigned short id;
+	struct netfront_info *np = netdev_priv(dev);
+	struct netif_tx_request *tx;
+	struct netif_extra_info *extra;
+	char *data = skb->data;
+	RING_IDX i;
+	grant_ref_t ref;
+	unsigned long mfn;
+	int notify;
+	int frags = skb_shinfo(skb)->nr_frags;
+	unsigned int offset = offset_in_page(data);
+	unsigned int len = skb_headlen(skb);
+
+	/* Check the fast path, if hooks are available */
+ 	if (np->accel_vif_state.hooks && 
+ 	    np->accel_vif_state.hooks->start_xmit(skb, dev)) { 
+ 		/* Fast path has sent this packet */ 
+ 		return 0; 
+ 	} 
+
+	frags += (offset + len + PAGE_SIZE - 1) / PAGE_SIZE;
+	if (unlikely(frags > MAX_SKB_FRAGS + 1)) {
+		printk(KERN_ALERT "xennet: skb rides the rocket: %d frags\n",
+		       frags);
+		dump_stack();
+		goto drop;
+	}
+
+	spin_lock_irq(&np->tx_lock);
+
+	if (unlikely(!netfront_carrier_ok(np) ||
+		     (frags > 1 && !xennet_can_sg(dev)) ||
+		     netif_needs_gso(dev, skb))) {
+		spin_unlock_irq(&np->tx_lock);
+		goto drop;
+	}
+
+	i = np->tx.req_prod_pvt;
+
+	id = get_id_from_freelist(np->tx_skbs);
+	np->tx_skbs[id] = skb;
+
+	tx = RING_GET_REQUEST(&np->tx, i);
+
+	tx->id   = id;
+	ref = gnttab_claim_grant_reference(&np->gref_tx_head);
+	BUG_ON((signed short)ref < 0);
+	mfn = virt_to_mfn(data);
+	gnttab_grant_foreign_access_ref(
+		ref, np->xbdev->otherend_id, mfn, GTF_readonly);
+	tx->gref = np->grant_tx_ref[id] = ref;
+	tx->offset = offset;
+	tx->size = len;
+
+	tx->flags = 0;
+	extra = NULL;
+
+	if (skb->ip_summed == CHECKSUM_HW) /* local packet? */
+		tx->flags |= NETTXF_csum_blank | NETTXF_data_validated;
+#ifdef CONFIG_XEN
+	if (skb->proto_data_valid) /* remote but checksummed? */
+		tx->flags |= NETTXF_data_validated;
+#endif
+
+#if HAVE_TSO
+	if (skb_shinfo(skb)->gso_size) {
+		struct netif_extra_info *gso = (struct netif_extra_info *)
+			RING_GET_REQUEST(&np->tx, ++i);
+
+		if (extra)
+			extra->flags |= XEN_NETIF_EXTRA_FLAG_MORE;
+		else
+			tx->flags |= NETTXF_extra_info;
+
+		gso->u.gso.size = skb_shinfo(skb)->gso_size;
+		gso->u.gso.type = XEN_NETIF_GSO_TYPE_TCPV4;
+		gso->u.gso.pad = 0;
+		gso->u.gso.features = 0;
+
+		gso->type = XEN_NETIF_EXTRA_TYPE_GSO;
+		gso->flags = 0;
+		extra = gso;
+	}
+#endif
+
+	np->tx.req_prod_pvt = i + 1;
+
+	xennet_make_frags(skb, dev, tx);
+	tx->size = skb->len;
+
+	RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&np->tx, notify);
+	if (notify)
+		notify_remote_via_irq(np->irq);
+
+	np->stats.tx_bytes += skb->len;
+	np->stats.tx_packets++;
+	dev->trans_start = jiffies;
+
+	/* Note: It is not safe to access skb after network_tx_buf_gc()! */
+	network_tx_buf_gc(dev);
+
+	if (!netfront_tx_slot_available(np))
+		netif_stop_queue(dev);
+
+	spin_unlock_irq(&np->tx_lock);
+
+	return 0;
+
+ drop:
+	np->stats.tx_dropped++;
+	dev_kfree_skb(skb);
+	return 0;
+}
+
+static irqreturn_t netif_int(int irq, void *dev_id, struct pt_regs *ptregs)
+{
+	struct net_device *dev = dev_id;
+	struct netfront_info *np = netdev_priv(dev);
+	unsigned long flags;
+
+	spin_lock_irqsave(&np->tx_lock, flags);
+
+	if (likely(netfront_carrier_ok(np))) {
+		network_tx_buf_gc(dev);
+		/* Under tx_lock: protects access to rx shared-ring indexes. */
+		if (RING_HAS_UNCONSUMED_RESPONSES(&np->rx)) {
+			netfront_accelerator_call_stop_napi_irq(np, dev);
+
+			netif_rx_schedule(dev);
+			dev->last_rx = jiffies;
+		}
+	}
+
+	spin_unlock_irqrestore(&np->tx_lock, flags);
+
+	return IRQ_HANDLED;
+}
+
+static void xennet_move_rx_slot(struct netfront_info *np, struct sk_buff *skb,
+				grant_ref_t ref)
+{
+	int new = xennet_rxidx(np->rx.req_prod_pvt);
+
+	BUG_ON(np->rx_skbs[new]);
+	np->rx_skbs[new] = skb;
+	np->grant_rx_ref[new] = ref;
+	RING_GET_REQUEST(&np->rx, np->rx.req_prod_pvt)->id = new;
+	RING_GET_REQUEST(&np->rx, np->rx.req_prod_pvt)->gref = ref;
+	np->rx.req_prod_pvt++;
+}
+
+int xennet_get_extras(struct netfront_info *np,
+		      struct netif_extra_info *extras, RING_IDX rp)
+
+{
+	struct netif_extra_info *extra;
+	RING_IDX cons = np->rx.rsp_cons;
+	int err = 0;
+
+	do {
+		struct sk_buff *skb;
+		grant_ref_t ref;
+
+		if (unlikely(cons + 1 == rp)) {
+			if (net_ratelimit())
+				WPRINTK("Missing extra info\n");
+			err = -EBADR;
+			break;
+		}
+
+		extra = (struct netif_extra_info *)
+			RING_GET_RESPONSE(&np->rx, ++cons);
+
+		if (unlikely(!extra->type ||
+			     extra->type >= XEN_NETIF_EXTRA_TYPE_MAX)) {
+			if (net_ratelimit())
+				WPRINTK("Invalid extra type: %d\n",
+					extra->type);
+			err = -EINVAL;
+		} else {
+			memcpy(&extras[extra->type - 1], extra,
+			       sizeof(*extra));
+		}
+
+		skb = xennet_get_rx_skb(np, cons);
+		ref = xennet_get_rx_ref(np, cons);
+		xennet_move_rx_slot(np, skb, ref);
+	} while (extra->flags & XEN_NETIF_EXTRA_FLAG_MORE);
+
+	np->rx.rsp_cons = cons;
+	return err;
+}
+
+static int xennet_get_responses(struct netfront_info *np,
+				struct netfront_rx_info *rinfo, RING_IDX rp,
+				struct sk_buff_head *list,
+				int *pages_flipped_p)
+{
+	int pages_flipped = *pages_flipped_p;
+	struct mmu_update *mmu;
+	struct multicall_entry *mcl;
+	struct netif_rx_response *rx = &rinfo->rx;
+	struct netif_extra_info *extras = rinfo->extras;
+	RING_IDX cons = np->rx.rsp_cons;
+	struct sk_buff *skb = xennet_get_rx_skb(np, cons);
+	grant_ref_t ref = xennet_get_rx_ref(np, cons);
+	int max = MAX_SKB_FRAGS + (rx->status <= RX_COPY_THRESHOLD);
+	int frags = 1;
+	int err = 0;
+	unsigned long ret;
+
+	if (rx->flags & NETRXF_extra_info) {
+		err = xennet_get_extras(np, extras, rp);
+		cons = np->rx.rsp_cons;
+	}
+
+	for (;;) {
+		unsigned long mfn;
+
+		if (unlikely(rx->status < 0 ||
+			     rx->offset + rx->status > PAGE_SIZE)) {
+			if (net_ratelimit())
+				WPRINTK("rx->offset: %x, size: %u\n",
+					rx->offset, rx->status);
+			xennet_move_rx_slot(np, skb, ref);
+			err = -EINVAL;
+			goto next;
+		}
+
+		/*
+		 * This definitely indicates a bug, either in this driver or in
+		 * the backend driver. In future this should flag the bad
+		 * situation to the system controller to reboot the backed.
+		 */
+		if (ref == GRANT_INVALID_REF) {
+			if (net_ratelimit())
+				WPRINTK("Bad rx response id %d.\n", rx->id);
+			err = -EINVAL;
+			goto next;
+		}
+
+		if (!np->copying_receiver) {
+			/* Memory pressure, insufficient buffer
+			 * headroom, ... */
+			if (!(mfn = gnttab_end_foreign_transfer_ref(ref))) {
+				if (net_ratelimit())
+					WPRINTK("Unfulfilled rx req "
+						"(id=%d, st=%d).\n",
+						rx->id, rx->status);
+				xennet_move_rx_slot(np, skb, ref);
+				err = -ENOMEM;
+				goto next;
+			}
+
+			if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+				/* Remap the page. */
+				struct page *page =
+					skb_shinfo(skb)->frags[0].page;
+				unsigned long pfn = page_to_pfn(page);
+				void *vaddr = page_address(page);
+
+				mcl = np->rx_mcl + pages_flipped;
+				mmu = np->rx_mmu + pages_flipped;
+
+				MULTI_update_va_mapping(mcl,
+							(unsigned long)vaddr,
+							pfn_pte_ma(mfn,
+								   PAGE_KERNEL),
+							0);
+				mmu->ptr = ((maddr_t)mfn << PAGE_SHIFT)
+					| MMU_MACHPHYS_UPDATE;
+				mmu->val = pfn;
+
+				set_phys_to_machine(pfn, mfn);
+			}
+			pages_flipped++;
+		} else {
+			ret = gnttab_end_foreign_access_ref(ref);
+			BUG_ON(!ret);
+		}
+
+		gnttab_release_grant_reference(&np->gref_rx_head, ref);
+
+		__skb_queue_tail(list, skb);
+
+next:
+		if (!(rx->flags & NETRXF_more_data))
+			break;
+
+		if (cons + frags == rp) {
+			if (net_ratelimit())
+				WPRINTK("Need more frags\n");
+			err = -ENOENT;
+			break;
+		}
+
+		rx = RING_GET_RESPONSE(&np->rx, cons + frags);
+		skb = xennet_get_rx_skb(np, cons + frags);
+		ref = xennet_get_rx_ref(np, cons + frags);
+		frags++;
+	}
+
+	if (unlikely(frags > max)) {
+		if (net_ratelimit())
+			WPRINTK("Too many frags\n");
+		err = -E2BIG;
+	}
+
+	if (unlikely(err))
+		np->rx.rsp_cons = cons + frags;
+
+	*pages_flipped_p = pages_flipped;
+
+	return err;
+}
+
+static RING_IDX xennet_fill_frags(struct netfront_info *np,
+				  struct sk_buff *skb,
+				  struct sk_buff_head *list)
+{
+	struct skb_shared_info *shinfo = skb_shinfo(skb);
+	int nr_frags = shinfo->nr_frags;
+	RING_IDX cons = np->rx.rsp_cons;
+	skb_frag_t *frag = shinfo->frags + nr_frags;
+	struct sk_buff *nskb;
+
+	while ((nskb = __skb_dequeue(list))) {
+		struct netif_rx_response *rx =
+			RING_GET_RESPONSE(&np->rx, ++cons);
+
+		frag->page = skb_shinfo(nskb)->frags[0].page;
+		frag->page_offset = rx->offset;
+		frag->size = rx->status;
+
+		skb->data_len += rx->status;
+
+		skb_shinfo(nskb)->nr_frags = 0;
+		kfree_skb(nskb);
+
+		frag++;
+		nr_frags++;
+	}
+
+	shinfo->nr_frags = nr_frags;
+	return cons;
+}
+
+static int xennet_set_skb_gso(struct sk_buff *skb,
+			      struct netif_extra_info *gso)
+{
+	if (!gso->u.gso.size) {
+		if (net_ratelimit())
+			WPRINTK("GSO size must not be zero.\n");
+		return -EINVAL;
+	}
+
+	/* Currently only TCPv4 S.O. is supported. */
+	if (gso->u.gso.type != XEN_NETIF_GSO_TYPE_TCPV4) {
+		if (net_ratelimit())
+			WPRINTK("Bad GSO type %d.\n", gso->u.gso.type);
+		return -EINVAL;
+	}
+
+#if HAVE_TSO
+	skb_shinfo(skb)->gso_size = gso->u.gso.size;
+#if HAVE_GSO
+	skb_shinfo(skb)->gso_type = SKB_GSO_TCPV4;
+
+	/* Header must be checked, and gso_segs computed. */
+	skb_shinfo(skb)->gso_type |= SKB_GSO_DODGY;
+#endif
+	skb_shinfo(skb)->gso_segs = 0;
+
+	return 0;
+#else
+	if (net_ratelimit())
+		WPRINTK("GSO unsupported by this kernel.\n");
+	return -EINVAL;
+#endif
+}
+
+static int netif_poll(struct net_device *dev, int *pbudget)
+{
+	struct netfront_info *np = netdev_priv(dev);
+	struct sk_buff *skb;
+	struct netfront_rx_info rinfo;
+	struct netif_rx_response *rx = &rinfo.rx;
+	struct netif_extra_info *extras = rinfo.extras;
+	RING_IDX i, rp;
+	struct multicall_entry *mcl;
+	int work_done, budget, more_to_do = 1, accel_more_to_do = 1;
+	struct sk_buff_head rxq;
+	struct sk_buff_head errq;
+	struct sk_buff_head tmpq;
+	unsigned long flags;
+	unsigned int len;
+	int pages_flipped = 0;
+	int err;
+
+	spin_lock(&np->rx_lock); /* no need for spin_lock_bh() in ->poll() */
+
+	if (unlikely(!netfront_carrier_ok(np))) {
+		spin_unlock(&np->rx_lock);
+		return 0;
+	}
+
+	skb_queue_head_init(&rxq);
+	skb_queue_head_init(&errq);
+	skb_queue_head_init(&tmpq);
+
+	if ((budget = *pbudget) > dev->quota)
+		budget = dev->quota;
+	rp = np->rx.sring->rsp_prod;
+	rmb(); /* Ensure we see queued responses up to 'rp'. */
+
+	i = np->rx.rsp_cons;
+	work_done = 0;
+	while ((i != rp) && (work_done < budget)) {
+		memcpy(rx, RING_GET_RESPONSE(&np->rx, i), sizeof(*rx));
+		memset(extras, 0, sizeof(rinfo.extras));
+
+		err = xennet_get_responses(np, &rinfo, rp, &tmpq,
+					   &pages_flipped);
+
+		if (unlikely(err)) {
+err:	
+			while ((skb = __skb_dequeue(&tmpq)))
+				__skb_queue_tail(&errq, skb);
+			np->stats.rx_errors++;
+			i = np->rx.rsp_cons;
+			continue;
+		}
+
+		skb = __skb_dequeue(&tmpq);
+
+		if (extras[XEN_NETIF_EXTRA_TYPE_GSO - 1].type) {
+			struct netif_extra_info *gso;
+			gso = &extras[XEN_NETIF_EXTRA_TYPE_GSO - 1];
+
+			if (unlikely(xennet_set_skb_gso(skb, gso))) {
+				__skb_queue_head(&tmpq, skb);
+				np->rx.rsp_cons += skb_queue_len(&tmpq);
+				goto err;
+			}
+		}
+
+		NETFRONT_SKB_CB(skb)->page = skb_shinfo(skb)->frags[0].page;
+		NETFRONT_SKB_CB(skb)->offset = rx->offset;
+
+		len = rx->status;
+		if (len > RX_COPY_THRESHOLD)
+			len = RX_COPY_THRESHOLD;
+		skb_put(skb, len);
+
+		if (rx->status > len) {
+			skb_shinfo(skb)->frags[0].page_offset =
+				rx->offset + len;
+			skb_shinfo(skb)->frags[0].size = rx->status - len;
+			skb->data_len = rx->status - len;
+		} else {
+			skb_shinfo(skb)->frags[0].page = NULL;
+			skb_shinfo(skb)->nr_frags = 0;
+		}
+
+		i = xennet_fill_frags(np, skb, &tmpq);
+
+		/*
+		 * Truesize must approximates the size of true data plus
+		 * any supervisor overheads. Adding hypervisor overheads
+		 * has been shown to significantly reduce achievable
+		 * bandwidth with the default receive buffer size. It is
+		 * therefore not wise to account for it here.
+		 *
+		 * After alloc_skb(RX_COPY_THRESHOLD), truesize is set to
+		 * RX_COPY_THRESHOLD + the supervisor overheads. Here, we
+		 * add the size of the data pulled in xennet_fill_frags().
+		 *
+		 * We also adjust for any unused space in the main data
+		 * area by subtracting (RX_COPY_THRESHOLD - len). This is
+		 * especially important with drivers which split incoming
+		 * packets into header and data, using only 66 bytes of
+		 * the main data area (see the e1000 driver for example.)
+		 * On such systems, without this last adjustement, our
+		 * achievable receive throughout using the standard receive
+		 * buffer size was cut by 25%(!!!).
+		 */
+		skb->truesize += skb->data_len - (RX_COPY_THRESHOLD - len);
+		skb->len += skb->data_len;
+
+		/*
+		 * Old backends do not assert data_validated but we
+		 * can infer it from csum_blank so test both flags.
+		 */
+		if (rx->flags & (NETRXF_data_validated|NETRXF_csum_blank))
+			skb->ip_summed = CHECKSUM_UNNECESSARY;
+		else
+			skb->ip_summed = CHECKSUM_NONE;
+#ifdef CONFIG_XEN
+		skb->proto_data_valid = (skb->ip_summed != CHECKSUM_NONE);
+		skb->proto_csum_blank = !!(rx->flags & NETRXF_csum_blank);
+#endif
+		np->stats.rx_packets++;
+		np->stats.rx_bytes += skb->len;
+
+		__skb_queue_tail(&rxq, skb);
+
+		np->rx.rsp_cons = ++i;
+		work_done++;
+	}
+
+	if (pages_flipped) {
+		/* Some pages are no longer absent... */
+		balloon_update_driver_allowance(-pages_flipped);
+
+		/* Do all the remapping work and M2P updates. */
+		if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+			mcl = np->rx_mcl + pages_flipped;
+			mcl->op = __HYPERVISOR_mmu_update;
+			mcl->args[0] = (unsigned long)np->rx_mmu;
+			mcl->args[1] = pages_flipped;
+			mcl->args[2] = 0;
+			mcl->args[3] = DOMID_SELF;
+			err = HYPERVISOR_multicall_check(np->rx_mcl,
+							 pages_flipped + 1,
+							 NULL);
+			BUG_ON(err);
+		}
+	}
+
+	while ((skb = __skb_dequeue(&errq)))
+		kfree_skb(skb);
+
+	while ((skb = __skb_dequeue(&rxq)) != NULL) {
+		struct page *page = NETFRONT_SKB_CB(skb)->page;
+		void *vaddr = page_address(page);
+		unsigned offset = NETFRONT_SKB_CB(skb)->offset;
+
+		memcpy(skb->data, vaddr + offset, skb_headlen(skb));
+
+		if (page != skb_shinfo(skb)->frags[0].page)
+			__free_page(page);
+
+		/* Ethernet work: Delayed to here as it peeks the header. */
+		skb->protocol = eth_type_trans(skb, dev);
+
+		/* Pass it up. */
+		netif_receive_skb(skb);
+		dev->last_rx = jiffies;
+	}
+
+	/* If we get a callback with very few responses, reduce fill target. */
+	/* NB. Note exponential increase, linear decrease. */
+	if (((np->rx.req_prod_pvt - np->rx.sring->rsp_prod) >
+	     ((3*np->rx_target) / 4)) &&
+	    (--np->rx_target < np->rx_min_target))
+		np->rx_target = np->rx_min_target;
+
+	network_alloc_rx_buffers(dev);
+
+	if (work_done < budget) {
+		/* there's some spare capacity, try the accelerated path */
+		int accel_budget = budget - work_done;
+		int accel_budget_start = accel_budget;
+
+ 		if (np->accel_vif_state.hooks) { 
+ 			accel_more_to_do =  
+ 				np->accel_vif_state.hooks->netdev_poll 
+ 				(dev, &accel_budget); 
+ 			work_done += (accel_budget_start - accel_budget); 
+ 		} else
+			accel_more_to_do = 0;
+	}
+
+	*pbudget   -= work_done;
+	dev->quota -= work_done;
+
+	if (work_done < budget) {
+		local_irq_save(flags);
+
+		RING_FINAL_CHECK_FOR_RESPONSES(&np->rx, more_to_do);
+
+		if (!more_to_do && !accel_more_to_do && 
+		    np->accel_vif_state.hooks) {
+			/* 
+			 *  Slow path has nothing more to do, see if
+			 *  fast path is likewise
+			 */
+			accel_more_to_do = 
+				np->accel_vif_state.hooks->start_napi_irq(dev);
+		}
+
+		if (!more_to_do && !accel_more_to_do)
+			__netif_rx_complete(dev);
+
+		local_irq_restore(flags);
+	}
+
+	spin_unlock(&np->rx_lock);
+	
+	return more_to_do | accel_more_to_do;
+}
+
+static void netif_release_tx_bufs(struct netfront_info *np)
+{
+	struct sk_buff *skb;
+	int i;
+
+	for (i = 1; i <= NET_TX_RING_SIZE; i++) {
+		if ((unsigned long)np->tx_skbs[i] < PAGE_OFFSET)
+			continue;
+
+		skb = np->tx_skbs[i];
+		gnttab_end_foreign_access_ref(np->grant_tx_ref[i]);
+		gnttab_release_grant_reference(
+			&np->gref_tx_head, np->grant_tx_ref[i]);
+		np->grant_tx_ref[i] = GRANT_INVALID_REF;
+		add_id_to_freelist(np->tx_skbs, i);
+		dev_kfree_skb_irq(skb);
+	}
+}
+
+static void netif_release_rx_bufs_flip(struct netfront_info *np)
+{
+	struct mmu_update      *mmu = np->rx_mmu;
+	struct multicall_entry *mcl = np->rx_mcl;
+	struct sk_buff_head free_list;
+	struct sk_buff *skb;
+	unsigned long mfn;
+	int xfer = 0, noxfer = 0, unused = 0;
+	int id, ref, rc;
+
+	skb_queue_head_init(&free_list);
+
+	spin_lock_bh(&np->rx_lock);
+
+	for (id = 0; id < NET_RX_RING_SIZE; id++) {
+		if ((ref = np->grant_rx_ref[id]) == GRANT_INVALID_REF) {
+			unused++;
+			continue;
+		}
+
+		skb = np->rx_skbs[id];
+		mfn = gnttab_end_foreign_transfer_ref(ref);
+		gnttab_release_grant_reference(&np->gref_rx_head, ref);
+		np->grant_rx_ref[id] = GRANT_INVALID_REF;
+		add_id_to_freelist(np->rx_skbs, id);
+
+		if (0 == mfn) {
+			struct page *page = skb_shinfo(skb)->frags[0].page;
+			balloon_release_driver_page(page);
+			skb_shinfo(skb)->nr_frags = 0;
+			dev_kfree_skb(skb);
+			noxfer++;
+			continue;
+		}
+
+		if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+			/* Remap the page. */
+			struct page *page = skb_shinfo(skb)->frags[0].page;
+			unsigned long pfn = page_to_pfn(page);
+			void *vaddr = page_address(page);
+
+			MULTI_update_va_mapping(mcl, (unsigned long)vaddr,
+						pfn_pte_ma(mfn, PAGE_KERNEL),
+						0);
+			mcl++;
+			mmu->ptr = ((maddr_t)mfn << PAGE_SHIFT)
+				| MMU_MACHPHYS_UPDATE;
+			mmu->val = pfn;
+			mmu++;
+
+			set_phys_to_machine(pfn, mfn);
+		}
+		__skb_queue_tail(&free_list, skb);
+		xfer++;
+	}
+
+	DPRINTK("%s: %d xfer, %d noxfer, %d unused\n",
+		__FUNCTION__, xfer, noxfer, unused);
+
+	if (xfer) {
+		/* Some pages are no longer absent... */
+		balloon_update_driver_allowance(-xfer);
+
+		if (!xen_feature(XENFEAT_auto_translated_physmap)) {
+			/* Do all the remapping work and M2P updates. */
+			mcl->op = __HYPERVISOR_mmu_update;
+			mcl->args[0] = (unsigned long)np->rx_mmu;
+			mcl->args[1] = mmu - np->rx_mmu;
+			mcl->args[2] = 0;
+			mcl->args[3] = DOMID_SELF;
+			mcl++;
+			rc = HYPERVISOR_multicall_check(
+				np->rx_mcl, mcl - np->rx_mcl, NULL);
+			BUG_ON(rc);
+		}
+	}
+
+	while ((skb = __skb_dequeue(&free_list)) != NULL)
+		dev_kfree_skb(skb);
+
+	spin_unlock_bh(&np->rx_lock);
+}
+
+static void netif_release_rx_bufs_copy(struct netfront_info *np)
+{
+	struct sk_buff *skb;
+	int i, ref;
+	int busy = 0, inuse = 0;
+
+	spin_lock_bh(&np->rx_lock);
+
+	for (i = 0; i < NET_RX_RING_SIZE; i++) {
+		ref = np->grant_rx_ref[i];
+
+		if (ref == GRANT_INVALID_REF)
+			continue;
+
+		inuse++;
+
+		skb = np->rx_skbs[i];
+
+		if (!gnttab_end_foreign_access_ref(ref))
+		{
+			busy++;
+			continue;
+		}
+
+		gnttab_release_grant_reference(&np->gref_rx_head, ref);
+		np->grant_rx_ref[i] = GRANT_INVALID_REF;
+		add_id_to_freelist(np->rx_skbs, i);
+
+		dev_kfree_skb(skb);
+	}
+
+	if (busy)
+		DPRINTK("%s: Unable to release %d of %d inuse grant references out of %ld total.\n",
+			__FUNCTION__, busy, inuse, NET_RX_RING_SIZE);
+
+	spin_unlock_bh(&np->rx_lock);
+}
+
+static int network_close(struct net_device *dev)
+{
+	struct netfront_info *np = netdev_priv(dev);
+	netif_stop_queue(np->netdev);
+	return 0;
+}
+
+
+static struct net_device_stats *network_get_stats(struct net_device *dev)
+{
+	struct netfront_info *np = netdev_priv(dev);
+
+	netfront_accelerator_call_get_stats(np, dev);
+	return &np->stats;
+}
+
+static int xennet_set_mac_address(struct net_device *dev, void *p)
+{
+	struct netfront_info *np = netdev_priv(dev);
+	struct sockaddr *addr = p;
+
+	if (netif_running(dev))
+		return -EBUSY;
+
+	if (!is_valid_ether_addr(addr->sa_data))
+		return -EADDRNOTAVAIL;
+
+	memcpy(dev->dev_addr, addr->sa_data, dev->addr_len);
+	memcpy(np->mac, addr->sa_data, ETH_ALEN);
+
+	return 0;
+}
+
+static int xennet_change_mtu(struct net_device *dev, int mtu)
+{
+	int max = xennet_can_sg(dev) ? 65535 - ETH_HLEN : ETH_DATA_LEN;
+
+	if (mtu > max)
+		return -EINVAL;
+	dev->mtu = mtu;
+	return 0;
+}
+
+static int xennet_set_sg(struct net_device *dev, u32 data)
+{
+	if (data) {
+		struct netfront_info *np = netdev_priv(dev);
+		int val;
+
+		if (xenbus_scanf(XBT_NIL, np->xbdev->otherend, "feature-sg",
+				 "%d", &val) < 0)
+			val = 0;
+		if (!val)
+			return -ENOSYS;
+	} else if (dev->mtu > ETH_DATA_LEN)
+		dev->mtu = ETH_DATA_LEN;
+
+	return ethtool_op_set_sg(dev, data);
+}
+
+static int xennet_set_tso(struct net_device *dev, u32 data)
+{
+	if (data) {
+		struct netfront_info *np = netdev_priv(dev);
+		int val;
+
+		if (xenbus_scanf(XBT_NIL, np->xbdev->otherend,
+				 "feature-gso-tcpv4", "%d", &val) < 0)
+			val = 0;
+		if (!val)
+			return -ENOSYS;
+	}
+
+	return ethtool_op_set_tso(dev, data);
+}
+
+static void xennet_set_features(struct net_device *dev)
+{
+	dev_disable_gso_features(dev);
+	xennet_set_sg(dev, 0);
+
+	/* We need checksum offload to enable scatter/gather and TSO. */
+	if (!(dev->features & NETIF_F_IP_CSUM))
+		return;
+
+	if (xennet_set_sg(dev, 1))
+		return;
+
+	/* Before 2.6.9 TSO seems to be unreliable so do not enable it
+	 * on older kernels.
+	 */
+	if (LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,9))
+		xennet_set_tso(dev, 1);
+}
+
+static int network_connect(struct net_device *dev)
+{
+	struct netfront_info *np = netdev_priv(dev);
+	int i, requeue_idx, err;
+	struct sk_buff *skb;
+	grant_ref_t ref;
+	netif_rx_request_t *req;
+	unsigned int feature_rx_copy, feature_rx_flip;
+
+	err = xenbus_scanf(XBT_NIL, np->xbdev->otherend,
+			   "feature-rx-copy", "%u", &feature_rx_copy);
+	if (err != 1)
+		feature_rx_copy = 0;
+	err = xenbus_scanf(XBT_NIL, np->xbdev->otherend,
+			   "feature-rx-flip", "%u", &feature_rx_flip);
+	if (err != 1)
+		feature_rx_flip = 1;
+
+	/*
+	 * Copy packets on receive path if:
+	 *  (a) This was requested by user, and the backend supports it; or
+	 *  (b) Flipping was requested, but this is unsupported by the backend.
+	 */
+	np->copying_receiver = ((MODPARM_rx_copy && feature_rx_copy) ||
+				(MODPARM_rx_flip && !feature_rx_flip));
+
+	err = talk_to_backend(np->xbdev, np);
+	if (err)
+		return err;
+
+	xennet_set_features(dev);
+
+	DPRINTK("device %s has %sing receive path.\n",
+		dev->name, np->copying_receiver ? "copy" : "flipp");
+
+	spin_lock_bh(&np->rx_lock);
+	spin_lock_irq(&np->tx_lock);
+
+	/*
+	 * Recovery procedure:
+	 *  NB. Freelist index entries are always going to be less than
+	 *  PAGE_OFFSET, whereas pointers to skbs will always be equal or
+	 *  greater than PAGE_OFFSET: we use this property to distinguish
+	 *  them.
+	 */
+
+	/* Step 1: Discard all pending TX packet fragments. */
+	netif_release_tx_bufs(np);
+
+	/* Step 2: Rebuild the RX buffer freelist and the RX ring itself. */
+	for (requeue_idx = 0, i = 0; i < NET_RX_RING_SIZE; i++) {
+		if (!np->rx_skbs[i])
+			continue;
+
+		skb = np->rx_skbs[requeue_idx] = xennet_get_rx_skb(np, i);
+		ref = np->grant_rx_ref[requeue_idx] = xennet_get_rx_ref(np, i);
+		req = RING_GET_REQUEST(&np->rx, requeue_idx);
+
+		if (!np->copying_receiver) {
+			gnttab_grant_foreign_transfer_ref(
+				ref, np->xbdev->otherend_id,
+				page_to_pfn(skb_shinfo(skb)->frags->page));
+		} else {
+			gnttab_grant_foreign_access_ref(
+				ref, np->xbdev->otherend_id,
+				pfn_to_mfn(page_to_pfn(skb_shinfo(skb)->
+						       frags->page)),
+				0);
+		}
+		req->gref = ref;
+		req->id   = requeue_idx;
+
+		requeue_idx++;
+	}
+
+	np->rx.req_prod_pvt = requeue_idx;
+
+	/*
+	 * Step 3: All public and private state should now be sane.  Get
+	 * ready to start sending and receiving packets and give the driver
+	 * domain a kick because we've probably just requeued some
+	 * packets.
+	 */
+	netfront_carrier_on(np);
+	notify_remote_via_irq(np->irq);
+	network_tx_buf_gc(dev);
+	network_alloc_rx_buffers(dev);
+
+	spin_unlock_irq(&np->tx_lock);
+	spin_unlock_bh(&np->rx_lock);
+
+	return 0;
+}
+
+static void netif_uninit(struct net_device *dev)
+{
+	struct netfront_info *np = netdev_priv(dev);
+	netif_release_tx_bufs(np);
+	if (np->copying_receiver)
+		netif_release_rx_bufs_copy(np);
+	else
+		netif_release_rx_bufs_flip(np);
+	gnttab_free_grant_references(np->gref_tx_head);
+	gnttab_free_grant_references(np->gref_rx_head);
+}
+
+static struct ethtool_ops network_ethtool_ops =
+{
+	.get_tx_csum = ethtool_op_get_tx_csum,
+	.set_tx_csum = ethtool_op_set_tx_csum,
+	.get_sg = ethtool_op_get_sg,
+	.set_sg = xennet_set_sg,
+#if HAVE_TSO
+	.get_tso = ethtool_op_get_tso,
+	.set_tso = xennet_set_tso,
+#endif
+	.get_link = ethtool_op_get_link,
+};
+
+#ifdef CONFIG_SYSFS
+static ssize_t show_rxbuf_min(struct class_device *cd, char *buf)
+{
+	struct net_device *netdev = container_of(cd, struct net_device,
+						 class_dev);
+	struct netfront_info *info = netdev_priv(netdev);
+
+	return sprintf(buf, "%u\n", info->rx_min_target);
+}
+
+static ssize_t store_rxbuf_min(struct class_device *cd,
+			       const char *buf, size_t len)
+{
+	struct net_device *netdev = container_of(cd, struct net_device,
+						 class_dev);
+	struct netfront_info *np = netdev_priv(netdev);
+	char *endp;
+	unsigned long target;
+
+	if (!capable(CAP_NET_ADMIN))
+		return -EPERM;
+
+	target = simple_strtoul(buf, &endp, 0);
+	if (endp == buf)
+		return -EBADMSG;
+
+	if (target < RX_MIN_TARGET)
+		target = RX_MIN_TARGET;
+	if (target > RX_MAX_TARGET)
+		target = RX_MAX_TARGET;
+
+	spin_lock_bh(&np->rx_lock);
+	if (target > np->rx_max_target)
+		np->rx_max_target = target;
+	np->rx_min_target = target;
+	if (target > np->rx_target)
+		np->rx_target = target;
+
+	network_alloc_rx_buffers(netdev);
+
+	spin_unlock_bh(&np->rx_lock);
+	return len;
+}
+
+static ssize_t show_rxbuf_max(struct class_device *cd, char *buf)
+{
+	struct net_device *netdev = container_of(cd, struct net_device,
+						 class_dev);
+	struct netfront_info *info = netdev_priv(netdev);
+
+	return sprintf(buf, "%u\n", info->rx_max_target);
+}
+
+static ssize_t store_rxbuf_max(struct class_device *cd,
+			       const char *buf, size_t len)
+{
+	struct net_device *netdev = container_of(cd, struct net_device,
+						 class_dev);
+	struct netfront_info *np = netdev_priv(netdev);
+	char *endp;
+	unsigned long target;
+
+	if (!capable(CAP_NET_ADMIN))
+		return -EPERM;
+
+	target = simple_strtoul(buf, &endp, 0);
+	if (endp == buf)
+		return -EBADMSG;
+
+	if (target < RX_MIN_TARGET)
+		target = RX_MIN_TARGET;
+	if (target > RX_MAX_TARGET)
+		target = RX_MAX_TARGET;
+
+	spin_lock_bh(&np->rx_lock);
+	if (target < np->rx_min_target)
+		np->rx_min_target = target;
+	np->rx_max_target = target;
+	if (target < np->rx_target)
+		np->rx_target = target;
+
+	network_alloc_rx_buffers(netdev);
+
+	spin_unlock_bh(&np->rx_lock);
+	return len;
+}
+
+static ssize_t show_rxbuf_cur(struct class_device *cd, char *buf)
+{
+	struct net_device *netdev = container_of(cd, struct net_device,
+						 class_dev);
+	struct netfront_info *info = netdev_priv(netdev);
+
+	return sprintf(buf, "%u\n", info->rx_target);
+}
+
+static const struct class_device_attribute xennet_attrs[] = {
+	__ATTR(rxbuf_min, S_IRUGO|S_IWUSR, show_rxbuf_min, store_rxbuf_min),
+	__ATTR(rxbuf_max, S_IRUGO|S_IWUSR, show_rxbuf_max, store_rxbuf_max),
+	__ATTR(rxbuf_cur, S_IRUGO, show_rxbuf_cur, NULL),
+};
+
+static int xennet_sysfs_addif(struct net_device *netdev)
+{
+	int i;
+	int error = 0;
+
+	for (i = 0; i < ARRAY_SIZE(xennet_attrs); i++) {
+		error = class_device_create_file(&netdev->class_dev, 
+						 &xennet_attrs[i]);
+		if (error)
+			goto fail;
+	}
+	return 0;
+
+ fail:
+	while (--i >= 0)
+		class_device_remove_file(&netdev->class_dev,
+					 &xennet_attrs[i]);
+	return error;
+}
+
+static void xennet_sysfs_delif(struct net_device *netdev)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(xennet_attrs); i++) {
+		class_device_remove_file(&netdev->class_dev,
+					 &xennet_attrs[i]);
+	}
+}
+
+#endif /* CONFIG_SYSFS */
+
+
+/*
+ * Nothing to do here. Virtual interface is point-to-point and the
+ * physical interface is probably promiscuous anyway.
+ */
+static void network_set_multicast_list(struct net_device *dev)
+{
+}
+
+static struct net_device * __devinit create_netdev(struct xenbus_device *dev)
+{
+	int i, err = 0;
+	struct net_device *netdev = NULL;
+	struct netfront_info *np = NULL;
+
+	netdev = alloc_etherdev(sizeof(struct netfront_info));
+	if (!netdev) {
+		printk(KERN_WARNING "%s> alloc_etherdev failed.\n",
+		       __FUNCTION__);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	np                   = netdev_priv(netdev);
+	np->xbdev            = dev;
+
+	spin_lock_init(&np->tx_lock);
+	spin_lock_init(&np->rx_lock);
+
+	init_accelerator_vif(np, dev);
+
+	skb_queue_head_init(&np->rx_batch);
+	np->rx_target     = RX_DFL_MIN_TARGET;
+	np->rx_min_target = RX_DFL_MIN_TARGET;
+	np->rx_max_target = RX_MAX_TARGET;
+
+	init_timer(&np->rx_refill_timer);
+	np->rx_refill_timer.data = (unsigned long)netdev;
+	np->rx_refill_timer.function = rx_refill_timeout;
+
+	/* Initialise {tx,rx}_skbs as a free chain containing every entry. */
+	for (i = 0; i <= NET_TX_RING_SIZE; i++) {
+		np->tx_skbs[i] = (void *)((unsigned long) i+1);
+		np->grant_tx_ref[i] = GRANT_INVALID_REF;
+	}
+
+	for (i = 0; i < NET_RX_RING_SIZE; i++) {
+		np->rx_skbs[i] = NULL;
+		np->grant_rx_ref[i] = GRANT_INVALID_REF;
+	}
+
+	/* A grant for every tx ring slot */
+	if (gnttab_alloc_grant_references(TX_MAX_TARGET,
+					  &np->gref_tx_head) < 0) {
+		printk(KERN_ALERT "#### netfront can't alloc tx grant refs\n");
+		err = -ENOMEM;
+		goto exit;
+	}
+	/* A grant for every rx ring slot */
+	if (gnttab_alloc_grant_references(RX_MAX_TARGET,
+					  &np->gref_rx_head) < 0) {
+		printk(KERN_ALERT "#### netfront can't alloc rx grant refs\n");
+		err = -ENOMEM;
+		goto exit_free_tx;
+	}
+
+	netdev->open            = network_open;
+	netdev->hard_start_xmit = network_start_xmit;
+	netdev->stop            = network_close;
+	netdev->get_stats       = network_get_stats;
+	netdev->poll            = netif_poll;
+	netdev->set_multicast_list = network_set_multicast_list;
+	netdev->uninit          = netif_uninit;
+	netdev->set_mac_address	= xennet_set_mac_address;
+	netdev->change_mtu	= xennet_change_mtu;
+	netdev->weight          = 64;
+	netdev->features        = NETIF_F_IP_CSUM;
+
+	SET_ETHTOOL_OPS(netdev, &network_ethtool_ops);
+	SET_MODULE_OWNER(netdev);
+	SET_NETDEV_DEV(netdev, &dev->dev);
+
+	np->netdev = netdev;
+
+	netfront_carrier_off(np);
+
+	return netdev;
+
+ exit_free_tx:
+	gnttab_free_grant_references(np->gref_tx_head);
+ exit:
+	free_netdev(netdev);
+	return ERR_PTR(err);
+}
+
+#ifdef CONFIG_INET
+/*
+ * We use this notifier to send out a fake ARP reply to reset switches and
+ * router ARP caches when an IP interface is brought up on a VIF.
+ */
+static int
+inetdev_notify(struct notifier_block *this, unsigned long event, void *ptr)
+{
+	struct in_ifaddr  *ifa = (struct in_ifaddr *)ptr;
+	struct net_device *dev = ifa->ifa_dev->dev;
+
+	/* UP event and is it one of our devices? */
+	if (event == NETDEV_UP && dev->open == network_open)
+		send_fake_arp(dev);
+
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block notifier_inetdev = {
+	.notifier_call  = inetdev_notify,
+	.next           = NULL,
+	.priority       = 0
+};
+#endif
+
+
+static void netif_disconnect_backend(struct netfront_info *info)
+{
+	/* Stop old i/f to prevent errors whilst we rebuild the state. */
+	spin_lock_bh(&info->rx_lock);
+	spin_lock_irq(&info->tx_lock);
+	netfront_carrier_off(info);
+	spin_unlock_irq(&info->tx_lock);
+	spin_unlock_bh(&info->rx_lock);
+
+	if (info->irq)
+		unbind_from_irqhandler(info->irq, info->netdev);
+	info->irq = 0;
+
+	end_access(info->tx_ring_ref, info->tx.sring);
+	end_access(info->rx_ring_ref, info->rx.sring);
+	info->tx_ring_ref = GRANT_INVALID_REF;
+	info->rx_ring_ref = GRANT_INVALID_REF;
+	info->tx.sring = NULL;
+	info->rx.sring = NULL;
+}
+
+
+static void end_access(int ref, void *page)
+{
+	if (ref != GRANT_INVALID_REF)
+		gnttab_end_foreign_access(ref, (unsigned long)page);
+}
+
+
+/* ** Driver registration ** */
+
+
+static const struct xenbus_device_id netfront_ids[] = {
+	{ "vif" },
+	{ "" }
+};
+MODULE_ALIAS("xen:vif");
+
+
+static struct xenbus_driver netfront_driver = {
+	.name = "vif",
+	.owner = THIS_MODULE,
+	.ids = netfront_ids,
+	.probe = netfront_probe,
+	.remove = __devexit_p(netfront_remove),
+	.suspend = netfront_suspend,
+	.suspend_cancel = netfront_suspend_cancel,
+	.resume = netfront_resume,
+	.otherend_changed = backend_changed,
+};
+
+
+static int __init netif_init(void)
+{
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+#ifdef CONFIG_XEN
+	if (MODPARM_rx_flip && MODPARM_rx_copy) {
+		WPRINTK("Cannot specify both rx_copy and rx_flip.\n");
+		return -EINVAL;
+	}
+
+	if (!MODPARM_rx_flip && !MODPARM_rx_copy)
+		MODPARM_rx_flip = 1; /* Default is to flip. */
+#endif
+
+	netif_init_accel();
+
+	IPRINTK("Initialising virtual ethernet driver.\n");
+
+#ifdef CONFIG_INET
+	(void)register_inetaddr_notifier(&notifier_inetdev);
+#endif
+
+	return xenbus_register_frontend(&netfront_driver);
+}
+module_init(netif_init);
+
+
+static void __exit netif_exit(void)
+{
+#ifdef CONFIG_INET
+	unregister_inetaddr_notifier(&notifier_inetdev);
+#endif
+
+	netif_exit_accel();
+
+	return xenbus_unregister_driver(&netfront_driver);
+}
+module_exit(netif_exit);
+
+MODULE_LICENSE("Dual BSD/GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/netfront/netfront.h linux-2.6.18-xen-3.3.0/drivers/xen/netfront/netfront.h
--- linux-2.6.18.8/drivers/xen/netfront/netfront.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/netfront/netfront.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,274 @@
+/******************************************************************************
+ * Virtual network driver for conversing with remote driver backends.
+ *
+ * Copyright (c) 2002-2005, K A Fraser
+ * Copyright (c) 2005, XenSource Ltd
+ * Copyright (C) 2007 Solarflare Communications, Inc.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef NETFRONT_H
+#define NETFRONT_H
+
+#include <xen/interface/io/netif.h>
+#include <linux/netdevice.h>
+#include <linux/skbuff.h>
+#include <linux/list.h>
+
+#define NET_TX_RING_SIZE __RING_SIZE((struct netif_tx_sring *)0, PAGE_SIZE)
+#define NET_RX_RING_SIZE __RING_SIZE((struct netif_rx_sring *)0, PAGE_SIZE)
+
+#include <xen/xenbus.h>
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+/* 
+ * Function pointer table for hooks into a network acceleration
+ * plugin.  These are called at appropriate points from the netfront
+ * driver 
+ */
+struct netfront_accel_hooks {
+	/* 
+	 * new_device: Accelerator hook to ask the plugin to support a
+	 * new network interface
+	 */
+	int (*new_device)(struct net_device *net_dev, struct xenbus_device *dev);
+	/*
+	 * remove: Opposite of new_device
+	 */
+	int (*remove)(struct xenbus_device *dev);
+	/*
+	 * The net_device is being polled, check the accelerated
+	 * hardware for any pending packets
+	 */
+	int (*netdev_poll)(struct net_device *dev, int *pbudget);
+	/*
+	 * start_xmit: Used to give the accelerated plugin the option
+	 * of sending a packet.  Returns non-zero if has done so, or
+	 * zero to decline and force the packet onto normal send
+	 * path
+	 */
+	int (*start_xmit)(struct sk_buff *skb, struct net_device *dev);
+	/* 
+	 * start/stop_napi_interrupts Used by netfront to indicate
+	 * when napi interrupts should be enabled or disabled 
+	 */
+	int (*start_napi_irq)(struct net_device *dev);
+	void (*stop_napi_irq)(struct net_device *dev);
+	/* 
+	 * Called before re-enabling the TX queue to check the fast
+	 * path has slots too
+	 */
+	int (*check_ready)(struct net_device *dev);
+	/*
+	 * Get the fastpath network statistics
+	 */
+	int (*get_stats)(struct net_device *dev,
+			 struct net_device_stats *stats);
+};
+
+
+/* Version of API/protocol for communication between netfront and
+   acceleration plugin supported */
+#define NETFRONT_ACCEL_VERSION 0x00010003
+
+/* 
+ * Per-netfront device state for the accelerator.  This is used to
+ * allow efficient per-netfront device access to the accelerator
+ * hooks 
+ */
+struct netfront_accel_vif_state {
+	struct list_head link;
+
+	struct xenbus_device *dev;
+	struct netfront_info *np;
+	struct netfront_accel_hooks *hooks;
+
+	/* Watch on the accelerator configuration value */
+	struct xenbus_watch accel_watch;
+	/* Work item to process change in accelerator */
+	struct work_struct accel_work;
+	/* The string from xenbus last time accel_watch fired */
+	char *accel_frontend;
+}; 
+
+/* 
+ * Per-accelerator state stored in netfront.  These form a list that
+ * is used to track which devices are accelerated by which plugins,
+ * and what plugins are available/have been requested 
+ */
+struct netfront_accelerator {
+	/* Used to make a list */
+	struct list_head link;
+	/* ID of the accelerator */
+	int id;
+	/*
+	 * String describing the accelerator.  Currently this is the
+	 * name of the accelerator module.  This is provided by the
+	 * backend accelerator through xenstore 
+	 */
+	char *frontend;
+	/* The hooks into the accelerator plugin module */
+	struct netfront_accel_hooks *hooks;
+
+	/* 
+	 * List of per-netfront device state (struct
+	 * netfront_accel_vif_state) for each netfront device that is
+	 * using this accelerator
+	 */
+	struct list_head vif_states;
+	spinlock_t vif_states_lock;
+};
+
+struct netfront_info {
+	struct list_head list;
+	struct net_device *netdev;
+
+	struct net_device_stats stats;
+
+	struct netif_tx_front_ring tx;
+	struct netif_rx_front_ring rx;
+
+	spinlock_t   tx_lock;
+	spinlock_t   rx_lock;
+
+	unsigned int irq;
+	unsigned int copying_receiver;
+	unsigned int carrier;
+
+	/* Receive-ring batched refills. */
+#define RX_MIN_TARGET 8
+#define RX_DFL_MIN_TARGET 64
+#define RX_MAX_TARGET min_t(int, NET_RX_RING_SIZE, 256)
+	unsigned rx_min_target, rx_max_target, rx_target;
+	struct sk_buff_head rx_batch;
+
+	struct timer_list rx_refill_timer;
+
+	/*
+	 * {tx,rx}_skbs store outstanding skbuffs. The first entry in tx_skbs
+	 * is an index into a chain of free entries.
+	 */
+	struct sk_buff *tx_skbs[NET_TX_RING_SIZE+1];
+	struct sk_buff *rx_skbs[NET_RX_RING_SIZE];
+
+#define TX_MAX_TARGET min_t(int, NET_RX_RING_SIZE, 256)
+	grant_ref_t gref_tx_head;
+	grant_ref_t grant_tx_ref[NET_TX_RING_SIZE + 1];
+	grant_ref_t gref_rx_head;
+	grant_ref_t grant_rx_ref[NET_RX_RING_SIZE];
+
+	struct xenbus_device *xbdev;
+	int tx_ring_ref;
+	int rx_ring_ref;
+	u8 mac[ETH_ALEN];
+
+	unsigned long rx_pfn_array[NET_RX_RING_SIZE];
+	struct multicall_entry rx_mcl[NET_RX_RING_SIZE+1];
+	struct mmu_update rx_mmu[NET_RX_RING_SIZE];
+
+	/* Private pointer to state internal to accelerator module */
+	void *accel_priv;
+	/* The accelerator used by this netfront device */
+	struct netfront_accelerator *accelerator;
+	/* The accelerator state for this netfront device */
+	struct netfront_accel_vif_state accel_vif_state;
+};
+
+
+/* Exported Functions */
+
+/*
+ * Called by an accelerator plugin module when it has loaded.
+ *
+ * frontend: the string describing the accelerator, currently the module name 
+ * hooks: the hooks for netfront to use to call into the accelerator
+ * version: the version of API between frontend and plugin requested
+ * 
+ * return: 0 on success, <0 on error, >0 (with version supported) on
+ * version mismatch
+ */
+extern int netfront_accelerator_loaded(int version, const char *frontend, 
+				       struct netfront_accel_hooks *hooks);
+
+/* 
+ * Called by an accelerator plugin module when it is about to unload.
+ *
+ * frontend: the string describing the accelerator.  Must match the
+ * one passed to netfront_accelerator_loaded()
+ */ 
+extern void netfront_accelerator_stop(const char *frontend);
+
+/* 
+ * Called by an accelerator before waking the net device's TX queue to
+ * ensure the slow path has available slots.  Returns true if OK to
+ * wake, false if still busy 
+ */
+extern int netfront_check_queue_ready(struct net_device *net_dev);
+
+
+/* Internal-to-netfront Functions */
+
+/* 
+ * Call into accelerator and check to see if it has tx space before we
+ * wake the net device's TX queue.  Returns true if OK to wake, false
+ * if still busy
+ */ 
+extern 
+int netfront_check_accelerator_queue_ready(struct net_device *dev,
+					   struct netfront_info *np);
+extern
+int netfront_accelerator_call_remove(struct netfront_info *np,
+				     struct xenbus_device *dev);
+extern
+int netfront_accelerator_suspend(struct netfront_info *np,
+				 struct xenbus_device *dev);
+extern
+int netfront_accelerator_suspend_cancel(struct netfront_info *np,
+					struct xenbus_device *dev);
+extern
+void netfront_accelerator_resume(struct netfront_info *np,
+				 struct xenbus_device *dev);
+extern
+void netfront_accelerator_call_stop_napi_irq(struct netfront_info *np,
+					     struct net_device *dev);
+extern
+int netfront_accelerator_call_get_stats(struct netfront_info *np,
+					struct net_device *dev);
+extern
+void netfront_accelerator_add_watch(struct netfront_info *np);
+
+extern
+void netif_init_accel(void);
+extern
+void netif_exit_accel(void);
+
+extern
+void init_accelerator_vif(struct netfront_info *np,
+			  struct xenbus_device *dev);
+#endif /* NETFRONT_H */
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/conf_space.c linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space.c
--- linux-2.6.18.8/drivers/xen/pciback/conf_space.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,426 @@
+/*
+ * PCI Backend - Functions for creating a virtual configuration space for
+ *               exported PCI Devices.
+ *               It's dangerous to allow PCI Driver Domains to change their
+ *               device's resources (memory, i/o ports, interrupts). We need to
+ *               restrict changes to certain PCI Configuration registers:
+ *               BARs, INTERRUPT_PIN, most registers in the header...
+ *
+ * Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+
+#include <linux/kernel.h>
+#include <linux/pci.h>
+#include "pciback.h"
+#include "conf_space.h"
+#include "conf_space_quirks.h"
+
+#define DEFINE_PCI_CONFIG(op,size,type) 			\
+int pciback_##op##_config_##size 				\
+(struct pci_dev *dev, int offset, type value, void *data)	\
+{								\
+	return pci_##op##_config_##size (dev, offset, value);	\
+}
+
+DEFINE_PCI_CONFIG(read, byte, u8 *)
+DEFINE_PCI_CONFIG(read, word, u16 *)
+DEFINE_PCI_CONFIG(read, dword, u32 *)
+
+DEFINE_PCI_CONFIG(write, byte, u8)
+DEFINE_PCI_CONFIG(write, word, u16)
+DEFINE_PCI_CONFIG(write, dword, u32)
+
+static int conf_space_read(struct pci_dev *dev,
+			   struct config_field_entry *entry, int offset,
+			   u32 * value)
+{
+	int ret = 0;
+	struct config_field *field = entry->field;
+
+	*value = 0;
+
+	switch (field->size) {
+	case 1:
+		if (field->u.b.read)
+			ret = field->u.b.read(dev, offset, (u8 *) value,
+					      entry->data);
+		break;
+	case 2:
+		if (field->u.w.read)
+			ret = field->u.w.read(dev, offset, (u16 *) value,
+					      entry->data);
+		break;
+	case 4:
+		if (field->u.dw.read)
+			ret = field->u.dw.read(dev, offset, value, entry->data);
+		break;
+	}
+	return ret;
+}
+
+static int conf_space_write(struct pci_dev *dev,
+			    struct config_field_entry *entry, int offset,
+			    u32 value)
+{
+	int ret = 0;
+	struct config_field *field = entry->field;
+
+	switch (field->size) {
+	case 1:
+		if (field->u.b.write)
+			ret = field->u.b.write(dev, offset, (u8) value,
+					       entry->data);
+		break;
+	case 2:
+		if (field->u.w.write)
+			ret = field->u.w.write(dev, offset, (u16) value,
+					       entry->data);
+		break;
+	case 4:
+		if (field->u.dw.write)
+			ret = field->u.dw.write(dev, offset, value,
+						entry->data);
+		break;
+	}
+	return ret;
+}
+
+static inline u32 get_mask(int size)
+{
+	if (size == 1)
+		return 0xff;
+	else if (size == 2)
+		return 0xffff;
+	else
+		return 0xffffffff;
+}
+
+static inline int valid_request(int offset, int size)
+{
+	/* Validate request (no un-aligned requests) */
+	if ((size == 1 || size == 2 || size == 4) && (offset % size) == 0)
+		return 1;
+	return 0;
+}
+
+static inline u32 merge_value(u32 val, u32 new_val, u32 new_val_mask,
+			      int offset)
+{
+	if (offset >= 0) {
+		new_val_mask <<= (offset * 8);
+		new_val <<= (offset * 8);
+	} else {
+		new_val_mask >>= (offset * -8);
+		new_val >>= (offset * -8);
+	}
+	val = (val & ~new_val_mask) | (new_val & new_val_mask);
+
+	return val;
+}
+
+static int pcibios_err_to_errno(int err)
+{
+	switch (err) {
+	case PCIBIOS_SUCCESSFUL:
+		return XEN_PCI_ERR_success;
+	case PCIBIOS_DEVICE_NOT_FOUND:
+		return XEN_PCI_ERR_dev_not_found;
+	case PCIBIOS_BAD_REGISTER_NUMBER:
+		return XEN_PCI_ERR_invalid_offset;
+	case PCIBIOS_FUNC_NOT_SUPPORTED:
+		return XEN_PCI_ERR_not_implemented;
+	case PCIBIOS_SET_FAILED:
+		return XEN_PCI_ERR_access_denied;
+	}
+	return err;
+}
+
+int pciback_config_read(struct pci_dev *dev, int offset, int size,
+			u32 * ret_val)
+{
+	int err = 0;
+	struct pciback_dev_data *dev_data = pci_get_drvdata(dev);
+	struct config_field_entry *cfg_entry;
+	struct config_field *field;
+	int req_start, req_end, field_start, field_end;
+	/* if read fails for any reason, return 0 (as if device didn't respond) */
+	u32 value = 0, tmp_val;
+
+	if (unlikely(verbose_request))
+		printk(KERN_DEBUG "pciback: %s: read %d bytes at 0x%x\n",
+		       pci_name(dev), size, offset);
+
+	if (!valid_request(offset, size)) {
+		err = XEN_PCI_ERR_invalid_offset;
+		goto out;
+	}
+
+	/* Get the real value first, then modify as appropriate */
+	switch (size) {
+	case 1:
+		err = pci_read_config_byte(dev, offset, (u8 *) & value);
+		break;
+	case 2:
+		err = pci_read_config_word(dev, offset, (u16 *) & value);
+		break;
+	case 4:
+		err = pci_read_config_dword(dev, offset, &value);
+		break;
+	}
+
+	list_for_each_entry(cfg_entry, &dev_data->config_fields, list) {
+		field = cfg_entry->field;
+
+		req_start = offset;
+		req_end = offset + size;
+		field_start = OFFSET(cfg_entry);
+		field_end = OFFSET(cfg_entry) + field->size;
+
+		if ((req_start >= field_start && req_start < field_end)
+		    || (req_end > field_start && req_end <= field_end)) {
+			err = conf_space_read(dev, cfg_entry, field_start,
+					      &tmp_val);
+			if (err)
+				goto out;
+
+			value = merge_value(value, tmp_val,
+					    get_mask(field->size),
+					    field_start - req_start);
+		}
+	}
+
+      out:
+	if (unlikely(verbose_request))
+		printk(KERN_DEBUG "pciback: %s: read %d bytes at 0x%x = %x\n",
+		       pci_name(dev), size, offset, value);
+
+	*ret_val = value;
+	return pcibios_err_to_errno(err);
+}
+
+int pciback_config_write(struct pci_dev *dev, int offset, int size, u32 value)
+{
+	int err = 0, handled = 0;
+	struct pciback_dev_data *dev_data = pci_get_drvdata(dev);
+	struct config_field_entry *cfg_entry;
+	struct config_field *field;
+	u32 tmp_val;
+	int req_start, req_end, field_start, field_end;
+
+	if (unlikely(verbose_request))
+		printk(KERN_DEBUG
+		       "pciback: %s: write request %d bytes at 0x%x = %x\n",
+		       pci_name(dev), size, offset, value);
+
+	if (!valid_request(offset, size))
+		return XEN_PCI_ERR_invalid_offset;
+
+	list_for_each_entry(cfg_entry, &dev_data->config_fields, list) {
+		field = cfg_entry->field;
+
+		req_start = offset;
+		req_end = offset + size;
+		field_start = OFFSET(cfg_entry);
+		field_end = OFFSET(cfg_entry) + field->size;
+
+		if ((req_start >= field_start && req_start < field_end)
+		    || (req_end > field_start && req_end <= field_end)) {
+			tmp_val = 0;
+
+			err = pciback_config_read(dev, field_start,
+						  field->size, &tmp_val);
+			if (err)
+				break;
+
+			tmp_val = merge_value(tmp_val, value, get_mask(size),
+					      req_start - field_start);
+
+			err = conf_space_write(dev, cfg_entry, field_start,
+					       tmp_val);
+
+			/* handled is set true here, but not every byte
+			 * may have been written! Properly detecting if
+			 * every byte is handled is unnecessary as the
+			 * flag is used to detect devices that need
+			 * special helpers to work correctly.
+			 */
+			handled = 1;
+		}
+	}
+
+	if (!handled && !err) {
+		/* By default, anything not specificially handled above is
+		 * read-only. The permissive flag changes this behavior so
+		 * that anything not specifically handled above is writable.
+		 * This means that some fields may still be read-only because
+		 * they have entries in the config_field list that intercept
+		 * the write and do nothing. */
+		if (dev_data->permissive) {
+			switch (size) {
+			case 1:
+				err = pci_write_config_byte(dev, offset,
+							    (u8) value);
+				break;
+			case 2:
+				err = pci_write_config_word(dev, offset,
+							    (u16) value);
+				break;
+			case 4:
+				err = pci_write_config_dword(dev, offset,
+							     (u32) value);
+				break;
+			}
+		} else if (!dev_data->warned_on_write) {
+			dev_data->warned_on_write = 1;
+			dev_warn(&dev->dev, "Driver tried to write to a "
+				 "read-only configuration space field at offset "
+				 "0x%x, size %d. This may be harmless, but if "
+				 "you have problems with your device:\n"
+				 "1) see permissive attribute in sysfs\n"
+				 "2) report problems to the xen-devel "
+				 "mailing list along with details of your "
+				 "device obtained from lspci.\n", offset, size);
+		}
+	}
+
+	return pcibios_err_to_errno(err);
+}
+
+void pciback_config_free_dyn_fields(struct pci_dev *dev)
+{
+	struct pciback_dev_data *dev_data = pci_get_drvdata(dev);
+	struct config_field_entry *cfg_entry, *t;
+	struct config_field *field;
+
+	dev_dbg(&dev->dev,
+		"free-ing dynamically allocated virtual configuration space fields\n");
+
+	list_for_each_entry_safe(cfg_entry, t, &dev_data->config_fields, list) {
+		field = cfg_entry->field;
+
+		if (field->clean) {
+			field->clean(field);
+
+			if (cfg_entry->data)
+				kfree(cfg_entry->data);
+
+			list_del(&cfg_entry->list);
+			kfree(cfg_entry);
+		}
+
+	}
+}
+
+void pciback_config_reset_dev(struct pci_dev *dev)
+{
+	struct pciback_dev_data *dev_data = pci_get_drvdata(dev);
+	struct config_field_entry *cfg_entry;
+	struct config_field *field;
+
+	dev_dbg(&dev->dev, "resetting virtual configuration space\n");
+
+	list_for_each_entry(cfg_entry, &dev_data->config_fields, list) {
+		field = cfg_entry->field;
+
+		if (field->reset)
+			field->reset(dev, OFFSET(cfg_entry), cfg_entry->data);
+	}
+}
+
+void pciback_config_free_dev(struct pci_dev *dev)
+{
+	struct pciback_dev_data *dev_data = pci_get_drvdata(dev);
+	struct config_field_entry *cfg_entry, *t;
+	struct config_field *field;
+
+	dev_dbg(&dev->dev, "free-ing virtual configuration space fields\n");
+
+	list_for_each_entry_safe(cfg_entry, t, &dev_data->config_fields, list) {
+		list_del(&cfg_entry->list);
+
+		field = cfg_entry->field;
+
+		if (field->release)
+			field->release(dev, OFFSET(cfg_entry), cfg_entry->data);
+
+		kfree(cfg_entry);
+	}
+}
+
+int pciback_config_add_field_offset(struct pci_dev *dev,
+				    struct config_field *field,
+				    unsigned int base_offset)
+{
+	int err = 0;
+	struct pciback_dev_data *dev_data = pci_get_drvdata(dev);
+	struct config_field_entry *cfg_entry;
+	void *tmp;
+
+	cfg_entry = kmalloc(sizeof(*cfg_entry), GFP_KERNEL);
+	if (!cfg_entry) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	cfg_entry->data = NULL;
+	cfg_entry->field = field;
+	cfg_entry->base_offset = base_offset;
+
+	/* silently ignore duplicate fields */
+	err = pciback_field_is_dup(dev,OFFSET(cfg_entry));
+	if (err)
+		goto out;
+
+	if (field->init) {
+		tmp = field->init(dev, OFFSET(cfg_entry));
+
+		if (IS_ERR(tmp)) {
+			err = PTR_ERR(tmp);
+			goto out;
+		}
+
+		cfg_entry->data = tmp;
+	}
+
+	dev_dbg(&dev->dev, "added config field at offset 0x%02x\n",
+		OFFSET(cfg_entry));
+	list_add_tail(&cfg_entry->list, &dev_data->config_fields);
+
+      out:
+	if (err)
+		kfree(cfg_entry);
+
+	return err;
+}
+
+/* This sets up the device's virtual configuration space to keep track of 
+ * certain registers (like the base address registers (BARs) so that we can
+ * keep the client from manipulating them directly.
+ */
+int pciback_config_init_dev(struct pci_dev *dev)
+{
+	int err = 0;
+	struct pciback_dev_data *dev_data = pci_get_drvdata(dev);
+
+	dev_dbg(&dev->dev, "initializing virtual configuration space\n");
+
+	INIT_LIST_HEAD(&dev_data->config_fields);
+
+	err = pciback_config_header_add_fields(dev);
+	if (err)
+		goto out;
+
+	err = pciback_config_capability_add_fields(dev);
+	if (err)
+		goto out;
+
+	err = pciback_config_quirks_init(dev);
+
+      out:
+	return err;
+}
+
+int pciback_config_init(void)
+{
+	return pciback_config_capability_init();
+}
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/conf_space_capability.c linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space_capability.c
--- linux-2.6.18.8/drivers/xen/pciback/conf_space_capability.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space_capability.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,71 @@
+/*
+ * PCI Backend - Handles the virtual fields found on the capability lists
+ *               in the configuration space.
+ *
+ * Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+
+#include <linux/kernel.h>
+#include <linux/pci.h>
+#include "pciback.h"
+#include "conf_space.h"
+#include "conf_space_capability.h"
+
+static LIST_HEAD(capabilities);
+
+static struct config_field caplist_header[] = {
+	{
+	 .offset    = PCI_CAP_LIST_ID,
+	 .size      = 2, /* encompass PCI_CAP_LIST_ID & PCI_CAP_LIST_NEXT */
+	 .u.w.read  = pciback_read_config_word,
+	 .u.w.write = NULL,
+	},
+	{
+	 .size = 0,
+	},
+};
+
+static inline void register_capability(struct pciback_config_capability *cap)
+{
+	list_add_tail(&cap->cap_list, &capabilities);
+}
+
+int pciback_config_capability_add_fields(struct pci_dev *dev)
+{
+	int err = 0;
+	struct pciback_config_capability *cap;
+	int cap_offset;
+
+	list_for_each_entry(cap, &capabilities, cap_list) {
+		cap_offset = pci_find_capability(dev, cap->capability);
+		if (cap_offset) {
+			dev_dbg(&dev->dev, "Found capability 0x%x at 0x%x\n",
+				cap->capability, cap_offset);
+
+			err = pciback_config_add_fields_offset(dev,
+							       caplist_header,
+							       cap_offset);
+			if (err)
+				goto out;
+			err = pciback_config_add_fields_offset(dev,
+							       cap->fields,
+							       cap_offset);
+			if (err)
+				goto out;
+		}
+	}
+
+      out:
+	return err;
+}
+
+extern struct pciback_config_capability pciback_config_capability_vpd;
+extern struct pciback_config_capability pciback_config_capability_pm;
+
+int pciback_config_capability_init(void)
+{
+	register_capability(&pciback_config_capability_vpd);
+	register_capability(&pciback_config_capability_pm);
+
+	return 0;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/conf_space_capability.h linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space_capability.h
--- linux-2.6.18.8/drivers/xen/pciback/conf_space_capability.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space_capability.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,23 @@
+/*
+ * PCI Backend - Data structures for special overlays for structures on
+ *               the capability list.
+ *
+ * Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+
+#ifndef __PCIBACK_CONFIG_CAPABILITY_H__
+#define __PCIBACK_CONFIG_CAPABILITY_H__
+
+#include <linux/pci.h>
+#include <linux/list.h>
+
+struct pciback_config_capability {
+	struct list_head cap_list;
+
+	int capability;
+
+	/* If the device has the capability found above, add these fields */
+	struct config_field *fields;
+};
+
+#endif
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/conf_space_capability_msi.c linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space_capability_msi.c
--- linux-2.6.18.8/drivers/xen/pciback/conf_space_capability_msi.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space_capability_msi.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,69 @@
+/*
+ * PCI Backend -- Configuration overlay for MSI capability
+ */
+#include <linux/pci.h>
+#include "conf_space.h"
+#include "conf_space_capability.h"
+#include <xen/interface/io/pciif.h>
+#include "pciback.h"
+
+int pciback_enable_msi(struct pciback_device *pdev,
+		struct pci_dev *dev, struct xen_pci_op *op)
+{
+	int otherend = pdev->xdev->otherend_id;
+	int status;
+
+	status = pci_enable_msi(dev);
+
+	if (status) {
+		printk("error enable msi for guest %x status %x\n", otherend, status);
+		op->value = 0;
+		return XEN_PCI_ERR_op_failed;
+	}
+
+	op->value = dev->irq;
+	return 0;
+}
+
+int pciback_disable_msi(struct pciback_device *pdev,
+		struct pci_dev *dev, struct xen_pci_op *op)
+{
+	pci_disable_msi(dev);
+
+	op->value = dev->irq;
+	return 0;
+}
+
+int pciback_enable_msix(struct pciback_device *pdev,
+		struct pci_dev *dev, struct xen_pci_op *op)
+{
+	int result;
+
+	if (op->value > SH_INFO_MAX_VEC)
+		return -EINVAL;
+	else {
+		struct msix_entry entries[op->value];
+		int i;
+
+		for (i = 0; i < op->value; i++) {
+			entries[i].entry = op->msix_entries[i].entry;
+			entries[i].vector = op->msix_entries[i].vector;
+		}
+
+		result = pci_enable_msix(dev, entries, op->value);
+		op->value = result;
+	}
+
+	return result;
+}
+
+int pciback_disable_msix(struct pciback_device *pdev,
+		struct pci_dev *dev, struct xen_pci_op *op)
+{
+
+	pci_disable_msix(dev);
+
+	op->value = dev->irq;
+	return 0;
+}
+
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/conf_space_capability_pm.c linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space_capability_pm.c
--- linux-2.6.18.8/drivers/xen/pciback/conf_space_capability_pm.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space_capability_pm.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,128 @@
+/*
+ * PCI Backend - Configuration space overlay for power management
+ *
+ * Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+
+#include <linux/pci.h>
+#include "conf_space.h"
+#include "conf_space_capability.h"
+
+static int pm_caps_read(struct pci_dev *dev, int offset, u16 *value,
+			void *data)
+{
+	int err;
+	u16 real_value;
+
+	err = pci_read_config_word(dev, offset, &real_value);
+	if (err)
+		goto out;
+
+	*value = real_value & ~PCI_PM_CAP_PME_MASK;
+
+      out:
+	return err;
+}
+
+/* PM_OK_BITS specifies the bits that the driver domain is allowed to change.
+ * Can't allow driver domain to enable PMEs - they're shared */
+#define PM_OK_BITS (PCI_PM_CTRL_PME_STATUS|PCI_PM_CTRL_DATA_SEL_MASK)
+
+static int pm_ctrl_write(struct pci_dev *dev, int offset, u16 new_value,
+			 void *data)
+{
+	int err;
+	u16 old_value;
+	pci_power_t new_state, old_state;
+
+	err = pci_read_config_word(dev, offset, &old_value);
+	if (err)
+		goto out;
+
+	old_state = (pci_power_t)(old_value & PCI_PM_CTRL_STATE_MASK);
+	new_state = (pci_power_t)(new_value & PCI_PM_CTRL_STATE_MASK);
+
+	new_value &= PM_OK_BITS;
+	if ((old_value & PM_OK_BITS) != new_value) {
+		new_value = (old_value & ~PM_OK_BITS) | new_value;
+		err = pci_write_config_word(dev, offset, new_value);
+		if (err)
+			goto out;
+	}
+
+	/* Let pci core handle the power management change */
+	dev_dbg(&dev->dev, "set power state to %x\n", new_state);
+	err = pci_set_power_state(dev, new_state);
+	if (err) {
+		err = PCIBIOS_SET_FAILED;
+		goto out;
+	}
+
+	/*
+	 * Device may lose PCI config info on D3->D0 transition. This
+	 * is a problem for some guests which will not reset BARs. Even
+	 * those that have a go will be foiled by our BAR-write handler
+	 * which will discard the write! Since Linux won't re-init
+	 * the config space automatically in all cases, we do it here.
+	 * Future: Should we re-initialise all first 64 bytes of config space?
+	 */
+	if (new_state == PCI_D0 &&
+	    (old_state == PCI_D3hot || old_state == PCI_D3cold) &&
+	    !(old_value & PCI_PM_CTRL_NO_SOFT_RESET))
+		pci_restore_bars(dev);
+
+ out:
+	return err;
+}
+
+/* Ensure PMEs are disabled */
+static void *pm_ctrl_init(struct pci_dev *dev, int offset)
+{
+	int err;
+	u16 value;
+
+	err = pci_read_config_word(dev, offset, &value);
+	if (err)
+		goto out;
+
+	if (value & PCI_PM_CTRL_PME_ENABLE) {
+		value &= ~PCI_PM_CTRL_PME_ENABLE;
+		err = pci_write_config_word(dev, offset, value);
+	}
+
+      out:
+	return ERR_PTR(err);
+}
+
+static struct config_field caplist_pm[] = {
+	{
+		.offset     = PCI_PM_PMC,
+		.size       = 2,
+		.u.w.read   = pm_caps_read,
+	},
+	{
+		.offset     = PCI_PM_CTRL,
+		.size       = 2,
+		.init       = pm_ctrl_init,
+		.u.w.read   = pciback_read_config_word,
+		.u.w.write  = pm_ctrl_write,
+	},
+	{
+		.offset     = PCI_PM_PPB_EXTENSIONS,
+		.size       = 1,
+		.u.b.read   = pciback_read_config_byte,
+	},
+	{
+		.offset     = PCI_PM_DATA_REGISTER,
+		.size       = 1,
+		.u.b.read   = pciback_read_config_byte,
+	},
+	{
+		.size = 0,
+	},
+};
+
+struct pciback_config_capability pciback_config_capability_pm = {
+	.capability = PCI_CAP_ID_PM,
+	.fields = caplist_pm,
+};
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/conf_space_capability_vpd.c linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space_capability_vpd.c
--- linux-2.6.18.8/drivers/xen/pciback/conf_space_capability_vpd.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space_capability_vpd.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,42 @@
+/*
+ * PCI Backend - Configuration space overlay for Vital Product Data
+ *
+ * Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+
+#include <linux/pci.h>
+#include "conf_space.h"
+#include "conf_space_capability.h"
+
+static int vpd_address_write(struct pci_dev *dev, int offset, u16 value,
+			     void *data)
+{
+	/* Disallow writes to the vital product data */
+	if (value & PCI_VPD_ADDR_F)
+		return PCIBIOS_SET_FAILED;
+	else
+		return pci_write_config_word(dev, offset, value);
+}
+
+static struct config_field caplist_vpd[] = {
+	{
+	 .offset    = PCI_VPD_ADDR,
+	 .size      = 2,
+	 .u.w.read  = pciback_read_config_word,
+	 .u.w.write = vpd_address_write,
+	 },
+	{
+	 .offset     = PCI_VPD_DATA,
+	 .size       = 4,
+	 .u.dw.read  = pciback_read_config_dword,
+	 .u.dw.write = NULL,
+	 },
+	{
+	 .size = 0,
+	 },
+};
+ 
+struct pciback_config_capability pciback_config_capability_vpd = {
+	.capability = PCI_CAP_ID_VPD,
+	.fields = caplist_vpd,
+};
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/conf_space.h linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space.h
--- linux-2.6.18.8/drivers/xen/pciback/conf_space.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,126 @@
+/*
+ * PCI Backend - Common data structures for overriding the configuration space
+ *
+ * Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+
+#ifndef __XEN_PCIBACK_CONF_SPACE_H__
+#define __XEN_PCIBACK_CONF_SPACE_H__
+
+#include <linux/list.h>
+#include <linux/err.h>
+
+/* conf_field_init can return an errno in a ptr with ERR_PTR() */
+typedef void *(*conf_field_init) (struct pci_dev * dev, int offset);
+typedef void (*conf_field_reset) (struct pci_dev * dev, int offset, void *data);
+typedef void (*conf_field_free) (struct pci_dev * dev, int offset, void *data);
+
+typedef int (*conf_dword_write) (struct pci_dev * dev, int offset, u32 value,
+				 void *data);
+typedef int (*conf_word_write) (struct pci_dev * dev, int offset, u16 value,
+				void *data);
+typedef int (*conf_byte_write) (struct pci_dev * dev, int offset, u8 value,
+				void *data);
+typedef int (*conf_dword_read) (struct pci_dev * dev, int offset, u32 * value,
+				void *data);
+typedef int (*conf_word_read) (struct pci_dev * dev, int offset, u16 * value,
+			       void *data);
+typedef int (*conf_byte_read) (struct pci_dev * dev, int offset, u8 * value,
+			       void *data);
+
+/* These are the fields within the configuration space which we
+ * are interested in intercepting reads/writes to and changing their
+ * values.
+ */
+struct config_field {
+	unsigned int offset;
+	unsigned int size;
+	unsigned int mask;
+	conf_field_init init;
+	conf_field_reset reset;
+	conf_field_free release;
+	void (*clean) (struct config_field * field);
+	union {
+		struct {
+			conf_dword_write write;
+			conf_dword_read read;
+		} dw;
+		struct {
+			conf_word_write write;
+			conf_word_read read;
+		} w;
+		struct {
+			conf_byte_write write;
+			conf_byte_read read;
+		} b;
+	} u;
+	struct list_head list;
+};
+
+struct config_field_entry {
+	struct list_head list;
+	struct config_field *field;
+	unsigned int base_offset;
+	void *data;
+};
+
+#define OFFSET(cfg_entry) ((cfg_entry)->base_offset+(cfg_entry)->field->offset)
+
+/* Add fields to a device - the add_fields macro expects to get a pointer to
+ * the first entry in an array (of which the ending is marked by size==0)
+ */
+int pciback_config_add_field_offset(struct pci_dev *dev,
+				    struct config_field *field,
+				    unsigned int offset);
+
+static inline int pciback_config_add_field(struct pci_dev *dev,
+					   struct config_field *field)
+{
+	return pciback_config_add_field_offset(dev, field, 0);
+}
+
+static inline int pciback_config_add_fields(struct pci_dev *dev,
+					    struct config_field *field)
+{
+	int i, err = 0;
+	for (i = 0; field[i].size != 0; i++) {
+		err = pciback_config_add_field(dev, &field[i]);
+		if (err)
+			break;
+	}
+	return err;
+}
+
+static inline int pciback_config_add_fields_offset(struct pci_dev *dev,
+						   struct config_field *field,
+						   unsigned int offset)
+{
+	int i, err = 0;
+	for (i = 0; field[i].size != 0; i++) {
+		err = pciback_config_add_field_offset(dev, &field[i], offset);
+		if (err)
+			break;
+	}
+	return err;
+}
+
+/* Read/Write the real configuration space */
+int pciback_read_config_byte(struct pci_dev *dev, int offset, u8 * value,
+			     void *data);
+int pciback_read_config_word(struct pci_dev *dev, int offset, u16 * value,
+			     void *data);
+int pciback_read_config_dword(struct pci_dev *dev, int offset, u32 * value,
+			      void *data);
+int pciback_write_config_byte(struct pci_dev *dev, int offset, u8 value,
+			      void *data);
+int pciback_write_config_word(struct pci_dev *dev, int offset, u16 value,
+			      void *data);
+int pciback_write_config_dword(struct pci_dev *dev, int offset, u32 value,
+			       void *data);
+
+int pciback_config_capability_init(void);
+
+int pciback_config_header_add_fields(struct pci_dev *dev);
+int pciback_config_capability_add_fields(struct pci_dev *dev);
+
+#endif				/* __XEN_PCIBACK_CONF_SPACE_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/conf_space_header.c linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space_header.c
--- linux-2.6.18.8/drivers/xen/pciback/conf_space_header.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space_header.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,323 @@
+/*
+ * PCI Backend - Handles the virtual fields in the configuration space headers.
+ *
+ * Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+
+#include <linux/kernel.h>
+#include <linux/pci.h>
+#include "pciback.h"
+#include "conf_space.h"
+
+struct pci_bar_info {
+	u32 val;
+	u32 len_val;
+	int which;
+};
+
+#define is_enable_cmd(value) ((value)&(PCI_COMMAND_MEMORY|PCI_COMMAND_IO))
+#define is_master_cmd(value) ((value)&PCI_COMMAND_MASTER)
+
+static int command_write(struct pci_dev *dev, int offset, u16 value, void *data)
+{
+	int err;
+
+	if (!dev->is_enabled && is_enable_cmd(value)) {
+		if (unlikely(verbose_request))
+			printk(KERN_DEBUG "pciback: %s: enable\n",
+			       pci_name(dev));
+		err = pci_enable_device(dev);
+		if (err)
+			return err;
+	} else if (dev->is_enabled && !is_enable_cmd(value)) {
+		if (unlikely(verbose_request))
+			printk(KERN_DEBUG "pciback: %s: disable\n",
+			       pci_name(dev));
+		pci_disable_device(dev);
+	}
+
+	if (!dev->is_busmaster && is_master_cmd(value)) {
+		if (unlikely(verbose_request))
+			printk(KERN_DEBUG "pciback: %s: set bus master\n",
+			       pci_name(dev));
+		pci_set_master(dev);
+	}
+
+	if (value & PCI_COMMAND_INVALIDATE) {
+		if (unlikely(verbose_request))
+			printk(KERN_DEBUG
+			       "pciback: %s: enable memory-write-invalidate\n",
+			       pci_name(dev));
+		err = pci_set_mwi(dev);
+		if (err) {
+			printk(KERN_WARNING
+			       "pciback: %s: cannot enable memory-write-invalidate (%d)\n",
+			       pci_name(dev), err);
+			value &= ~PCI_COMMAND_INVALIDATE;
+		}
+	}
+
+	return pci_write_config_word(dev, offset, value);
+}
+
+static int rom_write(struct pci_dev *dev, int offset, u32 value, void *data)
+{
+	struct pci_bar_info *bar = data;
+
+	if (unlikely(!bar)) {
+		printk(KERN_WARNING "pciback: driver data not found for %s\n",
+		       pci_name(dev));
+		return XEN_PCI_ERR_op_failed;
+	}
+
+	/* A write to obtain the length must happen as a 32-bit write.
+	 * This does not (yet) support writing individual bytes
+	 */
+	if (value == ~PCI_ROM_ADDRESS_ENABLE)
+		bar->which = 1;
+	else {
+		u32 tmpval;
+		pci_read_config_dword(dev, offset, &tmpval);
+		if (tmpval != bar->val && value == bar->val) {
+			/* Allow restoration of bar value. */
+			pci_write_config_dword(dev, offset, bar->val);
+		}
+		bar->which = 0;
+	}
+
+	/* Do we need to support enabling/disabling the rom address here? */
+
+	return 0;
+}
+
+/* For the BARs, only allow writes which write ~0 or
+ * the correct resource information
+ * (Needed for when the driver probes the resource usage)
+ */
+static int bar_write(struct pci_dev *dev, int offset, u32 value, void *data)
+{
+	struct pci_bar_info *bar = data;
+
+	if (unlikely(!bar)) {
+		printk(KERN_WARNING "pciback: driver data not found for %s\n",
+		       pci_name(dev));
+		return XEN_PCI_ERR_op_failed;
+	}
+
+	/* A write to obtain the length must happen as a 32-bit write.
+	 * This does not (yet) support writing individual bytes
+	 */
+	if (value == ~0)
+		bar->which = 1;
+	else {
+		u32 tmpval;
+		pci_read_config_dword(dev, offset, &tmpval);
+		if (tmpval != bar->val && value == bar->val) {
+			/* Allow restoration of bar value. */
+			pci_write_config_dword(dev, offset, bar->val);
+		}
+		bar->which = 0;
+	}
+
+	return 0;
+}
+
+static int bar_read(struct pci_dev *dev, int offset, u32 * value, void *data)
+{
+	struct pci_bar_info *bar = data;
+
+	if (unlikely(!bar)) {
+		printk(KERN_WARNING "pciback: driver data not found for %s\n",
+		       pci_name(dev));
+		return XEN_PCI_ERR_op_failed;
+	}
+
+	*value = bar->which ? bar->len_val : bar->val;
+
+	return 0;
+}
+
+static inline void read_dev_bar(struct pci_dev *dev,
+				struct pci_bar_info *bar_info, int offset,
+				u32 len_mask)
+{
+	pci_read_config_dword(dev, offset, &bar_info->val);
+	pci_write_config_dword(dev, offset, len_mask);
+	pci_read_config_dword(dev, offset, &bar_info->len_val);
+	pci_write_config_dword(dev, offset, bar_info->val);
+}
+
+static void *bar_init(struct pci_dev *dev, int offset)
+{
+	struct pci_bar_info *bar = kmalloc(sizeof(*bar), GFP_KERNEL);
+
+	if (!bar)
+		return ERR_PTR(-ENOMEM);
+
+	read_dev_bar(dev, bar, offset, ~0);
+	bar->which = 0;
+
+	return bar;
+}
+
+static void *rom_init(struct pci_dev *dev, int offset)
+{
+	struct pci_bar_info *bar = kmalloc(sizeof(*bar), GFP_KERNEL);
+
+	if (!bar)
+		return ERR_PTR(-ENOMEM);
+
+	read_dev_bar(dev, bar, offset, ~PCI_ROM_ADDRESS_ENABLE);
+	bar->which = 0;
+
+	return bar;
+}
+
+static void bar_reset(struct pci_dev *dev, int offset, void *data)
+{
+	struct pci_bar_info *bar = data;
+
+	bar->which = 0;
+}
+
+static void bar_release(struct pci_dev *dev, int offset, void *data)
+{
+	kfree(data);
+}
+
+static int interrupt_read(struct pci_dev *dev, int offset, u8 * value,
+			  void *data)
+{
+	*value = (u8) dev->irq;
+
+	return 0;
+}
+
+static int bist_write(struct pci_dev *dev, int offset, u8 value, void *data)
+{
+	u8 cur_value;
+	int err;
+
+	err = pci_read_config_byte(dev, offset, &cur_value);
+	if (err)
+		goto out;
+
+	if ((cur_value & ~PCI_BIST_START) == (value & ~PCI_BIST_START)
+	    || value == PCI_BIST_START)
+		err = pci_write_config_byte(dev, offset, value);
+
+      out:
+	return err;
+}
+
+static struct config_field header_common[] = {
+	{
+	 .offset    = PCI_COMMAND,
+	 .size      = 2,
+	 .u.w.read  = pciback_read_config_word,
+	 .u.w.write = command_write,
+	},
+	{
+	 .offset    = PCI_INTERRUPT_LINE,
+	 .size      = 1,
+	 .u.b.read  = interrupt_read,
+	},
+	{
+	 .offset    = PCI_INTERRUPT_PIN,
+	 .size      = 1,
+	 .u.b.read  = pciback_read_config_byte,
+	},
+	{
+	 /* Any side effects of letting driver domain control cache line? */
+	 .offset    = PCI_CACHE_LINE_SIZE,
+	 .size      = 1,
+	 .u.b.read  = pciback_read_config_byte,
+	 .u.b.write = pciback_write_config_byte,
+	},
+	{
+	 .offset    = PCI_LATENCY_TIMER,
+	 .size      = 1,
+	 .u.b.read  = pciback_read_config_byte,
+	},
+	{
+	 .offset    = PCI_BIST,
+	 .size      = 1,
+	 .u.b.read  = pciback_read_config_byte,
+	 .u.b.write = bist_write,
+	},
+	{
+	 .size = 0,
+	},
+};
+
+#define CFG_FIELD_BAR(reg_offset) 			\
+	{ 						\
+	 .offset     = reg_offset, 			\
+	 .size       = 4, 				\
+	 .init       = bar_init, 			\
+	 .reset      = bar_reset, 			\
+	 .release    = bar_release, 			\
+	 .u.dw.read  = bar_read, 			\
+	 .u.dw.write = bar_write, 			\
+	 }
+
+#define CFG_FIELD_ROM(reg_offset) 			\
+	{ 						\
+	 .offset     = reg_offset, 			\
+	 .size       = 4, 				\
+	 .init       = rom_init, 			\
+	 .reset      = bar_reset, 			\
+	 .release    = bar_release, 			\
+	 .u.dw.read  = bar_read, 			\
+	 .u.dw.write = rom_write, 			\
+	 }
+
+static struct config_field header_0[] = {
+	CFG_FIELD_BAR(PCI_BASE_ADDRESS_0),
+	CFG_FIELD_BAR(PCI_BASE_ADDRESS_1),
+	CFG_FIELD_BAR(PCI_BASE_ADDRESS_2),
+	CFG_FIELD_BAR(PCI_BASE_ADDRESS_3),
+	CFG_FIELD_BAR(PCI_BASE_ADDRESS_4),
+	CFG_FIELD_BAR(PCI_BASE_ADDRESS_5),
+	CFG_FIELD_ROM(PCI_ROM_ADDRESS),
+	{
+	 .size = 0,
+	},
+};
+
+static struct config_field header_1[] = {
+	CFG_FIELD_BAR(PCI_BASE_ADDRESS_0),
+	CFG_FIELD_BAR(PCI_BASE_ADDRESS_1),
+	CFG_FIELD_ROM(PCI_ROM_ADDRESS1),
+	{
+	 .size = 0,
+	},
+};
+
+int pciback_config_header_add_fields(struct pci_dev *dev)
+{
+	int err;
+
+	err = pciback_config_add_fields(dev, header_common);
+	if (err)
+		goto out;
+
+	switch (dev->hdr_type) {
+	case PCI_HEADER_TYPE_NORMAL:
+		err = pciback_config_add_fields(dev, header_0);
+		break;
+
+	case PCI_HEADER_TYPE_BRIDGE:
+		err = pciback_config_add_fields(dev, header_1);
+		break;
+
+	default:
+		err = -EINVAL;
+		printk(KERN_ERR "pciback: %s: Unsupported header type %d!\n",
+		       pci_name(dev), dev->hdr_type);
+		break;
+	}
+
+      out:
+	return err;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/conf_space_quirks.c linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space_quirks.c
--- linux-2.6.18.8/drivers/xen/pciback/conf_space_quirks.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space_quirks.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,126 @@
+/*
+ * PCI Backend - Handle special overlays for broken devices.
+ *
+ * Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ * Author: Chris Bookholt <hap10@epoch.ncsc.mil>
+ */
+
+#include <linux/kernel.h>
+#include <linux/pci.h>
+#include "pciback.h"
+#include "conf_space.h"
+#include "conf_space_quirks.h"
+
+LIST_HEAD(pciback_quirks);
+
+struct pciback_config_quirk *pciback_find_quirk(struct pci_dev *dev)
+{
+	struct pciback_config_quirk *tmp_quirk;
+
+	list_for_each_entry(tmp_quirk, &pciback_quirks, quirks_list)
+	    if (pci_match_id(&tmp_quirk->devid, dev))
+		goto out;
+	tmp_quirk = NULL;
+	printk(KERN_DEBUG
+	       "quirk didn't match any device pciback knows about\n");
+      out:
+	return tmp_quirk;
+}
+
+static inline void register_quirk(struct pciback_config_quirk *quirk)
+{
+	list_add_tail(&quirk->quirks_list, &pciback_quirks);
+}
+
+int pciback_field_is_dup(struct pci_dev *dev, unsigned int reg)
+{
+	int ret = 0;
+	struct pciback_dev_data *dev_data = pci_get_drvdata(dev);
+	struct config_field_entry *cfg_entry;
+
+	list_for_each_entry(cfg_entry, &dev_data->config_fields, list) {
+		if ( OFFSET(cfg_entry) == reg) {
+			ret = 1;
+			break;
+		}
+	}
+	return ret;
+}
+
+int pciback_config_quirks_add_field(struct pci_dev *dev, struct config_field
+				    *field)
+{
+	int err = 0;
+
+	switch (field->size) {
+	case 1:
+		field->u.b.read = pciback_read_config_byte;
+		field->u.b.write = pciback_write_config_byte;
+		break;
+	case 2:
+		field->u.w.read = pciback_read_config_word;
+		field->u.w.write = pciback_write_config_word;
+		break;
+	case 4:
+		field->u.dw.read = pciback_read_config_dword;
+		field->u.dw.write = pciback_write_config_dword;
+		break;
+	default:
+		err = -EINVAL;
+		goto out;
+	}
+
+	pciback_config_add_field(dev, field);
+
+      out:
+	return err;
+}
+
+int pciback_config_quirks_init(struct pci_dev *dev)
+{
+	struct pciback_config_quirk *quirk;
+	int ret = 0;
+
+	quirk = kzalloc(sizeof(*quirk), GFP_ATOMIC);
+	if (!quirk) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	quirk->devid.vendor = dev->vendor;
+	quirk->devid.device = dev->device;
+	quirk->devid.subvendor = dev->subsystem_vendor;
+	quirk->devid.subdevice = dev->subsystem_device;
+	quirk->devid.class = 0;
+	quirk->devid.class_mask = 0;
+	quirk->devid.driver_data = 0UL;
+
+	quirk->pdev = dev;
+
+	register_quirk(quirk);
+      out:
+	return ret;
+}
+
+void pciback_config_field_free(struct config_field *field)
+{
+	kfree(field);
+}
+
+int pciback_config_quirk_release(struct pci_dev *dev)
+{
+	struct pciback_config_quirk *quirk;
+	int ret = 0;
+
+	quirk = pciback_find_quirk(dev);
+	if (!quirk) {
+		ret = -ENXIO;
+		goto out;
+	}
+
+	list_del(&quirk->quirks_list);
+	kfree(quirk);
+
+      out:
+	return ret;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/conf_space_quirks.h linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space_quirks.h
--- linux-2.6.18.8/drivers/xen/pciback/conf_space_quirks.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/conf_space_quirks.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,35 @@
+/*
+ * PCI Backend - Data structures for special overlays for broken devices.
+ *
+ * Ryan Wilson <hap9@epoch.ncsc.mil>
+ * Chris Bookholt <hap10@epoch.ncsc.mil>
+ */
+
+#ifndef __XEN_PCIBACK_CONF_SPACE_QUIRKS_H__
+#define __XEN_PCIBACK_CONF_SPACE_QUIRKS_H__
+
+#include <linux/pci.h>
+#include <linux/list.h>
+
+struct pciback_config_quirk {
+	struct list_head quirks_list;
+	struct pci_device_id devid;
+	struct pci_dev *pdev;
+};
+
+struct pciback_config_quirk *pciback_find_quirk(struct pci_dev *dev);
+
+int pciback_config_quirks_add_field(struct pci_dev *dev, struct config_field
+				    *field);
+
+int pciback_config_quirks_remove_field(struct pci_dev *dev, int reg);
+
+int pciback_config_quirks_init(struct pci_dev *dev);
+
+void pciback_config_field_free(struct config_field *field);
+
+int pciback_config_quirk_release(struct pci_dev *dev);
+
+int pciback_field_is_dup(struct pci_dev *dev, unsigned int reg);
+
+#endif
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/controller.c linux-2.6.18-xen-3.3.0/drivers/xen/pciback/controller.c
--- linux-2.6.18.8/drivers/xen/pciback/controller.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/controller.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,408 @@
+/*
+ * Copyright (C) 2007 Hewlett-Packard Development Company, L.P.
+ *      Alex Williamson <alex.williamson@hp.com>
+ *
+ * PCI "Controller" Backend - virtualize PCI bus topology based on PCI
+ * controllers.  Devices under the same PCI controller are exposed on the
+ * same virtual domain:bus.  Within a bus, device slots are virtualized
+ * to compact the bus.
+ *
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+ * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ */
+
+#include <linux/acpi.h>
+#include <linux/list.h>
+#include <linux/pci.h>
+#include <linux/spinlock.h>
+#include "pciback.h"
+
+#define PCI_MAX_BUSSES	255
+#define PCI_MAX_SLOTS	32
+
+struct controller_dev_entry {
+	struct list_head list;
+	struct pci_dev *dev;
+	unsigned int devfn;
+};
+
+struct controller_list_entry {
+	struct list_head list;
+	struct pci_controller *controller;
+	unsigned int domain;
+	unsigned int bus;
+	unsigned int next_devfn;
+	struct list_head dev_list;
+};
+
+struct controller_dev_data {
+	struct list_head list;
+	unsigned int next_domain;
+	unsigned int next_bus;
+	spinlock_t lock;
+};
+
+struct walk_info {
+	struct pciback_device *pdev;
+	int resource_count;
+	int root_num;
+};
+
+struct pci_dev *pciback_get_pci_dev(struct pciback_device *pdev,
+				    unsigned int domain, unsigned int bus,
+				    unsigned int devfn)
+{
+	struct controller_dev_data *dev_data = pdev->pci_dev_data;
+	struct controller_dev_entry *dev_entry;
+	struct controller_list_entry *cntrl_entry;
+	struct pci_dev *dev = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev_data->lock, flags);
+
+	list_for_each_entry(cntrl_entry, &dev_data->list, list) {
+		if (cntrl_entry->domain != domain ||
+		    cntrl_entry->bus != bus)
+			continue;
+
+		list_for_each_entry(dev_entry, &cntrl_entry->dev_list, list) {
+			if (devfn == dev_entry->devfn) {
+				dev = dev_entry->dev;
+				goto found;
+			}
+		}
+	}
+found:
+	spin_unlock_irqrestore(&dev_data->lock, flags);
+
+	return dev;
+}
+
+int pciback_add_pci_dev(struct pciback_device *pdev, struct pci_dev *dev,
+			int devid, publish_pci_dev_cb publish_cb)
+{
+	struct controller_dev_data *dev_data = pdev->pci_dev_data;
+	struct controller_dev_entry *dev_entry;
+	struct controller_list_entry *cntrl_entry;
+	struct pci_controller *dev_controller = PCI_CONTROLLER(dev);
+	unsigned long flags;
+	int ret = 0, found = 0;
+
+	spin_lock_irqsave(&dev_data->lock, flags);
+
+	/* Look to see if we already have a domain:bus for this controller */
+	list_for_each_entry(cntrl_entry, &dev_data->list, list) {
+		if (cntrl_entry->controller == dev_controller) {
+			found = 1;
+			break;
+		}
+	}
+
+	if (!found) {
+		cntrl_entry = kmalloc(sizeof(*cntrl_entry), GFP_ATOMIC);
+		if (!cntrl_entry) {
+			ret =  -ENOMEM;
+			goto out;
+		}
+
+		cntrl_entry->controller = dev_controller;
+		cntrl_entry->next_devfn = PCI_DEVFN(0, 0);
+
+		cntrl_entry->domain = dev_data->next_domain;
+		cntrl_entry->bus = dev_data->next_bus++;
+		if (dev_data->next_bus > PCI_MAX_BUSSES) {
+			dev_data->next_domain++;
+			dev_data->next_bus = 0;
+		}
+
+		INIT_LIST_HEAD(&cntrl_entry->dev_list);
+
+		list_add_tail(&cntrl_entry->list, &dev_data->list);
+	}
+
+	if (PCI_SLOT(cntrl_entry->next_devfn) > PCI_MAX_SLOTS) {
+		/*
+		 * While it seems unlikely, this can actually happen if
+		 * a controller has P2P bridges under it.
+		 */
+		xenbus_dev_fatal(pdev->xdev, -ENOSPC, "Virtual bus %04x:%02x "
+				 "is full, no room to export %04x:%02x:%02x.%x",
+				 cntrl_entry->domain, cntrl_entry->bus,
+				 pci_domain_nr(dev->bus), dev->bus->number,
+				 PCI_SLOT(dev->devfn), PCI_FUNC(dev->devfn));
+		ret = -ENOSPC;
+		goto out;
+	}
+
+	dev_entry = kmalloc(sizeof(*dev_entry), GFP_ATOMIC);
+	if (!dev_entry) {
+		if (list_empty(&cntrl_entry->dev_list)) {
+			list_del(&cntrl_entry->list);
+			kfree(cntrl_entry);
+		}
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	dev_entry->dev = dev;
+	dev_entry->devfn = cntrl_entry->next_devfn;
+
+	list_add_tail(&dev_entry->list, &cntrl_entry->dev_list);
+
+	cntrl_entry->next_devfn += PCI_DEVFN(1, 0);
+
+out:
+	spin_unlock_irqrestore(&dev_data->lock, flags);
+
+	/* TODO: Publish virtual domain:bus:slot.func here. */
+
+	return ret;
+}
+
+void pciback_release_pci_dev(struct pciback_device *pdev, struct pci_dev *dev)
+{
+	struct controller_dev_data *dev_data = pdev->pci_dev_data;
+	struct controller_list_entry *cntrl_entry;
+	struct controller_dev_entry *dev_entry = NULL;
+	struct pci_dev *found_dev = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev_data->lock, flags);
+
+	list_for_each_entry(cntrl_entry, &dev_data->list, list) {
+		if (cntrl_entry->controller != PCI_CONTROLLER(dev))
+			continue;
+
+		list_for_each_entry(dev_entry, &cntrl_entry->dev_list, list) {
+			if (dev_entry->dev == dev) {
+				found_dev = dev_entry->dev;
+				break;
+			}
+		}
+	}
+
+	if (!found_dev) {
+		spin_unlock_irqrestore(&dev_data->lock, flags);
+		return;
+	}
+
+	list_del(&dev_entry->list);
+	kfree(dev_entry);
+
+	if (list_empty(&cntrl_entry->dev_list)) {
+		list_del(&cntrl_entry->list);
+		kfree(cntrl_entry);
+	}
+
+	spin_unlock_irqrestore(&dev_data->lock, flags);
+	pcistub_put_pci_dev(found_dev);
+}
+
+int pciback_init_devices(struct pciback_device *pdev)
+{
+	struct controller_dev_data *dev_data;
+
+	dev_data = kmalloc(sizeof(*dev_data), GFP_KERNEL);
+	if (!dev_data)
+		return -ENOMEM;
+
+	spin_lock_init(&dev_data->lock);
+
+	INIT_LIST_HEAD(&dev_data->list);
+
+	/* Starting domain:bus numbers */
+	dev_data->next_domain = 0;
+	dev_data->next_bus = 0;
+
+	pdev->pci_dev_data = dev_data;
+
+	return 0;
+}
+
+static acpi_status write_xenbus_resource(struct acpi_resource *res, void *data)
+{
+	struct walk_info *info = data;
+	struct acpi_resource_address64 addr;
+	acpi_status status;
+	int i, len, err;
+	char str[32], tmp[3];
+	unsigned char *ptr, *buf;
+
+	status = acpi_resource_to_address64(res, &addr);
+
+	/* Do we care about this range?  Let's check. */
+	if (!ACPI_SUCCESS(status) ||
+	    !(addr.resource_type == ACPI_MEMORY_RANGE ||
+	      addr.resource_type == ACPI_IO_RANGE) ||
+	    !addr.address_length || addr.producer_consumer != ACPI_PRODUCER)
+		return AE_OK;
+
+	/*
+	 * Furthermore, we really only care to tell the guest about
+	 * address ranges that require address translation of some sort.
+	 */
+	if (!(addr.resource_type == ACPI_MEMORY_RANGE &&
+	      addr.info.mem.translation) &&
+	    !(addr.resource_type == ACPI_IO_RANGE &&
+	      addr.info.io.translation))
+		return AE_OK;
+	   
+	/* Store the resource in xenbus for the guest */
+	len = snprintf(str, sizeof(str), "root-%d-resource-%d",
+		       info->root_num, info->resource_count);
+	if (unlikely(len >= (sizeof(str) - 1)))
+		return AE_OK;
+
+	buf = kzalloc((sizeof(*res) * 2) + 1, GFP_KERNEL);
+	if (!buf)
+		return AE_OK;
+
+	/* Clean out resource_source */
+	res->data.address64.resource_source.index = 0xFF;
+	res->data.address64.resource_source.string_length = 0;
+	res->data.address64.resource_source.string_ptr = NULL;
+
+	ptr = (unsigned char *)res;
+
+	/* Turn the acpi_resource into an ASCII byte stream */
+	for (i = 0; i < sizeof(*res); i++) {
+		snprintf(tmp, sizeof(tmp), "%02x", ptr[i]);
+		strncat(buf, tmp, 2);
+	}
+
+	err = xenbus_printf(XBT_NIL, info->pdev->xdev->nodename,
+			    str, "%s", buf);
+
+	if (!err)
+		info->resource_count++;
+
+	kfree(buf);
+
+	return AE_OK;
+}
+
+int pciback_publish_pci_roots(struct pciback_device *pdev,
+			      publish_pci_root_cb publish_root_cb)
+{
+	struct controller_dev_data *dev_data = pdev->pci_dev_data;
+	struct controller_list_entry *cntrl_entry;
+	int i, root_num, len, err = 0;
+	unsigned int domain, bus;
+	char str[64];
+	struct walk_info info;
+
+	spin_lock(&dev_data->lock);
+
+	list_for_each_entry(cntrl_entry, &dev_data->list, list) {
+		/* First publish all the domain:bus info */
+		err = publish_root_cb(pdev, cntrl_entry->domain,
+				      cntrl_entry->bus);
+		if (err)
+			goto out;
+
+		/*
+ 		 * Now figure out which root-%d this belongs to
+		 * so we can associate resources with it.
+		 */
+		err = xenbus_scanf(XBT_NIL, pdev->xdev->nodename,
+				   "root_num", "%d", &root_num);
+
+		if (err != 1)
+			goto out;
+
+		for (i = 0; i < root_num; i++) {
+			len = snprintf(str, sizeof(str), "root-%d", i);
+			if (unlikely(len >= (sizeof(str) - 1))) {
+				err = -ENOMEM;
+				goto out;
+			}
+
+			err = xenbus_scanf(XBT_NIL, pdev->xdev->nodename,
+					   str, "%x:%x", &domain, &bus);
+			if (err != 2)
+				goto out;
+
+			/* Is this the one we just published? */
+			if (domain == cntrl_entry->domain &&
+			    bus == cntrl_entry->bus)
+				break;
+		}
+
+		if (i == root_num)
+			goto out;
+
+		info.pdev = pdev;
+		info.resource_count = 0;
+		info.root_num = i;
+
+		/* Let ACPI do the heavy lifting on decoding resources */
+		acpi_walk_resources(cntrl_entry->controller->acpi_handle,
+				    METHOD_NAME__CRS, write_xenbus_resource,
+				    &info);
+
+		/* No resouces.  OK.  On to the next one */
+		if (!info.resource_count)
+			continue;
+
+		/* Store the number of resources we wrote for this root-%d */
+		len = snprintf(str, sizeof(str), "root-%d-resources", i);
+		if (unlikely(len >= (sizeof(str) - 1))) {
+			err = -ENOMEM;
+			goto out;
+		}
+
+		err = xenbus_printf(XBT_NIL, pdev->xdev->nodename, str,
+				    "%d", info.resource_count);
+		if (err)
+			goto out;
+	}
+
+	/* Finally, write some magic to synchronize with the guest. */
+	len = snprintf(str, sizeof(str), "root-resource-magic");
+	if (unlikely(len >= (sizeof(str) - 1))) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	err = xenbus_printf(XBT_NIL, pdev->xdev->nodename, str,
+			    "%lx", (sizeof(struct acpi_resource) * 2) + 1);
+
+out:
+	spin_unlock(&dev_data->lock);
+
+	return err;
+}
+
+void pciback_release_devices(struct pciback_device *pdev)
+{
+	struct controller_dev_data *dev_data = pdev->pci_dev_data;
+	struct controller_list_entry *cntrl_entry, *c;
+	struct controller_dev_entry *dev_entry, *d;
+
+	list_for_each_entry_safe(cntrl_entry, c, &dev_data->list, list) {
+		list_for_each_entry_safe(dev_entry, d,
+					 &cntrl_entry->dev_list, list) {
+			list_del(&dev_entry->list);
+			pcistub_put_pci_dev(dev_entry->dev);
+			kfree(dev_entry);
+		}
+		list_del(&cntrl_entry->list);
+		kfree(cntrl_entry);
+	}
+
+	kfree(dev_data);
+	pdev->pci_dev_data = NULL;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/pciback/Makefile
--- linux-2.6.18.8/drivers/xen/pciback/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,17 @@
+obj-$(CONFIG_XEN_PCIDEV_BACKEND) += pciback.o
+
+pciback-y := pci_stub.o pciback_ops.o xenbus.o
+pciback-y += conf_space.o conf_space_header.o \
+	     conf_space_capability.o \
+	     conf_space_capability_vpd.o \
+	     conf_space_capability_pm.o \
+             conf_space_quirks.o
+pciback-$(CONFIG_PCI_MSI) += conf_space_capability_msi.o
+pciback-$(CONFIG_XEN_PCIDEV_BACKEND_VPCI) += vpci.o
+pciback-$(CONFIG_XEN_PCIDEV_BACKEND_SLOT) += slot.o
+pciback-$(CONFIG_XEN_PCIDEV_BACKEND_PASS) += passthrough.o
+pciback-$(CONFIG_XEN_PCIDEV_BACKEND_CONTROLLER) += controller.o
+
+ifeq ($(CONFIG_XEN_PCIDEV_BE_DEBUG),y)
+EXTRA_CFLAGS += -DDEBUG
+endif
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/passthrough.c linux-2.6.18-xen-3.3.0/drivers/xen/pciback/passthrough.c
--- linux-2.6.18.8/drivers/xen/pciback/passthrough.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/passthrough.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,166 @@
+/*
+ * PCI Backend - Provides restricted access to the real PCI bus topology
+ *               to the frontend
+ *
+ *   Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+
+#include <linux/list.h>
+#include <linux/pci.h>
+#include <linux/spinlock.h>
+#include "pciback.h"
+
+struct passthrough_dev_data {
+	/* Access to dev_list must be protected by lock */
+	struct list_head dev_list;
+	spinlock_t lock;
+};
+
+struct pci_dev *pciback_get_pci_dev(struct pciback_device *pdev,
+				    unsigned int domain, unsigned int bus,
+				    unsigned int devfn)
+{
+	struct passthrough_dev_data *dev_data = pdev->pci_dev_data;
+	struct pci_dev_entry *dev_entry;
+	struct pci_dev *dev = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev_data->lock, flags);
+
+	list_for_each_entry(dev_entry, &dev_data->dev_list, list) {
+		if (domain == (unsigned int)pci_domain_nr(dev_entry->dev->bus)
+		    && bus == (unsigned int)dev_entry->dev->bus->number
+		    && devfn == dev_entry->dev->devfn) {
+			dev = dev_entry->dev;
+			break;
+		}
+	}
+
+	spin_unlock_irqrestore(&dev_data->lock, flags);
+
+	return dev;
+}
+
+int pciback_add_pci_dev(struct pciback_device *pdev, struct pci_dev *dev,
+			int devid, publish_pci_dev_cb publish_cb)
+{
+	struct passthrough_dev_data *dev_data = pdev->pci_dev_data;
+	struct pci_dev_entry *dev_entry;
+	unsigned long flags;
+	unsigned int domain, bus, devfn;
+	int err;
+
+	dev_entry = kmalloc(sizeof(*dev_entry), GFP_KERNEL);
+	if (!dev_entry)
+		return -ENOMEM;
+	dev_entry->dev = dev;
+
+	spin_lock_irqsave(&dev_data->lock, flags);
+	list_add_tail(&dev_entry->list, &dev_data->dev_list);
+	spin_unlock_irqrestore(&dev_data->lock, flags);
+
+	/* Publish this device. */
+	domain = (unsigned int)pci_domain_nr(dev->bus);
+	bus = (unsigned int)dev->bus->number;
+	devfn = dev->devfn;
+	err = publish_cb(pdev, domain, bus, devfn, devid);
+
+	return err;
+}
+
+void pciback_release_pci_dev(struct pciback_device *pdev, struct pci_dev *dev)
+{
+	struct passthrough_dev_data *dev_data = pdev->pci_dev_data;
+	struct pci_dev_entry *dev_entry, *t;
+	struct pci_dev *found_dev = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev_data->lock, flags);
+
+	list_for_each_entry_safe(dev_entry, t, &dev_data->dev_list, list) {
+		if (dev_entry->dev == dev) {
+			list_del(&dev_entry->list);
+			found_dev = dev_entry->dev;
+			kfree(dev_entry);
+		}
+	}
+
+	spin_unlock_irqrestore(&dev_data->lock, flags);
+
+	if (found_dev)
+		pcistub_put_pci_dev(found_dev);
+}
+
+int pciback_init_devices(struct pciback_device *pdev)
+{
+	struct passthrough_dev_data *dev_data;
+
+	dev_data = kmalloc(sizeof(*dev_data), GFP_KERNEL);
+	if (!dev_data)
+		return -ENOMEM;
+
+	spin_lock_init(&dev_data->lock);
+
+	INIT_LIST_HEAD(&dev_data->dev_list);
+
+	pdev->pci_dev_data = dev_data;
+
+	return 0;
+}
+
+int pciback_publish_pci_roots(struct pciback_device *pdev,
+			      publish_pci_root_cb publish_root_cb)
+{
+	int err = 0;
+	struct passthrough_dev_data *dev_data = pdev->pci_dev_data;
+	struct pci_dev_entry *dev_entry, *e;
+	struct pci_dev *dev;
+	int found;
+	unsigned int domain, bus;
+
+	spin_lock(&dev_data->lock);
+
+	list_for_each_entry(dev_entry, &dev_data->dev_list, list) {
+		/* Only publish this device as a root if none of its
+		 * parent bridges are exported
+		 */
+		found = 0;
+		dev = dev_entry->dev->bus->self;
+		for (; !found && dev != NULL; dev = dev->bus->self) {
+			list_for_each_entry(e, &dev_data->dev_list, list) {
+				if (dev == e->dev) {
+					found = 1;
+					break;
+				}
+			}
+		}
+
+		domain = (unsigned int)pci_domain_nr(dev_entry->dev->bus);
+		bus = (unsigned int)dev_entry->dev->bus->number;
+
+		if (!found) {
+			err = publish_root_cb(pdev, domain, bus);
+			if (err)
+				break;
+		}
+	}
+
+	spin_unlock(&dev_data->lock);
+
+	return err;
+}
+
+void pciback_release_devices(struct pciback_device *pdev)
+{
+	struct passthrough_dev_data *dev_data = pdev->pci_dev_data;
+	struct pci_dev_entry *dev_entry, *t;
+
+	list_for_each_entry_safe(dev_entry, t, &dev_data->dev_list, list) {
+		list_del(&dev_entry->list);
+		pcistub_put_pci_dev(dev_entry->dev);
+		kfree(dev_entry);
+	}
+
+	kfree(dev_data);
+	pdev->pci_dev_data = NULL;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/pciback.h linux-2.6.18-xen-3.3.0/drivers/xen/pciback/pciback.h
--- linux-2.6.18.8/drivers/xen/pciback/pciback.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/pciback.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,111 @@
+/*
+ * PCI Backend Common Data Structures & Function Declarations
+ *
+ *   Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+#ifndef __XEN_PCIBACK_H__
+#define __XEN_PCIBACK_H__
+
+#include <linux/pci.h>
+#include <linux/interrupt.h>
+#include <xen/xenbus.h>
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/workqueue.h>
+#include <asm/atomic.h>
+#include <xen/interface/io/pciif.h>
+
+struct pci_dev_entry {
+	struct list_head list;
+	struct pci_dev *dev;
+};
+
+#define _PDEVF_op_active 	(0)
+#define PDEVF_op_active 	(1<<(_PDEVF_op_active))
+
+struct pciback_device {
+	void *pci_dev_data;
+	spinlock_t dev_lock;
+
+	struct xenbus_device *xdev;
+
+	struct xenbus_watch be_watch;
+	u8 be_watching;
+
+	int evtchn_irq;
+
+	struct vm_struct *sh_area;
+	struct xen_pci_sharedinfo *sh_info;
+
+	unsigned long flags;
+
+	struct work_struct op_work;
+};
+
+struct pciback_dev_data {
+	struct list_head config_fields;
+	int permissive;
+	int warned_on_write;
+};
+
+/* Get/Put PCI Devices that are hidden from the PCI Backend Domain */
+struct pci_dev *pcistub_get_pci_dev_by_slot(struct pciback_device *pdev,
+					    int domain, int bus,
+					    int slot, int func);
+struct pci_dev *pcistub_get_pci_dev(struct pciback_device *pdev,
+				    struct pci_dev *dev);
+void pcistub_put_pci_dev(struct pci_dev *dev);
+
+/* Ensure a device is turned off or reset */
+void pciback_reset_device(struct pci_dev *pdev);
+
+/* Access a virtual configuration space for a PCI device */
+int pciback_config_init(void);
+int pciback_config_init_dev(struct pci_dev *dev);
+void pciback_config_free_dyn_fields(struct pci_dev *dev);
+void pciback_config_reset_dev(struct pci_dev *dev);
+void pciback_config_free_dev(struct pci_dev *dev);
+int pciback_config_read(struct pci_dev *dev, int offset, int size,
+			u32 * ret_val);
+int pciback_config_write(struct pci_dev *dev, int offset, int size, u32 value);
+
+/* Handle requests for specific devices from the frontend */
+typedef int (*publish_pci_dev_cb) (struct pciback_device *pdev,
+				   unsigned int domain, unsigned int bus,
+				   unsigned int devfn, unsigned int devid);
+typedef int (*publish_pci_root_cb) (struct pciback_device * pdev,
+				    unsigned int domain, unsigned int bus);
+int pciback_add_pci_dev(struct pciback_device *pdev, struct pci_dev *dev,
+			int devid, publish_pci_dev_cb publish_cb);
+void pciback_release_pci_dev(struct pciback_device *pdev, struct pci_dev *dev);
+struct pci_dev *pciback_get_pci_dev(struct pciback_device *pdev,
+				    unsigned int domain, unsigned int bus,
+				    unsigned int devfn);
+int pciback_init_devices(struct pciback_device *pdev);
+int pciback_publish_pci_roots(struct pciback_device *pdev,
+			      publish_pci_root_cb cb);
+void pciback_release_devices(struct pciback_device *pdev);
+
+/* Handles events from front-end */
+irqreturn_t pciback_handle_event(int irq, void *dev_id, struct pt_regs *regs);
+void pciback_do_op(void *data);
+
+int pciback_xenbus_register(void);
+void pciback_xenbus_unregister(void);
+
+#ifdef CONFIG_PCI_MSI
+int pciback_enable_msi(struct pciback_device *pdev,
+                       struct pci_dev *dev, struct xen_pci_op *op);
+
+int pciback_disable_msi(struct pciback_device *pdev,
+                         struct pci_dev *dev, struct xen_pci_op *op);
+
+
+int pciback_enable_msix(struct pciback_device *pdev,
+                        struct pci_dev *dev, struct xen_pci_op *op);
+
+int pciback_disable_msix(struct pciback_device *pdev,
+                        struct pci_dev *dev, struct xen_pci_op *op);
+#endif
+extern int verbose_request;
+#endif
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/pciback_ops.c linux-2.6.18-xen-3.3.0/drivers/xen/pciback/pciback_ops.c
--- linux-2.6.18.8/drivers/xen/pciback/pciback_ops.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/pciback_ops.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,117 @@
+/*
+ * PCI Backend Operations - respond to PCI requests from Frontend
+ *
+ *   Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+#include <linux/module.h>
+#include <asm/bitops.h>
+#include <xen/evtchn.h>
+#include "pciback.h"
+
+int verbose_request = 0;
+module_param(verbose_request, int, 0644);
+
+/* Ensure a device is "turned off" and ready to be exported.
+ * (Also see pciback_config_reset to ensure virtual configuration space is
+ * ready to be re-exported)
+ */
+void pciback_reset_device(struct pci_dev *dev)
+{
+	u16 cmd;
+
+	/* Disable devices (but not bridges) */
+	if (dev->hdr_type == PCI_HEADER_TYPE_NORMAL) {
+		pci_disable_device(dev);
+
+		pci_write_config_word(dev, PCI_COMMAND, 0);
+
+		dev->is_enabled = 0;
+		dev->is_busmaster = 0;
+	} else {
+		pci_read_config_word(dev, PCI_COMMAND, &cmd);
+		if (cmd & (PCI_COMMAND_INVALIDATE)) {
+			cmd &= ~(PCI_COMMAND_INVALIDATE);
+			pci_write_config_word(dev, PCI_COMMAND, cmd);
+
+			dev->is_busmaster = 0;
+		}
+	}
+}
+
+static inline void test_and_schedule_op(struct pciback_device *pdev)
+{
+	/* Check that frontend is requesting an operation and that we are not
+	 * already processing a request */
+	if (test_bit(_XEN_PCIF_active, (unsigned long *)&pdev->sh_info->flags)
+	    && !test_and_set_bit(_PDEVF_op_active, &pdev->flags))
+		schedule_work(&pdev->op_work);
+}
+
+/* Performing the configuration space reads/writes must not be done in atomic
+ * context because some of the pci_* functions can sleep (mostly due to ACPI
+ * use of semaphores). This function is intended to be called from a work
+ * queue in process context taking a struct pciback_device as a parameter */
+void pciback_do_op(void *data)
+{
+	struct pciback_device *pdev = data;
+	struct pci_dev *dev;
+	struct xen_pci_op *op = &pdev->sh_info->op;
+
+	dev = pciback_get_pci_dev(pdev, op->domain, op->bus, op->devfn);
+
+	if (dev == NULL)
+		op->err = XEN_PCI_ERR_dev_not_found;
+	else
+	{
+		switch (op->cmd)
+		{
+			case XEN_PCI_OP_conf_read:
+				op->err = pciback_config_read(dev,
+					  op->offset, op->size, &op->value);
+				break;
+			case XEN_PCI_OP_conf_write:
+				op->err = pciback_config_write(dev,
+					  op->offset, op->size,	op->value);
+				break;
+#ifdef CONFIG_PCI_MSI
+			case XEN_PCI_OP_enable_msi:
+				op->err = pciback_enable_msi(pdev, dev, op);
+				break;
+			case XEN_PCI_OP_disable_msi:
+				op->err = pciback_disable_msi(pdev, dev, op);
+				break;
+			case XEN_PCI_OP_enable_msix:
+				op->err = pciback_enable_msix(pdev, dev, op);
+				break;
+			case XEN_PCI_OP_disable_msix:
+				op->err = pciback_disable_msix(pdev, dev, op);
+				break;
+#endif
+			default:
+				op->err = XEN_PCI_ERR_not_implemented;
+				break;
+		}
+	}
+	/* Tell the driver domain that we're done. */ 
+	wmb();
+	clear_bit(_XEN_PCIF_active, (unsigned long *)&pdev->sh_info->flags);
+	notify_remote_via_irq(pdev->evtchn_irq);
+
+	/* Mark that we're done. */
+	smp_mb__before_clear_bit(); /* /after/ clearing PCIF_active */
+	clear_bit(_PDEVF_op_active, &pdev->flags);
+	smp_mb__after_clear_bit(); /* /before/ final check for work */
+
+	/* Check to see if the driver domain tried to start another request in
+	 * between clearing _XEN_PCIF_active and clearing _PDEVF_op_active. */
+	test_and_schedule_op(pdev);
+}
+
+irqreturn_t pciback_handle_event(int irq, void *dev_id, struct pt_regs *regs)
+{
+	struct pciback_device *pdev = dev_id;
+
+	test_and_schedule_op(pdev);
+
+	return IRQ_HANDLED;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/pci_stub.c linux-2.6.18-xen-3.3.0/drivers/xen/pciback/pci_stub.c
--- linux-2.6.18.8/drivers/xen/pciback/pci_stub.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/pci_stub.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,948 @@
+/*
+ * PCI Stub Driver - Grabs devices in backend to be exported later
+ *
+ * Ryan Wilson <hap9@epoch.ncsc.mil>
+ * Chris Bookholt <hap10@epoch.ncsc.mil>
+ */
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/kref.h>
+#include <asm/atomic.h>
+#include "pciback.h"
+#include "conf_space.h"
+#include "conf_space_quirks.h"
+
+static char *pci_devs_to_hide = NULL;
+module_param_named(hide, pci_devs_to_hide, charp, 0444);
+
+struct pcistub_device_id {
+	struct list_head slot_list;
+	int domain;
+	unsigned char bus;
+	unsigned int devfn;
+};
+static LIST_HEAD(pcistub_device_ids);
+static DEFINE_SPINLOCK(device_ids_lock);
+
+struct pcistub_device {
+	struct kref kref;
+	struct list_head dev_list;
+	spinlock_t lock;
+
+	struct pci_dev *dev;
+	struct pciback_device *pdev;	/* non-NULL if struct pci_dev is in use */
+};
+
+/* Access to pcistub_devices & seized_devices lists and the initialize_devices
+ * flag must be locked with pcistub_devices_lock
+ */
+static DEFINE_SPINLOCK(pcistub_devices_lock);
+static LIST_HEAD(pcistub_devices);
+
+/* wait for device_initcall before initializing our devices
+ * (see pcistub_init_devices_late)
+ */
+static int initialize_devices = 0;
+static LIST_HEAD(seized_devices);
+
+static struct pcistub_device *pcistub_device_alloc(struct pci_dev *dev)
+{
+	struct pcistub_device *psdev;
+
+	dev_dbg(&dev->dev, "pcistub_device_alloc\n");
+
+	psdev = kzalloc(sizeof(*psdev), GFP_ATOMIC);
+	if (!psdev)
+		return NULL;
+
+	psdev->dev = pci_dev_get(dev);
+	if (!psdev->dev) {
+		kfree(psdev);
+		return NULL;
+	}
+
+	kref_init(&psdev->kref);
+	spin_lock_init(&psdev->lock);
+
+	return psdev;
+}
+
+/* Don't call this directly as it's called by pcistub_device_put */
+static void pcistub_device_release(struct kref *kref)
+{
+	struct pcistub_device *psdev;
+
+	psdev = container_of(kref, struct pcistub_device, kref);
+
+	dev_dbg(&psdev->dev->dev, "pcistub_device_release\n");
+
+	/* Clean-up the device */
+	pciback_reset_device(psdev->dev);
+	pciback_config_free_dyn_fields(psdev->dev);
+	pciback_config_free_dev(psdev->dev);
+	kfree(pci_get_drvdata(psdev->dev));
+	pci_set_drvdata(psdev->dev, NULL);
+
+	pci_dev_put(psdev->dev);
+
+	kfree(psdev);
+}
+
+static inline void pcistub_device_get(struct pcistub_device *psdev)
+{
+	kref_get(&psdev->kref);
+}
+
+static inline void pcistub_device_put(struct pcistub_device *psdev)
+{
+	kref_put(&psdev->kref, pcistub_device_release);
+}
+
+static struct pcistub_device *pcistub_device_find(int domain, int bus,
+						  int slot, int func)
+{
+	struct pcistub_device *psdev = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&pcistub_devices_lock, flags);
+
+	list_for_each_entry(psdev, &pcistub_devices, dev_list) {
+		if (psdev->dev != NULL
+		    && domain == pci_domain_nr(psdev->dev->bus)
+		    && bus == psdev->dev->bus->number
+		    && PCI_DEVFN(slot, func) == psdev->dev->devfn) {
+			pcistub_device_get(psdev);
+			goto out;
+		}
+	}
+
+	/* didn't find it */
+	psdev = NULL;
+
+      out:
+	spin_unlock_irqrestore(&pcistub_devices_lock, flags);
+	return psdev;
+}
+
+static struct pci_dev *pcistub_device_get_pci_dev(struct pciback_device *pdev,
+						  struct pcistub_device *psdev)
+{
+	struct pci_dev *pci_dev = NULL;
+	unsigned long flags;
+
+	pcistub_device_get(psdev);
+
+	spin_lock_irqsave(&psdev->lock, flags);
+	if (!psdev->pdev) {
+		psdev->pdev = pdev;
+		pci_dev = psdev->dev;
+	}
+	spin_unlock_irqrestore(&psdev->lock, flags);
+
+	if (!pci_dev)
+		pcistub_device_put(psdev);
+
+	return pci_dev;
+}
+
+struct pci_dev *pcistub_get_pci_dev_by_slot(struct pciback_device *pdev,
+					    int domain, int bus,
+					    int slot, int func)
+{
+	struct pcistub_device *psdev;
+	struct pci_dev *found_dev = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&pcistub_devices_lock, flags);
+
+	list_for_each_entry(psdev, &pcistub_devices, dev_list) {
+		if (psdev->dev != NULL
+		    && domain == pci_domain_nr(psdev->dev->bus)
+		    && bus == psdev->dev->bus->number
+		    && PCI_DEVFN(slot, func) == psdev->dev->devfn) {
+			found_dev = pcistub_device_get_pci_dev(pdev, psdev);
+			break;
+		}
+	}
+
+	spin_unlock_irqrestore(&pcistub_devices_lock, flags);
+	return found_dev;
+}
+
+struct pci_dev *pcistub_get_pci_dev(struct pciback_device *pdev,
+				    struct pci_dev *dev)
+{
+	struct pcistub_device *psdev;
+	struct pci_dev *found_dev = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&pcistub_devices_lock, flags);
+
+	list_for_each_entry(psdev, &pcistub_devices, dev_list) {
+		if (psdev->dev == dev) {
+			found_dev = pcistub_device_get_pci_dev(pdev, psdev);
+			break;
+		}
+	}
+
+	spin_unlock_irqrestore(&pcistub_devices_lock, flags);
+	return found_dev;
+}
+
+void pcistub_put_pci_dev(struct pci_dev *dev)
+{
+	struct pcistub_device *psdev, *found_psdev = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&pcistub_devices_lock, flags);
+
+	list_for_each_entry(psdev, &pcistub_devices, dev_list) {
+		if (psdev->dev == dev) {
+			found_psdev = psdev;
+			break;
+		}
+	}
+
+	spin_unlock_irqrestore(&pcistub_devices_lock, flags);
+
+	/* Cleanup our device
+	 * (so it's ready for the next domain)
+	 */
+	pciback_reset_device(found_psdev->dev);
+	pciback_config_free_dyn_fields(found_psdev->dev);
+	pciback_config_reset_dev(found_psdev->dev);
+
+	spin_lock_irqsave(&found_psdev->lock, flags);
+	found_psdev->pdev = NULL;
+	spin_unlock_irqrestore(&found_psdev->lock, flags);
+
+	pcistub_device_put(found_psdev);
+}
+
+static int __devinit pcistub_match_one(struct pci_dev *dev,
+				       struct pcistub_device_id *pdev_id)
+{
+	/* Match the specified device by domain, bus, slot, func and also if
+	 * any of the device's parent bridges match.
+	 */
+	for (; dev != NULL; dev = dev->bus->self) {
+		if (pci_domain_nr(dev->bus) == pdev_id->domain
+		    && dev->bus->number == pdev_id->bus
+		    && dev->devfn == pdev_id->devfn)
+			return 1;
+
+		/* Sometimes topmost bridge links to itself. */
+		if (dev == dev->bus->self)
+			break;
+	}
+
+	return 0;
+}
+
+static int __devinit pcistub_match(struct pci_dev *dev)
+{
+	struct pcistub_device_id *pdev_id;
+	unsigned long flags;
+	int found = 0;
+
+	spin_lock_irqsave(&device_ids_lock, flags);
+	list_for_each_entry(pdev_id, &pcistub_device_ids, slot_list) {
+		if (pcistub_match_one(dev, pdev_id)) {
+			found = 1;
+			break;
+		}
+	}
+	spin_unlock_irqrestore(&device_ids_lock, flags);
+
+	return found;
+}
+
+static int __devinit pcistub_init_device(struct pci_dev *dev)
+{
+	struct pciback_dev_data *dev_data;
+	int err = 0;
+
+	dev_dbg(&dev->dev, "initializing...\n");
+
+	/* The PCI backend is not intended to be a module (or to work with
+	 * removable PCI devices (yet). If it were, pciback_config_free()
+	 * would need to be called somewhere to free the memory allocated
+	 * here and then to call kfree(pci_get_drvdata(psdev->dev)).
+	 */
+	dev_data = kzalloc(sizeof(*dev_data), GFP_ATOMIC);
+	if (!dev_data) {
+		err = -ENOMEM;
+		goto out;
+	}
+	pci_set_drvdata(dev, dev_data);
+
+	dev_dbg(&dev->dev, "initializing config\n");
+	err = pciback_config_init_dev(dev);
+	if (err)
+		goto out;
+
+	/* HACK: Force device (& ACPI) to determine what IRQ it's on - we
+	 * must do this here because pcibios_enable_device may specify
+	 * the pci device's true irq (and possibly its other resources)
+	 * if they differ from what's in the configuration space.
+	 * This makes the assumption that the device's resources won't
+	 * change after this point (otherwise this code may break!)
+	 */
+	dev_dbg(&dev->dev, "enabling device\n");
+	err = pci_enable_device(dev);
+	if (err)
+		goto config_release;
+
+	/* Now disable the device (this also ensures some private device
+	 * data is setup before we export)
+	 */
+	dev_dbg(&dev->dev, "reset device\n");
+	pciback_reset_device(dev);
+
+	return 0;
+
+      config_release:
+	pciback_config_free_dev(dev);
+
+      out:
+	pci_set_drvdata(dev, NULL);
+	kfree(dev_data);
+	return err;
+}
+
+/*
+ * Because some initialization still happens on
+ * devices during fs_initcall, we need to defer
+ * full initialization of our devices until
+ * device_initcall.
+ */
+static int __init pcistub_init_devices_late(void)
+{
+	struct pcistub_device *psdev;
+	unsigned long flags;
+	int err = 0;
+
+	pr_debug("pciback: pcistub_init_devices_late\n");
+
+	spin_lock_irqsave(&pcistub_devices_lock, flags);
+
+	while (!list_empty(&seized_devices)) {
+		psdev = container_of(seized_devices.next,
+				     struct pcistub_device, dev_list);
+		list_del(&psdev->dev_list);
+
+		spin_unlock_irqrestore(&pcistub_devices_lock, flags);
+
+		err = pcistub_init_device(psdev->dev);
+		if (err) {
+			dev_err(&psdev->dev->dev,
+				"error %d initializing device\n", err);
+			kfree(psdev);
+			psdev = NULL;
+		}
+
+		spin_lock_irqsave(&pcistub_devices_lock, flags);
+
+		if (psdev)
+			list_add_tail(&psdev->dev_list, &pcistub_devices);
+	}
+
+	initialize_devices = 1;
+
+	spin_unlock_irqrestore(&pcistub_devices_lock, flags);
+
+	return 0;
+}
+
+static int __devinit pcistub_seize(struct pci_dev *dev)
+{
+	struct pcistub_device *psdev;
+	unsigned long flags;
+	int err = 0;
+
+	psdev = pcistub_device_alloc(dev);
+	if (!psdev)
+		return -ENOMEM;
+
+	spin_lock_irqsave(&pcistub_devices_lock, flags);
+
+	if (initialize_devices) {
+		spin_unlock_irqrestore(&pcistub_devices_lock, flags);
+
+		/* don't want irqs disabled when calling pcistub_init_device */
+		err = pcistub_init_device(psdev->dev);
+
+		spin_lock_irqsave(&pcistub_devices_lock, flags);
+
+		if (!err)
+			list_add(&psdev->dev_list, &pcistub_devices);
+	} else {
+		dev_dbg(&dev->dev, "deferring initialization\n");
+		list_add(&psdev->dev_list, &seized_devices);
+	}
+
+	spin_unlock_irqrestore(&pcistub_devices_lock, flags);
+
+	if (err)
+		pcistub_device_put(psdev);
+
+	return err;
+}
+
+static int __devinit pcistub_probe(struct pci_dev *dev,
+				   const struct pci_device_id *id)
+{
+	int err = 0;
+
+	dev_dbg(&dev->dev, "probing...\n");
+
+	if (pcistub_match(dev)) {
+
+		if (dev->hdr_type != PCI_HEADER_TYPE_NORMAL
+		    && dev->hdr_type != PCI_HEADER_TYPE_BRIDGE) {
+			dev_err(&dev->dev, "can't export pci devices that "
+				"don't have a normal (0) or bridge (1) "
+				"header type!\n");
+			err = -ENODEV;
+			goto out;
+		}
+
+		dev_info(&dev->dev, "seizing device\n");
+		err = pcistub_seize(dev);
+	} else
+		/* Didn't find the device */
+		err = -ENODEV;
+
+      out:
+	return err;
+}
+
+static void pcistub_remove(struct pci_dev *dev)
+{
+	struct pcistub_device *psdev, *found_psdev = NULL;
+	unsigned long flags;
+
+	dev_dbg(&dev->dev, "removing\n");
+
+	spin_lock_irqsave(&pcistub_devices_lock, flags);
+
+	pciback_config_quirk_release(dev);
+
+	list_for_each_entry(psdev, &pcistub_devices, dev_list) {
+		if (psdev->dev == dev) {
+			found_psdev = psdev;
+			break;
+		}
+	}
+
+	spin_unlock_irqrestore(&pcistub_devices_lock, flags);
+
+	if (found_psdev) {
+		dev_dbg(&dev->dev, "found device to remove - in use? %p\n",
+			found_psdev->pdev);
+
+		if (found_psdev->pdev) {
+			printk(KERN_WARNING "pciback: ****** removing device "
+			       "%s while still in-use! ******\n",
+			       pci_name(found_psdev->dev));
+			printk(KERN_WARNING "pciback: ****** driver domain may "
+			       "still access this device's i/o resources!\n");
+			printk(KERN_WARNING "pciback: ****** shutdown driver "
+			       "domain before binding device\n");
+			printk(KERN_WARNING "pciback: ****** to other drivers "
+			       "or domains\n");
+
+			pciback_release_pci_dev(found_psdev->pdev,
+						found_psdev->dev);
+		}
+
+		spin_lock_irqsave(&pcistub_devices_lock, flags);
+		list_del(&found_psdev->dev_list);
+		spin_unlock_irqrestore(&pcistub_devices_lock, flags);
+
+		/* the final put for releasing from the list */
+		pcistub_device_put(found_psdev);
+	}
+}
+
+static struct pci_device_id pcistub_ids[] = {
+	{
+	 .vendor = PCI_ANY_ID,
+	 .device = PCI_ANY_ID,
+	 .subvendor = PCI_ANY_ID,
+	 .subdevice = PCI_ANY_ID,
+	 },
+	{0,},
+};
+
+/*
+ * Note: There is no MODULE_DEVICE_TABLE entry here because this isn't
+ * for a normal device. I don't want it to be loaded automatically.
+ */
+
+static struct pci_driver pciback_pci_driver = {
+	.name = "pciback",
+	.id_table = pcistub_ids,
+	.probe = pcistub_probe,
+	.remove = pcistub_remove,
+};
+
+static inline int str_to_slot(const char *buf, int *domain, int *bus,
+			      int *slot, int *func)
+{
+	int err;
+
+	err = sscanf(buf, " %x:%x:%x.%x", domain, bus, slot, func);
+	if (err == 4)
+		return 0;
+	else if (err < 0)
+		return -EINVAL;
+
+	/* try again without domain */
+	*domain = 0;
+	err = sscanf(buf, " %x:%x.%x", bus, slot, func);
+	if (err == 3)
+		return 0;
+
+	return -EINVAL;
+}
+
+static inline int str_to_quirk(const char *buf, int *domain, int *bus, int
+			       *slot, int *func, int *reg, int *size, int *mask)
+{
+	int err;
+
+	err =
+	    sscanf(buf, " %04x:%02x:%02x.%1x-%08x:%1x:%08x", domain, bus, slot,
+		   func, reg, size, mask);
+	if (err == 7)
+		return 0;
+	return -EINVAL;
+}
+
+static int pcistub_device_id_add(int domain, int bus, int slot, int func)
+{
+	struct pcistub_device_id *pci_dev_id;
+	unsigned long flags;
+
+	pci_dev_id = kmalloc(sizeof(*pci_dev_id), GFP_KERNEL);
+	if (!pci_dev_id)
+		return -ENOMEM;
+
+	pci_dev_id->domain = domain;
+	pci_dev_id->bus = bus;
+	pci_dev_id->devfn = PCI_DEVFN(slot, func);
+
+	pr_debug("pciback: wants to seize %04x:%02x:%02x.%01x\n",
+		 domain, bus, slot, func);
+
+	spin_lock_irqsave(&device_ids_lock, flags);
+	list_add_tail(&pci_dev_id->slot_list, &pcistub_device_ids);
+	spin_unlock_irqrestore(&device_ids_lock, flags);
+
+	return 0;
+}
+
+static int pcistub_device_id_remove(int domain, int bus, int slot, int func)
+{
+	struct pcistub_device_id *pci_dev_id, *t;
+	int devfn = PCI_DEVFN(slot, func);
+	int err = -ENOENT;
+	unsigned long flags;
+
+	spin_lock_irqsave(&device_ids_lock, flags);
+	list_for_each_entry_safe(pci_dev_id, t, &pcistub_device_ids, slot_list) {
+
+		if (pci_dev_id->domain == domain
+		    && pci_dev_id->bus == bus && pci_dev_id->devfn == devfn) {
+			/* Don't break; here because it's possible the same
+			 * slot could be in the list more than once
+			 */
+			list_del(&pci_dev_id->slot_list);
+			kfree(pci_dev_id);
+
+			err = 0;
+
+			pr_debug("pciback: removed %04x:%02x:%02x.%01x from "
+				 "seize list\n", domain, bus, slot, func);
+		}
+	}
+	spin_unlock_irqrestore(&device_ids_lock, flags);
+
+	return err;
+}
+
+static int pcistub_reg_add(int domain, int bus, int slot, int func, int reg,
+			   int size, int mask)
+{
+	int err = 0;
+	struct pcistub_device *psdev;
+	struct pci_dev *dev;
+	struct config_field *field;
+
+	psdev = pcistub_device_find(domain, bus, slot, func);
+	if (!psdev || !psdev->dev) {
+		err = -ENODEV;
+		goto out;
+	}
+	dev = psdev->dev;
+
+	field = kzalloc(sizeof(*field), GFP_ATOMIC);
+	if (!field) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	field->offset = reg;
+	field->size = size;
+	field->mask = mask;
+	field->init = NULL;
+	field->reset = NULL;
+	field->release = NULL;
+	field->clean = pciback_config_field_free;
+
+	err = pciback_config_quirks_add_field(dev, field);
+	if (err)
+		kfree(field);
+      out:
+	return err;
+}
+
+static ssize_t pcistub_slot_add(struct device_driver *drv, const char *buf,
+				size_t count)
+{
+	int domain, bus, slot, func;
+	int err;
+
+	err = str_to_slot(buf, &domain, &bus, &slot, &func);
+	if (err)
+		goto out;
+
+	err = pcistub_device_id_add(domain, bus, slot, func);
+
+      out:
+	if (!err)
+		err = count;
+	return err;
+}
+
+DRIVER_ATTR(new_slot, S_IWUSR, NULL, pcistub_slot_add);
+
+static ssize_t pcistub_slot_remove(struct device_driver *drv, const char *buf,
+				   size_t count)
+{
+	int domain, bus, slot, func;
+	int err;
+
+	err = str_to_slot(buf, &domain, &bus, &slot, &func);
+	if (err)
+		goto out;
+
+	err = pcistub_device_id_remove(domain, bus, slot, func);
+
+      out:
+	if (!err)
+		err = count;
+	return err;
+}
+
+DRIVER_ATTR(remove_slot, S_IWUSR, NULL, pcistub_slot_remove);
+
+static ssize_t pcistub_slot_show(struct device_driver *drv, char *buf)
+{
+	struct pcistub_device_id *pci_dev_id;
+	size_t count = 0;
+	unsigned long flags;
+
+	spin_lock_irqsave(&device_ids_lock, flags);
+	list_for_each_entry(pci_dev_id, &pcistub_device_ids, slot_list) {
+		if (count >= PAGE_SIZE)
+			break;
+
+		count += scnprintf(buf + count, PAGE_SIZE - count,
+				   "%04x:%02x:%02x.%01x\n",
+				   pci_dev_id->domain, pci_dev_id->bus,
+				   PCI_SLOT(pci_dev_id->devfn),
+				   PCI_FUNC(pci_dev_id->devfn));
+	}
+	spin_unlock_irqrestore(&device_ids_lock, flags);
+
+	return count;
+}
+
+DRIVER_ATTR(slots, S_IRUSR, pcistub_slot_show, NULL);
+
+static ssize_t pcistub_quirk_add(struct device_driver *drv, const char *buf,
+				 size_t count)
+{
+	int domain, bus, slot, func, reg, size, mask;
+	int err;
+
+	err = str_to_quirk(buf, &domain, &bus, &slot, &func, &reg, &size,
+			   &mask);
+	if (err)
+		goto out;
+
+	err = pcistub_reg_add(domain, bus, slot, func, reg, size, mask);
+
+      out:
+	if (!err)
+		err = count;
+	return err;
+}
+
+static ssize_t pcistub_quirk_show(struct device_driver *drv, char *buf)
+{
+	int count = 0;
+	unsigned long flags;
+	extern struct list_head pciback_quirks;
+	struct pciback_config_quirk *quirk;
+	struct pciback_dev_data *dev_data;
+	struct config_field *field;
+	struct config_field_entry *cfg_entry;
+
+	spin_lock_irqsave(&device_ids_lock, flags);
+	list_for_each_entry(quirk, &pciback_quirks, quirks_list) {
+		if (count >= PAGE_SIZE)
+			goto out;
+
+		count += scnprintf(buf + count, PAGE_SIZE - count,
+				   "%02x:%02x.%01x\n\t%04x:%04x:%04x:%04x\n",
+				   quirk->pdev->bus->number,
+				   PCI_SLOT(quirk->pdev->devfn),
+				   PCI_FUNC(quirk->pdev->devfn),
+				   quirk->devid.vendor, quirk->devid.device,
+				   quirk->devid.subvendor,
+				   quirk->devid.subdevice);
+
+		dev_data = pci_get_drvdata(quirk->pdev);
+
+		list_for_each_entry(cfg_entry, &dev_data->config_fields, list) {
+			field = cfg_entry->field;
+			if (count >= PAGE_SIZE)
+				goto out;
+
+			count += scnprintf(buf + count, PAGE_SIZE - count,
+					   "\t\t%08x:%01x:%08x\n",
+					   cfg_entry->base_offset + field->offset, 
+					   field->size, field->mask);
+		}
+	}
+
+      out:
+	spin_unlock_irqrestore(&device_ids_lock, flags);
+
+	return count;
+}
+
+DRIVER_ATTR(quirks, S_IRUSR | S_IWUSR, pcistub_quirk_show, pcistub_quirk_add);
+
+static ssize_t permissive_add(struct device_driver *drv, const char *buf,
+			      size_t count)
+{
+	int domain, bus, slot, func;
+	int err;
+	struct pcistub_device *psdev;
+	struct pciback_dev_data *dev_data;
+	err = str_to_slot(buf, &domain, &bus, &slot, &func);
+	if (err)
+		goto out;
+	psdev = pcistub_device_find(domain, bus, slot, func);
+	if (!psdev) {
+		err = -ENODEV;
+		goto out;
+	}
+	if (!psdev->dev) {
+		err = -ENODEV;
+		goto release;
+	}
+	dev_data = pci_get_drvdata(psdev->dev);
+	/* the driver data for a device should never be null at this point */
+	if (!dev_data) {
+		err = -ENXIO;
+		goto release;
+	}
+	if (!dev_data->permissive) {
+		dev_data->permissive = 1;
+		/* Let user know that what they're doing could be unsafe */
+		dev_warn(&psdev->dev->dev,
+			 "enabling permissive mode configuration space accesses!\n");
+		dev_warn(&psdev->dev->dev,
+			 "permissive mode is potentially unsafe!\n");
+	}
+      release:
+	pcistub_device_put(psdev);
+      out:
+	if (!err)
+		err = count;
+	return err;
+}
+
+static ssize_t permissive_show(struct device_driver *drv, char *buf)
+{
+	struct pcistub_device *psdev;
+	struct pciback_dev_data *dev_data;
+	size_t count = 0;
+	unsigned long flags;
+	spin_lock_irqsave(&pcistub_devices_lock, flags);
+	list_for_each_entry(psdev, &pcistub_devices, dev_list) {
+		if (count >= PAGE_SIZE)
+			break;
+		if (!psdev->dev)
+			continue;
+		dev_data = pci_get_drvdata(psdev->dev);
+		if (!dev_data || !dev_data->permissive)
+			continue;
+		count +=
+		    scnprintf(buf + count, PAGE_SIZE - count, "%s\n",
+			      pci_name(psdev->dev));
+	}
+	spin_unlock_irqrestore(&pcistub_devices_lock, flags);
+	return count;
+}
+
+DRIVER_ATTR(permissive, S_IRUSR | S_IWUSR, permissive_show, permissive_add);
+
+#ifdef CONFIG_PCI_MSI
+
+int pciback_get_owner(struct pci_dev *dev)
+{
+	struct pcistub_device *psdev;
+
+	psdev = pcistub_device_find(pci_domain_nr(dev->bus), dev->bus->number,
+			PCI_SLOT(dev->devfn), PCI_FUNC(dev->devfn));
+
+	if (!psdev || !psdev->pdev)
+		return -1;
+
+	return psdev->pdev->xdev->otherend_id;
+}
+#endif
+
+static void pcistub_exit(void)
+{
+	driver_remove_file(&pciback_pci_driver.driver, &driver_attr_new_slot);
+	driver_remove_file(&pciback_pci_driver.driver,
+			   &driver_attr_remove_slot);
+	driver_remove_file(&pciback_pci_driver.driver, &driver_attr_slots);
+	driver_remove_file(&pciback_pci_driver.driver, &driver_attr_quirks);
+	driver_remove_file(&pciback_pci_driver.driver, &driver_attr_permissive);
+
+	pci_unregister_driver(&pciback_pci_driver);
+	WARN_ON(unregister_msi_get_owner(pciback_get_owner));
+}
+
+static int __init pcistub_init(void)
+{
+	int pos = 0;
+	int err = 0;
+	int domain, bus, slot, func;
+	int parsed;
+
+	if (pci_devs_to_hide && *pci_devs_to_hide) {
+		do {
+			parsed = 0;
+
+			err = sscanf(pci_devs_to_hide + pos,
+				     " (%x:%x:%x.%x) %n",
+				     &domain, &bus, &slot, &func, &parsed);
+			if (err != 4) {
+				domain = 0;
+				err = sscanf(pci_devs_to_hide + pos,
+					     " (%x:%x.%x) %n",
+					     &bus, &slot, &func, &parsed);
+				if (err != 3)
+					goto parse_error;
+			}
+
+			err = pcistub_device_id_add(domain, bus, slot, func);
+			if (err)
+				goto out;
+
+			/* if parsed<=0, we've reached the end of the string */
+			pos += parsed;
+		} while (parsed > 0 && pci_devs_to_hide[pos]);
+	}
+
+	/* If we're the first PCI Device Driver to register, we're the
+	 * first one to get offered PCI devices as they become
+	 * available (and thus we can be the first to grab them)
+	 */
+	err = pci_register_driver(&pciback_pci_driver);
+	if (err < 0)
+		goto out;
+
+	err = driver_create_file(&pciback_pci_driver.driver,
+				 &driver_attr_new_slot);
+	if (!err)
+		err = driver_create_file(&pciback_pci_driver.driver,
+					 &driver_attr_remove_slot);
+	if (!err)
+		err = driver_create_file(&pciback_pci_driver.driver,
+					 &driver_attr_slots);
+	if (!err)
+		err = driver_create_file(&pciback_pci_driver.driver,
+					 &driver_attr_quirks);
+	if (!err)
+		err = driver_create_file(&pciback_pci_driver.driver,
+					 &driver_attr_permissive);
+
+	if (!err)
+		err = register_msi_get_owner(pciback_get_owner);
+	if (err)
+		pcistub_exit();
+
+      out:
+	return err;
+
+      parse_error:
+	printk(KERN_ERR "pciback: Error parsing pci_devs_to_hide at \"%s\"\n",
+	       pci_devs_to_hide + pos);
+	return -EINVAL;
+}
+
+#ifndef MODULE
+/*
+ * fs_initcall happens before device_initcall
+ * so pciback *should* get called first (b/c we 
+ * want to suck up any device before other drivers
+ * get a chance by being the first pci device
+ * driver to register)
+ */
+fs_initcall(pcistub_init);
+#endif
+
+static int __init pciback_init(void)
+{
+	int err;
+
+	err = pciback_config_init();
+	if (err)
+		return err;
+
+#ifdef MODULE
+	err = pcistub_init();
+	if (err < 0)
+		return err;
+#endif
+
+	pcistub_init_devices_late();
+	err = pciback_xenbus_register();
+	if (err)
+		pcistub_exit();
+
+	return err;
+}
+
+static void __exit pciback_cleanup(void)
+{
+	pciback_xenbus_unregister();
+	pcistub_exit();
+}
+
+module_init(pciback_init);
+module_exit(pciback_cleanup);
+
+MODULE_LICENSE("Dual BSD/GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/slot.c linux-2.6.18-xen-3.3.0/drivers/xen/pciback/slot.c
--- linux-2.6.18.8/drivers/xen/pciback/slot.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/slot.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,157 @@
+/*
+ * PCI Backend - Provides a Virtual PCI bus (with real devices)
+ *               to the frontend
+ *
+ *   Author: Ryan Wilson <hap9@epoch.ncsc.mil> (vpci.c)
+ *   Author: Tristan Gingold <tristan.gingold@bull.net>, from vpci.c
+ */
+
+#include <linux/list.h>
+#include <linux/slab.h>
+#include <linux/pci.h>
+#include <linux/spinlock.h>
+#include "pciback.h"
+
+/* There are at most 32 slots in a pci bus.  */
+#define PCI_SLOT_MAX 32
+
+#define PCI_BUS_NBR 2
+
+struct slot_dev_data {
+	/* Access to dev_list must be protected by lock */
+	struct pci_dev *slots[PCI_BUS_NBR][PCI_SLOT_MAX];
+	spinlock_t lock;
+};
+
+struct pci_dev *pciback_get_pci_dev(struct pciback_device *pdev,
+				    unsigned int domain, unsigned int bus,
+				    unsigned int devfn)
+{
+	struct pci_dev *dev = NULL;
+	struct slot_dev_data *slot_dev = pdev->pci_dev_data;
+	unsigned long flags;
+
+	if (domain != 0 || PCI_FUNC(devfn) != 0)
+		return NULL;
+
+	if (PCI_SLOT(devfn) >= PCI_SLOT_MAX || bus >= PCI_BUS_NBR)
+		return NULL;
+
+	spin_lock_irqsave(&slot_dev->lock, flags);
+	dev = slot_dev->slots[bus][PCI_SLOT(devfn)];
+	spin_unlock_irqrestore(&slot_dev->lock, flags);
+
+	return dev;
+}
+
+int pciback_add_pci_dev(struct pciback_device *pdev, struct pci_dev *dev,
+			int devid, publish_pci_dev_cb publish_cb)
+{
+	int err = 0, slot, bus;
+	struct slot_dev_data *slot_dev = pdev->pci_dev_data;
+	unsigned long flags;
+
+	if ((dev->class >> 24) == PCI_BASE_CLASS_BRIDGE) {
+		err = -EFAULT;
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Can't export bridges on the virtual PCI bus");
+		goto out;
+	}
+
+	spin_lock_irqsave(&slot_dev->lock, flags);
+
+	/* Assign to a new slot on the virtual PCI bus */
+	for (bus = 0; bus < PCI_BUS_NBR; bus++)
+		for (slot = 0; slot < PCI_SLOT_MAX; slot++) {
+			if (slot_dev->slots[bus][slot] == NULL) {
+				printk(KERN_INFO
+				       "pciback: slot: %s: assign to virtual slot %d, bus %d\n",
+				       pci_name(dev), slot, bus);
+				slot_dev->slots[bus][slot] = dev;
+				goto unlock;
+			}
+		}
+
+	err = -ENOMEM;
+	xenbus_dev_fatal(pdev->xdev, err,
+			 "No more space on root virtual PCI bus");
+
+      unlock:
+	spin_unlock_irqrestore(&slot_dev->lock, flags);
+
+	/* Publish this device. */
+	if(!err)
+		err = publish_cb(pdev, 0, 0, PCI_DEVFN(slot, 0), devid);
+
+      out:
+	return err;
+}
+
+void pciback_release_pci_dev(struct pciback_device *pdev, struct pci_dev *dev)
+{
+	int slot, bus;
+	struct slot_dev_data *slot_dev = pdev->pci_dev_data;
+	struct pci_dev *found_dev = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&slot_dev->lock, flags);
+
+	for (bus = 0; bus < PCI_BUS_NBR; bus++)
+		for (slot = 0; slot < PCI_SLOT_MAX; slot++) {
+			if (slot_dev->slots[bus][slot] == dev) {
+				slot_dev->slots[bus][slot] = NULL;
+				found_dev = dev;
+				goto out;
+			}
+		}
+
+      out:
+	spin_unlock_irqrestore(&slot_dev->lock, flags);
+
+	if (found_dev)
+		pcistub_put_pci_dev(found_dev);
+}
+
+int pciback_init_devices(struct pciback_device *pdev)
+{
+	int slot, bus;
+	struct slot_dev_data *slot_dev;
+
+	slot_dev = kmalloc(sizeof(*slot_dev), GFP_KERNEL);
+	if (!slot_dev)
+		return -ENOMEM;
+
+	spin_lock_init(&slot_dev->lock);
+
+	for (bus = 0; bus < PCI_BUS_NBR; bus++)
+		for (slot = 0; slot < PCI_SLOT_MAX; slot++)
+			slot_dev->slots[bus][slot] = NULL;
+
+	pdev->pci_dev_data = slot_dev;
+
+	return 0;
+}
+
+int pciback_publish_pci_roots(struct pciback_device *pdev,
+			      publish_pci_root_cb publish_cb)
+{
+	/* The Virtual PCI bus has only one root */
+	return publish_cb(pdev, 0, 0);
+}
+
+void pciback_release_devices(struct pciback_device *pdev)
+{
+	int slot, bus;
+	struct slot_dev_data *slot_dev = pdev->pci_dev_data;
+	struct pci_dev *dev;
+
+	for (bus = 0; bus < PCI_BUS_NBR; bus++)
+		for (slot = 0; slot < PCI_SLOT_MAX; slot++) {
+			dev = slot_dev->slots[bus][slot];
+			if (dev != NULL)
+				pcistub_put_pci_dev(dev);
+		}
+
+	kfree(slot_dev);
+	pdev->pci_dev_data = NULL;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/vpci.c linux-2.6.18-xen-3.3.0/drivers/xen/pciback/vpci.c
--- linux-2.6.18.8/drivers/xen/pciback/vpci.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/vpci.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,212 @@
+/*
+ * PCI Backend - Provides a Virtual PCI bus (with real devices)
+ *               to the frontend
+ *
+ *   Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+
+#include <linux/list.h>
+#include <linux/slab.h>
+#include <linux/pci.h>
+#include <linux/spinlock.h>
+#include "pciback.h"
+
+#define PCI_SLOT_MAX 32
+
+struct vpci_dev_data {
+	/* Access to dev_list must be protected by lock */
+	struct list_head dev_list[PCI_SLOT_MAX];
+	spinlock_t lock;
+};
+
+static inline struct list_head *list_first(struct list_head *head)
+{
+	return head->next;
+}
+
+struct pci_dev *pciback_get_pci_dev(struct pciback_device *pdev,
+				    unsigned int domain, unsigned int bus,
+				    unsigned int devfn)
+{
+	struct pci_dev_entry *entry;
+	struct pci_dev *dev = NULL;
+	struct vpci_dev_data *vpci_dev = pdev->pci_dev_data;
+	unsigned long flags;
+
+	if (domain != 0 || bus != 0)
+		return NULL;
+
+	if (PCI_SLOT(devfn) < PCI_SLOT_MAX) {
+		spin_lock_irqsave(&vpci_dev->lock, flags);
+
+		list_for_each_entry(entry,
+				    &vpci_dev->dev_list[PCI_SLOT(devfn)],
+				    list) {
+			if (PCI_FUNC(entry->dev->devfn) == PCI_FUNC(devfn)) {
+				dev = entry->dev;
+				break;
+			}
+		}
+
+		spin_unlock_irqrestore(&vpci_dev->lock, flags);
+	}
+	return dev;
+}
+
+static inline int match_slot(struct pci_dev *l, struct pci_dev *r)
+{
+	if (pci_domain_nr(l->bus) == pci_domain_nr(r->bus)
+	    && l->bus == r->bus && PCI_SLOT(l->devfn) == PCI_SLOT(r->devfn))
+		return 1;
+
+	return 0;
+}
+
+int pciback_add_pci_dev(struct pciback_device *pdev, struct pci_dev *dev,
+			int devid, publish_pci_dev_cb publish_cb)
+{
+	int err = 0, slot, func;
+	struct pci_dev_entry *t, *dev_entry;
+	struct vpci_dev_data *vpci_dev = pdev->pci_dev_data;
+	unsigned long flags;
+
+	if ((dev->class >> 24) == PCI_BASE_CLASS_BRIDGE) {
+		err = -EFAULT;
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Can't export bridges on the virtual PCI bus");
+		goto out;
+	}
+
+	dev_entry = kmalloc(sizeof(*dev_entry), GFP_KERNEL);
+	if (!dev_entry) {
+		err = -ENOMEM;
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Error adding entry to virtual PCI bus");
+		goto out;
+	}
+
+	dev_entry->dev = dev;
+
+	spin_lock_irqsave(&vpci_dev->lock, flags);
+
+	/* Keep multi-function devices together on the virtual PCI bus */
+	for (slot = 0; slot < PCI_SLOT_MAX; slot++) {
+		if (!list_empty(&vpci_dev->dev_list[slot])) {
+			t = list_entry(list_first(&vpci_dev->dev_list[slot]),
+				       struct pci_dev_entry, list);
+
+			if (match_slot(dev, t->dev)) {
+				pr_info("pciback: vpci: %s: "
+					"assign to virtual slot %d func %d\n",
+					pci_name(dev), slot,
+					PCI_FUNC(dev->devfn));
+				list_add_tail(&dev_entry->list,
+					      &vpci_dev->dev_list[slot]);
+				func = PCI_FUNC(dev->devfn);
+				goto unlock;
+			}
+		}
+	}
+
+	/* Assign to a new slot on the virtual PCI bus */
+	for (slot = 0; slot < PCI_SLOT_MAX; slot++) {
+		if (list_empty(&vpci_dev->dev_list[slot])) {
+			printk(KERN_INFO
+			       "pciback: vpci: %s: assign to virtual slot %d\n",
+			       pci_name(dev), slot);
+			list_add_tail(&dev_entry->list,
+				      &vpci_dev->dev_list[slot]);
+			func = PCI_FUNC(dev->devfn);
+			goto unlock;
+		}
+	}
+
+	err = -ENOMEM;
+	xenbus_dev_fatal(pdev->xdev, err,
+			 "No more space on root virtual PCI bus");
+
+      unlock:
+	spin_unlock_irqrestore(&vpci_dev->lock, flags);
+
+	/* Publish this device. */
+	if(!err)
+		err = publish_cb(pdev, 0, 0, PCI_DEVFN(slot, func), devid);
+
+      out:
+	return err;
+}
+
+void pciback_release_pci_dev(struct pciback_device *pdev, struct pci_dev *dev)
+{
+	int slot;
+	struct vpci_dev_data *vpci_dev = pdev->pci_dev_data;
+	struct pci_dev *found_dev = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&vpci_dev->lock, flags);
+
+	for (slot = 0; slot < PCI_SLOT_MAX; slot++) {
+		struct pci_dev_entry *e, *tmp;
+		list_for_each_entry_safe(e, tmp, &vpci_dev->dev_list[slot],
+					 list) {
+			if (e->dev == dev) {
+				list_del(&e->list);
+				found_dev = e->dev;
+				kfree(e);
+				goto out;
+			}
+		}
+	}
+
+      out:
+	spin_unlock_irqrestore(&vpci_dev->lock, flags);
+
+	if (found_dev)
+		pcistub_put_pci_dev(found_dev);
+}
+
+int pciback_init_devices(struct pciback_device *pdev)
+{
+	int slot;
+	struct vpci_dev_data *vpci_dev;
+
+	vpci_dev = kmalloc(sizeof(*vpci_dev), GFP_KERNEL);
+	if (!vpci_dev)
+		return -ENOMEM;
+
+	spin_lock_init(&vpci_dev->lock);
+
+	for (slot = 0; slot < PCI_SLOT_MAX; slot++) {
+		INIT_LIST_HEAD(&vpci_dev->dev_list[slot]);
+	}
+
+	pdev->pci_dev_data = vpci_dev;
+
+	return 0;
+}
+
+int pciback_publish_pci_roots(struct pciback_device *pdev,
+			      publish_pci_root_cb publish_cb)
+{
+	/* The Virtual PCI bus has only one root */
+	return publish_cb(pdev, 0, 0);
+}
+
+void pciback_release_devices(struct pciback_device *pdev)
+{
+	int slot;
+	struct vpci_dev_data *vpci_dev = pdev->pci_dev_data;
+
+	for (slot = 0; slot < PCI_SLOT_MAX; slot++) {
+		struct pci_dev_entry *e, *tmp;
+		list_for_each_entry_safe(e, tmp, &vpci_dev->dev_list[slot],
+					 list) {
+			list_del(&e->list);
+			pcistub_put_pci_dev(e->dev);
+			kfree(e);
+		}
+	}
+
+	kfree(vpci_dev);
+	pdev->pci_dev_data = NULL;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/pciback/xenbus.c linux-2.6.18-xen-3.3.0/drivers/xen/pciback/xenbus.c
--- linux-2.6.18.8/drivers/xen/pciback/xenbus.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pciback/xenbus.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,704 @@
+/*
+ * PCI Backend Xenbus Setup - handles setup with frontend and xend
+ *
+ *   Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/list.h>
+#include <linux/vmalloc.h>
+#include <xen/xenbus.h>
+#include <xen/evtchn.h>
+#include "pciback.h"
+
+#define INVALID_EVTCHN_IRQ  (-1)
+
+static struct pciback_device *alloc_pdev(struct xenbus_device *xdev)
+{
+	struct pciback_device *pdev;
+
+	pdev = kzalloc(sizeof(struct pciback_device), GFP_KERNEL);
+	if (pdev == NULL)
+		goto out;
+	dev_dbg(&xdev->dev, "allocated pdev @ 0x%p\n", pdev);
+
+	pdev->xdev = xdev;
+	xdev->dev.driver_data = pdev;
+
+	spin_lock_init(&pdev->dev_lock);
+
+	pdev->sh_area = NULL;
+	pdev->sh_info = NULL;
+	pdev->evtchn_irq = INVALID_EVTCHN_IRQ;
+	pdev->be_watching = 0;
+
+	INIT_WORK(&pdev->op_work, pciback_do_op, pdev);
+
+	if (pciback_init_devices(pdev)) {
+		kfree(pdev);
+		pdev = NULL;
+	}
+      out:
+	return pdev;
+}
+
+static void pciback_disconnect(struct pciback_device *pdev)
+{
+	spin_lock(&pdev->dev_lock);
+
+	/* Ensure the guest can't trigger our handler before removing devices */
+	if (pdev->evtchn_irq != INVALID_EVTCHN_IRQ) {
+		unbind_from_irqhandler(pdev->evtchn_irq, pdev);
+		pdev->evtchn_irq = INVALID_EVTCHN_IRQ;
+	}
+
+	/* If the driver domain started an op, make sure we complete it or
+	 * delete it before releasing the shared memory */
+	cancel_delayed_work(&pdev->op_work);
+	flush_scheduled_work();
+
+	if (pdev->sh_info != NULL) {
+		xenbus_unmap_ring_vfree(pdev->xdev, pdev->sh_area);
+		pdev->sh_info = NULL;
+	}
+
+	spin_unlock(&pdev->dev_lock);
+}
+
+static void free_pdev(struct pciback_device *pdev)
+{
+	if (pdev->be_watching)
+		unregister_xenbus_watch(&pdev->be_watch);
+
+	pciback_disconnect(pdev);
+
+	pciback_release_devices(pdev);
+
+	pdev->xdev->dev.driver_data = NULL;
+	pdev->xdev = NULL;
+
+	kfree(pdev);
+}
+
+static int pciback_do_attach(struct pciback_device *pdev, int gnt_ref,
+			     int remote_evtchn)
+{
+	int err = 0;
+	struct vm_struct *area;
+
+	dev_dbg(&pdev->xdev->dev,
+		"Attaching to frontend resources - gnt_ref=%d evtchn=%d\n",
+		gnt_ref, remote_evtchn);
+
+	area = xenbus_map_ring_valloc(pdev->xdev, gnt_ref);
+	if (IS_ERR(area)) {
+		err = PTR_ERR(area);
+		goto out;
+	}
+	pdev->sh_area = area;
+	pdev->sh_info = area->addr;
+
+	err = bind_interdomain_evtchn_to_irqhandler(
+		pdev->xdev->otherend_id, remote_evtchn, pciback_handle_event,
+		SA_SAMPLE_RANDOM, "pciback", pdev);
+	if (err < 0) {
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Error binding event channel to IRQ");
+		goto out;
+	}
+	pdev->evtchn_irq = err;
+	err = 0;
+
+	dev_dbg(&pdev->xdev->dev, "Attached!\n");
+      out:
+	return err;
+}
+
+static int pciback_attach(struct pciback_device *pdev)
+{
+	int err = 0;
+	int gnt_ref, remote_evtchn;
+	char *magic = NULL;
+
+	spin_lock(&pdev->dev_lock);
+
+	/* Make sure we only do this setup once */
+	if (xenbus_read_driver_state(pdev->xdev->nodename) !=
+	    XenbusStateInitialised)
+		goto out;
+
+	/* Wait for frontend to state that it has published the configuration */
+	if (xenbus_read_driver_state(pdev->xdev->otherend) !=
+	    XenbusStateInitialised)
+		goto out;
+
+	dev_dbg(&pdev->xdev->dev, "Reading frontend config\n");
+
+	err = xenbus_gather(XBT_NIL, pdev->xdev->otherend,
+			    "pci-op-ref", "%u", &gnt_ref,
+			    "event-channel", "%u", &remote_evtchn,
+			    "magic", NULL, &magic, NULL);
+	if (err) {
+		/* If configuration didn't get read correctly, wait longer */
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Error reading configuration from frontend");
+		goto out;
+	}
+
+	if (magic == NULL || strcmp(magic, XEN_PCI_MAGIC) != 0) {
+		xenbus_dev_fatal(pdev->xdev, -EFAULT,
+				 "version mismatch (%s/%s) with pcifront - "
+				 "halting pciback",
+				 magic, XEN_PCI_MAGIC);
+		goto out;
+	}
+
+	err = pciback_do_attach(pdev, gnt_ref, remote_evtchn);
+	if (err)
+		goto out;
+
+	dev_dbg(&pdev->xdev->dev, "Connecting...\n");
+
+	err = xenbus_switch_state(pdev->xdev, XenbusStateConnected);
+	if (err)
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Error switching to connected state!");
+
+	dev_dbg(&pdev->xdev->dev, "Connected? %d\n", err);
+      out:
+	spin_unlock(&pdev->dev_lock);
+
+	if (magic)
+		kfree(magic);
+
+	return err;
+}
+
+static int pciback_publish_pci_dev(struct pciback_device *pdev,
+				   unsigned int domain, unsigned int bus,
+				   unsigned int devfn, unsigned int devid)
+{
+	int err;
+	int len;
+	char str[64];
+
+	len = snprintf(str, sizeof(str), "vdev-%d", devid);
+	if (unlikely(len >= (sizeof(str) - 1))) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	err = xenbus_printf(XBT_NIL, pdev->xdev->nodename, str,
+			    "%04x:%02x:%02x.%02x", domain, bus,
+			    PCI_SLOT(devfn), PCI_FUNC(devfn));
+
+      out:
+	return err;
+}
+
+static int pciback_export_device(struct pciback_device *pdev,
+				 int domain, int bus, int slot, int func,
+				 int devid)
+{
+	struct pci_dev *dev;
+	int err = 0;
+
+	dev_dbg(&pdev->xdev->dev, "exporting dom %x bus %x slot %x func %x\n",
+		domain, bus, slot, func);
+
+	dev = pcistub_get_pci_dev_by_slot(pdev, domain, bus, slot, func);
+	if (!dev) {
+		err = -EINVAL;
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Couldn't locate PCI device "
+				 "(%04x:%02x:%02x.%01x)! "
+				 "perhaps already in-use?",
+				 domain, bus, slot, func);
+		goto out;
+	}
+
+	err = pciback_add_pci_dev(pdev, dev, devid, pciback_publish_pci_dev);
+	if (err)
+		goto out;
+
+	/* TODO: It'd be nice to export a bridge and have all of its children
+	 * get exported with it. This may be best done in xend (which will
+	 * have to calculate resource usage anyway) but we probably want to
+	 * put something in here to ensure that if a bridge gets given to a
+	 * driver domain, that all devices under that bridge are not given
+	 * to other driver domains (as he who controls the bridge can disable
+	 * it and stop the other devices from working).
+	 */
+      out:
+	return err;
+}
+
+static int pciback_remove_device(struct pciback_device *pdev,
+				 int domain, int bus, int slot, int func)
+{
+	int err = 0;
+	struct pci_dev *dev;
+
+	dev_dbg(&pdev->xdev->dev, "removing dom %x bus %x slot %x func %x\n",
+		domain, bus, slot, func);
+
+	dev = pciback_get_pci_dev(pdev, domain, bus, PCI_DEVFN(slot, func));
+	if (!dev) {
+		err = -EINVAL;
+		dev_dbg(&pdev->xdev->dev, "Couldn't locate PCI device "
+			"(%04x:%02x:%02x.%01x)! not owned by this domain\n",
+			domain, bus, slot, func);
+		goto out;
+	}
+
+	pciback_release_pci_dev(pdev, dev);
+	
+      out:
+	return err;
+}
+
+static int pciback_publish_pci_root(struct pciback_device *pdev,
+				    unsigned int domain, unsigned int bus)
+{
+	unsigned int d, b;
+	int i, root_num, len, err;
+	char str[64];
+
+	dev_dbg(&pdev->xdev->dev, "Publishing pci roots\n");
+
+	err = xenbus_scanf(XBT_NIL, pdev->xdev->nodename,
+			   "root_num", "%d", &root_num);
+	if (err == 0 || err == -ENOENT)
+		root_num = 0;
+	else if (err < 0)
+		goto out;
+
+	/* Verify that we haven't already published this pci root */
+	for (i = 0; i < root_num; i++) {
+		len = snprintf(str, sizeof(str), "root-%d", i);
+		if (unlikely(len >= (sizeof(str) - 1))) {
+			err = -ENOMEM;
+			goto out;
+		}
+
+		err = xenbus_scanf(XBT_NIL, pdev->xdev->nodename,
+				   str, "%x:%x", &d, &b);
+		if (err < 0)
+			goto out;
+		if (err != 2) {
+			err = -EINVAL;
+			goto out;
+		}
+
+		if (d == domain && b == bus) {
+			err = 0;
+			goto out;
+		}
+	}
+
+	len = snprintf(str, sizeof(str), "root-%d", root_num);
+	if (unlikely(len >= (sizeof(str) - 1))) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	dev_dbg(&pdev->xdev->dev, "writing root %d at %04x:%02x\n",
+		root_num, domain, bus);
+
+	err = xenbus_printf(XBT_NIL, pdev->xdev->nodename, str,
+			    "%04x:%02x", domain, bus);
+	if (err)
+		goto out;
+
+	err = xenbus_printf(XBT_NIL, pdev->xdev->nodename,
+			    "root_num", "%d", (root_num + 1));
+
+      out:
+	return err;
+}
+
+static int pciback_reconfigure(struct pciback_device *pdev)
+{
+	int err = 0;
+	int num_devs;
+	int domain, bus, slot, func;
+	int substate;
+	int i, len;
+	char state_str[64];
+	char dev_str[64];
+
+	spin_lock(&pdev->dev_lock);
+
+	dev_dbg(&pdev->xdev->dev, "Reconfiguring device ...\n");
+
+	/* Make sure we only reconfigure once */
+	if (xenbus_read_driver_state(pdev->xdev->nodename) !=
+	    XenbusStateReconfiguring)
+		goto out;
+
+	err = xenbus_scanf(XBT_NIL, pdev->xdev->nodename, "num_devs", "%d",
+			   &num_devs);
+	if (err != 1) {
+		if (err >= 0)
+			err = -EINVAL;
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Error reading number of devices");
+		goto out;
+	}
+
+	for (i = 0; i < num_devs; i++) {
+		len = snprintf(state_str, sizeof(state_str), "state-%d", i);
+		if (unlikely(len >= (sizeof(state_str) - 1))) {
+			err = -ENOMEM;
+			xenbus_dev_fatal(pdev->xdev, err,
+					 "String overflow while reading "
+					 "configuration");
+			goto out;
+		}
+		err = xenbus_scanf(XBT_NIL, pdev->xdev->nodename, state_str,
+				   "%d", &substate);
+		if (err != 1) 
+			substate = XenbusStateUnknown;
+
+		switch (substate) {
+		case XenbusStateInitialising:
+			dev_dbg(&pdev->xdev->dev, "Attaching dev-%d ...\n", i);
+
+			len = snprintf(dev_str, sizeof(dev_str), "dev-%d", i);
+			if (unlikely(len >= (sizeof(dev_str) - 1))) {
+				err = -ENOMEM;
+				xenbus_dev_fatal(pdev->xdev, err,
+						 "String overflow while "
+						 "reading configuration");
+				goto out;
+			}
+			err = xenbus_scanf(XBT_NIL, pdev->xdev->nodename,
+					   dev_str, "%x:%x:%x.%x",
+					   &domain, &bus, &slot, &func);
+			if (err < 0) {
+				xenbus_dev_fatal(pdev->xdev, err,
+						 "Error reading device "
+						 "configuration");
+				goto out;
+			}
+			if (err != 4) {
+				err = -EINVAL;
+				xenbus_dev_fatal(pdev->xdev, err,
+						 "Error parsing pci device "
+						 "configuration");
+				goto out;
+			}
+	
+			err = pciback_export_device(pdev, domain, bus, slot,
+						    func, i);
+			if (err)
+				goto out;
+
+			/* Publish pci roots. */
+			err = pciback_publish_pci_roots(pdev, pciback_publish_pci_root);
+			if (err) {
+				xenbus_dev_fatal(pdev->xdev, err,
+						 "Error while publish PCI root"
+						 "buses for frontend");
+				goto out;
+			}
+
+			err = xenbus_printf(XBT_NIL, pdev->xdev->nodename,
+					    state_str, "%d",
+					    XenbusStateInitialised);
+			if (err) {
+				xenbus_dev_fatal(pdev->xdev, err,
+						 "Error switching substate of "
+						 "dev-%d\n", i);
+				goto out;
+			}	
+			break;
+
+		case XenbusStateClosing:
+			dev_dbg(&pdev->xdev->dev, "Detaching dev-%d ...\n", i);
+
+			len = snprintf(dev_str, sizeof(dev_str), "vdev-%d", i);
+			if (unlikely(len >= (sizeof(dev_str) - 1))) {
+				err = -ENOMEM;
+				xenbus_dev_fatal(pdev->xdev, err,
+						 "String overflow while "
+						 "reading configuration");
+				goto out;
+			}
+			err = xenbus_scanf(XBT_NIL, pdev->xdev->nodename,
+					   dev_str, "%x:%x:%x.%x",
+					   &domain, &bus, &slot, &func);
+			if (err < 0) {
+				xenbus_dev_fatal(pdev->xdev, err,
+						 "Error reading device "
+						 "configuration");
+				goto out;
+			}
+			if (err != 4) {
+				err = -EINVAL;
+				xenbus_dev_fatal(pdev->xdev, err,
+						 "Error parsing pci device "
+						 "configuration");
+				goto out;
+			}
+
+			err = pciback_remove_device(pdev, domain, bus, slot,
+						    func);
+			if(err)
+				goto out;
+
+			/* TODO: If at some point we implement support for pci
+			 * root hot-remove on pcifront side, we'll need to
+			 * remove unnecessary xenstore nodes of pci roots here.
+			 */
+
+			break;
+
+		default:
+			break;
+		}
+	}
+
+	err = xenbus_switch_state(pdev->xdev, XenbusStateReconfigured);
+	if (err) {
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Error switching to reconfigured state!");
+		goto out;
+	}
+	
+      out:
+	spin_unlock(&pdev->dev_lock);
+
+	return 0;
+}
+
+static void pciback_frontend_changed(struct xenbus_device *xdev,
+				     enum xenbus_state fe_state)
+{
+	struct pciback_device *pdev = xdev->dev.driver_data;
+
+	dev_dbg(&xdev->dev, "fe state changed %d\n", fe_state);
+
+	switch (fe_state) {
+	case XenbusStateInitialised:
+		pciback_attach(pdev);
+		break;
+
+	case XenbusStateReconfiguring:
+		pciback_reconfigure(pdev);
+		break;
+
+	case XenbusStateConnected:
+		/* pcifront switched its state from reconfiguring to connected.
+		 * Then switch to connected state.
+		 */
+		xenbus_switch_state(xdev, XenbusStateConnected);
+		break;
+
+	case XenbusStateClosing:
+		pciback_disconnect(pdev);
+		xenbus_switch_state(xdev, XenbusStateClosing);
+		break;
+
+	case XenbusStateClosed:
+		pciback_disconnect(pdev);
+		xenbus_switch_state(xdev, XenbusStateClosed);
+		if (xenbus_dev_is_online(xdev))
+			break;
+		/* fall through if not online */
+	case XenbusStateUnknown:
+		dev_dbg(&xdev->dev, "frontend is gone! unregister device\n");
+		device_unregister(&xdev->dev);
+		break;
+
+	default:
+		break;
+	}
+}
+
+static int pciback_setup_backend(struct pciback_device *pdev)
+{
+	/* Get configuration from xend (if available now) */
+	int domain, bus, slot, func;
+	int err = 0;
+	int i, num_devs;
+	char dev_str[64];
+	char state_str[64];
+
+	spin_lock(&pdev->dev_lock);
+
+	/* It's possible we could get the call to setup twice, so make sure
+	 * we're not already connected.
+	 */
+	if (xenbus_read_driver_state(pdev->xdev->nodename) !=
+	    XenbusStateInitWait)
+		goto out;
+
+	dev_dbg(&pdev->xdev->dev, "getting be setup\n");
+
+	err = xenbus_scanf(XBT_NIL, pdev->xdev->nodename, "num_devs", "%d",
+			   &num_devs);
+	if (err != 1) {
+		if (err >= 0)
+			err = -EINVAL;
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Error reading number of devices");
+		goto out;
+	}
+
+	for (i = 0; i < num_devs; i++) {
+		int l = snprintf(dev_str, sizeof(dev_str), "dev-%d", i);
+		if (unlikely(l >= (sizeof(dev_str) - 1))) {
+			err = -ENOMEM;
+			xenbus_dev_fatal(pdev->xdev, err,
+					 "String overflow while reading "
+					 "configuration");
+			goto out;
+		}
+
+		err = xenbus_scanf(XBT_NIL, pdev->xdev->nodename, dev_str,
+				   "%x:%x:%x.%x", &domain, &bus, &slot, &func);
+		if (err < 0) {
+			xenbus_dev_fatal(pdev->xdev, err,
+					 "Error reading device configuration");
+			goto out;
+		}
+		if (err != 4) {
+			err = -EINVAL;
+			xenbus_dev_fatal(pdev->xdev, err,
+					 "Error parsing pci device "
+					 "configuration");
+			goto out;
+		}
+
+		err = pciback_export_device(pdev, domain, bus, slot, func, i);
+		if (err)
+			goto out;
+
+		/* Switch substate of this device. */
+		l = snprintf(state_str, sizeof(state_str), "state-%d", i);
+		if (unlikely(l >= (sizeof(state_str) - 1))) {
+			err = -ENOMEM;
+			xenbus_dev_fatal(pdev->xdev, err,
+					 "String overflow while reading "
+					 "configuration");
+			goto out;
+		}
+		err = xenbus_printf(XBT_NIL, pdev->xdev->nodename, state_str,
+				    "%d", XenbusStateInitialised);
+		if (err) {
+			xenbus_dev_fatal(pdev->xdev, err, "Error switching "
+					 "substate of dev-%d\n", i);
+			goto out;
+		}	
+	}
+
+	err = pciback_publish_pci_roots(pdev, pciback_publish_pci_root);
+	if (err) {
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Error while publish PCI root buses "
+				 "for frontend");
+		goto out;
+	}
+
+	err = xenbus_switch_state(pdev->xdev, XenbusStateInitialised);
+	if (err)
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Error switching to initialised state!");
+
+      out:
+	spin_unlock(&pdev->dev_lock);
+
+	if (!err)
+		/* see if pcifront is already configured (if not, we'll wait) */
+		pciback_attach(pdev);
+
+	return err;
+}
+
+static void pciback_be_watch(struct xenbus_watch *watch,
+			     const char **vec, unsigned int len)
+{
+	struct pciback_device *pdev =
+	    container_of(watch, struct pciback_device, be_watch);
+
+	switch (xenbus_read_driver_state(pdev->xdev->nodename)) {
+	case XenbusStateInitWait:
+		pciback_setup_backend(pdev);
+		break;
+
+	default:
+		break;
+	}
+}
+
+static int pciback_xenbus_probe(struct xenbus_device *dev,
+				const struct xenbus_device_id *id)
+{
+	int err = 0;
+	struct pciback_device *pdev = alloc_pdev(dev);
+
+	if (pdev == NULL) {
+		err = -ENOMEM;
+		xenbus_dev_fatal(dev, err,
+				 "Error allocating pciback_device struct");
+		goto out;
+	}
+
+	/* wait for xend to configure us */
+	err = xenbus_switch_state(dev, XenbusStateInitWait);
+	if (err)
+		goto out;
+
+	/* watch the backend node for backend configuration information */
+	err = xenbus_watch_path(dev, dev->nodename, &pdev->be_watch,
+				pciback_be_watch);
+	if (err)
+		goto out;
+	pdev->be_watching = 1;
+
+	/* We need to force a call to our callback here in case
+	 * xend already configured us!
+	 */
+	pciback_be_watch(&pdev->be_watch, NULL, 0);
+
+      out:
+	return err;
+}
+
+static int pciback_xenbus_remove(struct xenbus_device *dev)
+{
+	struct pciback_device *pdev = dev->dev.driver_data;
+
+	if (pdev != NULL)
+		free_pdev(pdev);
+
+	return 0;
+}
+
+static const struct xenbus_device_id xenpci_ids[] = {
+	{"pci"},
+	{{0}},
+};
+
+static struct xenbus_driver xenbus_pciback_driver = {
+	.name 			= "pciback",
+	.owner 			= THIS_MODULE,
+	.ids 			= xenpci_ids,
+	.probe 			= pciback_xenbus_probe,
+	.remove 		= pciback_xenbus_remove,
+	.otherend_changed 	= pciback_frontend_changed,
+};
+
+int __init pciback_xenbus_register(void)
+{
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	return xenbus_register_backend(&xenbus_pciback_driver);
+}
+
+void __exit pciback_xenbus_unregister(void)
+{
+	xenbus_unregister_driver(&xenbus_pciback_driver);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/pcifront/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/pcifront/Makefile
--- linux-2.6.18.8/drivers/xen/pcifront/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pcifront/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,7 @@
+obj-y += pcifront.o
+
+pcifront-y := pci_op.o xenbus.o pci.o
+
+ifeq ($(CONFIG_XEN_PCIDEV_FE_DEBUG),y)
+EXTRA_CFLAGS += -DDEBUG
+endif
diff -rpuN linux-2.6.18.8/drivers/xen/pcifront/pci.c linux-2.6.18-xen-3.3.0/drivers/xen/pcifront/pci.c
--- linux-2.6.18.8/drivers/xen/pcifront/pci.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pcifront/pci.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,46 @@
+/*
+ * PCI Frontend Operations - ensure only one PCI frontend runs at a time
+ *
+ *   Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/pci.h>
+#include <linux/spinlock.h>
+#include "pcifront.h"
+
+DEFINE_SPINLOCK(pcifront_dev_lock);
+static struct pcifront_device *pcifront_dev = NULL;
+
+int pcifront_connect(struct pcifront_device *pdev)
+{
+	int err = 0;
+
+	spin_lock(&pcifront_dev_lock);
+
+	if (!pcifront_dev) {
+		dev_info(&pdev->xdev->dev, "Installing PCI frontend\n");
+		pcifront_dev = pdev;
+	}
+	else {
+		dev_err(&pdev->xdev->dev, "PCI frontend already installed!\n");
+		err = -EEXIST;
+	}
+
+	spin_unlock(&pcifront_dev_lock);
+
+	return err;
+}
+
+void pcifront_disconnect(struct pcifront_device *pdev)
+{
+	spin_lock(&pcifront_dev_lock);
+
+	if (pdev == pcifront_dev) {
+		dev_info(&pdev->xdev->dev,
+			 "Disconnecting PCI Frontend Buses\n");
+		pcifront_dev = NULL;
+	}
+
+	spin_unlock(&pcifront_dev_lock);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/pcifront/pcifront.h linux-2.6.18-xen-3.3.0/drivers/xen/pcifront/pcifront.h
--- linux-2.6.18.8/drivers/xen/pcifront/pcifront.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pcifront/pcifront.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,42 @@
+/*
+ * PCI Frontend - Common data structures & function declarations
+ *
+ *   Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+#ifndef __XEN_PCIFRONT_H__
+#define __XEN_PCIFRONT_H__
+
+#include <linux/spinlock.h>
+#include <linux/pci.h>
+#include <xen/xenbus.h>
+#include <xen/interface/io/pciif.h>
+#include <xen/pcifront.h>
+
+struct pci_bus_entry {
+	struct list_head list;
+	struct pci_bus *bus;
+};
+
+struct pcifront_device {
+	struct xenbus_device *xdev;
+	struct list_head root_buses;
+	spinlock_t dev_lock;
+
+	int evtchn;
+	int gnt_ref;
+
+	/* Lock this when doing any operations in sh_info */
+	spinlock_t sh_info_lock;
+	struct xen_pci_sharedinfo *sh_info;
+};
+
+int pcifront_connect(struct pcifront_device *pdev);
+void pcifront_disconnect(struct pcifront_device *pdev);
+
+int pcifront_scan_root(struct pcifront_device *pdev,
+		       unsigned int domain, unsigned int bus);
+int pcifront_rescan_root(struct pcifront_device *pdev,
+			 unsigned int domain, unsigned int bus);
+void pcifront_free_roots(struct pcifront_device *pdev);
+
+#endif	/* __XEN_PCIFRONT_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/pcifront/pci_op.c linux-2.6.18-xen-3.3.0/drivers/xen/pcifront/pci_op.c
--- linux-2.6.18.8/drivers/xen/pcifront/pci_op.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pcifront/pci_op.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,551 @@
+/*
+ * PCI Frontend Operations - Communicates with frontend
+ *
+ *   Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+#include <linux/module.h>
+#include <linux/version.h>
+#include <linux/init.h>
+#include <linux/pci.h>
+#include <linux/spinlock.h>
+#include <linux/time.h>
+#include <xen/evtchn.h>
+#include "pcifront.h"
+
+static int verbose_request = 0;
+module_param(verbose_request, int, 0644);
+
+#ifdef __ia64__
+static void pcifront_init_sd(struct pcifront_sd *sd,
+			     unsigned int domain, unsigned int bus,
+			     struct pcifront_device *pdev)
+{
+	int err, i, j, k, len, root_num, res_count;
+	struct acpi_resource res;
+	unsigned int d, b, byte;
+	unsigned long magic;
+	char str[64], tmp[3];
+	unsigned char *buf, *bufp;
+	u8 *ptr;
+
+	memset(sd, 0, sizeof(*sd));
+
+	sd->segment = domain;
+	sd->node = -1;	/* Revisit for NUMA */
+	sd->platform_data = pdev;
+
+	/* Look for resources for this controller in xenbus. */
+	err = xenbus_scanf(XBT_NIL, pdev->xdev->otherend, "root_num",
+			   "%d", &root_num);
+	if (err != 1)
+		return;
+
+	for (i = 0; i < root_num; i++) {
+		len = snprintf(str, sizeof(str), "root-%d", i);
+		if (unlikely(len >= (sizeof(str) - 1)))
+			return;
+
+		err = xenbus_scanf(XBT_NIL, pdev->xdev->otherend,
+				   str, "%x:%x", &d, &b);
+		if (err != 2)
+			return;
+
+		if (d == domain && b == bus)
+			break;
+	}
+
+	if (i == root_num)
+		return;
+
+	len = snprintf(str, sizeof(str), "root-resource-magic");
+
+	err = xenbus_scanf(XBT_NIL, pdev->xdev->otherend,
+			   str, "%lx", &magic);
+
+	if (err != 1)
+		return; /* No resources, nothing to do */
+
+	if (magic != (sizeof(res) * 2) + 1) {
+		printk(KERN_WARNING "pcifront: resource magic mismatch\n");
+		return;
+	}
+
+	len = snprintf(str, sizeof(str), "root-%d-resources", i);
+	if (unlikely(len >= (sizeof(str) - 1)))
+		return;
+
+	err = xenbus_scanf(XBT_NIL, pdev->xdev->otherend,
+			   str, "%d", &res_count);
+
+	if (err != 1)
+		return; /* No resources, nothing to do */
+
+	sd->window = kzalloc(sizeof(*sd->window) * res_count, GFP_KERNEL);
+	if (!sd->window)
+		return;
+
+	/* magic is also the size of the byte stream in xenbus */
+	buf = kmalloc(magic, GFP_KERNEL);
+	if (!buf) {
+		kfree(sd->window);
+		sd->window = NULL;
+		return;
+	}
+
+	/* Read the resources out of xenbus */
+	for (j = 0; j < res_count; j++) {
+		memset(&res, 0, sizeof(res));
+		memset(buf, 0, magic);
+
+		len = snprintf(str, sizeof(str), "root-%d-resource-%d", i, j);
+		if (unlikely(len >= (sizeof(str) - 1)))
+			return;
+
+		err = xenbus_scanf(XBT_NIL, pdev->xdev->otherend, str,
+				   "%s", buf);
+		if (err != 1) {
+			printk(KERN_WARNING "pcifront: error reading "
+			       "resource %d on bus %04x:%02x\n",
+			       j, domain, bus);
+			continue;
+		}
+
+		bufp = buf;
+		ptr = (u8 *)&res;
+		memset(tmp, 0, sizeof(tmp));
+
+		/* Copy ASCII byte stream into structure */
+		for (k = 0; k < magic - 1; k += 2) {
+			memcpy(tmp, bufp, 2);
+			bufp += 2;
+
+			sscanf(tmp, "%02x", &byte);
+			*ptr = byte;
+			ptr++;
+		}
+
+		xen_add_resource(sd, domain, bus, &res);
+		sd->windows++;
+	}
+	kfree(buf);
+}
+#endif
+
+static int errno_to_pcibios_err(int errno)
+{
+	switch (errno) {
+	case XEN_PCI_ERR_success:
+		return PCIBIOS_SUCCESSFUL;
+
+	case XEN_PCI_ERR_dev_not_found:
+		return PCIBIOS_DEVICE_NOT_FOUND;
+
+	case XEN_PCI_ERR_invalid_offset:
+	case XEN_PCI_ERR_op_failed:
+		return PCIBIOS_BAD_REGISTER_NUMBER;
+
+	case XEN_PCI_ERR_not_implemented:
+		return PCIBIOS_FUNC_NOT_SUPPORTED;
+
+	case XEN_PCI_ERR_access_denied:
+		return PCIBIOS_SET_FAILED;
+	}
+	return errno;
+}
+
+static int do_pci_op(struct pcifront_device *pdev, struct xen_pci_op *op)
+{
+	int err = 0;
+	struct xen_pci_op *active_op = &pdev->sh_info->op;
+	unsigned long irq_flags;
+	evtchn_port_t port = pdev->evtchn;
+	s64 ns, ns_timeout;
+	struct timeval tv;
+
+	spin_lock_irqsave(&pdev->sh_info_lock, irq_flags);
+
+	memcpy(active_op, op, sizeof(struct xen_pci_op));
+
+	/* Go */
+	wmb();
+	set_bit(_XEN_PCIF_active, (unsigned long *)&pdev->sh_info->flags);
+	notify_remote_via_evtchn(port);
+
+	/*
+	 * We set a poll timeout of 3 seconds but give up on return after
+	 * 2 seconds. It is better to time out too late rather than too early
+	 * (in the latter case we end up continually re-executing poll() with a
+	 * timeout in the past). 1s difference gives plenty of slack for error.
+	 */
+	do_gettimeofday(&tv);
+	ns_timeout = timeval_to_ns(&tv) + 2 * (s64)NSEC_PER_SEC;
+
+	clear_evtchn(port);
+
+	while (test_bit(_XEN_PCIF_active,
+			(unsigned long *)&pdev->sh_info->flags)) {
+		if (HYPERVISOR_poll(&port, 1, jiffies + 3*HZ))
+			BUG();
+		clear_evtchn(port);
+		do_gettimeofday(&tv);
+		ns = timeval_to_ns(&tv);
+		if (ns > ns_timeout) {
+			dev_err(&pdev->xdev->dev,
+				"pciback not responding!!!\n");
+			clear_bit(_XEN_PCIF_active,
+				  (unsigned long *)&pdev->sh_info->flags);
+			err = XEN_PCI_ERR_dev_not_found;
+			goto out;
+		}
+	}
+
+	memcpy(op, active_op, sizeof(struct xen_pci_op));
+
+	err = op->err;
+      out:
+	spin_unlock_irqrestore(&pdev->sh_info_lock, irq_flags);
+	return err;
+}
+
+/* Access to this function is spinlocked in drivers/pci/access.c */
+static int pcifront_bus_read(struct pci_bus *bus, unsigned int devfn,
+			     int where, int size, u32 * val)
+{
+	int err = 0;
+	struct xen_pci_op op = {
+		.cmd    = XEN_PCI_OP_conf_read,
+		.domain = pci_domain_nr(bus),
+		.bus    = bus->number,
+		.devfn  = devfn,
+		.offset = where,
+		.size   = size,
+	};
+	struct pcifront_sd *sd = bus->sysdata;
+	struct pcifront_device *pdev = pcifront_get_pdev(sd);
+
+	if (verbose_request)
+		dev_info(&pdev->xdev->dev,
+			 "read dev=%04x:%02x:%02x.%01x - offset %x size %d\n",
+			 pci_domain_nr(bus), bus->number, PCI_SLOT(devfn),
+			 PCI_FUNC(devfn), where, size);
+
+	err = do_pci_op(pdev, &op);
+
+	if (likely(!err)) {
+		if (verbose_request)
+			dev_info(&pdev->xdev->dev, "read got back value %x\n",
+				 op.value);
+
+		*val = op.value;
+	} else if (err == -ENODEV) {
+		/* No device here, pretend that it just returned 0 */
+		err = 0;
+		*val = 0;
+	}
+
+	return errno_to_pcibios_err(err);
+}
+
+/* Access to this function is spinlocked in drivers/pci/access.c */
+static int pcifront_bus_write(struct pci_bus *bus, unsigned int devfn,
+			      int where, int size, u32 val)
+{
+	struct xen_pci_op op = {
+		.cmd    = XEN_PCI_OP_conf_write,
+		.domain = pci_domain_nr(bus),
+		.bus    = bus->number,
+		.devfn  = devfn,
+		.offset = where,
+		.size   = size,
+		.value  = val,
+	};
+	struct pcifront_sd *sd = bus->sysdata;
+	struct pcifront_device *pdev = pcifront_get_pdev(sd);
+
+	if (verbose_request)
+		dev_info(&pdev->xdev->dev,
+			 "write dev=%04x:%02x:%02x.%01x - "
+			 "offset %x size %d val %x\n",
+			 pci_domain_nr(bus), bus->number,
+			 PCI_SLOT(devfn), PCI_FUNC(devfn), where, size, val);
+
+	return errno_to_pcibios_err(do_pci_op(pdev, &op));
+}
+
+struct pci_ops pcifront_bus_ops = {
+	.read = pcifront_bus_read,
+	.write = pcifront_bus_write,
+};
+
+#ifdef CONFIG_PCI_MSI
+int pci_frontend_enable_msix(struct pci_dev *dev,
+		struct msix_entry *entries,
+		int nvec)
+{
+	int err;
+	int i;
+	struct xen_pci_op op = {
+		.cmd    = XEN_PCI_OP_enable_msix,
+		.domain = pci_domain_nr(dev->bus),
+		.bus = dev->bus->number,
+		.devfn = dev->devfn,
+		.value = nvec,
+	};
+	struct pcifront_sd *sd = dev->bus->sysdata;
+	struct pcifront_device *pdev = pcifront_get_pdev(sd);
+
+	if (nvec > SH_INFO_MAX_VEC) {
+		printk("too much vector for pci frontend%x\n", nvec);
+		return -EINVAL;
+	}
+
+	for (i = 0; i < nvec; i++) {
+		op.msix_entries[i].entry = entries[i].entry;
+		op.msix_entries[i].vector = entries[i].vector;
+	}
+
+	err = do_pci_op(pdev, &op);
+
+	if (!err) {
+		if (!op.value) {
+			/* we get the result */
+			for ( i = 0; i < nvec; i++)
+				entries[i].vector = op.msix_entries[i].vector;
+			return 0;
+		}
+		else {
+            printk("enable msix get value %x\n", op.value);
+			return op.value;
+		}
+	}
+	else {
+        printk("enable msix get err %x\n", err);
+		return err;
+	}
+}
+
+void pci_frontend_disable_msix(struct pci_dev* dev)
+{
+	int err;
+	struct xen_pci_op op = {
+		.cmd    = XEN_PCI_OP_disable_msix,
+		.domain = pci_domain_nr(dev->bus),
+		.bus = dev->bus->number,
+		.devfn = dev->devfn,
+	};
+	struct pcifront_sd *sd = dev->bus->sysdata;
+	struct pcifront_device *pdev = pcifront_get_pdev(sd);
+
+	err = do_pci_op(pdev, &op);
+
+	/* What should do for error ? */
+	if (err)
+		printk("pci_disable_msix get err %x\n", err);
+}
+
+int pci_frontend_enable_msi(struct pci_dev *dev)
+{
+	int err;
+	struct xen_pci_op op = {
+		.cmd    = XEN_PCI_OP_enable_msi,
+		.domain = pci_domain_nr(dev->bus),
+		.bus = dev->bus->number,
+		.devfn = dev->devfn,
+	};
+	struct pcifront_sd *sd = dev->bus->sysdata;
+	struct pcifront_device *pdev = pcifront_get_pdev(sd);
+
+	err = do_pci_op(pdev, &op);
+	if (likely(!err)) {
+		dev->irq = op.value;
+	}
+	else {
+		printk("pci frontend enable msi failed for dev %x:%x \n",
+				op.bus, op.devfn);
+		err = -EINVAL;
+	}
+	return err;
+}
+
+void pci_frontend_disable_msi(struct pci_dev* dev)
+{
+	int err;
+	struct xen_pci_op op = {
+		.cmd    = XEN_PCI_OP_disable_msi,
+		.domain = pci_domain_nr(dev->bus),
+		.bus = dev->bus->number,
+		.devfn = dev->devfn,
+	};
+	struct pcifront_sd *sd = dev->bus->sysdata;
+	struct pcifront_device *pdev = pcifront_get_pdev(sd);
+
+	err = do_pci_op(pdev, &op);
+	if (err == XEN_PCI_ERR_dev_not_found) {
+		/* XXX No response from backend, what shall we do? */
+		printk("get no response from backend for disable MSI\n");
+		return;
+	}
+	if (likely(!err))
+		dev->irq = op.value;
+	else
+		/* how can pciback notify us fail? */
+		printk("get fake response frombackend \n");
+}
+#endif /* CONFIG_PCI_MSI */
+
+/* Claim resources for the PCI frontend as-is, backend won't allow changes */
+static void pcifront_claim_resource(struct pci_dev *dev, void *data)
+{
+	struct pcifront_device *pdev = data;
+	int i;
+	struct resource *r;
+
+	for (i = 0; i < PCI_NUM_RESOURCES; i++) {
+		r = &dev->resource[i];
+
+		if (!r->parent && r->start && r->flags) {
+			dev_dbg(&pdev->xdev->dev, "claiming resource %s/%d\n",
+				pci_name(dev), i);
+			pci_claim_resource(dev, i);
+		}
+	}
+}
+
+int __devinit pcifront_scan_root(struct pcifront_device *pdev,
+				 unsigned int domain, unsigned int bus)
+{
+	struct pci_bus *b;
+	struct pcifront_sd *sd = NULL;
+	struct pci_bus_entry *bus_entry = NULL;
+	int err = 0;
+
+#ifndef CONFIG_PCI_DOMAINS
+	if (domain != 0) {
+		dev_err(&pdev->xdev->dev,
+			"PCI Root in non-zero PCI Domain! domain=%d\n", domain);
+		dev_err(&pdev->xdev->dev,
+			"Please compile with CONFIG_PCI_DOMAINS\n");
+		err = -EINVAL;
+		goto err_out;
+	}
+#endif
+
+	dev_info(&pdev->xdev->dev, "Creating PCI Frontend Bus %04x:%02x\n",
+		 domain, bus);
+
+	bus_entry = kmalloc(sizeof(*bus_entry), GFP_KERNEL);
+	sd = kmalloc(sizeof(*sd), GFP_KERNEL);
+	if (!bus_entry || !sd) {
+		err = -ENOMEM;
+		goto err_out;
+	}
+	pcifront_init_sd(sd, domain, bus, pdev);
+
+	b = pci_scan_bus_parented(&pdev->xdev->dev, bus,
+				  &pcifront_bus_ops, sd);
+	if (!b) {
+		dev_err(&pdev->xdev->dev,
+			"Error creating PCI Frontend Bus!\n");
+		err = -ENOMEM;
+		goto err_out;
+	}
+
+	pcifront_setup_root_resources(b, sd);
+	bus_entry->bus = b;
+
+	list_add(&bus_entry->list, &pdev->root_buses);
+
+	/* Claim resources before going "live" with our devices */
+	pci_walk_bus(b, pcifront_claim_resource, pdev);
+
+	pci_bus_add_devices(b);
+
+	return 0;
+
+      err_out:
+	kfree(bus_entry);
+	kfree(sd);
+
+	return err;
+}
+
+int __devinit pcifront_rescan_root(struct pcifront_device *pdev,
+				   unsigned int domain, unsigned int bus)
+{
+	struct pci_bus *b;
+	struct pci_dev *d;
+	unsigned int devfn;
+
+#ifndef CONFIG_PCI_DOMAINS
+	if (domain != 0) {
+		dev_err(&pdev->xdev->dev,
+			"PCI Root in non-zero PCI Domain! domain=%d\n", domain);
+		dev_err(&pdev->xdev->dev,
+			"Please compile with CONFIG_PCI_DOMAINS\n");
+		return -EINVAL;
+	}
+#endif
+
+	dev_info(&pdev->xdev->dev, "Rescanning PCI Frontend Bus %04x:%02x\n",
+		 domain, bus);
+
+	b = pci_find_bus(domain, bus);
+	if(!b)
+		/* If the bus is unknown, create it. */
+		return pcifront_scan_root(pdev, domain, bus);
+
+	/* Rescan the bus for newly attached functions and add.
+	 * We omit handling of PCI bridge attachment because pciback prevents
+	 * bridges from being exported.
+	 */ 
+	for (devfn = 0; devfn < 0x100; devfn++) {
+		d = pci_get_slot(b, devfn);
+		if(d) {
+			/* Device is already known. */
+			pci_dev_put(d);
+			continue;
+		}
+
+		d = pci_scan_single_device(b, devfn);
+		if (d) {
+			dev_info(&pdev->xdev->dev, "New device on "
+				 "%04x:%02x:%02x.%02x found.\n", domain, bus,
+				 PCI_SLOT(devfn), PCI_FUNC(devfn));
+			pci_bus_add_device(d);
+		}
+	}
+
+	return 0;
+}
+
+static void free_root_bus_devs(struct pci_bus *bus)
+{
+	struct pci_dev *dev;
+
+	while (!list_empty(&bus->devices)) {
+		dev = container_of(bus->devices.next, struct pci_dev,
+				   bus_list);
+		dev_dbg(&dev->dev, "removing device\n");
+		pci_remove_bus_device(dev);
+	}
+}
+
+void pcifront_free_roots(struct pcifront_device *pdev)
+{
+	struct pci_bus_entry *bus_entry, *t;
+
+	dev_dbg(&pdev->xdev->dev, "cleaning up root buses\n");
+
+	list_for_each_entry_safe(bus_entry, t, &pdev->root_buses, list) {
+		list_del(&bus_entry->list);
+
+		free_root_bus_devs(bus_entry->bus);
+
+		kfree(bus_entry->bus->sysdata);
+
+		device_unregister(bus_entry->bus->bridge);
+		pci_remove_bus(bus_entry->bus);
+
+		kfree(bus_entry);
+	}
+}
diff -rpuN linux-2.6.18.8/drivers/xen/pcifront/xenbus.c linux-2.6.18-xen-3.3.0/drivers/xen/pcifront/xenbus.c
--- linux-2.6.18.8/drivers/xen/pcifront/xenbus.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/pcifront/xenbus.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,455 @@
+/*
+ * PCI Frontend Xenbus Setup - handles setup with backend (imports page/evtchn)
+ *
+ *   Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/mm.h>
+#include <xen/xenbus.h>
+#include <xen/gnttab.h>
+#include "pcifront.h"
+
+#ifndef __init_refok
+#define __init_refok
+#endif
+
+#define INVALID_GRANT_REF (0)
+#define INVALID_EVTCHN    (-1)
+
+static struct pcifront_device *alloc_pdev(struct xenbus_device *xdev)
+{
+	struct pcifront_device *pdev;
+
+	pdev = kzalloc(sizeof(struct pcifront_device), GFP_KERNEL);
+	if (pdev == NULL)
+		goto out;
+
+	pdev->sh_info =
+	    (struct xen_pci_sharedinfo *)__get_free_page(GFP_KERNEL);
+	if (pdev->sh_info == NULL) {
+		kfree(pdev);
+		pdev = NULL;
+		goto out;
+	}
+	pdev->sh_info->flags = 0;
+
+	xdev->dev.driver_data = pdev;
+	pdev->xdev = xdev;
+
+	INIT_LIST_HEAD(&pdev->root_buses);
+
+	spin_lock_init(&pdev->dev_lock);
+	spin_lock_init(&pdev->sh_info_lock);
+
+	pdev->evtchn = INVALID_EVTCHN;
+	pdev->gnt_ref = INVALID_GRANT_REF;
+
+	dev_dbg(&xdev->dev, "Allocated pdev @ 0x%p pdev->sh_info @ 0x%p\n",
+		pdev, pdev->sh_info);
+      out:
+	return pdev;
+}
+
+static void free_pdev(struct pcifront_device *pdev)
+{
+	dev_dbg(&pdev->xdev->dev, "freeing pdev @ 0x%p\n", pdev);
+
+	pcifront_free_roots(pdev);
+
+	if (pdev->evtchn != INVALID_EVTCHN)
+		xenbus_free_evtchn(pdev->xdev, pdev->evtchn);
+
+	if (pdev->gnt_ref != INVALID_GRANT_REF)
+		gnttab_end_foreign_access(pdev->gnt_ref,
+					  (unsigned long)pdev->sh_info);
+
+	pdev->xdev->dev.driver_data = NULL;
+
+	kfree(pdev);
+}
+
+static int pcifront_publish_info(struct pcifront_device *pdev)
+{
+	int err = 0;
+	struct xenbus_transaction trans;
+
+	err = xenbus_grant_ring(pdev->xdev, virt_to_mfn(pdev->sh_info));
+	if (err < 0)
+		goto out;
+
+	pdev->gnt_ref = err;
+
+	err = xenbus_alloc_evtchn(pdev->xdev, &pdev->evtchn);
+	if (err)
+		goto out;
+
+      do_publish:
+	err = xenbus_transaction_start(&trans);
+	if (err) {
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Error writing configuration for backend "
+				 "(start transaction)");
+		goto out;
+	}
+
+	err = xenbus_printf(trans, pdev->xdev->nodename,
+			    "pci-op-ref", "%u", pdev->gnt_ref);
+	if (!err)
+		err = xenbus_printf(trans, pdev->xdev->nodename,
+				    "event-channel", "%u", pdev->evtchn);
+	if (!err)
+		err = xenbus_printf(trans, pdev->xdev->nodename,
+				    "magic", XEN_PCI_MAGIC);
+
+	if (err) {
+		xenbus_transaction_end(trans, 1);
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Error writing configuration for backend");
+		goto out;
+	} else {
+		err = xenbus_transaction_end(trans, 0);
+		if (err == -EAGAIN)
+			goto do_publish;
+		else if (err) {
+			xenbus_dev_fatal(pdev->xdev, err,
+					 "Error completing transaction "
+					 "for backend");
+			goto out;
+		}
+	}
+
+	xenbus_switch_state(pdev->xdev, XenbusStateInitialised);
+
+	dev_dbg(&pdev->xdev->dev, "publishing successful!\n");
+
+      out:
+	return err;
+}
+
+static int __devinit pcifront_try_connect(struct pcifront_device *pdev)
+{
+	int err = -EFAULT;
+	int i, num_roots, len;
+	char str[64];
+	unsigned int domain, bus;
+
+	spin_lock(&pdev->dev_lock);
+
+	/* Only connect once */
+	if (xenbus_read_driver_state(pdev->xdev->nodename) !=
+	    XenbusStateInitialised)
+		goto out;
+
+	err = pcifront_connect(pdev);
+	if (err) {
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Error connecting PCI Frontend");
+		goto out;
+	}
+
+	err = xenbus_scanf(XBT_NIL, pdev->xdev->otherend,
+			   "root_num", "%d", &num_roots);
+	if (err == -ENOENT) {
+		xenbus_dev_error(pdev->xdev, err,
+				 "No PCI Roots found, trying 0000:00");
+		err = pcifront_scan_root(pdev, 0, 0);
+		num_roots = 0;
+	} else if (err != 1) {
+		if (err == 0)
+			err = -EINVAL;
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Error reading number of PCI roots");
+		goto out;
+	}
+
+	for (i = 0; i < num_roots; i++) {
+		len = snprintf(str, sizeof(str), "root-%d", i);
+		if (unlikely(len >= (sizeof(str) - 1))) {
+			err = -ENOMEM;
+			goto out;
+		}
+
+		err = xenbus_scanf(XBT_NIL, pdev->xdev->otherend, str,
+				   "%x:%x", &domain, &bus);
+		if (err != 2) {
+			if (err >= 0)
+				err = -EINVAL;
+			xenbus_dev_fatal(pdev->xdev, err,
+					 "Error reading PCI root %d", i);
+			goto out;
+		}
+
+		err = pcifront_scan_root(pdev, domain, bus);
+		if (err) {
+			xenbus_dev_fatal(pdev->xdev, err,
+					 "Error scanning PCI root %04x:%02x",
+					 domain, bus);
+			goto out;
+		}
+	}
+
+	err = xenbus_switch_state(pdev->xdev, XenbusStateConnected);
+	if (err)
+		goto out;
+
+      out:
+	spin_unlock(&pdev->dev_lock);
+	return err;
+}
+
+static int pcifront_try_disconnect(struct pcifront_device *pdev)
+{
+	int err = 0;
+	enum xenbus_state prev_state;
+
+	spin_lock(&pdev->dev_lock);
+
+	prev_state = xenbus_read_driver_state(pdev->xdev->nodename);
+
+	if (prev_state >= XenbusStateClosing)
+		goto out;
+
+	if(prev_state == XenbusStateConnected) {
+		pcifront_free_roots(pdev);
+		pcifront_disconnect(pdev);
+	}
+
+	err = xenbus_switch_state(pdev->xdev, XenbusStateClosed);
+
+      out:
+	spin_unlock(&pdev->dev_lock);
+
+	return err;
+}
+
+static int __devinit pcifront_attach_devices(struct pcifront_device *pdev)
+{
+	int err = -EFAULT;
+	int i, num_roots, len;
+	unsigned int domain, bus;
+	char str[64];
+
+	spin_lock(&pdev->dev_lock);
+
+	if (xenbus_read_driver_state(pdev->xdev->nodename) !=
+	    XenbusStateReconfiguring)
+		goto out;
+
+	err = xenbus_scanf(XBT_NIL, pdev->xdev->otherend,
+			   "root_num", "%d", &num_roots);
+	if (err == -ENOENT) {
+		xenbus_dev_error(pdev->xdev, err,
+				 "No PCI Roots found, trying 0000:00");
+		err = pcifront_rescan_root(pdev, 0, 0);
+		num_roots = 0;
+	} else if (err != 1) {
+		if (err == 0)
+			err = -EINVAL;
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Error reading number of PCI roots");
+		goto out;
+	}
+
+	for (i = 0; i < num_roots; i++) {
+		len = snprintf(str, sizeof(str), "root-%d", i);
+		if (unlikely(len >= (sizeof(str) - 1))) {
+			err = -ENOMEM;
+			goto out;
+		}
+
+		err = xenbus_scanf(XBT_NIL, pdev->xdev->otherend, str,
+				   "%x:%x", &domain, &bus);
+		if (err != 2) {
+			if (err >= 0)
+				err = -EINVAL;
+			xenbus_dev_fatal(pdev->xdev, err,
+					 "Error reading PCI root %d", i);
+			goto out;
+		}
+
+		err = pcifront_rescan_root(pdev, domain, bus);
+		if (err) {
+			xenbus_dev_fatal(pdev->xdev, err,
+					 "Error scanning PCI root %04x:%02x",
+					 domain, bus);
+			goto out;
+		}
+	}
+
+	xenbus_switch_state(pdev->xdev, XenbusStateConnected);
+
+      out:
+	spin_unlock(&pdev->dev_lock);
+	return err;
+}
+
+static int pcifront_detach_devices(struct pcifront_device *pdev)
+{
+	int err = 0;
+	int i, num_devs;
+	unsigned int domain, bus, slot, func;
+	struct pci_bus *pci_bus;
+	struct pci_dev *pci_dev;
+	char str[64];
+
+	spin_lock(&pdev->dev_lock);
+
+	if (xenbus_read_driver_state(pdev->xdev->nodename) !=
+	    XenbusStateConnected)
+		goto out;
+
+	err = xenbus_scanf(XBT_NIL, pdev->xdev->otherend, "num_devs", "%d",
+			   &num_devs);
+	if (err != 1) {
+		if (err >= 0)
+			err = -EINVAL;
+		xenbus_dev_fatal(pdev->xdev, err,
+				 "Error reading number of PCI devices");
+		goto out;
+	}
+
+	/* Find devices being detached and remove them. */
+	for (i = 0; i < num_devs; i++) {
+		int l, state;
+		l = snprintf(str, sizeof(str), "state-%d", i);
+		if (unlikely(l >= (sizeof(str) - 1))) {
+			err = -ENOMEM;
+			goto out;
+		}
+		err = xenbus_scanf(XBT_NIL, pdev->xdev->otherend, str, "%d",
+				   &state);
+		if (err != 1)
+			state = XenbusStateUnknown;
+
+		if (state != XenbusStateClosing)
+			continue;
+
+		/* Remove device. */
+		l = snprintf(str, sizeof(str), "vdev-%d", i);
+		if (unlikely(l >= (sizeof(str) - 1))) {
+			err = -ENOMEM;
+			goto out;
+		}
+		err = xenbus_scanf(XBT_NIL, pdev->xdev->otherend, str,
+			   	   "%x:%x:%x.%x", &domain, &bus, &slot, &func);
+		if (err != 4) {
+			if (err >= 0)
+				err = -EINVAL;
+			xenbus_dev_fatal(pdev->xdev, err,
+				 	 "Error reading PCI device %d", i);
+			goto out;
+		}
+
+		pci_bus = pci_find_bus(domain, bus);
+		if(!pci_bus) {
+			dev_dbg(&pdev->xdev->dev, "Cannot get bus %04x:%02x\n",
+				domain, bus);
+			continue;
+		}
+		pci_dev = pci_get_slot(pci_bus, PCI_DEVFN(slot, func));
+		if(!pci_dev) {
+			dev_dbg(&pdev->xdev->dev,
+				"Cannot get PCI device %04x:%02x:%02x.%02x\n",
+				domain, bus, slot, func);
+			continue;
+		}
+		pci_remove_bus_device(pci_dev);
+		pci_dev_put(pci_dev);
+
+		dev_dbg(&pdev->xdev->dev,
+			"PCI device %04x:%02x:%02x.%02x removed.\n",
+			domain, bus, slot, func);
+	}
+
+	err = xenbus_switch_state(pdev->xdev, XenbusStateReconfiguring);
+
+      out:
+	spin_unlock(&pdev->dev_lock);
+	return err;
+}
+
+static void __init_refok pcifront_backend_changed(struct xenbus_device *xdev,
+						  enum xenbus_state be_state)
+{
+	struct pcifront_device *pdev = xdev->dev.driver_data;
+
+	switch (be_state) {
+	case XenbusStateUnknown:
+	case XenbusStateInitialising:
+	case XenbusStateInitWait:
+	case XenbusStateInitialised:
+	case XenbusStateClosed:
+		break;
+
+	case XenbusStateConnected:
+		pcifront_try_connect(pdev);
+		break;
+
+	case XenbusStateClosing:
+		dev_warn(&xdev->dev, "backend going away!\n");
+		pcifront_try_disconnect(pdev);
+		break;
+
+	case XenbusStateReconfiguring:
+		pcifront_detach_devices(pdev);
+		break;
+
+	case XenbusStateReconfigured:
+		pcifront_attach_devices(pdev);
+		break;
+	}
+}
+
+static int pcifront_xenbus_probe(struct xenbus_device *xdev,
+				 const struct xenbus_device_id *id)
+{
+	int err = 0;
+	struct pcifront_device *pdev = alloc_pdev(xdev);
+
+	if (pdev == NULL) {
+		err = -ENOMEM;
+		xenbus_dev_fatal(xdev, err,
+				 "Error allocating pcifront_device struct");
+		goto out;
+	}
+
+	err = pcifront_publish_info(pdev);
+
+      out:
+	return err;
+}
+
+static int pcifront_xenbus_remove(struct xenbus_device *xdev)
+{
+	if (xdev->dev.driver_data)
+		free_pdev(xdev->dev.driver_data);
+
+	return 0;
+}
+
+static const struct xenbus_device_id xenpci_ids[] = {
+	{"pci"},
+	{{0}},
+};
+MODULE_ALIAS("xen:pci");
+
+static struct xenbus_driver xenbus_pcifront_driver = {
+	.name 			= "pcifront",
+	.owner 			= THIS_MODULE,
+	.ids 			= xenpci_ids,
+	.probe 			= pcifront_xenbus_probe,
+	.remove 		= pcifront_xenbus_remove,
+	.otherend_changed 	= pcifront_backend_changed,
+};
+
+static int __init pcifront_init(void)
+{
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	return xenbus_register_frontend(&xenbus_pcifront_driver);
+}
+
+/* Initialize after the Xen PCI Frontend Stub is initialized */
+subsys_initcall(pcifront_init);
diff -rpuN linux-2.6.18.8/drivers/xen/privcmd/compat_privcmd.c linux-2.6.18-xen-3.3.0/drivers/xen/privcmd/compat_privcmd.c
--- linux-2.6.18.8/drivers/xen/privcmd/compat_privcmd.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/privcmd/compat_privcmd.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,73 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ *
+ * Copyright (C) IBM Corp. 2006
+ *
+ * Authors: Jimi Xenidis <jimix@watson.ibm.com>
+ */
+
+#include <linux/config.h>
+#include <linux/compat.h>
+#include <linux/ioctl.h>
+#include <linux/syscalls.h>
+#include <asm/hypervisor.h>
+#include <asm/uaccess.h>
+#include <xen/public/privcmd.h>
+#include <xen/compat_ioctl.h>
+
+int privcmd_ioctl_32(int fd, unsigned int cmd, unsigned long arg)
+{
+	int ret;
+
+	switch (cmd) {
+	case IOCTL_PRIVCMD_MMAP_32: {
+		struct privcmd_mmap *p;
+		struct privcmd_mmap_32 *p32;
+		struct privcmd_mmap_32 n32;
+
+		p32 = compat_ptr(arg);
+		p = compat_alloc_user_space(sizeof(*p));
+		if (copy_from_user(&n32, p32, sizeof(n32)) ||
+		    put_user(n32.num, &p->num) ||
+		    put_user(n32.dom, &p->dom) ||
+		    put_user(compat_ptr(n32.entry), &p->entry))
+			return -EFAULT;
+		
+		ret = sys_ioctl(fd, IOCTL_PRIVCMD_MMAP, (unsigned long)p);
+	}
+		break;
+	case IOCTL_PRIVCMD_MMAPBATCH_32: {
+		struct privcmd_mmapbatch *p;
+		struct privcmd_mmapbatch_32 *p32;
+		struct privcmd_mmapbatch_32 n32;
+
+		p32 = compat_ptr(arg);
+		p = compat_alloc_user_space(sizeof(*p));
+		if (copy_from_user(&n32, p32, sizeof(n32)) ||
+		    put_user(n32.num, &p->num) ||
+		    put_user(n32.dom, &p->dom) ||
+		    put_user(n32.addr, &p->addr) ||
+		    put_user(compat_ptr(n32.arr), &p->arr))
+			return -EFAULT;
+		
+		ret = sys_ioctl(fd, IOCTL_PRIVCMD_MMAPBATCH, (unsigned long)p);
+	}
+		break;
+	default:
+		ret = -EINVAL;
+		break;
+	}
+	return ret;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/privcmd/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/privcmd/Makefile
--- linux-2.6.18.8/drivers/xen/privcmd/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/privcmd/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,3 @@
+
+obj-y	+= privcmd.o
+obj-$(CONFIG_COMPAT)	+= compat_privcmd.o
diff -rpuN linux-2.6.18.8/drivers/xen/privcmd/privcmd.c linux-2.6.18-xen-3.3.0/drivers/xen/privcmd/privcmd.c
--- linux-2.6.18.8/drivers/xen/privcmd/privcmd.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/privcmd/privcmd.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,356 @@
+/******************************************************************************
+ * privcmd.c
+ * 
+ * Interface to privileged domain-0 commands.
+ * 
+ * Copyright (c) 2002-2004, K A Fraser, B Dragovic
+ */
+
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/string.h>
+#include <linux/errno.h>
+#include <linux/mm.h>
+#include <linux/mman.h>
+#include <linux/swap.h>
+#include <linux/smp_lock.h>
+#include <linux/highmem.h>
+#include <linux/pagemap.h>
+#include <linux/seq_file.h>
+#include <asm/hypervisor.h>
+
+#include <asm/pgalloc.h>
+#include <asm/pgtable.h>
+#include <asm/uaccess.h>
+#include <asm/tlb.h>
+#include <asm/hypervisor.h>
+#include <xen/public/privcmd.h>
+#include <xen/interface/xen.h>
+#include <xen/xen_proc.h>
+#include <xen/features.h>
+
+static struct proc_dir_entry *privcmd_intf;
+static struct proc_dir_entry *capabilities_intf;
+
+#ifndef HAVE_ARCH_PRIVCMD_MMAP
+static int privcmd_enforce_singleshot_mapping(struct vm_area_struct *vma);
+#endif
+
+static long privcmd_ioctl(struct file *file,
+			  unsigned int cmd, unsigned long data)
+{
+	int ret = -ENOSYS;
+	void __user *udata = (void __user *) data;
+
+	switch (cmd) {
+	case IOCTL_PRIVCMD_HYPERCALL: {
+		privcmd_hypercall_t hypercall;
+  
+		if (copy_from_user(&hypercall, udata, sizeof(hypercall)))
+			return -EFAULT;
+
+#if defined(__i386__)
+		if (hypercall.op >= (PAGE_SIZE >> 5))
+			break;
+		__asm__ __volatile__ (
+			"pushl %%ebx; pushl %%ecx; pushl %%edx; "
+			"pushl %%esi; pushl %%edi; "
+			"movl  8(%%eax),%%ebx ;"
+			"movl 16(%%eax),%%ecx ;"
+			"movl 24(%%eax),%%edx ;"
+			"movl 32(%%eax),%%esi ;"
+			"movl 40(%%eax),%%edi ;"
+			"movl   (%%eax),%%eax ;"
+			"shll $5,%%eax ;"
+			"addl $hypercall_page,%%eax ;"
+			"call *%%eax ;"
+			"popl %%edi; popl %%esi; popl %%edx; "
+			"popl %%ecx; popl %%ebx"
+			: "=a" (ret) : "0" (&hypercall) : "memory" );
+#elif defined (__x86_64__)
+		if (hypercall.op < (PAGE_SIZE >> 5)) {
+			long ign1, ign2, ign3;
+			__asm__ __volatile__ (
+				"movq %8,%%r10; movq %9,%%r8;"
+				"shll $5,%%eax ;"
+				"addq $hypercall_page,%%rax ;"
+				"call *%%rax"
+				: "=a" (ret), "=D" (ign1),
+				  "=S" (ign2), "=d" (ign3)
+				: "0" ((unsigned int)hypercall.op),
+				"1" (hypercall.arg[0]),
+				"2" (hypercall.arg[1]),
+				"3" (hypercall.arg[2]),
+				"g" (hypercall.arg[3]),
+				"g" (hypercall.arg[4])
+				: "r8", "r10", "memory" );
+		}
+#else
+		ret = privcmd_hypercall(&hypercall);
+#endif
+	}
+	break;
+
+	case IOCTL_PRIVCMD_MMAP: {
+#define MMAP_NR_PER_PAGE (int)((PAGE_SIZE-sizeof(struct list_head))/sizeof(privcmd_mmap_entry_t))
+		privcmd_mmap_t mmapcmd;
+		privcmd_mmap_entry_t *msg;
+		privcmd_mmap_entry_t __user *p;
+		struct mm_struct *mm = current->mm;
+		struct vm_area_struct *vma;
+		unsigned long va;
+		int i, rc;
+		LIST_HEAD(pagelist);
+		struct list_head *l,*l2;
+
+		if (!is_initial_xendomain())
+			return -EPERM;
+
+		if (copy_from_user(&mmapcmd, udata, sizeof(mmapcmd)))
+			return -EFAULT;
+
+		p = mmapcmd.entry;
+		for (i = 0; i < mmapcmd.num;) {
+			int nr = min(mmapcmd.num - i, MMAP_NR_PER_PAGE);
+
+			rc = -ENOMEM;
+			l = (struct list_head *) __get_free_page(GFP_KERNEL);
+			if (l == NULL)
+				goto mmap_out;
+
+			INIT_LIST_HEAD(l);
+			list_add_tail(l, &pagelist);
+			msg = (privcmd_mmap_entry_t*)(l + 1);
+
+			rc = -EFAULT;
+			if (copy_from_user(msg, p, nr*sizeof(*msg)))
+				goto mmap_out;
+			i += nr;
+			p += nr;
+		}
+
+		l = pagelist.next;
+		msg = (privcmd_mmap_entry_t*)(l + 1);
+
+		down_write(&mm->mmap_sem);
+
+		vma = find_vma(mm, msg->va);
+		rc = -EINVAL;
+		if (!vma || (msg->va != vma->vm_start) ||
+		    !privcmd_enforce_singleshot_mapping(vma))
+			goto mmap_out;
+
+		va = vma->vm_start;
+
+		i = 0;
+		list_for_each(l, &pagelist) {
+			int nr = i + min(mmapcmd.num - i, MMAP_NR_PER_PAGE);
+
+			msg = (privcmd_mmap_entry_t*)(l + 1);
+			while (i<nr) {
+
+				/* Do not allow range to wrap the address space. */
+				rc = -EINVAL;
+				if ((msg->npages > (LONG_MAX >> PAGE_SHIFT)) ||
+				    ((unsigned long)(msg->npages << PAGE_SHIFT) >= -va))
+					goto mmap_out;
+
+				/* Range chunks must be contiguous in va space. */
+				if ((msg->va != va) ||
+				    ((msg->va+(msg->npages<<PAGE_SHIFT)) > vma->vm_end))
+					goto mmap_out;
+
+				if ((rc = direct_remap_pfn_range(
+					     vma,
+					     msg->va & PAGE_MASK,
+					     msg->mfn,
+					     msg->npages << PAGE_SHIFT,
+					     vma->vm_page_prot,
+					     mmapcmd.dom)) < 0)
+					goto mmap_out;
+
+				va += msg->npages << PAGE_SHIFT;
+				msg++;
+				i++;
+			}
+		}
+
+		rc = 0;
+
+	mmap_out:
+		up_write(&mm->mmap_sem);
+		list_for_each_safe(l,l2,&pagelist)
+			free_page((unsigned long)l);
+		ret = rc;
+	}
+#undef MMAP_NR_PER_PAGE
+	break;
+
+	case IOCTL_PRIVCMD_MMAPBATCH: {
+#define MMAPBATCH_NR_PER_PAGE (unsigned long)((PAGE_SIZE-sizeof(struct list_head))/sizeof(unsigned long))
+		privcmd_mmapbatch_t m;
+		struct mm_struct *mm = current->mm;
+		struct vm_area_struct *vma;
+		xen_pfn_t __user *p;
+		unsigned long addr, *mfn, nr_pages;
+		int i;
+		LIST_HEAD(pagelist);
+		struct list_head *l, *l2;
+
+		if (!is_initial_xendomain())
+			return -EPERM;
+
+		if (copy_from_user(&m, udata, sizeof(m)))
+			return -EFAULT;
+
+		nr_pages = m.num;
+		if ((m.num <= 0) || (nr_pages > (LONG_MAX >> PAGE_SHIFT)))
+			return -EINVAL;
+
+		p = m.arr;
+		for (i=0; i<nr_pages; )	{
+			int nr = min(nr_pages - i, MMAPBATCH_NR_PER_PAGE);
+
+			ret = -ENOMEM;
+			l = (struct list_head *)__get_free_page(GFP_KERNEL);
+			if (l == NULL)
+				goto mmapbatch_out;
+
+			INIT_LIST_HEAD(l);
+			list_add_tail(l, &pagelist);
+
+			mfn = (unsigned long*)(l + 1);
+			ret = -EFAULT;
+			if (copy_from_user(mfn, p, nr*sizeof(*mfn)))
+				goto mmapbatch_out;
+
+			i += nr; p+= nr;
+		}
+
+		down_write(&mm->mmap_sem);
+
+		vma = find_vma(mm, m.addr);
+		ret = -EINVAL;
+		if (!vma ||
+		    (m.addr != vma->vm_start) ||
+		    ((m.addr + (nr_pages << PAGE_SHIFT)) != vma->vm_end) ||
+		    !privcmd_enforce_singleshot_mapping(vma)) {
+			up_write(&mm->mmap_sem);
+			goto mmapbatch_out;
+		}
+
+		p = m.arr;
+		addr = m.addr;
+		i = 0;
+		ret = 0;
+		list_for_each(l, &pagelist) {
+			int nr = i + min(nr_pages - i, MMAPBATCH_NR_PER_PAGE);
+			mfn = (unsigned long *)(l + 1);
+
+			while (i<nr) {
+				if(direct_remap_pfn_range(vma, addr & PAGE_MASK,
+							  *mfn, PAGE_SIZE,
+							  vma->vm_page_prot, m.dom) < 0) {
+					*mfn |= 0xf0000000U;
+					ret++;
+				}
+				mfn++; i++; addr += PAGE_SIZE;
+			}
+		}
+
+		up_write(&mm->mmap_sem);
+		if (ret > 0) {
+			p = m.arr;
+			i = 0;
+			ret = 0;
+			list_for_each(l, &pagelist) {
+				int nr = min(nr_pages - i, MMAPBATCH_NR_PER_PAGE);
+				mfn = (unsigned long *)(l + 1);
+				if (copy_to_user(p, mfn, nr*sizeof(*mfn)))
+					ret = -EFAULT;
+				i += nr; p += nr;
+			}
+		}
+	mmapbatch_out:
+		list_for_each_safe(l,l2,&pagelist)
+			free_page((unsigned long)l);
+#undef MMAPBATCH_NR_PER_PAGE
+	}
+	break;
+
+	default:
+		ret = -EINVAL;
+		break;
+	}
+
+	return ret;
+}
+
+#ifndef HAVE_ARCH_PRIVCMD_MMAP
+static struct page *privcmd_nopage(struct vm_area_struct *vma,
+				   unsigned long address,
+				   int *type)
+{
+	return NOPAGE_SIGBUS;
+}
+
+static struct vm_operations_struct privcmd_vm_ops = {
+	.nopage = privcmd_nopage
+};
+
+static int privcmd_mmap(struct file * file, struct vm_area_struct * vma)
+{
+	/* Unsupported for auto-translate guests. */
+	if (xen_feature(XENFEAT_auto_translated_physmap))
+		return -ENOSYS;
+
+	/* DONTCOPY is essential for Xen as copy_page_range is broken. */
+	vma->vm_flags |= VM_RESERVED | VM_IO | VM_DONTCOPY;
+	vma->vm_ops = &privcmd_vm_ops;
+	vma->vm_private_data = NULL;
+
+	return 0;
+}
+
+static int privcmd_enforce_singleshot_mapping(struct vm_area_struct *vma)
+{
+	return (xchg(&vma->vm_private_data, (void *)1) == NULL);
+}
+#endif
+
+static const struct file_operations privcmd_file_ops = {
+	.unlocked_ioctl = privcmd_ioctl,
+	.mmap = privcmd_mmap,
+};
+
+static int capabilities_read(char *page, char **start, off_t off,
+			     int count, int *eof, void *data)
+{
+	int len = 0;
+	*page = 0;
+
+	if (is_initial_xendomain())
+		len = sprintf( page, "control_d\n" );
+
+	*eof = 1;
+	return len;
+}
+
+static int __init privcmd_init(void)
+{
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	privcmd_intf = create_xen_proc_entry("privcmd", 0400);
+	if (privcmd_intf != NULL)
+		privcmd_intf->proc_fops = &privcmd_file_ops;
+
+	capabilities_intf = create_xen_proc_entry("capabilities", 0400 );
+	if (capabilities_intf != NULL)
+		capabilities_intf->read_proc = capabilities_read;
+
+	return 0;
+}
+
+__initcall(privcmd_init);
diff -rpuN linux-2.6.18.8/drivers/xen/scsiback/common.h linux-2.6.18-xen-3.3.0/drivers/xen/scsiback/common.h
--- linux-2.6.18.8/drivers/xen/scsiback/common.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/scsiback/common.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,181 @@
+/*
+ * Copyright (c) 2008, FUJITSU Limited
+ *
+ * Based on the blkback driver code.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __SCSIIF__BACKEND__COMMON_H__
+#define __SCSIIF__BACKEND__COMMON_H__
+
+#include <linux/version.h>
+#include <linux/module.h>
+#include <linux/interrupt.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <linux/wait.h>
+#include <linux/sched.h>
+#include <linux/kthread.h>
+#include <linux/blkdev.h>
+#include <linux/list.h>
+#include <linux/kthread.h>
+#include <scsi/scsi.h>
+#include <scsi/scsi_cmnd.h>
+#include <scsi/scsi_host.h>
+#include <scsi/scsi_device.h>
+#include <scsi/scsi_dbg.h>
+#include <scsi/scsi_eh.h>
+#include <asm/io.h>
+#include <asm/setup.h>
+#include <asm/pgalloc.h>
+#include <asm/delay.h>
+#include <xen/evtchn.h>
+#include <asm/hypervisor.h>
+#include <xen/gnttab.h>
+#include <xen/driver_util.h>
+#include <xen/xenbus.h>
+#include <xen/interface/io/ring.h>
+#include <xen/interface/grant_table.h>
+#include <xen/interface/io/vscsiif.h>
+
+
+#define DPRINTK(_f, _a...)			\
+	pr_debug("(file=%s, line=%d) " _f,	\
+		 __FILE__ , __LINE__ , ## _a )
+
+struct ids_tuple {
+	unsigned int hst;		/* host    */
+	unsigned int chn;		/* channel */
+	unsigned int tgt;		/* target  */
+	unsigned int lun;		/* LUN     */
+};
+
+struct v2p_entry {
+	struct ids_tuple v;		/* translate from */
+	struct scsi_device *sdev;	/* translate to   */
+	struct list_head l;
+};
+
+struct vscsibk_info {
+	struct xenbus_device *dev;
+
+	domid_t domid;
+	unsigned int evtchn;
+	unsigned int irq;
+
+	struct vscsiif_back_ring  ring;
+	struct vm_struct *ring_area;
+	grant_handle_t shmem_handle;
+	grant_ref_t shmem_ref;
+
+	spinlock_t ring_lock;
+	atomic_t nr_unreplied_reqs;
+
+	spinlock_t v2p_lock;
+	struct list_head v2p_entry_lists;
+
+	struct task_struct *kthread;
+	wait_queue_head_t waiting_to_free;
+	wait_queue_head_t wq;
+	unsigned int waiting_reqs;
+	struct page **mmap_pages;
+
+};
+
+typedef struct {
+	unsigned char act;
+	struct vscsibk_info *info;
+	struct scsi_device *sdev;
+
+	uint16_t rqid;
+	
+	uint8_t nr_segments;
+	uint8_t cmnd[VSCSIIF_MAX_COMMAND_SIZE];
+	uint8_t cmd_len;
+
+	uint8_t sc_data_direction;
+	uint16_t timeout_per_command;
+	
+	uint32_t request_bufflen;
+	struct scatterlist *sgl;
+	grant_ref_t gref[VSCSIIF_SG_TABLESIZE];
+
+	int32_t rslt;
+	uint32_t resid;
+	uint8_t sense_buffer[VSCSIIF_SENSE_BUFFERSIZE];
+
+	struct list_head free_list;
+} pending_req_t;
+
+
+
+#define scsiback_get(_b) (atomic_inc(&(_b)->nr_unreplied_reqs))
+#define scsiback_put(_b)				\
+	do {						\
+		if (atomic_dec_and_test(&(_b)->nr_unreplied_reqs))	\
+			wake_up(&(_b)->waiting_to_free);\
+	} while (0)
+
+#define VSCSIIF_TIMEOUT		(900*HZ)
+
+
+irqreturn_t scsiback_intr(int, void *, struct pt_regs *);
+int scsiback_init_sring(struct vscsibk_info *info,
+		unsigned long ring_ref, unsigned int evtchn);
+int scsiback_schedule(void *data);
+
+
+struct vscsibk_info *vscsibk_info_alloc(domid_t domid);
+void scsiback_free(struct vscsibk_info *info);
+void scsiback_disconnect(struct vscsibk_info *info);
+int __init scsiback_interface_init(void);
+void scsiback_interface_exit(void);
+int scsiback_xenbus_init(void);
+void scsiback_xenbus_unregister(void);
+
+void scsiback_init_translation_table(struct vscsibk_info *info);
+
+int scsiback_add_translation_entry(struct vscsibk_info *info,
+			struct scsi_device *sdev, struct ids_tuple *v);
+
+int scsiback_del_translation_entry(struct vscsibk_info *info,
+				struct ids_tuple *v);
+struct scsi_device *scsiback_do_translation(struct vscsibk_info *info,
+			struct ids_tuple *v);
+void scsiback_release_translation_entry(struct vscsibk_info *info);
+
+
+void scsiback_cmd_exec(pending_req_t *pending_req);
+void scsiback_do_resp_with_sense(char *sense_buffer, int32_t result,
+			uint32_t resid, pending_req_t *pending_req);
+void scsiback_fast_flush_area(pending_req_t *req);
+
+void scsiback_rsp_emulation(pending_req_t *pending_req);
+void scsiback_req_emulation_or_cmdexec(pending_req_t *pending_req);
+void scsiback_emulation_init(void);
+
+
+#endif /* __SCSIIF__BACKEND__COMMON_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/scsiback/emulate.c linux-2.6.18-xen-3.3.0/drivers/xen/scsiback/emulate.c
--- linux-2.6.18.8/drivers/xen/scsiback/emulate.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/scsiback/emulate.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,454 @@
+/*
+ * Xen SCSI backend driver
+ *
+ * Copyright (c) 2008, FUJITSU Limited
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <scsi/scsi.h>
+#include <scsi/scsi_cmnd.h>
+#include <scsi/scsi_device.h>
+#include "common.h"
+
+/* Following SCSI commands are not defined in scsi/scsi.h */
+#define EXTENDED_COPY		0x83	/* EXTENDED COPY command        */
+#define REPORT_ALIASES		0xa3	/* REPORT ALIASES command       */
+#define CHANGE_ALIASES		0xa4	/* CHANGE ALIASES command       */
+#define SET_PRIORITY		0xa4	/* SET PRIORITY command         */
+
+
+/*
+  The bitmap in order to control emulation.
+  (Bit 3 to 7 are reserved for future use.)
+*/
+#define VSCSIIF_NEED_CMD_EXEC		0x01	/* If this bit is set, cmd exec	*/
+						/* is required.			*/
+#define VSCSIIF_NEED_EMULATE_REQBUF	0x02	/* If this bit is set, need	*/
+						/* emulation reqest buff before	*/
+						/* cmd exec.			*/
+#define VSCSIIF_NEED_EMULATE_RSPBUF	0x04	/* If this bit is set, need	*/
+						/* emulation resp buff after	*/
+						/* cmd exec.			*/
+
+/* Additional Sense Code (ASC) used */
+#define NO_ADDITIONAL_SENSE		0x0
+#define LOGICAL_UNIT_NOT_READY		0x4
+#define UNRECOVERED_READ_ERR		0x11
+#define PARAMETER_LIST_LENGTH_ERR	0x1a
+#define INVALID_OPCODE			0x20
+#define ADDR_OUT_OF_RANGE		0x21
+#define INVALID_FIELD_IN_CDB		0x24
+#define INVALID_FIELD_IN_PARAM_LIST	0x26
+#define POWERON_RESET			0x29
+#define SAVING_PARAMS_UNSUP		0x39
+#define THRESHOLD_EXCEEDED		0x5d
+#define LOW_POWER_COND_ON		0x5e
+
+
+
+/* Number os SCSI op_code	*/
+#define VSCSI_MAX_SCSI_OP_CODE		256
+static unsigned char bitmap[VSCSI_MAX_SCSI_OP_CODE];
+
+
+
+/*
+  Emulation routines for each SCSI op_code.
+*/
+static void (*pre_function[VSCSI_MAX_SCSI_OP_CODE])(pending_req_t *, void *);
+static void (*post_function[VSCSI_MAX_SCSI_OP_CODE])(pending_req_t *, void *);
+
+
+static const int check_condition_result =
+		(DRIVER_SENSE << 24) | SAM_STAT_CHECK_CONDITION;
+
+static void scsiback_mk_sense_buffer(uint8_t *data, uint8_t key,
+			uint8_t asc, uint8_t asq)
+{
+	data[0] = 0x70;  /* fixed, current */
+	data[2] = key;
+	data[7] = 0xa;	  /* implies 18 byte sense buffer */
+	data[12] = asc;
+	data[13] = asq;
+}
+
+static void resp_not_supported_cmd(pending_req_t *pending_req, void *data)
+{
+	scsiback_mk_sense_buffer(pending_req->sense_buffer, ILLEGAL_REQUEST,
+		INVALID_OPCODE, 0);
+	pending_req->resid = 0;
+	pending_req->rslt  = check_condition_result;
+}
+
+
+static int __copy_to_sg(struct scatterlist *sg, unsigned int nr_sg,
+	       void *buf, unsigned int buflen)
+{
+	void *from = buf;
+	void *to;
+	unsigned int from_rest = buflen;
+	unsigned int to_capa;
+	unsigned int copy_size = 0;
+	unsigned int i;
+	unsigned long pfn;
+
+	for (i = 0; i < nr_sg; i++) {
+		if (sg->page == NULL) {
+			printk(KERN_WARNING "%s: inconsistent length field in "
+			       "scatterlist\n", __FUNCTION__);
+			return -ENOMEM;
+		}
+
+		to_capa  = sg->length;
+		copy_size = min_t(unsigned int, to_capa, from_rest);
+
+		pfn = page_to_pfn(sg->page);
+		to = pfn_to_kaddr(pfn) + (sg->offset);
+		memcpy(to, from, copy_size);
+
+		from_rest  -= copy_size;
+		if (from_rest == 0) {
+			return 0;
+		}
+		
+		sg++;
+		from += copy_size;
+	}
+
+	printk(KERN_WARNING "%s: no space in scatterlist\n",
+	       __FUNCTION__);
+	return -ENOMEM;
+}
+
+static int __copy_from_sg(struct scatterlist *sg, unsigned int nr_sg,
+		 void *buf, unsigned int buflen)
+{
+	void *from;
+	void *to = buf;
+	unsigned int from_rest;
+	unsigned int to_capa = buflen;
+	unsigned int copy_size;
+	unsigned int i;
+	unsigned long pfn;
+
+	for (i = 0; i < nr_sg; i++) {
+		if (sg->page == NULL) {
+			printk(KERN_WARNING "%s: inconsistent length field in "
+			       "scatterlist\n", __FUNCTION__);
+			return -ENOMEM;
+		}
+
+		from_rest = sg->length;
+		if ((from_rest > 0) && (to_capa < from_rest)) {
+			printk(KERN_WARNING
+			       "%s: no space in destination buffer\n",
+			       __FUNCTION__);
+			return -ENOMEM;
+		}
+		copy_size = from_rest;
+
+		pfn = page_to_pfn(sg->page);
+		from = pfn_to_kaddr(pfn) + (sg->offset);
+		memcpy(to, from, copy_size);
+
+		to_capa  -= copy_size;
+		
+		sg++;
+		to += copy_size;
+	}
+
+	return 0;
+}
+
+static int __nr_luns_under_host(struct vscsibk_info *info)
+{
+	struct v2p_entry *entry;
+	struct list_head *head = &(info->v2p_entry_lists);
+	unsigned long flags;
+	int lun_cnt = 0;
+
+	spin_lock_irqsave(&info->v2p_lock, flags);
+	list_for_each_entry(entry, head, l) {
+			lun_cnt++;
+	}
+	spin_unlock_irqrestore(&info->v2p_lock, flags);
+
+	return (lun_cnt);
+}
+
+
+/* REPORT LUNS Define*/
+#define VSCSI_REPORT_LUNS_HEADER	8
+#define VSCSI_REPORT_LUNS_RETRY		3
+
+/* quoted scsi_debug.c/resp_report_luns() */
+static void __report_luns(pending_req_t *pending_req, void *data)
+{
+	struct vscsibk_info *info   = pending_req->info;
+	unsigned int        channel = pending_req->sdev->channel;
+	unsigned int        target  = pending_req->sdev->id;
+	unsigned int        nr_seg  = pending_req->nr_segments;
+	unsigned char *cmd = (unsigned char *)pending_req->cmnd;
+	
+	unsigned char *buff = NULL;
+	unsigned char alloc_len;
+	unsigned int alloc_luns = 0;
+	unsigned int req_bufflen = 0;
+	unsigned int actual_len = 0;
+	unsigned int retry_cnt = 0;
+	int select_report = (int)cmd[2];
+	int i, lun_cnt = 0, lun, upper, err = 0;
+	
+	struct v2p_entry *entry;
+	struct list_head *head = &(info->v2p_entry_lists);
+	unsigned long flags;
+	
+	struct scsi_lun *one_lun;
+
+	req_bufflen = cmd[9] + (cmd[8] << 8) + (cmd[7] << 16) + (cmd[6] << 24);
+	if ((req_bufflen < 4) || (select_report != 0))
+		goto fail;
+
+	alloc_luns = __nr_luns_under_host(info);
+	alloc_len  = sizeof(struct scsi_lun) * alloc_luns
+				+ VSCSI_REPORT_LUNS_HEADER;
+retry:
+	if ((buff = kmalloc(alloc_len, GFP_KERNEL)) == NULL) {
+		printk(KERN_ERR "scsiback:%s kmalloc err\n", __FUNCTION__);
+		goto fail;
+	}
+
+	memset(buff, 0, alloc_len);
+
+	one_lun = (struct scsi_lun *) &buff[8];
+	spin_lock_irqsave(&info->v2p_lock, flags);
+	list_for_each_entry(entry, head, l) {
+		if ((entry->v.chn == channel) &&
+		    (entry->v.tgt == target)) {
+			
+			/* check overflow */
+			if (lun_cnt >= alloc_luns) {
+				spin_unlock_irqrestore(&info->v2p_lock,
+							flags);
+
+				if (retry_cnt < VSCSI_REPORT_LUNS_RETRY) {
+					retry_cnt++;
+					if (buff)
+						kfree(buff);
+					goto retry;
+				}
+
+				goto fail;
+			}
+
+			lun = entry->v.lun;
+			upper = (lun >> 8) & 0x3f;
+			if (upper)
+				one_lun[lun_cnt].scsi_lun[0] = upper;
+			one_lun[lun_cnt].scsi_lun[1] = lun & 0xff;
+			lun_cnt++;
+		}
+	}
+
+	spin_unlock_irqrestore(&info->v2p_lock, flags);
+
+	buff[2] = ((sizeof(struct scsi_lun) * lun_cnt) >> 8) & 0xff;
+	buff[3] = (sizeof(struct scsi_lun) * lun_cnt) & 0xff;
+
+	actual_len = lun_cnt * sizeof(struct scsi_lun) 
+				+ VSCSI_REPORT_LUNS_HEADER;
+	req_bufflen = 0;
+	for (i = 0; i < nr_seg; i++)
+		req_bufflen += pending_req->sgl[i].length;
+
+	err = __copy_to_sg(pending_req->sgl, nr_seg, buff, 
+				min(req_bufflen, actual_len));
+	if (err)
+		goto fail;
+
+	memset(pending_req->sense_buffer, 0, VSCSIIF_SENSE_BUFFERSIZE);
+	pending_req->rslt = 0x00;
+	pending_req->resid = req_bufflen - min(req_bufflen, actual_len);
+
+	kfree(buff);
+	return;
+
+fail:
+	scsiback_mk_sense_buffer(pending_req->sense_buffer, ILLEGAL_REQUEST,
+		INVALID_FIELD_IN_CDB, 0);
+	pending_req->rslt  = check_condition_result;
+	pending_req->resid = 0;
+	if (buff)
+		kfree(buff);
+	return;
+}
+
+
+
+int __pre_do_emulation(pending_req_t *pending_req, void *data)
+{
+	uint8_t op_code = pending_req->cmnd[0];
+
+	if ((bitmap[op_code] & VSCSIIF_NEED_EMULATE_REQBUF) &&
+	    pre_function[op_code] != NULL) {
+		pre_function[op_code](pending_req, data);
+	}
+
+	/*
+	    0: no need for native driver call, so should return immediately.
+	    1: non emulation or should call native driver 
+	       after modifing the request buffer.
+	*/
+	return !!(bitmap[op_code] & VSCSIIF_NEED_CMD_EXEC);
+}
+
+void scsiback_rsp_emulation(pending_req_t *pending_req)
+{
+	uint8_t op_code = pending_req->cmnd[0];
+
+	if ((bitmap[op_code] & VSCSIIF_NEED_EMULATE_RSPBUF) &&
+	    post_function[op_code] != NULL) {
+		post_function[op_code](pending_req, NULL);
+	}
+
+	return;
+}
+
+
+void scsiback_req_emulation_or_cmdexec(pending_req_t *pending_req)
+{
+	if (__pre_do_emulation(pending_req, NULL)) {
+		scsiback_cmd_exec(pending_req);
+	}
+	else {
+		scsiback_fast_flush_area(pending_req);
+		scsiback_do_resp_with_sense(pending_req->sense_buffer,
+		  pending_req->rslt, pending_req->resid, pending_req);
+	}
+}
+
+
+/*
+  Following are not customizable functions.
+*/
+void scsiback_emulation_init(void)
+{
+	int i;
+
+	/* Initialize to default state */
+	for (i = 0; i < VSCSI_MAX_SCSI_OP_CODE; i++) {
+		bitmap[i]        = (VSCSIIF_NEED_EMULATE_REQBUF | 
+					VSCSIIF_NEED_EMULATE_RSPBUF);
+		pre_function[i]  = resp_not_supported_cmd;
+		post_function[i] = NULL;
+		/* means,
+		   - no need for pre-emulation
+		   - no need for post-emulation
+		   - call native driver
+		*/
+	}
+
+	/*
+	  Register appropriate functions below as you need.
+	  (See scsi/scsi.h for definition of SCSI op_code.)
+	*/
+
+	/*
+	  This command is Non emulation.
+	*/
+	bitmap[TEST_UNIT_READY] = VSCSIIF_NEED_CMD_EXEC;
+	pre_function[TEST_UNIT_READY] = NULL;
+	post_function[TEST_UNIT_READY] = NULL;
+
+	bitmap[REZERO_UNIT] = VSCSIIF_NEED_CMD_EXEC;
+	pre_function[REZERO_UNIT] = NULL;
+	post_function[REZERO_UNIT] = NULL;
+
+	bitmap[REQUEST_SENSE] = VSCSIIF_NEED_CMD_EXEC;
+	pre_function[REQUEST_SENSE] = NULL;
+	post_function[REQUEST_SENSE] = NULL;
+
+	bitmap[FORMAT_UNIT] = VSCSIIF_NEED_CMD_EXEC;
+	pre_function[FORMAT_UNIT] = NULL;
+	post_function[FORMAT_UNIT] = NULL;
+
+	bitmap[READ_BLOCK_LIMITS] = VSCSIIF_NEED_CMD_EXEC;
+	pre_function[READ_BLOCK_LIMITS] = NULL;
+	post_function[READ_BLOCK_LIMITS] = NULL;
+
+	bitmap[READ_6] = VSCSIIF_NEED_CMD_EXEC;
+	pre_function[READ_6] = NULL;
+	post_function[READ_6] = NULL;
+
+	bitmap[WRITE_6] = VSCSIIF_NEED_CMD_EXEC;
+	pre_function[WRITE_6] = NULL;
+	post_function[WRITE_6] = NULL;
+
+	bitmap[WRITE_FILEMARKS] = VSCSIIF_NEED_CMD_EXEC;
+	pre_function[WRITE_FILEMARKS] = NULL;
+	post_function[WRITE_FILEMARKS] = NULL;
+
+	bitmap[SPACE] = VSCSIIF_NEED_CMD_EXEC;
+	pre_function[SPACE] = NULL;
+	post_function[SPACE] = NULL;
+
+	bitmap[INQUIRY] = VSCSIIF_NEED_CMD_EXEC;
+	pre_function[INQUIRY] = NULL;
+	post_function[INQUIRY] = NULL;
+
+	bitmap[ERASE] = VSCSIIF_NEED_CMD_EXEC;
+	pre_function[ERASE] = NULL;
+	post_function[ERASE] = NULL;
+
+	bitmap[MODE_SENSE] = VSCSIIF_NEED_CMD_EXEC;
+	pre_function[MODE_SENSE] = NULL;
+	post_function[MODE_SENSE] = NULL;
+
+	bitmap[SEND_DIAGNOSTIC] = VSCSIIF_NEED_CMD_EXEC;
+	pre_function[SEND_DIAGNOSTIC] = NULL;
+	post_function[SEND_DIAGNOSTIC] = NULL;
+
+	bitmap[READ_CAPACITY] = VSCSIIF_NEED_CMD_EXEC;
+	pre_function[READ_CAPACITY] = NULL;
+	post_function[READ_CAPACITY] = NULL;
+
+	bitmap[READ_10] = VSCSIIF_NEED_CMD_EXEC;
+	pre_function[READ_10] = NULL;
+	post_function[READ_10] = NULL;
+
+	bitmap[WRITE_10] = VSCSIIF_NEED_CMD_EXEC;
+	pre_function[WRITE_10] = NULL;
+	post_function[WRITE_10] = NULL;
+
+	/*
+	  This command is Full emulation.
+	*/
+	pre_function[REPORT_LUNS] = __report_luns;
+	bitmap[REPORT_LUNS] = (VSCSIIF_NEED_EMULATE_REQBUF | 
+					VSCSIIF_NEED_EMULATE_RSPBUF);
+
+	return;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/scsiback/interface.c linux-2.6.18-xen-3.3.0/drivers/xen/scsiback/interface.c
--- linux-2.6.18.8/drivers/xen/scsiback/interface.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/scsiback/interface.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,182 @@
+/*
+ * interface management.
+ *
+ * Copyright (c) 2008, FUJITSU Limited
+ *
+ * Based on the blkback driver code.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <scsi/scsi.h>
+#include <scsi/scsi_host.h>
+#include <scsi/scsi_device.h>
+#include "common.h"
+
+#include <xen/evtchn.h>
+#include <linux/kthread.h>
+
+
+static kmem_cache_t *scsiback_cachep;
+
+struct vscsibk_info *vscsibk_info_alloc(domid_t domid)
+{
+	struct vscsibk_info *info;
+
+	info = kmem_cache_alloc(scsiback_cachep, GFP_KERNEL);
+	if (!info)
+		return ERR_PTR(-ENOMEM);
+
+	memset(info, 0, sizeof(*info));
+	info->domid = domid;
+	spin_lock_init(&info->ring_lock);
+	atomic_set(&info->nr_unreplied_reqs, 0);
+	init_waitqueue_head(&info->wq);
+	init_waitqueue_head(&info->waiting_to_free);
+
+	return info;
+}
+
+static int map_frontend_page( struct vscsibk_info *info,
+				unsigned long ring_ref)
+{
+	struct gnttab_map_grant_ref op;
+	int err;
+
+	gnttab_set_map_op(&op, (unsigned long)info->ring_area->addr,
+				GNTMAP_host_map, ring_ref,
+				info->domid);
+
+	err = HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, &op, 1);
+	BUG_ON(err);
+
+	if (op.status) {
+		printk(KERN_ERR "scsiback: Grant table operation failure !\n");
+		return op.status;
+	}
+
+	info->shmem_ref    = ring_ref;
+	info->shmem_handle = op.handle;
+
+	return (GNTST_okay);
+}
+
+static void unmap_frontend_page(struct vscsibk_info *info)
+{
+	struct gnttab_unmap_grant_ref op;
+	int err;
+
+	gnttab_set_unmap_op(&op, (unsigned long)info->ring_area->addr,
+				GNTMAP_host_map, info->shmem_handle);
+
+	err = HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1);
+	BUG_ON(err);
+
+}
+
+int scsiback_init_sring(struct vscsibk_info *info,
+		unsigned long ring_ref, unsigned int evtchn)
+{
+	struct vscsiif_sring *sring;
+	int err;
+
+	if (info->irq) {
+		printk(KERN_ERR "scsiback: Already connected through?\n");
+		return -1;
+	}
+
+	info->ring_area = alloc_vm_area(PAGE_SIZE);
+	if (!info)
+		return -ENOMEM;
+
+	err = map_frontend_page(info, ring_ref);
+	if (err)
+		goto free_vm;
+
+	sring = (struct vscsiif_sring *) info->ring_area->addr;
+	BACK_RING_INIT(&info->ring, sring, PAGE_SIZE);
+
+	err = bind_interdomain_evtchn_to_irqhandler(
+			info->domid, evtchn,
+			scsiback_intr, 0, "vscsiif-backend", info);
+
+	if (err < 0)
+		goto unmap_page;
+		
+	info->irq = err;
+
+	return 0;
+
+unmap_page:
+	unmap_frontend_page(info);
+free_vm:
+	free_vm_area(info->ring_area);
+
+	return err;
+}
+
+void scsiback_disconnect(struct vscsibk_info *info)
+{
+	if (info->kthread) {
+		kthread_stop(info->kthread);
+		info->kthread = NULL;
+	}
+
+	wait_event(info->waiting_to_free, 
+		atomic_read(&info->nr_unreplied_reqs) == 0);
+
+	if (info->irq) {
+		unbind_from_irqhandler(info->irq, info);
+		info->irq = 0;
+	}
+
+	if (info->ring.sring) {
+		unmap_frontend_page(info);
+		free_vm_area(info->ring_area);
+		info->ring.sring = NULL;
+	}
+}
+
+void scsiback_free(struct vscsibk_info *info)
+{
+	kmem_cache_free(scsiback_cachep, info);
+}
+
+int __init scsiback_interface_init(void)
+{
+	scsiback_cachep = kmem_cache_create("vscsiif_cache",
+		sizeof(struct vscsibk_info), 0, 0, NULL, NULL);
+	if (!scsiback_cachep) {
+		printk(KERN_ERR "scsiback: can't init scsi cache\n");
+		return -ENOMEM;
+	}
+	
+	return 0;
+}
+
+void scsiback_interface_exit(void)
+{
+	kmem_cache_destroy(scsiback_cachep);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/scsiback/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/scsiback/Makefile
--- linux-2.6.18.8/drivers/xen/scsiback/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/scsiback/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,4 @@
+obj-$(CONFIG_XEN_SCSI_BACKEND) := xen-scsibk.o
+
+xen-scsibk-y	:= interface.o scsiback.o xenbus.o translate.o emulate.o
+
diff -rpuN linux-2.6.18.8/drivers/xen/scsiback/scsiback.c linux-2.6.18-xen-3.3.0/drivers/xen/scsiback/scsiback.c
--- linux-2.6.18.8/drivers/xen/scsiback/scsiback.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/scsiback/scsiback.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,717 @@
+/*
+ * Xen SCSI backend driver
+ *
+ * Copyright (c) 2008, FUJITSU Limited
+ *
+ * Based on the blkback driver code.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/spinlock.h>
+#include <linux/kthread.h>
+#include <linux/list.h>
+#include <linux/delay.h>
+#include <xen/balloon.h>
+#include <asm/hypervisor.h>
+#include <scsi/scsi.h>
+#include <scsi/scsi_cmnd.h>
+#include <scsi/scsi_host.h>
+#include <scsi/scsi_device.h>
+#include <scsi/scsi_dbg.h>
+#include <scsi/scsi_eh.h>
+
+#include "common.h"
+
+
+struct list_head pending_free;
+DEFINE_SPINLOCK(pending_free_lock);
+DECLARE_WAIT_QUEUE_HEAD(pending_free_wq);
+
+int vscsiif_reqs = VSCSIIF_BACK_MAX_PENDING_REQS;
+module_param_named(reqs, vscsiif_reqs, int, 0);
+MODULE_PARM_DESC(reqs, "Number of scsiback requests to allocate");
+
+static unsigned int log_print_stat = 0;
+module_param(log_print_stat, int, 0644);
+
+#define SCSIBACK_INVALID_HANDLE (~0)
+
+static pending_req_t *pending_reqs;
+static struct page **pending_pages;
+static grant_handle_t *pending_grant_handles;
+
+static int vaddr_pagenr(pending_req_t *req, int seg)
+{
+	return (req - pending_reqs) * VSCSIIF_SG_TABLESIZE + seg;
+}
+
+static unsigned long vaddr(pending_req_t *req, int seg)
+{
+	unsigned long pfn = page_to_pfn(pending_pages[vaddr_pagenr(req, seg)]);
+	return (unsigned long)pfn_to_kaddr(pfn);
+}
+
+#define pending_handle(_req, _seg) \
+	(pending_grant_handles[vaddr_pagenr(_req, _seg)])
+
+
+void scsiback_fast_flush_area(pending_req_t *req)
+{
+	struct gnttab_unmap_grant_ref unmap[VSCSIIF_SG_TABLESIZE];
+	unsigned int i, invcount = 0;
+	grant_handle_t handle;
+	int err;
+
+	if (req->nr_segments) {
+		for (i = 0; i < req->nr_segments; i++) {
+			handle = pending_handle(req, i);
+			if (handle == SCSIBACK_INVALID_HANDLE)
+				continue;
+			gnttab_set_unmap_op(&unmap[i], vaddr(req, i),
+						GNTMAP_host_map, handle);
+			pending_handle(req, i) = SCSIBACK_INVALID_HANDLE;
+			invcount++;
+		}
+
+		err = HYPERVISOR_grant_table_op(
+			GNTTABOP_unmap_grant_ref, unmap, invcount);
+		BUG_ON(err);
+		kfree(req->sgl);
+	}
+
+	return;
+}
+
+
+static pending_req_t * alloc_req(struct vscsibk_info *info)
+{
+	pending_req_t *req = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&pending_free_lock, flags);
+	if (!list_empty(&pending_free)) {
+		req = list_entry(pending_free.next, pending_req_t, free_list);
+		list_del(&req->free_list);
+	}
+	spin_unlock_irqrestore(&pending_free_lock, flags);
+	return req;
+}
+
+
+static void free_req(pending_req_t *req)
+{
+	unsigned long flags;
+	int was_empty;
+
+	spin_lock_irqsave(&pending_free_lock, flags);
+	was_empty = list_empty(&pending_free);
+	list_add(&req->free_list, &pending_free);
+	spin_unlock_irqrestore(&pending_free_lock, flags);
+	if (was_empty)
+		wake_up(&pending_free_wq);
+}
+
+
+static void scsiback_notify_work(struct vscsibk_info *info)
+{
+	info->waiting_reqs = 1;
+	wake_up(&info->wq);
+}
+
+void scsiback_do_resp_with_sense(char *sense_buffer, int32_t result,
+			uint32_t resid, pending_req_t *pending_req)
+{
+	vscsiif_response_t *ring_res;
+	struct vscsibk_info *info = pending_req->info;
+	int notify;
+	int more_to_do = 1;
+	unsigned long flags;
+
+	DPRINTK("%s\n",__FUNCTION__);
+
+	spin_lock_irqsave(&info->ring_lock, flags);
+
+	ring_res = RING_GET_RESPONSE(&info->ring, info->ring.rsp_prod_pvt);
+	info->ring.rsp_prod_pvt++;
+
+	ring_res->rslt   = result;
+	ring_res->rqid   = pending_req->rqid;
+
+	if (sense_buffer != NULL) {
+		memcpy(ring_res->sense_buffer, sense_buffer,
+				VSCSIIF_SENSE_BUFFERSIZE);
+		ring_res->sense_len = VSCSIIF_SENSE_BUFFERSIZE;
+	} else {
+		ring_res->sense_len = 0;
+	}
+
+	ring_res->residual_len = resid;
+
+	RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(&info->ring, notify);
+	if (info->ring.rsp_prod_pvt == info->ring.req_cons) {
+		RING_FINAL_CHECK_FOR_REQUESTS(&info->ring, more_to_do);
+	} else if (RING_HAS_UNCONSUMED_REQUESTS(&info->ring)) {
+		more_to_do = 1;
+	}
+	
+	spin_unlock_irqrestore(&info->ring_lock, flags);
+
+	if (more_to_do)
+		scsiback_notify_work(info);
+
+	if (notify)
+		notify_remote_via_irq(info->irq);
+
+	free_req(pending_req);
+}
+
+static void scsiback_print_status(char *sense_buffer, int errors,
+					pending_req_t *pending_req)
+{
+	struct scsi_device *sdev = pending_req->sdev;
+	
+	printk(KERN_ERR "scsiback: %d:%d:%d:%d ",sdev->host->host_no,
+			sdev->channel, sdev->id, sdev->lun);
+	printk(KERN_ERR "status = 0x%02x, message = 0x%02x, host = 0x%02x, driver = 0x%02x\n",
+			status_byte(errors), msg_byte(errors),
+			host_byte(errors), driver_byte(errors));
+
+	printk(KERN_ERR "scsiback: cmnd[0]=0x%02X\n",
+			pending_req->cmnd[0]);
+
+	if (CHECK_CONDITION & status_byte(errors))
+		__scsi_print_sense("scsiback", sense_buffer, SCSI_SENSE_BUFFERSIZE);
+}
+
+
+static void scsiback_cmd_done(struct request *req, int errors)
+{
+	pending_req_t *pending_req = req->end_io_data;
+	unsigned char *sense_buffer;
+	unsigned int resid;
+
+	sense_buffer = req->sense;
+	resid        = req->data_len;
+
+	if (errors != 0) {
+		if (log_print_stat)
+			scsiback_print_status(sense_buffer, errors, pending_req);
+	}
+
+	scsiback_rsp_emulation(pending_req);
+
+	scsiback_fast_flush_area(pending_req);
+	scsiback_do_resp_with_sense(sense_buffer, errors, resid, pending_req);
+	scsiback_put(pending_req->info);
+
+	__blk_put_request(req->q, req);
+}
+
+
+static int scsiback_gnttab_data_map(vscsiif_request_t *ring_req,
+					pending_req_t *pending_req)
+{
+	u32 flags;
+	int write;
+	int i, err = 0;
+	unsigned int data_len = 0;
+	struct gnttab_map_grant_ref map[VSCSIIF_SG_TABLESIZE];
+	struct vscsibk_info *info   = pending_req->info;
+
+	int data_dir = (int)pending_req->sc_data_direction;
+	unsigned int nr_segments = (unsigned int)pending_req->nr_segments;
+
+	write = (data_dir == DMA_TO_DEVICE);
+
+	if (nr_segments) {
+		/* free of (sgl) in fast_flush_area()*/
+		pending_req->sgl = kmalloc(sizeof(struct scatterlist) * nr_segments,
+						GFP_KERNEL);
+		if (!pending_req->sgl) {
+			printk(KERN_ERR "scsiback: %s: kmalloc() error.\n", __FUNCTION__);
+			return -ENOMEM;
+		}
+
+		for (i = 0; i < nr_segments; i++) {
+			flags = GNTMAP_host_map;
+			if (write)
+				flags |= GNTMAP_readonly;
+			gnttab_set_map_op(&map[i], vaddr(pending_req, i), flags,
+						ring_req->seg[i].gref,
+						info->domid);
+		}
+
+		err = HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, map, nr_segments);
+		BUG_ON(err);
+
+		for (i = 0; i < nr_segments; i++) {
+			if (unlikely(map[i].status != 0)) {
+				printk(KERN_ERR "scsiback: invalid buffer -- could not remap it\n");
+				map[i].handle = SCSIBACK_INVALID_HANDLE;
+				err |= 1;
+			}
+
+			pending_handle(pending_req, i) = map[i].handle;
+
+			if (err)
+				continue;
+
+			set_phys_to_machine(__pa(vaddr(
+				pending_req, i)) >> PAGE_SHIFT,
+				FOREIGN_FRAME(map[i].dev_bus_addr >> PAGE_SHIFT));
+
+			pending_req->sgl[i].page   = virt_to_page(vaddr(pending_req, i));
+			pending_req->sgl[i].offset = ring_req->seg[i].offset;
+			pending_req->sgl[i].length = ring_req->seg[i].length;
+			data_len += pending_req->sgl[i].length;
+
+			barrier();
+			if (pending_req->sgl[i].offset >= PAGE_SIZE ||
+			    pending_req->sgl[i].length > PAGE_SIZE ||
+			    pending_req->sgl[i].offset + pending_req->sgl[i].length > PAGE_SIZE)
+				err |= 1;
+
+		}
+
+		if (err)
+			goto fail_flush;
+	}
+	
+	pending_req->request_bufflen = data_len;
+	
+	return 0;
+	
+fail_flush:
+	scsiback_fast_flush_area(pending_req);
+	return -ENOMEM;
+}
+
+/* quoted scsi_lib.c/scsi_merge_bio */
+static int scsiback_merge_bio(struct request *rq, struct bio *bio)
+{
+	struct request_queue *q = rq->q;
+
+	bio->bi_flags &= ~(1 << BIO_SEG_VALID);
+	if (rq_data_dir(rq) == WRITE)
+		bio->bi_rw |= (1 << BIO_RW);
+
+	blk_queue_bounce(q, &bio);
+
+	if (!rq->bio)
+		blk_rq_bio_prep(q, rq, bio);
+	else if (!q->back_merge_fn(q, rq, bio))
+		return -EINVAL;
+	else {
+		rq->biotail->bi_next = bio;
+		rq->biotail          = bio;
+		rq->hard_nr_sectors += bio_sectors(bio);
+		rq->nr_sectors       = rq->hard_nr_sectors;
+	}
+
+	return 0;
+}
+
+
+/* quoted scsi_lib.c/scsi_bi_endio */
+static int scsiback_bi_endio(struct bio *bio, unsigned int bytes_done, int error)
+{
+	if (bio->bi_size)
+		return 1;
+
+	bio_put(bio);
+	return 0;
+}
+
+
+
+/* quoted scsi_lib.c/scsi_req_map_sg . */
+static int request_map_sg(struct request *rq, pending_req_t *pending_req, unsigned int count)
+{
+	struct request_queue *q = rq->q;
+	int nr_pages;
+	unsigned int nsegs = count;
+
+	unsigned int data_len = 0, len, bytes, off;
+	struct page *page;
+	struct bio *bio = NULL;
+	int i, err, nr_vecs = 0;
+
+	for (i = 0; i < nsegs; i++) {
+		page = pending_req->sgl[i].page;
+		off = (unsigned int)pending_req->sgl[i].offset;
+		len = (unsigned int)pending_req->sgl[i].length;
+		data_len += len;
+
+		nr_pages = (len + off + PAGE_SIZE - 1) >> PAGE_SHIFT;
+		while (len > 0) {
+			bytes = min_t(unsigned int, len, PAGE_SIZE - off);
+
+			if (!bio) {
+				nr_vecs = min_t(int, BIO_MAX_PAGES, nr_pages);
+				nr_pages -= nr_vecs;
+				bio = bio_alloc(GFP_KERNEL, nr_vecs);
+				if (!bio) {
+					err = -ENOMEM;
+					goto free_bios;
+				}
+				bio->bi_end_io = scsiback_bi_endio;
+			}
+
+			if (bio_add_pc_page(q, bio, page, bytes, off) !=
+						bytes) {
+				bio_put(bio);
+				err = -EINVAL;
+				goto free_bios;
+			}
+
+			if (bio->bi_vcnt >= nr_vecs) {
+				err = scsiback_merge_bio(rq, bio);
+				if (err) {
+					bio_endio(bio, bio->bi_size, 0);
+					goto free_bios;
+				}
+				bio = NULL;
+			}
+
+			page++;
+			len -= bytes;
+			off = 0;
+		}
+	}
+
+	rq->buffer   = rq->data = NULL;
+	rq->data_len = data_len;
+
+	return 0;
+
+free_bios:
+	while ((bio = rq->bio) != NULL) {
+		rq->bio = bio->bi_next;
+		/*
+		 * call endio instead of bio_put incase it was bounced
+		 */
+		bio_endio(bio, bio->bi_size, 0);
+	}
+
+	return err;
+}
+
+
+void scsiback_cmd_exec(pending_req_t *pending_req)
+{
+	int cmd_len  = (int)pending_req->cmd_len;
+	int data_dir = (int)pending_req->sc_data_direction;
+	unsigned int nr_segments = (unsigned int)pending_req->nr_segments;
+	unsigned int timeout;
+	struct request *rq;
+	int write;
+
+	DPRINTK("%s\n",__FUNCTION__);
+
+	/* because it doesn't timeout backend earlier than frontend.*/
+	if (pending_req->timeout_per_command)
+		timeout = pending_req->timeout_per_command * HZ;
+	else
+		timeout = VSCSIIF_TIMEOUT;
+
+	write = (data_dir == DMA_TO_DEVICE);
+	rq = blk_get_request(pending_req->sdev->request_queue, write, GFP_KERNEL);
+
+	rq->flags  |= REQ_BLOCK_PC;
+	rq->cmd_len = cmd_len;
+	memcpy(rq->cmd, pending_req->cmnd, cmd_len);
+
+	memset(pending_req->sense_buffer, 0, VSCSIIF_SENSE_BUFFERSIZE);
+	rq->sense       = pending_req->sense_buffer;
+	rq->sense_len = 0;
+
+	/* not allowed to retry in backend.                   */
+	rq->retries   = 0;
+	rq->timeout   = timeout;
+	rq->end_io_data = pending_req;
+
+	if (nr_segments) {
+
+		if (request_map_sg(rq, pending_req, nr_segments)) {
+			printk(KERN_ERR "scsiback: SG Request Map Error\n");
+			return;
+		}
+	}
+
+	scsiback_get(pending_req->info);
+	blk_execute_rq_nowait(rq->q, NULL, rq, 1, scsiback_cmd_done);
+
+	return ;
+}
+
+
+static void scsiback_device_reset_exec(pending_req_t *pending_req)
+{
+	struct vscsibk_info *info = pending_req->info;
+	int err;
+	struct scsi_device *sdev = pending_req->sdev;
+
+	scsiback_get(info);
+	err = scsi_reset_provider(sdev, SCSI_TRY_RESET_DEVICE);
+
+	scsiback_do_resp_with_sense(NULL, err, 0, pending_req);
+	scsiback_put(info);
+
+	return;
+}
+
+
+irqreturn_t scsiback_intr(int irq, void *dev_id, struct pt_regs *regs)
+{
+	scsiback_notify_work((struct vscsibk_info *)dev_id);
+	return IRQ_HANDLED;
+}
+
+static int prepare_pending_reqs(struct vscsibk_info *info,
+		vscsiif_request_t *ring_req, pending_req_t *pending_req)
+{
+	struct scsi_device *sdev;
+	struct ids_tuple vir;
+	int err = -EINVAL;
+
+	DPRINTK("%s\n",__FUNCTION__);
+
+	pending_req->rqid       = ring_req->rqid;
+	pending_req->act        = ring_req->act;
+
+	pending_req->info       = info;
+
+	vir.chn = ring_req->channel;
+	vir.tgt = ring_req->id;
+	vir.lun = ring_req->lun;
+
+	rmb();
+	sdev = scsiback_do_translation(info, &vir);
+	if (!sdev) {
+		pending_req->sdev = NULL;
+		DPRINTK("scsiback: doesn't exist.\n");
+		err = -ENODEV;
+		goto invalid_value;
+	}
+	pending_req->sdev = sdev;
+
+	/* request range check from frontend */
+	pending_req->sc_data_direction = ring_req->sc_data_direction;
+	barrier();
+	if ((pending_req->sc_data_direction != DMA_BIDIRECTIONAL) &&
+		(pending_req->sc_data_direction != DMA_TO_DEVICE) &&
+		(pending_req->sc_data_direction != DMA_FROM_DEVICE) &&
+		(pending_req->sc_data_direction != DMA_NONE)) {
+		DPRINTK("scsiback: invalid parameter data_dir = %d\n",
+			pending_req->sc_data_direction);
+		err = -EINVAL;
+		goto invalid_value;
+	}
+
+	pending_req->nr_segments = ring_req->nr_segments;
+	barrier();
+	if (pending_req->nr_segments > VSCSIIF_SG_TABLESIZE) {
+		DPRINTK("scsiback: invalid parameter nr_seg = %d\n",
+			pending_req->nr_segments);
+		err = -EINVAL;
+		goto invalid_value;
+	}
+
+	pending_req->cmd_len = ring_req->cmd_len;
+	barrier();
+	if (pending_req->cmd_len > VSCSIIF_MAX_COMMAND_SIZE) {
+		DPRINTK("scsiback: invalid parameter cmd_len = %d\n",
+			pending_req->cmd_len);
+		err = -EINVAL;
+		goto invalid_value;
+	}
+	memcpy(pending_req->cmnd, ring_req->cmnd, pending_req->cmd_len);
+	
+	pending_req->timeout_per_command = ring_req->timeout_per_command;
+
+	if(scsiback_gnttab_data_map(ring_req, pending_req)) {
+		DPRINTK("scsiback: invalid buffer\n");
+		err = -EINVAL;
+		goto invalid_value;
+	}
+
+	return 0;
+
+invalid_value:
+	return err;
+}
+
+
+static int scsiback_do_cmd_fn(struct vscsibk_info *info)
+{
+	struct vscsiif_back_ring *ring = &info->ring;
+	vscsiif_request_t  *ring_req;
+
+	pending_req_t *pending_req;
+	RING_IDX rc, rp;
+	int err, more_to_do = 0;
+
+	DPRINTK("%s\n",__FUNCTION__);
+
+	rc = ring->req_cons;
+	rp = ring->sring->req_prod;
+	rmb();
+
+	while ((rc != rp)) {
+		if (RING_REQUEST_CONS_OVERFLOW(ring, rc))
+			break;
+		pending_req = alloc_req(info);
+		if (NULL == pending_req) {
+			more_to_do = 1;
+			break;
+		}
+
+		ring_req = RING_GET_REQUEST(ring, rc);
+		ring->req_cons = ++rc;
+
+		err = prepare_pending_reqs(info, ring_req,
+						pending_req);
+		if (err == -EINVAL) {
+			scsiback_do_resp_with_sense(NULL, (DRIVER_ERROR << 24),
+				0, pending_req);
+			continue;
+		} else if (err == -ENODEV) {
+			scsiback_do_resp_with_sense(NULL, (DID_NO_CONNECT << 16),
+				0, pending_req);
+			continue;
+		}
+
+		if (pending_req->act == VSCSIIF_ACT_SCSI_CDB) {
+			scsiback_req_emulation_or_cmdexec(pending_req);
+		} else if (pending_req->act == VSCSIIF_ACT_SCSI_RESET) {
+			scsiback_device_reset_exec(pending_req);
+		} else {
+			printk(KERN_ERR "scsiback: invalid parameter for request\n");
+			scsiback_do_resp_with_sense(NULL, (DRIVER_ERROR << 24),
+				0, pending_req);
+			continue;
+		}
+	}
+
+	if (RING_HAS_UNCONSUMED_REQUESTS(ring))
+		more_to_do = 1;
+
+	/* Yield point for this unbounded loop. */
+	cond_resched();
+
+	return more_to_do;
+}
+
+
+int scsiback_schedule(void *data)
+{
+	struct vscsibk_info *info = (struct vscsibk_info *)data;
+
+	DPRINTK("%s\n",__FUNCTION__);
+
+	while (!kthread_should_stop()) {
+		wait_event_interruptible(
+			info->wq,
+			info->waiting_reqs || kthread_should_stop());
+		wait_event_interruptible(
+			pending_free_wq,
+			!list_empty(&pending_free) || kthread_should_stop());
+
+		info->waiting_reqs = 0;
+		smp_mb();
+
+		if (scsiback_do_cmd_fn(info))
+			info->waiting_reqs = 1;
+	}
+
+	return 0;
+}
+
+
+static int __init scsiback_init(void)
+{
+	int i, mmap_pages;
+
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	mmap_pages = vscsiif_reqs * VSCSIIF_SG_TABLESIZE;
+
+	pending_reqs          = kmalloc(sizeof(pending_reqs[0]) *
+					vscsiif_reqs, GFP_KERNEL);
+	pending_grant_handles = kmalloc(sizeof(pending_grant_handles[0]) *
+					mmap_pages, GFP_KERNEL);
+	pending_pages         = alloc_empty_pages_and_pagevec(mmap_pages);
+
+	if (!pending_reqs || !pending_grant_handles || !pending_pages)
+		goto out_of_memory;
+
+	for (i = 0; i < mmap_pages; i++)
+		pending_grant_handles[i] = SCSIBACK_INVALID_HANDLE;
+
+	if (scsiback_interface_init() < 0)
+		goto out_of_kmem;
+
+	memset(pending_reqs, 0, sizeof(pending_reqs));
+	INIT_LIST_HEAD(&pending_free);
+
+	for (i = 0; i < vscsiif_reqs; i++)
+		list_add_tail(&pending_reqs[i].free_list, &pending_free);
+
+	if (scsiback_xenbus_init())
+		goto out_of_xenbus;
+
+	scsiback_emulation_init();
+
+	return 0;
+
+out_of_xenbus:
+	scsiback_xenbus_unregister();
+out_of_kmem:
+	scsiback_interface_exit();
+out_of_memory:
+	kfree(pending_reqs);
+	kfree(pending_grant_handles);
+	free_empty_pages_and_pagevec(pending_pages, mmap_pages);
+	printk(KERN_ERR "scsiback: %s: out of memory\n", __FUNCTION__);
+	return -ENOMEM;
+}
+
+static void __exit scsiback_exit(void)
+{
+	scsiback_xenbus_unregister();
+	scsiback_interface_exit();
+	kfree(pending_reqs);
+	kfree(pending_grant_handles);
+	free_empty_pages_and_pagevec(pending_pages, (vscsiif_reqs * VSCSIIF_SG_TABLESIZE));
+
+}
+
+module_init(scsiback_init);
+module_exit(scsiback_exit);
+
+MODULE_DESCRIPTION("Xen SCSI backend driver");
+MODULE_LICENSE("Dual BSD/GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/scsiback/translate.c linux-2.6.18-xen-3.3.0/drivers/xen/scsiback/translate.c
--- linux-2.6.18.8/drivers/xen/scsiback/translate.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/scsiback/translate.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,168 @@
+/*
+ * Xen SCSI backend driver
+ *
+ * Copyright (c) 2008, FUJITSU Limited
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/list.h>
+#include <linux/gfp.h>
+
+#include "common.h"
+
+/*
+  Initialize the translation entry list
+*/
+void scsiback_init_translation_table(struct vscsibk_info *info)
+{
+	INIT_LIST_HEAD(&info->v2p_entry_lists);
+	spin_lock_init(&info->v2p_lock);
+}
+
+
+/*
+  Add a new translation entry
+*/
+int scsiback_add_translation_entry(struct vscsibk_info *info,
+			struct scsi_device *sdev, struct ids_tuple *v)
+{
+	int err = 0;
+	struct v2p_entry *entry;
+	struct v2p_entry *new;
+	struct list_head *head = &(info->v2p_entry_lists);
+	unsigned long flags;
+	
+	spin_lock_irqsave(&info->v2p_lock, flags);
+
+	/* Check double assignment to identical virtual ID */
+	list_for_each_entry(entry, head, l) {
+		if ((entry->v.chn == v->chn) &&
+		    (entry->v.tgt == v->tgt) &&
+		    (entry->v.lun == v->lun)) {
+			printk(KERN_WARNING "scsiback: Virtual ID is already used. "
+			       "Assignment was not performed.\n");
+			err = -EEXIST;
+			goto out;
+		}
+
+	}
+
+	/* Create a new translation entry and add to the list */
+	if ((new = kmalloc(sizeof(struct v2p_entry), GFP_ATOMIC)) == NULL) {
+		printk(KERN_ERR "scsiback: %s: kmalloc() error.\n", __FUNCTION__);
+		err = -ENOMEM;
+		goto out;
+	}
+	new->v = *v;
+	new->sdev = sdev;
+	list_add_tail(&new->l, head);
+
+out:	
+	spin_unlock_irqrestore(&info->v2p_lock, flags);
+	return err;
+}
+
+
+/*
+  Delete the translation entry specfied
+*/
+int scsiback_del_translation_entry(struct vscsibk_info *info,
+				struct ids_tuple *v)
+{
+	struct v2p_entry *entry;
+	struct list_head *head = &(info->v2p_entry_lists);
+	unsigned long flags;
+
+	spin_lock_irqsave(&info->v2p_lock, flags);
+	/* Find out the translation entry specified */
+	list_for_each_entry(entry, head, l) {
+		if ((entry->v.chn == v->chn) &&
+		    (entry->v.tgt == v->tgt) &&
+		    (entry->v.lun == v->lun)) {
+			goto found;
+		}
+	}
+
+	spin_unlock_irqrestore(&info->v2p_lock, flags);
+	return 1;
+
+found:
+	/* Delete the translation entry specfied */
+	scsi_device_put(entry->sdev);
+	list_del(&entry->l);
+	kfree(entry);
+
+	spin_unlock_irqrestore(&info->v2p_lock, flags);
+	return 0;
+}
+
+
+/*
+  Perform virtual to physical translation
+*/
+struct scsi_device *scsiback_do_translation(struct vscsibk_info *info,
+			struct ids_tuple *v)
+{
+	struct v2p_entry *entry;
+	struct list_head *head = &(info->v2p_entry_lists);
+	struct scsi_device *sdev = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&info->v2p_lock, flags);
+	list_for_each_entry(entry, head, l) {
+		if ((entry->v.chn == v->chn) &&
+		    (entry->v.tgt == v->tgt) &&
+		    (entry->v.lun == v->lun)) {
+			sdev = entry->sdev;
+			goto out;
+		}
+	}
+out:
+	spin_unlock_irqrestore(&info->v2p_lock, flags);
+	return sdev;
+}
+
+
+/*
+  Release the translation entry specfied
+*/
+void scsiback_release_translation_entry(struct vscsibk_info *info)
+{
+	struct v2p_entry *entry, *tmp;
+	struct list_head *head = &(info->v2p_entry_lists);
+	unsigned long flags;
+
+	spin_lock_irqsave(&info->v2p_lock, flags);
+	list_for_each_entry_safe(entry, tmp, head, l) {
+		scsi_device_put(entry->sdev);
+		list_del(&entry->l);
+		kfree(entry);
+	}
+
+	spin_unlock_irqrestore(&info->v2p_lock, flags);
+	return;
+
+}
diff -rpuN linux-2.6.18.8/drivers/xen/scsiback/xenbus.c linux-2.6.18-xen-3.3.0/drivers/xen/scsiback/xenbus.c
--- linux-2.6.18.8/drivers/xen/scsiback/xenbus.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/scsiback/xenbus.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,368 @@
+/*
+ * Xen SCSI backend driver
+ *
+ * Copyright (c) 2008, FUJITSU Limited
+ *
+ * Based on the blkback driver code.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <stdarg.h>
+#include <linux/module.h>
+#include <linux/kthread.h>
+#include <scsi/scsi.h>
+#include <scsi/scsi_host.h>
+#include <scsi/scsi_device.h>
+
+#include "common.h"
+
+struct backend_info
+{
+	struct xenbus_device *dev;
+	struct vscsibk_info *info;
+};
+
+
+static int __vscsiif_name(struct backend_info *be, char *buf)
+{
+	struct xenbus_device *dev = be->dev;
+	unsigned int domid, id;
+
+	sscanf(dev->nodename, "backend/vscsi/%u/%u", &domid, &id);
+	snprintf(buf, TASK_COMM_LEN, "vscsi.%u.%u", be->info->domid, id);
+
+	return 0;
+}
+
+static int scsiback_map(struct backend_info *be)
+{
+	struct xenbus_device *dev = be->dev;
+	unsigned long ring_ref;
+	unsigned int evtchn;
+	int err;
+	char name[TASK_COMM_LEN];
+
+	err = xenbus_gather(XBT_NIL, dev->otherend,
+			"ring-ref", "%lu", &ring_ref,
+			"event-channel", "%u", &evtchn, NULL);
+	if (err) {
+		xenbus_dev_fatal(dev, err, "reading %s ring", dev->otherend);
+		return err;
+	}
+
+	err = scsiback_init_sring(be->info, ring_ref, evtchn);
+	if (err)
+		return err;
+
+	err = __vscsiif_name(be, name);
+	if (err) {
+		xenbus_dev_error(dev, err, "get scsiback dev name");
+		return err;
+	}
+
+	be->info->kthread = kthread_run(scsiback_schedule, be->info, name);
+	if (IS_ERR(be->info->kthread)) {
+		err = PTR_ERR(be->info->kthread);
+		be->info->kthread = NULL;
+		xenbus_dev_error(be->dev, err, "start vscsiif");
+		return err;
+	}
+
+	return 0;
+}
+
+
+struct scsi_device *scsiback_get_scsi_device(struct ids_tuple *phy)
+{
+	struct Scsi_Host *shost;
+	struct scsi_device *sdev = NULL;
+
+	shost = scsi_host_lookup(phy->hst);
+	if (IS_ERR(shost)) {
+		printk(KERN_ERR "scsiback: host%d doesn't exist.\n",
+			phy->hst);
+		return NULL;
+	}
+	sdev   = scsi_device_lookup(shost, phy->chn, phy->tgt, phy->lun);
+	if (!sdev) {
+		printk(KERN_ERR "scsiback: %d:%d:%d:%d doesn't exist.\n",
+			phy->hst, phy->chn, phy->tgt, phy->lun);
+		scsi_host_put(shost);
+		return NULL;
+	}
+
+	scsi_host_put(shost);
+	return (sdev);
+}
+
+#define VSCSIBACK_OP_ADD_OR_DEL_LUN	1
+#define VSCSIBACK_OP_UPDATEDEV_STATE	2
+
+
+static void scsiback_do_lun_hotplug(struct backend_info *be, int op)
+{
+	int i, err = 0;
+	struct ids_tuple phy, vir;
+	int device_state;
+	char str[64], state_str[64];
+	char **dir;
+	unsigned int dir_n = 0;
+	struct xenbus_device *dev = be->dev;
+	struct scsi_device *sdev;
+
+	dir = xenbus_directory(XBT_NIL, dev->nodename, "vscsi-devs", &dir_n);
+	if (IS_ERR(dir))
+		return;
+
+	for (i = 0; i < dir_n; i++) {
+		
+		/* read status */
+		snprintf(state_str, sizeof(state_str), "vscsi-devs/%s/state", dir[i]);
+		err = xenbus_scanf(XBT_NIL, dev->nodename, state_str, "%u",
+			&device_state);
+		if (XENBUS_EXIST_ERR(err))
+			continue;
+
+		/* physical SCSI device */
+		snprintf(str, sizeof(str), "vscsi-devs/%s/p-dev", dir[i]);
+		err = xenbus_scanf(XBT_NIL, dev->nodename, str,
+			"%u:%u:%u:%u", &phy.hst, &phy.chn, &phy.tgt, &phy.lun);
+		if (XENBUS_EXIST_ERR(err)) {
+			xenbus_printf(XBT_NIL, dev->nodename, state_str,
+					"%d", XenbusStateClosed);
+			continue;
+		}
+
+		/* virtual SCSI device */
+		snprintf(str, sizeof(str), "vscsi-devs/%s/v-dev", dir[i]);
+		err = xenbus_scanf(XBT_NIL, dev->nodename, str,
+			"%u:%u:%u:%u", &vir.hst, &vir.chn, &vir.tgt, &vir.lun);
+		if (XENBUS_EXIST_ERR(err)) {
+			xenbus_printf(XBT_NIL, dev->nodename, state_str,
+					"%d", XenbusStateClosed);
+			continue;
+		}
+
+		switch (op) {
+		case VSCSIBACK_OP_ADD_OR_DEL_LUN:
+			if (device_state == XenbusStateInitialising) {
+				sdev = scsiback_get_scsi_device(&phy);
+				if (!sdev)
+					xenbus_printf(XBT_NIL, dev->nodename, state_str, 
+							    "%d", XenbusStateClosed);
+				else {
+					err = scsiback_add_translation_entry(be->info, sdev, &vir);
+					if (!err) {
+						if (xenbus_printf(XBT_NIL, dev->nodename, state_str, 
+								    "%d", XenbusStateInitialised)) {
+							printk(KERN_ERR "scsiback: xenbus_printf error %s\n", state_str);
+							scsiback_del_translation_entry(be->info, &vir);
+						}
+					} else {
+						scsi_device_put(sdev);
+						xenbus_printf(XBT_NIL, dev->nodename, state_str, 
+								    "%d", XenbusStateClosed);
+					}
+				}
+			}
+
+			if (device_state == XenbusStateClosing) {
+				if (!scsiback_del_translation_entry(be->info, &vir)) {
+					if (xenbus_printf(XBT_NIL, dev->nodename, state_str, 
+							    "%d", XenbusStateClosed))
+						printk(KERN_ERR "scsiback: xenbus_printf error %s\n", state_str);
+				}
+			}
+			break;
+
+		case VSCSIBACK_OP_UPDATEDEV_STATE:
+			if (device_state == XenbusStateInitialised) {
+				/* modify vscsi-devs/dev-x/state */
+				if (xenbus_printf(XBT_NIL, dev->nodename, state_str, 
+						    "%d", XenbusStateConnected)) {
+					printk(KERN_ERR "scsiback: xenbus_printf error %s\n", state_str);
+					scsiback_del_translation_entry(be->info, &vir);
+					xenbus_printf(XBT_NIL, dev->nodename, state_str, 
+							    "%d", XenbusStateClosed);
+				}
+			}
+			break;
+		/*When it is necessary, processing is added here.*/
+		default:
+			break;
+		}
+	}
+
+	kfree(dir);
+	return ;
+}
+
+
+static void scsiback_frontend_changed(struct xenbus_device *dev,
+					enum xenbus_state frontend_state)
+{
+	struct backend_info *be = dev->dev.driver_data;
+	int err;
+
+	switch (frontend_state) {
+	case XenbusStateInitialising:
+		break;
+	case XenbusStateInitialised:
+		err = scsiback_map(be);
+		if (err)
+			break;
+
+		scsiback_do_lun_hotplug(be, VSCSIBACK_OP_ADD_OR_DEL_LUN);
+		xenbus_switch_state(dev, XenbusStateConnected);
+
+		break;
+	case XenbusStateConnected:
+
+		scsiback_do_lun_hotplug(be, VSCSIBACK_OP_UPDATEDEV_STATE);
+
+		if (dev->state == XenbusStateConnected)
+			break;
+
+		xenbus_switch_state(dev, XenbusStateConnected);
+
+		break;
+
+	case XenbusStateClosing:
+		scsiback_disconnect(be->info);
+		xenbus_switch_state(dev, XenbusStateClosing);
+		break;
+
+	case XenbusStateClosed:
+		xenbus_switch_state(dev, XenbusStateClosed);
+		if (xenbus_dev_is_online(dev))
+			break;
+		/* fall through if not online */
+	case XenbusStateUnknown:
+		device_unregister(&dev->dev);
+		break;
+
+	case XenbusStateReconfiguring:
+		scsiback_do_lun_hotplug(be, VSCSIBACK_OP_ADD_OR_DEL_LUN);
+
+		xenbus_switch_state(dev, XenbusStateReconfigured);
+
+		break;
+
+	default:
+		xenbus_dev_fatal(dev, -EINVAL, "saw state %d at frontend",
+					frontend_state);
+		break;
+	}
+}
+
+
+static int scsiback_remove(struct xenbus_device *dev)
+{
+	struct backend_info *be = dev->dev.driver_data;
+
+	if (be->info) {
+		scsiback_disconnect(be->info);
+		scsiback_release_translation_entry(be->info);
+		scsiback_free(be->info);
+		be->info = NULL;
+	}
+
+	kfree(be);
+	dev->dev.driver_data = NULL;
+
+	return 0;
+}
+
+
+static int scsiback_probe(struct xenbus_device *dev,
+			   const struct xenbus_device_id *id)
+{
+	int err;
+
+	struct backend_info *be = kzalloc(sizeof(struct backend_info),
+					  GFP_KERNEL);
+
+	DPRINTK("%p %d\n", dev, dev->otherend_id);
+
+	if (!be) {
+		xenbus_dev_fatal(dev, -ENOMEM,
+				 "allocating backend structure");
+		return -ENOMEM;
+	}
+	be->dev = dev;
+	dev->dev.driver_data = be;
+
+	be->info = vscsibk_info_alloc(dev->otherend_id);
+	if (IS_ERR(be->info)) {
+		err = PTR_ERR(be->info);
+		be->info = NULL;
+		xenbus_dev_fatal(dev, err, "creating scsihost interface");
+		goto fail;
+	}
+
+	be->info->dev = dev;
+	be->info->irq = 0;
+
+	scsiback_init_translation_table(be->info);
+
+	err = xenbus_switch_state(dev, XenbusStateInitWait);
+	if (err)
+		goto fail;
+
+	return 0;
+
+
+fail:
+	printk(KERN_WARNING "scsiback: %s failed\n",__FUNCTION__);
+	scsiback_remove(dev);
+
+	return err;
+}
+
+
+static struct xenbus_device_id scsiback_ids[] = {
+	{ "vscsi" },
+	{ "" }
+};
+
+static struct xenbus_driver scsiback = {
+	.name			= "vscsi",
+	.owner			= THIS_MODULE,
+	.ids			= scsiback_ids,
+	.probe			= scsiback_probe,
+	.remove			= scsiback_remove,
+	.otherend_changed	= scsiback_frontend_changed
+};
+
+int scsiback_xenbus_init(void)
+{
+	return xenbus_register_backend(&scsiback);
+}
+
+void scsiback_xenbus_unregister(void)
+{
+	xenbus_unregister_driver(&scsiback);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/scsifront/common.h linux-2.6.18-xen-3.3.0/drivers/xen/scsifront/common.h
--- linux-2.6.18.8/drivers/xen/scsifront/common.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/scsifront/common.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,129 @@
+/*
+ * Xen SCSI frontend driver
+ *
+ * Copyright (c) 2008, FUJITSU Limited
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_DRIVERS_SCSIFRONT_H__
+#define __XEN_DRIVERS_SCSIFRONT_H__
+
+#include <linux/version.h>
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/device.h>
+#include <linux/kthread.h>
+#include <linux/wait.h>
+#include <linux/interrupt.h>
+#include <linux/spinlock.h>
+#include <linux/sched.h>
+#include <linux/blkdev.h>
+#include <scsi/scsi_cmnd.h>
+#include <scsi/scsi_device.h>
+#include <scsi/scsi.h>
+#include <scsi/scsi_host.h>
+#include <xen/xenbus.h>
+#include <xen/gnttab.h>
+#include <xen/evtchn.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/io/ring.h>
+#include <xen/interface/io/vscsiif.h>
+#include <asm/delay.h>
+
+
+#define GRANT_INVALID_REF	0
+#define VSCSI_IN_ABORT		1
+#define VSCSI_IN_RESET		2
+
+/* tuning point*/
+#define VSCSIIF_DEFAULT_CMD_PER_LUN 10
+#define VSCSIIF_MAX_TARGET          64
+#define VSCSIIF_MAX_LUN             255
+
+#define VSCSIIF_RING_SIZE	\
+    __RING_SIZE((struct vscsiif_sring *)0, PAGE_SIZE)
+#define VSCSIIF_MAX_REQS	VSCSIIF_RING_SIZE
+
+struct vscsifrnt_shadow {
+	uint16_t next_free;
+	
+	/* command between backend and frontend
+	 * VSCSIIF_ACT_SCSI_CDB or VSCSIIF_ACT_SCSI_RESET */
+	unsigned char act;
+	
+	/* do reset function */
+	wait_queue_head_t wq_reset;	/* reset work queue           */
+	int wait_reset;			/* reset work queue condition */
+	int32_t rslt_reset;		/* reset response status      */
+					/* (SUCESS or FAILED)         */
+
+	/* for DMA_TO_DEVICE(1), DMA_FROM_DEVICE(2), DMA_NONE(3) 
+	   requests */
+	unsigned int sc_data_direction;
+	
+	/* Number of pieces of scatter-gather */
+	unsigned int nr_segments;
+
+	/* requested struct scsi_cmnd is stored from kernel */
+	unsigned long req_scsi_cmnd;
+	int gref[VSCSIIF_SG_TABLESIZE];
+};
+
+struct vscsifrnt_info {
+	struct xenbus_device *dev;
+
+	struct Scsi_Host *host;
+
+	spinlock_t io_lock;
+	spinlock_t shadow_lock;
+	unsigned int evtchn;
+	unsigned int irq;
+
+	grant_ref_t ring_ref;
+	struct vscsiif_front_ring ring;
+	struct vscsiif_response	ring_res;
+
+	struct vscsifrnt_shadow shadow[VSCSIIF_MAX_REQS];
+	uint32_t shadow_free;
+
+	struct task_struct *kthread;
+	wait_queue_head_t wq;
+	unsigned int waiting_resp;
+
+};
+
+#define DPRINTK(_f, _a...)				\
+	pr_debug("(file=%s, line=%d) " _f,	\
+		 __FILE__ , __LINE__ , ## _a )
+
+int scsifront_xenbus_init(void);
+void scsifront_xenbus_unregister(void);
+int scsifront_schedule(void *data);
+irqreturn_t scsifront_intr(int irq, void *dev_id, struct pt_regs *ptregs);
+int scsifront_cmd_done(struct vscsifrnt_info *info);
+
+
+#endif /* __XEN_DRIVERS_SCSIFRONT_H__  */
diff -rpuN linux-2.6.18.8/drivers/xen/scsifront/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/scsifront/Makefile
--- linux-2.6.18.8/drivers/xen/scsifront/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/scsifront/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,3 @@
+
+obj-$(CONFIG_XEN_SCSI_FRONTEND)	:= xenscsi.o
+xenscsi-objs := scsifront.o xenbus.o
diff -rpuN linux-2.6.18.8/drivers/xen/scsifront/scsifront.c linux-2.6.18-xen-3.3.0/drivers/xen/scsifront/scsifront.c
--- linux-2.6.18.8/drivers/xen/scsifront/scsifront.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/scsifront/scsifront.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,511 @@
+/*
+ * Xen SCSI frontend driver
+ *
+ * Copyright (c) 2008, FUJITSU Limited
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+ 
+
+#include <linux/version.h>
+#include "common.h"
+
+static int get_id_from_freelist(struct vscsifrnt_info *info)
+{
+	unsigned long flags;
+	uint32_t free;
+
+	spin_lock_irqsave(&info->shadow_lock, flags);
+
+	free = info->shadow_free;
+	BUG_ON(free > VSCSIIF_MAX_REQS);
+	info->shadow_free = info->shadow[free].next_free;
+	info->shadow[free].next_free = 0x0fff;
+
+	info->shadow[free].wait_reset = 0;
+
+	spin_unlock_irqrestore(&info->shadow_lock, flags);
+
+	return free;
+}
+
+static void add_id_to_freelist(struct vscsifrnt_info *info, uint32_t id)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&info->shadow_lock, flags);
+
+	info->shadow[id].next_free  = info->shadow_free;
+	info->shadow[id].req_scsi_cmnd = 0;
+	info->shadow_free = id;
+
+	spin_unlock_irqrestore(&info->shadow_lock, flags);
+}
+
+
+struct vscsiif_request * scsifront_pre_request(struct vscsifrnt_info *info)
+{
+	struct vscsiif_front_ring *ring = &(info->ring);
+	vscsiif_request_t *ring_req;
+	uint32_t id;
+
+	ring_req = RING_GET_REQUEST(&(info->ring), ring->req_prod_pvt);
+
+	ring->req_prod_pvt++;
+	
+	id = get_id_from_freelist(info);	/* use id by response */
+	ring_req->rqid = (uint16_t)id;
+
+	return ring_req;
+}
+
+
+static void scsifront_notify_work(struct vscsifrnt_info *info)
+{
+	info->waiting_resp = 1;
+	wake_up(&info->wq);
+}
+
+
+static void scsifront_do_request(struct vscsifrnt_info *info)
+{
+	struct vscsiif_front_ring *ring = &(info->ring);
+	unsigned int irq = info->irq;
+	int notify;
+
+	RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(ring, notify);
+	if (notify)
+		notify_remote_via_irq(irq);
+}
+
+irqreturn_t scsifront_intr(int irq, void *dev_id, struct pt_regs *ptregs)
+{
+	scsifront_notify_work((struct vscsifrnt_info *)dev_id);
+	return IRQ_HANDLED;
+}
+
+
+static void scsifront_gnttab_done(struct vscsifrnt_shadow *s, uint32_t id)
+{
+	int i;
+
+	if (s->sc_data_direction == DMA_NONE)
+		return;
+
+	if (s->nr_segments) {
+		for (i = 0; i < s->nr_segments; i++) {
+			if (unlikely(gnttab_query_foreign_access(
+				s->gref[i]) != 0)) {
+				printk(KERN_ALERT "scsifront: "
+					"grant still in use by backend.\n");
+				BUG();
+			}
+			gnttab_end_foreign_access(s->gref[i], 0UL);
+		}
+	}
+
+	return;
+}
+
+
+static void scsifront_cdb_cmd_done(struct vscsifrnt_info *info,
+		       vscsiif_response_t *ring_res)
+{
+	struct scsi_cmnd *sc;
+	uint32_t id;
+	uint8_t sense_len;
+
+	id = ring_res->rqid;
+	sc = (struct scsi_cmnd *)info->shadow[id].req_scsi_cmnd;
+
+	if (sc == NULL)
+		BUG();
+
+	scsifront_gnttab_done(&info->shadow[id], id);
+	add_id_to_freelist(info, id);
+
+	sc->result = ring_res->rslt;
+	sc->resid  = ring_res->residual_len;
+
+	if (ring_res->sense_len > VSCSIIF_SENSE_BUFFERSIZE)
+		sense_len = VSCSIIF_SENSE_BUFFERSIZE;
+	else
+		sense_len = ring_res->sense_len;
+
+	if (sense_len)
+		memcpy(sc->sense_buffer, ring_res->sense_buffer, sense_len);
+
+	sc->scsi_done(sc);
+
+	return;
+}
+
+
+static void scsifront_sync_cmd_done(struct vscsifrnt_info *info,
+				vscsiif_response_t *ring_res)
+{
+	uint16_t id = ring_res->rqid;
+	unsigned long flags;
+	
+	spin_lock_irqsave(&info->shadow_lock, flags);
+	info->shadow[id].wait_reset = 1;
+	info->shadow[id].rslt_reset = ring_res->rslt;
+	spin_unlock_irqrestore(&info->shadow_lock, flags);
+
+	wake_up(&(info->shadow[id].wq_reset));
+}
+
+
+int scsifront_cmd_done(struct vscsifrnt_info *info)
+{
+	vscsiif_response_t *ring_res;
+
+	RING_IDX i, rp;
+	int more_to_do = 0;
+	unsigned long flags;
+
+	spin_lock_irqsave(&info->io_lock, flags);
+
+	rp = info->ring.sring->rsp_prod;
+	rmb();
+	for (i = info->ring.rsp_cons; i != rp; i++) {
+		
+		ring_res = RING_GET_RESPONSE(&info->ring, i);
+
+		if (info->shadow[ring_res->rqid].act == VSCSIIF_ACT_SCSI_CDB)
+			scsifront_cdb_cmd_done(info, ring_res);
+		else
+			scsifront_sync_cmd_done(info, ring_res);
+	}
+
+	info->ring.rsp_cons = i;
+
+	if (i != info->ring.req_prod_pvt) {
+		RING_FINAL_CHECK_FOR_RESPONSES(&info->ring, more_to_do);
+	} else {
+		info->ring.sring->rsp_event = i + 1;
+	}
+
+	spin_unlock_irqrestore(&info->io_lock, flags);
+
+
+	/* Yield point for this unbounded loop. */
+	cond_resched();
+
+	return more_to_do;
+}
+
+
+
+
+int scsifront_schedule(void *data)
+{
+	struct vscsifrnt_info *info = (struct vscsifrnt_info *)data;
+
+	while (!kthread_should_stop()) {
+		wait_event_interruptible(
+			info->wq,
+			info->waiting_resp || kthread_should_stop());
+
+		info->waiting_resp = 0;
+		smp_mb();
+
+		if (scsifront_cmd_done(info))
+			info->waiting_resp = 1;
+	}
+
+	return 0;
+}
+
+
+
+static int map_data_for_request(struct vscsifrnt_info *info,
+		struct scsi_cmnd *sc, vscsiif_request_t *ring_req, uint32_t id)
+{
+	grant_ref_t gref_head;
+	struct page *page;
+	int err, i, ref, ref_cnt = 0;
+	int write = (sc->sc_data_direction == DMA_TO_DEVICE);
+	int nr_pages, off, len, bytes;
+	unsigned long buffer_pfn;
+	unsigned int data_len = 0;
+
+	if (sc->sc_data_direction == DMA_NONE)
+		return 0;
+
+	err = gnttab_alloc_grant_references(VSCSIIF_SG_TABLESIZE, &gref_head);
+	if (err) {
+		printk(KERN_ERR "scsifront: gnttab_alloc_grant_references() error\n");
+		return -ENOMEM;
+	}
+
+	if (sc->use_sg) {
+		/* quoted scsi_lib.c/scsi_req_map_sg . */
+		struct scatterlist *sg = (struct scatterlist *)sc->request_buffer;
+		nr_pages = (sc->request_bufflen + sg[0].offset + PAGE_SIZE - 1) >> PAGE_SHIFT;
+
+		if (nr_pages > VSCSIIF_SG_TABLESIZE) {
+			printk(KERN_ERR "scsifront: Unable to map request_buffer for command!\n");
+			ref_cnt = (-E2BIG);
+			goto big_to_sg;
+		}
+
+		for (i = 0; i < sc->use_sg; i++) {
+			page = sg[i].page;
+			off = sg[i].offset;
+			len = sg[i].length;
+			data_len += len;
+
+			buffer_pfn = page_to_phys(page) >> PAGE_SHIFT;
+
+			while (len > 0) {
+				bytes = min_t(unsigned int, len, PAGE_SIZE - off);
+				
+				ref = gnttab_claim_grant_reference(&gref_head);
+				BUG_ON(ref == -ENOSPC);
+
+				gnttab_grant_foreign_access_ref(ref, info->dev->otherend_id,
+					buffer_pfn, write);
+
+				info->shadow[id].gref[ref_cnt]  = ref;
+				ring_req->seg[ref_cnt].gref     = ref;
+				ring_req->seg[ref_cnt].offset   = (uint16_t)off;
+				ring_req->seg[ref_cnt].length   = (uint16_t)bytes;
+
+				buffer_pfn++;
+				len -= bytes;
+				off = 0;
+				ref_cnt++;
+			}
+		}
+	} else if (sc->request_bufflen) {
+		unsigned long end   = ((unsigned long)sc->request_buffer
+					+ sc->request_bufflen + PAGE_SIZE - 1) >> PAGE_SHIFT;
+		unsigned long start = (unsigned long)sc->request_buffer >> PAGE_SHIFT;
+
+		page = virt_to_page(sc->request_buffer);
+		nr_pages = end - start;
+		len = sc->request_bufflen;
+
+		if (nr_pages > VSCSIIF_SG_TABLESIZE) {
+			ref_cnt = (-E2BIG);
+			goto big_to_sg;
+		}
+
+		buffer_pfn = page_to_phys(page) >> PAGE_SHIFT;
+
+		off = offset_in_page((unsigned long)sc->request_buffer);
+		for (i = 0; i < nr_pages; i++) {
+			bytes = PAGE_SIZE - off;
+
+			if (bytes > len)
+				bytes = len;
+
+			ref = gnttab_claim_grant_reference(&gref_head);
+			BUG_ON(ref == -ENOSPC);
+
+			gnttab_grant_foreign_access_ref(ref, info->dev->otherend_id,
+				buffer_pfn, write);
+
+			info->shadow[id].gref[i] = ref;
+			ring_req->seg[i].gref     = ref;
+			ring_req->seg[i].offset   = (uint16_t)off;
+			ring_req->seg[i].length   = (uint16_t)bytes;
+
+			buffer_pfn++;
+			len -= bytes;
+			off = 0;
+			ref_cnt++;
+		}
+	}
+
+big_to_sg:
+
+	gnttab_free_grant_references(gref_head);
+
+	return ref_cnt;
+}
+
+static int scsifront_queuecommand(struct scsi_cmnd *sc,
+				  void (*done)(struct scsi_cmnd *))
+{
+	struct vscsifrnt_info *info =
+		(struct vscsifrnt_info *) sc->device->host->hostdata;
+	vscsiif_request_t *ring_req;
+	int ref_cnt;
+	uint16_t rqid;
+
+	if (RING_FULL(&info->ring)) {
+		goto out_host_busy;
+	}
+
+	sc->scsi_done = done;
+	sc->result    = 0;
+
+	ring_req          = scsifront_pre_request(info);
+	rqid              = ring_req->rqid;
+	ring_req->act     = VSCSIIF_ACT_SCSI_CDB;
+
+	ring_req->id      = sc->device->id;
+	ring_req->lun     = sc->device->lun;
+	ring_req->channel = sc->device->channel;
+	ring_req->cmd_len = sc->cmd_len;
+
+	BUG_ON(sc->cmd_len > VSCSIIF_MAX_COMMAND_SIZE);
+
+	if ( sc->cmd_len )
+		memcpy(ring_req->cmnd, sc->cmnd, sc->cmd_len);
+	else
+		memset(ring_req->cmnd, 0, VSCSIIF_MAX_COMMAND_SIZE);
+
+	ring_req->sc_data_direction   = (uint8_t)sc->sc_data_direction;
+	ring_req->timeout_per_command = (sc->timeout_per_command / HZ);
+
+	info->shadow[rqid].req_scsi_cmnd     = (unsigned long)sc;
+	info->shadow[rqid].sc_data_direction = sc->sc_data_direction;
+	info->shadow[rqid].act               = ring_req->act;
+
+	ref_cnt = map_data_for_request(info, sc, ring_req, rqid);
+	if (ref_cnt < 0) {
+		add_id_to_freelist(info, rqid);
+		if (ref_cnt == (-ENOMEM))
+			goto out_host_busy;
+		else {
+			sc->result = (DID_ERROR << 16);
+			goto out_fail_command;
+		}
+	}
+
+	ring_req->nr_segments          = (uint8_t)ref_cnt;
+	info->shadow[rqid].nr_segments = ref_cnt;
+
+	scsifront_do_request(info);
+
+	return 0;
+
+out_host_busy:
+	return SCSI_MLQUEUE_HOST_BUSY;
+
+out_fail_command:
+	done(sc);
+	return 0;
+}
+
+
+static int scsifront_eh_abort_handler(struct scsi_cmnd *sc)
+{
+	return (FAILED);
+}
+
+/* vscsi supports only device_reset, because it is each of LUNs */
+static int scsifront_dev_reset_handler(struct scsi_cmnd *sc)
+{
+	struct Scsi_Host *host = sc->device->host;
+	struct vscsifrnt_info *info =
+		(struct vscsifrnt_info *) sc->device->host->hostdata;
+
+	vscsiif_request_t *ring_req;
+	uint16_t rqid;
+	int err;
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,12)
+	spin_lock_irq(host->host_lock);
+#endif
+
+	ring_req      = scsifront_pre_request(info);
+	ring_req->act = VSCSIIF_ACT_SCSI_RESET;
+
+	rqid          = ring_req->rqid;
+	info->shadow[rqid].act = VSCSIIF_ACT_SCSI_RESET;
+
+	ring_req->channel = sc->device->channel;
+	ring_req->id      = sc->device->id;
+	ring_req->lun     = sc->device->lun;
+	ring_req->cmd_len = sc->cmd_len;
+
+	if ( sc->cmd_len )
+		memcpy(ring_req->cmnd, sc->cmnd, sc->cmd_len);
+	else
+		memset(ring_req->cmnd, 0, VSCSIIF_MAX_COMMAND_SIZE);
+
+	ring_req->sc_data_direction   = (uint8_t)sc->sc_data_direction;
+	ring_req->timeout_per_command = (sc->timeout_per_command / HZ);
+	ring_req->nr_segments         = 0;
+
+	scsifront_do_request(info);	
+
+	spin_unlock_irq(host->host_lock);
+	wait_event_interruptible(info->shadow[rqid].wq_reset,
+			 info->shadow[rqid].wait_reset);
+	spin_lock_irq(host->host_lock);
+
+	err = info->shadow[rqid].rslt_reset;
+
+	add_id_to_freelist(info, rqid);
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,12)
+	spin_unlock_irq(host->host_lock);
+#endif
+	return (err);
+}
+
+
+struct scsi_host_template scsifront_sht = {
+	.module			= THIS_MODULE,
+	.name			= "Xen SCSI frontend driver",
+	.queuecommand		= scsifront_queuecommand,
+	.eh_abort_handler	= scsifront_eh_abort_handler,
+	.eh_device_reset_handler= scsifront_dev_reset_handler,
+	.cmd_per_lun		= VSCSIIF_DEFAULT_CMD_PER_LUN,
+	.can_queue		= VSCSIIF_MAX_REQS,
+	.this_id 		= -1,
+	.sg_tablesize		= VSCSIIF_SG_TABLESIZE,
+	.use_clustering		= DISABLE_CLUSTERING,
+	.proc_name		= "scsifront",
+};
+
+
+static int __init scsifront_init(void)
+{
+	int err;
+
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	err = scsifront_xenbus_init();
+
+	return err;
+}
+
+static void __exit scsifront_exit(void)
+{
+	scsifront_xenbus_unregister();
+}
+
+module_init(scsifront_init);
+module_exit(scsifront_exit);
+
+MODULE_DESCRIPTION("Xen SCSI frontend driver");
+MODULE_LICENSE("GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/scsifront/xenbus.c linux-2.6.18-xen-3.3.0/drivers/xen/scsifront/xenbus.c
--- linux-2.6.18.8/drivers/xen/scsifront/xenbus.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/scsifront/xenbus.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,421 @@
+/*
+ * Xen SCSI frontend driver
+ *
+ * Copyright (c) 2008, FUJITSU Limited
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+ 
+
+#include <linux/version.h>
+#include "common.h"
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,11)
+  #define DEFAULT_TASK_COMM_LEN	16
+#else
+  #define DEFAULT_TASK_COMM_LEN	TASK_COMM_LEN
+#endif
+
+extern struct scsi_host_template scsifront_sht;
+
+static void scsifront_free(struct vscsifrnt_info *info)
+{
+	struct Scsi_Host *host = info->host;
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,14)
+	if (host->shost_state != SHOST_DEL) {
+#else
+	if (!test_bit(SHOST_DEL, &host->shost_state)) {
+#endif
+		scsi_remove_host(info->host);
+	}
+
+	if (info->ring_ref != GRANT_INVALID_REF) {
+		gnttab_end_foreign_access(info->ring_ref,
+					(unsigned long)info->ring.sring);
+		info->ring_ref = GRANT_INVALID_REF;
+		info->ring.sring = NULL;
+	}
+
+	if (info->irq)
+		unbind_from_irqhandler(info->irq, info);
+	info->irq = 0;
+
+	scsi_host_put(info->host);
+}
+
+
+static int scsifront_alloc_ring(struct vscsifrnt_info *info)
+{
+	struct xenbus_device *dev = info->dev;
+	struct vscsiif_sring *sring;
+	int err = -ENOMEM;
+
+
+	info->ring_ref = GRANT_INVALID_REF;
+
+	/***** Frontend to Backend ring start *****/
+	sring = (struct vscsiif_sring *) __get_free_page(GFP_KERNEL);
+	if (!sring) {
+		xenbus_dev_fatal(dev, err, "fail to allocate shared ring (Front to Back)");
+		return err;
+	}
+	SHARED_RING_INIT(sring);
+	FRONT_RING_INIT(&info->ring, sring, PAGE_SIZE);
+
+	err = xenbus_grant_ring(dev, virt_to_mfn(sring));
+	if (err < 0) {
+		free_page((unsigned long) sring);
+		info->ring.sring = NULL;
+		xenbus_dev_fatal(dev, err, "fail to grant shared ring (Front to Back)");
+		goto free_sring;
+	}
+	info->ring_ref = err;
+
+	err = bind_listening_port_to_irqhandler(
+			dev->otherend_id, scsifront_intr,
+			SA_SAMPLE_RANDOM, "scsifront", info);
+
+	if (err <= 0) {
+		xenbus_dev_fatal(dev, err, "bind_listening_port_to_irqhandler");
+		goto free_sring;
+	}
+	info->irq = err;
+
+	return 0;
+
+/* free resource */
+free_sring:
+	scsifront_free(info);
+
+	return err;
+}
+
+
+static int scsifront_init_ring(struct vscsifrnt_info *info)
+{
+	struct xenbus_device *dev = info->dev;
+	struct xenbus_transaction xbt;
+	int err;
+
+	DPRINTK("%s\n",__FUNCTION__);
+
+	err = scsifront_alloc_ring(info);
+	if (err)
+		return err;
+	DPRINTK("%u %u\n", info->ring_ref, info->evtchn);
+
+again:
+	err = xenbus_transaction_start(&xbt);
+	if (err) {
+		xenbus_dev_fatal(dev, err, "starting transaction");
+	}
+
+	err = xenbus_printf(xbt, dev->nodename, "ring-ref", "%u",
+				info->ring_ref);
+	if (err) {
+		xenbus_dev_fatal(dev, err, "%s", "writing ring-ref");
+		goto fail;
+	}
+
+	err = xenbus_printf(xbt, dev->nodename, "event-channel", "%u",
+				irq_to_evtchn_port(info->irq));
+
+	if (err) {
+		xenbus_dev_fatal(dev, err, "%s", "writing event-channel");
+		goto fail;
+	}
+
+	err = xenbus_transaction_end(xbt, 0);
+	if (err) {
+		if (err == -EAGAIN)
+			goto again;
+		xenbus_dev_fatal(dev, err, "completing transaction");
+		goto free_sring;
+	}
+
+	return 0;
+
+fail:
+	xenbus_transaction_end(xbt, 1);
+free_sring:
+	/* free resource */
+	scsifront_free(info);
+	
+	return err;
+}
+
+
+static int scsifront_probe(struct xenbus_device *dev,
+				const struct xenbus_device_id *id)
+{
+	struct vscsifrnt_info *info;
+	struct Scsi_Host *host;
+	int i, err = -ENOMEM;
+	char name[DEFAULT_TASK_COMM_LEN];
+
+	host = scsi_host_alloc(&scsifront_sht, sizeof(*info));
+	if (!host) {
+		xenbus_dev_fatal(dev, err, "fail to allocate scsi host");
+		return err;
+	}
+	info = (struct vscsifrnt_info *) host->hostdata;
+	info->host = host;
+
+
+	dev->dev.driver_data = info;
+	info->dev  = dev;
+
+	for (i = 0; i < VSCSIIF_MAX_REQS; i++) {
+		info->shadow[i].next_free = i + 1;
+		init_waitqueue_head(&(info->shadow[i].wq_reset));
+		info->shadow[i].wait_reset = 0;
+	}
+	info->shadow[VSCSIIF_MAX_REQS - 1].next_free = 0x0fff;
+
+	err = scsifront_init_ring(info);
+	if (err) {
+		scsi_host_put(host);
+		return err;
+	}
+
+	init_waitqueue_head(&info->wq);
+	spin_lock_init(&info->io_lock);
+	spin_lock_init(&info->shadow_lock);
+
+	snprintf(name, DEFAULT_TASK_COMM_LEN, "vscsiif.%d", info->host->host_no);
+
+	info->kthread = kthread_run(scsifront_schedule, info, name);
+	if (IS_ERR(info->kthread)) {
+		err = PTR_ERR(info->kthread);
+		info->kthread = NULL;
+		printk(KERN_ERR "scsifront: kthread start err %d\n", err);
+		goto free_sring;
+	}
+
+	host->max_id      = VSCSIIF_MAX_TARGET;
+	host->max_channel = 0;
+	host->max_lun     = VSCSIIF_MAX_LUN;
+	host->max_sectors = (VSCSIIF_SG_TABLESIZE - 1) * PAGE_SIZE / 512;
+
+	err = scsi_add_host(host, &dev->dev);
+	if (err) {
+		printk(KERN_ERR "scsifront: fail to add scsi host %d\n", err);
+		goto free_sring;
+	}
+
+	xenbus_switch_state(dev, XenbusStateInitialised);
+
+	return 0;
+
+free_sring:
+	/* free resource */
+	scsifront_free(info);
+	return err;
+}
+
+static int scsifront_remove(struct xenbus_device *dev)
+{
+	struct vscsifrnt_info *info = dev->dev.driver_data;
+
+	DPRINTK("%s: %s removed\n",__FUNCTION__ ,dev->nodename);
+
+	if (info->kthread) {
+		kthread_stop(info->kthread);
+		info->kthread = NULL;
+	}
+
+	scsifront_free(info);
+	
+	return 0;
+}
+
+
+static int scsifront_disconnect(struct vscsifrnt_info *info)
+{
+	struct xenbus_device *dev = info->dev;
+	struct Scsi_Host *host = info->host;
+
+	DPRINTK("%s: %s disconnect\n",__FUNCTION__ ,dev->nodename);
+
+	/* 
+	  When this function is executed,  all devices of 
+	  Frontend have been deleted. 
+	  Therefore, it need not block I/O before remove_host.
+	*/
+
+	scsi_remove_host(host);
+	xenbus_frontend_closed(dev);
+
+	return 0;
+}
+
+#define VSCSIFRONT_OP_ADD_LUN	1
+#define VSCSIFRONT_OP_DEL_LUN	2
+
+static void scsifront_do_lun_hotplug(struct vscsifrnt_info *info, int op)
+{
+	struct xenbus_device *dev = info->dev;
+	int i, err = 0;
+	char str[64], state_str[64];
+	char **dir;
+	unsigned int dir_n = 0;
+	unsigned int device_state;
+	unsigned int hst, chn, tgt, lun;
+	struct scsi_device *sdev;
+
+	dir = xenbus_directory(XBT_NIL, dev->otherend, "vscsi-devs", &dir_n);
+	if (IS_ERR(dir))
+		return;
+
+	for (i = 0; i < dir_n; i++) {
+		/* read status */
+		snprintf(str, sizeof(str), "vscsi-devs/%s/state", dir[i]);
+		err = xenbus_scanf(XBT_NIL, dev->otherend, str, "%u",
+			&device_state);
+		if (XENBUS_EXIST_ERR(err))
+			continue;
+		
+		/* virtual SCSI device */
+		snprintf(str, sizeof(str), "vscsi-devs/%s/v-dev", dir[i]);
+		err = xenbus_scanf(XBT_NIL, dev->otherend, str,
+			"%u:%u:%u:%u", &hst, &chn, &tgt, &lun);
+		if (XENBUS_EXIST_ERR(err))
+			continue;
+
+		/* front device state path */
+		snprintf(state_str, sizeof(state_str), "vscsi-devs/%s/state", dir[i]);
+
+		switch (op) {
+		case VSCSIFRONT_OP_ADD_LUN:
+			if (device_state == XenbusStateInitialised) {
+				sdev = scsi_device_lookup(info->host, chn, tgt, lun);
+				if (sdev) {
+					printk(KERN_ERR "scsifront: Device already in use.\n");
+					scsi_device_put(sdev);
+					xenbus_printf(XBT_NIL, dev->nodename,
+						state_str, "%d", XenbusStateClosed);
+				} else {
+					scsi_add_device(info->host, chn, tgt, lun);
+					xenbus_printf(XBT_NIL, dev->nodename,
+						state_str, "%d", XenbusStateConnected);
+				}
+			}
+			break;
+		case VSCSIFRONT_OP_DEL_LUN:
+			if (device_state == XenbusStateClosing) {
+				sdev = scsi_device_lookup(info->host, chn, tgt, lun);
+				if (sdev) {
+					scsi_remove_device(sdev);
+					scsi_device_put(sdev);
+					xenbus_printf(XBT_NIL, dev->nodename,
+						state_str, "%d", XenbusStateClosed);
+				}
+			}
+			break;
+		default:
+			break;
+		}
+	}
+	
+	kfree(dir);
+	return;
+}
+
+
+
+
+static void scsifront_backend_changed(struct xenbus_device *dev,
+				enum xenbus_state backend_state)
+{
+	struct vscsifrnt_info *info = dev->dev.driver_data;
+
+	DPRINTK("%p %u %u\n", dev, dev->state, backend_state);
+
+	switch (backend_state) {
+	case XenbusStateUnknown:
+	case XenbusStateInitialising:
+	case XenbusStateInitWait:
+	case XenbusStateClosed:
+		break;
+
+	case XenbusStateInitialised:
+		break;
+
+	case XenbusStateConnected:
+		if (xenbus_read_driver_state(dev->nodename) ==
+			XenbusStateInitialised) {
+			scsifront_do_lun_hotplug(info, VSCSIFRONT_OP_ADD_LUN);
+		}
+		
+		if (dev->state == XenbusStateConnected)
+			break;
+			
+		xenbus_switch_state(dev, XenbusStateConnected);
+		break;
+
+	case XenbusStateClosing:
+		scsifront_disconnect(info);
+		break;
+
+	case XenbusStateReconfiguring:
+		scsifront_do_lun_hotplug(info, VSCSIFRONT_OP_DEL_LUN);
+		xenbus_switch_state(dev, XenbusStateReconfiguring);
+		break;
+
+	case XenbusStateReconfigured:
+		scsifront_do_lun_hotplug(info, VSCSIFRONT_OP_ADD_LUN);
+		xenbus_switch_state(dev, XenbusStateConnected);
+		break;
+	}
+}
+
+
+static struct xenbus_device_id scsifront_ids[] = {
+	{ "vscsi" },
+	{ "" }
+};
+
+
+static struct xenbus_driver scsifront_driver = {
+	.name			= "vscsi",
+	.owner			= THIS_MODULE,
+	.ids			= scsifront_ids,
+	.probe			= scsifront_probe,
+	.remove			= scsifront_remove,
+/* 	.resume			= scsifront_resume, */
+	.otherend_changed	= scsifront_backend_changed,
+};
+
+int scsifront_xenbus_init(void)
+{
+	return xenbus_register_frontend(&scsifront_driver);
+}
+
+void scsifront_xenbus_unregister(void)
+{
+	xenbus_unregister_driver(&scsifront_driver);
+}
+
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/accel.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/accel.c
--- linux-2.6.18.8/drivers/xen/sfc_netback/accel.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/accel.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,129 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include "accel.h"
+#include "accel_msg_iface.h"
+#include "accel_solarflare.h"
+
+#include <linux/notifier.h>
+
+#ifdef EFX_GCOV
+#include "gcov.h"
+#endif
+
+static int netback_accel_netdev_event(struct notifier_block *nb,
+				      unsigned long event, void *ptr)
+{
+	struct net_device *net_dev = (struct net_device *)ptr;
+	struct netback_accel *bend;
+
+	if ((event == NETDEV_UP) || (event == NETDEV_DOWN)) {
+		mutex_lock(&bend_list_mutex);
+		bend = bend_list;
+		while (bend != NULL) {
+			mutex_lock(&bend->bend_mutex);
+			/*
+			 * This happens when the shared pages have
+			 * been unmapped, but the bend not yet removed
+			 * from list
+			 */
+			if (bend->shared_page == NULL)
+				goto next;
+
+			if (bend->net_dev->ifindex == net_dev->ifindex)
+				netback_accel_set_interface_state
+					(bend, event == NETDEV_UP);
+
+		next:
+			mutex_unlock(&bend->bend_mutex);
+			bend = bend->next_bend;
+		}
+		mutex_unlock(&bend_list_mutex);
+	}
+
+	return NOTIFY_DONE;
+}
+
+
+static struct notifier_block netback_accel_netdev_notifier = {
+	.notifier_call = netback_accel_netdev_event,
+};
+
+
+unsigned sfc_netback_max_pages = NETBACK_ACCEL_DEFAULT_MAX_BUF_PAGES;
+module_param_named(max_pages, sfc_netback_max_pages, uint, 0644);
+MODULE_PARM_DESC(max_pages, 
+		 "The number of buffer pages to enforce on each guest");
+
+/* Initialise subsystems need for the accelerated fast path */
+static int __init netback_accel_init(void)
+{
+	int rc = 0;
+
+#ifdef EFX_GCOV
+	gcov_provider_init(THIS_MODULE);
+#endif
+
+	rc = netback_accel_init_fwd();
+
+	if (rc == 0)
+		netback_accel_debugfs_init();
+
+	if (rc == 0)
+		rc = netback_accel_sf_init();
+
+	if (rc == 0)
+		rc = register_netdevice_notifier
+			(&netback_accel_netdev_notifier);
+
+	/*
+	 * What if no device was found, shouldn't we clean up stuff
+	 * we've allocated for acceleration subsystem?
+	 */
+
+	return rc;
+}
+
+module_init(netback_accel_init);
+
+static void __exit netback_accel_exit(void)
+{
+	unregister_netdevice_notifier(&netback_accel_netdev_notifier);
+
+	netback_accel_sf_shutdown();
+
+	netback_accel_shutdown_bends();
+
+	netback_accel_debugfs_fini();
+
+	netback_accel_shutdown_fwd();
+
+#ifdef EFX_GCOV
+	gcov_provider_fini(THIS_MODULE);
+#endif
+}
+
+module_exit(netback_accel_exit);
+
+MODULE_LICENSE("GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/accel_debugfs.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/accel_debugfs.c
--- linux-2.6.18.8/drivers/xen/sfc_netback/accel_debugfs.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/accel_debugfs.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,148 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/fs.h>
+#include <linux/debugfs.h>
+
+#include "accel.h"
+
+#if defined(CONFIG_DEBUG_FS)
+static struct dentry *sfc_debugfs_root = NULL;
+#endif
+
+#if NETBACK_ACCEL_STATS
+struct netback_accel_global_stats global_stats;
+#if defined(CONFIG_DEBUG_FS)
+static struct netback_accel_global_dbfs  global_dbfs;
+#endif
+#endif
+
+void netback_accel_debugfs_init(void) 
+{
+#if defined(CONFIG_DEBUG_FS)
+	sfc_debugfs_root = debugfs_create_dir("sfc_netback", NULL);
+	if (sfc_debugfs_root == NULL)
+		return;
+
+	global_dbfs.num_fwds = debugfs_create_u32
+		("num_fwds", S_IRUSR | S_IRGRP | S_IROTH,
+		 sfc_debugfs_root, &global_stats.num_fwds);
+	global_dbfs.dl_tx_packets = debugfs_create_u64
+		("dl_tx_packets", S_IRUSR | S_IRGRP | S_IROTH,
+		 sfc_debugfs_root, &global_stats.dl_tx_packets);
+	global_dbfs.dl_rx_packets = debugfs_create_u64
+		("dl_rx_packets", S_IRUSR | S_IRGRP | S_IROTH,
+		 sfc_debugfs_root, &global_stats.dl_rx_packets);
+	global_dbfs.dl_tx_bad_packets = debugfs_create_u64
+		("dl_tx_bad_packets", S_IRUSR | S_IRGRP | S_IROTH,
+		 sfc_debugfs_root, &global_stats.dl_tx_bad_packets);
+#endif
+}
+
+
+void netback_accel_debugfs_fini(void)
+{
+#if defined(CONFIG_DEBUG_FS)
+	debugfs_remove(global_dbfs.num_fwds);
+	debugfs_remove(global_dbfs.dl_tx_packets);
+	debugfs_remove(global_dbfs.dl_rx_packets);
+	debugfs_remove(global_dbfs.dl_tx_bad_packets);
+
+	debugfs_remove(sfc_debugfs_root);
+#endif
+}
+
+
+int netback_accel_debugfs_create(struct netback_accel *bend)
+{
+#if defined(CONFIG_DEBUG_FS)
+	/* Smallest length is 7 (vif0.0\n) */
+	int length = 7, temp;
+
+	if (sfc_debugfs_root == NULL)
+		return -ENOENT;
+
+	/* Work out length of string representation of far_end and vif_num */
+	temp = bend->far_end;
+	while (temp > 9) {
+		length++;
+		temp = temp / 10;
+	}
+	temp = bend->vif_num;
+	while (temp > 9) {
+		length++;
+		temp = temp / 10;
+	}
+
+	bend->dbfs_dir_name = kmalloc(length, GFP_KERNEL);
+	if (bend->dbfs_dir_name == NULL)
+		return -ENOMEM;
+	sprintf(bend->dbfs_dir_name, "vif%d.%d", bend->far_end, bend->vif_num);
+
+	bend->dbfs_dir = debugfs_create_dir(bend->dbfs_dir_name, 
+					    sfc_debugfs_root);
+	if (bend->dbfs_dir == NULL) {
+		kfree(bend->dbfs_dir_name);
+		return -ENOMEM;
+	}
+
+#if NETBACK_ACCEL_STATS
+	bend->dbfs.evq_wakeups = debugfs_create_u64
+		("evq_wakeups", S_IRUSR | S_IRGRP | S_IROTH,
+		 bend->dbfs_dir, &bend->stats.evq_wakeups);
+	bend->dbfs.evq_timeouts = debugfs_create_u64
+		("evq_timeouts", S_IRUSR | S_IRGRP | S_IROTH,
+		 bend->dbfs_dir, &bend->stats.evq_timeouts);
+	bend->dbfs.num_filters = debugfs_create_u32
+		("num_filters", S_IRUSR | S_IRGRP | S_IROTH,
+		 bend->dbfs_dir, &bend->stats.num_filters);
+	bend->dbfs.num_buffer_pages = debugfs_create_u32
+		("num_buffer_pages", S_IRUSR | S_IRGRP | S_IROTH,
+		 bend->dbfs_dir, &bend->stats.num_buffer_pages);
+#endif
+#endif
+        return 0;
+}
+
+
+int netback_accel_debugfs_remove(struct netback_accel *bend)
+{
+#if defined(CONFIG_DEBUG_FS)
+	if (bend->dbfs_dir != NULL) {
+#if NETBACK_ACCEL_STATS
+		debugfs_remove(bend->dbfs.evq_wakeups);
+		debugfs_remove(bend->dbfs.evq_timeouts);
+		debugfs_remove(bend->dbfs.num_filters);
+		debugfs_remove(bend->dbfs.num_buffer_pages);
+#endif
+		debugfs_remove(bend->dbfs_dir);
+	}
+
+	if (bend->dbfs_dir_name)
+		kfree(bend->dbfs_dir_name);
+#endif
+        return 0;
+}
+
+
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/accel_fwd.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/accel_fwd.c
--- linux-2.6.18.8/drivers/xen/sfc_netback/accel_fwd.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/accel_fwd.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,420 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include "accel.h"
+#include "accel_cuckoo_hash.h"
+#include "accel_util.h"
+#include "accel_solarflare.h"
+
+#include "driverlink_api.h"
+
+#include <linux/if_arp.h>
+#include <linux/skbuff.h>
+#include <linux/list.h>
+
+/* State stored in the forward table */
+struct fwd_struct {
+	struct list_head link; /* Forms list */
+	void * context;
+	__u8 valid;
+	__u8 mac[ETH_ALEN];
+};
+
+/* Max value we support */
+#define NUM_FWDS_BITS 8
+#define NUM_FWDS (1 << NUM_FWDS_BITS)
+#define FWD_MASK (NUM_FWDS - 1)
+
+struct port_fwd {
+	/* Make a list */
+	struct list_head link;
+	/* Hash table to store the fwd_structs */
+	cuckoo_hash_table fwd_hash_table;
+	/* The array of fwd_structs */
+	struct fwd_struct *fwd_array;
+	/* Linked list of entries in use. */
+	struct list_head fwd_list;
+	/* Could do something clever with a reader/writer lock. */
+	spinlock_t fwd_lock;
+	/* Make find_free_entry() a bit faster by caching this */
+	int last_free_index;
+};
+
+/*
+ * This is unlocked as it's only called from dl probe and remove,
+ * which are themselves synchronised.  Could get rid of it entirely as
+ * it's never iterated, but useful for debug
+ */
+static struct list_head port_fwds;
+
+
+/* Search the fwd_array for an unused entry */
+static int fwd_find_free_entry(struct port_fwd *fwd_set)
+{
+	int index = fwd_set->last_free_index;
+
+	do {
+		if (!fwd_set->fwd_array[index].valid) {
+			fwd_set->last_free_index = index;
+			return index;
+		}
+		index++;
+		if (index >= NUM_FWDS)
+			index = 0;
+	} while (index != fwd_set->last_free_index);
+
+	return -ENOMEM;
+}
+
+
+/* Look up a MAC in the hash table. Caller should hold table lock. */
+static inline struct fwd_struct *fwd_find_entry(const __u8 *mac,
+						struct port_fwd *fwd_set)
+{
+	cuckoo_hash_value value;
+	cuckoo_hash_mac_key key = cuckoo_mac_to_key(mac);
+
+	if (cuckoo_hash_lookup(&fwd_set->fwd_hash_table,
+			       (cuckoo_hash_key *)(&key),
+			       &value)) {
+		struct fwd_struct *fwd = &fwd_set->fwd_array[value];
+		DPRINTK_ON(memcmp(fwd->mac, mac, ETH_ALEN) != 0);
+		return fwd;
+	}
+
+	return NULL;
+}
+
+
+/* Initialise each nic port's fowarding table */
+void *netback_accel_init_fwd_port(void) 
+{	
+	struct port_fwd *fwd_set;
+
+	fwd_set = kzalloc(sizeof(struct port_fwd), GFP_KERNEL);
+	if (fwd_set == NULL) {
+		return NULL;
+	}
+
+	spin_lock_init(&fwd_set->fwd_lock);
+	
+	fwd_set->fwd_array = kzalloc(sizeof (struct fwd_struct) * NUM_FWDS,
+				     GFP_KERNEL);
+	if (fwd_set->fwd_array == NULL) {
+		kfree(fwd_set);
+		return NULL;
+	}
+	
+	if (cuckoo_hash_init(&fwd_set->fwd_hash_table, NUM_FWDS_BITS, 8) != 0) {
+		kfree(fwd_set->fwd_array);
+		kfree(fwd_set);
+		return NULL;
+	}
+	
+	INIT_LIST_HEAD(&fwd_set->fwd_list);
+	
+	list_add(&fwd_set->link, &port_fwds);
+
+	return fwd_set;
+}
+
+
+void netback_accel_shutdown_fwd_port(void *fwd_priv)
+{
+	struct port_fwd *fwd_set = (struct port_fwd *)fwd_priv;
+
+	BUG_ON(fwd_priv == NULL);
+	
+	BUG_ON(list_empty(&port_fwds));
+	list_del(&fwd_set->link);
+
+	BUG_ON(!list_empty(&fwd_set->fwd_list));
+
+	cuckoo_hash_destroy(&fwd_set->fwd_hash_table);
+	kfree(fwd_set->fwd_array);
+	kfree(fwd_set);
+}
+
+
+int netback_accel_init_fwd()
+{
+	INIT_LIST_HEAD(&port_fwds);
+	return 0;
+}
+
+
+void netback_accel_shutdown_fwd()
+{
+	BUG_ON(!list_empty(&port_fwds));
+}
+
+
+/*
+ * Add an entry to the forwarding table.  Returns -ENOMEM if no
+ * space.
+ */
+int netback_accel_fwd_add(const __u8 *mac, void *context, void *fwd_priv)
+{
+	struct fwd_struct *fwd;
+	int rc = 0, index;
+	unsigned long flags;
+	cuckoo_hash_mac_key key = cuckoo_mac_to_key(mac);
+	struct port_fwd *fwd_set = (struct port_fwd *)fwd_priv;
+
+	BUG_ON(fwd_priv == NULL);
+
+	DPRINTK("Adding mac " MAC_FMT "\n", MAC_ARG(mac));
+       
+	spin_lock_irqsave(&fwd_set->fwd_lock, flags);
+	
+	if ((rc = fwd_find_free_entry(fwd_set)) < 0 ) {
+		spin_unlock_irqrestore(&fwd_set->fwd_lock, flags);
+		return rc;
+	}
+
+	index = rc;
+
+	/* Shouldn't already be in the table */
+	if (cuckoo_hash_lookup(&fwd_set->fwd_hash_table,
+			       (cuckoo_hash_key *)(&key), &rc) != 0) {
+		spin_unlock_irqrestore(&fwd_set->fwd_lock, flags);
+		EPRINTK("MAC address " MAC_FMT " already accelerated.\n",
+			MAC_ARG(mac));
+		return -EEXIST;
+	}
+
+	if ((rc = cuckoo_hash_add(&fwd_set->fwd_hash_table,
+				  (cuckoo_hash_key *)(&key), index, 1)) == 0) {
+		fwd = &fwd_set->fwd_array[index];
+		fwd->valid = 1;
+		fwd->context = context;
+		memcpy(fwd->mac, mac, ETH_ALEN);
+		list_add(&fwd->link, &fwd_set->fwd_list);
+		NETBACK_ACCEL_STATS_OP(global_stats.num_fwds++);
+	}
+
+	spin_unlock_irqrestore(&fwd_set->fwd_lock, flags);
+
+	/*
+	 * No need to tell frontend that this mac address is local -
+	 * it should auto-discover through packets on fastpath what is
+	 * local and what is not, and just being on same server
+	 * doesn't make it local (it could be on a different
+	 * bridge)
+	 */
+
+	return rc;
+}
+
+
+/* remove an entry from the forwarding tables. */
+void netback_accel_fwd_remove(const __u8 *mac, void *fwd_priv)
+{
+	struct fwd_struct *fwd;
+	unsigned long flags;
+	cuckoo_hash_mac_key key = cuckoo_mac_to_key(mac);
+	struct port_fwd *fwd_set = (struct port_fwd *)fwd_priv;
+
+	DPRINTK("Removing mac " MAC_FMT "\n", MAC_ARG(mac));
+
+	BUG_ON(fwd_priv == NULL);
+
+	spin_lock_irqsave(&fwd_set->fwd_lock, flags);
+
+	fwd = fwd_find_entry(mac, fwd_set);
+	if (fwd != NULL) {
+		BUG_ON(list_empty(&fwd_set->fwd_list));
+		list_del(&fwd->link);
+
+		fwd->valid = 0;
+		cuckoo_hash_remove(&fwd_set->fwd_hash_table, 
+				   (cuckoo_hash_key *)(&key));
+		NETBACK_ACCEL_STATS_OP(global_stats.num_fwds--);
+	}
+	spin_unlock_irqrestore(&fwd_set->fwd_lock, flags);
+
+	/*
+	 * No need to tell frontend that this is no longer present -
+	 * the frontend is currently only interested in remote
+	 * addresses and it works these out (mostly) by itself
+	 */
+}
+
+
+/* Set the context pointer for a hash table entry. */
+int netback_accel_fwd_set_context(const __u8 *mac, void *context, 
+				  void *fwd_priv)
+{
+	struct fwd_struct *fwd;
+	unsigned long flags;
+	int rc = -ENOENT;
+	struct port_fwd *fwd_set = (struct port_fwd *)fwd_priv;
+
+	BUG_ON(fwd_priv == NULL);
+
+	spin_lock_irqsave(&fwd_set->fwd_lock, flags);
+	fwd = fwd_find_entry(mac, fwd_set);
+	if (fwd != NULL) {
+		fwd->context = context;
+		rc = 0;
+	}
+	spin_unlock_irqrestore(&fwd_set->fwd_lock, flags);
+	return rc;
+}
+
+
+/**************************************************************************
+ * Process a received packet
+ **************************************************************************/
+
+/*
+ * Returns whether or not we have a match in our forward table for the
+ * this skb. Must be called with appropriate fwd_lock already held
+ */
+static struct netback_accel *for_a_vnic(struct netback_pkt_buf *skb, 
+					struct port_fwd *fwd_set)
+{
+	struct fwd_struct *fwd;
+	struct netback_accel *retval = NULL;
+
+	fwd = fwd_find_entry(skb->mac.raw, fwd_set);
+	if (fwd != NULL)
+		retval = fwd->context;
+	return retval;
+}
+
+
+static inline int packet_is_arp_reply(struct sk_buff *skb)
+{
+	return skb->protocol == ntohs(ETH_P_ARP) 
+		&& skb->nh.arph->ar_op == ntohs(ARPOP_REPLY);
+}
+
+
+static inline void hdr_to_filt(struct ethhdr *ethhdr, struct iphdr *ip,
+			       struct netback_accel_filter_spec *spec)
+{
+	spec->proto = ip->protocol;
+	spec->destip_be = ip->daddr;
+	memcpy(spec->mac, ethhdr->h_source, ETH_ALEN);
+
+	if (ip->protocol == IPPROTO_TCP) {
+		struct tcphdr *tcp = (struct tcphdr *)((char *)ip + 4 * ip->ihl);
+		spec->destport_be = tcp->dest;
+	} else {
+		struct udphdr *udp = (struct udphdr *)((char *)ip + 4 * ip->ihl);
+		EPRINTK_ON(ip->protocol != IPPROTO_UDP);
+		spec->destport_be = udp->dest;
+	}
+}
+
+
+static inline int netback_accel_can_filter(struct netback_pkt_buf *skb) 
+{
+	return (skb->protocol == htons(ETH_P_IP) && 
+		((skb->nh.iph->protocol == IPPROTO_TCP) ||
+		 (skb->nh.iph->protocol == IPPROTO_UDP)));
+}
+
+
+static inline void netback_accel_filter_packet(struct netback_accel *bend,
+					       struct netback_pkt_buf *skb)
+{
+	struct netback_accel_filter_spec fs;
+	struct ethhdr *eh = (struct ethhdr *)(skb->mac.raw);
+
+	hdr_to_filt(eh, skb->nh.iph, &fs);
+	
+	netback_accel_filter_check_add(bend, &fs);
+}
+
+
+/*
+ * Receive a packet and do something appropriate with it. Return true
+ * to take exclusive ownership of the packet.  This is verging on
+ * solarflare specific
+ */
+void netback_accel_rx_packet(struct netback_pkt_buf *skb, void *fwd_priv)
+{
+	struct netback_accel *bend;
+	struct port_fwd *fwd_set = (struct port_fwd *)fwd_priv;
+	unsigned long flags;
+
+	BUG_ON(fwd_priv == NULL);
+
+	/* Checking for bcast is cheaper so do that first */
+	if (is_broadcast_ether_addr(skb->mac.raw)) {
+		/* pass through the slow path by not claiming ownership */
+		return;
+	} else if (is_multicast_ether_addr(skb->mac.raw)) {
+		/* pass through the slow path by not claiming ownership */
+		return;
+	} else {
+		/* It is unicast */
+		spin_lock_irqsave(&fwd_set->fwd_lock, flags);
+		/* We insert filter to pass it off to a VNIC */
+		if ((bend = for_a_vnic(skb, fwd_set)) != NULL)
+			if (netback_accel_can_filter(skb))
+				netback_accel_filter_packet(bend, skb);
+		spin_unlock_irqrestore(&fwd_set->fwd_lock, flags);
+	}
+	return;
+}
+
+
+void netback_accel_tx_packet(struct sk_buff *skb, void *fwd_priv) 
+{
+	__u8 *mac;
+	unsigned long flags;
+	struct port_fwd *fwd_set = (struct port_fwd *)fwd_priv;
+	struct fwd_struct *fwd;
+
+	BUG_ON(fwd_priv == NULL);
+
+	if (is_broadcast_ether_addr(skb->mac.raw) && packet_is_arp_reply(skb)) {
+		/*
+		 * update our fast path forwarding to reflect this
+		 * gratuitous ARP
+		 */ 
+		mac = skb->mac.raw+ETH_ALEN;
+
+		DPRINTK("%s: found gratuitous ARP for " MAC_FMT "\n",
+			__FUNCTION__, MAC_ARG(mac));
+
+		spin_lock_irqsave(&fwd_set->fwd_lock, flags);
+		/*
+		 * Might not be local, but let's tell them all it is,
+		 * and they can restore the fastpath if they continue
+		 * to get packets that way
+		 */
+		list_for_each_entry(fwd, &fwd_set->fwd_list, link) {
+			struct netback_accel *bend = fwd->context;
+			if (bend != NULL)
+				netback_accel_msg_tx_new_localmac(bend, mac);
+		}
+
+		spin_unlock_irqrestore(&fwd_set->fwd_lock, flags);
+	}
+	return;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/accel.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/accel.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/accel.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/accel.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,393 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef NETBACK_ACCEL_H
+#define NETBACK_ACCEL_H
+
+#include <linux/slab.h>
+#include <linux/ip.h>
+#include <linux/tcp.h>
+#include <linux/udp.h>
+#include <linux/in.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/mutex.h>
+#include <linux/wait.h>
+
+#include <xen/xenbus.h>
+
+#include "accel_shared_fifo.h"
+#include "accel_msg_iface.h"
+#include "accel_util.h"
+
+/**************************************************************************
+ * Datatypes
+ **************************************************************************/
+
+#define NETBACK_ACCEL_DEFAULT_MAX_FILTERS (8)
+#define NETBACK_ACCEL_DEFAULT_MAX_MCASTS (8)
+#define NETBACK_ACCEL_DEFAULT_MAX_BUF_PAGES (384)
+/* Variable to store module parameter for max_buf_pages */
+extern unsigned sfc_netback_max_pages;
+
+#define NETBACK_ACCEL_STATS 1
+
+#if NETBACK_ACCEL_STATS
+#define NETBACK_ACCEL_STATS_OP(x) x
+#else
+#define NETBACK_ACCEL_STATS_OP(x)
+#endif
+
+/*! Statistics for a given backend */
+struct netback_accel_stats {
+	/*! Number of eventq wakeup events */
+	u64 evq_wakeups;
+	/*! Number of eventq timeout events */
+	u64 evq_timeouts;
+	/*! Number of filters used */
+	u32 num_filters;
+	/*! Number of buffer pages registered */
+	u32 num_buffer_pages;
+};
+
+
+/* Debug fs nodes for each of the above stats */
+struct netback_accel_dbfs {
+	struct dentry *evq_wakeups;
+	struct dentry *evq_timeouts;
+	struct dentry *num_filters;
+	struct dentry *num_buffer_pages;
+};
+
+
+/*! Resource limits for a given NIC */
+struct netback_accel_limits {
+	int max_filters;	    /*!< Max. number of filters to use. */
+	int max_mcasts;	     /*!< Max. number  of mcast subscriptions */
+	int max_buf_pages;	  /*!< Max. number of pages of NIC buffers */
+};
+
+
+/*! The state for an instance of the back end driver. */
+struct netback_accel {
+	/*! mutex to protect this state */
+	struct mutex bend_mutex;
+
+	/*! Watches on xenstore */
+	struct xenbus_watch domu_accel_watch;
+	struct xenbus_watch config_accel_watch;
+
+	/*! Pointer to whatever device cookie ties us in to the hypervisor */
+	void *hdev_data;
+
+	/*! FIFO indices. Next page is msg FIFOs */
+	struct net_accel_shared_page *shared_page;
+
+	/*! Defer control message processing */
+	struct work_struct handle_msg;
+
+	/*! Identifies other end VM and interface.*/
+	int far_end;
+	int vif_num;
+
+	/*!< To unmap the shared pages */
+	void *sh_pages_unmap;
+
+	/* Resource tracking */
+	/*! Limits on H/W & Dom0 resources */
+	struct netback_accel_limits quotas;
+
+	/* Hardware resources */
+	/*! The H/W type of associated NIC */
+	enum net_accel_hw_type hw_type;
+	/*! State of allocation */	       
+	int hw_state;
+	/*! Index into ci_driver.nics[] for this interface */
+	int nic_index;
+	/*! How to set up the acceleration for this hardware */
+	int (*accel_setup)(struct netback_accel *); 
+	/*! And how to stop it. */
+	void (*accel_shutdown)(struct netback_accel *);
+
+	/*! The physical/real net_dev for this interface */
+	struct net_device *net_dev;
+
+	/*! Magic pointer to locate state in fowarding table */
+	void *fwd_priv;
+
+	/*! Message FIFO */
+	sh_msg_fifo2 to_domU;
+	/*! Message FIFO */
+	sh_msg_fifo2 from_domU;
+
+	/*! General notification channel id */
+	int msg_channel;
+	/*! General notification channel irq */
+	int msg_channel_irq;
+
+	/*! Event channel id dedicated to network packet interrupts. */
+	int net_channel; 
+	/*! Event channel irq dedicated to network packets interrupts */
+	int net_channel_irq; 
+
+	/*! The MAC address the frontend goes by. */
+	u8 mac[ETH_ALEN];
+	/*! Driver name of associated NIC */
+	char *nicname;    
+
+	/*! Array of pointers to buffer pages mapped */
+	grant_handle_t *buffer_maps; 
+	u64 *buffer_addrs;
+	/*! Index into buffer_maps */
+	int buffer_maps_index; 
+	/*! Max number of pages that domU is allowed/will request to map */
+	int max_pages; 
+
+	/*! Pointer to hardware specific private area */
+	void *accel_hw_priv; 
+
+	/*! Wait queue for changes in accelstate. */
+	wait_queue_head_t state_wait_queue;
+
+	/*! Current state of the frontend according to the xenbus
+	 *  watch. */
+	XenbusState frontend_state;
+
+	/*! Current state of this backend. */
+	XenbusState backend_state;
+
+	/*! Non-zero if the backend is being removed. */
+	int removing;
+
+	/*! Non-zero if the setup_vnic has been called. */
+	int vnic_is_setup;
+
+#if NETBACK_ACCEL_STATS
+	struct netback_accel_stats stats;
+#endif	
+#if defined(CONFIG_DEBUG_FS)
+	char *dbfs_dir_name;
+	struct dentry *dbfs_dir;
+	struct netback_accel_dbfs dbfs;
+#endif
+
+	/*! List */
+	struct netback_accel *next_bend;
+};
+
+
+/*
+ * Values for netback_accel.hw_state.  States of resource allocation
+ * we can go through
+ */
+/*! No hardware has yet been allocated. */
+#define NETBACK_ACCEL_RES_NONE  (0)
+/*! Hardware has been allocated. */
+#define NETBACK_ACCEL_RES_ALLOC (1)
+#define NETBACK_ACCEL_RES_FILTER (2)
+#define NETBACK_ACCEL_RES_HWINFO (3)
+
+/*! Filtering specification. This assumes that for VNIC support we
+ *  will always want wildcard entries, so only specifies the
+ *  destination IP/port
+ */
+struct netback_accel_filter_spec {
+	/*! Internal, used to access efx_vi API */
+	void *filter_handle; 
+
+	/*! Destination IP in network order */
+	u32 destip_be;
+	/*! Destination port in network order */
+	u16 destport_be;
+	/*! Mac address */
+	u8  mac[ETH_ALEN];
+	/*! TCP or UDP */
+	u8  proto;	
+};
+
+
+/**************************************************************************
+ * From accel.c
+ **************************************************************************/
+
+/*! \brief Start up all the acceleration plugins 
+ *
+ * \return 0 on success, an errno on failure
+ */
+extern int netback_accel_init_accel(void);
+
+/*! \brief Shut down all the acceleration plugins 
+ */
+extern void netback_accel_shutdown_accel(void);
+
+
+/**************************************************************************
+ * From accel_fwd.c
+ **************************************************************************/
+
+/*! \brief Init the forwarding infrastructure
+ * \return 0 on success, or -ENOMEM if it couldn't get memory for the
+ * forward table 
+ */
+extern int netback_accel_init_fwd(void);
+
+/*! \brief Shut down the forwarding and free memory. */
+extern void netback_accel_shutdown_fwd(void);
+
+/*! Initialise each nic port's fowarding table */
+extern void *netback_accel_init_fwd_port(void);
+extern void netback_accel_shutdown_fwd_port(void *fwd_priv);
+
+/*! \brief Add an entry to the forwarding table. 
+ * \param mac : MAC address, used as hash key
+ * \param ctxt : value to associate with key (can be NULL, see
+ * netback_accel_fwd_set_context)
+ * \return 0 on success, -ENOMEM if table was full and could no grow it
+ */
+extern int netback_accel_fwd_add(const __u8 *mac, void *context,
+				 void *fwd_priv);
+
+/*! \brief Remove an entry from the forwarding table. 
+ * \param mac : the MAC address to remove
+ * \return nothing: it is not an error if the mac was not in the table
+ */
+extern void netback_accel_fwd_remove(const __u8 *mac, void *fwd_priv);
+
+/*! \brief Set the context pointer for an existing fwd table entry.
+ * \param mac : key that is already present in the table
+ * \param context : new value to associate with key
+ * \return 0 on success, -ENOENT if mac not present in table.
+ */
+extern int netback_accel_fwd_set_context(const __u8 *mac, void *context,
+					 void *fwd_priv);
+
+/**************************************************************************
+ * From accel_msg.c
+ **************************************************************************/
+
+
+/*! \brief Send the start-of-day message that handshakes with the VNIC
+ *  and tells it its MAC address.
+ *
+ * \param bend The back end driver data structure
+ * \param version The version of communication to use, e.g. NET_ACCEL_MSG_VERSION
+ */
+extern void netback_accel_msg_tx_hello(struct netback_accel *bend,
+				       unsigned version);
+
+/*! \brief Send a "there's a new local mac address" message 
+ *
+ * \param bend The back end driver data structure for the vnic to send
+ * the message to 
+ * \param mac Pointer to the new mac address
+ */
+extern void netback_accel_msg_tx_new_localmac(struct netback_accel *bend,
+					      const void *mac);
+
+/*! \brief Send a "a mac address that was local has gone away" message 
+ *
+ * \param bend The back end driver data structure for the vnic to send
+ * the message to 
+ * \param mac Pointer to the old mac address
+ */
+extern void netback_accel_msg_tx_old_localmac(struct netback_accel *bend,
+					      const void *mac);
+
+extern void netback_accel_set_interface_state(struct netback_accel *bend,
+					      int up);
+
+/*! \brief Process the message queue for a bend that has just
+ * interrupted.
+ * 
+ * Demultiplexs an interrupt from the front end driver, taking
+ * messages from the fifo and taking appropriate action.
+ * 
+ * \param bend The back end driver data structure
+ */
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20)
+extern void netback_accel_msg_rx_handler(struct work_struct *arg);
+#else
+extern void netback_accel_msg_rx_handler(void *bend_void);
+#endif
+
+/**************************************************************************
+ * From accel_xenbus.c
+ **************************************************************************/
+/*! List of all the bends currently in existence. */
+extern struct netback_accel *bend_list;
+extern struct mutex bend_list_mutex;
+
+/*! \brief Probe a new network interface. */
+extern int netback_accel_probe(struct xenbus_device *dev);
+
+/*! \brief Remove a network interface. */
+extern int netback_accel_remove(struct xenbus_device *dev);
+
+/*! \brief Shutdown all accelerator backends */
+extern void netback_accel_shutdown_bends(void);
+
+/*! \brief Initiate the xenbus state teardown handshake */
+extern void netback_accel_set_closing(struct netback_accel *bend);
+
+/**************************************************************************
+ * From accel_debugfs.c
+ **************************************************************************/
+/*! Global statistics */
+struct netback_accel_global_stats {
+	/*! Number of TX packets seen through driverlink */
+	u64 dl_tx_packets;
+	/*! Number of TX packets seen through driverlink we didn't like */
+	u64 dl_tx_bad_packets;
+	/*! Number of RX packets seen through driverlink */
+	u64 dl_rx_packets;
+	/*! Number of mac addresses we are forwarding to */
+	u32 num_fwds;
+};
+
+/*! Debug fs entries for each of the above stats */
+struct netback_accel_global_dbfs {
+	struct dentry *dl_tx_packets;
+	struct dentry *dl_tx_bad_packets;
+	struct dentry *dl_rx_packets;
+	struct dentry *num_fwds;
+};
+
+#if NETBACK_ACCEL_STATS
+extern struct netback_accel_global_stats global_stats;
+#endif
+
+/*! \brief Initialise the debugfs root and populate with global stats */
+extern void netback_accel_debugfs_init(void);
+
+/*! \brief Remove our debugfs root directory */
+extern void netback_accel_debugfs_fini(void);
+
+/*! \brief Add per-bend statistics to debug fs */
+extern int netback_accel_debugfs_create(struct netback_accel *bend);
+/*! \brief Remove per-bend statistics from debug fs */
+extern int netback_accel_debugfs_remove(struct netback_accel *bend);
+
+#endif /* NETBACK_ACCEL_H */
+
+
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/accel_msg.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/accel_msg.c
--- linux-2.6.18.8/drivers/xen/sfc_netback/accel_msg.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/accel_msg.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,392 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <xen/evtchn.h>
+
+#include "accel.h"
+#include "accel_msg_iface.h"
+#include "accel_util.h"
+#include "accel_solarflare.h"
+
+/* Send a HELLO to front end to start things off */
+void netback_accel_msg_tx_hello(struct netback_accel *bend, unsigned version)
+{
+	unsigned long lock_state;
+	struct net_accel_msg *msg = 
+		net_accel_msg_start_send(bend->shared_page,
+					 &bend->to_domU, &lock_state);
+	/* The queue _cannot_ be full, we're the first users. */
+	EPRINTK_ON(msg == NULL);
+
+	if (msg != NULL) {
+		net_accel_msg_init(msg, NET_ACCEL_MSG_HELLO);
+		msg->u.hello.version = version;
+		msg->u.hello.max_pages = bend->quotas.max_buf_pages; 
+		VPRINTK("Sending hello to channel %d\n", bend->msg_channel);
+		net_accel_msg_complete_send_notify(bend->shared_page, 
+						   &bend->to_domU,
+						   &lock_state, 
+						   bend->msg_channel_irq);
+	}
+}
+
+/* Send a local mac message to vnic */
+static void netback_accel_msg_tx_localmac(struct netback_accel *bend, 
+					  int type, const void *mac)
+{
+	unsigned long lock_state;
+	struct net_accel_msg *msg;
+
+	BUG_ON(bend == NULL || mac == NULL);
+
+	VPRINTK("Sending local mac message: " MAC_FMT "\n", 
+		MAC_ARG((const char *)mac));  
+	
+	msg = net_accel_msg_start_send(bend->shared_page, &bend->to_domU,
+				       &lock_state);
+	
+	if (msg != NULL) {
+		net_accel_msg_init(msg, NET_ACCEL_MSG_LOCALMAC);
+		msg->u.localmac.flags = type;
+		memcpy(msg->u.localmac.mac, mac, ETH_ALEN);
+		net_accel_msg_complete_send_notify(bend->shared_page, 
+						   &bend->to_domU,
+						   &lock_state, 
+						   bend->msg_channel_irq);
+	} else {
+		/*
+		 * TODO if this happens we may leave a domU
+		 * fastpathing packets when they should be delivered
+		 * locally.  Solution is get domU to timeout entries
+		 * in its fastpath lookup table when it receives no RX
+		 * traffic
+		 */
+		EPRINTK("%s: saw full queue, may need ARP timer to recover\n",
+			__FUNCTION__);
+	}
+}
+
+/* Send an add local mac message to vnic */
+void netback_accel_msg_tx_new_localmac(struct netback_accel *bend,
+				       const void *mac)
+{
+	netback_accel_msg_tx_localmac(bend, NET_ACCEL_MSG_ADD, mac);
+}
+
+
+static int netback_accel_msg_rx_buffer_map(struct netback_accel *bend, 
+					   struct net_accel_msg *msg)
+{
+	int log2_pages, rc;
+
+	/* Can only allocate in power of two */
+	log2_pages = log2_ge(msg->u.mapbufs.pages, 0);
+	if (msg->u.mapbufs.pages != pow2(log2_pages)) {
+		EPRINTK("%s: Can only alloc bufs in power of 2 sizes (%d)\n",
+			__FUNCTION__, msg->u.mapbufs.pages);
+		rc = -EINVAL;
+		goto err_out;
+	}
+  
+	/*
+	 * Sanity.  Assumes NET_ACCEL_MSG_MAX_PAGE_REQ is same for
+	 * both directions/domains
+	 */
+	if (msg->u.mapbufs.pages > NET_ACCEL_MSG_MAX_PAGE_REQ) {
+		EPRINTK("%s: too many pages in a single message: %d %d\n", 
+			__FUNCTION__, msg->u.mapbufs.pages,
+			NET_ACCEL_MSG_MAX_PAGE_REQ);
+		rc = -EINVAL;
+		goto err_out;
+	}
+  
+	if ((rc = netback_accel_add_buffers(bend, msg->u.mapbufs.pages, 
+					    log2_pages, msg->u.mapbufs.grants, 
+					    &msg->u.mapbufs.buf)) < 0) {
+		goto err_out;
+	}
+
+	msg->id |= NET_ACCEL_MSG_REPLY;
+  
+	return 0;
+
+ err_out:
+	EPRINTK("%s: err_out\n", __FUNCTION__);
+	msg->id |= NET_ACCEL_MSG_ERROR | NET_ACCEL_MSG_REPLY;
+	return rc;
+}
+
+
+/* Hint from frontend that one of our filters is out of date */
+static int netback_accel_process_fastpath(struct netback_accel *bend, 
+					  struct net_accel_msg *msg)
+{
+	struct netback_accel_filter_spec spec;
+
+	if (msg->u.fastpath.flags & NET_ACCEL_MSG_REMOVE) {
+		/* 
+		 * Would be nice to BUG() this but would leave us
+		 * vulnerable to naughty frontend
+		 */
+		EPRINTK_ON(msg->u.fastpath.flags & NET_ACCEL_MSG_ADD);
+		
+		memcpy(spec.mac, msg->u.fastpath.mac, ETH_ALEN);
+		spec.destport_be = msg->u.fastpath.port;
+		spec.destip_be = msg->u.fastpath.ip;
+		spec.proto = msg->u.fastpath.proto;
+
+		netback_accel_filter_remove_spec(bend, &spec);
+	}
+
+	return 0;
+}
+
+
+/* Flow control for message queues */
+inline void set_queue_not_full(struct netback_accel *bend)
+{
+	if (!test_and_set_bit(NET_ACCEL_MSG_AFLAGS_QUEUEUNOTFULL_B, 
+			      (unsigned long *)&bend->shared_page->aflags))
+		notify_remote_via_irq(bend->msg_channel_irq);
+	else
+		VPRINTK("queue not full bit already set, not signalling\n");
+}
+
+
+/* Flow control for message queues */
+inline void set_queue_full(struct netback_accel *bend)
+{
+	if (!test_and_set_bit(NET_ACCEL_MSG_AFLAGS_QUEUE0FULL_B,
+			      (unsigned long *)&bend->shared_page->aflags))
+		notify_remote_via_irq(bend->msg_channel_irq);
+	else
+		VPRINTK("queue full bit already set, not signalling\n");
+}
+
+
+void netback_accel_set_interface_state(struct netback_accel *bend, int up)
+{
+	bend->shared_page->net_dev_up = up;
+	if (!test_and_set_bit(NET_ACCEL_MSG_AFLAGS_NETUPDOWN_B, 
+			     (unsigned long *)&bend->shared_page->aflags))
+		notify_remote_via_irq(bend->msg_channel_irq);
+	else
+		VPRINTK("interface up/down bit already set, not signalling\n");
+}
+
+
+static int check_rx_hello_version(unsigned version) 
+{
+	/* Should only happen if there's been a version mismatch */
+	BUG_ON(version == NET_ACCEL_MSG_VERSION);
+
+	if (version > NET_ACCEL_MSG_VERSION) {
+		/* Newer protocol, we must refuse */
+		return -EPROTO;
+	}
+
+	if (version < NET_ACCEL_MSG_VERSION) {
+		/*
+		 * We are newer, so have discretion to accept if we
+		 * wish.  For now however, just reject
+		 */
+		return -EPROTO;
+	}
+
+	return -EINVAL;
+}
+
+
+static int process_rx_msg(struct netback_accel *bend,
+			  struct net_accel_msg *msg)
+{
+	int err = 0;
+		      
+	switch (msg->id) {
+	case NET_ACCEL_MSG_REPLY | NET_ACCEL_MSG_HELLO:
+		/* Reply to a HELLO; mark ourselves as connected */
+		DPRINTK("got Hello reply, version %.8x\n",
+			msg->u.hello.version);
+		
+		/*
+		 * Check that we've not successfully done this
+		 * already.  NB no check at the moment that this reply
+		 * comes after we've actually sent a HELLO as that's
+		 * not possible with the current code structure
+		 */
+		if (bend->hw_state != NETBACK_ACCEL_RES_NONE)
+			return -EPROTO;
+
+		/* Store max_pages for accel_setup */
+		if (msg->u.hello.max_pages > bend->quotas.max_buf_pages) {
+			EPRINTK("More pages than quota allows (%d > %d)\n",
+				msg->u.hello.max_pages, 
+				bend->quotas.max_buf_pages);
+			/* Force it down to the quota */
+			msg->u.hello.max_pages = bend->quotas.max_buf_pages;
+		}
+		bend->max_pages = msg->u.hello.max_pages;
+		
+		/* Set up the hardware visible to the other end */
+		err = bend->accel_setup(bend);
+		if (err) {
+			/* This is fatal */
+			DPRINTK("Hello gave accel_setup error %d\n", err);
+			netback_accel_set_closing(bend);
+		} else {
+			/*
+			 * Now add the context so that packet
+			 * forwarding will commence
+			 */
+			netback_accel_fwd_set_context(bend->mac, bend, 
+						      bend->fwd_priv);
+		}
+		break;
+	case NET_ACCEL_MSG_REPLY | NET_ACCEL_MSG_HELLO | NET_ACCEL_MSG_ERROR:
+		EPRINTK("got Hello error, versions us:%.8x them:%.8x\n",
+			NET_ACCEL_MSG_VERSION, msg->u.hello.version);
+
+		if (bend->hw_state != NETBACK_ACCEL_RES_NONE)
+			return -EPROTO;
+
+		if (msg->u.hello.version != NET_ACCEL_MSG_VERSION) {
+			/* Error is due to version mismatch */
+			err = check_rx_hello_version(msg->u.hello.version);
+			if (err == 0) {
+				/*
+				 * It's OK to be compatible, send
+				 * another hello with compatible version
+				 */
+				netback_accel_msg_tx_hello
+					(bend, msg->u.hello.version);
+			} else {
+				/*
+				 * Tell frontend that we're not going to
+				 * send another HELLO by going to Closing.
+				 */
+				netback_accel_set_closing(bend);
+			}
+		} 
+		break;
+	case NET_ACCEL_MSG_MAPBUF:
+		VPRINTK("Got mapped buffers request %d\n",
+			msg->u.mapbufs.reqid);
+
+		if (bend->hw_state == NETBACK_ACCEL_RES_NONE)
+			return -EPROTO;
+
+		/*
+		 * Frontend wants a buffer table entry for the
+		 * supplied pages
+		 */
+		err = netback_accel_msg_rx_buffer_map(bend, msg);
+		if (net_accel_msg_reply_notify(bend->shared_page,
+					       bend->msg_channel_irq, 
+					       &bend->to_domU, msg)) {
+			/*
+			 * This is fatal as we can't tell the frontend
+			 * about the problem through the message
+			 * queue, and so would otherwise stalemate
+			 */
+			netback_accel_set_closing(bend);
+		}
+		break;
+	case NET_ACCEL_MSG_FASTPATH:
+		DPRINTK("Got fastpath request\n");
+
+		if (bend->hw_state == NETBACK_ACCEL_RES_NONE)
+			return -EPROTO;
+
+		err = netback_accel_process_fastpath(bend, msg);
+		break;
+	default:
+		EPRINTK("Huh? Message code is %x\n", msg->id);
+		err = -EPROTO;
+		break;
+	}
+	return err;
+}
+
+
+/*  Demultiplex an IRQ from the frontend driver.  */
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20)
+void netback_accel_msg_rx_handler(struct work_struct *arg)
+#else
+void netback_accel_msg_rx_handler(void *bend_void)
+#endif
+{
+	struct net_accel_msg msg;
+	int err, queue_was_full = 0;
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20)
+	struct netback_accel *bend = 
+		container_of(arg, struct netback_accel, handle_msg);
+#else
+	struct netback_accel *bend = (struct netback_accel *)bend_void;
+#endif
+
+	mutex_lock(&bend->bend_mutex);
+
+	/*
+	 * This happens when the shared pages have been unmapped, but
+	 * the workqueue not flushed yet
+	 */
+	if (bend->shared_page == NULL)
+		goto done;
+
+	if ((bend->shared_page->aflags &
+	     NET_ACCEL_MSG_AFLAGS_TO_DOM0_MASK) != 0) {
+		if (bend->shared_page->aflags &
+		    NET_ACCEL_MSG_AFLAGS_QUEUE0NOTFULL) {
+			/* We've been told there may now be space. */
+			clear_bit(NET_ACCEL_MSG_AFLAGS_QUEUE0NOTFULL_B, 
+				  (unsigned long *)&bend->shared_page->aflags);
+		}
+
+		if (bend->shared_page->aflags &
+		    NET_ACCEL_MSG_AFLAGS_QUEUEUFULL) {
+			clear_bit(NET_ACCEL_MSG_AFLAGS_QUEUEUFULL_B, 
+				  (unsigned long *)&bend->shared_page->aflags);
+			queue_was_full = 1;
+		}
+	}
+
+	while ((err = net_accel_msg_recv(bend->shared_page, &bend->from_domU,
+					 &msg)) == 0) {
+		err = process_rx_msg(bend, &msg);
+		
+		if (err != 0) {
+			EPRINTK("%s: Error %d\n", __FUNCTION__, err);
+			goto err;
+		}
+	}
+
+ err:
+	/* There will be space now if we can make any. */
+	if (queue_was_full) 
+		set_queue_not_full(bend);
+ done:
+	mutex_unlock(&bend->bend_mutex);
+
+	return;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/accel_solarflare.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/accel_solarflare.c
--- linux-2.6.18.8/drivers/xen/sfc_netback/accel_solarflare.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/accel_solarflare.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1253 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include "common.h"
+
+#include "accel.h"
+#include "accel_solarflare.h"
+#include "accel_msg_iface.h"
+#include "accel_util.h"
+
+#include "accel_cuckoo_hash.h"
+
+#include "ci/driver/resource/efx_vi.h"
+
+#include "ci/efrm/nic_table.h" 
+#include "ci/efhw/public.h"
+
+#include <xen/evtchn.h>
+#include <xen/driver_util.h>
+#include <linux/list.h>
+#include <linux/mutex.h>
+
+#include "driverlink_api.h"
+
+#define SF_XEN_RX_USR_BUF_SIZE 2048
+
+struct falcon_bend_accel_priv {
+	struct efx_vi_state *efx_vih;
+
+	/*! Array of pointers to dma_map state, used so VNIC can
+	 *  request their removal in a single message
+	 */
+	struct efx_vi_dma_map_state **dma_maps;
+	/*! Index into dma_maps */
+	int dma_maps_index; 
+
+	/*! Serialises access to filters */
+	spinlock_t filter_lock;	     
+	/*! Bitmap of which filters are free */
+	unsigned long free_filters;	 
+	/*! Used for index normalisation */
+	u32 filter_idx_mask;		
+	struct netback_accel_filter_spec *fspecs; 
+	cuckoo_hash_table filter_hash_table;
+
+	u32 txdmaq_gnt;
+	u32 rxdmaq_gnt;
+	u32 doorbell_gnt;
+	u32 evq_rptr_gnt;
+	u32 evq_mem_gnts[EF_HW_FALCON_EVQ_PAGES];
+	u32 evq_npages;
+};
+
+/* Forward declaration */
+static int netback_accel_filter_init(struct netback_accel *);
+static void netback_accel_filter_shutdown(struct netback_accel *);
+
+/**************************************************************************
+ * 
+ * Driverlink stuff
+ *
+ **************************************************************************/
+
+struct driverlink_port {
+	struct list_head link;
+	enum net_accel_hw_type type;
+	struct net_device *net_dev;
+	struct efx_dl_device *efx_dl_dev;
+	int nic_index;
+	void *fwd_priv;
+};
+
+static struct list_head dl_ports;
+
+/* This mutex protects global state, such as the dl_ports list */
+DEFINE_MUTEX(accel_mutex);
+
+static int init_done = 0;
+
+/* The DL callbacks */
+
+
+#if defined(EFX_USE_FASTCALL)
+static enum efx_veto fastcall
+#else
+static enum efx_veto
+#endif
+bend_dl_tx_packet(struct efx_dl_device *efx_dl_dev,
+		  struct sk_buff *skb)
+{
+	struct driverlink_port *port = efx_dl_dev->priv;
+
+	BUG_ON(port == NULL);
+
+	NETBACK_ACCEL_STATS_OP(global_stats.dl_tx_packets++);
+	if (skb->mac.raw != NULL)
+		netback_accel_tx_packet(skb, port->fwd_priv);
+	else {
+		DPRINTK("Ignoring packet with missing mac address\n");
+		NETBACK_ACCEL_STATS_OP(global_stats.dl_tx_bad_packets++);
+	}
+	return EFX_ALLOW_PACKET;
+}
+
+/* EFX_USE_FASTCALL */
+#if defined(EFX_USE_FASTCALL)
+static enum efx_veto fastcall
+#else
+static enum efx_veto
+#endif
+bend_dl_rx_packet(struct efx_dl_device *efx_dl_dev,
+		  const char *pkt_buf, int pkt_len)
+{
+	struct driverlink_port *port = efx_dl_dev->priv;
+	struct netback_pkt_buf pkt;
+	struct ethhdr *eh;
+
+	BUG_ON(port == NULL);
+
+	pkt.mac.raw = (char *)pkt_buf;
+	pkt.nh.raw = (char *)pkt_buf + ETH_HLEN;
+	eh = (struct ethhdr *)pkt_buf;
+	pkt.protocol = eh->h_proto;
+
+	NETBACK_ACCEL_STATS_OP(global_stats.dl_rx_packets++);
+	netback_accel_rx_packet(&pkt, port->fwd_priv);
+	return EFX_ALLOW_PACKET;
+}
+
+
+/* Callbacks we'd like to get from the netdriver through driverlink */
+struct efx_dl_callbacks bend_dl_callbacks =
+	{
+		.tx_packet = bend_dl_tx_packet,
+		.rx_packet = bend_dl_rx_packet,
+	};
+
+
+static struct netback_accel_hooks accel_hooks = {
+	THIS_MODULE,
+	&netback_accel_probe,
+	&netback_accel_remove
+};
+
+
+/*
+ * Handy helper which given an efx_dl_device works out which
+ * efab_nic_t index into efrm_nic_table.nics[] it corresponds to 
+ */
+static int efx_device_to_efab_nic_index(struct efx_dl_device *efx_dl_dev) 
+{
+	int i;
+
+	for (i = 0; i < EFHW_MAX_NR_DEVS; i++) {
+		struct efhw_nic *nic = efrm_nic_table.nic[i];
+
+		/*
+		 * It's possible for the nic structure to have not
+		 * been initialised if the resource driver failed its
+		 * driverlink probe
+		 */ 
+		if (nic == NULL || nic->net_driver_dev == NULL)
+			continue;
+
+		/* Work out if these are talking about the same NIC */
+		if (nic->net_driver_dev->pci_dev == efx_dl_dev->pci_dev)
+			return i;
+	}
+
+	return -1;
+}
+
+
+/* Driver link probe - register our callbacks */
+static int bend_dl_probe(struct efx_dl_device *efx_dl_dev,
+			 const struct net_device *net_dev,
+			 const struct efx_dl_device_info *dev_info,
+			 const char* silicon_rev)
+{
+	int rc;
+	enum net_accel_hw_type type;
+	struct driverlink_port *port;
+
+	DPRINTK("%s: %s\n", __FUNCTION__, silicon_rev);
+
+	if (strcmp(silicon_rev, "falcon/a1") == 0)
+		type = NET_ACCEL_MSG_HWTYPE_FALCON_A;
+	else if (strcmp(silicon_rev, "falcon/b0") == 0)
+		type = NET_ACCEL_MSG_HWTYPE_FALCON_B;
+	else {
+		EPRINTK("%s: unsupported silicon %s\n", __FUNCTION__,
+			silicon_rev);
+		rc = -EINVAL;
+		goto fail1;
+	}
+	
+	port = kmalloc(sizeof(struct driverlink_port), GFP_KERNEL);
+	if (port == NULL) {
+		EPRINTK("%s: no memory for dl probe\n", __FUNCTION__);
+		rc = -ENOMEM;
+		goto fail1;
+	}
+
+	port->efx_dl_dev = efx_dl_dev;
+	efx_dl_dev->priv = port;
+
+	port->nic_index = efx_device_to_efab_nic_index(efx_dl_dev);
+	if (port->nic_index < 0) {
+		/*
+		 * This can happen in theory if the resource driver
+		 * failed to initialise properly
+		 */
+		EPRINTK("%s: nic structure not found\n", __FUNCTION__);
+		rc = -EINVAL;
+		goto fail2;
+	}
+
+	port->fwd_priv = netback_accel_init_fwd_port();
+	if (port->fwd_priv == NULL) {
+		EPRINTK("%s: failed to set up forwarding for port\n",
+			__FUNCTION__);
+		rc = -ENOMEM;
+		goto fail2;
+	}
+
+	rc = efx_dl_register_callbacks(efx_dl_dev, &bend_dl_callbacks);
+	if (rc != 0) {
+		EPRINTK("%s: register_callbacks failed\n", __FUNCTION__);
+		goto fail3;
+	}
+
+	port->type = type;
+	port->net_dev = (struct net_device *)net_dev;
+
+	mutex_lock(&accel_mutex);
+	list_add(&port->link, &dl_ports);
+	mutex_unlock(&accel_mutex);
+
+	rc = netback_connect_accelerator(NETBACK_ACCEL_VERSION, 0,
+					 port->net_dev->name, &accel_hooks);
+
+	if (rc < 0) {
+		EPRINTK("Xen netback accelerator version mismatch\n");
+		goto fail4;
+	} else if (rc > 0) {
+		/*
+		 * In future may want to add backwards compatibility
+		 * and accept certain subsets of previous versions
+		 */
+		EPRINTK("Xen netback accelerator version mismatch\n");
+		goto fail4;
+	} 
+
+	return 0;
+
+ fail4:
+	mutex_lock(&accel_mutex);
+	list_del(&port->link);
+	mutex_unlock(&accel_mutex);
+
+	efx_dl_unregister_callbacks(efx_dl_dev, &bend_dl_callbacks);
+ fail3: 
+	netback_accel_shutdown_fwd_port(port->fwd_priv);
+ fail2:
+	efx_dl_dev->priv = NULL;
+	kfree(port);
+ fail1:
+	return rc;
+}
+
+
+static void bend_dl_remove(struct efx_dl_device *efx_dl_dev)
+{
+	struct driverlink_port *port;
+
+	DPRINTK("Unregistering driverlink callbacks.\n");
+
+	mutex_lock(&accel_mutex);
+
+	port = (struct driverlink_port *)efx_dl_dev->priv;
+
+	BUG_ON(list_empty(&dl_ports));
+	BUG_ON(port == NULL);
+	BUG_ON(port->efx_dl_dev != efx_dl_dev);
+
+	netback_disconnect_accelerator(0, port->net_dev->name);
+
+	list_del(&port->link);
+
+	mutex_unlock(&accel_mutex);
+
+	efx_dl_unregister_callbacks(efx_dl_dev, &bend_dl_callbacks);
+	netback_accel_shutdown_fwd_port(port->fwd_priv);
+
+	efx_dl_dev->priv = NULL;
+	kfree(port);
+
+	return;
+}
+
+
+static struct efx_dl_driver bend_dl_driver = 
+	{
+		.name = "SFC Xen backend",
+		.probe = bend_dl_probe,
+		.remove = bend_dl_remove,
+	};
+
+
+int netback_accel_sf_init(void)
+{
+	int rc, nic_i;
+	struct efhw_nic *nic;
+
+	INIT_LIST_HEAD(&dl_ports);
+
+	rc = efx_dl_register_driver(&bend_dl_driver);
+	/* If we couldn't find the NET driver, give up */
+	if (rc == -ENOENT)
+		return rc;
+	
+	if (rc == 0) {
+		EFRM_FOR_EACH_NIC(nic_i, nic)
+			falcon_nic_set_rx_usr_buf_size(nic, 
+						       SF_XEN_RX_USR_BUF_SIZE);
+	}
+
+	init_done = (rc == 0);
+	return rc;
+}
+
+
+void netback_accel_sf_shutdown(void)
+{
+	if (!init_done)
+		return;
+	DPRINTK("Unregistering driverlink driver\n");
+
+	/*
+	 * This will trigger removal callbacks for all the devices, which
+	 * will unregister their callbacks, disconnect from netfront, etc.
+	 */
+	efx_dl_unregister_driver(&bend_dl_driver);
+}
+
+
+int netback_accel_sf_hwtype(struct netback_accel *bend)
+{
+	struct driverlink_port *port;
+
+	mutex_lock(&accel_mutex);
+
+	list_for_each_entry(port, &dl_ports, link) {
+		if (strcmp(bend->nicname, port->net_dev->name) == 0) {
+			bend->hw_type = port->type;
+			bend->accel_setup = netback_accel_setup_vnic_hw;
+			bend->accel_shutdown = netback_accel_shutdown_vnic_hw;
+			bend->fwd_priv = port->fwd_priv;
+			/* This is just needed to pass to efx_vi_alloc */
+			bend->nic_index = port->nic_index;
+			bend->net_dev = port->net_dev;
+			mutex_unlock(&accel_mutex);
+			return 0;
+		}
+	}
+
+	mutex_unlock(&accel_mutex);
+
+	EPRINTK("Failed to identify backend device '%s' with a NIC\n",
+		bend->nicname);
+
+	return -ENOENT;
+}
+
+
+/****************************************************************************
+ * Resource management code
+ ***************************************************************************/
+
+static int alloc_page_state(struct netback_accel *bend, int max_pages)
+{
+	struct falcon_bend_accel_priv *accel_hw_priv;
+
+	if (max_pages < 0 || max_pages > bend->quotas.max_buf_pages) {
+		EPRINTK("%s: invalid max_pages: %d\n", __FUNCTION__, max_pages);
+		return -EINVAL;
+	}
+
+	accel_hw_priv = kzalloc(sizeof(struct falcon_bend_accel_priv),
+				GFP_KERNEL);
+	if (accel_hw_priv == NULL) {
+		EPRINTK("%s: no memory for accel_hw_priv\n", __FUNCTION__);
+		return -ENOMEM;
+	}
+
+	accel_hw_priv->dma_maps = kzalloc
+		(sizeof(struct efx_vi_dma_map_state **) * 
+		 (max_pages / NET_ACCEL_MSG_MAX_PAGE_REQ), GFP_KERNEL);
+	if (accel_hw_priv->dma_maps == NULL) {
+		EPRINTK("%s: no memory for dma_maps\n", __FUNCTION__);
+		kfree(accel_hw_priv);
+		return -ENOMEM;
+	}
+
+	bend->buffer_maps = kzalloc(sizeof(struct vm_struct *) * max_pages, 
+				    GFP_KERNEL);
+	if (bend->buffer_maps == NULL) {
+		EPRINTK("%s: no memory for buffer_maps\n", __FUNCTION__);
+		kfree(accel_hw_priv->dma_maps);
+		kfree(accel_hw_priv);
+		return -ENOMEM;
+	}
+
+	bend->buffer_addrs = kzalloc(sizeof(u64) * max_pages, GFP_KERNEL);
+	if (bend->buffer_addrs == NULL) {
+		kfree(bend->buffer_maps);
+		kfree(accel_hw_priv->dma_maps);
+		kfree(accel_hw_priv);
+		return -ENOMEM;
+	}
+
+	bend->accel_hw_priv = accel_hw_priv;
+
+	return 0;
+}
+
+
+static int free_page_state(struct netback_accel *bend)
+{
+	struct falcon_bend_accel_priv *accel_hw_priv;
+
+	DPRINTK("%s: %p\n", __FUNCTION__, bend);
+
+	accel_hw_priv = bend->accel_hw_priv;
+
+	if (accel_hw_priv) {
+		kfree(accel_hw_priv->dma_maps);
+		kfree(bend->buffer_maps);
+		kfree(bend->buffer_addrs);
+		kfree(accel_hw_priv);
+		bend->accel_hw_priv = NULL;
+		bend->max_pages = 0;
+	}
+
+	return 0;
+}
+
+
+/* The timeout event callback for the event q */
+static void bend_evq_timeout(void *context, int is_timeout)
+{
+	struct netback_accel *bend = (struct netback_accel *)context;
+	if (is_timeout) {
+		/* Pass event to vnic front end driver */
+		VPRINTK("timeout event to %d\n", bend->net_channel);
+		NETBACK_ACCEL_STATS_OP(bend->stats.evq_timeouts++);
+		notify_remote_via_irq(bend->net_channel_irq);
+	} else {
+		/* It's a wakeup event, used by Falcon */
+		VPRINTK("wakeup to %d\n", bend->net_channel);
+		NETBACK_ACCEL_STATS_OP(bend->stats.evq_wakeups++);
+		notify_remote_via_irq(bend->net_channel_irq);
+	}
+}
+
+
+/*
+ * Create the eventq and associated gubbins for communication with the
+ * front end vnic driver
+ */
+static int ef_get_vnic(struct netback_accel *bend)
+{
+	struct falcon_bend_accel_priv *accel_hw_priv;
+	int rc = 0;
+
+	BUG_ON(bend->hw_state != NETBACK_ACCEL_RES_NONE);
+
+	/* Allocate page related state and accel_hw_priv */
+	rc = alloc_page_state(bend, bend->max_pages);
+	if (rc != 0) {
+		EPRINTK("Failed to allocate page state: %d\n", rc);
+		return rc;
+	}
+
+	accel_hw_priv = bend->accel_hw_priv;
+
+	rc = efx_vi_alloc(&accel_hw_priv->efx_vih, bend->nic_index);
+	if (rc != 0) {
+		EPRINTK("%s: efx_vi_alloc failed %d\n", __FUNCTION__, rc);
+		free_page_state(bend);
+		return rc;
+	}
+
+	rc = efx_vi_eventq_register_callback(accel_hw_priv->efx_vih,
+					     bend_evq_timeout,
+					     bend);
+	if (rc != 0) {
+		EPRINTK("%s: register_callback failed %d\n", __FUNCTION__, rc);
+		efx_vi_free(accel_hw_priv->efx_vih);
+		free_page_state(bend);
+		return rc;
+	}
+
+	bend->hw_state = NETBACK_ACCEL_RES_ALLOC;
+	
+	return 0;
+}
+
+
+static void ef_free_vnic(struct netback_accel *bend)
+{
+	struct falcon_bend_accel_priv *accel_hw_priv = bend->accel_hw_priv;
+
+	BUG_ON(bend->hw_state != NETBACK_ACCEL_RES_ALLOC);
+
+	efx_vi_eventq_kill_callback(accel_hw_priv->efx_vih);
+
+	DPRINTK("Hardware is freeable. Will proceed.\n");
+
+	efx_vi_free(accel_hw_priv->efx_vih);
+	accel_hw_priv->efx_vih = NULL;
+
+	VPRINTK("Free page state...\n");
+	free_page_state(bend);
+
+	bend->hw_state = NETBACK_ACCEL_RES_NONE;
+}
+
+
+static inline void ungrant_or_crash(grant_ref_t gntref, int domain) {
+	if (net_accel_ungrant_page(gntref) == -EBUSY)
+		net_accel_shutdown_remote(domain);
+}
+
+
+static void netback_accel_release_hwinfo(struct netback_accel *bend)
+{
+	struct falcon_bend_accel_priv *accel_hw_priv = bend->accel_hw_priv;
+	int i;
+
+	DPRINTK("Remove dma q grants %d %d\n", accel_hw_priv->txdmaq_gnt,
+		accel_hw_priv->rxdmaq_gnt);
+	ungrant_or_crash(accel_hw_priv->txdmaq_gnt, bend->far_end);
+	ungrant_or_crash(accel_hw_priv->rxdmaq_gnt, bend->far_end);
+
+	DPRINTK("Remove doorbell grant %d\n", accel_hw_priv->doorbell_gnt);
+	ungrant_or_crash(accel_hw_priv->doorbell_gnt, bend->far_end);
+
+	if (bend->hw_type == NET_ACCEL_MSG_HWTYPE_FALCON_A) {
+		DPRINTK("Remove rptr grant %d\n", accel_hw_priv->evq_rptr_gnt);
+		ungrant_or_crash(accel_hw_priv->evq_rptr_gnt, bend->far_end);
+	}
+
+	for (i = 0; i < accel_hw_priv->evq_npages; i++) {
+		DPRINTK("Remove evq grant %d\n", accel_hw_priv->evq_mem_gnts[i]);
+		ungrant_or_crash(accel_hw_priv->evq_mem_gnts[i], bend->far_end);
+	}
+
+	bend->hw_state = NETBACK_ACCEL_RES_FILTER;
+
+	return;
+}
+
+
+static int ef_bend_hwinfo_falcon_common(struct netback_accel *bend, 
+					struct net_accel_hw_falcon_b *hwinfo)
+{
+	struct falcon_bend_accel_priv *accel_hw_priv = bend->accel_hw_priv;
+	struct efx_vi_hw_resource_metadata res_mdata;
+	struct efx_vi_hw_resource res_array[EFX_VI_HW_RESOURCE_MAXSIZE];
+	int rc, len = EFX_VI_HW_RESOURCE_MAXSIZE, i, pfn = 0;
+	unsigned long txdmaq_pfn = 0, rxdmaq_pfn = 0;
+
+	rc = efx_vi_hw_resource_get_phys(accel_hw_priv->efx_vih, &res_mdata,
+					 res_array, &len);
+	if (rc != 0) {
+		DPRINTK("%s: resource_get_phys returned %d\n",
+			__FUNCTION__, rc);
+		return rc;
+	}
+
+	if (res_mdata.version != 0)
+		return -EPROTO;
+
+	hwinfo->nic_arch = res_mdata.nic_arch;
+	hwinfo->nic_variant = res_mdata.nic_variant;
+	hwinfo->nic_revision = res_mdata.nic_revision;
+
+	hwinfo->evq_order = res_mdata.evq_order;
+	hwinfo->evq_offs = res_mdata.evq_offs;
+	hwinfo->evq_capacity = res_mdata.evq_capacity;
+	hwinfo->instance = res_mdata.instance;
+	hwinfo->rx_capacity = res_mdata.rx_capacity;
+	hwinfo->tx_capacity = res_mdata.tx_capacity;
+
+	VPRINTK("evq_order %d evq_offs %d evq_cap %d inst %d rx_cap %d tx_cap %d\n",
+		hwinfo->evq_order, hwinfo->evq_offs, hwinfo->evq_capacity,
+		hwinfo->instance, hwinfo->rx_capacity, hwinfo->tx_capacity);
+
+	for (i = 0; i < len; i++) {
+		struct efx_vi_hw_resource *res = &(res_array[i]);
+		switch (res->type) {
+		case EFX_VI_HW_RESOURCE_TXDMAQ:
+			txdmaq_pfn = page_to_pfn(virt_to_page(res->address));
+			break;
+		case EFX_VI_HW_RESOURCE_RXDMAQ: 
+			rxdmaq_pfn = page_to_pfn(virt_to_page(res->address));
+			break;
+		case EFX_VI_HW_RESOURCE_EVQTIMER:
+			break;
+		case EFX_VI_HW_RESOURCE_EVQRPTR:
+		case EFX_VI_HW_RESOURCE_EVQRPTR_OFFSET:
+			hwinfo->evq_rptr = res->address;
+			break;
+		case EFX_VI_HW_RESOURCE_EVQMEMKVA: 
+			accel_hw_priv->evq_npages =  1 << res_mdata.evq_order;
+			pfn = page_to_pfn(virt_to_page(res->address));
+			break;
+		case EFX_VI_HW_RESOURCE_BELLPAGE:
+			hwinfo->doorbell_mfn  = res->address;
+			break;
+		default:
+			EPRINTK("%s: Unknown hardware resource type %d\n",
+				__FUNCTION__, res->type);
+			break;
+		}
+	}
+
+	VPRINTK("Passing txdmaq page pfn %lx\n", txdmaq_pfn);
+	accel_hw_priv->txdmaq_gnt = hwinfo->txdmaq_gnt = 
+		net_accel_grant_page(bend->hdev_data, pfn_to_mfn(txdmaq_pfn), 
+				     0);
+
+	VPRINTK("Passing rxdmaq page pfn %lx\n", rxdmaq_pfn);
+	accel_hw_priv->rxdmaq_gnt = hwinfo->rxdmaq_gnt = 
+		net_accel_grant_page(bend->hdev_data, pfn_to_mfn(rxdmaq_pfn), 
+				     0);
+
+	VPRINTK("Passing doorbell page mfn %x\n", hwinfo->doorbell_mfn);
+	/* Make the relevant H/W pages mappable by the far end */
+	accel_hw_priv->doorbell_gnt = hwinfo->doorbell_gnt = 
+		net_accel_grant_page(bend->hdev_data, hwinfo->doorbell_mfn, 1);
+	
+	/* Now do the same for the memory pages */
+	/* Convert the page + length we got back for the evq to grants. */
+	for (i = 0; i < accel_hw_priv->evq_npages; i++) {
+		accel_hw_priv->evq_mem_gnts[i] = hwinfo->evq_mem_gnts[i] =
+			net_accel_grant_page(bend->hdev_data, pfn_to_mfn(pfn), 0);
+		VPRINTK("Got grant %u for evq pfn %x\n", hwinfo->evq_mem_gnts[i], 
+			pfn);
+		pfn++;
+	}
+
+	return 0;
+}
+
+
+static int ef_bend_hwinfo_falcon_a(struct netback_accel *bend, 
+				   struct net_accel_hw_falcon_a *hwinfo)
+{
+	int rc;
+	struct falcon_bend_accel_priv *accel_hw_priv = bend->accel_hw_priv;
+
+	if ((rc = ef_bend_hwinfo_falcon_common(bend, &hwinfo->common)) != 0)
+		return rc;
+
+	/*
+	 * Note that unlike the above, where the message field is the
+	 * page number, here evq_rptr is the entire address because
+	 * it is currently a pointer into the densely mapped timer page.
+	 */
+	VPRINTK("Passing evq_rptr pfn %x for rptr %x\n", 
+		hwinfo->common.evq_rptr >> PAGE_SHIFT,
+		hwinfo->common.evq_rptr);
+	rc = net_accel_grant_page(bend->hdev_data, 
+				  hwinfo->common.evq_rptr >> PAGE_SHIFT, 0);
+	if (rc < 0)
+		return rc;
+
+	accel_hw_priv->evq_rptr_gnt = hwinfo->evq_rptr_gnt = rc;
+	VPRINTK("evq_rptr_gnt got %d\n", hwinfo->evq_rptr_gnt);
+	
+	return 0;
+}
+
+
+static int ef_bend_hwinfo_falcon_b(struct netback_accel *bend, 
+				   struct net_accel_hw_falcon_b *hwinfo)
+{
+	return ef_bend_hwinfo_falcon_common(bend, hwinfo);
+}
+
+
+/*
+ * Fill in the message with a description of the hardware resources, based on
+ * the H/W type
+ */
+static int netback_accel_hwinfo(struct netback_accel *bend, 
+				struct net_accel_msg_hw *msgvi)
+{
+	int rc = 0;
+	
+	BUG_ON(bend->hw_state != NETBACK_ACCEL_RES_FILTER);
+
+	msgvi->type = bend->hw_type;
+	switch (bend->hw_type) {
+	case NET_ACCEL_MSG_HWTYPE_FALCON_A:
+		rc = ef_bend_hwinfo_falcon_a(bend, &msgvi->resources.falcon_a);
+		break;
+	case NET_ACCEL_MSG_HWTYPE_FALCON_B:
+		rc = ef_bend_hwinfo_falcon_b(bend, &msgvi->resources.falcon_b);
+		break;
+	case NET_ACCEL_MSG_HWTYPE_NONE:
+		/* Nothing to do. The slow path should just work. */
+		break;
+	}
+
+	if (rc == 0)
+		bend->hw_state = NETBACK_ACCEL_RES_HWINFO;
+		
+	return rc;
+}
+
+
+/* Allocate hardware resources and make them available to the client domain */
+int netback_accel_setup_vnic_hw(struct netback_accel *bend)
+{
+	struct net_accel_msg msg;
+	int err;
+
+	/* Allocate the event queue, VI and so on. */
+	err = ef_get_vnic(bend);
+	if (err) {
+		EPRINTK("Failed to allocate hardware resource for bend:"
+			"error %d\n", err);
+		return err;
+	}
+
+	/* Set up the filter management */
+	err = netback_accel_filter_init(bend);
+	if (err) {
+		EPRINTK("Filter setup failed, error %d", err);
+		ef_free_vnic(bend);
+		return err;
+	}
+
+	net_accel_msg_init(&msg, NET_ACCEL_MSG_SETHW);
+
+	/*
+	 * Extract the low-level hardware info we will actually pass to the
+	 * other end, and set up the grants/ioremap permissions needed
+	 */
+	err = netback_accel_hwinfo(bend, &msg.u.hw);
+
+	if (err != 0) {
+		netback_accel_filter_shutdown(bend);
+		ef_free_vnic(bend);
+		return err;
+	}
+
+	/* Send the message, this is a reply to a hello-reply */
+	err = net_accel_msg_reply_notify(bend->shared_page, 
+					 bend->msg_channel_irq, 
+					 &bend->to_domU, &msg);
+
+	/*
+	 * The message should succeed as it's logically a reply and we
+	 * guarantee space for replies, but a misbehaving frontend
+	 * could result in that behaviour, so be tolerant
+	 */
+	if (err != 0) {
+		netback_accel_release_hwinfo(bend);
+		netback_accel_filter_shutdown(bend);
+		ef_free_vnic(bend);
+	}
+
+	return err;
+}
+
+
+/* Free hardware resources  */
+void netback_accel_shutdown_vnic_hw(struct netback_accel *bend)
+{
+	/*
+	 * Only try and release resources if accel_hw_priv was setup,
+	 * otherwise there is nothing to do as we're on "null-op"
+	 * acceleration
+	 */
+	switch (bend->hw_state) {
+	case NETBACK_ACCEL_RES_HWINFO:
+		VPRINTK("Release hardware resources\n");
+		netback_accel_release_hwinfo(bend);
+		/* deliberate drop through */
+	case NETBACK_ACCEL_RES_FILTER:		
+		VPRINTK("Free filters...\n");
+		netback_accel_filter_shutdown(bend);
+		/* deliberate drop through */
+	case NETBACK_ACCEL_RES_ALLOC:
+		VPRINTK("Free vnic...\n");
+		ef_free_vnic(bend);
+		/* deliberate drop through */
+	case NETBACK_ACCEL_RES_NONE:
+		break;
+	default:
+		BUG();
+	}
+}
+
+/**************************************************************************
+ * 
+ * Buffer table stuff
+ *
+ **************************************************************************/
+
+/*
+ * Undo any allocation that netback_accel_msg_rx_buffer_map() has made
+ * if it fails half way through
+ */
+static inline void buffer_map_cleanup(struct netback_accel *bend, int i)
+{
+	while (i > 0) {
+		i--;
+		bend->buffer_maps_index--;
+		net_accel_unmap_device_page(bend->hdev_data, 
+					    bend->buffer_maps[bend->buffer_maps_index],
+					    bend->buffer_addrs[bend->buffer_maps_index]);
+	}
+}
+
+
+int netback_accel_add_buffers(struct netback_accel *bend, int pages, int log2_pages,
+			      u32 *grants, u32 *buf_addr_out)
+{
+	struct falcon_bend_accel_priv *accel_hw_priv = bend->accel_hw_priv;
+	unsigned long long addr_array[NET_ACCEL_MSG_MAX_PAGE_REQ];
+	int rc, i, index;
+	u64 dev_bus_addr;
+
+	/* Make sure we can't overflow the dma_maps array */
+	if (accel_hw_priv->dma_maps_index >= 
+	    bend->max_pages / NET_ACCEL_MSG_MAX_PAGE_REQ) {
+		EPRINTK("%s: too many buffer table allocations: %d %d\n",
+			__FUNCTION__, accel_hw_priv->dma_maps_index, 
+			bend->max_pages / NET_ACCEL_MSG_MAX_PAGE_REQ);
+		return -EINVAL;
+	}
+
+	/* Make sure we can't overflow the buffer_maps array */
+	if (bend->buffer_maps_index + pages > bend->max_pages) {
+		EPRINTK("%s: too many pages mapped: %d + %d > %d\n", 
+			__FUNCTION__, bend->buffer_maps_index,
+			pages, bend->max_pages);
+		return -EINVAL;
+	}
+
+	for (i = 0; i < pages; i++) {
+		VPRINTK("%s: mapping page %d\n", __FUNCTION__, i);
+		rc = net_accel_map_device_page
+			(bend->hdev_data, grants[i],
+			 &bend->buffer_maps[bend->buffer_maps_index],
+			 &dev_bus_addr);
+    
+		if (rc != 0) {
+			EPRINTK("error in net_accel_map_device_page\n");
+			buffer_map_cleanup(bend, i);
+			return rc;
+		}
+		
+		bend->buffer_addrs[bend->buffer_maps_index] = dev_bus_addr;
+
+		bend->buffer_maps_index++;
+
+		addr_array[i] = dev_bus_addr;
+	}
+
+	VPRINTK("%s: mapping dma addresses to vih %p\n", __FUNCTION__, 
+		accel_hw_priv->efx_vih);
+
+	index = accel_hw_priv->dma_maps_index;
+	if ((rc = efx_vi_dma_map_addrs(accel_hw_priv->efx_vih, addr_array, pages,
+				       &(accel_hw_priv->dma_maps[index]))) < 0) {
+		EPRINTK("error in dma_map_pages\n");
+		buffer_map_cleanup(bend, i);
+		return rc;
+	}
+
+	accel_hw_priv->dma_maps_index++;
+	NETBACK_ACCEL_STATS_OP(bend->stats.num_buffer_pages += pages);
+
+	//DPRINTK("%s: getting map address\n", __FUNCTION__);
+
+	*buf_addr_out = efx_vi_dma_get_map_addr(accel_hw_priv->efx_vih, 
+						accel_hw_priv->dma_maps[index]);
+
+	//DPRINTK("%s: done\n", __FUNCTION__);
+
+	return 0;
+}
+
+
+int netback_accel_remove_buffers(struct netback_accel *bend)
+{
+	/* Only try to free buffers if accel_hw_priv was setup */
+	if (bend->hw_state != NETBACK_ACCEL_RES_NONE) {
+		struct falcon_bend_accel_priv *accel_hw_priv = bend->accel_hw_priv;
+		int i;
+
+		efx_vi_reset(accel_hw_priv->efx_vih);
+
+		while (accel_hw_priv->dma_maps_index > 0) {
+			accel_hw_priv->dma_maps_index--;
+			i = accel_hw_priv->dma_maps_index;
+			efx_vi_dma_unmap_addrs(accel_hw_priv->efx_vih, 
+					       accel_hw_priv->dma_maps[i]);
+		}
+		
+		while (bend->buffer_maps_index > 0) {
+			VPRINTK("Unmapping granted buffer %d\n", 
+				bend->buffer_maps_index);
+			bend->buffer_maps_index--;
+			i = bend->buffer_maps_index;
+			net_accel_unmap_device_page(bend->hdev_data, 
+						    bend->buffer_maps[i],
+						    bend->buffer_addrs[i]);
+		}
+
+		NETBACK_ACCEL_STATS_OP(bend->stats.num_buffer_pages = 0);
+	}
+
+	return 0;
+}
+
+/**************************************************************************
+ * 
+ * Filter stuff
+ *
+ **************************************************************************/
+
+static int netback_accel_filter_init(struct netback_accel *bend)
+{
+	struct falcon_bend_accel_priv *accel_hw_priv = bend->accel_hw_priv;
+	int i, rc;
+
+	BUG_ON(bend->hw_state != NETBACK_ACCEL_RES_ALLOC);
+
+	spin_lock_init(&accel_hw_priv->filter_lock);
+
+	if ((rc = cuckoo_hash_init(&accel_hw_priv->filter_hash_table, 
+				   5 /* space for 32 filters */, 8)) != 0) {
+		EPRINTK("Failed to initialise filter hash table\n");
+		return rc;
+	}
+
+	accel_hw_priv->fspecs = kzalloc(sizeof(struct netback_accel_filter_spec) *
+					bend->quotas.max_filters,
+					GFP_KERNEL);
+
+	if (accel_hw_priv->fspecs == NULL) {
+		EPRINTK("No memory for filter specs.\n");
+		cuckoo_hash_destroy(&accel_hw_priv->filter_hash_table);
+		return -ENOMEM;
+	}
+
+	for (i = 0; i < bend->quotas.max_filters; i++) {
+		accel_hw_priv->free_filters |= (1 << i);
+	}
+
+	/* Base mask on highest set bit in max_filters  */
+	accel_hw_priv->filter_idx_mask = (1 << fls(bend->quotas.max_filters)) - 1;
+	VPRINTK("filter setup: max is %x mask is %x\n",
+		bend->quotas.max_filters, accel_hw_priv->filter_idx_mask);
+
+	bend->hw_state = NETBACK_ACCEL_RES_FILTER;
+
+	return 0;
+}
+
+
+static inline void make_filter_key(cuckoo_hash_ip_key *key,  
+				   struct netback_accel_filter_spec *filt)
+
+{
+	key->local_ip = filt->destip_be;
+	key->local_port = filt->destport_be;
+	key->proto = filt->proto;
+}
+
+
+static inline 
+void netback_accel_free_filter(struct falcon_bend_accel_priv *accel_hw_priv,
+			       int filter)
+{
+	cuckoo_hash_ip_key filter_key;
+
+	if (!(accel_hw_priv->free_filters & (1 << filter))) {
+		efx_vi_filter_stop(accel_hw_priv->efx_vih, 
+				   accel_hw_priv->fspecs[filter].filter_handle);
+		make_filter_key(&filter_key, &(accel_hw_priv->fspecs[filter]));
+		if (cuckoo_hash_remove(&accel_hw_priv->filter_hash_table,
+				       (cuckoo_hash_key *)&filter_key)) {
+			EPRINTK("%s: Couldn't find filter to remove from table\n",
+				__FUNCTION__);
+			BUG();
+		}
+	}
+}
+
+
+static void netback_accel_filter_shutdown(struct netback_accel *bend)
+{
+	struct falcon_bend_accel_priv *accel_hw_priv = bend->accel_hw_priv;
+	int i;
+	unsigned long flags;
+
+	BUG_ON(bend->hw_state != NETBACK_ACCEL_RES_FILTER);
+
+	spin_lock_irqsave(&accel_hw_priv->filter_lock, flags);
+
+	BUG_ON(accel_hw_priv->fspecs == NULL);
+
+	for (i = 0; i < bend->quotas.max_filters; i++) {
+		netback_accel_free_filter(accel_hw_priv, i);
+	}
+	
+	kfree(accel_hw_priv->fspecs);
+	accel_hw_priv->fspecs = NULL;
+	accel_hw_priv->free_filters = 0;
+	
+	cuckoo_hash_destroy(&accel_hw_priv->filter_hash_table);
+
+	spin_unlock_irqrestore(&accel_hw_priv->filter_lock, flags);
+
+	bend->hw_state = NETBACK_ACCEL_RES_ALLOC;
+}
+
+
+/*! Suggest a filter to replace when we want to insert a new one and have
+ *  none free.
+ */
+static unsigned get_victim_filter(struct netback_accel *bend)
+{
+	/*
+	 * We could attempt to get really clever, and may do at some
+	 * point, but random replacement is v. cheap and low on
+	 * pathological worst cases.
+	 */
+	unsigned index, cycles;
+
+	rdtscl(cycles);
+
+	/*
+	 * Some doubt about the quality of the bottom few bits, so
+	 * throw 'em * away
+	 */
+	index = (cycles >> 4) & ((struct falcon_bend_accel_priv *)
+				 bend->accel_hw_priv)->filter_idx_mask;
+	/*
+	 * We don't enforce that the number of filters is a power of
+	 * two, but the masking gets us to within one subtraction of a
+	 * valid index
+	 */
+	if (index >= bend->quotas.max_filters)
+		index -= bend->quotas.max_filters;
+	DPRINTK("backend %s->%d has no free filters. Filter %d will be evicted\n",
+		bend->nicname, bend->far_end, index);
+	return index;
+}
+
+
+/* Add a filter for the specified IP/port to the backend */
+int 
+netback_accel_filter_check_add(struct netback_accel *bend, 
+			       struct netback_accel_filter_spec *filt)
+{
+	struct falcon_bend_accel_priv *accel_hw_priv = bend->accel_hw_priv;
+	struct netback_accel_filter_spec *fs;
+	unsigned filter_index;
+	unsigned long flags;
+	int rc, recycling = 0;
+	cuckoo_hash_ip_key filter_key, evict_key;
+
+	BUG_ON(filt->proto != IPPROTO_TCP && filt->proto != IPPROTO_UDP);
+
+	DPRINTK("Will add %s filter for dst ip %08x and dst port %d\n", 
+		(filt->proto == IPPROTO_TCP) ? "TCP" : "UDP",
+		be32_to_cpu(filt->destip_be), be16_to_cpu(filt->destport_be));
+
+	spin_lock_irqsave(&accel_hw_priv->filter_lock, flags);
+	/*
+	 * Check to see if we're already filtering this IP address and
+	 * port. Happens if you insert a filter mid-stream as there
+	 * are many packets backed up to be delivered to dom0 already
+	 */
+	make_filter_key(&filter_key, filt);
+	if (cuckoo_hash_lookup(&accel_hw_priv->filter_hash_table, 
+			       (cuckoo_hash_key *)(&filter_key), 
+			       &filter_index)) {
+		DPRINTK("Found matching filter %d already in table\n", 
+			filter_index);
+		rc = -1;
+		goto out;
+	}
+
+	if (accel_hw_priv->free_filters == 0) {
+		filter_index = get_victim_filter(bend);
+		recycling = 1;
+	} else {
+		filter_index = __ffs(accel_hw_priv->free_filters);
+		clear_bit(filter_index, &accel_hw_priv->free_filters);
+	}
+
+	fs = &accel_hw_priv->fspecs[filter_index];
+
+	if (recycling) {
+		DPRINTK("Removing filter index %d handle %p\n", filter_index,
+			fs->filter_handle);
+
+		if ((rc = efx_vi_filter_stop(accel_hw_priv->efx_vih, 
+					     fs->filter_handle)) != 0) {
+			EPRINTK("Couldn't clear NIC filter table entry %d\n", rc);
+		}
+
+		make_filter_key(&evict_key, fs);
+		if (cuckoo_hash_remove(&accel_hw_priv->filter_hash_table,
+				       (cuckoo_hash_key *)&evict_key)) {
+			EPRINTK("Couldn't find filter to remove from table\n");
+			BUG();
+		}
+		NETBACK_ACCEL_STATS_OP(bend->stats.num_filters--);
+	}
+
+	/* Update the filter spec with new details */
+	*fs = *filt;
+
+	if ((rc = cuckoo_hash_add(&accel_hw_priv->filter_hash_table, 
+				  (cuckoo_hash_key *)&filter_key, filter_index,
+				  1)) != 0) {
+		EPRINTK("Error (%d) adding filter to table\n", rc);
+		accel_hw_priv->free_filters |= (1 << filter_index);
+		goto out;
+	}
+
+	rc = efx_vi_filter(accel_hw_priv->efx_vih, filt->proto, filt->destip_be,
+			   filt->destport_be, 
+			   (struct filter_resource_t **)&fs->filter_handle);
+
+	if (rc != 0) {
+		EPRINTK("Hardware filter insertion failed. Error %d\n", rc);
+		accel_hw_priv->free_filters |= (1 << filter_index);
+		cuckoo_hash_remove(&accel_hw_priv->filter_hash_table, 
+				   (cuckoo_hash_key *)&filter_key);
+		rc = -1;
+		goto out;
+	}
+
+	NETBACK_ACCEL_STATS_OP(bend->stats.num_filters++);
+
+	VPRINTK("%s: success index %d handle %p\n", __FUNCTION__, filter_index, 
+		fs->filter_handle);
+
+	rc = filter_index;
+ out:
+	spin_unlock_irqrestore(&accel_hw_priv->filter_lock, flags);
+	return rc;
+}
+
+
+/* Remove a filter entry for the specific device and IP/port */
+static void netback_accel_filter_remove(struct netback_accel *bend, 
+					int filter_index)
+{
+	struct falcon_bend_accel_priv *accel_hw_priv = bend->accel_hw_priv;
+
+	BUG_ON(accel_hw_priv->free_filters & (1 << filter_index));
+	netback_accel_free_filter(accel_hw_priv, filter_index);
+	accel_hw_priv->free_filters |= (1 << filter_index);
+}
+
+
+/* Remove a filter entry for the specific device and IP/port */
+void netback_accel_filter_remove_spec(struct netback_accel *bend, 
+				      struct netback_accel_filter_spec *filt)
+{
+	struct falcon_bend_accel_priv *accel_hw_priv = bend->accel_hw_priv;
+	unsigned filter_found;
+	unsigned long flags;
+	cuckoo_hash_ip_key filter_key;
+	struct netback_accel_filter_spec *fs;
+
+	if (filt->proto == IPPROTO_TCP) {
+		DPRINTK("Remove TCP filter for dst ip %08x and dst port %d\n",
+			be32_to_cpu(filt->destip_be),
+			be16_to_cpu(filt->destport_be));
+	} else if (filt->proto == IPPROTO_UDP) {
+		DPRINTK("Remove UDP filter for dst ip %08x and dst port %d\n",
+			be32_to_cpu(filt->destip_be),
+			be16_to_cpu(filt->destport_be));
+	} else {
+		/*
+		 * This could be provoked by an evil frontend, so can't
+		 * BUG(), but harmless as it should fail tests below 
+		 */
+		DPRINTK("Non-TCP/UDP filter dst ip %08x and dst port %d\n",
+			be32_to_cpu(filt->destip_be),
+			be16_to_cpu(filt->destport_be));
+	}
+
+	spin_lock_irqsave(&accel_hw_priv->filter_lock, flags);
+
+	make_filter_key(&filter_key, filt);
+	if (!cuckoo_hash_lookup(&accel_hw_priv->filter_hash_table, 
+			       (cuckoo_hash_key *)(&filter_key), 
+			       &filter_found)) {
+		EPRINTK("Couldn't find matching filter already in table\n");
+		goto out;
+	}
+	
+	/* Do a full check to make sure we've not had a hash collision */
+	fs = &accel_hw_priv->fspecs[filter_found];
+	if (fs->destip_be == filt->destip_be &&
+	    fs->destport_be == filt->destport_be &&
+	    fs->proto == filt->proto &&
+	    !memcmp(fs->mac, filt->mac, ETH_ALEN)) {
+		netback_accel_filter_remove(bend, filter_found);
+	} else {
+		EPRINTK("Entry in hash table does not match filter spec\n");
+		goto out;
+	}
+
+ out:
+	spin_unlock_irqrestore(&accel_hw_priv->filter_lock, flags);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/accel_solarflare.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/accel_solarflare.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/accel_solarflare.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/accel_solarflare.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,88 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef NETBACK_ACCEL_SOLARFLARE_H
+#define NETBACK_ACCEL_SOLARFLARE_H
+
+#include "accel.h"
+#include "accel_msg_iface.h"
+
+#include "driverlink_api.h"
+
+#define MAX_NICS 5
+#define MAX_PORTS 2
+
+
+extern int netback_accel_sf_init(void);
+extern void netback_accel_sf_shutdown(void);
+extern int netback_accel_sf_hwtype(struct netback_accel *bend);
+
+extern int netback_accel_sf_char_init(void);
+extern void netback_accel_sf_char_shutdown(void);
+
+extern int netback_accel_setup_vnic_hw(struct netback_accel *bend);
+extern void netback_accel_shutdown_vnic_hw(struct netback_accel *bend);
+
+extern int netback_accel_add_buffers(struct netback_accel *bend, int pages, 
+				     int log2_pages, u32 *grants,
+				     u32 *buf_addr_out);
+extern int netback_accel_remove_buffers(struct netback_accel *bend);
+
+
+/* Add a filter for the specified IP/port to the backend */
+extern int
+netback_accel_filter_check_add(struct netback_accel *bend, 
+			       struct netback_accel_filter_spec *filt);
+/* Remove a filter entry for the specific device and IP/port */
+extern
+void netback_accel_filter_remove_index(struct netback_accel *bend, 
+				       int filter_index);
+extern
+void netback_accel_filter_remove_spec(struct netback_accel *bend, 
+				      struct netback_accel_filter_spec *filt);
+
+/* This is designed to look a bit like a skb */
+struct netback_pkt_buf {
+	union {
+		unsigned char *raw;
+	} mac;
+	union {
+		struct iphdr  *iph;
+		struct arphdr *arph;
+		unsigned char *raw;
+	} nh;
+	int protocol;
+};
+
+/*! \brief Handle a received packet: insert fast path filters as necessary
+ * \param skb The packet buffer
+ */
+extern void netback_accel_rx_packet(struct netback_pkt_buf *skb, void *fwd_priv);
+
+/*! \brief Handle a transmitted packet: update fast path filters as necessary
+ * \param skb The packet buffer
+ */
+extern void netback_accel_tx_packet(struct sk_buff *skb, void *fwd_priv);
+
+#endif /* NETBACK_ACCEL_SOLARFLARE_H */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/accel_xenbus.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/accel_xenbus.c
--- linux-2.6.18.8/drivers/xen/sfc_netback/accel_xenbus.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/accel_xenbus.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,831 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <xen/evtchn.h>
+#include <linux/mutex.h>
+
+/* drivers/xen/netback/common.h */
+#include "common.h"
+
+#include "accel.h"
+#include "accel_solarflare.h"
+#include "accel_util.h"
+
+#define NODENAME_PATH_FMT "backend/vif/%d/%d"
+
+#define NETBACK_ACCEL_FROM_XENBUS_DEVICE(_dev) (struct netback_accel *) \
+	((struct backend_info *)(_dev)->dev.driver_data)->netback_accel_priv
+
+/* List of all the bends currently in existence. */
+struct netback_accel *bend_list = NULL;
+DEFINE_MUTEX(bend_list_mutex);
+
+/* Put in bend_list.  Must hold bend_list_mutex */
+static void link_bend(struct netback_accel *bend)
+{
+	bend->next_bend = bend_list;
+	bend_list = bend;
+}
+
+/* Remove from bend_list,  Must hold bend_list_mutex */
+static void unlink_bend(struct netback_accel *bend)
+{
+	struct netback_accel *tmp = bend_list;
+	struct netback_accel *prev = NULL;
+	while (tmp != NULL) {
+		if (tmp == bend) {
+			if (prev != NULL)
+				prev->next_bend = bend->next_bend;
+			else
+				bend_list = bend->next_bend;
+			return;
+		}
+		prev = tmp;
+		tmp = tmp->next_bend;
+	}
+}
+
+
+/* Demultiplex a message IRQ from the frontend driver.  */
+static irqreturn_t msgirq_from_frontend(int irq, void *context, 
+				     struct pt_regs *unused)
+{
+	struct xenbus_device *dev = context;
+	struct netback_accel *bend = NETBACK_ACCEL_FROM_XENBUS_DEVICE(dev);
+	VPRINTK("irq %d from device %s\n", irq, dev->nodename);
+	schedule_work(&bend->handle_msg);
+	return IRQ_HANDLED;
+}
+
+
+/*
+ * Demultiplex an IRQ from the frontend driver.  This is never used
+ * functionally, but we need it to pass to the bind function, and may
+ * get called spuriously
+ */
+static irqreturn_t netirq_from_frontend(int irq, void *context, 
+					struct pt_regs *unused)
+{
+	VPRINTK("netirq %d from device %s\n", irq,
+		((struct xenbus_device *)context)->nodename);
+	
+	return IRQ_HANDLED;
+}
+
+
+/* Read the limits values of the xenbus structure. */
+static 
+void cfg_hw_quotas(struct xenbus_device *dev, struct netback_accel *bend)
+{
+	int err = xenbus_gather
+		(XBT_NIL, dev->nodename,
+		 "limits/max-filters", "%d", &bend->quotas.max_filters,
+		 "limits/max-buf-pages", "%d", &bend->quotas.max_buf_pages,
+		 "limits/max-mcasts", "%d", &bend->quotas.max_mcasts,
+		 NULL);
+	if (err) {
+		/*
+		 * TODO what if they have previously been set by the
+		 * user?  This will overwrite with defaults.  Maybe
+		 * not what we want to do, but useful in startup
+		 * case 
+		 */
+		DPRINTK("Failed to read quotas from xenbus, using defaults\n");
+		bend->quotas.max_filters = NETBACK_ACCEL_DEFAULT_MAX_FILTERS;
+		bend->quotas.max_buf_pages = sfc_netback_max_pages;
+		bend->quotas.max_mcasts = NETBACK_ACCEL_DEFAULT_MAX_MCASTS;
+	}
+
+	return;
+}
+
+
+static void bend_config_accel_change(struct xenbus_watch *watch,
+				     const char **vec, unsigned int len)
+{
+	struct netback_accel *bend;
+
+	bend = container_of(watch, struct netback_accel, config_accel_watch);
+
+	mutex_lock(&bend->bend_mutex);
+	if (bend->config_accel_watch.node != NULL) {
+		struct xenbus_device *dev = 
+			(struct xenbus_device *)bend->hdev_data;
+		DPRINTK("Watch matched, got dev %p otherend %p\n",
+			dev, dev->otherend);
+		if(!xenbus_exists(XBT_NIL, watch->node, "")) {
+			DPRINTK("Ignoring watch as otherend seems invalid\n");
+			goto out;
+		}
+		
+		cfg_hw_quotas(dev, bend);
+	}
+ out:
+	mutex_unlock(&bend->bend_mutex);
+	return;
+}
+
+
+/*
+ * Setup watch on "limits" in the backend vif info to know when
+ * configuration has been set
+ */
+static int setup_config_accel_watch(struct xenbus_device *dev,
+				    struct netback_accel *bend)
+{
+	int err;
+
+	VPRINTK("Setting watch on %s/%s\n", dev->nodename, "limits");
+
+	err = xenbus_watch_path2(dev, dev->nodename, "limits", 
+				 &bend->config_accel_watch, 
+				 bend_config_accel_change);
+
+	if (err) {
+		EPRINTK("%s: Failed to register xenbus watch: %d\n",
+			__FUNCTION__, err);
+		bend->config_accel_watch.node = NULL;
+		return err;
+	}
+	return 0;
+}
+
+
+static int 
+cfg_frontend_info(struct xenbus_device *dev, struct netback_accel *bend,
+		  int *grants)
+{
+	/* Get some info from xenbus on the event channel and shmem grant */
+	int err = xenbus_gather(XBT_NIL, dev->otherend, 
+				"accel-msg-channel", "%u", &bend->msg_channel, 
+				"accel-ctrl-page", "%d", &(grants[0]),
+				"accel-msg-page", "%d", &(grants[1]),
+				"accel-net-channel", "%u", &bend->net_channel,
+				NULL);
+	if (err)
+		EPRINTK("failed to read event channels or shmem grant: %d\n",
+			err);
+	else
+		DPRINTK("got event chan %d and net chan %d from frontend\n",
+			bend->msg_channel, bend->net_channel);
+	return err;
+}
+
+
+/* Setup all the comms needed to chat with the front end driver */
+static int setup_vnic(struct xenbus_device *dev)
+{
+	struct netback_accel *bend;
+	int grants[2], err, msgs_per_queue;
+
+	bend = NETBACK_ACCEL_FROM_XENBUS_DEVICE(dev);
+
+	err = cfg_frontend_info(dev, bend, grants);
+	if (err)
+		goto fail1;
+
+	/*
+	 * If we get here, both frontend Connected and configuration
+	 * options available.  All is well.
+	 */
+
+	/* Get the hardware quotas for the VNIC in question.  */
+	cfg_hw_quotas(dev, bend);
+
+	/* Set up the deferred work handlers */
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20)
+	INIT_WORK(&bend->handle_msg, 
+		  netback_accel_msg_rx_handler);
+#else
+	INIT_WORK(&bend->handle_msg, 
+		  netback_accel_msg_rx_handler,
+		  (void*)bend);
+#endif
+
+	/* Request the frontend mac */
+	err = net_accel_xen_net_read_mac(dev, bend->mac);
+	if (err)
+		goto fail2;
+
+	/* Set up the shared page. */
+	bend->shared_page = net_accel_map_grants_contig(dev, grants, 2, 
+							&bend->sh_pages_unmap);
+
+	if (bend->shared_page == NULL) {
+		EPRINTK("failed to map shared page for %s\n", dev->otherend);
+		err = -ENOMEM;
+		goto fail2;
+	}
+
+	/* Initialise the shared page(s) used for comms */
+	net_accel_msg_init_page(bend->shared_page, PAGE_SIZE, 
+				bend->net_dev->flags & IFF_UP);
+
+	msgs_per_queue = (PAGE_SIZE/2) / sizeof(struct net_accel_msg);
+
+	net_accel_msg_init_queue
+		(&bend->to_domU, &bend->shared_page->queue0,
+		 (struct net_accel_msg *)((__u8*)bend->shared_page + PAGE_SIZE),
+		 msgs_per_queue);
+
+	net_accel_msg_init_queue
+		(&bend->from_domU, &bend->shared_page->queue1, 
+		 (struct net_accel_msg *)((__u8*)bend->shared_page + 
+					  (3 * PAGE_SIZE / 2)),
+		 msgs_per_queue);
+
+	/* Bind the message event channel to a handler
+	 *
+	 * Note that we will probably get a spurious interrupt when we
+	 * do this, so it must not be done until we have set up
+	 * everything we need to handle it.
+	 */
+	err = bind_interdomain_evtchn_to_irqhandler(dev->otherend_id,
+						    bend->msg_channel,
+						    msgirq_from_frontend,
+						    0,
+						    "netback_accel",
+						    dev);
+	if (err < 0) {
+		EPRINTK("failed to bind event channel: %d\n", err);
+		goto fail3;
+	}
+	else
+		bend->msg_channel_irq = err;
+
+	/* TODO: No need to bind this evtchn to an irq. */
+	err = bind_interdomain_evtchn_to_irqhandler(dev->otherend_id,
+						    bend->net_channel,
+						    netirq_from_frontend,
+						    0,
+						    "netback_accel",
+						    dev);
+	if (err < 0) {
+		EPRINTK("failed to bind net channel: %d\n", err);
+		goto fail4;
+	}  
+	else
+		bend->net_channel_irq = err;
+
+	/*
+	 * Grab ourselves an entry in the forwarding hash table. We do
+	 * this now so we don't have the embarassmesnt of sorting out
+	 * an allocation failure while at IRQ. Because we pass NULL as
+	 * the context, the actual hash lookup will succeed for this
+	 * NIC, but the check for somewhere to forward to will
+	 * fail. This is necessary to prevent forwarding before
+	 * hardware resources are set up
+	 */
+	err = netback_accel_fwd_add(bend->mac, NULL, bend->fwd_priv);
+	if (err) {
+		EPRINTK("failed to add to fwd hash table\n");
+		goto fail5;
+	}
+
+	/*
+	 * Say hello to frontend.  Important to do this straight after
+	 * obtaining the message queue as otherwise we are vulnerable
+	 * to an evil frontend sending a HELLO-REPLY before we've sent
+	 * the HELLO and confusing us
+	 */
+	netback_accel_msg_tx_hello(bend, NET_ACCEL_MSG_VERSION);
+	return 0;
+
+ fail5:
+	unbind_from_irqhandler(bend->net_channel_irq, dev);
+ fail4:
+	unbind_from_irqhandler(bend->msg_channel_irq, dev);
+ fail3:
+	net_accel_unmap_grants_contig(dev, bend->sh_pages_unmap);
+	bend->shared_page = NULL;
+	bend->sh_pages_unmap = NULL;
+ fail2:
+ fail1:
+	return err;
+}
+
+
+static int read_nicname(struct xenbus_device *dev, struct netback_accel *bend)
+{
+	int len;
+
+	/* nic name used to select interface used for acceleration */
+	bend->nicname = xenbus_read(XBT_NIL, dev->nodename, "accel", &len);
+	if (IS_ERR(bend->nicname))
+		return PTR_ERR(bend->nicname);
+
+	return 0;
+}
+
+static const char *frontend_name = "sfc_netfront";
+
+static int publish_frontend_name(struct xenbus_device *dev)
+{
+	struct xenbus_transaction tr;
+	int err;
+	
+	/* Publish the name of the frontend driver */
+	do {
+		err = xenbus_transaction_start(&tr);
+		if (err != 0) { 
+			EPRINTK("%s: transaction start failed\n", __FUNCTION__);
+			return err;
+		}
+		err = xenbus_printf(tr, dev->nodename, "accel-frontend", 
+				    "%s", frontend_name);
+		if (err != 0) {
+			EPRINTK("%s: xenbus_printf failed\n", __FUNCTION__);
+			xenbus_transaction_end(tr, 1);
+			return err;
+		}
+		err = xenbus_transaction_end(tr, 0);
+	} while (err == -EAGAIN);
+	
+	if (err != 0) {
+		EPRINTK("failed to end frontend name transaction\n");
+		return err;
+	}
+	return 0;
+}
+
+
+static int unpublish_frontend_name(struct xenbus_device *dev)
+{
+	struct xenbus_transaction tr;
+	int err;
+
+	do {
+		err = xenbus_transaction_start(&tr);
+		if (err != 0)
+			break;
+		err = xenbus_rm(tr, dev->nodename, "accel-frontend");
+		if (err != 0) {
+			xenbus_transaction_end(tr, 1);
+			break;
+		}
+		err = xenbus_transaction_end(tr, 0);
+	} while (err == -EAGAIN);
+
+	return err;
+}
+
+
+static void cleanup_vnic(struct netback_accel *bend)
+{
+	struct xenbus_device *dev;
+
+	dev = (struct xenbus_device *)bend->hdev_data;
+
+	DPRINTK("%s: bend %p dev %p\n", __FUNCTION__, bend, dev);
+
+	DPRINTK("%s: Remove %p's mac from fwd table...\n", 
+		__FUNCTION__, bend);
+	netback_accel_fwd_remove(bend->mac, bend->fwd_priv);
+
+	/* Free buffer table allocations */
+	netback_accel_remove_buffers(bend);
+
+	DPRINTK("%s: Release hardware resources...\n", __FUNCTION__);
+	if (bend->accel_shutdown)
+		bend->accel_shutdown(bend);
+
+	if (bend->net_channel_irq) {
+		unbind_from_irqhandler(bend->net_channel_irq, dev);
+		bend->net_channel_irq = 0;
+	}
+
+	if (bend->msg_channel_irq) {
+		unbind_from_irqhandler(bend->msg_channel_irq, dev);
+		bend->msg_channel_irq = 0;
+	}
+
+	if (bend->sh_pages_unmap) {
+		DPRINTK("%s: Unmap grants %p\n", __FUNCTION__, 
+			bend->sh_pages_unmap);
+		net_accel_unmap_grants_contig(dev, bend->sh_pages_unmap);
+		bend->sh_pages_unmap = NULL;
+		bend->shared_page = NULL;
+	}
+}
+
+
+/*************************************************************************/
+
+/*
+ * The following code handles accelstate changes between the frontend
+ * and the backend.  It calls setup_vnic and cleanup_vnic in matching
+ * pairs in response to transitions.
+ *
+ * Valid state transitions for Dom0 are as follows:
+ *
+ * Closed->Init       on probe or in response to Init from domU
+ * Closed->Closing    on error/remove
+ *
+ * Init->Connected    in response to Connected from domU
+ * Init->Closing      on error/remove or in response to Closing from domU
+ *
+ * Connected->Closing on error/remove or in response to Closing from domU
+ *
+ * Closing->Closed    in response to Closed from domU
+ *
+ */
+
+
+static void netback_accel_frontend_changed(struct xenbus_device *dev,
+					   XenbusState frontend_state)
+{
+	struct netback_accel *bend = NETBACK_ACCEL_FROM_XENBUS_DEVICE(dev);
+	XenbusState backend_state;
+
+	DPRINTK("%s: changing from %s to %s. nodename %s, otherend %s\n",
+		__FUNCTION__, xenbus_strstate(bend->frontend_state),
+		xenbus_strstate(frontend_state),dev->nodename, dev->otherend);
+
+	/*
+	 * Ignore duplicate state changes.  This can happen if the
+	 * frontend changes state twice in quick succession and the
+	 * first watch fires in the backend after the second
+	 * transition has completed.
+	 */
+	if (bend->frontend_state == frontend_state)
+		return;
+
+	bend->frontend_state = frontend_state;
+	backend_state = bend->backend_state;
+
+	switch (frontend_state) {
+	case XenbusStateInitialising:
+		if (backend_state == XenbusStateClosed &&
+		    !bend->removing)
+			backend_state = XenbusStateInitialising;
+		break;
+
+	case XenbusStateConnected:
+		if (backend_state == XenbusStateInitialising) {
+			if (!bend->vnic_is_setup &&
+			    setup_vnic(dev) == 0) {
+				bend->vnic_is_setup = 1;
+				backend_state = XenbusStateConnected;
+			} else {
+				backend_state = XenbusStateClosing;
+			}
+		}
+		break;
+
+	case XenbusStateInitWait:
+	case XenbusStateInitialised:
+	default:
+		DPRINTK("Unknown state %s (%d) from frontend.\n",
+			xenbus_strstate(frontend_state), frontend_state);
+		/* Unknown state.  Fall through. */
+	case XenbusStateClosing:
+		if (backend_state != XenbusStateClosed)
+			backend_state = XenbusStateClosing;
+
+		/*
+		 * The bend will now persist (with watches active) in
+		 * case the frontend comes back again, eg. after
+		 * frontend module reload or suspend/resume
+		 */
+
+		break;
+
+	case XenbusStateUnknown:
+	case XenbusStateClosed:
+		if (bend->vnic_is_setup) {
+			bend->vnic_is_setup = 0;
+			cleanup_vnic(bend);
+		}
+
+		if (backend_state == XenbusStateClosing)
+			backend_state = XenbusStateClosed;
+		break;
+	}
+
+	if (backend_state != bend->backend_state) {
+		DPRINTK("Switching from state %s (%d) to %s (%d)\n",
+			xenbus_strstate(bend->backend_state),
+			bend->backend_state,
+			xenbus_strstate(backend_state), backend_state);
+		bend->backend_state = backend_state;
+		net_accel_update_state(dev, backend_state);
+	}
+
+	wake_up(&bend->state_wait_queue);
+}
+
+
+/* accelstate on the frontend's xenbus node has changed */
+static void bend_domu_accel_change(struct xenbus_watch *watch,
+				   const char **vec, unsigned int len)
+{
+	int state;
+	struct netback_accel *bend;
+
+	bend = container_of(watch, struct netback_accel, domu_accel_watch);
+	if (bend->domu_accel_watch.node != NULL) {
+		struct xenbus_device *dev = 
+			(struct xenbus_device *)bend->hdev_data;
+		VPRINTK("Watch matched, got dev %p otherend %p\n",
+			dev, dev->otherend);
+		/*
+		 * dev->otherend != NULL check to protect against
+		 * watch firing when domain goes away and we haven't
+		 * yet cleaned up
+		 */
+		if (!dev->otherend ||
+		    !xenbus_exists(XBT_NIL, watch->node, "") ||
+		    strncmp(dev->otherend, vec[XS_WATCH_PATH],
+			    strlen(dev->otherend))) {
+			DPRINTK("Ignoring watch as otherend seems invalid\n");
+			return;
+		}
+
+		mutex_lock(&bend->bend_mutex);
+
+		xenbus_scanf(XBT_NIL, dev->otherend, "accelstate", "%d", 
+			     &state);
+		netback_accel_frontend_changed(dev, state);
+
+		mutex_unlock(&bend->bend_mutex);
+	}
+}
+
+/* Setup watch on frontend's accelstate */
+static int setup_domu_accel_watch(struct xenbus_device *dev,
+				  struct netback_accel *bend)
+{
+	int err;
+
+	VPRINTK("Setting watch on %s/%s\n", dev->otherend, "accelstate");
+
+	err = xenbus_watch_path2(dev, dev->otherend, "accelstate", 
+				 &bend->domu_accel_watch, 
+				 bend_domu_accel_change);
+	if (err) {
+		EPRINTK("%s: Failed to register xenbus watch: %d\n",
+			__FUNCTION__, err);
+		goto fail;
+	}
+	return 0;
+ fail:
+	bend->domu_accel_watch.node = NULL;
+	return err;
+}
+
+
+int netback_accel_probe(struct xenbus_device *dev)
+{
+	struct netback_accel *bend;
+	struct backend_info *binfo;
+	int err;
+
+	DPRINTK("%s: passed device %s\n", __FUNCTION__, dev->nodename);
+
+	/* Allocate structure to store all our state... */
+	bend = kzalloc(sizeof(struct netback_accel), GFP_KERNEL);
+	if (bend == NULL) {
+		DPRINTK("%s: no memory for bend\n", __FUNCTION__);
+		return -ENOMEM;
+	}
+	
+	mutex_init(&bend->bend_mutex);
+
+	mutex_lock(&bend->bend_mutex);
+
+	/* ...and store it where we can get at it */
+	binfo = (struct backend_info *) dev->dev.driver_data;
+	binfo->netback_accel_priv = bend;
+	/* And vice-versa */
+	bend->hdev_data = dev;
+
+	DPRINTK("%s: Adding bend %p to list\n", __FUNCTION__, bend);
+	
+	init_waitqueue_head(&bend->state_wait_queue);
+	bend->vnic_is_setup = 0;
+	bend->frontend_state = XenbusStateUnknown;
+	bend->backend_state = XenbusStateClosed;
+	bend->removing = 0;
+
+	sscanf(dev->nodename, NODENAME_PATH_FMT, &bend->far_end, 
+	       &bend->vif_num);
+
+	err = read_nicname(dev, bend);
+	if (err) {
+		/*
+		 * Technically not an error, just means we're not 
+		 * supposed to accelerate this
+		 */
+		DPRINTK("failed to get device name\n");
+		goto fail_nicname;
+	}
+
+	/*
+	 * Look up the device name in the list of NICs provided by
+	 * driverlink to get the hardware type.
+	 */
+	err = netback_accel_sf_hwtype(bend);
+	if (err) {
+		/*
+		 * Technically not an error, just means we're not
+		 * supposed to accelerate this, probably belongs to
+		 * some other backend
+		 */
+		DPRINTK("failed to match device name\n");
+		goto fail_init_type;
+	}
+
+	err = publish_frontend_name(dev);
+	if (err)
+		goto fail_publish;
+
+	err = netback_accel_debugfs_create(bend);
+	if (err)
+		goto fail_debugfs;
+	
+	mutex_unlock(&bend->bend_mutex);
+
+	err = setup_config_accel_watch(dev, bend);
+	if (err)
+		goto fail_config_watch;
+
+	err = setup_domu_accel_watch(dev, bend);
+	if (err)
+		goto fail_domu_watch;
+
+	/*
+	 * Indicate to the other end that we're ready to start unless
+	 * the watch has already fired.
+	 */
+	mutex_lock(&bend->bend_mutex);
+	if (bend->backend_state == XenbusStateClosed) {
+		bend->backend_state = XenbusStateInitialising;
+		net_accel_update_state(dev, XenbusStateInitialising);
+	}
+	mutex_unlock(&bend->bend_mutex);
+
+	mutex_lock(&bend_list_mutex);
+	link_bend(bend);
+	mutex_unlock(&bend_list_mutex);
+
+	return 0;
+
+fail_domu_watch:
+
+	unregister_xenbus_watch(&bend->config_accel_watch);
+	kfree(bend->config_accel_watch.node);
+fail_config_watch:
+
+	/*
+	 * Flush the scheduled work queue before freeing bend to get
+	 * rid of any pending netback_accel_msg_rx_handler()
+	 */
+	flush_scheduled_work();
+
+	mutex_lock(&bend->bend_mutex);
+	net_accel_update_state(dev, XenbusStateUnknown);
+	netback_accel_debugfs_remove(bend);
+fail_debugfs:
+
+	unpublish_frontend_name(dev);
+fail_publish:
+
+	/* No need to reverse netback_accel_sf_hwtype. */
+fail_init_type:
+
+	kfree(bend->nicname);
+fail_nicname:
+	binfo->netback_accel_priv = NULL;
+	mutex_unlock(&bend->bend_mutex);
+	kfree(bend);
+	return err;
+}
+
+
+int netback_accel_remove(struct xenbus_device *dev)
+{
+	struct backend_info *binfo;
+	struct netback_accel *bend; 
+	int frontend_state;
+
+	binfo = (struct backend_info *) dev->dev.driver_data;
+	bend = (struct netback_accel *) binfo->netback_accel_priv;
+
+	DPRINTK("%s: dev %p bend %p\n", __FUNCTION__, dev, bend);
+	
+	BUG_ON(bend == NULL);
+	
+	mutex_lock(&bend_list_mutex);
+	unlink_bend(bend);
+	mutex_unlock(&bend_list_mutex);
+
+	mutex_lock(&bend->bend_mutex);
+
+	/* Reject any requests to connect. */
+	bend->removing = 1;
+
+	/*
+	 * Switch to closing to tell the other end that we're going
+	 * away.
+	 */
+	if (bend->backend_state != XenbusStateClosing) {
+		bend->backend_state = XenbusStateClosing;
+		net_accel_update_state(dev, XenbusStateClosing);
+	}
+
+	frontend_state = (int)XenbusStateUnknown;
+	xenbus_scanf(XBT_NIL, dev->otherend, "accelstate", "%d",
+		     &frontend_state);
+
+	mutex_unlock(&bend->bend_mutex);
+
+	/*
+	 * Wait until this end goes to the closed state.  This happens
+	 * in response to the other end going to the closed state.
+	 * Don't bother doing this if the other end is already closed
+	 * because if it is then there is nothing to do.
+	 */
+	if (frontend_state != (int)XenbusStateClosed &&
+	    frontend_state != (int)XenbusStateUnknown)
+		wait_event(bend->state_wait_queue,
+			   bend->backend_state == XenbusStateClosed);
+
+	unregister_xenbus_watch(&bend->domu_accel_watch);
+	kfree(bend->domu_accel_watch.node);
+
+	unregister_xenbus_watch(&bend->config_accel_watch);
+	kfree(bend->config_accel_watch.node);
+
+	/*
+	 * Flush the scheduled work queue before freeing bend to get
+	 * rid of any pending netback_accel_msg_rx_handler()
+	 */
+	flush_scheduled_work();
+
+	mutex_lock(&bend->bend_mutex);
+
+	/* Tear down the vnic if it was set up. */
+	if (bend->vnic_is_setup) {
+		bend->vnic_is_setup = 0;
+		cleanup_vnic(bend);
+	}
+
+	bend->backend_state = XenbusStateUnknown;
+	net_accel_update_state(dev, XenbusStateUnknown);
+
+	netback_accel_debugfs_remove(bend);
+
+	unpublish_frontend_name(dev);
+
+	kfree(bend->nicname);
+
+	binfo->netback_accel_priv = NULL;
+
+	mutex_unlock(&bend->bend_mutex);
+
+	kfree(bend);
+
+	return 0;
+}
+
+
+void netback_accel_shutdown_bends(void)
+{
+	mutex_lock(&bend_list_mutex);
+	/*
+	 * I think we should have had a remove callback for all
+	 * interfaces before being allowed to unload the module
+	 */
+	BUG_ON(bend_list != NULL);
+	mutex_unlock(&bend_list_mutex);
+}
+
+
+void netback_accel_set_closing(struct netback_accel *bend) 
+{
+
+	bend->backend_state = XenbusStateClosing;
+	net_accel_update_state((struct xenbus_device *)bend->hdev_data,
+			       XenbusStateClosing);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/compat/gcc.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/compat/gcc.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/compat/gcc.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/compat/gcc.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,158 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*! \cidoxg_include_ci_compat  */
+
+#ifndef __CI_COMPAT_GCC_H__
+#define __CI_COMPAT_GCC_H__
+
+
+#define CI_HAVE_INT64
+
+
+#if defined(__linux__) && defined(__KERNEL__)
+
+# include <linux/types.h>
+
+typedef __u64                 ci_uint64;
+typedef __s64                 ci_int64;
+# if BITS_PER_LONG == 32
+typedef __s32                 ci_ptr_arith_t;
+typedef __u32                 ci_uintptr_t;
+# else
+typedef __s64                 ci_ptr_arith_t;
+typedef __u64                 ci_uintptr_t;
+# endif
+
+
+/* it's not obvious to me why the below is wrong for x64_64, but
+ * gcc seems to complain on this platform
+ */
+# if defined(__ia64__)
+#  define CI_PRId64            "ld"
+#  define CI_PRIi64            "li"
+#  define CI_PRIo64            "lo"
+#  define CI_PRIu64            "lu"
+#  define CI_PRIx64            "lx"
+#  define CI_PRIX64            "lX"
+# else
+#  define CI_PRId64            "lld"
+#  define CI_PRIi64            "lli"
+#  define CI_PRIo64            "llo"
+#  define CI_PRIu64            "llu"
+#  define CI_PRIx64            "llx"
+#  define CI_PRIX64            "llX"
+# endif
+
+# define CI_PRId32            "d"
+# define CI_PRIi32            "i"
+# define CI_PRIo32            "o"
+# define CI_PRIu32            "u"
+# define CI_PRIx32            "x"
+# define CI_PRIX32            "X"
+
+#else
+
+# include <stdint.h>
+# include <inttypes.h>
+
+typedef uint64_t              ci_uint64;
+typedef int64_t               ci_int64;
+typedef intptr_t              ci_ptr_arith_t;
+typedef uintptr_t             ci_uintptr_t;
+
+# define CI_PRId64            PRId64
+# define CI_PRIi64            PRIi64
+# define CI_PRIo64            PRIo64
+# define CI_PRIu64            PRIu64
+# define CI_PRIx64            PRIx64
+# define CI_PRIX64            PRIX64
+
+# define CI_PRId32            PRId32
+# define CI_PRIi32            PRIi32
+# define CI_PRIo32            PRIo32
+# define CI_PRIu32            PRIu32
+# define CI_PRIx32            PRIx32
+# define CI_PRIX32            PRIX32
+
+#endif
+
+
+typedef ci_uint64                       ci_fixed_descriptor_t;
+
+#define from_fixed_descriptor(desc) ((ci_uintptr_t)(desc))
+#define to_fixed_descriptor(desc) ((ci_fixed_descriptor_t)(ci_uintptr_t)(desc))
+
+
+#if __GNUC__ >= 3 && !defined(__cplusplus)
+/*
+** Checks that [p_mbr] has the same type as [&c_type::mbr_name].
+*/
+# define CI_CONTAINER(c_type, mbr_name, p_mbr)				\
+   __builtin_choose_expr(						\
+     __builtin_types_compatible_p(__typeof__(&((c_type*)0)->mbr_name),	\
+				 __typeof__(p_mbr)),			\
+     __CI_CONTAINER(c_type, mbr_name, p_mbr), (void)0)
+
+# define ci_restrict  __restrict__
+#endif
+
+
+#if !defined(__KERNEL__) || defined(__unix__)
+#define CI_HAVE_NPRINTF  1
+#endif
+
+
+/* At what version was this introduced? */
+#if __GNUC__ >= 3 || (__GNUC__ == 2 && __GNUC_MINOR__ > 91)
+# define CI_LIKELY(t)    __builtin_expect((t), 1)
+# define CI_UNLIKELY(t)  __builtin_expect((t), 0)
+#endif
+
+/**********************************************************************
+ * Attributes
+ */
+#if __GNUC__ >= 3 && defined(NDEBUG)
+# define CI_HF __attribute__((visibility("hidden")))
+# define CI_HV __attribute__((visibility("hidden")))
+#else
+# define CI_HF
+# define CI_HV
+#endif
+
+#if __GNUC__ >= 4 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)
+# define ci_noinline  static __attribute__((__noinline__))
+/* (Linux 2.6 defines its own "noinline", so we use the "__noinline__" form) */
+#else
+# define ci_noinline  static
+#endif
+
+#define CI_ALIGN(x) __attribute__ ((aligned (x)))
+
+#define CI_PRINTF_LIKE(a,b) __attribute__((format(printf,a,b)))
+
+#endif  /* __CI_COMPAT_GCC_H__ */
+
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/compat/gcc_x86.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/compat/gcc_x86.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/compat/gcc_x86.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/compat/gcc_x86.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,115 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*! \cidoxg_include_ci_compat  */
+
+#ifndef __CI_COMPAT_GCC_X86_H__
+#define __CI_COMPAT_GCC_X86_H__
+
+/*
+** The facts:
+**
+**   SSE   sfence
+**   SSE2  lfence, mfence, pause
+*/
+
+/* 
+   Barriers to enforce ordering with respect to:
+
+   normal memory use: ci_wmb, ci_rmb, ci_wmb
+   IO bus access use: ci_wiob, ci_riob, ci_iob
+*/
+#if defined(__x86_64__)
+# define ci_x86_mb() __asm__ __volatile__ ("lock; addl $0,0(%%rsp)":::"memory")
+#else
+# define ci_x86_mb() __asm__ __volatile__ ("lock; addl $0,0(%%esp)":::"memory")
+#endif
+
+/* ?? measure the impact of latency of sfence on a modern processor before we
+   take a decision on how to integrate with respect to writecombining */
+
+/* DJR: I don't think we need to add "memory" here.  It means the asm does
+** something to memory that GCC doesn't understand.  But all this does is
+** commit changes that GCC thinks have already happened.  NB. GCC will not
+** reorder across a __volatile__ __asm__ anyway.
+*/
+#define ci_gcc_fence()    __asm__ __volatile__ ("")
+
+#if __GNUC__ >= 3 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 96)
+# define ci_x86_sfence()  __asm__ __volatile__ ("sfence")
+# define ci_x86_lfence()  __asm__ __volatile__ ("lfence")
+# define ci_x86_mfence()  __asm__ __volatile__ ("mfence")
+#else
+# define ci_x86_sfence()  __asm__ __volatile__ (".byte 0x0F, 0xAE, 0xF8")
+# define ci_x86_lfence()  __asm__ __volatile__ (".byte 0x0F, 0xAE, 0xE8")
+# define ci_x86_mfence()  __asm__ __volatile__ (".byte 0x0F, 0xAE, 0xF0")
+#endif
+
+
+/* x86 processors to P4 Xeon store in-order unless executing streaming
+   extensions or when using writecombining 
+
+   Hence we do not define ci_wmb to use sfence by default. Requirement is that
+   we do not use writecombining to memory and any code which uses SSE
+   extensions must call sfence directly 
+
+   We need to track non intel clones which may support out of order store.
+
+*/
+
+#if CI_CPU_OOS
+# if CI_CPU_HAS_SSE
+#  define ci_wmb()	ci_x86_sfence()
+# else
+#  define ci_wmb()	ci_x86_mb()
+# endif
+#else
+# define ci_wmb()       ci_gcc_fence()
+#endif
+
+#if CI_CPU_HAS_SSE2
+# define ci_rmb()	ci_x86_lfence()
+# define ci_mb()	ci_x86_mfence()
+# define ci_riob()	ci_x86_lfence()
+# define ci_wiob()	ci_x86_sfence()
+# define ci_iob()	ci_x86_mfence()
+#else
+# if CI_CPU_HAS_SSE
+#  define ci_wiob()	ci_x86_sfence()
+# else
+#  define ci_wiob()	ci_x86_mb()
+# endif
+# define ci_rmb()	ci_x86_mb()
+# define ci_mb()   	ci_x86_mb()
+# define ci_riob()  	ci_x86_mb()
+# define ci_iob()  	ci_x86_mb()
+#endif
+
+typedef unsigned long   ci_phys_addr_t;
+#define ci_phys_addr_fmt  "%lx"
+
+#endif  /* __CI_COMPAT_GCC_X86_H__ */
+
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/compat/primitive.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/compat/primitive.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/compat/primitive.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/compat/primitive.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,77 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+/*! \cidoxg_include_ci_compat  */
+
+#ifndef __CI_COMPAT_PRIMITIVE_H__
+#define __CI_COMPAT_PRIMITIVE_H__
+
+
+/**********************************************************************
+ * Primitive types.
+ */
+
+typedef unsigned char                   ci_uint8;
+typedef char                            ci_int8;
+
+typedef unsigned short                  ci_uint16;
+typedef short                           ci_int16;
+
+typedef unsigned int                    ci_uint32;
+typedef int                             ci_int32;
+
+/* 64-bit support is platform dependent. */
+
+
+/**********************************************************************
+ * Other fancy types.
+ */
+
+typedef ci_uint8                        ci_octet;
+
+typedef enum {
+  CI_FALSE = 0,
+  CI_TRUE
+} ci_boolean_t;
+
+
+/**********************************************************************
+ * Some nice types you'd always assumed were standards.
+ * (Really, they are SYSV "standards".)
+ */
+
+#ifdef _WIN32
+typedef unsigned long                   ulong;              
+typedef unsigned int                    uint;
+typedef char*                           caddr_t;
+#elif defined(__linux__) && defined(__KERNEL__)
+#include <linux/types.h>
+#elif defined(__linux__)
+#include <sys/types.h>
+#endif
+
+
+#endif  /* __CI_COMPAT_PRIMITIVE_H__ */
+
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/compat/sysdep.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/compat/sysdep.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/compat/sysdep.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/compat/sysdep.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,166 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*! \cidoxg_include_ci_compat  */
+
+#ifndef __CI_COMPAT_SYSDEP_H__
+#define __CI_COMPAT_SYSDEP_H__
+
+
+/**********************************************************************
+ * Platform definition fixups.
+ */
+
+#if defined(__ci_ul_driver__) && !defined(__ci_driver__)
+# define __ci_driver__
+#endif
+
+#if defined(__ci_driver__) && !defined(__ci_ul_driver__) && \
+   !defined(__KERNEL__)
+# define __KERNEL__
+#endif
+
+
+/**********************************************************************
+ * Sanity checks (no cheating!)
+ */
+
+#if defined(__KERNEL__) && !defined(__ci_driver__)
+# error Insane.
+#endif
+
+#if defined(__KERNEL__) && defined(__ci_ul_driver__)
+# error Madness.
+#endif
+
+#if defined(__unix__) && defined(_WIN32)
+# error Strange.
+#endif
+
+#if defined(__GNUC__) && defined(_MSC_VER)
+# error Crazy.
+#endif
+
+
+/**********************************************************************
+ * Compiler and processor dependencies.
+ */
+
+#if defined(__GNUC__)
+
+# include <ci/compat/gcc.h>
+
+# if defined(__i386__)
+#  include <ci/compat/x86.h>
+#  include <ci/compat/gcc_x86.h>
+# elif defined(__x86_64__)
+#  include <ci/compat/x86_64.h>
+#  include <ci/compat/gcc_x86.h>
+# elif defined(__PPC__)
+#  include <ci/compat/ppc.h>
+#  include <ci/compat/gcc_ppc.h>
+# elif defined(__ia64__)
+#  include <ci/compat/ia64.h>
+#  include <ci/compat/gcc_ia64.h>
+# else
+#  error Unknown processor - GNU C
+# endif
+
+#elif defined(_MSC_VER)
+
+# include <ci/compat/msvc.h>
+
+# if defined(__i386__)
+#  include <ci/compat/x86.h>
+#  include <ci/compat/msvc_x86.h>
+# elif defined(__x86_64__)
+#  include <ci/compat/x86_64.h>
+#  include <ci/compat/msvc_x86_64.h>
+# else
+#  error Unknown processor MSC
+# endif
+
+#elif defined(__PGI)
+
+# include <ci/compat/x86.h>
+# include <ci/compat/pg_x86.h>
+
+#elif defined(__INTEL_COMPILER)
+
+/* Intel compilers v7 claim to be very gcc compatible. */
+# if __INTEL_COMPILER >= 700
+#  include <ci/compat/gcc.h>
+#  include <ci/compat/x86.h>
+#  include <ci/compat/gcc_x86.h>
+# else
+#  error Old Intel compiler not supported.  Yet.
+# endif
+
+#else
+# error Unknown compiler.
+#endif
+
+
+/**********************************************************************
+ * Misc stuff (that probably shouldn't be here).
+ */
+
+#ifdef __sun
+# ifdef __KERNEL__
+#  define _KERNEL
+#  define _SYSCALL32
+#  ifdef _LP64
+#   define _SYSCALL32_IMPL
+#  endif
+# else
+#  define _REENTRANT
+# endif
+#endif
+
+
+/**********************************************************************
+ * Defaults for anything left undefined.
+ */
+
+#ifndef  CI_LIKELY
+# define CI_LIKELY(t)    (t)
+# define CI_UNLIKELY(t)  (t)
+#endif
+
+#ifndef  ci_restrict
+# define ci_restrict
+#endif
+
+#ifndef  ci_inline
+# define ci_inline  static inline
+#endif
+
+#ifndef  ci_noinline
+# define ci_noinline  static
+#endif
+
+#endif  /* __CI_COMPAT_SYSDEP_H__ */
+
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/compat/utils.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/compat/utils.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/compat/utils.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/compat/utils.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,269 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*
+ * \author  djr
+ *  \brief  Handy utility macros.
+ *   \date  2003/01/17
+ */
+
+/*! \cidoxg_include_ci_compat  */
+
+#ifndef __CI_COMPAT_UTILS_H__
+#define __CI_COMPAT_UTILS_H__
+
+
+/**********************************************************************
+ * Alignment -- [align] must be a power of 2.
+ **********************************************************************/
+
+  /*! Align forward onto next boundary. */
+
+#define CI_ALIGN_FWD(p, align)               (((p)+(align)-1u) & ~((align)-1u))
+
+
+  /*! Align back onto prev boundary. */
+
+#define CI_ALIGN_BACK(p, align)              ((p) & ~((align)-1u))
+
+
+  /*! How far to next boundary? */
+
+#define CI_ALIGN_NEEDED(p, align, signed_t)  (-(signed_t)(p) & ((align)-1u))
+
+
+  /*! How far beyond prev boundary? */
+
+#define CI_OFFSET(p, align)                  ((p) & ((align)-1u))
+
+
+  /*! Does object fit in gap before next boundary? */
+
+#define CI_FITS(p, size, align, signed_t)			\
+  (CI_ALIGN_NEEDED((p) + 1, (align), signed_t) + 1 >= (size))
+
+
+  /*! Align forward onto next boundary. */
+
+#define CI_PTR_ALIGN_FWD(p, align)					   \
+  ((char*) CI_ALIGN_FWD(((ci_ptr_arith_t)(p)), ((ci_ptr_arith_t)(align))))
+
+  /*! Align back onto prev boundary. */
+
+#define CI_PTR_ALIGN_BACK(p, align)					    \
+  ((char*) CI_ALIGN_BACK(((ci_ptr_arith_t)(p)), ((ci_ptr_arith_t)(align))))
+
+  /*! How far to next boundary? */
+
+#define CI_PTR_ALIGN_NEEDED(p, align)					\
+  CI_ALIGN_NEEDED(((ci_ptr_arith_t)(p)), ((ci_ptr_arith_t)(align)),	\
+		  ci_ptr_arith_t)
+
+  /*! How far to next boundary? NZ = not zero i.e. give align if on boundary  */
+
+#define CI_PTR_ALIGN_NEEDED_NZ(p, align)					\
+  ((align) - (((char*)p) -                                                      \
+  ((char*) CI_ALIGN_BACK(((ci_ptr_arith_t)(p)), ((ci_ptr_arith_t)(align))))))
+
+  /*! How far beyond prev boundary? */
+
+#define CI_PTR_OFFSET(p, align)					\
+  CI_OFFSET(((ci_ptr_arith_t)(p)), ((ci_ptr_arith_t)(align)))
+
+
+  /* Same as CI_ALIGN_FWD and CI_ALIGN_BACK. */
+
+#define CI_ROUND_UP(i, align)      (((i)+(align)-1u) & ~((align)-1u))
+
+#define CI_ROUND_DOWN(i, align)    ((i) & ~((align)-1u))
+
+
+/**********************************************************************
+ * Byte-order
+ **********************************************************************/
+
+/* These are not flags.  They are enumeration values for use with
+ * CI_MY_BYTE_ORDER. */
+#define CI_BIG_ENDIAN          1
+#define CI_LITTLE_ENDIAN       0
+
+/*
+** Note that these byte-swapping primitives may leave junk in bits above
+** the range they operate on.
+**
+** The CI_BSWAP_nn() routines require that bits above [nn] are zero.  Use
+** CI_BSWAPM_nn(x) if this cannot be guaranteed.
+*/
+
+/* ?? May be able to improve on some of these with inline assembler on some
+** platforms.
+*/
+
+#define CI_BSWAP_16(v)    ((((v) & 0xff) << 8) | ((v) >> 8))
+#define CI_BSWAPM_16(v)   ((((v) & 0xff) << 8) | (((v) & 0xff00) >> 8))
+
+#define CI_BSWAP_32(v)    (((v) >> 24)               | 	\
+			   (((v) & 0x00ff0000) >> 8) |	\
+			   (((v) & 0x0000ff00) << 8) |	\
+			   ((v) << 24))
+#define CI_BSWAPM_32(v)   ((((v) & 0xff000000) >> 24) |	\
+			   (((v) & 0x00ff0000) >> 8)  |	\
+			   (((v) & 0x0000ff00) << 8)  |	\
+			   ((v) << 24))
+
+#define CI_BSWAP_64(v)    (((v) >> 56)                        |	\
+			   (((v) & 0x00ff000000000000) >> 40) |	\
+			   (((v) & 0x0000ff0000000000) >> 24) |	\
+			   (((v) & 0x000000ff00000000) >> 8)  |	\
+			   (((v) & 0x00000000ff000000) << 8)  |	\
+			   (((v) & 0x0000000000ff0000) << 24) |	\
+			   (((v) & 0x000000000000ff00) << 40) |	\
+			   ((v) << 56))
+
+# define CI_BSWAPPED_16_IF(c,v)  ((c) ? CI_BSWAP_16(v) : (v))
+# define CI_BSWAPPED_32_IF(c,v)  ((c) ? CI_BSWAP_32(v) : (v))
+# define CI_BSWAPPED_64_IF(c,v)  ((c) ? CI_BSWAP_64(v) : (v))
+# define CI_BSWAP_16_IF(c,v)     do{ if((c)) (v) = CI_BSWAP_16(v); }while(0)
+# define CI_BSWAP_32_IF(c,v)     do{ if((c)) (v) = CI_BSWAP_32(v); }while(0)
+# define CI_BSWAP_64_IF(c,v)     do{ if((c)) (v) = CI_BSWAP_64(v); }while(0)
+
+#if (CI_MY_BYTE_ORDER == CI_LITTLE_ENDIAN)
+# define CI_BSWAP_LE16(v)    (v)
+# define CI_BSWAP_LE32(v)    (v)
+# define CI_BSWAP_LE64(v)    (v)
+# define CI_BSWAP_BE16(v)    CI_BSWAP_16(v)
+# define CI_BSWAP_BE32(v)    CI_BSWAP_32(v)
+# define CI_BSWAP_BE64(v)    CI_BSWAP_64(v)
+# define CI_BSWAPM_LE16(v)   (v)
+# define CI_BSWAPM_LE32(v)   (v)
+# define CI_BSWAPM_LE64(v)   (v)
+# define CI_BSWAPM_BE16(v)   CI_BSWAPM_16(v)
+# define CI_BSWAPM_BE32(v)   CI_BSWAPM_32(v)
+#elif (CI_MY_BYTE_ORDER == CI_BIG_ENDIAN)
+# define CI_BSWAP_BE16(v)    (v)
+# define CI_BSWAP_BE32(v)    (v)
+# define CI_BSWAP_BE64(v)    (v)
+# define CI_BSWAP_LE16(v)    CI_BSWAP_16(v)
+# define CI_BSWAP_LE32(v)    CI_BSWAP_32(v)
+# define CI_BSWAP_LE64(v)    CI_BSWAP_64(v)
+# define CI_BSWAPM_BE16(v)   (v)
+# define CI_BSWAPM_BE32(v)   (v)
+# define CI_BSWAPM_BE64(v)   (v)
+# define CI_BSWAPM_LE16(v)   CI_BSWAPM_16(v)
+# define CI_BSWAPM_LE32(v)   CI_BSWAPM_32(v)
+#else
+# error Bad endian.
+#endif
+
+
+/**********************************************************************
+ * Get pointer to struct from pointer to member
+ **********************************************************************/
+
+#define CI_MEMBER_OFFSET(c_type, mbr_name)  \
+  ((ci_uint32) (ci_uintptr_t)(&((c_type*)0)->mbr_name))
+
+#define CI_MEMBER_SIZE(c_type, mbr_name)        \
+  sizeof(((c_type*)0)->mbr_name)
+
+#define __CI_CONTAINER(c_type, mbr_name, p_mbr)  \
+  ( (c_type*) ((char*)(p_mbr) - CI_MEMBER_OFFSET(c_type, mbr_name)) )
+
+#ifndef CI_CONTAINER
+# define CI_CONTAINER(t,m,p)  __CI_CONTAINER(t,m,p)
+#endif
+
+
+/**********************************************************************
+ * Structure member initialiser.
+ **********************************************************************/
+
+#ifndef CI_STRUCT_MBR
+# define CI_STRUCT_MBR(name, val)	.name = val
+#endif
+
+
+/**********************************************************************
+ * min / max
+ **********************************************************************/ 
+
+#define CI_MIN(x,y) (((x) < (y)) ? (x) : (y))
+#define CI_MAX(x,y) (((x) > (y)) ? (x) : (y))
+
+/**********************************************************************
+ * abs
+ **********************************************************************/ 
+
+#define CI_ABS(x) (((x) < 0) ? -(x) : (x))
+
+/**********************************************************************
+ * Conditional debugging
+ **********************************************************************/ 
+
+#ifdef NDEBUG
+# define CI_DEBUG(x)
+# define CI_NDEBUG(x)      x
+# define CI_IF_DEBUG(y,n)  (n)
+# define CI_DEBUG_ARG(x)
+#else
+# define CI_DEBUG(x)       x
+# define CI_NDEBUG(x)
+# define CI_IF_DEBUG(y,n)  (y)
+# define CI_DEBUG_ARG(x)   ,x
+#endif
+
+#ifdef __KERNEL__
+#define CI_KERNEL_ARG(x)   ,x
+#else
+#define CI_KERNEL_ARG(x)
+#endif
+
+#ifdef _WIN32
+# define CI_KERNEL_ARG_WIN(x) CI_KERNEL_ARG(x)
+# define CI_ARG_WIN(x) ,x
+#else
+# define CI_KERNEL_ARG_WIN(x)
+# define CI_ARG_WIN(x) 
+#endif
+
+#ifdef __unix__
+# define CI_KERNEL_ARG_UNIX(x) CI_KERNEL_ARG(x)
+# define CI_ARG_UNIX(x) ,x
+#else
+# define CI_KERNEL_ARG_UNIX(x)
+# define CI_ARG_UNIX(x) 
+#endif
+
+#ifdef __linux__
+# define CI_KERNEL_ARG_LINUX(x) CI_KERNEL_ARG(x)
+# define CI_ARG_LINUX(x) ,x
+#else
+# define CI_KERNEL_ARG_LINUX(x)
+# define CI_ARG_LINUX(x) 
+#endif
+
+
+#endif  /* __CI_COMPAT_UTILS_H__ */
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/compat/x86_64.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/compat/x86_64.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/compat/x86_64.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/compat/x86_64.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,54 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*
+ * \author  djr
+ *  \brief  Arch stuff for AMD x86_64.
+ *   \date  2004/08/17
+ */
+
+/*! \cidoxg_include_ci_compat  */
+#ifndef __CI_COMPAT_X86_64_H__
+#define __CI_COMPAT_X86_64_H__
+
+
+#define CI_MY_BYTE_ORDER	CI_LITTLE_ENDIAN
+
+#define CI_WORD_SIZE		8
+#define CI_PTR_SIZE		8
+
+#define CI_PAGE_SIZE		4096
+#define CI_PAGE_SHIFT		12
+#define CI_PAGE_MASK		(~(CI_PAGE_SIZE - 1))
+
+#define CI_CPU_HAS_SSE		1	/* SSE extensions supported */
+
+/* SSE2 disabled while investigating BUG1060 */
+#define CI_CPU_HAS_SSE2		0	/* SSE2 extensions supported */
+#define CI_CPU_OOS		0	/* CPU does out of order stores */
+
+
+#endif  /* __CI_COMPAT_X86_64_H__ */
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/compat/x86.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/compat/x86.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/compat/x86.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/compat/x86.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,48 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*! \cidoxg_include_ci_compat  */
+
+#ifndef __CI_COMPAT_X86_H__
+#define __CI_COMPAT_X86_H__
+
+
+#define CI_MY_BYTE_ORDER   CI_LITTLE_ENDIAN
+
+#define CI_WORD_SIZE       4
+#define CI_PTR_SIZE        4
+
+#define CI_PAGE_SIZE       4096
+#define CI_PAGE_SHIFT      12
+#define CI_PAGE_MASK       (~(CI_PAGE_SIZE - 1))
+
+#define CI_CPU_HAS_SSE	   1	/* SSE extensions supported */
+#define CI_CPU_HAS_SSE2	   0	/* SSE2 extensions supported */
+#define CI_CPU_OOS	   0	/* CPU does out of order stores */
+
+
+#endif  /* __CI_COMPAT_X86_H__ */
+
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/compat.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/compat.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/compat.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/compat.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,53 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*
+ * \author  djr
+ *  \brief  Compatability layer.  Provides definitions of fundamental
+ *          types and definitions that are used throughout CI source
+ *          code.  It does not introduce any link time dependencies,
+ *          or include any unnecessary system headers.
+ */
+/*! \cidoxg_include_ci */
+
+#ifndef __CI_COMPAT_H__
+#define __CI_COMPAT_H__
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#include <ci/compat/primitive.h>
+#include <ci/compat/sysdep.h>
+#include <ci/compat/utils.h>
+
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif  /* __CI_COMPAT_H__ */
+
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/driver/resource/efx_vi.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/driver/resource/efx_vi.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/driver/resource/efx_vi.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/driver/resource/efx_vi.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,276 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file contains public EFX VI API to Solarflare resource manager.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_DRIVER_RESOURCE_EFX_VI_H__
+#define __CI_DRIVER_RESOURCE_EFX_VI_H__
+
+/* Default size of event queue in the efx_vi resource.  Copied from
+ * CI_CFG_NETIF_EVENTQ_SIZE */
+#define EFX_VI_EVENTQ_SIZE_DEFAULT 1024
+
+extern int efx_vi_eventq_size;
+
+/**************************************************************************
+ * efx_vi_state types, allocation and free
+ **************************************************************************/
+
+/*! Handle for refering to a efx_vi */
+struct efx_vi_state;
+
+/*!
+ * Allocate an efx_vi, including event queue and pt_endpoint
+ *
+ * \param vih_out Pointer to a handle that is set on success
+ * \param nic_index Index of NIC to apply this resource to
+ * \return Zero on success (and vih_out set), non-zero on failure.
+ */
+extern int
+efx_vi_alloc(struct efx_vi_state **vih_out, int nic_index);
+
+/*!
+ * Free a previously allocated efx_vi
+ *
+ * \param vih The handle of the efx_vi to free
+ */
+extern void
+efx_vi_free(struct efx_vi_state *vih);
+
+/*!
+ * Reset a previously allocated efx_vi
+ *
+ * \param vih The handle of the efx_vi to reset
+ */
+extern void
+efx_vi_reset(struct efx_vi_state *vih);
+
+/**************************************************************************
+ * efx_vi_eventq types and functions
+ **************************************************************************/
+
+/*!
+ * Register a function to receive callbacks when event queue timeouts
+ * or wakeups occur.  Only one function per efx_vi can be registered
+ * at once.
+ *
+ * \param vih The handle to identify the efx_vi
+ * \param callback The function to callback
+ * \param context An argument to pass to the callback function
+ * \return Zero on success, non-zero on failure.
+ */
+extern int
+efx_vi_eventq_register_callback(struct efx_vi_state *vih,
+				void (*callback)(void *context, int is_timeout),
+				void *context);
+
+/*!
+ * Remove the current eventq timeout or wakeup callback function
+ *
+ * \param vih The handle to identify the efx_vi
+ * \return Zero on success, non-zero on failure
+ */
+extern int
+efx_vi_eventq_kill_callback(struct efx_vi_state *vih);
+
+/**************************************************************************
+ * efx_vi_dma_map types and functions
+ **************************************************************************/
+
+/*!
+ * Handle for refering to a efx_vi
+ */
+struct efx_vi_dma_map_state;
+
+/*!
+ * Map a list of buffer pages so they are registered with the hardware
+ *
+ * \param vih The handle to identify the efx_vi
+ * \param addrs An array of page pointers to map
+ * \param n_addrs Length of the page pointer array.  Must be a power of two.
+ * \param dmh_out Set on success to a handle used to refer to this mapping
+ * \return Zero on success, non-zero on failure.
+ */
+extern int
+efx_vi_dma_map_pages(struct efx_vi_state *vih, struct page **pages,
+			 int n_pages, struct efx_vi_dma_map_state **dmh_out);
+extern int
+efx_vi_dma_map_addrs(struct efx_vi_state *vih,
+		     unsigned long long *dev_bus_addrs, int n_pages,
+		     struct efx_vi_dma_map_state **dmh_out);
+
+/*!
+ * Unmap a previously mapped set of pages so they are no longer registered
+ * with the hardware.
+ *
+ * \param vih The handle to identify the efx_vi
+ * \param dmh The handle to identify the dma mapping
+ */
+extern void
+efx_vi_dma_unmap_pages(struct efx_vi_state *vih,
+		       struct efx_vi_dma_map_state *dmh);
+extern void
+efx_vi_dma_unmap_addrs(struct efx_vi_state *vih,
+		       struct efx_vi_dma_map_state *dmh);
+
+/*!
+ * Retrieve the buffer address of the mapping
+ *
+ * \param vih The handle to identify the efx_vi
+ * \param dmh The handle to identify the buffer mapping
+ * \return The buffer address on success, or zero on failure
+ */
+extern unsigned
+efx_vi_dma_get_map_addr(struct efx_vi_state *vih,
+			struct efx_vi_dma_map_state *dmh);
+
+/**************************************************************************
+ * efx_vi filter functions
+ **************************************************************************/
+
+#define EFX_VI_STATIC_FILTERS 32
+
+/*! Handle to refer to a filter instance */
+struct filter_resource_t;
+
+/*!
+ * Allocate and add a filter
+ *
+ * \param vih The handle to identify the efx_vi
+ * \param protocol The protocol of the new filter: UDP or TCP
+ * \param ip_addr_be32 The local ip address of the filter
+ * \param port_le16 The local port of the filter
+ * \param fh_out Set on success to be a handle to refer to this filter
+ * \return Zero on success, non-zero on failure.
+ */
+extern int
+efx_vi_filter(struct efx_vi_state *vih, int protocol, unsigned ip_addr_be32,
+	      int port_le16, struct filter_resource_t **fh_out);
+
+/*!
+ * Remove a filter and free resources associated with it
+ *
+ * \param vih The handle to identify the efx_vi
+ * \param fh The handle to identify the filter
+ * \return Zero on success, non-zero on failure
+ */
+extern int
+efx_vi_filter_stop(struct efx_vi_state *vih, struct filter_resource_t *fh);
+
+/**************************************************************************
+ * efx_vi hw resources types and functions
+ **************************************************************************/
+
+/*! Constants for the type field in efx_vi_hw_resource */
+#define EFX_VI_HW_RESOURCE_TXDMAQ    0x0	/* PFN of TX DMA Q */
+#define EFX_VI_HW_RESOURCE_RXDMAQ    0x1	/* PFN of RX DMA Q */
+#define EFX_VI_HW_RESOURCE_TXBELL    0x2	/* PFN of TX Doorbell (EF1) */
+#define EFX_VI_HW_RESOURCE_RXBELL    0x3	/* PFN of RX Doorbell (EF1) */
+#define EFX_VI_HW_RESOURCE_EVQTIMER  0x4	/* Address of event q timer */
+
+/* Address of event q pointer (EF1) */
+#define EFX_VI_HW_RESOURCE_EVQPTR    0x5
+/* Address of register pointer (Falcon A) */
+#define EFX_VI_HW_RESOURCE_EVQRPTR   0x6
+/* Offset of register pointer (Falcon B) */
+#define EFX_VI_HW_RESOURCE_EVQRPTR_OFFSET 0x7
+/* Address of mem KVA */
+#define EFX_VI_HW_RESOURCE_EVQMEMKVA 0x8
+/* PFN of doorbell page (Falcon) */
+#define EFX_VI_HW_RESOURCE_BELLPAGE  0x9
+
+/*! How large an array to allocate for the get_() functions - smaller
+  than the total number of constants as some are mutually exclusive */
+#define EFX_VI_HW_RESOURCE_MAXSIZE   0x7
+
+/*! Constants for the mem_type field in efx_vi_hw_resource */
+#define EFX_VI_HW_RESOURCE_IOBUFFER   0	/* Host memory */
+#define EFX_VI_HW_RESOURCE_PERIPHERAL 1	/* Card memory/registers */
+
+/*!
+ * Data structure providing information on a hardware resource mapping
+ */
+struct efx_vi_hw_resource {
+	u8 type;		/*!< What this resource represents */
+	u8 mem_type;		/*!< What type of memory is it in, eg,
+				 * host or iomem */
+	u8 more_to_follow;	/*!< Is this part of a multi-region resource */
+	u32 length;		/*!< Length of the resource in bytes */
+	unsigned long address;	/*!< Address of this resource */
+};
+
+/*!
+ * Metadata concerning the list of hardware resource mappings
+ */
+struct efx_vi_hw_resource_metadata {
+	int version;
+	int evq_order;
+	int evq_offs;
+	int evq_capacity;
+	int instance;
+	unsigned rx_capacity;
+	unsigned tx_capacity;
+	int nic_arch;
+	int nic_revision;
+	char nic_variant;
+};
+
+/*!
+ * Obtain a list of hardware resource mappings, using virtual addresses
+ *
+ * \param vih The handle to identify the efx_vi
+ * \param mdata Pointer to a structure to receive the metadata
+ * \param hw_res_array An array to receive the list of hardware resources
+ * \param length The length of hw_res_array.  Updated on success to contain
+ * the number of entries in the supplied array that were used.
+ * \return Zero on success, non-zero on failure
+ */
+extern int
+efx_vi_hw_resource_get_virt(struct efx_vi_state *vih,
+			    struct efx_vi_hw_resource_metadata *mdata,
+			    struct efx_vi_hw_resource *hw_res_array,
+			    int *length);
+
+/*!
+ * Obtain a list of hardware resource mappings, using physical addresses
+ *
+ * \param vih The handle to identify the efx_vi
+ * \param mdata Pointer to a structure to receive the metadata
+ * \param hw_res_array An array to receive the list of hardware resources
+ * \param length The length of hw_res_array.  Updated on success to contain
+ * the number of entries in the supplied array that were used.
+ * \return Zero on success, non-zero on failure
+ */
+extern int
+efx_vi_hw_resource_get_phys(struct efx_vi_state *vih,
+			    struct efx_vi_hw_resource_metadata *mdata,
+			    struct efx_vi_hw_resource *hw_res_array,
+			    int *length);
+
+#endif /* __CI_DRIVER_RESOURCE_EFX_VI_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/common.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/common.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/common.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/common.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,102 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides API of the efhw library which may be used both from
+ * the kernel and from the user-space code.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_COMMON_H__
+#define __CI_EFHW_COMMON_H__
+
+#include <ci/efhw/common_sysdep.h>
+
+enum efhw_arch {
+	EFHW_ARCH_FALCON,
+	EFHW_ARCH_SIENA,
+};
+
+typedef uint32_t efhw_buffer_addr_t;
+#define EFHW_BUFFER_ADDR_FMT	"[ba:%"PRIx32"]"
+
+/*! Comment? */
+typedef union {
+	uint64_t u64;
+	struct {
+		uint32_t a;
+		uint32_t b;
+	} opaque;
+	struct {
+		uint32_t code;
+		uint32_t status;
+	} ev1002;
+} efhw_event_t;
+
+/* Flags for TX/RX queues */
+#define EFHW_VI_JUMBO_EN           0x01  /*! scatter RX over multiple desc */
+#define EFHW_VI_ISCSI_RX_HDIG_EN   0x02  /*! iscsi rx header digest */
+#define EFHW_VI_ISCSI_TX_HDIG_EN   0x04  /*! iscsi tx header digest */
+#define EFHW_VI_ISCSI_RX_DDIG_EN   0x08  /*! iscsi rx data digest */
+#define EFHW_VI_ISCSI_TX_DDIG_EN   0x10  /*! iscsi tx data digest */
+#define EFHW_VI_TX_PHYS_ADDR_EN    0x20  /*! TX physical address mode */
+#define EFHW_VI_RX_PHYS_ADDR_EN    0x40  /*! RX physical address mode */
+#define EFHW_VI_RM_WITH_INTERRUPT  0x80  /*! VI with an interrupt */
+#define EFHW_VI_TX_IP_CSUM_DIS     0x100 /*! enable ip checksum generation */
+#define EFHW_VI_TX_TCPUDP_CSUM_DIS 0x200 /*! enable tcp/udp checksum
+					   generation */
+#define EFHW_VI_TX_TCPUDP_ONLY     0x400 /*! drop non-tcp/udp packets */
+
+/* Types of hardware filter */
+/* Each of these values implicitly selects scatter filters on B0 - or in
+   EFHW_IP_FILTER_TYPE_NOSCAT_B0_MASK if a non-scatter filter is required */
+#define EFHW_IP_FILTER_TYPE_UDP_WILDCARD  (0)	/* dest host only */
+#define EFHW_IP_FILTER_TYPE_UDP_FULL      (1)	/* dest host and port */
+#define EFHW_IP_FILTER_TYPE_TCP_WILDCARD  (2)	/* dest based filter */
+#define EFHW_IP_FILTER_TYPE_TCP_FULL      (3)	/* src  filter */
+/* Same again, but with RSS (for B0 only) */
+#define EFHW_IP_FILTER_TYPE_UDP_WILDCARD_RSS_B0  (4)
+#define EFHW_IP_FILTER_TYPE_UDP_FULL_RSS_B0      (5)
+#define EFHW_IP_FILTER_TYPE_TCP_WILDCARD_RSS_B0  (6)
+#define EFHW_IP_FILTER_TYPE_TCP_FULL_RSS_B0      (7)
+
+#define EFHW_IP_FILTER_TYPE_FULL_MASK      (0x1) /* Mask for full / wildcard */
+#define EFHW_IP_FILTER_TYPE_TCP_MASK       (0x2) /* Mask for TCP type */
+#define EFHW_IP_FILTER_TYPE_RSS_B0_MASK    (0x4) /* Mask for B0 RSS enable */
+#define EFHW_IP_FILTER_TYPE_NOSCAT_B0_MASK (0x8) /* Mask for B0 SCATTER dsbl */
+
+#define EFHW_IP_FILTER_TYPE_MASK	(0xffff) /* Mask of types above */
+
+#define EFHW_IP_FILTER_BROADCAST	(0x10000) /* driverlink filter
+						     support */
+
+#endif /* __CI_EFHW_COMMON_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/common_sysdep.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/common_sysdep.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/common_sysdep.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/common_sysdep.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,67 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides version-independent Linux kernel API for
+ * userland-to-kernel interfaces.
+ * Only kernels >=2.6.9 are supported.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_COMMON_LINUX_H__
+#define __CI_EFHW_COMMON_LINUX_H__
+
+#include <linux/types.h>
+#include <linux/version.h>
+
+/* Dirty hack, but Linux kernel does not provide DMA_ADDR_T_FMT */
+#if BITS_PER_LONG == 64 || defined(CONFIG_HIGHMEM64G)
+#define DMA_ADDR_T_FMT "%llx"
+#else
+#define DMA_ADDR_T_FMT "%x"
+#endif
+
+/* Linux kernel also does not provide PRIx32... Sigh. */
+#define PRIx32 "x"
+#define PRIx64 "llx"
+
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,19)
+enum {
+	false = 0,
+	true = 1
+};
+
+typedef _Bool bool;
+#endif /* LINUX_VERSION_CODE < 2.6.19 */
+
+#endif /* __CI_EFHW_COMMON_LINUX_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/debug.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/debug.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/debug.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/debug.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,84 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides debug-related API for efhw library using Linux kernel
+ * primitives.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_DEBUG_LINUX_H__
+#define __CI_EFHW_DEBUG_LINUX_H__
+
+#define EFHW_PRINTK_PREFIX "[sfc efhw] "
+
+#define EFHW_PRINTK(level, fmt, ...) \
+	printk(level EFHW_PRINTK_PREFIX fmt "\n", __VA_ARGS__)
+
+/* Following macros should be used with non-zero format parameters
+ * due to __VA_ARGS__ limitations.  Use "%s" with __FUNCTION__ if you can't
+ * find better parameters. */
+#define EFHW_ERR(fmt, ...)     EFHW_PRINTK(KERN_ERR, fmt, __VA_ARGS__)
+#define EFHW_WARN(fmt, ...)    EFHW_PRINTK(KERN_WARNING, fmt, __VA_ARGS__)
+#define EFHW_NOTICE(fmt, ...)  EFHW_PRINTK(KERN_NOTICE, fmt, __VA_ARGS__)
+#if 0 && !defined(NDEBUG)
+#define EFHW_TRACE(fmt, ...) EFHW_PRINTK(KERN_DEBUG, fmt, __VA_ARGS__)
+#else
+#define EFHW_TRACE(fmt, ...)
+#endif
+
+#ifndef NDEBUG
+#define EFHW_ASSERT(cond)  BUG_ON((cond) == 0)
+#define EFHW_DO_DEBUG(expr) expr
+#else
+#define EFHW_ASSERT(cond)
+#define EFHW_DO_DEBUG(expr)
+#endif
+
+#define EFHW_TEST(expr)			\
+	do {				\
+		if (unlikely(!(expr)))	\
+		BUG();			\
+	} while (0)
+
+/* Build time asserts. We paste the line number into the type name
+ * so that the macro can be used more than once per file even if the
+ * compiler objects to multiple identical typedefs. Collisions
+ * between use in different header files is still possible. */
+#ifndef EFHW_BUILD_ASSERT
+#define __EFHW_BUILD_ASSERT_NAME(_x) __EFHW_BUILD_ASSERT_ILOATHECPP(_x)
+#define __EFHW_BUILD_ASSERT_ILOATHECPP(_x)  __EFHW_BUILD_ASSERT__ ##_x
+#define EFHW_BUILD_ASSERT(e) \
+	typedef char __EFHW_BUILD_ASSERT_NAME(__LINE__)[(e) ? 1 : -1]
+#endif
+
+#endif /* __CI_EFHW_DEBUG_LINUX_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/efhw_config.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/efhw_config.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/efhw_config.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/efhw_config.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,43 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides some limits used in both kernel and userland code.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_EFAB_CONFIG_H__
+#define __CI_EFHW_EFAB_CONFIG_H__
+
+#define EFHW_MAX_NR_DEVS 5	/* max number of efhw devices supported */
+
+#endif /* __CI_EFHW_EFAB_CONFIG_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/efhw_types.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/efhw_types.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/efhw_types.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/efhw_types.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,342 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides struct efhw_nic and some related types.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_EFAB_TYPES_H__
+#define __CI_EFHW_EFAB_TYPES_H__
+
+#include <ci/efhw/efhw_config.h>
+#include <ci/efhw/hardware_sysdep.h>
+#include <ci/efhw/iopage_types.h>
+#include <ci/efhw/sysdep.h>
+
+/*--------------------------------------------------------------------
+ *
+ * hardware limits used in the types
+ *
+ *--------------------------------------------------------------------*/
+
+#define EFHW_KEVENTQ_MAX    8
+
+/*--------------------------------------------------------------------
+ *
+ * forward type declarations
+ *
+ *--------------------------------------------------------------------*/
+
+struct efhw_nic;
+
+/*--------------------------------------------------------------------
+ *
+ * Managed interface
+ *
+ *--------------------------------------------------------------------*/
+
+struct efhw_buffer_table_allocation{
+	unsigned base;
+	unsigned order;
+};
+
+struct eventq_resource_hardware {
+	/*!iobuffer allocated for eventq - can be larger than eventq */
+	efhw_iopages_t iobuff;
+	unsigned iobuff_off;
+	struct efhw_buffer_table_allocation buf_tbl_alloc;
+	int capacity;		/*!< capacity of event queue */
+};
+
+/*--------------------------------------------------------------------
+ *
+ * event queues and event driven callbacks
+ *
+ *--------------------------------------------------------------------*/
+
+struct efhw_keventq {
+	volatile int lock;
+	caddr_t evq_base;
+	int32_t evq_ptr;
+	uint32_t evq_mask;
+	unsigned instance;
+	struct eventq_resource_hardware hw;
+	struct efhw_ev_handler *ev_handlers;
+};
+
+/**********************************************************************
+ * Portable HW interface. ***************************************
+ **********************************************************************/
+
+/*--------------------------------------------------------------------
+ *
+ * EtherFabric Functional units - configuration and control
+ *
+ *--------------------------------------------------------------------*/
+
+struct efhw_func_ops {
+
+  /*-------------- Initialisation ------------ */
+
+	/*! close down all hardware functional units - leaves NIC in a safe
+	   state for driver unload */
+	void (*close_hardware) (struct efhw_nic *nic);
+
+	/*! initialise all hardware functional units */
+	int (*init_hardware) (struct efhw_nic *nic,
+			      struct efhw_ev_handler *,
+			      const uint8_t *mac_addr);
+
+  /*-------------- Interrupt support  ------------ */
+
+	/*! Main interrupt routine
+	 **        This function returns,
+	 **  - zero,       if the IRQ was not generated by EF1
+	 **  - non-zero,   if EF1 was the source of the IRQ
+	 **
+	 **
+	 ** opaque is an OS provided pointer for use by the OS callbacks
+	 ** e.g in Windows used to indicate DPC scheduled
+	 */
+	int (*interrupt) (struct efhw_nic *nic);
+
+	/*! Enable given interrupt mask for the given IRQ unit */
+	void (*interrupt_enable) (struct efhw_nic *nic, uint idx);
+
+	/*! Disable given interrupt mask for the given IRQ unit */
+	void (*interrupt_disable) (struct efhw_nic *nic, uint idx);
+
+	/*! Set interrupt moderation strategy for the given IRQ unit
+	 ** val is in usec
+	 */
+	void (*set_interrupt_moderation)(struct efhw_nic *nic,
+					 uint idx, uint val);
+
+  /*-------------- Event support  ------------ */
+
+	/*! Enable the given event queue
+	   depending on the underlying implementation (EF1 or Falcon) then
+	   either a q_base_addr in host memory, or a buffer base id should
+	   be proivded
+	 */
+	void (*event_queue_enable) (struct efhw_nic *nic,
+				    uint evq,	/* evnt queue index */
+				    uint evq_size,	/* units of #entries */
+				    dma_addr_t q_base_addr, uint buf_base_id);
+
+	/*! Disable the given event queue (and any associated timer) */
+	void (*event_queue_disable) (struct efhw_nic *nic, uint evq,
+				     int timer_only);
+
+	/*! request wakeup from the NIC on a given event Q */
+	void (*wakeup_request) (struct efhw_nic *nic, dma_addr_t q_base_addr,
+				int next_i, int evq);
+
+	/*! Push a SW event on a given eventQ */
+	void (*sw_event) (struct efhw_nic *nic, int data, int evq);
+
+  /*-------------- Filter support  ------------ */
+
+	/*! Setup a given filter - The software can request a filter_i,
+	 * but some EtherFabric implementations will override with
+	 * a more suitable index
+	 */
+	int (*ipfilter_set) (struct efhw_nic *nic, int type,
+			     int *filter_i, int dmaq,
+			     unsigned saddr_be32, unsigned sport_be16,
+			     unsigned daddr_be32, unsigned dport_be16);
+
+	/*! Attach a given filter to a DMAQ */
+	void (*ipfilter_attach) (struct efhw_nic *nic, int filter_idx,
+				 int dmaq_idx);
+
+	/*! Detach a filter from its DMAQ */
+	void (*ipfilter_detach) (struct efhw_nic *nic, int filter_idx);
+
+	/*! Clear down a given filter */
+	void (*ipfilter_clear) (struct efhw_nic *nic, int filter_idx);
+
+  /*-------------- DMA support  ------------ */
+
+	/*! Initialise NIC state for a given TX DMAQ */
+	void (*dmaq_tx_q_init) (struct efhw_nic *nic,
+				uint dmaq, uint evq, uint owner, uint tag,
+				uint dmaq_size, uint buf_idx, uint flags);
+
+	/*! Initialise NIC state for a given RX DMAQ */
+	void (*dmaq_rx_q_init) (struct efhw_nic *nic,
+				uint dmaq, uint evq, uint owner, uint tag,
+				uint dmaq_size, uint buf_idx, uint flags);
+
+	/*! Disable a given TX DMAQ */
+	void (*dmaq_tx_q_disable) (struct efhw_nic *nic, uint dmaq);
+
+	/*! Disable a given RX DMAQ */
+	void (*dmaq_rx_q_disable) (struct efhw_nic *nic, uint dmaq);
+
+	/*! Flush a given TX DMA channel */
+	int (*flush_tx_dma_channel) (struct efhw_nic *nic, uint dmaq);
+
+	/*! Flush a given RX DMA channel */
+	int (*flush_rx_dma_channel) (struct efhw_nic *nic, uint dmaq);
+
+  /*-------------- Buffer table Support ------------ */
+
+	/*! Initialise a buffer table page */
+	void (*buffer_table_set) (struct efhw_nic *nic,
+				  dma_addr_t dma_addr,
+				  uint bufsz, uint region,
+				  int own_id, int buffer_id);
+
+	/*! Initialise a block of buffer table pages */
+	void (*buffer_table_set_n) (struct efhw_nic *nic, int buffer_id,
+				    dma_addr_t dma_addr,
+				    uint bufsz, uint region,
+				    int n_pages, int own_id);
+
+	/*! Clear a block of buffer table pages */
+	void (*buffer_table_clear) (struct efhw_nic *nic, int buffer_id,
+				    int num);
+
+	/*! Commit a buffer table update  */
+	void (*buffer_table_commit) (struct efhw_nic *nic);
+
+};
+
+
+/*----------------------------------------------------------------------------
+ *
+ * NIC type
+ *
+ *---------------------------------------------------------------------------*/
+
+struct efhw_device_type {
+	int  arch;            /* enum efhw_arch */
+	char variant;         /* 'A', 'B', ... */
+	int  revision;        /* 0, 1, ... */
+};
+
+
+/*----------------------------------------------------------------------------
+ *
+ * EtherFabric NIC instance - nic.c for HW independent functions
+ *
+ *---------------------------------------------------------------------------*/
+
+/*! */
+struct efhw_nic {
+	/*! zero base index in efrm_nic_table.nic array */
+	volatile int index;
+	int ifindex;		/*!< OS level nic index */
+#ifdef HAS_NET_NAMESPACE
+	struct net *nd_net;
+#endif
+
+	struct efhw_device_type devtype;
+
+	/*! Options that can be set by user. */
+	unsigned options;
+# define NIC_OPT_EFTEST             0x1	/* owner is an eftest app */
+
+# define NIC_OPT_DEFAULT            0
+
+	/*! Internal flags that indicate hardware properties at runtime. */
+	unsigned flags;
+# define NIC_FLAG_NO_INTERRUPT          0x01 /* to be set at init time only */
+# define NIC_FLAG_TRY_MSI               0x02
+# define NIC_FLAG_MSI                   0x04
+# define NIC_FLAG_OS_IRQ_EN             0x08
+# define NIC_FLAG_10G                   0x10
+
+	unsigned mtu;		/*!< MAC MTU (includes MAC hdr) */
+
+	/* hardware resources */
+
+	/*! I/O address of the start of the bar */
+	efhw_ioaddr_t bar_ioaddr;
+
+	/*! Bar number of control aperture. */
+	unsigned ctr_ap_bar;
+	/*! Length of control aperture in bytes. */
+	unsigned ctr_ap_bytes;
+
+	uint8_t mac_addr[ETH_ALEN];	/*!< mac address  */
+
+	/*! EtherFabric Functional Units -- functions */
+	const struct efhw_func_ops *efhw_func;
+
+	/* Value read from FPGA version register.  Zero for asic. */
+	unsigned fpga_version;
+
+	/*! This lock protects a number of misc NIC resources.  It should
+	 * only be used for things that can be at the bottom of the lock
+	 * order.  ie. You mustn't attempt to grab any other lock while
+	 * holding this one.
+	 */
+	spinlock_t *reg_lock;
+	spinlock_t the_reg_lock;
+
+	int buf_commit_outstanding;	/*!< outstanding buffer commits */
+
+	/*! interrupt callbacks (hard-irq) */
+	void (*irq_handler) (struct efhw_nic *, int unit);
+
+	/*! event queues per driver */
+	struct efhw_keventq evq[EFHW_KEVENTQ_MAX];
+
+/* for marking when we are not using an IRQ unit
+      - 0 is a valid offset to an IRQ unit on EF1! */
+#define EFHW_IRQ_UNIT_UNUSED  0xffff
+	/*! interrupt unit in use  */
+	unsigned int irq_unit[EFHW_KEVENTQ_MAX];
+	efhw_iopage_t irq_iobuff;	/*!<  Falcon SYSERR interrupt */
+
+	/* The new driverlink infrastructure. */
+	struct efx_dl_device *net_driver_dev;
+	struct efx_dlfilt_cb_s *dlfilter_cb;
+
+	/*! Bit masks of the sizes of event queues and dma queues supported
+	 * by the nic. */
+	unsigned evq_sizes;
+	unsigned rxq_sizes;
+	unsigned txq_sizes;
+
+	/* Size of filter table (including odd and even banks). */
+	unsigned filter_tbl_size;
+};
+
+
+#define EFHW_KVA(nic)       ((nic)->bar_ioaddr)
+
+
+#endif /* __CI_EFHW_EFHW_TYPES_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/hardware_sysdep.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/hardware_sysdep.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/hardware_sysdep.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/hardware_sysdep.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,84 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides version-independent Linux kernel API for header files
+ * with hardware-related definitions (in ci/driver/efab/hardware*).
+ * Only kernels >=2.6.9 are supported.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_HARDWARE_LINUX_H__
+#define __CI_EFHW_HARDWARE_LINUX_H__
+
+#include <asm/io.h>
+
+#ifdef __LITTLE_ENDIAN
+#define EFHW_IS_LITTLE_ENDIAN
+#elif __BIG_ENDIAN
+#define EFHW_IS_BIG_ENDIAN
+#else
+#error Unknown endianness
+#endif
+
+#ifndef mmiowb
+	#if defined(__i386__) || defined(__x86_64__)
+		#define mmiowb()
+	#elif defined(__ia64__)
+		#ifndef ia64_mfa
+			#define ia64_mfa() asm volatile ("mf.a" ::: "memory")
+		#endif
+	#define mmiowb ia64_mfa
+	#else
+	#error "Need definition for mmiowb()"
+	#endif
+#endif
+
+typedef char *efhw_ioaddr_t;
+
+#ifndef readq
+static inline uint64_t __readq(void __iomem *addr)
+{
+	return *(volatile uint64_t *)addr;
+}
+#define readq(x) __readq(x)
+#endif
+
+#ifndef writeq
+static inline void __writeq(uint64_t v, void __iomem *addr)
+{
+	*(volatile uint64_t *)addr = v;
+}
+#define writeq(val, addr) __writeq((val), (addr))
+#endif
+
+#endif /* __CI_EFHW_HARDWARE_LINUX_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/iopage_types.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/iopage_types.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/iopage_types.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/iopage_types.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,188 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides efhw_page_t and efhw_iopage_t for Linux kernel.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_IOPAGE_LINUX_H__
+#define __CI_EFHW_IOPAGE_LINUX_H__
+
+#include <linux/gfp.h>
+#include <linux/hardirq.h>
+#include <ci/efhw/debug.h>
+
+/*--------------------------------------------------------------------
+ *
+ * efhw_page_t: A single page of memory.  Directly mapped in the driver,
+ * and can be mapped to userlevel.
+ *
+ *--------------------------------------------------------------------*/
+
+typedef struct {
+	unsigned long kva;
+} efhw_page_t;
+
+static inline int efhw_page_alloc(efhw_page_t *p)
+{
+	p->kva = __get_free_page(in_interrupt()? GFP_ATOMIC : GFP_KERNEL);
+	return p->kva ? 0 : -ENOMEM;
+}
+
+static inline int efhw_page_alloc_zeroed(efhw_page_t *p)
+{
+	p->kva = get_zeroed_page(in_interrupt()? GFP_ATOMIC : GFP_KERNEL);
+	return p->kva ? 0 : -ENOMEM;
+}
+
+static inline void efhw_page_free(efhw_page_t *p)
+{
+	free_page(p->kva);
+	EFHW_DO_DEBUG(memset(p, 0, sizeof(*p)));
+}
+
+static inline char *efhw_page_ptr(efhw_page_t *p)
+{
+	return (char *)p->kva;
+}
+
+static inline unsigned efhw_page_pfn(efhw_page_t *p)
+{
+	return (unsigned)(__pa(p->kva) >> PAGE_SHIFT);
+}
+
+static inline void efhw_page_mark_invalid(efhw_page_t *p)
+{
+	p->kva = 0;
+}
+
+static inline int efhw_page_is_valid(efhw_page_t *p)
+{
+	return p->kva != 0;
+}
+
+static inline void efhw_page_init_from_va(efhw_page_t *p, void *va)
+{
+	p->kva = (unsigned long)va;
+}
+
+/*--------------------------------------------------------------------
+ *
+ * efhw_iopage_t: A single page of memory.  Directly mapped in the driver,
+ * and can be mapped to userlevel.  Can also be accessed by the NIC.
+ *
+ *--------------------------------------------------------------------*/
+
+typedef struct {
+	efhw_page_t p;
+	dma_addr_t dma_addr;
+} efhw_iopage_t;
+
+static inline dma_addr_t efhw_iopage_dma_addr(efhw_iopage_t *p)
+{
+	return p->dma_addr;
+}
+
+#define efhw_iopage_ptr(iop)		efhw_page_ptr(&(iop)->p)
+#define efhw_iopage_pfn(iop)		efhw_page_pfn(&(iop)->p)
+#define efhw_iopage_mark_invalid(iop)	efhw_page_mark_invalid(&(iop)->p)
+#define efhw_iopage_is_valid(iop)	efhw_page_is_valid(&(iop)->p)
+
+/*--------------------------------------------------------------------
+ *
+ * efhw_iopages_t: A set of pages that are contiguous in physical memory.
+ * Directly mapped in the driver, and can be mapped to userlevel.  Can also
+ * be accessed by the NIC.
+ *
+ * NB. The O/S may be unwilling to allocate many, or even any of these.  So
+ * only use this type where the NIC really needs a physically contiguous
+ * buffer.
+ *
+ *--------------------------------------------------------------------*/
+
+typedef struct {
+	caddr_t kva;
+	unsigned order;
+	dma_addr_t dma_addr;
+} efhw_iopages_t;
+
+static inline caddr_t efhw_iopages_ptr(efhw_iopages_t *p)
+{
+	return p->kva;
+}
+
+static inline unsigned efhw_iopages_pfn(efhw_iopages_t *p)
+{
+	return (unsigned)(__pa(p->kva) >> PAGE_SHIFT);
+}
+
+static inline dma_addr_t efhw_iopages_dma_addr(efhw_iopages_t *p)
+{
+	return p->dma_addr;
+}
+
+static inline unsigned efhw_iopages_size(efhw_iopages_t *p)
+{
+	return 1u << (p->order + PAGE_SHIFT);
+}
+
+/* efhw_iopage_t <-> efhw_iopages_t conversions for handling physically
+ * contiguous allocations in iobufsets for iSCSI.  This allows the
+ * essential information about contiguous allocations from
+ * efhw_iopages_alloc() to be saved away in the efhw_iopage_t array in an
+ * iobufset.  (Changing the iobufset resource to use a union type would
+ * involve a lot of code changes, and make the iobufset's metadata larger
+ * which could be bad as it's supposed to fit into a single page on some
+ * platforms.)
+ */
+static inline void
+efhw_iopage_init_from_iopages(efhw_iopage_t *iopage,
+			    efhw_iopages_t *iopages, unsigned pageno)
+{
+	iopage->p.kva = ((unsigned long)efhw_iopages_ptr(iopages))
+	    + (pageno * PAGE_SIZE);
+	iopage->dma_addr = efhw_iopages_dma_addr(iopages) +
+	    (pageno * PAGE_SIZE);
+}
+
+static inline void
+efhw_iopages_init_from_iopage(efhw_iopages_t *iopages,
+			    efhw_iopage_t *iopage, unsigned order)
+{
+	iopages->kva = (caddr_t) efhw_iopage_ptr(iopage);
+	EFHW_ASSERT(iopages->kva);
+	iopages->order = order;
+	iopages->dma_addr = efhw_iopage_dma_addr(iopage);
+}
+
+#endif /* __CI_EFHW_IOPAGE_LINUX_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/public.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/public.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/public.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/public.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,83 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides public API of efhw library exported from the SFC
+ * resource driver.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_PUBLIC_H__
+#define __CI_EFHW_PUBLIC_H__
+
+#include <ci/efhw/common.h>
+#include <ci/efhw/efhw_types.h>
+
+/*! Returns true if we have some EtherFabric functional units -
+  whether configured or not */
+static inline int efhw_nic_have_functional_units(struct efhw_nic *nic)
+{
+	return nic->efhw_func != 0;
+}
+
+/*! Returns true if the EtherFabric functional units have been configured  */
+static inline int efhw_nic_have_hw(struct efhw_nic *nic)
+{
+	return efhw_nic_have_functional_units(nic) && (EFHW_KVA(nic) != 0);
+}
+
+/*! Helper function to allocate the iobuffer needed by an eventq
+ *   - it ensures the eventq has the correct alignment for the NIC
+ *
+ * \param rm        Event-queue resource manager
+ * \param instance  Event-queue instance (index)
+ * \param buf_bytes Requested size of eventq
+ * \return          < 0 if iobuffer allocation fails
+ */
+int efhw_nic_event_queue_alloc_iobuffer(struct efhw_nic *nic,
+					struct eventq_resource_hardware *h,
+					int evq_instance, unsigned buf_bytes);
+
+extern void falcon_nic_set_rx_usr_buf_size(struct efhw_nic *,
+					   int rx_usr_buf_size);
+
+extern void
+falcon_nic_rx_filter_ctl_set(struct efhw_nic *nic, uint32_t tcp_full,
+			     uint32_t tcp_wild,
+			     uint32_t udp_full, uint32_t udp_wild);
+
+extern void
+falcon_nic_rx_filter_ctl_get(struct efhw_nic *nic, uint32_t *tcp_full,
+			     uint32_t *tcp_wild,
+			     uint32_t *udp_full, uint32_t *udp_wild);
+
+#endif /* __CI_EFHW_PUBLIC_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/sysdep.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/sysdep.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/efhw/sysdep.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efhw/sysdep.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,72 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides version-independent Linux kernel API for efhw library.
+ * Only kernels >=2.6.9 are supported.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFHW_SYSDEP_LINUX_H__
+#define __CI_EFHW_SYSDEP_LINUX_H__
+
+#include <linux/version.h>
+#include <linux/module.h>
+#include <linux/spinlock.h>
+#include <linux/delay.h>
+#include <linux/if_ether.h>
+
+#include <linux/netdevice.h> /* necessary for etherdevice.h on some kernels */
+#include <linux/etherdevice.h>
+
+#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,21)
+static inline int is_local_ether_addr(const u8 *addr)
+{
+	return (0x02 & addr[0]);
+}
+#endif
+
+typedef unsigned long irq_flags_t;
+
+#define spin_lock_destroy(l_)  do {} while (0)
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,24)
+#define HAS_NET_NAMESPACE
+#endif
+
+/* Funny, but linux has round_up for x86 only, defined in
+ * x86-specific header */
+#ifndef round_up
+#define round_up(x, y) (((x) + (y) - 1) & ~((y)-1))
+#endif
+
+#endif /* __CI_EFHW_SYSDEP_LINUX_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/efrm/nic_table.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efrm/nic_table.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/efrm/nic_table.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efrm/nic_table.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,98 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides public API for NIC table.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_NIC_TABLE_H__
+#define __CI_EFRM_NIC_TABLE_H__
+
+#include <ci/efhw/efhw_types.h>
+#include <ci/efrm/sysdep.h>
+
+/*--------------------------------------------------------------------
+ *
+ * struct efrm_nic_table - top level driver object keeping all NICs -
+ * implemented in driver_object.c
+ *
+ *--------------------------------------------------------------------*/
+
+/*! Comment? */
+struct efrm_nic_table {
+	/*! nics attached to this driver */
+	struct efhw_nic *nic[EFHW_MAX_NR_DEVS];
+	/*! pointer to an arbitrary struct efhw_nic if one exists;
+	 * for code which does not care which NIC it wants but
+	 * still needs one. Note you cannot assume nic[0] exists. */
+	struct efhw_nic *a_nic;
+	uint32_t nic_count;	/*!< number of nics attached to this driver */
+	spinlock_t lock;	/*!< lock for table modifications */
+	atomic_t ref_count;	/*!< refcount for users of nic table */
+};
+
+/* Resource driver structures used by other drivers as well */
+extern struct efrm_nic_table efrm_nic_table;
+
+static inline void efrm_nic_table_hold(void)
+{
+	atomic_inc(&efrm_nic_table.ref_count);
+}
+
+static inline void efrm_nic_table_rele(void)
+{
+	atomic_dec(&efrm_nic_table.ref_count);
+}
+
+static inline int efrm_nic_table_held(void)
+{
+	return (atomic_read(&efrm_nic_table.ref_count) != 0);
+}
+
+/* Run code block _x multiple times with variable nic set to each
+ * registered NIC in turn.
+ * DO NOT "break" out of this loop early. */
+#define EFRM_FOR_EACH_NIC(_nic_i, _nic)					\
+	for ((_nic_i) = (efrm_nic_table_hold(), 0);			\
+	     (_nic_i) < EFHW_MAX_NR_DEVS || (efrm_nic_table_rele(), 0);	\
+	     (_nic_i)++)						\
+		if (((_nic) = efrm_nic_table.nic[_nic_i]))
+
+#define EFRM_FOR_EACH_NIC_IN_SET(_set, _i, _nic)			\
+	for ((_i) = (efrm_nic_table_hold(), 0);				\
+	     (_i) < EFHW_MAX_NR_DEVS || (efrm_nic_table_rele(), 0);	\
+	     ++(_i))							\
+		if (((_nic) = efrm_nic_table.nic[_i]) &&		\
+		    efrm_nic_set_read((_set), (_i)))
+
+#endif /* __CI_EFRM_NIC_TABLE_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/efrm/sysdep.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efrm/sysdep.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/efrm/sysdep.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efrm/sysdep.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,54 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides Linux-like system-independent API for efrm library.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_SYSDEP_H__
+#define __CI_EFRM_SYSDEP_H__
+
+/* Spinlocks are defined in efhw/sysdep.h */
+#include <ci/efhw/sysdep.h>
+
+#if defined(__linux__) && defined(__KERNEL__)
+
+# include <ci/efrm/sysdep_linux.h>
+
+#else
+
+# include <ci/efrm/sysdep_ci2linux.h>
+
+#endif
+
+#endif /* __CI_EFRM_SYSDEP_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/efrm/sysdep_linux.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efrm/sysdep_linux.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/efrm/sysdep_linux.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/efrm/sysdep_linux.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,248 @@
+/****************************************************************************
+ * Driver for Solarflare network controllers -
+ *          resource management for Xen backend, OpenOnload, etc
+ *           (including support for SFE4001 10GBT NIC)
+ *
+ * This file provides version-independent Linux kernel API for efrm library.
+ * Only kernels >=2.6.9 are supported.
+ *
+ * Copyright 2005-2007: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Kfifo API is partially stolen from linux-2.6.22/include/linux/list.h
+ * Copyright (C) 2004 Stelian Pop <stelian@popies.net>
+ *
+ * Developed and maintained by Solarflare Communications:
+ *                      <linux-xen-drivers@solarflare.com>
+ *                      <onload-dev@solarflare.com>
+ *
+ * Certain parts of the driver were implemented by
+ *          Alexandra Kossovsky <Alexandra.Kossovsky@oktetlabs.ru>
+ *          OKTET Labs Ltd, Russia,
+ *          http://oktetlabs.ru, <info@oktetlabs.ru>
+ *          by request of Solarflare Communications
+ *
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef __CI_EFRM_SYSDEP_LINUX_H__
+#define __CI_EFRM_SYSDEP_LINUX_H__
+
+#include <linux/version.h>
+#include <linux/list.h>
+#include <linux/vmalloc.h>
+#include <linux/errno.h>
+#include <linux/string.h>
+#include <linux/workqueue.h>
+#include <linux/gfp.h>
+#include <linux/slab.h>
+#include <linux/hardirq.h>
+#include <linux/kernel.h>
+#include <linux/if_ether.h>
+#include <linux/completion.h>
+#include <linux/in.h>
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20)
+/* get roundup_pow_of_two(), which was in kernel.h in early kernel versions */
+#include <linux/log2.h>
+#endif
+
+/********************************************************************
+ *
+ * List API
+ *
+ ********************************************************************/
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,18)
+static inline void
+list_replace_init(struct list_head *old, struct list_head *new)
+{
+	new->next = old->next;
+	new->next->prev = new;
+	new->prev = old->prev;
+	new->prev->next = new;
+	INIT_LIST_HEAD(old);
+}
+#endif
+
+static inline struct list_head *list_pop(struct list_head *list)
+{
+	struct list_head *link = list->next;
+	list_del(link);
+	return link;
+}
+
+static inline struct list_head *list_pop_tail(struct list_head *list)
+{
+	struct list_head *link = list->prev;
+	list_del(link);
+	return link;
+}
+
+/********************************************************************
+ *
+ * Workqueue API
+ *
+ ********************************************************************/
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,20)
+#define NEED_OLD_WORK_API
+
+/**
+ * The old and new work function prototypes just change
+ * the type of the pointer in the only argument, so it's
+ * safe to cast one function type to the other
+ */
+typedef void (*efrm_old_work_func_t) (void *p);
+
+#undef INIT_WORK
+#define INIT_WORK(_work, _func)					\
+	do {							\
+		INIT_LIST_HEAD(&(_work)->entry);		\
+		(_work)->pending = 0;				\
+		PREPARE_WORK((_work),				\
+			     (efrm_old_work_func_t) (_func),	\
+			     (_work));				\
+	} while (0)
+
+#endif
+
+/********************************************************************
+ *
+ * Kfifo API
+ *
+ ********************************************************************/
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,10)
+
+#if !defined(RHEL_RELEASE_CODE) || (RHEL_RELEASE_CODE < 1029)
+typedef unsigned gfp_t;
+#endif
+
+#define HAS_NO_KFIFO
+
+struct kfifo {
+	unsigned char *buffer;	/* the buffer holding the data */
+	unsigned int size;	/* the size of the allocated buffer */
+	unsigned int in;	/* data is added at offset (in % size) */
+	unsigned int out;	/* data is extracted from off. (out % size) */
+	spinlock_t *lock;	/* protects concurrent modifications */
+};
+
+extern struct kfifo *kfifo_init(unsigned char *buffer, unsigned int size,
+				gfp_t gfp_mask, spinlock_t *lock);
+extern struct kfifo *kfifo_alloc(unsigned int size, gfp_t gfp_mask,
+				 spinlock_t *lock);
+extern void kfifo_free(struct kfifo *fifo);
+extern unsigned int __kfifo_put(struct kfifo *fifo,
+				unsigned char *buffer, unsigned int len);
+extern unsigned int __kfifo_get(struct kfifo *fifo,
+				unsigned char *buffer, unsigned int len);
+
+/**
+ * kfifo_put - puts some data into the FIFO
+ * @fifo: the fifo to be used.
+ * @buffer: the data to be added.
+ * @len: the length of the data to be added.
+ *
+ * This function copies at most @len bytes from the @buffer into
+ * the FIFO depending on the free space, and returns the number of
+ * bytes copied.
+ */
+static inline unsigned int
+kfifo_put(struct kfifo *fifo, unsigned char *buffer, unsigned int len)
+{
+	unsigned long flags;
+	unsigned int ret;
+
+	spin_lock_irqsave(fifo->lock, flags);
+
+	ret = __kfifo_put(fifo, buffer, len);
+
+	spin_unlock_irqrestore(fifo->lock, flags);
+
+	return ret;
+}
+
+/**
+ * kfifo_get - gets some data from the FIFO
+ * @fifo: the fifo to be used.
+ * @buffer: where the data must be copied.
+ * @len: the size of the destination buffer.
+ *
+ * This function copies at most @len bytes from the FIFO into the
+ * @buffer and returns the number of copied bytes.
+ */
+static inline unsigned int
+kfifo_get(struct kfifo *fifo, unsigned char *buffer, unsigned int len)
+{
+	unsigned long flags;
+	unsigned int ret;
+
+	spin_lock_irqsave(fifo->lock, flags);
+
+	ret = __kfifo_get(fifo, buffer, len);
+
+	/*
+	 * optimization: if the FIFO is empty, set the indices to 0
+	 * so we don't wrap the next time
+	 */
+	if (fifo->in == fifo->out)
+		fifo->in = fifo->out = 0;
+
+	spin_unlock_irqrestore(fifo->lock, flags);
+
+	return ret;
+}
+
+/**
+ * __kfifo_len - returns the number of bytes available in the FIFO, no locking version
+ * @fifo: the fifo to be used.
+ */
+static inline unsigned int __kfifo_len(struct kfifo *fifo)
+{
+	return fifo->in - fifo->out;
+}
+
+/**
+ * kfifo_len - returns the number of bytes available in the FIFO
+ * @fifo: the fifo to be used.
+ */
+static inline unsigned int kfifo_len(struct kfifo *fifo)
+{
+	unsigned long flags;
+	unsigned int ret;
+
+	spin_lock_irqsave(fifo->lock, flags);
+
+	ret = __kfifo_len(fifo);
+
+	spin_unlock_irqrestore(fifo->lock, flags);
+
+	return ret;
+}
+
+#else
+#include <linux/kfifo.h>
+#endif
+
+static inline void kfifo_vfree(struct kfifo *fifo)
+{
+	vfree(fifo->buffer);
+	kfree(fifo);
+}
+
+#endif /* __CI_EFRM_SYSDEP_LINUX_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/tools/config.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/tools/config.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/tools/config.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/tools/config.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,49 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*! \cidoxg_include_ci_tools */
+
+#ifndef __CI_TOOLS_CONFIG_H__
+#define __CI_TOOLS_CONFIG_H__
+
+
+/**********************************************************************
+ * Debugging.
+ */
+
+#define CI_INCLUDE_ASSERT_VALID           0
+
+/* Set non-zero to allow info about who has allocated what to appear in
+ * /proc/drivers/level5/mem.
+ * However - Note that doing so can lead to segfault when you unload the
+ * driver, and other weirdness.  i.e. I don't think the code for is quite
+ * right (written by Oktet, hacked by gel), but it does work well enough to be
+ * useful.
+ */
+#define CI_MEMLEAK_DEBUG_ALLOC_TABLE	  0
+
+
+#endif  /* __CI_TOOLS_CONFIG_H__ */
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/tools/debug.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/tools/debug.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/tools/debug.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/tools/debug.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,336 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*! \cidoxg_include_ci_tools */
+
+#ifndef __CI_TOOLS_DEBUG_H__
+#define __CI_TOOLS_DEBUG_H__
+
+#define CI_LOG_E(x)       x              /* errors      */
+#define CI_LOG_W(x)       x              /* warnings    */
+#define CI_LOG_I(x)       x              /* information */
+#define CI_LOG_V(x)       x              /* verbose     */
+
+/* Build time asserts. We paste the line number into the type name
+ * so that the macro can be used more than once per file even if the
+ * compiler objects to multiple identical typedefs. Collisions
+ * between use in different header files is still possible. */
+#ifndef CI_BUILD_ASSERT
+#define __CI_BUILD_ASSERT_NAME(_x) __CI_BUILD_ASSERT_ILOATHECPP(_x)
+#define __CI_BUILD_ASSERT_ILOATHECPP(_x)  __CI_BUILD_ASSERT__ ##_x
+#define CI_BUILD_ASSERT(e)\
+ typedef char  __CI_BUILD_ASSERT_NAME(__LINE__)[(e)?1:-1]
+#endif
+
+
+#ifdef NDEBUG
+
+# define _ci_check(exp, file, line)
+# define _ci_assert2(e, x, y, file, line)
+# define _ci_assert(exp, file, line)
+# define _ci_assert_equal(exp1, exp2, file, line)
+# define _ci_assert_equiv(exp1, exp2, file, line)
+# define _ci_assert_nequal(exp1, exp2, file, line)
+# define _ci_assert_le(exp1, exp2, file, line)
+# define _ci_assert_lt(exp1, exp2, file, line)
+# define _ci_assert_ge(exp1, exp2, file, line)
+# define _ci_assert_gt(exp1, exp2, file, line)
+# define _ci_assert_impl(exp1, exp2, file, line)
+
+# define _ci_verify(exp, file, line) \
+  do { \
+    (void)(exp); \
+  } while (0)
+
+# define CI_DEBUG_TRY(exp) \
+  do { \
+    (void)(exp); \
+  } while (0)
+
+#define CI_TRACE(exp,fmt)
+#define CI_TRACE_INT(integer)
+#define CI_TRACE_INT32(integer)
+#define CI_TRACE_INT64(integer)
+#define CI_TRACE_UINT(integer)
+#define CI_TRACE_UINT32(integer)
+#define CI_TRACE_UINT64(integer)
+#define CI_TRACE_HEX(integer)
+#define CI_TRACE_HEX32(integer)
+#define CI_TRACE_HEX64(integer)
+#define CI_TRACE_PTR(pointer)
+#define CI_TRACE_STRING(string)
+#define CI_TRACE_MAC(mac)
+#define CI_TRACE_IP(ip_be32)
+#define CI_TRACE_ARP(arp_pkt)
+
+#else
+
+# define _CI_ASSERT_FMT   "\nfrom %s:%d"
+
+# define _ci_check(exp, file, line)                             \
+  do {                                                          \
+    if (CI_UNLIKELY(!(exp)))                                    \
+      ci_warn(("ci_check(%s)"_CI_ASSERT_FMT, #exp,              \
+               (file), (line)));                                \
+  } while (0)
+
+/*
+ * NOTE: ci_fail() emits the file and line where the assert is actually
+ *       coded.
+ */
+
+# define _ci_assert(exp, file, line)                            \
+  do {                                                          \
+    if (CI_UNLIKELY(!(exp)))                                    \
+      ci_fail(("ci_assert(%s)"_CI_ASSERT_FMT, #exp,		\
+               (file), (line)));                                \
+  } while (0)
+
+# define _ci_assert2(e, x, y, file, line)  do {                 \
+    if(CI_UNLIKELY( ! (e) ))                                    \
+      ci_fail(("ci_assert(%s)\nwhere [%s=%"CI_PRIx64"] "        \
+               "[%s=%"CI_PRIx64"]\nat %s:%d\nfrom %s:%d", #e    \
+               , #x, (ci_uint64)(ci_uintptr_t)(x)               \
+               , #y, (ci_uint64)(ci_uintptr_t)(y),              \
+               __FILE__, __LINE__, (file), (line)));            \
+  } while (0)
+
+# define _ci_verify(exp, file, line)                            \
+  do {                                                          \
+    if (CI_UNLIKELY(!(exp)))                                    \
+      ci_fail(("ci_verify(%s)"_CI_ASSERT_FMT, #exp,             \
+               (file), (line)));                                \
+  } while (0)
+
+# define _ci_assert_equal(x, y, f, l)  _ci_assert2((x)==(y), x, y, (f), (l))
+# define _ci_assert_nequal(x, y, f, l) _ci_assert2((x)!=(y), x, y, (f), (l))
+# define _ci_assert_le(x, y, f, l)     _ci_assert2((x)<=(y), x, y, (f), (l))
+# define _ci_assert_lt(x, y, f, l)     _ci_assert2((x)< (y), x, y, (f), (l))
+# define _ci_assert_ge(x, y, f, l)     _ci_assert2((x)>=(y), x, y, (f), (l))
+# define _ci_assert_gt(x, y, f, l)     _ci_assert2((x)> (y), x, y, (f), (l))
+# define _ci_assert_or(x, y, f, l)     _ci_assert2((x)||(y), x, y, (f), (l))
+# define _ci_assert_impl(x, y, f, l)   _ci_assert2(!(x) || (y), x, y, (f), (l))
+# define _ci_assert_equiv(x, y, f, l)  _ci_assert2(!(x)== !(y), x, y, (f), (l))
+
+#define _ci_assert_equal_msg(exp1, exp2, msg, file, line)       \
+  do {                                                          \
+    if (CI_UNLIKELY((exp1)!=(exp2)))                            \
+      ci_fail(("ci_assert_equal_msg(%s == %s) were "            \
+               "(%"CI_PRIx64":%"CI_PRIx64") with msg[%c%c%c%c]" \
+               _CI_ASSERT_FMT, #exp1, #exp2,                    \
+               (ci_uint64)(ci_uintptr_t)(exp1),                 \
+               (ci_uint64)(ci_uintptr_t)(exp2),                 \
+               (((ci_uint32)msg) >> 24) && 0xff,                \
+               (((ci_uint32)msg) >> 16) && 0xff,                \
+               (((ci_uint32)msg) >> 8 ) && 0xff,                \
+               (((ci_uint32)msg)      ) && 0xff,                \
+               (file), (line)));                                \
+  } while (0)
+
+# define CI_DEBUG_TRY(exp)  CI_TRY(exp)
+
+#define CI_TRACE(exp,fmt)						\
+  ci_log("%s:%d:%s] " #exp "=" fmt,                                     \
+         __FILE__, __LINE__, __FUNCTION__, (exp))
+
+
+#define CI_TRACE_INT(integer)						\
+  ci_log("%s:%d:%s] " #integer "=%d",                                   \
+         __FILE__, __LINE__, __FUNCTION__, (integer))
+
+
+#define CI_TRACE_INT32(integer)						\
+  ci_log("%s:%d:%s] " #integer "=%d",                                   \
+         __FILE__, __LINE__, __FUNCTION__, ((ci_int32)integer))
+
+
+#define CI_TRACE_INT64(integer)						\
+  ci_log("%s:%d:%s] " #integer "=%lld",                                 \
+         __FILE__, __LINE__, __FUNCTION__, ((ci_int64)integer))
+
+
+#define CI_TRACE_UINT(integer)						\
+  ci_log("%s:%d:%s] " #integer "=%ud",                                  \
+         __FILE__, __LINE__, __FUNCTION__, (integer))
+
+
+#define CI_TRACE_UINT32(integer)			  	        \
+  ci_log("%s:%d:%s] " #integer "=%ud",                                  \
+         __FILE__, __LINE__, __FUNCTION__, ((ci_uint32)integer))
+
+
+#define CI_TRACE_UINT64(integer)			  	        \
+  ci_log("%s:%d:%s] " #integer "=%ulld",                                \
+         __FILE__, __LINE__, __FUNCTION__, ((ci_uint64)integer))
+
+
+#define CI_TRACE_HEX(integer)						\
+  ci_log("%s:%d:%s] " #integer "=0x%x",                                 \
+         __FILE__, __LINE__, __FUNCTION__, (integer))
+
+
+#define CI_TRACE_HEX32(integer)						\
+  ci_log("%s:%d:%s] " #integer "=0x%x",                                 \
+         __FILE__, __LINE__, __FUNCTION__, ((ci_uint32)integer))
+
+
+#define CI_TRACE_HEX64(integer)						\
+  ci_log("%s:%d:%s] " #integer "=0x%llx",                               \
+         __FILE__, __LINE__, __FUNCTION__, ((ci_uint64)integer))
+
+
+#define CI_TRACE_PTR(pointer)				                \
+  ci_log("%s:%d:%s] " #pointer "=0x%p",                                 \
+         __FILE__, __LINE__, __FUNCTION__, (pointer))
+
+
+#define CI_TRACE_STRING(string)					        \
+  ci_log("%s:%d:%s] " #string "=%s",                                    \
+         __FILE__, __LINE__, __FUNCTION__, (string))
+
+
+#define CI_TRACE_MAC(mac)						\
+  ci_log("%s:%d:%s] " #mac "=" CI_MAC_PRINTF_FORMAT,                    \
+         __FILE__, __LINE__, __FUNCTION__, CI_MAC_PRINTF_ARGS(mac))
+
+
+#define CI_TRACE_IP(ip_be32)						\
+  ci_log("%s:%d:%s] " #ip_be32 "=" CI_IP_PRINTF_FORMAT, __FILE__,       \
+         __LINE__, __FUNCTION__, CI_IP_PRINTF_ARGS(&(ip_be32)))
+
+
+#define CI_TRACE_ARP(arp_pkt)                                           \
+  ci_log("%s:%d:%s]\n"CI_ARP_PRINTF_FORMAT,                             \
+         __FILE__, __LINE__, __FUNCTION__, CI_ARP_PRINTF_ARGS(arp_pkt))
+
+#endif  /* NDEBUG */
+
+#define ci_check(exp) \
+        _ci_check(exp, __FILE__, __LINE__)
+
+#define ci_assert(exp) \
+        _ci_assert(exp, __FILE__, __LINE__)
+
+#define ci_verify(exp) \
+        _ci_verify(exp, __FILE__, __LINE__)
+
+#define ci_assert_equal(exp1, exp2) \
+        _ci_assert_equal(exp1, exp2, __FILE__, __LINE__)
+
+#define ci_assert_equal_msg(exp1, exp2, msg) \
+        _ci_assert_equal_msg(exp1, exp2, msg, __FILE__, __LINE__)
+
+#define ci_assert_nequal(exp1, exp2) \
+        _ci_assert_nequal(exp1, exp2, __FILE__, __LINE__)
+
+#define ci_assert_le(exp1, exp2) \
+        _ci_assert_le(exp1, exp2, __FILE__, __LINE__)
+
+#define ci_assert_lt(exp1, exp2) \
+        _ci_assert_lt(exp1, exp2, __FILE__, __LINE__)
+
+#define ci_assert_ge(exp1, exp2) \
+        _ci_assert_ge(exp1, exp2, __FILE__, __LINE__)
+
+#define ci_assert_gt(exp1, exp2) \
+        _ci_assert_gt(exp1, exp2, __FILE__, __LINE__)
+
+#define ci_assert_impl(exp1, exp2) \
+        _ci_assert_impl(exp1, exp2, __FILE__, __LINE__)
+
+#define ci_assert_equiv(exp1, exp2) \
+        _ci_assert_equiv(exp1, exp2, __FILE__, __LINE__)
+
+
+#define CI_TEST(exp)                            \
+  do{                                           \
+    if( CI_UNLIKELY(!(exp)) )                   \
+      ci_fail(("CI_TEST(%s)", #exp));           \
+  }while(0)
+
+
+#define CI_TRY(exp)				\
+  do{						\
+    int _trc;					\
+    _trc=(exp);					\
+    if( CI_UNLIKELY(_trc < 0) )			\
+      ci_sys_fail(#exp, _trc);			\
+  }while(0)
+
+
+#define CI_TRY_RET(exp)							 \
+  do{									 \
+    int _trc;								 \
+    _trc=(exp);								 \
+    if( CI_UNLIKELY(_trc < 0) ) {					 \
+      ci_log("%s returned %d at %s:%d", #exp, _trc, __FILE__, __LINE__); \
+      return _trc;							 \
+    }									 \
+  }while(0)
+
+#define CI_LOGLEVEL_TRY_RET(logfn, exp)                                    \
+  do{									 \
+    int _trc;								 \
+    _trc=(exp);								 \
+    if( CI_UNLIKELY(_trc < 0) ) {					 \
+      logfn (ci_log("%s returned %d at %s:%d", #exp, _trc, __FILE__, __LINE__)); \
+      return _trc;							 \
+    }									 \
+  }while(0)
+
+
+#define CI_SOCK_TRY(exp)			\
+  do{						\
+    ci_sock_err_t _trc;				\
+    _trc=(exp);					\
+    if( CI_UNLIKELY(!ci_sock_errok(_trc)) )	\
+      ci_sys_fail(#exp, _trc.val);		\
+  }while(0)
+
+
+#define CI_SOCK_TRY_RET(exp)						     \
+  do{									     \
+    ci_sock_err_t _trc;							     \
+    _trc=(exp);								     \
+    if( CI_UNLIKELY(!ci_sock_errok(_trc)) ) {		  		     \
+      ci_log("%s returned %d at %s:%d", #exp, _trc.val, __FILE__, __LINE__); \
+      return ci_sock_errcode(_trc);					     \
+    }									     \
+  }while(0)
+
+
+#define CI_SOCK_TRY_SOCK_RET(exp)					     \
+  do{									     \
+    ci_sock_err_t _trc;							     \
+    _trc=(exp);								     \
+    if( CI_UNLIKELY(!ci_sock_errok(_trc)) ) {		  		     \
+      ci_log("%s returned %d at %s:%d", #exp, _trc.val, __FILE__, __LINE__); \
+      return _trc;							     \
+    }									     \
+  }while(0)
+
+#endif  /* __CI_TOOLS_DEBUG_H__ */
+
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/tools/log.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/tools/log.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/tools/log.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/tools/log.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,262 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*
+ * \author  djr
+ *  \brief  Functions for logging and pretty-printing.
+ *   \date  2002/08/07
+ */
+
+/*! \cidoxg_include_ci_tools */
+
+#ifndef __CI_TOOLS_LOG_H__
+#define __CI_TOOLS_LOG_H__
+
+#include <stdarg.h>
+
+
+/**********************************************************************
+ * Logging.
+ */
+
+/* size of internal log buffer */ 
+#define  CI_LOG_MAX_LINE        512
+/* uses of ci_log must ensure that all trace messages are shorter than this */ 
+#define  CI_LOG_MAX_MSG_LENGTH        (CI_LOG_MAX_LINE-50)
+
+extern void ci_vlog(const char* fmt, va_list args)  CI_HF;
+extern void ci_log(const char* fmt, ...) CI_PRINTF_LIKE(1,2) CI_HF;
+
+  /*! Set the prefix for log messages.
+  **
+  ** Uses the storage pointed to by \em prefix.  Therefore \em prefix must
+  ** be allocated on the heap, or statically.
+  */
+extern void ci_set_log_prefix(const char* prefix)  CI_HF;
+
+typedef void (*ci_log_fn_t)(const char* msg);
+extern ci_log_fn_t  ci_log_fn  CI_HV;
+
+/* Log functions. */
+extern void ci_log_null(const char* msg) CI_HF;
+extern void ci_log_stderr(const char* msg) CI_HF;
+extern void ci_log_stdout(const char* msg) CI_HF;
+extern void ci_log_syslog(const char* msg) CI_HF;
+
+/*! Call the following to install special logging behaviours. */
+extern void ci_log_buffer_till_fail(void) CI_HF;
+extern void ci_log_buffer_till_exit(void) CI_HF;
+
+extern void __ci_log_unique(const char* msg) CI_HF;
+extern ci_log_fn_t __ci_log_unique_fn CI_HV;
+ci_inline void ci_log_uniquify(void) {
+  if( ci_log_fn != __ci_log_unique ) {
+    __ci_log_unique_fn = ci_log_fn;
+    ci_log_fn = __ci_log_unique;
+  }
+}
+
+extern void ci_log_file(const char* msg) CI_HF;
+extern int  ci_log_file_fd CI_HV;
+
+extern void __ci_log_nth(const char* msg) CI_HF;
+extern ci_log_fn_t __ci_log_nth_fn CI_HV;
+extern int  ci_log_nth_n CI_HV;  /* default 100 */
+ci_inline void ci_log_nth(void) {
+  if( ci_log_fn != __ci_log_nth ) {
+    __ci_log_nth_fn = ci_log_fn;
+    ci_log_fn = __ci_log_nth;
+  }
+}
+
+extern int  ci_log_level  CI_HV;
+
+extern int  ci_log_options  CI_HV;
+#define CI_LOG_PID		0x1
+#define CI_LOG_TID		0x2
+#define CI_LOG_TIME		0x4
+#define CI_LOG_DELTA		0x8
+
+/**********************************************************************
+ * Used to define which mode we are in
+ */
+#if (defined(_WIN32) && !defined(__KERNEL__))
+typedef enum {
+  ci_log_md_NULL=0,
+    ci_log_md_ioctl,
+    ci_log_md_stderr,
+    ci_log_md_stdout,
+    ci_log_md_file,
+    ci_log_md_serial,
+    ci_log_md_syslog,
+    ci_log_md_pidfile
+} ci_log_mode_t;
+extern ci_log_mode_t ci_log_mode;
+#endif
+
+/**********************************************************************
+ * Pretty-printing.
+ */
+
+extern char ci_printable_char(char c) CI_HF;
+
+extern void (*ci_hex_dump_formatter)(char* buf, const ci_octet* s,
+				     int i, int off, int len) CI_HV;
+extern void ci_hex_dump_format_octets(char*,const ci_octet*,int,int,int) CI_HF;
+extern void ci_hex_dump_format_dwords(char*,const ci_octet*,int,int,int) CI_HF;
+
+extern void ci_hex_dump_row(char* buf, volatile const void* s, int len,
+			    ci_ptr_arith_t address) CI_HF;
+  /*!< A row contains up to 16 bytes.  Row starts at [address & 15u], so
+  ** therefore [len + (address & 15u)] must be <= 16.
+  */
+
+extern void ci_hex_dump(ci_log_fn_t, volatile const void*,
+			int len, ci_ptr_arith_t address) CI_HF;
+
+extern int  ci_hex_dump_to_raw(const char* src_hex, void* buf,
+			       unsigned* addr_out_opt, int* skip)  CI_HF;
+  /*!< Recovers raw data from a single line of a hex dump.  [buf] must be at
+  ** least 16 bytes long.  Returns the number of bytes written to [buf] (in
+  ** range 1 -> 16), or -1 if [src_hex] doesn't contain hex data.  Does not
+  ** cope with missing bytes at the start of a line.
+  */
+
+extern int ci_format_eth_addr(char* buf, const void* eth_mac_addr,
+			      char sep)  CI_HF;
+  /*!< This will write 18 characters to <buf> including terminating null.
+  ** Returns number of bytes written excluding null.  If [sep] is zero, ':'
+  ** is used.
+  */
+
+extern int ci_parse_eth_addr(void* eth_mac_addr,
+			     const char* str, char sep) CI_HF;
+  /*!< If [sep] is zero, absolutely any separator is accepted (even
+  ** inconsistent separators).  Returns 0 on success, -1 on error.
+  */
+
+extern int ci_format_ip4_addr(char* buf, unsigned addr_be32) CI_HF;
+  /*!< Formats the IP address (in network endian) in dotted-quad.  Returns
+  ** the number of bytes written (up to 15), excluding the null.  [buf]
+  ** must be at least 16 bytes long.
+  */
+
+
+/**********************************************************************
+ * Error checking.
+ */
+
+extern void (*ci_fail_stop_fn)(void) CI_HV;
+
+extern void ci_fail_stop(void) CI_HF;
+extern void ci_fail_hang(void) CI_HF;
+extern void ci_fail_bomb(void) CI_HF;
+extern void ci_backtrace(void) CI_HF;
+
+#if defined __linux__ && !defined __KERNEL__
+extern void ci_fail_abort (void) CI_HF;
+#endif
+
+#ifdef __GNUC__
+extern void
+__ci_fail(const char*, ...) CI_PRINTF_LIKE(1,2) CI_HF;
+#else
+# if _PREFAST_
+  extern void _declspec(noreturn) __ci_fail(const char* fmt, ...);
+# else 
+  extern void __ci_fail(const char* fmt, ...);
+# endif
+
+#endif
+
+#define ci_warn(x)							   \
+  do{ ci_log("WARN at %s:%d", __FILE__, __LINE__); }while(0)
+
+#define ci_fail(x)							   \
+  do{ ci_log("FAIL at %s:%d", __FILE__, __LINE__);  __ci_fail x; }while(0)
+
+extern void __ci_sys_fail(const char* fn, int rc,
+			  const char* file, int line) CI_HF;
+#define ci_sys_fail(fn, rc)  __ci_sys_fail(fn, rc, __FILE__, __LINE__)
+
+/**********************************************************************
+ * Logging to buffer (src/citools/log_buffer.c)
+ */
+
+/*! Divert ci_log() messages to the log buffer
+ *  normally they go to the  system console */
+extern void ci_log_buffer_till_fail(void) CI_HF;
+
+/*! Dump the contents of the log buffer to the system console */
+extern void ci_log_buffer_dump(void) CI_HF;
+
+
+/**********************************************************************
+ * Some useful pretty-printing.
+ */
+
+#ifdef  __linux__
+# define CI_SOCKCALL_FLAGS_FMT	"%s%s%s%s%s%s%s%s%s%s%s"
+
+# define CI_SOCKCALL_FLAGS_PRI_ARG(x)		\
+  (((x) & MSG_OOB         ) ? "OOB "         :""),	\
+  (((x) & MSG_PEEK        ) ? "PEEK "        :""),	\
+  (((x) & MSG_DONTROUTE   ) ? "DONTROUTE "   :""),	\
+  (((x) & MSG_EOR         ) ? "EOR "         :""),	\
+  (((x) & MSG_CTRUNC      ) ? "CTRUNC "      :""),	\
+  (((x) & MSG_TRUNC       ) ? "TRUNC "       :""),	\
+  (((x) & MSG_WAITALL     ) ? "WAITALL "     :""),	\
+  (((x) & MSG_DONTWAIT    ) ? "DONTWAIT "    :""),	\
+  (((x) & MSG_NOSIGNAL    ) ? "NOSIGNAL "    :""),	\
+  (((x) & MSG_ERRQUEUE    ) ? "ERRQUEUE "    :""),	\
+  (((x) & MSG_CONFIRM     ) ? "CONFIRM "     :"")
+#endif
+
+#ifdef  _WIN32
+# define CI_SOCKCALL_FLAGS_FMT	"%s%s%s"
+
+# define CI_SOCKCALL_FLAGS_PRI_ARG(x)		\
+  (((x) & MSG_OOB         ) ? "OOB "         :""),	\
+  (((x) & MSG_PEEK        ) ? "PEEK "        :""),	\
+  (((x) & MSG_DONTROUTE   ) ? "DONTROUTE "   :"")
+#endif
+
+#ifdef  __sun__
+# define CI_SOCKCALL_FLAGS_FMT	"%s%s%s%s%s%s%s%s%s"
+
+# define CI_SOCKCALL_FLAGS_PRI_ARG(x)		\
+  (((x) & MSG_OOB         ) ? "OOB "         :""),	\
+  (((x) & MSG_PEEK        ) ? "PEEK "        :""),	\
+  (((x) & MSG_DONTROUTE   ) ? "DONTROUTE "   :""),	\
+  (((x) & MSG_EOR         ) ? "EOR "         :""),	\
+  (((x) & MSG_CTRUNC      ) ? "CTRUNC "      :""),	\
+  (((x) & MSG_TRUNC       ) ? "TRUNC "       :""),	\
+  (((x) & MSG_WAITALL     ) ? "WAITALL "     :""),	\
+  (((x) & MSG_DONTWAIT    ) ? "DONTWAIT "    :""),	\
+  (((x) & MSG_NOTIFICATION) ? "NOTIFICATION" :"")
+#endif
+
+#endif  /* __CI_TOOLS_LOG_H__ */
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/tools/platform/gcc_x86.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/tools/platform/gcc_x86.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/tools/platform/gcc_x86.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/tools/platform/gcc_x86.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,361 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*! \cidoxg_include_ci_tools_platform  */
+
+#ifndef __CI_TOOLS_GCC_X86_H__
+#define __CI_TOOLS_GCC_X86_H__
+
+
+/**********************************************************************
+ * Free-running cycle counters.
+ */
+
+#define CI_HAVE_FRC64
+#define CI_HAVE_FRC32
+
+#define ci_frc32(pval)  __asm__ __volatile__("rdtsc" : "=a" (*pval) : : "edx")
+
+#if defined(__x86_64__)
+ci_inline void ci_frc64(ci_uint64* pval) {
+  /* temp fix until we figure how to get this out in one bite */	   
+  ci_uint64 low, high;
+  __asm__ __volatile__("rdtsc" : "=a" (low) , "=d" (high));	 	
+  *pval = (high << 32) | low;
+}
+
+#else
+#define ci_frc64(pval)  __asm__ __volatile__("rdtsc" : "=A" (*pval))
+#endif
+
+#define ci_frc_flush()  /* ?? Need a pipeline barrier. */
+
+
+/**********************************************************************
+ * Atomic integer.
+ */
+
+/*
+** int  ci_atomic_read(a)         { return a->n;        }
+** void ci_atomic_set(a, v)       { a->n = v;           }
+** void ci_atomic_inc(a)          { ++a->n;             }
+** void ci_atomic_dec(a)          { --a->n;             }
+** int  ci_atomic_inc_and_test(a) { return ++a->n == 0; }
+** int  ci_atomic_dec_and_test(a) { return --a->n == 0; }
+** void ci_atomic_and(a, v)       { a->n &= v;          }
+** void ci_atomic_or(a, v)        { a->n |= v;          }
+*/
+
+typedef struct { volatile ci_int32 n; } ci_atomic_t;
+
+#define CI_ATOMIC_INITIALISER(i)  {(i)}
+
+static inline ci_int32  ci_atomic_read(const ci_atomic_t* a) { return a->n; }
+static inline void ci_atomic_set(ci_atomic_t* a, int v) { a->n = v; ci_wmb();   }
+
+static inline void ci_atomic_inc(ci_atomic_t* a)
+{ __asm__ __volatile__("lock; incl %0" : "+m" (a->n)); }
+
+ 
+static inline void ci_atomic_dec(ci_atomic_t* a)
+{ __asm__ __volatile__("lock; decl %0" : "+m" (a->n)); }
+
+static inline int ci_atomic_inc_and_test(ci_atomic_t* a) {
+  char r;
+  __asm__ __volatile__("lock; incl %0; sete %1"
+		       : "+m" (a->n), "=qm" (r));
+  return r;
+}
+
+static inline int ci_atomic_dec_and_test(ci_atomic_t* a) {
+  char r;
+  __asm__ __volatile__("lock; decl %0; sete %1"
+		       : "+m" (a->n), "=qm" (r));
+  return r;
+}
+
+ci_inline int
+ci_atomic_xadd (ci_atomic_t *a, int v) {
+   __asm__ ("lock xadd %0, %1" : "=r" (v), "+m" (a->n) : "0" (v));
+  return v;
+}
+ci_inline int
+ci_atomic_xchg (ci_atomic_t *a, int v) {
+   __asm__ ("lock xchg %0, %1" : "=r" (v), "+m" (a->n) : "0" (v));
+  return v;
+}
+
+ci_inline void ci_atomic32_or(volatile ci_uint32* p, ci_uint32 mask)
+{ __asm__ __volatile__("lock; orl %1, %0" : "+m" (*p) : "ir" (mask)); }
+
+ci_inline void ci_atomic32_and(volatile ci_uint32* p, ci_uint32 mask)
+{ __asm__ __volatile__("lock; andl %1, %0" : "+m" (*p) : "ir" (mask)); }
+
+ci_inline void ci_atomic32_add(volatile ci_uint32* p, ci_uint32 v)
+{ __asm__ __volatile__("lock; addl %1, %0" : "+m" (*p) : "ir" (v)); }
+
+#define ci_atomic_or(a, v)   ci_atomic32_or ((ci_uint32*) &(a)->n, (v))
+#define ci_atomic_and(a, v)  ci_atomic32_and((ci_uint32*) &(a)->n, (v))
+#define ci_atomic_add(a, v)  ci_atomic32_add((ci_uint32*) &(a)->n, (v))
+
+extern int ci_glibc_uses_nptl (void) CI_HF;
+extern int ci_glibc_nptl_broken(void) CI_HF;
+extern int ci_glibc_gs_get_is_multihreaded_offset (void) CI_HF;
+extern int ci_glibc_gs_is_multihreaded_offset CI_HV;
+
+#if !defined(__x86_64__)
+#ifdef __GLIBC__
+/* Returns non-zero if the calling process might be mulithreaded, returns 0 if
+ * it definitely isn't (i.e. if reimplementing this function for other
+ * architectures and platforms, you can safely just return 1).
+ */
+static inline int ci_is_multithreaded (void) {
+
+  while (1) {
+    if (ci_glibc_gs_is_multihreaded_offset >= 0) {
+      /* NPTL keeps a variable that tells us this hanging off gs (i.e. in thread-
+       * local storage); just return this
+       */
+      int r;
+      __asm__ __volatile__ ("movl %%gs:(%1), %0"
+                            : "=r" (r)
+                            : "r" (ci_glibc_gs_is_multihreaded_offset));
+      return r;
+    }
+
+    if (ci_glibc_gs_is_multihreaded_offset == -2) {
+      /* This means we've already determined that the libc version is NOT good
+       * for our funky "is multithreaded" hack
+       */
+      return 1;
+    }
+
+    /* If we get here, it means this is the first time the function has been
+     * called -- detect the libc version and go around again.
+     */
+    ci_glibc_gs_is_multihreaded_offset = ci_glibc_gs_get_is_multihreaded_offset ();
+
+    /* Go around again.  We do the test here rather than at the top so that we go
+     * quicker in the common the case
+     */
+  }
+}
+
+#else    /* def __GLIBC__ */
+
+#define ci_is_multithreaded() 1 /* ?? Is the the POSIX way of finding out */
+                                /*    whether the appication is single */
+                                /*    threaded? */
+
+#endif   /* def __GLIBC__ */
+
+#else    /* defined __x86_64__ */
+
+static inline int ci_is_multithreaded (void) {
+  /* Now easy way to tell on x86_64; so assume we're multithreaded */
+  return 1;
+}
+
+#endif    /* defined __x86_64__ */
+
+
+/**********************************************************************
+ * Compare and swap.
+ */
+
+#define CI_HAVE_COMPARE_AND_SWAP
+
+ci_inline int ci_cas32_succeed(volatile ci_int32* p, ci_int32 oldval,
+                               ci_int32 newval) {
+  char ret;
+  ci_int32 prevval;
+  __asm__ __volatile__("lock; cmpxchgl %3, %1; sete %0"
+		       : "=q"(ret), "+m"(*p), "=a"(prevval)
+		       : "r"(newval), "a"(oldval));
+  return ret;
+}
+
+ci_inline int ci_cas32_fail(volatile ci_int32* p, ci_int32 oldval,
+                            ci_int32 newval) {
+  char ret;
+  ci_int32 prevval;
+  __asm__ __volatile__("lock; cmpxchgl %3, %1; setne %0"
+		       : "=q"(ret), "+m"(*p), "=a"(prevval)
+		       : "r"(newval), "a"(oldval));
+  return ret;
+}
+
+#ifdef __x86_64__
+ci_inline int ci_cas64_succeed(volatile ci_int64* p, ci_int64 oldval,
+			       ci_int64 newval) {
+  char ret;
+  ci_int64 prevval;
+  __asm__ __volatile__("lock; cmpxchgq %3, %1; sete %0"
+		       : "=q"(ret), "+m"(*p), "=a"(prevval)
+		       : "r"(newval), "a"(oldval));
+  return ret;
+}
+
+ci_inline int ci_cas64_fail(volatile ci_int64* p, ci_int64 oldval,
+			    ci_int64 newval) {
+  char ret;
+  ci_int64 prevval;
+  __asm__ __volatile__("lock; cmpxchgq %3, %1; setne %0"
+		       : "=q"(ret), "+m"(*p), "=a"(prevval)
+		       : "r"(newval), "a"(oldval));
+  return ret;
+}
+#endif
+
+ci_inline int ci_cas32u_succeed(volatile ci_uint32* p, ci_uint32 oldval, ci_uint32 newval) {
+  char ret;
+  ci_uint32 prevval;
+  __asm__ __volatile__("lock; cmpxchgl %3, %1; sete %0"
+		       : "=q"(ret), "+m"(*p), "=a"(prevval)
+		       : "r"(newval), "a"(oldval));
+  return ret;
+}
+
+ci_inline int ci_cas32u_fail(volatile ci_uint32* p, ci_uint32 oldval, ci_uint32 newval) {
+  char ret;
+  ci_uint32 prevval;
+  __asm__ __volatile__("lock; cmpxchgl %3, %1; setne %0"
+		       : "=q"(ret), "+m"(*p), "=a"(prevval)
+		       : "r"(newval), "a"(oldval));
+  return ret;
+}
+
+ci_inline int ci_cas64u_succeed(volatile ci_uint64* p, ci_uint64 oldval,
+			       ci_uint64 newval) {
+  char ret;
+  ci_uint64 prevval;
+  __asm__ __volatile__("lock; cmpxchgq %3, %1; sete %0"
+		       : "=q"(ret), "+m"(*p), "=a"(prevval)
+		       : "r"(newval), "a"(oldval));
+  return ret;
+}
+
+ci_inline int ci_cas64u_fail(volatile ci_uint64* p, ci_uint64 oldval,
+			    ci_uint64 newval) {
+  char ret;
+  ci_uint64 prevval;
+  __asm__ __volatile__("lock; cmpxchgq %3, %1; setne %0"
+		       : "=q"(ret), "+m"(*p), "=a"(prevval)
+		       : "r"(newval), "a"(oldval));
+  return ret;
+}
+
+#ifdef __x86_64__
+
+# define ci_cas_uintptr_succeed(p,o,n)				\
+    ci_cas64u_succeed((volatile ci_uint64*) (p), (o), (n))
+# define ci_cas_uintptr_fail(p,o,n)				\
+    ci_cas64u_fail((volatile ci_uint64*) (p), (o), (n))
+
+#else
+
+# define ci_cas_uintptr_succeed(p,o,n)				\
+    ci_cas32u_succeed((volatile ci_uint32*) (p), (o), (n))
+# define ci_cas_uintptr_fail(p,o,n)				\
+    ci_cas32u_fail((volatile ci_uint32*) (p), (o), (n))
+
+#endif
+
+
+/**********************************************************************
+ * Atomic bit field.
+ */
+
+typedef ci_uint32  ci_bits;
+#define CI_BITS_N			32u
+
+#define CI_BITS_DECLARE(name, n)			\
+  ci_bits name[((n) + CI_BITS_N - 1u) / CI_BITS_N]
+
+ci_inline void ci_bits_clear_all(volatile ci_bits* b, int n_bits)
+{ memset((void*) b, 0, (n_bits+CI_BITS_N-1u) / CI_BITS_N * sizeof(ci_bits)); }
+
+ci_inline void ci_bit_set(volatile ci_bits* b, int i) {
+  __asm__ __volatile__("lock; btsl %1, %0"
+		       : "=m" (*b)
+		       : "Ir" (i));
+}
+
+ci_inline void ci_bit_clear(volatile ci_bits* b, int i) {
+  __asm__ __volatile__("lock; btrl %1, %0"
+		       : "=m" (*b)
+		       : "Ir" (i));
+}
+
+ci_inline int  ci_bit_test(volatile ci_bits* b, int i) {
+  char rc;
+  __asm__("btl %2, %1; setc %0"
+	  : "=r" (rc)
+	  : "m" (*b), "Ir" (i));
+  return rc;
+}
+
+ci_inline int ci_bit_test_and_set(volatile ci_bits* b, int i) {
+  char rc;
+  __asm__ __volatile__("lock; btsl %2, %1; setc %0"
+		       : "=r" (rc), "+m" (*b)
+		       : "Ir" (i));
+  return rc;
+}
+
+ci_inline int ci_bit_test_and_clear(volatile ci_bits* b, int i) {
+  char rc;
+  __asm__ __volatile__("lock; btrl %2, %1; setc %0"
+		       : "=r" (rc), "+m" (*b)
+		       : "Ir" (i));
+  return rc;
+}
+
+/* These mask ops only work within a single ci_bits word. */
+#define ci_bit_mask_set(b,m)	ci_atomic32_or((b), (m))
+#define ci_bit_mask_clear(b,m)	ci_atomic32_and((b), ~(m))
+
+
+/**********************************************************************
+ * Misc.
+ */
+
+#if __GNUC__ >= 3
+# define ci_spinloop_pause()  __asm__("pause") 
+#else
+# define ci_spinloop_pause()  __asm__(".byte 0xf3, 0x90")
+#endif
+
+
+#define CI_HAVE_ADDC32
+#define ci_add_carry32(sum, v)  __asm__("addl %1, %0 ;"			  \
+					"adcl $0, %0 ;"			  \
+					: "=r" (sum)			  \
+					: "g" ((ci_uint32) v), "0" (sum))
+
+
+#endif  /* __CI_TOOLS_GCC_X86_H__ */
+
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/tools/platform/linux_kernel.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/tools/platform/linux_kernel.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/tools/platform/linux_kernel.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/tools/platform/linux_kernel.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,362 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+
+/*! \cidoxg_include_ci_tools_platform  */
+
+#ifndef __CI_TOOLS_LINUX_KERNEL_H__
+#define __CI_TOOLS_LINUX_KERNEL_H__
+
+/**********************************************************************
+ * Need to know the kernel version.
+ */
+
+#ifndef LINUX_VERSION_CODE
+# include <linux/version.h>
+# ifndef UTS_RELEASE
+   /* 2.6.18 onwards defines UTS_RELEASE in a separate header */
+#  include <linux/utsrelease.h>
+# endif
+#endif
+
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,0) || \
+    LINUX_VERSION_CODE >= KERNEL_VERSION(2,7,0)
+# error "Linux 2.6 required"
+#endif
+
+
+#include <linux/slab.h>     /* kmalloc / kfree */
+#include <linux/vmalloc.h>  /* vmalloc / vfree */
+#include <linux/interrupt.h>/* in_interrupt()  */
+#include <linux/in.h>
+#include <linux/in6.h>
+#include <linux/spinlock.h>
+#include <linux/highmem.h>
+#include <linux/smp_lock.h>
+#include <linux/ctype.h>
+#include <linux/uio.h>
+#include <asm/current.h>
+#include <asm/errno.h>
+#include <asm/kmap_types.h>
+#include <asm/semaphore.h>
+
+#include <ci/tools/config.h>
+
+#define ci_in_irq        in_irq
+#define ci_in_interrupt  in_interrupt
+#define ci_in_atomic     in_atomic
+
+
+/**********************************************************************
+ * Misc stuff.
+ */
+
+#ifdef BUG
+# define  CI_BOMB     BUG
+#endif
+
+ci_inline void* __ci_alloc(size_t n)
+{ return kmalloc(n, (in_interrupt() ? GFP_ATOMIC : GFP_KERNEL)); }
+
+ci_inline void* __ci_atomic_alloc(size_t n)
+{ return kmalloc(n, GFP_ATOMIC ); }
+
+ci_inline void  __ci_free(void* p)     { return kfree(p);   }
+ci_inline void* __ci_vmalloc(size_t n) { return vmalloc(n); }
+ci_inline void  __ci_vfree(void* p)    { return vfree(p);   }
+
+
+#if CI_MEMLEAK_DEBUG_ALLOC_TABLE
+  #define ci_alloc(s)     ci_alloc_memleak_debug (s, __FILE__, __LINE__)
+  #define ci_atomic_alloc(s)  ci_atomic_alloc_memleak_debug(s, __FILE__, __LINE__)
+  #define ci_free         ci_free_memleak_debug
+  #define ci_vmalloc(s)   ci_vmalloc_memleak_debug (s, __FILE__,__LINE__)
+  #define ci_vfree        ci_vfree_memleak_debug
+  #define ci_alloc_fn     ci_alloc_fn_memleak_debug
+  #define ci_vmalloc_fn   ci_vmalloc_fn_memleak_debug
+#else /* !CI_MEMLEAK_DEBUG_ALLOC_TABLE */
+  #define ci_alloc_fn     __ci_alloc
+  #define ci_vmalloc_fn   __ci_vmalloc
+#endif 
+
+#ifndef ci_alloc
+  #define ci_atomic_alloc __ci_atomic_alloc
+  #define ci_alloc        __ci_alloc
+  #define ci_free         __ci_free
+  #define ci_vmalloc      __ci_vmalloc
+  #define ci_vmalloc_fn   __ci_vmalloc
+  #define ci_vfree        __ci_vfree
+#endif
+
+#define ci_sprintf        sprintf
+#define ci_vsprintf       vsprintf
+#define ci_snprintf       snprintf
+#define ci_vsnprintf      vsnprintf
+#define ci_sscanf         sscanf
+
+
+#define CI_LOG_FN_DEFAULT  ci_log_syslog
+
+
+/*--------------------------------------------------------------------
+ *
+ * irqs_disabled - needed for kmap helpers on some kernels 
+ *
+ *--------------------------------------------------------------------*/
+#ifdef irqs_disabled
+# define ci_irqs_disabled irqs_disabled
+#else
+# if defined(__i386__) | defined(__x86_64__)
+#   define ci_irqs_disabled(x)                  \
+  ({                                            \
+    unsigned long flags;                        \
+    local_save_flags(flags);                    \
+    !(flags & (1<<9));                          \
+  })
+# else
+#  error "Need to implement irqs_disabled() for your architecture"
+# endif
+#endif
+
+
+/**********************************************************************
+ * kmap helpers. 
+ *
+ * Use ci_k(un)map for code paths which are not in an atomic context.
+ * For atomic code you need to use ci_k(un)map_in_atomic. This will grab
+ * one of the per-CPU kmap slots.
+ *
+ * NB in_interrupt != in_irq. If you don't know the difference then
+ * don't use kmap_in_atomic
+ *
+ * 2.4 allocates kmap slots by function. We are going to re-use the
+ * skb module's slot - we also use the same interlock
+ * 
+ * 2.6 allocates kmap slots by type as well as by function. We are
+ * going to use the currently (2.6.10) unsused SOFTIRQ slot 
+ *
+ */
+
+ci_inline void* ci_kmap(struct page *page) {
+  CI_DEBUG(if( ci_in_atomic() | ci_in_interrupt() | ci_in_irq() )  BUG());
+  return kmap(page);
+}
+
+ci_inline void ci_kunmap(struct page *page) {
+  kunmap(page);
+}
+
+#define CI_KM_SLOT KM_SOFTIRQ0
+
+
+typedef struct semaphore ci_semaphore_t;
+
+ci_inline void
+ci_sem_init (ci_semaphore_t *sem, int val) {
+  sema_init (sem, val);
+}
+
+ci_inline void
+ci_sem_down (ci_semaphore_t *sem) {
+  down (sem);
+}
+
+ci_inline int
+ci_sem_trydown (ci_semaphore_t *sem) {
+  return down_trylock (sem);
+}
+
+ci_inline void
+ci_sem_up (ci_semaphore_t *sem) {
+  up (sem);
+}
+
+ci_inline int
+ci_sem_get_count(ci_semaphore_t *sem) {
+  return sem->count.counter;
+}
+
+ci_inline void* ci_kmap_in_atomic(struct page *page) 
+{
+  CI_DEBUG(if( ci_in_irq() )  BUG());
+
+  /* iSCSI can call without in_interrupt() but with irqs_disabled()
+     and in a context that can't sleep, so we need to check that
+     too */
+  if(ci_in_interrupt() || ci_irqs_disabled())
+    return kmap_atomic(page, CI_KM_SLOT);
+  else
+    return kmap(page);
+}
+
+ci_inline void ci_kunmap_in_atomic(struct page *page, void* kaddr) 
+{
+  CI_DEBUG(if( ci_in_irq() )  BUG());
+
+  /* iSCSI can call without in_interrupt() but with irqs_disabled()
+     and in a context that can't sleep, so we need to check that
+     too */
+  if(ci_in_interrupt() || ci_irqs_disabled())
+    kunmap_atomic(kaddr, CI_KM_SLOT);
+  else
+    kunmap(page);
+}
+
+/**********************************************************************
+ * spinlock implementation: used by <ci/tools/spinlock.h>
+ */
+
+#define CI_HAVE_SPINLOCKS
+
+typedef ci_uintptr_t    			ci_lock_holder_t;
+#define ci_lock_thisthread 		(ci_lock_holder_t)current		       	
+#define ci_lock_no_holder     (ci_lock_holder_t)NULL
+
+typedef spinlock_t			ci_lock_i;
+typedef spinlock_t			ci_irqlock_i;
+typedef unsigned long			ci_irqlock_state_t;
+
+#define IRQLOCK_CYCLES  500000
+
+#define ci_lock_ctor_i(l)		spin_lock_init(l)
+#define ci_lock_dtor_i(l)		do{}while(0)
+#define ci_lock_lock_i(l)		spin_lock(l)
+#define ci_lock_trylock_i(l)		spin_trylock(l)
+#define ci_lock_unlock_i(l)		spin_unlock(l)
+
+#define ci_irqlock_ctor_i(l)		spin_lock_init(l)
+#define ci_irqlock_dtor_i(l)		do{}while(0)
+#define ci_irqlock_lock_i(l,s)		spin_lock_irqsave(l,*(s))
+#define ci_irqlock_unlock_i(l,s)	spin_unlock_irqrestore(l, *(s))
+
+
+/**********************************************************************
+ * register access
+ */
+
+#include <asm/io.h>
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,9)
+typedef volatile void __iomem*	ioaddr_t;
+#else
+typedef unsigned long ioaddr_t;
+#endif
+
+
+
+/**********************************************************************
+ * thread implementation -- kernel dependancies probably should be
+ * moved to driver/linux_kernel.h
+ */
+
+#define ci_linux_daemonize(name) daemonize(name)
+
+#include <linux/workqueue.h>
+
+
+typedef struct {
+  void*			(*fn)(void* arg);
+  void*			arg;
+  const char*		name;
+  int			thrd_id;
+  struct completion	exit_event;
+  struct work_struct	keventd_witem;
+} ci_kernel_thread_t;
+
+
+typedef ci_kernel_thread_t* cithread_t;
+
+
+extern int cithread_create(cithread_t* tid, void* (*fn)(void*), void* arg,
+			   const char* name);
+extern int cithread_detach(cithread_t kt);
+extern int cithread_join(cithread_t kt);
+
+
+/* Kernel sysctl variables. */
+extern int sysctl_tcp_wmem[3];
+extern int sysctl_tcp_rmem[3];
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,24)
+#define LINUX_HAS_SYSCTL_MEM_MAX
+extern ci_uint32 sysctl_wmem_max;
+extern ci_uint32 sysctl_rmem_max;
+#endif
+
+
+/*--------------------------------------------------------------------
+ *
+ * ci_bigbuf_t: An abstraction of a large buffer.  Needed because in the
+ * Linux kernel, large buffers need to be allocated with vmalloc(), whereas
+ * smaller buffers should use kmalloc().  This abstraction chooses the
+ * appropriate mechansim.
+ *
+ *--------------------------------------------------------------------*/
+
+typedef struct {
+  char*		p;
+  int		is_vmalloc;
+} ci_bigbuf_t;
+
+
+ci_inline int ci_bigbuf_alloc(ci_bigbuf_t* bb, size_t bytes) {
+  if( bytes >= CI_PAGE_SIZE && ! ci_in_atomic() ) {
+    bb->is_vmalloc = 1;
+    if( (bb->p = vmalloc(bytes)) )  return 0;
+  }
+  bb->is_vmalloc = 0;
+  bb->p = kmalloc(bytes, ci_in_interrupt() ? GFP_ATOMIC : GFP_KERNEL);
+  return bb->p ? 0 : -ENOMEM;
+}
+
+ci_inline void ci_bigbuf_free(ci_bigbuf_t* bb) {
+  if( bb->is_vmalloc )  vfree(bb->p);
+  else                  kfree(bb->p);
+}
+
+ci_inline char* ci_bigbuf_ptr(ci_bigbuf_t* bb)
+{ return bb->p; }
+
+/**********************************************************************
+ * struct iovec abstraction (for Windows port)
+ */
+
+typedef struct iovec ci_iovec;
+
+/* Accessors for buffer/length */
+#define CI_IOVEC_BASE(i) ((i)->iov_base)
+#define CI_IOVEC_LEN(i)  ((i)->iov_len)
+
+/**********************************************************************
+ * Signals
+ */
+
+ci_inline void
+ci_send_sig(int signum)
+{
+  send_sig(signum, current, 0);
+}
+
+#endif  /* __CI_TOOLS_LINUX_KERNEL_H__ */
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/ci/tools/sysdep.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/tools/sysdep.h
--- linux-2.6.18.8/drivers/xen/sfc_netback/ci/tools/sysdep.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/ci/tools/sysdep.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,132 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*! \cidoxg_include_ci_tools */
+
+#ifndef __CI_TOOLS_SYSDEP_H__
+#define __CI_TOOLS_SYSDEP_H__
+
+/* Make this header self-sufficient */
+#include <ci/compat.h>
+#include <ci/tools/log.h>
+#include <ci/tools/debug.h>
+
+
+/**********************************************************************
+ * Platform dependencies.
+ */
+
+#if defined(__KERNEL__)
+
+# if defined(__linux__)
+#  include <ci/tools/platform/linux_kernel.h>
+# elif defined(_WIN32)
+#  include <ci/tools/platform/win32_kernel.h>
+# elif defined(__sun__)
+#  include <ci/tools/platform/sunos_kernel.h>
+# else
+#  error Unknown platform.
+# endif
+
+#elif defined(_WIN32)
+
+# include <ci/tools/platform/win32.h>
+
+#elif defined(__unix__)
+
+# include <ci/tools/platform/unix.h>
+
+#else
+
+# error Unknown platform.
+
+#endif
+
+#if defined(__linux__)
+/*! Linux sendfile() support enable/disable. */
+# define CI_HAVE_SENDFILE            /* provide sendfile i/f */
+
+# define CI_HAVE_OS_NOPAGE
+#endif
+
+#if defined(__sun__)
+# define CI_HAVE_SENDFILE	     /* provide sendfile i/f */
+# define CI_HAVE_SENDFILEV           /* provide sendfilev i/f */
+
+# define CI_IOCTL_SENDFILE           /*  use efrm CI_SENDFILEV ioctl */
+#endif
+
+#if defined(_WIN32)
+typedef ci_uint32 ci_uerr_t; /* range of OS user-mode return codes */
+typedef ci_uint32 ci_kerr_t; /* range of OS kernel-mode return codes */
+#elif defined(__unix__)
+typedef ci_int32 ci_uerr_t; /* range of OS user-mode return codes */
+typedef ci_int32 ci_kerr_t; /* range of OS kernel-mode return codes */
+#endif
+
+
+/**********************************************************************
+ * Compiler and processor dependencies.
+ */
+
+#if defined(__GNUC__)
+
+#if defined(__i386__) || defined(__x86_64__)
+# include <ci/tools/platform/gcc_x86.h>
+#elif defined(__PPC__)
+#  include <ci/tools/platform/gcc_ppc.h>
+#elif defined(__ia64__)
+#  include <ci/tools/platform/gcc_ia64.h>
+#else
+# error Unknown processor.
+#endif
+
+#elif defined(_MSC_VER)
+
+#if defined(__i386__)
+# include <ci/tools/platform/msvc_x86.h>
+# elif defined(__x86_64__)
+# include <ci/tools/platform/msvc_x86_64.h>
+#else
+# error Unknown processor.
+#endif
+
+#elif defined(__PGI)
+
+# include <ci/tools/platform/pg_x86.h>
+
+#elif defined(__INTEL_COMPILER)
+
+/* Intel compilers v7 claim to be very gcc compatible. */
+# include <ci/tools/platform/gcc_x86.h>
+
+#else
+# error Unknown compiler.
+#endif
+
+
+#endif  /* __CI_TOOLS_SYSDEP_H__ */
+
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netback/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/Makefile
--- linux-2.6.18.8/drivers/xen/sfc_netback/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netback/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,12 @@
+EXTRA_CFLAGS += -Idrivers/xen/sfc_netback -Idrivers/xen/sfc_netutil -Idrivers/xen/netback -Idrivers/net/sfc
+EXTRA_CFLAGS += -D__ci_driver__ 
+EXTRA_CFLAGS += -DEFX_USE_KCOMPAT
+EXTRA_CFLAGS += -Werror
+
+ifdef GCOV
+EXTRA_CFLAGS += -fprofile-arcs -ftest-coverage -DEFX_GCOV
+endif
+
+obj-$(CONFIG_XEN_NETDEV_ACCEL_SFC_BACKEND) := sfc_netback.o
+
+sfc_netback-objs   := accel.o accel_fwd.o accel_msg.o accel_solarflare.o accel_xenbus.o accel_debugfs.o
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/accel_bufs.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_bufs.c
--- linux-2.6.18.8/drivers/xen/sfc_netfront/accel_bufs.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_bufs.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,393 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <xen/gnttab.h>
+
+#include "accel_bufs.h"
+#include "accel_util.h"
+
+#include "accel.h"
+
+
+static int 
+netfront_accel_alloc_buf_desc_blocks(struct netfront_accel_bufinfo *manager,
+				     int pages)
+{
+	manager->desc_blocks = 
+		kzalloc(sizeof(struct netfront_accel_pkt_desc *) * 
+			NETFRONT_ACCEL_BUF_NUM_BLOCKS(pages), GFP_KERNEL);
+	if (manager->desc_blocks == NULL) {
+		return -ENOMEM;
+	}
+	
+	return 0;
+}
+
+static int 
+netfront_accel_alloc_buf_lists(struct netfront_accel_bufpages *bufpages,
+			       int pages)
+{
+	bufpages->page_list = kmalloc(pages * sizeof(void *), GFP_KERNEL);
+	if (bufpages->page_list == NULL) {
+		return -ENOMEM;
+	}
+
+	bufpages->grant_list = kzalloc(pages * sizeof(grant_ref_t), GFP_KERNEL);
+	if (bufpages->grant_list == NULL) {
+		kfree(bufpages->page_list);
+		bufpages->page_list = NULL;
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+
+int netfront_accel_alloc_buffer_mem(struct netfront_accel_bufpages *bufpages,
+				    struct netfront_accel_bufinfo *rx_manager,
+				    struct netfront_accel_bufinfo *tx_manager,
+				    int pages)
+{
+	int n, rc;
+
+	if ((rc = netfront_accel_alloc_buf_desc_blocks
+	     (rx_manager, pages - (pages / sfc_netfront_buffer_split))) < 0) {
+		goto rx_fail;
+	}
+
+	if ((rc = netfront_accel_alloc_buf_desc_blocks
+	     (tx_manager, pages / sfc_netfront_buffer_split)) < 0) {
+		goto tx_fail;
+	}
+
+	if ((rc = netfront_accel_alloc_buf_lists(bufpages, pages)) < 0) {
+		goto lists_fail;
+	}
+
+	for (n = 0; n < pages; n++) {
+		void *tmp = (void*)__get_free_page(GFP_KERNEL);
+		if (tmp == NULL)
+			break;
+
+		bufpages->page_list[n] = tmp;
+	}
+
+	if (n != pages) {
+		EPRINTK("%s: not enough pages: %d != %d\n", __FUNCTION__, n, 
+			pages);
+		for (; n >= 0; n--)
+			free_page((unsigned long)(bufpages->page_list[n]));
+		rc = -ENOMEM;
+		goto pages_fail;
+	}
+
+	bufpages->max_pages = pages;
+	bufpages->page_reqs = 0;
+
+	return 0;
+
+ pages_fail:
+	kfree(bufpages->page_list);
+	kfree(bufpages->grant_list);
+
+	bufpages->page_list = NULL;
+	bufpages->grant_list = NULL;
+ lists_fail:
+	kfree(tx_manager->desc_blocks);
+	tx_manager->desc_blocks = NULL;
+
+ tx_fail:
+	kfree(rx_manager->desc_blocks);
+	rx_manager->desc_blocks = NULL;
+ rx_fail:
+	return rc;
+}
+
+
+void netfront_accel_free_buffer_mem(struct netfront_accel_bufpages *bufpages,
+				    struct netfront_accel_bufinfo *rx_manager,
+				    struct netfront_accel_bufinfo *tx_manager)
+{
+	int i;
+
+	for (i = 0; i < bufpages->max_pages; i++) {
+		if (bufpages->grant_list[i] != 0)
+			net_accel_ungrant_page(bufpages->grant_list[i]);
+		free_page((unsigned long)(bufpages->page_list[i]));
+	}
+
+	if (bufpages->max_pages) {
+		kfree(bufpages->page_list);
+		kfree(bufpages->grant_list);
+		kfree(rx_manager->desc_blocks);
+		kfree(tx_manager->desc_blocks);
+	}
+}
+
+
+/*
+ * Allocate memory for the buffer manager and create a lock.  If no
+ * lock is supplied its own is allocated.
+ */
+struct netfront_accel_bufinfo *netfront_accel_init_bufs(spinlock_t *lock)
+{
+	struct netfront_accel_bufinfo *res = kmalloc(sizeof(*res), GFP_KERNEL);
+	if (res != NULL) {
+		res->npages = res->nused = 0;
+		res->first_free = -1;
+
+		if (lock == NULL) {
+			res->lock = kmalloc(sizeof(*res->lock), GFP_KERNEL);
+			if (res->lock == NULL) {
+				kfree(res);
+				return NULL;
+			}
+			spin_lock_init(res->lock);
+			res->internally_locked = 1;
+		} else {
+			res->lock = lock;
+			res->internally_locked = 0;
+		}
+		
+		res->desc_blocks = NULL;
+	}
+
+	return res;
+}
+
+
+void netfront_accel_fini_bufs(struct netfront_accel_bufinfo *bufs)
+{
+	if (bufs->internally_locked)
+		kfree(bufs->lock);
+	kfree(bufs);
+}
+
+
+int netfront_accel_buf_map_request(struct xenbus_device *dev,
+				   struct netfront_accel_bufpages *bufpages,
+				   struct net_accel_msg *msg, 
+				   int pages, int offset)
+{
+	int i, mfn;
+	int err;
+
+	net_accel_msg_init(msg, NET_ACCEL_MSG_MAPBUF);
+
+	BUG_ON(pages > NET_ACCEL_MSG_MAX_PAGE_REQ);
+
+	msg->u.mapbufs.pages = pages;
+
+	for (i = 0; i < msg->u.mapbufs.pages; i++) {
+		/* 
+		 * This can happen if we tried to send this message
+		 * earlier but the queue was full.
+		 */
+		if (bufpages->grant_list[offset+i] != 0) {
+			msg->u.mapbufs.grants[i] = 
+				bufpages->grant_list[offset+i];
+			continue;
+		}
+
+		mfn = virt_to_mfn(bufpages->page_list[offset+i]);
+		VPRINTK("%s: Granting page %d, mfn %08x\n",
+			__FUNCTION__, i, mfn);
+
+		bufpages->grant_list[offset+i] =
+			net_accel_grant_page(dev, mfn, 0);
+		msg->u.mapbufs.grants[i] = bufpages->grant_list[offset+i];
+
+		if (msg->u.mapbufs.grants[i] < 0) {
+			EPRINTK("%s: Failed to grant buffer: %d\n",
+				__FUNCTION__, msg->u.mapbufs.grants[i]);
+			err = -EIO;
+			goto error;
+		}
+	}
+
+	/* This is interpreted on return as the offset in the the page_list */
+	msg->u.mapbufs.reqid = offset;
+
+	return 0;
+
+error:
+	/* Ungrant all the pages we've successfully granted. */
+	for (i--; i >= 0; i--) {
+		net_accel_ungrant_page(bufpages->grant_list[offset+i]);
+		bufpages->grant_list[offset+i] = 0;
+	}
+	return err;
+}
+
+
+/* Process a response to a buffer request. */
+int netfront_accel_add_bufs(struct netfront_accel_bufpages *bufpages,
+			    struct netfront_accel_bufinfo *manager, 
+			    struct net_accel_msg *msg)
+{
+	int msg_pages, page_offset, i, newtot;
+	int old_block_count, new_block_count;
+	u32 msg_buf;
+	unsigned long flags;
+
+	VPRINTK("%s: manager %p msg %p\n", __FUNCTION__, manager, msg);
+
+	BUG_ON(msg->id != (NET_ACCEL_MSG_MAPBUF | NET_ACCEL_MSG_REPLY));
+
+	msg_pages = msg->u.mapbufs.pages;
+	msg_buf = msg->u.mapbufs.buf;
+	page_offset = msg->u.mapbufs.reqid;
+
+	spin_lock_irqsave(manager->lock, flags);
+	newtot = manager->npages + msg_pages;
+	old_block_count = 
+		(manager->npages + NETFRONT_ACCEL_BUF_PAGES_PER_BLOCK - 1) >>
+		NETFRONT_ACCEL_BUF_PAGES_PER_BLOCK_SHIFT;
+	new_block_count = 
+		(newtot + NETFRONT_ACCEL_BUF_PAGES_PER_BLOCK - 1) >>
+		NETFRONT_ACCEL_BUF_PAGES_PER_BLOCK_SHIFT;
+
+	for (i = old_block_count; i < new_block_count; i++) {
+		struct netfront_accel_pkt_desc *block;
+		if (manager->desc_blocks[i] != NULL) {
+			VPRINTK("Not needed\n");
+			continue;
+		}
+		block = kzalloc(NETFRONT_ACCEL_BUFS_PER_BLOCK * 
+				sizeof(netfront_accel_pkt_desc), GFP_ATOMIC);
+		if (block == NULL) {
+			spin_unlock_irqrestore(manager->lock, flags);
+			return -ENOMEM;
+		}
+		manager->desc_blocks[i] = block;
+	}
+	for (i = manager->npages; i < newtot; i++) {
+		int k, j = i - manager->npages;
+		int block_num;
+		int block_idx;
+		struct netfront_accel_pkt_desc *pkt;
+
+		block_num = i >> NETFRONT_ACCEL_BUF_PAGES_PER_BLOCK_SHIFT;
+		block_idx = (NETFRONT_ACCEL_BUFS_PER_PAGE*i)
+			& (NETFRONT_ACCEL_BUFS_PER_BLOCK-1);
+
+		pkt = manager->desc_blocks[block_num] + block_idx;
+		
+		for (k = 0; k < NETFRONT_ACCEL_BUFS_PER_PAGE; k++) {
+			BUG_ON(page_offset + j >= bufpages->max_pages);
+
+			pkt[k].buf_id = NETFRONT_ACCEL_BUFS_PER_PAGE * i + k;
+			pkt[k].pkt_kva = bufpages->page_list[page_offset + j] +
+				(PAGE_SIZE/NETFRONT_ACCEL_BUFS_PER_PAGE) * k;
+			pkt[k].pkt_buff_addr = msg_buf +
+				(PAGE_SIZE/NETFRONT_ACCEL_BUFS_PER_PAGE) * 
+				(NETFRONT_ACCEL_BUFS_PER_PAGE * j + k);
+			pkt[k].next_free = manager->first_free;
+			manager->first_free = pkt[k].buf_id;
+			*(int*)(pkt[k].pkt_kva) = pkt[k].buf_id;
+
+			VPRINTK("buf %d desc %p kva %p buffaddr %x\n",
+				pkt[k].buf_id, &(pkt[k]), pkt[k].pkt_kva, 
+				pkt[k].pkt_buff_addr);
+		}
+	}
+	manager->npages = newtot;
+	spin_unlock_irqrestore(manager->lock, flags);
+	VPRINTK("Added %d pages. Total is now %d\n", msg_pages,
+		manager->npages);
+	return 0;
+}
+
+
+netfront_accel_pkt_desc *
+netfront_accel_buf_find(struct netfront_accel_bufinfo *manager, u16 id)
+{
+	netfront_accel_pkt_desc *pkt;
+	int block_num = id >> NETFRONT_ACCEL_BUFS_PER_BLOCK_SHIFT;
+	int block_idx = id & (NETFRONT_ACCEL_BUFS_PER_BLOCK - 1);
+	BUG_ON(id >= manager->npages * NETFRONT_ACCEL_BUFS_PER_PAGE);
+	BUG_ON(block_idx >= NETFRONT_ACCEL_BUFS_PER_BLOCK);
+	pkt = manager->desc_blocks[block_num] + block_idx;
+	return pkt;
+}
+
+
+/* Allocate a buffer from the buffer manager */
+netfront_accel_pkt_desc *
+netfront_accel_buf_get(struct netfront_accel_bufinfo *manager)
+{
+	int bufno = -1;
+	netfront_accel_pkt_desc *buf = NULL;
+	unsigned long flags = 0;
+
+	/* Any spare? */
+	if (manager->first_free == -1)
+		return NULL;
+	/* Take lock */
+	if (manager->internally_locked)
+		spin_lock_irqsave(manager->lock, flags);
+	bufno = manager->first_free;
+	if (bufno != -1) {
+		buf = netfront_accel_buf_find(manager, bufno);
+		manager->first_free = buf->next_free;
+		manager->nused++;
+	}
+	/* Release lock */
+	if (manager->internally_locked)
+		spin_unlock_irqrestore(manager->lock, flags);
+
+	/* Tell the world */
+	VPRINTK("Allocated buffer %i, buffaddr %x\n", bufno,
+		buf->pkt_buff_addr);
+
+	return buf;
+}
+
+
+/* Release a buffer back to the buffer manager pool */
+int netfront_accel_buf_put(struct netfront_accel_bufinfo *manager, u16 id)
+{
+	netfront_accel_pkt_desc *buf = netfront_accel_buf_find(manager, id);
+	unsigned long flags = 0;
+	unsigned was_empty = 0;
+	int bufno = id;
+
+	VPRINTK("Freeing buffer %i\n", id);
+	BUG_ON(id == (u16)-1);
+
+	if (manager->internally_locked)
+		spin_lock_irqsave(manager->lock, flags);
+
+	if (manager->first_free == -1)
+		was_empty = 1;
+
+	buf->next_free = manager->first_free;
+	manager->first_free = bufno;
+	manager->nused--;
+
+	if (manager->internally_locked)
+		spin_unlock_irqrestore(manager->lock, flags);
+
+	return was_empty;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/accel_bufs.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_bufs.h
--- linux-2.6.18.8/drivers/xen/sfc_netfront/accel_bufs.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_bufs.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,181 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef NETFRONT_ACCEL_BUFS_H
+#define NETFRONT_ACCEL_BUFS_H
+
+#include <linux/skbuff.h>
+#include <linux/spinlock.h>
+#include <xen/xenbus.h>
+
+#include "accel_msg_iface.h"
+
+
+/*! Buffer descriptor structure */
+typedef struct netfront_accel_pkt_desc {
+	int buf_id;
+	u32 pkt_buff_addr;
+	void *pkt_kva;
+	/* This is the socket buffer currently married to this buffer */
+	struct sk_buff *skb;
+	int next_free;
+} netfront_accel_pkt_desc;
+
+
+#define NETFRONT_ACCEL_DEFAULT_BUF_PAGES (384)
+#define NETFRONT_ACCEL_BUF_PAGES_PER_BLOCK_SHIFT (4)
+#define NETFRONT_ACCEL_BUF_PAGES_PER_BLOCK		\
+	(1 << (NETFRONT_ACCEL_BUF_PAGES_PER_BLOCK_SHIFT))
+#define NETFRONT_ACCEL_BUFS_PER_PAGE_SHIFT (1)
+#define NETFRONT_ACCEL_BUFS_PER_PAGE			\
+	(1 << (NETFRONT_ACCEL_BUFS_PER_PAGE_SHIFT))
+#define NETFRONT_ACCEL_BUFS_PER_BLOCK_SHIFT		\
+	(NETFRONT_ACCEL_BUF_PAGES_PER_BLOCK_SHIFT +     \
+	 NETFRONT_ACCEL_BUFS_PER_PAGE_SHIFT)
+#define NETFRONT_ACCEL_BUFS_PER_BLOCK			\
+	(1 << NETFRONT_ACCEL_BUFS_PER_BLOCK_SHIFT)
+#define NETFRONT_ACCEL_BUF_NUM_BLOCKS(max_pages)			\
+	(((max_pages)+NETFRONT_ACCEL_BUF_PAGES_PER_BLOCK-1) /		\
+	 NETFRONT_ACCEL_BUF_PAGES_PER_BLOCK)
+
+/*! Buffer management structure. */
+struct netfront_accel_bufinfo {
+	/* number added to this manager */
+	unsigned npages;
+	/* number currently used from this manager */
+	unsigned nused;
+
+	int first_free;
+
+	int internally_locked;
+	spinlock_t *lock;
+
+	/*
+	 * array of pointers (length NETFRONT_ACCEL_BUF_NUM_BLOCKS) to
+	 * pkt descs
+	 */
+	struct netfront_accel_pkt_desc **desc_blocks; 
+};
+
+
+struct netfront_accel_bufpages {
+	/* length of lists of pages/grants */
+	int max_pages;
+	/* list of pages allocated for network buffers */
+	void **page_list;
+	/* list of grants for the above pages */
+	grant_ref_t *grant_list;
+	
+	/* number of page requests that have been made */
+	unsigned page_reqs;
+};
+
+
+/*! Allocate memory for the buffer manager, set up locks etc.
+ * Optionally takes a lock to use, if not supplied it makes its own.
+ *
+ * \return pointer to netfront_accel_bufinfo structure that represents the
+ * buffer manager
+ */
+extern struct netfront_accel_bufinfo *
+netfront_accel_init_bufs(spinlock_t *lock);
+
+/*! Allocate memory for the buffers
+ */
+extern int
+netfront_accel_alloc_buffer_mem(struct netfront_accel_bufpages *bufpages,
+				struct netfront_accel_bufinfo *rx_res,
+				struct netfront_accel_bufinfo *tx_res,
+				int pages);
+extern void
+netfront_accel_free_buffer_mem(struct netfront_accel_bufpages *bufpages,
+			       struct netfront_accel_bufinfo *rx_res,
+			       struct netfront_accel_bufinfo *tx_res);
+
+/*! Release memory for the buffer manager, buffers, etc.
+ *
+ * \param manager pointer to netfront_accel_bufinfo structure that
+ * represents the buffer manager
+ */
+extern void netfront_accel_fini_bufs(struct netfront_accel_bufinfo *manager);
+
+/*! Release a buffer.
+ *
+ * \param manager  The buffer manager which owns the buffer.
+ * \param id   The buffer identifier.
+ */
+extern int netfront_accel_buf_put(struct netfront_accel_bufinfo *manager, 
+				  u16 id);
+
+/*! Get the packet descriptor associated with a buffer id.
+ *
+ * \param manager  The buffer manager which owns the buffer.
+ * \param id       The buffer identifier.
+ *
+ * The returned value is the packet descriptor for this buffer.
+ */
+extern netfront_accel_pkt_desc *
+netfront_accel_buf_find(struct netfront_accel_bufinfo *manager, u16 id);
+
+
+/*! Fill out a message request for some buffers to be mapped by the
+ * back end driver
+ * 
+ * \param manager The buffer manager 
+ * \param msg Pointer to an ef_msg to complete.
+ * \return 0 on success
+ */
+extern int 
+netfront_accel_buf_map_request(struct xenbus_device *dev,
+			       struct netfront_accel_bufpages *bufpages,
+			       struct net_accel_msg *msg, 
+			       int pages, int offset);
+
+/*! Process a response to a buffer request. 
+ * 
+ * Deal with a received message from the back end in response to our
+ * request for buffers
+ * 
+ * \param manager The buffer manager
+ * \param msg The received message from the back end describing new
+ * buffers
+ * \return 0 on success
+ */
+extern int 
+netfront_accel_add_bufs(struct netfront_accel_bufpages *bufpages,
+			struct netfront_accel_bufinfo *manager,
+			struct net_accel_msg *msg);
+
+
+/*! Allocate a buffer from the buffer manager 
+ *
+ * \param manager The buffer manager data structure
+ * \param id On exit, the id of the buffer allocated
+ * \return Pointer to buffer descriptor.
+ */
+struct netfront_accel_pkt_desc *
+netfront_accel_buf_get(struct netfront_accel_bufinfo *manager);
+
+#endif /* NETFRONT_ACCEL_BUFS_H */
+
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/accel_debugfs.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_debugfs.c
--- linux-2.6.18.8/drivers/xen/sfc_netfront/accel_debugfs.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_debugfs.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,211 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/fs.h>
+#include <linux/debugfs.h>
+
+#include "accel.h"
+
+#if defined(CONFIG_DEBUG_FS)
+static struct dentry *sfc_debugfs_root = NULL;
+#endif
+
+void netfront_accel_debugfs_init(void) 
+{
+#if defined(CONFIG_DEBUG_FS)
+	sfc_debugfs_root = debugfs_create_dir(frontend_name, NULL);
+#endif
+}
+
+
+void netfront_accel_debugfs_fini(void)
+{
+#if defined(CONFIG_DEBUG_FS)
+	if (sfc_debugfs_root)
+		debugfs_remove(sfc_debugfs_root);
+#endif
+}
+
+
+int netfront_accel_debugfs_create(netfront_accel_vnic *vnic)
+{
+#if defined(CONFIG_DEBUG_FS)
+	if (sfc_debugfs_root == NULL)
+		return -ENOENT;
+
+	vnic->dbfs_dir = debugfs_create_dir(vnic->net_dev->name, 
+					    sfc_debugfs_root);
+	if (vnic->dbfs_dir == NULL)
+		return -ENOMEM;
+
+	vnic->netdev_dbfs.fastpath_rx_pkts = debugfs_create_u32
+		("fastpath_rx_pkts", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->netdev_stats.fastpath_rx_pkts);
+	vnic->netdev_dbfs.fastpath_rx_bytes = debugfs_create_u32
+		("fastpath_rx_bytes", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->netdev_stats.fastpath_rx_bytes);
+	vnic->netdev_dbfs.fastpath_rx_errors = debugfs_create_u32
+		("fastpath_rx_errors", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->netdev_stats.fastpath_rx_errors);
+	vnic->netdev_dbfs.fastpath_tx_pkts = debugfs_create_u32
+		("fastpath_tx_pkts", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->netdev_stats.fastpath_tx_pkts);
+	vnic->netdev_dbfs.fastpath_tx_bytes = debugfs_create_u32
+		("fastpath_tx_bytes", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->netdev_stats.fastpath_tx_bytes);
+	vnic->netdev_dbfs.fastpath_tx_errors = debugfs_create_u32
+		("fastpath_tx_errors", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->netdev_stats.fastpath_tx_errors);
+
+#if NETFRONT_ACCEL_STATS
+	vnic->dbfs.irq_count = debugfs_create_u64
+		("irq_count", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.irq_count);
+	vnic->dbfs.useless_irq_count = debugfs_create_u64
+		("useless_irq_count", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.useless_irq_count);
+	vnic->dbfs.poll_schedule_count = debugfs_create_u64
+		("poll_schedule_count", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.poll_schedule_count);
+	vnic->dbfs.poll_call_count = debugfs_create_u64
+		("poll_call_count", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.poll_call_count);
+	vnic->dbfs.poll_reschedule_count = debugfs_create_u64
+		("poll_reschedule_count", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.poll_reschedule_count);
+	vnic->dbfs.queue_stops = debugfs_create_u64
+		("queue_stops", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.queue_stops);
+	vnic->dbfs.queue_wakes = debugfs_create_u64
+		("queue_wakes", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.queue_wakes);
+	vnic->dbfs.ssr_bursts = debugfs_create_u64
+		("ssr_bursts", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.ssr_bursts);
+	vnic->dbfs.ssr_drop_stream = debugfs_create_u64
+		("ssr_drop_stream", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.ssr_drop_stream);
+	vnic->dbfs.ssr_misorder = debugfs_create_u64
+		("ssr_misorder", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.ssr_misorder);
+	vnic->dbfs.ssr_slow_start = debugfs_create_u64
+		("ssr_slow_start", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.ssr_slow_start);
+	vnic->dbfs.ssr_merges = debugfs_create_u64
+		("ssr_merges", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.ssr_merges);
+	vnic->dbfs.ssr_too_many = debugfs_create_u64
+		("ssr_too_many", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.ssr_too_many);
+	vnic->dbfs.ssr_new_stream = debugfs_create_u64
+		("ssr_new_stream", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.ssr_new_stream);
+
+	vnic->dbfs.fastpath_tx_busy = debugfs_create_u64
+		("fastpath_tx_busy", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.fastpath_tx_busy);
+	vnic->dbfs.fastpath_tx_completions = debugfs_create_u64
+		("fastpath_tx_completions", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.fastpath_tx_completions);
+	vnic->dbfs.fastpath_tx_pending_max = debugfs_create_u32
+		("fastpath_tx_pending_max", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.fastpath_tx_pending_max);
+	vnic->dbfs.event_count = debugfs_create_u64
+		("event_count", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.event_count);
+	vnic->dbfs.bad_event_count = debugfs_create_u64
+		("bad_event_count", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.bad_event_count);
+	vnic->dbfs.event_count_since_irq = debugfs_create_u32
+		("event_count_since_irq", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.event_count_since_irq);
+	vnic->dbfs.events_per_irq_max = debugfs_create_u32
+		("events_per_irq_max", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.events_per_irq_max);
+	vnic->dbfs.fastpath_frm_trunc = debugfs_create_u64
+		("fastpath_frm_trunc", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.fastpath_frm_trunc);
+	vnic->dbfs.rx_no_desc_trunc = debugfs_create_u64
+		("rx_no_desc_trunc", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.rx_no_desc_trunc);
+	vnic->dbfs.events_per_poll_max = debugfs_create_u32
+		("events_per_poll_max", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.events_per_poll_max);
+	vnic->dbfs.events_per_poll_rx_max = debugfs_create_u32
+		("events_per_poll_rx_max", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.events_per_poll_rx_max);
+	vnic->dbfs.events_per_poll_tx_max = debugfs_create_u32
+		("events_per_poll_tx_max", S_IRUSR | S_IRGRP | S_IROTH,
+		 vnic->dbfs_dir, &vnic->stats.events_per_poll_tx_max);
+#endif
+#endif
+	return 0;
+}
+
+
+int netfront_accel_debugfs_remove(netfront_accel_vnic *vnic)
+{
+#if defined(CONFIG_DEBUG_FS)
+	if (vnic->dbfs_dir != NULL) {
+		debugfs_remove(vnic->netdev_dbfs.fastpath_rx_pkts);
+		debugfs_remove(vnic->netdev_dbfs.fastpath_rx_bytes);
+		debugfs_remove(vnic->netdev_dbfs.fastpath_rx_errors);
+		debugfs_remove(vnic->netdev_dbfs.fastpath_tx_pkts);
+		debugfs_remove(vnic->netdev_dbfs.fastpath_tx_bytes);
+		debugfs_remove(vnic->netdev_dbfs.fastpath_tx_errors);
+		
+#if NETFRONT_ACCEL_STATS
+		debugfs_remove(vnic->dbfs.irq_count);
+		debugfs_remove(vnic->dbfs.useless_irq_count);
+		debugfs_remove(vnic->dbfs.poll_schedule_count);
+		debugfs_remove(vnic->dbfs.poll_call_count);
+		debugfs_remove(vnic->dbfs.poll_reschedule_count);
+		debugfs_remove(vnic->dbfs.queue_stops);
+		debugfs_remove(vnic->dbfs.queue_wakes);
+		debugfs_remove(vnic->dbfs.ssr_bursts);
+		debugfs_remove(vnic->dbfs.ssr_drop_stream);
+		debugfs_remove(vnic->dbfs.ssr_misorder);
+		debugfs_remove(vnic->dbfs.ssr_slow_start);
+		debugfs_remove(vnic->dbfs.ssr_merges);
+		debugfs_remove(vnic->dbfs.ssr_too_many);
+		debugfs_remove(vnic->dbfs.ssr_new_stream);
+		
+		debugfs_remove(vnic->dbfs.fastpath_tx_busy);
+		debugfs_remove(vnic->dbfs.fastpath_tx_completions);
+		debugfs_remove(vnic->dbfs.fastpath_tx_pending_max);
+		debugfs_remove(vnic->dbfs.event_count);
+		debugfs_remove(vnic->dbfs.bad_event_count);
+		debugfs_remove(vnic->dbfs.event_count_since_irq);
+		debugfs_remove(vnic->dbfs.events_per_irq_max);
+		debugfs_remove(vnic->dbfs.fastpath_frm_trunc);
+		debugfs_remove(vnic->dbfs.rx_no_desc_trunc);
+		debugfs_remove(vnic->dbfs.events_per_poll_max);
+		debugfs_remove(vnic->dbfs.events_per_poll_rx_max);
+		debugfs_remove(vnic->dbfs.events_per_poll_tx_max);
+#endif
+		debugfs_remove(vnic->dbfs_dir);
+	}
+#endif
+	return 0;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/accel.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel.h
--- linux-2.6.18.8/drivers/xen/sfc_netfront/accel.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,477 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef NETFRONT_ACCEL_H
+#define NETFRONT_ACCEL_H
+
+#include "accel_msg_iface.h"
+#include "accel_cuckoo_hash.h"
+#include "accel_bufs.h"
+
+#include "etherfabric/ef_vi.h"
+
+#include <xen/xenbus.h>
+#include <xen/evtchn.h>
+
+#include <linux/kernel.h>
+#include <linux/list.h>
+
+enum netfront_accel_post_status {
+	NETFRONT_ACCEL_STATUS_GOOD,
+	NETFRONT_ACCEL_STATUS_BUSY,
+	NETFRONT_ACCEL_STATUS_CANT
+};
+
+#define NETFRONT_ACCEL_STATS 1
+#if NETFRONT_ACCEL_STATS
+#define NETFRONT_ACCEL_STATS_OP(x) x
+#else
+#define NETFRONT_ACCEL_STATS_OP(x)
+#endif
+
+
+enum netfront_accel_msg_state {
+	NETFRONT_ACCEL_MSG_NONE = 0,
+	NETFRONT_ACCEL_MSG_HELLO = 1,
+	NETFRONT_ACCEL_MSG_HW = 2
+};
+
+
+typedef struct {
+	u32 in_progress;
+	u32 total_len;
+	struct sk_buff *skb;
+} netfront_accel_jumbo_state;
+
+
+struct netfront_accel_ssr_state {
+	/** List of tracked connections. */
+	struct list_head conns;
+
+	/** Free efx_ssr_conn instances. */
+	struct list_head free_conns;
+};
+
+
+struct netfront_accel_netdev_stats {
+	/* Fastpath stats. */
+	u32 fastpath_rx_pkts;
+	u32 fastpath_rx_bytes;
+	u32 fastpath_rx_errors;
+	u32 fastpath_tx_pkts; 
+	u32 fastpath_tx_bytes;
+	u32 fastpath_tx_errors;
+};
+
+
+struct netfront_accel_netdev_dbfs {
+	struct dentry *fastpath_rx_pkts;
+	struct dentry *fastpath_rx_bytes;
+	struct dentry *fastpath_rx_errors;
+	struct dentry *fastpath_tx_pkts; 
+	struct dentry *fastpath_tx_bytes;
+	struct dentry *fastpath_tx_errors;
+};
+
+
+struct netfront_accel_stats {
+	/** Fast path events */
+	u64 fastpath_tx_busy;
+
+	/** TX DMA queue status */
+	u64 fastpath_tx_completions;
+
+	/** The number of events processed. */
+	u64 event_count;
+
+	/** Number of frame trunc events seen on fastpath */
+	u64 fastpath_frm_trunc;
+
+	/** Number of no rx descriptor trunc events seen on fastpath */
+	u64 rx_no_desc_trunc;
+
+	/** The number of misc bad events (e.g. RX_DISCARD) processed. */
+	u64 bad_event_count;
+
+	/** Number of events dealt with in poll loop */
+	u32 events_per_poll_max;
+	u32 events_per_poll_tx_max;
+	u32 events_per_poll_rx_max;
+
+	/** Largest number of concurrently outstanding tx descriptors */
+	u32 fastpath_tx_pending_max;
+
+	/** The number of events since the last interrupts. */
+	u32 event_count_since_irq;
+
+	/** The max number of events between interrupts. */
+	u32 events_per_irq_max;
+
+	/** The number of interrupts. */
+	u64 irq_count;
+
+	/** The number of useless interrupts. */
+	u64 useless_irq_count;
+
+	/** The number of polls scheduled. */
+	u64 poll_schedule_count;
+
+	/** The number of polls called. */
+	u64 poll_call_count;
+
+	/** The number of rechecks. */
+	u64 poll_reschedule_count;
+
+	/** Number of times we've called netif_stop_queue/netif_wake_queue */
+	u64 queue_stops;
+	u64 queue_wakes;
+
+	/** SSR stats */
+	u64 ssr_bursts;
+	u64 ssr_drop_stream;
+	u64 ssr_misorder;
+	u64 ssr_slow_start;
+	u64 ssr_merges;
+	u64 ssr_too_many;
+	u64 ssr_new_stream;
+};
+
+
+struct netfront_accel_dbfs {
+	struct dentry *fastpath_tx_busy;
+	struct dentry *fastpath_tx_completions;
+	struct dentry *fastpath_tx_pending_max;
+	struct dentry *fastpath_frm_trunc;
+	struct dentry *rx_no_desc_trunc;
+	struct dentry *event_count;
+	struct dentry *bad_event_count;
+	struct dentry *events_per_poll_max;
+	struct dentry *events_per_poll_rx_max;
+	struct dentry *events_per_poll_tx_max;
+	struct dentry *event_count_since_irq;
+	struct dentry *events_per_irq_max;
+	struct dentry *irq_count;
+	struct dentry *useless_irq_count;
+	struct dentry *poll_schedule_count;
+	struct dentry *poll_call_count;
+	struct dentry *poll_reschedule_count;
+	struct dentry *queue_stops;
+	struct dentry *queue_wakes;
+	struct dentry *ssr_bursts;
+	struct dentry *ssr_drop_stream;
+	struct dentry *ssr_misorder;
+	struct dentry *ssr_slow_start;
+	struct dentry *ssr_merges;
+	struct dentry *ssr_too_many;
+	struct dentry *ssr_new_stream;
+};
+
+
+typedef struct netfront_accel_vnic {
+	struct netfront_accel_vnic *next;
+	
+	struct mutex vnic_mutex;
+
+	spinlock_t tx_lock;
+
+	struct netfront_accel_bufpages bufpages;
+	struct netfront_accel_bufinfo *rx_bufs;
+	struct netfront_accel_bufinfo *tx_bufs;
+	
+	/** Hardware & VI state */
+	ef_vi vi;
+
+	ef_vi_state *vi_state;
+
+	ef_eventq_state evq_state;
+
+	void *evq_mapping;
+
+	/** Hardware dependant state */
+	union {
+		struct {
+			/** Falcon A or B */
+			enum net_accel_hw_type type; 
+			u32 *evq_rptr;
+			u32 *doorbell;
+			void *evq_rptr_mapping;
+			void *doorbell_mapping;
+			void *txdmaq_mapping;
+			void *rxdmaq_mapping;
+		} falcon;
+	} hw;
+  
+	/** RX DMA queue status */
+	u32 rx_dma_level;
+
+	/** Number of RX descriptors waiting to be pushed to the card. */
+	u32 rx_dma_batched;
+#define NETFRONT_ACCEL_RX_DESC_BATCH 16
+
+	/**
+	 * Hash table of remote mac addresses to decide whether to try
+	 * fast path
+	 */
+	cuckoo_hash_table fastpath_table;
+	spinlock_t table_lock;
+
+	/** the local mac address of virtual interface we're accelerating */
+	u8 mac[ETH_ALEN];
+
+	int rx_pkt_stride;
+	int rx_skb_stride;
+
+	/**
+	 * Keep track of fragments of jumbo packets as events are
+	 * delivered by NIC 
+	 */
+	netfront_accel_jumbo_state jumbo_state;
+
+	struct net_device *net_dev;
+
+	/** These two gate the enabling of fast path operations */
+	int frontend_ready;
+	int backend_netdev_up;
+
+	int irq_enabled;
+	spinlock_t irq_enabled_lock;
+
+	int tx_enabled;
+
+	int poll_enabled;
+
+	/** A spare slot for a TX packet.  This is treated as an extension
+	 * of the DMA queue. */
+	struct sk_buff *tx_skb;
+
+	/** Keep track of fragments of SSR packets */
+	struct netfront_accel_ssr_state ssr_state;
+
+	struct xenbus_device *dev;
+
+	/** Event channel for messages */
+	int msg_channel;
+	int msg_channel_irq;
+
+	/** Event channel for network interrupts. */
+	int net_channel;
+	int net_channel_irq;
+
+	struct net_accel_shared_page *shared_page;
+
+	grant_ref_t ctrl_page_gnt;
+	grant_ref_t msg_page_gnt;
+
+	/** Message Qs, 1 each way. */
+	sh_msg_fifo2 to_dom0;
+	sh_msg_fifo2 from_dom0;
+
+	enum netfront_accel_msg_state msg_state;
+
+	/** Watch on accelstate */
+	struct xenbus_watch backend_accel_watch;
+	/** Watch on frontend's MAC address */
+	struct xenbus_watch mac_address_watch;
+
+	/** Work to process received irq/msg */
+	struct work_struct msg_from_bend;
+
+	/** Wait queue for changes in accelstate. */
+	wait_queue_head_t state_wait_queue;
+
+	/** The current accelstate of this driver. */
+	XenbusState frontend_state;
+
+	/** The most recent accelstate seen by the xenbus watch. */
+	XenbusState backend_state;
+
+	/** Non-zero if we should reject requests to connect. */
+	int removing;
+
+	/** Non-zero if the domU shared state has been initialised. */
+	int domU_state_is_setup;
+
+	/** Non-zero if the dom0 shared state has been initialised. */
+	int dom0_state_is_setup;
+
+	/* Those statistics that are added to the netdev stats */
+	struct netfront_accel_netdev_stats netdev_stats;
+	struct netfront_accel_netdev_stats stats_last_read;
+#ifdef CONFIG_DEBUG_FS
+	struct netfront_accel_netdev_dbfs netdev_dbfs;
+#endif
+
+	/* These statistics are internal and optional */
+#if NETFRONT_ACCEL_STATS
+	struct netfront_accel_stats stats;
+#ifdef CONFIG_DEBUG_FS
+	struct netfront_accel_dbfs dbfs;
+#endif
+#endif
+
+	/** Debufs fs dir for this interface */
+	struct dentry *dbfs_dir;
+} netfront_accel_vnic;
+
+
+/* Module parameters */
+extern unsigned sfc_netfront_max_pages;
+extern unsigned sfc_netfront_buffer_split;
+
+extern const char *frontend_name;
+extern struct netfront_accel_hooks accel_hooks;
+extern struct workqueue_struct *netfront_accel_workqueue;
+
+
+extern
+void netfront_accel_vi_ctor(netfront_accel_vnic *vnic);
+
+extern
+int netfront_accel_vi_init(netfront_accel_vnic *vnic, 
+			   struct net_accel_msg_hw *hw_msg);
+
+extern
+void netfront_accel_vi_dtor(netfront_accel_vnic *vnic);
+
+
+/**
+ * Add new buffers which have been registered with the NIC.
+ *
+ * @v   vnic     The vnic instance to process the response.
+ *
+ * The buffers contained in the message are added to the buffer pool.
+ */
+extern
+void netfront_accel_vi_add_bufs(netfront_accel_vnic *vnic, int is_rx);
+
+/**
+ * Put a packet on the tx DMA queue.
+ *
+ * @v  vnic	 The vnic instance to accept the packet.
+ * @v  skb	 A sk_buff to send.
+ *
+ * Attempt to send a packet.  On success, the skb is owned by the DMA
+ * queue and will be released when the completion event arrives.
+ */
+extern enum netfront_accel_post_status
+netfront_accel_vi_tx_post(netfront_accel_vnic *vnic,
+			  struct sk_buff *skb);
+
+
+/**
+ * Process events in response to an interrupt.
+ *
+ * @v   vnic       The vnic instance to poll.
+ * @v   rx_packets The maximum number of rx packets to process.
+ * @ret rx_done    The number of rx packets processed.
+ *
+ * The vnic will process events until there are no more events
+ * remaining or the specified number of rx packets has been processed.
+ * The split from the interrupt call is to allow Linux NAPI
+ * polling.
+ */
+extern
+int netfront_accel_vi_poll(netfront_accel_vnic *vnic, int rx_packets);
+
+
+/**
+ * Iterate over the fragments of a packet buffer.
+ *
+ * @v   skb      The packet buffer to examine.
+ * @v   idx      A variable name for the fragment index.
+ * @v   data     A variable name for the address of the fragment data.
+ * @v   length   A variable name for the fragment length.
+ * @v   code     A section of code to execute for each fragment.
+ *
+ * This macro iterates over the fragments in a packet buffer and
+ * executes the code for each of them.
+ */
+#define NETFRONT_ACCEL_PKTBUFF_FOR_EACH_FRAGMENT(skb, frag_idx,		\
+						 frag_data, frag_len,   \
+						 code)			\
+	do {								\
+		int frag_idx;						\
+		void *frag_data;					\
+		unsigned int	  frag_len;				\
+									\
+		frag_data = skb->data;					\
+		frag_len = skb_headlen(skb);				\
+		frag_idx = 0;						\
+		while (1) { /* For each fragment */			\
+			code;						\
+			if (frag_idx >= skb_shinfo(skb)->nr_frags) {	\
+				break;					\
+			} else {					\
+				skb_frag_t *fragment;			\
+				fragment = &skb_shinfo(skb)->frags[frag_idx]; \
+				frag_len = fragment->size;		\
+				frag_data = ((void*)page_address(fragment->page) \
+					     + fragment->page_offset);	\
+			};						\
+			frag_idx++;					\
+		}							\
+	} while(0)
+
+static inline
+void netfront_accel_disable_net_interrupts(netfront_accel_vnic *vnic)
+{
+	mask_evtchn(vnic->net_channel);
+}
+
+static inline
+void netfront_accel_enable_net_interrupts(netfront_accel_vnic *vnic)
+{
+	unmask_evtchn(vnic->net_channel);
+}
+
+void netfront_accel_msg_tx_fastpath(netfront_accel_vnic *vnic, const void *mac,
+				    u32 ip, u16 port, u8 protocol);
+
+/* Process an IRQ received from back end driver */
+irqreturn_t netfront_accel_msg_channel_irq_from_bend(int irq, void *context, 
+						     struct pt_regs *unused);
+irqreturn_t netfront_accel_net_channel_irq_from_bend(int irq, void *context, 
+						     struct pt_regs *unused);
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20)
+extern void netfront_accel_msg_from_bend(struct work_struct *context);
+#else
+extern void netfront_accel_msg_from_bend(void *context);
+#endif
+
+extern void vnic_stop_fastpath(netfront_accel_vnic *vnic);
+
+extern int netfront_accel_probe(struct net_device *net_dev, 
+				struct xenbus_device *dev);
+extern int netfront_accel_remove(struct xenbus_device *dev);
+extern void netfront_accel_set_closing(netfront_accel_vnic *vnic);
+
+extern int netfront_accel_vi_enable_interrupts(netfront_accel_vnic *vnic);
+
+extern void netfront_accel_debugfs_init(void);
+extern void netfront_accel_debugfs_fini(void);
+extern int netfront_accel_debugfs_create(netfront_accel_vnic *vnic);
+extern int netfront_accel_debugfs_remove(netfront_accel_vnic *vnic);
+
+#endif /* NETFRONT_ACCEL_H */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/accel_msg.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_msg.c
--- linux-2.6.18.8/drivers/xen/sfc_netfront/accel_msg.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_msg.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,566 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/stddef.h>
+#include <linux/errno.h>
+
+#include <xen/xenbus.h>
+
+#include "accel.h"
+#include "accel_msg_iface.h"
+#include "accel_util.h"
+#include "accel_bufs.h"
+
+#include "netfront.h" /* drivers/xen/netfront/netfront.h */
+
+static void vnic_start_interrupts(netfront_accel_vnic *vnic)
+{
+	unsigned long flags;
+	
+	/* Prime our interrupt */
+	spin_lock_irqsave(&vnic->irq_enabled_lock, flags);
+	if (!netfront_accel_vi_enable_interrupts(vnic)) {
+		/* Cripes, that was quick, better pass it up */
+		netfront_accel_disable_net_interrupts(vnic);
+		vnic->irq_enabled = 0;
+		NETFRONT_ACCEL_STATS_OP(vnic->stats.poll_schedule_count++);
+		netif_rx_schedule(vnic->net_dev);
+	} else {
+		/*
+		 * Nothing yet, make sure we get interrupts through
+		 * back end 
+		 */
+		vnic->irq_enabled = 1;
+		netfront_accel_enable_net_interrupts(vnic);
+	}
+	spin_unlock_irqrestore(&vnic->irq_enabled_lock, flags);
+}
+
+
+static void vnic_stop_interrupts(netfront_accel_vnic *vnic)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&vnic->irq_enabled_lock, flags);
+	netfront_accel_disable_net_interrupts(vnic);
+	vnic->irq_enabled = 0;
+	spin_unlock_irqrestore(&vnic->irq_enabled_lock, flags);
+}
+
+
+static void vnic_start_fastpath(netfront_accel_vnic *vnic)
+{
+	struct net_device *net_dev = vnic->net_dev;
+	unsigned long flags;
+
+	DPRINTK("%s\n", __FUNCTION__);
+
+	spin_lock_irqsave(&vnic->tx_lock, flags);
+	vnic->tx_enabled = 1;
+	spin_unlock_irqrestore(&vnic->tx_lock, flags);
+	
+	netif_poll_disable(net_dev);
+	vnic->poll_enabled = 1;
+	netif_poll_enable(net_dev);
+	
+	vnic_start_interrupts(vnic);
+}
+
+
+void vnic_stop_fastpath(netfront_accel_vnic *vnic)
+{
+	struct net_device *net_dev = vnic->net_dev;
+	struct netfront_info *np = (struct netfront_info *)netdev_priv(net_dev);
+	unsigned long flags1, flags2;
+
+	DPRINTK("%s\n", __FUNCTION__);
+
+	vnic_stop_interrupts(vnic);
+	
+	spin_lock_irqsave(&vnic->tx_lock, flags1);
+	vnic->tx_enabled = 0;
+	spin_lock_irqsave(&np->tx_lock, flags2);
+	if (vnic->tx_skb != NULL) {
+		dev_kfree_skb_any(vnic->tx_skb);
+		vnic->tx_skb = NULL;
+		if (netfront_check_queue_ready(net_dev)) {
+			netif_wake_queue(net_dev);
+			NETFRONT_ACCEL_STATS_OP
+				(vnic->stats.queue_wakes++);
+		}
+	}
+	spin_unlock_irqrestore(&np->tx_lock, flags2);
+	spin_unlock_irqrestore(&vnic->tx_lock, flags1);
+	
+	/* Must prevent polls and hold lock to modify poll_enabled */
+	netif_poll_disable(net_dev);
+	spin_lock_irqsave(&vnic->irq_enabled_lock, flags1);
+	vnic->poll_enabled = 0;
+	spin_unlock_irqrestore(&vnic->irq_enabled_lock, flags1);
+	netif_poll_enable(net_dev);
+}
+
+
+static void netfront_accel_interface_up(netfront_accel_vnic *vnic)
+{
+
+	if (!vnic->backend_netdev_up) {
+		vnic->backend_netdev_up = 1;
+		
+		if (vnic->frontend_ready)
+			vnic_start_fastpath(vnic);
+	}
+}
+
+
+static void netfront_accel_interface_down(netfront_accel_vnic *vnic)
+{
+
+	if (vnic->backend_netdev_up) {
+		vnic->backend_netdev_up = 0;
+		
+		if (vnic->frontend_ready)
+			vnic_stop_fastpath(vnic);
+	}
+}
+
+
+static int vnic_add_bufs(netfront_accel_vnic *vnic, 
+			 struct net_accel_msg *msg)
+{
+	int rc, offset;
+	struct netfront_accel_bufinfo *bufinfo;
+  
+	BUG_ON(msg->u.mapbufs.pages > NET_ACCEL_MSG_MAX_PAGE_REQ);
+
+	offset = msg->u.mapbufs.reqid;
+
+	if (offset < vnic->bufpages.max_pages - 
+	    (vnic->bufpages.max_pages / sfc_netfront_buffer_split)) {
+		bufinfo = vnic->rx_bufs;
+	} else
+		bufinfo = vnic->tx_bufs;
+
+	/* Queue up some Rx buffers to start things off. */
+	if ((rc = netfront_accel_add_bufs(&vnic->bufpages, bufinfo, msg)) == 0) {
+		netfront_accel_vi_add_bufs(vnic, bufinfo == vnic->rx_bufs);
+
+		if (offset + msg->u.mapbufs.pages == vnic->bufpages.max_pages) {
+			VPRINTK("%s: got all buffers back\n", __FUNCTION__);
+			vnic->frontend_ready = 1;
+			if (vnic->backend_netdev_up)
+				vnic_start_fastpath(vnic);
+		} else {
+			VPRINTK("%s: got buffers back %d %d\n", __FUNCTION__, 
+				offset, msg->u.mapbufs.pages);
+		}
+	}
+
+	return rc;
+}
+
+
+/* The largest [o] such that (1u << o) <= n.  Requires n > 0. */
+
+inline unsigned log2_le(unsigned long n) {
+	unsigned order = 1;
+	while ((1ul << order) <= n) ++order;
+	return (order - 1);
+}
+
+static int vnic_send_buffer_requests(netfront_accel_vnic *vnic,
+				     struct netfront_accel_bufpages *bufpages)
+{
+	int pages, offset, rc = 0, sent = 0;
+	struct net_accel_msg msg;
+
+	while (bufpages->page_reqs < bufpages->max_pages) {
+		offset = bufpages->page_reqs;
+
+		pages = pow2(log2_le(bufpages->max_pages - 
+				     bufpages->page_reqs));
+		pages = pages < NET_ACCEL_MSG_MAX_PAGE_REQ ? 
+			pages : NET_ACCEL_MSG_MAX_PAGE_REQ;
+
+		BUG_ON(offset < 0);
+		BUG_ON(pages <= 0);
+
+		rc = netfront_accel_buf_map_request(vnic->dev, bufpages,
+						    &msg, pages, offset);
+		if (rc == 0) {
+			rc = net_accel_msg_send(vnic->shared_page, 
+						&vnic->to_dom0, &msg);
+			if (rc < 0) {
+				VPRINTK("%s: queue full, stopping for now\n",
+					__FUNCTION__);
+				break;
+			}
+			sent++;
+		} else {
+			EPRINTK("%s: problem with grant, stopping for now\n",
+				__FUNCTION__);
+			break;
+		}
+
+		bufpages->page_reqs += pages;
+	}
+
+	if (sent)
+		net_accel_msg_notify(vnic->msg_channel_irq);
+
+	return rc;
+}
+
+
+/*
+ * In response to dom0 saying "my queue is full", we reply with this
+ * when it is no longer full
+ */
+inline void vnic_set_queue_not_full(netfront_accel_vnic *vnic)
+{
+
+	if (test_and_set_bit(NET_ACCEL_MSG_AFLAGS_QUEUE0NOTFULL_B,
+			    (unsigned long *)&vnic->shared_page->aflags))
+		notify_remote_via_irq(vnic->msg_channel_irq);
+	else
+		VPRINTK("queue not full bit already set, not signalling\n");
+}
+
+/* 
+ * Notify dom0 that the queue we want to use is full, it should
+ * respond by setting MSG_AFLAGS_QUEUEUNOTFULL in due course
+ */
+inline void vnic_set_queue_full(netfront_accel_vnic *vnic)
+{
+
+	if (!test_and_set_bit(NET_ACCEL_MSG_AFLAGS_QUEUEUFULL_B,
+			     (unsigned long *)&vnic->shared_page->aflags))
+		notify_remote_via_irq(vnic->msg_channel_irq);
+	else
+		VPRINTK("queue full bit already set, not signalling\n");
+}
+
+
+static int vnic_check_hello_version(unsigned version) 
+{
+	if (version > NET_ACCEL_MSG_VERSION) {
+		/* Newer protocol, we must refuse */
+		return -EPROTO;
+	}
+
+	if (version < NET_ACCEL_MSG_VERSION) {
+		/*
+		 * We are newer, so have discretion to accept if we
+		 * wish.  For now however, just reject
+		 */
+		return -EPROTO;
+	}
+
+	BUG_ON(version != NET_ACCEL_MSG_VERSION);
+	return 0;
+}
+
+
+static int vnic_process_hello_msg(netfront_accel_vnic *vnic,
+				  struct net_accel_msg *msg)
+{
+	int err = 0;
+	unsigned pages = sfc_netfront_max_pages;
+
+	if (vnic_check_hello_version(msg->u.hello.version) < 0) {
+		msg->id = NET_ACCEL_MSG_HELLO | NET_ACCEL_MSG_REPLY 
+			| NET_ACCEL_MSG_ERROR;
+		msg->u.hello.version = NET_ACCEL_MSG_VERSION;
+	} else {
+		vnic->backend_netdev_up
+			= vnic->shared_page->net_dev_up;
+		
+		msg->id = NET_ACCEL_MSG_HELLO | NET_ACCEL_MSG_REPLY;
+		msg->u.hello.version = NET_ACCEL_MSG_VERSION;
+		if (msg->u.hello.max_pages &&
+		    msg->u.hello.max_pages < pages)
+			pages = msg->u.hello.max_pages;
+		msg->u.hello.max_pages = pages;
+		
+		/* Half of pages for rx, half for tx */ 
+		err = netfront_accel_alloc_buffer_mem(&vnic->bufpages,
+						      vnic->rx_bufs, 
+						      vnic->tx_bufs,
+						      pages);
+		if (err)
+			msg->id |= NET_ACCEL_MSG_ERROR;		
+	}
+	
+	/* Send reply */
+	net_accel_msg_reply_notify(vnic->shared_page, vnic->msg_channel_irq,
+				   &vnic->to_dom0, msg);
+	return err;
+}
+
+
+static int vnic_process_localmac_msg(netfront_accel_vnic *vnic,
+				     struct net_accel_msg *msg)
+{
+	unsigned long flags;
+	cuckoo_hash_mac_key key;
+
+	if (msg->u.localmac.flags & NET_ACCEL_MSG_ADD) {
+		DPRINTK("MAC has moved, could be local: " MAC_FMT "\n",
+			MAC_ARG(msg->u.localmac.mac));
+		key = cuckoo_mac_to_key(msg->u.localmac.mac);
+		spin_lock_irqsave(&vnic->table_lock, flags);
+		/* Try to remove it, not a big deal if not there */
+		cuckoo_hash_remove(&vnic->fastpath_table, 
+				   (cuckoo_hash_key *)&key);
+		spin_unlock_irqrestore(&vnic->table_lock, flags);
+	}
+	
+	return 0;
+}
+
+
+static 
+int vnic_process_rx_msg(netfront_accel_vnic *vnic,
+			struct net_accel_msg *msg)
+{
+	int err;
+
+	switch (msg->id) {
+	case NET_ACCEL_MSG_HELLO:
+		/* Hello, reply with Reply */
+		DPRINTK("got Hello, with version %.8x\n",
+			msg->u.hello.version);
+		BUG_ON(vnic->msg_state != NETFRONT_ACCEL_MSG_NONE);
+		err = vnic_process_hello_msg(vnic, msg);
+		if (err == 0)
+			vnic->msg_state = NETFRONT_ACCEL_MSG_HELLO;
+		break;
+	case NET_ACCEL_MSG_SETHW:
+		/* Hardware info message */
+		DPRINTK("got H/W info\n");
+		BUG_ON(vnic->msg_state != NETFRONT_ACCEL_MSG_HELLO);
+		err = netfront_accel_vi_init(vnic, &msg->u.hw);
+		if (err == 0)
+			vnic->msg_state = NETFRONT_ACCEL_MSG_HW;
+		break;
+	case NET_ACCEL_MSG_MAPBUF | NET_ACCEL_MSG_REPLY:
+		VPRINTK("Got mapped buffers back\n");
+		BUG_ON(vnic->msg_state != NETFRONT_ACCEL_MSG_HW);
+		err = vnic_add_bufs(vnic, msg);
+		break;
+	case NET_ACCEL_MSG_MAPBUF | NET_ACCEL_MSG_REPLY | NET_ACCEL_MSG_ERROR:
+		/* No buffers.  Can't use the fast path. */
+		EPRINTK("Got mapped buffers error.  Cannot accelerate.\n");
+		BUG_ON(vnic->msg_state != NETFRONT_ACCEL_MSG_HW);
+		err = -EIO;
+		break;
+	case NET_ACCEL_MSG_LOCALMAC:
+		/* Should be add, remove not currently used */
+		EPRINTK_ON(!(msg->u.localmac.flags & NET_ACCEL_MSG_ADD));
+		BUG_ON(vnic->msg_state != NETFRONT_ACCEL_MSG_HW);
+		err = vnic_process_localmac_msg(vnic, msg);
+		break;
+	default:
+		EPRINTK("Huh? Message code is 0x%x\n", msg->id);
+		err = -EPROTO;
+		break;
+	}
+
+	return err;
+}
+
+
+/* Process an IRQ received from back end driver */
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20)
+void netfront_accel_msg_from_bend(struct work_struct *context)
+#else
+void netfront_accel_msg_from_bend(void *context)
+#endif
+{
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20)
+	netfront_accel_vnic *vnic = 
+		container_of(context, netfront_accel_vnic, msg_from_bend);
+#else
+	netfront_accel_vnic *vnic = (netfront_accel_vnic *)context;
+#endif
+	struct net_accel_msg msg;
+	int err, queue_was_full = 0;
+	
+	mutex_lock(&vnic->vnic_mutex);
+
+	/*
+	 * This happens when the shared pages have been unmapped but
+	 * the workqueue has yet to be flushed 
+	 */
+	if (!vnic->dom0_state_is_setup) 
+		goto unlock_out;
+
+	while ((vnic->shared_page->aflags & NET_ACCEL_MSG_AFLAGS_TO_DOMU_MASK)
+	       != 0) {
+		if (vnic->shared_page->aflags &
+		    NET_ACCEL_MSG_AFLAGS_QUEUEUNOTFULL) {
+			/* We've been told there may now be space. */
+			clear_bit(NET_ACCEL_MSG_AFLAGS_QUEUEUNOTFULL_B,
+				  (unsigned long *)&vnic->shared_page->aflags);
+		}
+
+		if (vnic->shared_page->aflags &
+		    NET_ACCEL_MSG_AFLAGS_QUEUE0FULL) {
+			/*
+			 * There will be space at the end of this
+			 * function if we can make any.
+			 */
+			clear_bit(NET_ACCEL_MSG_AFLAGS_QUEUE0FULL_B,
+				  (unsigned long *)&vnic->shared_page->aflags);
+			queue_was_full = 1;
+		}
+
+		if (vnic->shared_page->aflags &
+		    NET_ACCEL_MSG_AFLAGS_NETUPDOWN) {
+			DPRINTK("%s: net interface change\n", __FUNCTION__);
+			clear_bit(NET_ACCEL_MSG_AFLAGS_NETUPDOWN_B,
+				  (unsigned long *)&vnic->shared_page->aflags);
+			if (vnic->shared_page->net_dev_up)
+				netfront_accel_interface_up(vnic);
+			else
+				netfront_accel_interface_down(vnic);
+		}
+	}
+
+	/* Pull msg out of shared memory */
+	while ((err = net_accel_msg_recv(vnic->shared_page, &vnic->from_dom0,
+					 &msg)) == 0) {
+		err = vnic_process_rx_msg(vnic, &msg);
+		
+		if (err != 0)
+			goto done;
+	}
+
+	/*
+	 * Send any pending buffer map request messages that we can,
+	 * and mark domU->dom0 as full if necessary.  
+	 */
+	if (vnic->msg_state == NETFRONT_ACCEL_MSG_HW &&
+	    vnic->bufpages.page_reqs < vnic->bufpages.max_pages) {
+		if (vnic_send_buffer_requests(vnic, &vnic->bufpages) == -ENOSPC)
+			vnic_set_queue_full(vnic);
+	}
+
+	/* 
+	 * If there are no messages then this is not an error.  It
+	 * just means that we've finished processing the queue.
+	 */
+	if (err == -ENOENT)
+		err = 0;
+ done:
+	/* We will now have made space in the dom0->domU queue if we can */
+	if (queue_was_full)
+		vnic_set_queue_not_full(vnic);
+
+	if (err != 0) {
+		EPRINTK("%s returned %d\n", __FUNCTION__, err);
+		netfront_accel_set_closing(vnic);
+	}
+
+ unlock_out:
+	mutex_unlock(&vnic->vnic_mutex);
+
+	return;
+}
+
+
+irqreturn_t netfront_accel_msg_channel_irq_from_bend(int irq, void *context, 
+						 struct pt_regs *unused)
+{
+	netfront_accel_vnic *vnic = (netfront_accel_vnic *)context;
+	VPRINTK("irq %d from device %s\n", irq, vnic->dev->nodename);
+
+	queue_work(netfront_accel_workqueue, &vnic->msg_from_bend);
+
+	return IRQ_HANDLED;
+}
+
+/* Process an interrupt received from the NIC via backend */
+irqreturn_t netfront_accel_net_channel_irq_from_bend(int irq, void *context, 
+						     struct pt_regs *unused)
+{
+	netfront_accel_vnic *vnic = (netfront_accel_vnic *)context;
+	struct net_device *net_dev = vnic->net_dev;
+	unsigned long flags;
+
+	VPRINTK("net irq %d from device %s\n", irq, vnic->dev->nodename);
+	
+	NETFRONT_ACCEL_STATS_OP(vnic->stats.irq_count++);
+
+	BUG_ON(net_dev==NULL);
+
+	spin_lock_irqsave(&vnic->irq_enabled_lock, flags);
+	if (vnic->irq_enabled) {
+		netfront_accel_disable_net_interrupts(vnic);
+		vnic->irq_enabled = 0;
+		spin_unlock_irqrestore(&vnic->irq_enabled_lock, flags);
+
+#if NETFRONT_ACCEL_STATS
+		vnic->stats.poll_schedule_count++;
+		if (vnic->stats.event_count_since_irq >
+		    vnic->stats.events_per_irq_max)
+			vnic->stats.events_per_irq_max = 
+				vnic->stats.event_count_since_irq;
+		vnic->stats.event_count_since_irq = 0;
+#endif
+		netif_rx_schedule(net_dev);
+	}
+	else {
+		spin_unlock_irqrestore(&vnic->irq_enabled_lock, flags);
+		NETFRONT_ACCEL_STATS_OP(vnic->stats.useless_irq_count++);
+		DPRINTK("%s: irq when disabled\n", __FUNCTION__);
+	}
+	
+	return IRQ_HANDLED;
+}
+
+
+void netfront_accel_msg_tx_fastpath(netfront_accel_vnic *vnic, const void *mac,
+				    u32 ip, u16 port, u8 protocol)
+{
+	unsigned long lock_state;
+	struct net_accel_msg *msg;
+
+	msg = net_accel_msg_start_send(vnic->shared_page, &vnic->to_dom0,
+				       &lock_state);
+
+	if (msg == NULL)
+		return;
+
+	net_accel_msg_init(msg, NET_ACCEL_MSG_FASTPATH);
+	msg->u.fastpath.flags = NET_ACCEL_MSG_REMOVE;
+	memcpy(msg->u.fastpath.mac, mac, ETH_ALEN);
+
+	msg->u.fastpath.port = port;
+	msg->u.fastpath.ip = ip;
+	msg->u.fastpath.proto = protocol;
+
+	net_accel_msg_complete_send_notify(vnic->shared_page, &vnic->to_dom0, 
+					   &lock_state, vnic->msg_channel_irq);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/accel_netfront.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_netfront.c
--- linux-2.6.18.8/drivers/xen/sfc_netfront/accel_netfront.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_netfront.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,319 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/skbuff.h>
+#include <linux/netdevice.h>
+
+/* drivers/xen/netfront/netfront.h */
+#include "netfront.h"
+
+#include "accel.h"
+#include "accel_bufs.h"
+#include "accel_util.h"
+#include "accel_msg_iface.h"
+#include "accel_ssr.h"
+ 
+#ifdef EFX_GCOV
+#include "gcov.h"
+#endif
+
+#define NETFRONT_ACCEL_VNIC_FROM_NETDEV(_nd)				\
+	((netfront_accel_vnic *)((struct netfront_info *)netdev_priv(net_dev))->accel_priv)
+
+static int netfront_accel_netdev_start_xmit(struct sk_buff *skb,
+					    struct net_device *net_dev)
+{
+	netfront_accel_vnic *vnic = NETFRONT_ACCEL_VNIC_FROM_NETDEV(net_dev);
+	struct netfront_info *np = 
+		(struct netfront_info *)netdev_priv(net_dev);
+	int handled, rc;
+	unsigned long flags1, flags2;
+
+	BUG_ON(vnic == NULL);
+
+	/* Take our tx lock and hold for the duration */
+	spin_lock_irqsave(&vnic->tx_lock, flags1);
+
+	if (!vnic->tx_enabled) {
+		rc = 0;
+		goto unlock_out;
+	}
+
+	handled = netfront_accel_vi_tx_post(vnic, skb);
+	if (handled == NETFRONT_ACCEL_STATUS_BUSY) {
+		BUG_ON(vnic->net_dev != net_dev);
+		DPRINTK("%s stopping queue\n", __FUNCTION__);
+
+		/* Netfront's lock protects tx_skb */
+		spin_lock_irqsave(&np->tx_lock, flags2);
+		BUG_ON(vnic->tx_skb != NULL);
+		vnic->tx_skb = skb;
+		netif_stop_queue(net_dev);
+		spin_unlock_irqrestore(&np->tx_lock, flags2);
+
+		NETFRONT_ACCEL_STATS_OP(vnic->stats.queue_stops++);
+	}
+
+	if (handled == NETFRONT_ACCEL_STATUS_CANT)
+		rc = 0;
+	else
+		rc = 1;
+
+unlock_out:
+	spin_unlock_irqrestore(&vnic->tx_lock, flags1);
+
+	return rc;
+}
+
+
+static int netfront_accel_netdev_poll(struct net_device *net_dev, int *budget)
+{
+	netfront_accel_vnic *vnic = NETFRONT_ACCEL_VNIC_FROM_NETDEV(net_dev);
+	int rx_allowed = *budget, rx_done;
+	
+	BUG_ON(vnic == NULL);
+
+	/* Can check this without lock as modifier excludes polls */ 
+	if (!vnic->poll_enabled)
+		return 0;
+
+	rx_done = netfront_accel_vi_poll(vnic, rx_allowed);
+	*budget -= rx_done;
+	
+	NETFRONT_ACCEL_STATS_OP(vnic->stats.poll_call_count++);
+
+	VPRINTK("%s: done %d allowed %d\n",
+		__FUNCTION__, rx_done, rx_allowed);
+
+	netfront_accel_ssr_end_of_burst(vnic, &vnic->ssr_state);
+
+	if (rx_done < rx_allowed) {
+		 return 0; /* Done */
+	}
+	
+	NETFRONT_ACCEL_STATS_OP(vnic->stats.poll_reschedule_count++);
+
+	return 1; /* More to do. */
+}
+
+
+/*
+ * Process request from netfront to start napi interrupt
+ * mode. (i.e. enable interrupts as it's finished polling)
+ */
+static int netfront_accel_start_napi_interrupts(struct net_device *net_dev) 
+{
+	netfront_accel_vnic *vnic = NETFRONT_ACCEL_VNIC_FROM_NETDEV(net_dev);
+	unsigned long flags;
+
+	BUG_ON(vnic == NULL);
+	
+	/*
+	 * Can check this without lock as writer excludes poll before
+	 * modifying
+	 */
+	if (!vnic->poll_enabled)
+		return 0;
+
+	if (!netfront_accel_vi_enable_interrupts(vnic)) {
+		/* 
+		 * There was something there, tell caller we had
+		 * something to do.
+		 */
+		return 1;
+	}
+
+	spin_lock_irqsave(&vnic->irq_enabled_lock, flags);
+	vnic->irq_enabled = 1;
+	netfront_accel_enable_net_interrupts(vnic);
+	spin_unlock_irqrestore(&vnic->irq_enabled_lock, flags);
+
+	return 0;
+}
+
+
+/*
+ * Process request from netfront to stop napi interrupt
+ * mode. (i.e. disable interrupts as it's starting to poll 
+ */
+static void netfront_accel_stop_napi_interrupts(struct net_device *net_dev) 
+{
+	netfront_accel_vnic *vnic = NETFRONT_ACCEL_VNIC_FROM_NETDEV(net_dev);
+	unsigned long flags;
+
+	BUG_ON(vnic == NULL);
+
+	spin_lock_irqsave(&vnic->irq_enabled_lock, flags);
+
+	if (!vnic->poll_enabled) {
+		spin_unlock_irqrestore(&vnic->irq_enabled_lock, flags);
+		return;
+	}
+
+	netfront_accel_disable_net_interrupts(vnic);
+	vnic->irq_enabled = 0;
+	spin_unlock_irqrestore(&vnic->irq_enabled_lock, flags);
+}
+
+
+static int netfront_accel_check_ready(struct net_device *net_dev)
+{
+	netfront_accel_vnic *vnic = NETFRONT_ACCEL_VNIC_FROM_NETDEV(net_dev);
+
+	BUG_ON(vnic == NULL);
+
+	/* This is protected by netfront's lock */ 
+	return vnic->tx_skb == NULL;
+}
+
+
+static int netfront_accel_get_stats(struct net_device *net_dev,
+				    struct net_device_stats *stats)
+{
+	netfront_accel_vnic *vnic = NETFRONT_ACCEL_VNIC_FROM_NETDEV(net_dev);
+	struct netfront_accel_netdev_stats now;
+
+	BUG_ON(vnic == NULL);
+
+	now.fastpath_rx_pkts   = vnic->netdev_stats.fastpath_rx_pkts;
+	now.fastpath_rx_bytes  = vnic->netdev_stats.fastpath_rx_bytes;
+	now.fastpath_rx_errors = vnic->netdev_stats.fastpath_rx_errors;
+	now.fastpath_tx_pkts   = vnic->netdev_stats.fastpath_tx_pkts;
+	now.fastpath_tx_bytes  = vnic->netdev_stats.fastpath_tx_bytes;
+	now.fastpath_tx_errors = vnic->netdev_stats.fastpath_tx_errors;
+	
+	stats->rx_packets += (now.fastpath_rx_pkts - 
+			      vnic->stats_last_read.fastpath_rx_pkts);
+	stats->rx_bytes   += (now.fastpath_rx_bytes -
+			      vnic->stats_last_read.fastpath_rx_bytes);
+	stats->rx_errors  += (now.fastpath_rx_errors - 
+			      vnic->stats_last_read.fastpath_rx_errors);
+	stats->tx_packets += (now.fastpath_tx_pkts - 
+			      vnic->stats_last_read.fastpath_tx_pkts);
+	stats->tx_bytes   += (now.fastpath_tx_bytes - 
+			      vnic->stats_last_read.fastpath_tx_bytes);
+	stats->tx_errors  += (now.fastpath_tx_errors - 
+			      vnic->stats_last_read.fastpath_tx_errors);
+	
+	vnic->stats_last_read = now;
+
+	return 0;
+}
+
+
+struct netfront_accel_hooks accel_hooks = {
+	.new_device	    = &netfront_accel_probe,
+	.remove		= &netfront_accel_remove,
+	.netdev_poll	   = &netfront_accel_netdev_poll,
+	.start_xmit	    = &netfront_accel_netdev_start_xmit,
+	.start_napi_irq	= &netfront_accel_start_napi_interrupts,
+	.stop_napi_irq	 = &netfront_accel_stop_napi_interrupts,
+	.check_ready	   = &netfront_accel_check_ready,
+	.get_stats	     = &netfront_accel_get_stats
+};
+
+
+unsigned sfc_netfront_max_pages = NETFRONT_ACCEL_DEFAULT_BUF_PAGES;
+module_param_named (max_pages, sfc_netfront_max_pages, uint, 0644);
+MODULE_PARM_DESC(max_pages, "Number of buffer pages to request");
+
+unsigned sfc_netfront_buffer_split = 2;
+module_param_named (buffer_split, sfc_netfront_buffer_split, uint, 0644);
+MODULE_PARM_DESC(buffer_split, 
+		 "Fraction of buffers to use for TX, rest for RX");
+
+
+const char *frontend_name = "sfc_netfront";
+
+struct workqueue_struct *netfront_accel_workqueue;
+
+static int __init netfront_accel_init(void)
+{
+	int rc;
+#ifdef EFX_GCOV	
+	gcov_provider_init(THIS_MODULE);
+#endif
+
+	/*
+	 * If we're running on dom0, netfront hasn't initialised
+	 * itself, so we need to keep away
+	 */
+	if (is_initial_xendomain())
+		return 0;
+
+	if (!is_pow2(sizeof(struct net_accel_msg)))
+		EPRINTK("%s: bad structure size\n", __FUNCTION__);
+
+	netfront_accel_workqueue = create_workqueue(frontend_name);
+
+	netfront_accel_debugfs_init();
+
+	rc = netfront_accelerator_loaded(NETFRONT_ACCEL_VERSION,
+					 frontend_name, &accel_hooks);
+
+	if (rc < 0) {
+		EPRINTK("Xen netfront accelerator version mismatch\n");
+		return -EINVAL;
+	}
+
+	if (rc > 0) {
+		/* 
+		 * In future may want to add backwards compatibility
+		 * and accept certain subsets of previous versions
+		 */
+		EPRINTK("Xen netfront accelerator version mismatch\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+module_init(netfront_accel_init);
+
+static void __exit netfront_accel_exit(void)
+{
+	if (is_initial_xendomain())
+		return;
+
+	DPRINTK("%s: unhooking\n", __FUNCTION__);
+
+	/* Unhook from normal netfront */
+	netfront_accelerator_stop(frontend_name);
+
+	DPRINTK("%s: done\n", __FUNCTION__);
+
+	netfront_accel_debugfs_fini();
+
+	flush_workqueue(netfront_accel_workqueue);
+
+	destroy_workqueue(netfront_accel_workqueue);
+
+#ifdef EFX_GCOV
+ 	gcov_provider_fini(THIS_MODULE);
+#endif
+	return;
+}
+module_exit(netfront_accel_exit);
+
+MODULE_LICENSE("GPL");
+
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/accel_ssr.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_ssr.c
--- linux-2.6.18.8/drivers/xen/sfc_netfront/accel_ssr.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_ssr.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,308 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/socket.h>
+#include <linux/in.h>
+#include <linux/ip.h>
+#include <linux/tcp.h>
+#include <linux/list.h>
+#include <net/ip.h>
+#include <net/checksum.h>
+
+#include "accel.h"
+#include "accel_util.h"
+#include "accel_bufs.h"
+
+#include "accel_ssr.h"
+
+static inline int list_valid(struct list_head *lh) {
+	return(lh->next != NULL);
+}
+
+static void netfront_accel_ssr_deliver (struct netfront_accel_vnic *vnic,
+					struct netfront_accel_ssr_state *st,
+					struct netfront_accel_ssr_conn *c);
+
+/** Construct an efx_ssr_state.
+ *
+ * @v st     The SSR state (per channel per port)
+ * @v port   The port.
+ */
+void netfront_accel_ssr_init(struct netfront_accel_ssr_state *st) {
+	unsigned i;
+
+	INIT_LIST_HEAD(&st->conns);
+	INIT_LIST_HEAD(&st->free_conns);
+	for (i = 0; i < 8; ++i) {
+		struct netfront_accel_ssr_conn *c = 
+			kmalloc(sizeof(*c), GFP_KERNEL);
+		if (c == NULL)  break;
+		c->n_in_order_pkts = 0;
+		c->skb = NULL;
+		list_add(&c->link, &st->free_conns);
+	}
+
+}
+
+
+/** Destructor for an efx_ssr_state.
+ *
+ * @v st     The SSR state (per channel per port)
+ */
+void netfront_accel_ssr_fini(netfront_accel_vnic *vnic, 
+			     struct netfront_accel_ssr_state *st) {
+	struct netfront_accel_ssr_conn *c;
+
+	/* Return cleanly if efx_ssr_init() not previously called */
+	BUG_ON(list_valid(&st->conns) != list_valid(&st->free_conns));
+	if (! list_valid(&st->conns))
+		return;
+
+	while ( ! list_empty(&st->free_conns)) {
+		c = list_entry(st->free_conns.prev, 
+			       struct netfront_accel_ssr_conn, link);
+		list_del(&c->link);
+		BUG_ON(c->skb != NULL);
+		kfree(c);
+	}
+	while ( ! list_empty(&st->conns)) {
+		c = list_entry(st->conns.prev, 
+			       struct netfront_accel_ssr_conn, link);
+		list_del(&c->link);
+		if (c->skb)
+			netfront_accel_ssr_deliver(vnic, st, c);
+		kfree(c);
+	}
+}
+
+
+/** Calc IP checksum and deliver to the OS
+ *
+ * @v st     The SSR state (per channel per port)
+ * @v c	     The SSR connection state
+ */
+static void netfront_accel_ssr_deliver(netfront_accel_vnic *vnic,
+				       struct netfront_accel_ssr_state *st,
+				       struct netfront_accel_ssr_conn *c) {
+	BUG_ON(c->skb == NULL);
+
+	/*
+	 * If we've chained packets together, recalculate the IP
+	 * checksum.
+	 */
+	if (skb_shinfo(c->skb)->frag_list) {
+		NETFRONT_ACCEL_STATS_OP(++vnic->stats.ssr_bursts);
+		c->iph->check = 0;
+		c->iph->check = ip_fast_csum((unsigned char *) c->iph, 
+					     c->iph->ihl);
+	}
+
+	VPRINTK("%s: %d\n", __FUNCTION__, c->skb->len);
+
+	netif_receive_skb(c->skb); 
+	c->skb = NULL;
+}
+
+
+/** Push held skbs down into network stack.
+ *
+ * @v st       SSR state
+ *
+ * Only called if we are tracking one or more connections.
+ */
+void __netfront_accel_ssr_end_of_burst(netfront_accel_vnic *vnic, 
+				       struct netfront_accel_ssr_state *st) {
+	struct netfront_accel_ssr_conn *c;
+
+	BUG_ON(list_empty(&st->conns));
+
+	list_for_each_entry(c, &st->conns, link)
+		if (c->skb)
+			netfront_accel_ssr_deliver(vnic, st, c);
+
+	/* Time-out connections that have received no traffic for 20ms. */
+	c = list_entry(st->conns.prev, struct netfront_accel_ssr_conn,
+		       link);
+	if (jiffies - c->last_pkt_jiffies > (HZ / 50 + 1)) {
+		NETFRONT_ACCEL_STATS_OP(++vnic->stats.ssr_drop_stream);
+		list_del(&c->link);
+		list_add(&c->link, &st->free_conns);
+	}
+}
+
+
+/** Process SKB and decide whether to dispatch it to the stack now or
+ * later.
+ *
+ * @v st	 SSR state
+ * @v skb	SKB to exmaine
+ * @ret rc       0 => deliver SKB to kernel now, otherwise the SKB belongs
+ *	       us.
+ */
+int netfront_accel_ssr_skb(struct netfront_accel_vnic *vnic,
+			   struct netfront_accel_ssr_state *st,
+			   struct sk_buff *skb) {
+	int data_length, dont_merge;
+	struct netfront_accel_ssr_conn *c;
+	struct iphdr *iph;
+	struct tcphdr *th;
+	unsigned th_seq;
+
+	BUG_ON(skb_shinfo(skb)->frag_list != NULL);
+	BUG_ON(skb->next != NULL);
+
+	/* We're not interested if it isn't TCP over IPv4. */
+	iph = (struct iphdr *) skb->data;
+	if (skb->protocol != htons(ETH_P_IP) ||
+	    iph->protocol != IPPROTO_TCP) {
+		return 0;
+	}
+
+	/* Ignore segments that fail csum or are fragmented. */
+	if (unlikely((skb->ip_summed - CHECKSUM_UNNECESSARY) |
+		     (iph->frag_off & htons(IP_MF | IP_OFFSET)))) {
+		return 0;
+	}
+
+	th = (struct tcphdr*)(skb->data + iph->ihl * 4);
+	data_length = ntohs(iph->tot_len) - iph->ihl * 4 - th->doff * 4;
+	th_seq = ntohl(th->seq);
+	dont_merge = (data_length == 0) | th->urg | th->syn | th->rst;
+
+	list_for_each_entry(c, &st->conns, link) {
+		if ((c->saddr  - iph->saddr) |
+		    (c->daddr  - iph->daddr) |
+		    (c->source - th->source) |
+		    (c->dest   - th->dest  ))
+			continue;
+
+		/* Re-insert at head of list to reduce lookup time. */
+		list_del(&c->link);
+		list_add(&c->link, &st->conns);
+		c->last_pkt_jiffies = jiffies;
+
+		if (unlikely(th_seq - c->next_seq)) {
+			/* Out-of-order, so start counting again. */
+			if (c->skb)
+				netfront_accel_ssr_deliver(vnic, st, c);
+			c->n_in_order_pkts = 0;
+			c->next_seq = th_seq + data_length;
+			NETFRONT_ACCEL_STATS_OP(++vnic->stats.ssr_misorder);
+			return 0;
+		}
+		c->next_seq = th_seq + data_length;
+
+		if (++c->n_in_order_pkts < 300) {
+			/* May be in slow-start, so don't merge. */
+			NETFRONT_ACCEL_STATS_OP(++vnic->stats.ssr_slow_start);
+			return 0;
+		}
+
+		if (unlikely(dont_merge)) {
+			if (c->skb)
+				netfront_accel_ssr_deliver(vnic, st, c);
+			return 0;
+		}
+
+		if (c->skb) {
+			c->iph->tot_len = ntohs(c->iph->tot_len);
+			c->iph->tot_len += data_length;
+			c->iph->tot_len = htons(c->iph->tot_len);
+			c->th->ack_seq = th->ack_seq;
+			c->th->fin |= th->fin;
+			c->th->psh |= th->psh;
+			c->th->window = th->window;
+
+			/* Remove the headers from this skb. */
+			skb_pull(skb, skb->len - data_length);
+
+			/*
+			 * Tack the new skb onto the head skb's frag_list.
+			 * This is exactly the format that fragmented IP
+			 * datagrams are reassembled into.
+			 */
+			BUG_ON(skb->next != 0);
+			if ( ! skb_shinfo(c->skb)->frag_list)
+				skb_shinfo(c->skb)->frag_list = skb;
+			else
+				c->skb_tail->next = skb;
+			c->skb_tail = skb;
+			c->skb->len += skb->len;
+			c->skb->data_len += skb->len;
+			c->skb->truesize += skb->truesize;
+
+			NETFRONT_ACCEL_STATS_OP(++vnic->stats.ssr_merges);
+
+			/*
+			 * If the next packet might push this super-packet
+			 * over the limit for an IP packet, deliver it now.
+			 * This is slightly conservative, but close enough.
+			 */
+			if (c->skb->len + 
+			    (PAGE_SIZE / NETFRONT_ACCEL_BUFS_PER_PAGE)
+			    > 16384)
+				netfront_accel_ssr_deliver(vnic, st, c);
+
+			return 1;
+		}
+		else {
+			c->iph = iph;
+			c->th = th;
+			c->skb = skb;
+			return 1;
+		}
+	}
+
+	/* We're not yet tracking this connection. */
+
+	if (dont_merge) {
+		return 0;
+	}
+
+	if (list_empty(&st->free_conns)) {
+		c = list_entry(st->conns.prev, 
+			       struct netfront_accel_ssr_conn,
+			       link);
+		if (c->skb) {
+			NETFRONT_ACCEL_STATS_OP(++vnic->stats.ssr_too_many);
+			return 0;
+		}
+	}
+	else {
+		c = list_entry(st->free_conns.next,
+			       struct netfront_accel_ssr_conn,
+			       link);
+	}
+	list_del(&c->link);
+	list_add(&c->link, &st->conns);
+	c->saddr = iph->saddr;
+	c->daddr = iph->daddr;
+	c->source = th->source;
+	c->dest = th->dest;
+	c->next_seq = th_seq + data_length;
+	c->n_in_order_pkts = 0;
+	BUG_ON(c->skb != NULL);
+	NETFRONT_ACCEL_STATS_OP(++vnic->stats.ssr_new_stream);
+	return 0;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/accel_ssr.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_ssr.h
--- linux-2.6.18.8/drivers/xen/sfc_netfront/accel_ssr.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_ssr.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,88 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef NETFRONT_ACCEL_SSR_H
+#define NETFRONT_ACCEL_SSR_H
+
+#include <linux/skbuff.h>
+#include <linux/ip.h>
+#include <linux/tcp.h>
+#include <linux/list.h>
+
+#include "accel.h"
+
+/** State for Soft Segment Reassembly (SSR). */
+
+struct netfront_accel_ssr_conn {
+	struct list_head link;
+
+	unsigned saddr, daddr;
+	unsigned short source, dest;
+
+	/** Number of in-order packets we've seen with payload. */
+	unsigned n_in_order_pkts;
+
+	/** Next in-order sequence number. */
+	unsigned next_seq;
+
+	/** Time we last saw a packet on this connection. */
+	unsigned long last_pkt_jiffies;
+
+	/** The SKB we are currently holding.  If NULL, then all following
+	 * fields are undefined.
+	 */
+	struct sk_buff *skb;
+
+	/** The tail of the frag_list of SKBs we're holding.  Only valid
+	 * after at least one merge.
+	 */
+	struct sk_buff *skb_tail;
+
+	/** The IP header of the skb we are holding. */
+	struct iphdr *iph;
+	
+	/** The TCP header of the skb we are holding. */
+	struct tcphdr *th;
+};
+
+extern void netfront_accel_ssr_init(struct netfront_accel_ssr_state *st);
+extern void netfront_accel_ssr_fini(netfront_accel_vnic *vnic,
+				    struct netfront_accel_ssr_state *st);
+
+extern void
+__netfront_accel_ssr_end_of_burst(netfront_accel_vnic *vnic,
+				  struct netfront_accel_ssr_state *st);
+
+extern int  netfront_accel_ssr_skb(netfront_accel_vnic *vnic,
+				   struct netfront_accel_ssr_state *st,
+				   struct sk_buff *skb);
+
+static inline void
+netfront_accel_ssr_end_of_burst (netfront_accel_vnic *vnic,
+				 struct netfront_accel_ssr_state *st) {
+	if ( ! list_empty(&st->conns) )
+		__netfront_accel_ssr_end_of_burst(vnic, st);
+}
+
+#endif /* NETFRONT_ACCEL_SSR_H */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/accel_tso.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_tso.c
--- linux-2.6.18.8/drivers/xen/sfc_netfront/accel_tso.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_tso.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,511 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/pci.h>
+#include <linux/tcp.h>
+#include <linux/ip.h>
+#include <linux/in.h>
+#include <linux/if_ether.h>
+
+#include "accel.h"
+#include "accel_util.h"
+
+#include "accel_tso.h"
+
+#define PTR_DIFF(p1, p2)  ((u8*)(p1) - (u8*)(p2))
+#define ETH_HDR_LEN(skb)  ((skb)->nh.raw - (skb)->data)
+#define SKB_TCP_OFF(skb)  PTR_DIFF ((skb)->h.th, (skb)->data)
+#define SKB_IP_OFF(skb)   PTR_DIFF ((skb)->nh.iph, (skb)->data)
+
+/*
+ * Set a maximum number of buffers in each output packet to make life
+ * a little simpler - if this is reached it will just move on to
+ * another packet 
+ */
+#define ACCEL_TSO_MAX_BUFFERS (6)
+
+/** TSO State.
+ *
+ * The state used during segmentation.  It is put into this data structure
+ * just to make it easy to pass into inline functions.
+ */
+struct netfront_accel_tso_state {
+	/** bytes of data we've yet to segment */
+	unsigned remaining_len;
+
+	/** current sequence number */
+	unsigned seqnum;
+
+	/** remaining space in current packet */
+	unsigned packet_space;
+
+	/** List of packets to be output, containing the buffers and
+	 *  iovecs to describe each packet 
+	 */
+	struct netfront_accel_tso_output_packet *output_packets;
+
+	/** Total number of buffers in output_packets */
+	unsigned buffers;
+
+	/** Total number of packets in output_packets */
+	unsigned packets;
+
+	/** Input Fragment Cursor.
+	 *
+	 * Where we are in the current fragment of the incoming SKB.  These
+	 * values get updated in place when we split a fragment over
+	 * multiple packets.
+	 */
+	struct {
+		/** address of current position */
+		void *addr;
+		/** remaining length */   
+		unsigned int len;
+	} ifc; /*  == ifc Input Fragment Cursor */
+
+	/** Parameters.
+	 *
+	 * These values are set once at the start of the TSO send and do
+	 * not get changed as the routine progresses.
+	 */
+	struct {
+		/* the number of bytes of header */
+		unsigned int header_length;
+
+		/* The number of bytes to put in each outgoing segment. */
+		int full_packet_size;
+		
+		/* Current IP ID, host endian. */
+		unsigned ip_id;
+
+		/* Max size of each output packet payload */
+		int gso_size;
+	} p;
+};
+
+
+/**
+ * Verify that our various assumptions about sk_buffs and the conditions
+ * under which TSO will be attempted hold true.
+ *
+ * @v skb	       The sk_buff to check.
+ */
+static inline void tso_check_safe(struct sk_buff *skb) {
+	EPRINTK_ON(skb->protocol != htons (ETH_P_IP));
+	EPRINTK_ON(((struct ethhdr*) skb->data)->h_proto != htons (ETH_P_IP));
+	EPRINTK_ON(skb->nh.iph->protocol != IPPROTO_TCP);
+	EPRINTK_ON((SKB_TCP_OFF(skb)
+		    + (skb->h.th->doff << 2u)) > skb_headlen(skb));
+}
+
+
+
+/** Parse the SKB header and initialise state. */
+static inline void tso_start(struct netfront_accel_tso_state *st, 
+			     struct sk_buff *skb) {
+
+	/*
+	 * All ethernet/IP/TCP headers combined size is TCP header size
+	 * plus offset of TCP header relative to start of packet.
+ 	 */
+	st->p.header_length = (skb->h.th->doff << 2u) + SKB_TCP_OFF(skb);
+	st->p.full_packet_size = (st->p.header_length
+				  + skb_shinfo(skb)->gso_size);
+	st->p.gso_size = skb_shinfo(skb)->gso_size;
+
+	st->p.ip_id = htons(skb->nh.iph->id);
+	st->seqnum = ntohl(skb->h.th->seq);
+
+	EPRINTK_ON(skb->h.th->urg);
+	EPRINTK_ON(skb->h.th->syn);
+	EPRINTK_ON(skb->h.th->rst);
+
+	st->remaining_len = skb->len - st->p.header_length;
+
+	st->output_packets = NULL;
+	st->buffers = 0;
+	st->packets = 0;
+
+	VPRINTK("Starting new TSO: hl %d ps %d gso %d seq %x len %d\n",
+		st->p.header_length, st->p.full_packet_size, st->p.gso_size,
+		st->seqnum, skb->len);
+}
+
+/**
+ * Add another NIC mapped buffer onto an output packet  
+ */ 
+static inline int tso_start_new_buffer(netfront_accel_vnic *vnic,
+				       struct netfront_accel_tso_state *st,
+				       int first)
+{
+	struct netfront_accel_tso_buffer *tso_buf;
+	struct netfront_accel_pkt_desc *buf;
+
+	/* Get a mapped packet buffer */
+	buf = netfront_accel_buf_get(vnic->tx_bufs);
+	if (buf == NULL) {
+		DPRINTK("%s: No buffer for TX\n", __FUNCTION__);
+		return -1;
+	}
+
+	/* Store a bit of meta-data at the end */
+	tso_buf =(struct netfront_accel_tso_buffer *)
+		(buf->pkt_kva + NETFRONT_ACCEL_TSO_BUF_LENGTH
+		 + sizeof(struct netfront_accel_tso_output_packet));
+
+	tso_buf->buf = buf;
+
+	tso_buf->length = 0;
+	
+	if (first) {
+		struct netfront_accel_tso_output_packet *output_packet 
+			= (struct netfront_accel_tso_output_packet *)
+			(buf->pkt_kva + NETFRONT_ACCEL_TSO_BUF_LENGTH);
+		output_packet->next = st->output_packets;
+		st->output_packets = output_packet;
+		tso_buf->next = NULL;
+		st->output_packets->tso_bufs = tso_buf;
+		st->output_packets->tso_bufs_len = 1;
+	} else {
+		tso_buf->next = st->output_packets->tso_bufs;
+		st->output_packets->tso_bufs = tso_buf;
+		st->output_packets->tso_bufs_len ++;
+	}
+
+	BUG_ON(st->output_packets->tso_bufs_len > ACCEL_TSO_MAX_BUFFERS);
+	
+	st->buffers ++;
+
+	/*
+	 * Store the context, set to NULL, last packet buffer will get
+	 * non-NULL later
+	 */
+	tso_buf->buf->skb = NULL;
+
+	return 0;
+}
+
+
+/* Generate a new header, and prepare for the new packet.
+ *
+ * @v vnic	      VNIC
+ * @v skb	       Socket buffer
+ * @v st		TSO state
+ * @ret rc	      0 on success, or -1 if failed to alloc header
+ */
+
+static inline 
+int tso_start_new_packet(netfront_accel_vnic *vnic,
+			 struct sk_buff *skb,
+			 struct netfront_accel_tso_state *st) 
+{
+	struct netfront_accel_tso_buffer *tso_buf;
+	struct iphdr *tsoh_iph;
+	struct tcphdr *tsoh_th;
+	unsigned ip_length;
+
+	if (tso_start_new_buffer(vnic, st, 1) < 0) {
+		NETFRONT_ACCEL_STATS_OP(vnic->stats.fastpath_tx_busy++);
+		return -1;		
+	}
+
+	/* This has been set up by tso_start_new_buffer() */
+	tso_buf = st->output_packets->tso_bufs;
+
+	/* Copy in the header */
+	memcpy(tso_buf->buf->pkt_kva, skb->data, st->p.header_length);
+	tso_buf->length = st->p.header_length;
+
+	tsoh_th = (struct tcphdr*) 
+		(tso_buf->buf->pkt_kva + SKB_TCP_OFF(skb));
+	tsoh_iph = (struct iphdr*) 
+		(tso_buf->buf->pkt_kva + SKB_IP_OFF(skb));
+
+	/* Set to zero to encourage falcon to fill these in */
+	tsoh_th->check  = 0;
+	tsoh_iph->check = 0;
+
+	tsoh_th->seq = htonl(st->seqnum);
+	st->seqnum += st->p.gso_size;
+
+	if (st->remaining_len > st->p.gso_size) {
+		/* This packet will not finish the TSO burst. */
+		ip_length = st->p.full_packet_size - ETH_HDR_LEN(skb);
+		tsoh_th->fin = 0;
+		tsoh_th->psh = 0;
+	} else {
+		/* This packet will be the last in the TSO burst. */
+		ip_length = (st->p.header_length - ETH_HDR_LEN(skb)
+			     + st->remaining_len);
+		tsoh_th->fin = skb->h.th->fin;
+		tsoh_th->psh = skb->h.th->psh;
+	}
+
+	tsoh_iph->tot_len = htons(ip_length);
+
+	/* Linux leaves suitable gaps in the IP ID space for us to fill. */
+	tsoh_iph->id = st->p.ip_id++;
+	tsoh_iph->id = htons(tsoh_iph->id);
+
+	st->packet_space = st->p.gso_size; 
+
+	st->packets++;
+
+	return 0;
+}
+
+
+
+static inline void tso_get_fragment(struct netfront_accel_tso_state *st, 
+				    int len, void *addr)
+{
+	st->ifc.len = len;
+	st->ifc.addr = addr;
+	return;
+}
+
+
+static inline void tso_unwind(netfront_accel_vnic *vnic, 
+			      struct netfront_accel_tso_state *st)
+{
+	struct netfront_accel_tso_buffer *tso_buf;
+	struct netfront_accel_tso_output_packet *output_packet;
+
+	DPRINTK("%s\n", __FUNCTION__);
+
+	while (st->output_packets != NULL) {
+		output_packet = st->output_packets;
+		st->output_packets = output_packet->next;
+		while (output_packet->tso_bufs != NULL) {
+			tso_buf = output_packet->tso_bufs;
+			output_packet->tso_bufs = tso_buf->next;
+
+			st->buffers --;
+			output_packet->tso_bufs_len --;
+
+			netfront_accel_buf_put(vnic->tx_bufs, 
+					       tso_buf->buf->buf_id);
+		}
+	}
+	BUG_ON(st->buffers != 0);
+}
+
+
+
+static inline
+void tso_fill_packet_with_fragment(netfront_accel_vnic *vnic,
+				   struct netfront_accel_tso_state *st) 
+{
+	struct netfront_accel_tso_buffer *tso_buf;
+	int n, space;
+
+	BUG_ON(st->output_packets == NULL);
+	BUG_ON(st->output_packets->tso_bufs == NULL);
+
+	tso_buf = st->output_packets->tso_bufs;
+
+	if (st->ifc.len == 0)  return;
+	if (st->packet_space == 0)  return;
+	if (tso_buf->length == NETFRONT_ACCEL_TSO_BUF_LENGTH) return;
+
+	n = min(st->ifc.len, st->packet_space);
+
+	space = NETFRONT_ACCEL_TSO_BUF_LENGTH - tso_buf->length;
+	n = min(n, space);
+
+	st->packet_space -= n;
+	st->remaining_len -= n;
+	st->ifc.len -= n;
+
+	memcpy(tso_buf->buf->pkt_kva + tso_buf->length, st->ifc.addr, n);
+
+	tso_buf->length += n;
+
+	BUG_ON(tso_buf->length > NETFRONT_ACCEL_TSO_BUF_LENGTH);
+
+	st->ifc.addr += n;
+
+	return;
+}
+
+
+int netfront_accel_enqueue_skb_tso(netfront_accel_vnic *vnic,
+				   struct sk_buff *skb)
+{
+	struct netfront_accel_tso_state state;
+	struct netfront_accel_tso_buffer *tso_buf = NULL;
+	struct netfront_accel_tso_output_packet *reversed_list = NULL;
+	struct netfront_accel_tso_output_packet	*tmp_pkt;
+	ef_iovec iovecs[ACCEL_TSO_MAX_BUFFERS];
+	int frag_i, rc, dma_id;
+	skb_frag_t *f;
+
+	tso_check_safe(skb);
+
+	if (skb->ip_summed != CHECKSUM_HW)
+		EPRINTK("Trying to TSO send a packet without HW checksum\n");
+
+	tso_start(&state, skb);
+
+	/*
+	 * Setup the first payload fragment.  If the skb header area
+	 * contains exactly the headers and all payload is in the frag
+	 * list things are little simpler
+	 */
+	if (skb_headlen(skb) == state.p.header_length) {
+		/* Grab the first payload fragment. */
+		BUG_ON(skb_shinfo(skb)->nr_frags < 1);
+		frag_i = 0;
+		f = &skb_shinfo(skb)->frags[frag_i];
+		tso_get_fragment(&state, f->size, 
+				 page_address(f->page) + f->page_offset);
+	} else {
+		int hl = state.p.header_length;
+		tso_get_fragment(&state,  skb_headlen(skb) - hl, 
+				 skb->data + hl);
+		frag_i = -1;
+	}
+
+	if (tso_start_new_packet(vnic, skb, &state) < 0) {
+		DPRINTK("%s: out of first start-packet memory\n",
+			__FUNCTION__);
+		goto unwind;
+	}
+
+	while (1) {
+		tso_fill_packet_with_fragment(vnic, &state);
+		
+		/* Move onto the next fragment? */
+		if (state.ifc.len == 0) {
+			if (++frag_i >= skb_shinfo(skb)->nr_frags)
+				/* End of payload reached. */
+				break;
+			f = &skb_shinfo(skb)->frags[frag_i];
+			tso_get_fragment(&state, f->size,
+					 page_address(f->page) +
+					 f->page_offset);
+		}
+
+		/* Start a new buffer? */
+		if ((state.output_packets->tso_bufs->length == 
+		     NETFRONT_ACCEL_TSO_BUF_LENGTH) &&
+		    tso_start_new_buffer(vnic, &state, 0)) {
+			DPRINTK("%s: out of start-buffer memory\n",
+				__FUNCTION__);
+			goto unwind;
+		}
+
+		/* Start at new packet? */
+		if ((state.packet_space == 0 || 
+		     ((state.output_packets->tso_bufs_len >=
+		       ACCEL_TSO_MAX_BUFFERS) &&
+		      (state.output_packets->tso_bufs->length >= 
+		       NETFRONT_ACCEL_TSO_BUF_LENGTH))) &&
+		    tso_start_new_packet(vnic, skb, &state) < 0) {
+			DPRINTK("%s: out of start-packet memory\n",
+				__FUNCTION__);
+			goto unwind;
+		}
+
+	}
+
+	/* Check for space */
+	if (ef_vi_transmit_space(&vnic->vi) < state.buffers) {
+		DPRINTK("%s: Not enough TX space (%d)\n",
+			__FUNCTION__, state.buffers);
+		goto unwind;
+	}
+
+	/*
+	 * Store the skb context in the most recent buffer (i.e. the
+	 * last buffer that will be sent)
+	 */
+	state.output_packets->tso_bufs->buf->skb = skb;
+
+	/* Reverse the list of packets as we construct it on a stack */
+	while (state.output_packets != NULL) {
+		tmp_pkt = state.output_packets;
+		state.output_packets = tmp_pkt->next;
+		tmp_pkt->next = reversed_list;
+		reversed_list = tmp_pkt;
+	}
+
+	/* Pass off to hardware */
+	while (reversed_list != NULL) {
+		tmp_pkt = reversed_list;
+		reversed_list = tmp_pkt->next;
+
+		BUG_ON(tmp_pkt->tso_bufs_len > ACCEL_TSO_MAX_BUFFERS);
+		BUG_ON(tmp_pkt->tso_bufs_len == 0);
+
+		dma_id = tmp_pkt->tso_bufs->buf->buf_id;
+
+		/*
+		 * Make an iovec of the buffers in the list, reversing
+		 * the buffers as we go as they are constructed on a
+		 * stack
+		 */
+		tso_buf = tmp_pkt->tso_bufs;
+		for (frag_i = tmp_pkt->tso_bufs_len - 1;
+		     frag_i >= 0;
+		     frag_i--) {
+			iovecs[frag_i].iov_base = tso_buf->buf->pkt_buff_addr;
+			iovecs[frag_i].iov_len = tso_buf->length;
+			tso_buf = tso_buf->next;
+		}
+
+		rc = ef_vi_transmitv(&vnic->vi, iovecs, tmp_pkt->tso_bufs_len,
+				     dma_id);
+		/*
+		 * We checked for space already, so it really should
+		 * succeed
+		 */
+		BUG_ON(rc != 0);
+	}
+
+	/* Track number of tx fastpath stats */
+	vnic->netdev_stats.fastpath_tx_bytes += skb->len;
+	vnic->netdev_stats.fastpath_tx_pkts += state.packets;
+#if NETFRONT_ACCEL_STATS
+	{
+		unsigned n;
+		n = vnic->netdev_stats.fastpath_tx_pkts -
+			vnic->stats.fastpath_tx_completions;
+		if (n > vnic->stats.fastpath_tx_pending_max)
+			vnic->stats.fastpath_tx_pending_max = n;
+	}
+#endif
+
+	return NETFRONT_ACCEL_STATUS_GOOD;
+ 
+ unwind:
+	tso_unwind(vnic, &state);
+
+	NETFRONT_ACCEL_STATS_OP(vnic->stats.fastpath_tx_busy++);
+
+	return NETFRONT_ACCEL_STATUS_BUSY;
+}
+
+
+
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/accel_tso.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_tso.h
--- linux-2.6.18.8/drivers/xen/sfc_netfront/accel_tso.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_tso.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,57 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef NETFRONT_ACCEL_TSO_H
+#define NETFRONT_ACCEL_TSO_H
+
+#include "accel_bufs.h"
+
+/* Track the buffers used in each output packet */
+struct netfront_accel_tso_buffer {
+	struct netfront_accel_tso_buffer *next;
+	struct netfront_accel_pkt_desc *buf;
+	unsigned length;
+};
+
+/* Track the output packets formed from each input packet */
+struct netfront_accel_tso_output_packet {
+	struct netfront_accel_tso_output_packet *next;
+	struct netfront_accel_tso_buffer *tso_bufs;
+	unsigned tso_bufs_len;
+};
+
+
+/*
+ * Max available space in a buffer for data once meta-data has taken
+ * its place 
+ */
+#define NETFRONT_ACCEL_TSO_BUF_LENGTH					\
+	((PAGE_SIZE / NETFRONT_ACCEL_BUFS_PER_PAGE)			\
+	 - sizeof(struct netfront_accel_tso_buffer)			\
+	 - sizeof(struct netfront_accel_tso_output_packet))
+
+int netfront_accel_enqueue_skb_tso(netfront_accel_vnic *vnic,
+				   struct sk_buff *skb);
+
+#endif /* NETFRONT_ACCEL_TSO_H */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/accel_vi.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_vi.c
--- linux-2.6.18.8/drivers/xen/sfc_netfront/accel_vi.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_vi.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1194 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/if_ether.h>
+#include <linux/ip.h>
+#include <net/checksum.h>
+#include <asm/io.h>
+
+#include "accel.h"
+#include "accel_util.h"
+#include "accel_bufs.h"
+#include "accel_tso.h"
+#include "accel_ssr.h"
+#include "netfront.h"
+
+#include "etherfabric/ef_vi.h"
+
+/*
+ * Max available space in a buffer for data once meta-data has taken
+ * its place
+ */
+#define NETFRONT_ACCEL_TX_BUF_LENGTH					\
+	((PAGE_SIZE / NETFRONT_ACCEL_BUFS_PER_PAGE)			\
+	 - sizeof(struct netfront_accel_tso_buffer))
+
+#define ACCEL_TX_MAX_BUFFERS (6)
+#define ACCEL_VI_POLL_EVENTS (8)
+
+static
+int netfront_accel_vi_init_fini(netfront_accel_vnic *vnic, 
+				struct net_accel_msg_hw *hw_msg)
+{
+	struct ef_vi_nic_type nic_type;
+	struct net_accel_hw_falcon_b *hw_info;
+	void *io_kva, *evq_base, *rx_dma_kva, *tx_dma_kva, *doorbell_kva;
+	u32 *evq_gnts;
+	u32 evq_order;
+	int vi_state_size;
+	u8 vi_data[VI_MAPPINGS_SIZE];
+
+	if (hw_msg == NULL)
+		goto fini;
+
+	/* And create the local macs table lock */
+	spin_lock_init(&vnic->table_lock);
+	
+	/* Create fastpath table, initial size 8, key length 8 */
+	if (cuckoo_hash_init(&vnic->fastpath_table, 3, 8)) {
+		EPRINTK("failed to allocate fastpath table\n");
+		goto fail_cuckoo;
+	}
+
+	vnic->hw.falcon.type = hw_msg->type;
+
+	switch (hw_msg->type) {
+	case NET_ACCEL_MSG_HWTYPE_FALCON_A:
+		hw_info = &hw_msg->resources.falcon_a.common;
+		/* Need the extra rptr register page on A1 */
+		io_kva = net_accel_map_iomem_page
+			(vnic->dev, hw_msg->resources.falcon_a.evq_rptr_gnt,
+			 &vnic->hw.falcon.evq_rptr_mapping);
+		if (io_kva == NULL) {
+			EPRINTK("%s: evq_rptr permission failed\n", __FUNCTION__);
+			goto evq_rptr_fail;
+		}
+
+		vnic->hw.falcon.evq_rptr = io_kva + 
+			(hw_info->evq_rptr & (PAGE_SIZE - 1));
+		break;
+	case NET_ACCEL_MSG_HWTYPE_FALCON_B:
+		hw_info = &hw_msg->resources.falcon_b;
+		break;
+	default:
+		goto bad_type;
+	}
+
+	/**** Event Queue ****/
+
+	/* Map the event queue pages */
+	evq_gnts = hw_info->evq_mem_gnts;
+	evq_order = hw_info->evq_order;
+
+	EPRINTK_ON(hw_info->evq_offs != 0);
+
+	DPRINTK("Will map evq %d pages\n", 1 << evq_order);
+
+	evq_base =
+		net_accel_map_grants_contig(vnic->dev, evq_gnts, 1 << evq_order,
+					    &vnic->evq_mapping);
+	if (evq_base == NULL) {
+		EPRINTK("%s: evq_base failed\n", __FUNCTION__);
+		goto evq_fail;
+	}
+
+	/**** Doorbells ****/
+	/* Set up the doorbell mappings. */
+	doorbell_kva = 
+		net_accel_map_iomem_page(vnic->dev, hw_info->doorbell_gnt,
+					 &vnic->hw.falcon.doorbell_mapping);
+	if (doorbell_kva == NULL) {
+		EPRINTK("%s: doorbell permission failed\n", __FUNCTION__);
+		goto doorbell_fail;
+	}
+	vnic->hw.falcon.doorbell = doorbell_kva;
+
+	/* On Falcon_B we get the rptr from the doorbell page */
+	if (hw_msg->type == NET_ACCEL_MSG_HWTYPE_FALCON_B) {
+		vnic->hw.falcon.evq_rptr = 
+			(u32 *)((char *)vnic->hw.falcon.doorbell 
+				+ hw_info->evq_rptr);
+	}
+
+	/**** DMA Queue ****/
+
+	/* Set up the DMA Queues from the message. */
+	tx_dma_kva = net_accel_map_grants_contig
+		(vnic->dev, &(hw_info->txdmaq_gnt), 1, 
+		 &vnic->hw.falcon.txdmaq_mapping);
+	if (tx_dma_kva == NULL) {
+		EPRINTK("%s: TX dma failed\n", __FUNCTION__);
+		goto tx_dma_fail;
+	}
+
+	rx_dma_kva = net_accel_map_grants_contig
+		(vnic->dev, &(hw_info->rxdmaq_gnt), 1, 
+		 &vnic->hw.falcon.rxdmaq_mapping);
+	if (rx_dma_kva == NULL) {
+		EPRINTK("%s: RX dma failed\n", __FUNCTION__);
+		goto rx_dma_fail;
+	}
+
+	/* Full confession */
+	DPRINTK("Mapped H/W"
+		"  Tx DMAQ grant %x -> %p\n"
+		"  Rx DMAQ grant %x -> %p\n"
+		"  EVQ grant %x -> %p\n",
+		hw_info->txdmaq_gnt, tx_dma_kva,
+		hw_info->rxdmaq_gnt, rx_dma_kva,
+		evq_gnts[0], evq_base
+		);
+
+	memset(vi_data, 0, sizeof(vi_data));
+	
+	/* TODO BUG11305: convert efhw_arch to ef_vi_arch
+	 * e.g.
+	 * arch = ef_vi_arch_from_efhw_arch(hw_info->nic_arch);
+	 * assert(arch >= 0);
+	 * nic_type.arch = arch;
+	 */
+	nic_type.arch = (unsigned char)hw_info->nic_arch;
+	nic_type.variant = (char)hw_info->nic_variant;
+	nic_type.revision = (unsigned char)hw_info->nic_revision;
+	
+	ef_vi_init_mapping_evq(vi_data, nic_type, hw_info->instance, 
+			       1 << (evq_order + PAGE_SHIFT), evq_base, 
+			       (void *)0xdeadbeef);
+
+	ef_vi_init_mapping_vi(vi_data, nic_type, hw_info->rx_capacity, 
+			      hw_info->tx_capacity, hw_info->instance, 
+			      doorbell_kva, rx_dma_kva, tx_dma_kva, 0);
+
+	vi_state_size = ef_vi_calc_state_bytes(hw_info->rx_capacity,
+					       hw_info->tx_capacity);
+	vnic->vi_state = (ef_vi_state *)kmalloc(vi_state_size, GFP_KERNEL);
+	if (vnic->vi_state == NULL) {
+		EPRINTK("%s: kmalloc for VI state failed\n", __FUNCTION__);
+		goto vi_state_fail;
+	}
+	ef_vi_init(&vnic->vi, vi_data, vnic->vi_state, &vnic->evq_state, 0);
+
+	ef_eventq_state_init(&vnic->vi);
+
+	ef_vi_state_init(&vnic->vi);
+
+	return 0;
+
+fini:
+	kfree(vnic->vi_state);
+	vnic->vi_state = NULL;
+vi_state_fail:
+	net_accel_unmap_grants_contig(vnic->dev, vnic->hw.falcon.rxdmaq_mapping);
+rx_dma_fail:
+	net_accel_unmap_grants_contig(vnic->dev, vnic->hw.falcon.txdmaq_mapping);
+tx_dma_fail:
+	net_accel_unmap_iomem_page(vnic->dev, vnic->hw.falcon.doorbell_mapping);
+	vnic->hw.falcon.doorbell = NULL;
+doorbell_fail:
+	net_accel_unmap_grants_contig(vnic->dev, vnic->evq_mapping);
+evq_fail:
+	if (vnic->hw.falcon.type == NET_ACCEL_MSG_HWTYPE_FALCON_A)
+		net_accel_unmap_iomem_page(vnic->dev, 
+					   vnic->hw.falcon.evq_rptr_mapping);
+	vnic->hw.falcon.evq_rptr = NULL;
+evq_rptr_fail:
+bad_type:
+	cuckoo_hash_destroy(&vnic->fastpath_table);
+fail_cuckoo:
+	return -EIO;
+}
+
+
+void netfront_accel_vi_ctor(netfront_accel_vnic *vnic)
+{
+	/* Just mark the VI as uninitialised. */
+	vnic->vi_state = NULL;
+}
+
+
+int netfront_accel_vi_init(netfront_accel_vnic *vnic, struct net_accel_msg_hw *hw_msg)
+{
+	BUG_ON(hw_msg == NULL);
+	return netfront_accel_vi_init_fini(vnic, hw_msg);
+}
+
+
+void netfront_accel_vi_dtor(netfront_accel_vnic *vnic)
+{
+	if (vnic->vi_state != NULL)
+		netfront_accel_vi_init_fini(vnic, NULL);
+}
+
+
+static
+void netfront_accel_vi_post_rx(netfront_accel_vnic *vnic, u16 id,
+			       netfront_accel_pkt_desc *buf)
+{
+
+	int idx = vnic->rx_dma_batched;
+
+#if 0
+	VPRINTK("Posting buffer %d (0x%08x) for rx at index %d, space is %d\n",
+		id, buf->pkt_buff_addr, idx, ef_vi_receive_space(&vnic->vi));
+#endif
+	/* Set up a virtual buffer descriptor */
+	ef_vi_receive_init(&vnic->vi, buf->pkt_buff_addr, id,
+			   /*rx_bytes=max*/0);
+
+	idx++;
+
+	vnic->rx_dma_level++;
+	
+	/* 
+	 * Only push the descriptor to the card if we've reached the
+	 * batch size.  Otherwise, the descriptors can sit around for
+	 * a while.  There will be plenty available.
+	 */
+	if (idx >= NETFRONT_ACCEL_RX_DESC_BATCH ||
+	    vnic->rx_dma_level < NETFRONT_ACCEL_RX_DESC_BATCH) {
+#if 0
+		VPRINTK("Flushing %d rx descriptors.\n", idx);
+#endif
+
+		/* Push buffer to hardware */
+		ef_vi_receive_push(&vnic->vi);
+		
+		idx = 0;
+	}
+	
+	vnic->rx_dma_batched = idx;
+}
+
+
+inline
+void netfront_accel_vi_post_rx_or_free(netfront_accel_vnic *vnic, u16 id,
+				       netfront_accel_pkt_desc *buf)
+{
+
+	VPRINTK("%s: %d\n", __FUNCTION__, id);
+
+	if (ef_vi_receive_space(&vnic->vi) <= vnic->rx_dma_batched) {
+		VPRINTK("RX space is full\n");
+		netfront_accel_buf_put(vnic->rx_bufs, id);
+		return;
+	}
+
+	VPRINTK("Completed buffer %d is reposted\n", id);
+	netfront_accel_vi_post_rx(vnic, id, buf);
+	
+	/*
+	 * Let's see if there's any more to be pushed out to the NIC
+	 * while we're here
+	 */
+	while (ef_vi_receive_space(&vnic->vi) > vnic->rx_dma_batched) {
+		/* Try to allocate a buffer. */
+		buf = netfront_accel_buf_get(vnic->rx_bufs);
+		if (buf == NULL)
+			break;
+		
+		/* Add it to the rx dma queue. */
+		netfront_accel_vi_post_rx(vnic, buf->buf_id, buf);	
+	}
+}
+
+
+void netfront_accel_vi_add_bufs(netfront_accel_vnic *vnic, int is_rx)
+{
+
+	while (is_rx && 
+	       ef_vi_receive_space(&vnic->vi) > vnic->rx_dma_batched) {
+		netfront_accel_pkt_desc *buf;
+		
+		VPRINTK("%s: %d\n", __FUNCTION__, vnic->rx_dma_level);
+		
+		/* Try to allocate a buffer. */
+		buf = netfront_accel_buf_get(vnic->rx_bufs);
+
+		if (buf == NULL)
+			break;
+		
+		/* Add it to the rx dma queue. */
+		netfront_accel_vi_post_rx(vnic, buf->buf_id, buf);
+	}
+
+	VPRINTK("%s: done\n", __FUNCTION__);
+}
+
+
+struct netfront_accel_multi_state {
+	unsigned remaining_len;
+
+	unsigned buffers;
+
+	struct netfront_accel_tso_buffer *output_buffers;
+
+	/* Where we are in the current fragment of the SKB. */
+	struct {
+		/* address of current position */
+		void *addr;
+		/* remaining length */	  
+		unsigned int len;
+	} ifc; /*  == Input Fragment Cursor */
+};
+
+
+static inline void multi_post_start(struct netfront_accel_multi_state *st, 
+				    struct sk_buff *skb)
+{
+	st->remaining_len = skb->len;
+	st->output_buffers = NULL;
+	st->buffers = 0;
+	st->ifc.len = skb_headlen(skb);
+	st->ifc.addr = skb->data;
+}
+
+static int multi_post_start_new_buffer(netfront_accel_vnic *vnic, 
+				       struct netfront_accel_multi_state *st)
+{
+	struct netfront_accel_tso_buffer *tso_buf;
+	struct netfront_accel_pkt_desc *buf;
+
+	/* Get a mapped packet buffer */
+	buf = netfront_accel_buf_get(vnic->tx_bufs);
+	if (buf == NULL) {
+		DPRINTK("%s: No buffer for TX\n", __FUNCTION__);
+		return -1;
+	}
+
+	/* Store a bit of meta-data at the end */
+	tso_buf = (struct netfront_accel_tso_buffer *)
+		(buf->pkt_kva + NETFRONT_ACCEL_TX_BUF_LENGTH);
+
+	tso_buf->buf = buf;
+
+	tso_buf->length = 0;
+	
+	tso_buf->next = st->output_buffers;
+	st->output_buffers = tso_buf;
+	st->buffers++;
+
+	BUG_ON(st->buffers >= ACCEL_TX_MAX_BUFFERS);
+
+	/*
+	 * Store the context, set to NULL, last packet buffer will get
+	 * non-NULL later
+	 */
+	tso_buf->buf->skb = NULL;
+	
+	return 0;
+}
+
+
+static void
+multi_post_fill_buffer_with_fragment(netfront_accel_vnic *vnic,
+				     struct netfront_accel_multi_state *st)
+{
+	struct netfront_accel_tso_buffer *tso_buf;
+	unsigned n, space;
+
+	BUG_ON(st->output_buffers == NULL);
+	tso_buf = st->output_buffers;
+
+	if (st->ifc.len == 0) return;
+	if (tso_buf->length == NETFRONT_ACCEL_TX_BUF_LENGTH) return;
+
+	BUG_ON(tso_buf->length > NETFRONT_ACCEL_TX_BUF_LENGTH);
+
+	space = NETFRONT_ACCEL_TX_BUF_LENGTH - tso_buf->length;
+	n = min(st->ifc.len, space);
+
+	memcpy(tso_buf->buf->pkt_kva + tso_buf->length, st->ifc.addr, n);
+
+	st->remaining_len -= n;
+	st->ifc.len -= n;
+	tso_buf->length += n;
+	st->ifc.addr += n;
+
+	BUG_ON(tso_buf->length > NETFRONT_ACCEL_TX_BUF_LENGTH);
+
+	return;
+}
+
+
+static inline void multi_post_unwind(netfront_accel_vnic *vnic,
+				     struct netfront_accel_multi_state *st)
+{
+	struct netfront_accel_tso_buffer *tso_buf;
+
+	DPRINTK("%s\n", __FUNCTION__);
+
+	while (st->output_buffers != NULL) {
+		tso_buf = st->output_buffers;
+		st->output_buffers = tso_buf->next;
+		st->buffers--;
+		netfront_accel_buf_put(vnic->tx_bufs, tso_buf->buf->buf_id);
+	}
+	BUG_ON(st->buffers != 0);
+}
+
+
+static enum netfront_accel_post_status
+netfront_accel_enqueue_skb_multi(netfront_accel_vnic *vnic, struct sk_buff *skb)
+{
+	struct netfront_accel_tso_buffer *tso_buf;
+	struct netfront_accel_multi_state state;
+	ef_iovec iovecs[ACCEL_TX_MAX_BUFFERS];
+	skb_frag_t *f;
+	int frag_i, rc, dma_id;
+
+	multi_post_start(&state, skb);
+
+	frag_i = -1;
+
+	if (skb->ip_summed == CHECKSUM_HW) {
+		/* Set to zero to encourage falcon to work it out for us */
+		*(u16*)(skb->h.raw + skb->csum) = 0;
+	}
+
+	if (multi_post_start_new_buffer(vnic, &state)) {
+		DPRINTK("%s: out of buffers\n", __FUNCTION__);
+		goto unwind;
+	}
+
+	while (1) {
+		multi_post_fill_buffer_with_fragment(vnic, &state);
+
+		/* Move onto the next fragment? */
+		if (state.ifc.len == 0) {
+			if (++frag_i >= skb_shinfo(skb)->nr_frags)
+				/* End of payload reached. */
+				break;
+			f = &skb_shinfo(skb)->frags[frag_i];
+			state.ifc.len = f->size;
+			state.ifc.addr = page_address(f->page) + f->page_offset;
+		}
+
+		/* Start a new buffer? */
+		if ((state.output_buffers->length == 
+		     NETFRONT_ACCEL_TX_BUF_LENGTH) &&
+		    multi_post_start_new_buffer(vnic, &state)) {
+			DPRINTK("%s: out of buffers\n", __FUNCTION__);
+			goto unwind;
+		}
+	}
+
+	/* Check for space */
+	if (ef_vi_transmit_space(&vnic->vi) < state.buffers) {
+		DPRINTK("%s: Not enough TX space (%d)\n", __FUNCTION__, state.buffers);
+		goto unwind;
+	}
+
+	/* Store the skb in what will be the last buffer's context */
+	state.output_buffers->buf->skb = skb;
+	/* Remember dma_id of what will be the last buffer */ 
+	dma_id = state.output_buffers->buf->buf_id;
+
+	/*
+	 * Make an iovec of the buffers in the list, reversing the
+	 * buffers as we go as they are constructed on a stack
+	 */
+	tso_buf = state.output_buffers;
+	for (frag_i = state.buffers-1; frag_i >= 0; frag_i--) {
+		iovecs[frag_i].iov_base = tso_buf->buf->pkt_buff_addr;
+		iovecs[frag_i].iov_len = tso_buf->length;
+		tso_buf = tso_buf->next;
+	}
+	
+	rc = ef_vi_transmitv(&vnic->vi, iovecs, state.buffers, dma_id);
+
+	/* Track number of tx fastpath stats */
+	vnic->netdev_stats.fastpath_tx_bytes += skb->len;
+	vnic->netdev_stats.fastpath_tx_pkts ++;
+#if NETFRONT_ACCEL_STATS
+	{
+		u32 n;
+		n = vnic->netdev_stats.fastpath_tx_pkts -
+			(u32)vnic->stats.fastpath_tx_completions;
+		if (n > vnic->stats.fastpath_tx_pending_max)
+			vnic->stats.fastpath_tx_pending_max = n;
+	}
+#endif
+	return NETFRONT_ACCEL_STATUS_GOOD;
+
+unwind:
+	multi_post_unwind(vnic, &state);
+
+	NETFRONT_ACCEL_STATS_OP(vnic->stats.fastpath_tx_busy++);
+
+	return NETFRONT_ACCEL_STATUS_BUSY;
+}
+
+
+static enum netfront_accel_post_status 
+netfront_accel_enqueue_skb_single(netfront_accel_vnic *vnic, struct sk_buff *skb)
+{
+	struct netfront_accel_tso_buffer *tso_buf;
+	struct netfront_accel_pkt_desc *buf;
+	u8 *kva;
+	int rc;
+
+	if (ef_vi_transmit_space(&vnic->vi) < 1) {
+		DPRINTK("%s: No TX space\n", __FUNCTION__);
+		NETFRONT_ACCEL_STATS_OP(vnic->stats.fastpath_tx_busy++);
+		return NETFRONT_ACCEL_STATUS_BUSY;
+	}
+
+	buf = netfront_accel_buf_get(vnic->tx_bufs);
+	if (buf == NULL) {
+		DPRINTK("%s: No buffer for TX\n", __FUNCTION__);
+		NETFRONT_ACCEL_STATS_OP(vnic->stats.fastpath_tx_busy++);
+		return NETFRONT_ACCEL_STATUS_BUSY;
+	}
+
+	/* Track number of tx fastpath stats */
+	vnic->netdev_stats.fastpath_tx_pkts++;
+	vnic->netdev_stats.fastpath_tx_bytes += skb->len;
+
+#if NETFRONT_ACCEL_STATS
+	{
+		u32 n;
+		n = vnic->netdev_stats.fastpath_tx_pkts - 
+			(u32)vnic->stats.fastpath_tx_completions;
+		if (n > vnic->stats.fastpath_tx_pending_max)
+			vnic->stats.fastpath_tx_pending_max = n;
+	}
+#endif
+	
+	/* Store the context */
+	buf->skb = skb;
+	
+	kva = buf->pkt_kva;
+
+	if (skb->ip_summed == CHECKSUM_HW) {
+		/* Set to zero to encourage falcon to work it out for us */
+		*(u16*)(skb->h.raw + skb->csum) = 0;
+	}
+	NETFRONT_ACCEL_PKTBUFF_FOR_EACH_FRAGMENT
+		(skb, idx, frag_data, frag_len, {
+			/* Copy in payload */
+			VPRINTK("*** Copying %d bytes to %p\n", frag_len, kva);
+			memcpy(kva, frag_data, frag_len);
+			kva += frag_len;
+		});
+
+	VPRINTK("%s: id %d pkt %p kva %p buff_addr 0x%08x\n", __FUNCTION__,
+		buf->buf_id, buf, buf->pkt_kva, buf->pkt_buff_addr);
+
+
+	/* Set up the TSO meta-data for a single buffer/packet */
+	tso_buf = (struct netfront_accel_tso_buffer *)
+		(buf->pkt_kva + NETFRONT_ACCEL_TX_BUF_LENGTH);
+	tso_buf->next = NULL;
+	tso_buf->buf = buf;
+	tso_buf->length = skb->len;
+
+	rc = ef_vi_transmit(&vnic->vi, buf->pkt_buff_addr, skb->len,
+			    buf->buf_id);
+	/* We checked for space already, so it really should succeed */
+	BUG_ON(rc != 0);
+
+	return NETFRONT_ACCEL_STATUS_GOOD;
+}
+
+
+enum netfront_accel_post_status 
+netfront_accel_vi_tx_post(netfront_accel_vnic *vnic, struct sk_buff *skb)
+{
+	struct ethhdr *pkt_eth_hdr;
+	struct iphdr *pkt_ipv4_hdr;
+	int value, try_fastpath;
+
+	/*
+	 * This assumes that the data field points to the dest mac
+	 * address.
+	 */
+	cuckoo_hash_mac_key key = cuckoo_mac_to_key(skb->data);
+
+	/*
+	 * NB very important that all things that could return "CANT"
+	 * are tested before things that return "BUSY" as if it it
+	 * returns "BUSY" it is assumed that it won't return "CANT"
+	 * next time it is tried
+	 */
+
+	/*
+	 * Do a fastpath send if fast path table lookup returns true.
+	 * We do this without the table lock and so may get the wrong
+	 * answer, but current opinion is that's not a big problem 
+	 */
+	try_fastpath = cuckoo_hash_lookup(&vnic->fastpath_table, 
+					  (cuckoo_hash_key *)(&key), &value);
+
+	if (!try_fastpath) {
+		VPRINTK("try fast path false for mac: " MAC_FMT "\n",
+			MAC_ARG(skb->data));
+		
+		return NETFRONT_ACCEL_STATUS_CANT;
+	}
+
+	/* Check to see if the packet can be sent. */
+	if (skb_headlen(skb) < sizeof(*pkt_eth_hdr) + sizeof(*pkt_ipv4_hdr)) {
+		EPRINTK("%s: Packet header is too small\n", __FUNCTION__);
+		return NETFRONT_ACCEL_STATUS_CANT;
+	}
+
+	pkt_eth_hdr  = (void*)skb->data;
+	pkt_ipv4_hdr = (void*)(pkt_eth_hdr+1);
+
+	if (be16_to_cpu(pkt_eth_hdr->h_proto) != ETH_P_IP) {
+		DPRINTK("%s: Packet is not IPV4 (ether_type=0x%04x)\n", __FUNCTION__,
+			be16_to_cpu(pkt_eth_hdr->h_proto));
+		return NETFRONT_ACCEL_STATUS_CANT;
+	}
+	
+	if (pkt_ipv4_hdr->protocol != IPPROTO_TCP &&
+	    pkt_ipv4_hdr->protocol != IPPROTO_UDP) {
+		DPRINTK("%s: Packet is not TCP/UDP (ip_protocol=0x%02x)\n",
+			__FUNCTION__, pkt_ipv4_hdr->protocol);
+		return NETFRONT_ACCEL_STATUS_CANT;
+	}
+	
+	VPRINTK("%s: %d bytes, gso %d\n", __FUNCTION__, skb->len, 
+		skb_shinfo(skb)->gso_size);
+	
+	if (skb_shinfo(skb)->gso_size) {
+		return netfront_accel_enqueue_skb_tso(vnic, skb);
+	}
+
+	if (skb->len <= NETFRONT_ACCEL_TX_BUF_LENGTH) {
+		return netfront_accel_enqueue_skb_single(vnic, skb);
+	}
+
+	return netfront_accel_enqueue_skb_multi(vnic, skb);
+}
+
+
+/*
+ * Copy the data to required end destination. NB. len is the total new
+ * length of the socket buffer, not the amount of data to copy
+ */
+inline
+int ef_vnic_copy_to_skb(netfront_accel_vnic *vnic, struct sk_buff *skb, 
+			struct netfront_accel_pkt_desc *buf, int len)
+{
+	int i, extra = len - skb->len;
+	char c;
+	int pkt_stride = vnic->rx_pkt_stride;
+	int skb_stride = vnic->rx_skb_stride;
+	char *skb_start;
+	
+	/*
+	 * This pulls stuff into the cache - have seen performance
+	 * benefit in this, but disabled by default
+	 */
+	skb_start = skb->data;
+	if (pkt_stride) {
+		for (i = 0; i < len; i += pkt_stride) {
+			c += ((volatile char*)(buf->pkt_kva))[i];
+		}
+	}
+	if (skb_stride) {
+		for (i = skb->len; i < len ; i += skb_stride) {
+			c += ((volatile char*)(skb_start))[i];
+		}
+	}
+
+	if (skb_tailroom(skb) >= extra) {
+		memcpy(skb_put(skb, extra), buf->pkt_kva, extra);
+		return 0;
+	}
+
+	return -ENOSPC;
+}
+
+
+static void discard_jumbo_state(netfront_accel_vnic *vnic) 
+{
+
+	if (vnic->jumbo_state.skb != NULL) {
+		dev_kfree_skb_any(vnic->jumbo_state.skb);
+
+		vnic->jumbo_state.skb = NULL;
+	}
+	vnic->jumbo_state.in_progress = 0;
+}
+
+
+static void  netfront_accel_vi_rx_complete(netfront_accel_vnic *vnic,
+					   struct sk_buff *skb)
+{
+	cuckoo_hash_mac_key key;
+	unsigned long flags;
+	int value;
+	struct net_device *net_dev;
+
+
+	key = cuckoo_mac_to_key(skb->data + ETH_ALEN);
+
+	/*
+	 * If this is a MAC address that we want to do fast path TX
+	 * to, and we don't already, add it to the fastpath table.
+	 * The initial lookup is done without the table lock and so
+	 * may get the wrong answer, but current opinion is that's not
+	 * a big problem
+	 */
+	if (is_valid_ether_addr(skb->data + ETH_ALEN) &&
+	    !cuckoo_hash_lookup(&vnic->fastpath_table, (cuckoo_hash_key *)&key,
+				&value)) {
+		spin_lock_irqsave(&vnic->table_lock, flags);
+		   
+		cuckoo_hash_add_check(&vnic->fastpath_table,
+				      (cuckoo_hash_key *)&key,
+				      1, 1);
+		
+		spin_unlock_irqrestore(&vnic->table_lock, flags);
+	}
+
+	if (compare_ether_addr(skb->data, vnic->mac)) {
+		struct iphdr *ip = (struct iphdr *)(skb->data + ETH_HLEN);
+		u16 port;
+
+		DPRINTK("%s: saw wrong MAC address " MAC_FMT "\n", 
+			__FUNCTION__, MAC_ARG(skb->data));
+
+		if (ip->protocol == IPPROTO_TCP) {
+			struct tcphdr *tcp = (struct tcphdr *)
+				((char *)ip + 4 * ip->ihl);
+			port = tcp->dest;
+		} else {
+			struct udphdr *udp = (struct udphdr *)
+				((char *)ip + 4 * ip->ihl);
+			EPRINTK_ON(ip->protocol != IPPROTO_UDP);
+			port = udp->dest;
+		}
+
+		netfront_accel_msg_tx_fastpath(vnic, skb->data,
+					       ip->daddr, port,
+					       ip->protocol);
+	}
+
+	net_dev = vnic->net_dev;
+	skb->dev = net_dev;
+	skb->protocol = eth_type_trans(skb, net_dev);
+	/* CHECKSUM_UNNECESSARY as hardware has done it already */
+	skb->ip_summed = CHECKSUM_UNNECESSARY;
+
+	if (!netfront_accel_ssr_skb(vnic, &vnic->ssr_state, skb))
+		netif_receive_skb(skb);
+}
+
+
+static int netfront_accel_vi_poll_process_rx(netfront_accel_vnic *vnic, 
+					     ef_event *ev)
+{
+	struct netfront_accel_bufinfo *bufinfo = vnic->rx_bufs;
+	struct netfront_accel_pkt_desc *buf = NULL;
+	struct sk_buff *skb;
+	int id, len, sop = 0, cont = 0;
+
+	VPRINTK("Rx event.\n");
+	/*
+	 * Complete the receive operation, and get the request id of
+	 * the buffer
+	 */
+	id = ef_vi_receive_done(&vnic->vi, ev);
+
+	if (id < 0 || id >= bufinfo->npages*NETFRONT_ACCEL_BUFS_PER_PAGE) {
+		EPRINTK("Rx packet %d is invalid\n", id);
+		/* Carry on round the loop if more events */
+		goto bad_packet;
+	}
+	/* Get our buffer descriptor */
+	buf = netfront_accel_buf_find(bufinfo, id);
+
+	len = EF_EVENT_RX_BYTES(*ev);
+
+	/* An RX buffer has been removed from the DMA ring. */
+	vnic->rx_dma_level--;
+
+	if (EF_EVENT_TYPE(*ev) == EF_EVENT_TYPE_RX) {
+		sop = EF_EVENT_RX_SOP(*ev);
+		cont = EF_EVENT_RX_CONT(*ev);
+
+		skb = vnic->jumbo_state.skb;
+
+		VPRINTK("Rx packet %d: %d bytes so far; sop %d; cont %d\n", 
+			id, len, sop, cont);
+
+		if (sop) {
+			if (!vnic->jumbo_state.in_progress) {
+				vnic->jumbo_state.in_progress = 1;
+				BUG_ON(vnic->jumbo_state.skb != NULL);
+			} else {
+				/*
+				 * This fragment shows a missing tail in 
+				 * previous one, but is itself possibly OK
+				 */
+				DPRINTK("sop and in_progress => no tail\n");
+
+				/* Release the socket buffer we already had */
+				discard_jumbo_state(vnic);
+
+				/* Now start processing this fragment */
+				vnic->jumbo_state.in_progress = 1;
+				skb = NULL;
+			}
+		} else if (!vnic->jumbo_state.in_progress) {
+			DPRINTK("!sop and !in_progress => missing head\n");
+			goto missing_head;
+		}
+
+		if (!cont) {
+			/* Update state for next time */
+			vnic->jumbo_state.in_progress = 0;
+			vnic->jumbo_state.skb = NULL;
+		} else if (!vnic->jumbo_state.in_progress) {
+			DPRINTK("cont and !in_progress => missing head\n");
+			goto missing_head;
+		}
+
+		if (skb == NULL) {
+			BUG_ON(!sop);
+
+			if (!cont)
+				skb = alloc_skb(len+NET_IP_ALIGN, GFP_ATOMIC);
+			else
+				skb = alloc_skb(vnic->net_dev->mtu+NET_IP_ALIGN, 
+						GFP_ATOMIC);
+
+			if (skb == NULL) {
+				DPRINTK("%s: Couldn't get an rx skb.\n",
+					__FUNCTION__);
+				netfront_accel_vi_post_rx_or_free(vnic, (u16)id, buf);
+				/*
+				 * Dropping this fragment means we
+				 * should discard the rest too
+				 */
+				discard_jumbo_state(vnic);
+
+				/* Carry on round the loop if more events */
+				return 0;
+			}
+
+		}
+		
+		/* Copy the data to required end destination */
+		if (ef_vnic_copy_to_skb(vnic, skb, buf, len) != 0) {
+			/*
+			 * No space in the skb - suggests > MTU packet
+			 * received
+			 */
+			EPRINTK("%s: Rx packet too large (%d)\n",
+				__FUNCTION__, len);
+			netfront_accel_vi_post_rx_or_free(vnic, (u16)id, buf);
+			discard_jumbo_state(vnic);
+			return 0;
+		}
+		
+		/* Put the buffer back in the DMA queue. */
+		netfront_accel_vi_post_rx_or_free(vnic, (u16)id, buf);
+
+		if (cont) {
+			vnic->jumbo_state.skb = skb;
+
+			return 0;
+		} else {
+			/* Track number of rx fastpath packets */
+			vnic->netdev_stats.fastpath_rx_pkts++;
+			vnic->netdev_stats.fastpath_rx_bytes += len;
+
+			netfront_accel_vi_rx_complete(vnic, skb);
+
+			return 1;
+		}
+	} else {
+		BUG_ON(EF_EVENT_TYPE(*ev) != EF_EVENT_TYPE_RX_DISCARD);
+
+		if (EF_EVENT_RX_DISCARD_TYPE(*ev) 
+		    == EF_EVENT_RX_DISCARD_TRUNC) {
+			DPRINTK("%s: " EF_EVENT_FMT 
+				" buffer %d FRM_TRUNC q_id %d\n",
+				__FUNCTION__, EF_EVENT_PRI_ARG(*ev), id,
+				EF_EVENT_RX_DISCARD_Q_ID(*ev) );
+			NETFRONT_ACCEL_STATS_OP(++vnic->stats.fastpath_frm_trunc);
+		} else if (EF_EVENT_RX_DISCARD_TYPE(*ev) 
+			  == EF_EVENT_RX_DISCARD_OTHER) {
+			DPRINTK("%s: " EF_EVENT_FMT 
+				" buffer %d RX_DISCARD_OTHER q_id %d\n",
+				__FUNCTION__, EF_EVENT_PRI_ARG(*ev), id,
+				EF_EVENT_RX_DISCARD_Q_ID(*ev) );
+			/*
+			 * Probably tail of packet for which error has
+			 * already been logged, so don't count in
+			 * stats
+			 */
+		} else {
+			EPRINTK("%s: " EF_EVENT_FMT 
+				" buffer %d rx discard type %d q_id %d\n",
+				__FUNCTION__, EF_EVENT_PRI_ARG(*ev), id,
+				EF_EVENT_RX_DISCARD_TYPE(*ev), 
+				EF_EVENT_RX_DISCARD_Q_ID(*ev) );
+			NETFRONT_ACCEL_STATS_OP(++vnic->stats.bad_event_count);
+		}
+	}
+
+	/* discard type drops through here */
+
+bad_packet:
+	/* Release the socket buffer we already had */
+	discard_jumbo_state(vnic);
+
+missing_head:
+	BUG_ON(vnic->jumbo_state.in_progress != 0);
+	BUG_ON(vnic->jumbo_state.skb != NULL);
+
+	if (id >= 0 && id < bufinfo->npages*NETFRONT_ACCEL_BUFS_PER_PAGE)
+		/* Put the buffer back in the DMA queue. */
+		netfront_accel_vi_post_rx_or_free(vnic, (u16)id, buf);
+
+	vnic->netdev_stats.fastpath_rx_errors++;
+
+	DPRINTK("%s experienced bad packet/missing fragment error: %d \n",
+		__FUNCTION__, ev->rx.flags);
+
+	return 0;
+}
+
+
+static void netfront_accel_vi_not_busy(netfront_accel_vnic *vnic)
+{
+	struct netfront_info *np = ((struct netfront_info *)
+				    netdev_priv(vnic->net_dev));
+	struct sk_buff *skb;
+	int handled;
+	unsigned long flags;
+	
+	/*
+	 * TODO if we could safely check tx_skb == NULL and return
+	 * early without taking the lock, that would obviously help
+	 * performance
+	 */
+
+	/* Take the netfront lock which protects tx_skb. */
+	spin_lock_irqsave(&np->tx_lock, flags);
+	if (vnic->tx_skb != NULL) {
+		DPRINTK("%s trying to send spare buffer\n", __FUNCTION__);
+		
+		skb = vnic->tx_skb;
+		vnic->tx_skb = NULL;
+
+		spin_unlock_irqrestore(&np->tx_lock, flags);
+
+		handled = netfront_accel_vi_tx_post(vnic, skb);
+		
+		spin_lock_irqsave(&np->tx_lock, flags);
+
+		if (handled != NETFRONT_ACCEL_STATUS_BUSY) {
+			DPRINTK("%s restarting tx\n", __FUNCTION__);
+			if (netfront_check_queue_ready(vnic->net_dev)) {
+				netif_wake_queue(vnic->net_dev);
+				NETFRONT_ACCEL_STATS_OP
+					(vnic->stats.queue_wakes++);
+			}
+		} else {
+			vnic->tx_skb = skb;
+		}
+		
+		/*
+		 * Should never get a CANT, as it checks that before
+		 * deciding it was BUSY first time round 
+		 */
+		BUG_ON(handled == NETFRONT_ACCEL_STATUS_CANT);
+	}
+	spin_unlock_irqrestore(&np->tx_lock, flags);
+}
+
+
+static void netfront_accel_vi_tx_complete(netfront_accel_vnic *vnic, 
+					  struct netfront_accel_tso_buffer *tso_buf,
+					  int is_last)
+{
+	struct netfront_accel_tso_buffer *next;
+
+	/* 
+	 * We get a single completion for every call to
+	 * ef_vi_transmitv so handle any other buffers which are part
+	 * of the same packet 
+	 */
+	while (tso_buf != NULL) {
+		if (tso_buf->buf->skb != NULL) {
+			dev_kfree_skb_any(tso_buf->buf->skb);
+			tso_buf->buf->skb = NULL;
+		}
+
+		next = tso_buf->next;
+
+		netfront_accel_buf_put(vnic->tx_bufs, tso_buf->buf->buf_id);
+
+		tso_buf = next;
+	}
+
+	/*
+	 * If this was the last one in the batch, we try and send any
+	 * pending tx_skb. There should now be buffers and
+	 * descriptors
+	 */
+	if (is_last)
+		netfront_accel_vi_not_busy(vnic);
+}
+
+
+static void netfront_accel_vi_poll_process_tx(netfront_accel_vnic *vnic,
+					      ef_event *ev)
+{
+	struct netfront_accel_pkt_desc *buf;
+	struct netfront_accel_tso_buffer *tso_buf;
+	ef_request_id ids[EF_VI_TRANSMIT_BATCH];
+	int i, n_ids;
+	unsigned long flags;
+
+	/* Get the request ids for this tx completion event. */
+	n_ids = ef_vi_transmit_unbundle(&vnic->vi, ev, ids);
+
+	/* Take the tx buffer spin lock and hold for the duration */
+	spin_lock_irqsave(&vnic->tx_lock, flags);
+
+	for (i = 0; i < n_ids; ++i) {
+		VPRINTK("Tx packet %d complete\n", ids[i]);
+		buf = netfront_accel_buf_find(vnic->tx_bufs, ids[i]);
+		NETFRONT_ACCEL_STATS_OP(vnic->stats.fastpath_tx_completions++);
+
+		tso_buf = (struct netfront_accel_tso_buffer *)
+			(buf->pkt_kva + NETFRONT_ACCEL_TX_BUF_LENGTH);
+		BUG_ON(tso_buf->buf != buf);
+
+		netfront_accel_vi_tx_complete(vnic, tso_buf, i == (n_ids-1));
+	}
+
+	spin_unlock_irqrestore(&vnic->tx_lock, flags);
+}
+
+
+int netfront_accel_vi_poll(netfront_accel_vnic *vnic, int rx_packets)
+{
+	ef_event ev[ACCEL_VI_POLL_EVENTS];
+	int rx_remain = rx_packets, rc, events, i;
+#if NETFRONT_ACCEL_STATS
+	int n_evs_polled = 0, rx_evs_polled = 0, tx_evs_polled = 0;
+#endif
+	BUG_ON(rx_packets <= 0);
+
+	events = ef_eventq_poll(&vnic->vi, ev, 
+				min(rx_remain, ACCEL_VI_POLL_EVENTS));
+	i = 0;
+	NETFRONT_ACCEL_STATS_OP(n_evs_polled += events);
+
+	VPRINTK("%s: %d events\n", __FUNCTION__, events);
+
+	/* Loop over each event */
+	while (events) {
+		VPRINTK("%s: Event "EF_EVENT_FMT", index %lu\n", __FUNCTION__, 
+			EF_EVENT_PRI_ARG(ev[i]),	
+			(unsigned long)(vnic->vi.evq_state->evq_ptr));
+
+		if ((EF_EVENT_TYPE(ev[i]) == EF_EVENT_TYPE_RX) ||
+		    (EF_EVENT_TYPE(ev[i]) == EF_EVENT_TYPE_RX_DISCARD)) {
+			rc = netfront_accel_vi_poll_process_rx(vnic, &ev[i]);
+			rx_remain -= rc;
+			BUG_ON(rx_remain < 0);
+			NETFRONT_ACCEL_STATS_OP(rx_evs_polled++);
+		} else if (EF_EVENT_TYPE(ev[i]) == EF_EVENT_TYPE_TX) {
+			netfront_accel_vi_poll_process_tx(vnic, &ev[i]);
+			NETFRONT_ACCEL_STATS_OP(tx_evs_polled++);
+		} else if (EF_EVENT_TYPE(ev[i]) == 
+			   EF_EVENT_TYPE_RX_NO_DESC_TRUNC) {
+			DPRINTK("%s: RX_NO_DESC_TRUNC " EF_EVENT_FMT "\n",
+				__FUNCTION__, EF_EVENT_PRI_ARG(ev[i]));
+			discard_jumbo_state(vnic);
+			NETFRONT_ACCEL_STATS_OP(vnic->stats.rx_no_desc_trunc++);
+		} else {
+			EPRINTK("Unexpected event " EF_EVENT_FMT "\n", 
+				EF_EVENT_PRI_ARG(ev[i]));
+			NETFRONT_ACCEL_STATS_OP(vnic->stats.bad_event_count++);
+		}
+
+		i++;
+
+		/* Carry on round the loop if more events and more space */
+		if (i == events) {
+			if (rx_remain == 0)
+				break;
+
+			events = ef_eventq_poll(&vnic->vi, ev, 
+						min(rx_remain, 
+						    ACCEL_VI_POLL_EVENTS));
+			i = 0;
+			NETFRONT_ACCEL_STATS_OP(n_evs_polled += events);
+		}
+	}
+	
+#if NETFRONT_ACCEL_STATS
+	vnic->stats.event_count += n_evs_polled;
+	vnic->stats.event_count_since_irq += n_evs_polled;
+	if (n_evs_polled > vnic->stats.events_per_poll_max)
+		vnic->stats.events_per_poll_max = n_evs_polled;
+	if (rx_evs_polled > vnic->stats.events_per_poll_rx_max)
+		vnic->stats.events_per_poll_rx_max = rx_evs_polled;
+	if (tx_evs_polled > vnic->stats.events_per_poll_tx_max)
+		vnic->stats.events_per_poll_tx_max = tx_evs_polled;
+#endif
+
+	return rx_packets - rx_remain;
+}
+
+
+int netfront_accel_vi_enable_interrupts(netfront_accel_vnic *vnic)
+{
+	u32 sw_evq_ptr;
+
+	VPRINTK("%s: checking for event on %p\n", __FUNCTION__, &vnic->vi.evq_state);
+
+	BUG_ON(vnic == NULL);
+	BUG_ON(vnic->vi.evq_state == NULL);
+
+	/* Do a quick check for an event. */
+	if (ef_eventq_has_event(&vnic->vi)) {
+		VPRINTK("%s: found event\n",  __FUNCTION__);
+		return 0;
+	}
+
+	VPRINTK("evq_ptr=0x%08x	 evq_mask=0x%08x\n",
+		vnic->evq_state.evq_ptr, vnic->vi.evq_mask);
+  
+	/* Request a wakeup from the hardware. */
+	sw_evq_ptr = vnic->evq_state.evq_ptr & vnic->vi.evq_mask;
+
+	BUG_ON(vnic->hw.falcon.evq_rptr == NULL);
+
+	VPRINTK("Requesting wakeup at 0x%08x, rptr %p\n", sw_evq_ptr,
+		vnic->hw.falcon.evq_rptr);
+	*(volatile u32 *)(vnic->hw.falcon.evq_rptr) = (sw_evq_ptr >> 3);
+
+	return 1;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/accel_xenbus.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_xenbus.c
--- linux-2.6.18.8/drivers/xen/sfc_netfront/accel_xenbus.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/accel_xenbus.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,776 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/stddef.h>
+#include <linux/errno.h>
+
+#include <xen/xenbus.h>
+#include <xen/evtchn.h>
+#include <xen/gnttab.h>
+
+#include "accel.h"
+#include "accel_util.h"
+#include "accel_msg_iface.h"
+#include "accel_bufs.h"
+#include "accel_ssr.h"
+/* drivers/xen/netfront/netfront.h */
+#include "netfront.h"
+
+void netfront_accel_set_closing(netfront_accel_vnic *vnic) 
+{
+
+	vnic->frontend_state = XenbusStateClosing;
+	net_accel_update_state(vnic->dev, XenbusStateClosing);
+}
+	
+
+static void mac_address_change(struct xenbus_watch *watch,
+			       const char **vec, unsigned int len)
+{
+	netfront_accel_vnic *vnic;
+	struct xenbus_device *dev;
+	int rc;
+
+	DPRINTK("%s\n", __FUNCTION__);
+	
+	vnic = container_of(watch, netfront_accel_vnic, 
+				mac_address_watch);
+	dev = vnic->dev;
+
+	rc = net_accel_xen_net_read_mac(dev, vnic->mac);
+
+	if (rc != 0)
+		EPRINTK("%s: failed to read mac (%d)\n", __FUNCTION__, rc);
+}
+
+
+static int setup_mac_address_watch(struct xenbus_device *dev,
+				   netfront_accel_vnic *vnic)
+{
+	int err;
+
+	DPRINTK("Setting watch on %s/%s\n", dev->nodename, "mac");
+
+	err = xenbus_watch_path2(dev, dev->nodename, "mac", 
+				 &vnic->mac_address_watch, 
+				 mac_address_change);
+	if (err) {
+		EPRINTK("%s: Failed to register xenbus watch: %d\n",
+			__FUNCTION__, err);
+		goto fail;
+	}
+
+	return 0;
+ fail:
+	vnic->mac_address_watch.node = NULL;
+	return err;
+}
+
+
+/* Grant access to some pages and publish through xenbus */
+static int make_named_grant(struct xenbus_device *dev, void *page, 
+			    const char *name, grant_ref_t *gnt_ref)
+{
+	struct xenbus_transaction tr;
+	int err;
+	grant_ref_t gnt;
+
+	gnt = net_accel_grant_page(dev, virt_to_mfn(page), 0);
+	if (gnt < 0)
+		return gnt;
+
+	do {
+		err = xenbus_transaction_start(&tr);
+		if (err != 0) {
+			EPRINTK("%s: transaction start failed %d\n",
+				__FUNCTION__, err);
+			return err;
+		}
+		err = xenbus_printf(tr, dev->nodename, name, "%d", gnt);
+		if (err != 0) {
+			EPRINTK("%s: xenbus_printf failed %d\n", __FUNCTION__,
+				err);
+			xenbus_transaction_end(tr, 1);
+			return err;
+		}
+		err = xenbus_transaction_end(tr, 0);
+	} while (err == -EAGAIN);
+	
+	if (err != 0) {
+		EPRINTK("%s: transaction end failed %d\n", __FUNCTION__, err);
+		return err;
+	}
+	
+	*gnt_ref = gnt;
+
+	return 0;
+}
+
+
+static int remove_named_grant(struct xenbus_device *dev,
+			      const char *name, grant_ref_t gnt_ref)
+{
+	struct xenbus_transaction tr;
+	int err;
+
+	net_accel_ungrant_page(gnt_ref);
+
+	do {
+		err = xenbus_transaction_start(&tr);
+		if (err != 0) {
+			EPRINTK("%s: transaction start failed %d\n",
+				__FUNCTION__, err);
+			return err;
+		}
+		err = xenbus_rm(tr, dev->nodename, name);
+		if (err != 0) {
+			EPRINTK("%s: xenbus_rm failed %d\n", __FUNCTION__,
+				err);
+			xenbus_transaction_end(tr, 1);
+			return err;
+		}
+		err = xenbus_transaction_end(tr, 0);
+	} while (err == -EAGAIN);
+	
+	if (err != 0) {
+		EPRINTK("%s: transaction end failed %d\n", __FUNCTION__, err);
+		return err;
+	}
+
+	return 0;
+}
+
+
+static 
+netfront_accel_vnic *netfront_accel_vnic_ctor(struct net_device *net_dev,
+					      struct xenbus_device *dev)
+{
+	struct netfront_info *np =
+		(struct netfront_info *)netdev_priv(net_dev);
+	netfront_accel_vnic *vnic;
+	int err;
+
+	/*
+	 * A bug in earlier versions of Xen accel plugin system meant
+	 * you could be probed twice for the same device on suspend
+	 * cancel.  Be tolerant of that.
+	 */ 
+	if (np->accel_priv != NULL)
+		return ERR_PTR(-EALREADY);
+
+	/* Alloc mem for state */
+	vnic = kzalloc(sizeof(netfront_accel_vnic), GFP_KERNEL);
+	if (vnic == NULL) {
+		EPRINTK("%s: no memory for vnic state\n", __FUNCTION__);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	spin_lock_init(&vnic->tx_lock);
+
+	mutex_init(&vnic->vnic_mutex);
+	mutex_lock(&vnic->vnic_mutex);
+
+	/* Store so state can be retrieved from device */
+	BUG_ON(np->accel_priv != NULL);
+	np->accel_priv = vnic;
+	vnic->dev = dev;
+	vnic->net_dev = net_dev;
+	spin_lock_init(&vnic->irq_enabled_lock);
+	netfront_accel_ssr_init(&vnic->ssr_state);
+
+	init_waitqueue_head(&vnic->state_wait_queue);
+	vnic->backend_state = XenbusStateUnknown;
+	vnic->frontend_state = XenbusStateClosed;
+	vnic->removing = 0;
+	vnic->domU_state_is_setup = 0;
+	vnic->dom0_state_is_setup = 0;
+	vnic->poll_enabled = 0;
+	vnic->tx_enabled = 0;
+	vnic->tx_skb = NULL;
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,20)
+	INIT_WORK(&vnic->msg_from_bend, netfront_accel_msg_from_bend);
+#else
+	INIT_WORK(&vnic->msg_from_bend, netfront_accel_msg_from_bend, vnic);
+#endif
+
+	netfront_accel_debugfs_create(vnic);
+
+	mutex_unlock(&vnic->vnic_mutex);
+
+	err = net_accel_xen_net_read_mac(dev, vnic->mac);
+	if (err) 
+		goto fail_mac;
+
+	/* Setup a watch on the frontend's MAC address */
+	err = setup_mac_address_watch(dev, vnic);
+	if (err)
+		goto fail_mac;
+
+	return vnic;
+
+fail_mac:
+
+	mutex_lock(&vnic->vnic_mutex);
+
+	netfront_accel_debugfs_remove(vnic);
+
+	netfront_accel_ssr_fini(vnic, &vnic->ssr_state);
+
+	EPRINTK_ON(vnic->tx_skb != NULL);
+
+	vnic->frontend_state = XenbusStateUnknown;
+	net_accel_update_state(dev, XenbusStateUnknown);
+
+	mutex_unlock(&vnic->vnic_mutex);
+
+	np->accel_priv = NULL;
+	kfree(vnic);
+
+	return ERR_PTR(err);
+}
+
+
+static void netfront_accel_vnic_dtor(netfront_accel_vnic *vnic)
+{
+	struct net_device *net_dev = vnic->net_dev;
+	struct netfront_info *np = 
+		(struct netfront_info *)netdev_priv(net_dev);
+
+	/*
+	 * Now we don't hold the lock any more it is safe to remove
+	 * this watch and synchonrise with the completion of
+	 * watches
+	 */
+	DPRINTK("%s: unregistering xenbus mac watch\n", __FUNCTION__);
+	unregister_xenbus_watch(&vnic->mac_address_watch);
+	kfree(vnic->mac_address_watch.node);
+
+	flush_workqueue(netfront_accel_workqueue);
+
+	mutex_lock(&vnic->vnic_mutex);
+
+	netfront_accel_debugfs_remove(vnic);
+
+	netfront_accel_ssr_fini(vnic, &vnic->ssr_state);
+
+	EPRINTK_ON(vnic->tx_skb != NULL);
+
+	vnic->frontend_state = XenbusStateUnknown;
+	net_accel_update_state(vnic->dev, XenbusStateUnknown);
+
+	mutex_unlock(&vnic->vnic_mutex);
+
+	np->accel_priv = NULL;
+	kfree(vnic);
+}
+
+
+static int vnic_setup_domU_shared_state(struct xenbus_device *dev,
+					netfront_accel_vnic *vnic)
+{
+	struct xenbus_transaction tr;
+	int err;
+	int msgs_per_queue;
+
+
+	DPRINTK("Setting up domU shared state.\n");
+
+	msgs_per_queue = (PAGE_SIZE/2) / sizeof(struct net_accel_msg);
+
+	/* Allocate buffer state */
+	vnic->tx_bufs = netfront_accel_init_bufs(&vnic->tx_lock);
+	if (vnic->tx_bufs == NULL) {
+		err = -ENOMEM;
+		EPRINTK("%s: Failed to allocate tx buffers\n", __FUNCTION__);
+		goto fail_tx_bufs;
+	}
+
+	vnic->rx_bufs = netfront_accel_init_bufs(NULL);
+	if (vnic->rx_bufs == NULL) {
+		err = -ENOMEM;
+		EPRINTK("%s: Failed to allocate rx buffers\n", __FUNCTION__);
+		goto fail_rx_bufs;
+	}
+
+	/* 
+	 * This allocates two pages, one for the shared page and one
+	 * for the message queue.
+	 */
+	vnic->shared_page = (struct net_accel_shared_page *)
+		__get_free_pages(GFP_KERNEL, 1);
+	if (vnic->shared_page == NULL) {
+		EPRINTK("%s: no memory for shared pages\n", __FUNCTION__);
+		err = -ENOMEM;
+		goto fail_shared_page;
+	}
+
+	net_accel_msg_init_queue
+		(&vnic->from_dom0, &vnic->shared_page->queue0, 
+		 (struct net_accel_msg *)((u8*)vnic->shared_page + PAGE_SIZE),
+		 msgs_per_queue);
+
+	net_accel_msg_init_queue
+		(&vnic->to_dom0, &vnic->shared_page->queue1,
+		 (struct net_accel_msg *)((u8*)vnic->shared_page +
+					  (3 * PAGE_SIZE / 2)),
+		 msgs_per_queue);
+	
+	vnic->msg_state = NETFRONT_ACCEL_MSG_NONE;
+
+	err = make_named_grant(dev, vnic->shared_page, "accel-ctrl-page",
+			       &vnic->ctrl_page_gnt);
+	if (err) {
+		EPRINTK("couldn't make ctrl-page named grant\n");
+		goto fail_ctrl_page_grant;
+	}
+
+	err = make_named_grant(dev, (u8*)vnic->shared_page + PAGE_SIZE,
+			       "accel-msg-page", &vnic->msg_page_gnt);
+	if (err) {
+		EPRINTK("couldn't make msg-page named grant\n");
+		goto fail_msg_page_grant;
+	}
+
+	/* Create xenbus msg event channel */
+	err = bind_listening_port_to_irqhandler
+		(dev->otherend_id, netfront_accel_msg_channel_irq_from_bend,
+		 SA_SAMPLE_RANDOM, "vnicctrl", vnic);
+	if (err < 0) {
+		EPRINTK("Couldn't bind msg event channel\n");
+		goto fail_msg_irq;
+	}
+	vnic->msg_channel_irq = err;
+	vnic->msg_channel = irq_to_evtchn_port(vnic->msg_channel_irq);
+	
+	/* Create xenbus net event channel */
+	err = bind_listening_port_to_irqhandler
+		(dev->otherend_id, netfront_accel_net_channel_irq_from_bend,
+		 SA_SAMPLE_RANDOM, "vnicfront", vnic);
+	if (err < 0) {
+		EPRINTK("Couldn't bind net event channel\n");
+		goto fail_net_irq;
+	}
+	vnic->net_channel_irq = err;
+	vnic->net_channel = irq_to_evtchn_port(vnic->net_channel_irq);
+	/* Want to ensure we don't get interrupts before we're ready */
+	netfront_accel_disable_net_interrupts(vnic);
+
+	DPRINTK("otherend %d has msg ch %u (%u) and net ch %u (%u)\n",
+		dev->otherend_id, vnic->msg_channel, vnic->msg_channel_irq, 
+		vnic->net_channel, vnic->net_channel_irq);
+
+	do {
+		err = xenbus_transaction_start(&tr);
+		if (err != 0) {
+			EPRINTK("%s: Transaction start failed %d\n",
+				__FUNCTION__, err);
+			goto fail_transaction;
+		}
+
+		err = xenbus_printf(tr, dev->nodename, "accel-msg-channel",
+				    "%u", vnic->msg_channel);
+		if (err != 0) {
+			EPRINTK("%s: event channel xenbus write failed %d\n",
+				__FUNCTION__, err);
+			xenbus_transaction_end(tr, 1);
+			goto fail_transaction;
+		}
+
+		err = xenbus_printf(tr, dev->nodename, "accel-net-channel",
+				    "%u", vnic->net_channel);
+		if (err != 0) {
+			EPRINTK("%s: net channel xenbus write failed %d\n",
+				__FUNCTION__, err);
+			xenbus_transaction_end(tr, 1);
+			goto fail_transaction;
+		}
+
+		err = xenbus_transaction_end(tr, 0);
+	} while (err == -EAGAIN);
+
+	if (err != 0) {
+		EPRINTK("%s: Transaction end failed %d\n", __FUNCTION__, err);
+		goto fail_transaction;
+	}
+
+	DPRINTK("Completed setting up domU shared state\n");
+
+	return 0;
+
+fail_transaction:
+
+	unbind_from_irqhandler(vnic->net_channel_irq, vnic);
+fail_net_irq:
+
+	unbind_from_irqhandler(vnic->msg_channel_irq, vnic);
+fail_msg_irq:
+
+	remove_named_grant(dev, "accel-ctrl-page", vnic->ctrl_page_gnt);
+fail_msg_page_grant:
+
+	remove_named_grant(dev, "accel-msg-page", vnic->msg_page_gnt);
+fail_ctrl_page_grant:
+
+	free_pages((unsigned long)vnic->shared_page, 1);
+	vnic->shared_page = NULL;
+fail_shared_page:
+
+	netfront_accel_fini_bufs(vnic->rx_bufs);
+fail_rx_bufs:
+
+	netfront_accel_fini_bufs(vnic->tx_bufs);
+fail_tx_bufs:
+
+	/* Undo the memory allocation created when we got the HELLO */
+	netfront_accel_free_buffer_mem(&vnic->bufpages,
+				       vnic->rx_bufs,
+				       vnic->tx_bufs);
+
+	DPRINTK("Failed to setup domU shared state with code %d\n", err);
+
+	return err;
+}
+
+
+static void vnic_remove_domU_shared_state(struct xenbus_device *dev, 
+					  netfront_accel_vnic *vnic)
+{
+	struct xenbus_transaction tr;
+	
+	/*
+	 * Don't remove any watches because we currently hold the
+	 * mutex and the watches take the mutex.
+	 */
+
+	DPRINTK("%s: removing event channel irq handlers %d %d\n",
+		__FUNCTION__, vnic->net_channel_irq, vnic->msg_channel_irq);
+	do {
+		if (xenbus_transaction_start(&tr) != 0)
+			break;
+		xenbus_rm(tr, dev->nodename, "accel-msg-channel");
+		xenbus_rm(tr, dev->nodename, "accel-net-channel");
+	} while (xenbus_transaction_end(tr, 0) == -EAGAIN);
+
+	unbind_from_irqhandler(vnic->net_channel_irq, vnic);
+	unbind_from_irqhandler(vnic->msg_channel_irq, vnic);
+
+	/* ungrant pages for msg channel */
+	remove_named_grant(dev, "accel-ctrl-page", vnic->ctrl_page_gnt);
+	remove_named_grant(dev, "accel-msg-page", vnic->msg_page_gnt);
+	free_pages((unsigned long)vnic->shared_page, 1);
+	vnic->shared_page = NULL;
+
+	/* ungrant pages for buffers, and free buffer memory */
+	netfront_accel_free_buffer_mem(&vnic->bufpages,
+				       vnic->rx_bufs,
+				       vnic->tx_bufs);
+	netfront_accel_fini_bufs(vnic->rx_bufs);
+	netfront_accel_fini_bufs(vnic->tx_bufs);
+}
+
+
+static void vnic_setup_dom0_shared_state(struct xenbus_device *dev,
+					netfront_accel_vnic *vnic)
+{
+	DPRINTK("Setting up dom0 shared state\n");
+
+	netfront_accel_vi_ctor(vnic);
+
+	/*
+	 * Message processing will be enabled when this function
+	 * returns, but we might have missed an interrupt.  Schedule a
+	 * check just in case.
+	 */
+	queue_work(netfront_accel_workqueue, &vnic->msg_from_bend);
+}
+
+
+static void vnic_remove_dom0_shared_state(struct xenbus_device *dev,
+					  netfront_accel_vnic *vnic)
+{
+	DPRINTK("Removing dom0 shared state\n");
+
+	vnic_stop_fastpath(vnic);
+
+	netfront_accel_vi_dtor(vnic);
+}
+
+
+/*************************************************************************/
+
+/*
+ * The following code handles accelstate changes between the frontend
+ * and the backend.  In response to transitions, calls the following
+ * functions in matching pairs:
+ *
+ *   vnic_setup_domU_shared_state
+ *   vnic_remove_domU_shared_state
+ *
+ *   vnic_setup_dom0_shared_state
+ *   vnic_remove_dom0_shared_state
+ *
+ * Valid state transitions for DomU are as follows:
+ *
+ * Closed->Init       on probe or in response to Init from dom0
+ *
+ * Init->Connected    in response to Init from dom0
+ * Init->Closing      on error providing dom0 is in Init
+ * Init->Closed       on remove or in response to Closing from dom0
+ *
+ * Connected->Closing on error/remove
+ * Connected->Closed  in response to Closing from dom0
+ *
+ * Closing->Closed    in response to Closing from dom0
+ *
+ */
+
+
+/* Function to deal with Xenbus accel state change in backend */
+static void netfront_accel_backend_accel_changed(netfront_accel_vnic *vnic,
+						 XenbusState backend_state)
+{
+	struct xenbus_device *dev = vnic->dev;
+	XenbusState frontend_state;
+	int state;
+
+	DPRINTK("%s: changing from %s to %s. nodename %s, otherend %s\n",
+		__FUNCTION__, xenbus_strstate(vnic->backend_state),
+		xenbus_strstate(backend_state), dev->nodename, dev->otherend);
+
+	/*
+	 * Ignore duplicate state changes.  This can happen if the
+	 * backend changes state twice in quick succession and the
+	 * first watch fires in the frontend after the second
+	 * transition has completed.
+	 */
+	if (vnic->backend_state == backend_state)
+		return;
+
+	vnic->backend_state = backend_state;
+	frontend_state = vnic->frontend_state;
+
+	switch (backend_state) {
+	case XenbusStateInitialising:
+		/*
+		 * It's possible for us to miss the closed state from
+		 * dom0, so do the work here.
+		 */
+		if (vnic->domU_state_is_setup) {
+			vnic_remove_domU_shared_state(dev, vnic);
+			vnic->domU_state_is_setup = 0;
+		}
+
+		if (frontend_state != XenbusStateInitialising) {
+			/* Make sure the backend doesn't go away. */
+			frontend_state = XenbusStateInitialising;
+			net_accel_update_state(dev, frontend_state);
+			xenbus_scanf(XBT_NIL, dev->otherend, "accelstate", "%d", &state);
+			backend_state = (XenbusState)state;
+			if (backend_state != XenbusStateInitialising)
+				break;
+		}
+
+		/* Start the new connection. */
+		if (!vnic->removing) {
+			BUG_ON(vnic->domU_state_is_setup);
+			if (vnic_setup_domU_shared_state(dev, vnic) == 0) {
+				vnic->domU_state_is_setup = 1;
+				frontend_state = XenbusStateConnected;
+			} else
+				frontend_state = XenbusStateClosing;
+		}
+		break;
+	case XenbusStateConnected:
+		if (vnic->domU_state_is_setup &&
+		    !vnic->dom0_state_is_setup) {
+			vnic_setup_dom0_shared_state(dev, vnic);
+			vnic->dom0_state_is_setup = 1;
+		}
+		break;
+	default:
+	case XenbusStateClosing:
+		if (vnic->dom0_state_is_setup) {
+			vnic_remove_dom0_shared_state(dev, vnic);
+			vnic->dom0_state_is_setup = 0;
+		}
+		frontend_state = XenbusStateClosed;
+		break;
+	case XenbusStateUnknown:
+	case XenbusStateClosed:
+		if (vnic->domU_state_is_setup) {
+			vnic_remove_domU_shared_state(dev, vnic);
+			vnic->domU_state_is_setup = 0;
+		}
+		break;
+	}
+
+	if (frontend_state != vnic->frontend_state) {
+		DPRINTK("Switching from state %s (%d) to %s (%d)\n",
+			xenbus_strstate(vnic->frontend_state),
+			vnic->frontend_state,
+			xenbus_strstate(frontend_state), frontend_state);
+		vnic->frontend_state = frontend_state;
+		net_accel_update_state(dev, frontend_state);
+	}
+
+	wake_up(&vnic->state_wait_queue);
+}
+
+
+static void backend_accel_state_change(struct xenbus_watch *watch,
+				       const char **vec, unsigned int len)
+{
+	int state;
+	netfront_accel_vnic *vnic;
+	struct xenbus_device *dev;
+
+	DPRINTK("%s\n", __FUNCTION__);
+
+	vnic = container_of(watch, struct netfront_accel_vnic,
+				backend_accel_watch);
+
+	mutex_lock(&vnic->vnic_mutex);
+
+	dev = vnic->dev;
+
+	state = (int)XenbusStateUnknown;
+	xenbus_scanf(XBT_NIL, dev->otherend, "accelstate", "%d", &state);
+	netfront_accel_backend_accel_changed(vnic, state);
+
+	mutex_unlock(&vnic->vnic_mutex);
+}
+
+
+static int setup_dom0_accel_watch(struct xenbus_device *dev,
+				  netfront_accel_vnic *vnic)
+{
+	int err;
+
+	DPRINTK("Setting watch on %s/%s\n", dev->otherend, "accelstate");
+
+	err = xenbus_watch_path2(dev, dev->otherend, "accelstate", 
+				 &vnic->backend_accel_watch, 
+				 backend_accel_state_change);
+	if (err) {
+		EPRINTK("%s: Failed to register xenbus watch: %d\n",
+			__FUNCTION__, err);
+		goto fail;
+	}
+	return 0;
+ fail:
+	vnic->backend_accel_watch.node = NULL;
+	return err;
+}
+
+
+int netfront_accel_probe(struct net_device *net_dev, struct xenbus_device *dev)
+{
+	netfront_accel_vnic *vnic;
+	int err;
+
+	DPRINTK("Probe passed device %s\n", dev->nodename);
+
+	vnic = netfront_accel_vnic_ctor(net_dev, dev);
+	if (IS_ERR(vnic))
+		return PTR_ERR(vnic);
+
+	/*
+	 * Setup a watch on the backend accel state.  This sets things
+	 * going.
+	 */
+	err = setup_dom0_accel_watch(dev, vnic);
+	if (err) {
+		netfront_accel_vnic_dtor(vnic);
+		EPRINTK("%s: probe failed with code %d\n", __FUNCTION__, err);
+		return err;
+	}
+
+	/*
+	 * Indicate to the other end that we're ready to start unless
+	 * the watch has already fired.
+	 */
+	mutex_lock(&vnic->vnic_mutex);
+	VPRINTK("setup success, updating accelstate\n");
+	if (vnic->frontend_state == XenbusStateClosed) {
+		vnic->frontend_state = XenbusStateInitialising;
+		net_accel_update_state(dev, XenbusStateInitialising);
+	}
+	mutex_unlock(&vnic->vnic_mutex);
+
+	DPRINTK("Probe done device %s\n", dev->nodename);
+
+	return 0;
+}
+
+
+int netfront_accel_remove(struct xenbus_device *dev)
+{
+	struct netfront_info *np =
+		(struct netfront_info *)dev->dev.driver_data;
+	netfront_accel_vnic *vnic = (netfront_accel_vnic *)np->accel_priv;
+
+	DPRINTK("%s %s\n", __FUNCTION__, dev->nodename);
+
+	BUG_ON(vnic == NULL);
+
+	mutex_lock(&vnic->vnic_mutex);
+
+	/* Reject any attempts to connect. */
+	vnic->removing = 1;
+
+	/* Close any existing connection. */
+	if (vnic->frontend_state == XenbusStateConnected) {
+		vnic->frontend_state = XenbusStateClosing;
+		net_accel_update_state(dev, XenbusStateClosing);
+	}
+
+	mutex_unlock(&vnic->vnic_mutex);
+
+	DPRINTK("%s waiting for release of %s\n", __FUNCTION__, dev->nodename);
+
+	/*
+	 * Wait for the xenbus watch to release the shared resources.
+	 * This indicates that dom0 has made the transition
+	 * Closing->Closed or that dom0 was in Closed or Init and no
+	 * resources were mapped.
+	 */
+	wait_event(vnic->state_wait_queue,
+		   !vnic->domU_state_is_setup);
+
+	/*
+	 * Now we don't need this watch anymore it is safe to remove
+	 * it (and so synchronise with it completing if outstanding)
+	 */
+	DPRINTK("%s: unregistering xenbus accel watch\n",
+		__FUNCTION__);
+	unregister_xenbus_watch(&vnic->backend_accel_watch);
+	kfree(vnic->backend_accel_watch.node);
+
+	netfront_accel_vnic_dtor(vnic);
+
+	DPRINTK("%s done %s\n", __FUNCTION__, dev->nodename);
+
+	return 0;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/ef_vi_falcon_core.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/ef_vi_falcon_core.h
--- linux-2.6.18.8/drivers/xen/sfc_netfront/ef_vi_falcon_core.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/ef_vi_falcon_core.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1075 @@
+
+#define  EFVI_FALCON_EXTENDED_P_BAR 1
+
+//////////////---- Bus Interface Unit Registers C Header ----//////////////
+#define IOM_IND_ADR_REG_OFST 0x0 // IO-mapped indirect access address register
+  #define IOM_AUTO_ADR_INC_EN_LBN 16
+  #define IOM_AUTO_ADR_INC_EN_WIDTH 1
+  #define IOM_IND_ADR_LBN 0
+  #define IOM_IND_ADR_WIDTH 16
+#define IOM_IND_DAT_REG_OFST 0x4 // IO-mapped indirect access data register
+  #define IOM_IND_DAT_LBN 0
+  #define IOM_IND_DAT_WIDTH 32
+#define ADR_REGION_REG_KER_OFST 0x0 // Address region register
+#define ADR_REGION_REG_OFST 0x0 // Address region register
+  #define ADR_REGION3_LBN 96
+  #define ADR_REGION3_WIDTH 18
+  #define ADR_REGION2_LBN 64
+  #define ADR_REGION2_WIDTH 18
+  #define ADR_REGION1_LBN 32
+  #define ADR_REGION1_WIDTH 18
+  #define ADR_REGION0_LBN 0
+  #define ADR_REGION0_WIDTH 18
+#define INT_EN_REG_KER_OFST 0x10 // Kernel driver Interrupt enable register
+  #define KER_INT_CHAR_LBN 4
+  #define KER_INT_CHAR_WIDTH 1
+  #define KER_INT_KER_LBN 3
+  #define KER_INT_KER_WIDTH 1
+  #define ILL_ADR_ERR_INT_EN_KER_LBN 2
+  #define ILL_ADR_ERR_INT_EN_KER_WIDTH 1
+  #define SRM_PERR_INT_EN_KER_LBN 1
+  #define SRM_PERR_INT_EN_KER_WIDTH 1
+  #define DRV_INT_EN_KER_LBN 0
+  #define DRV_INT_EN_KER_WIDTH 1
+#define INT_EN_REG_CHAR_OFST 0x20 // Char Driver interrupt enable register
+  #define CHAR_INT_CHAR_LBN 4
+  #define CHAR_INT_CHAR_WIDTH 1
+  #define CHAR_INT_KER_LBN 3
+  #define CHAR_INT_KER_WIDTH 1
+  #define ILL_ADR_ERR_INT_EN_CHAR_LBN 2
+  #define ILL_ADR_ERR_INT_EN_CHAR_WIDTH 1
+  #define SRM_PERR_INT_EN_CHAR_LBN 1
+  #define SRM_PERR_INT_EN_CHAR_WIDTH 1
+  #define DRV_INT_EN_CHAR_LBN 0
+  #define DRV_INT_EN_CHAR_WIDTH 1
+#define INT_ADR_REG_KER_OFST 0x30 // Interrupt host address for Kernel driver
+  #define INT_ADR_KER_LBN 0
+  #define INT_ADR_KER_WIDTH 64
+  #define DRV_INT_KER_LBN 32
+  #define DRV_INT_KER_WIDTH 1
+  #define EV_FF_HALF_INT_KER_LBN 3
+  #define EV_FF_HALF_INT_KER_WIDTH 1
+  #define EV_FF_FULL_INT_KER_LBN 2
+  #define EV_FF_FULL_INT_KER_WIDTH 1
+  #define ILL_ADR_ERR_INT_KER_LBN 1
+  #define ILL_ADR_ERR_INT_KER_WIDTH 1
+  #define SRAM_PERR_INT_KER_LBN 0
+  #define SRAM_PERR_INT_KER_WIDTH 1
+#define INT_ADR_REG_CHAR_OFST 0x40 // Interrupt host address for Char driver
+  #define INT_ADR_CHAR_LBN 0
+  #define INT_ADR_CHAR_WIDTH 64
+  #define DRV_INT_CHAR_LBN 32
+  #define DRV_INT_CHAR_WIDTH 1
+  #define EV_FF_HALF_INT_CHAR_LBN 3
+  #define EV_FF_HALF_INT_CHAR_WIDTH 1
+  #define EV_FF_FULL_INT_CHAR_LBN 2
+  #define EV_FF_FULL_INT_CHAR_WIDTH 1
+  #define ILL_ADR_ERR_INT_CHAR_LBN 1
+  #define ILL_ADR_ERR_INT_CHAR_WIDTH 1
+  #define SRAM_PERR_INT_CHAR_LBN 0
+  #define SRAM_PERR_INT_CHAR_WIDTH 1
+#define INT_ISR0_B0_OFST 0x90 // B0 only
+#define INT_ISR1_B0_OFST 0xA0
+#define INT_ACK_REG_KER_A1_OFST 0x50 // Kernel interrupt acknowledge register
+  #define RESERVED_LBN 0
+  #define RESERVED_WIDTH 32
+#define INT_ACK_REG_CHAR_A1_OFST 0x60 // CHAR interrupt acknowledge register
+  #define RESERVED_LBN 0
+  #define RESERVED_WIDTH 32
+//////////////---- Global CSR Registers C Header ----//////////////
+#define STRAP_REG_KER_OFST 0x200 // ASIC strap status register
+#define STRAP_REG_OFST 0x200 // ASIC strap status register
+  #define ONCHIP_SRAM_LBN 16
+  #define ONCHIP_SRAM_WIDTH 0
+  #define STRAP_ISCSI_EN_LBN 3
+  #define STRAP_ISCSI_EN_WIDTH 1
+  #define STRAP_PINS_LBN 0
+  #define STRAP_PINS_WIDTH 3
+#define GPIO_CTL_REG_KER_OFST 0x210 // GPIO control register
+#define GPIO_CTL_REG_OFST 0x210 // GPIO control register
+  #define GPIO_OEN_LBN 24
+  #define GPIO_OEN_WIDTH 4
+  #define GPIO_OUT_LBN 16
+  #define GPIO_OUT_WIDTH 4
+  #define GPIO_IN_LBN 8
+  #define GPIO_IN_WIDTH 4
+  #define GPIO_PWRUP_VALUE_LBN 0
+  #define GPIO_PWRUP_VALUE_WIDTH 4
+#define GLB_CTL_REG_KER_OFST 0x220 // Global control register
+#define GLB_CTL_REG_OFST 0x220 // Global control register
+  #define SWRST_LBN 0
+  #define SWRST_WIDTH 1
+#define FATAL_INTR_REG_KER_OFST 0x230 // Fatal interrupt register for Kernel
+  #define PCI_BUSERR_INT_KER_EN_LBN 43
+  #define PCI_BUSERR_INT_KER_EN_WIDTH 1
+  #define SRAM_OOB_INT_KER_EN_LBN 42
+  #define SRAM_OOB_INT_KER_EN_WIDTH 1
+  #define BUFID_OOB_INT_KER_EN_LBN 41
+  #define BUFID_OOB_INT_KER_EN_WIDTH 1
+  #define MEM_PERR_INT_KER_EN_LBN 40
+  #define MEM_PERR_INT_KER_EN_WIDTH 1
+  #define RBUF_OWN_INT_KER_EN_LBN 39
+  #define RBUF_OWN_INT_KER_EN_WIDTH 1
+  #define TBUF_OWN_INT_KER_EN_LBN 38
+  #define TBUF_OWN_INT_KER_EN_WIDTH 1
+  #define RDESCQ_OWN_INT_KER_EN_LBN 37
+  #define RDESCQ_OWN_INT_KER_EN_WIDTH 1
+  #define TDESCQ_OWN_INT_KER_EN_LBN 36
+  #define TDESCQ_OWN_INT_KER_EN_WIDTH 1
+  #define EVQ_OWN_INT_KER_EN_LBN 35
+  #define EVQ_OWN_INT_KER_EN_WIDTH 1
+  #define EVFF_OFLO_INT_KER_EN_LBN 34
+  #define EVFF_OFLO_INT_KER_EN_WIDTH 1
+  #define ILL_ADR_INT_KER_EN_LBN 33
+  #define ILL_ADR_INT_KER_EN_WIDTH 1
+  #define SRM_PERR_INT_KER_EN_LBN 32
+  #define SRM_PERR_INT_KER_EN_WIDTH 1
+  #define PCI_BUSERR_INT_KER_LBN 11
+  #define PCI_BUSERR_INT_KER_WIDTH 1
+  #define SRAM_OOB_INT_KER_LBN 10
+  #define SRAM_OOB_INT_KER_WIDTH 1
+  #define BUFID_OOB_INT_KER_LBN 9
+  #define BUFID_OOB_INT_KER_WIDTH 1
+  #define MEM_PERR_INT_KER_LBN 8
+  #define MEM_PERR_INT_KER_WIDTH 1
+  #define RBUF_OWN_INT_KER_LBN 7
+  #define RBUF_OWN_INT_KER_WIDTH 1
+  #define TBUF_OWN_INT_KER_LBN 6
+  #define TBUF_OWN_INT_KER_WIDTH 1
+  #define RDESCQ_OWN_INT_KER_LBN 5
+  #define RDESCQ_OWN_INT_KER_WIDTH 1
+  #define TDESCQ_OWN_INT_KER_LBN 4
+  #define TDESCQ_OWN_INT_KER_WIDTH 1
+  #define EVQ_OWN_INT_KER_LBN 3
+  #define EVQ_OWN_INT_KER_WIDTH 1
+  #define EVFF_OFLO_INT_KER_LBN 2
+  #define EVFF_OFLO_INT_KER_WIDTH 1
+  #define ILL_ADR_INT_KER_LBN 1
+  #define ILL_ADR_INT_KER_WIDTH 1
+  #define SRM_PERR_INT_KER_LBN 0
+  #define SRM_PERR_INT_KER_WIDTH 1
+#define FATAL_INTR_REG_OFST 0x240 // Fatal interrupt register for Char
+  #define PCI_BUSERR_INT_CHAR_EN_LBN 43
+  #define PCI_BUSERR_INT_CHAR_EN_WIDTH 1
+  #define SRAM_OOB_INT_CHAR_EN_LBN 42
+  #define SRAM_OOB_INT_CHAR_EN_WIDTH 1
+  #define BUFID_OOB_INT_CHAR_EN_LBN 41
+  #define BUFID_OOB_INT_CHAR_EN_WIDTH 1
+  #define MEM_PERR_INT_CHAR_EN_LBN 40
+  #define MEM_PERR_INT_CHAR_EN_WIDTH 1
+  #define RBUF_OWN_INT_CHAR_EN_LBN 39
+  #define RBUF_OWN_INT_CHAR_EN_WIDTH 1
+  #define TBUF_OWN_INT_CHAR_EN_LBN 38
+  #define TBUF_OWN_INT_CHAR_EN_WIDTH 1
+  #define RDESCQ_OWN_INT_CHAR_EN_LBN 37
+  #define RDESCQ_OWN_INT_CHAR_EN_WIDTH 1
+  #define TDESCQ_OWN_INT_CHAR_EN_LBN 36
+  #define TDESCQ_OWN_INT_CHAR_EN_WIDTH 1
+  #define EVQ_OWN_INT_CHAR_EN_LBN 35
+  #define EVQ_OWN_INT_CHAR_EN_WIDTH 1
+  #define EVFF_OFLO_INT_CHAR_EN_LBN 34
+  #define EVFF_OFLO_INT_CHAR_EN_WIDTH 1
+  #define ILL_ADR_INT_CHAR_EN_LBN 33
+  #define ILL_ADR_INT_CHAR_EN_WIDTH 1
+  #define SRM_PERR_INT_CHAR_EN_LBN 32
+  #define SRM_PERR_INT_CHAR_EN_WIDTH 1
+  #define FATAL_INTR_REG_EN_BITS    0xffffffffffffffffULL
+  #define PCI_BUSERR_INT_CHAR_LBN 11
+  #define PCI_BUSERR_INT_CHAR_WIDTH 1
+  #define SRAM_OOB_INT_CHAR_LBN 10
+  #define SRAM_OOB_INT_CHAR_WIDTH 1
+  #define BUFID_OOB_INT_CHAR_LBN 9
+  #define BUFID_OOB_INT_CHAR_WIDTH 1
+  #define MEM_PERR_INT_CHAR_LBN 8
+  #define MEM_PERR_INT_CHAR_WIDTH 1
+  #define RBUF_OWN_INT_CHAR_LBN 7
+  #define RBUF_OWN_INT_CHAR_WIDTH 1
+  #define TBUF_OWN_INT_CHAR_LBN 6
+  #define TBUF_OWN_INT_CHAR_WIDTH 1
+  #define RDESCQ_OWN_INT_CHAR_LBN 5
+  #define RDESCQ_OWN_INT_CHAR_WIDTH 1
+  #define TDESCQ_OWN_INT_CHAR_LBN 4
+  #define TDESCQ_OWN_INT_CHAR_WIDTH 1
+  #define EVQ_OWN_INT_CHAR_LBN 3
+  #define EVQ_OWN_INT_CHAR_WIDTH 1
+  #define EVFF_OFLO_INT_CHAR_LBN 2
+  #define EVFF_OFLO_INT_CHAR_WIDTH 1
+  #define ILL_ADR_INT_CHAR_LBN 1
+  #define ILL_ADR_INT_CHAR_WIDTH 1
+  #define SRM_PERR_INT_CHAR_LBN 0
+  #define SRM_PERR_INT_CHAR_WIDTH 1
+#define DP_CTRL_REG_OFST 0x250 // Datapath control register
+  #define FLS_EVQ_ID_LBN 0
+  #define FLS_EVQ_ID_WIDTH 12
+#define MEM_STAT_REG_KER_OFST 0x260 // Memory status register
+#define MEM_STAT_REG_OFST 0x260 // Memory status register
+  #define MEM_PERR_VEC_LBN 53
+  #define MEM_PERR_VEC_WIDTH 38
+  #define MBIST_CORR_LBN 38
+  #define MBIST_CORR_WIDTH 15
+  #define MBIST_ERR_LBN 0
+  #define MBIST_ERR_WIDTH 38
+#define DEBUG_REG_KER_OFST 0x270 // Debug register
+#define DEBUG_REG_OFST 0x270 // Debug register
+  #define DEBUG_BLK_SEL2_LBN 47
+  #define DEBUG_BLK_SEL2_WIDTH 3
+  #define DEBUG_BLK_SEL1_LBN 44
+  #define DEBUG_BLK_SEL1_WIDTH 3
+  #define DEBUG_BLK_SEL0_LBN 41
+  #define DEBUG_BLK_SEL0_WIDTH 3
+  #define MISC_DEBUG_ADDR_LBN 36
+  #define MISC_DEBUG_ADDR_WIDTH 5
+  #define SERDES_DEBUG_ADDR_LBN 31
+  #define SERDES_DEBUG_ADDR_WIDTH 5
+  #define EM_DEBUG_ADDR_LBN 26
+  #define EM_DEBUG_ADDR_WIDTH 5
+  #define SR_DEBUG_ADDR_LBN 21
+  #define SR_DEBUG_ADDR_WIDTH 5
+  #define EV_DEBUG_ADDR_LBN 16
+  #define EV_DEBUG_ADDR_WIDTH 5
+  #define RX_DEBUG_ADDR_LBN 11
+  #define RX_DEBUG_ADDR_WIDTH 5
+  #define TX_DEBUG_ADDR_LBN 6
+  #define TX_DEBUG_ADDR_WIDTH 5
+  #define BIU_DEBUG_ADDR_LBN 1
+  #define BIU_DEBUG_ADDR_WIDTH 5
+  #define DEBUG_EN_LBN 0
+  #define DEBUG_EN_WIDTH 1
+#define DRIVER_REG0_KER_OFST 0x280 // Driver scratch register 0
+#define DRIVER_REG0_OFST 0x280 // Driver scratch register 0
+  #define DRIVER_DW0_LBN 0
+  #define DRIVER_DW0_WIDTH 32
+#define DRIVER_REG1_KER_OFST 0x290 // Driver scratch register 1
+#define DRIVER_REG1_OFST 0x290 // Driver scratch register 1
+  #define DRIVER_DW1_LBN 0
+  #define DRIVER_DW1_WIDTH 32
+#define DRIVER_REG2_KER_OFST 0x2A0 // Driver scratch register 2
+#define DRIVER_REG2_OFST 0x2A0 // Driver scratch register 2
+  #define DRIVER_DW2_LBN 0
+  #define DRIVER_DW2_WIDTH 32
+#define DRIVER_REG3_KER_OFST 0x2B0 // Driver scratch register 3
+#define DRIVER_REG3_OFST 0x2B0 // Driver scratch register 3
+  #define DRIVER_DW3_LBN 0
+  #define DRIVER_DW3_WIDTH 32
+#define DRIVER_REG4_KER_OFST 0x2C0 // Driver scratch register 4
+#define DRIVER_REG4_OFST 0x2C0 // Driver scratch register 4
+  #define DRIVER_DW4_LBN 0
+  #define DRIVER_DW4_WIDTH 32
+#define DRIVER_REG5_KER_OFST 0x2D0 // Driver scratch register 5
+#define DRIVER_REG5_OFST 0x2D0 // Driver scratch register 5
+  #define DRIVER_DW5_LBN 0
+  #define DRIVER_DW5_WIDTH 32
+#define DRIVER_REG6_KER_OFST 0x2E0 // Driver scratch register 6
+#define DRIVER_REG6_OFST 0x2E0 // Driver scratch register 6
+  #define DRIVER_DW6_LBN 0
+  #define DRIVER_DW6_WIDTH 32
+#define DRIVER_REG7_KER_OFST 0x2F0 // Driver scratch register 7
+#define DRIVER_REG7_OFST 0x2F0 // Driver scratch register 7
+  #define DRIVER_DW7_LBN 0
+  #define DRIVER_DW7_WIDTH 32
+#define ALTERA_BUILD_REG_OFST 0x300 // Altera build register
+#define ALTERA_BUILD_REG_OFST 0x300 // Altera build register
+  #define ALTERA_BUILD_VER_LBN 0
+  #define ALTERA_BUILD_VER_WIDTH 32
+
+/* so called CSR spare register 
+    - contains separate parity enable bits for the various internal memory blocks */
+#define MEM_PARITY_ERR_EN_REG_KER 0x310 
+#define MEM_PARITY_ALL_BLOCKS_EN_LBN 64
+#define MEM_PARITY_ALL_BLOCKS_EN_WIDTH 38
+#define MEM_PARITY_TX_DATA_EN_LBN   72
+#define MEM_PARITY_TX_DATA_EN_WIDTH 2
+
+//////////////---- Event & Timer Module Registers C Header ----//////////////
+
+#if EFVI_FALCON_EXTENDED_P_BAR
+#define EVQ_RPTR_REG_KER_OFST 0x11B00 // Event queue read pointer register
+#else
+#define EVQ_RPTR_REG_KER_OFST 0x1B00 // Event queue read pointer register
+#endif
+
+#define EVQ_RPTR_REG_OFST 0xFA0000 // Event queue read pointer register array.
+  #define EVQ_RPTR_LBN 0
+  #define EVQ_RPTR_WIDTH 15
+
+#if EFVI_FALCON_EXTENDED_P_BAR
+#define EVQ_PTR_TBL_KER_OFST 0x11A00 // Event queue pointer table for kernel access
+#else
+#define EVQ_PTR_TBL_KER_OFST 0x1A00 // Event queue pointer table for kernel access
+#endif
+
+#define EVQ_PTR_TBL_CHAR_OFST 0xF60000 // Event queue pointer table for char direct access
+  #define EVQ_WKUP_OR_INT_EN_LBN 39
+  #define EVQ_WKUP_OR_INT_EN_WIDTH 1
+  #define EVQ_NXT_WPTR_LBN 24
+  #define EVQ_NXT_WPTR_WIDTH 15
+  #define EVQ_EN_LBN 23
+  #define EVQ_EN_WIDTH 1
+  #define EVQ_SIZE_LBN 20
+  #define EVQ_SIZE_WIDTH 3
+  #define EVQ_BUF_BASE_ID_LBN 0
+  #define EVQ_BUF_BASE_ID_WIDTH 20
+#define TIMER_CMD_REG_KER_OFST 0x420 // Timer table for kernel access. Page-mapped
+#define TIMER_CMD_REG_PAGE4_OFST 0x8420 // Timer table for user-level access. Page-mapped. For lowest 1K queues.
+#define TIMER_CMD_REG_PAGE123K_OFST 0x1000420 // Timer table for user-level access. Page-mapped. For upper 3K queues.
+#define TIMER_TBL_OFST 0xF70000 // Timer table for char driver direct access
+  #define TIMER_MODE_LBN 12
+  #define TIMER_MODE_WIDTH 2
+  #define TIMER_VAL_LBN 0
+  #define TIMER_VAL_WIDTH 12
+  #define TIMER_MODE_INT_HLDOFF 2
+  #define EVQ_BUF_SIZE_LBN 0
+  #define EVQ_BUF_SIZE_WIDTH 1
+#define DRV_EV_REG_KER_OFST 0x440 // Driver generated event register
+#define DRV_EV_REG_OFST 0x440 // Driver generated event register
+  #define DRV_EV_QID_LBN 64
+  #define DRV_EV_QID_WIDTH 12
+  #define DRV_EV_DATA_LBN 0
+  #define DRV_EV_DATA_WIDTH 64
+#define EVQ_CTL_REG_KER_OFST 0x450 // Event queue control register
+#define EVQ_CTL_REG_OFST 0x450 // Event queue control register
+  #define RX_EVQ_WAKEUP_MASK_B0_LBN 15
+  #define RX_EVQ_WAKEUP_MASK_B0_WIDTH 6
+  #define EVQ_OWNERR_CTL_LBN 14
+  #define EVQ_OWNERR_CTL_WIDTH 1
+  #define EVQ_FIFO_AF_TH_LBN 8
+  #define EVQ_FIFO_AF_TH_WIDTH 6
+  #define EVQ_FIFO_NOTAF_TH_LBN 0
+  #define EVQ_FIFO_NOTAF_TH_WIDTH 6
+//////////////---- SRAM Module Registers C Header ----//////////////
+#define BUF_TBL_CFG_REG_KER_OFST 0x600 // Buffer table configuration register
+#define BUF_TBL_CFG_REG_OFST 0x600 // Buffer table configuration register
+  #define BUF_TBL_MODE_LBN 3
+  #define BUF_TBL_MODE_WIDTH 1
+#define SRM_RX_DC_CFG_REG_KER_OFST 0x610 // SRAM receive descriptor cache configuration register
+#define SRM_RX_DC_CFG_REG_OFST 0x610 // SRAM receive descriptor cache configuration register
+  #define SRM_RX_DC_BASE_ADR_LBN 0
+  #define SRM_RX_DC_BASE_ADR_WIDTH 21
+#define SRM_TX_DC_CFG_REG_KER_OFST 0x620 // SRAM transmit descriptor cache configuration register
+#define SRM_TX_DC_CFG_REG_OFST 0x620 // SRAM transmit descriptor cache configuration register
+  #define SRM_TX_DC_BASE_ADR_LBN 0
+  #define SRM_TX_DC_BASE_ADR_WIDTH 21
+#define SRM_CFG_REG_KER_OFST 0x630 // SRAM configuration register
+#define SRM_CFG_REG_OFST 0x630 // SRAM configuration register
+  #define SRAM_OOB_ADR_INTEN_LBN 5
+  #define SRAM_OOB_ADR_INTEN_WIDTH 1
+  #define SRAM_OOB_BUF_INTEN_LBN 4
+  #define SRAM_OOB_BUF_INTEN_WIDTH 1
+  #define SRAM_BT_INIT_EN_LBN 3
+  #define SRAM_BT_INIT_EN_WIDTH 1
+  #define SRM_NUM_BANK_LBN 2
+  #define SRM_NUM_BANK_WIDTH 1
+  #define SRM_BANK_SIZE_LBN 0
+  #define SRM_BANK_SIZE_WIDTH 2
+#define BUF_TBL_UPD_REG_KER_OFST 0x650 // Buffer table update register
+#define BUF_TBL_UPD_REG_OFST 0x650 // Buffer table update register
+  #define BUF_UPD_CMD_LBN 63
+  #define BUF_UPD_CMD_WIDTH 1
+  #define BUF_CLR_CMD_LBN 62
+  #define BUF_CLR_CMD_WIDTH 1
+  #define BUF_CLR_END_ID_LBN 32
+  #define BUF_CLR_END_ID_WIDTH 20
+  #define BUF_CLR_START_ID_LBN 0
+  #define BUF_CLR_START_ID_WIDTH 20
+#define SRM_UPD_EVQ_REG_KER_OFST 0x660 // Buffer table update register
+#define SRM_UPD_EVQ_REG_OFST 0x660 // Buffer table update register
+  #define SRM_UPD_EVQ_ID_LBN 0
+  #define SRM_UPD_EVQ_ID_WIDTH 12
+#define SRAM_PARITY_REG_KER_OFST 0x670 // SRAM parity register.
+#define SRAM_PARITY_REG_OFST 0x670 // SRAM parity register.
+  #define FORCE_SRAM_PERR_LBN 0
+  #define FORCE_SRAM_PERR_WIDTH 1
+
+#if EFVI_FALCON_EXTENDED_P_BAR
+#define BUF_HALF_TBL_KER_OFST 0x18000 // Buffer table in half buffer table mode direct access by kernel driver
+#else
+#define BUF_HALF_TBL_KER_OFST 0x8000 // Buffer table in half buffer table mode direct access by kernel driver
+#endif
+
+
+#define BUF_HALF_TBL_OFST 0x800000 // Buffer table in half buffer table mode direct access by char driver
+  #define BUF_ADR_HBUF_ODD_LBN 44
+  #define BUF_ADR_HBUF_ODD_WIDTH 20
+  #define BUF_OWNER_ID_HBUF_ODD_LBN 32
+  #define BUF_OWNER_ID_HBUF_ODD_WIDTH 12
+  #define BUF_ADR_HBUF_EVEN_LBN 12
+  #define BUF_ADR_HBUF_EVEN_WIDTH 20
+  #define BUF_OWNER_ID_HBUF_EVEN_LBN 0
+  #define BUF_OWNER_ID_HBUF_EVEN_WIDTH 12
+
+
+#if EFVI_FALCON_EXTENDED_P_BAR
+#define BUF_FULL_TBL_KER_OFST 0x18000 // Buffer table in full buffer table mode direct access by kernel driver
+#else
+#define BUF_FULL_TBL_KER_OFST 0x8000 // Buffer table in full buffer table mode direct access by kernel driver
+#endif
+
+
+
+
+#define BUF_FULL_TBL_OFST 0x800000 // Buffer table in full buffer table mode direct access by char driver
+  #define IP_DAT_BUF_SIZE_LBN 50
+  #define IP_DAT_BUF_SIZE_WIDTH 1
+  #define BUF_ADR_REGION_LBN 48
+  #define BUF_ADR_REGION_WIDTH 2
+  #define BUF_ADR_FBUF_LBN 14
+  #define BUF_ADR_FBUF_WIDTH 34
+  #define BUF_OWNER_ID_FBUF_LBN 0
+  #define BUF_OWNER_ID_FBUF_WIDTH 14
+#define SRM_DBG_REG_OFST 0x3000000 // SRAM debug access
+  #define SRM_DBG_LBN 0
+  #define SRM_DBG_WIDTH 64
+//////////////---- RX Datapath Registers C Header ----//////////////
+
+#define RX_CFG_REG_KER_OFST 0x800 // Receive configuration register
+#define RX_CFG_REG_OFST 0x800 // Receive configuration register
+
+#if !defined(FALCON_64K_RXFIFO) && !defined(FALCON_PRE_02020029)
+# if !defined(FALCON_128K_RXFIFO)
+#  define FALCON_128K_RXFIFO
+# endif
+#endif
+
+#if defined(FALCON_128K_RXFIFO)
+
+/* new for B0 */
+  #define RX_TOEP_TCP_SUPPRESS_B0_LBN 48
+  #define RX_TOEP_TCP_SUPPRESS_B0_WIDTH 1
+  #define RX_INGR_EN_B0_LBN 47
+  #define RX_INGR_EN_B0_WIDTH 1
+  #define RX_TOEP_IPV4_B0_LBN 46
+  #define RX_TOEP_IPV4_B0_WIDTH 1
+  #define RX_HASH_ALG_B0_LBN 45
+  #define RX_HASH_ALG_B0_WIDTH 1
+  #define RX_HASH_INSERT_HDR_B0_LBN 44
+  #define RX_HASH_INSERT_HDR_B0_WIDTH 1
+/* moved for B0 */
+  #define RX_DESC_PUSH_EN_B0_LBN 43
+  #define RX_DESC_PUSH_EN_B0_WIDTH 1
+  #define RX_RDW_PATCH_EN_LBN 42 /* Non head of line blocking */
+  #define RX_RDW_PATCH_EN_WIDTH 1
+  #define RX_PCI_BURST_SIZE_B0_LBN 39
+  #define RX_PCI_BURST_SIZE_B0_WIDTH 3
+  #define RX_OWNERR_CTL_B0_LBN 38
+  #define RX_OWNERR_CTL_B0_WIDTH 1
+  #define RX_XON_TX_TH_B0_LBN 33 
+  #define RX_XON_TX_TH_B0_WIDTH 5
+  #define RX_XOFF_TX_TH_B0_LBN 28 
+  #define RX_XOFF_TX_TH_B0_WIDTH 5
+  #define RX_USR_BUF_SIZE_B0_LBN 19
+  #define RX_USR_BUF_SIZE_B0_WIDTH 9
+  #define RX_XON_MAC_TH_B0_LBN 10
+  #define RX_XON_MAC_TH_B0_WIDTH 9
+  #define RX_XOFF_MAC_TH_B0_LBN 1
+  #define RX_XOFF_MAC_TH_B0_WIDTH 9
+  #define RX_XOFF_MAC_EN_B0_LBN 0
+  #define RX_XOFF_MAC_EN_B0_WIDTH 1
+
+#elif !defined(FALCON_PRE_02020029)
+/* new for B0 */
+  #define RX_TOEP_TCP_SUPPRESS_B0_LBN 46
+  #define RX_TOEP_TCP_SUPPRESS_B0_WIDTH 1
+  #define RX_INGR_EN_B0_LBN 45
+  #define RX_INGR_EN_B0_WIDTH 1
+  #define RX_TOEP_IPV4_B0_LBN 44
+  #define RX_TOEP_IPV4_B0_WIDTH 1
+  #define RX_HASH_ALG_B0_LBN 43
+  #define RX_HASH_ALG_B0_WIDTH 41
+  #define RX_HASH_INSERT_HDR_B0_LBN 42
+  #define RX_HASH_INSERT_HDR_B0_WIDTH 1
+/* moved for B0 */
+  #define RX_DESC_PUSH_EN_B0_LBN 41
+  #define RX_DESC_PUSH_EN_B0_WIDTH 1
+  #define RX_PCI_BURST_SIZE_B0_LBN 37
+  #define RX_PCI_BURST_SIZE_B0_WIDTH 3
+  #define RX_OWNERR_CTL_B0_LBN 36
+  #define RX_OWNERR_CTL_B0_WIDTH 1
+  #define RX_XON_TX_TH_B0_LBN 31
+  #define RX_XON_TX_TH_B0_WIDTH 5
+  #define RX_XOFF_TX_TH_B0_LBN 26
+  #define RX_XOFF_TX_TH_B0_WIDTH 5
+  #define RX_USR_BUF_SIZE_B0_LBN 17
+  #define RX_USR_BUF_SIZE_B0_WIDTH 9
+  #define RX_XON_MAC_TH_B0_LBN 9
+  #define RX_XON_MAC_TH_B0_WIDTH 8
+  #define RX_XOFF_MAC_TH_B0_LBN 1
+  #define RX_XOFF_MAC_TH_B0_WIDTH 8
+  #define RX_XOFF_MAC_EN_B0_LBN 0
+  #define RX_XOFF_MAC_EN_B0_WIDTH 1
+
+#else
+/* new for B0 */
+  #define RX_TOEP_TCP_SUPPRESS_B0_LBN 44
+  #define RX_TOEP_TCP_SUPPRESS_B0_WIDTH 1
+  #define RX_INGR_EN_B0_LBN 43
+  #define RX_INGR_EN_B0_WIDTH 1
+  #define RX_TOEP_IPV4_B0_LBN 42
+  #define RX_TOEP_IPV4_B0_WIDTH 1
+  #define RX_HASH_ALG_B0_LBN 41
+  #define RX_HASH_ALG_B0_WIDTH 41
+  #define RX_HASH_INSERT_HDR_B0_LBN 40
+  #define RX_HASH_INSERT_HDR_B0_WIDTH 1
+/* moved for B0 */
+  #define RX_DESC_PUSH_EN_B0_LBN 35
+  #define RX_DESC_PUSH_EN_B0_WIDTH 1
+  #define RX_PCI_BURST_SIZE_B0_LBN 35
+  #define RX_PCI_BURST_SIZE_B0_WIDTH 2
+  #define RX_OWNERR_CTL_B0_LBN 34
+  #define RX_OWNERR_CTL_B0_WIDTH 1
+  #define RX_XON_TX_TH_B0_LBN 29
+  #define RX_XON_TX_TH_B0_WIDTH 5
+  #define RX_XOFF_TX_TH_B0_LBN 24
+  #define RX_XOFF_TX_TH_B0_WIDTH 5
+  #define RX_USR_BUF_SIZE_B0_LBN 15
+  #define RX_USR_BUF_SIZE_B0_WIDTH 9
+  #define RX_XON_MAC_TH_B0_LBN 8
+  #define RX_XON_MAC_TH_B0_WIDTH 7
+  #define RX_XOFF_MAC_TH_B0_LBN 1
+  #define RX_XOFF_MAC_TH_B0_WIDTH 7
+  #define RX_XOFF_MAC_EN_B0_LBN 0
+  #define RX_XOFF_MAC_EN_B0_WIDTH 1
+
+#endif
+
+/* A0/A1 */
+  #define RX_PUSH_EN_A1_LBN 35
+  #define RX_PUSH_EN_A1_WIDTH 1
+  #define RX_PCI_BURST_SIZE_A1_LBN 31
+  #define RX_PCI_BURST_SIZE_A1_WIDTH 3
+  #define RX_OWNERR_CTL_A1_LBN 30
+  #define RX_OWNERR_CTL_A1_WIDTH 1
+  #define RX_XON_TX_TH_A1_LBN 25
+  #define RX_XON_TX_TH_A1_WIDTH 5
+  #define RX_XOFF_TX_TH_A1_LBN 20
+  #define RX_XOFF_TX_TH_A1_WIDTH 5
+  #define RX_USR_BUF_SIZE_A1_LBN 11
+  #define RX_USR_BUF_SIZE_A1_WIDTH 9
+  #define RX_XON_MAC_TH_A1_LBN 6
+  #define RX_XON_MAC_TH_A1_WIDTH 5
+  #define RX_XOFF_MAC_TH_A1_LBN 1
+  #define RX_XOFF_MAC_TH_A1_WIDTH 5
+  #define RX_XOFF_MAC_EN_A1_LBN 0
+  #define RX_XOFF_MAC_EN_A1_WIDTH 1
+
+#define RX_FILTER_CTL_REG_OFST 0x810 // Receive filter control registers
+  #define SCATTER_ENBL_NO_MATCH_Q_B0_LBN 40
+  #define SCATTER_ENBL_NO_MATCH_Q_B0_WIDTH 1
+  #define UDP_FULL_SRCH_LIMIT_LBN 32
+  #define UDP_FULL_SRCH_LIMIT_WIDTH 8
+  #define NUM_KER_LBN 24
+  #define NUM_KER_WIDTH 2
+  #define UDP_WILD_SRCH_LIMIT_LBN 16
+  #define UDP_WILD_SRCH_LIMIT_WIDTH 8
+  #define TCP_WILD_SRCH_LIMIT_LBN 8
+  #define TCP_WILD_SRCH_LIMIT_WIDTH 8
+  #define TCP_FULL_SRCH_LIMIT_LBN 0
+  #define TCP_FULL_SRCH_LIMIT_WIDTH 8
+#define RX_FLUSH_DESCQ_REG_KER_OFST 0x820 // Receive flush descriptor queue register
+#define RX_FLUSH_DESCQ_REG_OFST 0x820 // Receive flush descriptor queue register
+  #define RX_FLUSH_DESCQ_CMD_LBN 24
+  #define RX_FLUSH_DESCQ_CMD_WIDTH 1
+  #define RX_FLUSH_EVQ_ID_LBN 12
+  #define RX_FLUSH_EVQ_ID_WIDTH 12
+  #define RX_FLUSH_DESCQ_LBN 0
+  #define RX_FLUSH_DESCQ_WIDTH 12
+#define RX_DESC_UPD_REG_KER_OFST 0x830 // Kernel  receive descriptor update register. Page-mapped
+#define RX_DESC_UPD_REG_PAGE4_OFST 0x8830 // Char & user receive descriptor update register. Page-mapped. For lowest 1K queues.
+#define RX_DESC_UPD_REG_PAGE123K_OFST 0x1000830 // Char & user receive descriptor update register. Page-mapped. For upper 3K queues.
+  #define RX_DESC_WPTR_LBN 96
+  #define RX_DESC_WPTR_WIDTH 12
+  #define RX_DESC_PUSH_CMD_LBN 95
+  #define RX_DESC_PUSH_CMD_WIDTH 1
+  #define RX_DESC_LBN 0
+  #define RX_DESC_WIDTH 64
+  #define RX_KER_DESC_LBN 0
+  #define RX_KER_DESC_WIDTH 64
+  #define RX_USR_DESC_LBN 0
+  #define RX_USR_DESC_WIDTH 32
+#define RX_DC_CFG_REG_KER_OFST 0x840 // Receive descriptor cache configuration register
+#define RX_DC_CFG_REG_OFST 0x840 // Receive descriptor cache configuration register
+  #define RX_DC_SIZE_LBN 0
+  #define RX_DC_SIZE_WIDTH 2
+#define RX_DC_PF_WM_REG_KER_OFST 0x850 // Receive descriptor cache pre-fetch watermark register
+#define RX_DC_PF_WM_REG_OFST 0x850 // Receive descriptor cache pre-fetch watermark register
+  #define RX_DC_PF_LWM_LO_LBN 0
+  #define RX_DC_PF_LWM_LO_WIDTH 6
+
+#define RX_RSS_TKEY_B0_OFST 0x860 // RSS Toeplitz hash key (B0 only)
+
+#define RX_NODESC_DROP_REG 0x880
+  #define RX_NODESC_DROP_CNT_LBN 0
+  #define RX_NODESC_DROP_CNT_WIDTH 16
+
+#define XM_TX_CFG_REG_OFST 0x1230
+  #define XM_AUTO_PAD_LBN 5
+  #define XM_AUTO_PAD_WIDTH 1
+
+#define RX_FILTER_TBL0_OFST 0xF00000 // Receive filter table - even entries
+  #define RSS_EN_0_B0_LBN 110
+  #define RSS_EN_0_B0_WIDTH 1
+  #define SCATTER_EN_0_B0_LBN 109
+  #define SCATTER_EN_0_B0_WIDTH 1
+  #define TCP_UDP_0_LBN 108
+  #define TCP_UDP_0_WIDTH 1
+  #define RXQ_ID_0_LBN 96
+  #define RXQ_ID_0_WIDTH 12
+  #define DEST_IP_0_LBN 64
+  #define DEST_IP_0_WIDTH 32
+  #define DEST_PORT_TCP_0_LBN 48
+  #define DEST_PORT_TCP_0_WIDTH 16
+  #define SRC_IP_0_LBN 16
+  #define SRC_IP_0_WIDTH 32
+  #define SRC_TCP_DEST_UDP_0_LBN 0
+  #define SRC_TCP_DEST_UDP_0_WIDTH 16
+#define RX_FILTER_TBL1_OFST 0xF00010 // Receive filter table - odd entries
+  #define RSS_EN_1_B0_LBN 110
+  #define RSS_EN_1_B0_WIDTH 1
+  #define SCATTER_EN_1_B0_LBN 109
+  #define SCATTER_EN_1_B0_WIDTH 1
+  #define TCP_UDP_1_LBN 108
+  #define TCP_UDP_1_WIDTH 1
+  #define RXQ_ID_1_LBN 96
+  #define RXQ_ID_1_WIDTH 12
+  #define DEST_IP_1_LBN 64
+  #define DEST_IP_1_WIDTH 32
+  #define DEST_PORT_TCP_1_LBN 48
+  #define DEST_PORT_TCP_1_WIDTH 16
+  #define SRC_IP_1_LBN 16
+  #define SRC_IP_1_WIDTH 32
+  #define SRC_TCP_DEST_UDP_1_LBN 0
+  #define SRC_TCP_DEST_UDP_1_WIDTH 16
+
+#if EFVI_FALCON_EXTENDED_P_BAR
+#define RX_DESC_PTR_TBL_KER_OFST 0x11800 // Receive descriptor pointer kernel access
+#else
+#define RX_DESC_PTR_TBL_KER_OFST 0x1800 // Receive descriptor pointer kernel access
+#endif
+
+
+#define RX_DESC_PTR_TBL_OFST 0xF40000 // Receive descriptor pointer table
+  #define RX_ISCSI_DDIG_EN_LBN 88
+  #define RX_ISCSI_DDIG_EN_WIDTH 1
+  #define RX_ISCSI_HDIG_EN_LBN 87
+  #define RX_ISCSI_HDIG_EN_WIDTH 1
+  #define RX_DESC_PREF_ACT_LBN 86
+  #define RX_DESC_PREF_ACT_WIDTH 1
+  #define RX_DC_HW_RPTR_LBN 80
+  #define RX_DC_HW_RPTR_WIDTH 6
+  #define RX_DESCQ_HW_RPTR_LBN 68
+  #define RX_DESCQ_HW_RPTR_WIDTH 12
+  #define RX_DESCQ_SW_WPTR_LBN 56
+  #define RX_DESCQ_SW_WPTR_WIDTH 12
+  #define RX_DESCQ_BUF_BASE_ID_LBN 36
+  #define RX_DESCQ_BUF_BASE_ID_WIDTH 20
+  #define RX_DESCQ_EVQ_ID_LBN 24
+  #define RX_DESCQ_EVQ_ID_WIDTH 12
+  #define RX_DESCQ_OWNER_ID_LBN 10
+  #define RX_DESCQ_OWNER_ID_WIDTH 14
+  #define RX_DESCQ_LABEL_LBN 5
+  #define RX_DESCQ_LABEL_WIDTH 5
+  #define RX_DESCQ_SIZE_LBN 3
+  #define RX_DESCQ_SIZE_WIDTH 2
+  #define RX_DESCQ_TYPE_LBN 2
+  #define RX_DESCQ_TYPE_WIDTH 1
+  #define RX_DESCQ_JUMBO_LBN 1
+  #define RX_DESCQ_JUMBO_WIDTH 1
+  #define RX_DESCQ_EN_LBN 0
+  #define RX_DESCQ_EN_WIDTH 1
+
+
+#define RX_RSS_INDIR_TBL_B0_OFST 0xFB0000 // RSS indirection table (B0 only)
+  #define RX_RSS_INDIR_ENT_B0_LBN 0
+  #define RX_RSS_INDIR_ENT_B0_WIDTH 6
+
+//////////////---- TX Datapath Registers C Header ----//////////////
+#define TX_FLUSH_DESCQ_REG_KER_OFST 0xA00 // Transmit flush descriptor queue register
+#define TX_FLUSH_DESCQ_REG_OFST 0xA00 // Transmit flush descriptor queue register
+  #define TX_FLUSH_DESCQ_CMD_LBN 12
+  #define TX_FLUSH_DESCQ_CMD_WIDTH 1
+  #define TX_FLUSH_DESCQ_LBN 0
+  #define TX_FLUSH_DESCQ_WIDTH 12
+#define TX_DESC_UPD_REG_KER_OFST 0xA10 // Kernel transmit descriptor update register. Page-mapped
+#define TX_DESC_UPD_REG_PAGE4_OFST 0x8A10 // Char & user transmit descriptor update register. Page-mapped
+#define TX_DESC_UPD_REG_PAGE123K_OFST 0x1000A10 // Char & user transmit descriptor update register. Page-mapped
+  #define TX_DESC_WPTR_LBN 96
+  #define TX_DESC_WPTR_WIDTH 12
+  #define TX_DESC_PUSH_CMD_LBN 95
+  #define TX_DESC_PUSH_CMD_WIDTH 1
+  #define TX_DESC_LBN 0
+  #define TX_DESC_WIDTH 95
+  #define TX_KER_DESC_LBN 0
+  #define TX_KER_DESC_WIDTH 64
+  #define TX_USR_DESC_LBN 0
+  #define TX_USR_DESC_WIDTH 64
+#define TX_DC_CFG_REG_KER_OFST 0xA20 // Transmit descriptor cache configuration register
+#define TX_DC_CFG_REG_OFST 0xA20 // Transmit descriptor cache configuration register
+  #define TX_DC_SIZE_LBN 0
+  #define TX_DC_SIZE_WIDTH 2
+
+#if EFVI_FALCON_EXTENDED_P_BAR
+#define TX_DESC_PTR_TBL_KER_OFST 0x11900 // Transmit descriptor pointer.
+#else
+#define TX_DESC_PTR_TBL_KER_OFST 0x1900 // Transmit descriptor pointer.
+#endif
+
+
+#define TX_DESC_PTR_TBL_OFST 0xF50000 // Transmit descriptor pointer
+  #define TX_NON_IP_DROP_DIS_B0_LBN 91
+  #define TX_NON_IP_DROP_DIS_B0_WIDTH 1
+  #define TX_IP_CHKSM_DIS_B0_LBN 90
+  #define TX_IP_CHKSM_DIS_B0_WIDTH 1
+  #define TX_TCP_CHKSM_DIS_B0_LBN 89
+  #define TX_TCP_CHKSM_DIS_B0_WIDTH 1
+  #define TX_DESCQ_EN_LBN 88
+  #define TX_DESCQ_EN_WIDTH 1
+  #define TX_ISCSI_DDIG_EN_LBN 87
+  #define TX_ISCSI_DDIG_EN_WIDTH 1
+  #define TX_ISCSI_HDIG_EN_LBN 86
+  #define TX_ISCSI_HDIG_EN_WIDTH 1
+  #define TX_DC_HW_RPTR_LBN 80
+  #define TX_DC_HW_RPTR_WIDTH 6
+  #define TX_DESCQ_HW_RPTR_LBN 68
+  #define TX_DESCQ_HW_RPTR_WIDTH 12
+  #define TX_DESCQ_SW_WPTR_LBN 56
+  #define TX_DESCQ_SW_WPTR_WIDTH 12
+  #define TX_DESCQ_BUF_BASE_ID_LBN 36
+  #define TX_DESCQ_BUF_BASE_ID_WIDTH 20
+  #define TX_DESCQ_EVQ_ID_LBN 24
+  #define TX_DESCQ_EVQ_ID_WIDTH 12
+  #define TX_DESCQ_OWNER_ID_LBN 10
+  #define TX_DESCQ_OWNER_ID_WIDTH 14
+  #define TX_DESCQ_LABEL_LBN 5
+  #define TX_DESCQ_LABEL_WIDTH 5
+  #define TX_DESCQ_SIZE_LBN 3
+  #define TX_DESCQ_SIZE_WIDTH 2
+  #define TX_DESCQ_TYPE_LBN 1
+  #define TX_DESCQ_TYPE_WIDTH 2
+  #define TX_DESCQ_FLUSH_LBN 0
+  #define TX_DESCQ_FLUSH_WIDTH 1
+#define TX_CFG_REG_KER_OFST 0xA50 // Transmit configuration register
+#define TX_CFG_REG_OFST 0xA50 // Transmit configuration register
+  #define TX_IP_ID_P1_OFS_LBN 32
+  #define TX_IP_ID_P1_OFS_WIDTH 15
+  #define TX_IP_ID_P0_OFS_LBN 16
+  #define TX_IP_ID_P0_OFS_WIDTH 15
+  #define TX_TURBO_EN_LBN 3
+  #define TX_TURBO_EN_WIDTH 1 
+  #define TX_OWNERR_CTL_LBN 2
+  #define TX_OWNERR_CTL_WIDTH 2
+  #define TX_NON_IP_DROP_DIS_LBN 1
+  #define TX_NON_IP_DROP_DIS_WIDTH 1
+  #define TX_IP_ID_REP_EN_LBN 0
+  #define TX_IP_ID_REP_EN_WIDTH 1
+#define TX_RESERVED_REG_KER_OFST 0xA80 // Transmit configuration register
+#define TX_RESERVED_REG_OFST 0xA80 // Transmit configuration register
+  #define TX_CSR_PUSH_EN_LBN 89
+  #define TX_CSR_PUSH_EN_WIDTH 1
+  #define TX_RX_SPACER_LBN 64
+  #define TX_RX_SPACER_WIDTH 8
+  #define TX_SW_EV_EN_LBN 59
+  #define TX_SW_EV_EN_WIDTH 1
+  #define TX_RX_SPACER_EN_LBN 57
+  #define TX_RX_SPACER_EN_WIDTH 1
+  #define TX_CSR_PREF_WD_TMR_LBN 24
+  #define TX_CSR_PREF_WD_TMR_WIDTH 16
+  #define TX_CSR_ONLY1TAG_LBN 21
+  #define TX_CSR_ONLY1TAG_WIDTH 1
+  #define TX_PREF_THRESHOLD_LBN 19
+  #define TX_PREF_THRESHOLD_WIDTH 2
+  #define TX_ONE_PKT_PER_Q_LBN 18
+  #define TX_ONE_PKT_PER_Q_WIDTH 1
+  #define TX_DIS_NON_IP_EV_LBN 17
+  #define TX_DIS_NON_IP_EV_WIDTH 1
+  #define TX_DMA_SPACER_LBN 8
+  #define TX_DMA_SPACER_WIDTH 8
+  #define TX_FLUSH_MIN_LEN_EN_B0_LBN 7
+  #define TX_FLUSH_MIN_LEN_EN_B0_WIDTH 1
+  #define TX_TCP_DIS_A1_LBN 7
+  #define TX_TCP_DIS_A1_WIDTH 1
+  #define TX_IP_DIS_A1_LBN 6
+  #define TX_IP_DIS_A1_WIDTH 1
+  #define TX_MAX_CPL_LBN 2
+  #define TX_MAX_CPL_WIDTH 2
+  #define TX_MAX_PREF_LBN 0
+  #define TX_MAX_PREF_WIDTH 2
+#define TX_VLAN_REG_OFST 0xAE0 // Transmit VLAN tag register
+  #define TX_VLAN_EN_LBN 127
+  #define TX_VLAN_EN_WIDTH 1
+  #define TX_VLAN7_PORT1_EN_LBN 125
+  #define TX_VLAN7_PORT1_EN_WIDTH 1
+  #define TX_VLAN7_PORT0_EN_LBN 124
+  #define TX_VLAN7_PORT0_EN_WIDTH 1
+  #define TX_VLAN7_LBN 112
+  #define TX_VLAN7_WIDTH 12
+  #define TX_VLAN6_PORT1_EN_LBN 109
+  #define TX_VLAN6_PORT1_EN_WIDTH 1
+  #define TX_VLAN6_PORT0_EN_LBN 108
+  #define TX_VLAN6_PORT0_EN_WIDTH 1
+  #define TX_VLAN6_LBN 96
+  #define TX_VLAN6_WIDTH 12
+  #define TX_VLAN5_PORT1_EN_LBN 93
+  #define TX_VLAN5_PORT1_EN_WIDTH 1
+  #define TX_VLAN5_PORT0_EN_LBN 92
+  #define TX_VLAN5_PORT0_EN_WIDTH 1
+  #define TX_VLAN5_LBN 80
+  #define TX_VLAN5_WIDTH 12
+  #define TX_VLAN4_PORT1_EN_LBN 77
+  #define TX_VLAN4_PORT1_EN_WIDTH 1
+  #define TX_VLAN4_PORT0_EN_LBN 76
+  #define TX_VLAN4_PORT0_EN_WIDTH 1
+  #define TX_VLAN4_LBN 64
+  #define TX_VLAN4_WIDTH 12
+  #define TX_VLAN3_PORT1_EN_LBN 61
+  #define TX_VLAN3_PORT1_EN_WIDTH 1
+  #define TX_VLAN3_PORT0_EN_LBN 60
+  #define TX_VLAN3_PORT0_EN_WIDTH 1
+  #define TX_VLAN3_LBN 48
+  #define TX_VLAN3_WIDTH 12
+  #define TX_VLAN2_PORT1_EN_LBN 45
+  #define TX_VLAN2_PORT1_EN_WIDTH 1
+  #define TX_VLAN2_PORT0_EN_LBN 44
+  #define TX_VLAN2_PORT0_EN_WIDTH 1
+  #define TX_VLAN2_LBN 32
+  #define TX_VLAN2_WIDTH 12
+  #define TX_VLAN1_PORT1_EN_LBN 29
+  #define TX_VLAN1_PORT1_EN_WIDTH 1
+  #define TX_VLAN1_PORT0_EN_LBN 28
+  #define TX_VLAN1_PORT0_EN_WIDTH 1
+  #define TX_VLAN1_LBN 16
+  #define TX_VLAN1_WIDTH 12
+  #define TX_VLAN0_PORT1_EN_LBN 13
+  #define TX_VLAN0_PORT1_EN_WIDTH 1
+  #define TX_VLAN0_PORT0_EN_LBN 12
+  #define TX_VLAN0_PORT0_EN_WIDTH 1
+  #define TX_VLAN0_LBN 0
+  #define TX_VLAN0_WIDTH 12
+#define TX_FIL_CTL_REG_OFST 0xAF0 // Transmit filter control register
+  #define TX_MADR1_FIL_EN_LBN 65
+  #define TX_MADR1_FIL_EN_WIDTH 1
+  #define TX_MADR0_FIL_EN_LBN 64
+  #define TX_MADR0_FIL_EN_WIDTH 1
+  #define TX_IPFIL31_PORT1_EN_LBN 63
+  #define TX_IPFIL31_PORT1_EN_WIDTH 1
+  #define TX_IPFIL31_PORT0_EN_LBN 62
+  #define TX_IPFIL31_PORT0_EN_WIDTH 1
+  #define TX_IPFIL30_PORT1_EN_LBN 61
+  #define TX_IPFIL30_PORT1_EN_WIDTH 1
+  #define TX_IPFIL30_PORT0_EN_LBN 60
+  #define TX_IPFIL30_PORT0_EN_WIDTH 1
+  #define TX_IPFIL29_PORT1_EN_LBN 59
+  #define TX_IPFIL29_PORT1_EN_WIDTH 1
+  #define TX_IPFIL29_PORT0_EN_LBN 58
+  #define TX_IPFIL29_PORT0_EN_WIDTH 1
+  #define TX_IPFIL28_PORT1_EN_LBN 57
+  #define TX_IPFIL28_PORT1_EN_WIDTH 1
+  #define TX_IPFIL28_PORT0_EN_LBN 56
+  #define TX_IPFIL28_PORT0_EN_WIDTH 1
+  #define TX_IPFIL27_PORT1_EN_LBN 55
+  #define TX_IPFIL27_PORT1_EN_WIDTH 1
+  #define TX_IPFIL27_PORT0_EN_LBN 54
+  #define TX_IPFIL27_PORT0_EN_WIDTH 1
+  #define TX_IPFIL26_PORT1_EN_LBN 53
+  #define TX_IPFIL26_PORT1_EN_WIDTH 1
+  #define TX_IPFIL26_PORT0_EN_LBN 52
+  #define TX_IPFIL26_PORT0_EN_WIDTH 1
+  #define TX_IPFIL25_PORT1_EN_LBN 51
+  #define TX_IPFIL25_PORT1_EN_WIDTH 1
+  #define TX_IPFIL25_PORT0_EN_LBN 50
+  #define TX_IPFIL25_PORT0_EN_WIDTH 1
+  #define TX_IPFIL24_PORT1_EN_LBN 49
+  #define TX_IPFIL24_PORT1_EN_WIDTH 1
+  #define TX_IPFIL24_PORT0_EN_LBN 48
+  #define TX_IPFIL24_PORT0_EN_WIDTH 1
+  #define TX_IPFIL23_PORT1_EN_LBN 47
+  #define TX_IPFIL23_PORT1_EN_WIDTH 1
+  #define TX_IPFIL23_PORT0_EN_LBN 46
+  #define TX_IPFIL23_PORT0_EN_WIDTH 1
+  #define TX_IPFIL22_PORT1_EN_LBN 45
+  #define TX_IPFIL22_PORT1_EN_WIDTH 1
+  #define TX_IPFIL22_PORT0_EN_LBN 44
+  #define TX_IPFIL22_PORT0_EN_WIDTH 1
+  #define TX_IPFIL21_PORT1_EN_LBN 43
+  #define TX_IPFIL21_PORT1_EN_WIDTH 1
+  #define TX_IPFIL21_PORT0_EN_LBN 42
+  #define TX_IPFIL21_PORT0_EN_WIDTH 1
+  #define TX_IPFIL20_PORT1_EN_LBN 41
+  #define TX_IPFIL20_PORT1_EN_WIDTH 1
+  #define TX_IPFIL20_PORT0_EN_LBN 40
+  #define TX_IPFIL20_PORT0_EN_WIDTH 1
+  #define TX_IPFIL19_PORT1_EN_LBN 39
+  #define TX_IPFIL19_PORT1_EN_WIDTH 1
+  #define TX_IPFIL19_PORT0_EN_LBN 38
+  #define TX_IPFIL19_PORT0_EN_WIDTH 1
+  #define TX_IPFIL18_PORT1_EN_LBN 37
+  #define TX_IPFIL18_PORT1_EN_WIDTH 1
+  #define TX_IPFIL18_PORT0_EN_LBN 36
+  #define TX_IPFIL18_PORT0_EN_WIDTH 1
+  #define TX_IPFIL17_PORT1_EN_LBN 35
+  #define TX_IPFIL17_PORT1_EN_WIDTH 1
+  #define TX_IPFIL17_PORT0_EN_LBN 34
+  #define TX_IPFIL17_PORT0_EN_WIDTH 1
+  #define TX_IPFIL16_PORT1_EN_LBN 33
+  #define TX_IPFIL16_PORT1_EN_WIDTH 1
+  #define TX_IPFIL16_PORT0_EN_LBN 32
+  #define TX_IPFIL16_PORT0_EN_WIDTH 1
+  #define TX_IPFIL15_PORT1_EN_LBN 31
+  #define TX_IPFIL15_PORT1_EN_WIDTH 1
+  #define TX_IPFIL15_PORT0_EN_LBN 30
+  #define TX_IPFIL15_PORT0_EN_WIDTH 1
+  #define TX_IPFIL14_PORT1_EN_LBN 29
+  #define TX_IPFIL14_PORT1_EN_WIDTH 1
+  #define TX_IPFIL14_PORT0_EN_LBN 28
+  #define TX_IPFIL14_PORT0_EN_WIDTH 1
+  #define TX_IPFIL13_PORT1_EN_LBN 27
+  #define TX_IPFIL13_PORT1_EN_WIDTH 1
+  #define TX_IPFIL13_PORT0_EN_LBN 26
+  #define TX_IPFIL13_PORT0_EN_WIDTH 1
+  #define TX_IPFIL12_PORT1_EN_LBN 25
+  #define TX_IPFIL12_PORT1_EN_WIDTH 1
+  #define TX_IPFIL12_PORT0_EN_LBN 24
+  #define TX_IPFIL12_PORT0_EN_WIDTH 1
+  #define TX_IPFIL11_PORT1_EN_LBN 23
+  #define TX_IPFIL11_PORT1_EN_WIDTH 1
+  #define TX_IPFIL11_PORT0_EN_LBN 22
+  #define TX_IPFIL11_PORT0_EN_WIDTH 1
+  #define TX_IPFIL10_PORT1_EN_LBN 21
+  #define TX_IPFIL10_PORT1_EN_WIDTH 1
+  #define TX_IPFIL10_PORT0_EN_LBN 20
+  #define TX_IPFIL10_PORT0_EN_WIDTH 1
+  #define TX_IPFIL9_PORT1_EN_LBN 19
+  #define TX_IPFIL9_PORT1_EN_WIDTH 1
+  #define TX_IPFIL9_PORT0_EN_LBN 18
+  #define TX_IPFIL9_PORT0_EN_WIDTH 1
+  #define TX_IPFIL8_PORT1_EN_LBN 17
+  #define TX_IPFIL8_PORT1_EN_WIDTH 1
+  #define TX_IPFIL8_PORT0_EN_LBN 16
+  #define TX_IPFIL8_PORT0_EN_WIDTH 1
+  #define TX_IPFIL7_PORT1_EN_LBN 15
+  #define TX_IPFIL7_PORT1_EN_WIDTH 1
+  #define TX_IPFIL7_PORT0_EN_LBN 14
+  #define TX_IPFIL7_PORT0_EN_WIDTH 1
+  #define TX_IPFIL6_PORT1_EN_LBN 13
+  #define TX_IPFIL6_PORT1_EN_WIDTH 1
+  #define TX_IPFIL6_PORT0_EN_LBN 12
+  #define TX_IPFIL6_PORT0_EN_WIDTH 1
+  #define TX_IPFIL5_PORT1_EN_LBN 11
+  #define TX_IPFIL5_PORT1_EN_WIDTH 1
+  #define TX_IPFIL5_PORT0_EN_LBN 10
+  #define TX_IPFIL5_PORT0_EN_WIDTH 1
+  #define TX_IPFIL4_PORT1_EN_LBN 9
+  #define TX_IPFIL4_PORT1_EN_WIDTH 1
+  #define TX_IPFIL4_PORT0_EN_LBN 8
+  #define TX_IPFIL4_PORT0_EN_WIDTH 1
+  #define TX_IPFIL3_PORT1_EN_LBN 7
+  #define TX_IPFIL3_PORT1_EN_WIDTH 1
+  #define TX_IPFIL3_PORT0_EN_LBN 6
+  #define TX_IPFIL3_PORT0_EN_WIDTH 1
+  #define TX_IPFIL2_PORT1_EN_LBN 5
+  #define TX_IPFIL2_PORT1_EN_WIDTH 1
+  #define TX_IPFIL2_PORT0_EN_LBN 4
+  #define TX_IPFIL2_PORT0_EN_WIDTH 1
+  #define TX_IPFIL1_PORT1_EN_LBN 3
+  #define TX_IPFIL1_PORT1_EN_WIDTH 1
+  #define TX_IPFIL1_PORT0_EN_LBN 2
+  #define TX_IPFIL1_PORT0_EN_WIDTH 1
+  #define TX_IPFIL0_PORT1_EN_LBN 1
+  #define TX_IPFIL0_PORT1_EN_WIDTH 1
+  #define TX_IPFIL0_PORT0_EN_LBN 0
+  #define TX_IPFIL0_PORT0_EN_WIDTH 1
+#define TX_IPFIL_TBL_OFST 0xB00 // Transmit IP source address filter table
+  #define TX_IPFIL_MASK_LBN 32
+  #define TX_IPFIL_MASK_WIDTH 32
+  #define TX_IP_SRC_ADR_LBN 0
+  #define TX_IP_SRC_ADR_WIDTH 32
+#define TX_PACE_REG_A1_OFST 0xF80000 // Transmit pace control register
+#define TX_PACE_REG_B0_OFST 0xA90    // Transmit pace control register
+  #define TX_PACE_SB_AF_LBN 19
+  #define TX_PACE_SB_AF_WIDTH 10
+  #define TX_PACE_SB_NOTAF_LBN 9
+  #define TX_PACE_SB_NOTAF_WIDTH 10
+  #define TX_PACE_FB_BASE_LBN 5
+  #define TX_PACE_FB_BASE_WIDTH 4
+  #define TX_PACE_BIN_TH_LBN 0
+  #define TX_PACE_BIN_TH_WIDTH 5
+#define TX_PACE_TBL_A1_OFST 0xF80040 // Transmit pacing table
+#define TX_PACE_TBL_FIRST_QUEUE_A1 4
+#define TX_PACE_TBL_B0_OFST 0xF80000 // Transmit pacing table
+#define TX_PACE_TBL_FIRST_QUEUE_B0 0
+  #define TX_PACE_LBN 0
+  #define TX_PACE_WIDTH 5
+
+//////////////---- EE/Flash Registers C Header ----//////////////
+#define EE_SPI_HCMD_REG_KER_OFST 0x100 // SPI host command register
+#define EE_SPI_HCMD_REG_OFST 0x100 // SPI host command register
+  #define EE_SPI_HCMD_CMD_EN_LBN 31
+  #define EE_SPI_HCMD_CMD_EN_WIDTH 1
+  #define EE_WR_TIMER_ACTIVE_LBN 28
+  #define EE_WR_TIMER_ACTIVE_WIDTH 1
+  #define EE_SPI_HCMD_SF_SEL_LBN 24
+  #define EE_SPI_HCMD_SF_SEL_WIDTH 1
+  #define EE_SPI_HCMD_DABCNT_LBN 16
+  #define EE_SPI_HCMD_DABCNT_WIDTH 5
+  #define EE_SPI_HCMD_READ_LBN 15
+  #define EE_SPI_HCMD_READ_WIDTH 1
+  #define EE_SPI_HCMD_DUBCNT_LBN 12
+  #define EE_SPI_HCMD_DUBCNT_WIDTH 2
+  #define EE_SPI_HCMD_ADBCNT_LBN 8
+  #define EE_SPI_HCMD_ADBCNT_WIDTH 2
+  #define EE_SPI_HCMD_ENC_LBN 0
+  #define EE_SPI_HCMD_ENC_WIDTH 8
+#define EE_SPI_HADR_REG_KER_OFST 0X110 // SPI host address register
+#define EE_SPI_HADR_REG_OFST 0X110 // SPI host address register
+  #define EE_SPI_HADR_DUBYTE_LBN 24
+  #define EE_SPI_HADR_DUBYTE_WIDTH 8
+  #define EE_SPI_HADR_ADR_LBN 0
+  #define EE_SPI_HADR_ADR_WIDTH 24
+#define EE_SPI_HDATA_REG_KER_OFST 0x120 // SPI host data register
+#define EE_SPI_HDATA_REG_OFST 0x120 // SPI host data register
+  #define EE_SPI_HDATA3_LBN 96
+  #define EE_SPI_HDATA3_WIDTH 32
+  #define EE_SPI_HDATA2_LBN 64
+  #define EE_SPI_HDATA2_WIDTH 32
+  #define EE_SPI_HDATA1_LBN 32
+  #define EE_SPI_HDATA1_WIDTH 32
+  #define EE_SPI_HDATA0_LBN 0
+  #define EE_SPI_HDATA0_WIDTH 32
+#define EE_BASE_PAGE_REG_KER_OFST 0x130 // Expansion ROM base mirror register
+#define EE_BASE_PAGE_REG_OFST 0x130 // Expansion ROM base mirror register
+  #define EE_EXP_ROM_WINDOW_BASE_LBN 16
+  #define EE_EXP_ROM_WINDOW_BASE_WIDTH 13
+  #define EE_EXPROM_MASK_LBN 0
+  #define EE_EXPROM_MASK_WIDTH 13
+#define EE_VPD_CFG0_REG_KER_OFST 0X140 // SPI/VPD configuration register
+#define EE_VPD_CFG0_REG_OFST 0X140 // SPI/VPD configuration register
+  #define EE_SF_FASTRD_EN_LBN 127
+  #define EE_SF_FASTRD_EN_WIDTH 1
+  #define EE_SF_CLOCK_DIV_LBN 120
+  #define EE_SF_CLOCK_DIV_WIDTH 7
+  #define EE_VPD_WIP_POLL_LBN 119
+  #define EE_VPD_WIP_POLL_WIDTH 1
+  #define EE_VPDW_LENGTH_LBN 80
+  #define EE_VPDW_LENGTH_WIDTH 15
+  #define EE_VPDW_BASE_LBN 64
+  #define EE_VPDW_BASE_WIDTH 15
+  #define EE_VPD_WR_CMD_EN_LBN 56
+  #define EE_VPD_WR_CMD_EN_WIDTH 8
+  #define EE_VPD_BASE_LBN 32
+  #define EE_VPD_BASE_WIDTH 24
+  #define EE_VPD_LENGTH_LBN 16
+  #define EE_VPD_LENGTH_WIDTH 13
+  #define EE_VPD_AD_SIZE_LBN 8
+  #define EE_VPD_AD_SIZE_WIDTH 5
+  #define EE_VPD_ACCESS_ON_LBN 5
+  #define EE_VPD_ACCESS_ON_WIDTH 1
+#define EE_VPD_SW_CNTL_REG_KER_OFST 0X150 // VPD access SW control register
+#define EE_VPD_SW_CNTL_REG_OFST 0X150 // VPD access SW control register
+  #define EE_VPD_CYCLE_PENDING_LBN 31
+  #define EE_VPD_CYCLE_PENDING_WIDTH 1
+  #define EE_VPD_CYC_WRITE_LBN 28
+  #define EE_VPD_CYC_WRITE_WIDTH 1
+  #define EE_VPD_CYC_ADR_LBN 0
+  #define EE_VPD_CYC_ADR_WIDTH 15
+#define EE_VPD_SW_DATA_REG_KER_OFST 0x160 // VPD access SW data register
+#define EE_VPD_SW_DATA_REG_OFST 0x160 // VPD access SW data register
+  #define EE_VPD_CYC_DAT_LBN 0
+  #define EE_VPD_CYC_DAT_WIDTH 32
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/ef_vi_falcon_desc.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/ef_vi_falcon_desc.h
--- linux-2.6.18.8/drivers/xen/sfc_netfront/ef_vi_falcon_desc.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/ef_vi_falcon_desc.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,43 @@
+//////////////---- Descriptors C Headers ----//////////////
+// Receive Kernel IP Descriptor
+  #define RX_KER_BUF_SIZE_LBN 48
+  #define RX_KER_BUF_SIZE_WIDTH 14
+  #define RX_KER_BUF_REGION_LBN 46
+  #define RX_KER_BUF_REGION_WIDTH 2
+      #define RX_KER_BUF_REGION0_DECODE 0
+      #define RX_KER_BUF_REGION1_DECODE 1
+      #define RX_KER_BUF_REGION2_DECODE 2
+      #define RX_KER_BUF_REGION3_DECODE 3
+  #define RX_KER_BUF_ADR_LBN 0
+  #define RX_KER_BUF_ADR_WIDTH 46
+// Receive User IP Descriptor
+  #define RX_USR_2BYTE_OFS_LBN 20
+  #define RX_USR_2BYTE_OFS_WIDTH 12
+  #define RX_USR_BUF_ID_LBN 0
+  #define RX_USR_BUF_ID_WIDTH 20
+// Transmit Kernel IP Descriptor
+  #define TX_KER_PORT_LBN 63
+  #define TX_KER_PORT_WIDTH 1
+  #define TX_KER_CONT_LBN 62
+  #define TX_KER_CONT_WIDTH 1
+  #define TX_KER_BYTE_CNT_LBN 48
+  #define TX_KER_BYTE_CNT_WIDTH 14
+  #define TX_KER_BUF_REGION_LBN 46
+  #define TX_KER_BUF_REGION_WIDTH 2
+      #define TX_KER_BUF_REGION0_DECODE 0
+      #define TX_KER_BUF_REGION1_DECODE 1
+      #define TX_KER_BUF_REGION2_DECODE 2
+      #define TX_KER_BUF_REGION3_DECODE 3
+  #define TX_KER_BUF_ADR_LBN 0
+  #define TX_KER_BUF_ADR_WIDTH 46
+// Transmit User IP Descriptor
+  #define TX_USR_PORT_LBN 47
+  #define TX_USR_PORT_WIDTH 1
+  #define TX_USR_CONT_LBN 46
+  #define TX_USR_CONT_WIDTH 1
+  #define TX_USR_BYTE_CNT_LBN 33
+  #define TX_USR_BYTE_CNT_WIDTH 13
+  #define TX_USR_BUF_ID_LBN 13
+  #define TX_USR_BUF_ID_WIDTH 20
+  #define TX_USR_BYTE_OFS_LBN 0
+  #define TX_USR_BYTE_OFS_WIDTH 13
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/ef_vi_falcon_event.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/ef_vi_falcon_event.h
--- linux-2.6.18.8/drivers/xen/sfc_netfront/ef_vi_falcon_event.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/ef_vi_falcon_event.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,123 @@
+//////////////---- Events Format C Header ----//////////////
+//////////////---- Event entry ----//////////////
+  #define EV_CODE_LBN 60
+  #define EV_CODE_WIDTH 4
+      #define RX_IP_EV_DECODE 0
+      #define TX_IP_EV_DECODE 2
+      #define DRIVER_EV_DECODE 5
+      #define GLOBAL_EV_DECODE 6
+      #define DRV_GEN_EV_DECODE 7
+  #define EV_DATA_LBN 0
+  #define EV_DATA_WIDTH 60
+//////////////---- Receive IP events for both Kernel & User event queues ----//////////////
+  #define RX_EV_PKT_OK_LBN 56
+  #define RX_EV_PKT_OK_WIDTH 1
+  #define RX_EV_BUF_OWNER_ID_ERR_LBN 54
+  #define RX_EV_BUF_OWNER_ID_ERR_WIDTH 1
+  #define RX_EV_IP_HDR_CHKSUM_ERR_LBN 52
+  #define RX_EV_IP_HDR_CHKSUM_ERR_WIDTH 1
+  #define RX_EV_TCP_UDP_CHKSUM_ERR_LBN 51
+  #define RX_EV_TCP_UDP_CHKSUM_ERR_WIDTH 1
+  #define RX_EV_ETH_CRC_ERR_LBN 50
+  #define RX_EV_ETH_CRC_ERR_WIDTH 1
+  #define RX_EV_FRM_TRUNC_LBN 49
+  #define RX_EV_FRM_TRUNC_WIDTH 1
+  #define RX_EV_DRIB_NIB_LBN 48
+  #define RX_EV_DRIB_NIB_WIDTH 1
+  #define RX_EV_TOBE_DISC_LBN 47
+  #define RX_EV_TOBE_DISC_WIDTH 1
+  #define RX_EV_PKT_TYPE_LBN 44
+  #define RX_EV_PKT_TYPE_WIDTH 3
+      #define RX_EV_PKT_TYPE_ETH_DECODE 0
+      #define RX_EV_PKT_TYPE_LLC_DECODE 1
+      #define RX_EV_PKT_TYPE_JUMBO_DECODE 2
+      #define RX_EV_PKT_TYPE_VLAN_DECODE 3
+      #define RX_EV_PKT_TYPE_VLAN_LLC_DECODE 4
+      #define RX_EV_PKT_TYPE_VLAN_JUMBO_DECODE 5
+  #define RX_EV_HDR_TYPE_LBN 42
+  #define RX_EV_HDR_TYPE_WIDTH 2
+      #define RX_EV_HDR_TYPE_TCP_IPV4_DECODE 0
+      #define RX_EV_HDR_TYPE_UDP_IPV4_DECODE 1
+      #define RX_EV_HDR_TYPE_OTHER_IP_DECODE 2
+      #define RX_EV_HDR_TYPE_NON_IP_DECODE 3
+  #define RX_EV_DESC_Q_EMPTY_LBN 41
+  #define RX_EV_DESC_Q_EMPTY_WIDTH 1
+  #define RX_EV_MCAST_HASH_MATCH_LBN 40
+  #define RX_EV_MCAST_HASH_MATCH_WIDTH 1
+  #define RX_EV_MCAST_PKT_LBN 39
+  #define RX_EV_MCAST_PKT_WIDTH 1
+  #define RX_EV_Q_LABEL_LBN 32
+  #define RX_EV_Q_LABEL_WIDTH 5
+  #define RX_JUMBO_CONT_LBN 31
+  #define RX_JUMBO_CONT_WIDTH 1
+  #define RX_SOP_LBN 15
+  #define RX_SOP_WIDTH 1
+  #define RX_PORT_LBN 30
+  #define RX_PORT_WIDTH 1
+  #define RX_EV_BYTE_CNT_LBN 16
+  #define RX_EV_BYTE_CNT_WIDTH 14
+  #define RX_iSCSI_PKT_OK_LBN 14
+  #define RX_iSCSI_PKT_OK_WIDTH 1
+  #define RX_ISCSI_DDIG_ERR_LBN 13
+  #define RX_ISCSI_DDIG_ERR_WIDTH 1
+  #define RX_ISCSI_HDIG_ERR_LBN 12
+  #define RX_ISCSI_HDIG_ERR_WIDTH 1
+  #define RX_EV_DESC_PTR_LBN 0
+  #define RX_EV_DESC_PTR_WIDTH 12
+//////////////---- Transmit IP events for both Kernel & User event queues ----//////////////
+  #define TX_EV_PKT_ERR_LBN 38
+  #define TX_EV_PKT_ERR_WIDTH 1
+  #define TX_EV_PKT_TOO_BIG_LBN 37
+  #define TX_EV_PKT_TOO_BIG_WIDTH 1
+  #define TX_EV_Q_LABEL_LBN 32
+  #define TX_EV_Q_LABEL_WIDTH 5
+  #define TX_EV_PORT_LBN 16
+  #define TX_EV_PORT_WIDTH 1
+  #define TX_EV_WQ_FF_FULL_LBN 15
+  #define TX_EV_WQ_FF_FULL_WIDTH 1
+  #define TX_EV_BUF_OWNER_ID_ERR_LBN 14
+  #define TX_EV_BUF_OWNER_ID_ERR_WIDTH 1
+  #define TX_EV_COMP_LBN 12
+  #define TX_EV_COMP_WIDTH 1
+  #define TX_EV_DESC_PTR_LBN 0
+  #define TX_EV_DESC_PTR_WIDTH 12
+//////////////---- Char or Kernel driver events ----//////////////
+  #define DRIVER_EV_SUB_CODE_LBN 56
+  #define DRIVER_EV_SUB_CODE_WIDTH 4
+      #define TX_DESCQ_FLS_DONE_EV_DECODE 0x0
+      #define RX_DESCQ_FLS_DONE_EV_DECODE 0x1
+      #define EVQ_INIT_DONE_EV_DECODE 0x2
+      #define EVQ_NOT_EN_EV_DECODE 0x3
+      #define RX_DESCQ_FLSFF_OVFL_EV_DECODE 0x4
+      #define SRM_UPD_DONE_EV_DECODE 0x5
+      #define WAKE_UP_EV_DECODE 0x6
+      #define TX_PKT_NON_TCP_UDP_DECODE 0x9
+      #define TIMER_EV_DECODE 0xA
+      #define RX_DSC_ERROR_EV_DECODE 0xE
+  #define DRIVER_EV_TX_DESCQ_ID_LBN 0
+  #define DRIVER_EV_TX_DESCQ_ID_WIDTH 12
+  #define DRIVER_EV_RX_DESCQ_ID_LBN 0
+  #define DRIVER_EV_RX_DESCQ_ID_WIDTH 12
+  #define DRIVER_EV_EVQ_ID_LBN 0
+  #define DRIVER_EV_EVQ_ID_WIDTH 12
+  #define DRIVER_TMR_ID_LBN 0
+  #define DRIVER_TMR_ID_WIDTH 12
+  #define DRIVER_EV_SRM_UPD_LBN 0
+  #define DRIVER_EV_SRM_UPD_WIDTH 2
+      #define SRM_CLR_EV_DECODE 0
+      #define SRM_UPD_EV_DECODE 1
+      #define SRM_ILLCLR_EV_DECODE 2
+//////////////---- Global events. Sent to both event queue 0 and 4. ----//////////////
+  #define XFP_PHY_INTR_LBN 10
+  #define XFP_PHY_INTR_WIDTH 1
+  #define XG_PHY_INTR_LBN 9
+  #define XG_PHY_INTR_WIDTH 1
+  #define G_PHY1_INTR_LBN 8
+  #define G_PHY1_INTR_WIDTH 1
+  #define G_PHY0_INTR_LBN 7
+  #define G_PHY0_INTR_WIDTH 1
+//////////////---- Driver generated events ----//////////////
+  #define DRV_GEN_EV_CODE_LBN 60
+  #define DRV_GEN_EV_CODE_WIDTH 4
+  #define DRV_GEN_EV_DATA_LBN 0
+  #define DRV_GEN_EV_DATA_WIDTH 60
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/ef_vi_falcon.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/ef_vi_falcon.h
--- linux-2.6.18.8/drivers/xen/sfc_netfront/ef_vi_falcon.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/ef_vi_falcon.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,172 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*
+ * \author  slp
+ *  \brief  Falcon specific definitions
+ *   \date  2004/08
+ */
+
+#ifndef __EF_VI_FALCON_H__
+#define __EF_VI_FALCON_H__    
+
+#define EFHW_4K		0x00001000u
+#define EFHW_8K		0x00002000u
+
+/* include the autogenerated register definitions */
+
+#include "ef_vi_falcon_core.h"
+#include "ef_vi_falcon_desc.h"
+#include "ef_vi_falcon_event.h"
+
+
+/*----------------------------------------------------------------------------
+ *
+ * Helpers to turn bit shifts into dword shifts and check that the bit fields 
+ * haven't overflown the dword etc. Aim is to preserve consistency with the 
+ * autogenerated headers - once stable we could hard code.
+ *
+ *---------------------------------------------------------------------------*/
+
+/* mask constructors */
+#define __FALCON_MASK(WIDTH,T)  ((((T)1) << (WIDTH)) - 1)
+#define __EFVI_MASK32(WIDTH)  __FALCON_MASK((WIDTH),uint32_t)
+#define __EFVI_MASK64(WIDTH)  __FALCON_MASK((WIDTH),uint64_t)
+
+#define __EFVI_FALCON_MASKFIELD32(LBN, WIDTH)   ((uint32_t)  \
+                             (__EFVI_MASK32(WIDTH) << (LBN)))
+
+/* constructors for fields which span the first and second dwords */
+#define __LW(LBN) (32 - LBN)
+#define LOW(v, LBN, WIDTH)   ((uint32_t)  \
+                               (((v) & __EFVI_MASK64(__LW((LBN)))) << (LBN)))
+#define HIGH(v, LBN, WIDTH)  ((uint32_t)(((v) >> __LW((LBN))) & \
+                                       __EFVI_MASK64((WIDTH - __LW((LBN))))))
+/* constructors for fields within the second dword */
+#define __DW2(LBN) 	  ((LBN) - 32)
+
+/* constructors for fields which span the second and third dwords */
+#define __LW2(LBN) (64 - LBN)
+#define LOW2(v, LBN, WIDTH) ((uint32_t) \
+                       (((v) & __EFVI_MASK64(__LW2((LBN)))) << ((LBN) - 32)))
+#define HIGH2(v, LBN, WIDTH)  ((uint32_t) \
+             (((v) >> __LW2((LBN))) & __EFVI_MASK64((WIDTH - __LW2((LBN))))))
+
+/* constructors for fields within the third dword */
+#define __DW3(LBN) 	  ((LBN) - 64)
+
+				
+/* constructors for fields which span the third and fourth dwords */
+#define __LW3(LBN) (96 - LBN)
+#define LOW3(v, LBN, WIDTH)   ((uint32_t)    \
+              (((v) & __EFVI_MASK64(__LW3((LBN)))) << ((LBN) - 64)))
+#define HIGH3(v, LBN, WIDTH)  ((unit32_t)    \
+             (((v) >> __LW3((LBN))) & __EFVI_MASK64((WIDTH - __LW3((LBN))))))
+
+/* constructors for fields within the fourth dword */
+#define __DW4(LBN) 	  ((LBN) - 96)
+
+/* checks that the autogenerated headers our consistent with our model */
+#define WIDTHCHCK(a, b) ef_assert((a) == (b))
+#define RANGECHCK(v, WIDTH) \
+                ef_assert(((uint64_t)(v) & ~(__EFVI_MASK64((WIDTH)))) == 0)
+
+/* fields within the first dword */
+#define DWCHCK(LBN, WIDTH) ef_assert(((LBN) >= 0) &&(((LBN)+(WIDTH)) <= 32))
+
+/* fields which span the first and second dwords */
+#define LWCHK(LBN, WIDTH)  ef_assert(WIDTH >= __LW(LBN))
+
+/*----------------------------------------------------------------------------
+ *
+ * Buffer virtual addresses (4K buffers) 
+ *
+ *---------------------------------------------------------------------------*/
+
+/* Form a buffer virtual address from buffer ID and offset.  If the offset
+** is larger than the buffer size, then the buffer indexed will be
+** calculated appropriately.  It is the responsibility of the caller to
+** ensure that they have valid buffers programmed at that address.
+*/
+#define EFVI_FALCON_VADDR_4K_S   	(12)	     
+#define EFVI_FALCON_VADDR_M       0xfffff		/* post shift mask  */
+
+
+#define EFVI_FALCON_BUFFER_4K_ADDR(id,off)      \
+  (((id) << EFVI_FALCON_VADDR_4K_S) + (off))
+
+#define EFVI_FALCON_BUFFER_4K_PAGE(vaddr)                       \
+  (((vaddr) >> EFVI_FALCON_VADDR_4K_S) & EFVI_FALCON_VADDR_M)
+
+#define EFVI_FALCON_BUFFER_4K_OFF(vaddr)                \
+  ((vaddr) & __EFVI_MASK32(EFVI_FALCON_VADDR_4K_S))
+
+
+/*----------------------------------------------------------------------------
+ *
+ * Masks
+ *
+ *---------------------------------------------------------------------------*/
+
+#define EFVI_FALCON_CLOCK_ASIC_HZ    (125000)
+#define EFVI_FALCON_CLOCK_FPGA_HZ    (62500)
+#define EFVI_FALCON_CLOCK_HZ         EFVI_FALCON_CLOCK_ASIC_HZ
+
+
+/*----------------------------------------------------------------------------
+ *
+ * Timers
+ *
+ *---------------------------------------------------------------------------*/
+
+/* Event-Queue Timer granularity - measured in us 
+   Given by: 4096 * 3 cycle * clock period */
+
+#define EFVI_FALCON_EVQTIMER_PERIOD_US   ((4096 * 3 * 1000) / EFVI_FALCON_CLOCK_HZ)
+
+/* mode bits */
+#define EFVI_FALCON_TIMER_MODE_DIS     0     /* disabled */
+#define EFVI_FALCON_TIMER_MODE_RUN     1     /* started counting right away */
+#define EFVI_FALCON_TIMER_MODE_HOLD    2     /* trigger mode (user queues) */
+
+#define EFVI_FALCON_EVQTIMER_HOLD     (EFVI_FALCON_TIMER_MODE_HOLD << TIMER_MODE_LBN)
+#define EFVI_FALCON_EVQTIMER_RUN      (EFVI_FALCON_TIMER_MODE_RUN  << TIMER_MODE_LBN)
+#define EFVI_FALCON_EVQTIMER_DISABLE  (EFVI_FALCON_TIMER_MODE_DIS  << TIMER_MODE_LBN) 
+
+
+/* ---- efhw_event_t helpers --- */
+
+#define EFVI_FALCON_EVENT_CODE(evp) \
+       ((evp)->u64 & EFVI_FALCON_EVENT_CODE_MASK)
+
+#define EFVI_FALCON_EVENT_SW_DATA_MASK    0x0000ffff
+
+#define __EFVI_FALCON_OPEN_MASK(WIDTH)  ((((uint64_t)1) << (WIDTH)) - 1)
+
+#define EFVI_FALCON_EVENT_CODE_MASK \
+           (__EFVI_FALCON_OPEN_MASK(EV_CODE_WIDTH) << EV_CODE_LBN)
+
+
+#endif  /* __EF_VI_FALCON_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/ef_vi_internal.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/ef_vi_internal.h
--- linux-2.6.18.8/drivers/xen/sfc_netfront/ef_vi_internal.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/ef_vi_internal.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,256 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*
+ * \author  djr
+ *  \brief  Really-and-truely-honestly internal stuff for libef.
+ *   \date  2004/06/13
+ */
+
+/*! \cidoxg_include_ci_ul */
+#ifndef __CI_EF_VI_INTERNAL_H__
+#define __CI_EF_VI_INTERNAL_H__
+
+
+/* These flags share space with enum ef_vi_flags. */
+#define EF_VI_BUG5692_WORKAROUND  0x10000
+
+
+/* ***********************************************************************
+ * COMPILATION CONTROL FLAGS (see ef_vi.h for "workaround" controls)
+ */
+
+#define EF_VI_DO_MAGIC_CHECKS 1
+
+
+/**********************************************************************
+ * Headers
+ */
+
+#include <etherfabric/ef_vi.h>
+#include "sysdep.h"
+#include "ef_vi_falcon.h"
+
+
+/**********************************************************************
+ * Debugging.
+ */
+
+#ifndef NDEBUG
+
+# define _ef_assert(exp, file, line) BUG_ON(!(exp));
+
+# define _ef_assert2(exp, x, y, file, line)  do {	\
+		if (unlikely(!(exp)))		\
+			BUG();				\
+	} while (0)
+
+#else
+
+# define _ef_assert(exp, file, line)
+# define _ef_assert2(e, x, y, file, line)
+
+#endif
+
+#define ef_assert(a)          do{ _ef_assert((a),__FILE__,__LINE__); } while(0)
+#define ef_assert_equal(a,b)  _ef_assert2((a)==(b),(a),(b),__FILE__,__LINE__)
+#define ef_assert_eq          ef_assert_equal
+#define ef_assert_lt(a,b)     _ef_assert2((a)<(b),(a),(b),__FILE__,__LINE__)
+#define ef_assert_le(a,b)     _ef_assert2((a)<=(b),(a),(b),__FILE__,__LINE__)
+#define ef_assert_nequal(a,b) _ef_assert2((a)!=(b),(a),(b),__FILE__,__LINE__)
+#define ef_assert_ne          ef_assert_nequal
+#define ef_assert_ge(a,b)     _ef_assert2((a)>=(b),(a),(b),__FILE__,__LINE__)
+#define ef_assert_gt(a,b)     _ef_assert2((a)>(b),(a),(b),__FILE__,__LINE__)
+
+/**********************************************************************
+ * Debug checks. ******************************************************
+ **********************************************************************/
+
+#ifdef NDEBUG
+# define EF_VI_MAGIC_SET(p, type)
+# define EF_VI_CHECK_VI(p)
+# define EF_VI_CHECK_EVENT_Q(p)
+# define EF_VI_CHECK_IOBUFSET(p)
+# define EF_VI_CHECK_FILTER(p)
+# define EF_VI_CHECK_SHMBUF(p)
+# define EF_VI_CHECK_PT_EP(p)
+#else
+# define EF_VI                    0x3
+# define EF_EPLOCK                0x6
+# define EF_IOBUFSET              0x9
+# define EF_FILTER                0xa
+# define EF_SHMBUF                0x11
+
+# define EF_VI_MAGIC(p, type)				\
+	(((unsigned)(type) << 28) |			\
+	 (((unsigned)(intptr_t)(p)) & 0x0fffffffu))
+
+# if !EF_VI_DO_MAGIC_CHECKS
+#  define EF_VI_MAGIC_SET(p, type)
+#  define EF_VI_MAGIC_CHECK(p, type)
+# else
+#  define EF_VI_MAGIC_SET(p, type)			\
+	do {						\
+		(p)->magic = EF_VI_MAGIC((p), (type));	\
+	} while (0)
+
+# define EF_VI_MAGIC_OKAY(p, type)                      \
+	((p)->magic == EF_VI_MAGIC((p), (type)))
+
+# define EF_VI_MAGIC_CHECK(p, type)                     \
+	ef_assert(EF_VI_MAGIC_OKAY((p), (type)))
+
+#endif /* EF_VI_DO_MAGIC_CHECKS */
+
+# define EF_VI_CHECK_VI(p)			\
+	ef_assert(p);				\
+	EF_VI_MAGIC_CHECK((p), EF_VI);
+
+# define EF_VI_CHECK_EVENT_Q(p)			\
+	ef_assert(p);				\
+	EF_VI_MAGIC_CHECK((p), EF_VI);		\
+	ef_assert((p)->evq_base);		\
+	ef_assert((p)->evq_mask);
+
+# define EF_VI_CHECK_PT_EP(p)			\
+	ef_assert(p);				\
+	EF_VI_MAGIC_CHECK((p), EF_VI);		\
+	ef_assert((p)->ep_state);
+
+# define EF_VI_CHECK_IOBUFSET(p)		\
+	ef_assert(p);				\
+	EF_VI_MAGIC_CHECK((p), EF_IOBUFSET)
+
+# define EF_VI_CHECK_FILTER(p)			\
+	ef_assert(p);				\
+	EF_VI_MAGIC_CHECK((p), EF_FILTER);
+
+# define EF_VI_CHECK_SHMBUF(p)			\
+	ef_assert(p);				\
+	EF_VI_MAGIC_CHECK((p), EF_SHMBUF);
+
+#endif
+
+#ifndef NDEBUG
+# define EF_DRIVER_MAGIC 0x00f00ba4
+# define EF_ASSERT_THIS_DRIVER_VALID(driver)				\
+	do{ ef_assert(driver);						\
+		EF_VI_MAGIC_CHECK((driver), EF_DRIVER_MAGIC);		\
+		ef_assert((driver)->init);               }while(0)
+
+# define EF_ASSERT_DRIVER_VALID() EF_ASSERT_THIS_DRIVER_VALID(&ci_driver)
+#else
+# define EF_ASSERT_THIS_DRIVER_VALID(driver)
+# define EF_ASSERT_DRIVER_VALID()
+#endif
+
+
+/* *************************************
+ * Power of 2 FIFO
+ */
+
+#define EF_VI_FIFO2_M(f, x)  ((x) & ((f)->fifo_mask))
+#define ef_vi_fifo2_valid(f) ((f) && (f)->fifo && (f)->fifo_mask > 0 &&	\
+			      (f)->fifo_rd_i <= (f)->fifo_mask       &&	\
+			      (f)->fifo_wr_i <= (f)->fifo_mask       &&	\
+			      EF_VI_IS_POW2((f)->fifo_mask+1u))
+
+#define ef_vi_fifo2_init(f, cap)			\
+	do{ ef_assert(EF_VI_IS_POW2((cap) + 1));	\
+		(f)->fifo_rd_i = (f)->fifo_wr_i = 0u;	\
+		(f)->fifo_mask = (cap);			\
+	}while(0)
+
+#define ef_vi_fifo2_is_empty(f) ((f)->fifo_rd_i == (f)->fifo_wr_i)
+#define ef_vi_fifo2_capacity(f) ((f)->fifo_mask)
+#define ef_vi_fifo2_buf_size(f) ((f)->fifo_mask + 1u)
+#define ef_vi_fifo2_end(f)      ((f)->fifo + ef_vi_fifo2_buf_size(f))
+#define ef_vi_fifo2_peek(f)     ((f)->fifo[(f)->fifo_rd_i])
+#define ef_vi_fifo2_poke(f)     ((f)->fifo[(f)->fifo_wr_i])
+#define ef_vi_fifo2_num(f)   EF_VI_FIFO2_M((f),(f)->fifo_wr_i-(f)->fifo_rd_i)
+
+#define ef_vi_fifo2_wr_prev(f)						\
+	do{ (f)->fifo_wr_i = EF_VI_FIFO2_M((f), (f)->fifo_wr_i - 1u); }while(0)
+#define ef_vi_fifo2_wr_next(f)						\
+	do{ (f)->fifo_wr_i = EF_VI_FIFO2_M((f), (f)->fifo_wr_i + 1u); }while(0)
+#define ef_vi_fifo2_rd_adv(f, n)					\
+	do{ (f)->fifo_rd_i = EF_VI_FIFO2_M((f), (f)->fifo_rd_i + (n)); }while(0)
+#define ef_vi_fifo2_rd_prev(f)						\
+	do{ (f)->fifo_rd_i = EF_VI_FIFO2_M((f), (f)->fifo_rd_i - 1u); }while(0)
+#define ef_vi_fifo2_rd_next(f)						\
+	do{ (f)->fifo_rd_i = EF_VI_FIFO2_M((f), (f)->fifo_rd_i + 1u); }while(0)
+
+#define ef_vi_fifo2_put(f, v)						\
+	do{ ef_vi_fifo2_poke(f) = (v); ef_vi_fifo2_wr_next(f); }while(0)
+#define ef_vi_fifo2_get(f, pv)						\
+	do{ *(pv) = ef_vi_fifo2_peek(f); ef_vi_fifo2_rd_next(f); }while(0)
+
+
+/* *********************************************************************
+ * Eventq handling
+ */
+
+typedef union {
+	uint64_t    u64;
+	struct {
+		uint32_t  a;
+		uint32_t  b;
+	} opaque;
+} ef_vi_event;
+
+
+#define EF_VI_EVENT_OFFSET(q, i)					\
+	(((q)->evq_state->evq_ptr - (i) * sizeof(ef_vi_event)) & (q)->evq_mask)
+
+#define EF_VI_EVENT_PTR(q, i)                                           \
+	((ef_vi_event*) ((q)->evq_base + EF_VI_EVENT_OFFSET((q), (i))))
+
+/* *********************************************************************
+ * Miscellaneous goodies
+ */
+#ifdef NDEBUG
+# define EF_VI_DEBUG(x)
+#else
+# define EF_VI_DEBUG(x)            x
+#endif
+
+#define EF_VI_ROUND_UP(i, align)   (((i)+(align)-1u) & ~((align)-1u))
+#define EF_VI_ALIGN_FWD(p, align)  (((p)+(align)-1u) & ~((align)-1u))
+#define EF_VI_ALIGN_BACK(p, align) ((p) & ~((align)-1u))
+#define EF_VI_PTR_ALIGN_BACK(p, align)					\
+	((char*)EF_VI_ALIGN_BACK(((intptr_t)(p)), ((intptr_t)(align))))
+#define EF_VI_IS_POW2(x)           ((x) && ! ((x) & ((x) - 1)))
+
+
+/* ******************************************************************** 
+ */
+
+extern void falcon_vi_init(ef_vi*, void* vvis ) EF_VI_HF;
+extern void ef_eventq_state_init(ef_vi* evq) EF_VI_HF;
+extern void __ef_init(void) EF_VI_HF;
+
+
+#endif  /* __CI_EF_VI_INTERNAL_H__ */
+
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/etherfabric/ef_vi.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/etherfabric/ef_vi.h
--- linux-2.6.18.8/drivers/xen/sfc_netfront/etherfabric/ef_vi.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/etherfabric/ef_vi.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,665 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*
+ *  \brief  Virtual Interface
+ *   \date  2007/05/16
+ */
+
+#ifndef __EFAB_EF_VI_H__
+#define __EFAB_EF_VI_H__
+
+
+/**********************************************************************
+ * Primitive types ****************************************************
+ **********************************************************************/
+
+/* We standardise on the types from stdint.h and synthesise these types
+ * for compilers/platforms that don't provide them */
+
+#  include <linux/types.h>
+# define EF_VI_ALIGN(x) __attribute__ ((aligned (x)))
+# define ef_vi_inline static inline
+
+
+
+/**********************************************************************
+ * Types **************************************************************
+ **********************************************************************/
+
+typedef uint32_t                ef_eventq_ptr;
+
+typedef uint64_t                ef_addr;
+typedef char*                   ef_vi_ioaddr_t;
+
+/**********************************************************************
+ * ef_event ***********************************************************
+ **********************************************************************/
+
+/*! \i_ef_vi A DMA request identifier.
+**
+** This is an integer token specified by the transport and associated
+** with a DMA request.  It is returned to the VI user with DMA completion
+** events.  It is typically used to identify the buffer associated with
+** the transfer.
+*/
+typedef int			ef_request_id;
+
+typedef union {
+	uint64_t  u64[1];
+	uint32_t  u32[2];
+} ef_vi_qword;
+
+typedef ef_vi_qword             ef_hw_event;
+
+#define EF_REQUEST_ID_BITS      16u
+#define EF_REQUEST_ID_MASK      ((1u << EF_REQUEST_ID_BITS) - 1u)
+
+/*! \i_ef_event An [ef_event] is a token that identifies something that
+** has happened.  Examples include packets received, packets transmitted
+** and errors.
+*/
+typedef union {
+	struct {
+		ef_hw_event    ev;
+		unsigned       type       :16;
+	} generic;
+	struct {
+		ef_hw_event    ev;
+		unsigned       type       :16;
+		/*ef_request_id  request_id :EF_REQUEST_ID_BITS;*/
+		unsigned       q_id       :16;
+		unsigned       len        :16;
+		unsigned       flags      :16;
+	} rx;
+	struct {  /* This *must* have same layout as [rx]. */
+		ef_hw_event    ev;
+		unsigned       type       :16;
+		/*ef_request_id  request_id :EF_REQUEST_ID_BITS;*/
+		unsigned       q_id       :16;
+		unsigned       len        :16;
+		unsigned       flags      :16;
+		unsigned       subtype    :16;
+	} rx_discard;
+	struct {
+		ef_hw_event    ev;
+		unsigned       type       :16;
+		/*ef_request_id  request_id :EF_REQUEST_ID_BITS;*/
+		unsigned       q_id       :16;
+	} tx;
+	struct {
+		ef_hw_event    ev;
+		unsigned       type       :16;
+		/*ef_request_id  request_id :EF_REQUEST_ID_BITS;*/
+		unsigned       q_id       :16;
+		unsigned       subtype    :16;
+	} tx_error;
+	struct {
+		ef_hw_event    ev;
+		unsigned       type       :16;
+		unsigned       q_id       :16;
+	} rx_no_desc_trunc;
+	struct {
+		ef_hw_event    ev;
+		unsigned       type       :16;
+		unsigned       data;
+	} sw;
+} ef_event;
+
+
+#define EF_EVENT_TYPE(e)        ((e).generic.type)
+enum {
+	/** Good data was received. */
+	EF_EVENT_TYPE_RX,
+	/** Packets have been sent. */
+	EF_EVENT_TYPE_TX,
+	/** Data received and buffer consumed, but something is wrong. */
+	EF_EVENT_TYPE_RX_DISCARD,
+	/** Transmit of packet failed. */
+	EF_EVENT_TYPE_TX_ERROR,
+	/** Received packet was truncated due to lack of descriptors. */
+	EF_EVENT_TYPE_RX_NO_DESC_TRUNC,
+	/** Software generated event. */
+	EF_EVENT_TYPE_SW,
+	/** Event queue overflow. */
+	EF_EVENT_TYPE_OFLOW,
+};
+
+#define EF_EVENT_RX_BYTES(e)    ((e).rx.len)
+#define EF_EVENT_RX_Q_ID(e)     ((e).rx.q_id)
+#define EF_EVENT_RX_CONT(e)     ((e).rx.flags & EF_EVENT_FLAG_CONT)
+#define EF_EVENT_RX_SOP(e)      ((e).rx.flags & EF_EVENT_FLAG_SOP)
+#define EF_EVENT_RX_ISCSI_OKAY(e) ((e).rx.flags & EF_EVENT_FLAG_ISCSI_OK)
+#define EF_EVENT_FLAG_SOP       0x1
+#define EF_EVENT_FLAG_CONT      0x2
+#define EF_EVENT_FLAG_ISCSI_OK  0x4
+
+#define EF_EVENT_TX_Q_ID(e)     ((e).tx.q_id)
+
+#define EF_EVENT_RX_DISCARD_Q_ID(e)  ((e).rx_discard.q_id)
+#define EF_EVENT_RX_DISCARD_LEN(e)   ((e).rx_discard.len)
+#define EF_EVENT_RX_DISCARD_TYPE(e)  ((e).rx_discard.subtype)
+enum {
+	EF_EVENT_RX_DISCARD_CSUM_BAD,
+	EF_EVENT_RX_DISCARD_CRC_BAD,
+	EF_EVENT_RX_DISCARD_TRUNC,
+	EF_EVENT_RX_DISCARD_RIGHTS,
+	EF_EVENT_RX_DISCARD_OTHER,
+};
+
+#define EF_EVENT_TX_ERROR_Q_ID(e)    ((e).tx_error.q_id)
+#define EF_EVENT_TX_ERROR_TYPE(e)    ((e).tx_error.subtype)
+enum {
+	EF_EVENT_TX_ERROR_RIGHTS,
+	EF_EVENT_TX_ERROR_OFLOW,
+	EF_EVENT_TX_ERROR_2BIG,
+	EF_EVENT_TX_ERROR_BUS,
+};
+
+#define EF_EVENT_RX_NO_DESC_TRUNC_Q_ID(e)  ((e).rx_no_desc_trunc.q_id)
+
+#define EF_EVENT_SW_DATA_MASK   0xffff
+#define EF_EVENT_SW_DATA(e)     ((e).sw.data)
+
+#define EF_EVENT_FMT            "[ev:%x:%08x:%08x]"
+#define EF_EVENT_PRI_ARG(e)     (unsigned) (e).generic.type,    \
+		(unsigned) (e).generic.ev.u32[1],		\
+		(unsigned) (e).generic.ev.u32[0]
+
+#define EF_GET_HW_EV(e)         ((e).generic.ev)
+#define EF_GET_HW_EV_PTR(e)     (&(e).generic.ev)
+#define EF_GET_HW_EV_U64(e)     ((e).generic.ev.u64[0])
+
+
+/* ***************** */
+
+/*! Used by netif shared state. Must use types of explicit size. */
+typedef struct {
+	uint16_t              rx_last_desc_ptr;   /* for RX duplicates       */
+	uint8_t               bad_sop;            /* bad SOP detected        */
+	uint8_t               frag_num;           /* next fragment #, 0=>SOP */
+} ef_rx_dup_state_t;
+
+
+/* Max number of ports on any SF NIC. */
+#define EFAB_DMAQS_PER_EVQ_MAX 32
+
+typedef struct {
+	ef_eventq_ptr	        evq_ptr;
+	int32_t               trashed;
+	ef_rx_dup_state_t     rx_dup_state[EFAB_DMAQS_PER_EVQ_MAX];
+} ef_eventq_state;
+
+
+/*! \i_ef_base [ef_iovec] is similar the standard [struct iovec].  An
+** array of these is used to designate a scatter/gather list of I/O
+** buffers.
+*/
+typedef struct {
+	ef_addr                       iov_base EF_VI_ALIGN(8);
+	unsigned                      iov_len;
+} ef_iovec;
+
+/* Falcon constants */
+#define TX_EV_DESC_PTR_LBN 0
+
+/**********************************************************************
+ * ef_iobufset ********************************************************
+ **********************************************************************/
+
+/*! \i_ef_bufs An [ef_iobufset] is a collection of buffers to be used
+** with the NIC.
+*/
+typedef struct ef_iobufset {
+	unsigned                      magic;
+	unsigned                      bufs_mmap_bytes;
+	unsigned                      bufs_handle;
+	int                           bufs_ptr_off;
+	ef_addr                       bufs_addr;
+	unsigned                      bufs_size; /* size rounded to pow2 */
+	int                           bufs_num;
+	int                           faultonaccess;
+} ef_iobufset;
+
+
+/**********************************************************************
+ * ef_vi **************************************************************
+ **********************************************************************/
+
+enum ef_vi_flags {
+	EF_VI_RX_SCATTER        = 0x1,
+	EF_VI_ISCSI_RX_HDIG     = 0x2,
+	EF_VI_ISCSI_TX_HDIG     = 0x4,
+	EF_VI_ISCSI_RX_DDIG     = 0x8,
+	EF_VI_ISCSI_TX_DDIG     = 0x10,
+	EF_VI_TX_PHYS_ADDR      = 0x20,
+	EF_VI_RX_PHYS_ADDR      = 0x40,
+	EF_VI_TX_IP_CSUM_DIS    = 0x80,
+	EF_VI_TX_TCPUDP_CSUM_DIS= 0x100,
+	EF_VI_TX_TCPUDP_ONLY    = 0x200,
+	/* Flags in range 0xXXXX0000 are for internal use. */
+};
+
+typedef struct {
+	uint32_t  added;
+	uint32_t  removed;
+} ef_vi_txq_state;
+
+typedef struct {
+	uint32_t  added;
+	uint32_t  removed;
+} ef_vi_rxq_state;
+
+typedef struct {
+	uint32_t         mask;
+	void*            doorbell;
+	void*            descriptors;
+	uint16_t*        ids;
+	unsigned         misalign_mask;
+} ef_vi_txq;
+
+typedef struct {
+	uint32_t         mask;
+	void*            doorbell;
+	void*            descriptors;
+	uint16_t*        ids;
+} ef_vi_rxq;
+
+typedef struct {
+	ef_eventq_state  evq;
+	ef_vi_txq_state  txq;
+	ef_vi_rxq_state  rxq;
+	/* Followed by request id fifos. */
+} ef_vi_state;
+
+/*! \i_ef_vi  A virtual interface.
+**
+** An [ef_vi] represents a virtual interface on a specific NIC.  A
+** virtual interface is a collection of an event queue and two DMA queues
+** used to pass Ethernet frames between the transport implementation and
+** the network.
+*/
+typedef struct ef_vi {
+	unsigned			magic;
+
+	unsigned                      vi_resource_id;
+	unsigned                      vi_resource_handle_hack;
+	unsigned                      vi_i;
+
+	char*				vi_mem_mmap_ptr;
+	int                           vi_mem_mmap_bytes;
+	char*				vi_io_mmap_ptr;
+	int                           vi_io_mmap_bytes;
+
+	ef_eventq_state*              evq_state;
+	char*                         evq_base;
+	unsigned                      evq_mask;
+	ef_vi_ioaddr_t                evq_timer_reg;
+
+	ef_vi_txq                     vi_txq;
+	ef_vi_rxq                     vi_rxq;
+	ef_vi_state*                  ep_state;
+	enum ef_vi_flags              vi_flags;
+} ef_vi;
+
+
+enum ef_vi_arch {
+	EF_VI_ARCH_FALCON,
+};
+
+
+struct ef_vi_nic_type {
+	unsigned char  arch;
+	char           variant;
+	unsigned char  revision;
+};
+
+
+/* This structure is opaque to the client & used to pass mapping data
+ * from the resource manager to the ef_vi lib. for ef_vi_init().
+ */
+struct vi_mappings {
+	uint32_t         signature;
+# define VI_MAPPING_VERSION   0x02  /*Byte: Increment me if struct altered*/
+# define VI_MAPPING_SIGNATURE (0xBA1150 + VI_MAPPING_VERSION)
+
+	struct ef_vi_nic_type nic_type;
+
+	int              vi_instance;
+
+	unsigned         evq_bytes;
+	char*            evq_base;
+	ef_vi_ioaddr_t   evq_timer_reg;
+
+	unsigned         rx_queue_capacity;
+	ef_vi_ioaddr_t   rx_dma_ef1;
+	char*            rx_dma_falcon;
+	ef_vi_ioaddr_t   rx_bell;
+
+	unsigned         tx_queue_capacity;
+	ef_vi_ioaddr_t   tx_dma_ef1;
+	char*            tx_dma_falcon;
+	ef_vi_ioaddr_t   tx_bell;
+};
+/* This is used by clients to allocate a suitably sized buffer for the 
+ * resource manager to fill & ef_vi_init() to use. */
+#define VI_MAPPINGS_SIZE (sizeof(struct vi_mappings))
+
+
+/**********************************************************************
+ * ef_config **********************************************************
+ **********************************************************************/
+
+struct ef_config_t {
+	int   log;                    /* debug logging level          */
+};
+
+extern struct ef_config_t  ef_config;
+
+
+/**********************************************************************
+ * ef_vi **************************************************************
+ **********************************************************************/
+
+/* Initialise [data_area] with information required to initialise an ef_vi.
+ * In the following, an unused param should be set to NULL. Note the case
+ * marked (*) of [iobuf_mmap] for falcon/driver; for normal driver this
+ * must be NULL.
+ *
+ * \param  data_area     [in,out] required, must ref at least VI_MAPPINGS_SIZE 
+ *                                bytes
+ * \param  evq_capacity  [in] number of events in event queue.  Specify 0 for
+ *                            no event queue.
+ * \param  rxq_capacity  [in] number of descriptors in RX DMA queue.  Specify
+ *                            0 for no RX queue.
+ * \param  txq_capacity  [in] number of descriptors in TX DMA queue.  Specify
+ *                            0 for no TX queue.
+ * \param  mmap_info     [in] mem-map info for resource
+ * \param  io_mmap       [in] ef1,    required
+ *                            falcon, required
+ * \param  iobuf_mmap    [in] ef1,    UL: unused
+ *                            falcon, UL: required
+ */
+extern void ef_vi_init_mapping_vi(void* data_area, struct ef_vi_nic_type,
+                                  unsigned rxq_capacity,
+                                  unsigned txq_capacity, int instance,
+                                  void* io_mmap, void* iobuf_mmap_rx,
+                                  void* iobuf_mmap_tx, enum ef_vi_flags);
+
+
+extern void ef_vi_init_mapping_evq(void* data_area, struct ef_vi_nic_type,
+                                   int instance, unsigned evq_bytes,
+                                   void* base, void* timer_reg);
+
+ef_vi_inline unsigned ef_vi_resource_id(ef_vi* vi)
+{ 
+	return vi->vi_resource_id; 
+}
+
+ef_vi_inline enum ef_vi_flags ef_vi_flags(ef_vi* vi)
+{ 
+	return vi->vi_flags; 
+}
+
+
+/**********************************************************************
+ * Receive interface **************************************************
+ **********************************************************************/
+
+/*! \i_ef_vi Returns the amount of space in the RX descriptor ring.
+**
+** \return the amount of space in the queue.
+*/
+ef_vi_inline int ef_vi_receive_space(ef_vi* vi) 
+{
+	ef_vi_rxq_state* qs = &vi->ep_state->rxq;
+	return vi->vi_rxq.mask - (qs->added - qs->removed);
+}
+
+
+/*! \i_ef_vi Returns the fill level of the RX descriptor ring.
+**
+** \return the fill level of the queue.
+*/
+ef_vi_inline int ef_vi_receive_fill_level(ef_vi* vi) 
+{
+	ef_vi_rxq_state* qs = &vi->ep_state->rxq;
+	return qs->added - qs->removed;
+}
+
+
+ef_vi_inline int ef_vi_receive_capacity(ef_vi* vi)
+{ 
+	return vi->vi_rxq.mask;
+}
+
+/*! \i_ef_vi  Complete a receive operation.
+**
+** When a receive completion event is received, it should be passed to
+** this function.  The request-id for the buffer that the packet was
+** delivered to is returned.
+**
+** After this function returns, more space may be available in the
+** receive queue.
+*/
+extern ef_request_id ef_vi_receive_done(const ef_vi*, const ef_event*);
+
+/*! \i_ef_vi  Return request ID indicated by a receive event
+ */
+ef_vi_inline ef_request_id ef_vi_receive_request_id(const ef_vi* vi,
+                                                    const ef_event* ef_ev)
+{
+	const ef_vi_qword* ev = EF_GET_HW_EV_PTR(*ef_ev);
+	return ev->u32[0] & vi->vi_rxq.mask;
+}
+  
+
+/*! \i_ef_vi  Form a receive descriptor.
+**
+** If \c initial_rx_bytes is zero use a reception size at least as large
+** as an MTU.
+*/
+extern int ef_vi_receive_init(ef_vi* vi, ef_addr addr, ef_request_id dma_id,
+                              int intial_rx_bytes);
+
+/*! \i_ef_vi  Submit initialised receive descriptors to the NIC. */
+extern void ef_vi_receive_push(ef_vi* vi);
+
+/*! \i_ef_vi  Post a buffer on the receive queue.
+**
+**   \return 0 on success, or -EAGAIN if the receive queue is full
+*/
+extern int ef_vi_receive_post(ef_vi*, ef_addr addr,
+			      ef_request_id dma_id);
+
+/**********************************************************************
+ * Transmit interface *************************************************
+ **********************************************************************/
+
+/*! \i_ef_vi Return the amount of space (in descriptors) in the transmit
+**           queue.
+**
+** \return the amount of space in the queue (in descriptors)
+*/
+ef_vi_inline int ef_vi_transmit_space(ef_vi* vi) 
+{
+	ef_vi_txq_state* qs = &vi->ep_state->txq;
+	return vi->vi_txq.mask - (qs->added - qs->removed);
+}
+
+
+/*! \i_ef_vi Returns the fill level of the TX descriptor ring.
+**
+** \return the fill level of the queue.
+*/
+ef_vi_inline int ef_vi_transmit_fill_level(ef_vi* vi)
+{
+	ef_vi_txq_state* qs = &vi->ep_state->txq;
+	return qs->added - qs->removed;
+}
+
+
+/*! \i_ef_vi Returns the total capacity of the TX descriptor ring.
+**
+** \return the capacity of the queue.
+*/
+ef_vi_inline int ef_vi_transmit_capacity(ef_vi* vi)
+{ 
+	return vi->vi_txq.mask;
+}
+
+
+/*! \i_ef_vi  Transmit a packet.
+**
+**   \param bytes must be greater than ETH_ZLEN.
+**   \return -EAGAIN if the transmit queue is full, or 0 on success
+*/
+extern int ef_vi_transmit(ef_vi*, ef_addr, int bytes, ef_request_id dma_id);
+
+/*! \i_ef_vi  Transmit a packet using a gather list.
+**
+**   \param iov_len must be greater than zero
+**   \param iov the first must be non-zero in length (but others need not)
+**
+**   \return -EAGAIN if the queue is full, or 0 on success
+*/
+extern int ef_vi_transmitv(ef_vi*, const ef_iovec* iov, int iov_len,
+                           ef_request_id dma_id);
+
+/*! \i_ef_vi  Initialise a DMA request.
+**
+** \return -EAGAIN if the queue is full, or 0 on success
+*/
+extern int ef_vi_transmit_init(ef_vi*, ef_addr, int bytes,
+                               ef_request_id dma_id);
+
+/*! \i_ef_vi  Initialise a DMA request.
+**
+** \return -EAGAIN if the queue is full, or 0 on success
+*/
+extern int ef_vi_transmitv_init(ef_vi*, const ef_iovec*, int iov_len,
+                                ef_request_id dma_id);
+
+/*! \i_ef_vi  Submit DMA requests to the NIC.
+**
+** The DMA requests must have been initialised using
+** ef_vi_transmit_init() or ef_vi_transmitv_init().
+*/
+extern void ef_vi_transmit_push(ef_vi*);
+
+
+/*! \i_ef_vi Maximum number of transmit completions per transmit event. */
+#define EF_VI_TRANSMIT_BATCH  64
+
+/*! \i_ef_vi Determine the set of [ef_request_id]s for each DMA request
+**           which has been completed by a given transmit completion
+**           event.
+**
+** \param ids must point to an array of length EF_VI_TRANSMIT_BATCH
+** \return the number of valid [ef_request_id]s (can be zero)
+*/
+extern int ef_vi_transmit_unbundle(ef_vi* ep, const ef_event*,
+                                   ef_request_id* ids);
+
+
+/*! \i_ef_event Returns true if ef_eventq_poll() will return event(s). */
+extern int ef_eventq_has_event(ef_vi* vi);
+
+/*! \i_ef_event Returns true if there are quite a few events in the event
+** queue.
+**
+** This looks ahead in the event queue, so has the property that it will
+** not ping-pong a cache-line when it is called concurrently with events
+** being delivered.
+*/
+extern int ef_eventq_has_many_events(ef_vi* evq, int look_ahead);
+
+/*! Type of function to handle unknown events arriving on event queue
+**  Return CI_TRUE iff the event has been handled.
+*/
+typedef int/*bool*/ ef_event_handler_fn(void* priv, ef_vi* evq, ef_event* ev);
+
+/*! Standard poll exception routine */
+extern int/*bool*/ ef_eventq_poll_exception(void* priv, ef_vi* evq,
+                                            ef_event* ev);
+
+/*! \i_ef_event  Retrieve events from the event queue, handle RX/TX events
+**  and pass any others to an exception handler function
+**
+**   \return The number of events retrieved.
+*/
+extern int ef_eventq_poll_evs(ef_vi* evq, ef_event* evs, int evs_len,
+                              ef_event_handler_fn *exception, void *expt_priv);
+
+/*! \i_ef_event  Retrieve events from the event queue.
+**
+**   \return The number of events retrieved.
+*/
+ef_vi_inline int ef_eventq_poll(ef_vi* evq, ef_event* evs, int evs_len)
+{
+	return ef_eventq_poll_evs(evq, evs, evs_len,
+                            &ef_eventq_poll_exception, (void*)0);
+}
+
+/*! \i_ef_event Returns the capacity of an event queue. */
+ef_vi_inline int ef_eventq_capacity(ef_vi* vi) 
+{
+	return (vi->evq_mask + 1u) / sizeof(ef_hw_event);
+}
+
+/* Returns the instance ID of [vi] */
+ef_vi_inline unsigned ef_vi_instance(ef_vi* vi)
+{ return vi->vi_i; }
+
+
+/**********************************************************************
+ * Initialisation *****************************************************
+ **********************************************************************/
+
+/*! Return size of state buffer of an initialised VI. */
+extern int ef_vi_state_bytes(ef_vi*);
+
+/*! Return size of buffer needed for VI state given sizes of RX and TX
+** DMA queues.  Queue sizes must be legal sizes (power of 2), or 0 (no
+** queue).
+*/
+extern int ef_vi_calc_state_bytes(int rxq_size, int txq_size);
+
+/*! Initialise [ef_vi] from the provided resources. [vvis] must have been
+** created by ef_make_vi_data() & remains owned by the caller.
+*/
+extern void ef_vi_init(ef_vi*, void* vi_info, ef_vi_state* state,
+                       ef_eventq_state* evq_state, enum ef_vi_flags);
+
+extern void ef_vi_state_init(ef_vi*);
+extern void ef_eventq_state_init(ef_vi*);
+
+/*! Convert an efhw device arch to ef_vi_arch, or returns -1 if not
+** recognised.
+*/
+extern int  ef_vi_arch_from_efhw_arch(int efhw_arch);
+
+
+#endif /* __EFAB_EF_VI_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/falcon_event.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/falcon_event.c
--- linux-2.6.18.8/drivers/xen/sfc_netfront/falcon_event.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/falcon_event.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,346 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*
+ * \author  djr
+ *  \brief  Routine to poll event queues.
+ *   \date  2003/03/04
+ */
+
+/*! \cidoxg_lib_ef */
+#include "ef_vi_internal.h"
+
+/* Be worried about this on byteswapped machines */
+/* Due to crazy chipsets, we see the event words being written in
+** arbitrary order (bug4539).  So test for presence of event must ensure
+** that both halves have changed from the null.
+*/
+# define EF_VI_IS_EVENT(evp)						\
+	( (((evp)->opaque.a != (uint32_t)-1) &&				\
+	   ((evp)->opaque.b != (uint32_t)-1)) )
+
+
+#ifdef NDEBUG
+# define IS_DEBUG 0
+#else
+# define IS_DEBUG 1
+#endif
+
+
+/*! Check for RX events with inconsistent SOP/CONT
+**
+** Returns true if this event should be discarded
+*/
+ef_vi_inline int ef_eventq_is_rx_sop_cont_bad_efab(ef_vi* vi,
+						   const ef_vi_qword* ev)
+{
+	ef_rx_dup_state_t* rx_dup_state;
+	uint8_t* bad_sop;
+
+	unsigned label = QWORD_GET_U(RX_EV_Q_LABEL, *ev);
+	unsigned sop   = QWORD_TEST_BIT(RX_SOP, *ev);
+  
+	ef_assert(vi);
+	ef_assert_lt(label, EFAB_DMAQS_PER_EVQ_MAX);
+
+	rx_dup_state = &vi->evq_state->rx_dup_state[label];
+	bad_sop = &rx_dup_state->bad_sop;
+
+	if( ! ((vi->vi_flags & EF_VI_BUG5692_WORKAROUND) || IS_DEBUG) ) {
+		*bad_sop = (*bad_sop && !sop);
+	}
+	else {
+		unsigned cont  = QWORD_TEST_BIT(RX_JUMBO_CONT, *ev);
+		uint8_t *frag_num = &rx_dup_state->frag_num;
+
+		/* bad_sop should latch till the next sop */
+		*bad_sop = (*bad_sop && !sop) || ( !!sop != (*frag_num==0) );
+
+		/* we do not check the number of bytes relative to the
+		 * fragment number and size of the user rx buffer here
+		 * because we don't know the size of the user rx
+		 * buffer - we probably should perform this check in
+		 * the nearest code calling this though.
+		 */
+		*frag_num = cont ? (*frag_num + 1) : 0;
+	}
+
+	return *bad_sop;
+}
+
+
+ef_vi_inline int falcon_rx_check_dup(ef_vi* evq, ef_event* ev_out,
+				     const ef_vi_qword* ev)
+{
+	unsigned q_id = QWORD_GET_U(RX_EV_Q_LABEL, *ev);
+	unsigned desc_ptr = QWORD_GET_U(RX_EV_DESC_PTR, *ev);
+	ef_rx_dup_state_t* rx_dup_state = &evq->evq_state->rx_dup_state[q_id];
+
+	if(likely( desc_ptr != rx_dup_state->rx_last_desc_ptr )) {
+		rx_dup_state->rx_last_desc_ptr = desc_ptr;
+		return 0;
+	}
+
+	rx_dup_state->rx_last_desc_ptr = desc_ptr;
+	rx_dup_state->bad_sop = 1;
+#ifndef NDEBUG
+	rx_dup_state->frag_num = 0;
+#endif
+	BUG_ON(!QWORD_TEST_BIT(RX_EV_FRM_TRUNC, *ev));
+	BUG_ON( QWORD_TEST_BIT(RX_EV_PKT_OK, *ev));
+	BUG_ON(!QWORD_GET_U(RX_EV_BYTE_CNT, *ev) == 0);
+	ev_out->rx_no_desc_trunc.type = EF_EVENT_TYPE_RX_NO_DESC_TRUNC;
+	ev_out->rx_no_desc_trunc.q_id = q_id;
+	return 1;
+}
+
+
+ef_vi_inline void falcon_rx_event(ef_event* ev_out, const ef_vi_qword* ev)
+{
+	if(likely( QWORD_TEST_BIT(RX_EV_PKT_OK, *ev) )) {
+		ev_out->rx.type = EF_EVENT_TYPE_RX;
+		ev_out->rx.q_id = QWORD_GET_U(RX_EV_Q_LABEL, *ev);
+		ev_out->rx.len  = QWORD_GET_U(RX_EV_BYTE_CNT, *ev);
+		if( QWORD_TEST_BIT(RX_SOP, *ev) )
+			ev_out->rx.flags = EF_EVENT_FLAG_SOP;
+		else
+			ev_out->rx.flags = 0;
+		if( QWORD_TEST_BIT(RX_JUMBO_CONT, *ev) )
+			ev_out->rx.flags |= EF_EVENT_FLAG_CONT;
+		if( QWORD_TEST_BIT(RX_iSCSI_PKT_OK, *ev) )
+			ev_out->rx.flags |= EF_EVENT_FLAG_ISCSI_OK;
+	}
+	else {
+		ev_out->rx_discard.type = EF_EVENT_TYPE_RX_DISCARD;
+		ev_out->rx_discard.q_id = QWORD_GET_U(RX_EV_Q_LABEL, *ev);
+		ev_out->rx_discard.len  = QWORD_GET_U(RX_EV_BYTE_CNT, *ev);
+#if 1  /* hack for ptloop compatability: ?? TODO purge */
+		if( QWORD_TEST_BIT(RX_SOP, *ev) )
+			ev_out->rx_discard.flags = EF_EVENT_FLAG_SOP;
+		else
+			ev_out->rx_discard.flags = 0;
+		if( QWORD_TEST_BIT(RX_JUMBO_CONT, *ev) )
+			ev_out->rx_discard.flags |= EF_EVENT_FLAG_CONT;
+		if( QWORD_TEST_BIT(RX_iSCSI_PKT_OK, *ev) )
+			ev_out->rx_discard.flags |= EF_EVENT_FLAG_ISCSI_OK;
+#endif
+		/* Order matters here: more fundamental errors first. */
+		if( QWORD_TEST_BIT(RX_EV_BUF_OWNER_ID_ERR, *ev) )
+			ev_out->rx_discard.subtype = 
+				EF_EVENT_RX_DISCARD_RIGHTS;
+		else if( QWORD_TEST_BIT(RX_EV_FRM_TRUNC, *ev) )
+			ev_out->rx_discard.subtype = 
+				EF_EVENT_RX_DISCARD_TRUNC;
+		else if( QWORD_TEST_BIT(RX_EV_ETH_CRC_ERR, *ev) )
+			ev_out->rx_discard.subtype = 
+				EF_EVENT_RX_DISCARD_CRC_BAD;
+		else if( QWORD_TEST_BIT(RX_EV_IP_HDR_CHKSUM_ERR, *ev) )
+			ev_out->rx_discard.subtype = 
+				EF_EVENT_RX_DISCARD_CSUM_BAD;
+		else if( QWORD_TEST_BIT(RX_EV_TCP_UDP_CHKSUM_ERR, *ev) )
+			ev_out->rx_discard.subtype = 
+				EF_EVENT_RX_DISCARD_CSUM_BAD;
+		else
+			ev_out->rx_discard.subtype = 
+				EF_EVENT_RX_DISCARD_OTHER;
+	}
+}
+
+
+ef_vi_inline void falcon_tx_event(ef_event* ev_out, const ef_vi_qword* ev)
+{
+	/* Danger danger!  No matter what we ask for wrt batching, we
+	** will get a batched event every 16 descriptors, and we also
+	** get dma-queue-empty events.  i.e. Duplicates are expected.
+	**
+	** In addition, if it's been requested in the descriptor, we
+	** get an event per descriptor.  (We don't currently request
+	** this).
+	*/
+	if(likely( QWORD_TEST_BIT(TX_EV_COMP, *ev) )) {
+		ev_out->tx.type = EF_EVENT_TYPE_TX;
+		ev_out->tx.q_id = QWORD_GET_U(TX_EV_Q_LABEL, *ev);
+	}
+	else {
+		ev_out->tx_error.type = EF_EVENT_TYPE_TX_ERROR;
+		ev_out->tx_error.q_id = QWORD_GET_U(TX_EV_Q_LABEL, *ev);
+		if(likely( QWORD_TEST_BIT(TX_EV_BUF_OWNER_ID_ERR, *ev) ))
+			ev_out->tx_error.subtype = EF_EVENT_TX_ERROR_RIGHTS;
+		else if(likely( QWORD_TEST_BIT(TX_EV_WQ_FF_FULL, *ev) ))
+			ev_out->tx_error.subtype = EF_EVENT_TX_ERROR_OFLOW;
+		else if(likely( QWORD_TEST_BIT(TX_EV_PKT_TOO_BIG, *ev) ))
+			ev_out->tx_error.subtype = EF_EVENT_TX_ERROR_2BIG;
+		else if(likely( QWORD_TEST_BIT(TX_EV_PKT_ERR, *ev) ))
+			ev_out->tx_error.subtype = EF_EVENT_TX_ERROR_BUS;
+	}
+}
+
+
+static void mark_bad(ef_event* ev)
+{
+	ev->generic.ev.u64[0] &=~ ((uint64_t) 1u << RX_EV_PKT_OK_LBN);
+}
+
+
+int ef_eventq_poll_evs(ef_vi* evq, ef_event* evs, int evs_len,
+		       ef_event_handler_fn *exception, void *expt_priv)
+{
+	int evs_len_orig = evs_len;
+
+	EF_VI_CHECK_EVENT_Q(evq);
+	ef_assert(evs);
+	ef_assert_gt(evs_len, 0);
+
+	if(unlikely( EF_VI_IS_EVENT(EF_VI_EVENT_PTR(evq, 1)) ))
+		goto overflow;
+
+	do {
+		{ /* Read the event out of the ring, then fiddle with
+		   * copied version.  Reason is that the ring is
+		   * likely to get pushed out of cache by another
+		   * event being delivered by hardware. */
+			ef_vi_event* ev = EF_VI_EVENT_PTR(evq, 0);
+			if( ! EF_VI_IS_EVENT(ev) )
+				break;
+			evs->generic.ev.u64[0] = cpu_to_le64 (ev->u64);
+			evq->evq_state->evq_ptr += sizeof(ef_vi_event);
+			ev->u64 = (uint64_t)(int64_t) -1;
+		}
+
+		/* Ugly: Exploit the fact that event code lies in top
+		 * bits of event. */
+		ef_assert_ge(EV_CODE_LBN, 32u);
+		switch( evs->generic.ev.u32[1] >> (EV_CODE_LBN - 32u) ) {
+		case RX_IP_EV_DECODE:
+			/* Look for duplicate desc_ptr: it signals
+			 * that a jumbo frame was truncated because we
+			 * ran out of descriptors. */
+			if(unlikely( falcon_rx_check_dup
+					   (evq, evs, &evs->generic.ev) )) {
+				--evs_len;
+				++evs;
+				break;
+			}
+			else {
+				/* Cope with FalconA1 bugs where RX
+				 * gives inconsistent RX events Mark
+				 * events as bad until SOP becomes
+				 * consistent again
+				 * ef_eventq_is_rx_sop_cont_bad() has
+				 * side effects - order is important
+				 */
+				if(unlikely
+				   (ef_eventq_is_rx_sop_cont_bad_efab
+				    (evq, &evs->generic.ev) )) {
+					mark_bad(evs);
+				}
+			}
+			falcon_rx_event(evs, &evs->generic.ev);
+			--evs_len;	
+			++evs;
+			break;
+
+		case TX_IP_EV_DECODE:
+			falcon_tx_event(evs, &evs->generic.ev);
+			--evs_len;
+			++evs;
+			break;
+
+		default:
+			break;
+		}
+	} while( evs_len );
+
+	return evs_len_orig - evs_len;
+
+
+ overflow:
+	evs->generic.type = EF_EVENT_TYPE_OFLOW;
+	evs->generic.ev.u64[0] = (uint64_t)((int64_t)-1);
+	return 1;
+}
+
+
+int/*bool*/ ef_eventq_poll_exception(void* priv, ef_vi* evq, ef_event* ev)
+{
+	int /*bool*/ handled = 0;
+  
+	switch( ev->generic.ev.u32[1] >> (EV_CODE_LBN - 32u) ) {
+	case DRIVER_EV_DECODE:
+		if( QWORD_GET_U(DRIVER_EV_SUB_CODE, ev->generic.ev) ==
+		    EVQ_INIT_DONE_EV_DECODE )
+			/* EVQ initialised event: ignore. */
+			handled = 1;
+		break;
+	}
+	return handled;
+}
+
+
+void ef_eventq_iterate(ef_vi* vi,
+		       void (*fn)(void* arg, ef_vi*, int rel_pos,
+				  int abs_pos, void* event),
+		       void* arg, int stop_at_end)
+{
+	int i, size_evs = (vi->evq_mask + 1) / sizeof(ef_vi_event);
+
+	for( i = 0; i < size_evs; ++i ) {
+		ef_vi_event* e = EF_VI_EVENT_PTR(vi, -i);
+		if( EF_VI_IS_EVENT(e) )
+			fn(arg, vi, i, 
+			   EF_VI_EVENT_OFFSET(vi, -i) / sizeof(ef_vi_event),
+			   e);
+		else if( stop_at_end )
+			break;
+	}
+}
+
+
+int ef_eventq_has_event(ef_vi* vi)
+{
+	return EF_VI_IS_EVENT(EF_VI_EVENT_PTR(vi, 0));
+}
+
+
+int ef_eventq_has_many_events(ef_vi* vi, int look_ahead)
+{
+	ef_assert_ge(look_ahead, 0);
+	return EF_VI_IS_EVENT(EF_VI_EVENT_PTR(vi, -look_ahead));
+}
+
+
+int ef_eventq_has_rx_event(ef_vi* vi)
+{
+	ef_vi_event* ev;
+	int i, n_evs = 0;
+
+	for( i = 0;  EF_VI_IS_EVENT(EF_VI_EVENT_PTR(vi, i)); --i ) {
+		ev = EF_VI_EVENT_PTR(vi, i);
+		if( EFVI_FALCON_EVENT_CODE(ev) == EF_EVENT_TYPE_RX )  n_evs++;
+	}
+	return n_evs;
+}
+
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/falcon_vi.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/falcon_vi.c
--- linux-2.6.18.8/drivers/xen/sfc_netfront/falcon_vi.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/falcon_vi.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,465 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*
+ * \author  djr, stg
+ *  \brief  Falcon-specific VI
+ *   \date  2006/11/30
+ */
+
+#include "ef_vi_internal.h"
+
+
+#define EFVI_FALCON_DMA_TX_FRAG		1
+
+
+/* TX descriptor for both physical and virtual packet transfers */
+typedef union {
+	uint32_t	dword[2];
+} ef_vi_falcon_dma_tx_buf_desc;
+typedef ef_vi_falcon_dma_tx_buf_desc ef_vi_falcon_dma_tx_phys_desc;
+
+
+/* RX descriptor for physical addressed transfers */
+typedef union {
+	uint32_t	dword[2];
+} ef_vi_falcon_dma_rx_phys_desc;
+
+
+/* RX descriptor for virtual packet transfers */
+typedef struct {
+	uint32_t	dword[1];
+} ef_vi_falcon_dma_rx_buf_desc;
+
+/* Buffer table index */
+typedef uint32_t		ef_vi_buffer_addr_t;
+
+ef_vi_inline int64_t dma_addr_to_u46(int64_t src_dma_addr)
+{
+	return (src_dma_addr & __FALCON_MASK(46, int64_t));
+}
+
+/*! Setup a physical address based descriptor with a specified length */
+ef_vi_inline void
+__falcon_dma_rx_calc_ip_phys(ef_vi_dma_addr_t dest_pa, 
+			     ef_vi_falcon_dma_rx_phys_desc *desc,
+			     int bytes)
+{
+	int region = 0; 		/* TODO fixme */
+	int64_t dest    = dma_addr_to_u46(dest_pa); /* lower 46 bits */
+
+	DWCHCK(__DW2(RX_KER_BUF_SIZE_LBN),  RX_KER_BUF_SIZE_WIDTH);
+	DWCHCK(__DW2(RX_KER_BUF_REGION_LBN),RX_KER_BUF_REGION_WIDTH);
+
+	LWCHK(RX_KER_BUF_ADR_LBN, RX_KER_BUF_ADR_WIDTH);
+
+	RANGECHCK(bytes,  RX_KER_BUF_SIZE_WIDTH);
+	RANGECHCK(region, RX_KER_BUF_REGION_WIDTH);
+
+	ef_assert(desc);
+
+	desc->dword[1] = ((bytes << __DW2(RX_KER_BUF_SIZE_LBN)) |
+			  (region << __DW2(RX_KER_BUF_REGION_LBN)) |
+			  (HIGH(dest,
+				RX_KER_BUF_ADR_LBN, 
+				RX_KER_BUF_ADR_WIDTH)));
+
+	desc->dword[0] = LOW(dest, 
+			     RX_KER_BUF_ADR_LBN, 
+			     RX_KER_BUF_ADR_WIDTH);
+}
+
+/*! Setup a virtual buffer descriptor for an IPMODE transfer */
+ef_vi_inline void
+__falcon_dma_tx_calc_ip_buf(unsigned buf_id, unsigned buf_ofs, unsigned bytes,
+			    int port, int frag, 
+			    ef_vi_falcon_dma_tx_buf_desc *desc)
+{
+	DWCHCK(__DW2(TX_USR_PORT_LBN), TX_USR_PORT_WIDTH);
+	DWCHCK(__DW2(TX_USR_CONT_LBN), TX_USR_CONT_WIDTH);
+	DWCHCK(__DW2(TX_USR_BYTE_CNT_LBN), TX_USR_BYTE_CNT_WIDTH);
+	LWCHK(RX_KER_BUF_ADR_LBN, RX_KER_BUF_ADR_WIDTH);
+	DWCHCK(TX_USR_BYTE_OFS_LBN, TX_USR_BYTE_OFS_WIDTH);
+
+	RANGECHCK(bytes,   TX_USR_BYTE_CNT_WIDTH);
+	RANGECHCK(port,    TX_USR_PORT_WIDTH);
+	RANGECHCK(frag,    TX_USR_CONT_WIDTH);
+	RANGECHCK(buf_id,  TX_USR_BUF_ID_WIDTH);
+	RANGECHCK(buf_ofs, TX_USR_BYTE_OFS_WIDTH);
+
+	ef_assert(desc);
+
+	desc->dword[1] = ((port   <<  __DW2(TX_USR_PORT_LBN))      | 
+			  (frag   <<  __DW2(TX_USR_CONT_LBN))      | 
+			  (bytes  <<  __DW2(TX_USR_BYTE_CNT_LBN))  |
+			  (HIGH(buf_id, 
+				TX_USR_BUF_ID_LBN,
+				TX_USR_BUF_ID_WIDTH)));
+
+	desc->dword[0] =  ((LOW(buf_id,
+				TX_USR_BUF_ID_LBN,
+				(TX_USR_BUF_ID_WIDTH))) |
+			   (buf_ofs << TX_USR_BYTE_OFS_LBN));
+}
+
+ef_vi_inline void
+falcon_dma_tx_calc_ip_buf_4k(unsigned buf_vaddr, unsigned bytes,
+			     int port, int frag, 
+			     ef_vi_falcon_dma_tx_buf_desc *desc)
+{
+	/* TODO FIXME [buf_vaddr] consists of the buffer index in the
+	** high bits, and an offset in the low bits. Assumptions
+	** permate the code that these can be rolled into one 32bit
+	** value, so this is currently preserved for Falcon. But we
+	** should change to support 8K pages
+	*/
+	unsigned buf_id =  EFVI_FALCON_BUFFER_4K_PAGE(buf_vaddr);
+	unsigned buf_ofs = EFVI_FALCON_BUFFER_4K_OFF(buf_vaddr);
+
+	__falcon_dma_tx_calc_ip_buf( buf_id, buf_ofs, bytes, port, frag, desc);
+}
+
+ef_vi_inline void
+falcon_dma_tx_calc_ip_buf(unsigned buf_vaddr, unsigned bytes, int port, 
+			  int frag, ef_vi_falcon_dma_tx_buf_desc *desc)
+{
+	falcon_dma_tx_calc_ip_buf_4k(buf_vaddr, bytes, port, frag, desc);
+}
+
+/*! Setup a virtual buffer based descriptor */
+ef_vi_inline void
+__falcon_dma_rx_calc_ip_buf(unsigned buf_id, unsigned buf_ofs, 
+			    ef_vi_falcon_dma_rx_buf_desc *desc)
+{ 
+	/* check alignment of buffer offset and pack */
+	ef_assert((buf_ofs & 0x1) == 0);
+
+	buf_ofs >>= 1;
+
+	DWCHCK(RX_USR_2BYTE_OFS_LBN, RX_USR_2BYTE_OFS_WIDTH);
+	DWCHCK(RX_USR_BUF_ID_LBN, RX_USR_BUF_ID_WIDTH);
+
+	RANGECHCK(buf_ofs, RX_USR_2BYTE_OFS_WIDTH);
+	RANGECHCK(buf_id,  RX_USR_BUF_ID_WIDTH);
+
+	ef_assert(desc);
+
+	desc->dword[0] = ((buf_ofs << RX_USR_2BYTE_OFS_LBN) | 
+			  (buf_id  << RX_USR_BUF_ID_LBN));
+}
+
+ef_vi_inline void
+falcon_dma_rx_calc_ip_buf_4k(unsigned buf_vaddr, 
+			     ef_vi_falcon_dma_rx_buf_desc *desc)
+{ 
+	/* TODO FIXME [buf_vaddr] consists of the buffer index in the
+	** high bits, and an offset in the low bits. Assumptions
+	** permeate the code that these can be rolled into one 32bit
+	** value, so this is currently preserved for Falcon. But we
+	** should change to support 8K pages
+	*/
+	unsigned buf_id =  EFVI_FALCON_BUFFER_4K_PAGE(buf_vaddr);
+	unsigned buf_ofs = EFVI_FALCON_BUFFER_4K_OFF(buf_vaddr);
+
+	__falcon_dma_rx_calc_ip_buf(buf_id, buf_ofs, desc);
+}
+
+ef_vi_inline void
+falcon_dma_rx_calc_ip_buf(unsigned buf_vaddr, 
+			  ef_vi_falcon_dma_rx_buf_desc *desc)
+{ 
+	falcon_dma_rx_calc_ip_buf_4k(buf_vaddr, desc);
+}
+
+
+ef_vi_inline ef_vi_dma_addr_t ef_physaddr(ef_addr efaddr)
+{
+	return (ef_vi_dma_addr_t) efaddr;
+}
+
+
+/*! Convert between an ef_addr and a buffer table index
+**  Assert that this was not a physical address
+*/
+ef_vi_inline ef_vi_buffer_addr_t ef_bufaddr(ef_addr efaddr)
+{
+	ef_assert(efaddr < ((uint64_t)1 << 32) );
+
+	return (ef_vi_buffer_addr_t) efaddr;
+}
+
+
+/*! Setup an physical address based descriptor for an IPMODE transfer */
+ef_vi_inline void
+falcon_dma_tx_calc_ip_phys(ef_vi_dma_addr_t src_dma_addr, unsigned bytes, 
+			   int port, int frag,
+			   ef_vi_falcon_dma_tx_phys_desc *desc)
+{
+
+	int region = 0; /* FIXME */
+	int64_t src    = dma_addr_to_u46(src_dma_addr); /* lower 46 bits */
+
+	DWCHCK(__DW2(TX_KER_PORT_LBN),      TX_KER_PORT_WIDTH);
+	DWCHCK(__DW2(TX_KER_CONT_LBN),      TX_KER_CONT_WIDTH);
+	DWCHCK(__DW2(TX_KER_BYTE_CNT_LBN),  TX_KER_BYTE_CNT_WIDTH);
+	DWCHCK(__DW2(TX_KER_BUF_REGION_LBN),TX_KER_BUF_REGION_WIDTH);
+
+	LWCHK(TX_KER_BUF_ADR_LBN, TX_KER_BUF_ADR_WIDTH);
+
+	RANGECHCK(port,   TX_KER_PORT_WIDTH);
+	RANGECHCK(frag,   TX_KER_CONT_WIDTH);
+	RANGECHCK(bytes,  TX_KER_BYTE_CNT_WIDTH);
+	RANGECHCK(region, TX_KER_BUF_REGION_WIDTH);
+
+	desc->dword[1] = ((port   <<  __DW2(TX_KER_PORT_LBN))      | 
+			  (frag   <<  __DW2(TX_KER_CONT_LBN))      | 
+			  (bytes  <<  __DW2(TX_KER_BYTE_CNT_LBN))  | 
+			  (region << __DW2(TX_KER_BUF_REGION_LBN)) |
+			  (HIGH(src,
+				TX_KER_BUF_ADR_LBN, 
+				TX_KER_BUF_ADR_WIDTH)));
+
+	ef_assert_equal(TX_KER_BUF_ADR_LBN, 0);
+	desc->dword[0] = (uint32_t) src_dma_addr;
+}
+
+
+void falcon_vi_init(ef_vi* vi, void* vvis)
+{
+	struct vi_mappings *vm = (struct vi_mappings*)vvis;
+	uint16_t* ids;
+
+	ef_assert(vi);
+	ef_assert(vvis);
+	ef_assert_equal(vm->signature, VI_MAPPING_SIGNATURE);
+	ef_assert_equal(vm->nic_type.arch, EF_VI_ARCH_FALCON);
+
+	/* Initialise masks to zero, so that ef_vi_state_init() will
+	** not do any harm when we don't have DMA queues. */
+	vi->vi_rxq.mask = vi->vi_txq.mask = 0;
+
+	/* Used for BUG5391_WORKAROUND. */
+	vi->vi_txq.misalign_mask = 0;
+
+	/* Initialise doorbell addresses to a distinctive small value
+	** which will cause a segfault, to trap doorbell pushes to VIs
+	** without DMA queues. */
+	vi->vi_rxq.doorbell = vi->vi_txq.doorbell = (ef_vi_ioaddr_t)0xdb;
+
+	ids = (uint16_t*) (vi->ep_state + 1);
+
+	if( vm->tx_queue_capacity ) {
+		vi->vi_txq.mask = vm->tx_queue_capacity - 1;
+		vi->vi_txq.doorbell = vm->tx_bell + 12;
+		vi->vi_txq.descriptors = vm->tx_dma_falcon;
+		vi->vi_txq.ids = ids;
+		ids += vi->vi_txq.mask + 1;
+		/* Check that the id fifo fits in the space allocated. */
+		ef_assert_le((char*) (vi->vi_txq.ids + vm->tx_queue_capacity),
+			     (char*) vi->ep_state
+			     + ef_vi_calc_state_bytes(vm->rx_queue_capacity,
+						      vm->tx_queue_capacity));
+	}
+	if( vm->rx_queue_capacity ) {
+		vi->vi_rxq.mask = vm->rx_queue_capacity - 1;
+		vi->vi_rxq.doorbell = vm->rx_bell + 12;
+		vi->vi_rxq.descriptors = vm->rx_dma_falcon;
+		vi->vi_rxq.ids = ids;
+		/* Check that the id fifo fits in the space allocated. */
+		ef_assert_le((char*) (vi->vi_rxq.ids + vm->rx_queue_capacity),
+			     (char*) vi->ep_state
+			     + ef_vi_calc_state_bytes(vm->rx_queue_capacity,
+						      vm->tx_queue_capacity));
+	}
+
+	if( vm->nic_type.variant == 'A' ) {
+		vi->vi_txq.misalign_mask = 15;    /* BUG5391_WORKAROUND */
+		vi->vi_flags |= EF_VI_BUG5692_WORKAROUND;
+	}
+}
+
+
+int ef_vi_transmitv_init(ef_vi* vi, const ef_iovec* iov, int iov_len,
+			 ef_request_id dma_id)
+{
+	ef_vi_txq* q = &vi->vi_txq;
+	ef_vi_txq_state* qs = &vi->ep_state->txq;
+	ef_vi_falcon_dma_tx_buf_desc* dp;
+	unsigned len, dma_len, di;
+	unsigned added_save = qs->added;
+	ef_addr dma_addr;
+	unsigned last_len = 0;
+
+	ef_assert(iov_len > 0);
+	ef_assert(iov);
+	ef_assert_equal((dma_id & EF_REQUEST_ID_MASK), dma_id);
+	ef_assert_nequal(dma_id, 0xffff);
+
+	dma_addr = iov->iov_base;
+	len = iov->iov_len;
+
+	if( vi->vi_flags & EF_VI_ISCSI_TX_DDIG ) {
+		/* Last 4 bytes of placeholder for digest must be
+		 * removed for h/w */
+		ef_assert(len > 4);
+		last_len = iov[iov_len - 1].iov_len;
+		if( last_len <= 4 ) {
+			ef_assert(iov_len > 1);
+			--iov_len;
+			last_len = iov[iov_len - 1].iov_len - (4 - last_len);
+		}
+		else {
+			last_len = iov[iov_len - 1].iov_len - 4;
+		}
+		if( iov_len == 1 )
+			len = last_len;
+	}
+
+	while( 1 ) {
+		if( qs->added - qs->removed >= q->mask ) {
+			qs->added = added_save;
+			return -EAGAIN;
+		}
+
+		dma_len = (~((unsigned) dma_addr) & 0xfff) + 1;
+		if( dma_len > len )  dma_len = len;
+		{ /* BUG5391_WORKAROUND */
+			unsigned misalign = 
+				(unsigned) dma_addr & q->misalign_mask;
+			if( misalign && dma_len + misalign > 512 )
+				dma_len = 512 - misalign;
+		}
+
+		di = qs->added++ & q->mask;
+		dp = (ef_vi_falcon_dma_tx_buf_desc*) q->descriptors + di;
+		if( vi->vi_flags & EF_VI_TX_PHYS_ADDR )
+			falcon_dma_tx_calc_ip_phys
+				(ef_physaddr(dma_addr), dma_len, /*port*/ 0,
+				 (iov_len == 1 && dma_len == len) ? 0 :
+				 EFVI_FALCON_DMA_TX_FRAG, dp);
+		else
+			falcon_dma_tx_calc_ip_buf
+				(ef_bufaddr(dma_addr), dma_len, /*port*/ 0,
+				 (iov_len == 1 && dma_len == len) ? 0 :
+				 EFVI_FALCON_DMA_TX_FRAG, dp);
+
+		dma_addr += dma_len;
+		len -= dma_len;
+
+		if( len == 0 ) {
+			if( --iov_len == 0 )  break;
+			++iov;
+			dma_addr = iov->iov_base;
+			len = iov->iov_len;
+			if( (vi->vi_flags & EF_VI_ISCSI_TX_DDIG) &&
+			    (iov_len == 1) )
+				len = last_len;
+		}
+	}
+
+	q->ids[di] = (uint16_t) dma_id;
+	return 0;
+}
+
+
+void ef_vi_transmit_push(ef_vi* vi)
+{
+	ef_vi_wiob();
+	writel((vi->ep_state->txq.added & vi->vi_txq.mask) <<
+		__DW4(TX_DESC_WPTR_LBN),
+        	vi->vi_txq.doorbell);
+}
+
+
+/*! The value of initial_rx_bytes is used to set RX_KER_BUF_SIZE in an initial
+**  receive descriptor here if physical addressing is being used. A value of
+**  zero represents 16384 bytes.  This is okay, because caller must provide a
+**  buffer than is > MTU, and mac should filter anything bigger than that.
+*/
+int ef_vi_receive_init(ef_vi* vi, ef_addr addr, ef_request_id dma_id,
+		       int initial_rx_bytes)
+{
+	ef_vi_rxq* q = &vi->vi_rxq;
+	ef_vi_rxq_state* qs = &vi->ep_state->rxq;
+	unsigned di;
+
+	if( ef_vi_receive_space(vi) ) {
+		di = qs->added++ & q->mask;
+		ef_assert_equal(q->ids[di], 0xffff);
+		q->ids[di] = (uint16_t) dma_id;
+
+		if( ! (vi->vi_flags & EF_VI_RX_PHYS_ADDR) ) {
+			ef_vi_falcon_dma_rx_buf_desc* dp;
+			dp = (ef_vi_falcon_dma_rx_buf_desc*) 
+				q->descriptors + di;
+			falcon_dma_rx_calc_ip_buf(ef_bufaddr(addr), dp);
+		}
+		else {
+			ef_vi_falcon_dma_rx_phys_desc* dp;
+			dp = (ef_vi_falcon_dma_rx_phys_desc*) 
+				q->descriptors + di;
+			__falcon_dma_rx_calc_ip_phys(addr, dp,
+						     initial_rx_bytes);
+		}
+
+		return 0;
+	}
+
+	return -EAGAIN;
+}
+
+
+void ef_vi_receive_push(ef_vi* vi)
+{
+	ef_vi_wiob();
+	writel ((vi->ep_state->rxq.added & vi->vi_rxq.mask) <<
+		__DW4(RX_DESC_WPTR_LBN),
+		vi->vi_rxq.doorbell);
+}
+
+
+ef_request_id ef_vi_receive_done(const ef_vi* vi, const ef_event* ef_ev)
+{
+	const ef_vi_qword* ev = EF_GET_HW_EV_PTR(*ef_ev);
+	unsigned di = ev->u32[0] & vi->vi_rxq.mask;
+	ef_request_id rq_id;
+
+	ef_assert(EF_EVENT_TYPE(*ef_ev) == EF_EVENT_TYPE_RX ||
+		  EF_EVENT_TYPE(*ef_ev) == EF_EVENT_TYPE_RX_DISCARD);
+
+	/* Detect spurious / duplicate RX events.  We may need to modify this
+	** code so that we are robust if they happen. */
+	ef_assert_equal(di, vi->ep_state->rxq.removed & vi->vi_rxq.mask);
+
+	/* We only support 1 port: so events should be in order. */
+	ef_assert(vi->vi_rxq.ids[di] != 0xffff);
+
+	rq_id = vi->vi_rxq.ids[di];
+	vi->vi_rxq.ids[di] = 0xffff;
+	++vi->ep_state->rxq.removed;
+	return rq_id;
+}
+
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/Makefile
--- linux-2.6.18.8/drivers/xen/sfc_netfront/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,11 @@
+EXTRA_CFLAGS += -Idrivers/xen/sfc_netfront -Idrivers/xen/sfc_netutil -Idrivers/xen/netfront
+EXTRA_CFLAGS += -D__ci_driver__
+EXTRA_CFLAGS += -Werror
+
+ifdef GCOV
+EXTRA_CFLAGS += -fprofile-arcs -ftest-coverage -DEFX_GCOV
+endif
+
+obj-$(CONFIG_XEN_NETDEV_ACCEL_SFC_FRONTEND)	:= sfc_netfront.o
+
+sfc_netfront-objs := accel_msg.o accel_bufs.o accel_netfront.o accel_vi.o accel_xenbus.o accel_tso.o accel_ssr.o accel_debugfs.o falcon_event.o falcon_vi.o pt_tx.o vi_init.o
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/pt_tx.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/pt_tx.c
--- linux-2.6.18.8/drivers/xen/sfc_netfront/pt_tx.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/pt_tx.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,91 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*
+ * \author  djr
+ *  \brief  Packet-mode transmit interface.
+ *   \date  2003/04/02
+ */
+
+/*! \cidoxg_lib_ef */
+#include "ef_vi_internal.h"
+
+
+int ef_vi_transmit_init(ef_vi* vi, ef_addr base, int len, ef_request_id dma_id)
+{
+	ef_iovec iov = { base, len };
+	return ef_vi_transmitv_init(vi, &iov, 1, dma_id);
+}
+
+
+int ef_vi_transmit(ef_vi* vi, ef_addr base, int len, ef_request_id dma_id)
+{
+	ef_iovec iov = { base, len };
+	int rc = ef_vi_transmitv_init(vi, &iov, 1, dma_id);
+	if( rc == 0 )  ef_vi_transmit_push(vi);
+	return rc;
+}
+
+
+int ef_vi_transmitv(ef_vi* vi, const ef_iovec* iov, int iov_len,
+                    ef_request_id dma_id)
+{
+	int rc = ef_vi_transmitv_init(vi, iov, iov_len, dma_id);
+	if( rc == 0 )  ef_vi_transmit_push(vi);
+	return rc;
+}
+
+
+int ef_vi_transmit_unbundle(ef_vi* vi, const ef_event* __ev,
+			    ef_request_id* ids)
+{
+	ef_request_id* ids_in = ids;
+	ef_vi_txq* q = &vi->vi_txq;
+	ef_vi_txq_state* qs = &vi->ep_state->txq;
+	const ef_vi_qword* ev = EF_GET_HW_EV_PTR(*__ev);
+	unsigned i, stop = (ev->u32[0] + 1) & q->mask;
+
+	ef_assert(EF_EVENT_TYPE(*__ev) == EF_EVENT_TYPE_TX ||
+		  EF_EVENT_TYPE(*__ev) == EF_EVENT_TYPE_TX_ERROR);
+
+	/* Shouldn't be batching more than 64 descriptors, and should not go
+	** backwards. */
+	ef_assert_le((((ev->u32[0] + 1) - qs->removed) & q->mask), 64);
+	/* Should not complete more than we've posted. */
+	ef_assert_le((((ev->u32[0] + 1) - qs->removed) & q->mask),
+		     qs->added - qs->removed);
+
+	for( i = qs->removed & q->mask; i != stop; i = ++qs->removed & q->mask )
+		if( q->ids[i] != 0xffff ) {
+			*ids++ = q->ids[i];
+			q->ids[i] = 0xffff;
+		}
+
+	ef_assert_le(ids - ids_in, EF_VI_TRANSMIT_BATCH);
+
+	return (int) (ids - ids_in);
+}
+
+/*! \cidoxg_end */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/sysdep.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/sysdep.h
--- linux-2.6.18.8/drivers/xen/sfc_netfront/sysdep.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/sysdep.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,184 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*
+ * \author  stg
+ *  \brief  System dependent support for ef vi lib
+ *   \date  2007/05/10
+ */
+
+/*! \cidoxg_include_ci_ul */
+#ifndef __CI_CIUL_SYSDEP_LINUX_H__
+#define __CI_CIUL_SYSDEP_LINUX_H__
+
+/**********************************************************************
+ * Kernel version compatability
+ */
+
+#if defined(__GNUC__)
+
+/* Linux kernel doesn't have stdint.h or [u]intptr_t. */
+# if !defined(LINUX_VERSION_CODE)
+#  include <linux/version.h>
+# endif
+# include <asm/io.h>
+
+/* In Linux 2.6.24, linux/types.h has uintptr_t */
+# if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,24)
+#  if BITS_PER_LONG == 32
+   typedef __u32         uintptr_t;
+#  else
+   typedef __u64         uintptr_t;
+#  endif
+# endif
+
+/* But even 2.6.24 doesn't define intptr_t */
+# if BITS_PER_LONG == 32
+   typedef __s32         intptr_t;
+# else
+   typedef __s64         intptr_t;
+# endif
+
+# if defined(__ia64__)
+#  define EF_VI_PRIx64  "lx"
+# else
+#  define EF_VI_PRIx64  "llx"
+# endif
+
+# define EF_VI_HF __attribute__((visibility("hidden")))
+# define EF_VI_HV __attribute__((visibility("hidden")))
+
+# if defined(__i386__) || defined(__x86_64__)  /* GCC x86/x64 */
+   typedef unsigned long long ef_vi_dma_addr_t; 
+#  if __GNUC__ >= 3 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 96)
+#   define ef_vi_wiob()  __asm__ __volatile__ ("sfence")
+#  else
+#   define ef_vi_wiob()  __asm__ __volatile__ (".byte 0x0F, 0xAE, 0xF8")
+#  endif
+
+# endif
+#endif
+
+#ifdef EFX_NOT_UPSTREAM
+
+/* Stuff for architectures/compilers not officially supported */
+
+#if !defined(__GNUC__)
+# if defined(__PPC__)  /* GCC, PPC */
+   typedef unsigned long     ef_vi_dma_addr_t;
+#  define ef_vi_wiob()  wmb()
+
+#  ifdef __powerpc64__
+#   ifdef CONFIG_SMP
+#    define CI_SMP_SYNC        "\n   eieio     \n"         /* memory cache sync */
+#    define CI_SMP_ISYNC       "\n   isync     \n"         /* instr cache sync */
+#   else
+#    define CI_SMP_SYNC
+#    define CI_SMP_ISYNC
+#   endif
+#  else	 /* for ppc32 systems */
+#   ifdef CONFIG_SMP
+#    define CI_SMP_SYNC        "\n   eieio     \n"
+#    define CI_SMP_ISYNC       "\n   sync      \n"
+#   else
+#    define CI_SMP_SYNC
+#    define CI_SMP_ISYNC
+#   endif
+#  endif
+
+# elif defined(__ia64__)  /* GCC, IA64 */
+   typedef unsigned long     ef_vi_dma_addr_t;
+#  define ef_vi_wiob()  __asm__ __volatile__("mf.a": : :"memory")
+
+# else
+#  error Unknown processor - GNU C
+# endif
+
+#elif defined(__PGI)
+# error PGI not supported 
+
+#elif defined(__INTEL_COMPILER)
+
+/* Intel compilers v7 claim to be very gcc compatible. */
+# if __INTEL_COMPILER >= 700
+#  if __GNUC__ >= 3 || (__GNUC__ == 2 && __GNUC_MINOR__ > 91)
+#   define EF_VI_LIKELY(t)    __builtin_expect((t), 1)
+#   define EF_VI_UNLIKELY(t)  __builtin_expect((t), 0)
+#  endif
+
+#  if __GNUC__ >= 3 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 96)
+#   define ef_vi_wiob()  __asm__ __volatile__ ("sfence")
+#  else
+#   define ef_vi_wiob()  __asm__ __volatile__ (".byte 0x0F, 0xAE, 0xF8")
+#  endif
+
+# else
+#  error Old Intel compiler not supported.
+# endif
+
+#else
+# error Unknown compiler.
+#endif
+
+#endif
+
+
+# include <linux/errno.h>
+
+
+/**********************************************************************
+ * Extracting bit fields.
+ */
+
+#define _QWORD_GET_LOW(f, v)                                    \
+  (((v).u32[0] >> (f##_LBN)) & ((1u << f##_WIDTH) - 1u))
+#define _QWORD_GET_HIGH(f, v)                                           \
+  (((v).u32[1] >> (f##_LBN - 32u)) & ((1u << f##_WIDTH) - 1u))
+#define _QWORD_GET_ANY(f, v)                                            \
+  (((v).u64[0] >> f##_LBN) & (((uint64_t) 1u << f##_WIDTH) - 1u))
+
+#define QWORD_GET(f, v)                                                     \
+  ((f##_LBN + f##_WIDTH) <= 32u                                             \
+   ? _QWORD_GET_LOW(f, (v))                                                 \
+   : ((f##_LBN >= 32u) ? _QWORD_GET_HIGH(f, (v)) : _QWORD_GET_ANY(f, (v))))
+
+#define QWORD_GET_U(f, v)  ((unsigned) QWORD_GET(f, (v)))
+
+#define _QWORD_TEST_BIT_LOW(f, v)   ((v).u32[0] & (1u << (f##_LBN)))
+#define _QWORD_TEST_BIT_HIGH(f, v)  ((v).u32[1] & (1u << (f##_LBN - 32u)))
+
+#define QWORD_TEST_BIT(f, v)                                                  \
+  (f##_LBN < 32 ? _QWORD_TEST_BIT_LOW(f, (v)) : _QWORD_TEST_BIT_HIGH(f, (v)))
+
+
+
+
+#ifndef DECLSPEC_NORETURN
+/* normally defined on Windows to expand to a declaration that the
+   function will not return */
+# define DECLSPEC_NORETURN
+#endif
+
+#endif  /* __CI_CIUL_SYSDEP_LINUX_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netfront/vi_init.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/vi_init.c
--- linux-2.6.18.8/drivers/xen/sfc_netfront/vi_init.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netfront/vi_init.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,183 @@
+/****************************************************************************
+ * Copyright 2002-2005: Level 5 Networks Inc.
+ * Copyright 2005-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications
+ *  <linux-xen-drivers@solarflare.com>
+ *  <onload-dev@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*
+ * \author  djr
+ *  \brief  Initialisation of VIs.
+ *   \date  2007/06/08
+ */
+
+#include "ef_vi_internal.h"
+
+#define EF_VI_STATE_BYTES(rxq_sz, txq_sz)			\
+	(sizeof(ef_vi_state) + (rxq_sz) * sizeof(uint16_t)	\
+	 + (txq_sz) * sizeof(uint16_t))
+
+int ef_vi_calc_state_bytes(int rxq_sz, int txq_sz)
+{
+	ef_assert(rxq_sz == 0 || EF_VI_IS_POW2(rxq_sz));
+	ef_assert(txq_sz == 0 || EF_VI_IS_POW2(txq_sz));
+
+	return EF_VI_STATE_BYTES(rxq_sz, txq_sz);
+}
+
+
+int ef_vi_state_bytes(ef_vi* vi)
+{
+	int rxq_sz = 0, txq_sz = 0;
+	if( ef_vi_receive_capacity(vi) )
+		rxq_sz = ef_vi_receive_capacity(vi) + 1;
+	if( ef_vi_transmit_capacity(vi) )
+		txq_sz = ef_vi_transmit_capacity(vi) + 1;
+
+	ef_assert(rxq_sz == 0 || EF_VI_IS_POW2(rxq_sz));
+	ef_assert(txq_sz == 0 || EF_VI_IS_POW2(txq_sz));
+
+	return EF_VI_STATE_BYTES(rxq_sz, txq_sz);
+}
+
+
+void ef_eventq_state_init(ef_vi* evq)
+{
+	int j;
+
+	for (j = 0; j<EFAB_DMAQS_PER_EVQ_MAX; j++) {
+		ef_rx_dup_state_t *rx_dup_state =
+			&evq->evq_state->rx_dup_state[j];
+		rx_dup_state->bad_sop = 0;
+		rx_dup_state->rx_last_desc_ptr = -1;
+		rx_dup_state->frag_num = 0;
+	}
+
+	evq->evq_state->evq_ptr = 0;
+}
+
+
+void ef_vi_state_init(ef_vi* vi)
+{
+	ef_vi_state* state = vi->ep_state;
+	unsigned i;
+
+	state->txq.added = state->txq.removed = 0;
+	state->rxq.added = state->rxq.removed = 0;
+
+	if( vi->vi_rxq.mask )
+		for( i = 0; i <= vi->vi_rxq.mask; ++i )
+			vi->vi_rxq.ids[i] = (uint16_t) -1;
+	if( vi->vi_txq.mask )
+		for( i = 0; i <= vi->vi_txq.mask; ++i )
+			vi->vi_txq.ids[i] = (uint16_t) -1;
+}
+
+
+void ef_vi_init_mapping_evq(void* data_area, struct ef_vi_nic_type nic_type,
+                            int instance, unsigned evq_bytes, void* base,
+                            void* timer_reg)
+{
+	struct vi_mappings* vm = (struct vi_mappings*) data_area;
+
+	vm->signature = VI_MAPPING_SIGNATURE;
+	vm->vi_instance = instance;
+	vm->nic_type = nic_type;
+	vm->evq_bytes = evq_bytes;
+	vm->evq_base = base;
+	vm->evq_timer_reg = timer_reg;
+}
+
+
+void ef_vi_init(ef_vi* vi, void* vvis, ef_vi_state* state,
+                ef_eventq_state* evq_state, enum ef_vi_flags vi_flags)
+{
+	struct vi_mappings* vm = (struct vi_mappings*) vvis;
+
+	vi->vi_i = vm->vi_instance;
+	vi->ep_state = state;
+	vi->vi_flags = vi_flags;
+
+	switch( vm->nic_type.arch ) {
+	case EF_VI_ARCH_FALCON:
+		falcon_vi_init(vi, vvis);
+		break;
+	default:
+		/* ?? TODO: We should return an error code. */
+		ef_assert(0);
+		break;
+	}
+
+	if( vm->evq_bytes ) {
+		vi->evq_state = evq_state;
+		vi->evq_mask = vm->evq_bytes - 1u;
+		vi->evq_base = vm->evq_base;
+		vi->evq_timer_reg = vm->evq_timer_reg;
+	}
+
+	EF_VI_MAGIC_SET(vi, EF_VI);
+}
+
+
+/* Initialise [data_area] with information required to initialise an ef_vi.
+ * In the following, an unused param should be set to NULL. Note the case
+ * marked (*) of [iobuf_mmap] for falcon/driver; for the normal driver this
+ * must be NULL.
+ *
+ * \param  data_area     [in,out] required, must ref at least VI_MAPPING_SIZE 
+ *                                bytes
+ * \param  io_mmap       [in] ef1,    required
+ *                            falcon, required
+ * \param  iobuf_mmap    [in] ef1,    unused
+ *                            falcon, required
+ */
+void ef_vi_init_mapping_vi(void* data_area, struct ef_vi_nic_type nic_type,
+                           unsigned rxq_capacity, unsigned txq_capacity,
+                           int instance, void* io_mmap,
+                           void* iobuf_mmap_rx, void* iobuf_mmap_tx,
+                           enum ef_vi_flags vi_flags)
+{
+	struct vi_mappings* vm = (struct vi_mappings*) data_area;
+	int rx_desc_bytes, rxq_bytes;
+
+	ef_assert(rxq_capacity > 0 || txq_capacity > 0);
+	ef_assert(vm);
+	ef_assert(io_mmap);
+	ef_assert(iobuf_mmap_rx || iobuf_mmap_tx);
+
+	vm->signature = VI_MAPPING_SIGNATURE;
+	vm->vi_instance = instance;
+	vm->nic_type = nic_type;
+
+	rx_desc_bytes = (vi_flags & EF_VI_RX_PHYS_ADDR) ? 8 : 4;
+	rxq_bytes = rxq_capacity * rx_desc_bytes;
+	rxq_bytes = (rxq_bytes + PAGE_SIZE - 1) & ~(PAGE_SIZE - 1);
+
+	if( iobuf_mmap_rx == iobuf_mmap_tx )
+		iobuf_mmap_tx = (char*) iobuf_mmap_rx + rxq_bytes;
+
+	vm->rx_queue_capacity = rxq_capacity;
+	vm->rx_dma_falcon = iobuf_mmap_rx;
+	vm->rx_bell       = (char*) io_mmap + (RX_DESC_UPD_REG_KER_OFST & 4095);
+	vm->tx_queue_capacity = txq_capacity;
+	vm->tx_dma_falcon = iobuf_mmap_tx;
+	vm->tx_bell       = (char*) io_mmap + (TX_DESC_UPD_REG_KER_OFST & 4095);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netutil/accel_cuckoo_hash.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netutil/accel_cuckoo_hash.c
--- linux-2.6.18.8/drivers/xen/sfc_netutil/accel_cuckoo_hash.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netutil/accel_cuckoo_hash.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,651 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/types.h> /* needed for linux/random.h */
+#include <linux/random.h>
+
+#include "accel_cuckoo_hash.h"
+#include "accel_util.h"
+
+static inline int cuckoo_hash_key_compare(cuckoo_hash_table *hashtab,
+					  cuckoo_hash_key *key1, 
+					  cuckoo_hash_key *key2)
+{
+	return !memcmp(key1, key2, hashtab->key_length);
+}
+
+
+static inline void cuckoo_hash_key_set(cuckoo_hash_key *key1, 
+				       cuckoo_hash_key *key2)
+{
+	*key1 = *key2;
+}
+
+
+/*
+ * Sets hash function parameters.  Chooses "a" to be odd, 0 < a < 2^w
+ * where w is the length of the key
+ */
+static void set_hash_parameters(cuckoo_hash_table *hashtab)
+{
+ again:
+	hashtab->a0 = hashtab->a1 = 0;
+
+	/* Make sure random */
+	get_random_bytes(&hashtab->a0, hashtab->key_length);
+	get_random_bytes(&hashtab->a1, hashtab->key_length);
+
+	/* Make sure odd */
+	hashtab->a0 |= 1;
+	hashtab->a1 |= 1;
+
+	/* Being different is good */
+	if (hashtab->a0 != hashtab->a1)
+		return;
+		       
+	goto again;
+}
+
+int cuckoo_hash_init(cuckoo_hash_table *hashtab, unsigned length_bits,
+		     unsigned key_length)
+{
+	char *table_mem;
+	unsigned length = 1 << length_bits;
+
+	BUG_ON(length_bits >= sizeof(unsigned) * 8);
+	BUG_ON(key_length > sizeof(cuckoo_hash_key));
+
+	table_mem = kmalloc(sizeof(cuckoo_hash_entry) * 2 * length, GFP_KERNEL);
+
+	if (table_mem == NULL)
+		return -ENOMEM;
+
+	hashtab->length = length;
+	hashtab->length_bits = length_bits;
+	hashtab->key_length = key_length;
+	hashtab->entries = 0;
+
+	hashtab->table0 = (cuckoo_hash_entry *)table_mem;
+	hashtab->table1 = (cuckoo_hash_entry *)
+		(table_mem + length * sizeof(cuckoo_hash_entry));
+
+	set_hash_parameters(hashtab);
+
+	/* Zero the table */
+	memset(hashtab->table0, 0, length * 2 * sizeof(cuckoo_hash_entry));
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(cuckoo_hash_init);
+
+void cuckoo_hash_destroy(cuckoo_hash_table *hashtab)
+{
+	if (hashtab->table0 != NULL)
+		kfree(hashtab->table0);
+}
+
+EXPORT_SYMBOL_GPL(cuckoo_hash_destroy);
+
+/* 
+ * This computes sizeof(cuckoo_hash) bits of hash, not all will be
+ * necessarily used, but the hash function throws away any that
+ * aren't
+ */ 
+static inline void cuckoo_compute_hash_helper(cuckoo_hash_table *hashtab,
+					      cuckoo_hash_key *a,
+					      cuckoo_hash_key *x,
+					      cuckoo_hash *result) 
+{
+	u64 multiply_result = 0, a_temp, x_temp;
+	u32 carry = 0;
+	u32 *a_words;
+	u32 *x_words;
+	int i;
+
+	/*
+	 * As the mod and div operations in the function effectively
+	 * reduce and shift the bits of the product down to just the
+	 * third word, we need only compute that and return it as a
+	 * result.
+	 *
+	 * Do enough long multiplication to get the word we need
+	 */
+
+	/* This assumes things about the sizes of the key and hash */
+	BUG_ON(hashtab->key_length % sizeof(u32) != 0);
+	BUG_ON(sizeof(cuckoo_hash) != sizeof(u32));
+
+	a_words = (u32 *)a;
+	x_words = (u32 *)x;
+
+	for (i = 0; i < hashtab->key_length / sizeof(u32); i++) {
+		a_temp = a_words[i];
+		x_temp = x_words[i];
+		
+		multiply_result = (a_temp * x_temp) + carry;
+		carry = (multiply_result >> 32) & 0xffffffff;
+	}
+	
+	*result = multiply_result & 0xffffffff;
+}
+
+
+/*
+ * Want to implement (ax mod 2^w) div 2^(w-q) for odd a, 0 < a < 2^w;
+ * w is the length of the key, q is the length of the hash, I think.
+ * See http://www.it-c.dk/people/pagh/papers/cuckoo-jour.pdf 
+ */
+static cuckoo_hash cuckoo_compute_hash(cuckoo_hash_table *hashtab, 
+				       cuckoo_hash_key *key, 
+				       cuckoo_hash_key *a)
+{
+	unsigned q = hashtab->length_bits;
+	unsigned shift = 32 - q;
+	unsigned mask = ((1 << q) - 1) << shift;
+	cuckoo_hash hash;
+
+	cuckoo_compute_hash_helper(hashtab, a, key, &hash);
+
+	/* 
+	 * Take the top few bits to get the right length for this
+	 * hash table 
+	 */
+	hash = (hash & mask) >> shift;
+
+	BUG_ON(hash >= hashtab->length);
+
+	return hash;
+}
+
+
+static int cuckoo_hash_lookup0(cuckoo_hash_table *hashtab,
+			       cuckoo_hash_key *key,
+			       cuckoo_hash_value *value)
+{
+	cuckoo_hash hash = cuckoo_compute_hash(hashtab, key, &hashtab->a0);
+
+	if ((hashtab->table0[hash].state == CUCKOO_HASH_STATE_OCCUPIED)
+	    && cuckoo_hash_key_compare(hashtab, &(hashtab->table0[hash].key),
+				       key)) {
+		*value = hashtab->table0[hash].value;
+		return 1;
+	}
+
+	return 0;
+}
+
+static int cuckoo_hash_lookup1(cuckoo_hash_table *hashtab,
+			       cuckoo_hash_key *key,
+			       cuckoo_hash_value *value)
+{
+	cuckoo_hash hash = cuckoo_compute_hash(hashtab, key, &hashtab->a1);
+
+	if ((hashtab->table1[hash].state == CUCKOO_HASH_STATE_OCCUPIED)
+	    && cuckoo_hash_key_compare(hashtab, &(hashtab->table1[hash].key),
+				       key)) {
+		*value = hashtab->table1[hash].value;
+		return 1;
+	}
+
+	return 0;
+}
+
+
+int cuckoo_hash_lookup(cuckoo_hash_table *hashtab, cuckoo_hash_key *key,
+		       cuckoo_hash_value *value)
+{
+	return cuckoo_hash_lookup0(hashtab, key, value)
+		|| cuckoo_hash_lookup1(hashtab, key, value);
+}
+EXPORT_SYMBOL_GPL(cuckoo_hash_lookup);
+
+
+/* Transfer any active entries from "old_table" into hashtab */
+static int cuckoo_hash_transfer_entries(cuckoo_hash_table *hashtab,
+					cuckoo_hash_entry *old_table,
+					unsigned capacity)
+{
+	int i, rc;
+	cuckoo_hash_entry *entry;
+
+	hashtab->entries = 0;
+
+	for (i = 0; i < capacity; i++) {
+		entry = &old_table[i];
+		if (entry->state == CUCKOO_HASH_STATE_OCCUPIED) {
+			rc = cuckoo_hash_add(hashtab, &(entry->key), 
+					     entry->value, 0);
+			if (rc != 0) {
+				return rc;
+			}
+		}
+	}
+  
+	return 0;
+}
+
+
+int cuckoo_hash_rehash(cuckoo_hash_table *hashtab)
+{
+	cuckoo_hash_entry *new_table;
+	cuckoo_hash_table old_hashtab;
+	int resize = 0, rc, rehash_count;
+
+	/*
+	 * Store old tables so we can access the existing values and
+	 * copy across
+	 */
+	memcpy(&old_hashtab, hashtab, sizeof(cuckoo_hash_table));
+
+	/* resize if hashtable is more than half full */
+	if (old_hashtab.entries > old_hashtab.length &&
+	    old_hashtab.length_bits < 32)
+		resize = 1;
+
+ resize:
+	if (resize) {
+		new_table = kmalloc(sizeof(cuckoo_hash_entry) * 4 * hashtab->length,
+				    GFP_ATOMIC);
+		if (new_table == NULL) {
+			rc = -ENOMEM;
+			goto err;
+		}
+
+		hashtab->length = 2 * hashtab->length;
+		hashtab->length_bits++;
+	} else {
+		new_table = kmalloc(sizeof(cuckoo_hash_entry) * 2 * hashtab->length,
+				    GFP_ATOMIC);
+		if (new_table == NULL) {
+			rc = -ENOMEM;
+			goto err;
+		}
+	}
+    
+	/*
+	 * Point hashtab to new memory region so we can try to
+	 * construct new table
+	 */
+	hashtab->table0 = new_table;
+	hashtab->table1 = (cuckoo_hash_entry *)
+		((char *)new_table + hashtab->length * sizeof(cuckoo_hash_entry));
+  
+	rehash_count = 0;
+
+ again:
+	/* Zero the new tables */
+	memset(new_table, 0, hashtab->length * 2 * sizeof(cuckoo_hash_entry));
+
+	/* Choose new parameters for the hash functions */
+	set_hash_parameters(hashtab);
+
+	/*
+	 * Multiply old_table_length by 2 as the length refers to each
+	 * table, and there are two of them.  This assumes that they
+	 * are arranged sequentially in memory, so assert it 
+	 */
+	BUG_ON(((char *)old_hashtab.table1) != 
+	       ((char *)old_hashtab.table0 + old_hashtab.length
+		* sizeof(cuckoo_hash_entry)));
+	rc = cuckoo_hash_transfer_entries(hashtab, old_hashtab.table0, 
+					  old_hashtab.length * 2);
+	if (rc < 0) {
+		/* Problem */
+		if (rc == -ENOSPC) {
+			++rehash_count;
+			if (rehash_count < CUCKOO_HASH_MAX_LOOP) {
+				/*
+				 * Wanted to rehash, but rather than
+				 * recurse we can just do it here
+				 */
+				goto again;
+			} else {
+				/*
+				 * Didn't manage to rehash, so let's
+				 * go up a size (if we haven't already
+				 * and there's space)
+				 */
+				if (!resize && hashtab->length_bits < 32) {
+					resize = 1;
+					kfree(new_table);
+					goto resize;
+				}
+				else
+					goto err;
+			}
+		}
+		else
+			goto err;
+	}
+
+	/* Success, I think.  Free up the old table */
+	kfree(old_hashtab.table0);
+  
+	/* We should have put all the entries from old table in the new one */
+	BUG_ON(hashtab->entries != old_hashtab.entries);
+
+	return 0;
+ err:
+	EPRINTK("%s: Rehash failed, giving up\n", __FUNCTION__);
+	/* Some other error, give up, at least restore table to how it was */
+	memcpy(hashtab, &old_hashtab, sizeof(cuckoo_hash_table));
+	if (new_table)
+		kfree(new_table);
+	return rc;
+}
+EXPORT_SYMBOL_GPL(cuckoo_hash_rehash);
+
+
+static int 
+cuckoo_hash_insert_or_displace(cuckoo_hash_entry *table, unsigned hash,
+			       cuckoo_hash_key *key, 
+			       cuckoo_hash_value value,
+			       cuckoo_hash_key *displaced_key, 
+			       cuckoo_hash_value *displaced_value)
+{
+	if (table[hash].state == CUCKOO_HASH_STATE_VACANT) {
+		cuckoo_hash_key_set(&(table[hash].key), key);
+		table[hash].value = value;
+		table[hash].state = CUCKOO_HASH_STATE_OCCUPIED;
+
+		return 1;
+	} else {
+		cuckoo_hash_key_set(displaced_key, &(table[hash].key));
+		*displaced_value = table[hash].value;
+		cuckoo_hash_key_set(&(table[hash].key), key);
+		table[hash].value = value;
+
+		return 0;
+	}
+}
+
+
+int cuckoo_hash_add(cuckoo_hash_table *hashtab, cuckoo_hash_key *key,
+		     cuckoo_hash_value value, int can_rehash)
+{
+	cuckoo_hash hash0, hash1;
+	int i, rc;
+	cuckoo_hash_key key1, key2;
+
+	cuckoo_hash_key_set(&key1, key);
+
+ again:
+	i = 0;
+	do {
+		hash0 = cuckoo_compute_hash(hashtab, &key1, &hashtab->a0);
+		if (cuckoo_hash_insert_or_displace(hashtab->table0, hash0, 
+						   &key1, value, &key2,
+						   &value)) {
+			/* Success */
+			hashtab->entries++;
+			return 0;
+		}
+	
+		hash1 = cuckoo_compute_hash(hashtab, &key2, &hashtab->a1);
+		if (cuckoo_hash_insert_or_displace(hashtab->table1, hash1,
+						   &key2, value, &key1,
+						   &value)) {
+			/* Success */
+			hashtab->entries++;
+			return 0;
+		}
+	} while (++i < CUCKOO_HASH_MAX_LOOP);
+
+	if (can_rehash) {
+		if ((rc = cuckoo_hash_rehash(hashtab)) < 0) {
+			/*
+			 * Give up - this will drop whichever
+			 * key/value pair we have currently displaced
+			 * on the floor
+			 */
+			return rc;
+		}
+		goto again;
+	}
+  
+	EPRINTK("%s: failed hash add\n", __FUNCTION__);
+	/*
+	 * Couldn't do it - bad as we've now removed some random thing
+	 * from the table, and will just drop it on the floor.  Better
+	 * would be to somehow revert the table to the state it was in
+	 * at the start
+	 */
+	return -ENOSPC;
+}
+EXPORT_SYMBOL_GPL(cuckoo_hash_add);
+
+
+int cuckoo_hash_add_check(cuckoo_hash_table *hashtab,
+			  cuckoo_hash_key *key, cuckoo_hash_value value,
+			  int can_rehash)
+{
+	int stored_value;
+
+	if (cuckoo_hash_lookup(hashtab, key, &stored_value))
+		return -EBUSY;
+
+	return cuckoo_hash_add(hashtab, key, value, can_rehash);
+}
+EXPORT_SYMBOL_GPL(cuckoo_hash_add_check);
+
+
+int cuckoo_hash_remove(cuckoo_hash_table *hashtab, cuckoo_hash_key *key)
+{
+	cuckoo_hash hash;
+
+	hash = cuckoo_compute_hash(hashtab, key, &hashtab->a0);
+	if ((hashtab->table0[hash].state == CUCKOO_HASH_STATE_OCCUPIED) &&
+	    cuckoo_hash_key_compare(hashtab, &(hashtab->table0[hash].key),
+				    key)) {
+		hashtab->table0[hash].state = CUCKOO_HASH_STATE_VACANT;
+		hashtab->entries--;
+		return 0;
+	}
+  
+	hash = cuckoo_compute_hash(hashtab, key, &hashtab->a1);
+	if ((hashtab->table1[hash].state == CUCKOO_HASH_STATE_OCCUPIED) &&
+	    cuckoo_hash_key_compare(hashtab, &(hashtab->table1[hash].key),
+				    key)) {
+		hashtab->table1[hash].state = CUCKOO_HASH_STATE_VACANT;
+		hashtab->entries--;
+		return 0;
+	}
+ 
+	return -EINVAL;
+}
+EXPORT_SYMBOL_GPL(cuckoo_hash_remove);
+
+
+int cuckoo_hash_update(cuckoo_hash_table *hashtab, cuckoo_hash_key *key,
+		       cuckoo_hash_value value)
+{
+	cuckoo_hash hash;
+
+	hash = cuckoo_compute_hash(hashtab, key, &hashtab->a0);
+	if ((hashtab->table0[hash].state == CUCKOO_HASH_STATE_OCCUPIED) &&
+	    cuckoo_hash_key_compare(hashtab, &(hashtab->table0[hash].key),
+				    key)) {
+		hashtab->table0[hash].value = value;
+		return 0;
+	}
+
+	hash = cuckoo_compute_hash(hashtab, key, &hashtab->a1);
+	if ((hashtab->table1[hash].state == CUCKOO_HASH_STATE_OCCUPIED) &&
+	    cuckoo_hash_key_compare(hashtab, &(hashtab->table1[hash].key),
+				    key)) {
+		hashtab->table1[hash].value = value;
+		return 0;
+	}
+
+	return -EINVAL;
+}
+EXPORT_SYMBOL_GPL(cuckoo_hash_update);
+
+
+void cuckoo_hash_iterate_reset(cuckoo_hash_table *hashtab)
+{
+	hashtab->iterate_index = 0;
+}
+EXPORT_SYMBOL_GPL(cuckoo_hash_iterate_reset);
+
+
+int cuckoo_hash_iterate(cuckoo_hash_table *hashtab,
+			cuckoo_hash_key *key, cuckoo_hash_value *value)
+{
+	unsigned index;
+
+	while (hashtab->iterate_index < hashtab->length) {
+		index = hashtab->iterate_index;
+		++hashtab->iterate_index;
+		if (hashtab->table0[index].state == CUCKOO_HASH_STATE_OCCUPIED) {
+			*key = hashtab->table0[index].key;
+			*value = hashtab->table0[index].value;
+			return 0;
+		}
+	}
+
+	while (hashtab->iterate_index >= hashtab->length &&
+	       hashtab->iterate_index < hashtab->length * 2) {
+		index = hashtab->iterate_index - hashtab->length;
+		++hashtab->iterate_index;		
+		if (hashtab->table1[index].state == CUCKOO_HASH_STATE_OCCUPIED) {
+			*key = hashtab->table1[index].key;
+			*value = hashtab->table1[index].value;
+			return 0;
+		}
+	}
+
+	return -ENOSPC;
+}
+EXPORT_SYMBOL_GPL(cuckoo_hash_iterate);
+
+
+#if 0
+void cuckoo_hash_valid(cuckoo_hash_table *hashtab)
+{
+	int i, entry_count = 0;
+
+	for (i=0; i < hashtab->length; i++) {
+		EPRINTK_ON(hashtab->table0[i].state != CUCKOO_HASH_STATE_VACANT &&
+			   hashtab->table0[i].state != CUCKOO_HASH_STATE_OCCUPIED);
+		if (hashtab->table0[i].state == CUCKOO_HASH_STATE_OCCUPIED)
+			entry_count++;
+		EPRINTK_ON(hashtab->table1[i].state != CUCKOO_HASH_STATE_VACANT &&
+			   hashtab->table1[i].state != CUCKOO_HASH_STATE_OCCUPIED);
+		if (hashtab->table1[i].state == CUCKOO_HASH_STATE_OCCUPIED)
+			entry_count++;	
+	}
+	
+	if (entry_count != hashtab->entries) {
+		EPRINTK("%s: bad count\n", __FUNCTION__);
+		cuckoo_hash_dump(hashtab);
+		return;
+	}
+
+	for (i=0; i< hashtab->length; i++) {
+		if (hashtab->table0[i].state == CUCKOO_HASH_STATE_OCCUPIED)
+			if (i != cuckoo_compute_hash(hashtab, 
+						     &hashtab->table0[i].key, 
+						     &hashtab->a0)) {
+				EPRINTK("%s: Bad key table 0 index %d\n",
+					__FUNCTION__, i);
+				cuckoo_hash_dump(hashtab);
+				return;
+			}
+		if (hashtab->table1[i].state == CUCKOO_HASH_STATE_OCCUPIED)
+			if (i != cuckoo_compute_hash(hashtab, 
+						     &hashtab->table1[i].key, 
+						     &hashtab->a1)) {
+				EPRINTK("%s: Bad key table 1 index %d\n",
+					__FUNCTION__, i);
+				cuckoo_hash_dump(hashtab);
+				return;
+			}
+	}
+
+}
+EXPORT_SYMBOL_GPL(cuckoo_hash_valid);
+
+
+void cuckoo_hash_dump(cuckoo_hash_table *hashtab)
+{
+	int i, entry_count;
+
+	entry_count = 0;
+	for (i=0; i < hashtab->length; i++) {
+		EPRINTK_ON(hashtab->table0[i].state != CUCKOO_HASH_STATE_VACANT &&
+			   hashtab->table0[i].state != CUCKOO_HASH_STATE_OCCUPIED);
+		if (hashtab->table0[i].state == CUCKOO_HASH_STATE_OCCUPIED)
+			entry_count++;
+		EPRINTK_ON(hashtab->table1[i].state != CUCKOO_HASH_STATE_VACANT &&
+			   hashtab->table1[i].state != CUCKOO_HASH_STATE_OCCUPIED);
+		if (hashtab->table1[i].state == CUCKOO_HASH_STATE_OCCUPIED)
+			entry_count++;	
+	}
+
+	EPRINTK("======================\n");
+	EPRINTK("Cuckoo hash table dump\n");
+	EPRINTK("======================\n");
+	EPRINTK("length: %d; length_bits: %d; key_length: %d\n", hashtab->length,
+		hashtab->length_bits, hashtab->key_length);
+	EPRINTK("Recorded entries: %d\n", hashtab->entries);
+	EPRINTK("Counted entries: %d\n", entry_count);
+	EPRINTK("a0: %llx; a1: %llx\n", hashtab->a0, hashtab->a1);
+	EPRINTK("-----------------------------------------\n");
+	EPRINTK("Index  Occupied  Key  Value Index0 Index1\n");
+	EPRINTK("-----------------------------------------\n");		
+	for (i=0; i< hashtab->length; i++) {
+		if (hashtab->table0[i].state == CUCKOO_HASH_STATE_OCCUPIED)
+		EPRINTK("%d %d %llx %d %d %d\n", i,
+			hashtab->table0[i].state == CUCKOO_HASH_STATE_OCCUPIED,
+			hashtab->table0[i].key, hashtab->table0[i].value,
+			cuckoo_compute_hash(hashtab, &hashtab->table0[i].key, 
+					    &hashtab->a0),
+			cuckoo_compute_hash(hashtab, &hashtab->table0[i].key, 
+					    &hashtab->a1));
+		else
+		EPRINTK("%d %d - - - -\n", i,
+			hashtab->table0[i].state == CUCKOO_HASH_STATE_OCCUPIED);
+			
+	}
+	EPRINTK("-----------------------------------------\n");
+	EPRINTK("Index  Occupied  Key  Value Index0 Index1\n");
+	EPRINTK("-----------------------------------------\n");
+	for (i=0; i< hashtab->length; i++) {
+		if (hashtab->table1[i].state == CUCKOO_HASH_STATE_OCCUPIED)
+		EPRINTK("%d %d %llx %d %d %d\n", i,
+			hashtab->table1[i].state == CUCKOO_HASH_STATE_OCCUPIED,
+			hashtab->table1[i].key, hashtab->table1[i].value,
+			cuckoo_compute_hash(hashtab, &hashtab->table1[i].key, 
+					    &hashtab->a0),
+			cuckoo_compute_hash(hashtab, &hashtab->table1[i].key, 
+					    &hashtab->a1));
+		else
+		EPRINTK("%d %d - - - -\n", i,
+			hashtab->table1[i].state == CUCKOO_HASH_STATE_OCCUPIED);
+	} 
+	EPRINTK("======================\n");
+}
+EXPORT_SYMBOL_GPL(cuckoo_hash_dump);
+#endif
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netutil/accel_cuckoo_hash.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netutil/accel_cuckoo_hash.h
--- linux-2.6.18.8/drivers/xen/sfc_netutil/accel_cuckoo_hash.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netutil/accel_cuckoo_hash.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,227 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+/*
+ * A cuckoo hash table consists of two sub tables.  Each entry can
+ * hash to a position in each table.  If, on entry, its position is
+ * found to be occupied, the existing element is moved to it's other
+ * location.  This recurses until success or a loop is found.  If a
+ * loop is found the table is rehashed.
+ *
+ *  See http://www.it-c.dk/people/pagh/papers/cuckoo-jour.pdf
+ */
+
+#ifndef NET_ACCEL_CUCKOO_HASH_H
+#define NET_ACCEL_CUCKOO_HASH_H
+
+/*! Type used for hash table keys of ip pairs */
+typedef struct {
+	u32 local_ip;
+	//u32 remote_ip;
+	u16 local_port;
+	//u16 remote_port;
+	/* Technically only 1 bit, but use 16 to make key a round
+	   number size */
+	u16 proto;
+} cuckoo_hash_ip_key;
+
+/*! Type used for hash table keys of mac addresses */
+typedef u64 cuckoo_hash_mac_key;
+
+/*! This type is designed to be large enough to hold all supported key
+ *  sizes to avoid having to malloc storage for them.
+ */
+typedef u64 cuckoo_hash_key;
+
+/*! Type used for the values stored in the hash table */
+typedef int cuckoo_hash_value;
+
+/*! Type used for the hash used to index the table */
+typedef u32 cuckoo_hash;
+
+/*! How long to spend displacing values when adding before giving up
+ *  and rehashing */
+#define CUCKOO_HASH_MAX_LOOP (hashtab->length)
+
+/*! State of hash table entry */
+typedef enum {
+	CUCKOO_HASH_STATE_VACANT = 0,
+	CUCKOO_HASH_STATE_OCCUPIED 
+} cuckoo_hash_state;
+
+/*! An entry in the hash table */
+typedef struct {
+	cuckoo_hash_state state;
+	cuckoo_hash_key key;
+	cuckoo_hash_value value;
+} cuckoo_hash_entry;
+
+/*! A cuckoo hash table */
+typedef struct {
+	/*! The length of each table (NB. there are two tables of this
+	 *  length) */
+	unsigned length; 
+	/*! The length of each table in bits */
+	unsigned length_bits;
+	/*! The length of the key in bytes */ 
+	unsigned key_length; 
+	/*! The number of entries currently stored in the table */
+	unsigned entries;
+	/*! Index into table used by cuckoo_hash_iterate */
+	unsigned iterate_index; 
+
+	/* parameter of hash functions */
+	/*! The "a" parameter of the first hash function */
+	cuckoo_hash_key a0; 
+	/*! The "a" parameter of the second hash function */
+	cuckoo_hash_key a1; 
+
+	/*! The first table */
+	cuckoo_hash_entry *table0; 
+	/*! The second table */
+	cuckoo_hash_entry *table1; 
+} cuckoo_hash_table;
+
+/*! Initialise the cuckoo has table 
+ *
+ * \param hashtab A pointer to an unitialised hash table structure
+ * \param length_bits The number of elements in each table equals
+ * 2**length_bits
+ * \param key_length The length of the key in bytes
+ *
+ * \return 0 on success, -ENOMEM if it couldn't allocate the tables
+ */
+extern
+int cuckoo_hash_init(cuckoo_hash_table *hashtab, unsigned length_bits,
+		     unsigned key_length);
+
+
+/*! Destroy a hash table
+ *
+ * \param hashtab A hash table that has previously been passed to a
+ * successful call of cuckoo_hash_init()
+ */
+extern
+void cuckoo_hash_destroy(cuckoo_hash_table *hashtab);
+
+
+/*! Lookup an entry in the hash table 
+ *
+ * \param hashtab The hash table in which to look.
+ * \param key Pointer to a mac address to use as the key
+ * \param value On exit set to the value stored if key was present
+ *
+ * \return 0 if not present in the table, non-zero if it is (and value
+ * is set accordingly)
+ */
+extern
+int cuckoo_hash_lookup(cuckoo_hash_table *hashtab,
+		       cuckoo_hash_key *key,
+		       cuckoo_hash_value *value);
+
+/*! Add an entry to the hash table.  Key must not be a duplicate of
+ * anything already in the table.  If this is a risk, see
+ * cuckoo_hash_add_check
+ *
+ * \param hashtab The hash table to add the entry to
+ * \param key Pointer to a mac address to use as a key
+ * \param value The value to store 
+ * \param can_rehash Flag to allow the add function to rehash the
+ * table if necessary
+ *
+ * \return 0 on success, non-zero on failure.  -ENOSPC means it just
+ * couldn't find anywhere to put it - this is bad and probably means
+ * an entry has been dropped on the floor (but the entry you just
+ * tried to add may now be included)
+ */
+extern
+int cuckoo_hash_add(cuckoo_hash_table *hashtab,
+		    cuckoo_hash_key *key, 
+		    cuckoo_hash_value value,
+		    int can_rehash);
+
+/*! Same as cuckoo_hash_add but first checks to ensure entry is not
+ * already there
+ * \return -EBUSY if already there
+ */
+
+extern
+int cuckoo_hash_add_check(cuckoo_hash_table *hashtab,
+			  cuckoo_hash_key *key, 
+			  cuckoo_hash_value value,
+			  int can_rehash);
+/*! Remove an entry from the table 
+ *
+ * \param hashtab The hash table to remove the entry from
+ * \param key The key that was used to previously add the entry
+ *
+ * \return 0 on success, -EINVAL if the entry couldn't be found 
+ */
+extern
+int cuckoo_hash_remove(cuckoo_hash_table *hashtab, cuckoo_hash_key *key);
+
+
+/*! Helper for those using mac addresses to convert to a key for the
+ *  hash table
+ */
+static inline cuckoo_hash_mac_key cuckoo_mac_to_key(const u8 *mac)
+{
+	return (cuckoo_hash_mac_key)(mac[0])
+		| (cuckoo_hash_mac_key)(mac[1]) << 8
+		| (cuckoo_hash_mac_key)(mac[2]) << 16
+		| (cuckoo_hash_mac_key)(mac[3]) << 24
+		| (cuckoo_hash_mac_key)(mac[4]) << 32
+		| (cuckoo_hash_mac_key)(mac[5]) << 40;
+}
+
+
+/*! Update an entry already in the hash table to take a new value 
+ *
+ * \param hashtab The hash table to add the entry to
+ * \param key Pointer to a mac address to use as a key
+ * \param value The value to store 
+ *
+ * \return 0 on success, non-zero on failure. 
+ */
+int cuckoo_hash_update(cuckoo_hash_table *hashtab, cuckoo_hash_key *key,
+		       cuckoo_hash_value value);
+
+
+/*! Go through the hash table and return all used entries (one per call)
+ *
+ * \param hashtab The hash table to iterate over 
+ * \param key Pointer to a key to take the returned key
+ * \param value Pointer to a value to take the returned value
+ *
+ * \return 0 on success (key, value set), non-zero on failure.
+ */
+int cuckoo_hash_iterate(cuckoo_hash_table *hashtab,
+			cuckoo_hash_key *key, cuckoo_hash_value *value);
+void cuckoo_hash_iterate_reset(cuckoo_hash_table *hashtab);
+
+/* debug, not compiled by default */
+void cuckoo_hash_valid(cuckoo_hash_table *hashtab);
+void cuckoo_hash_dump(cuckoo_hash_table *hashtab);
+
+#endif /* NET_ACCEL_CUCKOO_HASH_H */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netutil/accel_msg_iface.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netutil/accel_msg_iface.c
--- linux-2.6.18.8/drivers/xen/sfc_netutil/accel_msg_iface.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netutil/accel_msg_iface.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,301 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <xen/evtchn.h>
+
+#include "accel_util.h"
+#include "accel_msg_iface.h"
+
+#define NET_ACCEL_MSG_Q_SIZE (1024)
+#define NET_ACCEL_MSG_Q_MASK (NET_ACCEL_MSG_Q_SIZE - 1)
+
+#ifdef NDEBUG
+#define NET_ACCEL_CHECK_MAGIC(_p, _errval)
+#define NET_ACCEL_SHOW_QUEUE(_t, _q, _id)
+#else
+#define NET_ACCEL_CHECK_MAGIC(_p, _errval)				\
+	if (_p->magic != NET_ACCEL_MSG_MAGIC) {				\
+		printk(KERN_ERR "%s: passed invalid shared page %p!\n", \
+		       __FUNCTION__, _p);				\
+		return _errval;						\
+	}
+#define NET_ACCEL_SHOW_QUEUE(_t, _q, _id)				\
+	printk(_t ": queue %d write %x read %x base %x limit %x\n",     \
+	       _id, _q->write, _q->read, _q->base, _q->limit);
+#endif
+
+/*
+ * We've been passed at least 2 pages. 1 control page and 1 or more
+ * data pages.
+ */
+int net_accel_msg_init_page(void *mem, int len, int up)
+{
+	struct net_accel_shared_page *shared_page = 
+		(struct net_accel_shared_page*)mem;
+
+	if ((unsigned long)shared_page & NET_ACCEL_MSG_Q_MASK)
+		return -EINVAL;
+
+	shared_page->magic = NET_ACCEL_MSG_MAGIC;
+
+	shared_page->aflags = 0;
+
+	shared_page->net_dev_up = up;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(net_accel_msg_init_page);
+
+
+void net_accel_msg_init_queue(sh_msg_fifo2 *queue,
+			      struct net_accel_msg_queue *indices,
+			      struct net_accel_msg *base, int size)
+{
+	queue->fifo = base;
+	spin_lock_init(&queue->lock);
+	sh_fifo2_init(queue, size-1, &indices->read, &indices->write);
+}
+EXPORT_SYMBOL_GPL(net_accel_msg_init_queue);
+
+
+static inline int _net_accel_msg_send(struct net_accel_shared_page *sp,
+				      sh_msg_fifo2 *queue,
+				      struct net_accel_msg *msg,
+				      int is_reply)
+{
+	int rc = 0;
+	NET_ACCEL_CHECK_MAGIC(sp, -EINVAL);
+	rmb();
+	if (is_reply) {
+		EPRINTK_ON(sh_fifo2_is_full(queue));
+		sh_fifo2_put(queue, *msg);
+	} else {
+		if (sh_fifo2_not_half_full(queue)) {
+			sh_fifo2_put(queue, *msg);
+		} else {
+			rc = -ENOSPC;
+		}
+	}
+	wmb();
+	return rc;
+}
+
+/* Notify after a batch of messages have been sent */
+void net_accel_msg_notify(int irq)
+{
+	notify_remote_via_irq(irq);
+}
+EXPORT_SYMBOL_GPL(net_accel_msg_notify);
+
+/* 
+ * Send a message on the specified FIFO. Returns 0 on success, -errno
+ * on failure. The message in msg is copied to the current slot of the
+ * FIFO.
+ */
+int net_accel_msg_send(struct net_accel_shared_page *sp, sh_msg_fifo2 *q, 
+		       struct net_accel_msg *msg)
+{
+	unsigned long flags;
+	int rc;
+	net_accel_msg_lock_queue(q, &flags);
+	rc = _net_accel_msg_send(sp, q, msg, 0);
+	net_accel_msg_unlock_queue(q, &flags);
+	return rc;
+}
+EXPORT_SYMBOL_GPL(net_accel_msg_send);
+
+
+/* As net_accel_msg_send but also posts a notification to the far end. */
+int net_accel_msg_send_notify(struct net_accel_shared_page *sp, int irq, 
+			      sh_msg_fifo2 *q, struct net_accel_msg *msg)
+{
+	unsigned long flags;
+	int rc;
+	net_accel_msg_lock_queue(q, &flags);
+	rc = _net_accel_msg_send(sp, q, msg, 0);
+	net_accel_msg_unlock_queue(q, &flags);
+	if (rc >= 0)
+		notify_remote_via_irq(irq);
+	return rc;
+}
+EXPORT_SYMBOL_GPL(net_accel_msg_send_notify);
+
+
+int net_accel_msg_reply(struct net_accel_shared_page *sp, sh_msg_fifo2 *q, 
+		       struct net_accel_msg *msg)
+{
+	unsigned long flags;
+	int rc;
+	net_accel_msg_lock_queue(q, &flags);
+	rc = _net_accel_msg_send(sp, q, msg, 1);
+	net_accel_msg_unlock_queue(q, &flags);
+	return rc;
+}
+EXPORT_SYMBOL_GPL(net_accel_msg_reply);
+
+
+/* As net_accel_msg_send but also posts a notification to the far end. */
+int net_accel_msg_reply_notify(struct net_accel_shared_page *sp, int irq, 
+			      sh_msg_fifo2 *q, struct net_accel_msg *msg)
+{
+	unsigned long flags;
+	int rc;
+	net_accel_msg_lock_queue(q, &flags);
+	rc = _net_accel_msg_send(sp, q, msg, 1);
+	net_accel_msg_unlock_queue(q, &flags);
+	if (rc >= 0)
+		notify_remote_via_irq(irq);
+	return rc;
+}
+EXPORT_SYMBOL_GPL(net_accel_msg_reply_notify);
+
+
+/*
+ * Look at a received message, if any, so a decision can be made about
+ * whether to read it now or not.  Cookie is a bit of debug which is
+ * set here and checked when passed to net_accel_msg_recv_next()
+ */
+int net_accel_msg_peek(struct net_accel_shared_page *sp, 
+		       sh_msg_fifo2 *queue, 
+		       struct net_accel_msg *msg, int *cookie)
+{
+	unsigned long flags;
+	int rc = 0;
+	NET_ACCEL_CHECK_MAGIC(sp, -EINVAL);
+	net_accel_msg_lock_queue(queue, &flags);
+	rmb();
+	if (sh_fifo2_is_empty(queue)) {
+		rc = -ENOENT;
+	} else {
+		*msg = sh_fifo2_peek(queue);
+		*cookie = *(queue->fifo_rd_i);
+	}
+	net_accel_msg_unlock_queue(queue, &flags);
+	return rc;
+}
+EXPORT_SYMBOL_GPL(net_accel_msg_peek);
+
+
+/*
+ * Move the queue onto the next element, used after finished with a
+ * peeked msg 
+ */
+int net_accel_msg_recv_next(struct net_accel_shared_page *sp, 
+			    sh_msg_fifo2 *queue, int cookie)
+{
+	unsigned long flags;
+	NET_ACCEL_CHECK_MAGIC(sp, -EINVAL);
+	net_accel_msg_lock_queue(queue, &flags);
+	rmb();
+	/* Mustn't be empty */
+	BUG_ON(sh_fifo2_is_empty(queue));
+	/* 
+	 * Check cookie matches, i.e. we're advancing over the same message
+	 * as was got using peek 
+	 */
+	BUG_ON(cookie != *(queue->fifo_rd_i));
+	sh_fifo2_rd_next(queue);
+	wmb();
+	net_accel_msg_unlock_queue(queue, &flags);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(net_accel_msg_recv_next);
+
+
+/* 
+ * Receive a message on the specified FIFO. Returns 0 on success,
+ * -errno on failure.
+ */
+int net_accel_msg_recv(struct net_accel_shared_page *sp, sh_msg_fifo2 *queue, 
+		       struct net_accel_msg *msg)
+{
+	unsigned long flags;
+	int rc = 0;
+	NET_ACCEL_CHECK_MAGIC(sp, -EINVAL);
+	net_accel_msg_lock_queue(queue, &flags);
+	rmb();
+	if (sh_fifo2_is_empty(queue)) {
+		rc = -ENOENT;
+	} else {
+		sh_fifo2_get(queue, msg);
+	}
+	wmb();
+	net_accel_msg_unlock_queue(queue, &flags);
+	return rc;
+}
+EXPORT_SYMBOL_GPL(net_accel_msg_recv);
+
+
+/* 
+ * Start sending a message without copying. returns a pointer to a message
+ * that will be filled out in place. The queue is locked until the message 
+ * is sent.
+ */
+struct net_accel_msg *net_accel_msg_start_send(struct net_accel_shared_page *sp,
+					       sh_msg_fifo2 *queue, unsigned long *flags)
+{
+	struct net_accel_msg *msg;
+	NET_ACCEL_CHECK_MAGIC(sp, NULL);
+	net_accel_msg_lock_queue(queue, flags);
+	rmb();
+	if (sh_fifo2_not_half_full(queue)) {
+		msg = sh_fifo2_pokep(queue);
+	} else {
+		net_accel_msg_unlock_queue(queue, flags);
+		msg = NULL;
+	}
+	return msg;
+}
+EXPORT_SYMBOL_GPL(net_accel_msg_start_send);
+
+
+static inline void _msg_complete(struct net_accel_shared_page *sp,
+				 sh_msg_fifo2 *queue,
+				 unsigned long *flags)
+{
+	sh_fifo2_wr_next(queue);
+	net_accel_msg_unlock_queue(queue, flags);
+}
+
+/*
+ * Complete the sending of a message started with net_accel_msg_start_send. The 
+ * message is implicit since the queue was locked by _start
+ */
+void net_accel_msg_complete_send(struct net_accel_shared_page *sp,
+				 sh_msg_fifo2 *queue,
+				 unsigned long *flags)
+{
+	_msg_complete(sp, queue, flags);
+}
+EXPORT_SYMBOL_GPL(net_accel_msg_complete_send);
+
+/* As net_accel_msg_complete_send but does the notify. */
+void net_accel_msg_complete_send_notify(struct net_accel_shared_page *sp, 
+					sh_msg_fifo2 *queue, 
+					unsigned long *flags, int irq)
+{
+	_msg_complete(sp, queue, flags);
+	notify_remote_via_irq(irq);
+}
+EXPORT_SYMBOL_GPL(net_accel_msg_complete_send_notify);
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netutil/accel_msg_iface.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netutil/accel_msg_iface.h
--- linux-2.6.18.8/drivers/xen/sfc_netutil/accel_msg_iface.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netutil/accel_msg_iface.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,414 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef NET_ACCEL_MSG_IFACE_H
+#define NET_ACCEL_MSG_IFACE_H
+
+#include <linux/ip.h>
+#include <linux/tcp.h>
+#include <linux/udp.h>
+#include <linux/in.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+
+#include "accel_shared_fifo.h"
+
+#define NET_ACCEL_MSG_MAGIC (0x85465479)
+
+/*! We talk version 0.010 of the interdomain protocol */
+#define NET_ACCEL_MSG_VERSION (0x00001000)
+
+/*! Shared memory portion of inter-domain FIFO */
+struct net_accel_msg_queue {
+	u32 read;
+	u32 write;
+};
+
+
+/*
+ * The aflags in the following structure is used as follows:
+ *
+ *  - each bit is set when one of the corresponding variables is
+ *  changed by either end.
+ *
+ *  - the end that has made the change then forwards an IRQ to the
+ *  other
+ *
+ *  - the IRQ handler deals with these bits either on the fast path, or
+ *  for less common changes, by jumping onto the slow path.
+ *
+ *  - once it has seen a change, it clears the relevant bit.
+ *
+ * aflags is accessed atomically using clear_bit, test_bit,
+ * test_and_set_bit etc
+ */
+
+/*
+ * The following used to signify to the other domain when the queue
+ * they want to use is full, and when it is no longer full.  Could be
+ * compressed to use fewer bits but done this way for simplicity and
+ * clarity
+ */
+
+/* "dom0->domU queue" is full */
+#define NET_ACCEL_MSG_AFLAGS_QUEUE0FULL      0x1 
+#define NET_ACCEL_MSG_AFLAGS_QUEUE0FULL_B    0
+/* "dom0->domU queue" is not full */
+#define NET_ACCEL_MSG_AFLAGS_QUEUE0NOTFULL   0x2 
+#define NET_ACCEL_MSG_AFLAGS_QUEUE0NOTFULL_B 1
+/* "domU->dom0 queue" is full */
+#define NET_ACCEL_MSG_AFLAGS_QUEUEUFULL      0x4 
+#define NET_ACCEL_MSG_AFLAGS_QUEUEUFULL_B    2
+/* "domU->dom0 queue" is not full */
+#define NET_ACCEL_MSG_AFLAGS_QUEUEUNOTFULL   0x8
+#define NET_ACCEL_MSG_AFLAGS_QUEUEUNOTFULL_B 3
+/* dom0 -> domU net_dev up/down events */
+#define NET_ACCEL_MSG_AFLAGS_NETUPDOWN	 0x10
+#define NET_ACCEL_MSG_AFLAGS_NETUPDOWN_B       4
+
+/*
+ * Masks used to test if there are any messages for domU and dom0
+ * respectively
+ */
+#define NET_ACCEL_MSG_AFLAGS_TO_DOMU_MASK	\
+	(NET_ACCEL_MSG_AFLAGS_QUEUE0FULL    |	\
+	 NET_ACCEL_MSG_AFLAGS_QUEUEUNOTFULL |	\
+	 NET_ACCEL_MSG_AFLAGS_NETUPDOWN)
+#define NET_ACCEL_MSG_AFLAGS_TO_DOM0_MASK	\
+	(NET_ACCEL_MSG_AFLAGS_QUEUE0NOTFULL |	\
+	 NET_ACCEL_MSG_AFLAGS_QUEUEUFULL)
+
+/*! The shared data structure used for inter-VM communication. */
+struct net_accel_shared_page {
+	/*! Sanity check */
+	u32 magic;	    
+	/*! Used by host/Dom0 */
+	struct net_accel_msg_queue queue0;
+	/*! Used by guest/DomU */
+	struct net_accel_msg_queue queue1;
+	/*! Atomic flags, used to communicate simple state changes */
+	u32 aflags;     
+	/*! State of net_dev used for acceleration */     
+	u32 net_dev_up; 
+};
+
+
+enum net_accel_hw_type {
+	/*! Not a virtualisable NIC: use slow path. */
+	NET_ACCEL_MSG_HWTYPE_NONE = 0,
+	/*! NIC is Falcon-based */
+	NET_ACCEL_MSG_HWTYPE_FALCON_A = 1,
+	NET_ACCEL_MSG_HWTYPE_FALCON_B = 2,
+};
+
+/*! The maximum number of pages used by an event queue. */
+#define EF_HW_FALCON_EVQ_PAGES 8
+
+struct net_accel_hw_falcon_b {
+	/* VI */
+	/*! Grant for Tx DMA Q */
+	u32 txdmaq_gnt;   
+	/*! Grant for Rx DMA Q */
+	u32 rxdmaq_gnt;   
+	/*! Machine frame number for Tx/Rx doorbell page */
+	u32 doorbell_mfn; 
+	/*! Grant for Tx/Rx doorbell page */
+	u32 doorbell_gnt;
+
+	/* Event Q */
+	/*! Grants for the pages of the EVQ */
+	u32 evq_mem_gnts[EF_HW_FALCON_EVQ_PAGES]; 
+	u32 evq_offs;
+	/*! log2(pages in event Q) */
+	u32 evq_order;    
+	/*! Capacity in events */
+	u32 evq_capacity; 
+	/*! Eventq pointer register physical address */
+	u32 evq_rptr; 
+	/*! Interface instance */
+	u32 instance; 
+	/*! Capacity of RX queue */
+	u32 rx_capacity;
+	/*! Capacity of TX queue */
+	u32 tx_capacity;
+
+	/* NIC */
+	s32 nic_arch;
+	s32 nic_revision;
+	u8 nic_variant;
+};
+
+struct net_accel_hw_falcon_a {
+	struct net_accel_hw_falcon_b common;
+	u32 evq_rptr_gnt;
+};
+
+
+/*! Description of the hardware that the DomU is being given. */
+struct net_accel_msg_hw {
+	u32 type;		/*!< Hardware type */
+	union {
+		struct net_accel_hw_falcon_a falcon_a;
+		struct net_accel_hw_falcon_b falcon_b;
+	} resources;
+};
+
+/*! Start-of-day handshake message. Dom0 fills in its version and
+ * sends, DomU checks, inserts its version and replies
+ */
+struct net_accel_msg_hello {
+	/*! Sender's version (set by each side in turn) */
+	u32 version;	
+	/*! max pages allocated/allowed for buffers */
+	u32 max_pages;      
+};
+
+/*! Maximum number of page requests that can fit in a message. */
+#define NET_ACCEL_MSG_MAX_PAGE_REQ (8)
+
+/*! Request for NIC buffers. DomU fils out pages and grants (and
+ *  optionally) reqid, dom0 fills out buf and sends reply 
+ */
+struct net_accel_msg_map_buffers {
+	u32 reqid;	/*!< Optional request ID */
+	u32 pages;	/*!< Number of pages to map */
+	u32 grants[NET_ACCEL_MSG_MAX_PAGE_REQ];  /*!< Grant ids to map */ 
+	u32 buf;	  /*!< NIC buffer address of pages obtained */
+};
+
+/*! Notification of a change to local mac address, used to filter
+  locally destined packets off the fast path */
+struct net_accel_msg_localmac {
+	u32 flags;	/*!< Should this be added or removed? */
+	u8 mac[ETH_ALEN]; /*!< The mac address to filter onto slow path */
+};
+
+struct net_accel_msg_fastpath {
+	u32 flags;	/*!< Should this be added or removed? */
+	u8  mac[ETH_ALEN];/*!< The mac address to filter onto fast path */
+	u16 port;	 /*!< The port of the connection */
+	u32 ip;	   /*!< The IP address of the connection */
+	u8  proto;	/*!< The protocol of connection (TCP/UDP) */
+};
+
+/*! Values for struct ef_msg_localmac/fastpath.flags */
+#define NET_ACCEL_MSG_ADD    0x1
+#define NET_ACCEL_MSG_REMOVE 0x2
+
+/*! Overall message structure */
+struct net_accel_msg {
+	/*! ID specifying type of messge */
+	u32 id;		     
+	union {
+		/*! handshake */
+		struct net_accel_msg_hello hello;  
+		/*! hardware description */
+		struct net_accel_msg_hw hw;	
+		/*! buffer map request */
+		struct net_accel_msg_map_buffers mapbufs; 
+		/*! mac address of a local interface */
+		struct net_accel_msg_localmac localmac; 
+		/*! address of a new fastpath connection */
+		struct net_accel_msg_fastpath fastpath; 
+		/*! make the message a fixed size */
+		u8 pad[128 - sizeof(u32)]; 
+	}  u;
+};
+
+
+#define NET_ACCEL_MSG_HW_TO_MSG(_u) container_of(_u, struct net_accel_msg, u.hw)
+
+/*! Inter-domain message FIFO */
+typedef struct {
+	struct net_accel_msg *fifo;
+	u32 fifo_mask;
+	u32 *fifo_rd_i;
+	u32 *fifo_wr_i;
+	spinlock_t lock;
+	u32 is_locked; /* Debug flag */
+} sh_msg_fifo2;
+
+
+#define NET_ACCEL_MSG_OFFSET_MASK PAGE_MASK
+
+/* Modifiers */
+#define NET_ACCEL_MSG_REPLY    (0x80000000)
+#define NET_ACCEL_MSG_ERROR    (0x40000000)
+
+/* Dom0 -> DomU and reply. Handshake/version check. */
+#define NET_ACCEL_MSG_HELLO    (0x00000001)
+/* Dom0 -> DomU : hardware setup (VI info.) */
+#define NET_ACCEL_MSG_SETHW    (0x00000002)
+/*
+ * Dom0 -> DomU. Notification of a local mac to add/remove from slow
+ * path filter
+ */
+#define NET_ACCEL_MSG_LOCALMAC (0x00000003)
+/* 
+ * DomU -> Dom0 and reply. Request for buffer table entries for
+ * preallocated pages.
+ */
+#define NET_ACCEL_MSG_MAPBUF   (0x00000004)
+/* 
+ * Dom0 -> DomU. Notification of a local mac to add/remove from fast
+ * path filter
+ */
+#define NET_ACCEL_MSG_FASTPATH (0x00000005)
+
+/*! Initialise a message and set the type
+ * \param message : the message
+ * \param code : the message type 
+ */
+static inline void net_accel_msg_init(struct net_accel_msg *msg, int code) {
+	msg->id = (u32)code;
+}
+
+/*! initialise a shared page structure
+ * \param shared_page : mapped memory in which the structure resides
+ * \param len : size of the message FIFO area that follows
+ * \param up : initial up/down state of netdev 
+ * \return 0 or an error code
+ */
+extern int net_accel_msg_init_page(void *shared_page, int len, int up);
+
+/*! initialise a message queue 
+ * \param queue : the message FIFO to initialise 
+ * \param indices : the read and write indices in shared memory
+ * \param base : the start of the memory area for the FIFO
+ * \param size : the size of the FIFO in bytes
+ */
+extern void net_accel_msg_init_queue(sh_msg_fifo2 *queue,
+				     struct net_accel_msg_queue *indices,
+				     struct net_accel_msg *base, int size);
+
+/* Notify after a batch of messages have been sent */
+extern void net_accel_msg_notify(int irq);
+
+/*! Send a message on the specified FIFO. The message is copied to the 
+ *  current slot of the FIFO.
+ * \param sp : pointer to shared page
+ * \param q : pointer to message FIFO to use
+ * \param msg : pointer to message 
+ * \return 0 on success, -errno on
+ */ 
+extern int net_accel_msg_send(struct net_accel_shared_page *sp,
+			      sh_msg_fifo2 *q, 
+			      struct net_accel_msg *msg);
+extern int net_accel_msg_reply(struct net_accel_shared_page *sp,
+			      sh_msg_fifo2 *q, 
+			      struct net_accel_msg *msg);
+
+/*! As net_accel_msg_send but also posts a notification to the far end. */
+extern int net_accel_msg_send_notify(struct net_accel_shared_page *sp, 
+				     int irq, sh_msg_fifo2 *q, 
+				     struct net_accel_msg *msg);
+/*! As net_accel_msg_send but also posts a notification to the far end. */
+extern int net_accel_msg_reply_notify(struct net_accel_shared_page *sp, 
+				      int irq, sh_msg_fifo2 *q, 
+				      struct net_accel_msg *msg);
+
+/*! Receive a message on the specified FIFO. Returns 0 on success,
+ *  -errno on failure.
+ */
+extern int net_accel_msg_recv(struct net_accel_shared_page *sp,
+			      sh_msg_fifo2 *q,
+			      struct net_accel_msg *msg);
+
+/*! Look at a received message, if any, so a decision can be made
+ *  about whether to read it now or not.  Cookie is a bit of debug
+ *  which is set here and checked when passed to
+ *  net_accel_msg_recv_next()
+ */
+extern int net_accel_msg_peek(struct net_accel_shared_page *sp,
+			      sh_msg_fifo2 *queue, 
+			      struct net_accel_msg *msg, int *cookie);
+/*! Move the queue onto the next element, used after finished with a
+ *  peeked msg 
+ */
+extern int net_accel_msg_recv_next(struct net_accel_shared_page *sp,
+				   sh_msg_fifo2 *queue, int cookie);
+
+/*! Start sending a message without copying. returns a pointer to a
+ *  message that will be filled out in place. The queue is locked
+ *  until the message is sent.
+ */
+extern 
+struct net_accel_msg *net_accel_msg_start_send(struct net_accel_shared_page *sp,
+					       sh_msg_fifo2 *queue,
+					       unsigned long *flags);
+
+
+/*! Complete the sending of a message started with
+ *  net_accel_msg_start_send. The message is implicit since the queue
+ *  was locked by _start 
+ */
+extern void net_accel_msg_complete_send(struct net_accel_shared_page *sp,
+					sh_msg_fifo2 *queue,
+					unsigned long *flags);
+
+/*! As net_accel_msg_complete_send but does the notify. */
+extern void net_accel_msg_complete_send_notify(struct net_accel_shared_page *sp, 
+					       sh_msg_fifo2 *queue,
+					       unsigned long *flags, int irq);
+
+/*! Lock the queue so that multiple "_locked" functions can be called
+ *  without the queue being modified by others 
+ */
+static inline
+void net_accel_msg_lock_queue(sh_msg_fifo2 *queue, unsigned long *flags)
+{
+	spin_lock_irqsave(&queue->lock, (*flags));
+	rmb();
+	BUG_ON(queue->is_locked);
+	queue->is_locked = 1;
+}
+
+/*! Unlock the queue */
+static inline
+void net_accel_msg_unlock_queue(sh_msg_fifo2 *queue, unsigned long *flags)
+{
+	BUG_ON(!queue->is_locked);
+	queue->is_locked = 0;
+	wmb();
+	spin_unlock_irqrestore(&queue->lock, (*flags));
+}
+
+/*! Give up without sending a message that was started with
+ *  net_accel_msg_start_send() 
+ */
+static inline 
+void net_accel_msg_abort_send(struct net_accel_shared_page *sp,
+			      sh_msg_fifo2 *queue, unsigned long *flags)
+{
+	net_accel_msg_unlock_queue(queue, flags);
+}
+
+/*! Test the queue to ensure there is sufficient space */
+static inline
+int net_accel_msg_check_space(sh_msg_fifo2 *queue, unsigned space)
+{
+	return sh_fifo2_space(queue) >= space;
+}
+
+#endif /* NET_ACCEL_MSG_IFACE_H */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netutil/accel_shared_fifo.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netutil/accel_shared_fifo.h
--- linux-2.6.18.8/drivers/xen/sfc_netutil/accel_shared_fifo.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netutil/accel_shared_fifo.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,127 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef NET_ACCEL_SHARED_FIFO_H
+#define NET_ACCEL_SHARED_FIFO_H
+
+/*
+ * This is based on fifo.h, but handles sharing between address spaces
+ * that don't trust each other, by splitting out the read and write
+ * indices. This costs at least one pointer indirection more than the
+ * vanilla version per access.
+ */
+
+typedef struct {
+	char*	 fifo;
+	unsigned      fifo_mask;
+	unsigned      *fifo_rd_i;
+	unsigned      *fifo_wr_i;
+} sh_byte_fifo2;
+
+#define SH_FIFO2_M(f, x)     ((x) & ((f)->fifo_mask))
+
+static inline unsigned log2_ge(unsigned long n, unsigned min_order) {
+	unsigned order = min_order;
+	while((1ul << order) < n) ++order;
+	return order;
+}
+
+static inline unsigned long pow2(unsigned order) {
+	return (1ul << order);
+}
+
+#define is_pow2(x)  (pow2(log2_ge((x), 0)) == (x))
+
+#define sh_fifo2_valid(f)  ((f) && (f)->fifo && (f)->fifo_mask > 0 &&   \
+			    is_pow2((f)->fifo_mask+1u))
+
+#define sh_fifo2_init(f, cap, _rptr, _wptr)		\
+	do {						\
+		BUG_ON(!is_pow2((cap) + 1));		\
+		(f)->fifo_rd_i = _rptr;			\
+		(f)->fifo_wr_i = _wptr;			\
+		*(f)->fifo_rd_i = *(f)->fifo_wr_i = 0u; \
+		(f)->fifo_mask = (cap);			\
+	} while(0)
+
+#define sh_fifo2_num(f)      SH_FIFO2_M((f),*(f)->fifo_wr_i - *(f)->fifo_rd_i)
+#define sh_fifo2_space(f)    SH_FIFO2_M((f),*(f)->fifo_rd_i - *(f)->fifo_wr_i-1u)
+#define sh_fifo2_is_empty(f)  (sh_fifo2_num(f)==0)
+#define sh_fifo2_not_empty(f) (sh_fifo2_num(f)!=0)
+#define sh_fifo2_is_full(f)   (sh_fifo2_space(f)==0u)
+#define sh_fifo2_not_full(f)  (sh_fifo2_space(f)!=0u)
+#define sh_fifo2_buf_size(f) ((f)->fifo_mask + 1u)
+#define sh_fifo2_capacity(f) ((f)->fifo_mask)
+#define sh_fifo2_end(f)      ((f)->fifo + sh_fifo2_buf_size(f))
+#define sh_fifo2_not_half_full(f) (sh_fifo2_space(f) > (sh_fifo2_capacity(f) >> 1))
+
+#define sh_fifo2_peek(f)     ((f)->fifo[SH_FIFO2_M((f), *(f)->fifo_rd_i)])
+#define sh_fifo2_peekp(f)    ((f)->fifo + SH_FIFO2_M((f), *(f)->fifo_rd_i))
+#define sh_fifo2_poke(f)     ((f)->fifo[SH_FIFO2_M((f), *(f)->fifo_wr_i)])
+#define sh_fifo2_pokep(f)    ((f)->fifo + SH_FIFO2_M((f), *(f)->fifo_wr_i))
+#define sh_fifo2_peek_i(f,i) ((f)->fifo[SH_FIFO2_M((f), *(f)->fifo_rd_i+(i))])
+#define sh_fifo2_poke_i(f,i) ((f)->fifo[SH_FIFO2_M((f), *(f)->fifo_wr_i+(i))])
+
+#define sh_fifo2_rd_next(f)					\
+	do {*(f)->fifo_rd_i = *(f)->fifo_rd_i + 1u;} while(0)
+#define sh_fifo2_wr_next(f)					\
+	do {*(f)->fifo_wr_i = *(f)->fifo_wr_i + 1u;} while(0)
+#define sh_fifo2_rd_adv(f, n)					\
+	do {*(f)->fifo_rd_i = *(f)->fifo_rd_i + (n);} while(0)
+#define sh_fifo2_wr_adv(f, n)					\
+	do {*(f)->fifo_wr_i = *(f)->fifo_wr_i + (n);} while(0)
+
+#define sh_fifo2_put(f, v)						\
+	do {sh_fifo2_poke(f) = (v); wmb(); sh_fifo2_wr_next(f);} while(0)
+
+#define sh_fifo2_get(f, pv)						\
+	do {*(pv) = sh_fifo2_peek(f); mb(); sh_fifo2_rd_next(f);} while(0)
+
+static inline unsigned sh_fifo2_contig_num(sh_byte_fifo2 *f)
+{
+	unsigned fifo_wr_i = SH_FIFO2_M(f, *f->fifo_wr_i);
+	unsigned fifo_rd_i = SH_FIFO2_M(f, *f->fifo_rd_i);
+
+	return (fifo_wr_i >= fifo_rd_i)
+		? fifo_wr_i - fifo_rd_i
+		: f->fifo_mask + 1u - *(f)->fifo_rd_i;
+}
+
+static inline unsigned sh_fifo2_contig_space(sh_byte_fifo2 *f)
+{
+	unsigned fifo_wr_i = SH_FIFO2_M(f, *f->fifo_wr_i);
+	unsigned fifo_rd_i = SH_FIFO2_M(f, *f->fifo_rd_i);
+
+	return (fifo_rd_i > fifo_wr_i)
+		? fifo_rd_i - fifo_wr_i - 1
+		: (f->fifo_mask + 1u - fifo_wr_i
+		   /*
+		    * The last byte can't be used if the read pointer
+		    * is at zero.
+		    */
+		   - (fifo_rd_i==0));
+}
+
+
+#endif /* NET_ACCEL_SHARED_FIFO_H */
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netutil/accel_util.c linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netutil/accel_util.c
--- linux-2.6.18.8/drivers/xen/sfc_netutil/accel_util.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netutil/accel_util.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,333 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#include <linux/if_ether.h>
+#include <asm/io.h>
+#include <asm/pgtable.h>
+#include <asm/hypercall.h>
+#include <xen/xenbus.h>
+#include <xen/driver_util.h>
+#include <xen/gnttab.h>
+
+#include "accel_util.h"
+
+#ifdef EFX_GCOV
+#include "gcov.h"
+
+static int __init net_accel_init(void)
+{
+	gcov_provider_init(THIS_MODULE);
+	return 0;
+}
+module_init(net_accel_init);
+
+static void __exit net_accel_exit(void)
+{
+	gcov_provider_fini(THIS_MODULE);
+}
+module_exit(net_accel_exit);
+#endif
+
+/* Shutdown remote domain that is misbehaving */
+int net_accel_shutdown_remote(int domain)
+{
+	struct sched_remote_shutdown sched_shutdown = {
+		.domain_id = domain,
+		.reason = SHUTDOWN_crash
+	};
+
+	EPRINTK("Crashing domain %d\n", domain);
+
+	return HYPERVISOR_sched_op(SCHEDOP_remote_shutdown, &sched_shutdown);
+}
+EXPORT_SYMBOL(net_accel_shutdown_remote);
+
+
+/* Based on xenbus_backend_client.c:xenbus_map_ring() */
+static int net_accel_map_grant(struct xenbus_device *dev, int gnt_ref,
+			       grant_handle_t *handle, void *vaddr, 
+			       u64 *dev_bus_addr, unsigned flags)
+{
+	struct gnttab_map_grant_ref op;
+	
+	gnttab_set_map_op(&op, (unsigned long)vaddr, flags,
+			  gnt_ref, dev->otherend_id);
+
+	BUG_ON(HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, &op, 1));
+
+	if (op.status != GNTST_okay) {
+		xenbus_dev_error
+			(dev, op.status,
+			 "failed mapping in shared page %d from domain %d\n",
+			 gnt_ref, dev->otherend_id);
+	} else {
+		*handle = op.handle;
+		if (dev_bus_addr)
+			*dev_bus_addr = op.dev_bus_addr;
+	}
+
+	return op.status;
+}
+
+
+/* Based on xenbus_backend_client.c:xenbus_unmap_ring() */
+static int net_accel_unmap_grant(struct xenbus_device *dev, 
+				 grant_handle_t handle,
+				 void *vaddr, u64 dev_bus_addr,
+				 unsigned flags)
+{
+	struct gnttab_unmap_grant_ref op;
+
+	gnttab_set_unmap_op(&op, (unsigned long)vaddr, flags, handle);
+	
+	if (dev_bus_addr)
+		op.dev_bus_addr = dev_bus_addr;
+
+	BUG_ON(HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1));
+
+	if (op.status != GNTST_okay)
+		xenbus_dev_error(dev, op.status,
+				 "failed unmapping page at handle %d error %d\n",
+				 handle, op.status);
+
+	return op.status;
+}
+
+
+int net_accel_map_device_page(struct xenbus_device *dev,  
+			      int gnt_ref, grant_handle_t *handle,
+			      u64 *dev_bus_addr)
+{
+	return net_accel_map_grant(dev, gnt_ref, handle, 0, dev_bus_addr,
+				   GNTMAP_device_map);
+}
+EXPORT_SYMBOL_GPL(net_accel_map_device_page);
+
+ 
+int net_accel_unmap_device_page(struct xenbus_device *dev,
+				grant_handle_t handle, u64 dev_bus_addr)
+{
+	return net_accel_unmap_grant(dev, handle, 0, dev_bus_addr, 
+				     GNTMAP_device_map);
+}
+EXPORT_SYMBOL_GPL(net_accel_unmap_device_page);
+
+
+struct net_accel_valloc_grant_mapping {
+	struct vm_struct *vm;
+	int pages;
+	grant_handle_t grant_handles[0];
+};
+
+/* Map a series of grants into a contiguous virtual area */
+static void *net_accel_map_grants_valloc(struct xenbus_device *dev, 
+					 unsigned *grants, int npages, 
+					 unsigned flags, void **priv)
+{
+	struct net_accel_valloc_grant_mapping *map;
+	struct vm_struct *vm;
+	void *addr;
+	int i, j, rc;
+
+	vm  = alloc_vm_area(PAGE_SIZE * npages);
+	if (vm == NULL) {
+		EPRINTK("No memory from alloc_vm_area.\n");
+		return NULL;
+	}
+	/* 
+	 * Get a structure in which we will record all the info needed
+	 * to undo the mapping.
+	 */
+	map = kzalloc(sizeof(struct net_accel_valloc_grant_mapping)  + 
+		      npages * sizeof(grant_handle_t), GFP_KERNEL);
+	if (map == NULL) {
+		EPRINTK("No memory for net_accel_valloc_grant_mapping\n");
+		free_vm_area(vm);
+		return NULL;
+	}
+	map->vm = vm;
+	map->pages = npages;
+
+	/* Do the actual mapping */
+	addr = vm->addr;
+	for (i = 0; i < npages; i++) {
+		rc = net_accel_map_grant(dev, grants[i], map->grant_handles + i, 
+					 addr, NULL, flags);
+		if (rc != 0)
+			goto undo;
+		addr = (void*)((unsigned long)addr + PAGE_SIZE);
+	}
+
+	if (priv)
+		*priv = (void *)map;
+	else
+		kfree(map);
+
+	return vm->addr;
+
+ undo:
+	EPRINTK("Aborting contig map due to single map failure %d (%d of %d)\n",
+		rc, i+1, npages);
+	for (j = 0; j < i; j++) {
+		addr = (void*)((unsigned long)vm->addr + (j * PAGE_SIZE));
+		net_accel_unmap_grant(dev, map->grant_handles[j], addr, 0,
+				      flags);
+	}
+	free_vm_area(vm);
+	kfree(map);
+	return NULL;
+}
+
+/* Undo the result of the mapping */
+static void net_accel_unmap_grants_vfree(struct xenbus_device *dev, 
+					 unsigned flags, void *priv)
+{
+	struct net_accel_valloc_grant_mapping *map = 
+		(struct net_accel_valloc_grant_mapping *)priv;
+
+	void *addr = map->vm->addr;
+	int npages = map->pages;
+	int i;
+
+	for (i = 0; i < npages; i++) {
+		net_accel_unmap_grant(dev, map->grant_handles[i], addr, 0,
+				      flags);
+		addr = (void*)((unsigned long)addr + PAGE_SIZE);
+	}
+	free_vm_area(map->vm);
+	kfree(map);
+}
+
+
+void *net_accel_map_grants_contig(struct xenbus_device *dev,
+				unsigned *grants, int npages, 
+				void **priv)
+{
+	return net_accel_map_grants_valloc(dev, grants, npages,
+					   GNTMAP_host_map, priv);
+}
+EXPORT_SYMBOL(net_accel_map_grants_contig);
+
+
+void net_accel_unmap_grants_contig(struct xenbus_device *dev,
+				   void *priv)
+{
+	net_accel_unmap_grants_vfree(dev, GNTMAP_host_map, priv);
+}
+EXPORT_SYMBOL(net_accel_unmap_grants_contig);
+
+
+void *net_accel_map_iomem_page(struct xenbus_device *dev, int gnt_ref,
+			     void **priv)
+{
+	return net_accel_map_grants_valloc(dev, &gnt_ref, 1, 
+					   GNTMAP_host_map, priv);
+}
+EXPORT_SYMBOL(net_accel_map_iomem_page);
+
+
+void net_accel_unmap_iomem_page(struct xenbus_device *dev, void *priv)
+{
+	net_accel_unmap_grants_vfree(dev, GNTMAP_host_map, priv);
+}
+EXPORT_SYMBOL(net_accel_unmap_iomem_page);
+
+
+int net_accel_grant_page(struct xenbus_device *dev, unsigned long mfn, 
+			 int is_iomem)
+{
+	int err = gnttab_grant_foreign_access(dev->otherend_id, mfn,
+					      is_iomem ? GTF_PCD : 0);
+	if (err < 0)
+		xenbus_dev_error(dev, err, "failed granting access to page\n");
+	return err;
+}
+EXPORT_SYMBOL_GPL(net_accel_grant_page);
+
+
+int net_accel_ungrant_page(grant_ref_t gntref)
+{
+	if (unlikely(gnttab_query_foreign_access(gntref) != 0)) {
+		EPRINTK("%s: remote domain still using grant %d\n", __FUNCTION__, 
+			gntref);
+		return -EBUSY;
+	}
+
+	gnttab_end_foreign_access(gntref, 0);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(net_accel_ungrant_page);
+
+
+int net_accel_xen_net_read_mac(struct xenbus_device *dev, u8 mac[])
+{
+	char *s, *e, *macstr;
+	int i;
+
+	macstr = s = xenbus_read(XBT_NIL, dev->nodename, "mac", NULL);
+	if (IS_ERR(macstr))
+		return PTR_ERR(macstr);
+
+	for (i = 0; i < ETH_ALEN; i++) {
+		mac[i] = simple_strtoul(s, &e, 16);
+		if ((s == e) || (*e != ((i == ETH_ALEN-1) ? '\0' : ':'))) {
+			kfree(macstr);
+			return -ENOENT;
+		}
+		s = e+1;
+	}
+
+	kfree(macstr);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(net_accel_xen_net_read_mac);
+
+
+void net_accel_update_state(struct xenbus_device *dev, int state)
+{
+	struct xenbus_transaction tr;
+	int err;
+
+	DPRINTK("%s: setting accelstate to %s\n", __FUNCTION__,
+		xenbus_strstate(state));
+
+	if (xenbus_exists(XBT_NIL, dev->nodename, "")) {
+		VPRINTK("%s: nodename %s\n", __FUNCTION__, dev->nodename);
+	again:
+		err = xenbus_transaction_start(&tr);
+		if (err == 0)
+			err = xenbus_printf(tr, dev->nodename, "accelstate",
+					    "%d", state);
+		if (err != 0) {
+			xenbus_transaction_end(tr, 1);
+		} else {
+			err = xenbus_transaction_end(tr, 0);
+			if (err == -EAGAIN)
+				goto again;
+		}
+	}
+}
+EXPORT_SYMBOL_GPL(net_accel_update_state);
+
+MODULE_LICENSE("GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netutil/accel_util.h linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netutil/accel_util.h
--- linux-2.6.18.8/drivers/xen/sfc_netutil/accel_util.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netutil/accel_util.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,127 @@
+/****************************************************************************
+ * Solarflare driver for Xen network acceleration
+ *
+ * Copyright 2006-2008: Solarflare Communications Inc,
+ *                      9501 Jeronimo Road, Suite 250,
+ *                      Irvine, CA 92618, USA
+ *
+ * Maintained by Solarflare Communications <linux-xen-drivers@solarflare.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation, incorporated herein by reference.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ ****************************************************************************
+ */
+
+#ifndef NETBACK_ACCEL_UTIL_H
+#define NETBACK_ACCEL_UTIL_H
+
+#ifdef DPRINTK
+#undef DPRINTK
+#endif
+
+#define FILE_LEAF strrchr(__FILE__, '/') ? strrchr(__FILE__, '/') + 1 : __FILE__
+
+#if 1
+#define VPRINTK(_f, _a...) 
+#else
+#define VPRINTK(_f, _a...)			\
+	printk("(file=%s, line=%d) " _f,	\
+	       FILE_LEAF , __LINE__ , ## _a )
+#endif
+
+#if 1
+#define DPRINTK(_f, _a...) 
+#else
+#define DPRINTK(_f, _a...)			\
+	printk("(file=%s, line=%d) " _f,	\
+	       FILE_LEAF , __LINE__ , ## _a )
+#endif
+
+#define EPRINTK(_f, _a...)			\
+	printk("(file=%s, line=%d) " _f,	\
+	       FILE_LEAF , __LINE__ , ## _a )
+
+#define EPRINTK_ON(exp)							\
+	do {								\
+		if (exp)						\
+			EPRINTK("%s at %s:%d\n", #exp, __FILE__, __LINE__); \
+	} while(0)
+
+#define DPRINTK_ON(exp)							\
+	do {								\
+		if (exp)						\
+			DPRINTK("%s at %s:%d\n", #exp, __FILE__, __LINE__); \
+	} while(0)
+
+#define MAC_FMT "%.2x:%.2x:%.2x:%.2x:%.2x:%.2x"
+#define MAC_ARG(_mac) (_mac)[0], (_mac)[1], (_mac)[2], (_mac)[3], (_mac)[4], (_mac)[5]
+
+#include <xen/xenbus.h>
+
+/*! Map a set of pages from another domain
+ * \param dev The xenbus device context
+ * \param priv The private data returned by the mapping function 
+ */
+extern 
+void *net_accel_map_grants_contig(struct xenbus_device *dev, 
+				  unsigned *grants, int npages, 
+				  void **priv);
+
+/*! Unmap a set of pages mapped using net_accel_map_grants_contig.
+ * \param dev The xenbus device context
+ * \param priv The private data returned by the mapping function 
+ */
+extern 
+void net_accel_unmap_grants_contig(struct xenbus_device *dev, void *priv);
+
+/*! Read the MAC address of a device from xenstore */
+extern
+int net_accel_xen_net_read_mac(struct xenbus_device *dev, u8 mac[]);
+
+/*! Update the accelstate field for a device in xenstore */
+extern
+void net_accel_update_state(struct xenbus_device *dev, int state);
+
+/* These four map/unmap functions are based on
+ * xenbus_backend_client.c:xenbus_map_ring().  However, they are not
+ * used for ring buffers, instead just to map pages between domains,
+ * or to map a page so that it is accessible by a device
+ */
+extern
+int net_accel_map_device_page(struct xenbus_device *dev,  
+			      int gnt_ref, grant_handle_t *handle,
+			      u64 *dev_bus_addr);
+extern
+int net_accel_unmap_device_page(struct xenbus_device *dev,
+				grant_handle_t handle, u64 dev_bus_addr);
+extern
+void *net_accel_map_iomem_page(struct xenbus_device *dev, int gnt_ref,
+			     void **priv);
+extern
+void net_accel_unmap_iomem_page(struct xenbus_device *dev, void *priv);
+
+/*! Grrant a page to remote domain */
+extern
+int net_accel_grant_page(struct xenbus_device *dev, unsigned long mfn, 
+			 int is_iomem);
+/*! Undo a net_accel_grant_page */
+extern
+int net_accel_ungrant_page(grant_ref_t gntref);
+
+
+/*! Shutdown remote domain that is misbehaving */
+extern
+int net_accel_shutdown_remote(int domain);
+
+
+#endif
diff -rpuN linux-2.6.18.8/drivers/xen/sfc_netutil/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netutil/Makefile
--- linux-2.6.18.8/drivers/xen/sfc_netutil/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/sfc_netutil/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,11 @@
+EXTRA_CFLAGS += -Idrivers/xen/sfc_netutil
+EXTRA_CFLAGS += -Werror
+
+ifdef GGOV
+EXTRA_CFLAGS += -fprofile-arcs -ftest-coverage -DEFX_GCOV
+endif
+
+obj-$(CONFIG_XEN_NETDEV_ACCEL_SFC_UTIL) := sfc_netutil.o
+
+sfc_netutil-objs := accel_cuckoo_hash.o accel_msg_iface.o accel_util.o 
+
diff -rpuN linux-2.6.18.8/drivers/xen/tpmback/common.h linux-2.6.18-xen-3.3.0/drivers/xen/tpmback/common.h
--- linux-2.6.18.8/drivers/xen/tpmback/common.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/tpmback/common.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,85 @@
+/******************************************************************************
+ * drivers/xen/tpmback/common.h
+ */
+
+#ifndef __TPM__BACKEND__COMMON_H__
+#define __TPM__BACKEND__COMMON_H__
+
+#include <linux/version.h>
+#include <linux/module.h>
+#include <linux/interrupt.h>
+#include <linux/slab.h>
+#include <xen/evtchn.h>
+#include <xen/driver_util.h>
+#include <xen/interface/grant_table.h>
+#include <xen/interface/io/tpmif.h>
+#include <asm/io.h>
+#include <asm/pgalloc.h>
+
+#define DPRINTK(_f, _a...)			\
+	pr_debug("(file=%s, line=%d) " _f,	\
+		 __FILE__ , __LINE__ , ## _a )
+
+struct backend_info;
+
+typedef struct tpmif_st {
+	struct list_head tpmif_list;
+	/* Unique identifier for this interface. */
+	domid_t domid;
+	unsigned int handle;
+
+	/* Physical parameters of the comms window. */
+	unsigned int irq;
+
+	/* The shared rings and indexes. */
+	tpmif_tx_interface_t *tx;
+	struct vm_struct *tx_area;
+
+	/* Miscellaneous private stuff. */
+	enum { DISCONNECTED, DISCONNECTING, CONNECTED } status;
+	int active;
+
+	struct tpmif_st *hash_next;
+	struct list_head list;	/* scheduling list */
+	atomic_t refcnt;
+
+	struct backend_info *bi;
+
+	grant_handle_t shmem_handle;
+	grant_ref_t shmem_ref;
+	struct page **mmap_pages;
+
+	char devname[20];
+} tpmif_t;
+
+void tpmif_disconnect_complete(tpmif_t * tpmif);
+tpmif_t *tpmif_find(domid_t domid, struct backend_info *bi);
+void tpmif_interface_init(void);
+void tpmif_interface_exit(void);
+void tpmif_schedule_work(tpmif_t * tpmif);
+void tpmif_deschedule_work(tpmif_t * tpmif);
+void tpmif_xenbus_init(void);
+void tpmif_xenbus_exit(void);
+int tpmif_map(tpmif_t *tpmif, unsigned long shared_page, unsigned int evtchn);
+irqreturn_t tpmif_be_int(int irq, void *dev_id, struct pt_regs *regs);
+
+long int tpmback_get_instance(struct backend_info *bi);
+
+int vtpm_release_packets(tpmif_t * tpmif, int send_msgs);
+
+
+#define tpmif_get(_b) (atomic_inc(&(_b)->refcnt))
+#define tpmif_put(_b)					\
+	do {						\
+		if (atomic_dec_and_test(&(_b)->refcnt))	\
+			tpmif_disconnect_complete(_b);	\
+	} while (0)
+
+extern int num_frontends;
+
+static inline unsigned long idx_to_kaddr(tpmif_t *t, unsigned int idx)
+{
+	return (unsigned long)pfn_to_kaddr(page_to_pfn(t->mmap_pages[idx]));
+}
+
+#endif /* __TPMIF__BACKEND__COMMON_H__ */
diff -rpuN linux-2.6.18.8/drivers/xen/tpmback/interface.c linux-2.6.18-xen-3.3.0/drivers/xen/tpmback/interface.c
--- linux-2.6.18.8/drivers/xen/tpmback/interface.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/tpmback/interface.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,168 @@
+ /*****************************************************************************
+ * drivers/xen/tpmback/interface.c
+ *
+ * Vritual TPM interface management.
+ *
+ * Copyright (c) 2005, IBM Corporation
+ *
+ * Author: Stefan Berger, stefanb@us.ibm.com
+ *
+ * This code has been derived from drivers/xen/netback/interface.c
+ * Copyright (c) 2004, Keir Fraser
+ */
+
+#include "common.h"
+#include <xen/balloon.h>
+#include <xen/gnttab.h>
+
+static kmem_cache_t *tpmif_cachep;
+int num_frontends = 0;
+
+LIST_HEAD(tpmif_list);
+
+static tpmif_t *alloc_tpmif(domid_t domid, struct backend_info *bi)
+{
+	tpmif_t *tpmif;
+
+	tpmif = kmem_cache_alloc(tpmif_cachep, GFP_KERNEL);
+	if (tpmif == NULL)
+		goto out_of_memory;
+
+	memset(tpmif, 0, sizeof (*tpmif));
+	tpmif->domid = domid;
+	tpmif->status = DISCONNECTED;
+	tpmif->bi = bi;
+	snprintf(tpmif->devname, sizeof(tpmif->devname), "tpmif%d", domid);
+	atomic_set(&tpmif->refcnt, 1);
+
+	tpmif->mmap_pages = alloc_empty_pages_and_pagevec(TPMIF_TX_RING_SIZE);
+	if (tpmif->mmap_pages == NULL)
+		goto out_of_memory;
+
+	list_add(&tpmif->tpmif_list, &tpmif_list);
+	num_frontends++;
+
+	return tpmif;
+
+ out_of_memory:
+	if (tpmif != NULL)
+		kmem_cache_free(tpmif_cachep, tpmif);
+	printk("%s: out of memory\n", __FUNCTION__);
+	return ERR_PTR(-ENOMEM);
+}
+
+static void free_tpmif(tpmif_t * tpmif)
+{
+	num_frontends--;
+	list_del(&tpmif->tpmif_list);
+	free_empty_pages_and_pagevec(tpmif->mmap_pages, TPMIF_TX_RING_SIZE);
+	kmem_cache_free(tpmif_cachep, tpmif);
+}
+
+tpmif_t *tpmif_find(domid_t domid, struct backend_info *bi)
+{
+	tpmif_t *tpmif;
+
+	list_for_each_entry(tpmif, &tpmif_list, tpmif_list) {
+		if (tpmif->bi == bi) {
+			if (tpmif->domid == domid) {
+				tpmif_get(tpmif);
+				return tpmif;
+			} else {
+				return ERR_PTR(-EEXIST);
+			}
+		}
+	}
+
+	return alloc_tpmif(domid, bi);
+}
+
+static int map_frontend_page(tpmif_t *tpmif, unsigned long shared_page)
+{
+	struct gnttab_map_grant_ref op;
+
+	gnttab_set_map_op(&op, (unsigned long)tpmif->tx_area->addr,
+			  GNTMAP_host_map, shared_page, tpmif->domid);
+
+	if (HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, &op, 1))
+		BUG();
+
+	if (op.status) {
+		DPRINTK(" Grant table operation failure !\n");
+		return op.status;
+	}
+
+	tpmif->shmem_ref = shared_page;
+	tpmif->shmem_handle = op.handle;
+
+	return 0;
+}
+
+static void unmap_frontend_page(tpmif_t *tpmif)
+{
+	struct gnttab_unmap_grant_ref op;
+
+	gnttab_set_unmap_op(&op, (unsigned long)tpmif->tx_area->addr,
+			    GNTMAP_host_map, tpmif->shmem_handle);
+
+	if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1))
+		BUG();
+}
+
+int tpmif_map(tpmif_t *tpmif, unsigned long shared_page, unsigned int evtchn)
+{
+	int err;
+
+	if (tpmif->irq)
+		return 0;
+
+	if ((tpmif->tx_area = alloc_vm_area(PAGE_SIZE)) == NULL)
+		return -ENOMEM;
+
+	err = map_frontend_page(tpmif, shared_page);
+	if (err) {
+		free_vm_area(tpmif->tx_area);
+		return err;
+	}
+
+	tpmif->tx = (tpmif_tx_interface_t *)tpmif->tx_area->addr;
+	memset(tpmif->tx, 0, PAGE_SIZE);
+
+	err = bind_interdomain_evtchn_to_irqhandler(
+		tpmif->domid, evtchn, tpmif_be_int, 0, tpmif->devname, tpmif);
+	if (err < 0) {
+		unmap_frontend_page(tpmif);
+		free_vm_area(tpmif->tx_area);
+		return err;
+	}
+	tpmif->irq = err;
+
+	tpmif->shmem_ref = shared_page;
+	tpmif->active = 1;
+
+	return 0;
+}
+
+void tpmif_disconnect_complete(tpmif_t *tpmif)
+{
+	if (tpmif->irq)
+		unbind_from_irqhandler(tpmif->irq, tpmif);
+
+	if (tpmif->tx) {
+		unmap_frontend_page(tpmif);
+		free_vm_area(tpmif->tx_area);
+	}
+
+	free_tpmif(tpmif);
+}
+
+void __init tpmif_interface_init(void)
+{
+	tpmif_cachep = kmem_cache_create("tpmif_cache", sizeof (tpmif_t),
+					 0, 0, NULL, NULL);
+}
+
+void __exit tpmif_interface_exit(void)
+{
+	kmem_cache_destroy(tpmif_cachep);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/tpmback/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/tpmback/Makefile
--- linux-2.6.18.8/drivers/xen/tpmback/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/tpmback/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,4 @@
+
+obj-$(CONFIG_XEN_TPMDEV_BACKEND)	+= tpmbk.o
+
+tpmbk-y += tpmback.o interface.o xenbus.o
diff -rpuN linux-2.6.18.8/drivers/xen/tpmback/tpmback.c linux-2.6.18-xen-3.3.0/drivers/xen/tpmback/tpmback.c
--- linux-2.6.18.8/drivers/xen/tpmback/tpmback.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/tpmback/tpmback.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,944 @@
+/******************************************************************************
+ * drivers/xen/tpmback/tpmback.c
+ *
+ * Copyright (c) 2005, IBM Corporation
+ *
+ * Author: Stefan Berger, stefanb@us.ibm.com
+ * Grant table support: Mahadevan Gomathisankaran
+ *
+ * This code has been derived from drivers/xen/netback/netback.c
+ * Copyright (c) 2002-2004, K A Fraser
+ *
+ */
+
+#include "common.h"
+#include <xen/evtchn.h>
+
+#include <linux/types.h>
+#include <linux/list.h>
+#include <linux/miscdevice.h>
+#include <linux/poll.h>
+#include <asm/uaccess.h>
+#include <xen/xenbus.h>
+#include <xen/interface/grant_table.h>
+#include <xen/gnttab.h>
+
+/* local data structures */
+struct data_exchange {
+	struct list_head pending_pak;
+	struct list_head current_pak;
+	unsigned int copied_so_far;
+	u8 has_opener:1;
+	u8 aborted:1;
+	rwlock_t pak_lock;	// protects all of the previous fields
+	wait_queue_head_t wait_queue;
+};
+
+struct vtpm_resp_hdr {
+	uint32_t instance_no;
+	uint16_t tag_no;
+	uint32_t len_no;
+	uint32_t ordinal_no;
+} __attribute__ ((packed));
+
+struct packet {
+	struct list_head next;
+	unsigned int data_len;
+	u8 *data_buffer;
+	tpmif_t *tpmif;
+	u32 tpm_instance;
+	u8 req_tag;
+	u32 last_read;
+	u8 flags;
+	struct timer_list processing_timer;
+};
+
+enum {
+	PACKET_FLAG_DISCARD_RESPONSE = 1,
+};
+
+/* local variables */
+static struct data_exchange dataex;
+
+/* local function prototypes */
+static int _packet_write(struct packet *pak,
+			 const char *data, size_t size, int userbuffer);
+static void processing_timeout(unsigned long ptr);
+static int packet_read_shmem(struct packet *pak,
+			     tpmif_t * tpmif,
+			     u32 offset,
+			     char *buffer, int isuserbuffer, u32 left);
+static int vtpm_queue_packet(struct packet *pak);
+
+/***************************************************************
+ Buffer copying fo user and kernel space buffes.
+***************************************************************/
+static inline int copy_from_buffer(void *to,
+				   const void *from, unsigned long size,
+				   int isuserbuffer)
+{
+	if (isuserbuffer) {
+		if (copy_from_user(to, (void __user *)from, size))
+			return -EFAULT;
+	} else {
+		memcpy(to, from, size);
+	}
+	return 0;
+}
+
+static inline int copy_to_buffer(void *to,
+				 const void *from, unsigned long size,
+				 int isuserbuffer)
+{
+	if (isuserbuffer) {
+		if (copy_to_user((void __user *)to, from, size))
+			return -EFAULT;
+	} else {
+		memcpy(to, from, size);
+	}
+	return 0;
+}
+
+
+static void dataex_init(struct data_exchange *dataex)
+{
+	INIT_LIST_HEAD(&dataex->pending_pak);
+	INIT_LIST_HEAD(&dataex->current_pak);
+	dataex->has_opener = 0;
+	rwlock_init(&dataex->pak_lock);
+	init_waitqueue_head(&dataex->wait_queue);
+}
+
+/***************************************************************
+ Packet-related functions
+***************************************************************/
+
+static struct packet *packet_find_instance(struct list_head *head,
+					   u32 tpm_instance)
+{
+	struct packet *pak;
+	struct list_head *p;
+
+	/*
+	 * traverse the list of packets and return the first
+	 * one with the given instance number
+	 */
+	list_for_each(p, head) {
+		pak = list_entry(p, struct packet, next);
+
+		if (pak->tpm_instance == tpm_instance) {
+			return pak;
+		}
+	}
+	return NULL;
+}
+
+static struct packet *packet_find_packet(struct list_head *head, void *packet)
+{
+	struct packet *pak;
+	struct list_head *p;
+
+	/*
+	 * traverse the list of packets and return the first
+	 * one with the given instance number
+	 */
+	list_for_each(p, head) {
+		pak = list_entry(p, struct packet, next);
+
+		if (pak == packet) {
+			return pak;
+		}
+	}
+	return NULL;
+}
+
+static struct packet *packet_alloc(tpmif_t * tpmif,
+				   u32 size, u8 req_tag, u8 flags)
+{
+	struct packet *pak = NULL;
+	pak = kzalloc(sizeof (struct packet), GFP_ATOMIC);
+	if (NULL != pak) {
+		if (tpmif) {
+			pak->tpmif = tpmif;
+			pak->tpm_instance = tpmback_get_instance(tpmif->bi);
+			tpmif_get(tpmif);
+		}
+		pak->data_len = size;
+		pak->req_tag = req_tag;
+		pak->last_read = 0;
+		pak->flags = flags;
+
+		/*
+		 * cannot do tpmif_get(tpmif); bad things happen
+		 * on the last tpmif_put()
+		 */
+		init_timer(&pak->processing_timer);
+		pak->processing_timer.function = processing_timeout;
+		pak->processing_timer.data = (unsigned long)pak;
+	}
+	return pak;
+}
+
+static void inline packet_reset(struct packet *pak)
+{
+	pak->last_read = 0;
+}
+
+static void packet_free(struct packet *pak)
+{
+	if (timer_pending(&pak->processing_timer)) {
+		BUG();
+	}
+
+	if (pak->tpmif)
+		tpmif_put(pak->tpmif);
+	kfree(pak->data_buffer);
+	/*
+	 * cannot do tpmif_put(pak->tpmif); bad things happen
+	 * on the last tpmif_put()
+	 */
+	kfree(pak);
+}
+
+
+/*
+ * Write data to the shared memory and send it to the FE.
+ */
+static int packet_write(struct packet *pak,
+			const char *data, size_t size, int isuserbuffer)
+{
+	int rc = 0;
+
+	if (0 != (pak->flags & PACKET_FLAG_DISCARD_RESPONSE)) {
+		/* Don't send a respone to this packet. Just acknowledge it. */
+		rc = size;
+	} else {
+		rc = _packet_write(pak, data, size, isuserbuffer);
+	}
+
+	return rc;
+}
+
+int _packet_write(struct packet *pak,
+		  const char *data, size_t size, int isuserbuffer)
+{
+	/*
+	 * Write into the shared memory pages directly
+	 * and send it to the front end.
+	 */
+	tpmif_t *tpmif = pak->tpmif;
+	grant_handle_t handle;
+	int rc = 0;
+	unsigned int i = 0;
+	unsigned int offset = 0;
+
+	if (tpmif == NULL) {
+		return -EFAULT;
+	}
+
+	if (tpmif->status == DISCONNECTED) {
+		return size;
+	}
+
+	while (offset < size && i < TPMIF_TX_RING_SIZE) {
+		unsigned int tocopy;
+		struct gnttab_map_grant_ref map_op;
+		struct gnttab_unmap_grant_ref unmap_op;
+		tpmif_tx_request_t *tx;
+
+		tx = &tpmif->tx->ring[i].req;
+
+		if (0 == tx->addr) {
+			DPRINTK("ERROR: Buffer for outgoing packet NULL?! i=%d\n", i);
+			return 0;
+		}
+
+		gnttab_set_map_op(&map_op, idx_to_kaddr(tpmif, i),
+				  GNTMAP_host_map, tx->ref, tpmif->domid);
+
+		if (unlikely(HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref,
+						       &map_op, 1))) {
+			BUG();
+		}
+
+		handle = map_op.handle;
+
+		if (map_op.status) {
+			DPRINTK(" Grant table operation failure !\n");
+			return 0;
+		}
+
+		tocopy = min_t(size_t, size - offset, PAGE_SIZE);
+
+		if (copy_from_buffer((void *)(idx_to_kaddr(tpmif, i) |
+					      (tx->addr & ~PAGE_MASK)),
+				     &data[offset], tocopy, isuserbuffer)) {
+			tpmif_put(tpmif);
+			return -EFAULT;
+		}
+		tx->size = tocopy;
+
+		gnttab_set_unmap_op(&unmap_op, idx_to_kaddr(tpmif, i),
+				    GNTMAP_host_map, handle);
+
+		if (unlikely
+		    (HYPERVISOR_grant_table_op
+		     (GNTTABOP_unmap_grant_ref, &unmap_op, 1))) {
+			BUG();
+		}
+
+		offset += tocopy;
+		i++;
+	}
+
+	rc = offset;
+	DPRINTK("Notifying frontend via irq %d\n", tpmif->irq);
+	notify_remote_via_irq(tpmif->irq);
+
+	return rc;
+}
+
+/*
+ * Read data from the shared memory and copy it directly into the
+ * provided buffer. Advance the read_last indicator which tells
+ * how many bytes have already been read.
+ */
+static int packet_read(struct packet *pak, size_t numbytes,
+		       char *buffer, size_t buffersize, int isuserbuffer)
+{
+	tpmif_t *tpmif = pak->tpmif;
+
+	/*
+	 * Read 'numbytes' of data from the buffer. The first 4
+	 * bytes are the instance number in network byte order,
+	 * after that come the data from the shared memory buffer.
+	 */
+	u32 to_copy;
+	u32 offset = 0;
+	u32 room_left = buffersize;
+
+	if (pak->last_read < 4) {
+		/*
+		 * copy the instance number into the buffer
+		 */
+		u32 instance_no = htonl(pak->tpm_instance);
+		u32 last_read = pak->last_read;
+
+		to_copy = min_t(size_t, 4 - last_read, numbytes);
+
+		if (copy_to_buffer(&buffer[0],
+				   &(((u8 *) & instance_no)[last_read]),
+				   to_copy, isuserbuffer)) {
+			return -EFAULT;
+		}
+
+		pak->last_read += to_copy;
+		offset += to_copy;
+		room_left -= to_copy;
+	}
+
+	/*
+	 * If the packet has a data buffer appended, read from it...
+	 */
+
+	if (room_left > 0) {
+		if (pak->data_buffer) {
+			u32 to_copy = min_t(u32, pak->data_len - offset, room_left);
+			u32 last_read = pak->last_read - 4;
+
+			if (copy_to_buffer(&buffer[offset],
+					   &pak->data_buffer[last_read],
+					   to_copy, isuserbuffer)) {
+				return -EFAULT;
+			}
+			pak->last_read += to_copy;
+			offset += to_copy;
+		} else {
+			offset = packet_read_shmem(pak,
+						   tpmif,
+						   offset,
+						   buffer,
+						   isuserbuffer, room_left);
+		}
+	}
+	return offset;
+}
+
+static int packet_read_shmem(struct packet *pak,
+			     tpmif_t * tpmif,
+			     u32 offset, char *buffer, int isuserbuffer,
+			     u32 room_left)
+{
+	u32 last_read = pak->last_read - 4;
+	u32 i = (last_read / PAGE_SIZE);
+	u32 pg_offset = last_read & (PAGE_SIZE - 1);
+	u32 to_copy;
+	grant_handle_t handle;
+
+	tpmif_tx_request_t *tx;
+
+	tx = &tpmif->tx->ring[0].req;
+	/*
+	 * Start copying data at the page with index 'index'
+	 * and within that page at offset 'offset'.
+	 * Copy a maximum of 'room_left' bytes.
+	 */
+	to_copy = min_t(u32, PAGE_SIZE - pg_offset, room_left);
+	while (to_copy > 0) {
+		void *src;
+		struct gnttab_map_grant_ref map_op;
+		struct gnttab_unmap_grant_ref unmap_op;
+
+		tx = &tpmif->tx->ring[i].req;
+
+		gnttab_set_map_op(&map_op, idx_to_kaddr(tpmif, i),
+				  GNTMAP_host_map, tx->ref, tpmif->domid);
+
+		if (unlikely(HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref,
+						       &map_op, 1))) {
+			BUG();
+		}
+
+		if (map_op.status) {
+			DPRINTK(" Grant table operation failure !\n");
+			return -EFAULT;
+		}
+
+		handle = map_op.handle;
+
+		if (to_copy > tx->size) {
+			/*
+			 * User requests more than what's available
+			 */
+			to_copy = min_t(u32, tx->size, to_copy);
+		}
+
+		DPRINTK("Copying from mapped memory at %08lx\n",
+			(unsigned long)(idx_to_kaddr(tpmif, i) |
+					(tx->addr & ~PAGE_MASK)));
+
+		src = (void *)(idx_to_kaddr(tpmif, i) |
+			       ((tx->addr & ~PAGE_MASK) + pg_offset));
+		if (copy_to_buffer(&buffer[offset],
+				   src, to_copy, isuserbuffer)) {
+			return -EFAULT;
+		}
+
+		DPRINTK("Data from TPM-FE of domain %d are %d %d %d %d\n",
+			tpmif->domid, buffer[offset], buffer[offset + 1],
+			buffer[offset + 2], buffer[offset + 3]);
+
+		gnttab_set_unmap_op(&unmap_op, idx_to_kaddr(tpmif, i),
+				    GNTMAP_host_map, handle);
+
+		if (unlikely
+		    (HYPERVISOR_grant_table_op
+		     (GNTTABOP_unmap_grant_ref, &unmap_op, 1))) {
+			BUG();
+		}
+
+		offset += to_copy;
+		pg_offset = 0;
+		last_read += to_copy;
+		room_left -= to_copy;
+
+		to_copy = min_t(u32, PAGE_SIZE, room_left);
+		i++;
+	}			/* while (to_copy > 0) */
+	/*
+	 * Adjust the last_read pointer
+	 */
+	pak->last_read = last_read + 4;
+	return offset;
+}
+
+/* ============================================================
+ * The file layer for reading data from this device
+ * ============================================================
+ */
+static int vtpm_op_open(struct inode *inode, struct file *f)
+{
+	int rc = 0;
+	unsigned long flags;
+
+	write_lock_irqsave(&dataex.pak_lock, flags);
+	if (dataex.has_opener == 0) {
+		dataex.has_opener = 1;
+	} else {
+		rc = -EPERM;
+	}
+	write_unlock_irqrestore(&dataex.pak_lock, flags);
+	return rc;
+}
+
+static ssize_t vtpm_op_read(struct file *file,
+			    char __user * data, size_t size, loff_t * offset)
+{
+	int ret_size = -ENODATA;
+	struct packet *pak = NULL;
+	unsigned long flags;
+
+	write_lock_irqsave(&dataex.pak_lock, flags);
+	if (dataex.aborted) {
+		dataex.aborted = 0;
+		dataex.copied_so_far = 0;
+		write_unlock_irqrestore(&dataex.pak_lock, flags);
+		return -EIO;
+	}
+
+	if (list_empty(&dataex.pending_pak)) {
+		write_unlock_irqrestore(&dataex.pak_lock, flags);
+		wait_event_interruptible(dataex.wait_queue,
+					 !list_empty(&dataex.pending_pak));
+		write_lock_irqsave(&dataex.pak_lock, flags);
+		dataex.copied_so_far = 0;
+	}
+
+	if (!list_empty(&dataex.pending_pak)) {
+		unsigned int left;
+
+		pak = list_entry(dataex.pending_pak.next, struct packet, next);
+		left = pak->data_len - dataex.copied_so_far;
+		list_del(&pak->next);
+		write_unlock_irqrestore(&dataex.pak_lock, flags);
+
+		DPRINTK("size given by app: %d, available: %d\n", size, left);
+
+		ret_size = min_t(size_t, size, left);
+
+		ret_size = packet_read(pak, ret_size, data, size, 1);
+
+		write_lock_irqsave(&dataex.pak_lock, flags);
+
+		if (ret_size < 0) {
+			del_singleshot_timer_sync(&pak->processing_timer);
+			packet_free(pak);
+			dataex.copied_so_far = 0;
+		} else {
+			DPRINTK("Copied %d bytes to user buffer\n", ret_size);
+
+			dataex.copied_so_far += ret_size;
+			if (dataex.copied_so_far >= pak->data_len + 4) {
+				DPRINTK("All data from this packet given to app.\n");
+				/* All data given to app */
+
+				del_singleshot_timer_sync(&pak->
+							  processing_timer);
+				list_add_tail(&pak->next, &dataex.current_pak);
+				/*
+				 * The more fontends that are handled at the same time,
+				 * the more time we give the TPM to process the request.
+				 */
+				mod_timer(&pak->processing_timer,
+					  jiffies + (num_frontends * 60 * HZ));
+				dataex.copied_so_far = 0;
+			} else {
+				list_add(&pak->next, &dataex.pending_pak);
+			}
+		}
+	}
+	write_unlock_irqrestore(&dataex.pak_lock, flags);
+
+	DPRINTK("Returning result from read to app: %d\n", ret_size);
+
+	return ret_size;
+}
+
+/*
+ * Write operation - only works after a previous read operation!
+ */
+static ssize_t vtpm_op_write(struct file *file,
+			     const char __user * data, size_t size,
+			     loff_t * offset)
+{
+	struct packet *pak;
+	int rc = 0;
+	unsigned int off = 4;
+	unsigned long flags;
+	struct vtpm_resp_hdr vrh;
+
+	/*
+	 * Minimum required packet size is:
+	 * 4 bytes for instance number
+	 * 2 bytes for tag
+	 * 4 bytes for paramSize
+	 * 4 bytes for the ordinal
+	 * sum: 14 bytes
+	 */
+	if (size < sizeof (vrh))
+		return -EFAULT;
+
+	if (copy_from_user(&vrh, data, sizeof (vrh)))
+		return -EFAULT;
+
+	/* malformed packet? */
+	if ((off + ntohl(vrh.len_no)) != size)
+		return -EFAULT;
+
+	write_lock_irqsave(&dataex.pak_lock, flags);
+	pak = packet_find_instance(&dataex.current_pak,
+				   ntohl(vrh.instance_no));
+
+	if (pak == NULL) {
+		write_unlock_irqrestore(&dataex.pak_lock, flags);
+		DPRINTK(KERN_ALERT "No associated packet! (inst=%d)\n",
+		        ntohl(vrh.instance_no));
+		return -EFAULT;
+	}
+
+	del_singleshot_timer_sync(&pak->processing_timer);
+	list_del(&pak->next);
+
+	write_unlock_irqrestore(&dataex.pak_lock, flags);
+
+	/*
+	 * The first 'offset' bytes must be the instance number - skip them.
+	 */
+	size -= off;
+
+	rc = packet_write(pak, &data[off], size, 1);
+
+	if (rc > 0) {
+		/* I neglected the first 4 bytes */
+		rc += off;
+	}
+	packet_free(pak);
+	return rc;
+}
+
+static int vtpm_op_release(struct inode *inode, struct file *file)
+{
+	unsigned long flags;
+
+	vtpm_release_packets(NULL, 1);
+	write_lock_irqsave(&dataex.pak_lock, flags);
+	dataex.has_opener = 0;
+	write_unlock_irqrestore(&dataex.pak_lock, flags);
+	return 0;
+}
+
+static unsigned int vtpm_op_poll(struct file *file,
+				 struct poll_table_struct *pts)
+{
+	unsigned int flags = POLLOUT | POLLWRNORM;
+
+	poll_wait(file, &dataex.wait_queue, pts);
+	if (!list_empty(&dataex.pending_pak)) {
+		flags |= POLLIN | POLLRDNORM;
+	}
+	return flags;
+}
+
+static const struct file_operations vtpm_ops = {
+	.owner = THIS_MODULE,
+	.llseek = no_llseek,
+	.open = vtpm_op_open,
+	.read = vtpm_op_read,
+	.write = vtpm_op_write,
+	.release = vtpm_op_release,
+	.poll = vtpm_op_poll,
+};
+
+static struct miscdevice vtpms_miscdevice = {
+	.minor = 225,
+	.name = "vtpm",
+	.fops = &vtpm_ops,
+};
+
+/***************************************************************
+ Utility functions
+***************************************************************/
+
+static int tpm_send_fail_message(struct packet *pak, u8 req_tag)
+{
+	int rc;
+	static const unsigned char tpm_error_message_fail[] = {
+		0x00, 0x00,
+		0x00, 0x00, 0x00, 0x0a,
+		0x00, 0x00, 0x00, 0x09	/* TPM_FAIL */
+	};
+	unsigned char buffer[sizeof (tpm_error_message_fail)];
+
+	memcpy(buffer, tpm_error_message_fail,
+	       sizeof (tpm_error_message_fail));
+	/*
+	 * Insert the right response tag depending on the given tag
+	 * All response tags are '+3' to the request tag.
+	 */
+	buffer[1] = req_tag + 3;
+
+	/*
+	 * Write the data to shared memory and notify the front-end
+	 */
+	rc = packet_write(pak, buffer, sizeof (buffer), 0);
+
+	return rc;
+}
+
+static int _vtpm_release_packets(struct list_head *head,
+				 tpmif_t * tpmif, int send_msgs)
+{
+	int aborted = 0;
+	int c = 0;
+	struct packet *pak;
+	struct list_head *pos, *tmp;
+
+	list_for_each_safe(pos, tmp, head) {
+		pak = list_entry(pos, struct packet, next);
+		c += 1;
+
+		if (tpmif == NULL || pak->tpmif == tpmif) {
+			int can_send = 0;
+
+			del_singleshot_timer_sync(&pak->processing_timer);
+			list_del(&pak->next);
+
+			if (pak->tpmif && pak->tpmif->status == CONNECTED) {
+				can_send = 1;
+			}
+
+			if (send_msgs && can_send) {
+				tpm_send_fail_message(pak, pak->req_tag);
+			}
+			packet_free(pak);
+			if (c == 1)
+				aborted = 1;
+		}
+	}
+	return aborted;
+}
+
+int vtpm_release_packets(tpmif_t * tpmif, int send_msgs)
+{
+	unsigned long flags;
+
+	write_lock_irqsave(&dataex.pak_lock, flags);
+
+	dataex.aborted = _vtpm_release_packets(&dataex.pending_pak,
+					       tpmif,
+					       send_msgs);
+	_vtpm_release_packets(&dataex.current_pak, tpmif, send_msgs);
+
+	write_unlock_irqrestore(&dataex.pak_lock, flags);
+	return 0;
+}
+
+static int vtpm_queue_packet(struct packet *pak)
+{
+	int rc = 0;
+
+	if (dataex.has_opener) {
+		unsigned long flags;
+
+		write_lock_irqsave(&dataex.pak_lock, flags);
+		list_add_tail(&pak->next, &dataex.pending_pak);
+		/* give the TPM some time to pick up the request */
+		mod_timer(&pak->processing_timer, jiffies + (30 * HZ));
+		write_unlock_irqrestore(&dataex.pak_lock, flags);
+
+		wake_up_interruptible(&dataex.wait_queue);
+	} else {
+		rc = -EFAULT;
+	}
+	return rc;
+}
+
+static int vtpm_receive(tpmif_t * tpmif, u32 size)
+{
+	int rc = 0;
+	unsigned char buffer[10];
+	__be32 *native_size;
+	struct packet *pak = packet_alloc(tpmif, size, 0, 0);
+
+	if (!pak)
+		return -ENOMEM;
+	/*
+	 * Read 10 bytes from the received buffer to test its
+	 * content for validity.
+	 */
+	if (sizeof (buffer) != packet_read(pak,
+					   sizeof (buffer), buffer,
+					   sizeof (buffer), 0)) {
+		goto failexit;
+	}
+	/*
+	 * Reset the packet read pointer so we can read all its
+	 * contents again.
+	 */
+	packet_reset(pak);
+
+	native_size = (__force __be32 *) (&buffer[4 + 2]);
+	/*
+	 * Verify that the size of the packet is correct
+	 * as indicated and that there's actually someone reading packets.
+	 * The minimum size of the packet is '10' for tag, size indicator
+	 * and ordinal.
+	 */
+	if (size < 10 ||
+	    be32_to_cpu(*native_size) != size ||
+	    0 == dataex.has_opener || tpmif->status != CONNECTED) {
+		rc = -EINVAL;
+		goto failexit;
+	} else {
+		rc = vtpm_queue_packet(pak);
+		if (rc < 0)
+			goto failexit;
+	}
+	return 0;
+
+      failexit:
+	if (pak) {
+		tpm_send_fail_message(pak, buffer[4 + 1]);
+		packet_free(pak);
+	}
+	return rc;
+}
+
+/*
+ * Timeout function that gets invoked when a packet has not been processed
+ * during the timeout period.
+ * The packet must be on a list when this function is invoked. This
+ * also means that once its taken off a list, the timer must be
+ * destroyed as well.
+ */
+static void processing_timeout(unsigned long ptr)
+{
+	struct packet *pak = (struct packet *)ptr;
+	unsigned long flags;
+
+	write_lock_irqsave(&dataex.pak_lock, flags);
+	/*
+	 * The packet needs to be searched whether it
+	 * is still on the list.
+	 */
+	if (pak == packet_find_packet(&dataex.pending_pak, pak) ||
+	    pak == packet_find_packet(&dataex.current_pak, pak)) {
+		if ((pak->flags & PACKET_FLAG_DISCARD_RESPONSE) == 0) {
+			tpm_send_fail_message(pak, pak->req_tag);
+		}
+		/* discard future responses */
+		pak->flags |= PACKET_FLAG_DISCARD_RESPONSE;
+	}
+
+	write_unlock_irqrestore(&dataex.pak_lock, flags);
+}
+
+static void tpm_tx_action(unsigned long unused);
+static DECLARE_TASKLET(tpm_tx_tasklet, tpm_tx_action, 0);
+
+static struct list_head tpm_schedule_list;
+static spinlock_t tpm_schedule_list_lock;
+
+static inline void maybe_schedule_tx_action(void)
+{
+	smp_mb();
+	tasklet_schedule(&tpm_tx_tasklet);
+}
+
+static inline int __on_tpm_schedule_list(tpmif_t * tpmif)
+{
+	return tpmif->list.next != NULL;
+}
+
+static void remove_from_tpm_schedule_list(tpmif_t * tpmif)
+{
+	spin_lock_irq(&tpm_schedule_list_lock);
+	if (likely(__on_tpm_schedule_list(tpmif))) {
+		list_del(&tpmif->list);
+		tpmif->list.next = NULL;
+		tpmif_put(tpmif);
+	}
+	spin_unlock_irq(&tpm_schedule_list_lock);
+}
+
+static void add_to_tpm_schedule_list_tail(tpmif_t * tpmif)
+{
+	if (__on_tpm_schedule_list(tpmif))
+		return;
+
+	spin_lock_irq(&tpm_schedule_list_lock);
+	if (!__on_tpm_schedule_list(tpmif) && tpmif->active) {
+		list_add_tail(&tpmif->list, &tpm_schedule_list);
+		tpmif_get(tpmif);
+	}
+	spin_unlock_irq(&tpm_schedule_list_lock);
+}
+
+void tpmif_schedule_work(tpmif_t * tpmif)
+{
+	add_to_tpm_schedule_list_tail(tpmif);
+	maybe_schedule_tx_action();
+}
+
+void tpmif_deschedule_work(tpmif_t * tpmif)
+{
+	remove_from_tpm_schedule_list(tpmif);
+}
+
+static void tpm_tx_action(unsigned long unused)
+{
+	struct list_head *ent;
+	tpmif_t *tpmif;
+	tpmif_tx_request_t *tx;
+
+	DPRINTK("%s: Getting data from front-end(s)!\n", __FUNCTION__);
+
+	while (!list_empty(&tpm_schedule_list)) {
+		/* Get a tpmif from the list with work to do. */
+		ent = tpm_schedule_list.next;
+		tpmif = list_entry(ent, tpmif_t, list);
+		tpmif_get(tpmif);
+		remove_from_tpm_schedule_list(tpmif);
+
+		tx = &tpmif->tx->ring[0].req;
+
+		/* pass it up */
+		vtpm_receive(tpmif, tx->size);
+
+		tpmif_put(tpmif);
+	}
+}
+
+irqreturn_t tpmif_be_int(int irq, void *dev_id, struct pt_regs *regs)
+{
+	tpmif_t *tpmif = (tpmif_t *) dev_id;
+
+	add_to_tpm_schedule_list_tail(tpmif);
+	maybe_schedule_tx_action();
+	return IRQ_HANDLED;
+}
+
+static int __init tpmback_init(void)
+{
+	int rc;
+
+	if ((rc = misc_register(&vtpms_miscdevice)) != 0) {
+		printk(KERN_ALERT
+		       "Could not register misc device for TPM BE.\n");
+		return rc;
+	}
+
+	dataex_init(&dataex);
+
+	spin_lock_init(&tpm_schedule_list_lock);
+	INIT_LIST_HEAD(&tpm_schedule_list);
+
+	tpmif_interface_init();
+	tpmif_xenbus_init();
+
+	printk(KERN_ALERT "Successfully initialized TPM backend driver.\n");
+
+	return 0;
+}
+
+module_init(tpmback_init);
+
+void __exit tpmback_exit(void)
+{
+	vtpm_release_packets(NULL, 0);
+	tpmif_xenbus_exit();
+	tpmif_interface_exit();
+	misc_deregister(&vtpms_miscdevice);
+}
+
+MODULE_LICENSE("Dual BSD/GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/tpmback/xenbus.c linux-2.6.18-xen-3.3.0/drivers/xen/tpmback/xenbus.c
--- linux-2.6.18.8/drivers/xen/tpmback/xenbus.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/tpmback/xenbus.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,289 @@
+/*  Xenbus code for tpmif backend
+    Copyright (C) 2005 IBM Corporation
+    Copyright (C) 2005 Rusty Russell <rusty@rustcorp.com.au>
+
+    This program is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program; if not, write to the Free Software
+    Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+*/
+#include <stdarg.h>
+#include <linux/module.h>
+#include <xen/xenbus.h>
+#include "common.h"
+
+struct backend_info
+{
+	struct xenbus_device *dev;
+
+	/* our communications channel */
+	tpmif_t *tpmif;
+
+	long int frontend_id;
+	long int instance; // instance of TPM
+	u8 is_instance_set;// whether instance number has been set
+
+	/* watch front end for changes */
+	struct xenbus_watch backend_watch;
+};
+
+static void maybe_connect(struct backend_info *be);
+static void connect(struct backend_info *be);
+static int connect_ring(struct backend_info *be);
+static void backend_changed(struct xenbus_watch *watch,
+			    const char **vec, unsigned int len);
+static void frontend_changed(struct xenbus_device *dev,
+			     enum xenbus_state frontend_state);
+
+long int tpmback_get_instance(struct backend_info *bi)
+{
+	long int res = -1;
+	if (bi && bi->is_instance_set)
+		res = bi->instance;
+	return res;
+}
+
+static int tpmback_remove(struct xenbus_device *dev)
+{
+	struct backend_info *be = dev->dev.driver_data;
+
+	if (!be) return 0;
+
+	if (be->backend_watch.node) {
+		unregister_xenbus_watch(&be->backend_watch);
+		kfree(be->backend_watch.node);
+		be->backend_watch.node = NULL;
+	}
+	if (be->tpmif) {
+		be->tpmif->bi = NULL;
+		vtpm_release_packets(be->tpmif, 0);
+		tpmif_put(be->tpmif);
+		be->tpmif = NULL;
+	}
+	kfree(be);
+	dev->dev.driver_data = NULL;
+	return 0;
+}
+
+static int tpmback_probe(struct xenbus_device *dev,
+			 const struct xenbus_device_id *id)
+{
+	int err;
+	struct backend_info *be = kzalloc(sizeof(struct backend_info),
+					  GFP_KERNEL);
+
+	if (!be) {
+		xenbus_dev_fatal(dev, -ENOMEM,
+				 "allocating backend structure");
+		return -ENOMEM;
+	}
+
+	be->is_instance_set = 0;
+	be->dev = dev;
+	dev->dev.driver_data = be;
+
+	err = xenbus_watch_path2(dev, dev->nodename,
+				 "instance", &be->backend_watch,
+				 backend_changed);
+	if (err) {
+		goto fail;
+	}
+
+	err = xenbus_switch_state(dev, XenbusStateInitWait);
+	if (err) {
+		goto fail;
+	}
+	return 0;
+fail:
+	tpmback_remove(dev);
+	return err;
+}
+
+
+static void backend_changed(struct xenbus_watch *watch,
+			    const char **vec, unsigned int len)
+{
+	int err;
+	long instance;
+	struct backend_info *be
+		= container_of(watch, struct backend_info, backend_watch);
+	struct xenbus_device *dev = be->dev;
+
+	err = xenbus_scanf(XBT_NIL, dev->nodename,
+			   "instance","%li", &instance);
+	if (XENBUS_EXIST_ERR(err)) {
+		return;
+	}
+
+	if (err != 1) {
+		xenbus_dev_fatal(dev, err, "reading instance");
+		return;
+	}
+
+	if (be->is_instance_set == 0) {
+		be->instance = instance;
+		be->is_instance_set = 1;
+	}
+}
+
+
+static void frontend_changed(struct xenbus_device *dev,
+			     enum xenbus_state frontend_state)
+{
+	struct backend_info *be = dev->dev.driver_data;
+	int err;
+
+	switch (frontend_state) {
+	case XenbusStateInitialising:
+	case XenbusStateInitialised:
+		break;
+
+	case XenbusStateConnected:
+		err = connect_ring(be);
+		if (err) {
+			return;
+		}
+		maybe_connect(be);
+		break;
+
+	case XenbusStateClosing:
+		be->instance = -1;
+		xenbus_switch_state(dev, XenbusStateClosing);
+		break;
+
+	case XenbusStateUnknown: /* keep it here */
+	case XenbusStateClosed:
+		xenbus_switch_state(dev, XenbusStateClosed);
+		device_unregister(&be->dev->dev);
+		tpmback_remove(dev);
+		break;
+
+	default:
+		xenbus_dev_fatal(dev, -EINVAL,
+				 "saw state %d at frontend",
+				 frontend_state);
+		break;
+	}
+}
+
+
+
+static void maybe_connect(struct backend_info *be)
+{
+	if (be->tpmif == NULL || be->tpmif->status == CONNECTED)
+		return;
+
+	connect(be);
+}
+
+
+static void connect(struct backend_info *be)
+{
+	struct xenbus_transaction xbt;
+	int err;
+	struct xenbus_device *dev = be->dev;
+	unsigned long ready = 1;
+
+again:
+	err = xenbus_transaction_start(&xbt);
+	if (err) {
+		xenbus_dev_fatal(be->dev, err, "starting transaction");
+		return;
+	}
+
+	err = xenbus_printf(xbt, be->dev->nodename,
+			    "ready", "%lu", ready);
+	if (err) {
+		xenbus_dev_fatal(be->dev, err, "writing 'ready'");
+		goto abort;
+	}
+
+	err = xenbus_transaction_end(xbt, 0);
+	if (err == -EAGAIN)
+		goto again;
+	if (err)
+		xenbus_dev_fatal(be->dev, err, "end of transaction");
+
+	err = xenbus_switch_state(dev, XenbusStateConnected);
+	if (!err)
+		be->tpmif->status = CONNECTED;
+	return;
+abort:
+	xenbus_transaction_end(xbt, 1);
+}
+
+
+static int connect_ring(struct backend_info *be)
+{
+	struct xenbus_device *dev = be->dev;
+	unsigned long ring_ref;
+	unsigned int evtchn;
+	int err;
+
+	err = xenbus_gather(XBT_NIL, dev->otherend,
+			    "ring-ref", "%lu", &ring_ref,
+			    "event-channel", "%u", &evtchn, NULL);
+	if (err) {
+		xenbus_dev_error(dev, err,
+				 "reading %s/ring-ref and event-channel",
+				 dev->otherend);
+		return err;
+	}
+
+	if (!be->tpmif) {
+		be->tpmif = tpmif_find(dev->otherend_id, be);
+		if (IS_ERR(be->tpmif)) {
+			err = PTR_ERR(be->tpmif);
+			be->tpmif = NULL;
+			xenbus_dev_fatal(dev,err,"creating vtpm interface");
+			return err;
+		}
+	}
+
+	if (be->tpmif != NULL) {
+		err = tpmif_map(be->tpmif, ring_ref, evtchn);
+		if (err) {
+			xenbus_dev_error(dev, err,
+					 "mapping shared-frame %lu port %u",
+					 ring_ref, evtchn);
+			return err;
+		}
+	}
+	return 0;
+}
+
+
+static const struct xenbus_device_id tpmback_ids[] = {
+	{ "vtpm" },
+	{ "" }
+};
+
+
+static struct xenbus_driver tpmback = {
+	.name = "vtpm",
+	.owner = THIS_MODULE,
+	.ids = tpmback_ids,
+	.probe = tpmback_probe,
+	.remove = tpmback_remove,
+	.otherend_changed = frontend_changed,
+};
+
+
+void tpmif_xenbus_init(void)
+{
+	xenbus_register_backend(&tpmback);
+}
+
+void tpmif_xenbus_exit(void)
+{
+	xenbus_unregister_driver(&tpmback);
+}
diff -rpuN linux-2.6.18.8/drivers/xen/util.c linux-2.6.18-xen-3.3.0/drivers/xen/util.c
--- linux-2.6.18.8/drivers/xen/util.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/util.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,65 @@
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <asm/uaccess.h>
+#include <xen/driver_util.h>
+
+struct class *get_xen_class(void)
+{
+	static struct class *xen_class;
+
+	if (xen_class)
+		return xen_class;
+
+	xen_class = class_create(THIS_MODULE, "xen");
+	if (IS_ERR(xen_class)) {
+		printk("Failed to create xen sysfs class.\n");
+		xen_class = NULL;
+	}
+
+	return xen_class;
+}
+EXPORT_SYMBOL_GPL(get_xen_class);
+
+#ifdef CONFIG_X86
+static int f(pte_t *pte, struct page *pmd_page, unsigned long addr, void *data)
+{
+	/* apply_to_page_range() does all the hard work. */
+	return 0;
+}
+
+struct vm_struct *alloc_vm_area(unsigned long size)
+{
+	struct vm_struct *area;
+
+	area = get_vm_area(size, VM_IOREMAP);
+	if (area == NULL)
+		return NULL;
+
+	/*
+	 * This ensures that page tables are constructed for this region
+	 * of kernel virtual address space and mapped into init_mm.
+	 */
+	if (apply_to_page_range(&init_mm, (unsigned long)area->addr,
+				area->size, f, NULL)) {
+		free_vm_area(area);
+		return NULL;
+	}
+
+	/* Map page directories into every address space. */
+	vmalloc_sync_all();
+
+	return area;
+}
+EXPORT_SYMBOL_GPL(alloc_vm_area);
+
+void free_vm_area(struct vm_struct *area)
+{
+	struct vm_struct *ret;
+	ret = remove_vm_area(area->addr);
+	BUG_ON(ret != area);
+	kfree(area);
+}
+EXPORT_SYMBOL_GPL(free_vm_area);
+#endif /* CONFIG_X86 */
diff -rpuN linux-2.6.18.8/drivers/xen/xenbus/Makefile linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/Makefile
--- linux-2.6.18.8/drivers/xen/xenbus/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,9 @@
+obj-y += xenbus_client.o xenbus_comms.o xenbus_xs.o xenbus_probe.o
+obj-$(CONFIG_XEN_BACKEND) += xenbus_be.o
+
+xenbus_be-objs =
+xenbus_be-objs += xenbus_backend_client.o
+
+xenbus-$(CONFIG_XEN_BACKEND) += xenbus_probe_backend.o
+obj-y += $(xenbus-y) $(xenbus-m)
+obj-$(CONFIG_XEN_XENBUS_DEV) += xenbus_dev.o
diff -rpuN linux-2.6.18.8/drivers/xen/xenbus/xenbus_backend_client.c linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_backend_client.c
--- linux-2.6.18.8/drivers/xen/xenbus/xenbus_backend_client.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_backend_client.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,147 @@
+/******************************************************************************
+ * Backend-client-facing interface for the Xenbus driver.  In other words, the
+ * interface between the Xenbus and the device-specific code in the backend
+ * driver.
+ *
+ * Copyright (C) 2005-2006 XenSource Ltd
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/err.h>
+#include <xen/gnttab.h>
+#include <xen/xenbus.h>
+#include <xen/driver_util.h>
+
+/* Based on Rusty Russell's skeleton driver's map_page */
+struct vm_struct *xenbus_map_ring_valloc(struct xenbus_device *dev, int gnt_ref)
+{
+	struct gnttab_map_grant_ref op;
+	struct vm_struct *area;
+
+	area = alloc_vm_area(PAGE_SIZE);
+	if (!area)
+		return ERR_PTR(-ENOMEM);
+
+	gnttab_set_map_op(&op, (unsigned long)area->addr, GNTMAP_host_map,
+			  gnt_ref, dev->otherend_id);
+	
+	if (HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, &op, 1))
+		BUG();
+
+	if (op.status != GNTST_okay) {
+		free_vm_area(area);
+		xenbus_dev_fatal(dev, op.status,
+				 "mapping in shared page %d from domain %d",
+				 gnt_ref, dev->otherend_id);
+		BUG_ON(!IS_ERR(ERR_PTR(op.status)));
+		return ERR_PTR(op.status);
+	}
+
+	/* Stuff the handle in an unused field */
+	area->phys_addr = (unsigned long)op.handle;
+
+	return area;
+}
+EXPORT_SYMBOL_GPL(xenbus_map_ring_valloc);
+
+
+int xenbus_map_ring(struct xenbus_device *dev, int gnt_ref,
+		   grant_handle_t *handle, void *vaddr)
+{
+	struct gnttab_map_grant_ref op;
+	
+	gnttab_set_map_op(&op, (unsigned long)vaddr, GNTMAP_host_map,
+			  gnt_ref, dev->otherend_id);
+	if (HYPERVISOR_grant_table_op(GNTTABOP_map_grant_ref, &op, 1))
+		BUG();
+
+	if (op.status != GNTST_okay) {
+		xenbus_dev_fatal(dev, op.status,
+				 "mapping in shared page %d from domain %d",
+				 gnt_ref, dev->otherend_id);
+	} else
+		*handle = op.handle;
+
+	return op.status;
+}
+EXPORT_SYMBOL_GPL(xenbus_map_ring);
+
+
+/* Based on Rusty Russell's skeleton driver's unmap_page */
+int xenbus_unmap_ring_vfree(struct xenbus_device *dev, struct vm_struct *area)
+{
+	struct gnttab_unmap_grant_ref op;
+
+	gnttab_set_unmap_op(&op, (unsigned long)area->addr, GNTMAP_host_map,
+			    (grant_handle_t)area->phys_addr);
+
+	if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1))
+		BUG();
+
+	if (op.status == GNTST_okay)
+		free_vm_area(area);
+	else
+		xenbus_dev_error(dev, op.status,
+				 "unmapping page at handle %d error %d",
+				 (int16_t)area->phys_addr, op.status);
+
+	return op.status;
+}
+EXPORT_SYMBOL_GPL(xenbus_unmap_ring_vfree);
+
+
+int xenbus_unmap_ring(struct xenbus_device *dev,
+		     grant_handle_t handle, void *vaddr)
+{
+	struct gnttab_unmap_grant_ref op;
+
+	gnttab_set_unmap_op(&op, (unsigned long)vaddr, GNTMAP_host_map,
+			    handle);
+	if (HYPERVISOR_grant_table_op(GNTTABOP_unmap_grant_ref, &op, 1))
+		BUG();
+
+	if (op.status != GNTST_okay)
+		xenbus_dev_error(dev, op.status,
+				 "unmapping page at handle %d error %d",
+				 handle, op.status);
+
+	return op.status;
+}
+EXPORT_SYMBOL_GPL(xenbus_unmap_ring);
+
+int xenbus_dev_is_online(struct xenbus_device *dev)
+{
+	int rc, val;
+
+	rc = xenbus_scanf(XBT_NIL, dev->nodename, "online", "%d", &val);
+	if (rc != 1)
+		val = 0; /* no online node present */
+
+	return val;
+}
+EXPORT_SYMBOL_GPL(xenbus_dev_is_online);
+
+MODULE_LICENSE("Dual BSD/GPL");
diff -rpuN linux-2.6.18.8/drivers/xen/xenbus/xenbus_client.c linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_client.c
--- linux-2.6.18.8/drivers/xen/xenbus/xenbus_client.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_client.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,284 @@
+/******************************************************************************
+ * Client-facing interface for the Xenbus driver.  In other words, the
+ * interface between the Xenbus and the device-specific code, be it the
+ * frontend or the backend of that driver.
+ *
+ * Copyright (C) 2005 XenSource Ltd
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/slab.h>
+#include <xen/evtchn.h>
+#include <xen/gnttab.h>
+#include <xen/xenbus.h>
+#include <xen/driver_util.h>
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+#define DPRINTK(fmt, args...) \
+    pr_debug("xenbus_client (%s:%d) " fmt ".\n", __FUNCTION__, __LINE__, ##args)
+
+const char *xenbus_strstate(enum xenbus_state state)
+{
+	static const char *const name[] = {
+		[ XenbusStateUnknown      ] = "Unknown",
+		[ XenbusStateInitialising ] = "Initialising",
+		[ XenbusStateInitWait     ] = "InitWait",
+		[ XenbusStateInitialised  ] = "Initialised",
+		[ XenbusStateConnected    ] = "Connected",
+		[ XenbusStateClosing      ] = "Closing",
+		[ XenbusStateClosed	  ] = "Closed",
+	};
+	return (state < ARRAY_SIZE(name)) ? name[state] : "INVALID";
+}
+EXPORT_SYMBOL_GPL(xenbus_strstate);
+
+int xenbus_watch_path(struct xenbus_device *dev, const char *path,
+		      struct xenbus_watch *watch,
+		      void (*callback)(struct xenbus_watch *,
+				       const char **, unsigned int))
+{
+	int err;
+
+	watch->node = path;
+	watch->callback = callback;
+
+	err = register_xenbus_watch(watch);
+
+	if (err) {
+		watch->node = NULL;
+		watch->callback = NULL;
+		xenbus_dev_fatal(dev, err, "adding watch on %s", path);
+	}
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(xenbus_watch_path);
+
+
+int xenbus_watch_path2(struct xenbus_device *dev, const char *path,
+		       const char *path2, struct xenbus_watch *watch,
+		       void (*callback)(struct xenbus_watch *,
+					const char **, unsigned int))
+{
+	int err;
+	char *state = kasprintf(GFP_NOIO | __GFP_HIGH, "%s/%s", path, path2);
+	if (!state) {
+		xenbus_dev_fatal(dev, -ENOMEM, "allocating path for watch");
+		return -ENOMEM;
+	}
+	err = xenbus_watch_path(dev, state, watch, callback);
+
+	if (err)
+		kfree(state);
+	return err;
+}
+EXPORT_SYMBOL_GPL(xenbus_watch_path2);
+
+
+int xenbus_switch_state(struct xenbus_device *dev, enum xenbus_state state)
+{
+	/* We check whether the state is currently set to the given value, and
+	   if not, then the state is set.  We don't want to unconditionally
+	   write the given state, because we don't want to fire watches
+	   unnecessarily.  Furthermore, if the node has gone, we don't write
+	   to it, as the device will be tearing down, and we don't want to
+	   resurrect that directory.
+
+	   Note that, because of this cached value of our state, this function
+	   will not work inside a Xenstore transaction (something it was
+	   trying to in the past) because dev->state would not get reset if
+	   the transaction was aborted.
+
+	 */
+
+	int current_state;
+	int err;
+
+	if (state == dev->state)
+		return 0;
+
+	err = xenbus_scanf(XBT_NIL, dev->nodename, "state", "%d",
+			   &current_state);
+	if (err != 1)
+		return 0;
+
+	err = xenbus_printf(XBT_NIL, dev->nodename, "state", "%d", state);
+	if (err) {
+		if (state != XenbusStateClosing) /* Avoid looping */
+			xenbus_dev_fatal(dev, err, "writing new state");
+		return err;
+	}
+
+	dev->state = state;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(xenbus_switch_state);
+
+int xenbus_frontend_closed(struct xenbus_device *dev)
+{
+	xenbus_switch_state(dev, XenbusStateClosed);
+	complete(&dev->down);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(xenbus_frontend_closed);
+
+/**
+ * Return the path to the error node for the given device, or NULL on failure.
+ * If the value returned is non-NULL, then it is the caller's to kfree.
+ */
+static char *error_path(struct xenbus_device *dev)
+{
+	return kasprintf(GFP_KERNEL, "error/%s", dev->nodename);
+}
+
+
+void _dev_error(struct xenbus_device *dev, int err, const char *fmt,
+		va_list ap)
+{
+	int ret;
+	unsigned int len;
+	char *printf_buffer = NULL, *path_buffer = NULL;
+
+#define PRINTF_BUFFER_SIZE 4096
+	printf_buffer = kmalloc(PRINTF_BUFFER_SIZE, GFP_KERNEL);
+	if (printf_buffer == NULL)
+		goto fail;
+
+	len = sprintf(printf_buffer, "%i ", -err);
+	ret = vsnprintf(printf_buffer+len, PRINTF_BUFFER_SIZE-len, fmt, ap);
+
+	BUG_ON(len + ret > PRINTF_BUFFER_SIZE-1);
+
+	dev_err(&dev->dev, "%s\n", printf_buffer);
+
+	path_buffer = error_path(dev);
+
+	if (path_buffer == NULL) {
+		printk("xenbus: failed to write error node for %s (%s)\n",
+		       dev->nodename, printf_buffer);
+		goto fail;
+	}
+
+	if (xenbus_write(XBT_NIL, path_buffer, "error", printf_buffer) != 0) {
+		printk("xenbus: failed to write error node for %s (%s)\n",
+		       dev->nodename, printf_buffer);
+		goto fail;
+	}
+
+fail:
+	if (printf_buffer)
+		kfree(printf_buffer);
+	if (path_buffer)
+		kfree(path_buffer);
+}
+
+
+void xenbus_dev_error(struct xenbus_device *dev, int err, const char *fmt,
+		      ...)
+{
+	va_list ap;
+
+	va_start(ap, fmt);
+	_dev_error(dev, err, fmt, ap);
+	va_end(ap);
+}
+EXPORT_SYMBOL_GPL(xenbus_dev_error);
+
+
+void xenbus_dev_fatal(struct xenbus_device *dev, int err, const char *fmt,
+		      ...)
+{
+	va_list ap;
+
+	va_start(ap, fmt);
+	_dev_error(dev, err, fmt, ap);
+	va_end(ap);
+
+	xenbus_switch_state(dev, XenbusStateClosing);
+}
+EXPORT_SYMBOL_GPL(xenbus_dev_fatal);
+
+
+int xenbus_grant_ring(struct xenbus_device *dev, unsigned long ring_mfn)
+{
+	int err = gnttab_grant_foreign_access(dev->otherend_id, ring_mfn, 0);
+	if (err < 0)
+		xenbus_dev_fatal(dev, err, "granting access to ring page");
+	return err;
+}
+EXPORT_SYMBOL_GPL(xenbus_grant_ring);
+
+
+int xenbus_alloc_evtchn(struct xenbus_device *dev, int *port)
+{
+	struct evtchn_alloc_unbound alloc_unbound;
+	int err;
+
+	alloc_unbound.dom        = DOMID_SELF;
+	alloc_unbound.remote_dom = dev->otherend_id;
+
+	err = HYPERVISOR_event_channel_op(EVTCHNOP_alloc_unbound,
+					  &alloc_unbound);
+	if (err)
+		xenbus_dev_fatal(dev, err, "allocating event channel");
+	else
+		*port = alloc_unbound.port;
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(xenbus_alloc_evtchn);
+
+
+int xenbus_free_evtchn(struct xenbus_device *dev, int port)
+{
+	struct evtchn_close close;
+	int err;
+
+	close.port = port;
+
+	err = HYPERVISOR_event_channel_op(EVTCHNOP_close, &close);
+	if (err)
+		xenbus_dev_error(dev, err, "freeing event channel %d", port);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(xenbus_free_evtchn);
+
+
+enum xenbus_state xenbus_read_driver_state(const char *path)
+{
+	enum xenbus_state result;
+	int err = xenbus_gather(XBT_NIL, path, "state", "%d", &result, NULL);
+	if (err)
+		result = XenbusStateUnknown;
+
+	return result;
+}
+EXPORT_SYMBOL_GPL(xenbus_read_driver_state);
diff -rpuN linux-2.6.18.8/drivers/xen/xenbus/xenbus_comms.c linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_comms.c
--- linux-2.6.18.8/drivers/xen/xenbus/xenbus_comms.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_comms.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,233 @@
+/******************************************************************************
+ * xenbus_comms.c
+ *
+ * Low level code to talks to Xen Store: ringbuffer and event channel.
+ *
+ * Copyright (C) 2005 Rusty Russell, IBM Corporation
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/wait.h>
+#include <linux/interrupt.h>
+#include <linux/sched.h>
+#include <linux/err.h>
+#include <linux/ptrace.h>
+#include <linux/workqueue.h>
+#include <xen/evtchn.h>
+#include <xen/xenbus.h>
+
+#include <asm/hypervisor.h>
+
+#include "xenbus_comms.h"
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+static int xenbus_irq;
+
+extern void xenbus_probe(void *);
+extern int xenstored_ready;
+static DECLARE_WORK(probe_work, xenbus_probe, NULL);
+
+static DECLARE_WAIT_QUEUE_HEAD(xb_waitq);
+
+static irqreturn_t wake_waiting(int irq, void *unused, struct pt_regs *regs)
+{
+	if (unlikely(xenstored_ready == 0)) {
+		xenstored_ready = 1;
+		schedule_work(&probe_work);
+	}
+
+	wake_up(&xb_waitq);
+	return IRQ_HANDLED;
+}
+
+static int check_indexes(XENSTORE_RING_IDX cons, XENSTORE_RING_IDX prod)
+{
+	return ((prod - cons) <= XENSTORE_RING_SIZE);
+}
+
+static void *get_output_chunk(XENSTORE_RING_IDX cons,
+			      XENSTORE_RING_IDX prod,
+			      char *buf, uint32_t *len)
+{
+	*len = XENSTORE_RING_SIZE - MASK_XENSTORE_IDX(prod);
+	if ((XENSTORE_RING_SIZE - (prod - cons)) < *len)
+		*len = XENSTORE_RING_SIZE - (prod - cons);
+	return buf + MASK_XENSTORE_IDX(prod);
+}
+
+static const void *get_input_chunk(XENSTORE_RING_IDX cons,
+				   XENSTORE_RING_IDX prod,
+				   const char *buf, uint32_t *len)
+{
+	*len = XENSTORE_RING_SIZE - MASK_XENSTORE_IDX(cons);
+	if ((prod - cons) < *len)
+		*len = prod - cons;
+	return buf + MASK_XENSTORE_IDX(cons);
+}
+
+int xb_write(const void *data, unsigned len)
+{
+	struct xenstore_domain_interface *intf = xen_store_interface;
+	XENSTORE_RING_IDX cons, prod;
+	int rc;
+
+	while (len != 0) {
+		void *dst;
+		unsigned int avail;
+
+		rc = wait_event_interruptible(
+			xb_waitq,
+			(intf->req_prod - intf->req_cons) !=
+			XENSTORE_RING_SIZE);
+		if (rc < 0)
+			return rc;
+
+		/* Read indexes, then verify. */
+		cons = intf->req_cons;
+		prod = intf->req_prod;
+		if (!check_indexes(cons, prod)) {
+			intf->req_cons = intf->req_prod = 0;
+			return -EIO;
+		}
+
+		dst = get_output_chunk(cons, prod, intf->req, &avail);
+		if (avail == 0)
+			continue;
+		if (avail > len)
+			avail = len;
+
+		/* Must write data /after/ reading the consumer index. */
+		mb();
+
+		memcpy(dst, data, avail);
+		data += avail;
+		len -= avail;
+
+		/* Other side must not see new producer until data is there. */
+		wmb();
+		intf->req_prod += avail;
+
+		/* Implies mb(): other side will see the updated producer. */
+		notify_remote_via_evtchn(xen_store_evtchn);
+	}
+
+	return 0;
+}
+
+int xb_data_to_read(void)
+{
+	struct xenstore_domain_interface *intf = xen_store_interface;
+	return (intf->rsp_cons != intf->rsp_prod);
+}
+
+int xb_wait_for_data_to_read(void)
+{
+	return wait_event_interruptible(xb_waitq, xb_data_to_read());
+}
+
+int xb_read(void *data, unsigned len)
+{
+	struct xenstore_domain_interface *intf = xen_store_interface;
+	XENSTORE_RING_IDX cons, prod;
+	int rc;
+
+	while (len != 0) {
+		unsigned int avail;
+		const char *src;
+
+		rc = xb_wait_for_data_to_read();
+		if (rc < 0)
+			return rc;
+
+		/* Read indexes, then verify. */
+		cons = intf->rsp_cons;
+		prod = intf->rsp_prod;
+		if (!check_indexes(cons, prod)) {
+			intf->rsp_cons = intf->rsp_prod = 0;
+			return -EIO;
+		}
+
+		src = get_input_chunk(cons, prod, intf->rsp, &avail);
+		if (avail == 0)
+			continue;
+		if (avail > len)
+			avail = len;
+
+		/* Must read data /after/ reading the producer index. */
+		rmb();
+
+		memcpy(data, src, avail);
+		data += avail;
+		len -= avail;
+
+		/* Other side must not see free space until we've copied out */
+		mb();
+		intf->rsp_cons += avail;
+
+		pr_debug("Finished read of %i bytes (%i to go)\n", avail, len);
+
+		/* Implies mb(): other side will see the updated consumer. */
+		notify_remote_via_evtchn(xen_store_evtchn);
+	}
+
+	return 0;
+}
+
+/* Set up interrupt handler off store event channel. */
+int xb_init_comms(void)
+{
+	struct xenstore_domain_interface *intf = xen_store_interface;
+	int err;
+
+	if (intf->req_prod != intf->req_cons)
+		printk(KERN_ERR "XENBUS request ring is not quiescent "
+		       "(%08x:%08x)!\n", intf->req_cons, intf->req_prod);
+
+	if (intf->rsp_prod != intf->rsp_cons) {
+		printk(KERN_WARNING "XENBUS response ring is not quiescent "
+		       "(%08x:%08x): fixing up\n",
+		       intf->rsp_cons, intf->rsp_prod);
+		intf->rsp_cons = intf->rsp_prod;
+	}
+
+	if (xenbus_irq)
+		unbind_from_irqhandler(xenbus_irq, &xb_waitq);
+
+	err = bind_caller_port_to_irqhandler(
+		xen_store_evtchn, wake_waiting,
+		0, "xenbus", &xb_waitq);
+	if (err <= 0) {
+		printk(KERN_ERR "XENBUS request irq failed %i\n", err);
+		return err;
+	}
+
+	xenbus_irq = err;
+
+	return 0;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/xenbus/xenbus_comms.h linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_comms.h
--- linux-2.6.18.8/drivers/xen/xenbus/xenbus_comms.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_comms.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,46 @@
+/*
+ * Private include for xenbus communications.
+ * 
+ * Copyright (C) 2005 Rusty Russell, IBM Corporation
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef _XENBUS_COMMS_H
+#define _XENBUS_COMMS_H
+
+int xs_init(void);
+int xb_init_comms(void);
+
+/* Low level routines. */
+int xb_write(const void *data, unsigned len);
+int xb_read(void *data, unsigned len);
+int xb_data_to_read(void);
+int xb_wait_for_data_to_read(void);
+int xs_input_avail(void);
+extern struct xenstore_domain_interface *xen_store_interface;
+extern int xen_store_evtchn;
+
+#endif /* _XENBUS_COMMS_H */
diff -rpuN linux-2.6.18.8/drivers/xen/xenbus/xenbus_dev.c linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_dev.c
--- linux-2.6.18.8/drivers/xen/xenbus/xenbus_dev.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_dev.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,408 @@
+/*
+ * xenbus_dev.c
+ * 
+ * Driver giving user-space access to the kernel's xenbus connection
+ * to xenstore.
+ * 
+ * Copyright (c) 2005, Christian Limpach
+ * Copyright (c) 2005, Rusty Russell, IBM Corporation
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/uio.h>
+#include <linux/notifier.h>
+#include <linux/wait.h>
+#include <linux/fs.h>
+#include <linux/poll.h>
+#include <linux/mutex.h>
+
+#include "xenbus_comms.h"
+
+#include <asm/uaccess.h>
+#include <asm/hypervisor.h>
+#include <xen/xenbus.h>
+#include <xen/xen_proc.h>
+#include <asm/hypervisor.h>
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+struct xenbus_dev_transaction {
+	struct list_head list;
+	struct xenbus_transaction handle;
+};
+
+struct read_buffer {
+	struct list_head list;
+	unsigned int cons;
+	unsigned int len;
+	char msg[];
+};
+
+struct xenbus_dev_data {
+	/* In-progress transaction. */
+	struct list_head transactions;
+
+	/* Active watches. */
+	struct list_head watches;
+
+	/* Partial request. */
+	unsigned int len;
+	union {
+		struct xsd_sockmsg msg;
+		char buffer[PAGE_SIZE];
+	} u;
+
+	/* Response queue. */
+	struct list_head read_buffers;
+	wait_queue_head_t read_waitq;
+
+	struct mutex reply_mutex;
+};
+
+static struct proc_dir_entry *xenbus_dev_intf;
+
+static ssize_t xenbus_dev_read(struct file *filp,
+			       char __user *ubuf,
+			       size_t len, loff_t *ppos)
+{
+	struct xenbus_dev_data *u = filp->private_data;
+	struct read_buffer *rb;
+	int i, ret;
+
+	mutex_lock(&u->reply_mutex);
+	while (list_empty(&u->read_buffers)) {
+		mutex_unlock(&u->reply_mutex);
+		ret = wait_event_interruptible(u->read_waitq,
+					       !list_empty(&u->read_buffers));
+		if (ret)
+			return ret;
+		mutex_lock(&u->reply_mutex);
+	}
+
+	rb = list_entry(u->read_buffers.next, struct read_buffer, list);
+	for (i = 0; i < len;) {
+		put_user(rb->msg[rb->cons], ubuf + i);
+		i++;
+		rb->cons++;
+		if (rb->cons == rb->len) {
+			list_del(&rb->list);
+			kfree(rb);
+			if (list_empty(&u->read_buffers))
+				break;
+			rb = list_entry(u->read_buffers.next,
+					struct read_buffer, list);
+		}
+	}
+	mutex_unlock(&u->reply_mutex);
+
+	return i;
+}
+
+static void queue_reply(struct xenbus_dev_data *u,
+			char *data, unsigned int len)
+{
+	struct read_buffer *rb;
+
+	if (len == 0)
+		return;
+
+	rb = kmalloc(sizeof(*rb) + len, GFP_KERNEL);
+	BUG_ON(rb == NULL);
+
+	rb->cons = 0;
+	rb->len = len;
+
+	memcpy(rb->msg, data, len);
+
+	list_add_tail(&rb->list, &u->read_buffers);
+
+	wake_up(&u->read_waitq);
+}
+
+struct watch_adapter
+{
+	struct list_head list;
+	struct xenbus_watch watch;
+	struct xenbus_dev_data *dev_data;
+	char *token;
+};
+
+static void free_watch_adapter (struct watch_adapter *watch)
+{
+	kfree(watch->watch.node);
+	kfree(watch->token);
+	kfree(watch);
+}
+
+static void watch_fired(struct xenbus_watch *watch,
+			const char **vec,
+			unsigned int len)
+{
+	struct watch_adapter *adap =
+            container_of(watch, struct watch_adapter, watch);
+	struct xsd_sockmsg hdr;
+	const char *path, *token;
+	int path_len, tok_len, body_len, data_len = 0;
+	
+	path = vec[XS_WATCH_PATH];
+	token = adap->token;
+
+	path_len = strlen(path) + 1;
+	tok_len = strlen(token) + 1;
+	if (len > 2)
+		data_len = vec[len] - vec[2] + 1;
+	body_len = path_len + tok_len + data_len;
+
+	hdr.type = XS_WATCH_EVENT;
+	hdr.len = body_len;
+
+	mutex_lock(&adap->dev_data->reply_mutex);
+	queue_reply(adap->dev_data, (char *)&hdr, sizeof(hdr));
+	queue_reply(adap->dev_data, (char *)path, path_len);
+	queue_reply(adap->dev_data, (char *)token, tok_len);
+	if (len > 2)
+		queue_reply(adap->dev_data, (char *)vec[2], data_len);
+	mutex_unlock(&adap->dev_data->reply_mutex);
+}
+
+static LIST_HEAD(watch_list);
+
+static ssize_t xenbus_dev_write(struct file *filp,
+				const char __user *ubuf,
+				size_t len, loff_t *ppos)
+{
+	struct xenbus_dev_data *u = filp->private_data;
+	struct xenbus_dev_transaction *trans = NULL;
+	uint32_t msg_type;
+	void *reply;
+	char *path, *token;
+	struct watch_adapter *watch, *tmp_watch;
+	int err, rc = len;
+
+	if ((len + u->len) > sizeof(u->u.buffer)) {
+		rc = -EINVAL;
+		goto out;
+	}
+
+	if (copy_from_user(u->u.buffer + u->len, ubuf, len) != 0) {
+		rc = -EFAULT;
+		goto out;
+	}
+
+	u->len += len;
+	if ((u->len < sizeof(u->u.msg)) ||
+	    (u->len < (sizeof(u->u.msg) + u->u.msg.len)))
+		return rc;
+
+	msg_type = u->u.msg.type;
+
+	switch (msg_type) {
+	case XS_TRANSACTION_START:
+	case XS_TRANSACTION_END:
+	case XS_DIRECTORY:
+	case XS_READ:
+	case XS_GET_PERMS:
+	case XS_RELEASE:
+	case XS_GET_DOMAIN_PATH:
+	case XS_WRITE:
+	case XS_MKDIR:
+	case XS_RM:
+	case XS_SET_PERMS:
+		if (msg_type == XS_TRANSACTION_START) {
+			trans = kmalloc(sizeof(*trans), GFP_KERNEL);
+			if (!trans) {
+				rc = -ENOMEM;
+				goto out;
+			}
+		}
+
+		reply = xenbus_dev_request_and_reply(&u->u.msg);
+		if (IS_ERR(reply)) {
+			kfree(trans);
+			rc = PTR_ERR(reply);
+			goto out;
+		}
+
+		if (msg_type == XS_TRANSACTION_START) {
+			trans->handle.id = simple_strtoul(reply, NULL, 0);
+			list_add(&trans->list, &u->transactions);
+		} else if (msg_type == XS_TRANSACTION_END) {
+			list_for_each_entry(trans, &u->transactions, list)
+				if (trans->handle.id == u->u.msg.tx_id)
+					break;
+			BUG_ON(&trans->list == &u->transactions);
+			list_del(&trans->list);
+			kfree(trans);
+		}
+		mutex_lock(&u->reply_mutex);
+		queue_reply(u, (char *)&u->u.msg, sizeof(u->u.msg));
+		queue_reply(u, (char *)reply, u->u.msg.len);
+		mutex_unlock(&u->reply_mutex);
+		kfree(reply);
+		break;
+
+	case XS_WATCH:
+	case XS_UNWATCH: {
+		static const char *XS_RESP = "OK";
+		struct xsd_sockmsg hdr;
+
+		path = u->u.buffer + sizeof(u->u.msg);
+		token = memchr(path, 0, u->u.msg.len);
+		if (token == NULL) {
+			rc = -EILSEQ;
+			goto out;
+		}
+		token++;
+
+		if (msg_type == XS_WATCH) {
+			watch = kzalloc(sizeof(*watch), GFP_KERNEL);
+			watch->watch.node = kmalloc(strlen(path)+1,
+                                                    GFP_KERNEL);
+			strcpy((char *)watch->watch.node, path);
+			watch->watch.callback = watch_fired;
+			watch->token = kmalloc(strlen(token)+1, GFP_KERNEL);
+			strcpy(watch->token, token);
+			watch->dev_data = u;
+
+			err = register_xenbus_watch(&watch->watch);
+			if (err) {
+				free_watch_adapter(watch);
+				rc = err;
+				goto out;
+			}
+			
+			list_add(&watch->list, &u->watches);
+		} else {
+			list_for_each_entry_safe(watch, tmp_watch,
+                                                 &u->watches, list) {
+				if (!strcmp(watch->token, token) &&
+				    !strcmp(watch->watch.node, path))
+				{
+					unregister_xenbus_watch(&watch->watch);
+					list_del(&watch->list);
+					free_watch_adapter(watch);
+					break;
+				}
+			}
+		}
+
+		hdr.type = msg_type;
+		hdr.len = strlen(XS_RESP) + 1;
+		mutex_lock(&u->reply_mutex);
+		queue_reply(u, (char *)&hdr, sizeof(hdr));
+		queue_reply(u, (char *)XS_RESP, hdr.len);
+		mutex_unlock(&u->reply_mutex);
+		break;
+	}
+
+	default:
+		rc = -EINVAL;
+		break;
+	}
+
+ out:
+	u->len = 0;
+	return rc;
+}
+
+static int xenbus_dev_open(struct inode *inode, struct file *filp)
+{
+	struct xenbus_dev_data *u;
+
+	if (xen_store_evtchn == 0)
+		return -ENOENT;
+
+	nonseekable_open(inode, filp);
+
+	u = kzalloc(sizeof(*u), GFP_KERNEL);
+	if (u == NULL)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&u->transactions);
+	INIT_LIST_HEAD(&u->watches);
+	INIT_LIST_HEAD(&u->read_buffers);
+	init_waitqueue_head(&u->read_waitq);
+
+	mutex_init(&u->reply_mutex);
+
+	filp->private_data = u;
+
+	return 0;
+}
+
+static int xenbus_dev_release(struct inode *inode, struct file *filp)
+{
+	struct xenbus_dev_data *u = filp->private_data;
+	struct xenbus_dev_transaction *trans, *tmp;
+	struct watch_adapter *watch, *tmp_watch;
+
+	list_for_each_entry_safe(trans, tmp, &u->transactions, list) {
+		xenbus_transaction_end(trans->handle, 1);
+		list_del(&trans->list);
+		kfree(trans);
+	}
+
+	list_for_each_entry_safe(watch, tmp_watch, &u->watches, list) {
+		unregister_xenbus_watch(&watch->watch);
+		list_del(&watch->list);
+		free_watch_adapter(watch);
+	}
+
+	kfree(u);
+
+	return 0;
+}
+
+static unsigned int xenbus_dev_poll(struct file *file, poll_table *wait)
+{
+	struct xenbus_dev_data *u = file->private_data;
+
+	poll_wait(file, &u->read_waitq, wait);
+	if (!list_empty(&u->read_buffers))
+		return POLLIN | POLLRDNORM;
+	return 0;
+}
+
+static const struct file_operations xenbus_dev_file_ops = {
+	.read = xenbus_dev_read,
+	.write = xenbus_dev_write,
+	.open = xenbus_dev_open,
+	.release = xenbus_dev_release,
+	.poll = xenbus_dev_poll,
+};
+
+int xenbus_dev_init(void)
+{
+	xenbus_dev_intf = create_xen_proc_entry("xenbus", 0400);
+	if (xenbus_dev_intf)
+		xenbus_dev_intf->proc_fops = &xenbus_dev_file_ops;
+
+	return 0;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/xenbus/xenbus_probe_backend.c linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_probe_backend.c
--- linux-2.6.18.8/drivers/xen/xenbus/xenbus_probe_backend.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_probe_backend.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,292 @@
+/******************************************************************************
+ * Talks to Xen Store to figure out what devices we have (backend half).
+ *
+ * Copyright (C) 2005 Rusty Russell, IBM Corporation
+ * Copyright (C) 2005 Mike Wray, Hewlett-Packard
+ * Copyright (C) 2005, 2006 XenSource Ltd
+ * Copyright (C) 2007 Solarflare Communications, Inc.
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#define DPRINTK(fmt, args...)				\
+	pr_debug("xenbus_probe (%s:%d) " fmt ".\n",	\
+		 __FUNCTION__, __LINE__, ##args)
+
+#include <linux/kernel.h>
+#include <linux/err.h>
+#include <linux/string.h>
+#include <linux/ctype.h>
+#include <linux/fcntl.h>
+#include <linux/mm.h>
+#include <linux/notifier.h>
+
+#include <asm/io.h>
+#include <asm/page.h>
+#include <asm/maddr.h>
+#include <asm/pgtable.h>
+#include <asm/hypervisor.h>
+#include <xen/xenbus.h>
+#include <xen/xen_proc.h>
+#include <xen/evtchn.h>
+#include <xen/features.h>
+
+#include "xenbus_comms.h"
+#include "xenbus_probe.h"
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+static int xenbus_uevent_backend(struct device *dev, char **envp,
+				 int num_envp, char *buffer, int buffer_size);
+static int xenbus_probe_backend(const char *type, const char *domid);
+
+extern int read_otherend_details(struct xenbus_device *xendev,
+				 char *id_node, char *path_node);
+
+static int read_frontend_details(struct xenbus_device *xendev)
+{
+	return read_otherend_details(xendev, "frontend-id", "frontend");
+}
+
+/* backend/<type>/<fe-uuid>/<id> => <type>-<fe-domid>-<id> */
+static int backend_bus_id(char bus_id[BUS_ID_SIZE], const char *nodename)
+{
+	int domid, err;
+	const char *devid, *type, *frontend;
+	unsigned int typelen;
+
+	type = strchr(nodename, '/');
+	if (!type)
+		return -EINVAL;
+	type++;
+	typelen = strcspn(type, "/");
+	if (!typelen || type[typelen] != '/')
+		return -EINVAL;
+
+	devid = strrchr(nodename, '/') + 1;
+
+	err = xenbus_gather(XBT_NIL, nodename, "frontend-id", "%i", &domid,
+			    "frontend", NULL, &frontend,
+			    NULL);
+	if (err)
+		return err;
+	if (strlen(frontend) == 0)
+		err = -ERANGE;
+	if (!err && !xenbus_exists(XBT_NIL, frontend, ""))
+		err = -ENOENT;
+	kfree(frontend);
+
+	if (err)
+		return err;
+
+	if (snprintf(bus_id, BUS_ID_SIZE,
+		     "%.*s-%i-%s", typelen, type, domid, devid) >= BUS_ID_SIZE)
+		return -ENOSPC;
+	return 0;
+}
+
+static struct xen_bus_type xenbus_backend = {
+	.root = "backend",
+	.levels = 3, 		/* backend/type/<frontend>/<id> */
+	.get_bus_id = backend_bus_id,
+	.probe = xenbus_probe_backend,
+	.error = -ENODEV,
+	.bus = {
+		.name     = "xen-backend",
+		.match    = xenbus_match,
+		.probe    = xenbus_dev_probe,
+		.remove   = xenbus_dev_remove,
+//		.shutdown = xenbus_dev_shutdown,
+		.uevent   = xenbus_uevent_backend,
+	},
+	.dev = {
+		.bus_id = "xen-backend",
+	},
+};
+
+static int xenbus_uevent_backend(struct device *dev, char **envp,
+				 int num_envp, char *buffer, int buffer_size)
+{
+	struct xenbus_device *xdev;
+	struct xenbus_driver *drv;
+	int i = 0;
+	int length = 0;
+
+	DPRINTK("");
+
+	if (dev == NULL)
+		return -ENODEV;
+
+	xdev = to_xenbus_device(dev);
+	if (xdev == NULL)
+		return -ENODEV;
+
+	/* stuff we want to pass to /sbin/hotplug */
+	add_uevent_var(envp, num_envp, &i, buffer, buffer_size, &length,
+		       "XENBUS_TYPE=%s", xdev->devicetype);
+
+	add_uevent_var(envp, num_envp, &i, buffer, buffer_size, &length,
+		       "XENBUS_PATH=%s", xdev->nodename);
+
+	add_uevent_var(envp, num_envp, &i, buffer, buffer_size, &length,
+		       "XENBUS_BASE_PATH=%s", xenbus_backend.root);
+
+	/* terminate, set to next free slot, shrink available space */
+	envp[i] = NULL;
+	envp = &envp[i];
+	num_envp -= i;
+	buffer = &buffer[length];
+	buffer_size -= length;
+
+	if (dev->driver) {
+		drv = to_xenbus_driver(dev->driver);
+		if (drv && drv->uevent)
+			return drv->uevent(xdev, envp, num_envp, buffer,
+					   buffer_size);
+	}
+
+	return 0;
+}
+
+int xenbus_register_backend(struct xenbus_driver *drv)
+{
+	drv->read_otherend_details = read_frontend_details;
+
+	return xenbus_register_driver_common(drv, &xenbus_backend);
+}
+EXPORT_SYMBOL_GPL(xenbus_register_backend);
+
+/* backend/<typename>/<frontend-uuid>/<name> */
+static int xenbus_probe_backend_unit(const char *dir,
+				     const char *type,
+				     const char *name)
+{
+	char *nodename;
+	int err;
+
+	nodename = kasprintf(GFP_KERNEL, "%s/%s", dir, name);
+	if (!nodename)
+		return -ENOMEM;
+
+	DPRINTK("%s\n", nodename);
+
+	err = xenbus_probe_node(&xenbus_backend, type, nodename);
+	kfree(nodename);
+	return err;
+}
+
+/* backend/<typename>/<frontend-domid> */
+static int xenbus_probe_backend(const char *type, const char *domid)
+{
+	char *nodename;
+	int err = 0;
+	char **dir;
+	unsigned int i, dir_n = 0;
+
+	DPRINTK("");
+
+	nodename = kasprintf(GFP_KERNEL, "%s/%s/%s", xenbus_backend.root, type, domid);
+	if (!nodename)
+		return -ENOMEM;
+
+	dir = xenbus_directory(XBT_NIL, nodename, "", &dir_n);
+	if (IS_ERR(dir)) {
+		kfree(nodename);
+		return PTR_ERR(dir);
+	}
+
+	for (i = 0; i < dir_n; i++) {
+		err = xenbus_probe_backend_unit(nodename, type, dir[i]);
+		if (err)
+			break;
+	}
+	kfree(dir);
+	kfree(nodename);
+	return err;
+}
+
+static void backend_changed(struct xenbus_watch *watch,
+			    const char **vec, unsigned int len)
+{
+	DPRINTK("");
+
+	dev_changed(vec[XS_WATCH_PATH], &xenbus_backend);
+}
+
+static struct xenbus_watch be_watch = {
+	.node = "backend",
+	.callback = backend_changed,
+};
+
+void xenbus_backend_suspend(int (*fn)(struct device *, void *))
+{
+	DPRINTK("");
+	if (!xenbus_backend.error)
+		bus_for_each_dev(&xenbus_backend.bus, NULL, NULL, fn);
+}
+
+void xenbus_backend_resume(int (*fn)(struct device *, void *))
+{
+	DPRINTK("");
+	if (!xenbus_backend.error)
+		bus_for_each_dev(&xenbus_backend.bus, NULL, NULL, fn);
+}
+
+void xenbus_backend_probe_and_watch(void)
+{
+	xenbus_probe_devices(&xenbus_backend);
+	register_xenbus_watch(&be_watch);
+}
+
+void xenbus_backend_bus_register(void)
+{
+	xenbus_backend.error = bus_register(&xenbus_backend.bus);
+	if (xenbus_backend.error)
+		printk(KERN_WARNING
+		       "XENBUS: Error registering backend bus: %i\n",
+		       xenbus_backend.error);
+}
+
+void xenbus_backend_device_register(void)
+{
+	if (xenbus_backend.error)
+		return;
+
+	xenbus_backend.error = device_register(&xenbus_backend.dev);
+	if (xenbus_backend.error) {
+		bus_unregister(&xenbus_backend.bus);
+		printk(KERN_WARNING
+		       "XENBUS: Error registering backend device: %i\n",
+		       xenbus_backend.error);
+	}
+}
+
+int xenbus_for_each_backend(void *arg, int (*fn)(struct device *, void *))
+{
+	return bus_for_each_dev(&xenbus_backend.bus, NULL, arg, fn);
+}
+EXPORT_SYMBOL_GPL(xenbus_for_each_backend);
diff -rpuN linux-2.6.18.8/drivers/xen/xenbus/xenbus_probe.c linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_probe.c
--- linux-2.6.18.8/drivers/xen/xenbus/xenbus_probe.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_probe.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,1153 @@
+/******************************************************************************
+ * Talks to Xen Store to figure out what devices we have.
+ *
+ * Copyright (C) 2005 Rusty Russell, IBM Corporation
+ * Copyright (C) 2005 Mike Wray, Hewlett-Packard
+ * Copyright (C) 2005, 2006 XenSource Ltd
+ * Copyright (C) 2007 Solarflare Communications, Inc.
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#define DPRINTK(fmt, args...)				\
+	pr_debug("xenbus_probe (%s:%d) " fmt ".\n",	\
+		 __FUNCTION__, __LINE__, ##args)
+
+#include <linux/kernel.h>
+#include <linux/err.h>
+#include <linux/string.h>
+#include <linux/ctype.h>
+#include <linux/fcntl.h>
+#include <linux/mm.h>
+#include <linux/notifier.h>
+#include <linux/mutex.h>
+#include <linux/module.h>
+
+#include <asm/io.h>
+#include <asm/page.h>
+#include <asm/maddr.h>
+#include <asm/pgtable.h>
+#include <asm/hypervisor.h>
+#include <xen/xenbus.h>
+#include <xen/xen_proc.h>
+#include <xen/evtchn.h>
+#include <xen/features.h>
+#ifdef MODULE
+#include <xen/hvm.h>
+#endif
+
+#include "xenbus_comms.h"
+#include "xenbus_probe.h"
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+int xen_store_evtchn;
+struct xenstore_domain_interface *xen_store_interface;
+static unsigned long xen_store_mfn;
+
+extern struct mutex xenwatch_mutex;
+
+static BLOCKING_NOTIFIER_HEAD(xenstore_chain);
+
+static void wait_for_devices(struct xenbus_driver *xendrv);
+
+static int xenbus_probe_frontend(const char *type, const char *name);
+
+static void xenbus_dev_shutdown(struct device *_dev);
+
+/* If something in array of ids matches this device, return it. */
+static const struct xenbus_device_id *
+match_device(const struct xenbus_device_id *arr, struct xenbus_device *dev)
+{
+	for (; *arr->devicetype != '\0'; arr++) {
+		if (!strcmp(arr->devicetype, dev->devicetype))
+			return arr;
+	}
+	return NULL;
+}
+
+int xenbus_match(struct device *_dev, struct device_driver *_drv)
+{
+	struct xenbus_driver *drv = to_xenbus_driver(_drv);
+
+	if (!drv->ids)
+		return 0;
+
+	return match_device(drv->ids, to_xenbus_device(_dev)) != NULL;
+}
+
+/* device/<type>/<id> => <type>-<id> */
+static int frontend_bus_id(char bus_id[BUS_ID_SIZE], const char *nodename)
+{
+	nodename = strchr(nodename, '/');
+	if (!nodename || strlen(nodename + 1) >= BUS_ID_SIZE) {
+		printk(KERN_WARNING "XENBUS: bad frontend %s\n", nodename);
+		return -EINVAL;
+	}
+
+	strlcpy(bus_id, nodename + 1, BUS_ID_SIZE);
+	if (!strchr(bus_id, '/')) {
+		printk(KERN_WARNING "XENBUS: bus_id %s no slash\n", bus_id);
+		return -EINVAL;
+	}
+	*strchr(bus_id, '/') = '-';
+	return 0;
+}
+
+
+static void free_otherend_details(struct xenbus_device *dev)
+{
+	kfree(dev->otherend);
+	dev->otherend = NULL;
+}
+
+
+static void free_otherend_watch(struct xenbus_device *dev)
+{
+	if (dev->otherend_watch.node) {
+		unregister_xenbus_watch(&dev->otherend_watch);
+		kfree(dev->otherend_watch.node);
+		dev->otherend_watch.node = NULL;
+	}
+}
+
+
+int read_otherend_details(struct xenbus_device *xendev,
+				 char *id_node, char *path_node)
+{
+	int err = xenbus_gather(XBT_NIL, xendev->nodename,
+				id_node, "%i", &xendev->otherend_id,
+				path_node, NULL, &xendev->otherend,
+				NULL);
+	if (err) {
+		xenbus_dev_fatal(xendev, err,
+				 "reading other end details from %s",
+				 xendev->nodename);
+		return err;
+	}
+	if (strlen(xendev->otherend) == 0 ||
+	    !xenbus_exists(XBT_NIL, xendev->otherend, "")) {
+		xenbus_dev_fatal(xendev, -ENOENT,
+				 "unable to read other end from %s.  "
+				 "missing or inaccessible.",
+				 xendev->nodename);
+		free_otherend_details(xendev);
+		return -ENOENT;
+	}
+
+	return 0;
+}
+
+
+static int read_backend_details(struct xenbus_device *xendev)
+{
+	return read_otherend_details(xendev, "backend-id", "backend");
+}
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,16)
+static int xenbus_uevent_frontend(struct device *dev, char **envp,
+				  int num_envp, char *buffer, int buffer_size)
+{
+	struct xenbus_device *xdev;
+	int length = 0, i = 0;
+
+	if (dev == NULL)
+		return -ENODEV;
+	xdev = to_xenbus_device(dev);
+	if (xdev == NULL)
+		return -ENODEV;
+
+	/* stuff we want to pass to /sbin/hotplug */
+	add_uevent_var(envp, num_envp, &i, buffer, buffer_size, &length,
+		       "XENBUS_TYPE=%s", xdev->devicetype);
+	add_uevent_var(envp, num_envp, &i, buffer, buffer_size, &length,
+		       "XENBUS_PATH=%s", xdev->nodename);
+	add_uevent_var(envp, num_envp, &i, buffer, buffer_size, &length,
+		       "MODALIAS=xen:%s", xdev->devicetype);
+
+	return 0;
+}
+#endif
+
+/* Bus type for frontend drivers. */
+static struct xen_bus_type xenbus_frontend = {
+	.root = "device",
+	.levels = 2, 		/* device/type/<id> */
+	.get_bus_id = frontend_bus_id,
+	.probe = xenbus_probe_frontend,
+	.error = -ENODEV,
+	.bus = {
+		.name     = "xen",
+		.match    = xenbus_match,
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,16)
+		.probe    = xenbus_dev_probe,
+		.remove   = xenbus_dev_remove,
+		.shutdown = xenbus_dev_shutdown,
+		.uevent   = xenbus_uevent_frontend,
+#endif
+	},
+	.dev = {
+		.bus_id = "xen",
+	},
+};
+
+static void otherend_changed(struct xenbus_watch *watch,
+			     const char **vec, unsigned int len)
+{
+	struct xenbus_device *dev =
+		container_of(watch, struct xenbus_device, otherend_watch);
+	struct xenbus_driver *drv = to_xenbus_driver(dev->dev.driver);
+	enum xenbus_state state;
+
+	/* Protect us against watches firing on old details when the otherend
+	   details change, say immediately after a resume. */
+	if (!dev->otherend ||
+	    strncmp(dev->otherend, vec[XS_WATCH_PATH],
+		    strlen(dev->otherend))) {
+		DPRINTK("Ignoring watch at %s", vec[XS_WATCH_PATH]);
+		return;
+	}
+
+	state = xenbus_read_driver_state(dev->otherend);
+
+	DPRINTK("state is %d (%s), %s, %s", state, xenbus_strstate(state),
+		dev->otherend_watch.node, vec[XS_WATCH_PATH]);
+
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,16)
+	/*
+	 * Ignore xenbus transitions during shutdown. This prevents us doing
+	 * work that can fail e.g., when the rootfs is gone.
+	 */
+	if (system_state > SYSTEM_RUNNING) {
+		struct xen_bus_type *bus = bus;
+		bus = container_of(dev->dev.bus, struct xen_bus_type, bus);
+		/* If we're frontend, drive the state machine to Closed. */
+		/* This should cause the backend to release our resources. */
+		if ((bus == &xenbus_frontend) && (state == XenbusStateClosing))
+			xenbus_frontend_closed(dev);
+		return;
+	}
+#endif
+
+	if (drv->otherend_changed)
+		drv->otherend_changed(dev, state);
+}
+
+
+static int talk_to_otherend(struct xenbus_device *dev)
+{
+	struct xenbus_driver *drv = to_xenbus_driver(dev->dev.driver);
+
+	free_otherend_watch(dev);
+	free_otherend_details(dev);
+
+	return drv->read_otherend_details(dev);
+}
+
+
+static int watch_otherend(struct xenbus_device *dev)
+{
+	return xenbus_watch_path2(dev, dev->otherend, "state",
+				  &dev->otherend_watch, otherend_changed);
+}
+
+
+int xenbus_dev_probe(struct device *_dev)
+{
+	struct xenbus_device *dev = to_xenbus_device(_dev);
+	struct xenbus_driver *drv = to_xenbus_driver(_dev->driver);
+	const struct xenbus_device_id *id;
+	int err;
+
+	DPRINTK("%s", dev->nodename);
+
+	if (!drv->probe) {
+		err = -ENODEV;
+		goto fail;
+	}
+
+	id = match_device(drv->ids, dev);
+	if (!id) {
+		err = -ENODEV;
+		goto fail;
+	}
+
+	err = talk_to_otherend(dev);
+	if (err) {
+		printk(KERN_WARNING
+		       "xenbus_probe: talk_to_otherend on %s failed.\n",
+		       dev->nodename);
+		return err;
+	}
+
+	err = drv->probe(dev, id);
+	if (err)
+		goto fail;
+
+	err = watch_otherend(dev);
+	if (err) {
+		printk(KERN_WARNING
+		       "xenbus_probe: watch_otherend on %s failed.\n",
+		       dev->nodename);
+		return err;
+	}
+
+	return 0;
+fail:
+	xenbus_dev_error(dev, err, "xenbus_dev_probe on %s", dev->nodename);
+	xenbus_switch_state(dev, XenbusStateClosed);
+	return -ENODEV;
+}
+
+int xenbus_dev_remove(struct device *_dev)
+{
+	struct xenbus_device *dev = to_xenbus_device(_dev);
+	struct xenbus_driver *drv = to_xenbus_driver(_dev->driver);
+
+	DPRINTK("%s", dev->nodename);
+
+	free_otherend_watch(dev);
+	free_otherend_details(dev);
+
+	if (drv->remove)
+		drv->remove(dev);
+
+	xenbus_switch_state(dev, XenbusStateClosed);
+	return 0;
+}
+
+static void xenbus_dev_shutdown(struct device *_dev)
+{
+	struct xenbus_device *dev = to_xenbus_device(_dev);
+	unsigned long timeout = 5*HZ;
+
+	DPRINTK("%s", dev->nodename);
+
+	if (is_initial_xendomain())
+		return;
+
+	get_device(&dev->dev);
+	if (dev->state != XenbusStateConnected) {
+		printk("%s: %s: %s != Connected, skipping\n", __FUNCTION__,
+		       dev->nodename, xenbus_strstate(dev->state));
+		goto out;
+	}
+	xenbus_switch_state(dev, XenbusStateClosing);
+	timeout = wait_for_completion_timeout(&dev->down, timeout);
+	if (!timeout)
+		printk("%s: %s timeout closing device\n", __FUNCTION__, dev->nodename);
+ out:
+	put_device(&dev->dev);
+}
+
+int xenbus_register_driver_common(struct xenbus_driver *drv,
+				  struct xen_bus_type *bus)
+{
+	int ret;
+
+	if (bus->error)
+		return bus->error;
+
+	drv->driver.name = drv->name;
+	drv->driver.bus = &bus->bus;
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,10)
+	drv->driver.owner = drv->owner;
+#endif
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,16)
+	drv->driver.probe = xenbus_dev_probe;
+	drv->driver.remove = xenbus_dev_remove;
+	drv->driver.shutdown = xenbus_dev_shutdown;
+#endif
+
+	mutex_lock(&xenwatch_mutex);
+	ret = driver_register(&drv->driver);
+	mutex_unlock(&xenwatch_mutex);
+	return ret;
+}
+
+int xenbus_register_frontend(struct xenbus_driver *drv)
+{
+	int ret;
+
+	drv->read_otherend_details = read_backend_details;
+
+	ret = xenbus_register_driver_common(drv, &xenbus_frontend);
+	if (ret)
+		return ret;
+
+	/* If this driver is loaded as a module wait for devices to attach. */
+	wait_for_devices(drv);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(xenbus_register_frontend);
+
+void xenbus_unregister_driver(struct xenbus_driver *drv)
+{
+	driver_unregister(&drv->driver);
+}
+EXPORT_SYMBOL_GPL(xenbus_unregister_driver);
+
+struct xb_find_info
+{
+	struct xenbus_device *dev;
+	const char *nodename;
+};
+
+static int cmp_dev(struct device *dev, void *data)
+{
+	struct xenbus_device *xendev = to_xenbus_device(dev);
+	struct xb_find_info *info = data;
+
+	if (!strcmp(xendev->nodename, info->nodename)) {
+		info->dev = xendev;
+		get_device(dev);
+		return 1;
+	}
+	return 0;
+}
+
+struct xenbus_device *xenbus_device_find(const char *nodename,
+					 struct bus_type *bus)
+{
+	struct xb_find_info info = { .dev = NULL, .nodename = nodename };
+
+	bus_for_each_dev(bus, NULL, &info, cmp_dev);
+	return info.dev;
+}
+
+static int cleanup_dev(struct device *dev, void *data)
+{
+	struct xenbus_device *xendev = to_xenbus_device(dev);
+	struct xb_find_info *info = data;
+	int len = strlen(info->nodename);
+
+	DPRINTK("%s", info->nodename);
+
+	/* Match the info->nodename path, or any subdirectory of that path. */
+	if (strncmp(xendev->nodename, info->nodename, len))
+		return 0;
+
+	/* If the node name is longer, ensure it really is a subdirectory. */
+	if ((strlen(xendev->nodename) > len) && (xendev->nodename[len] != '/'))
+		return 0;
+
+	info->dev = xendev;
+	get_device(dev);
+	return 1;
+}
+
+static void xenbus_cleanup_devices(const char *path, struct bus_type *bus)
+{
+	struct xb_find_info info = { .nodename = path };
+
+	do {
+		info.dev = NULL;
+		bus_for_each_dev(bus, NULL, &info, cleanup_dev);
+		if (info.dev) {
+			device_unregister(&info.dev->dev);
+			put_device(&info.dev->dev);
+		}
+	} while (info.dev);
+}
+
+static void xenbus_dev_release(struct device *dev)
+{
+	if (dev)
+		kfree(to_xenbus_device(dev));
+}
+
+static ssize_t xendev_show_nodename(struct device *dev,
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,13)
+				    struct device_attribute *attr,
+#endif
+				    char *buf)
+{
+	return sprintf(buf, "%s\n", to_xenbus_device(dev)->nodename);
+}
+DEVICE_ATTR(nodename, S_IRUSR | S_IRGRP | S_IROTH, xendev_show_nodename, NULL);
+
+static ssize_t xendev_show_devtype(struct device *dev,
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,13)
+				   struct device_attribute *attr,
+#endif
+				   char *buf)
+{
+	return sprintf(buf, "%s\n", to_xenbus_device(dev)->devicetype);
+}
+DEVICE_ATTR(devtype, S_IRUSR | S_IRGRP | S_IROTH, xendev_show_devtype, NULL);
+
+
+int xenbus_probe_node(struct xen_bus_type *bus,
+		      const char *type,
+		      const char *nodename)
+{
+	int err;
+	struct xenbus_device *xendev;
+	size_t stringlen;
+	char *tmpstring;
+
+	enum xenbus_state state = xenbus_read_driver_state(nodename);
+
+	if (bus->error)
+		return bus->error;
+
+	if (state != XenbusStateInitialising) {
+		/* Device is not new, so ignore it.  This can happen if a
+		   device is going away after switching to Closed.  */
+		return 0;
+	}
+
+	stringlen = strlen(nodename) + 1 + strlen(type) + 1;
+	xendev = kzalloc(sizeof(*xendev) + stringlen, GFP_KERNEL);
+	if (!xendev)
+		return -ENOMEM;
+
+	xendev->state = XenbusStateInitialising;
+
+	/* Copy the strings into the extra space. */
+
+	tmpstring = (char *)(xendev + 1);
+	strcpy(tmpstring, nodename);
+	xendev->nodename = tmpstring;
+
+	tmpstring += strlen(tmpstring) + 1;
+	strcpy(tmpstring, type);
+	xendev->devicetype = tmpstring;
+	init_completion(&xendev->down);
+
+	xendev->dev.parent = &bus->dev;
+	xendev->dev.bus = &bus->bus;
+	xendev->dev.release = xenbus_dev_release;
+
+	err = bus->get_bus_id(xendev->dev.bus_id, xendev->nodename);
+	if (err)
+		goto fail;
+
+	/* Register with generic device framework. */
+	err = device_register(&xendev->dev);
+	if (err)
+		goto fail;
+
+	err = device_create_file(&xendev->dev, &dev_attr_nodename);
+	if (err)
+		goto unregister;
+	err = device_create_file(&xendev->dev, &dev_attr_devtype);
+	if (err)
+		goto unregister;
+
+	return 0;
+unregister:
+	device_remove_file(&xendev->dev, &dev_attr_nodename);
+	device_remove_file(&xendev->dev, &dev_attr_devtype);
+	device_unregister(&xendev->dev);
+fail:
+	kfree(xendev);
+	return err;
+}
+
+/* device/<typename>/<name> */
+static int xenbus_probe_frontend(const char *type, const char *name)
+{
+	char *nodename;
+	int err;
+
+	nodename = kasprintf(GFP_KERNEL, "%s/%s/%s", xenbus_frontend.root, type, name);
+	if (!nodename)
+		return -ENOMEM;
+
+	DPRINTK("%s", nodename);
+
+	err = xenbus_probe_node(&xenbus_frontend, type, nodename);
+	kfree(nodename);
+	return err;
+}
+
+static int xenbus_probe_device_type(struct xen_bus_type *bus, const char *type)
+{
+	int err = 0;
+	char **dir;
+	unsigned int dir_n = 0;
+	int i;
+
+	dir = xenbus_directory(XBT_NIL, bus->root, type, &dir_n);
+	if (IS_ERR(dir))
+		return PTR_ERR(dir);
+
+	for (i = 0; i < dir_n; i++) {
+		err = bus->probe(type, dir[i]);
+		if (err)
+			break;
+	}
+	kfree(dir);
+	return err;
+}
+
+int xenbus_probe_devices(struct xen_bus_type *bus)
+{
+	int err = 0;
+	char **dir;
+	unsigned int i, dir_n;
+
+	if (bus->error)
+		return bus->error;
+
+	dir = xenbus_directory(XBT_NIL, bus->root, "", &dir_n);
+	if (IS_ERR(dir))
+		return PTR_ERR(dir);
+
+	for (i = 0; i < dir_n; i++) {
+		err = xenbus_probe_device_type(bus, dir[i]);
+		if (err)
+			break;
+	}
+	kfree(dir);
+	return err;
+}
+
+static unsigned int char_count(const char *str, char c)
+{
+	unsigned int i, ret = 0;
+
+	for (i = 0; str[i]; i++)
+		if (str[i] == c)
+			ret++;
+	return ret;
+}
+
+static int strsep_len(const char *str, char c, unsigned int len)
+{
+	unsigned int i;
+
+	for (i = 0; str[i]; i++)
+		if (str[i] == c) {
+			if (len == 0)
+				return i;
+			len--;
+		}
+	return (len == 0) ? i : -ERANGE;
+}
+
+void dev_changed(const char *node, struct xen_bus_type *bus)
+{
+	int exists, rootlen;
+	struct xenbus_device *dev;
+	char type[BUS_ID_SIZE];
+	const char *p, *root;
+
+	if (bus->error || char_count(node, '/') < 2)
+ 		return;
+
+	exists = xenbus_exists(XBT_NIL, node, "");
+	if (!exists) {
+		xenbus_cleanup_devices(node, &bus->bus);
+		return;
+	}
+
+	/* backend/<type>/... or device/<type>/... */
+	p = strchr(node, '/') + 1;
+	snprintf(type, BUS_ID_SIZE, "%.*s", (int)strcspn(p, "/"), p);
+	type[BUS_ID_SIZE-1] = '\0';
+
+	rootlen = strsep_len(node, '/', bus->levels);
+	if (rootlen < 0)
+		return;
+	root = kasprintf(GFP_KERNEL, "%.*s", rootlen, node);
+	if (!root)
+		return;
+
+	dev = xenbus_device_find(root, &bus->bus);
+	if (!dev)
+		xenbus_probe_node(bus, type, root);
+	else
+		put_device(&dev->dev);
+
+	kfree(root);
+}
+
+static void frontend_changed(struct xenbus_watch *watch,
+			     const char **vec, unsigned int len)
+{
+	DPRINTK("");
+
+	dev_changed(vec[XS_WATCH_PATH], &xenbus_frontend);
+}
+
+/* We watch for devices appearing and vanishing. */
+static struct xenbus_watch fe_watch = {
+	.node = "device",
+	.callback = frontend_changed,
+};
+
+static int suspend_dev(struct device *dev, void *data)
+{
+	int err = 0;
+	struct xenbus_driver *drv;
+	struct xenbus_device *xdev;
+
+	DPRINTK("");
+
+	if (dev->driver == NULL)
+		return 0;
+	drv = to_xenbus_driver(dev->driver);
+	xdev = container_of(dev, struct xenbus_device, dev);
+	if (drv->suspend)
+		err = drv->suspend(xdev);
+	if (err)
+		printk(KERN_WARNING
+		       "xenbus: suspend %s failed: %i\n", dev->bus_id, err);
+	return 0;
+}
+
+static int suspend_cancel_dev(struct device *dev, void *data)
+{
+	int err = 0;
+	struct xenbus_driver *drv;
+	struct xenbus_device *xdev;
+
+	DPRINTK("");
+
+	if (dev->driver == NULL)
+		return 0;
+	drv = to_xenbus_driver(dev->driver);
+	xdev = container_of(dev, struct xenbus_device, dev);
+	if (drv->suspend_cancel)
+		err = drv->suspend_cancel(xdev);
+	if (err)
+		printk(KERN_WARNING
+		       "xenbus: suspend_cancel %s failed: %i\n",
+		       dev->bus_id, err);
+	return 0;
+}
+
+static int resume_dev(struct device *dev, void *data)
+{
+	int err;
+	struct xenbus_driver *drv;
+	struct xenbus_device *xdev;
+
+	DPRINTK("");
+
+	if (dev->driver == NULL)
+		return 0;
+
+	drv = to_xenbus_driver(dev->driver);
+	xdev = container_of(dev, struct xenbus_device, dev);
+
+	err = talk_to_otherend(xdev);
+	if (err) {
+		printk(KERN_WARNING
+		       "xenbus: resume (talk_to_otherend) %s failed: %i\n",
+		       dev->bus_id, err);
+		return err;
+	}
+
+	xdev->state = XenbusStateInitialising;
+
+	if (drv->resume) {
+		err = drv->resume(xdev);
+		if (err) { 
+			printk(KERN_WARNING
+			       "xenbus: resume %s failed: %i\n", 
+			       dev->bus_id, err);
+			return err;
+		}
+	}
+
+	err = watch_otherend(xdev);
+	if (err) {
+		printk(KERN_WARNING
+		       "xenbus_probe: resume (watch_otherend) %s failed: "
+		       "%d.\n", dev->bus_id, err);
+		return err;
+	}
+
+	return 0;
+}
+
+void xenbus_suspend(void)
+{
+	DPRINTK("");
+
+	if (!xenbus_frontend.error)
+		bus_for_each_dev(&xenbus_frontend.bus, NULL, NULL, suspend_dev);
+	xenbus_backend_suspend(suspend_dev);
+	xs_suspend();
+}
+EXPORT_SYMBOL_GPL(xenbus_suspend);
+
+void xenbus_resume(void)
+{
+	xb_init_comms();
+	xs_resume();
+	if (!xenbus_frontend.error)
+		bus_for_each_dev(&xenbus_frontend.bus, NULL, NULL, resume_dev);
+	xenbus_backend_resume(resume_dev);
+}
+EXPORT_SYMBOL_GPL(xenbus_resume);
+
+void xenbus_suspend_cancel(void)
+{
+	xs_suspend_cancel();
+	if (!xenbus_frontend.error)
+		bus_for_each_dev(&xenbus_frontend.bus, NULL, NULL, suspend_cancel_dev);
+	xenbus_backend_resume(suspend_cancel_dev);
+}
+EXPORT_SYMBOL_GPL(xenbus_suspend_cancel);
+
+/* A flag to determine if xenstored is 'ready' (i.e. has started) */
+int xenstored_ready = 0;
+
+
+int register_xenstore_notifier(struct notifier_block *nb)
+{
+	int ret = 0;
+
+	if (xenstored_ready > 0)
+		ret = nb->notifier_call(nb, 0, NULL);
+	else
+		blocking_notifier_chain_register(&xenstore_chain, nb);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(register_xenstore_notifier);
+
+void unregister_xenstore_notifier(struct notifier_block *nb)
+{
+	blocking_notifier_chain_unregister(&xenstore_chain, nb);
+}
+EXPORT_SYMBOL_GPL(unregister_xenstore_notifier);
+
+
+void xenbus_probe(void *unused)
+{
+	BUG_ON((xenstored_ready <= 0));
+
+	/* Enumerate devices in xenstore and watch for changes. */
+	xenbus_probe_devices(&xenbus_frontend);
+	register_xenbus_watch(&fe_watch);
+	xenbus_backend_probe_and_watch();
+
+	/* Notify others that xenstore is up */
+	blocking_notifier_call_chain(&xenstore_chain, 0, NULL);
+}
+
+
+#if defined(CONFIG_PROC_FS) && defined(CONFIG_XEN_PRIVILEGED_GUEST)
+static struct file_operations xsd_kva_fops;
+static struct proc_dir_entry *xsd_kva_intf;
+static struct proc_dir_entry *xsd_port_intf;
+
+static int xsd_kva_mmap(struct file *file, struct vm_area_struct *vma)
+{
+	size_t size = vma->vm_end - vma->vm_start;
+
+	if ((size > PAGE_SIZE) || (vma->vm_pgoff != 0))
+		return -EINVAL;
+
+	if (remap_pfn_range(vma, vma->vm_start, mfn_to_pfn(xen_store_mfn),
+			    size, vma->vm_page_prot))
+		return -EAGAIN;
+
+	return 0;
+}
+
+static int xsd_kva_read(char *page, char **start, off_t off,
+			int count, int *eof, void *data)
+{
+	int len;
+
+	len  = sprintf(page, "0x%p", xen_store_interface);
+	*eof = 1;
+	return len;
+}
+
+static int xsd_port_read(char *page, char **start, off_t off,
+			 int count, int *eof, void *data)
+{
+	int len;
+
+	len  = sprintf(page, "%d", xen_store_evtchn);
+	*eof = 1;
+	return len;
+}
+#endif
+
+static int xenbus_probe_init(void)
+{
+	int err = 0;
+	unsigned long page = 0;
+
+	DPRINTK("");
+
+	if (!is_running_on_xen())
+		return -ENODEV;
+
+	/* Register ourselves with the kernel bus subsystem */
+	xenbus_frontend.error = bus_register(&xenbus_frontend.bus);
+	if (xenbus_frontend.error)
+		printk(KERN_WARNING
+		       "XENBUS: Error registering frontend bus: %i\n",
+		       xenbus_frontend.error);
+	xenbus_backend_bus_register();
+
+	/*
+	 * Domain0 doesn't have a store_evtchn or store_mfn yet.
+	 */
+	if (is_initial_xendomain()) {
+		struct evtchn_alloc_unbound alloc_unbound;
+
+		/* Allocate page. */
+		page = get_zeroed_page(GFP_KERNEL);
+		if (!page)
+			return -ENOMEM;
+
+		xen_store_mfn = xen_start_info->store_mfn =
+			pfn_to_mfn(virt_to_phys((void *)page) >>
+				   PAGE_SHIFT);
+
+		/* Next allocate a local port which xenstored can bind to */
+		alloc_unbound.dom        = DOMID_SELF;
+		alloc_unbound.remote_dom = 0;
+
+		err = HYPERVISOR_event_channel_op(EVTCHNOP_alloc_unbound,
+						  &alloc_unbound);
+		if (err == -ENOSYS)
+			goto err;
+		BUG_ON(err);
+		xen_store_evtchn = xen_start_info->store_evtchn =
+			alloc_unbound.port;
+
+#if defined(CONFIG_PROC_FS) && defined(CONFIG_XEN_PRIVILEGED_GUEST)
+		/* And finally publish the above info in /proc/xen */
+		xsd_kva_intf = create_xen_proc_entry("xsd_kva", 0600);
+		if (xsd_kva_intf) {
+			memcpy(&xsd_kva_fops, xsd_kva_intf->proc_fops,
+			       sizeof(xsd_kva_fops));
+			xsd_kva_fops.mmap = xsd_kva_mmap;
+			xsd_kva_intf->proc_fops = &xsd_kva_fops;
+			xsd_kva_intf->read_proc = xsd_kva_read;
+		}
+		xsd_port_intf = create_xen_proc_entry("xsd_port", 0400);
+		if (xsd_port_intf)
+			xsd_port_intf->read_proc = xsd_port_read;
+#endif
+		xen_store_interface = mfn_to_virt(xen_store_mfn);
+	} else {
+		xenstored_ready = 1;
+#ifdef CONFIG_XEN
+		xen_store_evtchn = xen_start_info->store_evtchn;
+		xen_store_mfn = xen_start_info->store_mfn;
+		xen_store_interface = mfn_to_virt(xen_store_mfn);
+#else
+		xen_store_evtchn = hvm_get_parameter(HVM_PARAM_STORE_EVTCHN);
+		xen_store_mfn = hvm_get_parameter(HVM_PARAM_STORE_PFN);
+		xen_store_interface = ioremap(xen_store_mfn << PAGE_SHIFT,
+					      PAGE_SIZE);
+#endif
+	}
+
+
+	xenbus_dev_init();
+
+	/* Initialize the interface to xenstore. */
+	err = xs_init();
+	if (err) {
+		printk(KERN_WARNING
+		       "XENBUS: Error initializing xenstore comms: %i\n", err);
+		goto err;
+	}
+
+	/* Register ourselves with the kernel device subsystem */
+	if (!xenbus_frontend.error) {
+		xenbus_frontend.error = device_register(&xenbus_frontend.dev);
+		if (xenbus_frontend.error) {
+			bus_unregister(&xenbus_frontend.bus);
+			printk(KERN_WARNING
+			       "XENBUS: Error registering frontend device: %i\n",
+			       xenbus_frontend.error);
+		}
+	}
+	xenbus_backend_device_register();
+
+	if (!is_initial_xendomain())
+		xenbus_probe(NULL);
+
+	return 0;
+
+ err:
+	if (page)
+		free_page(page);
+
+	/*
+	 * Do not unregister the xenbus front/backend buses here. The buses
+	 * must exist because front/backend drivers will use them when they are
+	 * registered.
+	 */
+
+	return err;
+}
+
+#ifdef CONFIG_XEN
+postcore_initcall(xenbus_probe_init);
+MODULE_LICENSE("Dual BSD/GPL");
+#else
+int xenbus_init(void)
+{
+	return xenbus_probe_init();
+}
+#endif
+
+static int is_disconnected_device(struct device *dev, void *data)
+{
+	struct xenbus_device *xendev = to_xenbus_device(dev);
+	struct device_driver *drv = data;
+	struct xenbus_driver *xendrv;
+
+	/*
+	 * A device with no driver will never connect. We care only about
+	 * devices which should currently be in the process of connecting.
+	 */
+	if (!dev->driver)
+		return 0;
+
+	/* Is this search limited to a particular driver? */
+	if (drv && (dev->driver != drv))
+		return 0;
+
+	xendrv = to_xenbus_driver(dev->driver);
+	return (xendev->state < XenbusStateConnected ||
+		(xendrv->is_ready && !xendrv->is_ready(xendev)));
+}
+
+static int exists_disconnected_device(struct device_driver *drv)
+{
+	if (xenbus_frontend.error)
+		return xenbus_frontend.error;
+	return bus_for_each_dev(&xenbus_frontend.bus, NULL, drv,
+				is_disconnected_device);
+}
+
+static int print_device_status(struct device *dev, void *data)
+{
+	struct xenbus_device *xendev = to_xenbus_device(dev);
+	struct device_driver *drv = data;
+	struct xenbus_driver *xendrv;
+
+	/* Is this operation limited to a particular driver? */
+	if (drv && (dev->driver != drv))
+		return 0;
+
+	if (!dev->driver) {
+		/* Information only: is this too noisy? */
+		printk(KERN_INFO "XENBUS: Device with no driver: %s\n",
+		       xendev->nodename);
+		return 0;
+	}
+
+	if (xendev->state < XenbusStateConnected) {
+		enum xenbus_state rstate = XenbusStateUnknown;
+		if (xendev->otherend)
+			rstate = xenbus_read_driver_state(xendev->otherend);
+		printk(KERN_WARNING "XENBUS: Timeout connecting "
+		       "to device: %s (local state %d, remote state %d)\n",
+		       xendev->nodename, xendev->state, rstate);
+	}
+
+	xendrv = to_xenbus_driver(dev->driver);
+	if (xendrv->is_ready && !xendrv->is_ready(xendev))
+		printk(KERN_WARNING "XENBUS: Device not ready: %s\n",
+		       xendev->nodename);
+
+	return 0;
+}
+
+/* We only wait for device setup after most initcalls have run. */
+static int ready_to_wait_for_devices;
+
+/*
+ * On a 5-minute timeout, wait for all devices currently configured.  We need
+ * to do this to guarantee that the filesystems and / or network devices
+ * needed for boot are available, before we can allow the boot to proceed.
+ *
+ * This needs to be on a late_initcall, to happen after the frontend device
+ * drivers have been initialised, but before the root fs is mounted.
+ *
+ * A possible improvement here would be to have the tools add a per-device
+ * flag to the store entry, indicating whether it is needed at boot time.
+ * This would allow people who knew what they were doing to accelerate their
+ * boot slightly, but of course needs tools or manual intervention to set up
+ * those flags correctly.
+ */
+static void wait_for_devices(struct xenbus_driver *xendrv)
+{
+	unsigned long start = jiffies;
+	struct device_driver *drv = xendrv ? &xendrv->driver : NULL;
+	unsigned int seconds_waited = 0;
+
+	if (!ready_to_wait_for_devices || !is_running_on_xen())
+		return;
+
+	while (exists_disconnected_device(drv)) {
+		if (time_after(jiffies, start + (seconds_waited+5)*HZ)) {
+			if (!seconds_waited)
+				printk(KERN_WARNING "XENBUS: Waiting for "
+				       "devices to initialise: ");
+			seconds_waited += 5;
+			printk("%us...", 300 - seconds_waited);
+			if (seconds_waited == 300)
+				break;
+		}
+		
+		schedule_timeout_interruptible(HZ/10);
+	}
+
+	if (seconds_waited)
+		printk("\n");
+
+	bus_for_each_dev(&xenbus_frontend.bus, NULL, drv,
+			 print_device_status);
+}
+
+#ifndef MODULE
+static int __init boot_wait_for_devices(void)
+{
+	if (!xenbus_frontend.error) {
+		ready_to_wait_for_devices = 1;
+		wait_for_devices(NULL);
+	}
+	return 0;
+}
+
+late_initcall(boot_wait_for_devices);
+#endif
+
+int xenbus_for_each_frontend(void *arg, int (*fn)(struct device *, void *))
+{
+	return bus_for_each_dev(&xenbus_frontend.bus, NULL, arg, fn);
+}
+EXPORT_SYMBOL_GPL(xenbus_for_each_frontend);
diff -rpuN linux-2.6.18.8/drivers/xen/xenbus/xenbus_probe.h linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_probe.h
--- linux-2.6.18.8/drivers/xen/xenbus/xenbus_probe.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_probe.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,75 @@
+/******************************************************************************
+ * xenbus_probe.h
+ *
+ * Talks to Xen Store to figure out what devices we have.
+ *
+ * Copyright (C) 2005 Rusty Russell, IBM Corporation
+ * Copyright (C) 2005 XenSource Ltd.
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef _XENBUS_PROBE_H
+#define _XENBUS_PROBE_H
+
+#if defined(CONFIG_XEN_BACKEND) || defined(CONFIG_XEN_BACKEND_MODULE)
+extern void xenbus_backend_suspend(int (*fn)(struct device *, void *));
+extern void xenbus_backend_resume(int (*fn)(struct device *, void *));
+extern void xenbus_backend_probe_and_watch(void);
+extern void xenbus_backend_bus_register(void);
+extern void xenbus_backend_device_register(void);
+#else
+static inline void xenbus_backend_suspend(int (*fn)(struct device *, void *)) {}
+static inline void xenbus_backend_resume(int (*fn)(struct device *, void *)) {}
+static inline void xenbus_backend_probe_and_watch(void) {}
+static inline void xenbus_backend_bus_register(void) {}
+static inline void xenbus_backend_device_register(void) {}
+#endif
+
+struct xen_bus_type
+{
+	char *root;
+	int error;
+	unsigned int levels;
+	int (*get_bus_id)(char bus_id[BUS_ID_SIZE], const char *nodename);
+	int (*probe)(const char *type, const char *dir);
+	struct bus_type bus;
+	struct device dev;
+};
+
+extern int xenbus_match(struct device *_dev, struct device_driver *_drv);
+extern int xenbus_dev_probe(struct device *_dev);
+extern int xenbus_dev_remove(struct device *_dev);
+extern int xenbus_register_driver_common(struct xenbus_driver *drv,
+					 struct xen_bus_type *bus);
+extern int xenbus_probe_node(struct xen_bus_type *bus,
+			     const char *type,
+			     const char *nodename);
+extern int xenbus_probe_devices(struct xen_bus_type *bus);
+
+extern void dev_changed(const char *node, struct xen_bus_type *bus);
+
+#endif
+
diff -rpuN linux-2.6.18.8/drivers/xen/xenbus/xenbus_xs.c linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_xs.c
--- linux-2.6.18.8/drivers/xen/xenbus/xenbus_xs.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/xenbus/xenbus_xs.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,902 @@
+/******************************************************************************
+ * xenbus_xs.c
+ *
+ * This is the kernel equivalent of the "xs" library.  We don't need everything
+ * and we use xenbus_comms for communication.
+ *
+ * Copyright (C) 2005 Rusty Russell, IBM Corporation
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/unistd.h>
+#include <linux/errno.h>
+#include <linux/types.h>
+#include <linux/uio.h>
+#include <linux/kernel.h>
+#include <linux/string.h>
+#include <linux/err.h>
+#include <linux/slab.h>
+#include <linux/fcntl.h>
+#include <linux/kthread.h>
+#include <linux/rwsem.h>
+#include <linux/module.h>
+#include <linux/mutex.h>
+#include <xen/xenbus.h>
+#include "xenbus_comms.h"
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+#ifndef PF_NOFREEZE /* Old kernel (pre-2.6.6). */
+#define PF_NOFREEZE	0
+#endif
+
+struct xs_stored_msg {
+	struct list_head list;
+
+	struct xsd_sockmsg hdr;
+
+	union {
+		/* Queued replies. */
+		struct {
+			char *body;
+		} reply;
+
+		/* Queued watch events. */
+		struct {
+			struct xenbus_watch *handle;
+			char **vec;
+			unsigned int vec_size;
+		} watch;
+	} u;
+};
+
+struct xs_handle {
+	/* A list of replies. Currently only one will ever be outstanding. */
+	struct list_head reply_list;
+	spinlock_t reply_lock;
+	wait_queue_head_t reply_waitq;
+
+	/*
+	 * Mutex ordering: transaction_mutex -> watch_mutex -> request_mutex.
+	 * response_mutex is never taken simultaneously with the other three.
+	 */
+
+	/* One request at a time. */
+	struct mutex request_mutex;
+
+	/* Protect xenbus reader thread against save/restore. */
+	struct mutex response_mutex;
+
+	/* Protect transactions against save/restore. */
+	struct rw_semaphore transaction_mutex;
+
+	/* Protect watch (de)register against save/restore. */
+	struct rw_semaphore watch_mutex;
+};
+
+static struct xs_handle xs_state;
+
+/* List of registered watches, and a lock to protect it. */
+static LIST_HEAD(watches);
+static DEFINE_SPINLOCK(watches_lock);
+
+/* List of pending watch callback events, and a lock to protect it. */
+static LIST_HEAD(watch_events);
+static DEFINE_SPINLOCK(watch_events_lock);
+
+/*
+ * Details of the xenwatch callback kernel thread. The thread waits on the
+ * watch_events_waitq for work to do (queued on watch_events list). When it
+ * wakes up it acquires the xenwatch_mutex before reading the list and
+ * carrying out work.
+ */
+static pid_t xenwatch_pid;
+/* static */ DEFINE_MUTEX(xenwatch_mutex);
+static DECLARE_WAIT_QUEUE_HEAD(watch_events_waitq);
+
+static int get_error(const char *errorstring)
+{
+	unsigned int i;
+
+	for (i = 0; strcmp(errorstring, xsd_errors[i].errstring) != 0; i++) {
+		if (i == ARRAY_SIZE(xsd_errors) - 1) {
+			printk(KERN_WARNING
+			       "XENBUS xen store gave: unknown error %s",
+			       errorstring);
+			return EINVAL;
+		}
+	}
+	return xsd_errors[i].errnum;
+}
+
+static void *read_reply(enum xsd_sockmsg_type *type, unsigned int *len)
+{
+	struct xs_stored_msg *msg;
+	char *body;
+
+	spin_lock(&xs_state.reply_lock);
+
+	while (list_empty(&xs_state.reply_list)) {
+		spin_unlock(&xs_state.reply_lock);
+		/* XXX FIXME: Avoid synchronous wait for response here. */
+		wait_event(xs_state.reply_waitq,
+			   !list_empty(&xs_state.reply_list));
+		spin_lock(&xs_state.reply_lock);
+	}
+
+	msg = list_entry(xs_state.reply_list.next,
+			 struct xs_stored_msg, list);
+	list_del(&msg->list);
+
+	spin_unlock(&xs_state.reply_lock);
+
+	*type = msg->hdr.type;
+	if (len)
+		*len = msg->hdr.len;
+	body = msg->u.reply.body;
+
+	kfree(msg);
+
+	return body;
+}
+
+void *xenbus_dev_request_and_reply(struct xsd_sockmsg *msg)
+{
+	void *ret;
+	struct xsd_sockmsg req_msg = *msg;
+	int err;
+
+	if (req_msg.type == XS_TRANSACTION_START)
+		down_read(&xs_state.transaction_mutex);
+
+	mutex_lock(&xs_state.request_mutex);
+
+	err = xb_write(msg, sizeof(*msg) + msg->len);
+	if (err) {
+		msg->type = XS_ERROR;
+		ret = ERR_PTR(err);
+	} else
+		ret = read_reply(&msg->type, &msg->len);
+
+	mutex_unlock(&xs_state.request_mutex);
+
+	if ((req_msg.type == XS_TRANSACTION_END) ||
+	    ((req_msg.type == XS_TRANSACTION_START) &&
+	     (msg->type == XS_ERROR)))
+		up_read(&xs_state.transaction_mutex);
+
+	return ret;
+}
+
+/* Send message to xs, get kmalloc'ed reply.  ERR_PTR() on error. */
+static void *xs_talkv(struct xenbus_transaction t,
+		      enum xsd_sockmsg_type type,
+		      const struct kvec *iovec,
+		      unsigned int num_vecs,
+		      unsigned int *len)
+{
+	struct xsd_sockmsg msg;
+	void *ret = NULL;
+	unsigned int i;
+	int err;
+
+	msg.tx_id = t.id;
+	msg.req_id = 0;
+	msg.type = type;
+	msg.len = 0;
+	for (i = 0; i < num_vecs; i++)
+		msg.len += iovec[i].iov_len;
+
+	mutex_lock(&xs_state.request_mutex);
+
+	err = xb_write(&msg, sizeof(msg));
+	if (err) {
+		mutex_unlock(&xs_state.request_mutex);
+		return ERR_PTR(err);
+	}
+
+	for (i = 0; i < num_vecs; i++) {
+		err = xb_write(iovec[i].iov_base, iovec[i].iov_len);;
+		if (err) {
+			mutex_unlock(&xs_state.request_mutex);
+			return ERR_PTR(err);
+		}
+	}
+
+	ret = read_reply(&msg.type, len);
+
+	mutex_unlock(&xs_state.request_mutex);
+
+	if (IS_ERR(ret))
+		return ret;
+
+	if (msg.type == XS_ERROR) {
+		err = get_error(ret);
+		kfree(ret);
+		return ERR_PTR(-err);
+	}
+
+	if (msg.type != type) {
+		if (printk_ratelimit())
+			printk(KERN_WARNING
+			       "XENBUS unexpected type [%d], expected [%d]\n",
+			       msg.type, type);
+		kfree(ret);
+		return ERR_PTR(-EINVAL);
+	}
+	return ret;
+}
+
+/* Simplified version of xs_talkv: single message. */
+static void *xs_single(struct xenbus_transaction t,
+		       enum xsd_sockmsg_type type,
+		       const char *string,
+		       unsigned int *len)
+{
+	struct kvec iovec;
+
+	iovec.iov_base = (void *)string;
+	iovec.iov_len = strlen(string) + 1;
+	return xs_talkv(t, type, &iovec, 1, len);
+}
+
+/* Many commands only need an ack, don't care what it says. */
+static int xs_error(char *reply)
+{
+	if (IS_ERR(reply))
+		return PTR_ERR(reply);
+	kfree(reply);
+	return 0;
+}
+
+static unsigned int count_strings(const char *strings, unsigned int len)
+{
+	unsigned int num;
+	const char *p;
+
+	for (p = strings, num = 0; p < strings + len; p += strlen(p) + 1)
+		num++;
+
+	return num;
+}
+
+/* Return the path to dir with /name appended. Buffer must be kfree()'ed. */
+static char *join(const char *dir, const char *name)
+{
+	char *buffer;
+
+	if (strlen(name) == 0)
+		buffer = kasprintf(GFP_NOIO | __GFP_HIGH, "%s", dir);
+	else
+		buffer = kasprintf(GFP_NOIO | __GFP_HIGH, "%s/%s", dir, name);
+	return (!buffer) ? ERR_PTR(-ENOMEM) : buffer;
+}
+
+static char **split(char *strings, unsigned int len, unsigned int *num)
+{
+	char *p, **ret;
+
+	/* Count the strings. */
+	*num = count_strings(strings, len) + 1;
+
+	/* Transfer to one big alloc for easy freeing. */
+	ret = kmalloc(*num * sizeof(char *) + len, GFP_NOIO | __GFP_HIGH);
+	if (!ret) {
+		kfree(strings);
+		return ERR_PTR(-ENOMEM);
+	}
+	memcpy(&ret[*num], strings, len);
+	kfree(strings);
+
+	strings = (char *)&ret[*num];
+	for (p = strings, *num = 0; p < strings + len; p += strlen(p) + 1)
+		ret[(*num)++] = p;
+	ret[*num] = strings + len;
+
+	return ret;
+}
+
+char **xenbus_directory(struct xenbus_transaction t,
+			const char *dir, const char *node, unsigned int *num)
+{
+	char *strings, *path;
+	unsigned int len;
+
+	path = join(dir, node);
+	if (IS_ERR(path))
+		return (char **)path;
+
+	strings = xs_single(t, XS_DIRECTORY, path, &len);
+	kfree(path);
+	if (IS_ERR(strings))
+		return (char **)strings;
+
+	return split(strings, len, num);
+}
+EXPORT_SYMBOL_GPL(xenbus_directory);
+
+/* Check if a path exists. Return 1 if it does. */
+int xenbus_exists(struct xenbus_transaction t,
+		  const char *dir, const char *node)
+{
+	char **d;
+	int dir_n;
+
+	d = xenbus_directory(t, dir, node, &dir_n);
+	if (IS_ERR(d))
+		return 0;
+	kfree(d);
+	return 1;
+}
+EXPORT_SYMBOL_GPL(xenbus_exists);
+
+/* Get the value of a single file.
+ * Returns a kmalloced value: call free() on it after use.
+ * len indicates length in bytes.
+ */
+void *xenbus_read(struct xenbus_transaction t,
+		  const char *dir, const char *node, unsigned int *len)
+{
+	char *path;
+	void *ret;
+
+	path = join(dir, node);
+	if (IS_ERR(path))
+		return (void *)path;
+
+	ret = xs_single(t, XS_READ, path, len);
+	kfree(path);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(xenbus_read);
+
+/* Write the value of a single file.
+ * Returns -err on failure.
+ */
+int xenbus_write(struct xenbus_transaction t,
+		 const char *dir, const char *node, const char *string)
+{
+	const char *path;
+	struct kvec iovec[2];
+	int ret;
+
+	path = join(dir, node);
+	if (IS_ERR(path))
+		return PTR_ERR(path);
+
+	iovec[0].iov_base = (void *)path;
+	iovec[0].iov_len = strlen(path) + 1;
+	iovec[1].iov_base = (void *)string;
+	iovec[1].iov_len = strlen(string);
+
+	ret = xs_error(xs_talkv(t, XS_WRITE, iovec, ARRAY_SIZE(iovec), NULL));
+	kfree(path);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(xenbus_write);
+
+/* Create a new directory. */
+int xenbus_mkdir(struct xenbus_transaction t,
+		 const char *dir, const char *node)
+{
+	char *path;
+	int ret;
+
+	path = join(dir, node);
+	if (IS_ERR(path))
+		return PTR_ERR(path);
+
+	ret = xs_error(xs_single(t, XS_MKDIR, path, NULL));
+	kfree(path);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(xenbus_mkdir);
+
+/* Destroy a file or directory (directories must be empty). */
+int xenbus_rm(struct xenbus_transaction t, const char *dir, const char *node)
+{
+	char *path;
+	int ret;
+
+	path = join(dir, node);
+	if (IS_ERR(path))
+		return PTR_ERR(path);
+
+	ret = xs_error(xs_single(t, XS_RM, path, NULL));
+	kfree(path);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(xenbus_rm);
+
+/* Start a transaction: changes by others will not be seen during this
+ * transaction, and changes will not be visible to others until end.
+ */
+int xenbus_transaction_start(struct xenbus_transaction *t)
+{
+	char *id_str;
+
+	down_read(&xs_state.transaction_mutex);
+
+	id_str = xs_single(XBT_NIL, XS_TRANSACTION_START, "", NULL);
+	if (IS_ERR(id_str)) {
+		up_read(&xs_state.transaction_mutex);
+		return PTR_ERR(id_str);
+	}
+
+	t->id = simple_strtoul(id_str, NULL, 0);
+	kfree(id_str);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(xenbus_transaction_start);
+
+/* End a transaction.
+ * If abandon is true, transaction is discarded instead of committed.
+ */
+int xenbus_transaction_end(struct xenbus_transaction t, int abort)
+{
+	char abortstr[2];
+	int err;
+
+	if (abort)
+		strcpy(abortstr, "F");
+	else
+		strcpy(abortstr, "T");
+
+	err = xs_error(xs_single(t, XS_TRANSACTION_END, abortstr, NULL));
+
+	up_read(&xs_state.transaction_mutex);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(xenbus_transaction_end);
+
+/* Single read and scanf: returns -errno or num scanned. */
+int xenbus_scanf(struct xenbus_transaction t,
+		 const char *dir, const char *node, const char *fmt, ...)
+{
+	va_list ap;
+	int ret;
+	char *val;
+
+	val = xenbus_read(t, dir, node, NULL);
+	if (IS_ERR(val))
+		return PTR_ERR(val);
+
+	va_start(ap, fmt);
+	ret = vsscanf(val, fmt, ap);
+	va_end(ap);
+	kfree(val);
+	/* Distinctive errno. */
+	if (ret == 0)
+		return -ERANGE;
+	return ret;
+}
+EXPORT_SYMBOL_GPL(xenbus_scanf);
+
+/* Single printf and write: returns -errno or 0. */
+int xenbus_printf(struct xenbus_transaction t,
+		  const char *dir, const char *node, const char *fmt, ...)
+{
+	va_list ap;
+	int ret;
+#define PRINTF_BUFFER_SIZE 4096
+	char *printf_buffer;
+
+	printf_buffer = kmalloc(PRINTF_BUFFER_SIZE, GFP_NOIO | __GFP_HIGH);
+	if (printf_buffer == NULL)
+		return -ENOMEM;
+
+	va_start(ap, fmt);
+	ret = vsnprintf(printf_buffer, PRINTF_BUFFER_SIZE, fmt, ap);
+	va_end(ap);
+
+	BUG_ON(ret > PRINTF_BUFFER_SIZE-1);
+	ret = xenbus_write(t, dir, node, printf_buffer);
+
+	kfree(printf_buffer);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(xenbus_printf);
+
+/* Takes tuples of names, scanf-style args, and void **, NULL terminated. */
+int xenbus_gather(struct xenbus_transaction t, const char *dir, ...)
+{
+	va_list ap;
+	const char *name;
+	int ret = 0;
+
+	va_start(ap, dir);
+	while (ret == 0 && (name = va_arg(ap, char *)) != NULL) {
+		const char *fmt = va_arg(ap, char *);
+		void *result = va_arg(ap, void *);
+		char *p;
+
+		p = xenbus_read(t, dir, name, NULL);
+		if (IS_ERR(p)) {
+			ret = PTR_ERR(p);
+			break;
+		}
+		if (fmt) {
+			if (sscanf(p, fmt, result) == 0)
+				ret = -EINVAL;
+			kfree(p);
+		} else
+			*(char **)result = p;
+	}
+	va_end(ap);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(xenbus_gather);
+
+static int xs_watch(const char *path, const char *token)
+{
+	struct kvec iov[2];
+
+	iov[0].iov_base = (void *)path;
+	iov[0].iov_len = strlen(path) + 1;
+	iov[1].iov_base = (void *)token;
+	iov[1].iov_len = strlen(token) + 1;
+
+	return xs_error(xs_talkv(XBT_NIL, XS_WATCH, iov,
+				 ARRAY_SIZE(iov), NULL));
+}
+
+static int xs_unwatch(const char *path, const char *token)
+{
+	struct kvec iov[2];
+
+	iov[0].iov_base = (char *)path;
+	iov[0].iov_len = strlen(path) + 1;
+	iov[1].iov_base = (char *)token;
+	iov[1].iov_len = strlen(token) + 1;
+
+	return xs_error(xs_talkv(XBT_NIL, XS_UNWATCH, iov,
+				 ARRAY_SIZE(iov), NULL));
+}
+
+static struct xenbus_watch *find_watch(const char *token)
+{
+	struct xenbus_watch *i, *cmp;
+
+	cmp = (void *)simple_strtoul(token, NULL, 16);
+
+	list_for_each_entry(i, &watches, list)
+		if (i == cmp)
+			return i;
+
+	return NULL;
+}
+
+/* Register callback to watch this node. */
+int register_xenbus_watch(struct xenbus_watch *watch)
+{
+	/* Pointer in ascii is the token. */
+	char token[sizeof(watch) * 2 + 1];
+	int err;
+
+	sprintf(token, "%lX", (long)watch);
+
+	down_read(&xs_state.watch_mutex);
+
+	spin_lock(&watches_lock);
+	BUG_ON(find_watch(token));
+	list_add(&watch->list, &watches);
+	spin_unlock(&watches_lock);
+
+	err = xs_watch(watch->node, token);
+
+	/* Ignore errors due to multiple registration. */
+	if ((err != 0) && (err != -EEXIST)) {
+		spin_lock(&watches_lock);
+		list_del(&watch->list);
+		spin_unlock(&watches_lock);
+	}
+
+	up_read(&xs_state.watch_mutex);
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(register_xenbus_watch);
+
+void unregister_xenbus_watch(struct xenbus_watch *watch)
+{
+	struct xs_stored_msg *msg, *tmp;
+	char token[sizeof(watch) * 2 + 1];
+	int err;
+
+	BUG_ON(watch->flags & XBWF_new_thread);
+
+	sprintf(token, "%lX", (long)watch);
+
+	down_read(&xs_state.watch_mutex);
+
+	spin_lock(&watches_lock);
+	BUG_ON(!find_watch(token));
+	list_del(&watch->list);
+	spin_unlock(&watches_lock);
+
+	err = xs_unwatch(watch->node, token);
+	if (err)
+		printk(KERN_WARNING
+		       "XENBUS Failed to release watch %s: %i\n",
+		       watch->node, err);
+
+	up_read(&xs_state.watch_mutex);
+
+	/* Cancel pending watch events. */
+	spin_lock(&watch_events_lock);
+	list_for_each_entry_safe(msg, tmp, &watch_events, list) {
+		if (msg->u.watch.handle != watch)
+			continue;
+		list_del(&msg->list);
+		kfree(msg->u.watch.vec);
+		kfree(msg);
+	}
+	spin_unlock(&watch_events_lock);
+
+	/* Flush any currently-executing callback, unless we are it. :-) */
+	if (current->pid != xenwatch_pid) {
+		mutex_lock(&xenwatch_mutex);
+		mutex_unlock(&xenwatch_mutex);
+	}
+}
+EXPORT_SYMBOL_GPL(unregister_xenbus_watch);
+
+void xs_suspend(void)
+{
+	down_write(&xs_state.transaction_mutex);
+	down_write(&xs_state.watch_mutex);
+	mutex_lock(&xs_state.request_mutex);
+	mutex_lock(&xs_state.response_mutex);
+}
+
+void xs_resume(void)
+{
+	struct xenbus_watch *watch;
+	char token[sizeof(watch) * 2 + 1];
+
+	mutex_unlock(&xs_state.response_mutex);
+	mutex_unlock(&xs_state.request_mutex);
+	up_write(&xs_state.transaction_mutex);
+
+	/* No need for watches_lock: the watch_mutex is sufficient. */
+	list_for_each_entry(watch, &watches, list) {
+		sprintf(token, "%lX", (long)watch);
+		xs_watch(watch->node, token);
+	}
+
+	up_write(&xs_state.watch_mutex);
+}
+
+void xs_suspend_cancel(void)
+{
+	mutex_unlock(&xs_state.response_mutex);
+	mutex_unlock(&xs_state.request_mutex);
+	up_write(&xs_state.watch_mutex);
+	up_write(&xs_state.transaction_mutex);
+}
+
+static int xenwatch_handle_callback(void *data)
+{
+	struct xs_stored_msg *msg = data;
+
+	msg->u.watch.handle->callback(msg->u.watch.handle,
+				      (const char **)msg->u.watch.vec,
+				      msg->u.watch.vec_size);
+
+	kfree(msg->u.watch.vec);
+	kfree(msg);
+
+	/* Kill this kthread if we were spawned just for this callback. */
+	if (current->pid != xenwatch_pid)
+		do_exit(0);
+
+	return 0;
+}
+
+static int xenwatch_thread(void *unused)
+{
+	struct list_head *ent;
+	struct xs_stored_msg *msg;
+
+	current->flags |= PF_NOFREEZE;
+	for (;;) {
+		wait_event_interruptible(watch_events_waitq,
+					 !list_empty(&watch_events));
+
+		if (kthread_should_stop())
+			break;
+
+		mutex_lock(&xenwatch_mutex);
+
+		spin_lock(&watch_events_lock);
+		ent = watch_events.next;
+		if (ent != &watch_events)
+			list_del(ent);
+		spin_unlock(&watch_events_lock);
+
+		if (ent == &watch_events) {
+			mutex_unlock(&xenwatch_mutex);
+			continue;
+		}
+
+		msg = list_entry(ent, struct xs_stored_msg, list);
+
+		/*
+		 * Unlock the mutex before running an XBWF_new_thread
+		 * handler. kthread_run can block which can deadlock
+		 * against unregister_xenbus_watch() if we need to
+		 * unregister other watches in order to make
+		 * progress. This can occur on resume before the swap
+		 * device is attached.
+		 */
+		if (msg->u.watch.handle->flags & XBWF_new_thread) {
+			mutex_unlock(&xenwatch_mutex);
+			kthread_run(xenwatch_handle_callback,
+				    msg, "xenwatch_cb");
+		} else {
+			xenwatch_handle_callback(msg);
+			mutex_unlock(&xenwatch_mutex);
+		}
+	}
+
+	return 0;
+}
+
+static int process_msg(void)
+{
+	struct xs_stored_msg *msg;
+	char *body;
+	int err;
+
+	/*
+	 * We must disallow save/restore while reading a xenstore message.
+	 * A partial read across s/r leaves us out of sync with xenstored.
+	 */
+	for (;;) {
+		err = xb_wait_for_data_to_read();
+		if (err)
+			return err;
+		mutex_lock(&xs_state.response_mutex);
+		if (xb_data_to_read())
+			break;
+		/* We raced with save/restore: pending data 'disappeared'. */
+		mutex_unlock(&xs_state.response_mutex);
+	}
+
+
+	msg = kmalloc(sizeof(*msg), GFP_NOIO | __GFP_HIGH);
+	if (msg == NULL) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	err = xb_read(&msg->hdr, sizeof(msg->hdr));
+	if (err) {
+		kfree(msg);
+		goto out;
+	}
+
+	body = kmalloc(msg->hdr.len + 1, GFP_NOIO | __GFP_HIGH);
+	if (body == NULL) {
+		kfree(msg);
+		err = -ENOMEM;
+		goto out;
+	}
+
+	err = xb_read(body, msg->hdr.len);
+	if (err) {
+		kfree(body);
+		kfree(msg);
+		goto out;
+	}
+	body[msg->hdr.len] = '\0';
+
+	if (msg->hdr.type == XS_WATCH_EVENT) {
+		msg->u.watch.vec = split(body, msg->hdr.len,
+					 &msg->u.watch.vec_size);
+		if (IS_ERR(msg->u.watch.vec)) {
+			err = PTR_ERR(msg->u.watch.vec);
+			kfree(msg);
+			goto out;
+		}
+
+		spin_lock(&watches_lock);
+		msg->u.watch.handle = find_watch(
+			msg->u.watch.vec[XS_WATCH_TOKEN]);
+		if (msg->u.watch.handle != NULL) {
+			spin_lock(&watch_events_lock);
+			list_add_tail(&msg->list, &watch_events);
+			wake_up(&watch_events_waitq);
+			spin_unlock(&watch_events_lock);
+		} else {
+			kfree(msg->u.watch.vec);
+			kfree(msg);
+		}
+		spin_unlock(&watches_lock);
+	} else {
+		msg->u.reply.body = body;
+		spin_lock(&xs_state.reply_lock);
+		list_add_tail(&msg->list, &xs_state.reply_list);
+		spin_unlock(&xs_state.reply_lock);
+		wake_up(&xs_state.reply_waitq);
+	}
+
+ out:
+	mutex_unlock(&xs_state.response_mutex);
+	return err;
+}
+
+static int xenbus_thread(void *unused)
+{
+	int err;
+
+	current->flags |= PF_NOFREEZE;
+	for (;;) {
+		err = process_msg();
+		if (err)
+			printk(KERN_WARNING "XENBUS error %d while reading "
+			       "message\n", err);
+		if (kthread_should_stop())
+			break;
+	}
+
+	return 0;
+}
+
+int xs_init(void)
+{
+	int err;
+	struct task_struct *task;
+
+	INIT_LIST_HEAD(&xs_state.reply_list);
+	spin_lock_init(&xs_state.reply_lock);
+	init_waitqueue_head(&xs_state.reply_waitq);
+
+	mutex_init(&xs_state.request_mutex);
+	mutex_init(&xs_state.response_mutex);
+	init_rwsem(&xs_state.transaction_mutex);
+	init_rwsem(&xs_state.watch_mutex);
+
+	/* Initialize the shared memory rings to talk to xenstored */
+	err = xb_init_comms();
+	if (err)
+		return err;
+
+	task = kthread_run(xenwatch_thread, NULL, "xenwatch");
+	if (IS_ERR(task))
+		return PTR_ERR(task);
+	xenwatch_pid = task->pid;
+
+	task = kthread_run(xenbus_thread, NULL, "xenbus");
+	if (IS_ERR(task))
+		return PTR_ERR(task);
+
+	return 0;
+}
diff -rpuN linux-2.6.18.8/drivers/xen/xenoprof/xenoprofile.c linux-2.6.18-xen-3.3.0/drivers/xen/xenoprof/xenoprofile.c
--- linux-2.6.18.8/drivers/xen/xenoprof/xenoprofile.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/drivers/xen/xenoprof/xenoprofile.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,546 @@
+/**
+ * @file xenoprofile.c
+ *
+ * @remark Copyright 2002 OProfile authors
+ * @remark Read the file COPYING
+ *
+ * @author John Levon <levon@movementarian.org>
+ *
+ * Modified by Aravind Menon and Jose Renato Santos for Xen
+ * These modifications are:
+ * Copyright (C) 2005 Hewlett-Packard Co.
+ *
+ * Separated out arch-generic part
+ * Copyright (c) 2006 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ */
+
+#include <linux/init.h>
+#include <linux/notifier.h>
+#include <linux/smp.h>
+#include <linux/oprofile.h>
+#include <linux/sysdev.h>
+#include <linux/slab.h>
+#include <linux/interrupt.h>
+#include <linux/vmalloc.h>
+#include <asm/pgtable.h>
+#include <xen/evtchn.h>
+#include <xen/xenoprof.h>
+#include <xen/driver_util.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/xenoprof.h>
+#include "../../../drivers/oprofile/cpu_buffer.h"
+#include "../../../drivers/oprofile/event_buffer.h"
+
+#define MAX_XENOPROF_SAMPLES 16
+
+/* sample buffers shared with Xen */
+xenoprof_buf_t * xenoprof_buf[MAX_VIRT_CPUS];
+/* Shared buffer area */
+struct xenoprof_shared_buffer shared_buffer;
+
+/* Passive sample buffers shared with Xen */
+xenoprof_buf_t *p_xenoprof_buf[MAX_OPROF_DOMAINS][MAX_VIRT_CPUS];
+/* Passive shared buffer area */
+struct xenoprof_shared_buffer p_shared_buffer[MAX_OPROF_DOMAINS];
+
+static int xenoprof_start(void);
+static void xenoprof_stop(void);
+
+static int xenoprof_enabled = 0;
+static int xenoprof_is_primary = 0;
+static int active_defined;
+
+extern unsigned long backtrace_depth;
+
+/* Number of buffers in shared area (one per VCPU) */
+int nbuf;
+/* Mappings of VIRQ_XENOPROF to irq number (per cpu) */
+int ovf_irq[NR_CPUS];
+/* cpu model type string - copied from Xen on XENOPROF_init command */
+char cpu_type[XENOPROF_CPU_TYPE_SIZE];
+
+#ifdef CONFIG_PM
+
+static int xenoprof_suspend(struct sys_device * dev, pm_message_t state)
+{
+	if (xenoprof_enabled == 1)
+		xenoprof_stop();
+	return 0;
+}
+
+
+static int xenoprof_resume(struct sys_device * dev)
+{
+	if (xenoprof_enabled == 1)
+		xenoprof_start();
+	return 0;
+}
+
+
+static struct sysdev_class oprofile_sysclass = {
+	set_kset_name("oprofile"),
+	.resume		= xenoprof_resume,
+	.suspend	= xenoprof_suspend
+};
+
+
+static struct sys_device device_oprofile = {
+	.id	= 0,
+	.cls	= &oprofile_sysclass,
+};
+
+
+static int __init init_driverfs(void)
+{
+	int error;
+	if (!(error = sysdev_class_register(&oprofile_sysclass)))
+		error = sysdev_register(&device_oprofile);
+	return error;
+}
+
+
+static void exit_driverfs(void)
+{
+	sysdev_unregister(&device_oprofile);
+	sysdev_class_unregister(&oprofile_sysclass);
+}
+
+#else
+#define init_driverfs() do { } while (0)
+#define exit_driverfs() do { } while (0)
+#endif /* CONFIG_PM */
+
+unsigned long long oprofile_samples = 0;
+unsigned long long p_oprofile_samples = 0;
+
+unsigned int pdomains;
+struct xenoprof_passive passive_domains[MAX_OPROF_DOMAINS];
+
+/* Check whether the given entry is an escape code */
+static int xenoprof_is_escape(xenoprof_buf_t * buf, int tail)
+{
+	return (buf->event_log[tail].eip == XENOPROF_ESCAPE_CODE);
+}
+
+/* Get the event at the given entry  */
+static uint8_t xenoprof_get_event(xenoprof_buf_t * buf, int tail)
+{
+	return (buf->event_log[tail].event);
+}
+
+static void xenoprof_add_pc(xenoprof_buf_t *buf, int is_passive)
+{
+	int head, tail, size;
+	int tracing = 0;
+
+	head = buf->event_head;
+	tail = buf->event_tail;
+	size = buf->event_size;
+
+	while (tail != head) {
+		if (xenoprof_is_escape(buf, tail) &&
+		    xenoprof_get_event(buf, tail) == XENOPROF_TRACE_BEGIN) {
+			tracing=1;
+			oprofile_add_pc(ESCAPE_CODE, buf->event_log[tail].mode, 
+					CPU_TRACE_BEGIN); 
+			if (!is_passive)
+				oprofile_samples++;
+			else
+				p_oprofile_samples++;
+			
+		} else {
+			oprofile_add_pc(buf->event_log[tail].eip,
+					buf->event_log[tail].mode,
+					buf->event_log[tail].event);
+			if (!tracing) {
+				if (!is_passive)
+					oprofile_samples++;
+				else
+					p_oprofile_samples++;
+			}
+       
+		}
+		tail++;
+		if(tail==size)
+		    tail=0;
+	}
+	buf->event_tail = tail;
+}
+
+static void xenoprof_handle_passive(void)
+{
+	int i, j;
+	int flag_domain, flag_switch = 0;
+	
+	for (i = 0; i < pdomains; i++) {
+		flag_domain = 0;
+		for (j = 0; j < passive_domains[i].nbuf; j++) {
+			xenoprof_buf_t *buf = p_xenoprof_buf[i][j];
+			if (buf->event_head == buf->event_tail)
+				continue;
+			if (!flag_domain) {
+				if (!oprofile_add_domain_switch(
+					passive_domains[i].domain_id))
+					goto done;
+				flag_domain = 1;
+			}
+			xenoprof_add_pc(buf, 1);
+			flag_switch = 1;
+		}
+	}
+done:
+	if (flag_switch)
+		oprofile_add_domain_switch(COORDINATOR_DOMAIN);
+}
+
+static irqreturn_t 
+xenoprof_ovf_interrupt(int irq, void * dev_id, struct pt_regs * regs)
+{
+	struct xenoprof_buf * buf;
+	static unsigned long flag;
+
+	buf = xenoprof_buf[smp_processor_id()];
+
+	xenoprof_add_pc(buf, 0);
+
+	if (xenoprof_is_primary && !test_and_set_bit(0, &flag)) {
+		xenoprof_handle_passive();
+		smp_mb__before_clear_bit();
+		clear_bit(0, &flag);
+	}
+
+	return IRQ_HANDLED;
+}
+
+
+static void unbind_virq(void)
+{
+	unsigned int i;
+
+	for_each_online_cpu(i) {
+		if (ovf_irq[i] >= 0) {
+			unbind_from_irqhandler(ovf_irq[i], NULL);
+			ovf_irq[i] = -1;
+		}
+	}
+}
+
+
+static int bind_virq(void)
+{
+	unsigned int i;
+	int result;
+
+	for_each_online_cpu(i) {
+		result = bind_virq_to_irqhandler(VIRQ_XENOPROF,
+						 i,
+						 xenoprof_ovf_interrupt,
+						 SA_INTERRUPT,
+						 "xenoprof",
+						 NULL);
+
+		if (result < 0) {
+			unbind_virq();
+			return result;
+		}
+
+		ovf_irq[i] = result;
+	}
+		
+	return 0;
+}
+
+
+static void unmap_passive_list(void)
+{
+	int i;
+	for (i = 0; i < pdomains; i++)
+		xenoprof_arch_unmap_shared_buffer(&p_shared_buffer[i]);
+	pdomains = 0;
+}
+
+
+static int map_xenoprof_buffer(int max_samples)
+{
+	struct xenoprof_get_buffer get_buffer;
+	struct xenoprof_buf *buf;
+	int ret, i;
+
+	if ( shared_buffer.buffer )
+		return 0;
+
+	get_buffer.max_samples = max_samples;
+	ret = xenoprof_arch_map_shared_buffer(&get_buffer, &shared_buffer);
+	if (ret)
+		return ret;
+	nbuf = get_buffer.nbuf;
+
+	for (i=0; i< nbuf; i++) {
+		buf = (struct xenoprof_buf*) 
+			&shared_buffer.buffer[i * get_buffer.bufsize];
+		BUG_ON(buf->vcpu_id >= MAX_VIRT_CPUS);
+		xenoprof_buf[buf->vcpu_id] = buf;
+	}
+
+	return 0;
+}
+
+
+static int xenoprof_setup(void)
+{
+	int ret;
+
+	if ( (ret = map_xenoprof_buffer(MAX_XENOPROF_SAMPLES)) )
+		return ret;
+
+	if ( (ret = bind_virq()) )
+		return ret;
+
+	if (xenoprof_is_primary) {
+		/* Define dom0 as an active domain if not done yet */
+		if (!active_defined) {
+			domid_t domid;
+			ret = HYPERVISOR_xenoprof_op(
+				XENOPROF_reset_active_list, NULL);
+			if (ret)
+				goto err;
+			domid = 0;
+			ret = HYPERVISOR_xenoprof_op(
+				XENOPROF_set_active, &domid);
+			if (ret)
+				goto err;
+			active_defined = 1;
+		}
+
+		if (backtrace_depth > 0) {
+			ret = HYPERVISOR_xenoprof_op(XENOPROF_set_backtrace, 
+						     &backtrace_depth);
+			if (ret)
+				backtrace_depth = 0;
+		}
+
+		ret = HYPERVISOR_xenoprof_op(XENOPROF_reserve_counters, NULL);
+		if (ret)
+			goto err;
+		
+		xenoprof_arch_counter();
+		ret = HYPERVISOR_xenoprof_op(XENOPROF_setup_events, NULL);
+		if (ret)
+			goto err;
+	}
+
+	ret = HYPERVISOR_xenoprof_op(XENOPROF_enable_virq, NULL);
+	if (ret)
+		goto err;
+
+	xenoprof_enabled = 1;
+	return 0;
+ err:
+	unbind_virq();
+	return ret;
+}
+
+
+static void xenoprof_shutdown(void)
+{
+	xenoprof_enabled = 0;
+
+	WARN_ON(HYPERVISOR_xenoprof_op(XENOPROF_disable_virq, NULL));
+
+	if (xenoprof_is_primary) {
+		WARN_ON(HYPERVISOR_xenoprof_op(XENOPROF_release_counters,
+					       NULL));
+		active_defined = 0;
+	}
+
+	unbind_virq();
+
+	xenoprof_arch_unmap_shared_buffer(&shared_buffer);
+	if (xenoprof_is_primary)
+		unmap_passive_list();
+}
+
+
+static int xenoprof_start(void)
+{
+	int ret = 0;
+
+	if (xenoprof_is_primary)
+		ret = HYPERVISOR_xenoprof_op(XENOPROF_start, NULL);
+	if (!ret)
+		xenoprof_arch_start();
+	return ret;
+}
+
+
+static void xenoprof_stop(void)
+{
+	if (xenoprof_is_primary)
+		WARN_ON(HYPERVISOR_xenoprof_op(XENOPROF_stop, NULL));
+	xenoprof_arch_stop();
+}
+
+
+static int xenoprof_set_active(int * active_domains,
+			       unsigned int adomains)
+{
+	int ret = 0;
+	int i;
+	int set_dom0 = 0;
+	domid_t domid;
+
+	if (!xenoprof_is_primary)
+		return 0;
+
+	if (adomains > MAX_OPROF_DOMAINS)
+		return -E2BIG;
+
+	ret = HYPERVISOR_xenoprof_op(XENOPROF_reset_active_list, NULL);
+	if (ret)
+		return ret;
+
+	for (i=0; i<adomains; i++) {
+		domid = active_domains[i];
+		if (domid != active_domains[i]) {
+			ret = -EINVAL;
+			goto out;
+		}
+		ret = HYPERVISOR_xenoprof_op(XENOPROF_set_active, &domid);
+		if (ret)
+			goto out;
+		if (active_domains[i] == 0)
+			set_dom0 = 1;
+	}
+	/* dom0 must always be active but may not be in the list */ 
+	if (!set_dom0) {
+		domid = 0;
+		ret = HYPERVISOR_xenoprof_op(XENOPROF_set_active, &domid);
+	}
+
+out:
+	if (ret)
+		WARN_ON(HYPERVISOR_xenoprof_op(XENOPROF_reset_active_list,
+					       NULL));
+	active_defined = !ret;
+	return ret;
+}
+
+static int xenoprof_set_passive(int * p_domains,
+                                unsigned int pdoms)
+{
+	int ret;
+	unsigned int i, j;
+	struct xenoprof_buf *buf;
+
+	if (!xenoprof_is_primary)
+        	return 0;
+
+	if (pdoms > MAX_OPROF_DOMAINS)
+		return -E2BIG;
+
+	ret = HYPERVISOR_xenoprof_op(XENOPROF_reset_passive_list, NULL);
+	if (ret)
+		return ret;
+	unmap_passive_list();
+
+	for (i = 0; i < pdoms; i++) {
+		passive_domains[i].domain_id = p_domains[i];
+		passive_domains[i].max_samples = 2048;
+		ret = xenoprof_arch_set_passive(&passive_domains[i],
+						&p_shared_buffer[i]);
+		if (ret)
+			goto out;
+		for (j = 0; j < passive_domains[i].nbuf; j++) {
+			buf = (struct xenoprof_buf *)
+				&p_shared_buffer[i].buffer[
+				j * passive_domains[i].bufsize];
+			BUG_ON(buf->vcpu_id >= MAX_VIRT_CPUS);
+			p_xenoprof_buf[i][buf->vcpu_id] = buf;
+		}
+	}
+
+	pdomains = pdoms;
+	return 0;
+
+out:
+	for (j = 0; j < i; j++)
+		xenoprof_arch_unmap_shared_buffer(&p_shared_buffer[i]);
+
+	return ret;
+}
+
+
+/* The dummy backtrace function to keep oprofile happy
+ * The real backtrace is done in xen
+ */
+static void xenoprof_dummy_backtrace(struct pt_regs * const regs, 
+				     unsigned int depth)
+{
+	/* this should never be called */
+	BUG();
+	return;
+}
+
+
+
+struct oprofile_operations xenoprof_ops = {
+#ifdef HAVE_XENOPROF_CREATE_FILES
+	.create_files 	= xenoprof_create_files,
+#endif
+	.set_active	= xenoprof_set_active,
+	.set_passive    = xenoprof_set_passive,
+	.setup 		= xenoprof_setup,
+	.shutdown	= xenoprof_shutdown,
+	.start		= xenoprof_start,
+	.stop		= xenoprof_stop,
+	.backtrace	= xenoprof_dummy_backtrace
+};
+
+
+/* in order to get driverfs right */
+static int using_xenoprof;
+
+int __init xenoprofile_init(struct oprofile_operations * ops)
+{
+	struct xenoprof_init init;
+	unsigned int i;
+	int ret;
+
+	ret = HYPERVISOR_xenoprof_op(XENOPROF_init, &init);
+	if (!ret) {
+		xenoprof_arch_init_counter(&init);
+		xenoprof_is_primary = init.is_primary;
+
+		/*  cpu_type is detected by Xen */
+		cpu_type[XENOPROF_CPU_TYPE_SIZE-1] = 0;
+		strncpy(cpu_type, init.cpu_type, XENOPROF_CPU_TYPE_SIZE - 1);
+		xenoprof_ops.cpu_type = cpu_type;
+
+		init_driverfs();
+		using_xenoprof = 1;
+		*ops = xenoprof_ops;
+
+		for (i=0; i<NR_CPUS; i++)
+			ovf_irq[i] = -1;
+
+		active_defined = 0;
+	}
+
+	printk(KERN_INFO "%s: ret %d, events %d, xenoprof_is_primary %d\n",
+	       __func__, ret, init.num_events, xenoprof_is_primary);
+	return ret;
+}
+
+
+void xenoprofile_exit(void)
+{
+	if (using_xenoprof)
+		exit_driverfs();
+
+	xenoprof_arch_unmap_shared_buffer(&shared_buffer);
+	if (xenoprof_is_primary) {
+		unmap_passive_list();
+		WARN_ON(HYPERVISOR_xenoprof_op(XENOPROF_shutdown, NULL));
+        }
+}
diff -rpuN linux-2.6.18.8/fs/aio.c linux-2.6.18-xen-3.3.0/fs/aio.c
--- linux-2.6.18.8/fs/aio.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/fs/aio.c	2008-08-21 11:36:07.000000000 +0200
@@ -34,6 +34,11 @@
 #include <asm/uaccess.h>
 #include <asm/mmu_context.h>
 
+#ifdef CONFIG_EPOLL
+#include <linux/poll.h>
+#include <linux/eventpoll.h>
+#endif
+
 #if DEBUG > 1
 #define dprintk		printk
 #else
@@ -1015,6 +1020,10 @@ put_rq:
 	if (waitqueue_active(&ctx->wait))
 		wake_up(&ctx->wait);
 
+#ifdef CONFIG_EPOLL
+	if (ctx->file && waitqueue_active(&ctx->poll_wait))
+		wake_up(&ctx->poll_wait);
+#endif
 	if (ret)
 		put_ioctx(ctx);
 
@@ -1024,6 +1033,8 @@ put_rq:
 /* aio_read_evt
  *	Pull an event off of the ioctx's event ring.  Returns the number of 
  *	events fetched (0 or 1 ;-)
+ *	If ent parameter is 0, just returns the number of events that would
+ *	be fetched.
  *	FIXME: make this use cmpxchg.
  *	TODO: make the ringbuffer user mmap()able (requires FIXME).
  */
@@ -1046,13 +1057,18 @@ static int aio_read_evt(struct kioctx *i
 
 	head = ring->head % info->nr;
 	if (head != ring->tail) {
-		struct io_event *evp = aio_ring_event(info, head, KM_USER1);
-		*ent = *evp;
-		head = (head + 1) % info->nr;
-		smp_mb(); /* finish reading the event before updatng the head */
-		ring->head = head;
-		ret = 1;
-		put_aio_ring_event(evp, KM_USER1);
+		if (ent) { /* event requested */
+			struct io_event *evp =
+				aio_ring_event(info, head, KM_USER1);
+			*ent = *evp;
+			head = (head + 1) % info->nr;
+			/* finish reading the event before updatng the head */
+			smp_mb();
+			ring->head = head;
+			ret = 1;
+			put_aio_ring_event(evp, KM_USER1);
+		} else /* only need to know availability */
+			ret = 1;
 	}
 	spin_unlock(&info->ring_lock);
 
@@ -1235,9 +1251,78 @@ static void io_destroy(struct kioctx *io
 
 	aio_cancel_all(ioctx);
 	wait_for_all_aios(ioctx);
+#ifdef CONFIG_EPOLL
+	/* forget the poll file, but it's up to the user to close it */
+	if (ioctx->file) {
+		ioctx->file->private_data = 0;
+		ioctx->file = 0;
+	}
+#endif
 	put_ioctx(ioctx);	/* once for the lookup */
 }
 
+#ifdef CONFIG_EPOLL
+
+static int aio_queue_fd_close(struct inode *inode, struct file *file)
+{
+	struct kioctx *ioctx = file->private_data;
+	if (ioctx) {
+		file->private_data = 0;
+		spin_lock_irq(&ioctx->ctx_lock);
+		ioctx->file = 0;
+		spin_unlock_irq(&ioctx->ctx_lock);
+	}
+	return 0;
+}
+
+static unsigned int aio_queue_fd_poll(struct file *file, poll_table *wait)
+{	unsigned int pollflags = 0;
+	struct kioctx *ioctx = file->private_data;
+
+	if (ioctx) {
+
+		spin_lock_irq(&ioctx->ctx_lock);
+		/* Insert inside our poll wait queue */
+		poll_wait(file, &ioctx->poll_wait, wait);
+
+		/* Check our condition */
+		if (aio_read_evt(ioctx, 0))
+			pollflags = POLLIN | POLLRDNORM;
+		spin_unlock_irq(&ioctx->ctx_lock);
+	}
+
+	return pollflags;
+}
+
+static const struct file_operations aioq_fops = {
+	.release	= aio_queue_fd_close,
+	.poll		= aio_queue_fd_poll
+};
+
+/* make_aio_fd:
+ *  Create a file descriptor that can be used to poll the event queue.
+ *  Based and piggybacked on the excellent epoll code.
+ */
+
+static int make_aio_fd(struct kioctx *ioctx)
+{
+	int error, fd;
+	struct inode *inode;
+	struct file *file;
+
+	error = ep_getfd(&fd, &inode, &file, NULL, &aioq_fops);
+	if (error)
+		return error;
+
+	/* associate the file with the IO context */
+	file->private_data = ioctx;
+	ioctx->file = file;
+	init_waitqueue_head(&ioctx->poll_wait);
+	return fd;
+}
+#endif
+
+
 /* sys_io_setup:
  *	Create an aio_context capable of receiving at least nr_events.
  *	ctxp must not point to an aio_context that already exists, and
@@ -1250,18 +1335,30 @@ static void io_destroy(struct kioctx *io
  *	resources are available.  May fail with -EFAULT if an invalid
  *	pointer is passed for ctxp.  Will fail with -ENOSYS if not
  *	implemented.
+ *
+ *	To request a selectable fd, the user context has to be initialized
+ *	to 1, instead of 0, and the return value is the fd.
+ *	This keeps the system call compatible, since a non-zero value
+ *	was not allowed so far.
  */
 asmlinkage long sys_io_setup(unsigned nr_events, aio_context_t __user *ctxp)
 {
 	struct kioctx *ioctx = NULL;
 	unsigned long ctx;
 	long ret;
+	int make_fd = 0;
 
 	ret = get_user(ctx, ctxp);
 	if (unlikely(ret))
 		goto out;
 
 	ret = -EINVAL;
+#ifdef CONFIG_EPOLL
+	if (ctx == 1) {
+		make_fd = 1;
+		ctx = 0;
+	}
+#endif
 	if (unlikely(ctx || nr_events == 0)) {
 		pr_debug("EINVAL: io_setup: ctx %lu nr_events %u\n",
 		         ctx, nr_events);
@@ -1272,8 +1369,12 @@ asmlinkage long sys_io_setup(unsigned nr
 	ret = PTR_ERR(ioctx);
 	if (!IS_ERR(ioctx)) {
 		ret = put_user(ioctx->user_id, ctxp);
-		if (!ret)
-			return 0;
+#ifdef CONFIG_EPOLL
+		if (make_fd && ret >= 0)
+			ret = make_aio_fd(ioctx);
+#endif
+		if (ret >= 0)
+			return ret;
 
 		get_ioctx(ioctx); /* io_destroy() expects us to hold a ref */
 		io_destroy(ioctx);
diff -rpuN linux-2.6.18.8/fs/binfmt_elf.c linux-2.6.18-xen-3.3.0/fs/binfmt_elf.c
--- linux-2.6.18.8/fs/binfmt_elf.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/fs/binfmt_elf.c	2008-08-21 11:36:07.000000000 +0200
@@ -1170,6 +1170,10 @@ static int dump_seek(struct file *file, 
  */
 static int maydump(struct vm_area_struct *vma)
 {
+	/* The vma can be set up to tell us the answer directly.  */
+	if (vma->vm_flags & VM_ALWAYSDUMP)
+		return 1;
+
 	/* Do not dump I/O mapped devices or special mappings */
 	if (vma->vm_flags & (VM_IO | VM_RESERVED))
 		return 0;
diff -rpuN linux-2.6.18.8/fs/bio.c linux-2.6.18-xen-3.3.0/fs/bio.c
--- linux-2.6.18.8/fs/bio.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/fs/bio.c	2008-08-21 11:36:07.000000000 +0200
@@ -112,7 +112,8 @@ void bio_free(struct bio *bio, struct bi
 
 	BIO_BUG_ON(pool_idx >= BIOVEC_NR_POOLS);
 
-	mempool_free(bio->bi_io_vec, bio_set->bvec_pools[pool_idx]);
+	if (bio->bi_io_vec)
+		mempool_free(bio->bi_io_vec, bio_set->bvec_pools[pool_idx]);
 	mempool_free(bio, bio_set->bio_pool);
 }
 
diff -rpuN linux-2.6.18.8/fs/compat_ioctl.c linux-2.6.18-xen-3.3.0/fs/compat_ioctl.c
--- linux-2.6.18.8/fs/compat_ioctl.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/fs/compat_ioctl.c	2008-08-21 11:36:07.000000000 +0200
@@ -124,6 +124,13 @@
 #include <linux/dvb/video.h>
 #include <linux/lp.h>
 
+#ifdef CONFIG_XEN
+#include <xen/interface/xen.h>
+#include <xen/public/evtchn.h>
+#include <xen/public/privcmd.h>
+#include <xen/compat_ioctl.h>
+#endif
+
 /* Aiee. Someone does not find a difference between int and long */
 #define EXT2_IOC32_GETFLAGS               _IOR('f', 1, int)
 #define EXT2_IOC32_SETFLAGS               _IOW('f', 2, int)
@@ -2948,6 +2955,18 @@ COMPATIBLE_IOCTL(LPRESET)
 /*LPGETSTATS not implemented, but no kernels seem to compile it in anyways*/
 COMPATIBLE_IOCTL(LPGETFLAGS)
 HANDLE_IOCTL(LPSETTIMEOUT, lp_timeout_trans)
+
+#ifdef CONFIG_XEN
+HANDLE_IOCTL(IOCTL_PRIVCMD_MMAP_32, privcmd_ioctl_32)
+HANDLE_IOCTL(IOCTL_PRIVCMD_MMAPBATCH_32, privcmd_ioctl_32)
+COMPATIBLE_IOCTL(IOCTL_PRIVCMD_HYPERCALL)
+COMPATIBLE_IOCTL(IOCTL_EVTCHN_BIND_VIRQ)
+COMPATIBLE_IOCTL(IOCTL_EVTCHN_BIND_INTERDOMAIN)
+COMPATIBLE_IOCTL(IOCTL_EVTCHN_BIND_UNBOUND_PORT)
+COMPATIBLE_IOCTL(IOCTL_EVTCHN_UNBIND)
+COMPATIBLE_IOCTL(IOCTL_EVTCHN_NOTIFY)
+COMPATIBLE_IOCTL(IOCTL_EVTCHN_RESET)
+#endif
 };
 
 int ioctl_table_size = ARRAY_SIZE(ioctl_start);
diff -rpuN linux-2.6.18.8/fs/debugfs/file.c linux-2.6.18-xen-3.3.0/fs/debugfs/file.c
--- linux-2.6.18.8/fs/debugfs/file.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/fs/debugfs/file.c	2008-08-21 11:36:07.000000000 +0200
@@ -170,6 +170,51 @@ struct dentry *debugfs_create_u32(const 
 }
 EXPORT_SYMBOL_GPL(debugfs_create_u32);
 
+
+static void debugfs_u64_set(void *data, u64 val)
+{
+	*(u64 *)data = val;
+}
+
+static u64 debugfs_u64_get(void *data)
+{
+	return *(u64 *)data;
+}
+DEFINE_SIMPLE_ATTRIBUTE(fops_u64, debugfs_u64_get, debugfs_u64_set, "%llu\n");
+
+/**
+ * debugfs_create_u64 - create a file in the debugfs filesystem that is used to read and write an unsigned 64 bit value.
+ *
+ * @name: a pointer to a string containing the name of the file to create.
+ * @mode: the permission that the file should have
+ * @parent: a pointer to the parent dentry for this file.  This should be a
+ *          directory dentry if set.  If this paramater is NULL, then the
+ *          file will be created in the root of the debugfs filesystem.
+ * @value: a pointer to the variable that the file should read to and write
+ *         from.
+ *
+ * This function creates a file in debugfs with the given name that
+ * contains the value of the variable @value.  If the @mode variable is so
+ * set, it can be read from, and written to.
+ *
+ * This function will return a pointer to a dentry if it succeeds.  This
+ * pointer must be passed to the debugfs_remove() function when the file is
+ * to be removed (no automatic cleanup happens if your module is unloaded,
+ * you are responsible here.)  If an error occurs, NULL will be returned.
+ *
+ * If debugfs is not enabled in the kernel, the value -ENODEV will be
+ * returned.  It is not wise to check for this value, but rather, check for
+ * NULL or !NULL instead as to eliminate the need for #ifdef in the calling
+ * code.
+ */
+struct dentry *debugfs_create_u64(const char *name, mode_t mode,
+				  struct dentry *parent, u64 *value)
+{
+	return debugfs_create_file(name, mode, parent, value, &fops_u64);
+}
+EXPORT_SYMBOL_GPL(debugfs_create_u64);
+
+
 static ssize_t read_file_bool(struct file *file, char __user *user_buf,
 			      size_t count, loff_t *ppos)
 {
diff -rpuN linux-2.6.18.8/fs/eventpoll.c linux-2.6.18-xen-3.3.0/fs/eventpoll.c
--- linux-2.6.18.8/fs/eventpoll.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/fs/eventpoll.c	2008-08-21 11:36:07.000000000 +0200
@@ -236,8 +236,6 @@ struct ep_pqueue {
 
 static void ep_poll_safewake_init(struct poll_safewake *psw);
 static void ep_poll_safewake(struct poll_safewake *psw, wait_queue_head_t *wq);
-static int ep_getfd(int *efd, struct inode **einode, struct file **efile,
-		    struct eventpoll *ep);
 static int ep_alloc(struct eventpoll **pep);
 static void ep_free(struct eventpoll *ep);
 static struct epitem *ep_find(struct eventpoll *ep, struct file *file, int fd);
@@ -267,7 +265,7 @@ static int ep_events_transfer(struct eve
 static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events,
 		   int maxevents, long timeout);
 static int eventpollfs_delete_dentry(struct dentry *dentry);
-static struct inode *ep_eventpoll_inode(void);
+static struct inode *ep_eventpoll_inode(const struct file_operations *fops);
 static int eventpollfs_get_sb(struct file_system_type *fs_type,
 			      int flags, const char *dev_name,
 			      void *data, struct vfsmount *mnt);
@@ -517,7 +515,7 @@ asmlinkage long sys_epoll_create(int siz
 	 * Creates all the items needed to setup an eventpoll file. That is,
 	 * a file structure, and inode and a free file descriptor.
 	 */
-	error = ep_getfd(&fd, &inode, &file, ep);
+	error = ep_getfd(&fd, &inode, &file, ep, &eventpoll_fops);
 	if (error)
 		goto eexit_2;
 
@@ -702,8 +700,8 @@ eexit_1:
 /*
  * Creates the file descriptor to be used by the epoll interface.
  */
-static int ep_getfd(int *efd, struct inode **einode, struct file **efile,
-		    struct eventpoll *ep)
+int ep_getfd(int *efd, struct inode **einode, struct file **efile,
+		    struct eventpoll *ep, const struct file_operations *fops)
 {
 	struct qstr this;
 	char name[32];
@@ -719,7 +717,7 @@ static int ep_getfd(int *efd, struct ino
 		goto eexit_1;
 
 	/* Allocates an inode from the eventpoll file system */
-	inode = ep_eventpoll_inode();
+	inode = ep_eventpoll_inode(fops);
 	error = PTR_ERR(inode);
 	if (IS_ERR(inode))
 		goto eexit_2;
@@ -750,7 +748,7 @@ static int ep_getfd(int *efd, struct ino
 
 	file->f_pos = 0;
 	file->f_flags = O_RDONLY;
-	file->f_op = &eventpoll_fops;
+	file->f_op = fops;
 	file->f_mode = FMODE_READ;
 	file->f_version = 0;
 	file->private_data = ep;
@@ -1569,7 +1567,7 @@ static int eventpollfs_delete_dentry(str
 }
 
 
-static struct inode *ep_eventpoll_inode(void)
+static struct inode *ep_eventpoll_inode(const struct file_operations *fops)
 {
 	int error = -ENOMEM;
 	struct inode *inode = new_inode(eventpoll_mnt->mnt_sb);
@@ -1577,7 +1575,7 @@ static struct inode *ep_eventpoll_inode(
 	if (!inode)
 		goto eexit_1;
 
-	inode->i_fop = &eventpoll_fops;
+	inode->i_fop = fops;
 
 	/*
 	 * Mark the inode dirty from the very beginning,
diff -rpuN linux-2.6.18.8/fs/Kconfig linux-2.6.18-xen-3.3.0/fs/Kconfig
--- linux-2.6.18.8/fs/Kconfig	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/fs/Kconfig	2008-08-21 11:36:07.000000000 +0200
@@ -826,6 +826,11 @@ config PROC_VMCORE
         help
         Exports the dump image of crashed kernel in ELF format.
 
+config PROC_IOMEM_MACHINE
+	bool
+	depends on PROC_FS && EXPERIMENTAL && KEXEC && XEN && IA64
+	default y
+
 config SYSFS
 	bool "sysfs file system support" if EMBEDDED
 	default y
@@ -865,6 +870,7 @@ config TMPFS
 config HUGETLBFS
 	bool "HugeTLB file system support"
 	depends X86 || IA64 || PPC64 || SPARC64 || SUPERH || BROKEN
+	depends on !XEN
 	help
 	  hugetlbfs is a filesystem backing for HugeTLB pages, based on
 	  ramfs. For architectures that support it, say Y here and read
diff -rpuN linux-2.6.18.8/fs/lockd/svc.c linux-2.6.18-xen-3.3.0/fs/lockd/svc.c
--- linux-2.6.18.8/fs/lockd/svc.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/fs/lockd/svc.c	2008-08-21 11:36:07.000000000 +0200
@@ -323,9 +323,6 @@ EXPORT_SYMBOL(lockd_down);
  * Sysctl parameters (same as module parameters, different interface).
  */
 
-/* Something that isn't CTL_ANY, CTL_NONE or a value that may clash. */
-#define CTL_UNNUMBERED		-2
-
 static ctl_table nlm_sysctls[] = {
 	{
 		.ctl_name	= CTL_UNNUMBERED,
diff -rpuN linux-2.6.18.8/fs/nfs/sysctl.c linux-2.6.18-xen-3.3.0/fs/nfs/sysctl.c
--- linux-2.6.18.8/fs/nfs/sysctl.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/fs/nfs/sysctl.c	2008-08-21 11:36:07.000000000 +0200
@@ -18,11 +18,6 @@
 static const int nfs_set_port_min = 0;
 static const int nfs_set_port_max = 65535;
 static struct ctl_table_header *nfs_callback_sysctl_table;
-/*
- * Something that isn't CTL_ANY, CTL_NONE or a value that may clash.
- * Use the same values as fs/lockd/svc.c
- */
-#define CTL_UNNUMBERED -2
 
 static ctl_table nfs_cb_sysctls[] = {
 #ifdef CONFIG_NFS_V4
diff -rpuN linux-2.6.18.8/fs/proc/proc_misc.c linux-2.6.18-xen-3.3.0/fs/proc/proc_misc.c
--- linux-2.6.18.8/fs/proc/proc_misc.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/fs/proc/proc_misc.c	2008-08-21 11:36:07.000000000 +0200
@@ -471,7 +471,7 @@ static int show_stat(struct seq_file *p,
 		(unsigned long long)cputime64_to_clock_t(irq),
 		(unsigned long long)cputime64_to_clock_t(softirq),
 		(unsigned long long)cputime64_to_clock_t(steal));
-	for_each_online_cpu(i) {
+	for_each_possible_cpu(i) {
 
 		/* Copy values here to work around gcc-2.95.3, gcc-2.96 */
 		user = kstat_cpu(i).cpustat.user;
diff -rpuN linux-2.6.18.8/fs/proc/vmcore.c linux-2.6.18-xen-3.3.0/fs/proc/vmcore.c
--- linux-2.6.18.8/fs/proc/vmcore.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/fs/proc/vmcore.c	2008-08-21 11:36:07.000000000 +0200
@@ -514,7 +514,7 @@ static int __init parse_crash_elf64_head
 	/* Do some basic Verification. */
 	if (memcmp(ehdr.e_ident, ELFMAG, SELFMAG) != 0 ||
 		(ehdr.e_type != ET_CORE) ||
-		!elf_check_arch(&ehdr) ||
+		!vmcore_elf_check_arch(&ehdr) ||
 		ehdr.e_ident[EI_CLASS] != ELFCLASS64 ||
 		ehdr.e_ident[EI_VERSION] != EV_CURRENT ||
 		ehdr.e_version != EV_CURRENT ||
diff -rpuN linux-2.6.18.8/fs/splice.c linux-2.6.18-xen-3.3.0/fs/splice.c
--- linux-2.6.18.8/fs/splice.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/fs/splice.c	2008-08-21 11:36:07.000000000 +0200
@@ -1141,6 +1141,9 @@ static int get_iovec_page_array(const st
 		if (unlikely(!base))
 			break;
 
+		if (unlikely(!access_ok(VERIFY_READ, base, len)))
+			break;
+
 		/*
 		 * Get this base offset and number of pages, then map
 		 * in the user pages.
diff -rpuN linux-2.6.18.8/fs/xfs/linux-2.6/xfs_buf.c linux-2.6.18-xen-3.3.0/fs/xfs/linux-2.6/xfs_buf.c
--- linux-2.6.18.8/fs/xfs/linux-2.6/xfs_buf.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/fs/xfs/linux-2.6/xfs_buf.c	2008-08-21 11:36:07.000000000 +0200
@@ -182,6 +182,19 @@ free_address(
 {
 	a_list_t	*aentry;
 
+#ifdef CONFIG_XEN
+	/*
+	 * Xen needs to be able to make sure it can get an exclusive
+	 * RO mapping of pages it wants to turn into a pagetable.  If
+	 * a newly allocated page is also still being vmap()ed by xfs,
+	 * it will cause pagetable construction to fail.  This is a
+	 * quick workaround to always eagerly unmap pages so that Xen
+	 * is happy.
+	 */
+	vunmap(addr);
+	return;
+#endif
+
 	aentry = kmalloc(sizeof(a_list_t), GFP_NOWAIT);
 	if (likely(aentry)) {
 		spin_lock(&as_lock);
diff -rpuN linux-2.6.18.8/.hg_archival.txt linux-2.6.18-xen-3.3.0/.hg_archival.txt
--- linux-2.6.18.8/.hg_archival.txt	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/.hg_archival.txt	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,2 @@
+repo: 831230e53067cb45d27b07d037b2e907b663c2db
+node: e86b8e0ac6fdc3144976c4ea0bd922532a1fe8b4
diff -rpuN linux-2.6.18.8/.hgignore linux-2.6.18-xen-3.3.0/.hgignore
--- linux-2.6.18.8/.hgignore	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/.hgignore	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,3 @@
+.*\.orig$
+.*\.rej$
+\.valid-src$
\ No newline at end of file
diff -rpuN linux-2.6.18.8/.hgtags linux-2.6.18-xen-3.3.0/.hgtags
--- linux-2.6.18.8/.hgtags	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/.hgtags	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,2 @@
+831230e53067cb45d27b07d037b2e907b663c2db v2.6.18
+08e85e79c65d0316bfda5e77e8a0dc7ab9ca181a xen-3.2.0
diff -rpuN linux-2.6.18.8/include/acpi/aclocal.h linux-2.6.18-xen-3.3.0/include/acpi/aclocal.h
--- linux-2.6.18.8/include/acpi/aclocal.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/acpi/aclocal.h	2008-08-21 11:36:07.000000000 +0200
@@ -708,7 +708,7 @@ struct acpi_bit_register_info {
  * must be preserved.
  */
 #define ACPI_PM1_STATUS_PRESERVED_BITS          0x0800	/* Bit 11 */
-#define ACPI_PM1_CONTROL_PRESERVED_BITS         0x0201	/* Bit 9, Bit 0 (SCI_EN) */
+#define ACPI_PM1_CONTROL_PRESERVED_BITS         0x0200	/* Bit 9 (whatever) */
 
 /*
  * Register IDs
diff -rpuN linux-2.6.18.8/include/acpi/pdc_intel.h linux-2.6.18-xen-3.3.0/include/acpi/pdc_intel.h
--- linux-2.6.18.8/include/acpi/pdc_intel.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/acpi/pdc_intel.h	2008-08-21 11:36:07.000000000 +0200
@@ -13,6 +13,7 @@
 #define ACPI_PDC_SMP_C_SWCOORD		(0x0040)
 #define ACPI_PDC_SMP_T_SWCOORD		(0x0080)
 #define ACPI_PDC_C_C1_FFH		(0x0100)
+#define ACPI_PDC_C_C2C3_FFH		(0x0200)
 
 #define ACPI_PDC_EST_CAPABILITY_SMP	(ACPI_PDC_SMP_C1PT | \
 					 ACPI_PDC_C_C1_HALT | \
@@ -25,6 +26,8 @@
 
 #define ACPI_PDC_C_CAPABILITY_SMP	(ACPI_PDC_SMP_C2C3 | \
 					 ACPI_PDC_SMP_C1PT | \
-					 ACPI_PDC_C_C1_HALT)
+					 ACPI_PDC_C_C1_HALT | \
+					 ACPI_PDC_C_C1_FFH | \
+					 ACPI_PDC_C_C2C3_FFH)
 
 #endif				/* __PDC_INTEL_H__ */
diff -rpuN linux-2.6.18.8/include/acpi/processor.h linux-2.6.18-xen-3.3.0/include/acpi/processor.h
--- linux-2.6.18.8/include/acpi/processor.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/acpi/processor.h	2008-08-21 11:36:07.000000000 +0200
@@ -21,6 +21,12 @@
 #define ACPI_PSD_REV0_REVISION		0 /* Support for _PSD as in ACPI 3.0 */
 #define ACPI_PSD_REV0_ENTRIES		5
 
+#ifdef CONFIG_XEN
+#define NR_ACPI_CPUS			(NR_CPUS < 256 ? 256 : NR_CPUS)
+#else
+#define NR_ACPI_CPUS			NR_CPUS
+#endif /* CONFIG_XEN */
+
 /*
  * Types of coordination defined in ACPI 3.0. Same macros can be used across
  * P, C and T states
@@ -33,6 +39,17 @@
 
 struct acpi_processor_cx;
 
+#ifdef CONFIG_PROCESSOR_EXTERNAL_CONTROL
+struct acpi_csd_package {
+	acpi_integer num_entries;
+	acpi_integer revision;
+	acpi_integer domain;
+	acpi_integer coord_type;
+	acpi_integer num_processors;
+	acpi_integer index;
+} __attribute__ ((packed));
+#endif
+
 struct acpi_power_register {
 	u8 descriptor;
 	u16 length;
@@ -63,6 +80,12 @@ struct acpi_processor_cx {
 	u32 power;
 	u32 usage;
 	u64 time;
+#ifdef CONFIG_PROCESSOR_EXTERNAL_CONTROL
+	/* Require raw information for external control logic */
+	struct acpi_power_register reg;
+	u32 csd_count;
+	struct acpi_csd_package *domain_info;
+#endif
 	struct acpi_processor_cx_policy promotion;
 	struct acpi_processor_cx_policy demotion;
 };
@@ -231,6 +254,9 @@ static inline void acpi_processor_ppc_ex
 {
 	return;
 }
+#ifdef CONFIG_PROCESSOR_EXTERNAL_CONTROL
+int acpi_processor_ppc_has_changed(struct acpi_processor *pr);
+#else
 static inline int acpi_processor_ppc_has_changed(struct acpi_processor *pr)
 {
 	static unsigned int printout = 1;
@@ -243,6 +269,7 @@ static inline int acpi_processor_ppc_has
 	}
 	return 0;
 }
+#endif				/* CONFIG_PROCESSOR_EXTERNAL_CONTROL */
 #endif				/* CONFIG_CPU_FREQ */
 
 /* in processor_throttling.c */
@@ -275,4 +302,79 @@ static inline void acpi_thermal_cpufreq_
 }
 #endif
 
+/* 
+ * Following are interfaces geared to external processor PM control
+ * logic like a VMM
+ */
+/* Events notified to external control logic */
+#define PROCESSOR_PM_INIT	1
+#define PROCESSOR_PM_CHANGE	2
+#define PROCESSOR_HOTPLUG	3
+
+/* Objects for the PM events */
+#define PM_TYPE_IDLE		0
+#define PM_TYPE_PERF		1
+#define PM_TYPE_THR		2
+#define PM_TYPE_MAX		3
+
+/* Processor hotplug events */
+#define HOTPLUG_TYPE_ADD	0
+#define HOTPLUG_TYPE_REMOVE	1
+
+#ifdef CONFIG_PROCESSOR_EXTERNAL_CONTROL
+struct processor_extcntl_ops {
+	/* Transfer processor PM events to external control logic */
+	int (*pm_ops[PM_TYPE_MAX])(struct acpi_processor *pr, int event);
+	/* Notify physical processor status to external control logic */
+	int (*hotplug)(struct acpi_processor *pr, int event);
+};
+extern const struct processor_extcntl_ops *processor_extcntl_ops;
+
+static inline int processor_cntl_external(void)
+{
+	return (processor_extcntl_ops != NULL);
+}
+
+static inline int processor_pm_external(void)
+{
+	return processor_cntl_external() &&
+		(processor_extcntl_ops->pm_ops[PM_TYPE_IDLE] != NULL);
+}
+
+static inline int processor_pmperf_external(void)
+{
+	return processor_cntl_external() &&
+		(processor_extcntl_ops->pm_ops[PM_TYPE_PERF] != NULL);
+}
+
+static inline int processor_pmthr_external(void)
+{
+	return processor_cntl_external() &&
+		(processor_extcntl_ops->pm_ops[PM_TYPE_THR] != NULL);
+}
+
+extern int processor_notify_external(struct acpi_processor *pr,
+			int event, int type);
+extern void processor_extcntl_init(void);
+extern int processor_extcntl_prepare(struct acpi_processor *pr);
+extern int acpi_processor_get_performance_info(struct acpi_processor *pr);
+extern int acpi_processor_get_psd(struct acpi_processor *pr);
+void arch_acpi_processor_init_extcntl(const struct processor_extcntl_ops **);
+#else
+static inline int processor_cntl_external(void) {return 0;}
+static inline int processor_pm_external(void) {return 0;}
+static inline int processor_pmperf_external(void) {return 0;}
+static inline int processor_pmthr_external(void) {return 0;}
+static inline int processor_notify_external(struct acpi_processor *pr,
+			int event, int type)
+{
+	return 0;
+}
+static inline void processor_extcntl_init(void) {}
+static inline int processor_extcntl_prepare(struct acpi_processor *pr)
+{
+	return 0;
+}
+#endif /* CONFIG_PROCESSOR_EXTERNAL_CONTROL */
+
 #endif
diff -rpuN linux-2.6.18.8/include/asm-generic/pci.h linux-2.6.18-xen-3.3.0/include/asm-generic/pci.h
--- linux-2.6.18.8/include/asm-generic/pci.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-generic/pci.h	2008-08-21 11:36:07.000000000 +0200
@@ -43,7 +43,9 @@ pcibios_select_root(struct pci_dev *pdev
 	return root;
 }
 
+#ifndef pcibios_scan_all_fns
 #define pcibios_scan_all_fns(a, b)	0
+#endif
 
 #ifndef HAVE_ARCH_PCI_GET_LEGACY_IDE_IRQ
 static inline int pci_get_legacy_ide_irq(struct pci_dev *dev, int channel)
diff -rpuN linux-2.6.18.8/include/asm-generic/pgtable.h linux-2.6.18-xen-3.3.0/include/asm-generic/pgtable.h
--- linux-2.6.18.8/include/asm-generic/pgtable.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-generic/pgtable.h	2008-08-21 11:36:07.000000000 +0200
@@ -188,6 +188,10 @@ static inline void ptep_set_wrprotect(st
 })
 #endif
 
+#ifndef arch_change_pte_range
+#define arch_change_pte_range(mm, pmd, addr, end, newprot) 0
+#endif
+
 #ifndef __ASSEMBLY__
 /*
  * When walking page tables, we usually want to skip any p?d_none entries;
diff -rpuN linux-2.6.18.8/include/asm-generic/vmlinux.lds.h linux-2.6.18-xen-3.3.0/include/asm-generic/vmlinux.lds.h
--- linux-2.6.18.8/include/asm-generic/vmlinux.lds.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-generic/vmlinux.lds.h	2008-08-21 11:36:07.000000000 +0200
@@ -194,3 +194,6 @@
 		.stab.index 0 : { *(.stab.index) }			\
 		.stab.indexstr 0 : { *(.stab.indexstr) }		\
 		.comment 0 : { *(.comment) }
+
+#define NOTES								\
+		.notes : { *(.note.*) } :note
diff -rpuN linux-2.6.18.8/include/asm-i386/acpi.h linux-2.6.18-xen-3.3.0/include/asm-i386/acpi.h
--- linux-2.6.18.8/include/asm-i386/acpi.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/acpi.h	2008-08-21 11:36:07.000000000 +0200
@@ -31,6 +31,9 @@
 #include <acpi/pdc_intel.h>
 
 #include <asm/system.h>		/* defines cmpxchg */
+#ifdef CONFIG_XEN
+#include <xen/interface/platform.h>
+#endif
 
 #define COMPILER_DEPENDENT_INT64   long long
 #define COMPILER_DEPENDENT_UINT64  unsigned long long
@@ -156,6 +159,27 @@ static inline void acpi_disable_pci(void
 }
 extern int acpi_irq_balance_set(char *str);
 
+#ifdef CONFIG_XEN
+static inline int acpi_notify_hypervisor_state(u8 sleep_state,
+					       u32 pm1a_cnt_val,
+					       u32 pm1b_cnt_val)
+{
+	struct xen_platform_op op = {
+		.cmd = XENPF_enter_acpi_sleep,
+		.interface_version = XENPF_INTERFACE_VERSION,
+		.u = {
+			.enter_acpi_sleep = {
+				.pm1a_cnt_val = pm1a_cnt_val,
+				.pm1b_cnt_val = pm1b_cnt_val,
+				.sleep_state = sleep_state,
+			},
+		},
+	};
+
+	return HYPERVISOR_platform_op(&op);
+}
+#endif /* CONFIG_XEN */
+
 #else	/* !CONFIG_ACPI */
 
 #define acpi_lapic 0
@@ -181,7 +205,9 @@ extern void acpi_reserve_bootmem(void);
 
 extern u8 x86_acpiid_to_apicid[];
 
+#ifndef CONFIG_XEN
 #define ARCH_HAS_POWER_INIT	1
+#endif
 
 #endif /*__KERNEL__*/
 
diff -rpuN linux-2.6.18.8/include/asm-i386/agp.h linux-2.6.18-xen-3.3.0/include/asm-i386/agp.h
--- linux-2.6.18.8/include/asm-i386/agp.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/agp.h	2008-08-21 11:36:07.000000000 +0200
@@ -12,8 +12,10 @@
  * data corruption on some CPUs.
  */
 
-int map_page_into_agp(struct page *page);
-int unmap_page_from_agp(struct page *page);
+/* Caller's responsibility to call global_flush_tlb() for
+ * performance reasons */
+#define map_page_into_agp(page) change_page_attr(page, 1, PAGE_KERNEL_NOCACHE)
+#define unmap_page_from_agp(page) change_page_attr(page, 1, PAGE_KERNEL)
 #define flush_agp_mappings() global_flush_tlb()
 
 /* Could use CLFLUSH here if the cpu supports it. But then it would
diff -rpuN linux-2.6.18.8/include/asm-i386/apic.h linux-2.6.18-xen-3.3.0/include/asm-i386/apic.h
--- linux-2.6.18.8/include/asm-i386/apic.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/apic.h	2008-08-21 11:36:07.000000000 +0200
@@ -119,10 +119,12 @@ extern void enable_NMI_through_LVT0 (voi
 
 extern int disable_timer_pin_1;
 
+#ifndef CONFIG_XEN
 void smp_send_timer_broadcast_ipi(struct pt_regs *regs);
 void switch_APIC_timer_to_ipi(void *cpumask);
 void switch_ipi_to_APIC_timer(void *cpumask);
 #define ARCH_APICTIMER_STOPS_ON_C3	1
+#endif
 
 extern int timer_over_8254;
 
diff -rpuN linux-2.6.18.8/include/asm-i386/e820.h linux-2.6.18-xen-3.3.0/include/asm-i386/e820.h
--- linux-2.6.18.8/include/asm-i386/e820.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/e820.h	2008-08-21 11:36:07.000000000 +0200
@@ -38,6 +38,7 @@ extern struct e820map e820;
 
 extern int e820_all_mapped(unsigned long start, unsigned long end,
 			   unsigned type);
+extern int e820_any_mapped(u64 start, u64 end, unsigned type);
 
 #endif/*!__ASSEMBLY__*/
 
diff -rpuN linux-2.6.18.8/include/asm-i386/elf.h linux-2.6.18-xen-3.3.0/include/asm-i386/elf.h
--- linux-2.6.18.8/include/asm-i386/elf.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/elf.h	2008-08-21 11:36:07.000000000 +0200
@@ -169,50 +169,6 @@ do if (vdso_enabled) {						\
 		NEW_AUX_ENT(AT_SYSINFO_EHDR, VDSO_COMPAT_BASE);	\
 } while (0)
 
-/*
- * These macros parameterize elf_core_dump in fs/binfmt_elf.c to write out
- * extra segments containing the vsyscall DSO contents.  Dumping its
- * contents makes post-mortem fully interpretable later without matching up
- * the same kernel and hardware config to see what PC values meant.
- * Dumping its extra ELF program headers includes all the other information
- * a debugger needs to easily find how the vsyscall DSO was being used.
- */
-#define ELF_CORE_EXTRA_PHDRS		(VDSO_HIGH_EHDR->e_phnum)
-#define ELF_CORE_WRITE_EXTRA_PHDRS					      \
-do {									      \
-	const struct elf_phdr *const vsyscall_phdrs =			      \
-		(const struct elf_phdr *) (VDSO_HIGH_BASE		      \
-					   + VDSO_HIGH_EHDR->e_phoff);    \
-	int i;								      \
-	Elf32_Off ofs = 0;						      \
-	for (i = 0; i < VDSO_HIGH_EHDR->e_phnum; ++i) {		      \
-		struct elf_phdr phdr = vsyscall_phdrs[i];		      \
-		if (phdr.p_type == PT_LOAD) {				      \
-			BUG_ON(ofs != 0);				      \
-			ofs = phdr.p_offset = offset;			      \
-			phdr.p_memsz = PAGE_ALIGN(phdr.p_memsz);	      \
-			phdr.p_filesz = phdr.p_memsz;			      \
-			offset += phdr.p_filesz;			      \
-		}							      \
-		else							      \
-			phdr.p_offset += ofs;				      \
-		phdr.p_paddr = 0; /* match other core phdrs */		      \
-		DUMP_WRITE(&phdr, sizeof(phdr));			      \
-	}								      \
-} while (0)
-#define ELF_CORE_WRITE_EXTRA_DATA					      \
-do {									      \
-	const struct elf_phdr *const vsyscall_phdrs =			      \
-		(const struct elf_phdr *) (VDSO_HIGH_BASE		      \
-					   + VDSO_HIGH_EHDR->e_phoff);    \
-	int i;								      \
-	for (i = 0; i < VDSO_HIGH_EHDR->e_phnum; ++i) {		      \
-		if (vsyscall_phdrs[i].p_type == PT_LOAD)		      \
-			DUMP_WRITE((void *) vsyscall_phdrs[i].p_vaddr,	      \
-				   PAGE_ALIGN(vsyscall_phdrs[i].p_memsz));    \
-	}								      \
-} while (0)
-
 #endif
 
 #endif
diff -rpuN linux-2.6.18.8/include/asm-i386/fixmap.h linux-2.6.18-xen-3.3.0/include/asm-i386/fixmap.h
--- linux-2.6.18.8/include/asm-i386/fixmap.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/fixmap.h	2008-08-21 11:36:07.000000000 +0200
@@ -19,7 +19,7 @@
  * Leave one empty page between vmalloc'ed areas and
  * the start of the fixmap.
  */
-#define __FIXADDR_TOP	0xfffff000
+extern unsigned long __FIXADDR_TOP;
 
 #ifndef __ASSEMBLY__
 #include <linux/kernel.h>
@@ -94,6 +94,8 @@ enum fixed_addresses {
 extern void __set_fixmap (enum fixed_addresses idx,
 					unsigned long phys, pgprot_t flags);
 
+extern void set_fixaddr_top(unsigned long top);
+
 #define set_fixmap(idx, phys) \
 		__set_fixmap(idx, phys, PAGE_KERNEL)
 /*
diff -rpuN linux-2.6.18.8/include/asm-i386/io_apic.h linux-2.6.18-xen-3.3.0/include/asm-i386/io_apic.h
--- linux-2.6.18.8/include/asm-i386/io_apic.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/io_apic.h	2008-08-21 11:36:07.000000000 +0200
@@ -12,7 +12,7 @@
 
 #ifdef CONFIG_X86_IO_APIC
 
-#ifdef CONFIG_PCI_MSI
+#if defined(CONFIG_PCI_MSI) && !defined(CONFIG_XEN)
 static inline int use_pci_vector(void)	{return 1;}
 static inline void disable_edge_ioapic_vector(unsigned int vector) { }
 static inline void mask_and_ack_level_ioapic_vector(unsigned int vector) { }
diff -rpuN linux-2.6.18.8/include/asm-i386/kexec.h linux-2.6.18-xen-3.3.0/include/asm-i386/kexec.h
--- linux-2.6.18.8/include/asm-i386/kexec.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/kexec.h	2008-08-21 11:36:07.000000000 +0200
@@ -1,6 +1,26 @@
 #ifndef _I386_KEXEC_H
 #define _I386_KEXEC_H
 
+#define PA_CONTROL_PAGE  0
+#define VA_CONTROL_PAGE  1
+#define PA_PGD           2
+#define VA_PGD           3
+#define PA_PTE_0         4
+#define VA_PTE_0         5
+#define PA_PTE_1         6
+#define VA_PTE_1         7
+#ifdef CONFIG_X86_PAE
+#define PA_PMD_0         8
+#define VA_PMD_0         9
+#define PA_PMD_1         10
+#define VA_PMD_1         11
+#define PAGES_NR         12
+#else
+#define PAGES_NR         8
+#endif
+
+#ifndef __ASSEMBLY__
+
 #include <asm/fixmap.h>
 #include <asm/ptrace.h>
 #include <asm/string.h>
@@ -27,6 +47,9 @@
 /* The native architecture */
 #define KEXEC_ARCH KEXEC_ARCH_386
 
+/* We can also handle crash dumps from 64 bit kernel. */
+#define vmcore_elf_check_arch_cross(x) ((x)->e_machine == EM_X86_64)
+
 #define MAX_NOTE_BYTES 1024
 
 /* CPU does not save ss and esp on stack if execution is already
@@ -72,5 +95,26 @@ static inline void crash_setup_regs(stru
                newregs->eip = (unsigned long)current_text_addr();
        }
 }
+asmlinkage NORET_TYPE void
+relocate_kernel(unsigned long indirection_page,
+		unsigned long control_page,
+		unsigned long start_address,
+		unsigned int has_pae) ATTRIB_NORET;
+
+
+/* Under Xen we need to work with machine addresses. These macros give the
+ * machine address of a certain page to the generic kexec code instead of 
+ * the pseudo physical address which would be given by the default macros.
+ */
+
+#ifdef CONFIG_XEN
+#define KEXEC_ARCH_HAS_PAGE_MACROS
+#define kexec_page_to_pfn(page)  pfn_to_mfn(page_to_pfn(page))
+#define kexec_pfn_to_page(pfn)   pfn_to_page(mfn_to_pfn(pfn))
+#define kexec_virt_to_phys(addr) virt_to_machine(addr)
+#define kexec_phys_to_virt(addr) phys_to_virt(machine_to_phys(addr))
+#endif
+
+#endif /* __ASSEMBLY__ */
 
 #endif /* _I386_KEXEC_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-default/mach_traps.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-default/mach_traps.h
--- linux-2.6.18.8/include/asm-i386/mach-default/mach_traps.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-default/mach_traps.h	2008-08-21 11:36:07.000000000 +0200
@@ -15,6 +15,18 @@ static inline void clear_mem_error(unsig
 	outb(reason, 0x61);
 }
 
+static inline void clear_io_check_error(unsigned char reason)
+{
+	unsigned long i;
+
+	reason = (reason & 0xf) | 8;
+	outb(reason, 0x61);
+	i = 2000;
+	while (--i) udelay(1000);
+	reason &= ~8;
+	outb(reason, 0x61);
+}
+
 static inline unsigned char get_nmi_reason(void)
 {
 	return inb(0x61);
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/agp.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/agp.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/agp.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/agp.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,44 @@
+#ifndef AGP_H
+#define AGP_H 1
+
+#include <asm/pgtable.h>
+#include <asm/cacheflush.h>
+#include <asm/system.h>
+
+/* 
+ * Functions to keep the agpgart mappings coherent with the MMU.
+ * The GART gives the CPU a physical alias of pages in memory. The alias region is
+ * mapped uncacheable. Make sure there are no conflicting mappings
+ * with different cachability attributes for the same page. This avoids
+ * data corruption on some CPUs.
+ */
+
+/* Caller's responsibility to call global_flush_tlb() for
+ * performance reasons */
+#define map_page_into_agp(page) ( \
+	xen_create_contiguous_region((unsigned long)page_address(page), 0, 32) \
+	?: change_page_attr(page, 1, PAGE_KERNEL_NOCACHE))
+#define unmap_page_from_agp(page) ( \
+	xen_destroy_contiguous_region((unsigned long)page_address(page), 0), \
+	/* only a fallback: xen_destroy_contiguous_region uses PAGE_KERNEL */ \
+	change_page_attr(page, 1, PAGE_KERNEL))
+#define flush_agp_mappings() global_flush_tlb()
+
+/* Could use CLFLUSH here if the cpu supports it. But then it would
+   need to be called for each cacheline of the whole page so it may not be 
+   worth it. Would need a page for it. */
+#define flush_agp_cache() wbinvd()
+
+/* Convert a physical address to an address suitable for the GART. */
+#define phys_to_gart(x) phys_to_machine(x)
+#define gart_to_phys(x) machine_to_phys(x)
+
+/* GATT allocation. Returns/accepts GATT kernel virtual address. */
+#define alloc_gatt_pages(order)	({                                          \
+	char *_t; dma_addr_t _d;                                            \
+	_t = dma_alloc_coherent(NULL,PAGE_SIZE<<(order),&_d,GFP_KERNEL);    \
+	_t; })
+#define free_gatt_pages(table, order)	\
+	dma_free_coherent(NULL,PAGE_SIZE<<(order),(table),virt_to_bus(table))
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/desc.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/desc.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/desc.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/desc.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,166 @@
+#ifndef __ARCH_DESC_H
+#define __ARCH_DESC_H
+
+#include <asm/ldt.h>
+#include <asm/segment.h>
+
+#define CPU_16BIT_STACK_SIZE 1024
+
+#ifndef __ASSEMBLY__
+
+#include <linux/preempt.h>
+#include <linux/smp.h>
+
+#include <asm/mmu.h>
+
+extern struct desc_struct cpu_gdt_table[GDT_ENTRIES];
+
+DECLARE_PER_CPU(unsigned char, cpu_16bit_stack[CPU_16BIT_STACK_SIZE]);
+
+struct Xgt_desc_struct {
+	unsigned short size;
+	unsigned long address __attribute__((packed));
+	unsigned short pad;
+} __attribute__ ((packed));
+
+extern struct Xgt_desc_struct idt_descr;
+DECLARE_PER_CPU(struct Xgt_desc_struct, cpu_gdt_descr);
+
+
+static inline struct desc_struct *get_cpu_gdt_table(unsigned int cpu)
+{
+	return (struct desc_struct *)per_cpu(cpu_gdt_descr, cpu).address;
+}
+
+#define load_TR_desc() __asm__ __volatile__("ltr %w0"::"q" (GDT_ENTRY_TSS*8))
+#define load_LDT_desc() __asm__ __volatile__("lldt %w0"::"q" (GDT_ENTRY_LDT*8))
+
+#define load_gdt(dtr) __asm__ __volatile("lgdt %0"::"m" (*dtr))
+#define load_idt(dtr) __asm__ __volatile("lidt %0"::"m" (*dtr))
+#define load_tr(tr) __asm__ __volatile("ltr %0"::"mr" (tr))
+#define load_ldt(ldt) __asm__ __volatile("lldt %0"::"mr" (ldt))
+
+#define store_gdt(dtr) __asm__ ("sgdt %0":"=m" (*dtr))
+#define store_idt(dtr) __asm__ ("sidt %0":"=m" (*dtr))
+#define store_tr(tr) __asm__ ("str %0":"=mr" (tr))
+#define store_ldt(ldt) __asm__ ("sldt %0":"=mr" (ldt))
+
+/*
+ * This is the ldt that every process will get unless we need
+ * something other than this.
+ */
+extern struct desc_struct default_ldt[];
+extern void set_intr_gate(unsigned int irq, void * addr);
+
+#define _set_tssldt_desc(n,addr,limit,type) \
+__asm__ __volatile__ ("movw %w3,0(%2)\n\t" \
+	"movw %w1,2(%2)\n\t" \
+	"rorl $16,%1\n\t" \
+	"movb %b1,4(%2)\n\t" \
+	"movb %4,5(%2)\n\t" \
+	"movb $0,6(%2)\n\t" \
+	"movb %h1,7(%2)\n\t" \
+	"rorl $16,%1" \
+	: "=m"(*(n)) : "q" (addr), "r"(n), "ir"(limit), "i"(type))
+
+#ifndef CONFIG_X86_NO_TSS
+static inline void __set_tss_desc(unsigned int cpu, unsigned int entry, void *addr)
+{
+	_set_tssldt_desc(&get_cpu_gdt_table(cpu)[entry], (int)addr,
+		offsetof(struct tss_struct, __cacheline_filler) - 1, 0x89);
+}
+
+#define set_tss_desc(cpu,addr) __set_tss_desc(cpu, GDT_ENTRY_TSS, addr)
+#endif
+
+static inline void set_ldt_desc(unsigned int cpu, void *addr, unsigned int size)
+{
+	_set_tssldt_desc(&get_cpu_gdt_table(cpu)[GDT_ENTRY_LDT], (int)addr, ((size << 3)-1), 0x82);
+}
+
+#define LDT_entry_a(info) \
+	((((info)->base_addr & 0x0000ffff) << 16) | ((info)->limit & 0x0ffff))
+
+#define LDT_entry_b(info) \
+	(((info)->base_addr & 0xff000000) | \
+	(((info)->base_addr & 0x00ff0000) >> 16) | \
+	((info)->limit & 0xf0000) | \
+	(((info)->read_exec_only ^ 1) << 9) | \
+	((info)->contents << 10) | \
+	(((info)->seg_not_present ^ 1) << 15) | \
+	((info)->seg_32bit << 22) | \
+	((info)->limit_in_pages << 23) | \
+	((info)->useable << 20) | \
+	0x7000)
+
+#define LDT_empty(info) (\
+	(info)->base_addr	== 0	&& \
+	(info)->limit		== 0	&& \
+	(info)->contents	== 0	&& \
+	(info)->read_exec_only	== 1	&& \
+	(info)->seg_32bit	== 0	&& \
+	(info)->limit_in_pages	== 0	&& \
+	(info)->seg_not_present	== 1	&& \
+	(info)->useable		== 0	)
+
+extern int write_ldt_entry(void *ldt, int entry, __u32 entry_a, __u32 entry_b);
+
+#if TLS_SIZE != 24
+# error update this code.
+#endif
+
+static inline void load_TLS(struct thread_struct *t, unsigned int cpu)
+{
+#define C(i) if (HYPERVISOR_update_descriptor(virt_to_machine(&get_cpu_gdt_table(cpu)[GDT_ENTRY_TLS_MIN + i]), \
+					       *(u64 *)&t->tls_array[i])) \
+		BUG();
+	C(0); C(1); C(2);
+#undef C
+}
+
+static inline void clear_LDT(void)
+{
+	int cpu = get_cpu();
+
+	/*
+	 * NB. We load the default_ldt for lcall7/27 handling on demand, as
+	 * it slows down context switching. Noone uses it anyway.
+	 */
+	cpu = cpu;		/* XXX avoid compiler warning */
+	xen_set_ldt(NULL, 0);
+	put_cpu();
+}
+
+/*
+ * load one particular LDT into the current CPU
+ */
+static inline void load_LDT_nolock(mm_context_t *pc, int cpu)
+{
+	void *segments = pc->ldt;
+	int count = pc->size;
+
+	if (likely(!count))
+		segments = NULL;
+
+	xen_set_ldt(segments, count);
+}
+
+static inline void load_LDT(mm_context_t *pc)
+{
+	int cpu = get_cpu();
+	load_LDT_nolock(pc, cpu);
+	put_cpu();
+}
+
+static inline unsigned long get_desc_base(unsigned long *desc)
+{
+	unsigned long base;
+	base = ((desc[0] >> 16)  & 0x0000ffff) |
+		((desc[1] << 16) & 0x00ff0000) |
+		(desc[1] & 0xff000000);
+	return base;
+}
+
+#endif /* !__ASSEMBLY__ */
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/dma-mapping.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/dma-mapping.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/dma-mapping.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/dma-mapping.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,151 @@
+#ifndef _ASM_I386_DMA_MAPPING_H
+#define _ASM_I386_DMA_MAPPING_H
+
+/*
+ * IOMMU interface. See Documentation/DMA-mapping.txt and DMA-API.txt for
+ * documentation.
+ */
+
+#include <linux/mm.h>
+#include <asm/cache.h>
+#include <asm/io.h>
+#include <asm/scatterlist.h>
+#include <asm/swiotlb.h>
+
+static inline int
+address_needs_mapping(struct device *hwdev, dma_addr_t addr)
+{
+	dma_addr_t mask = 0xffffffff;
+	/* If the device has a mask, use it, otherwise default to 32 bits */
+	if (hwdev && hwdev->dma_mask)
+		mask = *hwdev->dma_mask;
+	return (addr & ~mask) != 0;
+}
+
+extern int range_straddles_page_boundary(paddr_t p, size_t size);
+
+#define dma_alloc_noncoherent(d, s, h, f) dma_alloc_coherent(d, s, h, f)
+#define dma_free_noncoherent(d, s, v, h) dma_free_coherent(d, s, v, h)
+
+void *dma_alloc_coherent(struct device *dev, size_t size,
+			   dma_addr_t *dma_handle, gfp_t flag);
+
+void dma_free_coherent(struct device *dev, size_t size,
+			 void *vaddr, dma_addr_t dma_handle);
+
+extern dma_addr_t
+dma_map_single(struct device *dev, void *ptr, size_t size,
+	       enum dma_data_direction direction);
+
+extern void
+dma_unmap_single(struct device *dev, dma_addr_t dma_addr, size_t size,
+		 enum dma_data_direction direction);
+
+extern int dma_map_sg(struct device *hwdev, struct scatterlist *sg,
+		      int nents, enum dma_data_direction direction);
+extern void dma_unmap_sg(struct device *hwdev, struct scatterlist *sg,
+			 int nents, enum dma_data_direction direction);
+
+#ifdef CONFIG_HIGHMEM
+extern dma_addr_t
+dma_map_page(struct device *dev, struct page *page, unsigned long offset,
+	     size_t size, enum dma_data_direction direction);
+
+extern void
+dma_unmap_page(struct device *dev, dma_addr_t dma_address, size_t size,
+	       enum dma_data_direction direction);
+#else
+#define dma_map_page(dev, page, offset, size, dir) \
+	dma_map_single(dev, page_address(page) + (offset), (size), (dir))
+#define dma_unmap_page dma_unmap_single
+#endif
+
+extern void
+dma_sync_single_for_cpu(struct device *dev, dma_addr_t dma_handle, size_t size,
+			enum dma_data_direction direction);
+
+extern void
+dma_sync_single_for_device(struct device *dev, dma_addr_t dma_handle, size_t size,
+                           enum dma_data_direction direction);
+
+static inline void
+dma_sync_single_range_for_cpu(struct device *dev, dma_addr_t dma_handle,
+			      unsigned long offset, size_t size,
+			      enum dma_data_direction direction)
+{
+	dma_sync_single_for_cpu(dev, dma_handle+offset, size, direction);
+}
+
+static inline void
+dma_sync_single_range_for_device(struct device *dev, dma_addr_t dma_handle,
+				 unsigned long offset, size_t size,
+				 enum dma_data_direction direction)
+{
+	dma_sync_single_for_device(dev, dma_handle+offset, size, direction);
+}
+
+static inline void
+dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sg, int nelems,
+		    enum dma_data_direction direction)
+{
+	if (swiotlb)
+		swiotlb_sync_sg_for_cpu(dev,sg,nelems,direction);
+	flush_write_buffers();
+}
+
+static inline void
+dma_sync_sg_for_device(struct device *dev, struct scatterlist *sg, int nelems,
+		    enum dma_data_direction direction)
+{
+	if (swiotlb)
+		swiotlb_sync_sg_for_device(dev,sg,nelems,direction);
+	flush_write_buffers();
+}
+
+extern int
+dma_mapping_error(dma_addr_t dma_addr);
+
+extern int
+dma_supported(struct device *dev, u64 mask);
+
+static inline int
+dma_set_mask(struct device *dev, u64 mask)
+{
+	if(!dev->dma_mask || !dma_supported(dev, mask))
+		return -EIO;
+
+	*dev->dma_mask = mask;
+
+	return 0;
+}
+
+static inline int
+dma_get_cache_alignment(void)
+{
+	/* no easy way to get cache size on all x86, so return the
+	 * maximum possible, to be safe */
+	return (1 << INTERNODE_CACHE_SHIFT);
+}
+
+#define dma_is_consistent(d)	(1)
+
+static inline void
+dma_cache_sync(void *vaddr, size_t size,
+	       enum dma_data_direction direction)
+{
+	flush_write_buffers();
+}
+
+#define ARCH_HAS_DMA_DECLARE_COHERENT_MEMORY
+extern int
+dma_declare_coherent_memory(struct device *dev, dma_addr_t bus_addr,
+			    dma_addr_t device_addr, size_t size, int flags);
+
+extern void
+dma_release_declared_memory(struct device *dev);
+
+extern void *
+dma_mark_declared_memory_occupied(struct device *dev,
+				  dma_addr_t device_addr, size_t size);
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/fixmap.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/fixmap.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/fixmap.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/fixmap.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,155 @@
+/*
+ * fixmap.h: compile-time virtual memory allocation
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License.  See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Copyright (C) 1998 Ingo Molnar
+ *
+ * Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999
+ */
+
+#ifndef _ASM_FIXMAP_H
+#define _ASM_FIXMAP_H
+
+
+/* used by vmalloc.c, vsyscall.lds.S.
+ *
+ * Leave one empty page between vmalloc'ed areas and
+ * the start of the fixmap.
+ */
+extern unsigned long __FIXADDR_TOP;
+
+#ifndef __ASSEMBLY__
+#include <linux/kernel.h>
+#include <asm/acpi.h>
+#include <asm/apicdef.h>
+#include <asm/page.h>
+#ifdef CONFIG_HIGHMEM
+#include <linux/threads.h>
+#include <asm/kmap_types.h>
+#endif
+
+/*
+ * Here we define all the compile-time 'special' virtual
+ * addresses. The point is to have a constant address at
+ * compile time, but to set the physical address only
+ * in the boot process. We allocate these special addresses
+ * from the end of virtual memory (0xfffff000) backwards.
+ * Also this lets us do fail-safe vmalloc(), we
+ * can guarantee that these special addresses and
+ * vmalloc()-ed addresses never overlap.
+ *
+ * these 'compile-time allocated' memory buffers are
+ * fixed-size 4k pages. (or larger if used with an increment
+ * highger than 1) use fixmap_set(idx,phys) to associate
+ * physical memory with fixmap indices.
+ *
+ * TLB entries of such buffers will not be flushed across
+ * task switches.
+ */
+enum fixed_addresses {
+	FIX_HOLE,
+	FIX_VDSO,
+#ifdef CONFIG_X86_LOCAL_APIC
+	FIX_APIC_BASE,	/* local (CPU) APIC) -- required for SMP or not */
+#endif
+#ifdef CONFIG_X86_IO_APIC
+	FIX_IO_APIC_BASE_0,
+	FIX_IO_APIC_BASE_END = FIX_IO_APIC_BASE_0 + MAX_IO_APICS-1,
+#endif
+#ifdef CONFIG_X86_VISWS_APIC
+	FIX_CO_CPU,	/* Cobalt timer */
+	FIX_CO_APIC,	/* Cobalt APIC Redirection Table */ 
+	FIX_LI_PCIA,	/* Lithium PCI Bridge A */
+	FIX_LI_PCIB,	/* Lithium PCI Bridge B */
+#endif
+#ifdef CONFIG_X86_F00F_BUG
+	FIX_F00F_IDT,	/* Virtual mapping for IDT */
+#endif
+#ifdef CONFIG_X86_CYCLONE_TIMER
+	FIX_CYCLONE_TIMER, /*cyclone timer register*/
+#endif 
+#ifdef CONFIG_HIGHMEM
+	FIX_KMAP_BEGIN,	/* reserved pte's for temporary kernel mappings */
+	FIX_KMAP_END = FIX_KMAP_BEGIN+(KM_TYPE_NR*NR_CPUS)-1,
+#endif
+#ifdef CONFIG_ACPI
+	FIX_ACPI_BEGIN,
+	FIX_ACPI_END = FIX_ACPI_BEGIN + FIX_ACPI_PAGES - 1,
+#endif
+#ifdef CONFIG_PCI_MMCONFIG
+	FIX_PCIE_MCFG,
+#endif
+	FIX_SHARED_INFO,
+#define NR_FIX_ISAMAPS	256
+	FIX_ISAMAP_END,
+	FIX_ISAMAP_BEGIN = FIX_ISAMAP_END + NR_FIX_ISAMAPS - 1,
+	__end_of_permanent_fixed_addresses,
+	/* temporary boot-time mappings, used before ioremap() is functional */
+#define NR_FIX_BTMAPS	16
+	FIX_BTMAP_END = __end_of_permanent_fixed_addresses,
+	FIX_BTMAP_BEGIN = FIX_BTMAP_END + NR_FIX_BTMAPS - 1,
+	FIX_WP_TEST,
+	__end_of_fixed_addresses
+};
+
+extern void set_fixaddr_top(unsigned long top);
+
+extern void __set_fixmap(enum fixed_addresses idx,
+					maddr_t phys, pgprot_t flags);
+
+#define set_fixmap(idx, phys) \
+		__set_fixmap(idx, phys, PAGE_KERNEL)
+/*
+ * Some hardware wants to get fixmapped without caching.
+ */
+#define set_fixmap_nocache(idx, phys) \
+		__set_fixmap(idx, phys, PAGE_KERNEL_NOCACHE)
+
+#define clear_fixmap(idx) \
+		__set_fixmap(idx, 0, __pgprot(0))
+
+#define FIXADDR_TOP	((unsigned long)__FIXADDR_TOP)
+
+#define __FIXADDR_SIZE	(__end_of_permanent_fixed_addresses << PAGE_SHIFT)
+#define __FIXADDR_BOOT_SIZE	(__end_of_fixed_addresses << PAGE_SHIFT)
+#define FIXADDR_START		(FIXADDR_TOP - __FIXADDR_SIZE)
+#define FIXADDR_BOOT_START	(FIXADDR_TOP - __FIXADDR_BOOT_SIZE)
+
+#define __fix_to_virt(x)	(FIXADDR_TOP - ((x) << PAGE_SHIFT))
+#define __virt_to_fix(x)	((FIXADDR_TOP - ((x)&PAGE_MASK)) >> PAGE_SHIFT)
+
+extern void __this_fixmap_does_not_exist(void);
+
+/*
+ * 'index to address' translation. If anyone tries to use the idx
+ * directly without tranlation, we catch the bug with a NULL-deference
+ * kernel oops. Illegal ranges of incoming indices are caught too.
+ */
+static __always_inline unsigned long fix_to_virt(const unsigned int idx)
+{
+	/*
+	 * this branch gets completely eliminated after inlining,
+	 * except when someone tries to use fixaddr indices in an
+	 * illegal way. (such as mixing up address types or using
+	 * out-of-range indices).
+	 *
+	 * If it doesn't get removed, the linker will complain
+	 * loudly with a reasonably clear error message..
+	 */
+	if (idx >= __end_of_fixed_addresses)
+		__this_fixmap_does_not_exist();
+
+        return __fix_to_virt(idx);
+}
+
+static inline unsigned long virt_to_fix(const unsigned long vaddr)
+{
+	BUG_ON(vaddr >= FIXADDR_TOP || vaddr < FIXADDR_START);
+	return __virt_to_fix(vaddr);
+}
+
+#endif /* !__ASSEMBLY__ */
+#endif
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/gnttab_dma.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/gnttab_dma.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/gnttab_dma.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/gnttab_dma.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,41 @@
+/*
+ * Copyright (c) 2007 Herbert Xu <herbert@gondor.apana.org.au>
+ * Copyright (c) 2007 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#ifndef _ASM_I386_GNTTAB_DMA_H
+#define _ASM_I386_GNTTAB_DMA_H
+
+static inline int gnttab_dma_local_pfn(struct page *page)
+{
+	/* Has it become a local MFN? */
+	return pfn_valid(mfn_to_local_pfn(pfn_to_mfn(page_to_pfn(page))));
+}
+
+static inline maddr_t gnttab_dma_map_page(struct page *page)
+{
+	__gnttab_dma_map_page(page);
+	return ((maddr_t)pfn_to_mfn(page_to_pfn(page)) << PAGE_SHIFT);
+}
+
+static inline void gnttab_dma_unmap_page(maddr_t maddr)
+{
+	__gnttab_dma_unmap_page(virt_to_page(bus_to_virt(maddr)));
+}
+
+#endif /* _ASM_I386_GNTTAB_DMA_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/highmem.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/highmem.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/highmem.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/highmem.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,80 @@
+/*
+ * highmem.h: virtual kernel memory mappings for high memory
+ *
+ * Used in CONFIG_HIGHMEM systems for memory pages which
+ * are not addressable by direct kernel virtual addresses.
+ *
+ * Copyright (C) 1999 Gerhard Wichert, Siemens AG
+ *		      Gerhard.Wichert@pdb.siemens.de
+ *
+ *
+ * Redesigned the x86 32-bit VM architecture to deal with 
+ * up to 16 Terabyte physical memory. With current x86 CPUs
+ * we now support up to 64 Gigabytes physical RAM.
+ *
+ * Copyright (C) 1999 Ingo Molnar <mingo@redhat.com>
+ */
+
+#ifndef _ASM_HIGHMEM_H
+#define _ASM_HIGHMEM_H
+
+#ifdef __KERNEL__
+
+#include <linux/interrupt.h>
+#include <linux/threads.h>
+#include <asm/kmap_types.h>
+#include <asm/tlbflush.h>
+
+/* declarations for highmem.c */
+extern unsigned long highstart_pfn, highend_pfn;
+
+extern pte_t *kmap_pte;
+extern pgprot_t kmap_prot;
+extern pte_t *pkmap_page_table;
+
+/*
+ * Right now we initialize only a single pte table. It can be extended
+ * easily, subsequent pte tables have to be allocated in one physical
+ * chunk of RAM.
+ */
+#ifdef CONFIG_X86_PAE
+#define LAST_PKMAP 512
+#else
+#define LAST_PKMAP 1024
+#endif
+/*
+ * Ordering is:
+ *
+ * FIXADDR_TOP
+ * 			fixed_addresses
+ * FIXADDR_START
+ * 			temp fixed addresses
+ * FIXADDR_BOOT_START
+ * 			Persistent kmap area
+ * PKMAP_BASE
+ * VMALLOC_END
+ * 			Vmalloc area
+ * VMALLOC_START
+ * high_memory
+ */
+#define PKMAP_BASE ( (FIXADDR_BOOT_START - PAGE_SIZE*(LAST_PKMAP + 1)) & PMD_MASK )
+#define LAST_PKMAP_MASK (LAST_PKMAP-1)
+#define PKMAP_NR(virt)  ((virt-PKMAP_BASE) >> PAGE_SHIFT)
+#define PKMAP_ADDR(nr)  (PKMAP_BASE + ((nr) << PAGE_SHIFT))
+
+extern void * FASTCALL(kmap_high(struct page *page));
+extern void FASTCALL(kunmap_high(struct page *page));
+
+void *kmap(struct page *page);
+void kunmap(struct page *page);
+void *kmap_atomic(struct page *page, enum km_type type);
+void *kmap_atomic_pte(struct page *page, enum km_type type);
+void kunmap_atomic(void *kvaddr, enum km_type type);
+void *kmap_atomic_pfn(unsigned long pfn, enum km_type type);
+struct page *kmap_atomic_to_page(void *ptr);
+
+#define flush_cache_kmaps()	do { } while (0)
+
+#endif /* __KERNEL__ */
+
+#endif /* _ASM_HIGHMEM_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/hw_irq.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/hw_irq.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/hw_irq.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/hw_irq.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,72 @@
+#ifndef _ASM_HW_IRQ_H
+#define _ASM_HW_IRQ_H
+
+/*
+ *	linux/include/asm/hw_irq.h
+ *
+ *	(C) 1992, 1993 Linus Torvalds, (C) 1997 Ingo Molnar
+ *
+ *	moved some of the old arch/i386/kernel/irq.h to here. VY
+ *
+ *	IRQ/IPI changes taken from work by Thomas Radke
+ *	<tomsoft@informatik.tu-chemnitz.de>
+ */
+
+#include <linux/profile.h>
+#include <asm/atomic.h>
+#include <asm/irq.h>
+#include <asm/sections.h>
+
+struct hw_interrupt_type;
+
+#define NMI_VECTOR		0x02
+
+/*
+ * Various low-level irq details needed by irq.c, process.c,
+ * time.c, io_apic.c and smp.c
+ *
+ * Interrupt entry/exit code at both C and assembly level
+ */
+
+extern u8 irq_vector[NR_IRQ_VECTORS];
+#define IO_APIC_VECTOR(irq)	(irq_vector[irq])
+#define AUTO_ASSIGN		-1
+
+extern void (*interrupt[NR_IRQS])(void);
+
+#ifdef CONFIG_SMP
+fastcall void reschedule_interrupt(void);
+fastcall void invalidate_interrupt(void);
+fastcall void call_function_interrupt(void);
+#endif
+
+#ifdef CONFIG_X86_LOCAL_APIC
+fastcall void apic_timer_interrupt(void);
+fastcall void error_interrupt(void);
+fastcall void spurious_interrupt(void);
+fastcall void thermal_interrupt(struct pt_regs *);
+#define platform_legacy_irq(irq)	((irq) < 16)
+#endif
+
+void disable_8259A_irq(unsigned int irq);
+void enable_8259A_irq(unsigned int irq);
+int i8259A_irq_pending(unsigned int irq);
+void make_8259A_irq(unsigned int irq);
+void init_8259A(int aeoi);
+void FASTCALL(send_IPI_self(int vector));
+void init_VISWS_APIC_irqs(void);
+void setup_IO_APIC(void);
+void disable_IO_APIC(void);
+#define print_IO_APIC()
+int IO_APIC_get_PCI_irq_vector(int bus, int slot, int fn);
+void send_IPI(int dest, int vector);
+void setup_ioapic_dest(void);
+
+extern unsigned long io_apic_irqs;
+
+extern atomic_t irq_err_count;
+extern atomic_t irq_mis_count;
+
+#define IO_APIC_IRQ(x) (((x) >= 16) || ((1<<(x)) & io_apic_irqs))
+
+#endif /* _ASM_HW_IRQ_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/hypercall.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/hypercall.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/hypercall.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/hypercall.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,416 @@
+/******************************************************************************
+ * hypercall.h
+ * 
+ * Linux-specific hypervisor handling.
+ * 
+ * Copyright (c) 2002-2004, K A Fraser
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __HYPERCALL_H__
+#define __HYPERCALL_H__
+
+#include <linux/string.h> /* memcpy() */
+#include <linux/stringify.h>
+
+#ifndef __HYPERVISOR_H__
+# error "please don't include this file directly"
+#endif
+
+#ifdef CONFIG_XEN
+#define HYPERCALL_STR(name)					\
+	"call hypercall_page + ("__stringify(__HYPERVISOR_##name)" * 32)"
+#else
+#define HYPERCALL_STR(name)					\
+	"mov hypercall_stubs,%%eax; "				\
+	"add $("__stringify(__HYPERVISOR_##name)" * 32),%%eax; "\
+	"call *%%eax"
+#endif
+
+#define _hypercall0(type, name)			\
+({						\
+	type __res;				\
+	asm volatile (				\
+		HYPERCALL_STR(name)		\
+		: "=a" (__res)			\
+		:				\
+		: "memory" );			\
+	__res;					\
+})
+
+#define _hypercall1(type, name, a1)				\
+({								\
+	type __res;						\
+	long __ign1;						\
+	asm volatile (						\
+		HYPERCALL_STR(name)				\
+		: "=a" (__res), "=b" (__ign1)			\
+		: "1" ((long)(a1))				\
+		: "memory" );					\
+	__res;							\
+})
+
+#define _hypercall2(type, name, a1, a2)				\
+({								\
+	type __res;						\
+	long __ign1, __ign2;					\
+	asm volatile (						\
+		HYPERCALL_STR(name)				\
+		: "=a" (__res), "=b" (__ign1), "=c" (__ign2)	\
+		: "1" ((long)(a1)), "2" ((long)(a2))		\
+		: "memory" );					\
+	__res;							\
+})
+
+#define _hypercall3(type, name, a1, a2, a3)			\
+({								\
+	type __res;						\
+	long __ign1, __ign2, __ign3;				\
+	asm volatile (						\
+		HYPERCALL_STR(name)				\
+		: "=a" (__res), "=b" (__ign1), "=c" (__ign2), 	\
+		"=d" (__ign3)					\
+		: "1" ((long)(a1)), "2" ((long)(a2)),		\
+		"3" ((long)(a3))				\
+		: "memory" );					\
+	__res;							\
+})
+
+#define _hypercall4(type, name, a1, a2, a3, a4)			\
+({								\
+	type __res;						\
+	long __ign1, __ign2, __ign3, __ign4;			\
+	asm volatile (						\
+		HYPERCALL_STR(name)				\
+		: "=a" (__res), "=b" (__ign1), "=c" (__ign2),	\
+		"=d" (__ign3), "=S" (__ign4)			\
+		: "1" ((long)(a1)), "2" ((long)(a2)),		\
+		"3" ((long)(a3)), "4" ((long)(a4))		\
+		: "memory" );					\
+	__res;							\
+})
+
+#define _hypercall5(type, name, a1, a2, a3, a4, a5)		\
+({								\
+	type __res;						\
+	long __ign1, __ign2, __ign3, __ign4, __ign5;		\
+	asm volatile (						\
+		HYPERCALL_STR(name)				\
+		: "=a" (__res), "=b" (__ign1), "=c" (__ign2),	\
+		"=d" (__ign3), "=S" (__ign4), "=D" (__ign5)	\
+		: "1" ((long)(a1)), "2" ((long)(a2)),		\
+		"3" ((long)(a3)), "4" ((long)(a4)),		\
+		"5" ((long)(a5))				\
+		: "memory" );					\
+	__res;							\
+})
+
+static inline int __must_check
+HYPERVISOR_set_trap_table(
+	const trap_info_t *table)
+{
+	return _hypercall1(int, set_trap_table, table);
+}
+
+static inline int __must_check
+HYPERVISOR_mmu_update(
+	mmu_update_t *req, unsigned int count, unsigned int *success_count,
+	domid_t domid)
+{
+	return _hypercall4(int, mmu_update, req, count, success_count, domid);
+}
+
+static inline int __must_check
+HYPERVISOR_mmuext_op(
+	struct mmuext_op *op, unsigned int count, unsigned int *success_count,
+	domid_t domid)
+{
+	return _hypercall4(int, mmuext_op, op, count, success_count, domid);
+}
+
+static inline int __must_check
+HYPERVISOR_set_gdt(
+	unsigned long *frame_list, unsigned int entries)
+{
+	return _hypercall2(int, set_gdt, frame_list, entries);
+}
+
+static inline int __must_check
+HYPERVISOR_stack_switch(
+	unsigned long ss, unsigned long esp)
+{
+	return _hypercall2(int, stack_switch, ss, esp);
+}
+
+static inline int __must_check
+HYPERVISOR_set_callbacks(
+	unsigned long event_selector, unsigned long event_address,
+	unsigned long failsafe_selector, unsigned long failsafe_address)
+{
+	return _hypercall4(int, set_callbacks,
+			   event_selector, event_address,
+			   failsafe_selector, failsafe_address);
+}
+
+static inline int
+HYPERVISOR_fpu_taskswitch(
+	int set)
+{
+	return _hypercall1(int, fpu_taskswitch, set);
+}
+
+static inline int __must_check
+HYPERVISOR_sched_op_compat(
+	int cmd, unsigned long arg)
+{
+	return _hypercall2(int, sched_op_compat, cmd, arg);
+}
+
+static inline int __must_check
+HYPERVISOR_sched_op(
+	int cmd, void *arg)
+{
+	return _hypercall2(int, sched_op, cmd, arg);
+}
+
+static inline long __must_check
+HYPERVISOR_set_timer_op(
+	u64 timeout)
+{
+	unsigned long timeout_hi = (unsigned long)(timeout>>32);
+	unsigned long timeout_lo = (unsigned long)timeout;
+	return _hypercall2(long, set_timer_op, timeout_lo, timeout_hi);
+}
+
+static inline int __must_check
+HYPERVISOR_platform_op(
+	struct xen_platform_op *platform_op)
+{
+	platform_op->interface_version = XENPF_INTERFACE_VERSION;
+	return _hypercall1(int, platform_op, platform_op);
+}
+
+static inline int __must_check
+HYPERVISOR_set_debugreg(
+	unsigned int reg, unsigned long value)
+{
+	return _hypercall2(int, set_debugreg, reg, value);
+}
+
+static inline unsigned long __must_check
+HYPERVISOR_get_debugreg(
+	unsigned int reg)
+{
+	return _hypercall1(unsigned long, get_debugreg, reg);
+}
+
+static inline int __must_check
+HYPERVISOR_update_descriptor(
+	u64 ma, u64 desc)
+{
+	return _hypercall4(int, update_descriptor, ma, ma>>32, desc, desc>>32);
+}
+
+static inline int __must_check
+HYPERVISOR_memory_op(
+	unsigned int cmd, void *arg)
+{
+	return _hypercall2(int, memory_op, cmd, arg);
+}
+
+static inline int __must_check
+HYPERVISOR_multicall(
+	multicall_entry_t *call_list, unsigned int nr_calls)
+{
+	return _hypercall2(int, multicall, call_list, nr_calls);
+}
+
+static inline int __must_check
+HYPERVISOR_update_va_mapping(
+	unsigned long va, pte_t new_val, unsigned long flags)
+{
+	unsigned long pte_hi = 0;
+#ifdef CONFIG_X86_PAE
+	pte_hi = new_val.pte_high;
+#endif
+	return _hypercall4(int, update_va_mapping, va,
+			   new_val.pte_low, pte_hi, flags);
+}
+
+static inline int __must_check
+HYPERVISOR_event_channel_op(
+	int cmd, void *arg)
+{
+	int rc = _hypercall2(int, event_channel_op, cmd, arg);
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (unlikely(rc == -ENOSYS)) {
+		struct evtchn_op op;
+		op.cmd = cmd;
+		memcpy(&op.u, arg, sizeof(op.u));
+		rc = _hypercall1(int, event_channel_op_compat, &op);
+		memcpy(arg, &op.u, sizeof(op.u));
+	}
+#endif
+
+	return rc;
+}
+
+static inline int __must_check
+HYPERVISOR_acm_op(
+	int cmd, void *arg)
+{
+	return _hypercall2(int, acm_op, cmd, arg);
+}
+
+static inline int __must_check
+HYPERVISOR_xen_version(
+	int cmd, void *arg)
+{
+	return _hypercall2(int, xen_version, cmd, arg);
+}
+
+static inline int __must_check
+HYPERVISOR_console_io(
+	int cmd, unsigned int count, char *str)
+{
+	return _hypercall3(int, console_io, cmd, count, str);
+}
+
+static inline int __must_check
+HYPERVISOR_physdev_op(
+	int cmd, void *arg)
+{
+	int rc = _hypercall2(int, physdev_op, cmd, arg);
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (unlikely(rc == -ENOSYS)) {
+		struct physdev_op op;
+		op.cmd = cmd;
+		memcpy(&op.u, arg, sizeof(op.u));
+		rc = _hypercall1(int, physdev_op_compat, &op);
+		memcpy(arg, &op.u, sizeof(op.u));
+	}
+#endif
+
+	return rc;
+}
+
+static inline int __must_check
+HYPERVISOR_grant_table_op(
+	unsigned int cmd, void *uop, unsigned int count)
+{
+	return _hypercall3(int, grant_table_op, cmd, uop, count);
+}
+
+static inline int __must_check
+HYPERVISOR_update_va_mapping_otherdomain(
+	unsigned long va, pte_t new_val, unsigned long flags, domid_t domid)
+{
+	unsigned long pte_hi = 0;
+#ifdef CONFIG_X86_PAE
+	pte_hi = new_val.pte_high;
+#endif
+	return _hypercall5(int, update_va_mapping_otherdomain, va,
+			   new_val.pte_low, pte_hi, flags, domid);
+}
+
+static inline int __must_check
+HYPERVISOR_vm_assist(
+	unsigned int cmd, unsigned int type)
+{
+	return _hypercall2(int, vm_assist, cmd, type);
+}
+
+static inline int __must_check
+HYPERVISOR_vcpu_op(
+	int cmd, unsigned int vcpuid, void *extra_args)
+{
+	return _hypercall3(int, vcpu_op, cmd, vcpuid, extra_args);
+}
+
+static inline int __must_check
+HYPERVISOR_suspend(
+	unsigned long srec)
+{
+	struct sched_shutdown sched_shutdown = {
+		.reason = SHUTDOWN_suspend
+	};
+
+	int rc = _hypercall3(int, sched_op, SCHEDOP_shutdown,
+			     &sched_shutdown, srec);
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (rc == -ENOSYS)
+		rc = _hypercall3(int, sched_op_compat, SCHEDOP_shutdown,
+				 SHUTDOWN_suspend, srec);
+#endif
+
+	return rc;
+}
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+static inline int
+HYPERVISOR_nmi_op(
+	unsigned long op, void *arg)
+{
+	return _hypercall2(int, nmi_op, op, arg);
+}
+#endif
+
+#ifndef CONFIG_XEN
+static inline unsigned long __must_check
+HYPERVISOR_hvm_op(
+    int op, void *arg)
+{
+    return _hypercall2(unsigned long, hvm_op, op, arg);
+}
+#endif
+
+static inline int __must_check
+HYPERVISOR_callback_op(
+	int cmd, const void *arg)
+{
+	return _hypercall2(int, callback_op, cmd, arg);
+}
+
+static inline int __must_check
+HYPERVISOR_xenoprof_op(
+	int op, void *arg)
+{
+	return _hypercall2(int, xenoprof_op, op, arg);
+}
+
+static inline int __must_check
+HYPERVISOR_kexec_op(
+	unsigned long op, void *args)
+{
+	return _hypercall2(int, kexec_op, op, args);
+}
+
+
+
+#endif /* __HYPERCALL_H__ */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/hypervisor.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/hypervisor.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/hypervisor.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/hypervisor.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,259 @@
+/******************************************************************************
+ * hypervisor.h
+ * 
+ * Linux-specific hypervisor handling.
+ * 
+ * Copyright (c) 2002-2004, K A Fraser
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __HYPERVISOR_H__
+#define __HYPERVISOR_H__
+
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/version.h>
+#include <linux/errno.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/platform.h>
+#include <xen/interface/event_channel.h>
+#include <xen/interface/physdev.h>
+#include <xen/interface/sched.h>
+#include <xen/interface/nmi.h>
+#include <asm/ptrace.h>
+#include <asm/page.h>
+#if defined(__i386__)
+#  ifdef CONFIG_X86_PAE
+#   include <asm-generic/pgtable-nopud.h>
+#  else
+#   include <asm-generic/pgtable-nopmd.h>
+#  endif
+#elif defined(__x86_64__) && LINUX_VERSION_CODE < KERNEL_VERSION(2,6,11)
+#  include <asm-generic/pgtable-nopud.h>
+#endif
+
+extern shared_info_t *HYPERVISOR_shared_info;
+
+#define vcpu_info(cpu) (HYPERVISOR_shared_info->vcpu_info + (cpu))
+#ifdef CONFIG_SMP
+#define current_vcpu_info() vcpu_info(smp_processor_id())
+#else
+#define current_vcpu_info() vcpu_info(0)
+#endif
+
+#ifdef CONFIG_X86_32
+extern unsigned long hypervisor_virt_start;
+#endif
+
+/* arch/xen/i386/kernel/setup.c */
+extern start_info_t *xen_start_info;
+#ifdef CONFIG_XEN_PRIVILEGED_GUEST
+#define is_initial_xendomain() (xen_start_info->flags & SIF_INITDOMAIN)
+#else
+#define is_initial_xendomain() 0
+#endif
+
+/* arch/xen/kernel/evtchn.c */
+/* Force a proper event-channel callback from Xen. */
+void force_evtchn_callback(void);
+
+/* arch/xen/kernel/process.c */
+void xen_cpu_idle (void);
+
+/* arch/xen/i386/kernel/hypervisor.c */
+void do_hypervisor_callback(struct pt_regs *regs);
+
+/* arch/xen/i386/mm/hypervisor.c */
+/*
+ * NB. ptr values should be PHYSICAL, not MACHINE. 'vals' should be already
+ * be MACHINE addresses.
+ */
+
+void xen_pt_switch(unsigned long ptr);
+void xen_new_user_pt(unsigned long ptr); /* x86_64 only */
+void xen_load_gs(unsigned int selector); /* x86_64 only */
+void xen_tlb_flush(void);
+void xen_invlpg(unsigned long ptr);
+
+void xen_l1_entry_update(pte_t *ptr, pte_t val);
+void xen_l2_entry_update(pmd_t *ptr, pmd_t val);
+void xen_l3_entry_update(pud_t *ptr, pud_t val); /* x86_64/PAE */
+void xen_l4_entry_update(pgd_t *ptr, pgd_t val); /* x86_64 only */
+void xen_pgd_pin(unsigned long ptr);
+void xen_pgd_unpin(unsigned long ptr);
+
+void xen_set_ldt(const void *ptr, unsigned int ents);
+
+#ifdef CONFIG_SMP
+#include <linux/cpumask.h>
+void xen_tlb_flush_all(void);
+void xen_invlpg_all(unsigned long ptr);
+void xen_tlb_flush_mask(cpumask_t *mask);
+void xen_invlpg_mask(cpumask_t *mask, unsigned long ptr);
+#endif
+
+/* Returns zero on success else negative errno. */
+int xen_create_contiguous_region(
+    unsigned long vstart, unsigned int order, unsigned int address_bits);
+void xen_destroy_contiguous_region(
+    unsigned long vstart, unsigned int order);
+
+struct page;
+
+int xen_limit_pages_to_max_mfn(
+	struct page *pages, unsigned int order, unsigned int address_bits);
+
+/* Turn jiffies into Xen system time. */
+u64 jiffies_to_st(unsigned long jiffies);
+
+#ifdef CONFIG_XEN_SCRUB_PAGES
+void scrub_pages(void *, unsigned int);
+#else
+#define scrub_pages(_p,_n) ((void)0)
+#endif
+
+#include <xen/hypercall.h>
+
+#if defined(CONFIG_X86_64)
+#define MULTI_UVMFLAGS_INDEX 2
+#define MULTI_UVMDOMID_INDEX 3
+#else
+#define MULTI_UVMFLAGS_INDEX 3
+#define MULTI_UVMDOMID_INDEX 4
+#endif
+
+#ifdef CONFIG_XEN
+#define is_running_on_xen() 1
+#else
+extern char *hypercall_stubs;
+#define is_running_on_xen() (!!hypercall_stubs)
+#endif
+
+static inline int
+HYPERVISOR_yield(
+	void)
+{
+	int rc = HYPERVISOR_sched_op(SCHEDOP_yield, NULL);
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (rc == -ENOSYS)
+		rc = HYPERVISOR_sched_op_compat(SCHEDOP_yield, 0);
+#endif
+
+	return rc;
+}
+
+static inline int
+HYPERVISOR_block(
+	void)
+{
+	int rc = HYPERVISOR_sched_op(SCHEDOP_block, NULL);
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (rc == -ENOSYS)
+		rc = HYPERVISOR_sched_op_compat(SCHEDOP_block, 0);
+#endif
+
+	return rc;
+}
+
+static inline void /*__noreturn*/
+HYPERVISOR_shutdown(
+	unsigned int reason)
+{
+	struct sched_shutdown sched_shutdown = {
+		.reason = reason
+	};
+
+	VOID(HYPERVISOR_sched_op(SCHEDOP_shutdown, &sched_shutdown));
+#if CONFIG_XEN_COMPAT <= 0x030002
+	VOID(HYPERVISOR_sched_op_compat(SCHEDOP_shutdown, reason));
+#endif
+	/* Don't recurse needlessly. */
+	BUG_ON(reason != SHUTDOWN_crash);
+	for(;;);
+}
+
+static inline int __must_check
+HYPERVISOR_poll(
+	evtchn_port_t *ports, unsigned int nr_ports, u64 timeout)
+{
+	int rc;
+	struct sched_poll sched_poll = {
+		.nr_ports = nr_ports,
+		.timeout = jiffies_to_st(timeout)
+	};
+	set_xen_guest_handle(sched_poll.ports, ports);
+
+	rc = HYPERVISOR_sched_op(SCHEDOP_poll, &sched_poll);
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (rc == -ENOSYS)
+		rc = HYPERVISOR_sched_op_compat(SCHEDOP_yield, 0);
+#endif
+
+	return rc;
+}
+
+#ifdef CONFIG_XEN
+
+static inline void
+MULTI_update_va_mapping(
+    multicall_entry_t *mcl, unsigned long va,
+    pte_t new_val, unsigned long flags)
+{
+    mcl->op = __HYPERVISOR_update_va_mapping;
+    mcl->args[0] = va;
+#if defined(CONFIG_X86_64)
+    mcl->args[1] = new_val.pte;
+#elif defined(CONFIG_X86_PAE)
+    mcl->args[1] = new_val.pte_low;
+    mcl->args[2] = new_val.pte_high;
+#else
+    mcl->args[1] = new_val.pte_low;
+    mcl->args[2] = 0;
+#endif
+    mcl->args[MULTI_UVMFLAGS_INDEX] = flags;
+}
+
+static inline void
+MULTI_grant_table_op(multicall_entry_t *mcl, unsigned int cmd,
+		     void *uop, unsigned int count)
+{
+    mcl->op = __HYPERVISOR_grant_table_op;
+    mcl->args[0] = cmd;
+    mcl->args[1] = (unsigned long)uop;
+    mcl->args[2] = count;
+}
+
+#else /* !defined(CONFIG_XEN) */
+
+/* Multicalls not supported for HVM guests. */
+#define MULTI_update_va_mapping(a,b,c,d) ((void)0)
+#define MULTI_grant_table_op(a,b,c,d) ((void)0)
+
+#endif
+
+#endif /* __HYPERVISOR_H__ */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/io.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/io.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/io.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/io.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,389 @@
+#ifndef _ASM_IO_H
+#define _ASM_IO_H
+
+#include <linux/string.h>
+#include <linux/compiler.h>
+
+/*
+ * This file contains the definitions for the x86 IO instructions
+ * inb/inw/inl/outb/outw/outl and the "string versions" of the same
+ * (insb/insw/insl/outsb/outsw/outsl). You can also use "pausing"
+ * versions of the single-IO instructions (inb_p/inw_p/..).
+ *
+ * This file is not meant to be obfuscating: it's just complicated
+ * to (a) handle it all in a way that makes gcc able to optimize it
+ * as well as possible and (b) trying to avoid writing the same thing
+ * over and over again with slight variations and possibly making a
+ * mistake somewhere.
+ */
+
+/*
+ * Thanks to James van Artsdalen for a better timing-fix than
+ * the two short jumps: using outb's to a nonexistent port seems
+ * to guarantee better timings even on fast machines.
+ *
+ * On the other hand, I'd like to be sure of a non-existent port:
+ * I feel a bit unsafe about using 0x80 (should be safe, though)
+ *
+ *		Linus
+ */
+
+ /*
+  *  Bit simplified and optimized by Jan Hubicka
+  *  Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999.
+  *
+  *  isa_memset_io, isa_memcpy_fromio, isa_memcpy_toio added,
+  *  isa_read[wl] and isa_write[wl] fixed
+  *  - Arnaldo Carvalho de Melo <acme@conectiva.com.br>
+  */
+
+#define IO_SPACE_LIMIT 0xffff
+
+#define XQUAD_PORTIO_BASE 0xfe400000
+#define XQUAD_PORTIO_QUAD 0x40000  /* 256k per quad. */
+
+#ifdef __KERNEL__
+
+#include <asm-generic/iomap.h>
+
+#include <linux/vmalloc.h>
+#include <asm/fixmap.h>
+
+/*
+ * Convert a physical pointer to a virtual kernel pointer for /dev/mem
+ * access
+ */
+#define xlate_dev_mem_ptr(p)	__va(p)
+
+/*
+ * Convert a virtual cached pointer to an uncached pointer
+ */
+#define xlate_dev_kmem_ptr(p)	p
+
+/**
+ *	virt_to_phys	-	map virtual addresses to physical
+ *	@address: address to remap
+ *
+ *	The returned physical address is the physical (CPU) mapping for
+ *	the memory address given. It is only valid to use this function on
+ *	addresses directly mapped or allocated via kmalloc. 
+ *
+ *	This function does not give bus mappings for DMA transfers. In
+ *	almost all conceivable cases a device driver should not be using
+ *	this function
+ */
+ 
+static inline unsigned long virt_to_phys(volatile void * address)
+{
+	return __pa(address);
+}
+
+/**
+ *	phys_to_virt	-	map physical address to virtual
+ *	@address: address to remap
+ *
+ *	The returned virtual address is a current CPU mapping for
+ *	the memory address given. It is only valid to use this function on
+ *	addresses that have a kernel mapping
+ *
+ *	This function does not handle bus mappings for DMA transfers. In
+ *	almost all conceivable cases a device driver should not be using
+ *	this function
+ */
+
+static inline void * phys_to_virt(unsigned long address)
+{
+	return __va(address);
+}
+
+/*
+ * Change "struct page" to physical address.
+ */
+#define page_to_pseudophys(page) ((dma_addr_t)page_to_pfn(page) << PAGE_SHIFT)
+#define page_to_phys(page)	 (phys_to_machine(page_to_pseudophys(page)))
+#define page_to_bus(page)	 (phys_to_machine(page_to_pseudophys(page)))
+
+#define bio_to_pseudophys(bio)	 (page_to_pseudophys(bio_page((bio))) + \
+				  (unsigned long) bio_offset((bio)))
+#define bvec_to_pseudophys(bv)	 (page_to_pseudophys((bv)->bv_page) + \
+				  (unsigned long) (bv)->bv_offset)
+
+#define BIOVEC_PHYS_MERGEABLE(vec1, vec2)	\
+	(((bvec_to_phys((vec1)) + (vec1)->bv_len) == bvec_to_phys((vec2))) && \
+	 ((bvec_to_pseudophys((vec1)) + (vec1)->bv_len) == \
+	  bvec_to_pseudophys((vec2))))
+
+extern void __iomem * __ioremap(unsigned long offset, unsigned long size, unsigned long flags);
+
+/**
+ * ioremap     -   map bus memory into CPU space
+ * @offset:    bus address of the memory
+ * @size:      size of the resource to map
+ *
+ * ioremap performs a platform specific sequence of operations to
+ * make bus memory CPU accessible via the readb/readw/readl/writeb/
+ * writew/writel functions and the other mmio helpers. The returned
+ * address is not guaranteed to be usable directly as a virtual
+ * address. 
+ */
+
+static inline void __iomem * ioremap(unsigned long offset, unsigned long size)
+{
+	return __ioremap(offset, size, 0);
+}
+
+extern void __iomem * ioremap_nocache(unsigned long offset, unsigned long size);
+extern void iounmap(volatile void __iomem *addr);
+
+/*
+ * bt_ioremap() and bt_iounmap() are for temporary early boot-time
+ * mappings, before the real ioremap() is functional.
+ * A boot-time mapping is currently limited to at most 16 pages.
+ */
+extern void *bt_ioremap(unsigned long offset, unsigned long size);
+extern void bt_iounmap(void *addr, unsigned long size);
+
+/* Use early IO mappings for DMI because it's initialized early */
+#define dmi_ioremap bt_ioremap
+#define dmi_iounmap bt_iounmap
+#define dmi_alloc alloc_bootmem
+
+/*
+ * ISA I/O bus memory addresses are 1:1 with the physical address.
+ */
+#define isa_virt_to_bus(_x) ({ BUG(); virt_to_bus(_x); })
+#define isa_page_to_bus(_x) isa_page_to_bus_is_UNSUPPORTED->x
+#define isa_bus_to_virt(_x) (void *)(__fix_to_virt(FIX_ISAMAP_BEGIN) + (_x))
+
+/*
+ * However PCI ones are not necessarily 1:1 and therefore these interfaces
+ * are forbidden in portable PCI drivers.
+ *
+ * Allow them on x86 for legacy drivers, though.
+ */
+#define virt_to_bus(_x) phys_to_machine(__pa(_x))
+#define bus_to_virt(_x) __va(machine_to_phys(_x))
+
+/*
+ * readX/writeX() are used to access memory mapped devices. On some
+ * architectures the memory mapped IO stuff needs to be accessed
+ * differently. On the x86 architecture, we just read/write the
+ * memory location directly.
+ */
+
+static inline unsigned char readb(const volatile void __iomem *addr)
+{
+	return *(volatile unsigned char __force *) addr;
+}
+static inline unsigned short readw(const volatile void __iomem *addr)
+{
+	return *(volatile unsigned short __force *) addr;
+}
+static inline unsigned int readl(const volatile void __iomem *addr)
+{
+	return *(volatile unsigned int __force *) addr;
+}
+#define readb_relaxed(addr) readb(addr)
+#define readw_relaxed(addr) readw(addr)
+#define readl_relaxed(addr) readl(addr)
+#define __raw_readb readb
+#define __raw_readw readw
+#define __raw_readl readl
+
+static inline void writeb(unsigned char b, volatile void __iomem *addr)
+{
+	*(volatile unsigned char __force *) addr = b;
+}
+static inline void writew(unsigned short b, volatile void __iomem *addr)
+{
+	*(volatile unsigned short __force *) addr = b;
+}
+static inline void writel(unsigned int b, volatile void __iomem *addr)
+{
+	*(volatile unsigned int __force *) addr = b;
+}
+#define __raw_writeb writeb
+#define __raw_writew writew
+#define __raw_writel writel
+
+#define mmiowb()
+
+static inline void memset_io(volatile void __iomem *addr, unsigned char val, int count)
+{
+	memset((void __force *) addr, val, count);
+}
+static inline void memcpy_fromio(void *dst, const volatile void __iomem *src, int count)
+{
+	__memcpy(dst, (void __force *) src, count);
+}
+static inline void memcpy_toio(volatile void __iomem *dst, const void *src, int count)
+{
+	__memcpy((void __force *) dst, src, count);
+}
+
+/*
+ * ISA space is 'always mapped' on a typical x86 system, no need to
+ * explicitly ioremap() it. The fact that the ISA IO space is mapped
+ * to PAGE_OFFSET is pure coincidence - it does not mean ISA values
+ * are physical addresses. The following constant pointer can be
+ * used as the IO-area pointer (it can be iounmapped as well, so the
+ * analogy with PCI is quite large):
+ */
+#define __ISA_IO_base ((char __iomem *)(fix_to_virt(FIX_ISAMAP_BEGIN)))
+
+/*
+ * Again, i386 does not require mem IO specific function.
+ */
+
+#define eth_io_copy_and_sum(a,b,c,d)		eth_copy_and_sum((a),(void __force *)(b),(c),(d))
+
+/**
+ *	check_signature		-	find BIOS signatures
+ *	@io_addr: mmio address to check 
+ *	@signature:  signature block
+ *	@length: length of signature
+ *
+ *	Perform a signature comparison with the mmio address io_addr. This
+ *	address should have been obtained by ioremap.
+ *	Returns 1 on a match.
+ */
+ 
+static inline int check_signature(volatile void __iomem * io_addr,
+	const unsigned char *signature, int length)
+{
+	int retval = 0;
+	do {
+		if (readb(io_addr) != *signature)
+			goto out;
+		io_addr++;
+		signature++;
+		length--;
+	} while (length);
+	retval = 1;
+out:
+	return retval;
+}
+
+/*
+ *	Cache management
+ *
+ *	This needed for two cases
+ *	1. Out of order aware processors
+ *	2. Accidentally out of order processors (PPro errata #51)
+ */
+ 
+#if defined(CONFIG_X86_OOSTORE) || defined(CONFIG_X86_PPRO_FENCE)
+
+static inline void flush_write_buffers(void)
+{
+	__asm__ __volatile__ ("lock; addl $0,0(%%esp)": : :"memory");
+}
+
+#define dma_cache_inv(_start,_size)		flush_write_buffers()
+#define dma_cache_wback(_start,_size)		flush_write_buffers()
+#define dma_cache_wback_inv(_start,_size)	flush_write_buffers()
+
+#else
+
+/* Nothing to do */
+
+#define dma_cache_inv(_start,_size)		do { } while (0)
+#define dma_cache_wback(_start,_size)		do { } while (0)
+#define dma_cache_wback_inv(_start,_size)	do { } while (0)
+#define flush_write_buffers()
+
+#endif
+
+#endif /* __KERNEL__ */
+
+#ifdef SLOW_IO_BY_JUMPING
+#define __SLOW_DOWN_IO "jmp 1f; 1: jmp 1f; 1:"
+#else
+#define __SLOW_DOWN_IO "outb %%al,$0x80;"
+#endif
+
+static inline void slow_down_io(void) {
+	__asm__ __volatile__(
+		__SLOW_DOWN_IO
+#ifdef REALLY_SLOW_IO
+		__SLOW_DOWN_IO __SLOW_DOWN_IO __SLOW_DOWN_IO
+#endif
+		: : );
+}
+
+#ifdef CONFIG_X86_NUMAQ
+extern void *xquad_portio;    /* Where the IO area was mapped */
+#define XQUAD_PORT_ADDR(port, quad) (xquad_portio + (XQUAD_PORTIO_QUAD*quad) + port)
+#define __BUILDIO(bwl,bw,type) \
+static inline void out##bwl##_quad(unsigned type value, int port, int quad) { \
+	if (xquad_portio) \
+		write##bwl(value, XQUAD_PORT_ADDR(port, quad)); \
+	else \
+		out##bwl##_local(value, port); \
+} \
+static inline void out##bwl(unsigned type value, int port) { \
+	out##bwl##_quad(value, port, 0); \
+} \
+static inline unsigned type in##bwl##_quad(int port, int quad) { \
+	if (xquad_portio) \
+		return read##bwl(XQUAD_PORT_ADDR(port, quad)); \
+	else \
+		return in##bwl##_local(port); \
+} \
+static inline unsigned type in##bwl(int port) { \
+	return in##bwl##_quad(port, 0); \
+}
+#else
+#define __BUILDIO(bwl,bw,type) \
+static inline void out##bwl(unsigned type value, int port) { \
+	out##bwl##_local(value, port); \
+} \
+static inline unsigned type in##bwl(int port) { \
+	return in##bwl##_local(port); \
+}
+#endif
+
+
+#define BUILDIO(bwl,bw,type) \
+static inline void out##bwl##_local(unsigned type value, int port) { \
+	__asm__ __volatile__("out" #bwl " %" #bw "0, %w1" : : "a"(value), "Nd"(port)); \
+} \
+static inline unsigned type in##bwl##_local(int port) { \
+	unsigned type value; \
+	__asm__ __volatile__("in" #bwl " %w1, %" #bw "0" : "=a"(value) : "Nd"(port)); \
+	return value; \
+} \
+static inline void out##bwl##_local_p(unsigned type value, int port) { \
+	out##bwl##_local(value, port); \
+	slow_down_io(); \
+} \
+static inline unsigned type in##bwl##_local_p(int port) { \
+	unsigned type value = in##bwl##_local(port); \
+	slow_down_io(); \
+	return value; \
+} \
+__BUILDIO(bwl,bw,type) \
+static inline void out##bwl##_p(unsigned type value, int port) { \
+	out##bwl(value, port); \
+	slow_down_io(); \
+} \
+static inline unsigned type in##bwl##_p(int port) { \
+	unsigned type value = in##bwl(port); \
+	slow_down_io(); \
+	return value; \
+} \
+static inline void outs##bwl(int port, const void *addr, unsigned long count) { \
+	__asm__ __volatile__("rep; outs" #bwl : "+S"(addr), "+c"(count) : "d"(port)); \
+} \
+static inline void ins##bwl(int port, void *addr, unsigned long count) { \
+	__asm__ __volatile__("rep; ins" #bwl : "+D"(addr), "+c"(count) : "d"(port)); \
+}
+
+BUILDIO(b,b,char)
+BUILDIO(w,w,short)
+BUILDIO(l,,int)
+
+/* We will be supplying our own /dev/mem implementation */
+#define ARCH_HAS_DEV_MEM
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/irqflags.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/irqflags.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/irqflags.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/irqflags.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,127 @@
+/*
+ * include/asm-i386/irqflags.h
+ *
+ * IRQ flags handling
+ *
+ * This file gets included from lowlevel asm headers too, to provide
+ * wrapped versions of the local_irq_*() APIs, based on the
+ * raw_local_irq_*() functions from the lowlevel headers.
+ */
+#ifndef _ASM_IRQFLAGS_H
+#define _ASM_IRQFLAGS_H
+
+#ifndef __ASSEMBLY__
+
+/* 
+ * The use of 'barrier' in the following reflects their use as local-lock
+ * operations. Reentrancy must be prevented (e.g., __cli()) /before/ following
+ * critical operations are executed. All critical operations must complete
+ * /before/ reentrancy is permitted (e.g., __sti()). Alpha architecture also
+ * includes these barriers, for example.
+ */
+
+#define __raw_local_save_flags() (current_vcpu_info()->evtchn_upcall_mask)
+
+#define raw_local_save_flags(flags) \
+		do { (flags) = __raw_local_save_flags(); } while (0)
+
+#define raw_local_irq_restore(x)					\
+do {									\
+	vcpu_info_t *_vcpu;						\
+	barrier();							\
+	_vcpu = current_vcpu_info();					\
+	if ((_vcpu->evtchn_upcall_mask = (x)) == 0) {			\
+		barrier(); /* unmask then check (avoid races) */	\
+		if (unlikely(_vcpu->evtchn_upcall_pending))		\
+			force_evtchn_callback();			\
+	}								\
+} while (0)
+
+#define raw_local_irq_disable()						\
+do {									\
+	current_vcpu_info()->evtchn_upcall_mask = 1;			\
+	barrier();							\
+} while (0)
+
+#define raw_local_irq_enable()						\
+do {									\
+	vcpu_info_t *_vcpu;						\
+	barrier();							\
+	_vcpu = current_vcpu_info();					\
+	_vcpu->evtchn_upcall_mask = 0;					\
+	barrier(); /* unmask then check (avoid races) */		\
+	if (unlikely(_vcpu->evtchn_upcall_pending))			\
+		force_evtchn_callback();				\
+} while (0)
+
+/*
+ * Used in the idle loop; sti takes one instruction cycle
+ * to complete:
+ */
+void raw_safe_halt(void);
+
+/*
+ * Used when interrupts are already enabled or to
+ * shutdown the processor:
+ */
+void halt(void);
+
+static inline int raw_irqs_disabled_flags(unsigned long flags)
+{
+	return (flags != 0);
+}
+
+#define raw_irqs_disabled()						\
+({									\
+	unsigned long flags = __raw_local_save_flags();			\
+									\
+	raw_irqs_disabled_flags(flags);					\
+})
+
+/*
+ * For spinlocks, etc:
+ */
+#define __raw_local_irq_save()						\
+({									\
+	unsigned long flags = __raw_local_save_flags();			\
+									\
+	raw_local_irq_disable();					\
+									\
+	flags;								\
+})
+
+#define raw_local_irq_save(flags) \
+		do { (flags) = __raw_local_irq_save(); } while (0)
+
+#endif /* __ASSEMBLY__ */
+
+/*
+ * Do the CPU's IRQ-state tracing from assembly code. We call a
+ * C function, so save all the C-clobbered registers:
+ */
+#ifdef CONFIG_TRACE_IRQFLAGS
+
+# define TRACE_IRQS_ON				\
+	pushl %eax;				\
+	pushl %ecx;				\
+	pushl %edx;				\
+	call trace_hardirqs_on;			\
+	popl %edx;				\
+	popl %ecx;				\
+	popl %eax;
+
+# define TRACE_IRQS_OFF				\
+	pushl %eax;				\
+	pushl %ecx;				\
+	pushl %edx;				\
+	call trace_hardirqs_off;		\
+	popl %edx;				\
+	popl %ecx;				\
+	popl %eax;
+
+#else
+# define TRACE_IRQS_ON
+# define TRACE_IRQS_OFF
+#endif
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/maddr.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/maddr.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/maddr.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/maddr.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,193 @@
+#ifndef _I386_MADDR_H
+#define _I386_MADDR_H
+
+#include <xen/features.h>
+#include <xen/interface/xen.h>
+
+/**** MACHINE <-> PHYSICAL CONVERSION MACROS ****/
+#define INVALID_P2M_ENTRY	(~0UL)
+#define FOREIGN_FRAME_BIT	(1UL<<31)
+#define FOREIGN_FRAME(m)	((m) | FOREIGN_FRAME_BIT)
+
+/* Definitions for machine and pseudophysical addresses. */
+#ifdef CONFIG_X86_PAE
+typedef unsigned long long paddr_t;
+typedef unsigned long long maddr_t;
+#else
+typedef unsigned long paddr_t;
+typedef unsigned long maddr_t;
+#endif
+
+#ifdef CONFIG_XEN
+
+extern unsigned long *phys_to_machine_mapping;
+extern unsigned long  max_mapnr;
+
+#undef machine_to_phys_mapping
+extern unsigned long *machine_to_phys_mapping;
+extern unsigned int   machine_to_phys_order;
+
+static inline unsigned long pfn_to_mfn(unsigned long pfn)
+{
+	if (xen_feature(XENFEAT_auto_translated_physmap))
+		return pfn;
+	BUG_ON(max_mapnr && pfn >= max_mapnr);
+	return phys_to_machine_mapping[pfn] & ~FOREIGN_FRAME_BIT;
+}
+
+static inline int phys_to_machine_mapping_valid(unsigned long pfn)
+{
+	if (xen_feature(XENFEAT_auto_translated_physmap))
+		return 1;
+	BUG_ON(max_mapnr && pfn >= max_mapnr);
+	return (phys_to_machine_mapping[pfn] != INVALID_P2M_ENTRY);
+}
+
+static inline unsigned long mfn_to_pfn(unsigned long mfn)
+{
+	unsigned long pfn;
+
+	if (xen_feature(XENFEAT_auto_translated_physmap))
+		return mfn;
+
+	if (unlikely((mfn >> machine_to_phys_order) != 0))
+		return max_mapnr;
+
+	/* The array access can fail (e.g., device space beyond end of RAM). */
+	asm (
+		"1:	movl %1,%0\n"
+		"2:\n"
+		".section .fixup,\"ax\"\n"
+		"3:	movl %2,%0\n"
+		"	jmp  2b\n"
+		".previous\n"
+		".section __ex_table,\"a\"\n"
+		"	.align 4\n"
+		"	.long 1b,3b\n"
+		".previous"
+		: "=r" (pfn)
+		: "m" (machine_to_phys_mapping[mfn]), "m" (max_mapnr) );
+
+	return pfn;
+}
+
+/*
+ * We detect special mappings in one of two ways:
+ *  1. If the MFN is an I/O page then Xen will set the m2p entry
+ *     to be outside our maximum possible pseudophys range.
+ *  2. If the MFN belongs to a different domain then we will certainly
+ *     not have MFN in our p2m table. Conversely, if the page is ours,
+ *     then we'll have p2m(m2p(MFN))==MFN.
+ * If we detect a special mapping then it doesn't have a 'struct page'.
+ * We force !pfn_valid() by returning an out-of-range pointer.
+ *
+ * NB. These checks require that, for any MFN that is not in our reservation,
+ * there is no PFN such that p2m(PFN) == MFN. Otherwise we can get confused if
+ * we are foreign-mapping the MFN, and the other domain as m2p(MFN) == PFN.
+ * Yikes! Various places must poke in INVALID_P2M_ENTRY for safety.
+ *
+ * NB2. When deliberately mapping foreign pages into the p2m table, you *must*
+ *      use FOREIGN_FRAME(). This will cause pte_pfn() to choke on it, as we
+ *      require. In all the cases we care about, the FOREIGN_FRAME bit is
+ *      masked (e.g., pfn_to_mfn()) so behaviour there is correct.
+ */
+static inline unsigned long mfn_to_local_pfn(unsigned long mfn)
+{
+	unsigned long pfn = mfn_to_pfn(mfn);
+	if ((pfn < max_mapnr)
+	    && !xen_feature(XENFEAT_auto_translated_physmap)
+	    && (phys_to_machine_mapping[pfn] != mfn))
+		return max_mapnr; /* force !pfn_valid() */
+	return pfn;
+}
+
+static inline void set_phys_to_machine(unsigned long pfn, unsigned long mfn)
+{
+	BUG_ON(max_mapnr && pfn >= max_mapnr);
+	if (xen_feature(XENFEAT_auto_translated_physmap)) {
+		BUG_ON(pfn != mfn && mfn != INVALID_P2M_ENTRY);
+		return;
+	}
+	phys_to_machine_mapping[pfn] = mfn;
+}
+
+static inline maddr_t phys_to_machine(paddr_t phys)
+{
+	maddr_t machine = pfn_to_mfn(phys >> PAGE_SHIFT);
+	machine = (machine << PAGE_SHIFT) | (phys & ~PAGE_MASK);
+	return machine;
+}
+
+static inline paddr_t machine_to_phys(maddr_t machine)
+{
+	paddr_t phys = mfn_to_pfn(machine >> PAGE_SHIFT);
+	phys = (phys << PAGE_SHIFT) | (machine & ~PAGE_MASK);
+	return phys;
+}
+
+#ifdef CONFIG_X86_PAE
+static inline paddr_t pte_phys_to_machine(paddr_t phys)
+{
+	/*
+	 * In PAE mode, the NX bit needs to be dealt with in the value
+	 * passed to pfn_to_mfn(). On x86_64, we need to mask it off,
+	 * but for i386 the conversion to ulong for the argument will
+	 * clip it off.
+	 */
+	maddr_t machine = pfn_to_mfn(phys >> PAGE_SHIFT);
+	machine = (machine << PAGE_SHIFT) | (phys & ~PHYSICAL_PAGE_MASK);
+	return machine;
+}
+
+static inline paddr_t pte_machine_to_phys(maddr_t machine)
+{
+	/*
+	 * In PAE mode, the NX bit needs to be dealt with in the value
+	 * passed to mfn_to_pfn(). On x86_64, we need to mask it off,
+	 * but for i386 the conversion to ulong for the argument will
+	 * clip it off.
+	 */
+	paddr_t phys = mfn_to_pfn(machine >> PAGE_SHIFT);
+	phys = (phys << PAGE_SHIFT) | (machine & ~PHYSICAL_PAGE_MASK);
+	return phys;
+}
+#endif
+
+#ifdef CONFIG_X86_PAE
+#define __pte_ma(x)	((pte_t) { (x), (maddr_t)(x) >> 32 } )
+static inline pte_t pfn_pte_ma(unsigned long page_nr, pgprot_t pgprot)
+{
+	pte_t pte;
+
+	pte.pte_high = (page_nr >> (32 - PAGE_SHIFT)) | \
+					(pgprot_val(pgprot) >> 32);
+	pte.pte_high &= (__supported_pte_mask >> 32);
+	pte.pte_low = ((page_nr << PAGE_SHIFT) | pgprot_val(pgprot)) & \
+							__supported_pte_mask;
+	return pte;
+}
+#else
+#define __pte_ma(x)	((pte_t) { (x) } )
+#define pfn_pte_ma(pfn, prot)	__pte_ma(((pfn) << PAGE_SHIFT) | pgprot_val(prot))
+#endif
+
+#else /* !CONFIG_XEN */
+
+#define pfn_to_mfn(pfn) (pfn)
+#define mfn_to_pfn(mfn) (mfn)
+#define mfn_to_local_pfn(mfn) (mfn)
+#define set_phys_to_machine(pfn, mfn) ((void)0)
+#define phys_to_machine_mapping_valid(pfn) (1)
+#define phys_to_machine(phys) ((maddr_t)(phys))
+#define machine_to_phys(mach) ((paddr_t)(mach))
+#define pfn_pte_ma(pfn, prot) pfn_pte(pfn, prot)
+#define __pte_ma(x) __pte(x)
+
+#endif /* !CONFIG_XEN */
+
+/* VIRT <-> MACHINE conversion */
+#define virt_to_machine(v)	(phys_to_machine(__pa(v)))
+#define virt_to_mfn(v)		(pfn_to_mfn(__pa(v) >> PAGE_SHIFT))
+#define mfn_to_virt(m)		(__va(mfn_to_pfn(m) << PAGE_SHIFT))
+
+#endif /* _I386_MADDR_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/mmu_context.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/mmu_context.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/mmu_context.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/mmu_context.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,108 @@
+#ifndef __I386_SCHED_H
+#define __I386_SCHED_H
+
+#include <asm/desc.h>
+#include <asm/atomic.h>
+#include <asm/pgalloc.h>
+#include <asm/tlbflush.h>
+
+/*
+ * Used for LDT copy/destruction.
+ */
+int init_new_context(struct task_struct *tsk, struct mm_struct *mm);
+void destroy_context(struct mm_struct *mm);
+
+
+static inline void enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk)
+{
+#if 0 /* XEN: no lazy tlb */
+	unsigned cpu = smp_processor_id();
+	if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_OK)
+		per_cpu(cpu_tlbstate, cpu).state = TLBSTATE_LAZY;
+#endif
+}
+
+#define prepare_arch_switch(next)	__prepare_arch_switch()
+
+static inline void __prepare_arch_switch(void)
+{
+	/*
+	 * Save away %fs and %gs. No need to save %es and %ds, as those
+	 * are always kernel segments while inside the kernel. Must
+	 * happen before reload of cr3/ldt (i.e., not in __switch_to).
+	 */
+	asm volatile ( "mov %%fs,%0 ; mov %%gs,%1"
+		: "=m" (current->thread.fs),
+		  "=m" (current->thread.gs));
+	asm volatile ( "movl %0,%%fs ; movl %0,%%gs"
+		: : "r" (0) );
+}
+
+extern void mm_pin(struct mm_struct *mm);
+extern void mm_unpin(struct mm_struct *mm);
+void mm_pin_all(void);
+
+static inline void switch_mm(struct mm_struct *prev,
+			     struct mm_struct *next,
+			     struct task_struct *tsk)
+{
+	int cpu = smp_processor_id();
+	struct mmuext_op _op[2], *op = _op;
+
+	if (likely(prev != next)) {
+		BUG_ON(!xen_feature(XENFEAT_writable_page_tables) &&
+		       !test_bit(PG_pinned, &virt_to_page(next->pgd)->flags));
+
+		/* stop flush ipis for the previous mm */
+		cpu_clear(cpu, prev->cpu_vm_mask);
+#if 0 /* XEN: no lazy tlb */
+		per_cpu(cpu_tlbstate, cpu).state = TLBSTATE_OK;
+		per_cpu(cpu_tlbstate, cpu).active_mm = next;
+#endif
+		cpu_set(cpu, next->cpu_vm_mask);
+
+		/* Re-load page tables: load_cr3(next->pgd) */
+		op->cmd = MMUEXT_NEW_BASEPTR;
+		op->arg1.mfn = pfn_to_mfn(__pa(next->pgd) >> PAGE_SHIFT);
+		op++;
+
+		/*
+		 * load the LDT, if the LDT is different:
+		 */
+		if (unlikely(prev->context.ldt != next->context.ldt)) {
+			/* load_LDT_nolock(&next->context, cpu) */
+			op->cmd = MMUEXT_SET_LDT;
+			op->arg1.linear_addr = (unsigned long)next->context.ldt;
+			op->arg2.nr_ents     = next->context.size;
+			op++;
+		}
+
+		BUG_ON(HYPERVISOR_mmuext_op(_op, op-_op, NULL, DOMID_SELF));
+	}
+#if 0 /* XEN: no lazy tlb */
+	else {
+		per_cpu(cpu_tlbstate, cpu).state = TLBSTATE_OK;
+		BUG_ON(per_cpu(cpu_tlbstate, cpu).active_mm != next);
+
+		if (!cpu_test_and_set(cpu, next->cpu_vm_mask)) {
+			/* We were in lazy tlb mode and leave_mm disabled 
+			 * tlb flush IPI delivery. We must reload %cr3.
+			 */
+			load_cr3(next->pgd);
+			load_LDT_nolock(&next->context, cpu);
+		}
+	}
+#endif
+}
+
+#define deactivate_mm(tsk, mm) \
+	asm("movl %0,%%fs ; movl %0,%%gs": :"r" (0))
+
+static inline void activate_mm(struct mm_struct *prev, struct mm_struct *next)
+{
+	if (!test_bit(PG_pinned, &virt_to_page(next->pgd)->flags))
+		mm_pin(next);
+	switch_mm(prev, next, NULL);
+}
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/mmu.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/mmu.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/mmu.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/mmu.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,29 @@
+#ifndef __i386_MMU_H
+#define __i386_MMU_H
+
+#include <asm/semaphore.h>
+/*
+ * The i386 doesn't have a mmu context, but
+ * we put the segment information here.
+ *
+ * cpu_vm_mask is used to optimize ldt flushing.
+ */
+typedef struct { 
+	int size;
+	struct semaphore sem;
+	void *ldt;
+	void *vdso;
+#ifdef CONFIG_XEN
+	int has_foreign_mappings;
+#endif
+} mm_context_t;
+
+/* mm/memory.c:exit_mmap hook */
+extern void _arch_exit_mmap(struct mm_struct *mm);
+#define arch_exit_mmap(_mm) _arch_exit_mmap(_mm)
+
+/* kernel/fork.c:dup_mmap hook */
+extern void _arch_dup_mmap(struct mm_struct *mm);
+#define arch_dup_mmap(mm, oldmm) ((void)(oldmm), _arch_dup_mmap(mm))
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/page.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/page.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/page.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/page.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,233 @@
+#ifndef _I386_PAGE_H
+#define _I386_PAGE_H
+
+/* PAGE_SHIFT determines the page size */
+#define PAGE_SHIFT	12
+#define PAGE_SIZE	(1UL << PAGE_SHIFT)
+#define PAGE_MASK	(~(PAGE_SIZE-1))
+
+#ifdef CONFIG_X86_PAE
+#define __PHYSICAL_MASK_SHIFT	40
+#define __PHYSICAL_MASK		((1ULL << __PHYSICAL_MASK_SHIFT) - 1)
+#define PHYSICAL_PAGE_MASK	(~((1ULL << PAGE_SHIFT) - 1) & __PHYSICAL_MASK)
+#else
+#define __PHYSICAL_MASK_SHIFT	32
+#define __PHYSICAL_MASK		(~0UL)
+#define PHYSICAL_PAGE_MASK	(PAGE_MASK & __PHYSICAL_MASK)
+#endif
+
+#define LARGE_PAGE_MASK (~(LARGE_PAGE_SIZE-1))
+#define LARGE_PAGE_SIZE (1UL << PMD_SHIFT)
+
+#ifdef __KERNEL__
+
+/*
+ * Need to repeat this here in order to not include pgtable.h (which in turn
+ * depends on definitions made here), but to be able to use the symbolic
+ * below. The preprocessor will warn if the two definitions aren't identical.
+ */
+#define _PAGE_PRESENT	0x001
+#define _PAGE_IO	0x200
+
+#ifndef __ASSEMBLY__
+
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <asm/bug.h>
+#include <xen/interface/xen.h>
+#include <xen/features.h>
+
+#ifdef CONFIG_X86_USE_3DNOW
+
+#include <asm/mmx.h>
+
+#define clear_page(page)	mmx_clear_page((void *)(page))
+#define copy_page(to,from)	mmx_copy_page(to,from)
+
+#else
+
+#define alloc_zeroed_user_highpage(vma, vaddr) alloc_page_vma(GFP_HIGHUSER | __GFP_ZERO, vma, vaddr)
+#define __HAVE_ARCH_ALLOC_ZEROED_USER_HIGHPAGE
+
+/*
+ *	On older X86 processors it's not a win to use MMX here it seems.
+ *	Maybe the K6-III ?
+ */
+ 
+#define clear_page(page)	memset((void *)(page), 0, PAGE_SIZE)
+#define copy_page(to,from)	memcpy((void *)(to), (void *)(from), PAGE_SIZE)
+
+#endif
+
+#define clear_user_page(page, vaddr, pg)	clear_page(page)
+#define copy_user_page(to, from, vaddr, pg)	copy_page(to, from)
+
+/*
+ * These are used to make use of C type-checking..
+ */
+extern int nx_enabled;
+#ifdef CONFIG_X86_PAE
+extern unsigned long long __supported_pte_mask;
+typedef struct { unsigned long pte_low, pte_high; } pte_t;
+typedef struct { unsigned long long pmd; } pmd_t;
+typedef struct { unsigned long long pgd; } pgd_t;
+typedef struct { unsigned long long pgprot; } pgprot_t;
+#define pgprot_val(x)	((x).pgprot)
+#include <asm/maddr.h>
+#define __pte(x) ({ unsigned long long _x = (x);			\
+    if ((_x & (_PAGE_PRESENT|_PAGE_IO)) == _PAGE_PRESENT)		\
+        _x = pte_phys_to_machine(_x);					\
+    ((pte_t) {(unsigned long)(_x), (unsigned long)(_x>>32)}); })
+#define __pgd(x) ({ unsigned long long _x = (x); \
+    (pgd_t) {((_x) & _PAGE_PRESENT) ? pte_phys_to_machine(_x) : (_x)}; })
+#define __pmd(x) ({ unsigned long long _x = (x); \
+    (pmd_t) {((_x) & _PAGE_PRESENT) ? pte_phys_to_machine(_x) : (_x)}; })
+static inline unsigned long long __pte_val(pte_t x)
+{
+	return ((unsigned long long)x.pte_high << 32) | x.pte_low;
+}
+static inline unsigned long long pte_val(pte_t x)
+{
+	unsigned long long ret = __pte_val(x);
+	if ((x.pte_low & (_PAGE_PRESENT|_PAGE_IO)) == _PAGE_PRESENT)
+		ret = pte_machine_to_phys(ret);
+	return ret;
+}
+#define __pmd_val(x) ((x).pmd)
+static inline unsigned long long pmd_val(pmd_t x)
+{
+	unsigned long long ret = __pmd_val(x);
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (ret) ret = pte_machine_to_phys(ret) | _PAGE_PRESENT;
+#else
+	if (ret & _PAGE_PRESENT) ret = pte_machine_to_phys(ret);
+#endif
+	return ret;
+}
+#define __pud_val(x) __pgd_val((x).pgd)
+#define __pgd_val(x) ((x).pgd)
+static inline unsigned long long pgd_val(pgd_t x)
+{
+	unsigned long long ret = __pgd_val(x);
+	if (ret & _PAGE_PRESENT) ret = pte_machine_to_phys(ret);
+	return ret;
+}
+#define HPAGE_SHIFT	21
+#else
+typedef struct { unsigned long pte_low; } pte_t;
+typedef struct { unsigned long pgd; } pgd_t;
+typedef struct { unsigned long pgprot; } pgprot_t;
+#define pgprot_val(x)	((x).pgprot)
+#include <asm/maddr.h>
+#define boot_pte_t pte_t /* or would you rather have a typedef */
+#define __pte_val(x) ((x).pte_low)
+#define pte_val(x) ((__pte_val(x) & (_PAGE_PRESENT|_PAGE_IO))	\
+		    == _PAGE_PRESENT ?				\
+		    machine_to_phys(__pte_val(x)) :		\
+		    __pte_val(x))
+#define __pte(x) ({ unsigned long _x = (x);			\
+    if ((_x & (_PAGE_PRESENT|_PAGE_IO)) == _PAGE_PRESENT)	\
+        _x = phys_to_machine(_x);				\
+    ((pte_t) { _x }); })
+#define __pmd_val(x) __pud_val((x).pud)
+#define __pud_val(x) __pgd_val((x).pgd)
+#define __pgd(x) ({ unsigned long _x = (x); \
+    (pgd_t) {((_x) & _PAGE_PRESENT) ? phys_to_machine(_x) : (_x)}; })
+#define __pgd_val(x) ((x).pgd)
+static inline unsigned long pgd_val(pgd_t x)
+{
+	unsigned long ret = __pgd_val(x);
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (ret) ret = machine_to_phys(ret) | _PAGE_PRESENT;
+#else
+	if (ret & _PAGE_PRESENT) ret = machine_to_phys(ret);
+#endif
+	return ret;
+}
+#define HPAGE_SHIFT	22
+#endif
+#define PTE_MASK	PHYSICAL_PAGE_MASK
+
+#ifdef CONFIG_HUGETLB_PAGE
+#define HPAGE_SIZE	((1UL) << HPAGE_SHIFT)
+#define HPAGE_MASK	(~(HPAGE_SIZE - 1))
+#define HUGETLB_PAGE_ORDER	(HPAGE_SHIFT - PAGE_SHIFT)
+#define HAVE_ARCH_HUGETLB_UNMAPPED_AREA
+#endif
+
+#define __pgprot(x)	((pgprot_t) { (x) } )
+
+#endif /* !__ASSEMBLY__ */
+
+/* to align the pointer to the (next) page boundary */
+#define PAGE_ALIGN(addr)	(((addr)+PAGE_SIZE-1)&PAGE_MASK)
+
+/*
+ * This handles the memory map.. We could make this a config
+ * option, but too many people screw it up, and too few need
+ * it.
+ *
+ * A __PAGE_OFFSET of 0xC0000000 means that the kernel has
+ * a virtual address space of one gigabyte, which limits the
+ * amount of physical memory you can use to about 950MB. 
+ *
+ * If you want more physical memory than this then see the CONFIG_HIGHMEM4G
+ * and CONFIG_HIGHMEM64G options in the kernel configuration.
+ */
+
+#ifndef __ASSEMBLY__
+
+struct vm_area_struct;
+
+/*
+ * This much address space is reserved for vmalloc() and iomap()
+ * as well as fixmap mappings.
+ */
+extern unsigned int __VMALLOC_RESERVE;
+
+extern int sysctl_legacy_va_layout;
+
+extern int page_is_ram(unsigned long pagenr);
+
+#endif /* __ASSEMBLY__ */
+
+#ifdef __ASSEMBLY__
+#define __PAGE_OFFSET		CONFIG_PAGE_OFFSET
+#define __PHYSICAL_START	CONFIG_PHYSICAL_START
+#else
+#define __PAGE_OFFSET		((unsigned long)CONFIG_PAGE_OFFSET)
+#define __PHYSICAL_START	((unsigned long)CONFIG_PHYSICAL_START)
+#endif
+#define __KERNEL_START		(__PAGE_OFFSET + __PHYSICAL_START)
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+#undef LOAD_OFFSET
+#define LOAD_OFFSET		0
+#endif
+
+#define PAGE_OFFSET		((unsigned long)__PAGE_OFFSET)
+#define VMALLOC_RESERVE		((unsigned long)__VMALLOC_RESERVE)
+#define MAXMEM			(__FIXADDR_TOP-__PAGE_OFFSET-__VMALLOC_RESERVE)
+#define __pa(x)			((unsigned long)(x)-PAGE_OFFSET)
+#define __va(x)			((void *)((unsigned long)(x)+PAGE_OFFSET))
+#define pfn_to_kaddr(pfn)      __va((pfn) << PAGE_SHIFT)
+#ifdef CONFIG_FLATMEM
+#define pfn_valid(pfn)		((pfn) < max_mapnr)
+#endif /* CONFIG_FLATMEM */
+#define virt_to_page(kaddr)	pfn_to_page(__pa(kaddr) >> PAGE_SHIFT)
+
+#define virt_addr_valid(kaddr)	pfn_valid(__pa(kaddr) >> PAGE_SHIFT)
+
+#define VM_DATA_DEFAULT_FLAGS \
+	(VM_READ | VM_WRITE | \
+	((current->personality & READ_IMPLIES_EXEC) ? VM_EXEC : 0 ) | \
+		 VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC)
+
+#include <asm-generic/memory_model.h>
+#include <asm-generic/page.h>
+
+#define __HAVE_ARCH_GATE_AREA 1
+#endif /* __KERNEL__ */
+
+#endif /* _I386_PAGE_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/pci.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/pci.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/pci.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/pci.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,148 @@
+#ifndef __i386_PCI_H
+#define __i386_PCI_H
+
+
+#ifdef __KERNEL__
+#include <linux/mm.h>		/* for struct page */
+
+/* Can be used to override the logic in pci_scan_bus for skipping
+   already-configured bus numbers - to be used for buggy BIOSes
+   or architectures with incomplete PCI setup by the loader */
+
+#ifdef CONFIG_PCI
+extern unsigned int pcibios_assign_all_busses(void);
+#else
+#define pcibios_assign_all_busses()	0
+#endif
+
+#include <asm/hypervisor.h>
+#define pcibios_scan_all_fns(a, b)	(!is_initial_xendomain())
+
+extern unsigned long pci_mem_start;
+#define PCIBIOS_MIN_IO		0x1000
+#define PCIBIOS_MIN_MEM		(pci_mem_start)
+
+#define PCIBIOS_MIN_CARDBUS_IO	0x4000
+
+void pcibios_config_init(void);
+struct pci_bus * pcibios_scan_root(int bus);
+
+void pcibios_set_master(struct pci_dev *dev);
+void pcibios_penalize_isa_irq(int irq, int active);
+struct irq_routing_table *pcibios_get_irq_routing_table(void);
+int pcibios_set_irq_routing(struct pci_dev *dev, int pin, int irq);
+
+/* Dynamic DMA mapping stuff.
+ * i386 has everything mapped statically.
+ */
+
+#include <linux/types.h>
+#include <linux/slab.h>
+#include <asm/scatterlist.h>
+#include <linux/string.h>
+#include <asm/io.h>
+
+struct pci_dev;
+
+#ifdef CONFIG_SWIOTLB
+
+
+/* On Xen we use SWIOTLB instead of blk-specific bounce buffers. */
+#define PCI_DMA_BUS_IS_PHYS	(0)
+
+#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)	\
+	dma_addr_t ADDR_NAME;
+#define DECLARE_PCI_UNMAP_LEN(LEN_NAME)		\
+	__u32 LEN_NAME;
+#define pci_unmap_addr(PTR, ADDR_NAME)			\
+	((PTR)->ADDR_NAME)
+#define pci_unmap_addr_set(PTR, ADDR_NAME, VAL)		\
+	(((PTR)->ADDR_NAME) = (VAL))
+#define pci_unmap_len(PTR, LEN_NAME)			\
+	((PTR)->LEN_NAME)
+#define pci_unmap_len_set(PTR, LEN_NAME, VAL)		\
+	(((PTR)->LEN_NAME) = (VAL))
+
+#else
+
+/* The PCI address space does equal the physical memory
+ * address space.  The networking and block device layers use
+ * this boolean for bounce buffer decisions.
+ */
+#define PCI_DMA_BUS_IS_PHYS	(1)
+
+/* pci_unmap_{page,single} is a nop so... */
+#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)
+#define DECLARE_PCI_UNMAP_LEN(LEN_NAME)
+#define pci_unmap_addr(PTR, ADDR_NAME)		(0)
+#define pci_unmap_addr_set(PTR, ADDR_NAME, VAL)	do { } while (0)
+#define pci_unmap_len(PTR, LEN_NAME)		(0)
+#define pci_unmap_len_set(PTR, LEN_NAME, VAL)	do { } while (0)
+
+#endif
+
+/* This is always fine. */
+#define pci_dac_dma_supported(pci_dev, mask)	(1)
+
+static inline dma64_addr_t
+pci_dac_page_to_dma(struct pci_dev *pdev, struct page *page, unsigned long offset, int direction)
+{
+	return ((dma64_addr_t) page_to_phys(page) +
+		(dma64_addr_t) offset);
+}
+
+static inline struct page *
+pci_dac_dma_to_page(struct pci_dev *pdev, dma64_addr_t dma_addr)
+{
+	return pfn_to_page(dma_addr >> PAGE_SHIFT);
+}
+
+static inline unsigned long
+pci_dac_dma_to_offset(struct pci_dev *pdev, dma64_addr_t dma_addr)
+{
+	return (dma_addr & ~PAGE_MASK);
+}
+
+static inline void
+pci_dac_dma_sync_single_for_cpu(struct pci_dev *pdev, dma64_addr_t dma_addr, size_t len, int direction)
+{
+}
+
+static inline void
+pci_dac_dma_sync_single_for_device(struct pci_dev *pdev, dma64_addr_t dma_addr, size_t len, int direction)
+{
+	flush_write_buffers();
+}
+
+#define HAVE_PCI_MMAP
+extern int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
+			       enum pci_mmap_state mmap_state, int write_combine);
+
+
+static inline void pcibios_add_platform_entries(struct pci_dev *dev)
+{
+}
+
+#ifdef CONFIG_PCI
+static inline void pci_dma_burst_advice(struct pci_dev *pdev,
+					enum pci_dma_burst_strategy *strat,
+					unsigned long *strategy_parameter)
+{
+	*strat = PCI_DMA_BURST_INFINITY;
+	*strategy_parameter = ~0UL;
+}
+#endif
+
+#endif /* __KERNEL__ */
+
+#ifdef CONFIG_XEN_PCIDEV_FRONTEND
+#include <xen/pcifront.h>
+#endif /* CONFIG_XEN_PCIDEV_FRONTEND */
+
+/* implement the pci_ DMA API in terms of the generic device dma_ one */
+#include <asm-generic/pci-dma-compat.h>
+
+/* generic pci stuff */
+#include <asm-generic/pci.h>
+
+#endif /* __i386_PCI_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/pgalloc.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/pgalloc.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/pgalloc.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/pgalloc.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,59 @@
+#ifndef _I386_PGALLOC_H
+#define _I386_PGALLOC_H
+
+#include <asm/fixmap.h>
+#include <linux/threads.h>
+#include <linux/mm.h>		/* for struct page */
+#include <asm/io.h>		/* for phys_to_virt and page_to_pseudophys */
+
+#define pmd_populate_kernel(mm, pmd, pte) \
+		set_pmd(pmd, __pmd(_PAGE_TABLE + __pa(pte)))
+
+#define pmd_populate(mm, pmd, pte) 					\
+do {									\
+	unsigned long pfn = page_to_pfn(pte);				\
+	if (test_bit(PG_pinned, &virt_to_page((mm)->pgd)->flags)) {	\
+		if (!PageHighMem(pte))					\
+			BUG_ON(HYPERVISOR_update_va_mapping(		\
+			  (unsigned long)__va(pfn << PAGE_SHIFT),	\
+			  pfn_pte(pfn, PAGE_KERNEL_RO), 0));		\
+		else if (!test_and_set_bit(PG_pinned, &pte->flags))	\
+			kmap_flush_unused();				\
+		set_pmd(pmd,						\
+		        __pmd(_PAGE_TABLE + ((paddr_t)pfn << PAGE_SHIFT))); \
+	} else							\
+		*(pmd) = __pmd(_PAGE_TABLE + ((paddr_t)pfn << PAGE_SHIFT)); \
+} while (0)
+
+/*
+ * Allocate and free page tables.
+ */
+extern pgd_t *pgd_alloc(struct mm_struct *);
+extern void pgd_free(pgd_t *pgd);
+
+extern pte_t *pte_alloc_one_kernel(struct mm_struct *, unsigned long);
+extern struct page *pte_alloc_one(struct mm_struct *, unsigned long);
+
+static inline void pte_free_kernel(pte_t *pte)
+{
+	make_lowmem_page_writable(pte, XENFEAT_writable_page_tables);
+	free_page((unsigned long)pte);
+}
+
+extern void pte_free(struct page *pte);
+
+#define __pte_free_tlb(tlb,pte) tlb_remove_page((tlb),(pte))
+
+#ifdef CONFIG_X86_PAE
+/*
+ * In the PAE case we free the pmds as part of the pgd.
+ */
+#define pmd_alloc_one(mm, addr)		({ BUG(); ((pmd_t *)2); })
+#define pmd_free(x)			do { } while (0)
+#define __pmd_free_tlb(tlb,x)		do { } while (0)
+#define pud_populate(mm, pmd, pte)	BUG()
+#endif
+
+#define check_pgt_cache()	do { } while (0)
+
+#endif /* _I386_PGALLOC_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/pgtable-2level-defs.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/pgtable-2level-defs.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/pgtable-2level-defs.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/pgtable-2level-defs.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,20 @@
+#ifndef _I386_PGTABLE_2LEVEL_DEFS_H
+#define _I386_PGTABLE_2LEVEL_DEFS_H
+
+#define HAVE_SHARED_KERNEL_PMD 0
+
+/*
+ * traditional i386 two-level paging structure:
+ */
+
+#define PGDIR_SHIFT	22
+#define PTRS_PER_PGD	1024
+
+/*
+ * the i386 is two-level, so we don't really have any
+ * PMD directory physically.
+ */
+
+#define PTRS_PER_PTE	1024
+
+#endif /* _I386_PGTABLE_2LEVEL_DEFS_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/pgtable-2level.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/pgtable-2level.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/pgtable-2level.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/pgtable-2level.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,121 @@
+#ifndef _I386_PGTABLE_2LEVEL_H
+#define _I386_PGTABLE_2LEVEL_H
+
+#include <asm-generic/pgtable-nopmd.h>
+
+#define pte_ERROR(e) \
+	printk("%s:%d: bad pte %08lx (pfn %05lx).\n", __FILE__, __LINE__, \
+	       __pte_val(e), pte_pfn(e))
+#define pgd_ERROR(e) \
+	printk("%s:%d: bad pgd %08lx (pfn %05lx).\n", __FILE__, __LINE__, \
+	       __pgd_val(e), pgd_val(e) >> PAGE_SHIFT)
+
+/*
+ * Certain architectures need to do special things when PTEs
+ * within a page table are directly modified.  Thus, the following
+ * hook is made available.
+ */
+#define set_pte(pteptr, pteval) (*(pteptr) = pteval)
+
+#define set_pte_at(_mm,addr,ptep,pteval) do {				\
+	if (((_mm) != current->mm && (_mm) != &init_mm) ||		\
+	    HYPERVISOR_update_va_mapping((addr), (pteval), 0))		\
+		set_pte((ptep), (pteval));				\
+} while (0)
+
+#define set_pte_at_sync(_mm,addr,ptep,pteval) do {			\
+	if (((_mm) != current->mm && (_mm) != &init_mm) ||		\
+	    HYPERVISOR_update_va_mapping((addr), (pteval), UVMF_INVLPG)) { \
+		set_pte((ptep), (pteval));				\
+		xen_invlpg((addr));					\
+	}								\
+} while (0)
+
+#define set_pte_atomic(pteptr, pteval) set_pte(pteptr,pteval)
+
+#define set_pmd(pmdptr, pmdval) xen_l2_entry_update((pmdptr), (pmdval))
+
+#define pte_clear(mm,addr,xp)	do { set_pte_at(mm, addr, xp, __pte(0)); } while (0)
+#define pmd_clear(xp)	do { set_pmd(xp, __pmd(0)); } while (0)
+
+#define pte_none(x) (!(x).pte_low)
+
+static inline pte_t ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
+{
+	pte_t pte = *ptep;
+	if (!pte_none(pte)) {
+		if ((mm != &init_mm) ||
+		    HYPERVISOR_update_va_mapping(addr, __pte(0), 0))
+			pte = __pte_ma(xchg(&ptep->pte_low, 0));
+	}
+	return pte;
+}
+
+#define ptep_clear_flush(vma, addr, ptep)			\
+({								\
+	pte_t *__ptep = (ptep);					\
+	pte_t __res = *__ptep;					\
+	if (!pte_none(__res) &&					\
+	    ((vma)->vm_mm != current->mm ||			\
+	     HYPERVISOR_update_va_mapping(addr, __pte(0),	\
+			(unsigned long)(vma)->vm_mm->cpu_vm_mask.bits| \
+				UVMF_INVLPG|UVMF_MULTI))) {	\
+		__ptep->pte_low = 0;				\
+		flush_tlb_page(vma, addr);			\
+	}							\
+	__res;							\
+})
+
+#define pte_same(a, b)		((a).pte_low == (b).pte_low)
+
+#define __pte_mfn(_pte) ((_pte).pte_low >> PAGE_SHIFT)
+#define pte_mfn(_pte) ((_pte).pte_low & _PAGE_PRESENT ? \
+	__pte_mfn(_pte) : pfn_to_mfn(__pte_mfn(_pte)))
+#define pte_pfn(_pte) ((_pte).pte_low & _PAGE_IO ? max_mapnr :	\
+		       (_pte).pte_low & _PAGE_PRESENT ?		\
+		       mfn_to_local_pfn(__pte_mfn(_pte)) :	\
+		       __pte_mfn(_pte))
+
+#define pte_page(_pte) pfn_to_page(pte_pfn(_pte))
+
+#define pfn_pte(pfn, prot)	__pte(((pfn) << PAGE_SHIFT) | pgprot_val(prot))
+#define pfn_pmd(pfn, prot)	__pmd(((pfn) << PAGE_SHIFT) | pgprot_val(prot))
+
+/*
+ * All present user pages are user-executable:
+ */
+static inline int pte_exec(pte_t pte)
+{
+	return pte_user(pte);
+}
+
+/*
+ * All present pages are kernel-executable:
+ */
+static inline int pte_exec_kernel(pte_t pte)
+{
+	return 1;
+}
+
+/*
+ * Bits 0, 6 and 7 are taken, split up the 29 bits of offset
+ * into this range:
+ */
+#define PTE_FILE_MAX_BITS	29
+
+#define pte_to_pgoff(pte) \
+	((((pte).pte_low >> 1) & 0x1f ) + (((pte).pte_low >> 8) << 5 ))
+
+#define pgoff_to_pte(off) \
+	((pte_t) { (((off) & 0x1f) << 1) + (((off) >> 5) << 8) + _PAGE_FILE })
+
+/* Encode and de-code a swap entry */
+#define __swp_type(x)			(((x).val >> 1) & 0x1f)
+#define __swp_offset(x)			((x).val >> 8)
+#define __swp_entry(type, offset)	((swp_entry_t) { ((type) << 1) | ((offset) << 8) })
+#define __pte_to_swp_entry(pte)		((swp_entry_t) { (pte).pte_low })
+#define __swp_entry_to_pte(x)		((pte_t) { (x).val })
+
+void vmalloc_sync_all(void);
+
+#endif /* _I386_PGTABLE_2LEVEL_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/pgtable-3level-defs.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/pgtable-3level-defs.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/pgtable-3level-defs.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/pgtable-3level-defs.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,24 @@
+#ifndef _I386_PGTABLE_3LEVEL_DEFS_H
+#define _I386_PGTABLE_3LEVEL_DEFS_H
+
+#define HAVE_SHARED_KERNEL_PMD 0
+
+/*
+ * PGDIR_SHIFT determines what a top-level page table entry can map
+ */
+#define PGDIR_SHIFT	30
+#define PTRS_PER_PGD	4
+
+/*
+ * PMD_SHIFT determines the size of the area a middle-level
+ * page table can map
+ */
+#define PMD_SHIFT	21
+#define PTRS_PER_PMD	512
+
+/*
+ * entries per page directory level
+ */
+#define PTRS_PER_PTE	512
+
+#endif /* _I386_PGTABLE_3LEVEL_DEFS_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/pgtable-3level.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/pgtable-3level.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/pgtable-3level.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/pgtable-3level.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,211 @@
+#ifndef _I386_PGTABLE_3LEVEL_H
+#define _I386_PGTABLE_3LEVEL_H
+
+#include <asm-generic/pgtable-nopud.h>
+
+/*
+ * Intel Physical Address Extension (PAE) Mode - three-level page
+ * tables on PPro+ CPUs.
+ *
+ * Copyright (C) 1999 Ingo Molnar <mingo@redhat.com>
+ */
+
+#define pte_ERROR(e) \
+	printk("%s:%d: bad pte %p(%016Lx pfn %08lx).\n", __FILE__, __LINE__, \
+	       &(e), __pte_val(e), pte_pfn(e))
+#define pmd_ERROR(e) \
+	printk("%s:%d: bad pmd %p(%016Lx pfn %08Lx).\n", __FILE__, __LINE__, \
+	       &(e), __pmd_val(e), (pmd_val(e) & PTE_MASK) >> PAGE_SHIFT)
+#define pgd_ERROR(e) \
+	printk("%s:%d: bad pgd %p(%016Lx pfn %08Lx).\n", __FILE__, __LINE__, \
+	       &(e), __pgd_val(e), (pgd_val(e) & PTE_MASK) >> PAGE_SHIFT)
+
+#define pud_none(pud)				0
+#define pud_bad(pud)				0
+#define pud_present(pud)			1
+
+/*
+ * Is the pte executable?
+ */
+static inline int pte_x(pte_t pte)
+{
+	return !(__pte_val(pte) & _PAGE_NX);
+}
+
+/*
+ * All present user-pages with !NX bit are user-executable:
+ */
+static inline int pte_exec(pte_t pte)
+{
+	return pte_user(pte) && pte_x(pte);
+}
+/*
+ * All present pages with !NX bit are kernel-executable:
+ */
+static inline int pte_exec_kernel(pte_t pte)
+{
+	return pte_x(pte);
+}
+
+/* Rules for using set_pte: the pte being assigned *must* be
+ * either not present or in a state where the hardware will
+ * not attempt to update the pte.  In places where this is
+ * not possible, use pte_get_and_clear to obtain the old pte
+ * value and then use set_pte to update it.  -ben
+ */
+#define __HAVE_ARCH_SET_PTE_ATOMIC
+
+static inline void set_pte(pte_t *ptep, pte_t pte)
+{
+	ptep->pte_high = pte.pte_high;
+	smp_wmb();
+	ptep->pte_low = pte.pte_low;
+}
+#define set_pte_atomic(pteptr,pteval) \
+		set_64bit((unsigned long long *)(pteptr),__pte_val(pteval))
+
+#define set_pte_at(_mm,addr,ptep,pteval) do {				\
+	if (((_mm) != current->mm && (_mm) != &init_mm) ||		\
+	    HYPERVISOR_update_va_mapping((addr), (pteval), 0))		\
+		set_pte((ptep), (pteval));				\
+} while (0)
+
+#define set_pte_at_sync(_mm,addr,ptep,pteval) do {			\
+	if (((_mm) != current->mm && (_mm) != &init_mm) ||		\
+	    HYPERVISOR_update_va_mapping((addr), (pteval), UVMF_INVLPG)) { \
+		set_pte((ptep), (pteval));				\
+		xen_invlpg((addr));					\
+	}								\
+} while (0)
+
+#define set_pmd(pmdptr,pmdval)				\
+		xen_l2_entry_update((pmdptr), (pmdval))
+#define set_pud(pudptr,pudval) \
+		xen_l3_entry_update((pudptr), (pudval))
+
+/*
+ * Pentium-II erratum A13: in PAE mode we explicitly have to flush
+ * the TLB via cr3 if the top-level pgd is changed...
+ * We do not let the generic code free and clear pgd entries due to
+ * this erratum.
+ */
+static inline void pud_clear (pud_t * pud) { }
+
+#define pud_page(pud) \
+((struct page *) __va(pud_val(pud) & PAGE_MASK))
+
+#define pud_page_kernel(pud) \
+((unsigned long) __va(pud_val(pud) & PAGE_MASK))
+
+
+/* Find an entry in the second-level page table.. */
+#define pmd_offset(pud, address) ((pmd_t *) pud_page(*(pud)) + \
+			pmd_index(address))
+
+static inline int pte_none(pte_t pte)
+{
+	return !(pte.pte_low | pte.pte_high);
+}
+
+/*
+ * For PTEs and PDEs, we must clear the P-bit first when clearing a page table
+ * entry, so clear the bottom half first and enforce ordering with a compiler
+ * barrier.
+ */
+static inline void pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
+{
+	if ((mm != current->mm && mm != &init_mm)
+	    || HYPERVISOR_update_va_mapping(addr, __pte(0), 0)) {
+		ptep->pte_low = 0;
+		smp_wmb();
+		ptep->pte_high = 0;
+	}
+}
+
+#define pmd_clear(xp)	do { set_pmd(xp, __pmd(0)); } while (0)
+
+static inline pte_t ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
+{
+	pte_t pte = *ptep;
+	if (!pte_none(pte)) {
+		if ((mm != &init_mm) ||
+		    HYPERVISOR_update_va_mapping(addr, __pte(0), 0)) {
+			uint64_t val = __pte_val(pte);
+			if (__cmpxchg64(ptep, val, 0) != val) {
+				/* xchg acts as a barrier before the setting of the high bits */
+				pte.pte_low = xchg(&ptep->pte_low, 0);
+				pte.pte_high = ptep->pte_high;
+				ptep->pte_high = 0;
+			}
+		}
+	}
+	return pte;
+}
+
+#define ptep_clear_flush(vma, addr, ptep)			\
+({								\
+	pte_t *__ptep = (ptep);					\
+	pte_t __res = *__ptep;					\
+	if (!pte_none(__res) &&					\
+	    ((vma)->vm_mm != current->mm ||			\
+	     HYPERVISOR_update_va_mapping(addr,	__pte(0),	\
+			(unsigned long)(vma)->vm_mm->cpu_vm_mask.bits| \
+				UVMF_INVLPG|UVMF_MULTI))) {	\
+		__ptep->pte_low = 0;				\
+		smp_wmb();					\
+		__ptep->pte_high = 0;				\
+		flush_tlb_page(vma, addr);			\
+	}							\
+	__res;							\
+})
+
+static inline int pte_same(pte_t a, pte_t b)
+{
+	return a.pte_low == b.pte_low && a.pte_high == b.pte_high;
+}
+
+#define pte_page(x)	pfn_to_page(pte_pfn(x))
+
+#define __pte_mfn(_pte) (((_pte).pte_low >> PAGE_SHIFT) | \
+			 ((_pte).pte_high << (32-PAGE_SHIFT)))
+#define pte_mfn(_pte) ((_pte).pte_low & _PAGE_PRESENT ? \
+	__pte_mfn(_pte) : pfn_to_mfn(__pte_mfn(_pte)))
+#define pte_pfn(_pte) ((_pte).pte_low & _PAGE_IO ? max_mapnr :	\
+		       (_pte).pte_low & _PAGE_PRESENT ?		\
+		       mfn_to_local_pfn(__pte_mfn(_pte)) :	\
+		       __pte_mfn(_pte))
+
+extern unsigned long long __supported_pte_mask;
+
+static inline pte_t pfn_pte(unsigned long page_nr, pgprot_t pgprot)
+{
+	return __pte((((unsigned long long)page_nr << PAGE_SHIFT) |
+			pgprot_val(pgprot)) & __supported_pte_mask);
+}
+
+static inline pmd_t pfn_pmd(unsigned long page_nr, pgprot_t pgprot)
+{
+	return __pmd((((unsigned long long)page_nr << PAGE_SHIFT) |
+			pgprot_val(pgprot)) & __supported_pte_mask);
+}
+
+/*
+ * Bits 0, 6 and 7 are taken in the low part of the pte,
+ * put the 32 bits of offset into the high part.
+ */
+#define pte_to_pgoff(pte) ((pte).pte_high)
+#define pgoff_to_pte(off) ((pte_t) { _PAGE_FILE, (off) })
+#define PTE_FILE_MAX_BITS       32
+
+/* Encode and de-code a swap entry */
+#define __swp_type(x)			(((x).val) & 0x1f)
+#define __swp_offset(x)			((x).val >> 5)
+#define __swp_entry(type, offset)	((swp_entry_t){(type) | (offset) << 5})
+#define __pte_to_swp_entry(pte)		((swp_entry_t){ (pte).pte_high })
+#define __swp_entry_to_pte(x)		((pte_t){ 0, (x).val })
+
+#define __pmd_free_tlb(tlb, x)		do { } while (0)
+
+void vmalloc_sync_all(void);
+
+#endif /* _I386_PGTABLE_3LEVEL_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/pgtable.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/pgtable.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/pgtable.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/pgtable.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,537 @@
+#ifndef _I386_PGTABLE_H
+#define _I386_PGTABLE_H
+
+#include <asm/hypervisor.h>
+
+/*
+ * The Linux memory management assumes a three-level page table setup. On
+ * the i386, we use that, but "fold" the mid level into the top-level page
+ * table, so that we physically have the same two-level page table as the
+ * i386 mmu expects.
+ *
+ * This file contains the functions and defines necessary to modify and use
+ * the i386 page table tree.
+ */
+#ifndef __ASSEMBLY__
+#include <asm/processor.h>
+#include <asm/fixmap.h>
+#include <linux/threads.h>
+
+#ifndef _I386_BITOPS_H
+#include <asm/bitops.h>
+#endif
+
+#include <linux/slab.h>
+#include <linux/list.h>
+#include <linux/spinlock.h>
+
+/* Is this pagetable pinned? */
+#define PG_pinned	PG_arch_1
+
+struct mm_struct;
+struct vm_area_struct;
+
+/*
+ * ZERO_PAGE is a global shared page that is always zero: used
+ * for zero-mapped memory areas etc..
+ */
+#define ZERO_PAGE(vaddr) (virt_to_page(empty_zero_page))
+extern unsigned long empty_zero_page[1024];
+extern pgd_t *swapper_pg_dir;
+extern kmem_cache_t *pgd_cache;
+extern kmem_cache_t *pmd_cache;
+extern spinlock_t pgd_lock;
+extern struct page *pgd_list;
+
+void pmd_ctor(void *, kmem_cache_t *, unsigned long);
+void pgd_ctor(void *, kmem_cache_t *, unsigned long);
+void pgd_dtor(void *, kmem_cache_t *, unsigned long);
+void pgtable_cache_init(void);
+void paging_init(void);
+
+/*
+ * The Linux x86 paging architecture is 'compile-time dual-mode', it
+ * implements both the traditional 2-level x86 page tables and the
+ * newer 3-level PAE-mode page tables.
+ */
+#ifdef CONFIG_X86_PAE
+# include <asm/pgtable-3level-defs.h>
+# define PMD_SIZE	(1UL << PMD_SHIFT)
+# define PMD_MASK	(~(PMD_SIZE-1))
+#else
+# include <asm/pgtable-2level-defs.h>
+#endif
+
+#define PGDIR_SIZE	(1UL << PGDIR_SHIFT)
+#define PGDIR_MASK	(~(PGDIR_SIZE-1))
+
+#define USER_PTRS_PER_PGD	(TASK_SIZE/PGDIR_SIZE)
+#define FIRST_USER_ADDRESS	0
+
+#define USER_PGD_PTRS (PAGE_OFFSET >> PGDIR_SHIFT)
+#define KERNEL_PGD_PTRS (PTRS_PER_PGD-USER_PGD_PTRS)
+
+#define TWOLEVEL_PGDIR_SHIFT	22
+#define BOOT_USER_PGD_PTRS (__PAGE_OFFSET >> TWOLEVEL_PGDIR_SHIFT)
+#define BOOT_KERNEL_PGD_PTRS (1024-BOOT_USER_PGD_PTRS)
+
+/* Just any arbitrary offset to the start of the vmalloc VM area: the
+ * current 8MB value just means that there will be a 8MB "hole" after the
+ * physical memory until the kernel virtual memory starts.  That means that
+ * any out-of-bounds memory accesses will hopefully be caught.
+ * The vmalloc() routines leaves a hole of 4kB between each vmalloced
+ * area for the same reason. ;)
+ */
+#define VMALLOC_OFFSET	(8*1024*1024)
+#define VMALLOC_START	(((unsigned long) high_memory + vmalloc_earlyreserve + \
+			2*VMALLOC_OFFSET-1) & ~(VMALLOC_OFFSET-1))
+#ifdef CONFIG_HIGHMEM
+# define VMALLOC_END	(PKMAP_BASE-2*PAGE_SIZE)
+#else
+# define VMALLOC_END	(FIXADDR_START-2*PAGE_SIZE)
+#endif
+
+/*
+ * _PAGE_PSE set in the page directory entry just means that
+ * the page directory entry points directly to a 4MB-aligned block of
+ * memory. 
+ */
+#define _PAGE_BIT_PRESENT	0
+#define _PAGE_BIT_RW		1
+#define _PAGE_BIT_USER		2
+#define _PAGE_BIT_PWT		3
+#define _PAGE_BIT_PCD		4
+#define _PAGE_BIT_ACCESSED	5
+#define _PAGE_BIT_DIRTY		6
+#define _PAGE_BIT_PSE		7	/* 4 MB (or 2MB) page, Pentium+, if present.. */
+#define _PAGE_BIT_GLOBAL	8	/* Global TLB entry PPro+ */
+/*#define _PAGE_BIT_UNUSED1	9*/	/* available for programmer */
+#define _PAGE_BIT_UNUSED2	10
+#define _PAGE_BIT_UNUSED3	11
+#define _PAGE_BIT_NX		63
+
+#define _PAGE_PRESENT	0x001
+#define _PAGE_RW	0x002
+#define _PAGE_USER	0x004
+#define _PAGE_PWT	0x008
+#define _PAGE_PCD	0x010
+#define _PAGE_ACCESSED	0x020
+#define _PAGE_DIRTY	0x040
+#define _PAGE_PSE	0x080	/* 4 MB (or 2MB) page, Pentium+, if present.. */
+#define _PAGE_GLOBAL	0x100	/* Global TLB entry PPro+ */
+/*#define _PAGE_UNUSED1	0x200*/	/* available for programmer */
+#define _PAGE_UNUSED2	0x400
+#define _PAGE_UNUSED3	0x800
+
+/* If _PAGE_PRESENT is clear, we use these: */
+#define _PAGE_FILE	0x040	/* nonlinear file mapping, saved PTE; unset:swap */
+#define _PAGE_PROTNONE	0x080	/* if the user mapped it with PROT_NONE;
+				   pte_present gives true */
+#ifdef CONFIG_X86_PAE
+#define _PAGE_NX	(1ULL<<_PAGE_BIT_NX)
+#else
+#define _PAGE_NX	0
+#endif
+
+/* Mapped page is I/O or foreign and has no associated page struct. */
+#define _PAGE_IO	0x200
+
+#define _PAGE_TABLE	(_PAGE_PRESENT | _PAGE_RW | _PAGE_USER | _PAGE_ACCESSED | _PAGE_DIRTY)
+#define _KERNPG_TABLE	(_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY)
+#define _PAGE_CHG_MASK	(PTE_MASK | _PAGE_ACCESSED | _PAGE_DIRTY | _PAGE_IO)
+
+#define PAGE_NONE \
+	__pgprot(_PAGE_PROTNONE | _PAGE_ACCESSED)
+#define PAGE_SHARED \
+	__pgprot(_PAGE_PRESENT | _PAGE_RW | _PAGE_USER | _PAGE_ACCESSED)
+
+#define PAGE_SHARED_EXEC \
+	__pgprot(_PAGE_PRESENT | _PAGE_RW | _PAGE_USER | _PAGE_ACCESSED)
+#define PAGE_COPY_NOEXEC \
+	__pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED | _PAGE_NX)
+#define PAGE_COPY_EXEC \
+	__pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED)
+#define PAGE_COPY \
+	PAGE_COPY_NOEXEC
+#define PAGE_READONLY \
+	__pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED | _PAGE_NX)
+#define PAGE_READONLY_EXEC \
+	__pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED)
+
+#define _PAGE_KERNEL \
+	(_PAGE_PRESENT | _PAGE_RW | _PAGE_DIRTY | _PAGE_ACCESSED | _PAGE_NX)
+#define _PAGE_KERNEL_EXEC \
+	(_PAGE_PRESENT | _PAGE_RW | _PAGE_DIRTY | _PAGE_ACCESSED)
+
+extern unsigned long long __PAGE_KERNEL, __PAGE_KERNEL_EXEC;
+#define __PAGE_KERNEL_RO		(__PAGE_KERNEL & ~_PAGE_RW)
+#define __PAGE_KERNEL_NOCACHE		(__PAGE_KERNEL | _PAGE_PCD)
+#define __PAGE_KERNEL_LARGE		(__PAGE_KERNEL | _PAGE_PSE)
+#define __PAGE_KERNEL_LARGE_EXEC	(__PAGE_KERNEL_EXEC | _PAGE_PSE)
+
+#define PAGE_KERNEL		__pgprot(__PAGE_KERNEL)
+#define PAGE_KERNEL_RO		__pgprot(__PAGE_KERNEL_RO)
+#define PAGE_KERNEL_EXEC	__pgprot(__PAGE_KERNEL_EXEC)
+#define PAGE_KERNEL_NOCACHE	__pgprot(__PAGE_KERNEL_NOCACHE)
+#define PAGE_KERNEL_LARGE	__pgprot(__PAGE_KERNEL_LARGE)
+#define PAGE_KERNEL_LARGE_EXEC	__pgprot(__PAGE_KERNEL_LARGE_EXEC)
+
+/*
+ * The i386 can't do page protection for execute, and considers that
+ * the same are read. Also, write permissions imply read permissions.
+ * This is the closest we can get..
+ */
+#define __P000	PAGE_NONE
+#define __P001	PAGE_READONLY
+#define __P010	PAGE_COPY
+#define __P011	PAGE_COPY
+#define __P100	PAGE_READONLY_EXEC
+#define __P101	PAGE_READONLY_EXEC
+#define __P110	PAGE_COPY_EXEC
+#define __P111	PAGE_COPY_EXEC
+
+#define __S000	PAGE_NONE
+#define __S001	PAGE_READONLY
+#define __S010	PAGE_SHARED
+#define __S011	PAGE_SHARED
+#define __S100	PAGE_READONLY_EXEC
+#define __S101	PAGE_READONLY_EXEC
+#define __S110	PAGE_SHARED_EXEC
+#define __S111	PAGE_SHARED_EXEC
+
+/*
+ * Define this if things work differently on an i386 and an i486:
+ * it will (on an i486) warn about kernel memory accesses that are
+ * done without a 'access_ok(VERIFY_WRITE,..)'
+ */
+#undef TEST_ACCESS_OK
+
+/* The boot page tables (all created as a single array) */
+extern unsigned long pg0[];
+
+#define pte_present(x)	((x).pte_low & (_PAGE_PRESENT | _PAGE_PROTNONE))
+
+/* To avoid harmful races, pmd_none(x) should check only the lower when PAE */
+#define pmd_none(x)	(!(unsigned long)__pmd_val(x))
+#if CONFIG_XEN_COMPAT <= 0x030002
+/* pmd_present doesn't just test the _PAGE_PRESENT bit since wr.p.t.
+   can temporarily clear it. */
+#define pmd_present(x)	(__pmd_val(x))
+#define pmd_bad(x)	((__pmd_val(x) & (~PAGE_MASK & ~_PAGE_USER & ~_PAGE_PRESENT)) != (_KERNPG_TABLE & ~_PAGE_PRESENT))
+#else
+#define pmd_present(x)	(__pmd_val(x) & _PAGE_PRESENT)
+#define pmd_bad(x)	((__pmd_val(x) & (~PAGE_MASK & ~_PAGE_USER)) != _KERNPG_TABLE)
+#endif
+
+
+#define pages_to_mb(x) ((x) >> (20-PAGE_SHIFT))
+
+/*
+ * The following only work if pte_present() is true.
+ * Undefined behaviour if not..
+ */
+static inline int pte_user(pte_t pte)		{ return (pte).pte_low & _PAGE_USER; }
+static inline int pte_read(pte_t pte)		{ return (pte).pte_low & _PAGE_USER; }
+static inline int pte_dirty(pte_t pte)		{ return (pte).pte_low & _PAGE_DIRTY; }
+static inline int pte_young(pte_t pte)		{ return (pte).pte_low & _PAGE_ACCESSED; }
+static inline int pte_write(pte_t pte)		{ return (pte).pte_low & _PAGE_RW; }
+static inline int pte_huge(pte_t pte)		{ return (pte).pte_low & _PAGE_PSE; }
+
+/*
+ * The following only works if pte_present() is not true.
+ */
+static inline int pte_file(pte_t pte)		{ return (pte).pte_low & _PAGE_FILE; }
+
+static inline pte_t pte_rdprotect(pte_t pte)	{ (pte).pte_low &= ~_PAGE_USER; return pte; }
+static inline pte_t pte_exprotect(pte_t pte)	{ (pte).pte_low &= ~_PAGE_USER; return pte; }
+static inline pte_t pte_mkclean(pte_t pte)	{ (pte).pte_low &= ~_PAGE_DIRTY; return pte; }
+static inline pte_t pte_mkold(pte_t pte)	{ (pte).pte_low &= ~_PAGE_ACCESSED; return pte; }
+static inline pte_t pte_wrprotect(pte_t pte)	{ (pte).pte_low &= ~_PAGE_RW; return pte; }
+static inline pte_t pte_mkread(pte_t pte)	{ (pte).pte_low |= _PAGE_USER; return pte; }
+static inline pte_t pte_mkexec(pte_t pte)	{ (pte).pte_low |= _PAGE_USER; return pte; }
+static inline pte_t pte_mkdirty(pte_t pte)	{ (pte).pte_low |= _PAGE_DIRTY; return pte; }
+static inline pte_t pte_mkyoung(pte_t pte)	{ (pte).pte_low |= _PAGE_ACCESSED; return pte; }
+static inline pte_t pte_mkwrite(pte_t pte)	{ (pte).pte_low |= _PAGE_RW; return pte; }
+static inline pte_t pte_mkhuge(pte_t pte)	{ (pte).pte_low |= _PAGE_PSE; return pte; }
+
+#ifdef CONFIG_X86_PAE
+# include <asm/pgtable-3level.h>
+#else
+# include <asm/pgtable-2level.h>
+#endif
+
+#define ptep_test_and_clear_dirty(vma, addr, ptep)			\
+({									\
+	pte_t __pte = *(ptep);						\
+	int __ret = pte_dirty(__pte);					\
+	if (__ret) {							\
+		__pte = pte_mkclean(__pte);				\
+		if ((vma)->vm_mm != current->mm ||			\
+		    HYPERVISOR_update_va_mapping(addr, __pte, 0))	\
+			(ptep)->pte_low = __pte.pte_low;		\
+	}								\
+	__ret;								\
+})
+
+#define ptep_test_and_clear_young(vma, addr, ptep)			\
+({									\
+	pte_t __pte = *(ptep);						\
+	int __ret = pte_young(__pte);					\
+	if (__ret)							\
+		__pte = pte_mkold(__pte);				\
+		if ((vma)->vm_mm != current->mm ||			\
+		    HYPERVISOR_update_va_mapping(addr, __pte, 0))	\
+			(ptep)->pte_low = __pte.pte_low;		\
+	__ret;								\
+})
+
+#define ptep_get_and_clear_full(mm, addr, ptep, full)			\
+	((full) ? ({							\
+		pte_t __res = *(ptep);					\
+		if (test_bit(PG_pinned, &virt_to_page((mm)->pgd)->flags)) \
+			xen_l1_entry_update(ptep, __pte(0));		\
+		else							\
+			*(ptep) = __pte(0);				\
+		__res;							\
+	 }) :								\
+	 ptep_get_and_clear(mm, addr, ptep))
+
+static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
+{
+	pte_t pte = *ptep;
+	if (pte_write(pte))
+		set_pte_at(mm, addr, ptep, pte_wrprotect(pte));
+}
+
+/*
+ * clone_pgd_range(pgd_t *dst, pgd_t *src, int count);
+ *
+ *  dst - pointer to pgd range anwhere on a pgd page
+ *  src - ""
+ *  count - the number of pgds to copy.
+ *
+ * dst and src can be on the same page, but the range must not overlap,
+ * and must not cross a page boundary.
+ */
+static inline void clone_pgd_range(pgd_t *dst, pgd_t *src, int count)
+{
+       memcpy(dst, src, count * sizeof(pgd_t));
+}
+
+/*
+ * Macro to mark a page protection value as "uncacheable".  On processors which do not support
+ * it, this is a no-op.
+ */
+#define pgprot_noncached(prot)	((boot_cpu_data.x86 > 3)					  \
+				 ? (__pgprot(pgprot_val(prot) | _PAGE_PCD | _PAGE_PWT)) : (prot))
+
+/*
+ * Conversion functions: convert a page and protection to a page entry,
+ * and a page entry and page directory to the page they refer to.
+ */
+
+#define mk_pte(page, pgprot)	pfn_pte(page_to_pfn(page), (pgprot))
+
+static inline pte_t pte_modify(pte_t pte, pgprot_t newprot)
+{
+	/*
+	 * Since this might change the present bit (which controls whether
+	 * a pte_t object has undergone p2m translation), we must use
+	 * pte_val() on the input pte and __pte() for the return value.
+	 */
+	paddr_t pteval = pte_val(pte);
+
+	pteval &= _PAGE_CHG_MASK;
+	pteval |= pgprot_val(newprot);
+#ifdef CONFIG_X86_PAE
+	pteval &= __supported_pte_mask;
+#endif
+	return __pte(pteval);
+}
+
+#define pmd_large(pmd) \
+((__pmd_val(pmd) & (_PAGE_PSE|_PAGE_PRESENT)) == (_PAGE_PSE|_PAGE_PRESENT))
+
+/*
+ * the pgd page can be thought of an array like this: pgd_t[PTRS_PER_PGD]
+ *
+ * this macro returns the index of the entry in the pgd page which would
+ * control the given virtual address
+ */
+#define pgd_index(address) (((address) >> PGDIR_SHIFT) & (PTRS_PER_PGD-1))
+#define pgd_index_k(addr) pgd_index(addr)
+
+/*
+ * pgd_offset() returns a (pgd_t *)
+ * pgd_index() is used get the offset into the pgd page's array of pgd_t's;
+ */
+#define pgd_offset(mm, address) ((mm)->pgd+pgd_index(address))
+
+/*
+ * a shortcut which implies the use of the kernel's pgd, instead
+ * of a process's
+ */
+#define pgd_offset_k(address) pgd_offset(&init_mm, address)
+
+/*
+ * the pmd page can be thought of an array like this: pmd_t[PTRS_PER_PMD]
+ *
+ * this macro returns the index of the entry in the pmd page which would
+ * control the given virtual address
+ */
+#define pmd_index(address) \
+		(((address) >> PMD_SHIFT) & (PTRS_PER_PMD-1))
+
+/*
+ * the pte page can be thought of an array like this: pte_t[PTRS_PER_PTE]
+ *
+ * this macro returns the index of the entry in the pte page which would
+ * control the given virtual address
+ */
+#define pte_index(address) \
+		(((address) >> PAGE_SHIFT) & (PTRS_PER_PTE - 1))
+#define pte_offset_kernel(dir, address) \
+	((pte_t *) pmd_page_kernel(*(dir)) +  pte_index(address))
+
+#define pmd_page(pmd) (pfn_to_page(pmd_val(pmd) >> PAGE_SHIFT))
+
+#define pmd_page_kernel(pmd) \
+		((unsigned long) __va(pmd_val(pmd) & PAGE_MASK))
+
+/*
+ * Helper function that returns the kernel pagetable entry controlling
+ * the virtual address 'address'. NULL means no pagetable entry present.
+ * NOTE: the return type is pte_t but if the pmd is PSE then we return it
+ * as a pte too.
+ */
+extern pte_t *lookup_address(unsigned long address);
+
+/*
+ * Make a given kernel text page executable/non-executable.
+ * Returns the previous executability setting of that page (which
+ * is used to restore the previous state). Used by the SMP bootup code.
+ * NOTE: this is an __init function for security reasons.
+ */
+#ifdef CONFIG_X86_PAE
+ extern int set_kernel_exec(unsigned long vaddr, int enable);
+#else
+ static inline int set_kernel_exec(unsigned long vaddr, int enable) { return 0;}
+#endif
+
+extern void noexec_setup(const char *str);
+
+#if defined(CONFIG_HIGHPTE)
+#define pte_offset_map(dir, address) \
+	((pte_t *)kmap_atomic_pte(pmd_page(*(dir)),KM_PTE0) + \
+	 pte_index(address))
+#define pte_offset_map_nested(dir, address) \
+	((pte_t *)kmap_atomic_pte(pmd_page(*(dir)),KM_PTE1) + \
+	 pte_index(address))
+#define pte_unmap(pte) kunmap_atomic(pte, KM_PTE0)
+#define pte_unmap_nested(pte) kunmap_atomic(pte, KM_PTE1)
+#else
+#define pte_offset_map(dir, address) \
+	((pte_t *)page_address(pmd_page(*(dir))) + pte_index(address))
+#define pte_offset_map_nested(dir, address) pte_offset_map(dir, address)
+#define pte_unmap(pte) do { } while (0)
+#define pte_unmap_nested(pte) do { } while (0)
+#endif
+
+#define __HAVE_ARCH_PTEP_ESTABLISH
+#define ptep_establish(vma, address, ptep, pteval)			\
+	do {								\
+		if ( likely((vma)->vm_mm == current->mm) ) {		\
+			BUG_ON(HYPERVISOR_update_va_mapping(address,	\
+				pteval,					\
+				(unsigned long)(vma)->vm_mm->cpu_vm_mask.bits| \
+					UVMF_INVLPG|UVMF_MULTI));	\
+		} else {						\
+			xen_l1_entry_update(ptep, pteval);		\
+			flush_tlb_page(vma, address);			\
+		}							\
+	} while (0)
+
+/*
+ * The i386 doesn't have any external MMU info: the kernel page
+ * tables contain all the necessary information.
+ *
+ * Also, we only update the dirty/accessed state if we set
+ * the dirty bit by hand in the kernel, since the hardware
+ * will do the accessed bit for us, and we don't want to
+ * race with other CPU's that might be updating the dirty
+ * bit at the same time.
+ */
+#define update_mmu_cache(vma,address,pte) do { } while (0)
+#define  __HAVE_ARCH_PTEP_SET_ACCESS_FLAGS
+#define ptep_set_access_flags(vma, address, ptep, entry, dirty)		\
+	do {								\
+		if (dirty)						\
+			ptep_establish(vma, address, ptep, entry);	\
+	} while (0)
+
+#include <xen/features.h>
+void make_lowmem_page_readonly(void *va, unsigned int feature);
+void make_lowmem_page_writable(void *va, unsigned int feature);
+void make_page_readonly(void *va, unsigned int feature);
+void make_page_writable(void *va, unsigned int feature);
+void make_pages_readonly(void *va, unsigned int nr, unsigned int feature);
+void make_pages_writable(void *va, unsigned int nr, unsigned int feature);
+
+#define virt_to_ptep(va)						\
+({									\
+	pte_t *__ptep = lookup_address((unsigned long)(va));		\
+	BUG_ON(!__ptep || !pte_present(*__ptep));			\
+	__ptep;								\
+})
+
+#define arbitrary_virt_to_machine(va)					\
+	(((maddr_t)pte_mfn(*virt_to_ptep(va)) << PAGE_SHIFT)		\
+	 | ((unsigned long)(va) & (PAGE_SIZE - 1)))
+
+#endif /* !__ASSEMBLY__ */
+
+#ifdef CONFIG_FLATMEM
+#define kern_addr_valid(addr)	(1)
+#endif /* CONFIG_FLATMEM */
+
+int direct_remap_pfn_range(struct vm_area_struct *vma,
+                           unsigned long address, 
+                           unsigned long mfn,
+                           unsigned long size, 
+                           pgprot_t prot,
+                           domid_t  domid);
+int direct_kernel_remap_pfn_range(unsigned long address, 
+				  unsigned long mfn,
+				  unsigned long size, 
+				  pgprot_t prot,
+				  domid_t  domid);
+int create_lookup_pte_addr(struct mm_struct *mm,
+                           unsigned long address,
+                           uint64_t *ptep);
+int touch_pte_range(struct mm_struct *mm,
+                    unsigned long address,
+                    unsigned long size);
+
+int xen_change_pte_range(struct mm_struct *mm, pmd_t *pmd,
+		unsigned long addr, unsigned long end, pgprot_t newprot);
+
+#define arch_change_pte_range(mm, pmd, addr, end, newprot)	\
+		xen_change_pte_range(mm, pmd, addr, end, newprot)
+
+#define io_remap_pfn_range(vma,from,pfn,size,prot) \
+direct_remap_pfn_range(vma,from,pfn,size,prot,DOMID_IO)
+
+#define MK_IOSPACE_PFN(space, pfn)	(pfn)
+#define GET_IOSPACE(pfn)		0
+#define GET_PFN(pfn)			(pfn)
+
+#define __HAVE_ARCH_PTEP_TEST_AND_CLEAR_YOUNG
+#define __HAVE_ARCH_PTEP_TEST_AND_CLEAR_DIRTY
+#define __HAVE_ARCH_PTEP_GET_AND_CLEAR
+#define __HAVE_ARCH_PTEP_GET_AND_CLEAR_FULL
+#define __HAVE_ARCH_PTEP_CLEAR_FLUSH
+#define __HAVE_ARCH_PTEP_SET_WRPROTECT
+#define __HAVE_ARCH_PTE_SAME
+#include <asm-generic/pgtable.h>
+
+#endif /* _I386_PGTABLE_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/processor.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/processor.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/processor.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/processor.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,743 @@
+/*
+ * include/asm-i386/processor.h
+ *
+ * Copyright (C) 1994 Linus Torvalds
+ */
+
+#ifndef __ASM_I386_PROCESSOR_H
+#define __ASM_I386_PROCESSOR_H
+
+#include <asm/vm86.h>
+#include <asm/math_emu.h>
+#include <asm/segment.h>
+#include <asm/page.h>
+#include <asm/types.h>
+#include <asm/sigcontext.h>
+#include <asm/cpufeature.h>
+#include <asm/msr.h>
+#include <asm/system.h>
+#include <linux/cache.h>
+#include <linux/threads.h>
+#include <asm/percpu.h>
+#include <linux/cpumask.h>
+#include <xen/interface/physdev.h>
+
+/* flag for disabling the tsc */
+extern int tsc_disable;
+
+struct desc_struct {
+	unsigned long a,b;
+};
+
+#define desc_empty(desc) \
+		(!((desc)->a | (desc)->b))
+
+#define desc_equal(desc1, desc2) \
+		(((desc1)->a == (desc2)->a) && ((desc1)->b == (desc2)->b))
+/*
+ * Default implementation of macro that returns current
+ * instruction pointer ("program counter").
+ */
+#define current_text_addr() ({ void *pc; __asm__("movl $1f,%0\n1:":"=g" (pc)); pc; })
+
+/*
+ *  CPU type and hardware bug flags. Kept separately for each CPU.
+ *  Members of this structure are referenced in head.S, so think twice
+ *  before touching them. [mj]
+ */
+
+struct cpuinfo_x86 {
+	__u8	x86;		/* CPU family */
+	__u8	x86_vendor;	/* CPU vendor */
+	__u8	x86_model;
+	__u8	x86_mask;
+	char	wp_works_ok;	/* It doesn't on 386's */
+	char	hlt_works_ok;	/* Problems on some 486Dx4's and old 386's */
+	char	hard_math;
+	char	rfu;
+       	int	cpuid_level;	/* Maximum supported CPUID level, -1=no CPUID */
+	unsigned long	x86_capability[NCAPINTS];
+	char	x86_vendor_id[16];
+	char	x86_model_id[64];
+	int 	x86_cache_size;  /* in KB - valid for CPUS which support this
+				    call  */
+	int 	x86_cache_alignment;	/* In bytes */
+	char	fdiv_bug;
+	char	f00f_bug;
+	char	coma_bug;
+	char	pad0;
+	int	x86_power;
+	unsigned long loops_per_jiffy;
+#ifdef CONFIG_SMP
+	cpumask_t llc_shared_map;	/* cpus sharing the last level cache */
+#endif
+	unsigned char x86_max_cores;	/* cpuid returned max cores value */
+	unsigned char apicid;
+#ifdef CONFIG_SMP
+	unsigned char booted_cores;	/* number of cores as seen by OS */
+	__u8 phys_proc_id; 		/* Physical processor id. */
+	__u8 cpu_core_id;  		/* Core id */
+#endif
+} __attribute__((__aligned__(SMP_CACHE_BYTES)));
+
+#define X86_VENDOR_INTEL 0
+#define X86_VENDOR_CYRIX 1
+#define X86_VENDOR_AMD 2
+#define X86_VENDOR_UMC 3
+#define X86_VENDOR_NEXGEN 4
+#define X86_VENDOR_CENTAUR 5
+#define X86_VENDOR_RISE 6
+#define X86_VENDOR_TRANSMETA 7
+#define X86_VENDOR_NSC 8
+#define X86_VENDOR_NUM 9
+#define X86_VENDOR_UNKNOWN 0xff
+
+/*
+ * capabilities of CPUs
+ */
+
+extern struct cpuinfo_x86 boot_cpu_data;
+extern struct cpuinfo_x86 new_cpu_data;
+#ifndef CONFIG_X86_NO_TSS
+extern struct tss_struct doublefault_tss;
+DECLARE_PER_CPU(struct tss_struct, init_tss);
+#endif
+
+#ifdef CONFIG_SMP
+extern struct cpuinfo_x86 cpu_data[];
+#define current_cpu_data cpu_data[smp_processor_id()]
+#else
+#define cpu_data (&boot_cpu_data)
+#define current_cpu_data boot_cpu_data
+#endif
+
+extern	int cpu_llc_id[NR_CPUS];
+extern char ignore_fpu_irq;
+
+extern void identify_cpu(struct cpuinfo_x86 *);
+extern void print_cpu_info(struct cpuinfo_x86 *);
+extern unsigned int init_intel_cacheinfo(struct cpuinfo_x86 *c);
+extern unsigned short num_cache_leaves;
+
+#ifdef CONFIG_X86_HT
+extern void detect_ht(struct cpuinfo_x86 *c);
+#else
+static inline void detect_ht(struct cpuinfo_x86 *c) {}
+#endif
+
+/*
+ * EFLAGS bits
+ */
+#define X86_EFLAGS_CF	0x00000001 /* Carry Flag */
+#define X86_EFLAGS_PF	0x00000004 /* Parity Flag */
+#define X86_EFLAGS_AF	0x00000010 /* Auxillary carry Flag */
+#define X86_EFLAGS_ZF	0x00000040 /* Zero Flag */
+#define X86_EFLAGS_SF	0x00000080 /* Sign Flag */
+#define X86_EFLAGS_TF	0x00000100 /* Trap Flag */
+#define X86_EFLAGS_IF	0x00000200 /* Interrupt Flag */
+#define X86_EFLAGS_DF	0x00000400 /* Direction Flag */
+#define X86_EFLAGS_OF	0x00000800 /* Overflow Flag */
+#define X86_EFLAGS_IOPL	0x00003000 /* IOPL mask */
+#define X86_EFLAGS_NT	0x00004000 /* Nested Task */
+#define X86_EFLAGS_RF	0x00010000 /* Resume Flag */
+#define X86_EFLAGS_VM	0x00020000 /* Virtual Mode */
+#define X86_EFLAGS_AC	0x00040000 /* Alignment Check */
+#define X86_EFLAGS_VIF	0x00080000 /* Virtual Interrupt Flag */
+#define X86_EFLAGS_VIP	0x00100000 /* Virtual Interrupt Pending */
+#define X86_EFLAGS_ID	0x00200000 /* CPUID detection flag */
+
+/*
+ * Generic CPUID function
+ * clear %ecx since some cpus (Cyrix MII) do not set or clear %ecx
+ * resulting in stale register contents being returned.
+ */
+static inline void cpuid(unsigned int op, unsigned int *eax, unsigned int *ebx, unsigned int *ecx, unsigned int *edx)
+{
+	__asm__(XEN_CPUID
+		: "=a" (*eax),
+		  "=b" (*ebx),
+		  "=c" (*ecx),
+		  "=d" (*edx)
+		: "0" (op), "c"(0));
+}
+
+/* Some CPUID calls want 'count' to be placed in ecx */
+static inline void cpuid_count(int op, int count, int *eax, int *ebx, int *ecx,
+	       	int *edx)
+{
+	__asm__(XEN_CPUID
+		: "=a" (*eax),
+		  "=b" (*ebx),
+		  "=c" (*ecx),
+		  "=d" (*edx)
+		: "0" (op), "c" (count));
+}
+
+/*
+ * CPUID functions returning a single datum
+ */
+static inline unsigned int cpuid_eax(unsigned int op)
+{
+	unsigned int eax;
+
+	__asm__(XEN_CPUID
+		: "=a" (eax)
+		: "0" (op)
+		: "bx", "cx", "dx");
+	return eax;
+}
+static inline unsigned int cpuid_ebx(unsigned int op)
+{
+	unsigned int eax, ebx;
+
+	__asm__(XEN_CPUID
+		: "=a" (eax), "=b" (ebx)
+		: "0" (op)
+		: "cx", "dx" );
+	return ebx;
+}
+static inline unsigned int cpuid_ecx(unsigned int op)
+{
+	unsigned int eax, ecx;
+
+	__asm__(XEN_CPUID
+		: "=a" (eax), "=c" (ecx)
+		: "0" (op)
+		: "bx", "dx" );
+	return ecx;
+}
+static inline unsigned int cpuid_edx(unsigned int op)
+{
+	unsigned int eax, edx;
+
+	__asm__(XEN_CPUID
+		: "=a" (eax), "=d" (edx)
+		: "0" (op)
+		: "bx", "cx");
+	return edx;
+}
+
+#define load_cr3(pgdir) write_cr3(__pa(pgdir))
+
+/*
+ * Intel CPU features in CR4
+ */
+#define X86_CR4_VME		0x0001	/* enable vm86 extensions */
+#define X86_CR4_PVI		0x0002	/* virtual interrupts flag enable */
+#define X86_CR4_TSD		0x0004	/* disable time stamp at ipl 3 */
+#define X86_CR4_DE		0x0008	/* enable debugging extensions */
+#define X86_CR4_PSE		0x0010	/* enable page size extensions */
+#define X86_CR4_PAE		0x0020	/* enable physical address extensions */
+#define X86_CR4_MCE		0x0040	/* Machine check enable */
+#define X86_CR4_PGE		0x0080	/* enable global pages */
+#define X86_CR4_PCE		0x0100	/* enable performance counters at ipl 3 */
+#define X86_CR4_OSFXSR		0x0200	/* enable fast FPU save and restore */
+#define X86_CR4_OSXMMEXCPT	0x0400	/* enable unmasked SSE exceptions */
+
+/*
+ * Save the cr4 feature set we're using (ie
+ * Pentium 4MB enable and PPro Global page
+ * enable), so that any CPU's that boot up
+ * after us can get the correct flags.
+ */
+extern unsigned long mmu_cr4_features;
+
+static inline void set_in_cr4 (unsigned long mask)
+{
+	unsigned cr4;
+	mmu_cr4_features |= mask;
+	cr4 = read_cr4();
+	cr4 |= mask;
+	write_cr4(cr4);
+}
+
+static inline void clear_in_cr4 (unsigned long mask)
+{
+	unsigned cr4;
+	mmu_cr4_features &= ~mask;
+	cr4 = read_cr4();
+	cr4 &= ~mask;
+	write_cr4(cr4);
+}
+
+/*
+ *      NSC/Cyrix CPU configuration register indexes
+ */
+
+#define CX86_PCR0 0x20
+#define CX86_GCR  0xb8
+#define CX86_CCR0 0xc0
+#define CX86_CCR1 0xc1
+#define CX86_CCR2 0xc2
+#define CX86_CCR3 0xc3
+#define CX86_CCR4 0xe8
+#define CX86_CCR5 0xe9
+#define CX86_CCR6 0xea
+#define CX86_CCR7 0xeb
+#define CX86_PCR1 0xf0
+#define CX86_DIR0 0xfe
+#define CX86_DIR1 0xff
+#define CX86_ARR_BASE 0xc4
+#define CX86_RCR_BASE 0xdc
+
+/*
+ *      NSC/Cyrix CPU indexed register access macros
+ */
+
+#define getCx86(reg) ({ outb((reg), 0x22); inb(0x23); })
+
+#define setCx86(reg, data) do { \
+	outb((reg), 0x22); \
+	outb((data), 0x23); \
+} while (0)
+
+/* Stop speculative execution */
+static inline void sync_core(void)
+{
+	int tmp;
+	asm volatile("cpuid" : "=a" (tmp) : "0" (1) : "ebx","ecx","edx","memory");
+}
+
+static inline void __monitor(const void *eax, unsigned long ecx,
+		unsigned long edx)
+{
+	/* "monitor %eax,%ecx,%edx;" */
+	asm volatile(
+		".byte 0x0f,0x01,0xc8;"
+		: :"a" (eax), "c" (ecx), "d"(edx));
+}
+
+static inline void __mwait(unsigned long eax, unsigned long ecx)
+{
+	/* "mwait %eax,%ecx;" */
+	asm volatile(
+		".byte 0x0f,0x01,0xc9;"
+		: :"a" (eax), "c" (ecx));
+}
+
+/* from system description table in BIOS.  Mostly for MCA use, but
+others may find it useful. */
+extern unsigned int machine_id;
+extern unsigned int machine_submodel_id;
+extern unsigned int BIOS_revision;
+extern unsigned int mca_pentium_flag;
+
+/* Boot loader type from the setup header */
+extern int bootloader_type;
+
+/*
+ * User space process size: 3GB (default).
+ */
+#define TASK_SIZE	(PAGE_OFFSET)
+
+/* This decides where the kernel will search for a free chunk of vm
+ * space during mmap's.
+ */
+#define TASK_UNMAPPED_BASE	(PAGE_ALIGN(TASK_SIZE / 3))
+
+#define HAVE_ARCH_PICK_MMAP_LAYOUT
+
+/*
+ * Size of io_bitmap.
+ */
+#define IO_BITMAP_BITS  65536
+#define IO_BITMAP_BYTES (IO_BITMAP_BITS/8)
+#define IO_BITMAP_LONGS (IO_BITMAP_BYTES/sizeof(long))
+#ifndef CONFIG_X86_NO_TSS
+#define IO_BITMAP_OFFSET offsetof(struct tss_struct,io_bitmap)
+#endif
+#define INVALID_IO_BITMAP_OFFSET 0x8000
+#define INVALID_IO_BITMAP_OFFSET_LAZY 0x9000
+
+struct i387_fsave_struct {
+	long	cwd;
+	long	swd;
+	long	twd;
+	long	fip;
+	long	fcs;
+	long	foo;
+	long	fos;
+	long	st_space[20];	/* 8*10 bytes for each FP-reg = 80 bytes */
+	long	status;		/* software status information */
+};
+
+struct i387_fxsave_struct {
+	unsigned short	cwd;
+	unsigned short	swd;
+	unsigned short	twd;
+	unsigned short	fop;
+	long	fip;
+	long	fcs;
+	long	foo;
+	long	fos;
+	long	mxcsr;
+	long	mxcsr_mask;
+	long	st_space[32];	/* 8*16 bytes for each FP-reg = 128 bytes */
+	long	xmm_space[32];	/* 8*16 bytes for each XMM-reg = 128 bytes */
+	long	padding[56];
+} __attribute__ ((aligned (16)));
+
+struct i387_soft_struct {
+	long	cwd;
+	long	swd;
+	long	twd;
+	long	fip;
+	long	fcs;
+	long	foo;
+	long	fos;
+	long	st_space[20];	/* 8*10 bytes for each FP-reg = 80 bytes */
+	unsigned char	ftop, changed, lookahead, no_update, rm, alimit;
+	struct info	*info;
+	unsigned long	entry_eip;
+};
+
+union i387_union {
+	struct i387_fsave_struct	fsave;
+	struct i387_fxsave_struct	fxsave;
+	struct i387_soft_struct soft;
+};
+
+typedef struct {
+	unsigned long seg;
+} mm_segment_t;
+
+struct thread_struct;
+
+#ifndef CONFIG_X86_NO_TSS
+struct tss_struct {
+	unsigned short	back_link,__blh;
+	unsigned long	esp0;
+	unsigned short	ss0,__ss0h;
+	unsigned long	esp1;
+	unsigned short	ss1,__ss1h;	/* ss1 is used to cache MSR_IA32_SYSENTER_CS */
+	unsigned long	esp2;
+	unsigned short	ss2,__ss2h;
+	unsigned long	__cr3;
+	unsigned long	eip;
+	unsigned long	eflags;
+	unsigned long	eax,ecx,edx,ebx;
+	unsigned long	esp;
+	unsigned long	ebp;
+	unsigned long	esi;
+	unsigned long	edi;
+	unsigned short	es, __esh;
+	unsigned short	cs, __csh;
+	unsigned short	ss, __ssh;
+	unsigned short	ds, __dsh;
+	unsigned short	fs, __fsh;
+	unsigned short	gs, __gsh;
+	unsigned short	ldt, __ldth;
+	unsigned short	trace, io_bitmap_base;
+	/*
+	 * The extra 1 is there because the CPU will access an
+	 * additional byte beyond the end of the IO permission
+	 * bitmap. The extra byte must be all 1 bits, and must
+	 * be within the limit.
+	 */
+	unsigned long	io_bitmap[IO_BITMAP_LONGS + 1];
+	/*
+	 * Cache the current maximum and the last task that used the bitmap:
+	 */
+	unsigned long io_bitmap_max;
+	struct thread_struct *io_bitmap_owner;
+	/*
+	 * pads the TSS to be cacheline-aligned (size is 0x100)
+	 */
+	unsigned long __cacheline_filler[35];
+	/*
+	 * .. and then another 0x100 bytes for emergency kernel stack
+	 */
+	unsigned long stack[64];
+} __attribute__((packed));
+#endif
+
+#define ARCH_MIN_TASKALIGN	16
+
+struct thread_struct {
+/* cached TLS descriptors. */
+	struct desc_struct tls_array[GDT_ENTRY_TLS_ENTRIES];
+	unsigned long	esp0;
+	unsigned long	sysenter_cs;
+	unsigned long	eip;
+	unsigned long	esp;
+	unsigned long	fs;
+	unsigned long	gs;
+/* Hardware debugging registers */
+	unsigned long	debugreg[8];  /* %%db0-7 debug registers */
+/* fault info */
+	unsigned long	cr2, trap_no, error_code;
+/* floating point info */
+	union i387_union	i387;
+/* virtual 86 mode info */
+	struct vm86_struct __user * vm86_info;
+	unsigned long		screen_bitmap;
+	unsigned long		v86flags, v86mask, saved_esp0;
+	unsigned int		saved_fs, saved_gs;
+/* IO permissions */
+	unsigned long	*io_bitmap_ptr;
+ 	unsigned long	iopl;
+/* max allowed port in the bitmap, in bytes: */
+	unsigned long	io_bitmap_max;
+};
+
+#define INIT_THREAD  {							\
+	.vm86_info = NULL,						\
+	.sysenter_cs = __KERNEL_CS,					\
+	.io_bitmap_ptr = NULL,						\
+}
+
+#ifndef CONFIG_X86_NO_TSS
+/*
+ * Note that the .io_bitmap member must be extra-big. This is because
+ * the CPU will access an additional byte beyond the end of the IO
+ * permission bitmap. The extra byte must be all 1 bits, and must
+ * be within the limit.
+ */
+#define INIT_TSS  {							\
+	.esp0		= sizeof(init_stack) + (long)&init_stack,	\
+	.ss0		= __KERNEL_DS,					\
+	.ss1		= __KERNEL_CS,					\
+	.io_bitmap_base	= INVALID_IO_BITMAP_OFFSET,			\
+	.io_bitmap	= { [ 0 ... IO_BITMAP_LONGS] = ~0 },		\
+}
+
+static inline void __load_esp0(struct tss_struct *tss, struct thread_struct *thread)
+{
+	tss->esp0 = thread->esp0;
+	/* This can only happen when SEP is enabled, no need to test "SEP"arately */
+	if (unlikely(tss->ss1 != thread->sysenter_cs)) {
+		tss->ss1 = thread->sysenter_cs;
+		wrmsr(MSR_IA32_SYSENTER_CS, thread->sysenter_cs, 0);
+	}
+}
+#define load_esp0(tss, thread) \
+	__load_esp0(tss, thread)
+#else
+#define load_esp0(tss, thread) do { \
+	if (HYPERVISOR_stack_switch(__KERNEL_DS, (thread)->esp0)) \
+		BUG(); \
+} while (0)
+#endif
+
+#define start_thread(regs, new_eip, new_esp) do {		\
+	__asm__("movl %0,%%fs ; movl %0,%%gs": :"r" (0));	\
+	set_fs(USER_DS);					\
+	regs->xds = __USER_DS;					\
+	regs->xes = __USER_DS;					\
+	regs->xss = __USER_DS;					\
+	regs->xcs = __USER_CS;					\
+	regs->eip = new_eip;					\
+	regs->esp = new_esp;					\
+} while (0)
+
+/*
+ * These special macros can be used to get or set a debugging register
+ */
+#define get_debugreg(var, register)				\
+		(var) = HYPERVISOR_get_debugreg((register))
+#define set_debugreg(value, register)			\
+		WARN_ON(HYPERVISOR_set_debugreg((register), (value)))
+
+/*
+ * Set IOPL bits in EFLAGS from given mask
+ */
+static inline void set_iopl_mask(unsigned mask)
+{
+	struct physdev_set_iopl set_iopl;
+
+	/* Force the change at ring 0. */
+	set_iopl.iopl = (mask == 0) ? 1 : (mask >> 12) & 3;
+	WARN_ON(HYPERVISOR_physdev_op(PHYSDEVOP_set_iopl, &set_iopl));
+}
+
+/* Forward declaration, a strange C thing */
+struct task_struct;
+struct mm_struct;
+
+/* Free all resources held by a thread. */
+extern void release_thread(struct task_struct *);
+
+/* Prepare to copy thread state - unlazy all lazy status */
+extern void prepare_to_copy(struct task_struct *tsk);
+
+/*
+ * create a kernel thread without removing it from tasklists
+ */
+extern int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags);
+
+extern unsigned long thread_saved_pc(struct task_struct *tsk);
+void show_trace(struct task_struct *task, struct pt_regs *regs, unsigned long *stack);
+
+unsigned long get_wchan(struct task_struct *p);
+
+#define THREAD_SIZE_LONGS      (THREAD_SIZE/sizeof(unsigned long))
+#define KSTK_TOP(info)                                                 \
+({                                                                     \
+       unsigned long *__ptr = (unsigned long *)(info);                 \
+       (unsigned long)(&__ptr[THREAD_SIZE_LONGS]);                     \
+})
+
+/*
+ * The below -8 is to reserve 8 bytes on top of the ring0 stack.
+ * This is necessary to guarantee that the entire "struct pt_regs"
+ * is accessable even if the CPU haven't stored the SS/ESP registers
+ * on the stack (interrupt gate does not save these registers
+ * when switching to the same priv ring).
+ * Therefore beware: accessing the xss/esp fields of the
+ * "struct pt_regs" is possible, but they may contain the
+ * completely wrong values.
+ */
+#define task_pt_regs(task)                                             \
+({                                                                     \
+       struct pt_regs *__regs__;                                       \
+       __regs__ = (struct pt_regs *)(KSTK_TOP(task_stack_page(task))-8); \
+       __regs__ - 1;                                                   \
+})
+
+#define KSTK_EIP(task) (task_pt_regs(task)->eip)
+#define KSTK_ESP(task) (task_pt_regs(task)->esp)
+
+
+struct microcode_header {
+	unsigned int hdrver;
+	unsigned int rev;
+	unsigned int date;
+	unsigned int sig;
+	unsigned int cksum;
+	unsigned int ldrver;
+	unsigned int pf;
+	unsigned int datasize;
+	unsigned int totalsize;
+	unsigned int reserved[3];
+};
+
+struct microcode {
+	struct microcode_header hdr;
+	unsigned int bits[0];
+};
+
+typedef struct microcode microcode_t;
+typedef struct microcode_header microcode_header_t;
+
+/* microcode format is extended from prescott processors */
+struct extended_signature {
+	unsigned int sig;
+	unsigned int pf;
+	unsigned int cksum;
+};
+
+struct extended_sigtable {
+	unsigned int count;
+	unsigned int cksum;
+	unsigned int reserved[3];
+	struct extended_signature sigs[0];
+};
+
+/* REP NOP (PAUSE) is a good thing to insert into busy-wait loops. */
+static inline void rep_nop(void)
+{
+	__asm__ __volatile__("rep;nop": : :"memory");
+}
+
+#define cpu_relax()	rep_nop()
+
+/* generic versions from gas */
+#define GENERIC_NOP1	".byte 0x90\n"
+#define GENERIC_NOP2    	".byte 0x89,0xf6\n"
+#define GENERIC_NOP3        ".byte 0x8d,0x76,0x00\n"
+#define GENERIC_NOP4        ".byte 0x8d,0x74,0x26,0x00\n"
+#define GENERIC_NOP5        GENERIC_NOP1 GENERIC_NOP4
+#define GENERIC_NOP6	".byte 0x8d,0xb6,0x00,0x00,0x00,0x00\n"
+#define GENERIC_NOP7	".byte 0x8d,0xb4,0x26,0x00,0x00,0x00,0x00\n"
+#define GENERIC_NOP8	GENERIC_NOP1 GENERIC_NOP7
+
+/* Opteron nops */
+#define K8_NOP1 GENERIC_NOP1
+#define K8_NOP2	".byte 0x66,0x90\n" 
+#define K8_NOP3	".byte 0x66,0x66,0x90\n" 
+#define K8_NOP4	".byte 0x66,0x66,0x66,0x90\n" 
+#define K8_NOP5	K8_NOP3 K8_NOP2 
+#define K8_NOP6	K8_NOP3 K8_NOP3
+#define K8_NOP7	K8_NOP4 K8_NOP3
+#define K8_NOP8	K8_NOP4 K8_NOP4
+
+/* K7 nops */
+/* uses eax dependencies (arbitary choice) */
+#define K7_NOP1  GENERIC_NOP1
+#define K7_NOP2	".byte 0x8b,0xc0\n" 
+#define K7_NOP3	".byte 0x8d,0x04,0x20\n"
+#define K7_NOP4	".byte 0x8d,0x44,0x20,0x00\n"
+#define K7_NOP5	K7_NOP4 ASM_NOP1
+#define K7_NOP6	".byte 0x8d,0x80,0,0,0,0\n"
+#define K7_NOP7        ".byte 0x8D,0x04,0x05,0,0,0,0\n"
+#define K7_NOP8        K7_NOP7 ASM_NOP1
+
+#ifdef CONFIG_MK8
+#define ASM_NOP1 K8_NOP1
+#define ASM_NOP2 K8_NOP2
+#define ASM_NOP3 K8_NOP3
+#define ASM_NOP4 K8_NOP4
+#define ASM_NOP5 K8_NOP5
+#define ASM_NOP6 K8_NOP6
+#define ASM_NOP7 K8_NOP7
+#define ASM_NOP8 K8_NOP8
+#elif defined(CONFIG_MK7)
+#define ASM_NOP1 K7_NOP1
+#define ASM_NOP2 K7_NOP2
+#define ASM_NOP3 K7_NOP3
+#define ASM_NOP4 K7_NOP4
+#define ASM_NOP5 K7_NOP5
+#define ASM_NOP6 K7_NOP6
+#define ASM_NOP7 K7_NOP7
+#define ASM_NOP8 K7_NOP8
+#else
+#define ASM_NOP1 GENERIC_NOP1
+#define ASM_NOP2 GENERIC_NOP2
+#define ASM_NOP3 GENERIC_NOP3
+#define ASM_NOP4 GENERIC_NOP4
+#define ASM_NOP5 GENERIC_NOP5
+#define ASM_NOP6 GENERIC_NOP6
+#define ASM_NOP7 GENERIC_NOP7
+#define ASM_NOP8 GENERIC_NOP8
+#endif
+
+#define ASM_NOP_MAX 8
+
+/* Prefetch instructions for Pentium III and AMD Athlon */
+/* It's not worth to care about 3dnow! prefetches for the K6
+   because they are microcoded there and very slow.
+   However we don't do prefetches for pre XP Athlons currently
+   That should be fixed. */
+#define ARCH_HAS_PREFETCH
+static inline void prefetch(const void *x)
+{
+	alternative_input(ASM_NOP4,
+			  "prefetchnta (%1)",
+			  X86_FEATURE_XMM,
+			  "r" (x));
+}
+
+#define ARCH_HAS_PREFETCH
+#define ARCH_HAS_PREFETCHW
+#define ARCH_HAS_SPINLOCK_PREFETCH
+
+/* 3dnow! prefetch to get an exclusive cache line. Useful for 
+   spinlocks to avoid one state transition in the cache coherency protocol. */
+static inline void prefetchw(const void *x)
+{
+	alternative_input(ASM_NOP4,
+			  "prefetchw (%1)",
+			  X86_FEATURE_3DNOW,
+			  "r" (x));
+}
+#define spin_lock_prefetch(x)	prefetchw(x)
+
+extern void select_idle_routine(const struct cpuinfo_x86 *c);
+
+#define cache_line_size() (boot_cpu_data.x86_cache_alignment)
+
+extern unsigned long boot_option_idle_override;
+extern void enable_sep_cpu(void);
+extern int sysenter_setup(void);
+
+#endif /* __ASM_I386_PROCESSOR_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/ptrace.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/ptrace.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/ptrace.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/ptrace.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,90 @@
+#ifndef _I386_PTRACE_H
+#define _I386_PTRACE_H
+
+#define EBX 0
+#define ECX 1
+#define EDX 2
+#define ESI 3
+#define EDI 4
+#define EBP 5
+#define EAX 6
+#define DS 7
+#define ES 8
+#define FS 9
+#define GS 10
+#define ORIG_EAX 11
+#define EIP 12
+#define CS  13
+#define EFL 14
+#define UESP 15
+#define SS   16
+#define FRAME_SIZE 17
+
+/* this struct defines the way the registers are stored on the 
+   stack during a system call. */
+
+struct pt_regs {
+	long ebx;
+	long ecx;
+	long edx;
+	long esi;
+	long edi;
+	long ebp;
+	long eax;
+	int  xds;
+	int  xes;
+	long orig_eax;
+	long eip;
+	int  xcs;
+	long eflags;
+	long esp;
+	int  xss;
+};
+
+/* Arbitrarily choose the same ptrace numbers as used by the Sparc code. */
+#define PTRACE_GETREGS            12
+#define PTRACE_SETREGS            13
+#define PTRACE_GETFPREGS          14
+#define PTRACE_SETFPREGS          15
+#define PTRACE_GETFPXREGS         18
+#define PTRACE_SETFPXREGS         19
+
+#define PTRACE_OLDSETOPTIONS         21
+
+#define PTRACE_GET_THREAD_AREA    25
+#define PTRACE_SET_THREAD_AREA    26
+
+#define PTRACE_SYSEMU		  31
+#define PTRACE_SYSEMU_SINGLESTEP  32
+
+#ifdef __KERNEL__
+
+#include <asm/vm86.h>
+
+struct task_struct;
+extern void send_sigtrap(struct task_struct *tsk, struct pt_regs *regs, int error_code);
+
+/*
+ * user_mode_vm(regs) determines whether a register set came from user mode.
+ * This is true if V8086 mode was enabled OR if the register set was from
+ * protected mode with RPL-3 CS value.  This tricky test checks that with
+ * one comparison.  Many places in the kernel can bypass this full check
+ * if they have already ruled out V8086 mode, so user_mode(regs) can be used.
+ */
+static inline int user_mode(struct pt_regs *regs)
+{
+	return (regs->xcs & 2) != 0;
+}
+static inline int user_mode_vm(struct pt_regs *regs)
+{
+	return ((regs->xcs & 2) | (regs->eflags & VM_MASK)) != 0;
+}
+#define instruction_pointer(regs) ((regs)->eip)
+#if defined(CONFIG_SMP) && defined(CONFIG_FRAME_POINTER)
+extern unsigned long profile_pc(struct pt_regs *regs);
+#else
+#define profile_pc(regs) instruction_pointer(regs)
+#endif
+#endif /* __KERNEL__ */
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/scatterlist.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/scatterlist.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/scatterlist.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/scatterlist.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,22 @@
+#ifndef _I386_SCATTERLIST_H
+#define _I386_SCATTERLIST_H
+
+struct scatterlist {
+    struct page		*page;
+    unsigned int	offset;
+    unsigned int	length;
+    dma_addr_t		dma_address;
+    unsigned int	dma_length;
+};
+
+/* These macros should be used after a pci_map_sg call has been done
+ * to get bus addresses of each of the SG entries and their lengths.
+ * You should only work with the number of sg entries pci_map_sg
+ * returns.
+ */
+#define sg_dma_address(sg)	((sg)->dma_address)
+#define sg_dma_len(sg)		((sg)->dma_length)
+
+#define ISA_DMA_THRESHOLD (0x00ffffff)
+
+#endif /* !(_I386_SCATTERLIST_H) */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/segment.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/segment.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/segment.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/segment.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,117 @@
+#ifndef _ASM_SEGMENT_H
+#define _ASM_SEGMENT_H
+
+/*
+ * The layout of the per-CPU GDT under Linux:
+ *
+ *   0 - null
+ *   1 - reserved
+ *   2 - reserved
+ *   3 - reserved
+ *
+ *   4 - unused			<==== new cacheline
+ *   5 - unused
+ *
+ *  ------- start of TLS (Thread-Local Storage) segments:
+ *
+ *   6 - TLS segment #1			[ glibc's TLS segment ]
+ *   7 - TLS segment #2			[ Wine's %fs Win32 segment ]
+ *   8 - TLS segment #3
+ *   9 - reserved
+ *  10 - reserved
+ *  11 - reserved
+ *
+ *  ------- start of kernel segments:
+ *
+ *  12 - kernel code segment		<==== new cacheline
+ *  13 - kernel data segment
+ *  14 - default user CS
+ *  15 - default user DS
+ *  16 - TSS
+ *  17 - LDT
+ *  18 - PNPBIOS support (16->32 gate)
+ *  19 - PNPBIOS support
+ *  20 - PNPBIOS support
+ *  21 - PNPBIOS support
+ *  22 - PNPBIOS support
+ *  23 - APM BIOS support
+ *  24 - APM BIOS support
+ *  25 - APM BIOS support 
+ *
+ *  26 - ESPFIX small SS
+ *  27 - unused
+ *  28 - unused
+ *  29 - unused
+ *  30 - unused
+ *  31 - TSS for double fault handler
+ */
+#define GDT_ENTRY_TLS_ENTRIES	3
+#define GDT_ENTRY_TLS_MIN	6
+#define GDT_ENTRY_TLS_MAX 	(GDT_ENTRY_TLS_MIN + GDT_ENTRY_TLS_ENTRIES - 1)
+
+#define TLS_SIZE (GDT_ENTRY_TLS_ENTRIES * 8)
+
+#define GDT_ENTRY_DEFAULT_USER_CS	14
+#define __USER_CS (GDT_ENTRY_DEFAULT_USER_CS * 8 + 3)
+
+#define GDT_ENTRY_DEFAULT_USER_DS	15
+#define __USER_DS (GDT_ENTRY_DEFAULT_USER_DS * 8 + 3)
+
+#define GDT_ENTRY_KERNEL_BASE	12
+
+#define GDT_ENTRY_KERNEL_CS		(GDT_ENTRY_KERNEL_BASE + 0)
+#define __KERNEL_CS (GDT_ENTRY_KERNEL_CS * 8)
+#define GET_KERNEL_CS() (__KERNEL_CS | (xen_feature(XENFEAT_supervisor_mode_kernel)?0:1) )
+
+#define GDT_ENTRY_KERNEL_DS		(GDT_ENTRY_KERNEL_BASE + 1)
+#define __KERNEL_DS (GDT_ENTRY_KERNEL_DS * 8)
+#define GET_KERNEL_DS() (__KERNEL_DS | (xen_feature(XENFEAT_supervisor_mode_kernel)?0:1) )
+
+#define GDT_ENTRY_TSS			(GDT_ENTRY_KERNEL_BASE + 4)
+#define GDT_ENTRY_LDT			(GDT_ENTRY_KERNEL_BASE + 5)
+
+#define GDT_ENTRY_PNPBIOS_BASE		(GDT_ENTRY_KERNEL_BASE + 6)
+#define GDT_ENTRY_APMBIOS_BASE		(GDT_ENTRY_KERNEL_BASE + 11)
+
+#define GDT_ENTRY_ESPFIX_SS		(GDT_ENTRY_KERNEL_BASE + 14)
+#define __ESPFIX_SS (GDT_ENTRY_ESPFIX_SS * 8)
+
+#define GDT_ENTRY_DOUBLEFAULT_TSS	31
+
+/*
+ * The GDT has 32 entries
+ */
+#define GDT_ENTRIES 32
+
+#define GDT_SIZE (GDT_ENTRIES * 8)
+
+/* Simple and small GDT entries for booting only */
+
+#define GDT_ENTRY_BOOT_CS		2
+#define __BOOT_CS	(GDT_ENTRY_BOOT_CS * 8)
+
+#define GDT_ENTRY_BOOT_DS		(GDT_ENTRY_BOOT_CS + 1)
+#define __BOOT_DS	(GDT_ENTRY_BOOT_DS * 8)
+
+/* The PnP BIOS entries in the GDT */
+#define GDT_ENTRY_PNPBIOS_CS32		(GDT_ENTRY_PNPBIOS_BASE + 0)
+#define GDT_ENTRY_PNPBIOS_CS16		(GDT_ENTRY_PNPBIOS_BASE + 1)
+#define GDT_ENTRY_PNPBIOS_DS		(GDT_ENTRY_PNPBIOS_BASE + 2)
+#define GDT_ENTRY_PNPBIOS_TS1		(GDT_ENTRY_PNPBIOS_BASE + 3)
+#define GDT_ENTRY_PNPBIOS_TS2		(GDT_ENTRY_PNPBIOS_BASE + 4)
+
+/* The PnP BIOS selectors */
+#define PNP_CS32   (GDT_ENTRY_PNPBIOS_CS32 * 8)	/* segment for calling fn */
+#define PNP_CS16   (GDT_ENTRY_PNPBIOS_CS16 * 8)	/* code segment for BIOS */
+#define PNP_DS     (GDT_ENTRY_PNPBIOS_DS * 8)	/* data segment for BIOS */
+#define PNP_TS1    (GDT_ENTRY_PNPBIOS_TS1 * 8)	/* transfer data segment */
+#define PNP_TS2    (GDT_ENTRY_PNPBIOS_TS2 * 8)	/* another data segment */
+
+/*
+ * The interrupt descriptor table has room for 256 idt's,
+ * the global descriptor table is dependent on the number
+ * of tasks we can have..
+ */
+#define IDT_ENTRIES 256
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/setup.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/setup.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/setup.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/setup.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,81 @@
+/*
+ *	Just a place holder. We don't want to have to test x86 before
+ *	we include stuff
+ */
+
+#ifndef _i386_SETUP_H
+#define _i386_SETUP_H
+
+#ifdef __KERNEL__
+#include <linux/pfn.h>
+
+/*
+ * Reserved space for vmalloc and iomap - defined in asm/page.h
+ */
+#define MAXMEM_PFN	PFN_DOWN(MAXMEM)
+#define MAX_NONPAE_PFN	(1 << 20)
+#endif
+
+#define PARAM_SIZE 4096
+#define COMMAND_LINE_SIZE 256
+
+#define OLD_CL_MAGIC_ADDR	0x90020
+#define OLD_CL_MAGIC		0xA33F
+#define OLD_CL_BASE_ADDR	0x90000
+#define OLD_CL_OFFSET		0x90022
+#define NEW_CL_POINTER		0x228	/* Relative to real mode data */
+
+#ifndef __ASSEMBLY__
+/*
+ * This is set up by the setup-routine at boot-time
+ */
+extern unsigned char boot_params[PARAM_SIZE];
+
+#define PARAM	(boot_params)
+#define SCREEN_INFO (*(struct screen_info *) (PARAM+0))
+#define EXT_MEM_K (*(unsigned short *) (PARAM+2))
+#define ALT_MEM_K (*(unsigned long *) (PARAM+0x1e0))
+#define E820_MAP_NR (*(char*) (PARAM+E820NR))
+#define E820_MAP    ((struct e820entry *) (PARAM+E820MAP))
+#define APM_BIOS_INFO (*(struct apm_bios_info *) (PARAM+0x40))
+#define IST_INFO   (*(struct ist_info *) (PARAM+0x60))
+#define DRIVE_INFO (*(struct drive_info_struct *) (PARAM+0x80))
+#define SYS_DESC_TABLE (*(struct sys_desc_table_struct*)(PARAM+0xa0))
+#define EFI_SYSTAB ((efi_system_table_t *) *((unsigned long *)(PARAM+0x1c4)))
+#define EFI_MEMDESC_SIZE (*((unsigned long *) (PARAM+0x1c8)))
+#define EFI_MEMDESC_VERSION (*((unsigned long *) (PARAM+0x1cc)))
+#define EFI_MEMMAP ((void *) *((unsigned long *)(PARAM+0x1d0)))
+#define EFI_MEMMAP_SIZE (*((unsigned long *) (PARAM+0x1d4)))
+#define MOUNT_ROOT_RDONLY (*(unsigned short *) (PARAM+0x1F2))
+#define RAMDISK_FLAGS (*(unsigned short *) (PARAM+0x1F8))
+#define VIDEO_MODE (*(unsigned short *) (PARAM+0x1FA))
+#define ORIG_ROOT_DEV (*(unsigned short *) (PARAM+0x1FC))
+#define AUX_DEVICE_INFO (*(unsigned char *) (PARAM+0x1FF))
+#define LOADER_TYPE (*(unsigned char *) (PARAM+0x210))
+#define KERNEL_START (*(unsigned long *) (PARAM+0x214))
+#define INITRD_START (__pa(xen_start_info->mod_start))
+#define INITRD_SIZE (xen_start_info->mod_len)
+#define EDID_INFO   (*(struct edid_info *) (PARAM+0x440))
+#define EDD_NR     (*(unsigned char *) (PARAM+EDDNR))
+#define EDD_MBR_SIG_NR (*(unsigned char *) (PARAM+EDD_MBR_SIG_NR_BUF))
+#define EDD_MBR_SIGNATURE ((unsigned int *) (PARAM+EDD_MBR_SIG_BUF))
+#define EDD_BUF     ((struct edd_info *) (PARAM+EDDBUF))
+
+/*
+ * Do NOT EVER look at the BIOS memory size location.
+ * It does not work on many machines.
+ */
+#define LOWMEMSIZE()	(0x9f000)
+
+struct e820entry;
+
+char * __init machine_specific_memory_setup(void);
+
+int __init copy_e820_map(struct e820entry * biosmap, int nr_map);
+int __init sanitize_e820_map(struct e820entry * biosmap, char * pnr_map);
+void __init add_memory_region(unsigned long long start,
+			      unsigned long long size, int type);
+
+#endif /* __ASSEMBLY__ */
+
+#endif /* _i386_SETUP_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/smp.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/smp.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/smp.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/smp.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,103 @@
+#ifndef __ASM_SMP_H
+#define __ASM_SMP_H
+
+/*
+ * We need the APIC definitions automatically as part of 'smp.h'
+ */
+#ifndef __ASSEMBLY__
+#include <linux/kernel.h>
+#include <linux/threads.h>
+#include <linux/cpumask.h>
+#endif
+
+#ifdef CONFIG_X86_LOCAL_APIC
+#ifndef __ASSEMBLY__
+#include <asm/fixmap.h>
+#include <asm/bitops.h>
+#include <asm/mpspec.h>
+#ifdef CONFIG_X86_IO_APIC
+#include <asm/io_apic.h>
+#endif
+#include <asm/apic.h>
+#endif
+#endif
+
+#define BAD_APICID 0xFFu
+#ifdef CONFIG_SMP
+#ifndef __ASSEMBLY__
+
+/*
+ * Private routines/data
+ */
+ 
+extern void smp_alloc_memory(void);
+extern int pic_mode;
+extern int smp_num_siblings;
+extern cpumask_t cpu_sibling_map[];
+extern cpumask_t cpu_core_map[];
+
+extern void (*mtrr_hook) (void);
+extern void zap_low_mappings (void);
+extern void lock_ipi_call_lock(void);
+extern void unlock_ipi_call_lock(void);
+
+#define MAX_APICID 256
+extern u8 x86_cpu_to_apicid[];
+
+#define cpu_physical_id(cpu)	x86_cpu_to_apicid[cpu]
+
+#ifdef CONFIG_HOTPLUG_CPU
+extern void cpu_exit_clear(void);
+extern void cpu_uninit(void);
+#endif
+
+/*
+ * This function is needed by all SMP systems. It must _always_ be valid
+ * from the initial startup. We map APIC_BASE very early in page_setup(),
+ * so this is correct in the x86 case.
+ */
+#define raw_smp_processor_id() (current_thread_info()->cpu)
+
+extern cpumask_t cpu_possible_map;
+#define cpu_callin_map cpu_possible_map
+
+/* We don't mark CPUs online until __cpu_up(), so we need another measure */
+static inline int num_booting_cpus(void)
+{
+	return cpus_weight(cpu_possible_map);
+}
+
+#ifdef CONFIG_X86_LOCAL_APIC
+
+#ifdef APIC_DEFINITION
+extern int hard_smp_processor_id(void);
+#else
+#include <mach_apicdef.h>
+static inline int hard_smp_processor_id(void)
+{
+	/* we don't want to mark this access volatile - bad code generation */
+	return GET_APIC_ID(*(unsigned long *)(APIC_BASE+APIC_ID));
+}
+#endif
+
+static __inline int logical_smp_processor_id(void)
+{
+	/* we don't want to mark this access volatile - bad code generation */
+	return GET_APIC_LOGICAL_ID(*(unsigned long *)(APIC_BASE+APIC_LDR));
+}
+
+#endif
+
+extern int __cpu_disable(void);
+extern void __cpu_die(unsigned int cpu);
+extern void prefill_possible_map(void);
+#endif /* !__ASSEMBLY__ */
+
+#else /* CONFIG_SMP */
+
+#define cpu_physical_id(cpu)		boot_cpu_physical_apicid
+
+#define NO_PROC_ID		0xFF		/* No processor magic marker */
+
+#endif
+#endif
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/spinlock.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/spinlock.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/spinlock.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/spinlock.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,202 @@
+#ifndef __ASM_SPINLOCK_H
+#define __ASM_SPINLOCK_H
+
+#include <asm/atomic.h>
+#include <asm/rwlock.h>
+#include <asm/page.h>
+#include <linux/compiler.h>
+
+/*
+ * Your basic SMP spinlocks, allowing only a single CPU anywhere
+ *
+ * Simple spin lock operations.  There are two variants, one clears IRQ's
+ * on the local processor, one does not.
+ *
+ * We make no fairness assumptions. They have a cost.
+ *
+ * (the type definitions are in asm/spinlock_types.h)
+ */
+
+#define __raw_spin_is_locked(x) \
+		(*(volatile signed char *)(&(x)->slock) <= 0)
+
+#define __raw_spin_lock_string \
+	"\n1:\t" \
+	LOCK_PREFIX " ; decb %0\n\t" \
+	"jns 3f\n" \
+	"2:\t" \
+	"rep;nop\n\t" \
+	"cmpb $0,%0\n\t" \
+	"jle 2b\n\t" \
+	"jmp 1b\n" \
+	"3:\n\t"
+
+/*
+ * NOTE: there's an irqs-on section here, which normally would have to be
+ * irq-traced, but on CONFIG_TRACE_IRQFLAGS we never use
+ * __raw_spin_lock_string_flags().
+ */
+#define __raw_spin_lock_string_flags \
+	"\n1:\t" \
+	LOCK_PREFIX " ; decb %0\n\t" \
+	"jns 5f\n" \
+	"2:\t" \
+	"testl $0x200, %1\n\t" \
+	"jz 4f\n\t" \
+	"#sti\n" \
+	"3:\t" \
+	"rep;nop\n\t" \
+	"cmpb $0, %0\n\t" \
+	"jle 3b\n\t" \
+	"#cli\n\t" \
+	"jmp 1b\n" \
+	"4:\t" \
+	"rep;nop\n\t" \
+	"cmpb $0, %0\n\t" \
+	"jg 1b\n\t" \
+	"jmp 4b\n" \
+	"5:\n\t"
+
+static inline void __raw_spin_lock(raw_spinlock_t *lock)
+{
+	asm(__raw_spin_lock_string : "+m" (lock->slock) : : "memory");
+}
+
+/*
+ * It is easier for the lock validator if interrupts are not re-enabled
+ * in the middle of a lock-acquire. This is a performance feature anyway
+ * so we turn it off:
+ */
+#ifndef CONFIG_PROVE_LOCKING
+static inline void __raw_spin_lock_flags(raw_spinlock_t *lock, unsigned long flags)
+{
+	asm(__raw_spin_lock_string_flags : "+m" (lock->slock) : "r" (flags) : "memory");
+}
+#endif
+
+static inline int __raw_spin_trylock(raw_spinlock_t *lock)
+{
+	char oldval;
+	__asm__ __volatile__(
+		"xchgb %b0,%1"
+		:"=q" (oldval), "+m" (lock->slock)
+		:"0" (0) : "memory");
+	return oldval > 0;
+}
+
+/*
+ * __raw_spin_unlock based on writing $1 to the low byte.
+ * This method works. Despite all the confusion.
+ * (except on PPro SMP or if we are using OOSTORE, so we use xchgb there)
+ * (PPro errata 66, 92)
+ */
+
+#if !defined(CONFIG_X86_OOSTORE) && !defined(CONFIG_X86_PPRO_FENCE)
+
+#define __raw_spin_unlock_string \
+	"movb $1,%0" \
+		:"+m" (lock->slock) : : "memory"
+
+
+static inline void __raw_spin_unlock(raw_spinlock_t *lock)
+{
+	__asm__ __volatile__(
+		__raw_spin_unlock_string
+	);
+}
+
+#else
+
+#define __raw_spin_unlock_string \
+	"xchgb %b0, %1" \
+		:"=q" (oldval), "+m" (lock->slock) \
+		:"0" (oldval) : "memory"
+
+static inline void __raw_spin_unlock(raw_spinlock_t *lock)
+{
+	char oldval = 1;
+
+	__asm__ __volatile__(
+		__raw_spin_unlock_string
+	);
+}
+
+#endif
+
+#define __raw_spin_unlock_wait(lock) \
+	do { while (__raw_spin_is_locked(lock)) cpu_relax(); } while (0)
+
+/*
+ * Read-write spinlocks, allowing multiple readers
+ * but only one writer.
+ *
+ * NOTE! it is quite common to have readers in interrupts
+ * but no interrupt writers. For those circumstances we
+ * can "mix" irq-safe locks - any writer needs to get a
+ * irq-safe write-lock, but readers can get non-irqsafe
+ * read-locks.
+ *
+ * On x86, we implement read-write locks as a 32-bit counter
+ * with the high bit (sign) being the "contended" bit.
+ *
+ * The inline assembly is non-obvious. Think about it.
+ *
+ * Changed to use the same technique as rw semaphores.  See
+ * semaphore.h for details.  -ben
+ *
+ * the helpers are in arch/i386/kernel/semaphore.c
+ */
+
+/**
+ * read_can_lock - would read_trylock() succeed?
+ * @lock: the rwlock in question.
+ */
+#define __raw_read_can_lock(x)		((int)(x)->lock > 0)
+
+/**
+ * write_can_lock - would write_trylock() succeed?
+ * @lock: the rwlock in question.
+ */
+#define __raw_write_can_lock(x)		((x)->lock == RW_LOCK_BIAS)
+
+static inline void __raw_read_lock(raw_rwlock_t *rw)
+{
+	__build_read_lock(rw, "__read_lock_failed");
+}
+
+static inline void __raw_write_lock(raw_rwlock_t *rw)
+{
+	__build_write_lock(rw, "__write_lock_failed");
+}
+
+static inline int __raw_read_trylock(raw_rwlock_t *lock)
+{
+	atomic_t *count = (atomic_t *)lock;
+	atomic_dec(count);
+	if (atomic_read(count) >= 0)
+		return 1;
+	atomic_inc(count);
+	return 0;
+}
+
+static inline int __raw_write_trylock(raw_rwlock_t *lock)
+{
+	atomic_t *count = (atomic_t *)lock;
+	if (atomic_sub_and_test(RW_LOCK_BIAS, count))
+		return 1;
+	atomic_add(RW_LOCK_BIAS, count);
+	return 0;
+}
+
+static inline void __raw_read_unlock(raw_rwlock_t *rw)
+{
+	asm volatile(LOCK_PREFIX "incl %0" :"+m" (rw->lock) : : "memory");
+}
+
+static inline void __raw_write_unlock(raw_rwlock_t *rw)
+{
+	asm volatile(LOCK_PREFIX "addl $" RW_LOCK_BIAS_STR ", %0"
+				 : "+m" (rw->lock) : : "memory");
+}
+
+#endif /* __ASM_SPINLOCK_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/swiotlb.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/swiotlb.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/swiotlb.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/swiotlb.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,43 @@
+#ifndef _ASM_SWIOTLB_H
+#define _ASM_SWIOTLB_H 1
+
+/* SWIOTLB interface */
+
+extern dma_addr_t swiotlb_map_single(struct device *hwdev, void *ptr, size_t size,
+				      int dir);
+extern void swiotlb_unmap_single(struct device *hwdev, dma_addr_t dev_addr,
+				  size_t size, int dir);
+extern void swiotlb_sync_single_for_cpu(struct device *hwdev,
+					 dma_addr_t dev_addr,
+					 size_t size, int dir);
+extern void swiotlb_sync_single_for_device(struct device *hwdev,
+					    dma_addr_t dev_addr,
+					    size_t size, int dir);
+extern void swiotlb_sync_sg_for_cpu(struct device *hwdev,
+				     struct scatterlist *sg, int nelems,
+				     int dir);
+extern void swiotlb_sync_sg_for_device(struct device *hwdev,
+					struct scatterlist *sg, int nelems,
+					int dir);
+extern int swiotlb_map_sg(struct device *hwdev, struct scatterlist *sg,
+		      int nents, int direction);
+extern void swiotlb_unmap_sg(struct device *hwdev, struct scatterlist *sg,
+			 int nents, int direction);
+extern int swiotlb_dma_mapping_error(dma_addr_t dma_addr);
+#ifdef CONFIG_HIGHMEM
+extern dma_addr_t swiotlb_map_page(struct device *hwdev, struct page *page,
+                                   unsigned long offset, size_t size,
+                                   enum dma_data_direction direction);
+extern void swiotlb_unmap_page(struct device *hwdev, dma_addr_t dma_address,
+                               size_t size, enum dma_data_direction direction);
+#endif
+extern int swiotlb_dma_supported(struct device *hwdev, u64 mask);
+extern void swiotlb_init(void);
+
+#ifdef CONFIG_SWIOTLB
+extern int swiotlb;
+#else
+#define swiotlb 0
+#endif
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/synch_bitops.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/synch_bitops.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/synch_bitops.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/synch_bitops.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,126 @@
+#ifndef __XEN_SYNCH_BITOPS_H__
+#define __XEN_SYNCH_BITOPS_H__
+
+/*
+ * Copyright 1992, Linus Torvalds.
+ * Heavily modified to provide guaranteed strong synchronisation
+ * when communicating with Xen or other guest OSes running on other CPUs.
+ */
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+#define ADDR (*(volatile long *) addr)
+
+static __inline__ void synch_set_bit(int nr, volatile void * addr)
+{
+    __asm__ __volatile__ ( 
+        "lock btsl %1,%0"
+        : "+m" (ADDR) : "Ir" (nr) : "memory" );
+}
+
+static __inline__ void synch_clear_bit(int nr, volatile void * addr)
+{
+    __asm__ __volatile__ (
+        "lock btrl %1,%0"
+        : "+m" (ADDR) : "Ir" (nr) : "memory" );
+}
+
+static __inline__ void synch_change_bit(int nr, volatile void * addr)
+{
+    __asm__ __volatile__ (
+        "lock btcl %1,%0"
+        : "+m" (ADDR) : "Ir" (nr) : "memory" );
+}
+
+static __inline__ int synch_test_and_set_bit(int nr, volatile void * addr)
+{
+    int oldbit;
+    __asm__ __volatile__ (
+        "lock btsl %2,%1\n\tsbbl %0,%0"
+        : "=r" (oldbit), "+m" (ADDR) : "Ir" (nr) : "memory");
+    return oldbit;
+}
+
+static __inline__ int synch_test_and_clear_bit(int nr, volatile void * addr)
+{
+    int oldbit;
+    __asm__ __volatile__ (
+        "lock btrl %2,%1\n\tsbbl %0,%0"
+        : "=r" (oldbit), "+m" (ADDR) : "Ir" (nr) : "memory");
+    return oldbit;
+}
+
+static __inline__ int synch_test_and_change_bit(int nr, volatile void * addr)
+{
+    int oldbit;
+
+    __asm__ __volatile__ (
+        "lock btcl %2,%1\n\tsbbl %0,%0"
+        : "=r" (oldbit), "+m" (ADDR) : "Ir" (nr) : "memory");
+    return oldbit;
+}
+
+struct __synch_xchg_dummy { unsigned long a[100]; };
+#define __synch_xg(x) ((struct __synch_xchg_dummy *)(x))
+
+#define synch_cmpxchg(ptr, old, new) \
+((__typeof__(*(ptr)))__synch_cmpxchg((ptr),\
+                                     (unsigned long)(old), \
+                                     (unsigned long)(new), \
+                                     sizeof(*(ptr))))
+
+static inline unsigned long __synch_cmpxchg(volatile void *ptr,
+					    unsigned long old,
+					    unsigned long new, int size)
+{
+	unsigned long prev;
+	switch (size) {
+	case 1:
+		__asm__ __volatile__("lock; cmpxchgb %b1,%2"
+				     : "=a"(prev)
+				     : "q"(new), "m"(*__synch_xg(ptr)),
+				       "0"(old)
+				     : "memory");
+		return prev;
+	case 2:
+		__asm__ __volatile__("lock; cmpxchgw %w1,%2"
+				     : "=a"(prev)
+				     : "r"(new), "m"(*__synch_xg(ptr)),
+				       "0"(old)
+				     : "memory");
+		return prev;
+#ifdef CONFIG_X86_64
+	case 4:
+		__asm__ __volatile__("lock; cmpxchgl %k1,%2"
+				     : "=a"(prev)
+				     : "r"(new), "m"(*__synch_xg(ptr)),
+				       "0"(old)
+				     : "memory");
+		return prev;
+	case 8:
+		__asm__ __volatile__("lock; cmpxchgq %1,%2"
+				     : "=a"(prev)
+				     : "r"(new), "m"(*__synch_xg(ptr)),
+				       "0"(old)
+				     : "memory");
+		return prev;
+#else
+	case 4:
+		__asm__ __volatile__("lock; cmpxchgl %1,%2"
+				     : "=a"(prev)
+				     : "r"(new), "m"(*__synch_xg(ptr)),
+				       "0"(old)
+				     : "memory");
+		return prev;
+#endif
+	}
+	return old;
+}
+
+#define synch_test_bit test_bit
+
+#define synch_cmpxchg_subword synch_cmpxchg
+
+#endif /* __XEN_SYNCH_BITOPS_H__ */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/system.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/system.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/system.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/system.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,488 @@
+#ifndef __ASM_SYSTEM_H
+#define __ASM_SYSTEM_H
+
+#include <linux/kernel.h>
+#include <asm/segment.h>
+#include <asm/cpufeature.h>
+#include <linux/bitops.h> /* for LOCK_PREFIX */
+#include <asm/synch_bitops.h>
+#include <asm/hypervisor.h>
+
+#ifdef __KERNEL__
+
+struct task_struct;	/* one of the stranger aspects of C forward declarations.. */
+extern struct task_struct * FASTCALL(__switch_to(struct task_struct *prev, struct task_struct *next));
+
+/*
+ * Saving eflags is important. It switches not only IOPL between tasks,
+ * it also protects other tasks from NT leaking through sysenter etc.
+ */
+#define switch_to(prev,next,last) do {					\
+	unsigned long esi,edi;						\
+	asm volatile("pushfl\n\t"		/* Save flags */	\
+		     "pushl %%ebp\n\t"					\
+		     "movl %%esp,%0\n\t"	/* save ESP */		\
+		     "movl %5,%%esp\n\t"	/* restore ESP */	\
+		     "movl $1f,%1\n\t"		/* save EIP */		\
+		     "pushl %6\n\t"		/* restore EIP */	\
+		     "jmp __switch_to\n"				\
+		     "1:\t"						\
+		     "popl %%ebp\n\t"					\
+		     "popfl"						\
+		     :"=m" (prev->thread.esp),"=m" (prev->thread.eip),	\
+		      "=a" (last),"=S" (esi),"=D" (edi)			\
+		     :"m" (next->thread.esp),"m" (next->thread.eip),	\
+		      "2" (prev), "d" (next));				\
+} while (0)
+
+#define _set_base(addr,base) do { unsigned long __pr; \
+__asm__ __volatile__ ("movw %%dx,%1\n\t" \
+	"rorl $16,%%edx\n\t" \
+	"movb %%dl,%2\n\t" \
+	"movb %%dh,%3" \
+	:"=&d" (__pr) \
+	:"m" (*((addr)+2)), \
+	 "m" (*((addr)+4)), \
+	 "m" (*((addr)+7)), \
+         "0" (base) \
+        ); } while(0)
+
+#define _set_limit(addr,limit) do { unsigned long __lr; \
+__asm__ __volatile__ ("movw %%dx,%1\n\t" \
+	"rorl $16,%%edx\n\t" \
+	"movb %2,%%dh\n\t" \
+	"andb $0xf0,%%dh\n\t" \
+	"orb %%dh,%%dl\n\t" \
+	"movb %%dl,%2" \
+	:"=&d" (__lr) \
+	:"m" (*(addr)), \
+	 "m" (*((addr)+6)), \
+	 "0" (limit) \
+        ); } while(0)
+
+#define set_base(ldt,base) _set_base( ((char *)&(ldt)) , (base) )
+#define set_limit(ldt,limit) _set_limit( ((char *)&(ldt)) , ((limit)-1) )
+
+/*
+ * Load a segment. Fall back on loading the zero
+ * segment if something goes wrong..
+ */
+#define loadsegment(seg,value)			\
+	asm volatile("\n"			\
+		"1:\t"				\
+		"mov %0,%%" #seg "\n"		\
+		"2:\n"				\
+		".section .fixup,\"ax\"\n"	\
+		"3:\t"				\
+		"pushl $0\n\t"			\
+		"popl %%" #seg "\n\t"		\
+		"jmp 2b\n"			\
+		".previous\n"			\
+		".section __ex_table,\"a\"\n\t"	\
+		".align 4\n\t"			\
+		".long 1b,3b\n"			\
+		".previous"			\
+		: :"rm" (value))
+
+/*
+ * Save a segment register away
+ */
+#define savesegment(seg, value) \
+	asm volatile("mov %%" #seg ",%0":"=rm" (value))
+
+#define read_cr0() ({ \
+	unsigned int __dummy; \
+	__asm__ __volatile__( \
+		"movl %%cr0,%0\n\t" \
+		:"=r" (__dummy)); \
+	__dummy; \
+})
+#define write_cr0(x) \
+	__asm__ __volatile__("movl %0,%%cr0": :"r" (x))
+
+#define read_cr2() (current_vcpu_info()->arch.cr2)
+#define write_cr2(x) \
+	__asm__ __volatile__("movl %0,%%cr2": :"r" (x))
+
+#define read_cr3() ({ \
+	unsigned int __dummy; \
+	__asm__ ( \
+		"movl %%cr3,%0\n\t" \
+		:"=r" (__dummy)); \
+	__dummy = xen_cr3_to_pfn(__dummy); \
+	mfn_to_pfn(__dummy) << PAGE_SHIFT; \
+})
+#define write_cr3(x) ({						\
+	unsigned int __dummy = pfn_to_mfn((x) >> PAGE_SHIFT);	\
+	__dummy = xen_pfn_to_cr3(__dummy);			\
+	__asm__ __volatile__("movl %0,%%cr3": :"r" (__dummy));	\
+})
+#define read_cr4() ({ \
+	unsigned int __dummy; \
+	__asm__( \
+		"movl %%cr4,%0\n\t" \
+		:"=r" (__dummy)); \
+	__dummy; \
+})
+#define read_cr4_safe() ({			      \
+	unsigned int __dummy;			      \
+	/* This could fault if %cr4 does not exist */ \
+	__asm__("1: movl %%cr4, %0		\n"   \
+		"2:				\n"   \
+		".section __ex_table,\"a\"	\n"   \
+		".long 1b,2b			\n"   \
+		".previous			\n"   \
+		: "=r" (__dummy): "0" (0));	      \
+	__dummy;				      \
+})
+
+#define write_cr4(x) \
+	__asm__ __volatile__("movl %0,%%cr4": :"r" (x))
+
+/*
+ * Clear and set 'TS' bit respectively
+ */
+#define clts() (HYPERVISOR_fpu_taskswitch(0))
+#define stts() (HYPERVISOR_fpu_taskswitch(1))
+
+#endif	/* __KERNEL__ */
+
+#define wbinvd() \
+	__asm__ __volatile__ ("wbinvd": : :"memory")
+
+static inline unsigned long get_limit(unsigned long segment)
+{
+	unsigned long __limit;
+	__asm__("lsll %1,%0"
+		:"=r" (__limit):"r" (segment));
+	return __limit+1;
+}
+
+#define nop() __asm__ __volatile__ ("nop")
+
+#define xchg(ptr,v) ((__typeof__(*(ptr)))__xchg((unsigned long)(v),(ptr),sizeof(*(ptr))))
+
+#define tas(ptr) (xchg((ptr),1))
+
+struct __xchg_dummy { unsigned long a[100]; };
+#define __xg(x) ((struct __xchg_dummy *)(x))
+
+
+#ifdef CONFIG_X86_CMPXCHG64
+
+/*
+ * The semantics of XCHGCMP8B are a bit strange, this is why
+ * there is a loop and the loading of %%eax and %%edx has to
+ * be inside. This inlines well in most cases, the cached
+ * cost is around ~38 cycles. (in the future we might want
+ * to do an SIMD/3DNOW!/MMX/FPU 64-bit store here, but that
+ * might have an implicit FPU-save as a cost, so it's not
+ * clear which path to go.)
+ *
+ * cmpxchg8b must be used with the lock prefix here to allow
+ * the instruction to be executed atomically, see page 3-102
+ * of the instruction set reference 24319102.pdf. We need
+ * the reader side to see the coherent 64bit value.
+ */
+static inline void __set_64bit (unsigned long long * ptr,
+		unsigned int low, unsigned int high)
+{
+	__asm__ __volatile__ (
+		"\n1:\t"
+		"movl (%0), %%eax\n\t"
+		"movl 4(%0), %%edx\n\t"
+		"lock cmpxchg8b (%0)\n\t"
+		"jnz 1b"
+		: /* no outputs */
+		:	"D"(ptr),
+			"b"(low),
+			"c"(high)
+		:	"ax","dx","memory");
+}
+
+static inline void __set_64bit_constant (unsigned long long *ptr,
+						 unsigned long long value)
+{
+	__set_64bit(ptr,(unsigned int)(value), (unsigned int)((value)>>32ULL));
+}
+#define ll_low(x)	*(((unsigned int*)&(x))+0)
+#define ll_high(x)	*(((unsigned int*)&(x))+1)
+
+static inline void __set_64bit_var (unsigned long long *ptr,
+			 unsigned long long value)
+{
+	__set_64bit(ptr,ll_low(value), ll_high(value));
+}
+
+#define set_64bit(ptr,value) \
+(__builtin_constant_p(value) ? \
+ __set_64bit_constant(ptr, value) : \
+ __set_64bit_var(ptr, value) )
+
+#define _set_64bit(ptr,value) \
+(__builtin_constant_p(value) ? \
+ __set_64bit(ptr, (unsigned int)(value), (unsigned int)((value)>>32ULL) ) : \
+ __set_64bit(ptr, ll_low(value), ll_high(value)) )
+
+#endif
+
+/*
+ * Note: no "lock" prefix even on SMP: xchg always implies lock anyway
+ * Note 2: xchg has side effect, so that attribute volatile is necessary,
+ *	  but generally the primitive is invalid, *ptr is output argument. --ANK
+ */
+static inline unsigned long __xchg(unsigned long x, volatile void * ptr, int size)
+{
+	switch (size) {
+		case 1:
+			__asm__ __volatile__("xchgb %b0,%1"
+				:"=q" (x)
+				:"m" (*__xg(ptr)), "0" (x)
+				:"memory");
+			break;
+		case 2:
+			__asm__ __volatile__("xchgw %w0,%1"
+				:"=r" (x)
+				:"m" (*__xg(ptr)), "0" (x)
+				:"memory");
+			break;
+		case 4:
+			__asm__ __volatile__("xchgl %0,%1"
+				:"=r" (x)
+				:"m" (*__xg(ptr)), "0" (x)
+				:"memory");
+			break;
+	}
+	return x;
+}
+
+/*
+ * Atomic compare and exchange.  Compare OLD with MEM, if identical,
+ * store NEW in MEM.  Return the initial value in MEM.  Success is
+ * indicated by comparing RETURN with OLD.
+ */
+
+#ifdef CONFIG_X86_CMPXCHG
+#define __HAVE_ARCH_CMPXCHG 1
+#define cmpxchg(ptr,o,n)\
+	((__typeof__(*(ptr)))__cmpxchg((ptr),(unsigned long)(o),\
+					(unsigned long)(n),sizeof(*(ptr))))
+#endif
+
+static inline unsigned long __cmpxchg(volatile void *ptr, unsigned long old,
+				      unsigned long new, int size)
+{
+	unsigned long prev;
+	switch (size) {
+	case 1:
+		__asm__ __volatile__(LOCK_PREFIX "cmpxchgb %b1,%2"
+				     : "=a"(prev)
+				     : "q"(new), "m"(*__xg(ptr)), "0"(old)
+				     : "memory");
+		return prev;
+	case 2:
+		__asm__ __volatile__(LOCK_PREFIX "cmpxchgw %w1,%2"
+				     : "=a"(prev)
+				     : "r"(new), "m"(*__xg(ptr)), "0"(old)
+				     : "memory");
+		return prev;
+	case 4:
+		__asm__ __volatile__(LOCK_PREFIX "cmpxchgl %1,%2"
+				     : "=a"(prev)
+				     : "r"(new), "m"(*__xg(ptr)), "0"(old)
+				     : "memory");
+		return prev;
+	}
+	return old;
+}
+
+#ifndef CONFIG_X86_CMPXCHG
+/*
+ * Building a kernel capable running on 80386. It may be necessary to
+ * simulate the cmpxchg on the 80386 CPU. For that purpose we define
+ * a function for each of the sizes we support.
+ */
+
+extern unsigned long cmpxchg_386_u8(volatile void *, u8, u8);
+extern unsigned long cmpxchg_386_u16(volatile void *, u16, u16);
+extern unsigned long cmpxchg_386_u32(volatile void *, u32, u32);
+
+static inline unsigned long cmpxchg_386(volatile void *ptr, unsigned long old,
+				      unsigned long new, int size)
+{
+	switch (size) {
+	case 1:
+		return cmpxchg_386_u8(ptr, old, new);
+	case 2:
+		return cmpxchg_386_u16(ptr, old, new);
+	case 4:
+		return cmpxchg_386_u32(ptr, old, new);
+	}
+	return old;
+}
+
+#define cmpxchg(ptr,o,n)						\
+({									\
+	__typeof__(*(ptr)) __ret;					\
+	if (likely(boot_cpu_data.x86 > 3))				\
+		__ret = __cmpxchg((ptr), (unsigned long)(o),		\
+					(unsigned long)(n), sizeof(*(ptr))); \
+	else								\
+		__ret = cmpxchg_386((ptr), (unsigned long)(o),		\
+					(unsigned long)(n), sizeof(*(ptr))); \
+	__ret;								\
+})
+#endif
+
+#ifdef CONFIG_X86_CMPXCHG64
+
+static inline unsigned long long __cmpxchg64(volatile void *ptr, unsigned long long old,
+				      unsigned long long new)
+{
+	unsigned long long prev;
+	__asm__ __volatile__(LOCK_PREFIX "cmpxchg8b %3"
+			     : "=A"(prev)
+			     : "b"((unsigned long)new),
+			       "c"((unsigned long)(new >> 32)),
+			       "m"(*__xg(ptr)),
+			       "0"(old)
+			     : "memory");
+	return prev;
+}
+
+#define cmpxchg64(ptr,o,n)\
+	((__typeof__(*(ptr)))__cmpxchg64((ptr),(unsigned long long)(o),\
+					(unsigned long long)(n)))
+
+#endif
+    
+/*
+ * Force strict CPU ordering.
+ * And yes, this is required on UP too when we're talking
+ * to devices.
+ *
+ * For now, "wmb()" doesn't actually do anything, as all
+ * Intel CPU's follow what Intel calls a *Processor Order*,
+ * in which all writes are seen in the program order even
+ * outside the CPU.
+ *
+ * I expect future Intel CPU's to have a weaker ordering,
+ * but I'd also expect them to finally get their act together
+ * and add some real memory barriers if so.
+ *
+ * Some non intel clones support out of order store. wmb() ceases to be a
+ * nop for these.
+ */
+ 
+
+/* 
+ * Actually only lfence would be needed for mb() because all stores done 
+ * by the kernel should be already ordered. But keep a full barrier for now. 
+ */
+
+#define mb() alternative("lock; addl $0,0(%%esp)", "mfence", X86_FEATURE_XMM2)
+#define rmb() alternative("lock; addl $0,0(%%esp)", "lfence", X86_FEATURE_XMM2)
+
+/**
+ * read_barrier_depends - Flush all pending reads that subsequents reads
+ * depend on.
+ *
+ * No data-dependent reads from memory-like regions are ever reordered
+ * over this barrier.  All reads preceding this primitive are guaranteed
+ * to access memory (but not necessarily other CPUs' caches) before any
+ * reads following this primitive that depend on the data return by
+ * any of the preceding reads.  This primitive is much lighter weight than
+ * rmb() on most CPUs, and is never heavier weight than is
+ * rmb().
+ *
+ * These ordering constraints are respected by both the local CPU
+ * and the compiler.
+ *
+ * Ordering is not guaranteed by anything other than these primitives,
+ * not even by data dependencies.  See the documentation for
+ * memory_barrier() for examples and URLs to more information.
+ *
+ * For example, the following code would force ordering (the initial
+ * value of "a" is zero, "b" is one, and "p" is "&a"):
+ *
+ * <programlisting>
+ *	CPU 0				CPU 1
+ *
+ *	b = 2;
+ *	memory_barrier();
+ *	p = &b;				q = p;
+ *					read_barrier_depends();
+ *					d = *q;
+ * </programlisting>
+ *
+ * because the read of "*q" depends on the read of "p" and these
+ * two reads are separated by a read_barrier_depends().  However,
+ * the following code, with the same initial values for "a" and "b":
+ *
+ * <programlisting>
+ *	CPU 0				CPU 1
+ *
+ *	a = 2;
+ *	memory_barrier();
+ *	b = 3;				y = b;
+ *					read_barrier_depends();
+ *					x = a;
+ * </programlisting>
+ *
+ * does not enforce ordering, since there is no data dependency between
+ * the read of "a" and the read of "b".  Therefore, on some CPUs, such
+ * as Alpha, "y" could be set to 3 and "x" to 0.  Use rmb()
+ * in cases like this where there are no data dependencies.
+ **/
+
+#define read_barrier_depends()	do { } while(0)
+
+#ifdef CONFIG_X86_OOSTORE
+/* Actually there are no OOO store capable CPUs for now that do SSE, 
+   but make it already an possibility. */
+#define wmb() alternative("lock; addl $0,0(%%esp)", "sfence", X86_FEATURE_XMM)
+#else
+#define wmb()	__asm__ __volatile__ ("": : :"memory")
+#endif
+
+#ifdef CONFIG_SMP
+#define smp_mb()	mb()
+#define smp_rmb()	rmb()
+#define smp_wmb()	wmb()
+#define smp_read_barrier_depends()	read_barrier_depends()
+#define set_mb(var, value) do { (void) xchg(&var, value); } while (0)
+#else
+#define smp_mb()	barrier()
+#define smp_rmb()	barrier()
+#define smp_wmb()	barrier()
+#define smp_read_barrier_depends()	do { } while(0)
+#define set_mb(var, value) do { var = value; barrier(); } while (0)
+#endif
+
+#include <linux/irqflags.h>
+
+/*
+ * disable hlt during certain critical i/o operations
+ */
+#define HAVE_DISABLE_HLT
+void disable_hlt(void);
+void enable_hlt(void);
+
+extern int es7000_plat;
+void cpu_idle_wait(void);
+
+/*
+ * On SMP systems, when the scheduler does migration-cost autodetection,
+ * it needs a way to flush as much of the CPU's caches as possible:
+ */
+static inline void sched_cacheflush(void)
+{
+	wbinvd();
+}
+
+extern unsigned long arch_align_stack(unsigned long sp);
+extern void free_init_pages(char *what, unsigned long begin, unsigned long end);
+
+void default_idle(void);
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/tlbflush.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/tlbflush.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/tlbflush.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/tlbflush.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,101 @@
+#ifndef _I386_TLBFLUSH_H
+#define _I386_TLBFLUSH_H
+
+#include <linux/mm.h>
+#include <asm/processor.h>
+
+#define __flush_tlb() xen_tlb_flush()
+#define __flush_tlb_global() xen_tlb_flush()
+#define __flush_tlb_all() xen_tlb_flush()
+
+extern unsigned long pgkern_mask;
+
+#define cpu_has_invlpg	(boot_cpu_data.x86 > 3)
+
+#define __flush_tlb_single(addr) xen_invlpg(addr)
+
+#define __flush_tlb_one(addr) __flush_tlb_single(addr)
+
+/*
+ * TLB flushing:
+ *
+ *  - flush_tlb() flushes the current mm struct TLBs
+ *  - flush_tlb_all() flushes all processes TLBs
+ *  - flush_tlb_mm(mm) flushes the specified mm context TLB's
+ *  - flush_tlb_page(vma, vmaddr) flushes one page
+ *  - flush_tlb_range(vma, start, end) flushes a range of pages
+ *  - flush_tlb_kernel_range(start, end) flushes a range of kernel pages
+ *  - flush_tlb_pgtables(mm, start, end) flushes a range of page tables
+ *
+ * ..but the i386 has somewhat limited tlb flushing capabilities,
+ * and page-granular flushes are available only on i486 and up.
+ */
+
+#ifndef CONFIG_SMP
+
+#define flush_tlb() __flush_tlb()
+#define flush_tlb_all() __flush_tlb_all()
+#define local_flush_tlb() __flush_tlb()
+
+static inline void flush_tlb_mm(struct mm_struct *mm)
+{
+	if (mm == current->active_mm)
+		__flush_tlb();
+}
+
+static inline void flush_tlb_page(struct vm_area_struct *vma,
+	unsigned long addr)
+{
+	if (vma->vm_mm == current->active_mm)
+		__flush_tlb_one(addr);
+}
+
+static inline void flush_tlb_range(struct vm_area_struct *vma,
+	unsigned long start, unsigned long end)
+{
+	if (vma->vm_mm == current->active_mm)
+		__flush_tlb();
+}
+
+#else
+
+#include <asm/smp.h>
+
+#define local_flush_tlb() \
+	__flush_tlb()
+
+#define flush_tlb_all xen_tlb_flush_all
+#define flush_tlb_current_task() xen_tlb_flush_mask(&current->mm->cpu_vm_mask)
+#define flush_tlb_mm(mm) xen_tlb_flush_mask(&(mm)->cpu_vm_mask)
+#define flush_tlb_page(vma, va) xen_invlpg_mask(&(vma)->vm_mm->cpu_vm_mask, va)
+
+#define flush_tlb()	flush_tlb_current_task()
+
+static inline void flush_tlb_range(struct vm_area_struct * vma, unsigned long start, unsigned long end)
+{
+	flush_tlb_mm(vma->vm_mm);
+}
+
+#define TLBSTATE_OK	1
+#define TLBSTATE_LAZY	2
+
+struct tlb_state
+{
+	struct mm_struct *active_mm;
+	int state;
+	char __cacheline_padding[L1_CACHE_BYTES-8];
+};
+DECLARE_PER_CPU(struct tlb_state, cpu_tlbstate);
+
+
+#endif
+
+#define flush_tlb_kernel_range(start, end) flush_tlb_all()
+
+static inline void flush_tlb_pgtables(struct mm_struct *mm,
+				      unsigned long start, unsigned long end)
+{
+	/* i386 does not keep any page table caches in TLB */
+}
+
+#endif /* _I386_TLBFLUSH_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/vga.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/vga.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/vga.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/vga.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,20 @@
+/*
+ *	Access to VGA videoram
+ *
+ *	(c) 1998 Martin Mares <mj@ucw.cz>
+ */
+
+#ifndef _LINUX_ASM_VGA_H_
+#define _LINUX_ASM_VGA_H_
+
+/*
+ *	On the PC, we can just recalculate addresses and then
+ *	access the videoram directly without any black magic.
+ */
+
+#define VGA_MAP_MEM(x,s) (unsigned long)isa_bus_to_virt(x)
+
+#define vga_readb(x) (*(x))
+#define vga_writeb(x,y) (*(y) = (x))
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/asm/xenoprof.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/xenoprof.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/asm/xenoprof.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/asm/xenoprof.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,48 @@
+/******************************************************************************
+ * asm-i386/mach-xen/asm/xenoprof.h
+ *
+ * Copyright (c) 2006 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+#ifndef __ASM_XENOPROF_H__
+#define __ASM_XENOPROF_H__
+#ifdef CONFIG_XEN
+
+struct super_block;
+struct dentry;
+int xenoprof_create_files(struct super_block * sb, struct dentry * root);
+#define HAVE_XENOPROF_CREATE_FILES
+
+struct xenoprof_init;
+void xenoprof_arch_init_counter(struct xenoprof_init *init);
+void xenoprof_arch_counter(void);
+void xenoprof_arch_start(void);
+void xenoprof_arch_stop(void);
+
+struct xenoprof_arch_shared_buffer {
+	/* nothing */
+};
+struct xenoprof_shared_buffer;
+void xenoprof_arch_unmap_shared_buffer(struct xenoprof_shared_buffer* sbuf);
+struct xenoprof_get_buffer;
+int xenoprof_arch_map_shared_buffer(struct xenoprof_get_buffer* get_buffer, struct xenoprof_shared_buffer* sbuf);
+struct xenoprof_passive;
+int xenoprof_arch_set_passive(struct xenoprof_passive* pdomain, struct xenoprof_shared_buffer* sbuf);
+
+#endif /* CONFIG_XEN */
+#endif /* __ASM_XENOPROF_H__ */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/irq_vectors.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/irq_vectors.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/irq_vectors.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/irq_vectors.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,125 @@
+/*
+ * This file should contain #defines for all of the interrupt vector
+ * numbers used by this architecture.
+ *
+ * In addition, there are some standard defines:
+ *
+ *	FIRST_EXTERNAL_VECTOR:
+ *		The first free place for external interrupts
+ *
+ *	SYSCALL_VECTOR:
+ *		The IRQ vector a syscall makes the user to kernel transition
+ *		under.
+ *
+ *	TIMER_IRQ:
+ *		The IRQ number the timer interrupt comes in at.
+ *
+ *	NR_IRQS:
+ *		The total number of interrupt vectors (including all the
+ *		architecture specific interrupts) needed.
+ *
+ */			
+#ifndef _ASM_IRQ_VECTORS_H
+#define _ASM_IRQ_VECTORS_H
+
+/*
+ * IDT vectors usable for external interrupt sources start
+ * at 0x20:
+ */
+#define FIRST_EXTERNAL_VECTOR	0x20
+
+#define SYSCALL_VECTOR		0x80
+
+/*
+ * Vectors 0x20-0x2f are used for ISA interrupts.
+ */
+
+#if 0
+/*
+ * Special IRQ vectors used by the SMP architecture, 0xf0-0xff
+ *
+ *  some of the following vectors are 'rare', they are merged
+ *  into a single vector (CALL_FUNCTION_VECTOR) to save vector space.
+ *  TLB, reschedule and local APIC vectors are performance-critical.
+ *
+ *  Vectors 0xf0-0xfa are free (reserved for future Linux use).
+ */
+#define SPURIOUS_APIC_VECTOR	0xff
+#define ERROR_APIC_VECTOR	0xfe
+#define INVALIDATE_TLB_VECTOR	0xfd
+#define RESCHEDULE_VECTOR	0xfc
+#define CALL_FUNCTION_VECTOR	0xfb
+
+#define THERMAL_APIC_VECTOR	0xf0
+/*
+ * Local APIC timer IRQ vector is on a different priority level,
+ * to work around the 'lost local interrupt if more than 2 IRQ
+ * sources per level' errata.
+ */
+#define LOCAL_TIMER_VECTOR	0xef
+#endif
+
+#define SPURIOUS_APIC_VECTOR	0xff
+#define ERROR_APIC_VECTOR	0xfe
+
+/*
+ * First APIC vector available to drivers: (vectors 0x30-0xee)
+ * we start at 0x31 to spread out vectors evenly between priority
+ * levels. (0x80 is the syscall vector)
+ */
+#define FIRST_DEVICE_VECTOR	0x31
+#define FIRST_SYSTEM_VECTOR	0xef
+
+/*
+ * 16 8259A IRQ's, 208 potential APIC interrupt sources.
+ * Right now the APIC is mostly only used for SMP.
+ * 256 vectors is an architectural limit. (we can have
+ * more than 256 devices theoretically, but they will
+ * have to use shared interrupts)
+ * Since vectors 0x00-0x1f are used/reserved for the CPU,
+ * the usable vector space is 0x20-0xff (224 vectors)
+ */
+
+#define RESCHEDULE_VECTOR	0
+#define CALL_FUNCTION_VECTOR	1
+#define NR_IPIS			2
+
+/*
+ * The maximum number of vectors supported by i386 processors
+ * is limited to 256. For processors other than i386, NR_VECTORS
+ * should be changed accordingly.
+ */
+#define NR_VECTORS 256
+
+#define FPU_IRQ			13
+
+#define	FIRST_VM86_IRQ		3
+#define LAST_VM86_IRQ		15
+#define invalid_vm86_irq(irq)	((irq) < 3 || (irq) > 15)
+
+/*
+ * The flat IRQ space is divided into two regions:
+ *  1. A one-to-one mapping of real physical IRQs. This space is only used
+ *     if we have physical device-access privilege. This region is at the 
+ *     start of the IRQ space so that existing device drivers do not need
+ *     to be modified to translate physical IRQ numbers into our IRQ space.
+ *  3. A dynamic mapping of inter-domain and Xen-sourced virtual IRQs. These
+ *     are bound using the provided bind/unbind functions.
+ */
+
+#define PIRQ_BASE		0
+#define NR_PIRQS		256
+
+#define DYNIRQ_BASE		(PIRQ_BASE + NR_PIRQS)
+#define NR_DYNIRQS		256
+
+#define NR_IRQS			(NR_PIRQS + NR_DYNIRQS)
+#define NR_IRQ_VECTORS		NR_IRQS
+
+#define pirq_to_irq(_x)		((_x) + PIRQ_BASE)
+#define irq_to_pirq(_x)		((_x) - PIRQ_BASE)
+
+#define dynirq_to_irq(_x)	((_x) + DYNIRQ_BASE)
+#define irq_to_dynirq(_x)	((_x) - DYNIRQ_BASE)
+
+#endif /* _ASM_IRQ_VECTORS_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/mach_traps.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/mach_traps.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/mach_traps.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/mach_traps.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,33 @@
+/*
+ *  include/asm-xen/asm-i386/mach-xen/mach_traps.h
+ *
+ *  Machine specific NMI handling for Xen
+ */
+#ifndef _MACH_TRAPS_H
+#define _MACH_TRAPS_H
+
+#include <linux/bitops.h>
+#include <xen/interface/nmi.h>
+
+static inline void clear_mem_error(unsigned char reason) {}
+static inline void clear_io_check_error(unsigned char reason) {}
+
+static inline unsigned char get_nmi_reason(void)
+{
+	shared_info_t *s = HYPERVISOR_shared_info;
+	unsigned char reason = 0;
+
+	/* construct a value which looks like it came from
+	 * port 0x61.
+	 */
+	if (test_bit(_XEN_NMIREASON_io_error, &s->arch.nmi_reason))
+		reason |= 0x40;
+	if (test_bit(_XEN_NMIREASON_parity_error, &s->arch.nmi_reason))
+		reason |= 0x80;
+
+        return reason;
+}
+
+static inline void reassert_nmi(void) {}
+
+#endif /* !_MACH_TRAPS_H */
diff -rpuN linux-2.6.18.8/include/asm-i386/mach-xen/setup_arch.h linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/setup_arch.h
--- linux-2.6.18.8/include/asm-i386/mach-xen/setup_arch.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/mach-xen/setup_arch.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,5 @@
+/* Hook to call BIOS initialisation function */
+
+#define ARCH_SETUP machine_specific_arch_setup();
+
+void __init machine_specific_arch_setup(void);
diff -rpuN linux-2.6.18.8/include/asm-i386/page.h linux-2.6.18-xen-3.3.0/include/asm-i386/page.h
--- linux-2.6.18.8/include/asm-i386/page.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/page.h	2008-08-21 11:36:07.000000000 +0200
@@ -122,7 +122,7 @@ extern int page_is_ram(unsigned long pag
 
 #define PAGE_OFFSET		((unsigned long)__PAGE_OFFSET)
 #define VMALLOC_RESERVE		((unsigned long)__VMALLOC_RESERVE)
-#define MAXMEM			(-__PAGE_OFFSET-__VMALLOC_RESERVE)
+#define MAXMEM			(__FIXADDR_TOP-__PAGE_OFFSET-__VMALLOC_RESERVE)
 #define __pa(x)			((unsigned long)(x)-PAGE_OFFSET)
 #define __va(x)			((void *)((unsigned long)(x)+PAGE_OFFSET))
 #define pfn_to_kaddr(pfn)      __va((pfn) << PAGE_SHIFT)
diff -rpuN linux-2.6.18.8/include/asm-i386/pgtable-2level-defs.h linux-2.6.18-xen-3.3.0/include/asm-i386/pgtable-2level-defs.h
--- linux-2.6.18.8/include/asm-i386/pgtable-2level-defs.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/pgtable-2level-defs.h	2008-08-21 11:36:07.000000000 +0200
@@ -1,6 +1,8 @@
 #ifndef _I386_PGTABLE_2LEVEL_DEFS_H
 #define _I386_PGTABLE_2LEVEL_DEFS_H
 
+#define HAVE_SHARED_KERNEL_PMD 0
+
 /*
  * traditional i386 two-level paging structure:
  */
diff -rpuN linux-2.6.18.8/include/asm-i386/pgtable-3level-defs.h linux-2.6.18-xen-3.3.0/include/asm-i386/pgtable-3level-defs.h
--- linux-2.6.18.8/include/asm-i386/pgtable-3level-defs.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-i386/pgtable-3level-defs.h	2008-08-21 11:36:07.000000000 +0200
@@ -1,6 +1,8 @@
 #ifndef _I386_PGTABLE_3LEVEL_DEFS_H
 #define _I386_PGTABLE_3LEVEL_DEFS_H
 
+#define HAVE_SHARED_KERNEL_PMD 1
+
 /*
  * PGDIR_SHIFT determines what a top-level page table entry can map
  */
diff -rpuN linux-2.6.18.8/include/asm-ia64/agp.h linux-2.6.18-xen-3.3.0/include/asm-ia64/agp.h
--- linux-2.6.18.8/include/asm-ia64/agp.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/agp.h	2008-08-21 11:36:07.000000000 +0200
@@ -19,13 +19,44 @@
 #define flush_agp_cache()		mb()
 
 /* Convert a physical address to an address suitable for the GART. */
+#ifndef CONFIG_XEN
 #define phys_to_gart(x) (x)
 #define gart_to_phys(x) (x)
+#else
+#define phys_to_gart(x) phys_to_machine_for_dma(x)
+#define gart_to_phys(x) machine_to_phys_for_dma(x)
+#endif
 
 /* GATT allocation. Returns/accepts GATT kernel virtual address. */
+#ifndef CONFIG_XEN
 #define alloc_gatt_pages(order)		\
 	((char *)__get_free_pages(GFP_KERNEL, (order)))
 #define free_gatt_pages(table, order)	\
 	free_pages((unsigned long)(table), (order))
+#else
+#include <asm/hypervisor.h>
+static inline char*
+alloc_gatt_pages(unsigned int order)
+{
+	unsigned long error;
+	unsigned long ret = __get_free_pages(GFP_KERNEL, (order));
+	if (ret == 0) {
+		goto out;
+	}
+	error = xen_create_contiguous_region(ret, order, 0);
+	if (error) {
+		free_pages(ret, order);
+		ret = 0;
+	}
+out:
+	return (char*)ret;
+}
+static inline void
+free_gatt_pages(void* table, unsigned int order)
+{
+	xen_destroy_contiguous_region((unsigned long)table, order);
+	free_pages((unsigned long)table, order);
+}
+#endif /* CONFIG_XEN */
 
 #endif /* _ASM_IA64_AGP_H */
diff -rpuN linux-2.6.18.8/include/asm-ia64/gcc_intrin.h linux-2.6.18-xen-3.3.0/include/asm-ia64/gcc_intrin.h
--- linux-2.6.18.8/include/asm-ia64/gcc_intrin.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/gcc_intrin.h	2008-08-21 11:36:07.000000000 +0200
@@ -26,7 +26,7 @@ extern void ia64_bad_param_for_getreg (v
 
 register unsigned long ia64_r13 asm ("r13") __attribute_used__;
 
-#define ia64_setreg(regnum, val)						\
+#define __ia64_setreg(regnum, val)						\
 ({										\
 	switch (regnum) {							\
 	    case _IA64_REG_PSR_L:						\
@@ -55,7 +55,7 @@ register unsigned long ia64_r13 asm ("r1
 	}									\
 })
 
-#define ia64_getreg(regnum)							\
+#define __ia64_getreg(regnum)							\
 ({										\
 	__u64 ia64_intri_res;							\
 										\
@@ -92,7 +92,7 @@ register unsigned long ia64_r13 asm ("r1
 
 #define ia64_hint_pause 0
 
-#define ia64_hint(mode)						\
+#define __ia64_hint(mode)						\
 ({								\
 	switch (mode) {						\
 	case ia64_hint_pause:					\
@@ -374,7 +374,7 @@ register unsigned long ia64_r13 asm ("r1
 
 #define ia64_invala() asm volatile ("invala" ::: "memory")
 
-#define ia64_thash(addr)							\
+#define __ia64_thash(addr)							\
 ({										\
 	__u64 ia64_intri_res;							\
 	asm volatile ("thash %0=%1" : "=r"(ia64_intri_res) : "r" (addr));	\
@@ -394,18 +394,18 @@ register unsigned long ia64_r13 asm ("r1
 
 #define ia64_nop(x)	asm volatile ("nop %0"::"i"(x));
 
-#define ia64_itci(addr)	asm volatile ("itc.i %0;;" :: "r"(addr) : "memory")
+#define __ia64_itci(addr)	asm volatile ("itc.i %0;;" :: "r"(addr) : "memory")
 
-#define ia64_itcd(addr)	asm volatile ("itc.d %0;;" :: "r"(addr) : "memory")
+#define __ia64_itcd(addr)	asm volatile ("itc.d %0;;" :: "r"(addr) : "memory")
 
 
-#define ia64_itri(trnum, addr) asm volatile ("itr.i itr[%0]=%1"				\
+#define __ia64_itri(trnum, addr) asm volatile ("itr.i itr[%0]=%1"			\
 					     :: "r"(trnum), "r"(addr) : "memory")
 
-#define ia64_itrd(trnum, addr) asm volatile ("itr.d dtr[%0]=%1"				\
+#define __ia64_itrd(trnum, addr) asm volatile ("itr.d dtr[%0]=%1"			\
 					     :: "r"(trnum), "r"(addr) : "memory")
 
-#define ia64_tpa(addr)								\
+#define __ia64_tpa(addr)							\
 ({										\
 	__u64 ia64_pa;								\
 	asm volatile ("tpa %0 = %1" : "=r"(ia64_pa) : "r"(addr) : "memory");	\
@@ -415,22 +415,22 @@ register unsigned long ia64_r13 asm ("r1
 #define __ia64_set_dbr(index, val)						\
 	asm volatile ("mov dbr[%0]=%1" :: "r"(index), "r"(val) : "memory")
 
-#define ia64_set_ibr(index, val)						\
+#define __ia64_set_ibr(index, val)						\
 	asm volatile ("mov ibr[%0]=%1" :: "r"(index), "r"(val) : "memory")
 
-#define ia64_set_pkr(index, val)						\
+#define __ia64_set_pkr(index, val)						\
 	asm volatile ("mov pkr[%0]=%1" :: "r"(index), "r"(val) : "memory")
 
-#define ia64_set_pmc(index, val)						\
+#define __ia64_set_pmc(index, val)						\
 	asm volatile ("mov pmc[%0]=%1" :: "r"(index), "r"(val) : "memory")
 
-#define ia64_set_pmd(index, val)						\
+#define __ia64_set_pmd(index, val)						\
 	asm volatile ("mov pmd[%0]=%1" :: "r"(index), "r"(val) : "memory")
 
-#define ia64_set_rr(index, val)							\
+#define __ia64_set_rr(index, val)							\
 	asm volatile ("mov rr[%0]=%1" :: "r"(index), "r"(val) : "memory");
 
-#define ia64_get_cpuid(index)								\
+#define __ia64_get_cpuid(index)								\
 ({											\
 	__u64 ia64_intri_res;								\
 	asm volatile ("mov %0=cpuid[%r1]" : "=r"(ia64_intri_res) : "rO"(index));	\
@@ -444,21 +444,21 @@ register unsigned long ia64_r13 asm ("r1
 	ia64_intri_res;								\
 })
 
-#define ia64_get_ibr(index)							\
+#define __ia64_get_ibr(index)							\
 ({										\
 	__u64 ia64_intri_res;							\
 	asm volatile ("mov %0=ibr[%1]" : "=r"(ia64_intri_res) : "r"(index));	\
 	ia64_intri_res;								\
 })
 
-#define ia64_get_pkr(index)							\
+#define __ia64_get_pkr(index)							\
 ({										\
 	__u64 ia64_intri_res;							\
 	asm volatile ("mov %0=pkr[%1]" : "=r"(ia64_intri_res) : "r"(index));	\
 	ia64_intri_res;								\
 })
 
-#define ia64_get_pmc(index)							\
+#define __ia64_get_pmc(index)							\
 ({										\
 	__u64 ia64_intri_res;							\
 	asm volatile ("mov %0=pmc[%1]" : "=r"(ia64_intri_res) : "r"(index));	\
@@ -466,48 +466,48 @@ register unsigned long ia64_r13 asm ("r1
 })
 
 
-#define ia64_get_pmd(index)							\
+#define __ia64_get_pmd(index)							\
 ({										\
 	__u64 ia64_intri_res;							\
 	asm volatile ("mov %0=pmd[%1]" : "=r"(ia64_intri_res) : "r"(index));	\
 	ia64_intri_res;								\
 })
 
-#define ia64_get_rr(index)							\
+#define __ia64_get_rr(index)							\
 ({										\
 	__u64 ia64_intri_res;							\
 	asm volatile ("mov %0=rr[%1]" : "=r"(ia64_intri_res) : "r" (index));	\
 	ia64_intri_res;								\
 })
 
-#define ia64_fc(addr)	asm volatile ("fc %0" :: "r"(addr) : "memory")
+#define __ia64_fc(addr)	asm volatile ("fc %0" :: "r"(addr) : "memory")
 
 
 #define ia64_sync_i()	asm volatile (";; sync.i" ::: "memory")
 
-#define ia64_ssm(mask)	asm volatile ("ssm %0":: "i"((mask)) : "memory")
-#define ia64_rsm(mask)	asm volatile ("rsm %0":: "i"((mask)) : "memory")
+#define __ia64_ssm(mask)	asm volatile ("ssm %0":: "i"((mask)) : "memory")
+#define __ia64_rsm(mask)	asm volatile ("rsm %0":: "i"((mask)) : "memory")
 #define ia64_sum(mask)	asm volatile ("sum %0":: "i"((mask)) : "memory")
 #define ia64_rum(mask)	asm volatile ("rum %0":: "i"((mask)) : "memory")
 
-#define ia64_ptce(addr)	asm volatile ("ptc.e %0" :: "r"(addr))
+#define __ia64_ptce(addr)	asm volatile ("ptc.e %0" :: "r"(addr))
 
-#define ia64_ptcga(addr, size)							\
+#define __ia64_ptcga(addr, size)							\
 do {										\
 	asm volatile ("ptc.ga %0,%1" :: "r"(addr), "r"(size) : "memory");	\
 	ia64_dv_serialize_data();						\
 } while (0)
 
-#define ia64_ptcl(addr, size)							\
+#define __ia64_ptcl(addr, size)							\
 do {										\
 	asm volatile ("ptc.l %0,%1" :: "r"(addr), "r"(size) : "memory");	\
 	ia64_dv_serialize_data();						\
 } while (0)
 
-#define ia64_ptri(addr, size)						\
+#define __ia64_ptri(addr, size)						\
 	asm volatile ("ptr.i %0,%1" :: "r"(addr), "r"(size) : "memory")
 
-#define ia64_ptrd(addr, size)						\
+#define __ia64_ptrd(addr, size)						\
 	asm volatile ("ptr.d %0,%1" :: "r"(addr), "r"(size) : "memory")
 
 /* Values for lfhint in ia64_lfetch and ia64_lfetch_fault */
@@ -589,7 +589,7 @@ do {										\
         }								\
 })
 
-#define ia64_intrin_local_irq_restore(x)			\
+#define __ia64_intrin_local_irq_restore(x)			\
 do {								\
 	asm volatile (";;   cmp.ne p6,p7=%0,r0;;"		\
 		      "(p6) ssm psr.i;"				\
@@ -598,4 +598,6 @@ do {								\
 		      :: "r"((x)) : "p6", "p7", "memory");	\
 } while (0)
 
+#define __ia64_get_psr_i()	(__ia64_getreg(_IA64_REG_PSR) & 0x4000UL)
+
 #endif /* _ASM_IA64_GCC_INTRIN_H */
diff -rpuN linux-2.6.18.8/include/asm-ia64/gnttab_dma.h linux-2.6.18-xen-3.3.0/include/asm-ia64/gnttab_dma.h
--- linux-2.6.18.8/include/asm-ia64/gnttab_dma.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/gnttab_dma.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,51 @@
+/*
+ * Copyright (c) 2007 Herbert Xu <herbert@gondor.apana.org.au>
+ * Copyright (c) 2007 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#ifndef _ASM_IA64_GNTTAB_DMA_H
+#define _ASM_IA64_GNTTAB_DMA_H
+
+static inline int gnttab_dma_local_pfn(struct page *page)
+{
+	return 0;
+}
+
+/* caller must get dma address after calling this function */
+static inline void gnttab_dma_use_page(struct page *page)
+{
+	__gnttab_dma_map_page(page);
+}
+
+static inline dma_addr_t gnttab_dma_map_page(struct page *page)
+{
+	gnttab_dma_use_page(page);
+	return page_to_bus(page);
+}
+
+static inline dma_addr_t gnttab_dma_map_virt(void *ptr)
+{
+	return gnttab_dma_map_page(virt_to_page(ptr)) + offset_in_page(ptr);
+}
+
+static inline void gnttab_dma_unmap_page(dma_addr_t dma_address)
+{
+	__gnttab_dma_unmap_page(virt_to_page(bus_to_virt(dma_address)));
+}
+
+#endif /* _ASM_IA64_GNTTAB_DMA_H */
diff -rpuN linux-2.6.18.8/include/asm-ia64/hw_irq.h linux-2.6.18-xen-3.3.0/include/asm-ia64/hw_irq.h
--- linux-2.6.18.8/include/asm-ia64/hw_irq.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/hw_irq.h	2008-08-21 11:36:07.000000000 +0200
@@ -15,7 +15,11 @@
 #include <asm/ptrace.h>
 #include <asm/smp.h>
 
+#ifndef CONFIG_XEN
 typedef u8 ia64_vector;
+#else
+typedef u16 ia64_vector;
+#endif
 
 /*
  * 0 special
@@ -99,6 +103,12 @@ extern void register_percpu_irq (ia64_ve
 
 static inline void ia64_resend_irq(unsigned int vector)
 {
+#ifdef CONFIG_XEN
+	extern int resend_irq_on_evtchn(unsigned int i);
+	if (is_running_on_xen())
+		resend_irq_on_evtchn(vector);
+	else
+#endif /* CONFIG_XEN */
 	platform_send_ipi(smp_processor_id(), vector, IA64_IPI_DM_INT, 0);
 }
 
diff -rpuN linux-2.6.18.8/include/asm-ia64/hypercall.h linux-2.6.18-xen-3.3.0/include/asm-ia64/hypercall.h
--- linux-2.6.18.8/include/asm-ia64/hypercall.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/hypercall.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,458 @@
+/******************************************************************************
+ * hypercall.h
+ * 
+ * Linux-specific hypervisor handling.
+ * 
+ * Copyright (c) 2002-2004, K A Fraser
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __HYPERCALL_H__
+#define __HYPERCALL_H__
+
+#ifndef __HYPERVISOR_H__
+# error "please don't include this file directly"
+#endif
+
+#include <asm/xen/xcom_hcall.h>
+struct xencomm_handle;
+extern unsigned long __hypercall(unsigned long a1, unsigned long a2,
+                                 unsigned long a3, unsigned long a4,
+                                 unsigned long a5, unsigned long cmd);
+
+/*
+ * Assembler stubs for hyper-calls.
+ */
+
+#define _hypercall0(type, name)					\
+({								\
+	long __res;						\
+	__res=__hypercall(0, 0, 0, 0, 0, __HYPERVISOR_##name);	\
+	(type)__res;						\
+})
+
+#define _hypercall1(type, name, a1)				\
+({								\
+	long __res;						\
+	__res = __hypercall((unsigned long)a1,			\
+	                     0, 0, 0, 0, __HYPERVISOR_##name);	\
+	(type)__res;						\
+})
+
+#define _hypercall2(type, name, a1, a2)				\
+({								\
+	long __res;						\
+	__res = __hypercall((unsigned long)a1,			\
+	                    (unsigned long)a2,			\
+	                    0, 0, 0, __HYPERVISOR_##name);	\
+	(type)__res;						\
+})
+
+#define _hypercall3(type, name, a1, a2, a3)			\
+({								\
+	long __res;						\
+	__res = __hypercall((unsigned long)a1,			\
+	                    (unsigned long)a2,			\
+	                    (unsigned long)a3,			\
+	                    0, 0, __HYPERVISOR_##name);		\
+	(type)__res;						\
+})
+
+#define _hypercall4(type, name, a1, a2, a3, a4)			\
+({								\
+	long __res;						\
+	__res = __hypercall((unsigned long)a1,			\
+	                    (unsigned long)a2,			\
+	                    (unsigned long)a3,			\
+	                    (unsigned long)a4,			\
+	                    0, __HYPERVISOR_##name);		\
+	(type)__res;						\
+})
+
+#define _hypercall5(type, name, a1, a2, a3, a4, a5)		\
+({								\
+	long __res;						\
+	__res = __hypercall((unsigned long)a1,			\
+	                    (unsigned long)a2,			\
+	                    (unsigned long)a3,			\
+	                    (unsigned long)a4,			\
+	                    (unsigned long)a5,			\
+	                    __HYPERVISOR_##name);		\
+	(type)__res;						\
+})
+
+
+static inline int
+xencomm_arch_hypercall_sched_op(int cmd, struct xencomm_handle *arg)
+{
+	return _hypercall2(int, sched_op, cmd, arg);
+}
+
+static inline long
+HYPERVISOR_set_timer_op(u64 timeout)
+{
+	unsigned long timeout_hi = (unsigned long)(timeout >> 32);
+	unsigned long timeout_lo = (unsigned long)timeout;
+	return _hypercall2(long, set_timer_op, timeout_lo, timeout_hi);
+}
+
+static inline int
+xencomm_arch_hypercall_platform_op(struct xencomm_handle *op)
+{
+	return _hypercall1(int, platform_op, op);
+}
+
+static inline int
+xencomm_arch_hypercall_sysctl(struct xencomm_handle *op)
+{
+	return _hypercall1(int, sysctl, op);
+}
+
+static inline int
+xencomm_arch_hypercall_domctl(struct xencomm_handle *op)
+{
+	return _hypercall1(int, domctl, op);
+}
+
+static inline int
+xencomm_arch_hypercall_multicall(struct xencomm_handle *call_list,
+				 int nr_calls)
+{
+	return _hypercall2(int, multicall, call_list, nr_calls);
+}
+
+static inline int
+xencomm_arch_hypercall_memory_op(unsigned int cmd, struct xencomm_handle *arg)
+{
+	return _hypercall2(int, memory_op, cmd, arg);
+}
+
+static inline int
+xencomm_arch_hypercall_event_channel_op(int cmd, struct xencomm_handle *arg)
+{
+	return _hypercall2(int, event_channel_op, cmd, arg);
+}
+
+static inline int
+xencomm_arch_hypercall_xsm_op(struct xencomm_handle *arg)
+{
+	return _hypercall1(int, xsm_op, arg);
+}
+
+static inline int
+xencomm_arch_hypercall_xen_version(int cmd, struct xencomm_handle *arg)
+{
+	return _hypercall2(int, xen_version, cmd, arg);
+}
+
+static inline int
+xencomm_arch_hypercall_console_io(int cmd, int count,
+                                  struct xencomm_handle *str)
+{
+	return _hypercall3(int, console_io, cmd, count, str);
+}
+
+static inline int
+xencomm_arch_hypercall_physdev_op(int cmd, struct xencomm_handle *arg)
+{
+	return _hypercall2(int, physdev_op, cmd, arg);
+}
+
+static inline int
+xencomm_arch_hypercall_grant_table_op(unsigned int cmd,
+                                      struct xencomm_handle *uop,
+                                      unsigned int count)
+{
+	return _hypercall3(int, grant_table_op, cmd, uop, count);
+}
+
+int HYPERVISOR_grant_table_op(unsigned int cmd, void *uop, unsigned int count);
+
+extern int xencomm_arch_hypercall_suspend(struct xencomm_handle *arg);
+
+static inline int
+xencomm_arch_hypercall_callback_op(int cmd, struct xencomm_handle *arg)
+{
+	return _hypercall2(int, callback_op, cmd, arg);
+}
+
+static inline unsigned long
+xencomm_arch_hypercall_hvm_op(int cmd, void *arg)
+{
+	return _hypercall2(unsigned long, hvm_op, cmd, arg);
+}
+
+static inline long
+xencomm_arch_hypercall_vcpu_op(int cmd, int cpu, void *arg)
+{
+	return _hypercall3(long, vcpu_op, cmd, cpu, arg);
+}
+
+static inline int
+HYPERVISOR_physdev_op(int cmd, void *arg)
+{
+	switch (cmd) {
+	case PHYSDEVOP_eoi:
+		return _hypercall1(int, ia64_fast_eoi,
+		                   ((struct physdev_eoi *)arg)->irq);
+	default:
+		return xencomm_hypercall_physdev_op(cmd, arg);
+	}
+}
+
+static inline int
+xencomm_arch_hypercall_xenoprof_op(int op, struct xencomm_handle *arg)
+{
+	return _hypercall2(int, xenoprof_op, op, arg);
+}
+
+static inline long
+xencomm_arch_hypercall_opt_feature(struct xencomm_handle *arg)
+{
+	return _hypercall1(long, opt_feature, arg);
+}
+
+extern fastcall unsigned int __do_IRQ(unsigned int irq, struct pt_regs *regs);
+static inline void exit_idle(void) {}
+#define do_IRQ(irq, regs) ({			\
+	irq_enter();				\
+	__do_IRQ((irq), (regs));		\
+	irq_exit();				\
+})
+
+#include <linux/err.h>
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
+
+static inline unsigned long
+__HYPERVISOR_ioremap(unsigned long ioaddr, unsigned long size)
+{
+	return _hypercall3(unsigned long, ia64_dom0vp_op,
+	                   IA64_DOM0VP_ioremap, ioaddr, size);
+}
+
+static inline unsigned long
+HYPERVISOR_ioremap(unsigned long ioaddr, unsigned long size)
+{
+	unsigned long ret = ioaddr;
+	if (is_running_on_xen()) {
+		ret = __HYPERVISOR_ioremap(ioaddr, size);
+		if (unlikely(ret == -ENOSYS))
+			panic("hypercall %s failed with %ld. "
+			      "Please check Xen and Linux config mismatch\n",
+			      __func__, -ret);
+		else if (unlikely(IS_ERR_VALUE(ret)))
+			ret = ioaddr;
+	}
+	return ret;
+}
+
+static inline unsigned long
+__HYPERVISOR_phystomach(unsigned long gpfn)
+{
+	return _hypercall2(unsigned long, ia64_dom0vp_op,
+	                   IA64_DOM0VP_phystomach, gpfn);
+}
+
+static inline unsigned long
+HYPERVISOR_phystomach(unsigned long gpfn)
+{
+	unsigned long ret = gpfn;
+	if (is_running_on_xen()) {
+		ret = __HYPERVISOR_phystomach(gpfn);
+	}
+	return ret;
+}
+
+static inline unsigned long
+__HYPERVISOR_machtophys(unsigned long mfn)
+{
+	return _hypercall2(unsigned long, ia64_dom0vp_op,
+	                   IA64_DOM0VP_machtophys, mfn);
+}
+
+static inline unsigned long
+HYPERVISOR_machtophys(unsigned long mfn)
+{
+	unsigned long ret = mfn;
+	if (is_running_on_xen()) {
+		ret = __HYPERVISOR_machtophys(mfn);
+	}
+	return ret;
+}
+
+static inline unsigned long
+__HYPERVISOR_zap_physmap(unsigned long gpfn, unsigned int extent_order)
+{
+	return _hypercall3(unsigned long, ia64_dom0vp_op,
+	                   IA64_DOM0VP_zap_physmap, gpfn, extent_order);
+}
+
+static inline unsigned long
+HYPERVISOR_zap_physmap(unsigned long gpfn, unsigned int extent_order)
+{
+	unsigned long ret = 0;
+	if (is_running_on_xen()) {
+		ret = __HYPERVISOR_zap_physmap(gpfn, extent_order);
+	}
+	return ret;
+}
+
+static inline unsigned long
+__HYPERVISOR_add_physmap(unsigned long gpfn, unsigned long mfn,
+			 unsigned long flags, domid_t domid)
+{
+	return _hypercall5(unsigned long, ia64_dom0vp_op,
+	                   IA64_DOM0VP_add_physmap, gpfn, mfn, flags, domid);
+}
+
+static inline unsigned long
+HYPERVISOR_add_physmap(unsigned long gpfn, unsigned long mfn,
+		       unsigned long flags, domid_t domid)
+{
+	unsigned long ret = 0;
+	BUG_ON(!is_running_on_xen());//XXX
+	if (is_running_on_xen()) {
+		ret = __HYPERVISOR_add_physmap(gpfn, mfn, flags, domid);
+	}
+	return ret;
+}
+
+static inline unsigned long
+__HYPERVISOR_add_physmap_with_gmfn(unsigned long gpfn, unsigned long gmfn,
+                                   unsigned long flags, domid_t domid)
+{
+	return _hypercall5(unsigned long, ia64_dom0vp_op,
+	                   IA64_DOM0VP_add_physmap_with_gmfn,
+	                   gpfn, gmfn, flags, domid);
+}
+
+static inline unsigned long
+HYPERVISOR_add_physmap_with_gmfn(unsigned long gpfn, unsigned long gmfn,
+				 unsigned long flags, domid_t domid)
+{
+	unsigned long ret = 0;
+	BUG_ON(!is_running_on_xen());//XXX
+	if (is_running_on_xen()) {
+		ret = __HYPERVISOR_add_physmap_with_gmfn(gpfn, gmfn,
+		                                         flags, domid);
+	}
+	return ret;
+}
+
+#ifdef CONFIG_XEN_IA64_EXPOSE_P2M
+static inline unsigned long
+HYPERVISOR_expose_p2m(unsigned long conv_start_gpfn,
+                      unsigned long assign_start_gpfn,
+                      unsigned long expose_size, unsigned long granule_pfn)
+{
+	return _hypercall5(unsigned long, ia64_dom0vp_op,
+	                   IA64_DOM0VP_expose_p2m, conv_start_gpfn,
+	                   assign_start_gpfn, expose_size, granule_pfn);
+}
+
+static inline int
+xencomm_arch_expose_foreign_p2m(unsigned long gpfn,
+				domid_t domid, struct xencomm_handle *arg,
+				unsigned long flags)
+{
+	return _hypercall5(int, ia64_dom0vp_op,
+			   IA64_DOM0VP_expose_foreign_p2m,
+			   gpfn, domid, arg, flags);
+}
+
+static inline int
+HYPERVISOR_unexpose_foreign_p2m(unsigned long gpfn, domid_t domid)
+{
+	return _hypercall3(int, ia64_dom0vp_op,
+			   IA64_DOM0VP_unexpose_foreign_p2m, gpfn, domid);
+}
+#endif
+
+static inline int
+xencomm_arch_hypercall_perfmon_op(unsigned long cmd,
+                                  struct xencomm_handle *arg,
+                                  unsigned long count)
+{
+	return _hypercall4(int, ia64_dom0vp_op,
+			   IA64_DOM0VP_perfmon, cmd, arg, count);
+}
+
+static inline int
+xencomm_arch_hypercall_fpswa_revision(struct xencomm_handle *arg)
+{
+	return _hypercall2(int, ia64_dom0vp_op,
+			   IA64_DOM0VP_fpswa_revision, arg);
+}
+
+static inline int
+xencomm_arch_hypercall_ia64_debug_op(unsigned long cmd,
+				     unsigned long domain,
+				     struct xencomm_handle *arg)
+{
+	return _hypercall3(int, ia64_debug_op, cmd, domain, arg);
+}
+
+static inline int
+HYPERVISOR_add_io_space(unsigned long phys_base,
+			unsigned long sparse,
+			unsigned long space_number)
+{
+	return _hypercall4(int, ia64_dom0vp_op, IA64_DOM0VP_add_io_space,
+			   phys_base, sparse, space_number);
+}
+
+static inline int
+xencomm_arch_hypercall_kexec_op(int cmd, struct xencomm_handle *arg)
+{
+	return _hypercall2(int, kexec_op, cmd, arg);
+}
+
+// for balloon driver
+#define HYPERVISOR_update_va_mapping(va, new_val, flags) (0)
+
+/* Use xencomm to do hypercalls.  */
+#define HYPERVISOR_sched_op xencomm_hypercall_sched_op
+#define HYPERVISOR_event_channel_op xencomm_hypercall_event_channel_op
+#define HYPERVISOR_callback_op xencomm_hypercall_callback_op
+#define HYPERVISOR_multicall xencomm_hypercall_multicall
+#define HYPERVISOR_xen_version xencomm_hypercall_xen_version
+#define HYPERVISOR_console_io xencomm_hypercall_console_io
+#define HYPERVISOR_hvm_op xencomm_hypercall_hvm_op
+#define HYPERVISOR_memory_op xencomm_hypercall_memory_op
+#define HYPERVISOR_xenoprof_op xencomm_hypercall_xenoprof_op
+#define HYPERVISOR_perfmon_op xencomm_hypercall_perfmon_op
+#define HYPERVISOR_fpswa_revision xencomm_hypercall_fpswa_revision
+#define HYPERVISOR_suspend xencomm_hypercall_suspend
+#define HYPERVISOR_vcpu_op xencomm_hypercall_vcpu_op
+#define HYPERVISOR_opt_feature xencomm_hypercall_opt_feature
+#define HYPERVISOR_kexec_op xencomm_hypercall_kexec_op
+
+/* to compile gnttab_copy_grant_page() in drivers/xen/core/gnttab.c */
+#define HYPERVISOR_mmu_update(req, count, success_count, domid) ({BUG();0;})
+
+#endif /* __HYPERCALL_H__ */
diff -rpuN linux-2.6.18.8/include/asm-ia64/hypervisor.h linux-2.6.18-xen-3.3.0/include/asm-ia64/hypervisor.h
--- linux-2.6.18.8/include/asm-ia64/hypervisor.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/hypervisor.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,233 @@
+/******************************************************************************
+ * hypervisor.h
+ * 
+ * Linux-specific hypervisor handling.
+ * 
+ * Copyright (c) 2002-2004, K A Fraser
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __HYPERVISOR_H__
+#define __HYPERVISOR_H__
+
+#ifdef CONFIG_XEN
+/* running_on_xen is set before executing any C code by early_xen_setup */
+extern const int running_on_xen;
+#define is_running_on_xen()			(running_on_xen)
+#else /* CONFIG_XEN */
+# ifdef CONFIG_VMX_GUEST
+#  define is_running_on_xen()			(1)
+# else /* CONFIG_VMX_GUEST */
+#  define is_running_on_xen()			(0)
+#  define HYPERVISOR_ioremap(offset, size)	(offset)
+# endif /* CONFIG_VMX_GUEST */
+#endif /* CONFIG_XEN */
+
+#if defined(CONFIG_XEN) || defined(CONFIG_VMX_GUEST)
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/version.h>
+#include <linux/errno.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/platform.h>
+#include <xen/interface/event_channel.h>
+#include <xen/interface/physdev.h>
+#include <xen/interface/sched.h>
+#include <xen/hypercall.h>
+#include <asm/ptrace.h>
+#include <asm/page.h>
+
+extern shared_info_t *HYPERVISOR_shared_info;
+extern start_info_t *xen_start_info;
+
+void force_evtchn_callback(void);
+
+/* Turn jiffies into Xen system time. XXX Implement me. */
+#define jiffies_to_st(j)	0
+
+static inline int
+HYPERVISOR_yield(
+	void)
+{
+	int rc = HYPERVISOR_sched_op(SCHEDOP_yield, NULL);
+
+	return rc;
+}
+
+static inline int
+HYPERVISOR_block(
+	void)
+{
+	int rc = HYPERVISOR_sched_op(SCHEDOP_block, NULL);
+
+	return rc;
+}
+
+static inline int
+HYPERVISOR_shutdown(
+	unsigned int reason)
+{
+	struct sched_shutdown sched_shutdown = {
+		.reason = reason
+	};
+
+	int rc = HYPERVISOR_sched_op(SCHEDOP_shutdown, &sched_shutdown);
+
+	return rc;
+}
+
+static inline int
+HYPERVISOR_poll(
+	evtchn_port_t *ports, unsigned int nr_ports, u64 timeout)
+{
+	struct sched_poll sched_poll = {
+		.nr_ports = nr_ports,
+		.timeout = jiffies_to_st(timeout)
+	};
+
+	int rc;
+
+	set_xen_guest_handle(sched_poll.ports, ports);
+	rc = HYPERVISOR_sched_op(SCHEDOP_poll, &sched_poll);
+
+	return rc;
+}
+
+#ifndef CONFIG_VMX_GUEST
+/* for drivers/xen/privcmd/privcmd.c */
+#define machine_to_phys_mapping 0
+struct vm_area_struct;
+int direct_remap_pfn_range(struct vm_area_struct *vma,
+			   unsigned long address,
+			   unsigned long mfn,
+			   unsigned long size,
+			   pgprot_t prot,
+			   domid_t  domid);
+struct file;
+int privcmd_enforce_singleshot_mapping(struct vm_area_struct *vma);
+int privcmd_mmap(struct file * file, struct vm_area_struct * vma);
+#define HAVE_ARCH_PRIVCMD_MMAP
+
+/* for drivers/xen/balloon/balloon.c */
+#ifdef CONFIG_XEN_SCRUB_PAGES
+#define scrub_pages(_p,_n) memset((void *)(_p), 0, (_n) << PAGE_SHIFT)
+#else
+#define scrub_pages(_p,_n) ((void)0)
+#endif
+#define	pte_mfn(_x)	pte_pfn(_x)
+#define phys_to_machine_mapping_valid(_x)	(1)
+
+void xen_contiguous_bitmap_init(unsigned long end_pfn);
+int __xen_create_contiguous_region(unsigned long vstart, unsigned int order, unsigned int address_bits);
+static inline int
+xen_create_contiguous_region(unsigned long vstart,
+                             unsigned int order, unsigned int address_bits)
+{
+	int ret = 0;
+	if (is_running_on_xen()) {
+		ret = __xen_create_contiguous_region(vstart, order,
+		                                     address_bits);
+	}
+	return ret;
+}
+
+void __xen_destroy_contiguous_region(unsigned long vstart, unsigned int order);
+static inline void
+xen_destroy_contiguous_region(unsigned long vstart, unsigned int order)
+{
+	if (is_running_on_xen())
+		__xen_destroy_contiguous_region(vstart, order);
+}
+
+struct page;
+
+int xen_limit_pages_to_max_mfn(struct page *pages, unsigned int order,
+			       unsigned int address_bits);
+
+/* For drivers/xen/core/machine_reboot.c */
+#define HAVE_XEN_POST_SUSPEND
+void xen_post_suspend(int suspend_cancelled);
+
+/* For setup_arch() in arch/ia64/kernel/setup.c */
+void xen_ia64_enable_opt_feature(void);
+#endif /* !CONFIG_VMX_GUEST */
+
+#define __pte_ma(_x)	((pte_t) {(_x)})        /* unmodified use */
+#define pfn_pte_ma(_x,_y)	__pte_ma(0)     /* unmodified use */
+
+/* for netfront.c, netback.c */
+#define MULTI_UVMFLAGS_INDEX 0 /* XXX any value */
+
+static inline void
+MULTI_update_va_mapping(
+	multicall_entry_t *mcl, unsigned long va,
+	pte_t new_val, unsigned long flags)
+{
+	mcl->op = __HYPERVISOR_update_va_mapping;
+	mcl->result = 0;
+}
+
+static inline void
+MULTI_grant_table_op(multicall_entry_t *mcl, unsigned int cmd,
+	void *uop, unsigned int count)
+{
+	mcl->op = __HYPERVISOR_grant_table_op;
+	mcl->args[0] = cmd;
+	mcl->args[1] = (unsigned long)uop;
+	mcl->args[2] = count;
+}
+
+/*
+ * for blktap.c
+ * int create_lookup_pte_addr(struct mm_struct *mm, 
+ *                            unsigned long address,
+ *                            uint64_t *ptep);
+ */
+#define create_lookup_pte_addr(mm, address, ptep)			\
+	({								\
+		printk(KERN_EMERG					\
+		       "%s:%d "						\
+		       "create_lookup_pte_addr() isn't supported.\n",	\
+		       __func__, __LINE__);				\
+		BUG();							\
+		(-ENOSYS);						\
+	})
+
+/* for debug */
+asmlinkage int xprintk(const char *fmt, ...);
+#define xprintd(fmt, ...)	xprintk("%s:%d " fmt, __func__, __LINE__, \
+					##__VA_ARGS__)
+
+#endif /* CONFIG_XEN || CONFIG_VMX_GUEST */
+
+#ifdef CONFIG_XEN_PRIVILEGED_GUEST
+#define is_initial_xendomain()						\
+	(is_running_on_xen() ? xen_start_info->flags & SIF_INITDOMAIN : 0)
+#else
+#define is_initial_xendomain() 0
+#endif
+
+#endif /* __HYPERVISOR_H__ */
diff -rpuN linux-2.6.18.8/include/asm-ia64/intel_intrin.h linux-2.6.18-xen-3.3.0/include/asm-ia64/intel_intrin.h
--- linux-2.6.18.8/include/asm-ia64/intel_intrin.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/intel_intrin.h	2008-08-21 11:36:07.000000000 +0200
@@ -16,8 +16,8 @@
 		 	 * intrinsic
 		 	 */
 
-#define ia64_getreg		__getReg
-#define ia64_setreg		__setReg
+#define __ia64_getreg		__getReg
+#define __ia64_setreg		__setReg
 
 #define ia64_hint		__hint
 #define ia64_hint_pause		__hint_pause
@@ -33,16 +33,16 @@
 #define ia64_getf_exp		__getf_exp
 #define ia64_shrp		_m64_shrp
 
-#define ia64_tpa		__tpa
+#define __ia64_tpa		__tpa
 #define ia64_invala		__invala
 #define ia64_invala_gr		__invala_gr
 #define ia64_invala_fr		__invala_fr
 #define ia64_nop		__nop
 #define ia64_sum		__sum
-#define ia64_ssm		__ssm
+#define __ia64_ssm		__ssm
 #define ia64_rum		__rum
-#define ia64_rsm		__rsm
-#define ia64_fc 		__fc
+#define __ia64_rsm		__rsm
+#define __ia64_fc 		__fc
 
 #define ia64_ldfs		__ldfs
 #define ia64_ldfd		__ldfd
@@ -80,24 +80,24 @@
 
 #define __ia64_set_dbr(index, val)	\
 		__setIndReg(_IA64_REG_INDR_DBR, index, val)
-#define ia64_set_ibr(index, val)	\
+#define __ia64_set_ibr(index, val)	\
 		__setIndReg(_IA64_REG_INDR_IBR, index, val)
-#define ia64_set_pkr(index, val)	\
+#define __ia64_set_pkr(index, val)	\
 		__setIndReg(_IA64_REG_INDR_PKR, index, val)
-#define ia64_set_pmc(index, val)	\
+#define __ia64_set_pmc(index, val)	\
 		__setIndReg(_IA64_REG_INDR_PMC, index, val)
-#define ia64_set_pmd(index, val)	\
+#define __ia64_set_pmd(index, val)	\
 		__setIndReg(_IA64_REG_INDR_PMD, index, val)
-#define ia64_set_rr(index, val)	\
+#define __ia64_set_rr(index, val)	\
 		__setIndReg(_IA64_REG_INDR_RR, index, val)
 
-#define ia64_get_cpuid(index) 	__getIndReg(_IA64_REG_INDR_CPUID, index)
+#define __ia64_get_cpuid(index) 	__getIndReg(_IA64_REG_INDR_CPUID, index)
 #define __ia64_get_dbr(index) 	__getIndReg(_IA64_REG_INDR_DBR, index)
-#define ia64_get_ibr(index) 	__getIndReg(_IA64_REG_INDR_IBR, index)
-#define ia64_get_pkr(index) 	__getIndReg(_IA64_REG_INDR_PKR, index)
-#define ia64_get_pmc(index) 	__getIndReg(_IA64_REG_INDR_PMC, index)
-#define ia64_get_pmd(index)  	__getIndReg(_IA64_REG_INDR_PMD, index)
-#define ia64_get_rr(index) 	__getIndReg(_IA64_REG_INDR_RR, index)
+#define __ia64_get_ibr(index) 	__getIndReg(_IA64_REG_INDR_IBR, index)
+#define __ia64_get_pkr(index) 	__getIndReg(_IA64_REG_INDR_PKR, index)
+#define __ia64_get_pmc(index) 	__getIndReg(_IA64_REG_INDR_PMC, index)
+#define __ia64_get_pmd(index)  	__getIndReg(_IA64_REG_INDR_PMD, index)
+#define __ia64_get_rr(index) 	__getIndReg(_IA64_REG_INDR_RR, index)
 
 #define ia64_srlz_d		__dsrlz
 #define ia64_srlz_i		__isrlz
@@ -116,18 +116,18 @@
 #define ia64_ld8_acq		__ld8_acq
 
 #define ia64_sync_i		__synci
-#define ia64_thash		__thash
-#define ia64_ttag		__ttag
-#define ia64_itcd		__itcd
-#define ia64_itci		__itci
-#define ia64_itrd		__itrd
-#define ia64_itri		__itri
-#define ia64_ptce		__ptce
-#define ia64_ptcl		__ptcl
-#define ia64_ptcg		__ptcg
-#define ia64_ptcga		__ptcga
-#define ia64_ptri		__ptri
-#define ia64_ptrd		__ptrd
+#define __ia64_thash		__thash
+#define __ia64_ttag		__ttag
+#define __ia64_itcd		__itcd
+#define __ia64_itci		__itci
+#define __ia64_itrd		__itrd
+#define __ia64_itri		__itri
+#define __ia64_ptce		__ptce
+#define __ia64_ptcl		__ptcl
+#define __ia64_ptcg		__ptcg
+#define __ia64_ptcga		__ptcga
+#define __ia64_ptri		__ptri
+#define __ia64_ptrd		__ptrd
 #define ia64_dep_mi		_m64_dep_mi
 
 /* Values for lfhint in __lfetch and __lfetch_fault */
@@ -142,16 +142,18 @@
 #define ia64_lfetch_fault	__lfetch_fault
 #define ia64_lfetch_fault_excl	__lfetch_fault_excl
 
-#define ia64_intrin_local_irq_restore(x)		\
+#define __ia64_intrin_local_irq_restore(x)		\
 do {							\
 	if ((x) != 0) {					\
-		ia64_ssm(IA64_PSR_I);			\
+		__ia64_ssm(IA64_PSR_I);			\
 		ia64_srlz_d();				\
 	} else {					\
-		ia64_rsm(IA64_PSR_I);			\
+		__ia64_rsm(IA64_PSR_I);			\
 	}						\
 } while (0)
 
+#define __ia64_get_psr_i()	(__ia64_getreg(_IA64_REG_PSR) & 0x4000UL)
+
 #define __builtin_trap()	__break(0);
 
 #endif /* _ASM_IA64_INTEL_INTRIN_H */
diff -rpuN linux-2.6.18.8/include/asm-ia64/intrinsics.h linux-2.6.18-xen-3.3.0/include/asm-ia64/intrinsics.h
--- linux-2.6.18.8/include/asm-ia64/intrinsics.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/intrinsics.h	2008-08-21 11:36:07.000000000 +0200
@@ -18,6 +18,15 @@
 # include <asm/gcc_intrin.h>
 #endif
 
+#define __ia64_set_rr0_to_rr4(val0, val1, val2, val3, val4)	\
+do {								\
+	__ia64_set_rr(0x0000000000000000UL, (val0));		\
+	__ia64_set_rr(0x2000000000000000UL, (val1));		\
+	__ia64_set_rr(0x4000000000000000UL, (val2));		\
+	__ia64_set_rr(0x6000000000000000UL, (val3));		\
+	__ia64_set_rr(0x8000000000000000UL, (val4));		\
+} while (0)
+
 /*
  * Force an unresolved reference if someone tries to use
  * ia64_fetch_and_add() with a bad value.
@@ -177,4 +186,5 @@ extern long ia64_cmpxchg_called_with_bad
 #endif /* !CONFIG_IA64_DEBUG_CMPXCHG */
 
 #endif
+#include <asm/privop.h>
 #endif /* _ASM_IA64_INTRINSICS_H */
diff -rpuN linux-2.6.18.8/include/asm-ia64/io.h linux-2.6.18-xen-3.3.0/include/asm-ia64/io.h
--- linux-2.6.18.8/include/asm-ia64/io.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/io.h	2008-08-21 11:36:07.000000000 +0200
@@ -96,9 +96,39 @@ extern int valid_mmap_phys_addr_range (u
  * The following two macros are deprecated and scheduled for removal.
  * Please use the PCI-DMA interface defined in <asm/pci.h> instead.
  */
+#ifndef CONFIG_XEN
 #define bus_to_virt	phys_to_virt
 #define virt_to_bus	virt_to_phys
 #define page_to_bus	page_to_phys
+#else
+#define bus_to_virt(bus)	\
+	phys_to_virt(machine_to_phys_for_dma(bus))
+#define virt_to_bus(virt)	\
+	phys_to_machine_for_dma(virt_to_phys(virt))
+#define page_to_bus(page)	\
+	phys_to_machine_for_dma(page_to_pseudophys(page))
+
+#define page_to_pseudophys(page) \
+	((dma_addr_t)page_to_pfn(page) << PAGE_SHIFT)
+
+/*
+ * Drivers that use page_to_phys() for bus addresses are broken.
+ * This includes:
+ * drivers/ide/cris/ide-cris.c
+ * drivers/scsi/dec_esp.c
+ */
+#define page_to_phys(page)	(page_to_pseudophys(page))
+#define bvec_to_bus(bv)		(page_to_bus((bv)->bv_page) + \
+				(unsigned long) (bv)->bv_offset)
+#define bio_to_pseudophys(bio)	(page_to_pseudophys(bio_page((bio))) +	\
+				 (unsigned long) bio_offset((bio)))
+#define bvec_to_pseudophys(bv)  (page_to_pseudophys((bv)->bv_page) +	\
+				 (unsigned long) (bv)->bv_offset)
+#define BIOVEC_PHYS_MERGEABLE(vec1, vec2)				\
+	(((bvec_to_bus((vec1)) + (vec1)->bv_len) == bvec_to_bus((vec2))) && \
+	 ((bvec_to_pseudophys((vec1)) + (vec1)->bv_len) ==		\
+	  bvec_to_pseudophys((vec2))))
+#endif /* CONFIG_XEN */
 
 # endif /* KERNEL */
 
diff -rpuN linux-2.6.18.8/include/asm-ia64/iosapic.h linux-2.6.18-xen-3.3.0/include/asm-ia64/iosapic.h
--- linux-2.6.18.8/include/asm-ia64/iosapic.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/iosapic.h	2008-08-21 11:36:07.000000000 +0200
@@ -53,6 +53,7 @@
 
 #define NR_IOSAPICS			256
 
+#ifndef CONFIG_XEN
 static inline unsigned int iosapic_read(char __iomem *iosapic, unsigned int reg)
 {
 	writel(reg, iosapic + IOSAPIC_REG_SELECT);
@@ -64,6 +65,7 @@ static inline void iosapic_write(char __
 	writel(reg, iosapic + IOSAPIC_REG_SELECT);
 	writel(val, iosapic + IOSAPIC_WINDOW);
 }
+#endif
 
 static inline void iosapic_eoi(char __iomem *iosapic, u32 vector)
 {
diff -rpuN linux-2.6.18.8/include/asm-ia64/irq.h linux-2.6.18-xen-3.3.0/include/asm-ia64/irq.h
--- linux-2.6.18.8/include/asm-ia64/irq.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/irq.h	2008-08-21 11:36:07.000000000 +0200
@@ -11,8 +11,41 @@
  * 02/29/00     D.Mosberger	moved most things into hw_irq.h
  */
 
+#ifndef CONFIG_XEN
 #define NR_IRQS		256
 #define NR_IRQ_VECTORS	NR_IRQS
+#else
+/*
+ * The flat IRQ space is divided into two regions:
+ *  1. A one-to-one mapping of real physical IRQs. This space is only used
+ *     if we have physical device-access privilege. This region is at the 
+ *     start of the IRQ space so that existing device drivers do not need
+ *     to be modified to translate physical IRQ numbers into our IRQ space.
+ *  3. A dynamic mapping of inter-domain and Xen-sourced virtual IRQs. These
+ *     are bound using the provided bind/unbind functions.
+ */
+
+#define PIRQ_BASE		0
+#define NR_PIRQS		256
+
+#define DYNIRQ_BASE		(PIRQ_BASE + NR_PIRQS)
+#define NR_DYNIRQS		(CONFIG_NR_CPUS * 8)
+
+#define NR_IRQS			(NR_PIRQS + NR_DYNIRQS)
+#define NR_IRQ_VECTORS		NR_IRQS
+
+#define pirq_to_irq(_x)		((_x) + PIRQ_BASE)
+#define irq_to_pirq(_x)		((_x) - PIRQ_BASE)
+
+#define dynirq_to_irq(_x)	((_x) + DYNIRQ_BASE)
+#define irq_to_dynirq(_x)	((_x) - DYNIRQ_BASE)
+
+#define RESCHEDULE_VECTOR	0
+#define IPI_VECTOR		1
+#define CMCP_VECTOR		2
+#define CPEP_VECTOR		3
+#define NR_IPIS			4
+#endif /* CONFIG_XEN */
 
 static __inline__ int
 irq_canonicalize (int irq)
diff -rpuN linux-2.6.18.8/include/asm-ia64/kexec.h linux-2.6.18-xen-3.3.0/include/asm-ia64/kexec.h
--- linux-2.6.18.8/include/asm-ia64/kexec.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/kexec.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,58 @@
+#ifndef _ASM_IA64_KEXEC_H
+#define _ASM_IA64_KEXEC_H
+
+
+/* Maximum physical address we can use pages from */
+#define KEXEC_SOURCE_MEMORY_LIMIT (-1UL)
+/* Maximum address we can reach in physical address mode */
+#define KEXEC_DESTINATION_MEMORY_LIMIT (-1UL)
+/* Maximum address we can use for the control code buffer */
+#define KEXEC_CONTROL_MEMORY_LIMIT TASK_SIZE
+
+#define KEXEC_CONTROL_CODE_SIZE (8192 + 8192 + 4096)
+
+/* The native architecture */
+#define KEXEC_ARCH KEXEC_ARCH_IA_64
+
+#define MAX_NOTE_BYTES 1024
+
+#define kexec_flush_icache_page(page) do { \
+                unsigned long page_addr = (unsigned long)page_address(page); \
+                flush_icache_range(page_addr, page_addr + PAGE_SIZE); \
+        } while(0)
+
+extern struct kimage *ia64_kimage;
+DECLARE_PER_CPU(u64, ia64_mca_pal_base);
+const extern unsigned int relocate_new_kernel_size;
+extern void relocate_new_kernel(unsigned long, unsigned long,
+		struct ia64_boot_param *, unsigned long);
+static inline void
+crash_setup_regs(struct pt_regs *newregs, struct pt_regs *oldregs)
+{
+}
+extern struct resource efi_memmap_res;
+extern struct resource boot_param_res;
+extern void kdump_smp_send_stop(void);
+extern void kdump_smp_send_init(void);
+extern void kexec_disable_iosapic(void);
+extern void crash_save_this_cpu(void);
+struct rsvd_region;
+extern unsigned long kdump_find_rsvd_region(unsigned long size,
+		struct rsvd_region *rsvd_regions, int n);
+extern void kdump_cpu_freeze(struct unw_frame_info *info, void *arg);
+extern int kdump_status[];
+extern atomic_t kdump_cpu_freezed;
+extern atomic_t kdump_in_progress;
+
+/* Kexec needs to know about the actual physical addresss.
+ * But in xen, on some architectures, a physical address is a
+ * pseudo-physical addresss. */
+#ifdef CONFIG_XEN
+#define KEXEC_ARCH_HAS_PAGE_MACROS
+#define kexec_page_to_pfn(page)  pfn_to_mfn_for_dma(page_to_pfn(page))
+#define kexec_pfn_to_page(pfn)   pfn_to_page(mfn_to_pfn_for_dma(pfn))
+#define kexec_virt_to_phys(addr) phys_to_machine_for_dma(__pa(addr))
+#define kexec_phys_to_virt(addr) phys_to_virt(machine_to_phys_for_dma(addr))
+#endif
+
+#endif /* _ASM_IA64_KEXEC_H */
diff -rpuN linux-2.6.18.8/include/asm-ia64/machvec.h linux-2.6.18-xen-3.3.0/include/asm-ia64/machvec.h
--- linux-2.6.18.8/include/asm-ia64/machvec.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/machvec.h	2008-08-21 11:36:07.000000000 +0200
@@ -35,6 +35,7 @@ typedef int ia64_mv_pci_legacy_read_t (s
 typedef int ia64_mv_pci_legacy_write_t (struct pci_bus *, u16 port, u32 val,
 					u8 size);
 typedef void ia64_mv_migrate_t(struct task_struct * task);
+typedef void ia64_mv_kernel_launch_event_t(void);
 
 /* DMA-mapping interface: */
 typedef void ia64_mv_dma_init (void);
@@ -108,6 +109,8 @@ extern void machvec_tlb_migrate_finish (
 #  include <asm/machvec_hpzx1_swiotlb.h>
 # elif defined (CONFIG_IA64_SGI_SN2)
 #  include <asm/machvec_sn2.h>
+# elif defined (CONFIG_IA64_XEN)
+#  include <asm/machvec_xen.h>
 # elif defined (CONFIG_IA64_GENERIC)
 
 # ifdef MACHVEC_PLATFORM_HEADER
@@ -205,6 +208,7 @@ struct ia64_machine_vector {
 	ia64_mv_readq_relaxed_t *readq_relaxed;
 	ia64_mv_migrate_t *migrate;
 	ia64_mv_msi_init_t *msi_init;
+	ia64_mv_kernel_launch_event_t *kernel_launch_event;
 } __attribute__((__aligned__(16))); /* align attrib? see above comment */
 
 #define MACHVEC_INIT(name)			\
@@ -303,6 +307,9 @@ extern ia64_mv_dma_supported		swiotlb_dm
 #ifndef platform_tlb_migrate_finish
 # define platform_tlb_migrate_finish	machvec_noop_mm
 #endif
+#ifndef platform_kernel_launch_event
+# define platform_kernel_launch_event	machvec_noop
+#endif
 #ifndef platform_dma_init
 # define platform_dma_init		swiotlb_init
 #endif
diff -rpuN linux-2.6.18.8/include/asm-ia64/machvec_sn2.h linux-2.6.18-xen-3.3.0/include/asm-ia64/machvec_sn2.h
--- linux-2.6.18.8/include/asm-ia64/machvec_sn2.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/machvec_sn2.h	2008-08-21 11:36:07.000000000 +0200
@@ -67,6 +67,7 @@ extern ia64_mv_dma_sync_sg_for_device	sn
 extern ia64_mv_dma_mapping_error	sn_dma_mapping_error;
 extern ia64_mv_dma_supported		sn_dma_supported;
 extern ia64_mv_migrate_t		sn_migrate;
+extern ia64_mv_kernel_launch_event_t	sn_kernel_launch_event;
 extern ia64_mv_msi_init_t		sn_msi_init;
 
 
@@ -119,6 +120,7 @@ extern ia64_mv_msi_init_t		sn_msi_init;
 #define platform_dma_mapping_error		sn_dma_mapping_error
 #define platform_dma_supported		sn_dma_supported
 #define platform_migrate		sn_migrate
+#define platform_kernel_launch_event    sn_kernel_launch_event
 #ifdef CONFIG_PCI_MSI
 #define platform_msi_init		sn_msi_init
 #else
diff -rpuN linux-2.6.18.8/include/asm-ia64/machvec_xen.h linux-2.6.18-xen-3.3.0/include/asm-ia64/machvec_xen.h
--- linux-2.6.18.8/include/asm-ia64/machvec_xen.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/machvec_xen.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,43 @@
+#ifndef _ASM_IA64_MACHVEC_XEN_h
+#define _ASM_IA64_MACHVEC_XEN_h
+
+extern ia64_mv_setup_t			xen_setup;
+extern ia64_mv_cpu_init_t		xen_cpu_init;
+extern ia64_mv_irq_init_t		xen_irq_init;
+extern ia64_mv_send_ipi_t		xen_platform_send_ipi;
+extern ia64_mv_dma_alloc_coherent	xen_alloc_coherent;
+extern ia64_mv_dma_free_coherent	xen_free_coherent;
+extern ia64_mv_dma_map_single		xen_map_single;
+extern ia64_mv_dma_unmap_single		xen_unmap_single;
+extern ia64_mv_dma_map_sg		xen_map_sg;
+extern ia64_mv_dma_unmap_sg		xen_unmap_sg;
+extern ia64_mv_dma_supported		xen_dma_supported;
+extern ia64_mv_dma_mapping_error	xen_dma_mapping_error;
+
+/*
+ * This stuff has dual use!
+ *
+ * For a generic kernel, the macros are used to initialize the
+ * platform's machvec structure.  When compiling a non-generic kernel,
+ * the macros are used directly.
+ */
+#define platform_name				"xen"
+#define platform_setup				xen_setup
+#define platform_cpu_init			xen_cpu_init
+#define platform_irq_init			xen_irq_init
+#define platform_send_ipi			xen_platform_send_ipi
+#define platform_dma_init			machvec_noop
+#define platform_dma_alloc_coherent		xen_alloc_coherent
+#define platform_dma_free_coherent		xen_free_coherent
+#define platform_dma_map_single			xen_map_single
+#define platform_dma_unmap_single		xen_unmap_single
+#define platform_dma_map_sg			xen_map_sg
+#define platform_dma_unmap_sg			xen_unmap_sg
+#define platform_dma_sync_single_for_cpu	machvec_dma_sync_single
+#define platform_dma_sync_sg_for_cpu		machvec_dma_sync_sg
+#define platform_dma_sync_single_for_device	machvec_dma_sync_single
+#define platform_dma_sync_sg_for_device		machvec_dma_sync_sg
+#define platform_dma_supported			xen_dma_supported
+#define platform_dma_mapping_error		xen_dma_mapping_error
+
+#endif /* _ASM_IA64_MACHVEC_XEN_h */
diff -rpuN linux-2.6.18.8/include/asm-ia64/maddr.h linux-2.6.18-xen-3.3.0/include/asm-ia64/maddr.h
--- linux-2.6.18.8/include/asm-ia64/maddr.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/maddr.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,116 @@
+#ifndef _ASM_IA64_MADDR_H
+#define _ASM_IA64_MADDR_H
+
+#include <linux/kernel.h>
+#include <asm/hypervisor.h>
+#include <xen/features.h>
+#include <xen/interface/xen.h>
+
+#ifdef CONFIG_XEN
+
+#define INVALID_P2M_ENTRY       (~0UL)
+
+#ifdef CONFIG_XEN_IA64_EXPOSE_P2M
+extern int p2m_initialized;
+extern unsigned long p2m_min_low_pfn;
+extern unsigned long p2m_max_low_pfn;
+extern unsigned long p2m_convert_min_pfn;
+extern unsigned long p2m_convert_max_pfn;
+extern volatile const pte_t* p2m_pte;
+unsigned long p2m_phystomach(unsigned long gpfn);
+#else
+#define p2m_initialized		(0)
+#define p2m_phystomach(gpfn)	INVALID_MFN
+#endif
+
+/* XXX xen page size != page size */
+static inline unsigned long
+pfn_to_mfn_for_dma(unsigned long pfn)
+{
+	unsigned long mfn;
+	if (p2m_initialized)
+		return p2m_phystomach(pfn);
+	mfn = HYPERVISOR_phystomach(pfn);
+	BUG_ON(mfn == 0); /* XXX */
+	BUG_ON(mfn == INVALID_P2M_ENTRY); /* XXX */
+	BUG_ON(mfn == INVALID_MFN);
+	return mfn;
+}
+
+static inline unsigned long
+phys_to_machine_for_dma(unsigned long phys)
+{
+	unsigned long machine =
+	              pfn_to_mfn_for_dma(phys >> PAGE_SHIFT) << PAGE_SHIFT;
+	machine |= (phys & ~PAGE_MASK);
+	return machine;
+}
+
+static inline unsigned long
+mfn_to_pfn_for_dma(unsigned long mfn)
+{
+	unsigned long pfn;
+	pfn = HYPERVISOR_machtophys(mfn);
+	BUG_ON(pfn == 0);
+	/* BUG_ON(pfn == INVALID_M2P_ENTRY); */
+	return pfn;
+}
+
+static inline unsigned long
+machine_to_phys_for_dma(unsigned long machine)
+{
+	unsigned long phys =
+	              mfn_to_pfn_for_dma(machine >> PAGE_SHIFT) << PAGE_SHIFT;
+	phys |= (machine & ~PAGE_MASK);
+	return phys;
+}
+
+#ifdef CONFIG_SPARSEMEM
+/*
+ * When CONFIG_SPARSEMEM=y, pfn_valid() is defined in
+ * linux/include/linux/mmzone.h. Hoever #include <linux/mmzone.h>
+ * causes the header inclusion hell.
+ */
+static inline int pfn_valid(unsigned long pfn);
+#endif
+
+static inline unsigned long
+mfn_to_local_pfn(unsigned long mfn)
+{
+	unsigned long pfn = mfn_to_pfn_for_dma(mfn);
+	if (!pfn_valid(pfn))
+		return INVALID_P2M_ENTRY;
+	return pfn;
+}
+
+#else /* !CONFIG_XEN */
+
+#define pfn_to_mfn_for_dma(pfn) (pfn)
+#define mfn_to_pfn_for_dma(mfn) (mfn)
+#define phys_to_machine_for_dma(phys) (phys)
+#define machine_to_phys_for_dma(machine) (machine)
+#define mfn_to_local_pfn(mfn) (mfn)
+
+#endif /* !CONFIG_XEN */
+
+#define mfn_to_pfn(mfn) (mfn)
+#define pfn_to_mfn(pfn) (pfn)
+
+#define mfn_to_virt(mfn) (__va((mfn) << PAGE_SHIFT))
+#define virt_to_mfn(virt) (__pa(virt) >> PAGE_SHIFT)
+#define virt_to_machine(virt) __pa(virt) /* for tpmfront.c */
+
+#define set_phys_to_machine(pfn, mfn) do { } while (0)
+
+typedef unsigned long maddr_t;	/* to compile netback, netfront */
+#ifndef _ASM_IA64_SN_TYPES_H /* paddr_t is defined in asm-ia64/sn/types.h */
+typedef unsigned long paddr_t;
+#endif
+
+#ifdef CONFIG_XEN
+int range_straddles_page_boundary(paddr_t p, size_t size);
+#else
+#define range_straddles_page_boundary(addr, size)	(0)
+#endif
+
+#endif /* _ASM_IA64_MADDR_H */
diff -rpuN linux-2.6.18.8/include/asm-ia64/meminit.h linux-2.6.18-xen-3.3.0/include/asm-ia64/meminit.h
--- linux-2.6.18.8/include/asm-ia64/meminit.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/meminit.h	2008-08-21 11:36:07.000000000 +0200
@@ -15,11 +15,17 @@
  * 	- initrd (optional)
  * 	- command line string
  * 	- kernel code & data
+ * 	- crash dumping code reserved region
  * 	- Kernel memory map built from EFI memory map
+ *	- xen start info
  *
  * More could be added if necessary
  */
-#define IA64_MAX_RSVD_REGIONS 6
+#ifndef CONFIG_XEN
+#define IA64_MAX_RSVD_REGIONS 7
+#else
+#define IA64_MAX_RSVD_REGIONS 8
+#endif
 
 struct rsvd_region {
 	unsigned long start;	/* virtual address of beginning of element */
diff -rpuN linux-2.6.18.8/include/asm-ia64/mmu_context.h linux-2.6.18-xen-3.3.0/include/asm-ia64/mmu_context.h
--- linux-2.6.18.8/include/asm-ia64/mmu_context.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/mmu_context.h	2008-08-21 11:36:07.000000000 +0200
@@ -151,11 +151,7 @@ reload_context (nv_mm_context_t context)
 #  endif
 #endif
 
-	ia64_set_rr(0x0000000000000000UL, rr0);
-	ia64_set_rr(0x2000000000000000UL, rr1);
-	ia64_set_rr(0x4000000000000000UL, rr2);
-	ia64_set_rr(0x6000000000000000UL, rr3);
-	ia64_set_rr(0x8000000000000000UL, rr4);
+	ia64_set_rr0_to_rr4(rr0, rr1, rr2, rr3, rr4);
 	ia64_srlz_i();			/* srlz.i implies srlz.d */
 }
 
diff -rpuN linux-2.6.18.8/include/asm-ia64/page.h linux-2.6.18-xen-3.3.0/include/asm-ia64/page.h
--- linux-2.6.18.8/include/asm-ia64/page.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/page.h	2008-08-21 11:36:07.000000000 +0200
@@ -119,6 +119,7 @@ extern struct page *vmem_map;
 #endif
 
 #ifdef CONFIG_FLATMEM
+extern unsigned long max_mapnr;
 # define pfn_valid(pfn)		(((pfn) < max_mapnr) && ia64_pfn_valid(pfn))
 #elif defined(CONFIG_DISCONTIGMEM)
 extern unsigned long min_low_pfn;
@@ -126,7 +127,9 @@ extern unsigned long max_low_pfn;
 # define pfn_valid(pfn)		(((pfn) >= min_low_pfn) && ((pfn) < max_low_pfn) && ia64_pfn_valid(pfn))
 #endif
 
+#ifndef CONFIG_XEN
 #define page_to_phys(page)	(page_to_pfn(page) << PAGE_SHIFT)
+#endif
 #define virt_to_page(kaddr)	pfn_to_page(__pa(kaddr) >> PAGE_SHIFT)
 #define pfn_to_kaddr(pfn)	__va((pfn) << PAGE_SHIFT)
 
@@ -227,5 +230,17 @@ get_order (unsigned long size)
 					 (((current->personality & READ_IMPLIES_EXEC) != 0)	\
 					  ? VM_EXEC : 0))
 
+#ifndef __ASSEMBLY__
+
+#include <linux/kernel.h>
+#include <asm/hypervisor.h>	/* to compile ioremap.c */
+
+#ifdef CONFIG_XEN
+
+#include <asm/maddr.h>
+
+#endif /* CONFIG_XEN */
+#endif /* __ASSEMBLY__ */
+
 # endif /* __KERNEL__ */
 #endif /* _ASM_IA64_PAGE_H */
diff -rpuN linux-2.6.18.8/include/asm-ia64/pgalloc.h linux-2.6.18-xen-3.3.0/include/asm-ia64/pgalloc.h
--- linux-2.6.18.8/include/asm-ia64/pgalloc.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/pgalloc.h	2008-08-21 11:36:07.000000000 +0200
@@ -125,7 +125,11 @@ static inline void pmd_free(pmd_t * pmd)
 static inline void
 pmd_populate(struct mm_struct *mm, pmd_t * pmd_entry, struct page *pte)
 {
+#ifndef CONFIG_XEN
 	pmd_val(*pmd_entry) = page_to_phys(pte);
+#else
+	pmd_val(*pmd_entry) = page_to_pseudophys(pte);
+#endif
 }
 
 static inline void
diff -rpuN linux-2.6.18.8/include/asm-ia64/privop.h linux-2.6.18-xen-3.3.0/include/asm-ia64/privop.h
--- linux-2.6.18.8/include/asm-ia64/privop.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/privop.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,63 @@
+#ifndef _ASM_IA64_PRIVOP_H
+#define _ASM_IA64_PRIVOP_H
+
+#ifndef _ASM_IA64_INTRINSICS_H
+#error "don't include privop.h directly. instead include intrinsics.h"
+#endif
+/*
+ * Copyright (C) 2005 Hewlett-Packard Co
+ *	Dan Magenheimer <dan.magenheimer@hp.com>
+ *
+ */
+
+#ifdef CONFIG_XEN
+#include <asm/xen/privop.h>
+#endif
+
+#ifndef __ASSEMBLY
+
+#ifndef IA64_PARAVIRTUALIZED
+
+#define ia64_getreg			__ia64_getreg
+#define ia64_setreg			__ia64_setreg
+#define ia64_hint			__ia64_hint
+#define ia64_thash			__ia64_thash
+#define ia64_itci			__ia64_itci
+#define ia64_itcd			__ia64_itcd
+#define ia64_itri			__ia64_itri
+#define ia64_itrd			__ia64_itrd
+#define ia64_tpa			__ia64_tpa
+#define ia64_set_ibr			__ia64_set_ibr
+#define ia64_set_pkr			__ia64_set_pkr
+#define ia64_set_pmc			__ia64_set_pmc
+#define ia64_set_pmd			__ia64_set_pmd
+#define ia64_set_rr			__ia64_set_rr
+#define ia64_get_cpuid			__ia64_get_cpuid
+#define ia64_get_ibr			__ia64_get_ibr
+#define ia64_get_pkr			__ia64_get_pkr
+#define ia64_get_pmc			__ia64_get_pmc
+#define ia64_get_pmd			__ia64_get_pmd
+#define ia64_get_rr			__ia64_get_rr
+#define ia64_fc				__ia64_fc
+#define ia64_ssm			__ia64_ssm
+#define ia64_rsm			__ia64_rsm
+#define ia64_ptce			__ia64_ptce
+#define ia64_ptcga			__ia64_ptcga
+#define ia64_ptcl			__ia64_ptcl
+#define ia64_ptri			__ia64_ptri
+#define ia64_ptrd			__ia64_ptrd
+#define ia64_get_psr_i			__ia64_get_psr_i
+#define ia64_intrin_local_irq_restore	__ia64_intrin_local_irq_restore
+#define ia64_leave_kernel		__ia64_leave_kernel
+#define ia64_leave_syscall		__ia64_leave_syscall
+#define ia64_trace_syscall		__ia64_trace_syscall
+#define ia64_ret_from_clone		__ia64_ret_from_clone
+#define ia64_switch_to			__ia64_switch_to
+#define ia64_pal_call_static		__ia64_pal_call_static
+#define ia64_set_rr0_to_rr4		__ia64_set_rr0_to_rr4
+
+#endif /* !IA64_PARAVIRTUALIZED */
+
+#endif /* !__ASSEMBLY */
+
+#endif /* _ASM_IA64_PRIVOP_H */
diff -rpuN linux-2.6.18.8/include/asm-ia64/sal.h linux-2.6.18-xen-3.3.0/include/asm-ia64/sal.h
--- linux-2.6.18.8/include/asm-ia64/sal.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/sal.h	2008-08-21 11:36:07.000000000 +0200
@@ -42,6 +42,9 @@
 #include <asm/pal.h>
 #include <asm/system.h>
 #include <asm/fpu.h>
+#ifdef CONFIG_XEN
+#include <asm/xen/xencomm.h>
+#endif
 
 extern spinlock_t sal_lock;
 
@@ -686,10 +689,43 @@ ia64_sal_clear_state_info (u64 sal_info_
 /* Get the processor and platform information logged by SAL with respect to the machine
  * state at the time of the MCAs, INITs, CMCs, or CPEs.
  */
+#ifdef CONFIG_XEN
+static inline u64 ia64_sal_get_state_info_size (u64 sal_info_type);
+typedef struct ia64_mca_xencomm_t {
+	void *record;
+	struct xencomm_handle *handle;
+	struct list_head list;
+} ia64_mca_xencomm_t;
+extern struct list_head ia64_mca_xencomm_list;
+extern spinlock_t ia64_mca_xencomm_lock;
+#endif
+
 static inline u64
 ia64_sal_get_state_info (u64 sal_info_type, u64 *sal_info)
 {
 	struct ia64_sal_retval isrv;
+#ifdef CONFIG_XEN
+	if (is_running_on_xen()) {
+		ia64_mca_xencomm_t *entry;
+		struct xencomm_handle *desc = NULL;
+		unsigned long flags;
+
+		spin_lock_irqsave(&ia64_mca_xencomm_lock, flags);
+		list_for_each_entry(entry, &ia64_mca_xencomm_list, list) {
+			if (entry->record == sal_info) {
+				desc = entry->handle;
+				break;
+			}
+		}
+		spin_unlock_irqrestore(&ia64_mca_xencomm_lock, flags);
+
+		if (desc == NULL)
+			return 0;
+
+		SAL_CALL_REENTRANT(isrv, SAL_GET_STATE_INFO, sal_info_type, 0,
+		                   desc, 0, 0, 0, 0);
+	} else
+#endif
 	SAL_CALL_REENTRANT(isrv, SAL_GET_STATE_INFO, sal_info_type, 0,
 	              sal_info, 0, 0, 0, 0);
 	if (isrv.status)
diff -rpuN linux-2.6.18.8/include/asm-ia64/sn/sn_sal.h linux-2.6.18-xen-3.3.0/include/asm-ia64/sn/sn_sal.h
--- linux-2.6.18.8/include/asm-ia64/sn/sn_sal.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/sn/sn_sal.h	2008-08-21 11:36:07.000000000 +0200
@@ -87,6 +87,8 @@
 #define  SN_SAL_INJECT_ERROR			   0x02000067
 #define  SN_SAL_SET_CPU_NUMBER			   0x02000068
 
+#define  SN_SAL_KERNEL_LAUNCH_EVENT		   0x02000069
+
 /*
  * Service-specific constants
  */
@@ -1154,4 +1156,11 @@ ia64_sn_set_cpu_number(int cpu)
 	SAL_CALL_NOLOCK(rv, SN_SAL_SET_CPU_NUMBER, cpu, 0, 0, 0, 0, 0, 0);
 	return rv.status;
 }
+static inline int
+ia64_sn_kernel_launch_event(void)
+{
+ 	struct ia64_sal_retval rv;
+	SAL_CALL_NOLOCK(rv, SN_SAL_KERNEL_LAUNCH_EVENT, 0, 0, 0, 0, 0, 0, 0);
+	return rv.status;
+}
 #endif /* _ASM_IA64_SN_SN_SAL_H */
diff -rpuN linux-2.6.18.8/include/asm-ia64/sn/types.h linux-2.6.18-xen-3.3.0/include/asm-ia64/sn/types.h
--- linux-2.6.18.8/include/asm-ia64/sn/types.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/sn/types.h	2008-08-21 11:36:07.000000000 +0200
@@ -20,7 +20,9 @@ typedef unsigned char	slotid_t;	/* slot 
 typedef unsigned char	slabid_t;	/* slab (asic) within slot */
 typedef u64 nic_t;
 typedef unsigned long iopaddr_t;
+#ifndef _ASM_IA64_MADDR_H /* paddr_t is defined in asm-ia64/maddr.h */
 typedef unsigned long paddr_t;
+#endif
 typedef short cnodeid_t;
 
 #endif /* _ASM_IA64_SN_TYPES_H */
diff -rpuN linux-2.6.18.8/include/asm-ia64/swiotlb.h linux-2.6.18-xen-3.3.0/include/asm-ia64/swiotlb.h
--- linux-2.6.18.8/include/asm-ia64/swiotlb.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/swiotlb.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,41 @@
+#ifndef _ASM_SWIOTLB_H
+#define _ASM_SWIOTLB_H 1
+
+/* SWIOTLB interface */
+
+extern dma_addr_t swiotlb_map_single(struct device *hwdev, void *ptr, size_t size,
+				      int dir);
+extern void swiotlb_unmap_single(struct device *hwdev, dma_addr_t dev_addr,
+				  size_t size, int dir);
+extern void swiotlb_sync_single_for_cpu(struct device *hwdev,
+					 dma_addr_t dev_addr,
+					 size_t size, int dir);
+extern void swiotlb_sync_single_for_device(struct device *hwdev,
+					    dma_addr_t dev_addr,
+					    size_t size, int dir);
+extern void swiotlb_sync_sg_for_cpu(struct device *hwdev,
+				     struct scatterlist *sg, int nelems,
+				     int dir);
+extern void swiotlb_sync_sg_for_device(struct device *hwdev,
+					struct scatterlist *sg, int nelems,
+					int dir);
+extern int swiotlb_map_sg(struct device *hwdev, struct scatterlist *sg,
+		      int nents, int direction);
+extern void swiotlb_unmap_sg(struct device *hwdev, struct scatterlist *sg,
+			 int nents, int direction);
+extern int swiotlb_dma_mapping_error(dma_addr_t dma_addr);
+extern dma_addr_t swiotlb_map_page(struct device *hwdev, struct page *page,
+                                   unsigned long offset, size_t size,
+                                   enum dma_data_direction direction);
+extern void swiotlb_unmap_page(struct device *hwdev, dma_addr_t dma_address,
+                               size_t size, enum dma_data_direction direction);
+extern int swiotlb_dma_supported(struct device *hwdev, u64 mask);
+extern void swiotlb_init(void);
+
+#ifdef CONFIG_SWIOTLB
+extern int swiotlb;
+#else
+#define swiotlb 0
+#endif
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-ia64/synch_bitops.h linux-2.6.18-xen-3.3.0/include/asm-ia64/synch_bitops.h
--- linux-2.6.18.8/include/asm-ia64/synch_bitops.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/synch_bitops.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,61 @@
+#ifndef __XEN_SYNCH_BITOPS_H__
+#define __XEN_SYNCH_BITOPS_H__
+
+/*
+ * Copyright 1992, Linus Torvalds.
+ * Heavily modified to provide guaranteed strong synchronisation
+ * when communicating with Xen or other guest OSes running on other CPUs.
+ */
+
+#define ADDR (*(volatile long *) addr)
+
+static __inline__ void synch_set_bit(int nr, volatile void * addr)
+{
+	set_bit(nr, addr);
+}
+
+static __inline__ void synch_clear_bit(int nr, volatile void * addr)
+{
+	clear_bit(nr, addr);
+}
+
+static __inline__ void synch_change_bit(int nr, volatile void * addr)
+{
+	change_bit(nr, addr);
+}
+
+static __inline__ int synch_test_and_set_bit(int nr, volatile void * addr)
+{
+    return test_and_set_bit(nr, addr);
+}
+
+static __inline__ int synch_test_and_clear_bit(int nr, volatile void * addr)
+{
+    return test_and_clear_bit(nr, addr);
+}
+
+static __inline__ int synch_test_and_change_bit(int nr, volatile void * addr)
+{
+    return test_and_change_bit(nr, addr);
+}
+
+static __inline__ int synch_const_test_bit(int nr, const volatile void * addr)
+{
+    return test_bit(nr, addr);
+}
+
+static __inline__ int synch_var_test_bit(int nr, volatile void * addr)
+{
+    return test_bit(nr, addr);
+}
+
+#define synch_cmpxchg	ia64_cmpxchg4_acq
+
+#define synch_test_bit(nr,addr) \
+(__builtin_constant_p(nr) ? \
+ synch_const_test_bit((nr),(addr)) : \
+ synch_var_test_bit((nr),(addr)))
+
+#define synch_cmpxchg_subword synch_cmpxchg
+
+#endif /* __XEN_SYNCH_BITOPS_H__ */
diff -rpuN linux-2.6.18.8/include/asm-ia64/system.h linux-2.6.18-xen-3.3.0/include/asm-ia64/system.h
--- linux-2.6.18.8/include/asm-ia64/system.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/system.h	2008-08-21 11:36:07.000000000 +0200
@@ -123,7 +123,7 @@ extern struct ia64_boot_param {
 #define __local_irq_save(x)			\
 do {						\
 	ia64_stop();				\
-	(x) = ia64_getreg(_IA64_REG_PSR);	\
+	(x) = ia64_get_psr_i();			\
 	ia64_stop();				\
 	ia64_rsm(IA64_PSR_I);			\
 } while (0)
@@ -171,7 +171,7 @@ do {								\
 #endif /* !CONFIG_IA64_DEBUG_IRQ */
 
 #define local_irq_enable()	({ ia64_stop(); ia64_ssm(IA64_PSR_I); ia64_srlz_d(); })
-#define local_save_flags(flags)	({ ia64_stop(); (flags) = ia64_getreg(_IA64_REG_PSR); })
+#define local_save_flags(flags)	({ ia64_stop(); (flags) = ia64_get_psr_i(); })
 
 #define irqs_disabled()				\
 ({						\
diff -rpuN linux-2.6.18.8/include/asm-ia64/xen/privop.h linux-2.6.18-xen-3.3.0/include/asm-ia64/xen/privop.h
--- linux-2.6.18.8/include/asm-ia64/xen/privop.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/xen/privop.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,550 @@
+#ifndef _ASM_IA64_XEN_PRIVOP_H
+#define _ASM_IA64_XEN_PRIVOP_H
+
+/*
+ * Copyright (C) 2005 Hewlett-Packard Co
+ *	Dan Magenheimer <dan.magenheimer@hp.com>
+ *
+ * Paravirtualizations of privileged operations for Xen/ia64
+ *
+ */
+
+#ifndef __ASSEMBLY__
+#include <linux/types.h>		/* arch-ia64.h requires uint64_t */
+#include <linux/stringify.h>
+#endif
+#include <xen/interface/arch-ia64.h>
+
+#define IA64_PARAVIRTUALIZED
+
+/* At 1 MB, before per-cpu space but still addressable using addl instead
+   of movl. */
+#define XSI_BASE				0xfffffffffff00000
+
+/* Address of mapped regs.  */
+#define XMAPPEDREGS_BASE		(XSI_BASE + XSI_SIZE)
+
+#ifdef __ASSEMBLY__
+#define	XEN_HYPER_RFI			break HYPERPRIVOP_RFI
+#define	XEN_HYPER_RSM_PSR_DT		break HYPERPRIVOP_RSM_DT
+#define	XEN_HYPER_SSM_PSR_DT		break HYPERPRIVOP_SSM_DT
+#define	XEN_HYPER_COVER			break HYPERPRIVOP_COVER
+#define	XEN_HYPER_ITC_D			break HYPERPRIVOP_ITC_D
+#define	XEN_HYPER_ITC_I			break HYPERPRIVOP_ITC_I
+#define	XEN_HYPER_SSM_I			break HYPERPRIVOP_SSM_I
+#define	XEN_HYPER_GET_IVR		break HYPERPRIVOP_GET_IVR
+#define	XEN_HYPER_GET_TPR		break HYPERPRIVOP_GET_TPR
+#define	XEN_HYPER_SET_TPR		break HYPERPRIVOP_SET_TPR
+#define	XEN_HYPER_EOI			break HYPERPRIVOP_EOI
+#define	XEN_HYPER_SET_ITM		break HYPERPRIVOP_SET_ITM
+#define	XEN_HYPER_THASH			break HYPERPRIVOP_THASH
+#define	XEN_HYPER_PTC_GA		break HYPERPRIVOP_PTC_GA
+#define	XEN_HYPER_ITR_D			break HYPERPRIVOP_ITR_D
+#define	XEN_HYPER_GET_RR		break HYPERPRIVOP_GET_RR
+#define	XEN_HYPER_SET_RR		break HYPERPRIVOP_SET_RR
+#define	XEN_HYPER_SET_KR		break HYPERPRIVOP_SET_KR
+#define	XEN_HYPER_FC			break HYPERPRIVOP_FC
+#define	XEN_HYPER_GET_CPUID		break HYPERPRIVOP_GET_CPUID
+#define	XEN_HYPER_GET_PMD		break HYPERPRIVOP_GET_PMD
+#define	XEN_HYPER_GET_EFLAG		break HYPERPRIVOP_GET_EFLAG
+#define	XEN_HYPER_SET_EFLAG		break HYPERPRIVOP_SET_EFLAG
+#define	XEN_HYPER_GET_PSR		break HYPERPRIVOP_GET_PSR
+
+#define XSI_IFS			(XSI_BASE + XSI_IFS_OFS)
+#define XSI_PRECOVER_IFS	(XSI_BASE + XSI_PRECOVER_IFS_OFS)
+#define XSI_IFA			(XSI_BASE + XSI_IFA_OFS)
+#define XSI_ISR			(XSI_BASE + XSI_ISR_OFS)
+#define XSI_IIM			(XSI_BASE + XSI_IIM_OFS)
+#define XSI_ITIR		(XSI_BASE + XSI_ITIR_OFS)
+#define XSI_PSR_I_ADDR		(XSI_BASE + XSI_PSR_I_ADDR_OFS)
+#define XSI_PSR_IC		(XSI_BASE + XSI_PSR_IC_OFS)
+#define XSI_IPSR		(XSI_BASE + XSI_IPSR_OFS)
+#define XSI_IIP			(XSI_BASE + XSI_IIP_OFS)
+#define XSI_B1NAT		(XSI_BASE + XSI_B1NATS_OFS)
+#define XSI_BANK1_R16		(XSI_BASE + XSI_BANK1_R16_OFS)
+#define XSI_BANKNUM		(XSI_BASE + XSI_BANKNUM_OFS)
+#define XSI_IHA			(XSI_BASE + XSI_IHA_OFS)
+#endif
+
+#ifndef __ASSEMBLY__
+#define	XEN_HYPER_SSM_I		asm("break %0" : : "i" (HYPERPRIVOP_SSM_I): "memory")
+
+/************************************************/
+/* Instructions paravirtualized for correctness */
+/************************************************/
+
+/* "fc" and "thash" are privilege-sensitive instructions, meaning they
+ *  may have different semantics depending on whether they are executed
+ *  at PL0 vs PL!=0.  When paravirtualized, these instructions mustn't
+ *  be allowed to execute directly, lest incorrect semantics result. */
+#ifdef ASM_SUPPORTED
+static inline void
+xen_fc(unsigned long addr)
+{
+	register __u64 __addr asm ("r8") = addr;
+	asm volatile ("break %0":: "i"(HYPERPRIVOP_FC), "r"(__addr): "memory");
+}
+
+static inline unsigned long
+xen_thash(unsigned long addr)
+{
+	register __u64 ia64_intri_res asm ("r8");
+	register __u64 __addr asm ("r8") = addr;
+	asm volatile ("break %1":
+		      "=r"(ia64_intri_res):
+		      "i"(HYPERPRIVOP_THASH), "0"(__addr));
+	return ia64_intri_res;
+}
+#else
+extern void xen_fc(unsigned long addr);
+extern unsigned long xen_thash(unsigned long addr);
+#endif
+
+#define ia64_fc(addr)							\
+do {									\
+	if (is_running_on_xen())					\
+		xen_fc((unsigned long)(addr));				\
+	else								\
+		__ia64_fc(addr);					\
+} while (0)
+
+#define ia64_thash(addr)						\
+({									\
+	unsigned long ia64_intri_res;					\
+	if (is_running_on_xen())					\
+		ia64_intri_res =					\
+			xen_thash((unsigned long)(addr));		\
+	else								\
+		ia64_intri_res = __ia64_thash(addr);			\
+	ia64_intri_res;							\
+})
+
+/* Note that "ttag" and "cover" are also privilege-sensitive; "ttag"
+ * is not currently used (though it may be in a long-format VHPT system!)
+ * and the semantics of cover only change if psr.ic is off which is very
+ * rare (and currently non-existent outside of assembly code */
+
+/* There are also privilege-sensitive registers.  These registers are
+ * readable at any privilege level but only writable at PL0. */
+#ifdef ASM_SUPPORTED
+static inline unsigned long
+xen_get_cpuid(int index)
+{
+	register __u64 ia64_intri_res asm ("r8");
+	register __u64 __index asm ("r8") = index;
+	asm volatile ("break %1":
+		      "=r"(ia64_intri_res):
+		      "i"(HYPERPRIVOP_GET_CPUID), "0"(__index));
+	return ia64_intri_res;
+}
+
+static inline unsigned long
+xen_get_pmd(int index)
+{
+	register __u64 ia64_intri_res asm ("r8");
+	register __u64 __index asm ("r8") = index;
+	asm volatile ("break %1":
+		      "=r"(ia64_intri_res):
+		      "i"(HYPERPRIVOP_GET_PMD), "0O"(__index));
+	return ia64_intri_res;
+}
+#else
+extern unsigned long xen_get_cpuid(int index);
+extern unsigned long xen_get_pmd(int index);
+#endif
+
+#define ia64_get_cpuid(i)						\
+({									\
+	unsigned long ia64_intri_res;					\
+	if (is_running_on_xen())					\
+		ia64_intri_res = xen_get_cpuid(i);			\
+	else								\
+		ia64_intri_res = __ia64_get_cpuid(i);			\
+	ia64_intri_res;							\
+})
+
+#define ia64_get_pmd(i)						\
+({									\
+	unsigned long ia64_intri_res;					\
+	if (is_running_on_xen())					\
+		ia64_intri_res = xen_get_pmd(i);			\
+	else								\
+		ia64_intri_res = __ia64_get_pmd(i);			\
+	ia64_intri_res;							\
+})
+
+#ifdef ASM_SUPPORTED
+static inline unsigned long
+xen_get_eflag(void)
+{
+	register __u64 ia64_intri_res asm ("r8");
+	asm volatile ("break %1":
+		      "=r"(ia64_intri_res): "i"(HYPERPRIVOP_GET_EFLAG));
+	return ia64_intri_res;
+}
+
+static inline void
+xen_set_eflag(unsigned long val)
+{
+	register __u64 __val asm ("r8") = val;
+	asm volatile ("break %0":: "i"(HYPERPRIVOP_SET_EFLAG), "r"(__val): "memory");
+}
+#else
+extern unsigned long xen_get_eflag(void);	/* see xen_ia64_getreg */
+extern void xen_set_eflag(unsigned long);	/* see xen_ia64_setreg */
+#endif
+
+/************************************************/
+/* Instructions paravirtualized for performance */
+/************************************************/
+
+/* Xen uses memory-mapped virtual privileged registers for access to many
+ * performance-sensitive privileged registers.  Some, like the processor
+ * status register (psr), are broken up into multiple memory locations.
+ * Others, like "pend", are abstractions based on privileged registers.
+ * "Pend" is guaranteed to be set if reading cr.ivr would return a
+ * (non-spurious) interrupt. */
+#define XEN_MAPPEDREGS ((struct mapped_regs *)XMAPPEDREGS_BASE)
+#define XSI_PSR_I			\
+	(*XEN_MAPPEDREGS->interrupt_mask_addr)
+#define xen_get_virtual_psr_i()		\
+	(!XSI_PSR_I)
+#define xen_set_virtual_psr_i(_val)	\
+	({ XSI_PSR_I = (uint8_t)(_val) ? 0 : 1; })
+#define xen_set_virtual_psr_ic(_val)	\
+	({ XEN_MAPPEDREGS->interrupt_collection_enabled = _val ? 1 : 0; })
+#define xen_get_virtual_pend()		\
+	(*(((uint8_t *)XEN_MAPPEDREGS->interrupt_mask_addr) - 1))
+
+/* Hyperprivops are "break" instructions with a well-defined API.
+ * In particular, the virtual psr.ic bit must be off; in this way
+ * it is guaranteed to never conflict with a linux break instruction.
+ * Normally, this is done in a xen stub but this one is frequent enough
+ * that we inline it */
+#define xen_hyper_ssm_i()						\
+({									\
+	XEN_HYPER_SSM_I;						\
+})
+
+/* turning off interrupts can be paravirtualized simply by writing
+ * to a memory-mapped virtual psr.i bit (implemented as a 16-bit bool) */
+#define xen_rsm_i()							\
+{									\
+	xen_set_virtual_psr_i(0);					\
+	barrier();							\
+}
+
+/* turning on interrupts is a bit more complicated.. write to the
+ * memory-mapped virtual psr.i bit first (to avoid race condition),
+ * then if any interrupts were pending, we have to execute a hyperprivop
+ * to ensure the pending interrupt gets delivered; else we're done! */
+#define xen_ssm_i()							\
+({									\
+	int old = xen_get_virtual_psr_i();				\
+	xen_set_virtual_psr_i(1);					\
+	barrier();							\
+	if (!old && xen_get_virtual_pend())				\
+		xen_hyper_ssm_i();					\
+})
+
+#define xen_ia64_intrin_local_irq_restore(x)				\
+{									\
+     if (is_running_on_xen()) {						\
+	if ((x) & IA64_PSR_I) { xen_ssm_i(); }				\
+	else { xen_rsm_i(); }						\
+    }									\
+    else __ia64_intrin_local_irq_restore((x));				\
+}
+
+#define	xen_get_psr_i()							\
+(									\
+	(is_running_on_xen()) ?						\
+		(xen_get_virtual_psr_i() ? IA64_PSR_I : 0)		\
+		: __ia64_get_psr_i()					\
+)
+
+#define xen_ia64_ssm(mask)						\
+{									\
+	if ((mask)==IA64_PSR_I) {					\
+		if (is_running_on_xen()) { xen_ssm_i(); }		\
+		else { __ia64_ssm(mask); }				\
+	}								\
+	else { __ia64_ssm(mask); }					\
+}
+
+#define xen_ia64_rsm(mask)						\
+{									\
+	if ((mask)==IA64_PSR_I) {					\
+		if (is_running_on_xen()) { xen_rsm_i(); }		\
+		else { __ia64_rsm(mask); }				\
+	}								\
+	else { __ia64_rsm(mask); }					\
+}
+
+
+/* Although all privileged operations can be left to trap and will
+ * be properly handled by Xen, some are frequent enough that we use
+ * hyperprivops for performance. */
+
+#ifndef ASM_SUPPORTED 
+extern unsigned long xen_get_psr(void);
+extern unsigned long xen_get_ivr(void);
+extern unsigned long xen_get_tpr(void);
+extern void xen_set_itm(unsigned long);
+extern void xen_set_tpr(unsigned long);
+extern void xen_eoi(unsigned long);
+extern void xen_set_rr(unsigned long index, unsigned long val);
+extern unsigned long xen_get_rr(unsigned long index);
+extern void xen_set_kr(unsigned long index, unsigned long val);
+extern void xen_ptcga(unsigned long addr, unsigned long size);
+#else
+static inline unsigned long
+xen_get_psr(void)
+{
+	register __u64 ia64_intri_res asm ("r8");
+	asm volatile ("break %1":
+		      "=r"(ia64_intri_res): "i"(HYPERPRIVOP_GET_PSR));
+	return ia64_intri_res;
+}
+
+static inline unsigned long
+xen_get_ivr(void)
+{
+	register __u64 ia64_intri_res asm ("r8");
+	asm volatile ("break %1":
+		      "=r"(ia64_intri_res): "i"(HYPERPRIVOP_GET_IVR));
+	return ia64_intri_res;
+}
+
+static inline unsigned long
+xen_get_tpr(void)
+{
+	register __u64 ia64_intri_res asm ("r8");
+	asm volatile ("break %1":
+		      "=r"(ia64_intri_res): "i"(HYPERPRIVOP_GET_TPR));
+	return ia64_intri_res;
+}
+
+static inline void
+xen_set_tpr(unsigned long val)
+{
+	register __u64 __val asm ("r8") = val;
+	asm volatile ("break %0"::
+		      "i"(HYPERPRIVOP_GET_TPR), "r"(__val): "memory");
+}
+
+static inline void
+xen_eoi(unsigned long val)
+{
+	register __u64 __val asm ("r8") = val;
+	asm volatile ("break %0"::
+		      "i"(HYPERPRIVOP_EOI), "r"(__val): "memory");
+}
+
+static inline void
+xen_set_itm(unsigned long val)
+{
+	register __u64 __val asm ("r8") = val;
+	asm volatile ("break %0":: "i"(HYPERPRIVOP_SET_ITM), "r"(__val): "memory");
+}
+
+static inline void
+xen_ptcga(unsigned long addr, unsigned long size)
+{
+	register __u64 __addr asm ("r8") = addr;
+	register __u64 __size asm ("r9") = size;
+	asm volatile ("break %0"::
+		      "i"(HYPERPRIVOP_PTC_GA), "r"(__addr), "r"(__size): "memory");
+}
+
+static inline unsigned long
+xen_get_rr(unsigned long index)
+{
+	register __u64 ia64_intri_res asm ("r8");
+	register __u64 __index asm ("r8") = index;
+	asm volatile ("break %1":
+		      "=r"(ia64_intri_res):
+		      "i"(HYPERPRIVOP_GET_RR), "0"(__index));
+	return ia64_intri_res;
+}
+
+static inline void
+xen_set_rr(unsigned long index, unsigned long val)
+{
+	register __u64 __index asm ("r8") = index;
+	register __u64 __val asm ("r9") = val;
+	asm volatile ("break %0"::
+		      "i"(HYPERPRIVOP_SET_RR), "r"(__index), "r"(__val): "memory");
+}
+
+static inline void
+xen_set_rr0_to_rr4(unsigned long val0, unsigned long val1,
+		   unsigned long val2, unsigned long val3, unsigned long val4)
+{
+	register __u64 __val0 asm ("r8") = val0;
+	register __u64 __val1 asm ("r9") = val1;
+	register __u64 __val2 asm ("r10") = val2;
+	register __u64 __val3 asm ("r11") = val3;
+	register __u64 __val4 asm ("r14") = val4;
+	asm volatile ("break %0" ::
+		      "i"(HYPERPRIVOP_SET_RR0_TO_RR4),
+		      "r"(__val0), "r"(__val1),
+		      "r"(__val2), "r"(__val3), "r"(__val4): "memory");
+}
+
+static inline void
+xen_set_kr(unsigned long index, unsigned long val)
+{
+	register __u64 __index asm ("r8") = index;
+	register __u64 __val asm ("r9") = val;
+	asm volatile ("break %0"::
+		      "i"(HYPERPRIVOP_SET_KR), "r"(__index), "r"(__val): "memory");
+}
+#endif
+
+/* Note: It may look wrong to test for is_running_on_xen() in each case.
+ * However regnum is always a constant so, as written, the compiler
+ * eliminates the switch statement, whereas is_running_on_xen() must be
+ * tested dynamically. */
+#define xen_ia64_getreg(regnum)						\
+({									\
+	__u64 ia64_intri_res;						\
+									\
+	switch(regnum) {						\
+	case _IA64_REG_PSR:						\
+		ia64_intri_res = (is_running_on_xen()) ?		\
+			xen_get_psr() :					\
+			__ia64_getreg(regnum);				\
+		break;							\
+	case _IA64_REG_CR_IVR:						\
+		ia64_intri_res = (is_running_on_xen()) ?		\
+			xen_get_ivr() :					\
+			__ia64_getreg(regnum);				\
+		break;							\
+	case _IA64_REG_CR_TPR:						\
+		ia64_intri_res = (is_running_on_xen()) ?		\
+			xen_get_tpr() :					\
+			__ia64_getreg(regnum);				\
+		break;							\
+	case _IA64_REG_AR_EFLAG:					\
+		ia64_intri_res = (is_running_on_xen()) ?		\
+			xen_get_eflag() :				\
+			__ia64_getreg(regnum);				\
+		break;							\
+	default:							\
+		ia64_intri_res = __ia64_getreg(regnum);			\
+		break;							\
+	}								\
+	ia64_intri_res;							\
+})
+
+#define xen_ia64_setreg(regnum,val)					\
+({									\
+	switch(regnum) {						\
+	case _IA64_REG_AR_KR0 ... _IA64_REG_AR_KR7:			\
+		(is_running_on_xen()) ?					\
+			xen_set_kr((regnum-_IA64_REG_AR_KR0), val) :	\
+			__ia64_setreg(regnum,val);			\
+		break;							\
+	case _IA64_REG_CR_ITM:						\
+		(is_running_on_xen()) ?					\
+			xen_set_itm(val) :				\
+			__ia64_setreg(regnum,val);			\
+		break;							\
+	case _IA64_REG_CR_TPR:						\
+		(is_running_on_xen()) ?					\
+			xen_set_tpr(val) :				\
+			__ia64_setreg(regnum,val);			\
+		break;							\
+	case _IA64_REG_CR_EOI:						\
+		(is_running_on_xen()) ?					\
+			xen_eoi(val) :					\
+			__ia64_setreg(regnum,val);			\
+		break;							\
+	case _IA64_REG_AR_EFLAG:					\
+		(is_running_on_xen()) ?					\
+			xen_set_eflag(val) :				\
+			__ia64_setreg(regnum,val);			\
+		break;							\
+	default:							\
+		__ia64_setreg(regnum,val);				\
+		break;							\
+	}								\
+})
+
+#define ia64_ptcga(addr, size)						\
+do {									\
+	if (is_running_on_xen())					\
+		xen_ptcga((addr), (size));				\
+	else								\
+		__ia64_ptcga((addr), (size));				\
+} while (0)
+
+#define ia64_set_rr(index, val)						\
+do {									\
+	if (is_running_on_xen())					\
+		xen_set_rr((index), (val));				\
+	else								\
+		__ia64_set_rr((index), (val));				\
+} while (0)
+
+#define ia64_get_rr(index)						\
+({									\
+	__u64 ia64_intri_res;						\
+	if (is_running_on_xen())					\
+		ia64_intri_res = xen_get_rr((index));			\
+	else								\
+		ia64_intri_res = __ia64_get_rr((index));		\
+	ia64_intri_res;							\
+})
+
+#define ia64_set_rr0_to_rr4(val0, val1, val2, val3, val4)		\
+do {									\
+	if (is_running_on_xen())					\
+		xen_set_rr0_to_rr4((val0), (val1), (val2),		\
+				   (val3), (val4));			\
+	else								\
+		__ia64_set_rr0_to_rr4((val0), (val1), (val2),		\
+				      (val3), (val4));			\
+} while (0)
+
+#define ia64_getreg			xen_ia64_getreg
+#define ia64_setreg			xen_ia64_setreg
+#define ia64_ssm			xen_ia64_ssm
+#define ia64_rsm			xen_ia64_rsm
+#define ia64_intrin_local_irq_restore	xen_ia64_intrin_local_irq_restore
+#define	ia64_get_psr_i			xen_get_psr_i
+
+/* the remainder of these are not performance-sensitive so its
+ * OK to not paravirtualize and just take a privop trap and emulate */
+#define ia64_hint			__ia64_hint
+#define ia64_set_pmd			__ia64_set_pmd
+#define ia64_itci			__ia64_itci
+#define ia64_itcd			__ia64_itcd
+#define ia64_itri			__ia64_itri
+#define ia64_itrd			__ia64_itrd
+#define ia64_tpa			__ia64_tpa
+#define ia64_set_ibr			__ia64_set_ibr
+#define ia64_set_pkr			__ia64_set_pkr
+#define ia64_set_pmc			__ia64_set_pmc
+#define ia64_get_ibr			__ia64_get_ibr
+#define ia64_get_pkr			__ia64_get_pkr
+#define ia64_get_pmc			__ia64_get_pmc
+#define ia64_ptce			__ia64_ptce
+#define ia64_ptcl			__ia64_ptcl
+#define ia64_ptri			__ia64_ptri
+#define ia64_ptrd			__ia64_ptrd
+
+#endif /* !__ASSEMBLY__ */
+
+/* these routines utilize privilege-sensitive or performance-sensitive
+ * privileged instructions so the code must be replaced with
+ * paravirtualized versions */
+#define	ia64_leave_kernel		xen_leave_kernel
+#define	ia64_leave_syscall		xen_leave_syscall
+#define	ia64_trace_syscall		xen_trace_syscall
+#define	ia64_ret_from_clone		xen_ret_from_clone
+#define	ia64_switch_to			xen_switch_to
+#define	ia64_pal_call_static		xen_pal_call_static
+
+#endif /* _ASM_IA64_XEN_PRIVOP_H */
diff -rpuN linux-2.6.18.8/include/asm-ia64/xen/xcom_hcall.h linux-2.6.18-xen-3.3.0/include/asm-ia64/xen/xcom_hcall.h
--- linux-2.6.18.8/include/asm-ia64/xen/xcom_hcall.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/xen/xcom_hcall.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,67 @@
+/*
+ * Copyright (C) 2006 Tristan Gingold <tristan.gingold@bull.net>, Bull SAS
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ * 
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ * 
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ */
+
+#ifndef _LINUX_XENCOMM_HCALL_H_
+#define _LINUX_XENCOMM_HCALL_H_
+
+/* These function creates inline or mini descriptor for the parameters and
+   calls the corresponding xencomm_arch_hypercall_X.
+   Architectures should defines HYPERVISOR_xxx as xencomm_hypercall_xxx unless
+   they want to use their own wrapper.  */
+extern int xencomm_hypercall_console_io(int cmd, int count, char *str);
+
+extern int xencomm_hypercall_event_channel_op(int cmd, void *op);
+
+extern int xencomm_hypercall_xen_version(int cmd, void *arg);
+
+extern int xencomm_hypercall_physdev_op(int cmd, void *op);
+
+extern int xencomm_hypercall_grant_table_op(unsigned int cmd, void *op,
+                                            unsigned int count);
+
+extern int xencomm_hypercall_sched_op(int cmd, void *arg);
+
+extern int xencomm_hypercall_multicall(void *call_list, int nr_calls);
+
+extern int xencomm_hypercall_callback_op(int cmd, void *arg);
+
+extern int xencomm_hypercall_memory_op(unsigned int cmd, void *arg);
+
+extern unsigned long xencomm_hypercall_hvm_op(int cmd, void *arg);
+
+extern int xencomm_hypercall_suspend(unsigned long srec);
+
+extern int xencomm_hypercall_xenoprof_op(int op, void *arg);
+
+extern int xencomm_hypercall_perfmon_op(unsigned long cmd, void* arg,
+                                        unsigned long count);
+
+extern long xencomm_hypercall_vcpu_op(int cmd, int cpu, void *arg);
+
+extern long xencomm_hypercall_opt_feature(void *arg);
+
+/* For privcmd.  Locally declare argument type to avoid include storm.
+   Type coherency will be checked within privcmd.c  */
+struct privcmd_hypercall;
+extern int privcmd_hypercall(struct privcmd_hypercall *hypercall);
+
+extern int xen_foreign_p2m_expose(struct privcmd_hypercall *hypercall);
+
+extern int xencomm_hypercall_kexec_op(int cmd, void *arg);
+
+#endif /* _LINUX_XENCOMM_HCALL_H_ */
diff -rpuN linux-2.6.18.8/include/asm-ia64/xen/xencomm.h linux-2.6.18-xen-3.3.0/include/asm-ia64/xen/xencomm.h
--- linux-2.6.18.8/include/asm-ia64/xen/xencomm.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/xen/xencomm.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,33 @@
+/*
+ * Copyright (C) 2006 Hollis Blanchard <hollisb@us.ibm.com>, IBM Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ * 
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ * 
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ */
+
+#ifndef _ASM_IA64_XENCOMM_H_
+#define _ASM_IA64_XENCOMM_H_
+
+#define is_kernel_addr(x)					\
+	((PAGE_OFFSET <= (x) &&					\
+	  (x) < (PAGE_OFFSET + (1UL << IA64_MAX_PHYS_BITS))) ||	\
+	 (KERNEL_START <= (x) &&				\
+	  (x) < KERNEL_START + KERNEL_TR_PAGE_SIZE))
+
+/* Must be called before any hypercall.  */
+extern void xencomm_initialize (void);
+
+#include <xen/xencomm.h>
+
+#endif /* _ASM_IA64_XENCOMM_H_ */
diff -rpuN linux-2.6.18.8/include/asm-ia64/xenoprof.h linux-2.6.18-xen-3.3.0/include/asm-ia64/xenoprof.h
--- linux-2.6.18.8/include/asm-ia64/xenoprof.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-ia64/xenoprof.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,48 @@
+/******************************************************************************
+ * asm-ia64/xenoprof.h
+ *
+ * Copyright (c) 2006 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+#ifndef __ASM_XENOPROF_H__
+#define __ASM_XENOPROF_H__
+#ifdef CONFIG_XEN
+
+#undef HAVE_XENOPROF_CREATE_FILES
+
+struct xenoprof_init;
+void xenoprof_arch_init_counter(struct xenoprof_init *init);
+void xenoprof_arch_counter(void);
+void xenoprof_arch_start(void);
+void xenoprof_arch_stop(void);
+
+struct xenoprof_arch_shared_buffer {
+	struct resource*	res;
+};
+
+struct xenoprof_shared_buffer;
+void xenoprof_arch_unmap_shared_buffer(struct xenoprof_shared_buffer *sbuf);
+struct xenoprof_get_buffer;
+int xenoprof_arch_map_shared_buffer(struct xenoprof_get_buffer *get_buffer,
+                                    struct xenoprof_shared_buffer *sbuf);
+struct xenoprof_passive;
+int xenoprof_arch_set_passive(struct xenoprof_passive *pdomain,
+                              struct xenoprof_shared_buffer *sbuf);
+
+#endif /* CONFIG_XEN */
+#endif /* __ASM_XENOPROF_H__ */
diff -rpuN linux-2.6.18.8/include/asm-powerpc/mpic.h linux-2.6.18-xen-3.3.0/include/asm-powerpc/mpic.h
--- linux-2.6.18.8/include/asm-powerpc/mpic.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-powerpc/mpic.h	2008-08-21 11:36:07.000000000 +0200
@@ -305,6 +305,8 @@ struct mpic
 #define MPIC_SPV_EOI			0x00000020
 /* No passthrough disable */
 #define MPIC_NO_PTHROU_DIS		0x00000040
+/* Skip reset of IPI vectors during init */
+#define MPIC_SKIP_IPI_INIT		0x00000080
 
 /* MPIC HW modification ID */
 #define MPIC_REGSET_MASK		0xf0000000
diff -rpuN linux-2.6.18.8/include/asm-powerpc/udbg.h linux-2.6.18-xen-3.3.0/include/asm-powerpc/udbg.h
--- linux-2.6.18.8/include/asm-powerpc/udbg.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-powerpc/udbg.h	2008-08-21 11:36:07.000000000 +0200
@@ -42,6 +42,7 @@ extern void __init udbg_init_debug_lpar(
 extern void __init udbg_init_pmac_realmode(void);
 extern void __init udbg_init_maple_realmode(void);
 extern void __init udbg_init_iseries(void);
+extern void __init udbg_init_xen(void);
 extern void __init udbg_init_rtas_panel(void);
 extern void __init udbg_init_rtas_console(void);
 
diff -rpuN linux-2.6.18.8/include/asm-powerpc/xen/asm/gnttab_dma.h linux-2.6.18-xen-3.3.0/include/asm-powerpc/xen/asm/gnttab_dma.h
--- linux-2.6.18.8/include/asm-powerpc/xen/asm/gnttab_dma.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-powerpc/xen/asm/gnttab_dma.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,29 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ * Copyright 2007 IBM Corp.
+ *
+ * Authors: Hollis Blanchard <hollisb@us.ibm.com>
+ */
+
+#ifndef _ASM_PPC_GNTTAB_DMA_H
+#define _ASM_PPC_GNTTAB_DMA_H
+
+static inline int gnttab_dma_local_pfn(struct page *page)
+{
+	return 0;
+}
+
+#endif /* _ASM_PPC_GNTTAB_DMA_H */
diff -rpuN linux-2.6.18.8/include/asm-powerpc/xen/asm/hypercall.h linux-2.6.18-xen-3.3.0/include/asm-powerpc/xen/asm/hypercall.h
--- linux-2.6.18.8/include/asm-powerpc/xen/asm/hypercall.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-powerpc/xen/asm/hypercall.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,90 @@
+/******************************************************************************
+ * hypercall.h
+ * 
+ * Linux-specific hypervisor handling.
+ * 
+ * Copyright (c) 2002-2004, K A Fraser
+ * 
+ * This file may be distributed separately from the Linux kernel, or
+ * incorporated into other software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Copyright 2007 IBM Corp.
+ *
+ * Authors: Hollis Blanchard <hollisb@us.ibm.com>
+ *          Jimi Xenidis <jimix@watson.ibm.com>
+ */
+
+#ifndef __HYPERCALL_H__
+#define __HYPERCALL_H__
+
+#include <asm/hvcall.h>
+#include <asm/page.h>
+#include <xen/xencomm.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/sched.h>
+
+#define XEN_MARK(a)((a) | (~0UL << 16))
+
+extern int HYPERVISOR_console_io(int cmd, int count, char *str);
+extern int HYPERVISOR_event_channel_op(int cmd, void *op);
+extern int HYPERVISOR_xen_version(int cmd, void *arg);
+extern int HYPERVISOR_physdev_op(int cmd, void *op);
+extern int HYPERVISOR_grant_table_op(unsigned int cmd, void *uop,
+		unsigned int count);
+extern int HYPERVISOR_vcpu_op(int cmd, int vcpuid, void *extra_args);
+extern int HYPERVISOR_memory_op(unsigned int cmd, void *arg);
+extern int HYPERVISOR_multicall(void *call_list, int nr_calls);
+
+extern int HYPERVISOR_sched_op(int cmd, void *arg);
+extern int HYPERVISOR_poll(
+	evtchn_port_t *ports, unsigned int nr_ports, u64 timeout);
+
+static inline int HYPERVISOR_shutdown(unsigned int reason)
+{
+	struct sched_shutdown sched_shutdown = {
+		.reason = reason
+	};
+
+	return HYPERVISOR_sched_op(SCHEDOP_shutdown, &sched_shutdown);
+}
+
+static inline int HYPERVISOR_set_timer_op(unsigned long arg)
+{
+	return plpar_hcall_norets(XEN_MARK(__HYPERVISOR_set_timer_op), arg);
+}
+
+extern int HYPERVISOR_suspend(unsigned long srec);
+extern int HYPERVISOR_kexec_op(unsigned long op, void *args);
+static inline unsigned long HYPERVISOR_hvm_op(int op, void *arg) {
+	return -ENOSYS;
+}
+
+static inline int
+HYPERVISOR_mmu_update(
+	mmu_update_t *req, int count, int *success_count, domid_t domid)
+{
+	return -ENOSYS;
+}
+
+struct privcmd_hypercall;
+extern int privcmd_hypercall(struct privcmd_hypercall *hypercall);
+
+#endif	/*  __HYPERCALL_H__ */
diff -rpuN linux-2.6.18.8/include/asm-powerpc/xen/asm/hypervisor.h linux-2.6.18-xen-3.3.0/include/asm-powerpc/xen/asm/hypervisor.h
--- linux-2.6.18.8/include/asm-powerpc/xen/asm/hypervisor.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-powerpc/xen/asm/hypervisor.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,276 @@
+/******************************************************************************
+ * hypervisor.h
+ * 
+ * Linux-specific hypervisor handling.
+ * 
+ * Copyright (c) 2002-2004, K A Fraser
+ * 
+ * This file may be distributed separately from the Linux kernel, or
+ * incorporated into other software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __HYPERVISOR_H__
+#define __HYPERVISOR_H__
+
+#include <linux/config.h>
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/version.h>
+#include <xen/interface/xen.h>
+#include <asm/ptrace.h>
+#include <asm/page.h>
+#include <asm/irq.h>
+
+extern shared_info_t *HYPERVISOR_shared_info;
+
+/* arch/xen/i386/kernel/setup.c */
+extern start_info_t *xen_start_info;
+
+#ifdef CONFIG_XEN_PRIVILEGED_GUEST
+#define is_initial_xendomain() (xen_start_info && \
+				(xen_start_info->flags & SIF_INITDOMAIN))
+#else
+#define is_initial_xendomain() 0
+#endif
+
+/* arch/xen/kernel/evtchn.c */
+/* Force a proper event-channel callback from Xen. */
+void force_evtchn_callback(void);
+
+/* arch/xen/kernel/process.c */
+void xen_cpu_idle (void);
+
+/* arch/xen/i386/kernel/hypervisor.c */
+void do_hypervisor_callback(struct pt_regs *regs);
+
+/* arch/xen/i386/kernel/head.S */
+void lgdt_finish(void);
+
+/* arch/xen/i386/mm/hypervisor.c */
+/*
+ * NB. ptr values should be PHYSICAL, not MACHINE. 'vals' should be already
+ * be MACHINE addresses.
+ */
+
+void xen_pt_switch(unsigned long ptr);
+void xen_new_user_pt(unsigned long ptr); /* x86_64 only */
+void xen_load_gs(unsigned int selector); /* x86_64 only */
+void xen_tlb_flush(void);
+void xen_invlpg(unsigned long ptr);
+
+#ifndef CONFIG_XEN_SHADOW_MODE
+void xen_l1_entry_update(pte_t *ptr, pte_t val);
+void xen_l2_entry_update(pmd_t *ptr, pmd_t val);
+void xen_l3_entry_update(pud_t *ptr, pud_t val); /* x86_64/PAE */
+void xen_l4_entry_update(pgd_t *ptr, pgd_t val); /* x86_64 only */
+void xen_pgd_pin(unsigned long ptr);
+void xen_pgd_unpin(unsigned long ptr);
+void xen_pud_pin(unsigned long ptr); /* x86_64 only */
+void xen_pud_unpin(unsigned long ptr); /* x86_64 only */
+void xen_pmd_pin(unsigned long ptr); /* x86_64 only */
+void xen_pmd_unpin(unsigned long ptr); /* x86_64 only */
+void xen_pte_pin(unsigned long ptr);
+void xen_pte_unpin(unsigned long ptr);
+#else
+#define xen_l1_entry_update(_p, _v) set_pte((_p), (_v))
+#define xen_l2_entry_update(_p, _v) set_pgd((_p), (_v))
+#define xen_pgd_pin(_p)   ((void)0)
+#define xen_pgd_unpin(_p) ((void)0)
+#define xen_pte_pin(_p)   ((void)0)
+#define xen_pte_unpin(_p) ((void)0)
+#endif
+
+void xen_set_ldt(unsigned long ptr, unsigned long bytes);
+void xen_machphys_update(unsigned long mfn, unsigned long pfn);
+
+#ifdef CONFIG_SMP
+#include <linux/cpumask.h>
+void xen_tlb_flush_all(void);
+void xen_invlpg_all(unsigned long ptr);
+void xen_tlb_flush_mask(cpumask_t *mask);
+void xen_invlpg_mask(cpumask_t *mask, unsigned long ptr);
+#endif
+
+/* Returns zero on success else negative errno. */
+static inline int xen_create_contiguous_region(
+    unsigned long vstart, unsigned int order, unsigned int address_bits)
+{
+	return 0;
+}
+static inline void xen_destroy_contiguous_region(
+    unsigned long vstart, unsigned int order)
+{
+	return;
+}
+
+#include <asm/hypercall.h>
+
+/* BEGIN: all of these need a new home */
+struct vm_area_struct;
+int direct_remap_pfn_range(struct vm_area_struct *vma,  unsigned long address,
+			   unsigned long mfn, unsigned long size,
+			   pgprot_t prot, domid_t  domid);
+#define	pfn_to_mfn(x)	(x)
+#define	mfn_to_pfn(x)	(x)
+#define phys_to_machine(phys) ((maddr_t)(phys))
+#define phys_to_machine_mapping_valid(pfn) (1)
+
+/* VIRT <-> MACHINE conversion */
+#define virt_to_machine(v)	(phys_to_machine(__pa(v)))
+#define machine_to_virt(m)	(__va(m))
+#define virt_to_mfn(v)		(pfn_to_mfn(__pa(v) >> PAGE_SHIFT))
+#define mfn_to_virt(m)		(__va(mfn_to_pfn(m) << PAGE_SHIFT))
+
+
+#define PIRQ_BASE		0
+#define NR_PIRQS		256
+
+#define DYNIRQ_BASE		(PIRQ_BASE + NR_PIRQS)
+#define NR_DYNIRQS		256
+
+#define NR_IPIS 4		/* PPC_MSG_DEBUGGER_BREAK + 1 */
+
+#if NR_IRQS < (NR_PIRQS + NR_DYNIRQS)
+#error to many Xen IRQs
+#endif
+
+#define NR_IRQ_VECTORS		NR_IRQS
+
+#define pirq_to_irq(_x)		((_x) + PIRQ_BASE)
+#define irq_to_pirq(_x)		((_x) - PIRQ_BASE)
+
+#define dynirq_to_irq(_x)	((_x) + DYNIRQ_BASE)
+#define irq_to_dynirq(_x)	((_x) - DYNIRQ_BASE)
+
+
+/* END:  all of these need a new home */
+
+#if defined(CONFIG_X86_64)
+#define MULTI_UVMFLAGS_INDEX 2
+#define MULTI_UVMDOMID_INDEX 3
+#else
+#define MULTI_UVMFLAGS_INDEX 3
+#define MULTI_UVMDOMID_INDEX 4
+#endif
+
+extern int is_running_on_xen(void);
+
+static inline void
+MULTI_update_va_mapping(
+    multicall_entry_t *mcl, unsigned long va,
+    pte_t new_val, unsigned long flags)
+{
+    mcl->op = __HYPERVISOR_update_va_mapping;
+    mcl->args[0] = va;
+#if defined(CONFIG_X86_64)
+    mcl->args[1] = new_val.pte;
+    mcl->args[2] = flags;
+#elif defined(CONFIG_X86_PAE)
+    mcl->args[1] = new_val.pte_low;
+    mcl->args[2] = new_val.pte_high;
+    mcl->args[3] = flags;
+#elif defined(CONFIG_PPC64)
+    mcl->args[1] = pte_val(new_val);
+    mcl->args[2] = 0;
+    mcl->args[3] = flags;
+#else
+    mcl->args[1] = new_val.pte_low;
+    mcl->args[2] = 0;
+    mcl->args[3] = flags;
+#endif
+}
+
+static inline void
+MULTI_update_va_mapping_otherdomain(
+    multicall_entry_t *mcl, unsigned long va,
+    pte_t new_val, unsigned long flags, domid_t domid)
+{
+    mcl->op = __HYPERVISOR_update_va_mapping_otherdomain;
+    mcl->args[0] = va;
+#if defined(CONFIG_X86_64)
+    mcl->args[1] = new_val.pte;
+    mcl->args[2] = flags;
+    mcl->args[3] = domid;
+#elif defined(CONFIG_X86_PAE)
+    mcl->args[1] = new_val.pte_low;
+    mcl->args[2] = new_val.pte_high;
+    mcl->args[3] = flags;
+    mcl->args[4] = domid;
+#elif defined(CONFIG_PPC64)
+    mcl->args[1] = pte_val(new_val);
+    mcl->args[2] = 0;
+    mcl->args[3] = flags;
+    mcl->args[4] = domid;
+#else
+    mcl->args[1] = new_val.pte_low;
+    mcl->args[2] = 0;
+    mcl->args[3] = flags;
+    mcl->args[4] = domid;
+#endif
+}
+
+#define INVALID_P2M_ENTRY (~0UL)
+#define FOREIGN_FRAME(m) (INVALID_P2M_ENTRY)
+static inline void set_phys_to_machine(unsigned long pfn, unsigned long mfn)
+{
+	if (pfn != mfn && mfn != INVALID_P2M_ENTRY)
+		printk(KERN_EMERG "%s: pfn: 0x%lx mfn: 0x%lx\n",
+		       __func__, pfn, mfn);
+	
+	return;
+}
+#define pfn_pte_ma(pfn, prot)	__pte(((pfn) << PAGE_SHIFT) | pgprot_val(prot))
+
+typedef unsigned long maddr_t;
+typedef unsigned long paddr_t;
+
+#ifdef CONFIG_XEN_SCRUB_PAGES
+
+static inline void scrub_pages(void *p, unsigned n)
+{
+	unsigned i;
+
+	for (i = 0; i < n; i++) {
+		clear_page(p);
+		p += PAGE_SIZE;
+	}
+}
+#else
+#define scrub_pages(_p,_n) ((void)0)
+#endif
+
+/*
+ * for blktap.c
+ * int create_lookup_pte_addr(struct mm_struct *mm, 
+ *                            unsigned long address,
+ *                            uint64_t *ptep);
+ */
+#define create_lookup_pte_addr(mm, address, ptep)			\
+	({								\
+		printk(KERN_EMERG					\
+		       "%s:%d "						\
+		       "create_lookup_pte_addr() isn't supported.\n",	\
+		       __func__, __LINE__);				\
+		BUG();							\
+		(-ENOSYS);						\
+	})
+
+#endif /* __HYPERVISOR_H__ */
diff -rpuN linux-2.6.18.8/include/asm-powerpc/xen/asm/maddr.h linux-2.6.18-xen-3.3.0/include/asm-powerpc/xen/asm/maddr.h
--- linux-2.6.18.8/include/asm-powerpc/xen/asm/maddr.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-powerpc/xen/asm/maddr.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,7 @@
+#ifndef _POWERPC_MADDR_H
+#define _POWERPC_MADDR_H
+
+#include <xen/features.h>
+#include <xen/interface/xen.h>
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-powerpc/xen/asm/synch_bitops.h linux-2.6.18-xen-3.3.0/include/asm-powerpc/xen/asm/synch_bitops.h
--- linux-2.6.18.8/include/asm-powerpc/xen/asm/synch_bitops.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-powerpc/xen/asm/synch_bitops.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,100 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ * Copyright 2006 IBM Corp.
+ *
+ * Authors: Jimi Xenidis <jimix@watson.ibm.com>
+ */
+
+#ifndef  __SYNCH_BITOPS_H__
+#define __SYNCH_BITOPS_H__
+
+#include <linux/config.h>
+#include <xen/interface/xen.h>
+
+#ifdef CONFIG_SMP
+#include <asm/bitops.h>
+
+#define synch_change_bit(a,b) change_bit(a,b)
+#define synch_clear_bit(a,b) clear_bit(a,b)
+#define synch_const_test_bit(a,b) const_test_bit(a,b) 
+#define synch_set_bit(a,b) set_bit(a,b)
+#define synch_test_and_set_bit(a,b) test_and_set_bit(a,b)
+#define synch_test_and_change_bit(a,b) test_and_change_bit(a,b)
+#define synch_test_and_clear_bit(a,b) test_and_clear_bit(a,b)
+#define synch_test_bit(a,b) test_bit(a,b)
+
+static __inline__ unsigned long
+__synch_cmpxchg_u16(volatile unsigned short *p, unsigned long old, unsigned long new)
+{
+	int idx;
+	volatile unsigned int *xp = (unsigned int *)((ulong)p & ~(0x3UL));
+	union {
+		unsigned int word;
+		struct {
+			unsigned short s[2];
+		}s;
+	} xold, xnew;
+
+	/* we could start the reservation here and copy the u32
+	 * assembler, but I don't think it will gain us a whole
+	 * lot. */
+	xold.word = *xp;
+	xnew.word = xold.word;
+	idx = ((ulong)p >> 1) & 0x1;
+	xold.s.s[idx] = old;
+	xnew.s.s[idx] = new;
+
+	return __cmpxchg_u32(xp, xold.word, xnew.word);
+}
+
+/*
+ * This function doesn't exist, so you'll get a linker error
+ * if something tries to do an invalid xchg().
+ */
+extern void __synch_cmpxchg_called_with_bad_pointer(void);
+static __inline__ unsigned long
+__synch_cmpxchg(volatile void *ptr, unsigned long old, unsigned long new,
+	       unsigned int size)
+{
+	switch (size) {
+	case 2:
+		return __synch_cmpxchg_u16(ptr, old, new);
+	case 4:
+		return __cmpxchg_u32(ptr, old, new);
+#ifdef CONFIG_PPC64
+	case 8:
+		return __cmpxchg_u64(ptr, old, new);
+#endif
+	}
+	__synch_cmpxchg_called_with_bad_pointer();
+	return old;
+}
+
+#define synch_cmpxchg(ptr,o,n)						 \
+  ({									 \
+     __typeof__(*(ptr)) _o_ = (o);					 \
+     __typeof__(*(ptr)) _n_ = (n);					 \
+     (__typeof__(*(ptr))) __synch_cmpxchg((ptr), (unsigned long)_o_,		 \
+				    (unsigned long)_n_, sizeof(*(ptr))); \
+  })
+
+#define synch_cmpxchg_subword(ptr,o,n) __synch_cmpxchg_u16((ptr), (o), (n))
+
+#else
+#error "this only works for CONFIG_SMP"
+#endif
+
+#endif /* __SYNCH_BITOPS_H__ */
diff -rpuN linux-2.6.18.8/include/asm-x86_64/acpi.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/acpi.h
--- linux-2.6.18.8/include/asm-x86_64/acpi.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/acpi.h	2008-08-21 11:36:07.000000000 +0200
@@ -28,6 +28,9 @@
 
 #ifdef __KERNEL__
 
+#ifdef CONFIG_XEN
+#include <xen/interface/platform.h>
+#endif
 #include <acpi/pdc_intel.h>
 
 #define COMPILER_DEPENDENT_INT64   long long
@@ -129,6 +132,27 @@ static inline void acpi_disable_pci(void
 }
 extern int acpi_irq_balance_set(char *str);
 
+#ifdef CONFIG_XEN
+static inline int acpi_notify_hypervisor_state(u8 sleep_state,
+					       u32 pm1a_cnt_val,
+					       u32 pm1b_cnt_val)
+{
+	struct xen_platform_op op = {
+		.cmd = XENPF_enter_acpi_sleep,
+		.interface_version = XENPF_INTERFACE_VERSION,
+		.u = {
+			.enter_acpi_sleep = {
+				.pm1a_cnt_val = pm1a_cnt_val,
+				.pm1b_cnt_val = pm1b_cnt_val,
+				.sleep_state = sleep_state,
+			},
+		},
+	};
+
+	return HYPERVISOR_platform_op(&op);
+}
+#endif /* CONFIG_XEN */
+
 #else	/* !CONFIG_ACPI */
 
 #define acpi_lapic 0
@@ -152,7 +176,6 @@ extern unsigned long acpi_wakeup_address
 
 /* early initialization routine */
 extern void acpi_reserve_bootmem(void);
-
 #endif /*CONFIG_ACPI_SLEEP*/
 
 #define boot_cpu_physical_apicid boot_cpu_id
@@ -162,7 +185,9 @@ extern int acpi_pci_disabled;
 
 extern u8 x86_acpiid_to_apicid[];
 
+#ifndef CONFIG_XEN
 #define ARCH_HAS_POWER_INIT 1
+#endif
 
 extern int acpi_skip_timer_override;
 
diff -rpuN linux-2.6.18.8/include/asm-x86_64/agp.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/agp.h
--- linux-2.6.18.8/include/asm-x86_64/agp.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/agp.h	2008-08-21 11:36:07.000000000 +0200
@@ -10,8 +10,10 @@
  * with different cachability attributes for the same page.
  */
 
-int map_page_into_agp(struct page *page);
-int unmap_page_from_agp(struct page *page);
+/* Caller's responsibility to call global_flush_tlb() for
+ * performance reasons */
+#define map_page_into_agp(page) change_page_attr(page, 1, PAGE_KERNEL_NOCACHE)
+#define unmap_page_from_agp(page) change_page_attr(page, 1, PAGE_KERNEL)
 #define flush_agp_mappings() global_flush_tlb()
 
 /* Could use CLFLUSH here if the cpu supports it. But then it would
diff -rpuN linux-2.6.18.8/include/asm-x86_64/apic.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/apic.h
--- linux-2.6.18.8/include/asm-x86_64/apic.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/apic.h	2008-08-21 11:36:07.000000000 +0200
@@ -98,11 +98,13 @@ extern void setup_APIC_extened_lvt(unsig
 extern int disable_timer_pin_1;
 
 
+#ifndef CONFIG_XEN
 void smp_send_timer_broadcast_ipi(void);
 void switch_APIC_timer_to_ipi(void *cpumask);
 void switch_ipi_to_APIC_timer(void *cpumask);
 
 #define ARCH_APICTIMER_STOPS_ON_C3	1
+#endif
 
 #endif /* CONFIG_X86_LOCAL_APIC */
 
diff -rpuN linux-2.6.18.8/include/asm-x86_64/io_apic.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/io_apic.h
--- linux-2.6.18.8/include/asm-x86_64/io_apic.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/io_apic.h	2008-08-21 11:36:07.000000000 +0200
@@ -12,7 +12,7 @@
 
 #ifdef CONFIG_X86_IO_APIC
 
-#ifdef CONFIG_PCI_MSI
+#if defined(CONFIG_PCI_MSI) && !defined(CONFIG_XEN)
 static inline int use_pci_vector(void)	{return 1;}
 static inline void disable_edge_ioapic_vector(unsigned int vector) { }
 static inline void mask_and_ack_level_ioapic_vector(unsigned int vector) { }
diff -rpuN linux-2.6.18.8/include/asm-x86_64/kexec.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/kexec.h
--- linux-2.6.18.8/include/asm-x86_64/kexec.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/kexec.h	2008-08-21 11:36:07.000000000 +0200
@@ -1,6 +1,27 @@
 #ifndef _X86_64_KEXEC_H
 #define _X86_64_KEXEC_H
 
+#define PA_CONTROL_PAGE  0
+#define VA_CONTROL_PAGE  1
+#define PA_PGD           2
+#define VA_PGD           3
+#define PA_PUD_0         4
+#define VA_PUD_0         5
+#define PA_PMD_0         6
+#define VA_PMD_0         7
+#define PA_PTE_0         8
+#define VA_PTE_0         9
+#define PA_PUD_1         10
+#define VA_PUD_1         11
+#define PA_PMD_1         12
+#define VA_PMD_1         13
+#define PA_PTE_1         14
+#define VA_PTE_1         15
+#define PA_TABLE_PAGE    16
+#define PAGES_NR         17
+
+#ifndef __ASSEMBLY__
+
 #include <linux/string.h>
 
 #include <asm/page.h>
@@ -64,4 +85,25 @@ static inline void crash_setup_regs(stru
 		newregs->rip = (unsigned long)current_text_addr();
 	}
 }
+
+NORET_TYPE void
+relocate_kernel(unsigned long indirection_page,
+		unsigned long page_list,
+		unsigned long start_address) ATTRIB_NORET;
+
+/* Under Xen we need to work with machine addresses. These macros give the
+ * machine address of a certain page to the generic kexec code instead of 
+ * the pseudo physical address which would be given by the default macros.
+ */
+
+#ifdef CONFIG_XEN
+#define KEXEC_ARCH_HAS_PAGE_MACROS
+#define kexec_page_to_pfn(page)  pfn_to_mfn(page_to_pfn(page))
+#define kexec_pfn_to_page(pfn)   pfn_to_page(mfn_to_pfn(pfn))
+#define kexec_virt_to_phys(addr) virt_to_machine(addr)
+#define kexec_phys_to_virt(addr) phys_to_virt(machine_to_phys(addr))
+#endif
+
+#endif /* __ASSEMBLY__ */
+
 #endif /* _X86_64_KEXEC_H */
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/agp.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/agp.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/agp.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/agp.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,40 @@
+#ifndef AGP_H
+#define AGP_H 1
+
+#include <asm/cacheflush.h>
+#include <asm/system.h>
+
+/*
+ * Functions to keep the agpgart mappings coherent.
+ * The GART gives the CPU a physical alias of memory. The alias is
+ * mapped uncacheable. Make sure there are no conflicting mappings
+ * with different cachability attributes for the same page.
+ */
+
+#define map_page_into_agp(page) ( \
+	xen_create_contiguous_region((unsigned long)page_address(page), 0, 32) \
+	?: change_page_attr(page, 1, PAGE_KERNEL_NOCACHE))
+#define unmap_page_from_agp(page) ( \
+	xen_destroy_contiguous_region((unsigned long)page_address(page), 0), \
+	/* only a fallback: xen_destroy_contiguous_region uses PAGE_KERNEL */ \
+	change_page_attr(page, 1, PAGE_KERNEL))
+#define flush_agp_mappings() global_flush_tlb()
+
+/* Could use CLFLUSH here if the cpu supports it. But then it would
+   need to be called for each cacheline of the whole page so it may not be
+   worth it. Would need a page for it. */
+#define flush_agp_cache() wbinvd()
+
+/* Convert a physical address to an address suitable for the GART. */
+#define phys_to_gart(x) phys_to_machine(x)
+#define gart_to_phys(x) machine_to_phys(x)
+
+/* GATT allocation. Returns/accepts GATT kernel virtual address. */
+#define alloc_gatt_pages(order)	({                                          \
+	char *_t; dma_addr_t _d;                                            \
+	_t = dma_alloc_coherent(NULL,PAGE_SIZE<<(order),&_d,GFP_KERNEL);    \
+	_t; })
+#define free_gatt_pages(table, order)	\
+	dma_free_coherent(NULL,PAGE_SIZE<<(order),(table),virt_to_bus(table))
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/arch_hooks.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/arch_hooks.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/arch_hooks.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/arch_hooks.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,27 @@
+#ifndef _ASM_ARCH_HOOKS_H
+#define _ASM_ARCH_HOOKS_H
+
+#include <linux/interrupt.h>
+
+/*
+ *	linux/include/asm/arch_hooks.h
+ *
+ *	define the architecture specific hooks 
+ */
+
+/* these aren't arch hooks, they are generic routines
+ * that can be used by the hooks */
+extern void init_ISA_irqs(void);
+extern void apic_intr_init(void);
+extern void smp_intr_init(void);
+extern irqreturn_t timer_interrupt(int irq, void *dev_id, struct pt_regs *regs);
+
+/* these are the defined hooks */
+extern void intr_init_hook(void);
+extern void pre_intr_init_hook(void);
+extern void pre_setup_arch_hook(void);
+extern void trap_init_hook(void);
+extern void time_init_hook(void);
+extern void mca_nmi_hook(void);
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/bootsetup.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/bootsetup.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/bootsetup.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/bootsetup.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,42 @@
+
+#ifndef _X86_64_BOOTSETUP_H
+#define _X86_64_BOOTSETUP_H 1
+
+#define BOOT_PARAM_SIZE		4096
+extern char x86_boot_params[BOOT_PARAM_SIZE];
+
+/*
+ * This is set up by the setup-routine at boot-time
+ */
+#define PARAM	((unsigned char *)x86_boot_params)
+#define SCREEN_INFO (*(struct screen_info *) (PARAM+0))
+#define EXT_MEM_K (*(unsigned short *) (PARAM+2))
+#define ALT_MEM_K (*(unsigned int *) (PARAM+0x1e0))
+#define E820_MAP_NR (*(char*) (PARAM+E820NR))
+#define E820_MAP    ((struct e820entry *) (PARAM+E820MAP))
+#define APM_BIOS_INFO (*(struct apm_bios_info *) (PARAM+0x40))
+#define DRIVE_INFO (*(struct drive_info_struct *) (PARAM+0x80))
+#define SYS_DESC_TABLE (*(struct sys_desc_table_struct*)(PARAM+0xa0))
+#define MOUNT_ROOT_RDONLY (*(unsigned short *) (PARAM+0x1F2))
+#define RAMDISK_FLAGS (*(unsigned short *) (PARAM+0x1F8))
+#define SAVED_VIDEO_MODE (*(unsigned short *) (PARAM+0x1FA))
+#define ORIG_ROOT_DEV (*(unsigned short *) (PARAM+0x1FC))
+#define AUX_DEVICE_INFO (*(unsigned char *) (PARAM+0x1FF))
+#define LOADER_TYPE (*(unsigned char *) (PARAM+0x210))
+#define KERNEL_START (*(unsigned int *) (PARAM+0x214))
+
+#define INITRD_START (__pa(xen_start_info->mod_start))
+#define INITRD_SIZE (xen_start_info->mod_len)
+#define EDID_INFO   (*(struct edid_info *) (PARAM+0x440))
+
+#define EDD_NR     (*(unsigned char *) (PARAM+EDDNR))
+#define EDD_MBR_SIG_NR (*(unsigned char *) (PARAM+EDD_MBR_SIG_NR_BUF))
+#define EDD_MBR_SIGNATURE ((unsigned int *) (PARAM+EDD_MBR_SIG_BUF))
+#define EDD_BUF     ((struct edd_info *) (PARAM+EDDBUF))
+#define COMMAND_LINE saved_command_line
+
+#define RAMDISK_IMAGE_START_MASK  	0x07FF
+#define RAMDISK_PROMPT_FLAG		0x8000
+#define RAMDISK_LOAD_FLAG		0x4000	
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/desc.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/desc.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/desc.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/desc.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,265 @@
+/* Written 2000 by Andi Kleen */ 
+#ifndef __ARCH_DESC_H
+#define __ARCH_DESC_H
+
+#include <linux/threads.h>
+#include <asm/ldt.h>
+
+#ifndef __ASSEMBLY__
+
+#include <linux/string.h>
+#include <linux/smp.h>
+
+#include <asm/segment.h>
+#include <asm/mmu.h>
+
+// 8 byte segment descriptor
+struct desc_struct { 
+	u16 limit0;
+	u16 base0;
+	unsigned base1 : 8, type : 4, s : 1, dpl : 2, p : 1;
+	unsigned limit : 4, avl : 1, l : 1, d : 1, g : 1, base2 : 8;
+} __attribute__((packed)); 
+
+struct n_desc_struct { 
+	unsigned int a,b;
+}; 	
+
+enum { 
+	GATE_INTERRUPT = 0xE, 
+	GATE_TRAP = 0xF, 	
+	GATE_CALL = 0xC,
+}; 	
+
+// 16byte gate
+struct gate_struct {          
+	u16 offset_low;
+	u16 segment; 
+	unsigned ist : 3, zero0 : 5, type : 5, dpl : 2, p : 1;
+	u16 offset_middle;
+	u32 offset_high;
+	u32 zero1; 
+} __attribute__((packed));
+
+#define PTR_LOW(x) ((unsigned long)(x) & 0xFFFF) 
+#define PTR_MIDDLE(x) (((unsigned long)(x) >> 16) & 0xFFFF)
+#define PTR_HIGH(x) ((unsigned long)(x) >> 32)
+
+enum { 
+	DESC_TSS = 0x9,
+	DESC_LDT = 0x2,
+}; 
+
+// LDT or TSS descriptor in the GDT. 16 bytes.
+struct ldttss_desc { 
+	u16 limit0;
+	u16 base0;
+	unsigned base1 : 8, type : 5, dpl : 2, p : 1;
+	unsigned limit1 : 4, zero0 : 3, g : 1, base2 : 8;
+	u32 base3;
+	u32 zero1; 
+} __attribute__((packed)); 
+
+struct desc_ptr {
+	unsigned short size;
+	unsigned long address;
+} __attribute__((packed)) ;
+
+extern struct desc_ptr idt_descr, cpu_gdt_descr[NR_CPUS];
+
+extern struct desc_struct cpu_gdt_table[GDT_ENTRIES];
+
+#define load_TR_desc() asm volatile("ltr %w0"::"r" (GDT_ENTRY_TSS*8))
+#define load_LDT_desc() asm volatile("lldt %w0"::"r" (GDT_ENTRY_LDT*8))
+
+static inline void clear_LDT(void)
+{
+	int cpu = get_cpu();
+
+	/*
+	 * NB. We load the default_ldt for lcall7/27 handling on demand, as
+	 * it slows down context switching. Noone uses it anyway.
+	 */
+	cpu = cpu;              /* XXX avoid compiler warning */
+	xen_set_ldt(NULL, 0);
+	put_cpu();
+}
+
+/*
+ * This is the ldt that every process will get unless we need
+ * something other than this.
+ */
+extern struct desc_struct default_ldt[];
+#ifndef CONFIG_X86_NO_IDT
+extern struct gate_struct idt_table[]; 
+#endif
+extern struct desc_ptr cpu_gdt_descr[];
+
+/* the cpu gdt accessor */
+#define cpu_gdt(_cpu) ((struct desc_struct *)cpu_gdt_descr[_cpu].address)
+
+static inline void _set_gate(void *adr, unsigned type, unsigned long func, unsigned dpl, unsigned ist)  
+{
+	struct gate_struct s; 	
+	s.offset_low = PTR_LOW(func); 
+	s.segment = __KERNEL_CS;
+	s.ist = ist; 
+	s.p = 1;
+	s.dpl = dpl; 
+	s.zero0 = 0;
+	s.zero1 = 0; 
+	s.type = type; 
+	s.offset_middle = PTR_MIDDLE(func); 
+	s.offset_high = PTR_HIGH(func); 
+	/* does not need to be atomic because it is only done once at setup time */ 
+	memcpy(adr, &s, 16); 
+} 
+
+#ifndef CONFIG_X86_NO_IDT
+static inline void set_intr_gate(int nr, void *func) 
+{ 
+	BUG_ON((unsigned)nr > 0xFF);
+	_set_gate(&idt_table[nr], GATE_INTERRUPT, (unsigned long) func, 0, 0); 
+} 
+
+static inline void set_intr_gate_ist(int nr, void *func, unsigned ist) 
+{ 
+	BUG_ON((unsigned)nr > 0xFF);
+	_set_gate(&idt_table[nr], GATE_INTERRUPT, (unsigned long) func, 0, ist); 
+} 
+
+static inline void set_system_gate(int nr, void *func) 
+{ 
+	BUG_ON((unsigned)nr > 0xFF);
+	_set_gate(&idt_table[nr], GATE_INTERRUPT, (unsigned long) func, 3, 0); 
+} 
+
+static inline void set_system_gate_ist(int nr, void *func, unsigned ist)
+{
+	_set_gate(&idt_table[nr], GATE_INTERRUPT, (unsigned long) func, 3, ist);
+}
+#endif
+
+static inline void set_tssldt_descriptor(void *ptr, unsigned long tss, unsigned type, 
+					 unsigned size) 
+{ 
+	struct ldttss_desc d;
+	memset(&d,0,sizeof(d)); 
+	d.limit0 = size & 0xFFFF;
+	d.base0 = PTR_LOW(tss); 
+	d.base1 = PTR_MIDDLE(tss) & 0xFF; 
+	d.type = type;
+	d.p = 1; 
+	d.limit1 = (size >> 16) & 0xF;
+	d.base2 = (PTR_MIDDLE(tss) >> 8) & 0xFF; 
+	d.base3 = PTR_HIGH(tss); 
+	memcpy(ptr, &d, 16); 
+}
+
+#ifndef CONFIG_X86_NO_TSS
+static inline void set_tss_desc(unsigned cpu, void *addr)
+{ 
+	/*
+	 * sizeof(unsigned long) coming from an extra "long" at the end
+	 * of the iobitmap. See tss_struct definition in processor.h
+	 *
+	 * -1? seg base+limit should be pointing to the address of the
+	 * last valid byte
+	 */
+	set_tssldt_descriptor(&cpu_gdt(cpu)[GDT_ENTRY_TSS], 
+		(unsigned long)addr, DESC_TSS,
+		IO_BITMAP_OFFSET + IO_BITMAP_BYTES + sizeof(unsigned long) - 1);
+} 
+#endif
+
+static inline void set_ldt_desc(unsigned cpu, void *addr, int size)
+{ 
+	set_tssldt_descriptor(&cpu_gdt(cpu)[GDT_ENTRY_LDT], (unsigned long)addr,
+			      DESC_LDT, size * 8 - 1);
+}
+
+static inline void set_seg_base(unsigned cpu, int entry, void *base)
+{ 
+	struct desc_struct *d = &cpu_gdt(cpu)[entry];
+	u32 addr = (u32)(u64)base;
+	BUG_ON((u64)base >> 32); 
+	d->base0 = addr & 0xffff;
+	d->base1 = (addr >> 16) & 0xff;
+	d->base2 = (addr >> 24) & 0xff;
+} 
+
+#define LDT_entry_a(info) \
+	((((info)->base_addr & 0x0000ffff) << 16) | ((info)->limit & 0x0ffff))
+/* Don't allow setting of the lm bit. It is useless anyways because 
+   64bit system calls require __USER_CS. */ 
+#define LDT_entry_b(info) \
+	(((info)->base_addr & 0xff000000) | \
+	(((info)->base_addr & 0x00ff0000) >> 16) | \
+	((info)->limit & 0xf0000) | \
+	(((info)->read_exec_only ^ 1) << 9) | \
+	((info)->contents << 10) | \
+	(((info)->seg_not_present ^ 1) << 15) | \
+	((info)->seg_32bit << 22) | \
+	((info)->limit_in_pages << 23) | \
+	((info)->useable << 20) | \
+	/* ((info)->lm << 21) | */ \
+	0x7000)
+
+#define LDT_empty(info) (\
+	(info)->base_addr	== 0	&& \
+	(info)->limit		== 0	&& \
+	(info)->contents	== 0	&& \
+	(info)->read_exec_only	== 1	&& \
+	(info)->seg_32bit	== 0	&& \
+	(info)->limit_in_pages	== 0	&& \
+	(info)->seg_not_present	== 1	&& \
+	(info)->useable		== 0	&& \
+	(info)->lm		== 0)
+
+#if TLS_SIZE != 24
+# error update this code.
+#endif
+
+static inline void load_TLS(struct thread_struct *t, unsigned int cpu)
+{
+#if 0
+	u64 *gdt = (u64 *)(cpu_gdt(cpu) + GDT_ENTRY_TLS_MIN);
+	gdt[0] = t->tls_array[0];
+	gdt[1] = t->tls_array[1];
+	gdt[2] = t->tls_array[2];
+#endif
+#define C(i) \
+	if (HYPERVISOR_update_descriptor(virt_to_machine(&cpu_gdt(cpu)[GDT_ENTRY_TLS_MIN + i]), \
+						 t->tls_array[i])) \
+		BUG();
+
+	C(0); C(1); C(2);
+#undef C
+} 
+
+/*
+ * load one particular LDT into the current CPU
+ */
+static inline void load_LDT_nolock (mm_context_t *pc, int cpu)
+{
+	void *segments = pc->ldt;
+	int count = pc->size;
+
+	if (likely(!count))
+		segments = NULL;
+
+	xen_set_ldt(segments, count);
+}
+
+static inline void load_LDT(mm_context_t *pc)
+{
+	int cpu = get_cpu();
+	load_LDT_nolock(pc, cpu);
+	put_cpu();
+}
+
+extern struct desc_ptr idt_descr;
+
+#endif /* !__ASSEMBLY__ */
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/dma-mapping.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/dma-mapping.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/dma-mapping.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/dma-mapping.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,207 @@
+#ifndef _X8664_DMA_MAPPING_H
+#define _X8664_DMA_MAPPING_H 1
+
+/*
+ * IOMMU interface. See Documentation/DMA-mapping.txt and DMA-API.txt for
+ * documentation.
+ */
+
+
+#include <asm/scatterlist.h>
+#include <asm/io.h>
+#include <asm/swiotlb.h>
+
+struct dma_mapping_ops {
+	int             (*mapping_error)(dma_addr_t dma_addr);
+	void*           (*alloc_coherent)(struct device *dev, size_t size,
+                                dma_addr_t *dma_handle, gfp_t gfp);
+	void            (*free_coherent)(struct device *dev, size_t size,
+                                void *vaddr, dma_addr_t dma_handle);
+	dma_addr_t      (*map_single)(struct device *hwdev, void *ptr,
+                                size_t size, int direction);
+	/* like map_single, but doesn't check the device mask */
+	dma_addr_t      (*map_simple)(struct device *hwdev, char *ptr,
+                                size_t size, int direction);
+	void            (*unmap_single)(struct device *dev, dma_addr_t addr,
+		                size_t size, int direction);
+	void            (*sync_single_for_cpu)(struct device *hwdev,
+		                dma_addr_t dma_handle, size_t size,
+				int direction);
+	void            (*sync_single_for_device)(struct device *hwdev,
+                                dma_addr_t dma_handle, size_t size,
+				int direction);
+	void            (*sync_single_range_for_cpu)(struct device *hwdev,
+                                dma_addr_t dma_handle, unsigned long offset,
+		                size_t size, int direction);
+	void            (*sync_single_range_for_device)(struct device *hwdev,
+				dma_addr_t dma_handle, unsigned long offset,
+		                size_t size, int direction);
+	void            (*sync_sg_for_cpu)(struct device *hwdev,
+                                struct scatterlist *sg, int nelems,
+				int direction);
+	void            (*sync_sg_for_device)(struct device *hwdev,
+				struct scatterlist *sg, int nelems,
+				int direction);
+	int             (*map_sg)(struct device *hwdev, struct scatterlist *sg,
+		                int nents, int direction);
+	void            (*unmap_sg)(struct device *hwdev,
+				struct scatterlist *sg, int nents,
+				int direction);
+	int             (*dma_supported)(struct device *hwdev, u64 mask);
+	int		is_phys;
+};
+
+extern dma_addr_t bad_dma_address;
+extern struct dma_mapping_ops* dma_ops;
+extern int iommu_merge;
+
+static inline int valid_dma_direction(int dma_direction)
+{
+	return ((dma_direction == DMA_BIDIRECTIONAL) ||
+		(dma_direction == DMA_TO_DEVICE) ||
+		(dma_direction == DMA_FROM_DEVICE));
+}
+
+#if 0
+static inline int dma_mapping_error(dma_addr_t dma_addr)
+{
+	if (dma_ops->mapping_error)
+		return dma_ops->mapping_error(dma_addr);
+
+	return (dma_addr == bad_dma_address);
+}
+
+extern void *dma_alloc_coherent(struct device *dev, size_t size,
+				dma_addr_t *dma_handle, gfp_t gfp);
+extern void dma_free_coherent(struct device *dev, size_t size, void *vaddr,
+			      dma_addr_t dma_handle);
+
+static inline dma_addr_t
+dma_map_single(struct device *hwdev, void *ptr, size_t size,
+	       int direction)
+{
+	BUG_ON(!valid_dma_direction(direction));
+	return dma_ops->map_single(hwdev, ptr, size, direction);
+}
+
+static inline void
+dma_unmap_single(struct device *dev, dma_addr_t addr,size_t size,
+		 int direction)
+{
+	BUG_ON(!valid_dma_direction(direction));
+	dma_ops->unmap_single(dev, addr, size, direction);
+}
+
+#define dma_map_page(dev,page,offset,size,dir) \
+	dma_map_single((dev), page_address(page)+(offset), (size), (dir))
+
+#define dma_unmap_page dma_unmap_single
+
+static inline void
+dma_sync_single_for_cpu(struct device *hwdev, dma_addr_t dma_handle,
+			size_t size, int direction)
+{
+	BUG_ON(!valid_dma_direction(direction));
+	if (dma_ops->sync_single_for_cpu)
+		dma_ops->sync_single_for_cpu(hwdev, dma_handle, size,
+					     direction);
+	flush_write_buffers();
+}
+
+static inline void
+dma_sync_single_for_device(struct device *hwdev, dma_addr_t dma_handle,
+			   size_t size, int direction)
+{
+	BUG_ON(!valid_dma_direction(direction));
+	if (dma_ops->sync_single_for_device)
+		dma_ops->sync_single_for_device(hwdev, dma_handle, size,
+						direction);
+	flush_write_buffers();
+}
+
+static inline void
+dma_sync_single_range_for_cpu(struct device *hwdev, dma_addr_t dma_handle,
+			      unsigned long offset, size_t size, int direction)
+{
+	BUG_ON(!valid_dma_direction(direction));
+	if (dma_ops->sync_single_range_for_cpu) {
+		dma_ops->sync_single_range_for_cpu(hwdev, dma_handle, offset, size, direction);
+	}
+
+	flush_write_buffers();
+}
+
+static inline void
+dma_sync_single_range_for_device(struct device *hwdev, dma_addr_t dma_handle,
+				 unsigned long offset, size_t size, int direction)
+{
+	BUG_ON(!valid_dma_direction(direction));
+	if (dma_ops->sync_single_range_for_device)
+		dma_ops->sync_single_range_for_device(hwdev, dma_handle,
+						      offset, size, direction);
+
+	flush_write_buffers();
+}
+
+static inline void
+dma_sync_sg_for_cpu(struct device *hwdev, struct scatterlist *sg,
+		    int nelems, int direction)
+{
+	BUG_ON(!valid_dma_direction(direction));
+	if (dma_ops->sync_sg_for_cpu)
+		dma_ops->sync_sg_for_cpu(hwdev, sg, nelems, direction);
+	flush_write_buffers();
+}
+
+static inline void
+dma_sync_sg_for_device(struct device *hwdev, struct scatterlist *sg,
+		       int nelems, int direction)
+{
+	BUG_ON(!valid_dma_direction(direction));
+	if (dma_ops->sync_sg_for_device) {
+		dma_ops->sync_sg_for_device(hwdev, sg, nelems, direction);
+	}
+
+	flush_write_buffers();
+}
+
+static inline int
+dma_map_sg(struct device *hwdev, struct scatterlist *sg, int nents, int direction)
+{
+	BUG_ON(!valid_dma_direction(direction));
+	return dma_ops->map_sg(hwdev, sg, nents, direction);
+}
+
+static inline void
+dma_unmap_sg(struct device *hwdev, struct scatterlist *sg, int nents,
+	     int direction)
+{
+	BUG_ON(!valid_dma_direction(direction));
+	dma_ops->unmap_sg(hwdev, sg, nents, direction);
+}
+
+extern int dma_supported(struct device *hwdev, u64 mask);
+
+/* same for gart, swiotlb, and nommu */
+static inline int dma_get_cache_alignment(void)
+{
+	return boot_cpu_data.x86_clflush_size;
+}
+
+#define dma_is_consistent(h) 1
+
+extern int dma_set_mask(struct device *dev, u64 mask);
+
+static inline void
+dma_cache_sync(void *vaddr, size_t size, enum dma_data_direction dir)
+{
+	flush_write_buffers();
+}
+
+extern struct device fallback_dev;
+extern int panic_on_overflow;
+#endif
+
+#endif /* _X8664_DMA_MAPPING_H */
+
+#include <asm-i386/mach-xen/asm/dma-mapping.h>
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/e820.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/e820.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/e820.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/e820.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,66 @@
+/*
+ * structures and definitions for the int 15, ax=e820 memory map
+ * scheme.
+ *
+ * In a nutshell, setup.S populates a scratch table in the
+ * empty_zero_block that contains a list of usable address/size
+ * duples.  setup.c, this information is transferred into the e820map,
+ * and in init.c/numa.c, that new information is used to mark pages
+ * reserved or not.
+ */
+#ifndef __E820_HEADER
+#define __E820_HEADER
+
+#include <linux/mmzone.h>
+
+#define E820MAP	0x2d0		/* our map */
+#define E820MAX	128		/* number of entries in E820MAP */
+#define E820NR	0x1e8		/* # entries in E820MAP */
+
+#define E820_RAM	1
+#define E820_RESERVED	2
+#define E820_ACPI	3 /* usable as RAM once ACPI tables have been read */
+#define E820_NVS	4
+
+#define HIGH_MEMORY	(1024*1024)
+
+#define LOWMEMSIZE()	(0x9f000)
+
+#ifndef __ASSEMBLY__
+struct e820entry {
+	u64 addr;	/* start of memory segment */
+	u64 size;	/* size of memory segment */
+	u32 type;	/* type of memory segment */
+} __attribute__((packed));
+
+struct e820map {
+    int nr_map;
+	struct e820entry map[E820MAX];
+};
+
+extern unsigned long find_e820_area(unsigned long start, unsigned long end, 
+				    unsigned size);
+extern void add_memory_region(unsigned long start, unsigned long size, 
+			      int type);
+extern void setup_memory_region(void);
+extern void contig_e820_setup(void); 
+extern unsigned long e820_end_of_ram(void);
+extern void e820_reserve_resources(struct e820entry *e820, int nr_map);
+extern void e820_print_map(char *who);
+extern int e820_any_mapped(unsigned long start, unsigned long end, unsigned type);
+extern int e820_all_mapped(unsigned long start, unsigned long end, unsigned type);
+
+extern void e820_bootmem_free(pg_data_t *pgdat, unsigned long start,unsigned long end);
+extern void e820_setup_gap(struct e820entry *e820, int nr_map);
+extern unsigned long e820_hole_size(unsigned long start_pfn,
+				    unsigned long end_pfn);
+
+extern void __init parse_memopt(char *p, char **end);
+extern void __init parse_memmapopt(char *p, char **end);
+
+extern struct e820map e820;
+
+extern unsigned ebda_addr, ebda_size;
+#endif/*!__ASSEMBLY__*/
+
+#endif/*__E820_HEADER*/
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/fixmap.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/fixmap.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/fixmap.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/fixmap.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,112 @@
+/*
+ * fixmap.h: compile-time virtual memory allocation
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License.  See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Copyright (C) 1998 Ingo Molnar
+ */
+
+#ifndef _ASM_FIXMAP_H
+#define _ASM_FIXMAP_H
+
+#include <linux/kernel.h>
+#include <asm/apicdef.h>
+#include <asm/page.h>
+#include <asm/vsyscall.h>
+#include <asm/vsyscall32.h>
+#include <asm/acpi.h>
+
+/*
+ * Here we define all the compile-time 'special' virtual
+ * addresses. The point is to have a constant address at
+ * compile time, but to set the physical address only
+ * in the boot process.
+ *
+ * these 'compile-time allocated' memory buffers are
+ * fixed-size 4k pages. (or larger if used with an increment
+ * highger than 1) use fixmap_set(idx,phys) to associate
+ * physical memory with fixmap indices.
+ *
+ * TLB entries of such buffers will not be flushed across
+ * task switches.
+ */
+
+enum fixed_addresses {
+	VSYSCALL_LAST_PAGE,
+	VSYSCALL_FIRST_PAGE = VSYSCALL_LAST_PAGE + ((VSYSCALL_END-VSYSCALL_START) >> PAGE_SHIFT) - 1,
+	VSYSCALL_HPET,
+	FIX_HPET_BASE,
+#ifdef CONFIG_X86_LOCAL_APIC
+	FIX_APIC_BASE,	/* local (CPU) APIC) -- required for SMP or not */
+#endif
+#ifdef CONFIG_X86_IO_APIC
+	FIX_IO_APIC_BASE_0,
+	FIX_IO_APIC_BASE_END = FIX_IO_APIC_BASE_0 + MAX_IO_APICS-1,
+#endif
+#ifdef CONFIG_ACPI
+	FIX_ACPI_BEGIN,
+	FIX_ACPI_END = FIX_ACPI_BEGIN + FIX_ACPI_PAGES - 1,
+#endif
+	FIX_SHARED_INFO,
+#define NR_FIX_ISAMAPS	256
+	FIX_ISAMAP_END,
+	FIX_ISAMAP_BEGIN = FIX_ISAMAP_END + NR_FIX_ISAMAPS - 1,
+	__end_of_permanent_fixed_addresses,
+	/* temporary boot-time mappings, used before ioremap() is functional */
+#define NR_FIX_BTMAPS	16
+	FIX_BTMAP_END = __end_of_permanent_fixed_addresses,
+	FIX_BTMAP_BEGIN = FIX_BTMAP_END + NR_FIX_BTMAPS - 1,
+	__end_of_fixed_addresses
+};
+
+extern void __set_fixmap (enum fixed_addresses idx,
+					unsigned long phys, pgprot_t flags);
+
+#define set_fixmap(idx, phys) \
+		__set_fixmap(idx, phys, PAGE_KERNEL)
+/*
+ * Some hardware wants to get fixmapped without caching.
+ */
+#define set_fixmap_nocache(idx, phys) \
+		__set_fixmap(idx, phys, PAGE_KERNEL_NOCACHE)
+
+#define clear_fixmap(idx) \
+                __set_fixmap(idx, 0, __pgprot(0))
+
+#define FIXADDR_TOP	(VSYSCALL_END-PAGE_SIZE)
+#define FIXADDR_SIZE	(__end_of_fixed_addresses << PAGE_SHIFT)
+#define FIXADDR_START	(FIXADDR_TOP - FIXADDR_SIZE)
+
+/* Only covers 32bit vsyscalls currently. Need another set for 64bit. */
+#define FIXADDR_USER_START	((unsigned long)VSYSCALL32_VSYSCALL)
+#define FIXADDR_USER_END	(FIXADDR_USER_START + PAGE_SIZE)
+
+#define __fix_to_virt(x)	(FIXADDR_TOP - ((x) << PAGE_SHIFT))
+
+extern void __this_fixmap_does_not_exist(void);
+
+/*
+ * 'index to address' translation. If anyone tries to use the idx
+ * directly without translation, we catch the bug with a NULL-deference
+ * kernel oops. Illegal ranges of incoming indices are caught too.
+ */
+static __always_inline unsigned long fix_to_virt(const unsigned int idx)
+{
+	/*
+	 * this branch gets completely eliminated after inlining,
+	 * except when someone tries to use fixaddr indices in an
+	 * illegal way. (such as mixing up address types or using
+	 * out-of-range indices).
+	 *
+	 * If it doesn't get removed, the linker will complain
+	 * loudly with a reasonably clear error message..
+	 */
+	if (idx >= __end_of_fixed_addresses)
+		__this_fixmap_does_not_exist();
+
+        return __fix_to_virt(idx);
+}
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/gnttab_dma.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/gnttab_dma.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/gnttab_dma.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/gnttab_dma.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1 @@
+#include <asm-i386/mach-xen/asm/gnttab_dma.h>
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/hw_irq.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/hw_irq.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/hw_irq.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/hw_irq.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,136 @@
+#ifndef _ASM_HW_IRQ_H
+#define _ASM_HW_IRQ_H
+
+/*
+ *	linux/include/asm/hw_irq.h
+ *
+ *	(C) 1992, 1993 Linus Torvalds, (C) 1997 Ingo Molnar
+ *
+ *	moved some of the old arch/i386/kernel/irq.h to here. VY
+ *
+ *	IRQ/IPI changes taken from work by Thomas Radke
+ *	<tomsoft@informatik.tu-chemnitz.de>
+ *
+ *	hacked by Andi Kleen for x86-64.
+ */
+
+#ifndef __ASSEMBLY__
+#include <asm/atomic.h>
+#include <asm/irq.h>
+#include <linux/profile.h>
+#include <linux/smp.h>
+
+struct hw_interrupt_type;
+#endif
+
+#define NMI_VECTOR		0x02
+/*
+ * IDT vectors usable for external interrupt sources start
+ * at 0x20:
+ */
+#define FIRST_EXTERNAL_VECTOR	0x20
+
+#define IA32_SYSCALL_VECTOR	0x80
+
+
+/*
+ * Vectors 0x20-0x2f are used for ISA interrupts.
+ */
+
+/*
+ * Special IRQ vectors used by the SMP architecture, 0xf0-0xff
+ *
+ *  some of the following vectors are 'rare', they are merged
+ *  into a single vector (CALL_FUNCTION_VECTOR) to save vector space.
+ *  TLB, reschedule and local APIC vectors are performance-critical.
+ */
+#ifndef CONFIG_XEN
+#define SPURIOUS_APIC_VECTOR	0xff
+#define ERROR_APIC_VECTOR	0xfe
+#define RESCHEDULE_VECTOR	0xfd
+#define CALL_FUNCTION_VECTOR	0xfc
+/* fb free - please don't readd KDB here because it's useless
+   (hint - think what a NMI bit does to a vector) */
+#define THERMAL_APIC_VECTOR	0xfa
+#define THRESHOLD_APIC_VECTOR   0xf9
+/* f8 free */
+#define INVALIDATE_TLB_VECTOR_END	0xf7
+#define INVALIDATE_TLB_VECTOR_START	0xf0	/* f0-f7 used for TLB flush */
+
+#define NUM_INVALIDATE_TLB_VECTORS	8
+#endif
+
+/*
+ * Local APIC timer IRQ vector is on a different priority level,
+ * to work around the 'lost local interrupt if more than 2 IRQ
+ * sources per level' errata.
+ */
+#define LOCAL_TIMER_VECTOR	0xef
+
+/*
+ * First APIC vector available to drivers: (vectors 0x30-0xee)
+ * we start at 0x31 to spread out vectors evenly between priority
+ * levels. (0x80 is the syscall vector)
+ */
+#define FIRST_DEVICE_VECTOR	0x31
+#define FIRST_SYSTEM_VECTOR	0xef   /* duplicated in irq.h */
+
+
+#ifndef __ASSEMBLY__
+extern u8 irq_vector[NR_IRQ_VECTORS];
+#define IO_APIC_VECTOR(irq)	(irq_vector[irq])
+#define AUTO_ASSIGN		-1
+
+/*
+ * Various low-level irq details needed by irq.c, process.c,
+ * time.c, io_apic.c and smp.c
+ *
+ * Interrupt entry/exit code at both C and assembly level
+ */
+
+extern void disable_8259A_irq(unsigned int irq);
+extern void enable_8259A_irq(unsigned int irq);
+extern int i8259A_irq_pending(unsigned int irq);
+extern void make_8259A_irq(unsigned int irq);
+extern void init_8259A(int aeoi);
+extern void FASTCALL(send_IPI_self(int vector));
+extern void init_VISWS_APIC_irqs(void);
+extern void setup_IO_APIC(void);
+extern void disable_IO_APIC(void);
+#define print_IO_APIC()
+extern int IO_APIC_get_PCI_irq_vector(int bus, int slot, int fn);
+extern void send_IPI(int dest, int vector);
+extern void setup_ioapic_dest(void);
+
+extern unsigned long io_apic_irqs;
+
+extern atomic_t irq_err_count;
+extern atomic_t irq_mis_count;
+
+#define IO_APIC_IRQ(x) (((x) >= 16) || ((1<<(x)) & io_apic_irqs))
+
+#define __STR(x) #x
+#define STR(x) __STR(x)
+
+#include <asm/ptrace.h>
+
+#define IRQ_NAME2(nr) nr##_interrupt(void)
+#define IRQ_NAME(nr) IRQ_NAME2(IRQ##nr)
+
+/*
+ *	SMP has a few special interrupts for IPI messages
+ */
+
+#define BUILD_IRQ(nr) \
+asmlinkage void IRQ_NAME(nr); \
+__asm__( \
+"\n.p2align\n" \
+"IRQ" #nr "_interrupt:\n\t" \
+	"push $~(" #nr ") ; " \
+	"jmp common_interrupt");
+
+#define platform_legacy_irq(irq)	((irq) < 16)
+
+#endif
+
+#endif /* _ASM_HW_IRQ_H */
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/hypercall.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/hypercall.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/hypercall.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/hypercall.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,415 @@
+/******************************************************************************
+ * hypercall.h
+ * 
+ * Linux-specific hypervisor handling.
+ * 
+ * Copyright (c) 2002-2004, K A Fraser
+ * 
+ * 64-bit updates:
+ *   Benjamin Liu <benjamin.liu@intel.com>
+ *   Jun Nakajima <jun.nakajima@intel.com>
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __HYPERCALL_H__
+#define __HYPERCALL_H__
+
+#include <linux/string.h> /* memcpy() */
+#include <linux/stringify.h>
+
+#ifndef __HYPERVISOR_H__
+# error "please don't include this file directly"
+#endif
+
+#ifdef CONFIG_XEN
+#define HYPERCALL_STR(name)					\
+	"call hypercall_page + ("__stringify(__HYPERVISOR_##name)" * 32)"
+#else
+#define HYPERCALL_STR(name)					\
+	"mov $("__stringify(__HYPERVISOR_##name)" * 32),%%eax; "\
+	"add hypercall_stubs(%%rip),%%rax; "			\
+	"call *%%rax"
+#endif
+
+#define _hypercall0(type, name)			\
+({						\
+	type __res;				\
+	asm volatile (				\
+		HYPERCALL_STR(name)		\
+		: "=a" (__res)			\
+		:				\
+		: "memory" );			\
+	__res;					\
+})
+
+#define _hypercall1(type, name, a1)				\
+({								\
+	type __res;						\
+	long __ign1;						\
+	asm volatile (						\
+		HYPERCALL_STR(name)				\
+		: "=a" (__res), "=D" (__ign1)			\
+		: "1" ((long)(a1))				\
+		: "memory" );					\
+	__res;							\
+})
+
+#define _hypercall2(type, name, a1, a2)				\
+({								\
+	type __res;						\
+	long __ign1, __ign2;					\
+	asm volatile (						\
+		HYPERCALL_STR(name)				\
+		: "=a" (__res), "=D" (__ign1), "=S" (__ign2)	\
+		: "1" ((long)(a1)), "2" ((long)(a2))		\
+		: "memory" );					\
+	__res;							\
+})
+
+#define _hypercall3(type, name, a1, a2, a3)			\
+({								\
+	type __res;						\
+	long __ign1, __ign2, __ign3;				\
+	asm volatile (						\
+		HYPERCALL_STR(name)				\
+		: "=a" (__res), "=D" (__ign1), "=S" (__ign2), 	\
+		"=d" (__ign3)					\
+		: "1" ((long)(a1)), "2" ((long)(a2)),		\
+		"3" ((long)(a3))				\
+		: "memory" );					\
+	__res;							\
+})
+
+#define _hypercall4(type, name, a1, a2, a3, a4)			\
+({								\
+	type __res;						\
+	long __ign1, __ign2, __ign3;				\
+	register long __arg4 asm("r10") = (long)(a4);		\
+	asm volatile (						\
+		HYPERCALL_STR(name)				\
+		: "=a" (__res), "=D" (__ign1), "=S" (__ign2),	\
+		  "=d" (__ign3), "+r" (__arg4)			\
+		: "1" ((long)(a1)), "2" ((long)(a2)),		\
+		  "3" ((long)(a3))				\
+		: "memory" );					\
+	__res;							\
+})
+
+#define _hypercall5(type, name, a1, a2, a3, a4, a5)		\
+({								\
+	type __res;						\
+	long __ign1, __ign2, __ign3;				\
+	register long __arg4 asm("r10") = (long)(a4);		\
+	register long __arg5 asm("r8") = (long)(a5);		\
+	asm volatile (						\
+		HYPERCALL_STR(name)				\
+		: "=a" (__res), "=D" (__ign1), "=S" (__ign2),	\
+		  "=d" (__ign3), "+r" (__arg4), "+r" (__arg5)	\
+		: "1" ((long)(a1)), "2" ((long)(a2)),		\
+		  "3" ((long)(a3))				\
+		: "memory" );					\
+	__res;							\
+})
+
+static inline int __must_check
+HYPERVISOR_set_trap_table(
+	const trap_info_t *table)
+{
+	return _hypercall1(int, set_trap_table, table);
+}
+
+static inline int __must_check
+HYPERVISOR_mmu_update(
+	mmu_update_t *req, unsigned int count, unsigned int *success_count,
+	domid_t domid)
+{
+	return _hypercall4(int, mmu_update, req, count, success_count, domid);
+}
+
+static inline int __must_check
+HYPERVISOR_mmuext_op(
+	struct mmuext_op *op, unsigned int count, unsigned int *success_count,
+	domid_t domid)
+{
+	return _hypercall4(int, mmuext_op, op, count, success_count, domid);
+}
+
+static inline int __must_check
+HYPERVISOR_set_gdt(
+	unsigned long *frame_list, unsigned int entries)
+{
+	return _hypercall2(int, set_gdt, frame_list, entries);
+}
+
+static inline int __must_check
+HYPERVISOR_stack_switch(
+	unsigned long ss, unsigned long esp)
+{
+	return _hypercall2(int, stack_switch, ss, esp);
+}
+
+static inline int __must_check
+HYPERVISOR_set_callbacks(
+	unsigned long event_address, unsigned long failsafe_address, 
+	unsigned long syscall_address)
+{
+	return _hypercall3(int, set_callbacks,
+			   event_address, failsafe_address, syscall_address);
+}
+
+static inline int
+HYPERVISOR_fpu_taskswitch(
+	int set)
+{
+	return _hypercall1(int, fpu_taskswitch, set);
+}
+
+static inline int __must_check
+HYPERVISOR_sched_op_compat(
+	int cmd, unsigned long arg)
+{
+	return _hypercall2(int, sched_op_compat, cmd, arg);
+}
+
+static inline int __must_check
+HYPERVISOR_sched_op(
+	int cmd, void *arg)
+{
+	return _hypercall2(int, sched_op, cmd, arg);
+}
+
+static inline long __must_check
+HYPERVISOR_set_timer_op(
+	u64 timeout)
+{
+	return _hypercall1(long, set_timer_op, timeout);
+}
+
+static inline int __must_check
+HYPERVISOR_platform_op(
+	struct xen_platform_op *platform_op)
+{
+	platform_op->interface_version = XENPF_INTERFACE_VERSION;
+	return _hypercall1(int, platform_op, platform_op);
+}
+
+static inline int __must_check
+HYPERVISOR_set_debugreg(
+	unsigned int reg, unsigned long value)
+{
+	return _hypercall2(int, set_debugreg, reg, value);
+}
+
+static inline unsigned long __must_check
+HYPERVISOR_get_debugreg(
+	unsigned int reg)
+{
+	return _hypercall1(unsigned long, get_debugreg, reg);
+}
+
+static inline int __must_check
+HYPERVISOR_update_descriptor(
+	unsigned long ma, unsigned long word)
+{
+	return _hypercall2(int, update_descriptor, ma, word);
+}
+
+static inline int __must_check
+HYPERVISOR_memory_op(
+	unsigned int cmd, void *arg)
+{
+	return _hypercall2(int, memory_op, cmd, arg);
+}
+
+static inline int __must_check
+HYPERVISOR_multicall(
+	multicall_entry_t *call_list, unsigned int nr_calls)
+{
+	return _hypercall2(int, multicall, call_list, nr_calls);
+}
+
+static inline int __must_check
+HYPERVISOR_update_va_mapping(
+	unsigned long va, pte_t new_val, unsigned long flags)
+{
+	return _hypercall3(int, update_va_mapping, va, new_val.pte, flags);
+}
+
+static inline int __must_check
+HYPERVISOR_event_channel_op(
+	int cmd, void *arg)
+{
+	int rc = _hypercall2(int, event_channel_op, cmd, arg);
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (unlikely(rc == -ENOSYS)) {
+		struct evtchn_op op;
+		op.cmd = cmd;
+		memcpy(&op.u, arg, sizeof(op.u));
+		rc = _hypercall1(int, event_channel_op_compat, &op);
+		memcpy(arg, &op.u, sizeof(op.u));
+	}
+#endif
+
+	return rc;
+}
+
+static inline int __must_check
+HYPERVISOR_acm_op(
+	int cmd, void *arg)
+{
+	return _hypercall2(int, acm_op, cmd, arg);
+}
+
+static inline int __must_check
+HYPERVISOR_xen_version(
+	int cmd, void *arg)
+{
+	return _hypercall2(int, xen_version, cmd, arg);
+}
+
+static inline int __must_check
+HYPERVISOR_console_io(
+	int cmd, unsigned int count, char *str)
+{
+	return _hypercall3(int, console_io, cmd, count, str);
+}
+
+static inline int __must_check
+HYPERVISOR_physdev_op(
+	int cmd, void *arg)
+{
+	int rc = _hypercall2(int, physdev_op, cmd, arg);
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (unlikely(rc == -ENOSYS)) {
+		struct physdev_op op;
+		op.cmd = cmd;
+		memcpy(&op.u, arg, sizeof(op.u));
+		rc = _hypercall1(int, physdev_op_compat, &op);
+		memcpy(arg, &op.u, sizeof(op.u));
+	}
+#endif
+
+	return rc;
+}
+
+static inline int __must_check
+HYPERVISOR_grant_table_op(
+	unsigned int cmd, void *uop, unsigned int count)
+{
+	return _hypercall3(int, grant_table_op, cmd, uop, count);
+}
+
+static inline int __must_check
+HYPERVISOR_update_va_mapping_otherdomain(
+	unsigned long va, pte_t new_val, unsigned long flags, domid_t domid)
+{
+	return _hypercall4(int, update_va_mapping_otherdomain, va,
+			   new_val.pte, flags, domid);
+}
+
+static inline int __must_check
+HYPERVISOR_vm_assist(
+	unsigned int cmd, unsigned int type)
+{
+	return _hypercall2(int, vm_assist, cmd, type);
+}
+
+static inline int __must_check
+HYPERVISOR_vcpu_op(
+	int cmd, unsigned int vcpuid, void *extra_args)
+{
+	return _hypercall3(int, vcpu_op, cmd, vcpuid, extra_args);
+}
+
+static inline int __must_check
+HYPERVISOR_set_segment_base(
+	int reg, unsigned long value)
+{
+	return _hypercall2(int, set_segment_base, reg, value);
+}
+
+static inline int __must_check
+HYPERVISOR_suspend(
+	unsigned long srec)
+{
+	struct sched_shutdown sched_shutdown = {
+		.reason = SHUTDOWN_suspend
+	};
+
+	int rc = _hypercall3(int, sched_op, SCHEDOP_shutdown,
+			     &sched_shutdown, srec);
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (rc == -ENOSYS)
+		rc = _hypercall3(int, sched_op_compat, SCHEDOP_shutdown,
+				 SHUTDOWN_suspend, srec);
+#endif
+
+	return rc;
+}
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+static inline int
+HYPERVISOR_nmi_op(
+	unsigned long op, void *arg)
+{
+	return _hypercall2(int, nmi_op, op, arg);
+}
+#endif
+
+#ifndef CONFIG_XEN
+static inline unsigned long __must_check
+HYPERVISOR_hvm_op(
+    int op, void *arg)
+{
+    return _hypercall2(unsigned long, hvm_op, op, arg);
+}
+#endif
+
+static inline int __must_check
+HYPERVISOR_callback_op(
+	int cmd, const void *arg)
+{
+	return _hypercall2(int, callback_op, cmd, arg);
+}
+
+static inline int __must_check
+HYPERVISOR_xenoprof_op(
+	int op, void *arg)
+{
+	return _hypercall2(int, xenoprof_op, op, arg);
+}
+
+static inline int __must_check
+HYPERVISOR_kexec_op(
+	unsigned long op, void *args)
+{
+	return _hypercall2(int, kexec_op, op, args);
+}
+
+#endif /* __HYPERCALL_H__ */
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/hypervisor.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/hypervisor.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/hypervisor.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/hypervisor.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,2 @@
+
+#include <asm-i386/mach-xen/asm/hypervisor.h>
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/io.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/io.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/io.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/io.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,329 @@
+#ifndef _ASM_IO_H
+#define _ASM_IO_H
+
+#include <asm/fixmap.h>
+
+/*
+ * This file contains the definitions for the x86 IO instructions
+ * inb/inw/inl/outb/outw/outl and the "string versions" of the same
+ * (insb/insw/insl/outsb/outsw/outsl). You can also use "pausing"
+ * versions of the single-IO instructions (inb_p/inw_p/..).
+ *
+ * This file is not meant to be obfuscating: it's just complicated
+ * to (a) handle it all in a way that makes gcc able to optimize it
+ * as well as possible and (b) trying to avoid writing the same thing
+ * over and over again with slight variations and possibly making a
+ * mistake somewhere.
+ */
+
+/*
+ * Thanks to James van Artsdalen for a better timing-fix than
+ * the two short jumps: using outb's to a nonexistent port seems
+ * to guarantee better timings even on fast machines.
+ *
+ * On the other hand, I'd like to be sure of a non-existent port:
+ * I feel a bit unsafe about using 0x80 (should be safe, though)
+ *
+ *		Linus
+ */
+
+ /*
+  *  Bit simplified and optimized by Jan Hubicka
+  *  Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999.
+  *
+  *  isa_memset_io, isa_memcpy_fromio, isa_memcpy_toio added,
+  *  isa_read[wl] and isa_write[wl] fixed
+  *  - Arnaldo Carvalho de Melo <acme@conectiva.com.br>
+  */
+
+#define __SLOW_DOWN_IO "\noutb %%al,$0x80"
+
+#ifdef REALLY_SLOW_IO
+#define __FULL_SLOW_DOWN_IO __SLOW_DOWN_IO __SLOW_DOWN_IO __SLOW_DOWN_IO __SLOW_DOWN_IO
+#else
+#define __FULL_SLOW_DOWN_IO __SLOW_DOWN_IO
+#endif
+
+/*
+ * Talk about misusing macros..
+ */
+#define __OUT1(s,x) \
+static inline void out##s(unsigned x value, unsigned short port) {
+
+#define __OUT2(s,s1,s2) \
+__asm__ __volatile__ ("out" #s " %" s1 "0,%" s2 "1"
+
+#define __OUT(s,s1,x) \
+__OUT1(s,x) __OUT2(s,s1,"w") : : "a" (value), "Nd" (port)); } \
+__OUT1(s##_p,x) __OUT2(s,s1,"w") __FULL_SLOW_DOWN_IO : : "a" (value), "Nd" (port));} \
+
+#define __IN1(s) \
+static inline RETURN_TYPE in##s(unsigned short port) { RETURN_TYPE _v;
+
+#define __IN2(s,s1,s2) \
+__asm__ __volatile__ ("in" #s " %" s2 "1,%" s1 "0"
+
+#define __IN(s,s1,i...) \
+__IN1(s) __IN2(s,s1,"w") : "=a" (_v) : "Nd" (port) ,##i ); return _v; } \
+__IN1(s##_p) __IN2(s,s1,"w") __FULL_SLOW_DOWN_IO : "=a" (_v) : "Nd" (port) ,##i ); return _v; } \
+
+#define __INS(s) \
+static inline void ins##s(unsigned short port, void * addr, unsigned long count) \
+{ __asm__ __volatile__ ("rep ; ins" #s \
+: "=D" (addr), "=c" (count) : "d" (port),"0" (addr),"1" (count)); }
+
+#define __OUTS(s) \
+static inline void outs##s(unsigned short port, const void * addr, unsigned long count) \
+{ __asm__ __volatile__ ("rep ; outs" #s \
+: "=S" (addr), "=c" (count) : "d" (port),"0" (addr),"1" (count)); }
+
+#define RETURN_TYPE unsigned char
+__IN(b,"")
+#undef RETURN_TYPE
+#define RETURN_TYPE unsigned short
+__IN(w,"")
+#undef RETURN_TYPE
+#define RETURN_TYPE unsigned int
+__IN(l,"")
+#undef RETURN_TYPE
+
+__OUT(b,"b",char)
+__OUT(w,"w",short)
+__OUT(l,,int)
+
+__INS(b)
+__INS(w)
+__INS(l)
+
+__OUTS(b)
+__OUTS(w)
+__OUTS(l)
+
+#define IO_SPACE_LIMIT 0xffff
+
+#if defined(__KERNEL__) && __x86_64__
+
+#include <linux/vmalloc.h>
+
+#ifndef __i386__
+/*
+ * Change virtual addresses to physical addresses and vv.
+ * These are pretty trivial
+ */
+static inline unsigned long virt_to_phys(volatile void * address)
+{
+	return __pa(address);
+}
+
+static inline void * phys_to_virt(unsigned long address)
+{
+	return __va(address);
+}
+
+#define virt_to_bus(_x) phys_to_machine(__pa(_x))
+#define bus_to_virt(_x) __va(machine_to_phys(_x))
+#endif
+
+/*
+ * Change "struct page" to physical address.
+ */
+#define page_to_pseudophys(page) ((dma_addr_t)page_to_pfn(page) << PAGE_SHIFT)
+#define page_to_phys(page)	 (phys_to_machine(page_to_pseudophys(page)))
+#define page_to_bus(page)	 (phys_to_machine(page_to_pseudophys(page)))
+
+#define bio_to_pseudophys(bio)	 (page_to_pseudophys(bio_page((bio))) + \
+				  (unsigned long) bio_offset((bio)))
+#define bvec_to_pseudophys(bv)	 (page_to_pseudophys((bv)->bv_page) + \
+				  (unsigned long) (bv)->bv_offset)
+
+#define BIOVEC_PHYS_MERGEABLE(vec1, vec2)	\
+	(((bvec_to_phys((vec1)) + (vec1)->bv_len) == bvec_to_phys((vec2))) && \
+	 ((bvec_to_pseudophys((vec1)) + (vec1)->bv_len) == \
+	  bvec_to_pseudophys((vec2))))
+
+#include <asm-generic/iomap.h>
+
+extern void __iomem *__ioremap(unsigned long offset, unsigned long size, unsigned long flags);
+
+static inline void __iomem * ioremap (unsigned long offset, unsigned long size)
+{
+	return __ioremap(offset, size, 0);
+}
+
+extern void *bt_ioremap(unsigned long addr, unsigned long size);
+extern void bt_iounmap(void *addr, unsigned long size);
+#define early_ioremap bt_ioremap
+#define early_iounmap bt_iounmap
+
+/*
+ * This one maps high address device memory and turns off caching for that area.
+ * it's useful if some control registers are in such an area and write combining
+ * or read caching is not desirable:
+ */
+extern void __iomem * ioremap_nocache (unsigned long offset, unsigned long size);
+extern void iounmap(volatile void __iomem *addr);
+
+/*
+ * ISA I/O bus memory addresses are 1:1 with the physical address.
+ */
+
+#define isa_virt_to_bus(_x) ({ BUG(); virt_to_bus(_x); })
+#define isa_page_to_bus(_x) isa_page_to_bus_is_UNSUPPORTED->x
+#define isa_bus_to_virt(_x) (void *)(__fix_to_virt(FIX_ISAMAP_BEGIN) + (_x))
+
+/*
+ * However PCI ones are not necessarily 1:1 and therefore these interfaces
+ * are forbidden in portable PCI drivers.
+ *
+ * Allow them on x86 for legacy drivers, though.
+ */
+#define virt_to_bus(_x) phys_to_machine(__pa(_x))
+#define bus_to_virt(_x) __va(machine_to_phys(_x))
+
+/*
+ * readX/writeX() are used to access memory mapped devices. On some
+ * architectures the memory mapped IO stuff needs to be accessed
+ * differently. On the x86 architecture, we just read/write the
+ * memory location directly.
+ */
+
+static inline __u8 __readb(const volatile void __iomem *addr)
+{
+	return *(__force volatile __u8 *)addr;
+}
+static inline __u16 __readw(const volatile void __iomem *addr)
+{
+	return *(__force volatile __u16 *)addr;
+}
+static __always_inline __u32 __readl(const volatile void __iomem *addr)
+{
+	return *(__force volatile __u32 *)addr;
+}
+static inline __u64 __readq(const volatile void __iomem *addr)
+{
+	return *(__force volatile __u64 *)addr;
+}
+#define readb(x) __readb(x)
+#define readw(x) __readw(x)
+#define readl(x) __readl(x)
+#define readq(x) __readq(x)
+#define readb_relaxed(a) readb(a)
+#define readw_relaxed(a) readw(a)
+#define readl_relaxed(a) readl(a)
+#define readq_relaxed(a) readq(a)
+#define __raw_readb readb
+#define __raw_readw readw
+#define __raw_readl readl
+#define __raw_readq readq
+
+#define mmiowb()
+
+static inline void __writel(__u32 b, volatile void __iomem *addr)
+{
+	*(__force volatile __u32 *)addr = b;
+}
+static inline void __writeq(__u64 b, volatile void __iomem *addr)
+{
+	*(__force volatile __u64 *)addr = b;
+}
+static inline void __writeb(__u8 b, volatile void __iomem *addr)
+{
+	*(__force volatile __u8 *)addr = b;
+}
+static inline void __writew(__u16 b, volatile void __iomem *addr)
+{
+	*(__force volatile __u16 *)addr = b;
+}
+#define writeq(val,addr) __writeq((val),(addr))
+#define writel(val,addr) __writel((val),(addr))
+#define writew(val,addr) __writew((val),(addr))
+#define writeb(val,addr) __writeb((val),(addr))
+#define __raw_writeb writeb
+#define __raw_writew writew
+#define __raw_writel writel
+#define __raw_writeq writeq
+
+void __memcpy_fromio(void*,unsigned long,unsigned);
+void __memcpy_toio(unsigned long,const void*,unsigned);
+
+static inline void memcpy_fromio(void *to, const volatile void __iomem *from, unsigned len)
+{
+	__memcpy_fromio(to,(unsigned long)from,len);
+}
+static inline void memcpy_toio(volatile void __iomem *to, const void *from, unsigned len)
+{
+	__memcpy_toio((unsigned long)to,from,len);
+}
+
+void memset_io(volatile void __iomem *a, int b, size_t c);
+
+/*
+ * ISA space is 'always mapped' on a typical x86 system, no need to
+ * explicitly ioremap() it. The fact that the ISA IO space is mapped
+ * to PAGE_OFFSET is pure coincidence - it does not mean ISA values
+ * are physical addresses. The following constant pointer can be
+ * used as the IO-area pointer (it can be iounmapped as well, so the
+ * analogy with PCI is quite large):
+ */
+#define __ISA_IO_base ((char __iomem *)(fix_to_virt(FIX_ISAMAP_BEGIN)))
+
+/*
+ * Again, x86-64 does not require mem IO specific function.
+ */
+
+#define eth_io_copy_and_sum(a,b,c,d)		eth_copy_and_sum((a),(void *)(b),(c),(d))
+
+/**
+ *	check_signature		-	find BIOS signatures
+ *	@io_addr: mmio address to check 
+ *	@signature:  signature block
+ *	@length: length of signature
+ *
+ *	Perform a signature comparison with the mmio address io_addr. This
+ *	address should have been obtained by ioremap.
+ *	Returns 1 on a match.
+ */
+ 
+static inline int check_signature(void __iomem *io_addr,
+	const unsigned char *signature, int length)
+{
+	int retval = 0;
+	do {
+		if (readb(io_addr) != *signature)
+			goto out;
+		io_addr++;
+		signature++;
+		length--;
+	} while (length);
+	retval = 1;
+out:
+	return retval;
+}
+
+/* Nothing to do */
+
+#define dma_cache_inv(_start,_size)		do { } while (0)
+#define dma_cache_wback(_start,_size)		do { } while (0)
+#define dma_cache_wback_inv(_start,_size)	do { } while (0)
+
+#define flush_write_buffers() 
+
+extern int iommu_bio_merge;
+#define BIO_VMERGE_BOUNDARY iommu_bio_merge
+
+/*
+ * Convert a physical pointer to a virtual kernel pointer for /dev/mem
+ * access
+ */
+#define xlate_dev_mem_ptr(p)	__va(p)
+
+/*
+ * Convert a virtual cached pointer to an uncached pointer
+ */
+#define xlate_dev_kmem_ptr(p)	p
+
+#endif /* __KERNEL__ */
+
+#define ARCH_HAS_DEV_MEM
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/irqflags.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/irqflags.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/irqflags.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/irqflags.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,139 @@
+/*
+ * include/asm-x86_64/irqflags.h
+ *
+ * IRQ flags handling
+ *
+ * This file gets included from lowlevel asm headers too, to provide
+ * wrapped versions of the local_irq_*() APIs, based on the
+ * raw_local_irq_*() functions from the lowlevel headers.
+ */
+#ifndef _ASM_IRQFLAGS_H
+#define _ASM_IRQFLAGS_H
+
+#ifndef __ASSEMBLY__
+/*
+ * Interrupt control:
+ */
+
+/*
+ * The use of 'barrier' in the following reflects their use as local-lock
+ * operations. Reentrancy must be prevented (e.g., __cli()) /before/ following
+ * critical operations are executed. All critical operations must complete
+ * /before/ reentrancy is permitted (e.g., __sti()). Alpha architecture also
+ * includes these barriers, for example.
+ */
+
+#define __raw_local_save_flags() (current_vcpu_info()->evtchn_upcall_mask)
+
+#define raw_local_save_flags(flags) \
+		do { (flags) = __raw_local_save_flags(); } while (0)
+
+#define raw_local_irq_restore(x)					\
+do {									\
+	vcpu_info_t *_vcpu;						\
+	barrier();							\
+	_vcpu = current_vcpu_info();		\
+	if ((_vcpu->evtchn_upcall_mask = (x)) == 0) {			\
+		barrier(); /* unmask then check (avoid races) */	\
+		if ( unlikely(_vcpu->evtchn_upcall_pending) )		\
+			force_evtchn_callback();			\
+	}								\
+} while (0)
+
+#ifdef CONFIG_X86_VSMP
+
+/*
+ * Interrupt control for the VSMP architecture:
+ */
+
+static inline void raw_local_irq_disable(void)
+{
+	unsigned long flags = __raw_local_save_flags();
+
+	raw_local_irq_restore((flags & ~(1 << 9)) | (1 << 18));
+}
+
+static inline void raw_local_irq_enable(void)
+{
+	unsigned long flags = __raw_local_save_flags();
+
+	raw_local_irq_restore((flags | (1 << 9)) & ~(1 << 18));
+}
+
+static inline int raw_irqs_disabled_flags(unsigned long flags)
+{
+	return !(flags & (1<<9)) || (flags & (1 << 18));
+}
+
+#else /* CONFIG_X86_VSMP */
+
+#define raw_local_irq_disable()						\
+do {									\
+	current_vcpu_info()->evtchn_upcall_mask = 1;					\
+	barrier();							\
+} while (0)
+
+#define raw_local_irq_enable()						\
+do {									\
+	vcpu_info_t *_vcpu;						\
+	barrier();							\
+	_vcpu = current_vcpu_info();		\
+	_vcpu->evtchn_upcall_mask = 0;					\
+	barrier(); /* unmask then check (avoid races) */		\
+	if ( unlikely(_vcpu->evtchn_upcall_pending) )			\
+		force_evtchn_callback();				\
+} while (0)
+
+static inline int raw_irqs_disabled_flags(unsigned long flags)
+{
+	return (flags != 0);
+}
+
+#endif
+
+/*
+ * For spinlocks, etc.:
+ */
+
+#define __raw_local_irq_save()						\
+({									\
+	unsigned long flags = __raw_local_save_flags();			\
+									\
+	raw_local_irq_disable();					\
+									\
+	flags;								\
+})
+
+#define raw_local_irq_save(flags) \
+		do { (flags) = __raw_local_irq_save(); } while (0)
+
+#define raw_irqs_disabled()						\
+({									\
+	unsigned long flags = __raw_local_save_flags();			\
+									\
+	raw_irqs_disabled_flags(flags);					\
+})
+
+/*
+ * Used in the idle loop; sti takes one instruction cycle
+ * to complete:
+ */
+void raw_safe_halt(void);
+
+/*
+ * Used when interrupts are already enabled or to
+ * shutdown the processor:
+ */
+void halt(void);
+
+#else /* __ASSEMBLY__: */
+# ifdef CONFIG_TRACE_IRQFLAGS
+#  define TRACE_IRQS_ON		call trace_hardirqs_on_thunk
+#  define TRACE_IRQS_OFF	call trace_hardirqs_off_thunk
+# else
+#  define TRACE_IRQS_ON
+#  define TRACE_IRQS_OFF
+# endif
+#endif
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/irq.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/irq.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/irq.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/irq.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,38 @@
+#ifndef _ASM_IRQ_H
+#define _ASM_IRQ_H
+
+/*
+ *	linux/include/asm/irq.h
+ *
+ *	(C) 1992, 1993 Linus Torvalds, (C) 1997 Ingo Molnar
+ *
+ *	IRQ/IPI changes taken from work by Thomas Radke
+ *	<tomsoft@informatik.tu-chemnitz.de>
+ */
+
+#include <linux/sched.h>
+/* include comes from machine specific directory */
+#include "irq_vectors.h"
+#include <asm/thread_info.h>
+
+static __inline__ int irq_canonicalize(int irq)
+{
+	return ((irq == 2) ? 9 : irq);
+}
+
+#ifdef CONFIG_X86_LOCAL_APIC
+#define ARCH_HAS_NMI_WATCHDOG		/* See include/linux/nmi.h */
+#endif
+
+#define KDB_VECTOR	0xf9
+
+# define irq_ctx_init(cpu) do { } while (0)
+
+#ifdef CONFIG_HOTPLUG_CPU
+#include <linux/cpumask.h>
+extern void fixup_irqs(cpumask_t map);
+#endif
+
+#define __ARCH_HAS_DO_SOFTIRQ 1
+
+#endif /* _ASM_IRQ_H */
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/maddr.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/maddr.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/maddr.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/maddr.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,161 @@
+#ifndef _X86_64_MADDR_H
+#define _X86_64_MADDR_H
+
+#include <xen/features.h>
+#include <xen/interface/xen.h>
+
+/**** MACHINE <-> PHYSICAL CONVERSION MACROS ****/
+#define INVALID_P2M_ENTRY	(~0UL)
+#define FOREIGN_FRAME_BIT	(1UL<<63)
+#define FOREIGN_FRAME(m)	((m) | FOREIGN_FRAME_BIT)
+
+/* Definitions for machine and pseudophysical addresses. */
+typedef unsigned long paddr_t;
+typedef unsigned long maddr_t;
+
+#ifdef CONFIG_XEN
+
+extern unsigned long *phys_to_machine_mapping;
+
+#undef machine_to_phys_mapping
+extern unsigned long *machine_to_phys_mapping;
+extern unsigned int   machine_to_phys_order;
+
+static inline unsigned long pfn_to_mfn(unsigned long pfn)
+{
+	if (xen_feature(XENFEAT_auto_translated_physmap))
+		return pfn;
+	BUG_ON(end_pfn && pfn >= end_pfn);
+	return phys_to_machine_mapping[pfn] & ~FOREIGN_FRAME_BIT;
+}
+
+static inline int phys_to_machine_mapping_valid(unsigned long pfn)
+{
+	if (xen_feature(XENFEAT_auto_translated_physmap))
+		return 1;
+	BUG_ON(end_pfn && pfn >= end_pfn);
+	return (phys_to_machine_mapping[pfn] != INVALID_P2M_ENTRY);
+}
+
+static inline unsigned long mfn_to_pfn(unsigned long mfn)
+{
+	unsigned long pfn;
+
+	if (xen_feature(XENFEAT_auto_translated_physmap))
+		return mfn;
+
+	if (unlikely((mfn >> machine_to_phys_order) != 0))
+		return end_pfn;
+
+	/* The array access can fail (e.g., device space beyond end of RAM). */
+	asm (
+		"1:	movq %1,%0\n"
+		"2:\n"
+		".section .fixup,\"ax\"\n"
+		"3:	movq %2,%0\n"
+		"	jmp  2b\n"
+		".previous\n"
+		".section __ex_table,\"a\"\n"
+		"	.align 8\n"
+		"	.quad 1b,3b\n"
+		".previous"
+		: "=r" (pfn)
+		: "m" (machine_to_phys_mapping[mfn]), "m" (end_pfn) );
+
+	return pfn;
+}
+
+/*
+ * We detect special mappings in one of two ways:
+ *  1. If the MFN is an I/O page then Xen will set the m2p entry
+ *     to be outside our maximum possible pseudophys range.
+ *  2. If the MFN belongs to a different domain then we will certainly
+ *     not have MFN in our p2m table. Conversely, if the page is ours,
+ *     then we'll have p2m(m2p(MFN))==MFN.
+ * If we detect a special mapping then it doesn't have a 'struct page'.
+ * We force !pfn_valid() by returning an out-of-range pointer.
+ *
+ * NB. These checks require that, for any MFN that is not in our reservation,
+ * there is no PFN such that p2m(PFN) == MFN. Otherwise we can get confused if
+ * we are foreign-mapping the MFN, and the other domain as m2p(MFN) == PFN.
+ * Yikes! Various places must poke in INVALID_P2M_ENTRY for safety.
+ *
+ * NB2. When deliberately mapping foreign pages into the p2m table, you *must*
+ *      use FOREIGN_FRAME(). This will cause pte_pfn() to choke on it, as we
+ *      require. In all the cases we care about, the FOREIGN_FRAME bit is
+ *      masked (e.g., pfn_to_mfn()) so behaviour there is correct.
+ */
+static inline unsigned long mfn_to_local_pfn(unsigned long mfn)
+{
+	unsigned long pfn = mfn_to_pfn(mfn);
+	if ((pfn < end_pfn)
+	    && !xen_feature(XENFEAT_auto_translated_physmap)
+	    && (phys_to_machine_mapping[pfn] != mfn))
+		return end_pfn; /* force !pfn_valid() */
+	return pfn;
+}
+
+static inline void set_phys_to_machine(unsigned long pfn, unsigned long mfn)
+{
+	BUG_ON(end_pfn && pfn >= end_pfn);
+	if (xen_feature(XENFEAT_auto_translated_physmap)) {
+		BUG_ON(pfn != mfn && mfn != INVALID_P2M_ENTRY);
+		return;
+	}
+	phys_to_machine_mapping[pfn] = mfn;
+}
+
+static inline maddr_t phys_to_machine(paddr_t phys)
+{
+	maddr_t machine = pfn_to_mfn(phys >> PAGE_SHIFT);
+	machine = (machine << PAGE_SHIFT) | (phys & ~PAGE_MASK);
+	return machine;
+}
+
+static inline paddr_t machine_to_phys(maddr_t machine)
+{
+	paddr_t phys = mfn_to_pfn(machine >> PAGE_SHIFT);
+	phys = (phys << PAGE_SHIFT) | (machine & ~PAGE_MASK);
+	return phys;
+}
+
+static inline paddr_t pte_phys_to_machine(paddr_t phys)
+{
+	maddr_t machine;
+	machine = pfn_to_mfn((phys & PHYSICAL_PAGE_MASK) >> PAGE_SHIFT);
+	machine = (machine << PAGE_SHIFT) | (phys & ~PHYSICAL_PAGE_MASK);
+	return machine;
+}
+
+static inline paddr_t pte_machine_to_phys(maddr_t machine)
+{
+	paddr_t phys;
+	phys = mfn_to_pfn((machine & PHYSICAL_PAGE_MASK) >> PAGE_SHIFT);
+	phys = (phys << PAGE_SHIFT) | (machine & ~PHYSICAL_PAGE_MASK);
+	return phys;
+}
+
+#define __pte_ma(x)     ((pte_t) { (x) } )
+#define pfn_pte_ma(pfn, prot)	__pte_ma((((pfn) << PAGE_SHIFT) | pgprot_val(prot)) & __supported_pte_mask)
+
+#else /* !CONFIG_XEN */
+
+#define pfn_to_mfn(pfn) (pfn)
+#define mfn_to_pfn(mfn) (mfn)
+#define mfn_to_local_pfn(mfn) (mfn)
+#define set_phys_to_machine(pfn, mfn) ((void)0)
+#define phys_to_machine_mapping_valid(pfn) (1)
+#define phys_to_machine(phys) ((maddr_t)(phys))
+#define machine_to_phys(mach) ((paddr_t)(mach))
+#define pfn_pte_ma(pfn, prot) pfn_pte(pfn, prot)
+#define __pte_ma(x) __pte(x)
+
+#endif /* !CONFIG_XEN */
+
+/* VIRT <-> MACHINE conversion */
+#define virt_to_machine(v)	(phys_to_machine(__pa(v)))
+#define virt_to_mfn(v)		(pfn_to_mfn(__pa(v) >> PAGE_SHIFT))
+#define mfn_to_virt(m)		(__va(mfn_to_pfn(m) << PAGE_SHIFT))
+
+#endif /* _X86_64_MADDR_H */
+
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/mmu_context.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/mmu_context.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/mmu_context.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/mmu_context.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,136 @@
+#ifndef __X86_64_MMU_CONTEXT_H
+#define __X86_64_MMU_CONTEXT_H
+
+#include <asm/desc.h>
+#include <asm/atomic.h>
+#include <asm/pgalloc.h>
+#include <asm/page.h>
+#include <asm/pda.h>
+#include <asm/pgtable.h>
+#include <asm/tlbflush.h>
+
+/*
+ * possibly do the LDT unload here?
+ */
+int init_new_context(struct task_struct *tsk, struct mm_struct *mm);
+void destroy_context(struct mm_struct *mm);
+
+static inline void enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk)
+{
+#if defined(CONFIG_SMP) && !defined(CONFIG_XEN)
+	if (read_pda(mmu_state) == TLBSTATE_OK) 
+		write_pda(mmu_state, TLBSTATE_LAZY);
+#endif
+}
+
+#define prepare_arch_switch(next)	__prepare_arch_switch()
+
+static inline void __prepare_arch_switch(void)
+{
+	/*
+	 * Save away %es, %ds, %fs and %gs. Must happen before reload
+	 * of cr3/ldt (i.e., not in __switch_to).
+	 */
+	__asm__ __volatile__ (
+		"mov %%es,%0 ; mov %%ds,%1 ; mov %%fs,%2 ; mov %%gs,%3"
+		: "=m" (current->thread.es),
+		  "=m" (current->thread.ds),
+		  "=m" (current->thread.fsindex),
+		  "=m" (current->thread.gsindex) );
+
+	if (current->thread.ds)
+		__asm__ __volatile__ ( "movl %0,%%ds" : : "r" (0) );
+
+	if (current->thread.es)
+		__asm__ __volatile__ ( "movl %0,%%es" : : "r" (0) );
+
+	if (current->thread.fsindex) {
+		__asm__ __volatile__ ( "movl %0,%%fs" : : "r" (0) );
+		current->thread.fs = 0;
+	}
+
+	if (current->thread.gsindex) {
+		load_gs_index(0);
+		current->thread.gs = 0;
+	}
+}
+
+extern void mm_pin(struct mm_struct *mm);
+extern void mm_unpin(struct mm_struct *mm);
+void mm_pin_all(void);
+
+static inline void load_cr3(pgd_t *pgd)
+{
+	asm volatile("movq %0,%%cr3" :: "r" (phys_to_machine(__pa(pgd))) :
+		     "memory");
+}
+
+static inline void switch_mm(struct mm_struct *prev, struct mm_struct *next, 
+			     struct task_struct *tsk)
+{
+	unsigned cpu = smp_processor_id();
+	struct mmuext_op _op[3], *op = _op;
+
+	if (likely(prev != next)) {
+		BUG_ON(!xen_feature(XENFEAT_writable_page_tables) &&
+		       !next->context.pinned);
+
+		/* stop flush ipis for the previous mm */
+		cpu_clear(cpu, prev->cpu_vm_mask);
+#if defined(CONFIG_SMP) && !defined(CONFIG_XEN)
+		write_pda(mmu_state, TLBSTATE_OK);
+		write_pda(active_mm, next);
+#endif
+		cpu_set(cpu, next->cpu_vm_mask);
+
+		/* load_cr3(next->pgd) */
+		op->cmd = MMUEXT_NEW_BASEPTR;
+		op->arg1.mfn = pfn_to_mfn(__pa(next->pgd) >> PAGE_SHIFT);
+		op++;
+
+		/* xen_new_user_pt(__pa(__user_pgd(next->pgd))) */
+		op->cmd = MMUEXT_NEW_USER_BASEPTR;
+		op->arg1.mfn = pfn_to_mfn(__pa(__user_pgd(next->pgd)) >> PAGE_SHIFT);
+		op++;
+		
+		if (unlikely(next->context.ldt != prev->context.ldt)) {
+			/* load_LDT_nolock(&next->context, cpu) */
+			op->cmd = MMUEXT_SET_LDT;
+			op->arg1.linear_addr = (unsigned long)next->context.ldt;
+			op->arg2.nr_ents     = next->context.size;
+			op++;
+		}
+
+		BUG_ON(HYPERVISOR_mmuext_op(_op, op-_op, NULL, DOMID_SELF));
+	}
+#if defined(CONFIG_SMP) && !defined(CONFIG_XEN)
+	else {
+		write_pda(mmu_state, TLBSTATE_OK);
+		if (read_pda(active_mm) != next)
+			out_of_line_bug();
+		if (!cpu_test_and_set(cpu, next->cpu_vm_mask)) {
+			/* We were in lazy tlb mode and leave_mm disabled 
+			 * tlb flush IPI delivery. We must reload CR3
+			 * to make sure to use no freed page tables.
+			 */
+                        load_cr3(next->pgd);
+                        xen_new_user_pt(__pa(__user_pgd(next->pgd)));		
+			load_LDT_nolock(&next->context, cpu);
+		}
+	}
+#endif
+}
+
+#define deactivate_mm(tsk,mm)	do { \
+	load_gs_index(0); \
+	asm volatile("movl %0,%%fs"::"r"(0));  \
+} while(0)
+
+static inline void activate_mm(struct mm_struct *prev, struct mm_struct *next)
+{
+	if (!next->context.pinned)
+		mm_pin(next);
+	switch_mm(prev, next, NULL);
+}
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/mmu.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/mmu.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/mmu.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/mmu.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,38 @@
+#ifndef __x86_64_MMU_H
+#define __x86_64_MMU_H
+
+#include <linux/spinlock.h>
+#include <asm/semaphore.h>
+
+/*
+ * The x86_64 doesn't have a mmu context, but
+ * we put the segment information here.
+ *
+ * cpu_vm_mask is used to optimize ldt flushing.
+ */
+typedef struct { 
+	void *ldt;
+	rwlock_t ldtlock; 
+	int size;
+	struct semaphore sem; 
+#ifdef CONFIG_XEN
+	unsigned pinned:1;
+	unsigned has_foreign_mappings:1;
+	struct list_head unpinned;
+#endif
+} mm_context_t;
+
+#ifdef CONFIG_XEN
+extern struct list_head mm_unpinned;
+extern spinlock_t mm_unpinned_lock;
+
+/* mm/memory.c:exit_mmap hook */
+extern void _arch_exit_mmap(struct mm_struct *mm);
+#define arch_exit_mmap(_mm) _arch_exit_mmap(_mm)
+
+/* kernel/fork.c:dup_mmap hook */
+extern void _arch_dup_mmap(struct mm_struct *mm);
+#define arch_dup_mmap(mm, oldmm) ((void)(oldmm), _arch_dup_mmap(mm))
+#endif
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/msr.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/msr.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/msr.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/msr.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,399 @@
+#ifndef X86_64_MSR_H
+#define X86_64_MSR_H 1
+
+#ifndef __ASSEMBLY__
+/*
+ * Access to machine-specific registers (available on 586 and better only)
+ * Note: the rd* operations modify the parameters directly (without using
+ * pointer indirection), this allows gcc to optimize better
+ */
+
+#define rdmsr(msr,val1,val2) \
+       __asm__ __volatile__("rdmsr" \
+			    : "=a" (val1), "=d" (val2) \
+			    : "c" (msr))
+
+
+#define rdmsrl(msr,val) do { unsigned long a__,b__; \
+       __asm__ __volatile__("rdmsr" \
+			    : "=a" (a__), "=d" (b__) \
+			    : "c" (msr)); \
+       val = a__ | (b__<<32); \
+} while(0)
+
+#define wrmsr(msr,val1,val2) \
+     __asm__ __volatile__("wrmsr" \
+			  : /* no outputs */ \
+			  : "c" (msr), "a" (val1), "d" (val2))
+
+#define wrmsrl(msr,val) wrmsr(msr,(__u32)((__u64)(val)),((__u64)(val))>>32) 
+
+/* wrmsr with exception handling */
+#define wrmsr_safe(msr,a,b) ({ int ret__;			\
+	asm volatile("2: wrmsr ; xorl %0,%0\n"			\
+		     "1:\n\t"					\
+		     ".section .fixup,\"ax\"\n\t"		\
+		     "3:  movl %4,%0 ; jmp 1b\n\t"		\
+		     ".previous\n\t"				\
+ 		     ".section __ex_table,\"a\"\n"		\
+		     "   .align 8\n\t"				\
+		     "   .quad 	2b,3b\n\t"			\
+		     ".previous"				\
+		     : "=a" (ret__)				\
+		     : "c" (msr), "0" (a), "d" (b), "i" (-EFAULT)); \
+	ret__; })
+
+#define checking_wrmsrl(msr,val) wrmsr_safe(msr,(u32)(val),(u32)((val)>>32))
+
+#define rdmsr_safe(msr,a,b) \
+	({ int ret__;						\
+	  asm volatile ("1:       rdmsr\n"			\
+                      "2:\n"					\
+                      ".section .fixup,\"ax\"\n"		\
+                      "3:       movl %4,%0\n"			\
+                      " jmp 2b\n"				\
+                      ".previous\n"				\
+                      ".section __ex_table,\"a\"\n"		\
+                      " .align 8\n"				\
+                      " .quad 1b,3b\n"				\
+                      ".previous":"=&bDS" (ret__), "=a"(*(a)), "=d"(*(b))\
+                      :"c"(msr), "i"(-EIO), "0"(0));		\
+	  ret__; })		
+
+#define rdtsc(low,high) \
+     __asm__ __volatile__("rdtsc" : "=a" (low), "=d" (high))
+
+#define rdtscl(low) \
+     __asm__ __volatile__ ("rdtsc" : "=a" (low) : : "edx")
+
+#define rdtscll(val) do { \
+     unsigned int __a,__d; \
+     asm volatile("rdtsc" : "=a" (__a), "=d" (__d)); \
+     (val) = ((unsigned long)__a) | (((unsigned long)__d)<<32); \
+} while(0)
+
+#define write_tsc(val1,val2) wrmsr(0x10, val1, val2)
+
+#define rdpmc(counter,low,high) \
+     __asm__ __volatile__("rdpmc" \
+			  : "=a" (low), "=d" (high) \
+			  : "c" (counter))
+
+static inline void cpuid(int op, unsigned int *eax, unsigned int *ebx,
+			 unsigned int *ecx, unsigned int *edx)
+{
+	__asm__(XEN_CPUID
+		: "=a" (*eax),
+		  "=b" (*ebx),
+		  "=c" (*ecx),
+		  "=d" (*edx)
+		: "0" (op));
+}
+
+/* Some CPUID calls want 'count' to be placed in ecx */
+static inline void cpuid_count(int op, int count, int *eax, int *ebx, int *ecx,
+	       	int *edx)
+{
+	__asm__(XEN_CPUID
+		: "=a" (*eax),
+		  "=b" (*ebx),
+		  "=c" (*ecx),
+		  "=d" (*edx)
+		: "0" (op), "c" (count));
+}
+
+/*
+ * CPUID functions returning a single datum
+ */
+static inline unsigned int cpuid_eax(unsigned int op)
+{
+	unsigned int eax;
+
+	__asm__(XEN_CPUID
+		: "=a" (eax)
+		: "0" (op)
+		: "bx", "cx", "dx");
+	return eax;
+}
+static inline unsigned int cpuid_ebx(unsigned int op)
+{
+	unsigned int eax, ebx;
+
+	__asm__(XEN_CPUID
+		: "=a" (eax), "=b" (ebx)
+		: "0" (op)
+		: "cx", "dx" );
+	return ebx;
+}
+static inline unsigned int cpuid_ecx(unsigned int op)
+{
+	unsigned int eax, ecx;
+
+	__asm__(XEN_CPUID
+		: "=a" (eax), "=c" (ecx)
+		: "0" (op)
+		: "bx", "dx" );
+	return ecx;
+}
+static inline unsigned int cpuid_edx(unsigned int op)
+{
+	unsigned int eax, edx;
+
+	__asm__(XEN_CPUID
+		: "=a" (eax), "=d" (edx)
+		: "0" (op)
+		: "bx", "cx");
+	return edx;
+}
+
+#define MSR_IA32_UCODE_WRITE		0x79
+#define MSR_IA32_UCODE_REV		0x8b
+
+
+#endif
+
+/* AMD/K8 specific MSRs */ 
+#define MSR_EFER 0xc0000080		/* extended feature register */
+#define MSR_STAR 0xc0000081		/* legacy mode SYSCALL target */
+#define MSR_LSTAR 0xc0000082 		/* long mode SYSCALL target */
+#define MSR_CSTAR 0xc0000083		/* compatibility mode SYSCALL target */
+#define MSR_SYSCALL_MASK 0xc0000084	/* EFLAGS mask for syscall */
+#define MSR_FS_BASE 0xc0000100		/* 64bit GS base */
+#define MSR_GS_BASE 0xc0000101		/* 64bit FS base */
+#define MSR_KERNEL_GS_BASE  0xc0000102	/* SwapGS GS shadow (or USER_GS from kernel) */ 
+/* EFER bits: */ 
+#define _EFER_SCE 0  /* SYSCALL/SYSRET */
+#define _EFER_LME 8  /* Long mode enable */
+#define _EFER_LMA 10 /* Long mode active (read-only) */
+#define _EFER_NX 11  /* No execute enable */
+
+#define EFER_SCE (1<<_EFER_SCE)
+#define EFER_LME (1<<_EFER_LME)
+#define EFER_LMA (1<<_EFER_LMA)
+#define EFER_NX (1<<_EFER_NX)
+
+/* Intel MSRs. Some also available on other CPUs */
+#define MSR_IA32_TSC		0x10
+#define MSR_IA32_PLATFORM_ID	0x17
+
+#define MSR_IA32_PERFCTR0      0xc1
+#define MSR_IA32_PERFCTR1      0xc2
+
+#define MSR_MTRRcap		0x0fe
+#define MSR_IA32_BBL_CR_CTL        0x119
+
+#define MSR_IA32_SYSENTER_CS	0x174
+#define MSR_IA32_SYSENTER_ESP	0x175
+#define MSR_IA32_SYSENTER_EIP	0x176
+
+#define MSR_IA32_MCG_CAP       0x179
+#define MSR_IA32_MCG_STATUS        0x17a
+#define MSR_IA32_MCG_CTL       0x17b
+
+#define MSR_IA32_EVNTSEL0      0x186
+#define MSR_IA32_EVNTSEL1      0x187
+
+#define MSR_IA32_DEBUGCTLMSR       0x1d9
+#define MSR_IA32_LASTBRANCHFROMIP  0x1db
+#define MSR_IA32_LASTBRANCHTOIP        0x1dc
+#define MSR_IA32_LASTINTFROMIP     0x1dd
+#define MSR_IA32_LASTINTTOIP       0x1de
+
+#define MSR_MTRRfix64K_00000	0x250
+#define MSR_MTRRfix16K_80000	0x258
+#define MSR_MTRRfix16K_A0000	0x259
+#define MSR_MTRRfix4K_C0000	0x268
+#define MSR_MTRRfix4K_C8000	0x269
+#define MSR_MTRRfix4K_D0000	0x26a
+#define MSR_MTRRfix4K_D8000	0x26b
+#define MSR_MTRRfix4K_E0000	0x26c
+#define MSR_MTRRfix4K_E8000	0x26d
+#define MSR_MTRRfix4K_F0000	0x26e
+#define MSR_MTRRfix4K_F8000	0x26f
+#define MSR_MTRRdefType		0x2ff
+
+#define MSR_IA32_MC0_CTL       0x400
+#define MSR_IA32_MC0_STATUS        0x401
+#define MSR_IA32_MC0_ADDR      0x402
+#define MSR_IA32_MC0_MISC      0x403
+
+#define MSR_P6_PERFCTR0			0xc1
+#define MSR_P6_PERFCTR1			0xc2
+#define MSR_P6_EVNTSEL0			0x186
+#define MSR_P6_EVNTSEL1			0x187
+
+/* K7/K8 MSRs. Not complete. See the architecture manual for a more complete list. */
+#define MSR_K7_EVNTSEL0            0xC0010000
+#define MSR_K7_PERFCTR0            0xC0010004
+#define MSR_K7_EVNTSEL1            0xC0010001
+#define MSR_K7_PERFCTR1            0xC0010005
+#define MSR_K7_EVNTSEL2            0xC0010002
+#define MSR_K7_PERFCTR2            0xC0010006
+#define MSR_K7_EVNTSEL3            0xC0010003
+#define MSR_K7_PERFCTR3            0xC0010007
+#define MSR_K8_TOP_MEM1		   0xC001001A
+#define MSR_K8_TOP_MEM2		   0xC001001D
+#define MSR_K8_SYSCFG		   0xC0010010
+#define MSR_K8_HWCR		   0xC0010015
+
+/* K6 MSRs */
+#define MSR_K6_EFER			0xC0000080
+#define MSR_K6_STAR			0xC0000081
+#define MSR_K6_WHCR			0xC0000082
+#define MSR_K6_UWCCR			0xC0000085
+#define MSR_K6_PSOR			0xC0000087
+#define MSR_K6_PFIR			0xC0000088
+
+/* Centaur-Hauls/IDT defined MSRs. */
+#define MSR_IDT_FCR1			0x107
+#define MSR_IDT_FCR2			0x108
+#define MSR_IDT_FCR3			0x109
+#define MSR_IDT_FCR4			0x10a
+
+#define MSR_IDT_MCR0			0x110
+#define MSR_IDT_MCR1			0x111
+#define MSR_IDT_MCR2			0x112
+#define MSR_IDT_MCR3			0x113
+#define MSR_IDT_MCR4			0x114
+#define MSR_IDT_MCR5			0x115
+#define MSR_IDT_MCR6			0x116
+#define MSR_IDT_MCR7			0x117
+#define MSR_IDT_MCR_CTRL		0x120
+
+/* VIA Cyrix defined MSRs*/
+#define MSR_VIA_FCR			0x1107
+#define MSR_VIA_LONGHAUL		0x110a
+#define MSR_VIA_RNG			0x110b
+#define MSR_VIA_BCR2			0x1147
+
+/* Intel defined MSRs. */
+#define MSR_IA32_P5_MC_ADDR		0
+#define MSR_IA32_P5_MC_TYPE		1
+#define MSR_IA32_PLATFORM_ID		0x17
+#define MSR_IA32_EBL_CR_POWERON		0x2a
+
+#define MSR_IA32_APICBASE               0x1b
+#define MSR_IA32_APICBASE_BSP           (1<<8)
+#define MSR_IA32_APICBASE_ENABLE        (1<<11)
+#define MSR_IA32_APICBASE_BASE          (0xfffff<<12)
+
+/* P4/Xeon+ specific */
+#define MSR_IA32_MCG_EAX		0x180
+#define MSR_IA32_MCG_EBX		0x181
+#define MSR_IA32_MCG_ECX		0x182
+#define MSR_IA32_MCG_EDX		0x183
+#define MSR_IA32_MCG_ESI		0x184
+#define MSR_IA32_MCG_EDI		0x185
+#define MSR_IA32_MCG_EBP		0x186
+#define MSR_IA32_MCG_ESP		0x187
+#define MSR_IA32_MCG_EFLAGS		0x188
+#define MSR_IA32_MCG_EIP		0x189
+#define MSR_IA32_MCG_RESERVED		0x18A
+
+#define MSR_P6_EVNTSEL0			0x186
+#define MSR_P6_EVNTSEL1			0x187
+
+#define MSR_IA32_PERF_STATUS		0x198
+#define MSR_IA32_PERF_CTL		0x199
+
+#define MSR_IA32_THERM_CONTROL		0x19a
+#define MSR_IA32_THERM_INTERRUPT	0x19b
+#define MSR_IA32_THERM_STATUS		0x19c
+#define MSR_IA32_MISC_ENABLE		0x1a0
+
+#define MSR_IA32_DEBUGCTLMSR		0x1d9
+#define MSR_IA32_LASTBRANCHFROMIP	0x1db
+#define MSR_IA32_LASTBRANCHTOIP		0x1dc
+#define MSR_IA32_LASTINTFROMIP		0x1dd
+#define MSR_IA32_LASTINTTOIP		0x1de
+
+#define MSR_IA32_MC0_CTL		0x400
+#define MSR_IA32_MC0_STATUS		0x401
+#define MSR_IA32_MC0_ADDR		0x402
+#define MSR_IA32_MC0_MISC		0x403
+
+/* Pentium IV performance counter MSRs */
+#define MSR_P4_BPU_PERFCTR0 		0x300
+#define MSR_P4_BPU_PERFCTR1 		0x301
+#define MSR_P4_BPU_PERFCTR2 		0x302
+#define MSR_P4_BPU_PERFCTR3 		0x303
+#define MSR_P4_MS_PERFCTR0 		0x304
+#define MSR_P4_MS_PERFCTR1 		0x305
+#define MSR_P4_MS_PERFCTR2 		0x306
+#define MSR_P4_MS_PERFCTR3 		0x307
+#define MSR_P4_FLAME_PERFCTR0 		0x308
+#define MSR_P4_FLAME_PERFCTR1 		0x309
+#define MSR_P4_FLAME_PERFCTR2 		0x30a
+#define MSR_P4_FLAME_PERFCTR3 		0x30b
+#define MSR_P4_IQ_PERFCTR0 		0x30c
+#define MSR_P4_IQ_PERFCTR1 		0x30d
+#define MSR_P4_IQ_PERFCTR2 		0x30e
+#define MSR_P4_IQ_PERFCTR3 		0x30f
+#define MSR_P4_IQ_PERFCTR4 		0x310
+#define MSR_P4_IQ_PERFCTR5 		0x311
+#define MSR_P4_BPU_CCCR0 		0x360
+#define MSR_P4_BPU_CCCR1 		0x361
+#define MSR_P4_BPU_CCCR2 		0x362
+#define MSR_P4_BPU_CCCR3 		0x363
+#define MSR_P4_MS_CCCR0 		0x364
+#define MSR_P4_MS_CCCR1 		0x365
+#define MSR_P4_MS_CCCR2 		0x366
+#define MSR_P4_MS_CCCR3 		0x367
+#define MSR_P4_FLAME_CCCR0 		0x368
+#define MSR_P4_FLAME_CCCR1 		0x369
+#define MSR_P4_FLAME_CCCR2 		0x36a
+#define MSR_P4_FLAME_CCCR3 		0x36b
+#define MSR_P4_IQ_CCCR0 		0x36c
+#define MSR_P4_IQ_CCCR1 		0x36d
+#define MSR_P4_IQ_CCCR2 		0x36e
+#define MSR_P4_IQ_CCCR3 		0x36f
+#define MSR_P4_IQ_CCCR4 		0x370
+#define MSR_P4_IQ_CCCR5 		0x371
+#define MSR_P4_ALF_ESCR0 		0x3ca
+#define MSR_P4_ALF_ESCR1 		0x3cb
+#define MSR_P4_BPU_ESCR0 		0x3b2
+#define MSR_P4_BPU_ESCR1 		0x3b3
+#define MSR_P4_BSU_ESCR0 		0x3a0
+#define MSR_P4_BSU_ESCR1 		0x3a1
+#define MSR_P4_CRU_ESCR0 		0x3b8
+#define MSR_P4_CRU_ESCR1 		0x3b9
+#define MSR_P4_CRU_ESCR2 		0x3cc
+#define MSR_P4_CRU_ESCR3 		0x3cd
+#define MSR_P4_CRU_ESCR4 		0x3e0
+#define MSR_P4_CRU_ESCR5 		0x3e1
+#define MSR_P4_DAC_ESCR0 		0x3a8
+#define MSR_P4_DAC_ESCR1 		0x3a9
+#define MSR_P4_FIRM_ESCR0 		0x3a4
+#define MSR_P4_FIRM_ESCR1 		0x3a5
+#define MSR_P4_FLAME_ESCR0 		0x3a6
+#define MSR_P4_FLAME_ESCR1 		0x3a7
+#define MSR_P4_FSB_ESCR0 		0x3a2
+#define MSR_P4_FSB_ESCR1 		0x3a3
+#define MSR_P4_IQ_ESCR0 		0x3ba
+#define MSR_P4_IQ_ESCR1 		0x3bb
+#define MSR_P4_IS_ESCR0 		0x3b4
+#define MSR_P4_IS_ESCR1 		0x3b5
+#define MSR_P4_ITLB_ESCR0 		0x3b6
+#define MSR_P4_ITLB_ESCR1 		0x3b7
+#define MSR_P4_IX_ESCR0 		0x3c8
+#define MSR_P4_IX_ESCR1 		0x3c9
+#define MSR_P4_MOB_ESCR0 		0x3aa
+#define MSR_P4_MOB_ESCR1 		0x3ab
+#define MSR_P4_MS_ESCR0 		0x3c0
+#define MSR_P4_MS_ESCR1 		0x3c1
+#define MSR_P4_PMH_ESCR0 		0x3ac
+#define MSR_P4_PMH_ESCR1 		0x3ad
+#define MSR_P4_RAT_ESCR0 		0x3bc
+#define MSR_P4_RAT_ESCR1 		0x3bd
+#define MSR_P4_SAAT_ESCR0 		0x3ae
+#define MSR_P4_SAAT_ESCR1 		0x3af
+#define MSR_P4_SSU_ESCR0 		0x3be
+#define MSR_P4_SSU_ESCR1 		0x3bf    /* guess: not defined in manual */
+#define MSR_P4_TBPU_ESCR0 		0x3c2
+#define MSR_P4_TBPU_ESCR1 		0x3c3
+#define MSR_P4_TC_ESCR0 		0x3c4
+#define MSR_P4_TC_ESCR1 		0x3c5
+#define MSR_P4_U2L_ESCR0 		0x3b0
+#define MSR_P4_U2L_ESCR1 		0x3b1
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/nmi.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/nmi.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/nmi.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/nmi.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,93 @@
+/*
+ *  linux/include/asm-i386/nmi.h
+ */
+#ifndef ASM_NMI_H
+#define ASM_NMI_H
+
+#include <linux/pm.h>
+#include <asm/io.h>
+
+#include <xen/interface/nmi.h>
+
+struct pt_regs;
+
+typedef int (*nmi_callback_t)(struct pt_regs * regs, int cpu);
+
+/**
+ * set_nmi_callback
+ *
+ * Set a handler for an NMI. Only one handler may be
+ * set. Return 1 if the NMI was handled.
+ */
+void set_nmi_callback(nmi_callback_t callback);
+
+/**
+ * unset_nmi_callback
+ *
+ * Remove the handler previously set.
+ */
+void unset_nmi_callback(void);
+
+#ifdef CONFIG_PM
+ 
+/** Replace the PM callback routine for NMI. */
+struct pm_dev * set_nmi_pm_callback(pm_callback callback);
+
+/** Unset the PM callback routine back to the default. */
+void unset_nmi_pm_callback(struct pm_dev * dev);
+
+#else
+
+static inline struct pm_dev * set_nmi_pm_callback(pm_callback callback)
+{
+	return 0;
+} 
+ 
+static inline void unset_nmi_pm_callback(struct pm_dev * dev)
+{
+}
+
+#endif /* CONFIG_PM */
+ 
+extern void default_do_nmi(struct pt_regs *);
+extern void die_nmi(char *str, struct pt_regs *regs);
+
+static inline unsigned char get_nmi_reason(void)
+{
+        shared_info_t *s = HYPERVISOR_shared_info;
+        unsigned char reason = 0;
+
+        /* construct a value which looks like it came from
+         * port 0x61.
+         */
+        if (test_bit(_XEN_NMIREASON_io_error, &s->arch.nmi_reason))
+                reason |= 0x40;
+        if (test_bit(_XEN_NMIREASON_parity_error, &s->arch.nmi_reason))
+                reason |= 0x80;
+
+        return reason;
+}
+
+extern int panic_on_timeout;
+extern int unknown_nmi_panic;
+
+extern int check_nmi_watchdog(void);
+ 
+extern void setup_apic_nmi_watchdog (void);
+extern int reserve_lapic_nmi(void);
+extern void release_lapic_nmi(void);
+extern void disable_timer_nmi_watchdog(void);
+extern void enable_timer_nmi_watchdog(void);
+extern void nmi_watchdog_tick (struct pt_regs * regs, unsigned reason);
+
+extern void nmi_watchdog_default(void);
+extern int setup_nmi_watchdog(char *);
+
+extern unsigned int nmi_watchdog;
+#define NMI_DEFAULT	-1
+#define NMI_NONE	0
+#define NMI_IO_APIC	1
+#define NMI_LOCAL_APIC	2
+#define NMI_INVALID	3
+
+#endif /* ASM_NMI_H */
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/page.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/page.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/page.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/page.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,212 @@
+#ifndef _X86_64_PAGE_H
+#define _X86_64_PAGE_H
+
+/* #include <linux/string.h> */
+#ifndef __ASSEMBLY__
+#include <linux/kernel.h>
+#include <linux/types.h>
+#include <asm/bug.h>
+#endif
+#include <xen/interface/xen.h> 
+
+/*
+ * Need to repeat this here in order to not include pgtable.h (which in turn
+ * depends on definitions made here), but to be able to use the symbolic
+ * below. The preprocessor will warn if the two definitions aren't identical.
+ */
+#define _PAGE_PRESENT	0x001
+#define _PAGE_IO	0x200
+
+/* PAGE_SHIFT determines the page size */
+#define PAGE_SHIFT	12
+#ifdef __ASSEMBLY__
+#define PAGE_SIZE	(0x1 << PAGE_SHIFT)
+#else
+#define PAGE_SIZE	(1UL << PAGE_SHIFT)
+#endif
+#define PAGE_MASK	(~(PAGE_SIZE-1))
+
+/* See Documentation/x86_64/mm.txt for a description of the memory map. */
+#define __PHYSICAL_MASK_SHIFT	46
+#define __PHYSICAL_MASK		((1UL << __PHYSICAL_MASK_SHIFT) - 1)
+#define __VIRTUAL_MASK_SHIFT	48
+#define __VIRTUAL_MASK		((1UL << __VIRTUAL_MASK_SHIFT) - 1)
+
+#define PHYSICAL_PAGE_MASK	(~(PAGE_SIZE-1) & __PHYSICAL_MASK)
+
+#define THREAD_ORDER 1 
+#define THREAD_SIZE  (PAGE_SIZE << THREAD_ORDER)
+#define CURRENT_MASK (~(THREAD_SIZE-1))
+
+#define EXCEPTION_STACK_ORDER 0
+#define EXCEPTION_STKSZ (PAGE_SIZE << EXCEPTION_STACK_ORDER)
+
+#define DEBUG_STACK_ORDER (EXCEPTION_STACK_ORDER + 1)
+#define DEBUG_STKSZ (PAGE_SIZE << DEBUG_STACK_ORDER)
+
+#define IRQSTACK_ORDER 2
+#define IRQSTACKSIZE (PAGE_SIZE << IRQSTACK_ORDER)
+
+#define STACKFAULT_STACK 1
+#define DOUBLEFAULT_STACK 2
+#define NMI_STACK 3
+#define DEBUG_STACK 4
+#define MCE_STACK 5
+#define N_EXCEPTION_STACKS 5  /* hw limit: 7 */
+
+#define LARGE_PAGE_MASK (~(LARGE_PAGE_SIZE-1))
+#define LARGE_PAGE_SIZE (1UL << PMD_SHIFT)
+
+#define HPAGE_SHIFT PMD_SHIFT
+#define HPAGE_SIZE	((1UL) << HPAGE_SHIFT)
+#define HPAGE_MASK	(~(HPAGE_SIZE - 1))
+#define HUGETLB_PAGE_ORDER	(HPAGE_SHIFT - PAGE_SHIFT)
+
+#ifdef __KERNEL__
+#ifndef __ASSEMBLY__
+
+extern unsigned long end_pfn;
+
+#include <asm/maddr.h>
+
+void clear_page(void *);
+void copy_page(void *, void *);
+
+#define clear_user_page(page, vaddr, pg)	clear_page(page)
+#define copy_user_page(to, from, vaddr, pg)	copy_page(to, from)
+
+#define alloc_zeroed_user_highpage(vma, vaddr) alloc_page_vma(GFP_HIGHUSER | __GFP_ZERO, vma, vaddr)
+#define __HAVE_ARCH_ALLOC_ZEROED_USER_HIGHPAGE
+
+/*
+ * These are used to make use of C type-checking..
+ */
+typedef struct { unsigned long pte; } pte_t;
+typedef struct { unsigned long pmd; } pmd_t;
+typedef struct { unsigned long pud; } pud_t;
+typedef struct { unsigned long pgd; } pgd_t;
+#define PTE_MASK	PHYSICAL_PAGE_MASK
+
+typedef struct { unsigned long pgprot; } pgprot_t;
+
+#define __pte_val(x) ((x).pte)
+#define pte_val(x) ((__pte_val(x) & (_PAGE_PRESENT|_PAGE_IO))	\
+		    == _PAGE_PRESENT ?				\
+		    pte_machine_to_phys(__pte_val(x)) :		\
+		    __pte_val(x))
+
+#define __pmd_val(x) ((x).pmd)
+static inline unsigned long pmd_val(pmd_t x)
+{
+	unsigned long ret = __pmd_val(x);
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (ret) ret = pte_machine_to_phys(ret) | _PAGE_PRESENT;
+#else
+	if (ret & _PAGE_PRESENT) ret = pte_machine_to_phys(ret);
+#endif
+	return ret;
+}
+
+#define __pud_val(x) ((x).pud)
+static inline unsigned long pud_val(pud_t x)
+{
+	unsigned long ret = __pud_val(x);
+	if (ret & _PAGE_PRESENT) ret = pte_machine_to_phys(ret);
+	return ret;
+}
+
+#define __pgd_val(x) ((x).pgd)
+static inline unsigned long pgd_val(pgd_t x)
+{
+	unsigned long ret = __pgd_val(x);
+	if (ret & _PAGE_PRESENT) ret = pte_machine_to_phys(ret);
+	return ret;
+}
+
+#define pgprot_val(x)	((x).pgprot)
+
+static inline pte_t __pte(unsigned long x)
+{
+	if ((x & (_PAGE_PRESENT|_PAGE_IO)) == _PAGE_PRESENT)
+		x = pte_phys_to_machine(x);
+	return ((pte_t) { (x) });
+}
+
+static inline pmd_t __pmd(unsigned long x)
+{
+	if (x & _PAGE_PRESENT) x = pte_phys_to_machine(x);
+	return ((pmd_t) { (x) });
+}
+
+static inline pud_t __pud(unsigned long x)
+{
+	if (x & _PAGE_PRESENT) x = pte_phys_to_machine(x);
+	return ((pud_t) { (x) });
+}
+
+static inline pgd_t __pgd(unsigned long x)
+{
+	if (x & _PAGE_PRESENT) x = pte_phys_to_machine(x);
+	return ((pgd_t) { (x) });
+}
+
+#define __pgprot(x)	((pgprot_t) { (x) } )
+
+#define __PHYSICAL_START	((unsigned long)CONFIG_PHYSICAL_START)
+#define __START_KERNEL		(__START_KERNEL_map + __PHYSICAL_START)
+#define __START_KERNEL_map	0xffffffff80000000UL
+#define __PAGE_OFFSET           0xffff880000000000UL	
+
+#else
+#define __PHYSICAL_START	CONFIG_PHYSICAL_START
+#define __START_KERNEL		(__START_KERNEL_map + __PHYSICAL_START)
+#define __START_KERNEL_map	0xffffffff80000000
+#define __PAGE_OFFSET           0xffff880000000000
+#endif /* !__ASSEMBLY__ */
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+#undef LOAD_OFFSET
+#define LOAD_OFFSET		0
+#endif
+
+/* to align the pointer to the (next) page boundary */
+#define PAGE_ALIGN(addr)	(((addr)+PAGE_SIZE-1)&PAGE_MASK)
+
+#define KERNEL_TEXT_SIZE  (40UL*1024*1024)
+#define KERNEL_TEXT_START 0xffffffff80000000UL 
+
+#define PAGE_OFFSET		((unsigned long)__PAGE_OFFSET)
+
+/* Note: __pa(&symbol_visible_to_c) should be always replaced with __pa_symbol.
+   Otherwise you risk miscompilation. */ 
+#define __pa(x)			(((unsigned long)(x)>=__START_KERNEL_map)?(unsigned long)(x) - (unsigned long)__START_KERNEL_map:(unsigned long)(x) - PAGE_OFFSET)
+/* __pa_symbol should be used for C visible symbols.
+   This seems to be the official gcc blessed way to do such arithmetic. */ 
+#define __pa_symbol(x)		\
+	({unsigned long v;  \
+	  asm("" : "=r" (v) : "0" (x)); \
+	  __pa(v); })
+
+#define __va(x)			((void *)((unsigned long)(x)+PAGE_OFFSET))
+#define __boot_va(x)		__va(x)
+#define __boot_pa(x)		__pa(x)
+#ifdef CONFIG_FLATMEM
+#define pfn_valid(pfn)		((pfn) < end_pfn)
+#endif
+
+#define virt_to_page(kaddr)	pfn_to_page(__pa(kaddr) >> PAGE_SHIFT)
+#define virt_addr_valid(kaddr)	pfn_valid(__pa(kaddr) >> PAGE_SHIFT)
+#define pfn_to_kaddr(pfn)      __va((pfn) << PAGE_SHIFT)
+
+#define VM_DATA_DEFAULT_FLAGS \
+	(((current->personality & READ_IMPLIES_EXEC) ? VM_EXEC : 0 ) | \
+	 VM_READ | VM_WRITE | VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC)
+
+#define __HAVE_ARCH_GATE_AREA 1	
+
+#include <asm-generic/memory_model.h>
+#include <asm-generic/page.h>
+
+#endif /* __KERNEL__ */
+
+#endif /* _X86_64_PAGE_H */
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/pci.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/pci.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/pci.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/pci.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,168 @@
+#ifndef __x8664_PCI_H
+#define __x8664_PCI_H
+
+#include <asm/io.h>
+
+#ifdef __KERNEL__
+
+#include <linux/mm.h> /* for struct page */
+
+/* Can be used to override the logic in pci_scan_bus for skipping
+   already-configured bus numbers - to be used for buggy BIOSes
+   or architectures with incomplete PCI setup by the loader */
+
+#ifdef CONFIG_PCI
+extern unsigned int pcibios_assign_all_busses(void);
+#else
+#define pcibios_assign_all_busses()	0
+#endif
+
+#include <asm/hypervisor.h>
+#define pcibios_scan_all_fns(a, b)	(!is_initial_xendomain())
+
+extern unsigned long pci_mem_start;
+#define PCIBIOS_MIN_IO		0x1000
+#define PCIBIOS_MIN_MEM		(pci_mem_start)
+
+#define PCIBIOS_MIN_CARDBUS_IO	0x4000
+
+void pcibios_config_init(void);
+struct pci_bus * pcibios_scan_root(int bus);
+extern int (*pci_config_read)(int seg, int bus, int dev, int fn, int reg, int len, u32 *value);
+extern int (*pci_config_write)(int seg, int bus, int dev, int fn, int reg, int len, u32 value);
+
+void pcibios_set_master(struct pci_dev *dev);
+void pcibios_penalize_isa_irq(int irq, int active);
+struct irq_routing_table *pcibios_get_irq_routing_table(void);
+int pcibios_set_irq_routing(struct pci_dev *dev, int pin, int irq);
+
+#include <linux/types.h>
+#include <linux/slab.h>
+#include <asm/scatterlist.h>
+#include <linux/string.h>
+#include <asm/page.h>
+
+extern void pci_iommu_alloc(void);
+extern int iommu_setup(char *opt);
+
+/* The PCI address space does equal the physical memory
+ * address space.  The networking and block device layers use
+ * this boolean for bounce buffer decisions
+ *
+ * On AMD64 it mostly equals, but we set it to zero if a hardware
+ * IOMMU (gart) of sotware IOMMU (swiotlb) is available.
+ */
+#define PCI_DMA_BUS_IS_PHYS (dma_ops->is_phys)
+
+#if defined(CONFIG_IOMMU) || defined(CONFIG_CALGARY_IOMMU)
+
+/*
+ * x86-64 always supports DAC, but sometimes it is useful to force
+ * devices through the IOMMU to get automatic sg list merging.
+ * Optional right now.
+ */
+extern int iommu_sac_force;
+#define pci_dac_dma_supported(pci_dev, mask)	(!iommu_sac_force)
+
+#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)	\
+	dma_addr_t ADDR_NAME;
+#define DECLARE_PCI_UNMAP_LEN(LEN_NAME)		\
+	__u32 LEN_NAME;
+#define pci_unmap_addr(PTR, ADDR_NAME)			\
+	((PTR)->ADDR_NAME)
+#define pci_unmap_addr_set(PTR, ADDR_NAME, VAL)		\
+	(((PTR)->ADDR_NAME) = (VAL))
+#define pci_unmap_len(PTR, LEN_NAME)			\
+	((PTR)->LEN_NAME)
+#define pci_unmap_len_set(PTR, LEN_NAME, VAL)		\
+	(((PTR)->LEN_NAME) = (VAL))
+
+#elif defined(CONFIG_SWIOTLB)
+
+#define pci_dac_dma_supported(pci_dev, mask)    1
+
+#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)	\
+	dma_addr_t ADDR_NAME;
+#define DECLARE_PCI_UNMAP_LEN(LEN_NAME)		\
+	__u32 LEN_NAME;
+#define pci_unmap_addr(PTR, ADDR_NAME)			\
+	((PTR)->ADDR_NAME)
+#define pci_unmap_addr_set(PTR, ADDR_NAME, VAL)		\
+	(((PTR)->ADDR_NAME) = (VAL))
+#define pci_unmap_len(PTR, LEN_NAME)			\
+	((PTR)->LEN_NAME)
+#define pci_unmap_len_set(PTR, LEN_NAME, VAL)		\
+	(((PTR)->LEN_NAME) = (VAL))
+
+#else
+/* No IOMMU */
+
+#define pci_dac_dma_supported(pci_dev, mask)    1
+
+#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)
+#define DECLARE_PCI_UNMAP_LEN(LEN_NAME)
+#define pci_unmap_addr(PTR, ADDR_NAME)		(0)
+#define pci_unmap_addr_set(PTR, ADDR_NAME, VAL)	do { } while (0)
+#define pci_unmap_len(PTR, LEN_NAME)		(0)
+#define pci_unmap_len_set(PTR, LEN_NAME, VAL)	do { } while (0)
+
+#endif
+
+#include <asm-generic/pci-dma-compat.h>
+
+static inline dma64_addr_t
+pci_dac_page_to_dma(struct pci_dev *pdev, struct page *page, unsigned long offset, int direction)
+{
+	return ((dma64_addr_t) page_to_phys(page) +
+		(dma64_addr_t) offset);
+}
+
+static inline struct page *
+pci_dac_dma_to_page(struct pci_dev *pdev, dma64_addr_t dma_addr)
+{
+	return virt_to_page(__va(dma_addr)); 	
+}
+
+static inline unsigned long
+pci_dac_dma_to_offset(struct pci_dev *pdev, dma64_addr_t dma_addr)
+{
+	return (dma_addr & ~PAGE_MASK);
+}
+
+static inline void
+pci_dac_dma_sync_single_for_cpu(struct pci_dev *pdev, dma64_addr_t dma_addr, size_t len, int direction)
+{
+}
+
+static inline void
+pci_dac_dma_sync_single_for_device(struct pci_dev *pdev, dma64_addr_t dma_addr, size_t len, int direction)
+{
+	flush_write_buffers();
+}
+
+#ifdef CONFIG_PCI
+static inline void pci_dma_burst_advice(struct pci_dev *pdev,
+					enum pci_dma_burst_strategy *strat,
+					unsigned long *strategy_parameter)
+{
+	*strat = PCI_DMA_BURST_INFINITY;
+	*strategy_parameter = ~0UL;
+}
+#endif
+
+#define HAVE_PCI_MMAP
+extern int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
+			       enum pci_mmap_state mmap_state, int write_combine);
+
+static inline void pcibios_add_platform_entries(struct pci_dev *dev)
+{
+}
+
+#endif /* __KERNEL__ */
+
+/* generic pci stuff */
+#ifdef CONFIG_PCI
+#include <asm-generic/pci.h>
+#endif
+
+#endif /* __x8664_PCI_H */
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/pgalloc.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/pgalloc.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/pgalloc.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/pgalloc.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,204 @@
+#ifndef _X86_64_PGALLOC_H
+#define _X86_64_PGALLOC_H
+
+#include <asm/fixmap.h>
+#include <asm/pda.h>
+#include <linux/threads.h>
+#include <linux/mm.h>
+#include <asm/io.h>		/* for phys_to_virt and page_to_pseudophys */
+
+#include <xen/features.h>
+void make_page_readonly(void *va, unsigned int feature);
+void make_page_writable(void *va, unsigned int feature);
+void make_pages_readonly(void *va, unsigned int nr, unsigned int feature);
+void make_pages_writable(void *va, unsigned int nr, unsigned int feature);
+
+#define __user_pgd(pgd) ((pgd) + PTRS_PER_PGD)
+
+static inline void pmd_populate_kernel(struct mm_struct *mm, pmd_t *pmd, pte_t *pte)
+{
+	set_pmd(pmd, __pmd(_PAGE_TABLE | __pa(pte)));
+}
+
+static inline void pmd_populate(struct mm_struct *mm, pmd_t *pmd, struct page *pte)
+{
+	if (unlikely((mm)->context.pinned)) {
+		BUG_ON(HYPERVISOR_update_va_mapping(
+			       (unsigned long)__va(page_to_pfn(pte) << PAGE_SHIFT),
+			       pfn_pte(page_to_pfn(pte), PAGE_KERNEL_RO), 0));
+		set_pmd(pmd, __pmd(_PAGE_TABLE | (page_to_pfn(pte) << PAGE_SHIFT)));
+	} else {
+		*(pmd) = __pmd(_PAGE_TABLE | (page_to_pfn(pte) << PAGE_SHIFT));
+	}
+}
+
+static inline void pud_populate(struct mm_struct *mm, pud_t *pud, pmd_t *pmd)
+{
+	if (unlikely((mm)->context.pinned)) {
+		BUG_ON(HYPERVISOR_update_va_mapping(
+			       (unsigned long)pmd,
+			       pfn_pte(virt_to_phys(pmd)>>PAGE_SHIFT, 
+				       PAGE_KERNEL_RO), 0));
+		set_pud(pud, __pud(_PAGE_TABLE | __pa(pmd)));
+	} else {
+		*(pud) =  __pud(_PAGE_TABLE | __pa(pmd));
+	}
+}
+
+/*
+ * We need to use the batch mode here, but pgd_pupulate() won't be
+ * be called frequently.
+ */
+static inline void pgd_populate(struct mm_struct *mm, pgd_t *pgd, pud_t *pud)
+{
+	if (unlikely((mm)->context.pinned)) {
+		BUG_ON(HYPERVISOR_update_va_mapping(
+			       (unsigned long)pud,
+			       pfn_pte(virt_to_phys(pud)>>PAGE_SHIFT, 
+				       PAGE_KERNEL_RO), 0));
+		set_pgd(pgd, __pgd(_PAGE_TABLE | __pa(pud)));
+		set_pgd(__user_pgd(pgd), __pgd(_PAGE_TABLE | __pa(pud)));
+	} else {
+		*(pgd) =  __pgd(_PAGE_TABLE | __pa(pud));
+		*(__user_pgd(pgd)) = *(pgd);
+	}
+}
+
+extern struct page *pte_alloc_one(struct mm_struct *mm, unsigned long addr);
+extern void pte_free(struct page *pte);
+
+static inline pmd_t *pmd_alloc_one(struct mm_struct *mm, unsigned long addr)
+{
+	struct page *pg;
+
+	pg = pte_alloc_one(mm, addr);
+	return pg ? page_address(pg) : NULL;
+}
+
+static inline void pmd_free(pmd_t *pmd)
+{
+	BUG_ON((unsigned long)pmd & (PAGE_SIZE-1));
+	pte_free(virt_to_page(pmd));
+}
+
+static inline pud_t *pud_alloc_one(struct mm_struct *mm, unsigned long addr)
+{
+	struct page *pg;
+
+	pg = pte_alloc_one(mm, addr);
+	return pg ? page_address(pg) : NULL;
+}
+
+static inline void pud_free(pud_t *pud)
+{
+	BUG_ON((unsigned long)pud & (PAGE_SIZE-1));
+	pte_free(virt_to_page(pud));
+}
+
+static inline void pgd_list_add(pgd_t *pgd)
+{
+	struct page *page = virt_to_page(pgd);
+
+	spin_lock(&pgd_lock);
+	page->index = (pgoff_t)pgd_list;
+	if (pgd_list)
+		pgd_list->private = (unsigned long)&page->index;
+	pgd_list = page;
+	page->private = (unsigned long)&pgd_list;
+	spin_unlock(&pgd_lock);
+}
+
+static inline void pgd_list_del(pgd_t *pgd)
+{
+	struct page *next, **pprev, *page = virt_to_page(pgd);
+
+	spin_lock(&pgd_lock);
+	next = (struct page *)page->index;
+	pprev = (struct page **)page->private;
+	*pprev = next;
+	if (next)
+		next->private = (unsigned long)pprev;
+	spin_unlock(&pgd_lock);
+}
+
+static inline pgd_t *pgd_alloc(struct mm_struct *mm)
+{
+	/*
+	 * We allocate two contiguous pages for kernel and user.
+	 */
+	unsigned boundary;
+	pgd_t *pgd = (pgd_t *)__get_free_pages(GFP_KERNEL|__GFP_REPEAT, 1);
+	if (!pgd)
+		return NULL;
+	pgd_list_add(pgd);
+	/*
+	 * Copy kernel pointers in from init.
+	 * Could keep a freelist or slab cache of those because the kernel
+	 * part never changes.
+	 */
+	boundary = pgd_index(__PAGE_OFFSET);
+	memset(pgd, 0, boundary * sizeof(pgd_t));
+	memcpy(pgd + boundary,
+	       init_level4_pgt + boundary,
+	       (PTRS_PER_PGD - boundary) * sizeof(pgd_t));
+
+	memset(__user_pgd(pgd), 0, PAGE_SIZE); /* clean up user pgd */
+	/*
+	 * Set level3_user_pgt for vsyscall area
+	 */
+	__user_pgd(pgd)[pgd_index(VSYSCALL_START)] =
+		__pgd(__pa_symbol(level3_user_pgt) | _PAGE_TABLE);
+	return pgd;
+}
+
+static inline void pgd_free(pgd_t *pgd)
+{
+	pte_t *ptep = virt_to_ptep(pgd);
+
+	if (!pte_write(*ptep)) {
+		xen_pgd_unpin(__pa(pgd));
+		BUG_ON(HYPERVISOR_update_va_mapping(
+			       (unsigned long)pgd,
+			       pfn_pte(virt_to_phys(pgd)>>PAGE_SHIFT, PAGE_KERNEL),
+			       0));
+	}
+
+	ptep = virt_to_ptep(__user_pgd(pgd));
+
+	if (!pte_write(*ptep)) {
+		xen_pgd_unpin(__pa(__user_pgd(pgd)));
+		BUG_ON(HYPERVISOR_update_va_mapping(
+			       (unsigned long)__user_pgd(pgd),
+			       pfn_pte(virt_to_phys(__user_pgd(pgd))>>PAGE_SHIFT, 
+				       PAGE_KERNEL),
+			       0));
+	}
+
+	pgd_list_del(pgd);
+	free_pages((unsigned long)pgd, 1);
+}
+
+static inline pte_t *pte_alloc_one_kernel(struct mm_struct *mm, unsigned long address)
+{
+	pte_t *pte = (pte_t *)get_zeroed_page(GFP_KERNEL|__GFP_REPEAT);
+	if (pte)
+		make_page_readonly(pte, XENFEAT_writable_page_tables);
+
+	return pte;
+}
+
+/* Should really implement gc for free page table pages. This could be
+   done with a reference count in struct page. */
+
+static inline void pte_free_kernel(pte_t *pte)
+{
+	BUG_ON((unsigned long)pte & (PAGE_SIZE-1));
+	make_page_writable(pte, XENFEAT_writable_page_tables);
+	free_page((unsigned long)pte); 
+}
+
+#define __pte_free_tlb(tlb,pte) tlb_remove_page((tlb),(pte))
+#define __pmd_free_tlb(tlb,x)   tlb_remove_page((tlb),virt_to_page(x))
+#define __pud_free_tlb(tlb,x)   tlb_remove_page((tlb),virt_to_page(x))
+
+#endif /* _X86_64_PGALLOC_H */
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/pgtable.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/pgtable.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/pgtable.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/pgtable.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,583 @@
+#ifndef _X86_64_PGTABLE_H
+#define _X86_64_PGTABLE_H
+
+/*
+ * This file contains the functions and defines necessary to modify and use
+ * the x86-64 page table tree.
+ */
+#include <asm/processor.h>
+#include <asm/fixmap.h>
+#include <asm/bitops.h>
+#include <linux/threads.h>
+#include <linux/sched.h>
+#include <asm/pda.h>
+#ifdef CONFIG_XEN
+#include <asm/hypervisor.h>
+
+extern pud_t level3_user_pgt[512];
+
+extern void xen_init_pt(void);
+
+extern pte_t *lookup_address(unsigned long address);
+
+#define virt_to_ptep(va)						\
+({									\
+	pte_t *__ptep = lookup_address((unsigned long)(va));		\
+	BUG_ON(!__ptep || !pte_present(*__ptep));			\
+	__ptep;								\
+})
+
+#define arbitrary_virt_to_machine(va)					\
+	(((maddr_t)pte_mfn(*virt_to_ptep(va)) << PAGE_SHIFT)		\
+	 | ((unsigned long)(va) & (PAGE_SIZE - 1)))
+#endif
+
+extern pud_t level3_kernel_pgt[512];
+extern pud_t level3_physmem_pgt[512];
+extern pud_t level3_ident_pgt[512];
+extern pmd_t level2_kernel_pgt[512];
+extern pgd_t init_level4_pgt[];
+extern pgd_t boot_level4_pgt[];
+extern unsigned long __supported_pte_mask;
+
+#define swapper_pg_dir init_level4_pgt
+
+extern int nonx_setup(char *str);
+extern void paging_init(void);
+extern void clear_kernel_mapping(unsigned long addr, unsigned long size);
+
+extern unsigned long pgkern_mask;
+
+/*
+ * ZERO_PAGE is a global shared page that is always zero: used
+ * for zero-mapped memory areas etc..
+ */
+extern unsigned long empty_zero_page[PAGE_SIZE/sizeof(unsigned long)];
+#define ZERO_PAGE(vaddr) (virt_to_page(empty_zero_page))
+
+/*
+ * PGDIR_SHIFT determines what a top-level page table entry can map
+ */
+#define PGDIR_SHIFT	39
+#define PTRS_PER_PGD	512
+
+/*
+ * 3rd level page
+ */
+#define PUD_SHIFT	30
+#define PTRS_PER_PUD	512
+
+/*
+ * PMD_SHIFT determines the size of the area a middle-level
+ * page table can map
+ */
+#define PMD_SHIFT	21
+#define PTRS_PER_PMD	512
+
+/*
+ * entries per page directory level
+ */
+#define PTRS_PER_PTE	512
+
+#define pte_ERROR(e) \
+	printk("%s:%d: bad pte %p(%016lx pfn %010lx).\n", __FILE__, __LINE__, \
+	       &(e), __pte_val(e), pte_pfn(e))
+#define pmd_ERROR(e) \
+	printk("%s:%d: bad pmd %p(%016lx pfn %010lx).\n", __FILE__, __LINE__, \
+	       &(e), __pmd_val(e), pmd_pfn(e))
+#define pud_ERROR(e) \
+	printk("%s:%d: bad pud %p(%016lx pfn %010lx).\n", __FILE__, __LINE__, \
+	       &(e), __pud_val(e), (pud_val(e) & __PHYSICAL_MASK) >> PAGE_SHIFT)
+#define pgd_ERROR(e) \
+	printk("%s:%d: bad pgd %p(%016lx pfn %010lx).\n", __FILE__, __LINE__, \
+	       &(e), __pgd_val(e), (pgd_val(e) & __PHYSICAL_MASK) >> PAGE_SHIFT)
+
+#define pgd_none(x)	(!__pgd_val(x))
+#define pud_none(x)	(!__pud_val(x))
+
+static inline void set_pte(pte_t *dst, pte_t val)
+{
+	*dst = val;
+}
+
+#define set_pmd(pmdptr, pmdval) xen_l2_entry_update(pmdptr, (pmdval))
+#define set_pud(pudptr, pudval) xen_l3_entry_update(pudptr, (pudval))
+#define set_pgd(pgdptr, pgdval) xen_l4_entry_update(pgdptr, (pgdval))
+
+static inline void pud_clear (pud_t * pud)
+{
+	set_pud(pud, __pud(0));
+}
+
+#define __user_pgd(pgd) ((pgd) + PTRS_PER_PGD)
+
+static inline void pgd_clear (pgd_t * pgd)
+{
+        set_pgd(pgd, __pgd(0));
+        set_pgd(__user_pgd(pgd), __pgd(0));
+}
+
+#define pud_page(pud) \
+    ((unsigned long) __va(pud_val(pud) & PHYSICAL_PAGE_MASK))
+
+#define pte_same(a, b)		((a).pte == (b).pte)
+
+#define pte_pgprot(a)	(__pgprot((a).pte & ~PHYSICAL_PAGE_MASK))
+
+#define PMD_SIZE	(1UL << PMD_SHIFT)
+#define PMD_MASK	(~(PMD_SIZE-1))
+#define PUD_SIZE	(1UL << PUD_SHIFT)
+#define PUD_MASK	(~(PUD_SIZE-1))
+#define PGDIR_SIZE	(1UL << PGDIR_SHIFT)
+#define PGDIR_MASK	(~(PGDIR_SIZE-1))
+
+#define USER_PTRS_PER_PGD	((TASK_SIZE-1)/PGDIR_SIZE+1)
+#define FIRST_USER_ADDRESS	0
+
+#ifndef __ASSEMBLY__
+#define MAXMEM		 0x3fffffffffffUL
+#define VMALLOC_START    0xffffc20000000000UL
+#define VMALLOC_END      0xffffe1ffffffffffUL
+#define MODULES_VADDR    0xffffffff88000000UL
+#define MODULES_END      0xfffffffffff00000UL
+#define MODULES_LEN   (MODULES_END - MODULES_VADDR)
+
+#define _PAGE_BIT_PRESENT	0
+#define _PAGE_BIT_RW		1
+#define _PAGE_BIT_USER		2
+#define _PAGE_BIT_PWT		3
+#define _PAGE_BIT_PCD		4
+#define _PAGE_BIT_ACCESSED	5
+#define _PAGE_BIT_DIRTY		6
+#define _PAGE_BIT_PSE		7	/* 4 MB (or 2MB) page */
+#define _PAGE_BIT_GLOBAL	8	/* Global TLB entry PPro+ */
+#define _PAGE_BIT_NX           63       /* No execute: only valid after cpuid check */
+
+#define _PAGE_PRESENT	0x001
+#define _PAGE_RW	0x002
+#define _PAGE_USER	0x004
+#define _PAGE_PWT	0x008
+#define _PAGE_PCD	0x010
+#define _PAGE_ACCESSED	0x020
+#define _PAGE_DIRTY	0x040
+#define _PAGE_PSE	0x080	/* 2MB page */
+#define _PAGE_FILE	0x040	/* nonlinear file mapping, saved PTE; unset:swap */
+#define _PAGE_GLOBAL	0x100	/* Global TLB entry */
+
+#define _PAGE_PROTNONE	0x080	/* If not present */
+#define _PAGE_NX        (1UL<<_PAGE_BIT_NX)
+
+/* Mapped page is I/O or foreign and has no associated page struct. */
+#define _PAGE_IO	0x200
+
+#if CONFIG_XEN_COMPAT <= 0x030002
+extern unsigned int __kernel_page_user;
+#else
+#define __kernel_page_user 0
+#endif
+
+#define _PAGE_TABLE	(_PAGE_PRESENT | _PAGE_RW | _PAGE_USER | _PAGE_ACCESSED | _PAGE_DIRTY)
+#define _KERNPG_TABLE	(_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED | _PAGE_DIRTY | __kernel_page_user)
+
+#define _PAGE_CHG_MASK	(PTE_MASK | _PAGE_ACCESSED | _PAGE_DIRTY | _PAGE_IO)
+
+#define PAGE_NONE	__pgprot(_PAGE_PROTNONE | _PAGE_ACCESSED)
+#define PAGE_SHARED	__pgprot(_PAGE_PRESENT | _PAGE_RW | _PAGE_USER | _PAGE_ACCESSED | _PAGE_NX)
+#define PAGE_SHARED_EXEC __pgprot(_PAGE_PRESENT | _PAGE_RW | _PAGE_USER | _PAGE_ACCESSED)
+#define PAGE_COPY_NOEXEC __pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED | _PAGE_NX)
+#define PAGE_COPY PAGE_COPY_NOEXEC
+#define PAGE_COPY_EXEC __pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED)
+#define PAGE_READONLY	__pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED | _PAGE_NX)
+#define PAGE_READONLY_EXEC __pgprot(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED)
+#define __PAGE_KERNEL \
+	(_PAGE_PRESENT | _PAGE_RW | _PAGE_DIRTY | _PAGE_ACCESSED | _PAGE_NX | __kernel_page_user)
+#define __PAGE_KERNEL_EXEC \
+	(_PAGE_PRESENT | _PAGE_RW | _PAGE_DIRTY | _PAGE_ACCESSED | __kernel_page_user)
+#define __PAGE_KERNEL_NOCACHE \
+	(_PAGE_PRESENT | _PAGE_RW | _PAGE_DIRTY | _PAGE_PCD | _PAGE_ACCESSED | _PAGE_NX | __kernel_page_user)
+#define __PAGE_KERNEL_RO \
+	(_PAGE_PRESENT | _PAGE_DIRTY | _PAGE_ACCESSED | _PAGE_NX | __kernel_page_user)
+#define __PAGE_KERNEL_VSYSCALL \
+	(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED)
+#define __PAGE_KERNEL_VSYSCALL_NOCACHE \
+	(_PAGE_PRESENT | _PAGE_USER | _PAGE_ACCESSED | _PAGE_PCD)
+#define __PAGE_KERNEL_LARGE \
+	(__PAGE_KERNEL | _PAGE_PSE)
+#define __PAGE_KERNEL_LARGE_EXEC \
+	(__PAGE_KERNEL_EXEC | _PAGE_PSE)
+
+/*
+ * We don't support GLOBAL page in xenolinux64
+ */
+#define MAKE_GLOBAL(x) __pgprot((x))
+
+#define PAGE_KERNEL MAKE_GLOBAL(__PAGE_KERNEL)
+#define PAGE_KERNEL_EXEC MAKE_GLOBAL(__PAGE_KERNEL_EXEC)
+#define PAGE_KERNEL_RO MAKE_GLOBAL(__PAGE_KERNEL_RO)
+#define PAGE_KERNEL_NOCACHE MAKE_GLOBAL(__PAGE_KERNEL_NOCACHE)
+#define PAGE_KERNEL_VSYSCALL32 __pgprot(__PAGE_KERNEL_VSYSCALL)
+#define PAGE_KERNEL_VSYSCALL MAKE_GLOBAL(__PAGE_KERNEL_VSYSCALL)
+#define PAGE_KERNEL_LARGE MAKE_GLOBAL(__PAGE_KERNEL_LARGE)
+#define PAGE_KERNEL_VSYSCALL_NOCACHE MAKE_GLOBAL(__PAGE_KERNEL_VSYSCALL_NOCACHE)
+
+/*         xwr */
+#define __P000	PAGE_NONE
+#define __P001	PAGE_READONLY
+#define __P010	PAGE_COPY
+#define __P011	PAGE_COPY
+#define __P100	PAGE_READONLY_EXEC
+#define __P101	PAGE_READONLY_EXEC
+#define __P110	PAGE_COPY_EXEC
+#define __P111	PAGE_COPY_EXEC
+
+#define __S000	PAGE_NONE
+#define __S001	PAGE_READONLY
+#define __S010	PAGE_SHARED
+#define __S011	PAGE_SHARED
+#define __S100	PAGE_READONLY_EXEC
+#define __S101	PAGE_READONLY_EXEC
+#define __S110	PAGE_SHARED_EXEC
+#define __S111	PAGE_SHARED_EXEC
+
+static inline unsigned long pgd_bad(pgd_t pgd)
+{
+       unsigned long val = __pgd_val(pgd);
+       val &= ~PTE_MASK;
+       val &= ~(_PAGE_USER | _PAGE_DIRTY);
+       return val & ~(_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED);
+}
+
+static inline unsigned long pud_bad(pud_t pud) 
+{ 
+       unsigned long val = __pud_val(pud);
+       val &= ~PTE_MASK; 
+       val &= ~(_PAGE_USER | _PAGE_DIRTY); 
+       return val & ~(_PAGE_PRESENT | _PAGE_RW | _PAGE_ACCESSED);      
+} 
+
+#define set_pte_at(_mm,addr,ptep,pteval) do {				\
+	if (((_mm) != current->mm && (_mm) != &init_mm) ||		\
+	    HYPERVISOR_update_va_mapping((addr), (pteval), 0))		\
+		set_pte((ptep), (pteval));				\
+} while (0)
+
+#define pte_none(x)	(!(x).pte)
+#define pte_present(x)	((x).pte & (_PAGE_PRESENT | _PAGE_PROTNONE))
+#define pte_clear(mm,addr,xp)	do { set_pte_at(mm, addr, xp, __pte(0)); } while (0)
+
+#define pages_to_mb(x) ((x) >> (20-PAGE_SHIFT))
+
+#define __pte_mfn(_pte) (((_pte).pte & PTE_MASK) >> PAGE_SHIFT)
+#define pte_mfn(_pte) ((_pte).pte & _PAGE_PRESENT ? \
+	__pte_mfn(_pte) : pfn_to_mfn(__pte_mfn(_pte)))
+#define pte_pfn(_pte) ((_pte).pte & _PAGE_IO ? end_pfn :	\
+		       (_pte).pte & _PAGE_PRESENT ?		\
+		       mfn_to_local_pfn(__pte_mfn(_pte)) :	\
+		       __pte_mfn(_pte))
+
+#define pte_page(x)	pfn_to_page(pte_pfn(x))
+
+static inline pte_t pfn_pte(unsigned long page_nr, pgprot_t pgprot)
+{
+	unsigned long pte = page_nr << PAGE_SHIFT;
+	pte |= pgprot_val(pgprot);
+	pte &= __supported_pte_mask;
+	return __pte(pte);
+}
+
+static inline pte_t ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
+{
+	pte_t pte = *ptep;
+	if (!pte_none(pte)) {
+		if ((mm != &init_mm) ||
+		    HYPERVISOR_update_va_mapping(addr, __pte(0), 0))
+			pte = __pte_ma(xchg(&ptep->pte, 0));
+	}
+	return pte;
+}
+
+static inline pte_t ptep_get_and_clear_full(struct mm_struct *mm, unsigned long addr, pte_t *ptep, int full)
+{
+	if (full) {
+		pte_t pte = *ptep;
+		if (mm->context.pinned)
+			xen_l1_entry_update(ptep, __pte(0));
+		else
+			*ptep = __pte(0);
+		return pte;
+	}
+	return ptep_get_and_clear(mm, addr, ptep);
+}
+
+#define ptep_clear_flush(vma, addr, ptep)			\
+({								\
+	pte_t *__ptep = (ptep);					\
+	pte_t __res = *__ptep;					\
+	if (!pte_none(__res) &&					\
+	    ((vma)->vm_mm != current->mm ||			\
+	     HYPERVISOR_update_va_mapping(addr,	__pte(0), 	\
+			(unsigned long)(vma)->vm_mm->cpu_vm_mask.bits| \
+				UVMF_INVLPG|UVMF_MULTI))) {	\
+		__ptep->pte = 0;				\
+		flush_tlb_page(vma, addr);			\
+	}							\
+	__res;							\
+})
+
+/*
+ * The following only work if pte_present() is true.
+ * Undefined behaviour if not..
+ */
+#define __LARGE_PTE (_PAGE_PSE|_PAGE_PRESENT)
+static inline int pte_user(pte_t pte)		{ return __pte_val(pte) & _PAGE_USER; }
+static inline int pte_read(pte_t pte)		{ return __pte_val(pte) & _PAGE_USER; }
+static inline int pte_exec(pte_t pte)		{ return __pte_val(pte) & _PAGE_USER; }
+static inline int pte_dirty(pte_t pte)		{ return __pte_val(pte) & _PAGE_DIRTY; }
+static inline int pte_young(pte_t pte)		{ return __pte_val(pte) & _PAGE_ACCESSED; }
+static inline int pte_write(pte_t pte)		{ return __pte_val(pte) & _PAGE_RW; }
+static inline int pte_file(pte_t pte)		{ return __pte_val(pte) & _PAGE_FILE; }
+static inline int pte_huge(pte_t pte)		{ return __pte_val(pte) & _PAGE_PSE; }
+
+static inline pte_t pte_rdprotect(pte_t pte)	{ __pte_val(pte) &= ~_PAGE_USER; return pte; }
+static inline pte_t pte_exprotect(pte_t pte)	{ __pte_val(pte) &= ~_PAGE_USER; return pte; }
+static inline pte_t pte_mkclean(pte_t pte)	{ __pte_val(pte) &= ~_PAGE_DIRTY; return pte; }
+static inline pte_t pte_mkold(pte_t pte)	{ __pte_val(pte) &= ~_PAGE_ACCESSED; return pte; }
+static inline pte_t pte_wrprotect(pte_t pte)	{ __pte_val(pte) &= ~_PAGE_RW; return pte; }
+static inline pte_t pte_mkread(pte_t pte)	{ __pte_val(pte) |= _PAGE_USER; return pte; }
+static inline pte_t pte_mkexec(pte_t pte)	{ __pte_val(pte) |= _PAGE_USER; return pte; }
+static inline pte_t pte_mkdirty(pte_t pte)	{ __pte_val(pte) |= _PAGE_DIRTY; return pte; }
+static inline pte_t pte_mkyoung(pte_t pte)	{ __pte_val(pte) |= _PAGE_ACCESSED; return pte; }
+static inline pte_t pte_mkwrite(pte_t pte)	{ __pte_val(pte) |= _PAGE_RW; return pte; }
+static inline pte_t pte_mkhuge(pte_t pte)	{ __pte_val(pte) |= _PAGE_PSE; return pte; }
+
+#define ptep_test_and_clear_dirty(vma, addr, ptep)			\
+({									\
+	pte_t __pte = *(ptep);						\
+	int __ret = pte_dirty(__pte);					\
+	if (__ret)							\
+		set_pte_at((vma)->vm_mm, addr, ptep, pte_mkclean(__pte)); \
+	__ret;								\
+})
+
+#define ptep_test_and_clear_young(vma, addr, ptep)			\
+({									\
+	pte_t __pte = *(ptep);						\
+	int __ret = pte_young(__pte);					\
+	if (__ret)							\
+		set_pte_at((vma)->vm_mm, addr, ptep, pte_mkold(__pte)); \
+	__ret;								\
+})
+
+static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
+{
+	pte_t pte = *ptep;
+	if (pte_write(pte))
+		set_pte_at(mm, addr, ptep, pte_wrprotect(pte));
+}
+
+/*
+ * Macro to mark a page protection value as "uncacheable".
+ */
+#define pgprot_noncached(prot)	(__pgprot(pgprot_val(prot) | _PAGE_PCD | _PAGE_PWT))
+
+static inline int pmd_large(pmd_t pte) { 
+	return (__pmd_val(pte) & __LARGE_PTE) == __LARGE_PTE;
+} 	
+
+
+/*
+ * Conversion functions: convert a page and protection to a page entry,
+ * and a page entry and page directory to the page they refer to.
+ */
+
+/*
+ * Level 4 access.
+ * Never use these in the common code.
+ */
+#define pgd_page(pgd) ((unsigned long) __va(pgd_val(pgd) & PTE_MASK))
+#define pgd_index(address) (((address) >> PGDIR_SHIFT) & (PTRS_PER_PGD-1))
+#define pgd_offset(mm, addr) ((mm)->pgd + pgd_index(addr))
+#define pgd_offset_k(address) (init_level4_pgt + pgd_index(address))
+#define pgd_present(pgd) (__pgd_val(pgd) & _PAGE_PRESENT)
+#define mk_kernel_pgd(address) __pgd((address) | _KERNPG_TABLE)
+
+/* PUD - Level3 access */
+/* to find an entry in a page-table-directory. */
+#define pud_index(address) (((address) >> PUD_SHIFT) & (PTRS_PER_PUD-1))
+#define pud_offset(pgd, address) ((pud_t *) pgd_page(*(pgd)) + pud_index(address))
+#define pud_present(pud) (__pud_val(pud) & _PAGE_PRESENT)
+
+/* PMD  - Level 2 access */
+#define pmd_page_kernel(pmd) ((unsigned long) __va(pmd_val(pmd) & PTE_MASK))
+#define pmd_page(pmd)		(pfn_to_page(pmd_val(pmd) >> PAGE_SHIFT))
+
+#define pmd_index(address) (((address) >> PMD_SHIFT) & (PTRS_PER_PMD-1))
+#define pmd_offset(dir, address) ((pmd_t *) pud_page(*(dir)) + \
+                                  pmd_index(address))
+#define pmd_none(x)	(!__pmd_val(x))
+#if CONFIG_XEN_COMPAT <= 0x030002
+/* pmd_present doesn't just test the _PAGE_PRESENT bit since wr.p.t.
+   can temporarily clear it. */
+#define pmd_present(x)	(__pmd_val(x))
+#else
+#define pmd_present(x)	(__pmd_val(x) & _PAGE_PRESENT)
+#endif
+#define pmd_clear(xp)	do { set_pmd(xp, __pmd(0)); } while (0)
+#define pmd_bad(x) ((__pmd_val(x) & ~(PTE_MASK | _PAGE_USER | _PAGE_PRESENT)) \
+		    != (_KERNPG_TABLE & ~(_PAGE_USER | _PAGE_PRESENT)))
+#define pfn_pmd(nr,prot) (__pmd(((nr) << PAGE_SHIFT) | pgprot_val(prot)))
+#define pmd_pfn(x)  ((pmd_val(x) & __PHYSICAL_MASK) >> PAGE_SHIFT)
+
+#define pte_to_pgoff(pte) ((__pte_val(pte) & PHYSICAL_PAGE_MASK) >> PAGE_SHIFT)
+#define pgoff_to_pte(off) ((pte_t) { ((off) << PAGE_SHIFT) | _PAGE_FILE })
+#define PTE_FILE_MAX_BITS __PHYSICAL_MASK_SHIFT
+
+/* PTE - Level 1 access. */
+
+/* page, protection -> pte */
+#define mk_pte(page, pgprot)	pfn_pte(page_to_pfn(page), (pgprot))
+#define mk_pte_huge(entry) (__pte_val(entry) |= _PAGE_PRESENT | _PAGE_PSE)
+ 
+/* physical address -> PTE */
+static inline pte_t mk_pte_phys(unsigned long physpage, pgprot_t pgprot)
+{ 
+	unsigned long pteval;
+	pteval = physpage | pgprot_val(pgprot);
+	return __pte(pteval);
+}
+ 
+/* Change flags of a PTE */
+static inline pte_t pte_modify(pte_t pte, pgprot_t newprot)
+{ 
+	/*
+	 * Since this might change the present bit (which controls whether
+	 * a pte_t object has undergone p2m translation), we must use
+	 * pte_val() on the input pte and __pte() for the return value.
+	 */
+	unsigned long pteval = pte_val(pte);
+
+	pteval &= _PAGE_CHG_MASK;
+	pteval |= pgprot_val(newprot);
+	pteval &= __supported_pte_mask;
+	return __pte(pteval);
+}
+
+#define pte_index(address) \
+		(((address) >> PAGE_SHIFT) & (PTRS_PER_PTE - 1))
+#define pte_offset_kernel(dir, address) ((pte_t *) pmd_page_kernel(*(dir)) + \
+			pte_index(address))
+
+/* x86-64 always has all page tables mapped. */
+#define pte_offset_map(dir,address) pte_offset_kernel(dir,address)
+#define pte_offset_map_nested(dir,address) pte_offset_kernel(dir,address)
+#define pte_unmap(pte) /* NOP */
+#define pte_unmap_nested(pte) /* NOP */ 
+
+#define update_mmu_cache(vma,address,pte) do { } while (0)
+
+/*
+ * Rules for using ptep_establish: the pte MUST be a user pte, and
+ * must be a present->present transition.
+ */
+#define __HAVE_ARCH_PTEP_ESTABLISH
+#define ptep_establish(vma, address, ptep, pteval)			\
+	do {								\
+		if ( likely((vma)->vm_mm == current->mm) ) {		\
+			BUG_ON(HYPERVISOR_update_va_mapping(address,	\
+				pteval,					\
+				(unsigned long)(vma)->vm_mm->cpu_vm_mask.bits| \
+					UVMF_INVLPG|UVMF_MULTI));	\
+		} else {						\
+			xen_l1_entry_update(ptep, pteval);		\
+			flush_tlb_page(vma, address);			\
+		}							\
+	} while (0)
+
+/* We only update the dirty/accessed state if we set
+ * the dirty bit by hand in the kernel, since the hardware
+ * will do the accessed bit for us, and we don't want to
+ * race with other CPU's that might be updating the dirty
+ * bit at the same time. */
+#define  __HAVE_ARCH_PTEP_SET_ACCESS_FLAGS
+#define ptep_set_access_flags(vma, address, ptep, entry, dirty)		\
+	do {								\
+		if (dirty)						\
+			ptep_establish(vma, address, ptep, entry);	\
+	} while (0)
+
+/* Encode and de-code a swap entry */
+#define __swp_type(x)			(((x).val >> 1) & 0x3f)
+#define __swp_offset(x)			((x).val >> 8)
+#define __swp_entry(type, offset)	((swp_entry_t) { ((type) << 1) | ((offset) << 8) })
+#define __pte_to_swp_entry(pte)		((swp_entry_t) { __pte_val(pte) })
+#define __swp_entry_to_pte(x)		((pte_t) { (x).val })
+
+extern spinlock_t pgd_lock;
+extern struct page *pgd_list;
+void vmalloc_sync_all(void);
+
+#endif /* !__ASSEMBLY__ */
+
+extern int kern_addr_valid(unsigned long addr); 
+
+#define DOMID_LOCAL (0xFFFFU)
+
+struct vm_area_struct;
+
+int direct_remap_pfn_range(struct vm_area_struct *vma,
+                            unsigned long address,
+                            unsigned long mfn,
+                            unsigned long size,
+                            pgprot_t prot,
+                            domid_t  domid);
+
+int direct_kernel_remap_pfn_range(unsigned long address, 
+				  unsigned long mfn,
+				  unsigned long size, 
+				  pgprot_t prot,
+				  domid_t  domid);
+
+int create_lookup_pte_addr(struct mm_struct *mm,
+                           unsigned long address,
+                           uint64_t *ptep);
+
+int touch_pte_range(struct mm_struct *mm,
+                    unsigned long address,
+                    unsigned long size);
+
+int xen_change_pte_range(struct mm_struct *mm, pmd_t *pmd,
+		unsigned long addr, unsigned long end, pgprot_t newprot);
+
+#define arch_change_pte_range(mm, pmd, addr, end, newprot)	\
+		xen_change_pte_range(mm, pmd, addr, end, newprot)
+
+#define io_remap_pfn_range(vma, vaddr, pfn, size, prot)		\
+		direct_remap_pfn_range(vma,vaddr,pfn,size,prot,DOMID_IO)
+
+#define MK_IOSPACE_PFN(space, pfn)	(pfn)
+#define GET_IOSPACE(pfn)		0
+#define GET_PFN(pfn)			(pfn)
+
+#define HAVE_ARCH_UNMAPPED_AREA
+
+#define pgtable_cache_init()   do { } while (0)
+#define check_pgt_cache()      do { } while (0)
+
+#define PAGE_AGP    PAGE_KERNEL_NOCACHE
+#define HAVE_PAGE_AGP 1
+
+/* fs/proc/kcore.c */
+#define	kc_vaddr_to_offset(v) ((v) & __VIRTUAL_MASK)
+#define	kc_offset_to_vaddr(o) \
+   (((o) & (1UL << (__VIRTUAL_MASK_SHIFT-1))) ? ((o) | (~__VIRTUAL_MASK)) : (o))
+
+#define __HAVE_ARCH_PTEP_TEST_AND_CLEAR_YOUNG
+#define __HAVE_ARCH_PTEP_TEST_AND_CLEAR_DIRTY
+#define __HAVE_ARCH_PTEP_GET_AND_CLEAR
+#define __HAVE_ARCH_PTEP_GET_AND_CLEAR_FULL
+#define __HAVE_ARCH_PTEP_CLEAR_FLUSH
+#define __HAVE_ARCH_PTEP_SET_WRPROTECT
+#define __HAVE_ARCH_PTE_SAME
+#include <asm-generic/pgtable.h>
+
+#endif /* _X86_64_PGTABLE_H */
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/processor.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/processor.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/processor.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/processor.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,502 @@
+/*
+ * include/asm-x86_64/processor.h
+ *
+ * Copyright (C) 1994 Linus Torvalds
+ */
+
+#ifndef __ASM_X86_64_PROCESSOR_H
+#define __ASM_X86_64_PROCESSOR_H
+
+#include <asm/segment.h>
+#include <asm/page.h>
+#include <asm/types.h>
+#include <asm/sigcontext.h>
+#include <asm/cpufeature.h>
+#include <linux/threads.h>
+#include <asm/msr.h>
+#include <asm/current.h>
+#include <asm/system.h>
+#include <asm/mmsegment.h>
+#include <asm/percpu.h>
+#include <linux/personality.h>
+#include <linux/cpumask.h>
+
+#define TF_MASK		0x00000100
+#define IF_MASK		0x00000200
+#define IOPL_MASK	0x00003000
+#define NT_MASK		0x00004000
+#define VM_MASK		0x00020000
+#define AC_MASK		0x00040000
+#define VIF_MASK	0x00080000	/* virtual interrupt flag */
+#define VIP_MASK	0x00100000	/* virtual interrupt pending */
+#define ID_MASK		0x00200000
+
+#define desc_empty(desc) \
+               (!((desc)->a | (desc)->b))
+
+#define desc_equal(desc1, desc2) \
+               (((desc1)->a == (desc2)->a) && ((desc1)->b == (desc2)->b))
+
+/*
+ * Default implementation of macro that returns current
+ * instruction pointer ("program counter").
+ */
+#define current_text_addr() ({ void *pc; asm volatile("leaq 1f(%%rip),%0\n1:":"=r"(pc)); pc; })
+
+/*
+ *  CPU type and hardware bug flags. Kept separately for each CPU.
+ */
+
+struct cpuinfo_x86 {
+	__u8	x86;		/* CPU family */
+	__u8	x86_vendor;	/* CPU vendor */
+	__u8	x86_model;
+	__u8	x86_mask;
+	int	cpuid_level;	/* Maximum supported CPUID level, -1=no CPUID */
+	__u32	x86_capability[NCAPINTS];
+	char	x86_vendor_id[16];
+	char	x86_model_id[64];
+	int 	x86_cache_size;  /* in KB */
+	int	x86_clflush_size;
+	int	x86_cache_alignment;
+	int	x86_tlbsize;	/* number of 4K pages in DTLB/ITLB combined(in pages)*/
+        __u8    x86_virt_bits, x86_phys_bits;
+	__u8	x86_max_cores;	/* cpuid returned max cores value */
+        __u32   x86_power; 	
+	__u32   extended_cpuid_level;	/* Max extended CPUID function supported */
+	unsigned long loops_per_jiffy;
+#ifdef CONFIG_SMP
+	cpumask_t llc_shared_map;	/* cpus sharing the last level cache */
+#endif
+	__u8	apicid;
+#ifdef CONFIG_SMP
+	__u8	booted_cores;	/* number of cores as seen by OS */
+	__u8	phys_proc_id;	/* Physical Processor id. */
+	__u8	cpu_core_id;	/* Core id. */
+#endif
+} ____cacheline_aligned;
+
+#define X86_VENDOR_INTEL 0
+#define X86_VENDOR_CYRIX 1
+#define X86_VENDOR_AMD 2
+#define X86_VENDOR_UMC 3
+#define X86_VENDOR_NEXGEN 4
+#define X86_VENDOR_CENTAUR 5
+#define X86_VENDOR_RISE 6
+#define X86_VENDOR_TRANSMETA 7
+#define X86_VENDOR_NUM 8
+#define X86_VENDOR_UNKNOWN 0xff
+
+#ifdef CONFIG_SMP
+extern struct cpuinfo_x86 cpu_data[];
+#define current_cpu_data cpu_data[smp_processor_id()]
+#else
+#define cpu_data (&boot_cpu_data)
+#define current_cpu_data boot_cpu_data
+#endif
+
+extern char ignore_irq13;
+
+extern void identify_cpu(struct cpuinfo_x86 *);
+extern void print_cpu_info(struct cpuinfo_x86 *);
+extern unsigned int init_intel_cacheinfo(struct cpuinfo_x86 *c);
+extern unsigned short num_cache_leaves;
+
+/*
+ * EFLAGS bits
+ */
+#define X86_EFLAGS_CF	0x00000001 /* Carry Flag */
+#define X86_EFLAGS_PF	0x00000004 /* Parity Flag */
+#define X86_EFLAGS_AF	0x00000010 /* Auxillary carry Flag */
+#define X86_EFLAGS_ZF	0x00000040 /* Zero Flag */
+#define X86_EFLAGS_SF	0x00000080 /* Sign Flag */
+#define X86_EFLAGS_TF	0x00000100 /* Trap Flag */
+#define X86_EFLAGS_IF	0x00000200 /* Interrupt Flag */
+#define X86_EFLAGS_DF	0x00000400 /* Direction Flag */
+#define X86_EFLAGS_OF	0x00000800 /* Overflow Flag */
+#define X86_EFLAGS_IOPL	0x00003000 /* IOPL mask */
+#define X86_EFLAGS_NT	0x00004000 /* Nested Task */
+#define X86_EFLAGS_RF	0x00010000 /* Resume Flag */
+#define X86_EFLAGS_VM	0x00020000 /* Virtual Mode */
+#define X86_EFLAGS_AC	0x00040000 /* Alignment Check */
+#define X86_EFLAGS_VIF	0x00080000 /* Virtual Interrupt Flag */
+#define X86_EFLAGS_VIP	0x00100000 /* Virtual Interrupt Pending */
+#define X86_EFLAGS_ID	0x00200000 /* CPUID detection flag */
+
+/*
+ * Intel CPU features in CR4
+ */
+#define X86_CR4_VME		0x0001	/* enable vm86 extensions */
+#define X86_CR4_PVI		0x0002	/* virtual interrupts flag enable */
+#define X86_CR4_TSD		0x0004	/* disable time stamp at ipl 3 */
+#define X86_CR4_DE		0x0008	/* enable debugging extensions */
+#define X86_CR4_PSE		0x0010	/* enable page size extensions */
+#define X86_CR4_PAE		0x0020	/* enable physical address extensions */
+#define X86_CR4_MCE		0x0040	/* Machine check enable */
+#define X86_CR4_PGE		0x0080	/* enable global pages */
+#define X86_CR4_PCE		0x0100	/* enable performance counters at ipl 3 */
+#define X86_CR4_OSFXSR		0x0200	/* enable fast FPU save and restore */
+#define X86_CR4_OSXMMEXCPT	0x0400	/* enable unmasked SSE exceptions */
+
+/*
+ * Save the cr4 feature set we're using (ie
+ * Pentium 4MB enable and PPro Global page
+ * enable), so that any CPU's that boot up
+ * after us can get the correct flags.
+ */
+extern unsigned long mmu_cr4_features;
+
+static inline void set_in_cr4 (unsigned long mask)
+{
+	mmu_cr4_features |= mask;
+	__asm__("movq %%cr4,%%rax\n\t"
+		"orq %0,%%rax\n\t"
+		"movq %%rax,%%cr4\n"
+		: : "irg" (mask)
+		:"ax");
+}
+
+static inline void clear_in_cr4 (unsigned long mask)
+{
+	mmu_cr4_features &= ~mask;
+	__asm__("movq %%cr4,%%rax\n\t"
+		"andq %0,%%rax\n\t"
+		"movq %%rax,%%cr4\n"
+		: : "irg" (~mask)
+		:"ax");
+}
+
+
+/*
+ * User space process size. 47bits minus one guard page.
+ */
+#define TASK_SIZE64	(0x800000000000UL - 4096)
+
+/* This decides where the kernel will search for a free chunk of vm
+ * space during mmap's.
+ */
+#define IA32_PAGE_OFFSET ((current->personality & ADDR_LIMIT_3GB) ? 0xc0000000 : 0xFFFFe000)
+
+#define TASK_SIZE 		(test_thread_flag(TIF_IA32) ? IA32_PAGE_OFFSET : TASK_SIZE64)
+#define TASK_SIZE_OF(child) 	((test_tsk_thread_flag(child, TIF_IA32)) ? IA32_PAGE_OFFSET : TASK_SIZE64)
+
+#define TASK_UNMAPPED_BASE	PAGE_ALIGN(TASK_SIZE/3)
+
+/*
+ * Size of io_bitmap.
+ */
+#define IO_BITMAP_BITS  65536
+#define IO_BITMAP_BYTES (IO_BITMAP_BITS/8)
+#define IO_BITMAP_LONGS (IO_BITMAP_BYTES/sizeof(long))
+#ifndef CONFIG_X86_NO_TSS
+#define IO_BITMAP_OFFSET offsetof(struct tss_struct,io_bitmap)
+#endif
+#define INVALID_IO_BITMAP_OFFSET 0x8000
+
+struct i387_fxsave_struct {
+	u16	cwd;
+	u16	swd;
+	u16	twd;
+	u16	fop;
+	u64	rip;
+	u64	rdp; 
+	u32	mxcsr;
+	u32	mxcsr_mask;
+	u32	st_space[32];	/* 8*16 bytes for each FP-reg = 128 bytes */
+	u32	xmm_space[64];	/* 16*16 bytes for each XMM-reg = 128 bytes */
+	u32	padding[24];
+} __attribute__ ((aligned (16)));
+
+union i387_union {
+	struct i387_fxsave_struct	fxsave;
+};
+
+#ifndef CONFIG_X86_NO_TSS
+struct tss_struct {
+	u32 reserved1;
+	u64 rsp0;	
+	u64 rsp1;
+	u64 rsp2;
+	u64 reserved2;
+	u64 ist[7];
+	u32 reserved3;
+	u32 reserved4;
+	u16 reserved5;
+	u16 io_bitmap_base;
+	/*
+	 * The extra 1 is there because the CPU will access an
+	 * additional byte beyond the end of the IO permission
+	 * bitmap. The extra byte must be all 1 bits, and must
+	 * be within the limit. Thus we have:
+	 *
+	 * 128 bytes, the bitmap itself, for ports 0..0x3ff
+	 * 8 bytes, for an extra "long" of ~0UL
+	 */
+	unsigned long io_bitmap[IO_BITMAP_LONGS + 1];
+} __attribute__((packed)) ____cacheline_aligned;
+
+DECLARE_PER_CPU(struct tss_struct,init_tss);
+#endif
+
+
+extern struct cpuinfo_x86 boot_cpu_data;
+#ifndef CONFIG_X86_NO_TSS
+/* Save the original ist values for checking stack pointers during debugging */
+struct orig_ist {
+	unsigned long ist[7];
+};
+DECLARE_PER_CPU(struct orig_ist, orig_ist);
+#endif
+
+#ifdef CONFIG_X86_VSMP
+#define ARCH_MIN_TASKALIGN	(1 << INTERNODE_CACHE_SHIFT)
+#define ARCH_MIN_MMSTRUCT_ALIGN	(1 << INTERNODE_CACHE_SHIFT)
+#else
+#define ARCH_MIN_TASKALIGN	16
+#define ARCH_MIN_MMSTRUCT_ALIGN	0
+#endif
+
+struct thread_struct {
+	unsigned long	rsp0;
+	unsigned long	rsp;
+	unsigned long 	userrsp;	/* Copy from PDA */ 
+	unsigned long	fs;
+	unsigned long	gs;
+	unsigned short	es, ds, fsindex, gsindex;	
+/* Hardware debugging registers */
+	unsigned long	debugreg0;  
+	unsigned long	debugreg1;  
+	unsigned long	debugreg2;  
+	unsigned long	debugreg3;  
+	unsigned long	debugreg6;  
+	unsigned long	debugreg7;  
+/* fault info */
+	unsigned long	cr2, trap_no, error_code;
+/* floating point info */
+	union i387_union	i387  __attribute__((aligned(16)));
+/* IO permissions. the bitmap could be moved into the GDT, that would make
+   switch faster for a limited number of ioperm using tasks. -AK */
+	int		ioperm;
+	unsigned long	*io_bitmap_ptr;
+	unsigned io_bitmap_max;
+/* cached TLS descriptors. */
+	u64 tls_array[GDT_ENTRY_TLS_ENTRIES];
+	unsigned int	iopl;
+} __attribute__((aligned(16)));
+
+#define INIT_THREAD  { \
+	.rsp0 = (unsigned long)&init_stack + sizeof(init_stack) \
+}
+
+#ifndef CONFIG_X86_NO_TSS
+#define INIT_TSS  { \
+	.rsp0 = (unsigned long)&init_stack + sizeof(init_stack) \
+}
+#endif
+
+#define INIT_MMAP \
+{ &init_mm, 0, 0, NULL, PAGE_SHARED, VM_READ | VM_WRITE | VM_EXEC, 1, NULL, NULL }
+
+#define start_thread(regs,new_rip,new_rsp) do { \
+	asm volatile("movl %0,%%fs; movl %0,%%es; movl %0,%%ds": :"r" (0));	 \
+	load_gs_index(0);							\
+	(regs)->rip = (new_rip);						 \
+	(regs)->rsp = (new_rsp);						 \
+	write_pda(oldrsp, (new_rsp));						 \
+	(regs)->cs = __USER_CS;							 \
+	(regs)->ss = __USER_DS;							 \
+	(regs)->eflags = 0x200;							 \
+	set_fs(USER_DS);							 \
+} while(0) 
+
+#define get_debugreg(var, register)				\
+	var = HYPERVISOR_get_debugreg(register)
+#define set_debugreg(value, register) do {			\
+	if (HYPERVISOR_set_debugreg(register, value))		\
+		BUG();						\
+} while (0)
+
+struct task_struct;
+struct mm_struct;
+
+/* Free all resources held by a thread. */
+extern void release_thread(struct task_struct *);
+
+/* Prepare to copy thread state - unlazy all lazy status */
+extern void prepare_to_copy(struct task_struct *tsk);
+
+/*
+ * create a kernel thread without removing it from tasklists
+ */
+extern long kernel_thread(int (*fn)(void *), void * arg, unsigned long flags);
+
+/*
+ * Return saved PC of a blocked thread.
+ * What is this good for? it will be always the scheduler or ret_from_fork.
+ */
+#define thread_saved_pc(t) (*(unsigned long *)((t)->thread.rsp - 8))
+
+extern unsigned long get_wchan(struct task_struct *p);
+#define task_pt_regs(tsk) ((struct pt_regs *)(tsk)->thread.rsp0 - 1)
+#define KSTK_EIP(tsk) (task_pt_regs(tsk)->rip)
+#define KSTK_ESP(tsk) -1 /* sorry. doesn't work for syscall. */
+
+
+struct microcode_header {
+	unsigned int hdrver;
+	unsigned int rev;
+	unsigned int date;
+	unsigned int sig;
+	unsigned int cksum;
+	unsigned int ldrver;
+	unsigned int pf;
+	unsigned int datasize;
+	unsigned int totalsize;
+	unsigned int reserved[3];
+};
+
+struct microcode {
+	struct microcode_header hdr;
+	unsigned int bits[0];
+};
+
+typedef struct microcode microcode_t;
+typedef struct microcode_header microcode_header_t;
+
+/* microcode format is extended from prescott processors */
+struct extended_signature {
+	unsigned int sig;
+	unsigned int pf;
+	unsigned int cksum;
+};
+
+struct extended_sigtable {
+	unsigned int count;
+	unsigned int cksum;
+	unsigned int reserved[3];
+	struct extended_signature sigs[0];
+};
+
+
+#define ASM_NOP1 K8_NOP1
+#define ASM_NOP2 K8_NOP2
+#define ASM_NOP3 K8_NOP3
+#define ASM_NOP4 K8_NOP4
+#define ASM_NOP5 K8_NOP5
+#define ASM_NOP6 K8_NOP6
+#define ASM_NOP7 K8_NOP7
+#define ASM_NOP8 K8_NOP8
+
+/* Opteron nops */
+#define K8_NOP1 ".byte 0x90\n"
+#define K8_NOP2	".byte 0x66,0x90\n" 
+#define K8_NOP3	".byte 0x66,0x66,0x90\n" 
+#define K8_NOP4	".byte 0x66,0x66,0x66,0x90\n" 
+#define K8_NOP5	K8_NOP3 K8_NOP2 
+#define K8_NOP6	K8_NOP3 K8_NOP3
+#define K8_NOP7	K8_NOP4 K8_NOP3
+#define K8_NOP8	K8_NOP4 K8_NOP4
+
+#define ASM_NOP_MAX 8
+
+/* REP NOP (PAUSE) is a good thing to insert into busy-wait loops. */
+static inline void rep_nop(void)
+{
+	__asm__ __volatile__("rep;nop": : :"memory");
+}
+
+/* Stop speculative execution */
+static inline void sync_core(void)
+{ 
+	int tmp;
+	asm volatile("cpuid" : "=a" (tmp) : "0" (1) : "ebx","ecx","edx","memory");
+} 
+
+#define cpu_has_fpu 1
+
+#define ARCH_HAS_PREFETCH
+static inline void prefetch(void *x) 
+{ 
+	asm volatile("prefetcht0 %0" :: "m" (*(unsigned long *)x));
+} 
+
+#define ARCH_HAS_PREFETCHW 1
+static inline void prefetchw(void *x) 
+{ 
+	alternative_input("prefetcht0 (%1)",
+			  "prefetchw (%1)",
+			  X86_FEATURE_3DNOW,
+			  "r" (x));
+} 
+
+#define ARCH_HAS_SPINLOCK_PREFETCH 1
+
+#define spin_lock_prefetch(x)  prefetchw(x)
+
+#define cpu_relax()   rep_nop()
+
+/*
+ *      NSC/Cyrix CPU configuration register indexes
+ */
+#define CX86_CCR0 0xc0
+#define CX86_CCR1 0xc1
+#define CX86_CCR2 0xc2
+#define CX86_CCR3 0xc3
+#define CX86_CCR4 0xe8
+#define CX86_CCR5 0xe9
+#define CX86_CCR6 0xea
+#define CX86_CCR7 0xeb
+#define CX86_DIR0 0xfe
+#define CX86_DIR1 0xff
+#define CX86_ARR_BASE 0xc4
+#define CX86_RCR_BASE 0xdc
+
+/*
+ *      NSC/Cyrix CPU indexed register access macros
+ */
+
+#define getCx86(reg) ({ outb((reg), 0x22); inb(0x23); })
+
+#define setCx86(reg, data) do { \
+	outb((reg), 0x22); \
+	outb((data), 0x23); \
+} while (0)
+
+static inline void serialize_cpu(void)
+{
+	__asm__ __volatile__ ("cpuid" : : : "ax", "bx", "cx", "dx");
+}
+
+static inline void __monitor(const void *eax, unsigned long ecx,
+		unsigned long edx)
+{
+	/* "monitor %eax,%ecx,%edx;" */
+	asm volatile(
+		".byte 0x0f,0x01,0xc8;"
+		: :"a" (eax), "c" (ecx), "d"(edx));
+}
+
+static inline void __mwait(unsigned long eax, unsigned long ecx)
+{
+	/* "mwait %eax,%ecx;" */
+	asm volatile(
+		".byte 0x0f,0x01,0xc9;"
+		: :"a" (eax), "c" (ecx));
+}
+
+#define stack_current() \
+({								\
+	struct thread_info *ti;					\
+	asm("andq %%rsp,%0; ":"=r" (ti) : "0" (CURRENT_MASK));	\
+	ti->task;					\
+})
+
+#define cache_line_size() (boot_cpu_data.x86_cache_alignment)
+
+extern unsigned long boot_option_idle_override;
+/* Boot loader type from the setup header */
+extern int bootloader_type;
+
+#define HAVE_ARCH_PICK_MMAP_LAYOUT 1
+
+#endif /* __ASM_X86_64_PROCESSOR_H */
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/ptrace.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/ptrace.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/ptrace.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/ptrace.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,127 @@
+#ifndef _X86_64_PTRACE_H
+#define _X86_64_PTRACE_H
+
+#if defined(__ASSEMBLY__) || defined(__FRAME_OFFSETS) 
+#define R15 0
+#define R14 8
+#define R13 16
+#define R12 24
+#define RBP 32
+#define RBX 40
+/* arguments: interrupts/non tracing syscalls only save upto here*/
+#define R11 48
+#define R10 56	
+#define R9 64
+#define R8 72
+#define RAX 80
+#define RCX 88
+#define RDX 96
+#define RSI 104
+#define RDI 112
+#define ORIG_RAX 120       /* = ERROR */ 
+/* end of arguments */ 	
+/* cpu exception frame or undefined in case of fast syscall. */
+#define RIP 128
+#define CS 136
+#define EFLAGS 144
+#define RSP 152
+#define SS 160
+#define ARGOFFSET R11
+#endif /* __ASSEMBLY__ */
+
+/* top of stack page */ 
+#define FRAME_SIZE 168
+
+#define PTRACE_OLDSETOPTIONS         21
+
+#ifndef __ASSEMBLY__ 
+
+struct pt_regs {
+	unsigned long r15;
+	unsigned long r14;
+	unsigned long r13;
+	unsigned long r12;
+	unsigned long rbp;
+	unsigned long rbx;
+/* arguments: non interrupts/non tracing syscalls only save upto here*/
+ 	unsigned long r11;
+	unsigned long r10;	
+	unsigned long r9;
+	unsigned long r8;
+	unsigned long rax;
+	unsigned long rcx;
+	unsigned long rdx;
+	unsigned long rsi;
+	unsigned long rdi;
+	unsigned long orig_rax;
+/* end of arguments */ 	
+/* cpu exception frame or undefined */
+	unsigned long rip;
+	unsigned long cs;
+	unsigned long eflags; 
+	unsigned long rsp; 
+	unsigned long ss;
+/* top of stack page */ 
+};
+
+#endif
+
+/* Arbitrarily choose the same ptrace numbers as used by the Sparc code. */
+#define PTRACE_GETREGS            12
+#define PTRACE_SETREGS            13
+#define PTRACE_GETFPREGS          14
+#define PTRACE_SETFPREGS          15
+#define PTRACE_GETFPXREGS         18
+#define PTRACE_SETFPXREGS         19
+
+/* only useful for access 32bit programs */
+#define PTRACE_GET_THREAD_AREA    25
+#define PTRACE_SET_THREAD_AREA    26
+
+#define PTRACE_ARCH_PRCTL	  30	/* arch_prctl for child */
+
+#if defined(__KERNEL__) && !defined(__ASSEMBLY__) 
+#define user_mode(regs) (!!((regs)->cs & 3))
+#define user_mode_vm(regs) user_mode(regs)
+#define instruction_pointer(regs) ((regs)->rip)
+#if defined(CONFIG_SMP) && defined(CONFIG_FRAME_POINTER)
+extern unsigned long profile_pc(struct pt_regs *regs);
+#else
+#define profile_pc(regs) instruction_pointer(regs)
+#endif
+
+#include <linux/compiler.h>
+
+void signal_fault(struct pt_regs *regs, void __user *frame, char *where);
+
+struct task_struct;
+
+extern unsigned long
+convert_rip_to_linear(struct task_struct *child, struct pt_regs *regs);
+
+enum {
+        EF_CF   = 0x00000001,
+        EF_PF   = 0x00000004,
+        EF_AF   = 0x00000010,
+        EF_ZF   = 0x00000040,
+        EF_SF   = 0x00000080,
+        EF_TF   = 0x00000100,
+        EF_IE   = 0x00000200,
+        EF_DF   = 0x00000400,
+        EF_OF   = 0x00000800,
+        EF_IOPL = 0x00003000,
+        EF_IOPL_RING0 = 0x00000000,
+        EF_IOPL_RING1 = 0x00001000,
+        EF_IOPL_RING2 = 0x00002000,
+        EF_NT   = 0x00004000,   /* nested task */
+        EF_RF   = 0x00010000,   /* resume */
+        EF_VM   = 0x00020000,   /* virtual mode */
+        EF_AC   = 0x00040000,   /* alignment */
+        EF_VIF  = 0x00080000,   /* virtual interrupt */
+        EF_VIP  = 0x00100000,   /* virtual interrupt pending */
+        EF_ID   = 0x00200000,   /* id */
+};
+
+#endif
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/smp.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/smp.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/smp.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/smp.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,150 @@
+#ifndef __ASM_SMP_H
+#define __ASM_SMP_H
+
+/*
+ * We need the APIC definitions automatically as part of 'smp.h'
+ */
+#ifndef __ASSEMBLY__
+#include <linux/threads.h>
+#include <linux/cpumask.h>
+#include <linux/bitops.h>
+extern int disable_apic;
+#endif
+
+#ifdef CONFIG_X86_LOCAL_APIC
+#ifndef __ASSEMBLY__
+#include <asm/fixmap.h>
+#include <asm/mpspec.h>
+#ifdef CONFIG_X86_IO_APIC
+#include <asm/io_apic.h>
+#endif
+#include <asm/apic.h>
+#include <asm/thread_info.h>
+#endif
+#endif
+
+#ifdef CONFIG_SMP
+#ifndef ASSEMBLY
+
+#include <asm/pda.h>
+
+struct pt_regs;
+
+extern cpumask_t cpu_present_mask;
+extern cpumask_t cpu_possible_map;
+extern cpumask_t cpu_online_map;
+extern cpumask_t cpu_initialized;
+
+/*
+ * Private routines/data
+ */
+ 
+extern void smp_alloc_memory(void);
+extern volatile unsigned long smp_invalidate_needed;
+extern int pic_mode;
+extern void lock_ipi_call_lock(void);
+extern void unlock_ipi_call_lock(void);
+extern int smp_num_siblings;
+extern void smp_send_reschedule(int cpu);
+void smp_stop_cpu(void);
+extern int smp_call_function_single(int cpuid, void (*func) (void *info),
+				void *info, int retry, int wait);
+
+extern cpumask_t cpu_sibling_map[NR_CPUS];
+extern cpumask_t cpu_core_map[NR_CPUS];
+extern u8 cpu_llc_id[NR_CPUS];
+
+#define SMP_TRAMPOLINE_BASE 0x6000
+
+/*
+ * On x86 all CPUs are mapped 1:1 to the APIC space.
+ * This simplifies scheduling and IPI sending and
+ * compresses data structures.
+ */
+
+static inline int num_booting_cpus(void)
+{
+	return cpus_weight(cpu_possible_map);
+}
+
+#define raw_smp_processor_id() read_pda(cpunumber)
+
+#ifdef CONFIG_X86_LOCAL_APIC
+static inline int hard_smp_processor_id(void)
+{
+	/* we don't want to mark this access volatile - bad code generation */
+	return GET_APIC_ID(*(unsigned int *)(APIC_BASE+APIC_ID));
+}
+#endif
+
+extern int safe_smp_processor_id(void);
+extern int __cpu_disable(void);
+extern void __cpu_die(unsigned int cpu);
+extern void prefill_possible_map(void);
+extern unsigned num_processors;
+extern unsigned disabled_cpus;
+
+#endif /* !ASSEMBLY */
+
+#define NO_PROC_ID		0xFF		/* No processor magic marker */
+
+#endif
+
+#ifndef ASSEMBLY
+/*
+ * Some lowlevel functions might want to know about
+ * the real APIC ID <-> CPU # mapping.
+ */
+extern u8 x86_cpu_to_apicid[NR_CPUS];	/* physical ID */
+extern u8 x86_cpu_to_log_apicid[NR_CPUS];
+extern u8 bios_cpu_apicid[];
+
+#ifdef CONFIG_X86_LOCAL_APIC
+static inline unsigned int cpu_mask_to_apicid(cpumask_t cpumask)
+{
+	return cpus_addr(cpumask)[0];
+}
+
+static inline int cpu_present_to_apicid(int mps_cpu)
+{
+	if (mps_cpu < NR_CPUS)
+		return (int)bios_cpu_apicid[mps_cpu];
+	else
+		return BAD_APICID;
+}
+#endif
+
+#endif /* !ASSEMBLY */
+
+#ifndef CONFIG_SMP
+#define stack_smp_processor_id() 0
+#define safe_smp_processor_id() 0
+#define cpu_logical_map(x) (x)
+#else
+#include <asm/thread_info.h>
+#define stack_smp_processor_id() \
+({ 								\
+	struct thread_info *ti;					\
+	__asm__("andq %%rsp,%0; ":"=r" (ti) : "0" (CURRENT_MASK));	\
+	ti->cpu;						\
+})
+#endif
+
+#ifndef __ASSEMBLY__
+#ifdef CONFIG_X86_LOCAL_APIC
+static __inline int logical_smp_processor_id(void)
+{
+	/* we don't want to mark this access volatile - bad code generation */
+	return GET_APIC_LOGICAL_ID(*(unsigned long *)(APIC_BASE+APIC_LDR));
+}
+#endif
+#endif
+
+#ifdef CONFIG_SMP
+#define cpu_physical_id(cpu)		x86_cpu_to_apicid[cpu]
+#else
+#define cpu_physical_id(cpu)		boot_cpu_id
+#endif
+
+#endif
+
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/synch_bitops.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/synch_bitops.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/synch_bitops.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/synch_bitops.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,2 @@
+
+#include <asm-i386/mach-xen/asm/synch_bitops.h>
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/system.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/system.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/system.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/system.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,256 @@
+#ifndef __ASM_SYSTEM_H
+#define __ASM_SYSTEM_H
+
+#include <linux/kernel.h>
+#include <asm/segment.h>
+#include <asm/alternative.h>
+
+#include <asm/synch_bitops.h>
+#include <asm/hypervisor.h>
+#include <xen/interface/arch-x86_64.h>
+
+#ifdef __KERNEL__
+
+#define __STR(x) #x
+#define STR(x) __STR(x)
+
+#define __SAVE(reg,offset) "movq %%" #reg ",(14-" #offset ")*8(%%rsp)\n\t"
+#define __RESTORE(reg,offset) "movq (14-" #offset ")*8(%%rsp),%%" #reg "\n\t"
+
+/* frame pointer must be last for get_wchan */
+#define SAVE_CONTEXT    "pushf ; pushq %%rbp ; movq %%rsi,%%rbp\n\t"
+#define RESTORE_CONTEXT "movq %%rbp,%%rsi ; popq %%rbp ; popf\n\t"
+
+#define __EXTRA_CLOBBER  \
+	,"rcx","rbx","rdx","r8","r9","r10","r11","r12","r13","r14","r15"
+
+#define switch_to(prev,next,last) \
+	asm volatile(SAVE_CONTEXT						    \
+		     "movq %%rsp,%P[threadrsp](%[prev])\n\t" /* save RSP */	  \
+		     "movq %P[threadrsp](%[next]),%%rsp\n\t" /* restore RSP */	  \
+		     "call __switch_to\n\t"					  \
+		     ".globl thread_return\n"					\
+		     "thread_return:\n\t"					    \
+		     "movq %%gs:%P[pda_pcurrent],%%rsi\n\t"			  \
+		     "movq %P[thread_info](%%rsi),%%r8\n\t"			  \
+		     LOCK_PREFIX "btr  %[tif_fork],%P[ti_flags](%%r8)\n\t"	  \
+		     "movq %%rax,%%rdi\n\t" 					  \
+		     "jc   ret_from_fork\n\t"					  \
+		     RESTORE_CONTEXT						    \
+		     : "=a" (last)					  	  \
+		     : [next] "S" (next), [prev] "D" (prev),			  \
+		       [threadrsp] "i" (offsetof(struct task_struct, thread.rsp)), \
+		       [ti_flags] "i" (offsetof(struct thread_info, flags)),\
+		       [tif_fork] "i" (TIF_FORK),			  \
+		       [thread_info] "i" (offsetof(struct task_struct, thread_info)), \
+		       [pda_pcurrent] "i" (offsetof(struct x8664_pda, pcurrent))   \
+		     : "memory", "cc" __EXTRA_CLOBBER)
+    
+extern void load_gs_index(unsigned); 
+
+/*
+ * Load a segment. Fall back on loading the zero
+ * segment if something goes wrong..
+ */
+#define loadsegment(seg,value)	\
+	asm volatile("\n"			\
+		"1:\t"				\
+		"movl %k0,%%" #seg "\n"		\
+		"2:\n"				\
+		".section .fixup,\"ax\"\n"	\
+		"3:\t"				\
+		"movl %1,%%" #seg "\n\t" 	\
+		"jmp 2b\n"			\
+		".previous\n"			\
+		".section __ex_table,\"a\"\n\t"	\
+		".align 8\n\t"			\
+		".quad 1b,3b\n"			\
+		".previous"			\
+		: :"r" (value), "r" (0))
+
+/*
+ * Clear and set 'TS' bit respectively
+ */
+#define clts() (HYPERVISOR_fpu_taskswitch(0))
+
+static inline unsigned long read_cr0(void)
+{ 
+	unsigned long cr0;
+	asm volatile("movq %%cr0,%0" : "=r" (cr0));
+	return cr0;
+} 
+
+static inline void write_cr0(unsigned long val) 
+{ 
+	asm volatile("movq %0,%%cr0" :: "r" (val));
+} 
+
+#define read_cr3() ({ \
+	unsigned long __dummy; \
+	asm("movq %%cr3,%0" : "=r" (__dummy)); \
+	machine_to_phys(__dummy); \
+})
+
+static inline unsigned long read_cr4(void)
+{ 
+	unsigned long cr4;
+	asm("movq %%cr4,%0" : "=r" (cr4));
+	return cr4;
+} 
+
+static inline void write_cr4(unsigned long val)
+{ 
+	asm volatile("movq %0,%%cr4" :: "r" (val));
+} 
+
+#define stts() (HYPERVISOR_fpu_taskswitch(1))
+
+#define wbinvd() \
+	__asm__ __volatile__ ("wbinvd": : :"memory");
+
+/*
+ * On SMP systems, when the scheduler does migration-cost autodetection,
+ * it needs a way to flush as much of the CPU's caches as possible.
+ */
+static inline void sched_cacheflush(void)
+{
+	wbinvd();
+}
+
+#endif	/* __KERNEL__ */
+
+#define nop() __asm__ __volatile__ ("nop")
+
+#define xchg(ptr,v) ((__typeof__(*(ptr)))__xchg((unsigned long)(v),(ptr),sizeof(*(ptr))))
+
+#define tas(ptr) (xchg((ptr),1))
+
+#define __xg(x) ((volatile long *)(x))
+
+static inline void set_64bit(volatile unsigned long *ptr, unsigned long val)
+{
+	*ptr = val;
+}
+
+#define _set_64bit set_64bit
+
+/*
+ * Note: no "lock" prefix even on SMP: xchg always implies lock anyway
+ * Note 2: xchg has side effect, so that attribute volatile is necessary,
+ *	  but generally the primitive is invalid, *ptr is output argument. --ANK
+ */
+static inline unsigned long __xchg(unsigned long x, volatile void * ptr, int size)
+{
+	switch (size) {
+		case 1:
+			__asm__ __volatile__("xchgb %b0,%1"
+				:"=q" (x)
+				:"m" (*__xg(ptr)), "0" (x)
+				:"memory");
+			break;
+		case 2:
+			__asm__ __volatile__("xchgw %w0,%1"
+				:"=r" (x)
+				:"m" (*__xg(ptr)), "0" (x)
+				:"memory");
+			break;
+		case 4:
+			__asm__ __volatile__("xchgl %k0,%1"
+				:"=r" (x)
+				:"m" (*__xg(ptr)), "0" (x)
+				:"memory");
+			break;
+		case 8:
+			__asm__ __volatile__("xchgq %0,%1"
+				:"=r" (x)
+				:"m" (*__xg(ptr)), "0" (x)
+				:"memory");
+			break;
+	}
+	return x;
+}
+
+/*
+ * Atomic compare and exchange.  Compare OLD with MEM, if identical,
+ * store NEW in MEM.  Return the initial value in MEM.  Success is
+ * indicated by comparing RETURN with OLD.
+ */
+
+#define __HAVE_ARCH_CMPXCHG 1
+
+static inline unsigned long __cmpxchg(volatile void *ptr, unsigned long old,
+				      unsigned long new, int size)
+{
+	unsigned long prev;
+	switch (size) {
+	case 1:
+		__asm__ __volatile__(LOCK_PREFIX "cmpxchgb %b1,%2"
+				     : "=a"(prev)
+				     : "q"(new), "m"(*__xg(ptr)), "0"(old)
+				     : "memory");
+		return prev;
+	case 2:
+		__asm__ __volatile__(LOCK_PREFIX "cmpxchgw %w1,%2"
+				     : "=a"(prev)
+				     : "r"(new), "m"(*__xg(ptr)), "0"(old)
+				     : "memory");
+		return prev;
+	case 4:
+		__asm__ __volatile__(LOCK_PREFIX "cmpxchgl %k1,%2"
+				     : "=a"(prev)
+				     : "r"(new), "m"(*__xg(ptr)), "0"(old)
+				     : "memory");
+		return prev;
+	case 8:
+		__asm__ __volatile__(LOCK_PREFIX "cmpxchgq %1,%2"
+				     : "=a"(prev)
+				     : "r"(new), "m"(*__xg(ptr)), "0"(old)
+				     : "memory");
+		return prev;
+	}
+	return old;
+}
+
+#define cmpxchg(ptr,o,n)\
+	((__typeof__(*(ptr)))__cmpxchg((ptr),(unsigned long)(o),\
+					(unsigned long)(n),sizeof(*(ptr))))
+
+#ifdef CONFIG_SMP
+#define smp_mb()	mb()
+#define smp_rmb()	rmb()
+#define smp_wmb()	wmb()
+#define smp_read_barrier_depends()	do {} while(0)
+#else
+#define smp_mb()	barrier()
+#define smp_rmb()	barrier()
+#define smp_wmb()	barrier()
+#define smp_read_barrier_depends()	do {} while(0)
+#endif
+
+    
+/*
+ * Force strict CPU ordering.
+ * And yes, this is required on UP too when we're talking
+ * to devices.
+ */
+#define mb() 	asm volatile("mfence":::"memory")
+#define rmb()	asm volatile("lfence":::"memory")
+
+#ifdef CONFIG_UNORDERED_IO
+#define wmb()	asm volatile("sfence" ::: "memory")
+#else
+#define wmb()	asm volatile("" ::: "memory")
+#endif
+#define read_barrier_depends()	do {} while(0)
+#define set_mb(var, value) do { (void) xchg(&var, value); } while (0)
+
+#define warn_if_not_ulong(x) do { unsigned long foo; (void) (&(x) == &foo); } while (0)
+
+#include <linux/irqflags.h>
+
+void cpu_idle_wait(void);
+
+extern unsigned long arch_align_stack(unsigned long sp);
+extern void free_init_pages(char *what, unsigned long begin, unsigned long end);
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/timer.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/timer.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/timer.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/timer.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,67 @@
+#ifndef _ASMi386_TIMER_H
+#define _ASMi386_TIMER_H
+#include <linux/init.h>
+
+/**
+ * struct timer_ops - used to define a timer source
+ *
+ * @name: name of the timer.
+ * @init: Probes and initializes the timer. Takes clock= override 
+ *        string as an argument. Returns 0 on success, anything else
+ *        on failure.
+ * @mark_offset: called by the timer interrupt.
+ * @get_offset:  called by gettimeofday(). Returns the number of microseconds
+ *               since the last timer interupt.
+ * @monotonic_clock: returns the number of nanoseconds since the init of the
+ *                   timer.
+ * @delay: delays this many clock cycles.
+ */
+struct timer_opts {
+	char* name;
+	void (*mark_offset)(void);
+	unsigned long (*get_offset)(void);
+	unsigned long long (*monotonic_clock)(void);
+	void (*delay)(unsigned long);
+	unsigned long (*read_timer)(void);
+	int (*suspend)(pm_message_t state);
+	int (*resume)(void);
+};
+
+struct init_timer_opts {
+	int (*init)(char *override);
+	struct timer_opts *opts;
+};
+
+#define TICK_SIZE (tick_nsec / 1000)
+
+extern struct timer_opts* __init select_timer(void);
+extern void clock_fallback(void);
+void setup_pit_timer(void);
+
+/* Modifiers for buggy PIT handling */
+
+extern int pit_latch_buggy;
+
+extern struct timer_opts *cur_timer;
+extern int timer_ack;
+
+/* list of externed timers */
+extern struct timer_opts timer_none;
+extern struct timer_opts timer_pit;
+extern struct init_timer_opts timer_pit_init;
+extern struct init_timer_opts timer_tsc_init;
+#ifdef CONFIG_X86_CYCLONE_TIMER
+extern struct init_timer_opts timer_cyclone_init;
+#endif
+
+extern unsigned long calibrate_tsc(void);
+extern void init_cpu_khz(void);
+#ifdef CONFIG_HPET_TIMER
+extern struct init_timer_opts timer_hpet_init;
+extern unsigned long calibrate_tsc_hpet(unsigned long *tsc_hpet_quotient_ptr);
+#endif
+
+#ifdef CONFIG_X86_PM_TIMER
+extern struct init_timer_opts timer_pmtmr_init;
+#endif
+#endif
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/tlbflush.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/tlbflush.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/tlbflush.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/tlbflush.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,103 @@
+#ifndef _X8664_TLBFLUSH_H
+#define _X8664_TLBFLUSH_H
+
+#include <linux/mm.h>
+#include <asm/processor.h>
+
+#define __flush_tlb()	xen_tlb_flush()
+
+/*
+ * Global pages have to be flushed a bit differently. Not a real
+ * performance problem because this does not happen often.
+ */
+#define __flush_tlb_global()	xen_tlb_flush()
+
+
+extern unsigned long pgkern_mask;
+
+#define __flush_tlb_all() __flush_tlb_global()
+
+#define __flush_tlb_one(addr)	xen_invlpg((unsigned long)addr)
+
+
+/*
+ * TLB flushing:
+ *
+ *  - flush_tlb() flushes the current mm struct TLBs
+ *  - flush_tlb_all() flushes all processes TLBs
+ *  - flush_tlb_mm(mm) flushes the specified mm context TLB's
+ *  - flush_tlb_page(vma, vmaddr) flushes one page
+ *  - flush_tlb_range(vma, start, end) flushes a range of pages
+ *  - flush_tlb_kernel_range(start, end) flushes a range of kernel pages
+ *  - flush_tlb_pgtables(mm, start, end) flushes a range of page tables
+ *
+ * x86-64 can only flush individual pages or full VMs. For a range flush
+ * we always do the full VM. Might be worth trying if for a small
+ * range a few INVLPGs in a row are a win.
+ */
+
+#ifndef CONFIG_SMP
+
+#define flush_tlb() __flush_tlb()
+#define flush_tlb_all() __flush_tlb_all()
+#define local_flush_tlb() __flush_tlb()
+
+static inline void flush_tlb_mm(struct mm_struct *mm)
+{
+	if (mm == current->active_mm)
+		__flush_tlb();
+}
+
+static inline void flush_tlb_page(struct vm_area_struct *vma,
+	unsigned long addr)
+{
+	if (vma->vm_mm == current->active_mm)
+		__flush_tlb_one(addr);
+}
+
+static inline void flush_tlb_range(struct vm_area_struct *vma,
+	unsigned long start, unsigned long end)
+{
+	if (vma->vm_mm == current->active_mm)
+		__flush_tlb();
+}
+
+#else
+
+#include <asm/smp.h>
+
+#define local_flush_tlb() \
+	__flush_tlb()
+
+#define flush_tlb_all xen_tlb_flush_all
+#define flush_tlb_current_task() xen_tlb_flush_mask(&current->mm->cpu_vm_mask)
+#define flush_tlb_mm(mm) xen_tlb_flush_mask(&(mm)->cpu_vm_mask)
+#define flush_tlb_page(vma, va) xen_invlpg_mask(&(vma)->vm_mm->cpu_vm_mask, va)
+
+#define flush_tlb()	flush_tlb_current_task()
+
+static inline void flush_tlb_range(struct vm_area_struct * vma, unsigned long start, unsigned long end)
+{
+	flush_tlb_mm(vma->vm_mm);
+}
+
+#define TLBSTATE_OK	1
+#define TLBSTATE_LAZY	2
+
+/* Roughly an IPI every 20MB with 4k pages for freeing page table
+   ranges. Cost is about 42k of memory for each CPU. */
+#define ARCH_FREE_PTE_NR 5350	
+
+#endif
+
+#define flush_tlb_kernel_range(start, end) flush_tlb_all()
+
+static inline void flush_tlb_pgtables(struct mm_struct *mm,
+				      unsigned long start, unsigned long end)
+{
+	/* x86_64 does not keep any page table caches in a software TLB.
+	   The CPUs do in their hardware TLBs, but they are handled
+	   by the normal TLB flushing algorithms. */
+}
+
+#endif /* _X8664_TLBFLUSH_H */
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/vga.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/vga.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/vga.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/vga.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,20 @@
+/*
+ *	Access to VGA videoram
+ *
+ *	(c) 1998 Martin Mares <mj@ucw.cz>
+ */
+
+#ifndef _LINUX_ASM_VGA_H_
+#define _LINUX_ASM_VGA_H_
+
+/*
+ *	On the PC, we can just recalculate addresses and then
+ *	access the videoram directly without any black magic.
+ */
+
+#define VGA_MAP_MEM(x,s) (unsigned long)isa_bus_to_virt(x)
+
+#define vga_readb(x) (*(x))
+#define vga_writeb(x,y) (*(y) = (x))
+
+#endif
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/xenoprof.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/xenoprof.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/xenoprof.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/xenoprof.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1 @@
+#include <asm-i386/mach-xen/asm/xenoprof.h>
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/xor.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/xor.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/asm/xor.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/asm/xor.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,328 @@
+/*
+ * x86-64 changes / gcc fixes from Andi Kleen. 
+ * Copyright 2002 Andi Kleen, SuSE Labs.
+ *
+ * This hasn't been optimized for the hammer yet, but there are likely
+ * no advantages to be gotten from x86-64 here anyways.
+ */
+
+typedef struct { unsigned long a,b; } __attribute__((aligned(16))) xmm_store_t;
+
+/* Doesn't use gcc to save the XMM registers, because there is no easy way to 
+   tell it to do a clts before the register saving. */
+#define XMMS_SAVE do {				\
+	preempt_disable();			\
+	if (!(current_thread_info()->status & TS_USEDFPU))	\
+		clts();				\
+	__asm__ __volatile__ ( 			\
+		"movups %%xmm0,(%1)	;\n\t"	\
+		"movups %%xmm1,0x10(%1)	;\n\t"	\
+		"movups %%xmm2,0x20(%1)	;\n\t"	\
+		"movups %%xmm3,0x30(%1)	;\n\t"	\
+		: "=&r" (cr0)			\
+		: "r" (xmm_save) 		\
+		: "memory");			\
+} while(0)
+
+#define XMMS_RESTORE do {			\
+	asm volatile (				\
+		"sfence			;\n\t"	\
+		"movups (%1),%%xmm0	;\n\t"	\
+		"movups 0x10(%1),%%xmm1	;\n\t"	\
+		"movups 0x20(%1),%%xmm2	;\n\t"	\
+		"movups 0x30(%1),%%xmm3	;\n\t"	\
+		:				\
+		: "r" (cr0), "r" (xmm_save)	\
+		: "memory");			\
+	if (!(current_thread_info()->status & TS_USEDFPU))	\
+		stts();				\
+	preempt_enable();			\
+} while(0)
+
+#define OFFS(x)		"16*("#x")"
+#define PF_OFFS(x)	"256+16*("#x")"
+#define	PF0(x)		"	prefetchnta "PF_OFFS(x)"(%[p1])		;\n"
+#define LD(x,y)		"       movaps   "OFFS(x)"(%[p1]), %%xmm"#y"	;\n"
+#define ST(x,y)		"       movaps %%xmm"#y",   "OFFS(x)"(%[p1])	;\n"
+#define PF1(x)		"	prefetchnta "PF_OFFS(x)"(%[p2])		;\n"
+#define PF2(x)		"	prefetchnta "PF_OFFS(x)"(%[p3])		;\n"
+#define PF3(x)		"	prefetchnta "PF_OFFS(x)"(%[p4])		;\n"
+#define PF4(x)		"	prefetchnta "PF_OFFS(x)"(%[p5])		;\n"
+#define PF5(x)		"	prefetchnta "PF_OFFS(x)"(%[p6])		;\n"
+#define XO1(x,y)	"       xorps   "OFFS(x)"(%[p2]), %%xmm"#y"	;\n"
+#define XO2(x,y)	"       xorps   "OFFS(x)"(%[p3]), %%xmm"#y"	;\n"
+#define XO3(x,y)	"       xorps   "OFFS(x)"(%[p4]), %%xmm"#y"	;\n"
+#define XO4(x,y)	"       xorps   "OFFS(x)"(%[p5]), %%xmm"#y"	;\n"
+#define XO5(x,y)	"       xorps   "OFFS(x)"(%[p6]), %%xmm"#y"	;\n"
+
+
+static void
+xor_sse_2(unsigned long bytes, unsigned long *p1, unsigned long *p2)
+{
+        unsigned int lines = bytes >> 8;
+	unsigned long cr0;
+	xmm_store_t xmm_save[4];
+
+	XMMS_SAVE;
+
+        asm volatile (
+#undef BLOCK
+#define BLOCK(i) \
+		LD(i,0)					\
+			LD(i+1,1)			\
+		PF1(i)					\
+				PF1(i+2)		\
+				LD(i+2,2)		\
+					LD(i+3,3)	\
+		PF0(i+4)				\
+				PF0(i+6)		\
+		XO1(i,0)				\
+			XO1(i+1,1)			\
+				XO1(i+2,2)		\
+					XO1(i+3,3)	\
+		ST(i,0)					\
+			ST(i+1,1)			\
+				ST(i+2,2)		\
+					ST(i+3,3)	\
+
+
+		PF0(0)
+				PF0(2)
+
+	" .align 32			;\n"
+        " 1:                            ;\n"
+
+		BLOCK(0)
+		BLOCK(4)
+		BLOCK(8)
+		BLOCK(12)
+
+        "       addq %[inc], %[p1]           ;\n"
+        "       addq %[inc], %[p2]           ;\n"
+		"		decl %[cnt] ; jnz 1b"
+	: [p1] "+r" (p1), [p2] "+r" (p2), [cnt] "+r" (lines)
+	: [inc] "r" (256UL) 
+        : "memory");
+
+	XMMS_RESTORE;
+}
+
+static void
+xor_sse_3(unsigned long bytes, unsigned long *p1, unsigned long *p2,
+	  unsigned long *p3)
+{
+	unsigned int lines = bytes >> 8;
+	xmm_store_t xmm_save[4];
+	unsigned long cr0;
+
+	XMMS_SAVE;
+
+        __asm__ __volatile__ (
+#undef BLOCK
+#define BLOCK(i) \
+		PF1(i)					\
+				PF1(i+2)		\
+		LD(i,0)					\
+			LD(i+1,1)			\
+				LD(i+2,2)		\
+					LD(i+3,3)	\
+		PF2(i)					\
+				PF2(i+2)		\
+		PF0(i+4)				\
+				PF0(i+6)		\
+		XO1(i,0)				\
+			XO1(i+1,1)			\
+				XO1(i+2,2)		\
+					XO1(i+3,3)	\
+		XO2(i,0)				\
+			XO2(i+1,1)			\
+				XO2(i+2,2)		\
+					XO2(i+3,3)	\
+		ST(i,0)					\
+			ST(i+1,1)			\
+				ST(i+2,2)		\
+					ST(i+3,3)	\
+
+
+		PF0(0)
+				PF0(2)
+
+	" .align 32			;\n"
+        " 1:                            ;\n"
+
+		BLOCK(0)
+		BLOCK(4)
+		BLOCK(8)
+		BLOCK(12)
+
+        "       addq %[inc], %[p1]           ;\n"
+        "       addq %[inc], %[p2]          ;\n"
+        "       addq %[inc], %[p3]           ;\n"
+		"		decl %[cnt] ; jnz 1b"
+	: [cnt] "+r" (lines),
+	  [p1] "+r" (p1), [p2] "+r" (p2), [p3] "+r" (p3)
+	: [inc] "r" (256UL)
+	: "memory"); 
+	XMMS_RESTORE;
+}
+
+static void
+xor_sse_4(unsigned long bytes, unsigned long *p1, unsigned long *p2,
+	  unsigned long *p3, unsigned long *p4)
+{
+	unsigned int lines = bytes >> 8;
+	xmm_store_t xmm_save[4]; 
+	unsigned long cr0;
+
+	XMMS_SAVE;
+
+        __asm__ __volatile__ (
+#undef BLOCK
+#define BLOCK(i) \
+		PF1(i)					\
+				PF1(i+2)		\
+		LD(i,0)					\
+			LD(i+1,1)			\
+				LD(i+2,2)		\
+					LD(i+3,3)	\
+		PF2(i)					\
+				PF2(i+2)		\
+		XO1(i,0)				\
+			XO1(i+1,1)			\
+				XO1(i+2,2)		\
+					XO1(i+3,3)	\
+		PF3(i)					\
+				PF3(i+2)		\
+		PF0(i+4)				\
+				PF0(i+6)		\
+		XO2(i,0)				\
+			XO2(i+1,1)			\
+				XO2(i+2,2)		\
+					XO2(i+3,3)	\
+		XO3(i,0)				\
+			XO3(i+1,1)			\
+				XO3(i+2,2)		\
+					XO3(i+3,3)	\
+		ST(i,0)					\
+			ST(i+1,1)			\
+				ST(i+2,2)		\
+					ST(i+3,3)	\
+
+
+		PF0(0)
+				PF0(2)
+
+	" .align 32			;\n"
+        " 1:                            ;\n"
+
+		BLOCK(0)
+		BLOCK(4)
+		BLOCK(8)
+		BLOCK(12)
+
+        "       addq %[inc], %[p1]           ;\n"
+        "       addq %[inc], %[p2]           ;\n"
+        "       addq %[inc], %[p3]           ;\n"
+        "       addq %[inc], %[p4]           ;\n"
+	"	decl %[cnt] ; jnz 1b"
+	: [cnt] "+c" (lines),
+	  [p1] "+r" (p1), [p2] "+r" (p2), [p3] "+r" (p3), [p4] "+r" (p4)
+	: [inc] "r" (256UL)
+        : "memory" );
+
+	XMMS_RESTORE;
+}
+
+static void
+xor_sse_5(unsigned long bytes, unsigned long *p1, unsigned long *p2,
+	  unsigned long *p3, unsigned long *p4, unsigned long *p5)
+{
+        unsigned int lines = bytes >> 8;
+	xmm_store_t xmm_save[4];
+	unsigned long cr0;
+
+	XMMS_SAVE;
+
+        __asm__ __volatile__ (
+#undef BLOCK
+#define BLOCK(i) \
+		PF1(i)					\
+				PF1(i+2)		\
+		LD(i,0)					\
+			LD(i+1,1)			\
+				LD(i+2,2)		\
+					LD(i+3,3)	\
+		PF2(i)					\
+				PF2(i+2)		\
+		XO1(i,0)				\
+			XO1(i+1,1)			\
+				XO1(i+2,2)		\
+					XO1(i+3,3)	\
+		PF3(i)					\
+				PF3(i+2)		\
+		XO2(i,0)				\
+			XO2(i+1,1)			\
+				XO2(i+2,2)		\
+					XO2(i+3,3)	\
+		PF4(i)					\
+				PF4(i+2)		\
+		PF0(i+4)				\
+				PF0(i+6)		\
+		XO3(i,0)				\
+			XO3(i+1,1)			\
+				XO3(i+2,2)		\
+					XO3(i+3,3)	\
+		XO4(i,0)				\
+			XO4(i+1,1)			\
+				XO4(i+2,2)		\
+					XO4(i+3,3)	\
+		ST(i,0)					\
+			ST(i+1,1)			\
+				ST(i+2,2)		\
+					ST(i+3,3)	\
+
+
+		PF0(0)
+				PF0(2)
+
+	" .align 32			;\n"
+        " 1:                            ;\n"
+
+		BLOCK(0)
+		BLOCK(4)
+		BLOCK(8)
+		BLOCK(12)
+
+        "       addq %[inc], %[p1]           ;\n"
+        "       addq %[inc], %[p2]           ;\n"
+        "       addq %[inc], %[p3]           ;\n"
+        "       addq %[inc], %[p4]           ;\n"
+        "       addq %[inc], %[p5]           ;\n"
+	"	decl %[cnt] ; jnz 1b"
+	: [cnt] "+c" (lines),
+  	  [p1] "+r" (p1), [p2] "+r" (p2), [p3] "+r" (p3), [p4] "+r" (p4), 
+	  [p5] "+r" (p5)
+	: [inc] "r" (256UL)
+	: "memory");
+
+	XMMS_RESTORE;
+}
+
+static struct xor_block_template xor_block_sse = {
+        .name = "generic_sse",
+        .do_2 = xor_sse_2,
+        .do_3 = xor_sse_3,
+        .do_4 = xor_sse_4,
+        .do_5 = xor_sse_5,
+};
+
+#undef XOR_TRY_TEMPLATES
+#define XOR_TRY_TEMPLATES				\
+	do {						\
+		xor_speed(&xor_block_sse);	\
+	} while (0)
+
+/* We force the use of the SSE xor block because it can write around L2.
+   We may also be able to load into the L1 only depending on how the cpu
+   deals with a load to a line that is being prefetched.  */
+#define XOR_SELECT_TEMPLATE(FASTEST) (&xor_block_sse)
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/irq_vectors.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/irq_vectors.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/irq_vectors.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/irq_vectors.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,123 @@
+/*
+ * This file should contain #defines for all of the interrupt vector
+ * numbers used by this architecture.
+ *
+ * In addition, there are some standard defines:
+ *
+ *	FIRST_EXTERNAL_VECTOR:
+ *		The first free place for external interrupts
+ *
+ *	SYSCALL_VECTOR:
+ *		The IRQ vector a syscall makes the user to kernel transition
+ *		under.
+ *
+ *	TIMER_IRQ:
+ *		The IRQ number the timer interrupt comes in at.
+ *
+ *	NR_IRQS:
+ *		The total number of interrupt vectors (including all the
+ *		architecture specific interrupts) needed.
+ *
+ */			
+#ifndef _ASM_IRQ_VECTORS_H
+#define _ASM_IRQ_VECTORS_H
+
+/*
+ * IDT vectors usable for external interrupt sources start
+ * at 0x20:
+ */
+#define FIRST_EXTERNAL_VECTOR	0x20
+
+#define SYSCALL_VECTOR		0x80
+
+/*
+ * Vectors 0x20-0x2f are used for ISA interrupts.
+ */
+
+#if 0
+/*
+ * Special IRQ vectors used by the SMP architecture, 0xf0-0xff
+ *
+ *  some of the following vectors are 'rare', they are merged
+ *  into a single vector (CALL_FUNCTION_VECTOR) to save vector space.
+ *  TLB, reschedule and local APIC vectors are performance-critical.
+ *
+ *  Vectors 0xf0-0xfa are free (reserved for future Linux use).
+ */
+#define INVALIDATE_TLB_VECTOR	0xfd
+#define RESCHEDULE_VECTOR	0xfc
+#define CALL_FUNCTION_VECTOR	0xfb
+
+#define THERMAL_APIC_VECTOR	0xf0
+/*
+ * Local APIC timer IRQ vector is on a different priority level,
+ * to work around the 'lost local interrupt if more than 2 IRQ
+ * sources per level' errata.
+ */
+#define LOCAL_TIMER_VECTOR	0xef
+#endif
+
+#define SPURIOUS_APIC_VECTOR	0xff
+#define ERROR_APIC_VECTOR	0xfe
+
+/*
+ * First APIC vector available to drivers: (vectors 0x30-0xee)
+ * we start at 0x31 to spread out vectors evenly between priority
+ * levels. (0x80 is the syscall vector)
+ */
+#define FIRST_DEVICE_VECTOR	0x31
+#define FIRST_SYSTEM_VECTOR	0xef
+
+/*
+ * 16 8259A IRQ's, 208 potential APIC interrupt sources.
+ * Right now the APIC is mostly only used for SMP.
+ * 256 vectors is an architectural limit. (we can have
+ * more than 256 devices theoretically, but they will
+ * have to use shared interrupts)
+ * Since vectors 0x00-0x1f are used/reserved for the CPU,
+ * the usable vector space is 0x20-0xff (224 vectors)
+ */
+
+#define RESCHEDULE_VECTOR	0
+#define CALL_FUNCTION_VECTOR	1
+#define NR_IPIS			2
+
+/*
+ * The maximum number of vectors supported by i386 processors
+ * is limited to 256. For processors other than i386, NR_VECTORS
+ * should be changed accordingly.
+ */
+#define NR_VECTORS 256
+
+#define FPU_IRQ			13
+
+#define	FIRST_VM86_IRQ		3
+#define LAST_VM86_IRQ		15
+#define invalid_vm86_irq(irq)	((irq) < 3 || (irq) > 15)
+
+/*
+ * The flat IRQ space is divided into two regions:
+ *  1. A one-to-one mapping of real physical IRQs. This space is only used
+ *     if we have physical device-access privilege. This region is at the 
+ *     start of the IRQ space so that existing device drivers do not need
+ *     to be modified to translate physical IRQ numbers into our IRQ space.
+ *  3. A dynamic mapping of inter-domain and Xen-sourced virtual IRQs. These
+ *     are bound using the provided bind/unbind functions.
+ */
+
+#define PIRQ_BASE		0
+#define NR_PIRQS		256
+
+#define DYNIRQ_BASE		(PIRQ_BASE + NR_PIRQS)
+#define NR_DYNIRQS		256
+
+#define NR_IRQS			(NR_PIRQS + NR_DYNIRQS)
+#define NR_IRQ_VECTORS		NR_IRQS
+
+#define pirq_to_irq(_x)		((_x) + PIRQ_BASE)
+#define irq_to_pirq(_x)		((_x) - PIRQ_BASE)
+
+#define dynirq_to_irq(_x)	((_x) + DYNIRQ_BASE)
+#define irq_to_dynirq(_x)	((_x) - DYNIRQ_BASE)
+
+#endif /* _ASM_IRQ_VECTORS_H */
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/mach_time.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/mach_time.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/mach_time.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/mach_time.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,111 @@
+/*
+ *  include/asm-i386/mach-default/mach_time.h
+ *
+ *  Machine specific set RTC function for generic.
+ *  Split out from time.c by Osamu Tomita <tomita@cinet.co.jp>
+ */
+#ifndef _MACH_TIME_H
+#define _MACH_TIME_H
+
+#include <asm-i386/mc146818rtc.h>
+
+/* for check timing call set_rtc_mmss() 500ms     */
+/* used in arch/i386/time.c::do_timer_interrupt() */
+#define USEC_AFTER	500000
+#define USEC_BEFORE	500000
+
+/*
+ * In order to set the CMOS clock precisely, set_rtc_mmss has to be
+ * called 500 ms after the second nowtime has started, because when
+ * nowtime is written into the registers of the CMOS clock, it will
+ * jump to the next second precisely 500 ms later. Check the Motorola
+ * MC146818A or Dallas DS12887 data sheet for details.
+ *
+ * BUG: This routine does not handle hour overflow properly; it just
+ *      sets the minutes. Usually you'll only notice that after reboot!
+ */
+static inline int mach_set_rtc_mmss(unsigned long nowtime)
+{
+	int retval = 0;
+	int real_seconds, real_minutes, cmos_minutes;
+	unsigned char save_control, save_freq_select;
+
+	save_control = CMOS_READ(RTC_CONTROL); /* tell the clock it's being set */
+	CMOS_WRITE((save_control|RTC_SET), RTC_CONTROL);
+
+	save_freq_select = CMOS_READ(RTC_FREQ_SELECT); /* stop and reset prescaler */
+	CMOS_WRITE((save_freq_select|RTC_DIV_RESET2), RTC_FREQ_SELECT);
+
+	cmos_minutes = CMOS_READ(RTC_MINUTES);
+	if (!(save_control & RTC_DM_BINARY) || RTC_ALWAYS_BCD)
+		BCD_TO_BIN(cmos_minutes);
+
+	/*
+	 * since we're only adjusting minutes and seconds,
+	 * don't interfere with hour overflow. This avoids
+	 * messing with unknown time zones but requires your
+	 * RTC not to be off by more than 15 minutes
+	 */
+	real_seconds = nowtime % 60;
+	real_minutes = nowtime / 60;
+	if (((abs(real_minutes - cmos_minutes) + 15)/30) & 1)
+		real_minutes += 30;		/* correct for half hour time zone */
+	real_minutes %= 60;
+
+	if (abs(real_minutes - cmos_minutes) < 30) {
+		if (!(save_control & RTC_DM_BINARY) || RTC_ALWAYS_BCD) {
+			BIN_TO_BCD(real_seconds);
+			BIN_TO_BCD(real_minutes);
+		}
+		CMOS_WRITE(real_seconds,RTC_SECONDS);
+		CMOS_WRITE(real_minutes,RTC_MINUTES);
+	} else {
+		printk(KERN_WARNING
+		       "set_rtc_mmss: can't update from %d to %d\n",
+		       cmos_minutes, real_minutes);
+		retval = -1;
+	}
+
+	/* The following flags have to be released exactly in this order,
+	 * otherwise the DS12887 (popular MC146818A clone with integrated
+	 * battery and quartz) will not reset the oscillator and will not
+	 * update precisely 500 ms later. You won't find this mentioned in
+	 * the Dallas Semiconductor data sheets, but who believes data
+	 * sheets anyway ...                           -- Markus Kuhn
+	 */
+	CMOS_WRITE(save_control, RTC_CONTROL);
+	CMOS_WRITE(save_freq_select, RTC_FREQ_SELECT);
+
+	return retval;
+}
+
+static inline unsigned long mach_get_cmos_time(void)
+{
+	unsigned int year, mon, day, hour, min, sec;
+
+	do {
+		sec = CMOS_READ(RTC_SECONDS);
+		min = CMOS_READ(RTC_MINUTES);
+		hour = CMOS_READ(RTC_HOURS);
+		day = CMOS_READ(RTC_DAY_OF_MONTH);
+		mon = CMOS_READ(RTC_MONTH);
+		year = CMOS_READ(RTC_YEAR);
+	} while (sec != CMOS_READ(RTC_SECONDS));
+
+	if (!(CMOS_READ(RTC_CONTROL) & RTC_DM_BINARY) || RTC_ALWAYS_BCD) {
+		BCD_TO_BIN(sec);
+		BCD_TO_BIN(min);
+		BCD_TO_BIN(hour);
+		BCD_TO_BIN(day);
+		BCD_TO_BIN(mon);
+		BCD_TO_BIN(year);
+	}
+
+	year += 1900;
+	if (year < 1970)
+		year += 100;
+
+	return mktime(year, mon, day, hour, min, sec);
+}
+
+#endif /* !_MACH_TIME_H */
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/mach_timer.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/mach_timer.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/mach_timer.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/mach_timer.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,50 @@
+/*
+ *  include/asm-i386/mach-default/mach_timer.h
+ *
+ *  Machine specific calibrate_tsc() for generic.
+ *  Split out from timer_tsc.c by Osamu Tomita <tomita@cinet.co.jp>
+ */
+/* ------ Calibrate the TSC ------- 
+ * Return 2^32 * (1 / (TSC clocks per usec)) for do_fast_gettimeoffset().
+ * Too much 64-bit arithmetic here to do this cleanly in C, and for
+ * accuracy's sake we want to keep the overhead on the CTC speaker (channel 2)
+ * output busy loop as low as possible. We avoid reading the CTC registers
+ * directly because of the awkward 8-bit access mechanism of the 82C54
+ * device.
+ */
+#ifndef _MACH_TIMER_H
+#define _MACH_TIMER_H
+
+#define CALIBRATE_TIME_MSEC 30 /* 30 msecs */
+#define CALIBRATE_LATCH	\
+	((CLOCK_TICK_RATE * CALIBRATE_TIME_MSEC + 1000/2)/1000)
+
+static inline void mach_prepare_counter(void)
+{
+       /* Set the Gate high, disable speaker */
+	outb((inb(0x61) & ~0x02) | 0x01, 0x61);
+
+	/*
+	 * Now let's take care of CTC channel 2
+	 *
+	 * Set the Gate high, program CTC channel 2 for mode 0,
+	 * (interrupt on terminal count mode), binary count,
+	 * load 5 * LATCH count, (LSB and MSB) to begin countdown.
+	 *
+	 * Some devices need a delay here.
+	 */
+	outb(0xb0, 0x43);			/* binary, mode 0, LSB/MSB, Ch 2 */
+	outb_p(CALIBRATE_LATCH & 0xff, 0x42);	/* LSB of count */
+	outb_p(CALIBRATE_LATCH >> 8, 0x42);       /* MSB of count */
+}
+
+static inline void mach_countup(unsigned long *count_p)
+{
+	unsigned long count = 0;
+	do {
+		count++;
+	} while ((inb_p(0x61) & 0x20) == 0);
+	*count_p = count;
+}
+
+#endif /* !_MACH_TIMER_H */
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/setup_arch_post.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/setup_arch_post.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/setup_arch_post.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/setup_arch_post.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,63 @@
+/**
+ * machine_specific_* - Hooks for machine specific setup.
+ *
+ * Description:
+ *	This is included late in kernel/setup.c so that it can make
+ *	use of all of the static functions.
+ **/
+
+#include <xen/interface/callback.h>
+
+extern void hypervisor_callback(void);
+extern void failsafe_callback(void);
+extern void nmi(void);
+
+static void __init machine_specific_arch_setup(void)
+{
+	int ret;
+	static struct callback_register __initdata event = {
+		.type = CALLBACKTYPE_event,
+		.address = (unsigned long) hypervisor_callback,
+	};
+	static struct callback_register __initdata failsafe = {
+		.type = CALLBACKTYPE_failsafe,
+		.address = (unsigned long)failsafe_callback,
+	};
+	static struct callback_register __initdata syscall = {
+		.type = CALLBACKTYPE_syscall,
+		.address = (unsigned long)system_call,
+	};
+#ifdef CONFIG_X86_LOCAL_APIC
+	static struct callback_register __initdata nmi_cb = {
+		.type = CALLBACKTYPE_nmi,
+		.address = (unsigned long)nmi,
+	};
+#endif
+
+	ret = HYPERVISOR_callback_op(CALLBACKOP_register, &event);
+	if (ret == 0)
+		ret = HYPERVISOR_callback_op(CALLBACKOP_register, &failsafe);
+	if (ret == 0)
+		ret = HYPERVISOR_callback_op(CALLBACKOP_register, &syscall);
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (ret == -ENOSYS)
+		ret = HYPERVISOR_set_callbacks(
+			event.address,
+			failsafe.address,
+			syscall.address);
+#endif
+	BUG_ON(ret);
+
+#ifdef CONFIG_X86_LOCAL_APIC
+	ret = HYPERVISOR_callback_op(CALLBACKOP_register, &nmi_cb);
+#if CONFIG_XEN_COMPAT <= 0x030002
+	if (ret == -ENOSYS) {
+		static struct xennmi_callback __initdata cb = {
+			.handler_address = (unsigned long)nmi
+		};
+
+		HYPERVISOR_nmi_op(XENNMI_register_callback, &cb);
+	}
+#endif
+#endif
+}
diff -rpuN linux-2.6.18.8/include/asm-x86_64/mach-xen/setup_arch_pre.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/setup_arch_pre.h
--- linux-2.6.18.8/include/asm-x86_64/mach-xen/setup_arch_pre.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/mach-xen/setup_arch_pre.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,5 @@
+/* Hook to call BIOS initialisation function */
+
+#define ARCH_SETUP machine_specific_arch_setup();
+
+static void __init machine_specific_arch_setup(void);
diff -rpuN linux-2.6.18.8/include/asm-x86_64/msi.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/msi.h
--- linux-2.6.18.8/include/asm-x86_64/msi.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/msi.h	2008-08-21 11:36:07.000000000 +0200
@@ -7,14 +7,21 @@
 #define ASM_MSI_H
 
 #include <asm/desc.h>
+#ifndef CONFIG_XEN
 #include <asm/mach_apic.h>
+#endif
 #include <asm/smp.h>
 
+#ifndef CONFIG_XEN
 #define LAST_DEVICE_VECTOR	(FIRST_SYSTEM_VECTOR - 1)
+#else
+#define LAST_DYNAMIC_VECTOR 0xdf
+#define LAST_DEVICE_VECTOR	(LAST_DYNAMIC_VECTOR)
+#endif
+
 #define MSI_TARGET_CPU_SHIFT	12
 
 extern struct msi_ops msi_apic_ops;
-
 static inline int msi_arch_init(void)
 {
 	msi_register(&msi_apic_ops);
diff -rpuN linux-2.6.18.8/include/asm-x86_64/signal.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/signal.h
--- linux-2.6.18.8/include/asm-x86_64/signal.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/signal.h	2008-08-21 11:36:07.000000000 +0200
@@ -24,10 +24,6 @@ typedef struct {
 } sigset_t;
 
 
-struct pt_regs; 
-asmlinkage int do_signal(struct pt_regs *regs, sigset_t *oldset);
-
-
 #else
 /* Here we must cater to libcs that poke about in kernel headers.  */
 
diff -rpuN linux-2.6.18.8/include/asm-x86_64/thread_info.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/thread_info.h
--- linux-2.6.18.8/include/asm-x86_64/thread_info.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/thread_info.h	2008-08-21 11:36:07.000000000 +0200
@@ -114,6 +114,7 @@ static inline struct thread_info *stack_
 #define TIF_IRET		5	/* force IRET */
 #define TIF_SYSCALL_AUDIT	7	/* syscall auditing active */
 #define TIF_SECCOMP		8	/* secure computing */
+#define TIF_RESTORE_SIGMASK	9	/* restore signal mask in do_signal */
 /* 16 free */
 #define TIF_IA32		17	/* 32bit process */ 
 #define TIF_FORK		18	/* ret_from_fork */
@@ -128,6 +129,7 @@ static inline struct thread_info *stack_
 #define _TIF_IRET		(1<<TIF_IRET)
 #define _TIF_SYSCALL_AUDIT	(1<<TIF_SYSCALL_AUDIT)
 #define _TIF_SECCOMP		(1<<TIF_SECCOMP)
+#define _TIF_RESTORE_SIGMASK	(1<<TIF_RESTORE_SIGMASK)
 #define _TIF_IA32		(1<<TIF_IA32)
 #define _TIF_FORK		(1<<TIF_FORK)
 #define _TIF_ABI_PENDING	(1<<TIF_ABI_PENDING)
diff -rpuN linux-2.6.18.8/include/asm-x86_64/unistd.h linux-2.6.18-xen-3.3.0/include/asm-x86_64/unistd.h
--- linux-2.6.18.8/include/asm-x86_64/unistd.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/asm-x86_64/unistd.h	2008-08-21 11:36:07.000000000 +0200
@@ -600,9 +600,9 @@ __SYSCALL(__NR_fchmodat, sys_fchmodat)
 #define __NR_faccessat		269
 __SYSCALL(__NR_faccessat, sys_faccessat)
 #define __NR_pselect6		270
-__SYSCALL(__NR_pselect6, sys_ni_syscall)	/* for now */
+__SYSCALL(__NR_pselect6, sys_pselect6)
 #define __NR_ppoll		271
-__SYSCALL(__NR_ppoll,	sys_ni_syscall)		/* for now */
+__SYSCALL(__NR_ppoll,	sys_ppoll)
 #define __NR_unshare		272
 __SYSCALL(__NR_unshare,	sys_unshare)
 #define __NR_set_robust_list	273
@@ -658,6 +658,7 @@ do { \
 #define __ARCH_WANT_SYS_SIGPENDING
 #define __ARCH_WANT_SYS_SIGPROCMASK
 #define __ARCH_WANT_SYS_RT_SIGACTION
+#define __ARCH_WANT_SYS_RT_SIGSUSPEND
 #define __ARCH_WANT_SYS_TIME
 #define __ARCH_WANT_COMPAT_SYS_TIME
 
diff -rpuN linux-2.6.18.8/include/linux/aio.h linux-2.6.18-xen-3.3.0/include/linux/aio.h
--- linux-2.6.18.8/include/linux/aio.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/aio.h	2008-08-21 11:36:07.000000000 +0200
@@ -191,6 +191,11 @@ struct kioctx {
 	struct aio_ring_info	ring_info;
 
 	struct work_struct	wq;
+#ifdef CONFIG_EPOLL
+	// poll integration
+	wait_queue_head_t       poll_wait;
+	struct file		*file;
+#endif
 };
 
 /* prototypes */
diff -rpuN linux-2.6.18.8/include/linux/bio.h linux-2.6.18-xen-3.3.0/include/linux/bio.h
--- linux-2.6.18.8/include/linux/bio.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/bio.h	2008-08-21 11:36:07.000000000 +0200
@@ -172,12 +172,27 @@ struct bio {
 #define bio_offset(bio)		bio_iovec((bio))->bv_offset
 #define bio_segments(bio)	((bio)->bi_vcnt - (bio)->bi_idx)
 #define bio_sectors(bio)	((bio)->bi_size >> 9)
-#define bio_cur_sectors(bio)	(bio_iovec(bio)->bv_len >> 9)
-#define bio_data(bio)		(page_address(bio_page((bio))) + bio_offset((bio)))
 #define bio_barrier(bio)	((bio)->bi_rw & (1 << BIO_RW_BARRIER))
 #define bio_sync(bio)		((bio)->bi_rw & (1 << BIO_RW_SYNC))
 #define bio_failfast(bio)	((bio)->bi_rw & (1 << BIO_RW_FAILFAST))
 #define bio_rw_ahead(bio)	((bio)->bi_rw & (1 << BIO_RW_AHEAD))
+#define bio_empty_barrier(bio)	(bio_barrier(bio) && !(bio)->bi_size)
+
+static inline unsigned int bio_cur_sectors(struct bio *bio)
+{
+	if (bio->bi_vcnt)
+		return bio_iovec(bio)->bv_len >> 9;
+
+	return 0;
+}
+
+static inline void *bio_data(struct bio *bio)
+{
+	if (bio->bi_vcnt)
+		return page_address(bio_page(bio)) + bio_offset(bio);
+
+	return NULL;
+}
 
 /*
  * will die
diff -rpuN linux-2.6.18.8/include/linux/blkdev.h linux-2.6.18-xen-3.3.0/include/linux/blkdev.h
--- linux-2.6.18.8/include/linux/blkdev.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/blkdev.h	2008-08-21 11:36:07.000000000 +0200
@@ -506,6 +506,8 @@ enum {
 #define blk_barrier_rq(rq)	((rq)->flags & REQ_HARDBARRIER)
 #define blk_fua_rq(rq)		((rq)->flags & REQ_FUA)
 
+#define blk_empty_barrier(rq)   (blk_barrier_rq(rq) && blk_fs_request(rq) && !(rq)->hard_nr_sectors)
+
 #define list_entry_rq(ptr)	list_entry((ptr), struct request, queuelist)
 
 #define rq_data_dir(rq)		((rq)->flags & 1)
diff -rpuN linux-2.6.18.8/include/linux/crash_dump.h linux-2.6.18-xen-3.3.0/include/linux/crash_dump.h
--- linux-2.6.18.8/include/linux/crash_dump.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/crash_dump.h	2008-08-21 11:36:07.000000000 +0200
@@ -14,5 +14,13 @@ extern ssize_t copy_oldmem_page(unsigned
 extern const struct file_operations proc_vmcore_operations;
 extern struct proc_dir_entry *proc_vmcore;
 
+/* Architecture code defines this if there are other possible ELF
+ * machine types, e.g. on bi-arch capable hardware. */
+#ifndef vmcore_elf_check_arch_cross
+#define vmcore_elf_check_arch_cross(x) 0
+#endif
+
+#define vmcore_elf_check_arch(x) (elf_check_arch(x) || vmcore_elf_check_arch_cross(x))
+
 #endif /* CONFIG_CRASH_DUMP */
 #endif /* LINUX_CRASHDUMP_H */
diff -rpuN linux-2.6.18.8/include/linux/debugfs.h linux-2.6.18-xen-3.3.0/include/linux/debugfs.h
--- linux-2.6.18.8/include/linux/debugfs.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/debugfs.h	2008-08-21 11:36:07.000000000 +0200
@@ -41,6 +41,8 @@ struct dentry *debugfs_create_u16(const 
 				  struct dentry *parent, u16 *value);
 struct dentry *debugfs_create_u32(const char *name, mode_t mode,
 				  struct dentry *parent, u32 *value);
+struct dentry *debugfs_create_u64(const char *name, mode_t mode,
+				  struct dentry *parent, u64 *value);
 struct dentry *debugfs_create_bool(const char *name, mode_t mode,
 				  struct dentry *parent, u32 *value);
 
@@ -94,6 +96,13 @@ static inline struct dentry *debugfs_cre
 	return ERR_PTR(-ENODEV);
 }
 
+static inline struct dentry *debugfs_create_u64(const char *name, mode_t mode,
+						struct dentry *parent,
+						u64 *value)
+{
+	return ERR_PTR(-ENODEV);
+}
+
 static inline struct dentry *debugfs_create_bool(const char *name, mode_t mode,
 						 struct dentry *parent,
 						 u32 *value)
diff -rpuN linux-2.6.18.8/include/linux/elfnote.h linux-2.6.18-xen-3.3.0/include/linux/elfnote.h
--- linux-2.6.18.8/include/linux/elfnote.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/elfnote.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,104 @@
+#ifndef _LINUX_ELFNOTE_H
+#define _LINUX_ELFNOTE_H
+/*
+ * Helper macros to generate ELF Note structures, which are put into a
+ * PT_NOTE segment of the final vmlinux image.  These are useful for
+ * including name-value pairs of metadata into the kernel binary (or
+ * modules?) for use by external programs.
+ *
+ * Each note has three parts: a name, a type and a desc.  The name is
+ * intended to distinguish the note's originator, so it would be a
+ * company, project, subsystem, etc; it must be in a suitable form for
+ * use in a section name.  The type is an integer which is used to tag
+ * the data, and is considered to be within the "name" namespace (so
+ * "FooCo"'s type 42 is distinct from "BarProj"'s type 42).  The
+ * "desc" field is the actual data.  There are no constraints on the
+ * desc field's contents, though typically they're fairly small.
+ *
+ * All notes from a given NAME are put into a section named
+ * .note.NAME.  When the kernel image is finally linked, all the notes
+ * are packed into a single .notes section, which is mapped into the
+ * PT_NOTE segment.  Because notes for a given name are grouped into
+ * the same section, they'll all be adjacent the output file.
+ *
+ * This file defines macros for both C and assembler use.  Their
+ * syntax is slightly different, but they're semantically similar.
+ *
+ * See the ELF specification for more detail about ELF notes.
+ */
+
+#ifdef __ASSEMBLER__
+/*
+ * Generate a structure with the same shape as Elf{32,64}_Nhdr (which
+ * turn out to be the same size and shape), followed by the name and
+ * desc data with appropriate padding.  The 'desctype' argument is the
+ * assembler pseudo op defining the type of the data e.g. .asciz while
+ * 'descdata' is the data itself e.g.  "hello, world".
+ *
+ * e.g. ELFNOTE(XYZCo, 42, .asciz, "forty-two")
+ *      ELFNOTE(XYZCo, 12, .long, 0xdeadbeef)
+ */
+#ifdef __STDC__
+#define ELFNOTE(name, type, desctype, descdata...) \
+.pushsection .note.name			;	\
+  .align 4				;	\
+  .long 2f - 1f		/* namesz */	;	\
+  .long 4f - 3f		/* descsz */	;	\
+  .long type				;	\
+1:.asciz #name				;	\
+2:.align 4				;	\
+3:desctype descdata			;	\
+4:.align 4				;	\
+.popsection
+#else /* !__STDC__, i.e. -traditional */
+#define ELFNOTE(name, type, desctype, descdata) \
+.pushsection .note.name			;	\
+  .align 4				;	\
+  .long 2f - 1f		/* namesz */	;	\
+  .long 4f - 3f		/* descsz */	;	\
+  .long type				;	\
+1:.asciz "name"				;	\
+2:.align 4				;	\
+3:desctype descdata			;	\
+4:.align 4				;	\
+.popsection
+#endif /* __STDC__ */
+#else	/* !__ASSEMBLER__ */
+#include <linux/elf.h>
+/*
+ * Use an anonymous structure which matches the shape of
+ * Elf{32,64}_Nhdr, but includes the name and desc data.  The size and
+ * type of name and desc depend on the macro arguments.  "name" must
+ * be a literal string, and "desc" must be passed by value.  You may
+ * only define one note per line, since __LINE__ is used to generate
+ * unique symbols.
+ */
+#define _ELFNOTE_PASTE(a,b)	a##b
+#define _ELFNOTE(size, name, unique, type, desc)			\
+	static const struct {						\
+		struct elf##size##_note _nhdr;				\
+		unsigned char _name[sizeof(name)]			\
+		__attribute__((aligned(sizeof(Elf##size##_Word))));	\
+		typeof(desc) _desc					\
+			     __attribute__((aligned(sizeof(Elf##size##_Word)))); \
+	} _ELFNOTE_PASTE(_note_, unique)				\
+		__attribute_used__					\
+		__attribute__((section(".note." name),			\
+			       aligned(sizeof(Elf##size##_Word)),	\
+			       unused)) = {				\
+		{							\
+			sizeof(name),					\
+			sizeof(desc),					\
+			type,						\
+		},							\
+		name,							\
+		desc							\
+	}
+#define ELFNOTE(size, name, type, desc)		\
+	_ELFNOTE(size, name, __LINE__, type, desc)
+
+#define ELFNOTE32(name, type, desc) ELFNOTE(32, name, type, desc)
+#define ELFNOTE64(name, type, desc) ELFNOTE(64, name, type, desc)
+#endif	/* __ASSEMBLER__ */
+
+#endif /* _LINUX_ELFNOTE_H */
diff -rpuN linux-2.6.18.8/include/linux/eventpoll.h linux-2.6.18-xen-3.3.0/include/linux/eventpoll.h
--- linux-2.6.18.8/include/linux/eventpoll.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/eventpoll.h	2008-08-21 11:36:07.000000000 +0200
@@ -90,6 +90,12 @@ static inline void eventpoll_release(str
 	eventpoll_release_file(file);
 }
 
+/*
+ * called by aio code to create fd that can poll the  aio event queueQ
+ */
+struct eventpoll;
+int ep_getfd(int *efd, struct inode **einode, struct file **efile,
+             struct eventpoll *ep, const struct file_operations *fops);
 #else
 
 static inline void eventpoll_init_file(struct file *file) {}
diff -rpuN linux-2.6.18.8/include/linux/highmem.h linux-2.6.18-xen-3.3.0/include/linux/highmem.h
--- linux-2.6.18.8/include/linux/highmem.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/highmem.h	2008-08-21 11:36:07.000000000 +0200
@@ -24,10 +24,16 @@ static inline void flush_kernel_dcache_p
 
 /* declarations for linux/mm/highmem.c */
 unsigned int nr_free_highpages(void);
+#ifdef CONFIG_XEN
+void kmap_flush_unused(void);
+#endif
 
 #else /* CONFIG_HIGHMEM */
 
 static inline unsigned int nr_free_highpages(void) { return 0; }
+#ifdef CONFIG_XEN
+static inline void kmap_flush_unused(void) { }
+#endif
 
 static inline void *kmap(struct page *page)
 {
diff -rpuN linux-2.6.18.8/include/linux/interrupt.h linux-2.6.18-xen-3.3.0/include/linux/interrupt.h
--- linux-2.6.18.8/include/linux/interrupt.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/interrupt.h	2008-08-21 11:36:07.000000000 +0200
@@ -166,6 +166,12 @@ static inline int disable_irq_wake(unsig
 
 #endif /* CONFIG_GENERIC_HARDIRQS */
 
+#ifdef CONFIG_HAVE_IRQ_IGNORE_UNHANDLED
+int irq_ignore_unhandled(unsigned int irq);
+#else
+#define irq_ignore_unhandled(irq) 0
+#endif
+
 #ifndef __ARCH_SET_SOFTIRQ_PENDING
 #define set_softirq_pending(x) (local_softirq_pending() = (x))
 #define or_softirq_pending(x)  (local_softirq_pending() |= (x))
diff -rpuN linux-2.6.18.8/include/linux/ioport.h linux-2.6.18-xen-3.3.0/include/linux/ioport.h
--- linux-2.6.18.8/include/linux/ioport.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/ioport.h	2008-08-21 11:36:07.000000000 +0200
@@ -93,6 +93,9 @@ struct resource_list {
 /* PC/ISA/whatever - the normal PC address spaces: IO and memory */
 extern struct resource ioport_resource;
 extern struct resource iomem_resource;
+#ifdef CONFIG_PROC_IOMEM_MACHINE
+extern struct resource iomem_machine_resource;
+#endif
 
 extern int request_resource(struct resource *root, struct resource *new);
 extern struct resource * ____request_resource(struct resource *root, struct resource *new);
diff -rpuN linux-2.6.18.8/include/linux/kexec.h linux-2.6.18-xen-3.3.0/include/linux/kexec.h
--- linux-2.6.18.8/include/linux/kexec.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/kexec.h	2008-08-21 11:36:07.000000000 +0200
@@ -31,6 +31,13 @@
 #error KEXEC_ARCH not defined
 #endif
 
+#ifndef KEXEC_ARCH_HAS_PAGE_MACROS
+#define kexec_page_to_pfn(page)  page_to_pfn(page)
+#define kexec_pfn_to_page(pfn)   pfn_to_page(pfn)
+#define kexec_virt_to_phys(addr) virt_to_phys(addr)
+#define kexec_phys_to_virt(addr) phys_to_virt(addr)
+#endif
+
 /*
  * This structure is used to hold the arguments that are used when loading
  * kernel binaries.
@@ -91,6 +98,12 @@ struct kimage {
 extern NORET_TYPE void machine_kexec(struct kimage *image) ATTRIB_NORET;
 extern int machine_kexec_prepare(struct kimage *image);
 extern void machine_kexec_cleanup(struct kimage *image);
+#ifdef CONFIG_XEN
+extern int xen_machine_kexec_load(struct kimage *image);
+extern void xen_machine_kexec_unload(struct kimage *image);
+extern void xen_machine_kexec_setup_resources(void);
+extern void xen_machine_kexec_register_resources(struct resource *res);
+#endif
 extern asmlinkage long sys_kexec_load(unsigned long entry,
 					unsigned long nr_segments,
 					struct kexec_segment __user *segments,
@@ -108,6 +121,10 @@ int kexec_should_crash(struct task_struc
 extern struct kimage *kexec_image;
 extern struct kimage *kexec_crash_image;
 
+#ifndef kexec_flush_icache_page
+#define kexec_flush_icache_page(page)
+#endif
+
 #define KEXEC_ON_CRASH  0x00000001
 #define KEXEC_ARCH_MASK 0xffff0000
 
@@ -131,6 +148,7 @@ extern struct resource crashk_res;
 typedef u32 note_buf_t[MAX_NOTE_BYTES/4];
 extern note_buf_t *crash_notes;
 
+
 #else /* !CONFIG_KEXEC */
 struct pt_regs;
 struct task_struct;
diff -rpuN linux-2.6.18.8/include/linux/libata.h linux-2.6.18-xen-3.3.0/include/linux/libata.h
--- linux-2.6.18.8/include/linux/libata.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/libata.h	2008-08-21 11:36:07.000000000 +0200
@@ -162,6 +162,12 @@ enum {
 	ATA_FLAG_SKIP_D2H_BSY	= (1 << 12), /* can't wait for the first D2H
 					      * Register FIS clearing BSY */
 	ATA_FLAG_DEBUGMSG	= (1 << 13),
+	ATA_FLAG_IGN_SIMPLEX    = (1 << 15), /* ignore SIMPLEX */
+	ATA_FLAG_NO_IORDY       = (1 << 16), /* controller lacks iordy */
+	ATA_FLAG_ACPI_SATA      = (1 << 17), /* need native SATA ACPI layout */
+	ATA_FLAG_AN             = (1 << 18), /* controller supports AN */
+	ATA_FLAG_PMP            = (1 << 19), /* controller supports PMP */
+	ATA_FLAG_IPM            = (1 << 20), /* driver can handle IPM */
 
 	/* The following flag belongs to ap->pflags but is kept in
 	 * ap->flags because it's referenced in many LLDs and will be
diff -rpuN linux-2.6.18.8/include/linux/mm.h linux-2.6.18-xen-3.3.0/include/linux/mm.h
--- linux-2.6.18.8/include/linux/mm.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/mm.h	2008-08-21 11:36:07.000000000 +0200
@@ -164,6 +164,10 @@ extern unsigned int kobjsize(const void 
 #define VM_NONLINEAR	0x00800000	/* Is non-linear (remap_file_pages) */
 #define VM_MAPPED_COPY	0x01000000	/* T if mapped copy of data (nommu mmap) */
 #define VM_INSERTPAGE	0x02000000	/* The vma has had "vm_insert_page()" done on it */
+#ifdef CONFIG_XEN
+#define VM_FOREIGN	0x04000000	/* Has pages belonging to another VM */
+#endif
+#define VM_ALWAYSDUMP	0x08000000	/* Always include in core dumps */
 
 #ifndef VM_STACK_DEFAULT_FLAGS		/* arch can override this */
 #define VM_STACK_DEFAULT_FLAGS VM_DATA_DEFAULT_FLAGS
@@ -202,6 +206,10 @@ struct vm_operations_struct {
 	/* notification that a previously read-only page is about to become
 	 * writable, if an error is returned it will cause a SIGBUS */
 	int (*page_mkwrite)(struct vm_area_struct *vma, struct page *page);
+	/* Area-specific function for clearing the PTE at @ptep. Returns the
+	 * original value of @ptep. */
+	pte_t (*zap_pte)(struct vm_area_struct *vma, 
+			 unsigned long addr, pte_t *ptep, int is_fullmm);
 #ifdef CONFIG_NUMA
 	int (*set_policy)(struct vm_area_struct *vma, struct mempolicy *new);
 	struct mempolicy *(*get_policy)(struct vm_area_struct *vma,
@@ -1027,6 +1035,13 @@ struct page *follow_page(struct vm_area_
 #define FOLL_GET	0x04	/* do get_page on page */
 #define FOLL_ANON	0x08	/* give ZERO_PAGE if no pgtable */
 
+#ifdef CONFIG_XEN
+typedef int (*pte_fn_t)(pte_t *pte, struct page *pmd_page, unsigned long addr,
+			void *data);
+extern int apply_to_page_range(struct mm_struct *mm, unsigned long address,
+			       unsigned long size, pte_fn_t fn, void *data);
+#endif
+
 #ifdef CONFIG_PROC_FS
 void vm_stat_account(struct mm_struct *, unsigned long, struct file *, long);
 #else
diff -rpuN linux-2.6.18.8/include/linux/moduleparam.h linux-2.6.18-xen-3.3.0/include/linux/moduleparam.h
--- linux-2.6.18.8/include/linux/moduleparam.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/moduleparam.h	2008-08-21 11:36:07.000000000 +0200
@@ -58,13 +58,23 @@ struct kparam_array
 	void *elem;
 };
 
+/* On alpha, ia64 and ppc64 relocations to global data cannot go into
+   read-only sections (which is part of respective UNIX ABI on these
+   platforms). So 'const' makes no sense and even causes compile failures
+   with some compilers. */
+#if defined(CONFIG_ALPHA) || defined(CONFIG_IA64) || defined(CONFIG_PPC64)
+#define __moduleparam_const
+#else
+#define __moduleparam_const const
+#endif
+
 /* This is the fundamental function for registering boot/module
    parameters.  perm sets the visibility in driverfs: 000 means it's
    not there, read bits mean it's readable, write bits mean it's
    writable. */
 #define __module_param_call(prefix, name, set, get, arg, perm)		\
 	static char __param_str_##name[] = prefix #name;		\
-	static struct kernel_param const __param_##name			\
+	static struct kernel_param __moduleparam_const __param_##name	\
 	__attribute_used__						\
     __attribute__ ((unused,__section__ ("__param"),aligned(sizeof(void *)))) \
 	= { __param_str_##name, perm, set, get, arg }
diff -rpuN linux-2.6.18.8/include/linux/netfilter_bridge.h linux-2.6.18-xen-3.3.0/include/linux/netfilter_bridge.h
--- linux-2.6.18.8/include/linux/netfilter_bridge.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/netfilter_bridge.h	2008-08-21 11:36:07.000000000 +0200
@@ -7,6 +7,7 @@
 #include <linux/netfilter.h>
 #if defined(__KERNEL__) && defined(CONFIG_BRIDGE_NETFILTER)
 #include <linux/if_ether.h>
+#include <linux/if_vlan.h>
 #endif
 
 /* Bridge Hooks */
@@ -54,16 +55,19 @@ int nf_bridge_maybe_copy_header(struct s
 
 	if (skb->nf_bridge) {
 		if (skb->protocol == __constant_htons(ETH_P_8021Q)) {
-			err = skb_cow(skb, 18);
+			int header_size = ETH_HLEN + VLAN_HLEN;
+
+			err = skb_cow(skb, header_size);
 			if (err)
 				return err;
-			memcpy(skb->data - 18, skb->nf_bridge->data, 18);
-			skb_push(skb, 4);
+			memcpy(skb->data - header_size, skb->nf_bridge->data,
+			       header_size);
+			skb_push(skb, VLAN_HLEN);
 		} else {
-			err = skb_cow(skb, 16);
+			err = skb_cow(skb, ETH_HLEN);
 			if (err)
 				return err;
-			memcpy(skb->data - 16, skb->nf_bridge->data, 16);
+			memcpy(skb->data - ETH_HLEN, skb->nf_bridge->data, ETH_HLEN);
 		}
 	}
 	return 0;
diff -rpuN linux-2.6.18.8/include/linux/oprofile.h linux-2.6.18-xen-3.3.0/include/linux/oprofile.h
--- linux-2.6.18.8/include/linux/oprofile.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/oprofile.h	2008-08-21 11:36:07.000000000 +0200
@@ -16,6 +16,8 @@
 #include <linux/types.h>
 #include <linux/spinlock.h>
 #include <asm/atomic.h>
+
+#include <xen/interface/xenoprof.h>
  
 struct super_block;
 struct dentry;
@@ -27,6 +29,11 @@ struct oprofile_operations {
 	/* create any necessary configuration files in the oprofile fs.
 	 * Optional. */
 	int (*create_files)(struct super_block * sb, struct dentry * root);
+	/* setup active domains with Xen */
+	int (*set_active)(int *active_domains, unsigned int adomains);
+        /* setup passive domains with Xen */
+        int (*set_passive)(int *passive_domains, unsigned int pdomains);
+	
 	/* Do any necessary interrupt setup. Optional. */
 	int (*setup)(void);
 	/* Do any necessary interrupt shutdown. Optional. */
@@ -78,6 +85,8 @@ void oprofile_add_pc(unsigned long pc, i
 /* add a backtrace entry, to be called from the ->backtrace callback */
 void oprofile_add_trace(unsigned long eip);
 
+/* add a domain switch entry */
+int oprofile_add_domain_switch(int32_t domain_id);
 
 /**
  * Create a file of the given name as a child of the given root, with
diff -rpuN linux-2.6.18.8/include/linux/page-flags.h linux-2.6.18-xen-3.3.0/include/linux/page-flags.h
--- linux-2.6.18.8/include/linux/page-flags.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/page-flags.h	2008-08-21 11:36:07.000000000 +0200
@@ -98,6 +98,8 @@
 #define PG_uncached		31	/* Page has been mapped as uncached */
 #endif
 
+#define PG_foreign		20	/* Page is owned by foreign allocator. */
+
 /*
  * Manipulation of page state flags
  */
@@ -247,6 +249,19 @@
 #define SetPageUncached(page)	set_bit(PG_uncached, &(page)->flags)
 #define ClearPageUncached(page)	clear_bit(PG_uncached, &(page)->flags)
 
+#define PageForeign(page)	test_bit(PG_foreign, &(page)->flags)
+#define SetPageForeign(_page, dtor) do {		\
+	set_bit(PG_foreign, &(_page)->flags);		\
+	BUG_ON((dtor) == (void (*)(struct page *))0);	\
+	(_page)->index = (long)(dtor);			\
+} while (0)
+#define ClearPageForeign(page) do {			\
+	clear_bit(PG_foreign, &(page)->flags);		\
+	(page)->index = 0;				\
+} while (0)
+#define PageForeignDestructor(_page)			\
+	((void (*)(struct page *))(_page)->index)(_page)
+
 struct page;	/* forward declaration */
 
 int test_clear_page_dirty(struct page *page);
diff -rpuN linux-2.6.18.8/include/linux/pci.h linux-2.6.18-xen-3.3.0/include/linux/pci.h
--- linux-2.6.18.8/include/linux/pci.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/pci.h	2008-08-21 11:36:07.000000000 +0200
@@ -152,6 +152,9 @@ struct pci_dev {
 	 * directly, use the values stored here. They might be different!
 	 */
 	unsigned int	irq;
+#ifdef CONFIG_XEN
+	unsigned int    irq_old;
+#endif
 	struct resource resource[DEVICE_COUNT_RESOURCE]; /* I/O and memory regions + expansion ROMs */
 
 	/* These fields are used by common fixups */
@@ -596,6 +599,10 @@ static inline int pci_enable_msix(struct
 	struct msix_entry *entries, int nvec) {return -1;}
 static inline void pci_disable_msix(struct pci_dev *dev) {}
 static inline void msi_remove_pci_irq_vectors(struct pci_dev *dev) {}
+#ifdef CONFIG_XEN
+#define register_msi_get_owner(func) 0
+#define unregister_msi_get_owner(func) 0
+#endif
 #else
 extern void pci_scan_msi_device(struct pci_dev *dev);
 extern int pci_enable_msi(struct pci_dev *dev);
@@ -604,6 +611,10 @@ extern int pci_enable_msix(struct pci_de
 	struct msix_entry *entries, int nvec);
 extern void pci_disable_msix(struct pci_dev *dev);
 extern void msi_remove_pci_irq_vectors(struct pci_dev *dev);
+#ifdef CONFIG_XEN
+extern int register_msi_get_owner(int (*func)(struct pci_dev *dev));
+extern int unregister_msi_get_owner(int (*func)(struct pci_dev *dev));
+#endif
 #endif
 
 extern void pci_block_user_cfg_access(struct pci_dev *dev);
diff -rpuN linux-2.6.18.8/include/linux/pci_ids.h linux-2.6.18-xen-3.3.0/include/linux/pci_ids.h
--- linux-2.6.18.8/include/linux/pci_ids.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/pci_ids.h	2008-08-21 11:36:07.000000000 +0200
@@ -15,6 +15,8 @@
 #define PCI_CLASS_STORAGE_FLOPPY	0x0102
 #define PCI_CLASS_STORAGE_IPI		0x0103
 #define PCI_CLASS_STORAGE_RAID		0x0104
+#define PCI_CLASS_STORAGE_SATA          0x0106
+#define PCI_CLASS_STORAGE_SATA_AHCI     0x010601
 #define PCI_CLASS_STORAGE_SAS		0x0107
 #define PCI_CLASS_STORAGE_OTHER		0x0180
 
@@ -363,6 +365,8 @@
 #define PCI_DEVICE_ID_ATI_IXP600_SATA	0x4380
 #define PCI_DEVICE_ID_ATI_IXP600_SRAID	0x4381
 #define PCI_DEVICE_ID_ATI_IXP600_IDE	0x438c
+#define PCI_DEVICE_ID_ATI_IXP700_SATA	0x4390
+#define PCI_DEVICE_ID_ATI_IXP700_IDE	0x439c
 
 #define PCI_VENDOR_ID_VLSI		0x1004
 #define PCI_DEVICE_ID_VLSI_82C592	0x0005
@@ -2188,6 +2192,13 @@
 #define PCI_DEVICE_ID_INTEL_ICH8_4	0x2815
 #define PCI_DEVICE_ID_INTEL_ICH8_5	0x283e
 #define PCI_DEVICE_ID_INTEL_ICH8_6	0x2850
+#define PCI_DEVICE_ID_INTEL_ICH9_0	0x2910
+#define PCI_DEVICE_ID_INTEL_ICH9_1	0x2911
+#define PCI_DEVICE_ID_INTEL_ICH9_2	0x2912
+#define PCI_DEVICE_ID_INTEL_ICH9_3	0x2913
+#define PCI_DEVICE_ID_INTEL_ICH9_4	0x2914
+#define PCI_DEVICE_ID_INTEL_ICH9_5	0x2915
+#define PCI_DEVICE_ID_INTEL_ICH9_6	0x2930
 #define PCI_DEVICE_ID_INTEL_82855PM_HB	0x3340
 #define PCI_DEVICE_ID_INTEL_82830_HB	0x3575
 #define PCI_DEVICE_ID_INTEL_82830_CGC	0x3577
@@ -2202,6 +2213,12 @@
 #define PCI_DEVICE_ID_INTEL_MCH_PC	0x3599
 #define PCI_DEVICE_ID_INTEL_MCH_PC1	0x359a
 #define PCI_DEVICE_ID_INTEL_E7525_MCH	0x359e
+#define PCI_DEVICE_ID_INTEL_ICH10_0	0x3a14
+#define PCI_DEVICE_ID_INTEL_ICH10_1	0x3a16
+#define PCI_DEVICE_ID_INTEL_ICH10_2	0x3a18
+#define PCI_DEVICE_ID_INTEL_ICH10_3	0x3a1a
+#define PCI_DEVICE_ID_INTEL_ICH10_4	0x3a30
+#define PCI_DEVICE_ID_INTEL_ICH10_5	0x3a60
 #define PCI_DEVICE_ID_INTEL_82371SB_0	0x7000
 #define PCI_DEVICE_ID_INTEL_82371SB_1	0x7010
 #define PCI_DEVICE_ID_INTEL_82371SB_2	0x7020
diff -rpuN linux-2.6.18.8/include/linux/sched.h linux-2.6.18-xen-3.3.0/include/linux/sched.h
--- linux-2.6.18.8/include/linux/sched.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/sched.h	2008-08-21 11:36:07.000000000 +0200
@@ -211,10 +211,15 @@ extern void update_process_times(int use
 extern void scheduler_tick(void);
 
 #ifdef CONFIG_DETECT_SOFTLOCKUP
+extern unsigned long softlockup_get_next_event(void);
 extern void softlockup_tick(void);
 extern void spawn_softlockup_task(void);
 extern void touch_softlockup_watchdog(void);
 #else
+static inline unsigned long softlockup_get_next_event(void)
+{
+	return MAX_JIFFY_OFFSET;
+}
 static inline void softlockup_tick(void)
 {
 }
diff -rpuN linux-2.6.18.8/include/linux/skbuff.h linux-2.6.18-xen-3.3.0/include/linux/skbuff.h
--- linux-2.6.18.8/include/linux/skbuff.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/skbuff.h	2008-08-21 11:36:07.000000000 +0200
@@ -203,6 +203,8 @@ enum {
  *	@local_df: allow local fragmentation
  *	@cloned: Head may be cloned (check refcnt to be sure)
  *	@nohdr: Payload reference only, must not modify header
+ *	@proto_data_valid: Protocol data validated since arriving at localhost
+ *	@proto_csum_blank: Protocol csum must be added before leaving localhost
  *	@pkt_type: Packet class
  *	@fclone: skbuff clone status
  *	@ip_summed: Driver fed us an IP checksum
@@ -282,7 +284,13 @@ struct sk_buff {
 				nfctinfo:3;
 	__u8			pkt_type:3,
 				fclone:2,
+#ifndef CONFIG_XEN
 				ipvs_property:1;
+#else
+				ipvs_property:1,
+				proto_data_valid:1,
+				proto_csum_blank:1;
+#endif
 	__be16			protocol;
 
 	void			(*destructor)(struct sk_buff *skb);
diff -rpuN linux-2.6.18.8/include/linux/sysctl.h linux-2.6.18-xen-3.3.0/include/linux/sysctl.h
--- linux-2.6.18.8/include/linux/sysctl.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/sysctl.h	2008-08-21 11:36:07.000000000 +0200
@@ -6,10 +6,17 @@
  ****************************************************************
  ****************************************************************
  **
+ **  WARNING:
  **  The values in this file are exported to user space via 
- **  the sysctl() binary interface.  However this interface
- **  is unstable and deprecated and will be removed in the future. 
- **  For a stable interface use /proc/sys.
+ **  the sysctl() binary interface.  Do *NOT* change the
+ **  numbering of any existing values here, and do not change
+ **  any numbers within any one set of values.  If you have to
+ **  have to redefine an existing interface, use a new number for it.
+ **  The kernel will then return -ENOTDIR to any application using
+ **  the old binary interface.
+ **
+ **  For new interfaces unless you really need a binary number
+ **  please use CTL_UNNUMBERED.
  **
  ****************************************************************
  ****************************************************************
@@ -48,6 +55,7 @@ struct __sysctl_args {
 #ifdef __KERNEL__
 #define CTL_ANY		-1	/* Matches any name */
 #define CTL_NONE	0
+#define CTL_UNNUMBERED	CTL_NONE	/* sysctl without a binary number */
 #endif
 
 enum
diff -rpuN linux-2.6.18.8/include/linux/vermagic.h linux-2.6.18-xen-3.3.0/include/linux/vermagic.h
--- linux-2.6.18.8/include/linux/vermagic.h	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/linux/vermagic.h	2008-08-21 11:36:07.000000000 +0200
@@ -17,6 +17,11 @@
 #else
 #define MODULE_VERMAGIC_MODULE_UNLOAD ""
 #endif
+#ifdef CONFIG_XEN
+#define MODULE_VERMAGIC_XEN "Xen "
+#else
+#define MODULE_VERMAGIC_XEN
+#endif
 #ifndef MODULE_ARCH_VERMAGIC
 #define MODULE_ARCH_VERMAGIC ""
 #endif
@@ -24,5 +29,6 @@
 #define VERMAGIC_STRING 						\
 	UTS_RELEASE " "							\
 	MODULE_VERMAGIC_SMP MODULE_VERMAGIC_PREEMPT 			\
-	MODULE_VERMAGIC_MODULE_UNLOAD MODULE_ARCH_VERMAGIC 		\
+	MODULE_VERMAGIC_MODULE_UNLOAD MODULE_VERMAGIC_XEN		\
+	MODULE_ARCH_VERMAGIC						\
 	"gcc-" __stringify(__GNUC__) "." __stringify(__GNUC_MINOR__)
diff -rpuN linux-2.6.18.8/include/xen/balloon.h linux-2.6.18-xen-3.3.0/include/xen/balloon.h
--- linux-2.6.18.8/include/xen/balloon.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/balloon.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,57 @@
+/******************************************************************************
+ * balloon.h
+ *
+ * Xen balloon driver - enables returning/claiming memory to/from Xen.
+ *
+ * Copyright (c) 2003, B Dragovic
+ * Copyright (c) 2003-2004, M Williamson, K Fraser
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __ASM_BALLOON_H__
+#define __ASM_BALLOON_H__
+
+/*
+ * Inform the balloon driver that it should allow some slop for device-driver
+ * memory activities.
+ */
+void balloon_update_driver_allowance(long delta);
+
+/* Allocate/free a set of empty pages in low memory (i.e., no RAM mapped). */
+struct page **alloc_empty_pages_and_pagevec(int nr_pages);
+void free_empty_pages_and_pagevec(struct page **pagevec, int nr_pages);
+
+void balloon_release_driver_page(struct page *page);
+
+/*
+ * Prevent the balloon driver from changing the memory reservation during
+ * a driver critical region.
+ */
+extern spinlock_t balloon_lock;
+#define balloon_lock(__flags)   spin_lock_irqsave(&balloon_lock, __flags)
+#define balloon_unlock(__flags) spin_unlock_irqrestore(&balloon_lock, __flags)
+
+#endif /* __ASM_BALLOON_H__ */
diff -rpuN linux-2.6.18.8/include/xen/blkif.h linux-2.6.18-xen-3.3.0/include/xen/blkif.h
--- linux-2.6.18.8/include/xen/blkif.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/blkif.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,123 @@
+/* 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_BLKIF_H__
+#define __XEN_BLKIF_H__
+
+#include <xen/interface/io/ring.h>
+#include <xen/interface/io/blkif.h>
+#include <xen/interface/io/protocols.h>
+
+/* Not a real protocol.  Used to generate ring structs which contain
+ * the elements common to all protocols only.  This way we get a
+ * compiler-checkable way to use common struct elements, so we can
+ * avoid using switch(protocol) in a number of places.  */
+struct blkif_common_request {
+	char dummy;
+};
+struct blkif_common_response {
+	char dummy;
+};
+
+/* i386 protocol version */
+#pragma pack(push, 4)
+struct blkif_x86_32_request {
+	uint8_t        operation;    /* BLKIF_OP_???                         */
+	uint8_t        nr_segments;  /* number of segments                   */
+	blkif_vdev_t   handle;       /* only for read/write requests         */
+	uint64_t       id;           /* private guest value, echoed in resp  */
+	blkif_sector_t sector_number;/* start sector idx on disk (r/w only)  */
+	struct blkif_request_segment seg[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+};
+struct blkif_x86_32_response {
+	uint64_t        id;              /* copied from request */
+	uint8_t         operation;       /* copied from request */
+	int16_t         status;          /* BLKIF_RSP_???       */
+};
+typedef struct blkif_x86_32_request blkif_x86_32_request_t;
+typedef struct blkif_x86_32_response blkif_x86_32_response_t;
+#pragma pack(pop)
+
+/* x86_64 protocol version */
+struct blkif_x86_64_request {
+	uint8_t        operation;    /* BLKIF_OP_???                         */
+	uint8_t        nr_segments;  /* number of segments                   */
+	blkif_vdev_t   handle;       /* only for read/write requests         */
+	uint64_t       __attribute__((__aligned__(8))) id;
+	blkif_sector_t sector_number;/* start sector idx on disk (r/w only)  */
+	struct blkif_request_segment seg[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+};
+struct blkif_x86_64_response {
+	uint64_t       __attribute__((__aligned__(8))) id;
+	uint8_t         operation;       /* copied from request */
+	int16_t         status;          /* BLKIF_RSP_???       */
+};
+typedef struct blkif_x86_64_request blkif_x86_64_request_t;
+typedef struct blkif_x86_64_response blkif_x86_64_response_t;
+
+DEFINE_RING_TYPES(blkif_common, struct blkif_common_request, struct blkif_common_response);
+DEFINE_RING_TYPES(blkif_x86_32, struct blkif_x86_32_request, struct blkif_x86_32_response);
+DEFINE_RING_TYPES(blkif_x86_64, struct blkif_x86_64_request, struct blkif_x86_64_response);
+
+union blkif_back_rings {
+	blkif_back_ring_t        native;
+	blkif_common_back_ring_t common;
+	blkif_x86_32_back_ring_t x86_32;
+	blkif_x86_64_back_ring_t x86_64;
+};
+typedef union blkif_back_rings blkif_back_rings_t;
+
+enum blkif_protocol {
+	BLKIF_PROTOCOL_NATIVE = 1,
+	BLKIF_PROTOCOL_X86_32 = 2,
+	BLKIF_PROTOCOL_X86_64 = 3,
+};
+
+static void inline blkif_get_x86_32_req(blkif_request_t *dst, blkif_x86_32_request_t *src)
+{
+	int i, n = BLKIF_MAX_SEGMENTS_PER_REQUEST;
+	dst->operation = src->operation;
+	dst->nr_segments = src->nr_segments;
+	dst->handle = src->handle;
+	dst->id = src->id;
+	dst->sector_number = src->sector_number;
+	barrier();
+	if (n > dst->nr_segments)
+		n = dst->nr_segments;
+	for (i = 0; i < n; i++)
+		dst->seg[i] = src->seg[i];
+}
+
+static void inline blkif_get_x86_64_req(blkif_request_t *dst, blkif_x86_64_request_t *src)
+{
+	int i, n = BLKIF_MAX_SEGMENTS_PER_REQUEST;
+	dst->operation = src->operation;
+	dst->nr_segments = src->nr_segments;
+	dst->handle = src->handle;
+	dst->id = src->id;
+	dst->sector_number = src->sector_number;
+	barrier();
+	if (n > dst->nr_segments)
+		n = dst->nr_segments;
+	for (i = 0; i < n; i++)
+		dst->seg[i] = src->seg[i];
+}
+
+#endif /* __XEN_BLKIF_H__ */
diff -rpuN linux-2.6.18.8/include/xen/compat_ioctl.h linux-2.6.18-xen-3.3.0/include/xen/compat_ioctl.h
--- linux-2.6.18.8/include/xen/compat_ioctl.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/compat_ioctl.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,45 @@
+/*
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 of the
+ * License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
+ *
+ * Copyright IBM Corp. 2007
+ *
+ * Authors: Jimi Xenidis <jimix@watson.ibm.com>
+ *          Hollis Blanchard <hollisb@us.ibm.com>
+ */
+
+#ifndef __LINUX_XEN_COMPAT_H__ 
+#define __LINUX_XEN_COMPAT_H__ 
+
+#include <linux/compat.h>
+
+extern int privcmd_ioctl_32(int fd, unsigned int cmd, unsigned long arg);
+struct privcmd_mmap_32 {
+	int num;
+	domid_t dom;
+	compat_uptr_t entry;
+};
+
+struct privcmd_mmapbatch_32 {
+	int num;     /* number of pages to populate */
+	domid_t dom; /* target domain */
+	__u64 addr;  /* virtual address */
+	compat_uptr_t arr; /* array of mfns - top nibble set on err */
+};
+#define IOCTL_PRIVCMD_MMAP_32                   \
+	_IOC(_IOC_NONE, 'P', 2, sizeof(struct privcmd_mmap_32))
+#define IOCTL_PRIVCMD_MMAPBATCH_32                  \
+	_IOC(_IOC_NONE, 'P', 3, sizeof(struct privcmd_mmapbatch_32))
+
+#endif /* __LINUX_XEN_COMPAT_H__ */
diff -rpuN linux-2.6.18.8/include/xen/cpu_hotplug.h linux-2.6.18-xen-3.3.0/include/xen/cpu_hotplug.h
--- linux-2.6.18.8/include/xen/cpu_hotplug.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/cpu_hotplug.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,41 @@
+#ifndef __XEN_CPU_HOTPLUG_H__
+#define __XEN_CPU_HOTPLUG_H__
+
+#include <linux/kernel.h>
+#include <linux/cpumask.h>
+
+#if defined(CONFIG_X86) && defined(CONFIG_SMP)
+extern cpumask_t cpu_initialized_map;
+#endif
+
+#if defined(CONFIG_HOTPLUG_CPU)
+
+int cpu_up_check(unsigned int cpu);
+void init_xenbus_allowed_cpumask(void);
+int smp_suspend(void);
+void smp_resume(void);
+
+void cpu_bringup(void);
+
+#else /* !defined(CONFIG_HOTPLUG_CPU) */
+
+#define cpu_up_check(cpu)		(0)
+#define init_xenbus_allowed_cpumask()	((void)0)
+
+static inline int smp_suspend(void)
+{
+	if (num_online_cpus() > 1) {
+		printk(KERN_WARNING "Can't suspend SMP guests "
+		       "without CONFIG_HOTPLUG_CPU\n");
+		return -EOPNOTSUPP;
+	}
+	return 0;
+}
+
+static inline void smp_resume(void)
+{
+}
+
+#endif /* !defined(CONFIG_HOTPLUG_CPU) */
+
+#endif /* __XEN_CPU_HOTPLUG_H__ */
diff -rpuN linux-2.6.18.8/include/xen/driver_util.h linux-2.6.18-xen-3.3.0/include/xen/driver_util.h
--- linux-2.6.18.8/include/xen/driver_util.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/driver_util.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,14 @@
+
+#ifndef __ASM_XEN_DRIVER_UTIL_H__
+#define __ASM_XEN_DRIVER_UTIL_H__
+
+#include <linux/vmalloc.h>
+#include <linux/device.h>
+
+/* Allocate/destroy a 'vmalloc' VM area. */
+extern struct vm_struct *alloc_vm_area(unsigned long size);
+extern void free_vm_area(struct vm_struct *area);
+
+extern struct class *get_xen_class(void);
+
+#endif /* __ASM_XEN_DRIVER_UTIL_H__ */
diff -rpuN linux-2.6.18.8/include/xen/evtchn.h linux-2.6.18-xen-3.3.0/include/xen/evtchn.h
--- linux-2.6.18.8/include/xen/evtchn.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/evtchn.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,153 @@
+/******************************************************************************
+ * evtchn.h
+ * 
+ * Communication via Xen event channels.
+ * Also definitions for the device that demuxes notifications to userspace.
+ * 
+ * Copyright (c) 2004-2005, K A Fraser
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __ASM_EVTCHN_H__
+#define __ASM_EVTCHN_H__
+
+#include <linux/interrupt.h>
+#include <asm/hypervisor.h>
+#include <asm/ptrace.h>
+#include <asm/synch_bitops.h>
+#include <xen/interface/event_channel.h>
+#include <linux/smp.h>
+
+/*
+ * LOW-LEVEL DEFINITIONS
+ */
+
+/*
+ * Dynamically bind an event source to an IRQ-like callback handler.
+ * On some platforms this may not be implemented via the Linux IRQ subsystem.
+ * The IRQ argument passed to the callback handler is the same as returned
+ * from the bind call. It may not correspond to a Linux IRQ number.
+ * Returns IRQ or negative errno.
+ */
+int bind_caller_port_to_irqhandler(
+	unsigned int caller_port,
+	irqreturn_t (*handler)(int, void *, struct pt_regs *),
+	unsigned long irqflags,
+	const char *devname,
+	void *dev_id);
+int bind_listening_port_to_irqhandler(
+	unsigned int remote_domain,
+	irqreturn_t (*handler)(int, void *, struct pt_regs *),
+	unsigned long irqflags,
+	const char *devname,
+	void *dev_id);
+int bind_interdomain_evtchn_to_irqhandler(
+	unsigned int remote_domain,
+	unsigned int remote_port,
+	irqreturn_t (*handler)(int, void *, struct pt_regs *),
+	unsigned long irqflags,
+	const char *devname,
+	void *dev_id);
+int bind_virq_to_irqhandler(
+	unsigned int virq,
+	unsigned int cpu,
+	irqreturn_t (*handler)(int, void *, struct pt_regs *),
+	unsigned long irqflags,
+	const char *devname,
+	void *dev_id);
+int bind_ipi_to_irqhandler(
+	unsigned int ipi,
+	unsigned int cpu,
+	irqreturn_t (*handler)(int, void *, struct pt_regs *),
+	unsigned long irqflags,
+	const char *devname,
+	void *dev_id);
+
+/*
+ * Common unbind function for all event sources. Takes IRQ to unbind from.
+ * Automatically closes the underlying event channel (except for bindings
+ * made with bind_caller_port_to_irqhandler()).
+ */
+void unbind_from_irqhandler(unsigned int irq, void *dev_id);
+
+void irq_resume(void);
+
+/* Entry point for notifications into Linux subsystems. */
+asmlinkage void evtchn_do_upcall(struct pt_regs *regs);
+
+/* Entry point for notifications into the userland character device. */
+void evtchn_device_upcall(int port);
+
+void mask_evtchn(int port);
+void disable_all_local_evtchn(void);
+void unmask_evtchn(int port);
+
+#ifdef CONFIG_SMP
+void rebind_evtchn_to_cpu(int port, unsigned int cpu);
+#else
+#define rebind_evtchn_to_cpu(port, cpu)	((void)0)
+#endif
+
+static inline int test_and_set_evtchn_mask(int port)
+{
+	shared_info_t *s = HYPERVISOR_shared_info;
+	return synch_test_and_set_bit(port, s->evtchn_mask);
+}
+
+static inline void clear_evtchn(int port)
+{
+	shared_info_t *s = HYPERVISOR_shared_info;
+	synch_clear_bit(port, s->evtchn_pending);
+}
+
+static inline void notify_remote_via_evtchn(int port)
+{
+	struct evtchn_send send = { .port = port };
+	VOID(HYPERVISOR_event_channel_op(EVTCHNOP_send, &send));
+}
+
+/*
+ * Use these to access the event channel underlying the IRQ handle returned
+ * by bind_*_to_irqhandler().
+ */
+void notify_remote_via_irq(int irq);
+int irq_to_evtchn_port(int irq);
+
+#define PIRQ_SET_MAPPING 0x0
+#define PIRQ_CLEAR_MAPPING 0x1
+#define PIRQ_GET_MAPPING 0x3
+int pirq_mapstatus(int pirq, int action);
+int set_pirq_hw_action(int pirq, int (*action)(int pirq, int action));
+int clear_pirq_hw_action(int pirq);
+
+#define PIRQ_STARTUP 1
+#define PIRQ_SHUTDOWN 2
+#define PIRQ_ENABLE 3
+#define PIRQ_DISABLE 4
+#define PIRQ_END 5
+#define PIRQ_ACK 6
+
+#endif /* __ASM_EVTCHN_H__ */
diff -rpuN linux-2.6.18.8/include/xen/features.h linux-2.6.18-xen-3.3.0/include/xen/features.h
--- linux-2.6.18.8/include/xen/features.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/features.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,20 @@
+/******************************************************************************
+ * features.h
+ *
+ * Query the features reported by Xen.
+ *
+ * Copyright (c) 2006, Ian Campbell
+ */
+
+#ifndef __ASM_XEN_FEATURES_H__
+#define __ASM_XEN_FEATURES_H__
+
+#include <xen/interface/version.h>
+
+extern void setup_xen_features(void);
+
+extern u8 xen_features[XENFEAT_NR_SUBMAPS * 32];
+
+#define xen_feature(flag)	(xen_features[flag])
+
+#endif /* __ASM_XEN_FEATURES_H__ */
diff -rpuN linux-2.6.18.8/include/xen/firmware.h linux-2.6.18-xen-3.3.0/include/xen/firmware.h
--- linux-2.6.18.8/include/xen/firmware.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/firmware.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,10 @@
+#ifndef __XEN_FIRMWARE_H__
+#define __XEN_FIRMWARE_H__
+
+#if defined(CONFIG_EDD) || defined(CONFIG_EDD_MODULE)
+void copy_edd(void);
+#endif
+
+void copy_edid(void);
+
+#endif /* __XEN_FIRMWARE_H__ */
diff -rpuN linux-2.6.18.8/include/xen/gnttab.h linux-2.6.18-xen-3.3.0/include/xen/gnttab.h
--- linux-2.6.18.8/include/xen/gnttab.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/gnttab.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,163 @@
+/******************************************************************************
+ * gnttab.h
+ * 
+ * Two sets of functionality:
+ * 1. Granting foreign access to our memory reservation.
+ * 2. Accessing others' memory reservations via grant references.
+ * (i.e., mechanisms for both sender and recipient of grant references)
+ * 
+ * Copyright (c) 2004-2005, K A Fraser
+ * Copyright (c) 2005, Christopher Clark
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __ASM_GNTTAB_H__
+#define __ASM_GNTTAB_H__
+
+#include <asm/hypervisor.h>
+#include <asm/maddr.h> /* maddr_t */
+#include <linux/mm.h>
+#include <xen/interface/grant_table.h>
+#include <xen/features.h>
+
+struct gnttab_free_callback {
+	struct gnttab_free_callback *next;
+	void (*fn)(void *);
+	void *arg;
+	u16 count;
+};
+
+int gnttab_grant_foreign_access(domid_t domid, unsigned long frame,
+				int flags);
+
+/*
+ * End access through the given grant reference, iff the grant entry is no
+ * longer in use.  Return 1 if the grant entry was freed, 0 if it is still in
+ * use.
+ */
+int gnttab_end_foreign_access_ref(grant_ref_t ref);
+
+/*
+ * Eventually end access through the given grant reference, and once that
+ * access has been ended, free the given page too.  Access will be ended
+ * immediately iff the grant entry is not in use, otherwise it will happen
+ * some time later.  page may be 0, in which case no freeing will occur.
+ */
+void gnttab_end_foreign_access(grant_ref_t ref, unsigned long page);
+
+int gnttab_grant_foreign_transfer(domid_t domid, unsigned long pfn);
+
+unsigned long gnttab_end_foreign_transfer_ref(grant_ref_t ref);
+unsigned long gnttab_end_foreign_transfer(grant_ref_t ref);
+
+int gnttab_query_foreign_access(grant_ref_t ref);
+
+/*
+ * operations on reserved batches of grant references
+ */
+int gnttab_alloc_grant_references(u16 count, grant_ref_t *pprivate_head);
+
+void gnttab_free_grant_reference(grant_ref_t ref);
+
+void gnttab_free_grant_references(grant_ref_t head);
+
+int gnttab_empty_grant_references(const grant_ref_t *pprivate_head);
+
+int gnttab_claim_grant_reference(grant_ref_t *pprivate_head);
+
+void gnttab_release_grant_reference(grant_ref_t *private_head,
+				    grant_ref_t release);
+
+void gnttab_request_free_callback(struct gnttab_free_callback *callback,
+				  void (*fn)(void *), void *arg, u16 count);
+void gnttab_cancel_free_callback(struct gnttab_free_callback *callback);
+
+void gnttab_grant_foreign_access_ref(grant_ref_t ref, domid_t domid,
+				     unsigned long frame, int flags);
+
+void gnttab_grant_foreign_transfer_ref(grant_ref_t, domid_t domid,
+				       unsigned long pfn);
+
+int gnttab_copy_grant_page(grant_ref_t ref, struct page **pagep);
+void __gnttab_dma_map_page(struct page *page);
+static inline void __gnttab_dma_unmap_page(struct page *page)
+{
+}
+
+void gnttab_reset_grant_page(struct page *page);
+
+int gnttab_suspend(void);
+int gnttab_resume(void);
+
+void *arch_gnttab_alloc_shared(unsigned long *frames);
+
+static inline void
+gnttab_set_map_op(struct gnttab_map_grant_ref *map, maddr_t addr,
+		  uint32_t flags, grant_ref_t ref, domid_t domid)
+{
+	if (flags & GNTMAP_contains_pte)
+		map->host_addr = addr;
+	else if (xen_feature(XENFEAT_auto_translated_physmap))
+		map->host_addr = __pa(addr);
+	else
+		map->host_addr = addr;
+
+	map->flags = flags;
+	map->ref = ref;
+	map->dom = domid;
+}
+
+static inline void
+gnttab_set_unmap_op(struct gnttab_unmap_grant_ref *unmap, maddr_t addr,
+		    uint32_t flags, grant_handle_t handle)
+{
+	if (flags & GNTMAP_contains_pte)
+		unmap->host_addr = addr;
+	else if (xen_feature(XENFEAT_auto_translated_physmap))
+		unmap->host_addr = __pa(addr);
+	else
+		unmap->host_addr = addr;
+
+	unmap->handle = handle;
+	unmap->dev_bus_addr = 0;
+}
+
+static inline void
+gnttab_set_replace_op(struct gnttab_unmap_and_replace *unmap, maddr_t addr,
+		      maddr_t new_addr, grant_handle_t handle)
+{
+	if (xen_feature(XENFEAT_auto_translated_physmap)) {
+		unmap->host_addr = __pa(addr);
+		unmap->new_addr = __pa(new_addr);
+	} else {
+		unmap->host_addr = addr;
+		unmap->new_addr = new_addr;
+	}
+
+	unmap->handle = handle;
+}
+
+#endif /* __ASM_GNTTAB_H__ */
diff -rpuN linux-2.6.18.8/include/xen/hvm.h linux-2.6.18-xen-3.3.0/include/xen/hvm.h
--- linux-2.6.18.8/include/xen/hvm.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/hvm.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,23 @@
+/* Simple wrappers around HVM functions */
+#ifndef XEN_HVM_H__
+#define XEN_HVM_H__
+
+#include <xen/interface/hvm/params.h>
+
+static inline unsigned long hvm_get_parameter(int idx)
+{
+	struct xen_hvm_param xhv;
+	int r;
+
+	xhv.domid = DOMID_SELF;
+	xhv.index = idx;
+	r = HYPERVISOR_hvm_op(HVMOP_get_param, &xhv);
+	if (r < 0) {
+		printk(KERN_ERR "cannot get hvm parameter %d: %d.\n",
+		       idx, r);
+		return 0;
+	}
+	return xhv.value;
+}
+
+#endif /* XEN_HVM_H__ */
diff -rpuN linux-2.6.18.8/include/xen/hypercall.h linux-2.6.18-xen-3.3.0/include/xen/hypercall.h
--- linux-2.6.18.8/include/xen/hypercall.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/hypercall.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,30 @@
+#ifndef __XEN_HYPERCALL_H__
+#define __XEN_HYPERCALL_H__
+
+#include <asm/hypercall.h>
+
+static inline int __must_check
+HYPERVISOR_multicall_check(
+	multicall_entry_t *call_list, unsigned int nr_calls,
+	const unsigned long *rc_list)
+{
+	int rc = HYPERVISOR_multicall(call_list, nr_calls);
+
+	if (unlikely(rc < 0))
+		return rc;
+	BUG_ON(rc);
+	BUG_ON((int)nr_calls < 0);
+
+	for ( ; nr_calls > 0; --nr_calls, ++call_list)
+		if (unlikely(call_list->result != (rc_list ? *rc_list++ : 0)))
+			return nr_calls;
+
+	return 0;
+}
+
+/* A construct to ignore the return value of hypercall wrappers in a few
+ * exceptional cases (simply casting the function result to void doesn't
+ * avoid the compiler warning): */
+#define VOID(expr) ((void)((expr)?:0))
+
+#endif /* __XEN_HYPERCALL_H__ */
diff -rpuN linux-2.6.18.8/include/xen/hypervisor_sysfs.h linux-2.6.18-xen-3.3.0/include/xen/hypervisor_sysfs.h
--- linux-2.6.18.8/include/xen/hypervisor_sysfs.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/hypervisor_sysfs.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,30 @@
+/*
+ *  copyright (c) 2006 IBM Corporation
+ *  Authored by: Mike D. Day <ncmike@us.ibm.com>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License version 2 as
+ *  published by the Free Software Foundation.
+ */
+
+#ifndef _HYP_SYSFS_H_
+#define _HYP_SYSFS_H_
+
+#include <linux/kobject.h>
+#include <linux/sysfs.h>
+
+#define HYPERVISOR_ATTR_RO(_name) \
+static struct hyp_sysfs_attr  _name##_attr = __ATTR_RO(_name)
+
+#define HYPERVISOR_ATTR_RW(_name) \
+static struct hyp_sysfs_attr _name##_attr = \
+	__ATTR(_name, 0644, _name##_show, _name##_store)
+
+struct hyp_sysfs_attr {
+	struct attribute attr;
+	ssize_t (*show)(struct hyp_sysfs_attr *, char *);
+	ssize_t (*store)(struct hyp_sysfs_attr *, const char *, size_t);
+	void *hyp_attr_data;
+};
+
+#endif /* _HYP_SYSFS_H_ */
diff -rpuN linux-2.6.18.8/include/xen/interface/acm.h linux-2.6.18-xen-3.3.0/include/xen/interface/acm.h
--- linux-2.6.18.8/include/xen/interface/acm.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/acm.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,228 @@
+/*
+ * acm.h: Xen access control module interface defintions
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Reiner Sailer <sailer@watson.ibm.com>
+ * Copyright (c) 2005, International Business Machines Corporation.
+ */
+
+#ifndef _XEN_PUBLIC_ACM_H
+#define _XEN_PUBLIC_ACM_H
+
+#include "xen.h"
+
+/* if ACM_DEBUG defined, all hooks should
+ * print a short trace message (comment it out
+ * when not in testing mode )
+ */
+/* #define ACM_DEBUG */
+
+#ifdef ACM_DEBUG
+#  define printkd(fmt, args...) printk(fmt,## args)
+#else
+#  define printkd(fmt, args...)
+#endif
+
+/* default ssid reference value if not supplied */
+#define ACM_DEFAULT_SSID  0x0
+#define ACM_DEFAULT_LOCAL_SSID  0x0
+
+/* Internal ACM ERROR types */
+#define ACM_OK     0
+#define ACM_UNDEF   -1
+#define ACM_INIT_SSID_ERROR  -2
+#define ACM_INIT_SOID_ERROR  -3
+#define ACM_ERROR          -4
+
+/* External ACCESS DECISIONS */
+#define ACM_ACCESS_PERMITTED        0
+#define ACM_ACCESS_DENIED           -111
+#define ACM_NULL_POINTER_ERROR      -200
+
+/*
+   Error codes reported in when trying to test for a new policy
+   These error codes are reported in an array of tuples where
+   each error code is followed by a parameter describing the error
+   more closely, such as a domain id.
+*/
+#define ACM_EVTCHN_SHARING_VIOLATION       0x100
+#define ACM_GNTTAB_SHARING_VIOLATION       0x101
+#define ACM_DOMAIN_LOOKUP                  0x102
+#define ACM_CHWALL_CONFLICT                0x103
+#define ACM_SSIDREF_IN_USE                 0x104
+
+
+/* primary policy in lower 4 bits */
+#define ACM_NULL_POLICY 0
+#define ACM_CHINESE_WALL_POLICY 1
+#define ACM_SIMPLE_TYPE_ENFORCEMENT_POLICY 2
+#define ACM_POLICY_UNDEFINED 15
+
+/* combinations have secondary policy component in higher 4bit */
+#define ACM_CHINESE_WALL_AND_SIMPLE_TYPE_ENFORCEMENT_POLICY \
+    ((ACM_SIMPLE_TYPE_ENFORCEMENT_POLICY << 4) | ACM_CHINESE_WALL_POLICY)
+
+/* policy: */
+#define ACM_POLICY_NAME(X) \
+ ((X) == (ACM_NULL_POLICY)) ? "NULL" :                        \
+    ((X) == (ACM_CHINESE_WALL_POLICY)) ? "CHINESE WALL" :        \
+    ((X) == (ACM_SIMPLE_TYPE_ENFORCEMENT_POLICY)) ? "SIMPLE TYPE ENFORCEMENT" : \
+    ((X) == (ACM_CHINESE_WALL_AND_SIMPLE_TYPE_ENFORCEMENT_POLICY)) ? "CHINESE WALL AND SIMPLE TYPE ENFORCEMENT" : \
+     "UNDEFINED"
+
+/* the following policy versions must be increased
+ * whenever the interpretation of the related
+ * policy's data structure changes
+ */
+#define ACM_POLICY_VERSION 3
+#define ACM_CHWALL_VERSION 1
+#define ACM_STE_VERSION  1
+
+/* defines a ssid reference used by xen */
+typedef uint32_t ssidref_t;
+
+/* hooks that are known to domains */
+#define ACMHOOK_none    0
+#define ACMHOOK_sharing 1
+
+/* -------security policy relevant type definitions-------- */
+
+/* type identifier; compares to "equal" or "not equal" */
+typedef uint16_t domaintype_t;
+
+/* CHINESE WALL POLICY DATA STRUCTURES
+ *
+ * current accumulated conflict type set:
+ * When a domain is started and has a type that is in
+ * a conflict set, the conflicting types are incremented in
+ * the aggregate set. When a domain is destroyed, the 
+ * conflicting types to its type are decremented.
+ * If a domain has multiple types, this procedure works over
+ * all those types.
+ *
+ * conflict_aggregate_set[i] holds the number of
+ *   running domains that have a conflict with type i.
+ *
+ * running_types[i] holds the number of running domains
+ *        that include type i in their ssidref-referenced type set
+ *
+ * conflict_sets[i][j] is "0" if type j has no conflict
+ *    with type i and is "1" otherwise.
+ */
+/* high-16 = version, low-16 = check magic */
+#define ACM_MAGIC  0x0001debc
+
+/* each offset in bytes from start of the struct they
+ * are part of */
+
+/* V3 of the policy buffer aded a version structure */
+struct acm_policy_version
+{
+    uint32_t major;
+    uint32_t minor;
+};
+
+
+/* each buffer consists of all policy information for
+ * the respective policy given in the policy code
+ *
+ * acm_policy_buffer, acm_chwall_policy_buffer,
+ * and acm_ste_policy_buffer need to stay 32-bit aligned
+ * because we create binary policies also with external
+ * tools that assume packed representations (e.g. the java tool)
+ */
+struct acm_policy_buffer {
+    uint32_t policy_version; /* ACM_POLICY_VERSION */
+    uint32_t magic;
+    uint32_t len;
+    uint32_t policy_reference_offset;
+    uint32_t primary_policy_code;
+    uint32_t primary_buffer_offset;
+    uint32_t secondary_policy_code;
+    uint32_t secondary_buffer_offset;
+    struct acm_policy_version xml_pol_version; /* add in V3 */
+};
+
+
+struct acm_policy_reference_buffer {
+    uint32_t len;
+};
+
+struct acm_chwall_policy_buffer {
+    uint32_t policy_version; /* ACM_CHWALL_VERSION */
+    uint32_t policy_code;
+    uint32_t chwall_max_types;
+    uint32_t chwall_max_ssidrefs;
+    uint32_t chwall_max_conflictsets;
+    uint32_t chwall_ssid_offset;
+    uint32_t chwall_conflict_sets_offset;
+    uint32_t chwall_running_types_offset;
+    uint32_t chwall_conflict_aggregate_offset;
+};
+
+struct acm_ste_policy_buffer {
+    uint32_t policy_version; /* ACM_STE_VERSION */
+    uint32_t policy_code;
+    uint32_t ste_max_types;
+    uint32_t ste_max_ssidrefs;
+    uint32_t ste_ssid_offset;
+};
+
+struct acm_stats_buffer {
+    uint32_t magic;
+    uint32_t len;
+    uint32_t primary_policy_code;
+    uint32_t primary_stats_offset;
+    uint32_t secondary_policy_code;
+    uint32_t secondary_stats_offset;
+};
+
+struct acm_ste_stats_buffer {
+    uint32_t ec_eval_count;
+    uint32_t gt_eval_count;
+    uint32_t ec_denied_count;
+    uint32_t gt_denied_count;
+    uint32_t ec_cachehit_count;
+    uint32_t gt_cachehit_count;
+};
+
+struct acm_ssid_buffer {
+    uint32_t len;
+    ssidref_t ssidref;
+    uint32_t policy_reference_offset;
+    uint32_t primary_policy_code;
+    uint32_t primary_max_types;
+    uint32_t primary_types_offset;
+    uint32_t secondary_policy_code;
+    uint32_t secondary_max_types;
+    uint32_t secondary_types_offset;
+};
+
+#endif
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/acm_ops.h linux-2.6.18-xen-3.3.0/include/xen/interface/acm_ops.h
--- linux-2.6.18.8/include/xen/interface/acm_ops.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/acm_ops.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,159 @@
+/*
+ * acm_ops.h: Xen access control module hypervisor commands
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Reiner Sailer <sailer@watson.ibm.com>
+ * Copyright (c) 2005,2006 International Business Machines Corporation.
+ */
+
+#ifndef __XEN_PUBLIC_ACM_OPS_H__
+#define __XEN_PUBLIC_ACM_OPS_H__
+
+#include "xen.h"
+#include "acm.h"
+
+/*
+ * Make sure you increment the interface version whenever you modify this file!
+ * This makes sure that old versions of acm tools will stop working in a
+ * well-defined way (rather than crashing the machine, for instance).
+ */
+#define ACM_INTERFACE_VERSION   0xAAAA000A
+
+/************************************************************************/
+
+/*
+ * Prototype for this hypercall is:
+ *  int acm_op(int cmd, void *args)
+ * @cmd  == ACMOP_??? (access control module operation).
+ * @args == Operation-specific extra arguments (NULL if none).
+ */
+
+
+#define ACMOP_setpolicy         1
+struct acm_setpolicy {
+    /* IN */
+    XEN_GUEST_HANDLE_64(void) pushcache;
+    uint32_t pushcache_size;
+};
+
+
+#define ACMOP_getpolicy         2
+struct acm_getpolicy {
+    /* IN */
+    XEN_GUEST_HANDLE_64(void) pullcache;
+    uint32_t pullcache_size;
+};
+
+
+#define ACMOP_dumpstats         3
+struct acm_dumpstats {
+    /* IN */
+    XEN_GUEST_HANDLE_64(void) pullcache;
+    uint32_t pullcache_size;
+};
+
+
+#define ACMOP_getssid           4
+#define ACM_GETBY_ssidref  1
+#define ACM_GETBY_domainid 2
+struct acm_getssid {
+    /* IN */
+    uint32_t get_ssid_by; /* ACM_GETBY_* */
+    union {
+        domaintype_t domainid;
+        ssidref_t    ssidref;
+    } id;
+    XEN_GUEST_HANDLE_64(void) ssidbuf;
+    uint32_t ssidbuf_size;
+};
+
+#define ACMOP_getdecision      5
+struct acm_getdecision {
+    /* IN */
+    uint32_t get_decision_by1; /* ACM_GETBY_* */
+    uint32_t get_decision_by2; /* ACM_GETBY_* */
+    union {
+        domaintype_t domainid;
+        ssidref_t    ssidref;
+    } id1;
+    union {
+        domaintype_t domainid;
+        ssidref_t    ssidref;
+    } id2;
+    uint32_t hook;
+    /* OUT */
+    uint32_t acm_decision;
+};
+
+
+#define ACMOP_chgpolicy        6
+struct acm_change_policy {
+    /* IN */
+    XEN_GUEST_HANDLE_64(void) policy_pushcache;
+    uint32_t policy_pushcache_size;
+    XEN_GUEST_HANDLE_64(void) del_array;
+    uint32_t delarray_size;
+    XEN_GUEST_HANDLE_64(void) chg_array;
+    uint32_t chgarray_size;
+    /* OUT */
+    /* array with error code */
+    XEN_GUEST_HANDLE_64(void) err_array;
+    uint32_t errarray_size;
+};
+
+#define ACMOP_relabeldoms       7
+struct acm_relabel_doms {
+    /* IN */
+    XEN_GUEST_HANDLE_64(void) relabel_map;
+    uint32_t relabel_map_size;
+    /* OUT */
+    XEN_GUEST_HANDLE_64(void) err_array;
+    uint32_t errarray_size;
+};
+
+/* future interface to Xen */
+struct xen_acmctl {
+    uint32_t cmd;
+    uint32_t interface_version;
+    union {
+        struct acm_setpolicy     setpolicy;
+        struct acm_getpolicy     getpolicy;
+        struct acm_dumpstats     dumpstats;
+        struct acm_getssid       getssid;
+        struct acm_getdecision   getdecision;
+        struct acm_change_policy change_policy;
+        struct acm_relabel_doms  relabel_doms;
+    } u;
+};
+
+typedef struct xen_acmctl xen_acmctl_t;
+DEFINE_XEN_GUEST_HANDLE(xen_acmctl_t);
+
+#endif /* __XEN_PUBLIC_ACM_OPS_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/arch-ia64/debug_op.h linux-2.6.18-xen-3.3.0/include/xen/interface/arch-ia64/debug_op.h
--- linux-2.6.18.8/include/xen/interface/arch-ia64/debug_op.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/arch-ia64/debug_op.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,96 @@
+/******************************************************************************
+ * debug_op.h
+ *
+ * Copyright (c) 2007 Tristan Gingold <tgingold@free.fr>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+
+#ifndef __XEN_PUBLIC_IA64_DEBUG_OP_H__
+#define __XEN_PUBLIC_IA64_DEBUG_OP_H__
+
+/* Set/Get extra conditions to break.  */
+#define XEN_IA64_DEBUG_OP_SET_FLAGS 1
+#define XEN_IA64_DEBUG_OP_GET_FLAGS 2
+
+/* Break on kernel single step.  */
+#define XEN_IA64_DEBUG_ON_KERN_SSTEP   (1 << 0)
+
+/* Break on kernel debug (breakpoint or watch point).  */
+#define XEN_IA64_DEBUG_ON_KERN_DEBUG   (1 << 1)
+
+/* Break on kernel taken branch.  */
+#define XEN_IA64_DEBUG_ON_KERN_TBRANCH (1 << 2)
+
+/* Break on interrupt injection.  */
+#define XEN_IA64_DEBUG_ON_EXTINT       (1 << 3)
+
+/* Break on interrupt injection.  */
+#define XEN_IA64_DEBUG_ON_EXCEPT       (1 << 4)
+
+/* Break on event injection.  */
+#define XEN_IA64_DEBUG_ON_EVENT        (1 << 5)
+
+/* Break on privop/virtualized instruction (slow path only).  */
+#define XEN_IA64_DEBUG_ON_PRIVOP       (1 << 6)
+
+/* Break on emulated PAL call (at entry).  */
+#define XEN_IA64_DEBUG_ON_PAL          (1 << 7)
+
+/* Break on emulated SAL call (at entry).  */
+#define XEN_IA64_DEBUG_ON_SAL          (1 << 8)
+
+/* Break on emulated EFI call (at entry).  */
+#define XEN_IA64_DEBUG_ON_EFI          (1 << 9)
+
+/* Break on rfi emulation (slow path only, before exec).  */
+#define XEN_IA64_DEBUG_ON_RFI          (1 << 10)
+
+/* Break on address translation switch.  */
+#define XEN_IA64_DEBUG_ON_MMU          (1 << 11)
+
+/* Break on bad guest physical address.  */
+#define XEN_IA64_DEBUG_ON_BAD_MPA      (1 << 12)
+
+/* Force psr.ss bit.  */
+#define XEN_IA64_DEBUG_FORCE_SS        (1 << 13)
+
+/* Force psr.db bit.  */
+#define XEN_IA64_DEBUG_FORCE_DB        (1 << 14)
+
+/* Break on ITR/PTR.  */
+#define XEN_IA64_DEBUG_ON_TR           (1 << 15)
+
+/* Break on ITC/PTC.L/PTC.G/PTC.GA.  */
+#define XEN_IA64_DEBUG_ON_TC           (1 << 16)
+
+/* Get translation cache.  */
+#define XEN_IA64_DEBUG_OP_GET_TC   3
+
+/* Translate virtual address to guest physical address.  */
+#define XEN_IA64_DEBUG_OP_TRANSLATE 4
+
+union xen_ia64_debug_op {
+    uint64_t flags;
+    struct xen_ia64_debug_vtlb {
+        uint64_t nbr;                             /* IN/OUT */
+        XEN_GUEST_HANDLE_64(ia64_tr_entry_t) tr;  /* IN/OUT */
+    } vtlb;
+};
+typedef union xen_ia64_debug_op xen_ia64_debug_op_t;
+DEFINE_XEN_GUEST_HANDLE(xen_ia64_debug_op_t);
+
+#endif /* __XEN_PUBLIC_IA64_DEBUG_OP_H__ */
diff -rpuN linux-2.6.18.8/include/xen/interface/arch-ia64/hvm/memmap.h linux-2.6.18-xen-3.3.0/include/xen/interface/arch-ia64/hvm/memmap.h
--- linux-2.6.18.8/include/xen/interface/arch-ia64/hvm/memmap.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/arch-ia64/hvm/memmap.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,88 @@
+/******************************************************************************
+ * memmap.h
+ *
+ * Copyright (c) 2008 Tristan Gingold <tgingold AT free fr>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+
+#ifndef __XEN_PUBLIC_HVM_MEMMAP_IA64_H__
+#define __XEN_PUBLIC_HVM_MEMMAP_IA64_H__
+
+#define MEM_G  (1UL << 30)
+#define MEM_M  (1UL << 20)
+#define MEM_K  (1UL << 10)
+
+/* Guest physical address of IO ports space.  */
+#define MMIO_START  (3 * MEM_G)
+#define MMIO_SIZE   (512 * MEM_M)
+
+#define VGA_IO_START  0xA0000UL
+#define VGA_IO_SIZE   0x20000
+
+#define LEGACY_IO_START  (MMIO_START + MMIO_SIZE)
+#define LEGACY_IO_SIZE   (64 * MEM_M)
+
+#define IO_PAGE_START  (LEGACY_IO_START + LEGACY_IO_SIZE)
+#define IO_PAGE_SIZE   XEN_PAGE_SIZE
+
+#define STORE_PAGE_START  (IO_PAGE_START + IO_PAGE_SIZE)
+#define STORE_PAGE_SIZE   XEN_PAGE_SIZE
+
+#define BUFFER_IO_PAGE_START  (STORE_PAGE_START + STORE_PAGE_SIZE)
+#define BUFFER_IO_PAGE_SIZE   XEN_PAGE_SIZE
+
+#define BUFFER_PIO_PAGE_START  (BUFFER_IO_PAGE_START + BUFFER_IO_PAGE_SIZE)
+#define BUFFER_PIO_PAGE_SIZE   XEN_PAGE_SIZE
+
+#define IO_SAPIC_START  0xfec00000UL
+#define IO_SAPIC_SIZE   0x100000
+
+#define PIB_START  0xfee00000UL
+#define PIB_SIZE   0x200000
+
+#define GFW_START  (4 * MEM_G - 16 * MEM_M)
+#define GFW_SIZE   (16 * MEM_M)
+
+/* domVTI */
+#define GPFN_FRAME_BUFFER  0x1 /* VGA framebuffer */
+#define GPFN_LOW_MMIO      0x2 /* Low MMIO range */
+#define GPFN_PIB           0x3 /* PIB base */
+#define GPFN_IOSAPIC       0x4 /* IOSAPIC base */
+#define GPFN_LEGACY_IO     0x5 /* Legacy I/O base */
+#define GPFN_HIGH_MMIO     0x6 /* High MMIO range */
+
+/* Nvram belongs to GFW memory space  */
+#define NVRAM_SIZE   (MEM_K * 64)
+#define NVRAM_START  (GFW_START + 10 * MEM_M)
+
+#define NVRAM_VALID_SIG  0x4650494e45584948 /* "HIXENIPF" */
+struct nvram_save_addr {
+    unsigned long addr;
+    unsigned long signature;
+};
+
+#endif /* __XEN_PUBLIC_HVM_MEMMAP_IA64_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/arch-ia64/hvm/save.h linux-2.6.18-xen-3.3.0/include/xen/interface/arch-ia64/hvm/save.h
--- linux-2.6.18.8/include/xen/interface/arch-ia64/hvm/save.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/arch-ia64/hvm/save.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,201 @@
+/******************************************************************************
+ * save_types.h
+ *
+ * Copyright (c) 2007 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+
+#ifndef __XEN_PUBLIC_HVM_SAVE_IA64_H__
+#define __XEN_PUBLIC_HVM_SAVE_IA64_H__
+
+#include <public/hvm/save.h>
+#include <public/arch-ia64.h>
+
+/* 
+ * Save/restore header: general info about the save file. 
+ */
+
+/* x86 uses 0x54381286 */
+#define HVM_FILE_MAGIC   0x343641492f6e6558UL   /* "Xen/IA64" */
+#define HVM_FILE_VERSION 0x0000000000000001UL
+
+struct hvm_save_header {
+    uint64_t magic;             /* Must be HVM_FILE_MAGIC */
+    uint64_t version;           /* File format version */
+    uint64_t changeset;         /* Version of Xen that saved this file */
+    uint64_t cpuid[5];          /* CPUID[0x01][%eax] on the saving machine */
+};
+
+DECLARE_HVM_SAVE_TYPE(HEADER, 1, struct hvm_save_header);
+
+/*
+ * CPU
+ */
+struct hvm_hw_ia64_cpu {
+    uint64_t    ipsr;
+};
+DECLARE_HVM_SAVE_TYPE(CPU, 2, struct hvm_hw_ia64_cpu);
+
+/*
+ * CPU
+ */
+struct hvm_hw_ia64_vpd {
+    struct vpd      vpd;
+};
+DECLARE_HVM_SAVE_TYPE(VPD, 3, struct hvm_hw_ia64_vpd);
+
+/*
+ * device dependency
+ * vacpi => viosapic => vlsapic
+ */
+/*
+ * vlsapic
+ */
+struct hvm_hw_ia64_vlsapic {
+    uint64_t insvc[4];
+    uint64_t vhpi; // ??? should this be saved in vpd
+    uint8_t xtp;
+    uint8_t pal_init_pending;
+    uint8_t pad[2];
+};
+DECLARE_HVM_SAVE_TYPE(VLSAPIC, 4, struct hvm_hw_ia64_vlsapic);
+/* set
+ * unconditionaly set v->arch.irq_new_peding = 1 
+ * unconditionaly set v->arch.irq_new_condition = 0
+ */
+
+/*
+ * vtime
+ */
+/* itc, itm, itv are saved by arch vcpu context */
+struct hvm_hw_ia64_vtime {
+    uint64_t itc;
+    uint64_t itm;
+
+    uint64_t last_itc;
+    uint64_t pending;
+};
+DECLARE_HVM_SAVE_TYPE(VTIME, 5, struct hvm_hw_ia64_vtime);
+/*
+ * calculate v->vtm.vtm_offset
+ * ??? Or should vtm_offset be set by leave_hypervisor_tail()?
+ * start vtm_timer if necessary by vtm_set_itm().
+ * ??? Or should vtm_timer be set by leave_hypervisor_tail()?
+ *
+ * ??? or should be done by schedule_tail()
+ *        => schedule_tail() should do.
+ */
+
+/*
+ * viosapic
+ */
+#define VIOSAPIC_NUM_PINS     48
+
+union viosapic_rte
+{
+    uint64_t bits;
+    struct {
+        uint8_t vector;
+
+        uint8_t delivery_mode  : 3;
+        uint8_t reserve1       : 1;
+        uint8_t delivery_status: 1;
+        uint8_t polarity       : 1;
+        uint8_t reserve2       : 1;
+        uint8_t trig_mode      : 1;
+
+        uint8_t mask           : 1;
+        uint8_t reserve3       : 7;
+
+        uint8_t reserved[3];
+        uint16_t dest_id;
+    }; 
+};
+
+struct hvm_hw_ia64_viosapic {
+    uint64_t    irr;
+    uint64_t    isr;
+    uint32_t    ioregsel;
+    uint32_t    pad;
+    uint64_t    lowest_vcpu_id;
+    uint64_t    base_address;
+    union viosapic_rte  redirtbl[VIOSAPIC_NUM_PINS];
+};
+DECLARE_HVM_SAVE_TYPE(VIOSAPIC, 6, struct hvm_hw_ia64_viosapic);
+  
+/*
+ * vacpi
+ * PM timer
+ */
+struct vacpi_regs {
+    union {
+        struct {
+            uint32_t pm1a_sts:16;/* PM1a_EVT_BLK.PM1a_STS: status register */
+            uint32_t pm1a_en:16; /* PM1a_EVT_BLK.PM1a_EN: enable register */
+        };
+        uint32_t evt_blk;
+    };
+    uint32_t tmr_val;   /* PM_TMR_BLK.TMR_VAL: 32bit free-running counter */
+};
+
+struct hvm_hw_ia64_vacpi {
+    struct vacpi_regs   regs;
+};
+DECLARE_HVM_SAVE_TYPE(VACPI, 7, struct hvm_hw_ia64_vacpi);
+/* update last_gtime and setup timer of struct vacpi */
+
+/*
+ * opt_feature: identity mapping of region 4, 5 and 7.
+ * With the c/s 16396:d2935f9c217f of xen-ia64-devel.hg,
+ * opt_feature hypercall supports only region 4,5,7 identity mappings.
+ * structure hvm_hw_ia64_identity_mappings only supports them.
+ * The new structure, struct hvm_hw_ia64_identity_mappings, is created to
+ * avoid to keep up with change of the xen/ia64 internal structure, struct
+ * opt_feature.
+ *
+ * If it is enhanced in the future, new structure will be created.
+ */
+struct hvm_hw_ia64_identity_mapping {
+    uint64_t on;        /* on/off */
+    uint64_t pgprot;    /* The page protection bit mask of the pte. */
+    uint64_t key;       /* A protection key. */
+};
+
+struct hvm_hw_ia64_identity_mappings {
+    struct hvm_hw_ia64_identity_mapping im_reg4;/* Region 4 identity mapping */
+    struct hvm_hw_ia64_identity_mapping im_reg5;/* Region 5 identity mapping */
+    struct hvm_hw_ia64_identity_mapping im_reg7;/* Region 7 identity mapping */
+};
+DECLARE_HVM_SAVE_TYPE(OPT_FEATURE_IDENTITY_MAPPINGS, 8, struct hvm_hw_ia64_identity_mappings);
+
+/* 
+ * Largest type-code in use
+ */
+#define HVM_SAVE_CODE_MAX       8
+
+#endif /* __XEN_PUBLIC_HVM_SAVE_IA64_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/arch-ia64/sioemu.h linux-2.6.18-xen-3.3.0/include/xen/interface/arch-ia64/sioemu.h
--- linux-2.6.18.8/include/xen/interface/arch-ia64/sioemu.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/arch-ia64/sioemu.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,89 @@
+/******************************************************************************
+ * sioemu.h
+ *
+ * Copyright (c) 2008 Tristan Gingold <tgingold@free.fr>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+
+#ifndef __XEN_PUBLIC_IA64_SIOEMU_H__
+#define __XEN_PUBLIC_IA64_SIOEMU_H__
+
+/* SIOEMU specific hypercalls.
+   The numbers are the minor part of FW_HYPERCALL_SIOEMU.  */
+
+/* Defines the callback entry point.  r8=ip, r9=data.
+   Must be called per-vcpu.  */
+#define SIOEMU_HYPERCALL_SET_CALLBACK 0x01
+
+/* Finish sioemu fw initialization and start firmware.  r8=ip.  */
+#define SIOEMU_HYPERCALL_START_FW 0x02
+
+/* Add IO pages in physmap.  */
+#define SIOEMU_HYPERCALL_ADD_IO_PHYSMAP 0x03
+
+/* Get wallclock time.  */
+#define SIOEMU_HYPERCALL_GET_TIME 0x04
+
+/* Flush cache.  */
+#define SIOEMU_HYPERCALL_FLUSH_CACHE 0x07
+
+/* Get freq base.  */
+#define SIOEMU_HYPERCALL_FREQ_BASE 0x08
+
+/* Return from callback.  */
+#define SIOEMU_HYPERCALL_CALLBACK_RETURN 0x09
+
+/* Deliver an interrupt.  */
+#define SIOEMU_HYPERCALL_DELIVER_INT 0x0a
+
+/* SIOEMU callback reason.  */
+
+/* An event (from event channel) has to be delivered.  */
+#define SIOEMU_CB_EVENT       0x00
+
+/* Emulate an IO access.  */
+#define SIOEMU_CB_IO_EMULATE  0x01
+
+/* An IPI is sent to a dead vcpu.  */
+#define SIOEMU_CB_WAKEUP_VCPU 0x02
+
+/* A SAL hypercall is executed.  */
+#define SIOEMU_CB_SAL_ASSIST  0x03
+
+#ifndef __ASSEMBLY__
+struct sioemu_callback_info {
+    /* Saved registers.  */
+    unsigned long ip;
+    unsigned long psr;
+    unsigned long ifs;
+    unsigned long nats;
+    unsigned long r8;
+    unsigned long r9;
+    unsigned long r10;
+    unsigned long r11;
+
+    /* Callback parameters.  */
+    unsigned long cause;
+    unsigned long arg0;
+    unsigned long arg1;
+    unsigned long arg2;
+    unsigned long arg3;
+    unsigned long _pad2[2];
+    unsigned long r2;
+};
+#endif /* __ASSEMBLY__ */
+#endif /* __XEN_PUBLIC_IA64_SIOEMU_H__ */
diff -rpuN linux-2.6.18.8/include/xen/interface/arch-ia64.h linux-2.6.18-xen-3.3.0/include/xen/interface/arch-ia64.h
--- linux-2.6.18.8/include/xen/interface/arch-ia64.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/arch-ia64.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,621 @@
+/******************************************************************************
+ * arch-ia64/hypervisor-if.h
+ * 
+ * Guest OS interface to IA64 Xen.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "xen.h"
+
+#ifndef __HYPERVISOR_IF_IA64_H__
+#define __HYPERVISOR_IF_IA64_H__
+
+#if !defined(__GNUC__) || defined(__STRICT_ANSI__)
+#error "Anonymous structs/unions are a GNU extension."
+#endif
+
+/* Structural guest handles introduced in 0x00030201. */
+#if __XEN_INTERFACE_VERSION__ >= 0x00030201
+#define ___DEFINE_XEN_GUEST_HANDLE(name, type) \
+    typedef struct { type *p; } __guest_handle_ ## name
+#else
+#define ___DEFINE_XEN_GUEST_HANDLE(name, type) \
+    typedef type * __guest_handle_ ## name
+#endif
+
+#define __DEFINE_XEN_GUEST_HANDLE(name, type) \
+    ___DEFINE_XEN_GUEST_HANDLE(name, type);   \
+    ___DEFINE_XEN_GUEST_HANDLE(const_##name, const type)
+
+#define DEFINE_XEN_GUEST_HANDLE(name)   __DEFINE_XEN_GUEST_HANDLE(name, name)
+#define XEN_GUEST_HANDLE(name)          __guest_handle_ ## name
+#define XEN_GUEST_HANDLE_64(name)       XEN_GUEST_HANDLE(name)
+#define uint64_aligned_t                uint64_t
+#define set_xen_guest_handle(hnd, val)  do { (hnd).p = val; } while (0)
+#ifdef __XEN_TOOLS__
+#define get_xen_guest_handle(val, hnd)  do { val = (hnd).p; } while (0)
+#endif
+
+#ifndef __ASSEMBLY__
+typedef unsigned long xen_pfn_t;
+#define PRI_xen_pfn "lx"
+#endif
+
+/* Arch specific VIRQs definition */
+#define VIRQ_ITC        VIRQ_ARCH_0 /* V. Virtual itc timer */
+#define VIRQ_MCA_CMC    VIRQ_ARCH_1 /* MCA cmc interrupt */
+#define VIRQ_MCA_CPE    VIRQ_ARCH_2 /* MCA cpe interrupt */
+
+/* Maximum number of virtual CPUs in multi-processor guests. */
+/* WARNING: before changing this, check that shared_info fits on a page */
+#define MAX_VIRT_CPUS 64
+
+/* IO ports location for PV.  */
+#define IO_PORTS_PADDR          0x00000ffffc000000UL
+#define IO_PORTS_SIZE           0x0000000004000000UL
+
+#ifndef __ASSEMBLY__
+
+typedef unsigned long xen_ulong_t;
+
+#ifdef __XEN_TOOLS__
+#define XEN_PAGE_SIZE XC_PAGE_SIZE
+#else
+#define XEN_PAGE_SIZE PAGE_SIZE
+#endif
+
+#define INVALID_MFN       (~0UL)
+
+struct pt_fpreg {
+    union {
+        unsigned long bits[2];
+        long double __dummy;    /* force 16-byte alignment */
+    } u;
+};
+
+union vac {
+    unsigned long value;
+    struct {
+        int a_int:1;
+        int a_from_int_cr:1;
+        int a_to_int_cr:1;
+        int a_from_psr:1;
+        int a_from_cpuid:1;
+        int a_cover:1;
+        int a_bsw:1;
+        long reserved:57;
+    };
+};
+typedef union vac vac_t;
+
+union vdc {
+    unsigned long value;
+    struct {
+        int d_vmsw:1;
+        int d_extint:1;
+        int d_ibr_dbr:1;
+        int d_pmc:1;
+        int d_to_pmd:1;
+        int d_itm:1;
+        long reserved:58;
+    };
+};
+typedef union vdc vdc_t;
+
+struct mapped_regs {
+    union vac   vac;
+    union vdc   vdc;
+    unsigned long  virt_env_vaddr;
+    unsigned long  reserved1[29];
+    unsigned long  vhpi;
+    unsigned long  reserved2[95];
+    union {
+        unsigned long  vgr[16];
+        unsigned long bank1_regs[16]; // bank1 regs (r16-r31) when bank0 active
+    };
+    union {
+        unsigned long  vbgr[16];
+        unsigned long bank0_regs[16]; // bank0 regs (r16-r31) when bank1 active
+    };
+    unsigned long  vnat;
+    unsigned long  vbnat;
+    unsigned long  vcpuid[5];
+    unsigned long  reserved3[11];
+    unsigned long  vpsr;
+    unsigned long  vpr;
+    unsigned long  reserved4[76];
+    union {
+        unsigned long  vcr[128];
+        struct {
+            unsigned long dcr;  // CR0
+            unsigned long itm;
+            unsigned long iva;
+            unsigned long rsv1[5];
+            unsigned long pta;  // CR8
+            unsigned long rsv2[7];
+            unsigned long ipsr;  // CR16
+            unsigned long isr;
+            unsigned long rsv3;
+            unsigned long iip;
+            unsigned long ifa;
+            unsigned long itir;
+            unsigned long iipa;
+            unsigned long ifs;
+            unsigned long iim;  // CR24
+            unsigned long iha;
+            unsigned long rsv4[38];
+            unsigned long lid;  // CR64
+            unsigned long ivr;
+            unsigned long tpr;
+            unsigned long eoi;
+            unsigned long irr[4];
+            unsigned long itv;  // CR72
+            unsigned long pmv;
+            unsigned long cmcv;
+            unsigned long rsv5[5];
+            unsigned long lrr0;  // CR80
+            unsigned long lrr1;
+            unsigned long rsv6[46];
+        };
+    };
+    union {
+        unsigned long  reserved5[128];
+        struct {
+            unsigned long precover_ifs;
+            unsigned long unat;  // not sure if this is needed until NaT arch is done
+            int interrupt_collection_enabled; // virtual psr.ic
+            /* virtual interrupt deliverable flag is evtchn_upcall_mask in
+             * shared info area now. interrupt_mask_addr is the address
+             * of evtchn_upcall_mask for current vcpu
+             */
+            unsigned char *interrupt_mask_addr;
+            int pending_interruption;
+            unsigned char vpsr_pp;
+            unsigned char vpsr_dfh;
+            unsigned char hpsr_dfh;
+            unsigned char hpsr_mfh;
+            unsigned long reserved5_1[4];
+            int metaphysical_mode; // 1 = use metaphys mapping, 0 = use virtual
+            int banknum; // 0 or 1, which virtual register bank is active
+            unsigned long rrs[8]; // region registers
+            unsigned long krs[8]; // kernel registers
+            unsigned long tmp[16]; // temp registers (e.g. for hyperprivops)
+        };
+    };
+};
+typedef struct mapped_regs mapped_regs_t;
+
+struct vpd {
+    struct mapped_regs vpd_low;
+    unsigned long  reserved6[3456];
+    unsigned long  vmm_avail[128];
+    unsigned long  reserved7[4096];
+};
+typedef struct vpd vpd_t;
+
+struct arch_vcpu_info {
+};
+typedef struct arch_vcpu_info arch_vcpu_info_t;
+
+/*
+ * This structure is used for magic page in domain pseudo physical address
+ * space and the result of XENMEM_machine_memory_map.
+ * As the XENMEM_machine_memory_map result,
+ * xen_memory_map::nr_entries indicates the size in bytes 
+ * including struct xen_ia64_memmap_info. Not the number of entries.
+ */
+struct xen_ia64_memmap_info {
+    uint64_t efi_memmap_size;       /* size of EFI memory map */
+    uint64_t efi_memdesc_size;      /* size of an EFI memory map descriptor */
+    uint32_t efi_memdesc_version;   /* memory descriptor version */
+    void *memdesc[0];               /* array of efi_memory_desc_t */
+};
+typedef struct xen_ia64_memmap_info xen_ia64_memmap_info_t;
+
+struct arch_shared_info {
+    /* PFN of the start_info page.  */
+    unsigned long start_info_pfn;
+
+    /* Interrupt vector for event channel.  */
+    int evtchn_vector;
+
+    /* PFN of memmap_info page */
+    unsigned int memmap_info_num_pages;/* currently only = 1 case is
+                                          supported. */
+    unsigned long memmap_info_pfn;
+
+    uint64_t pad[31];
+};
+typedef struct arch_shared_info arch_shared_info_t;
+
+typedef unsigned long xen_callback_t;
+
+struct ia64_tr_entry {
+    unsigned long pte;
+    unsigned long itir;
+    unsigned long vadr;
+    unsigned long rid;
+};
+typedef struct ia64_tr_entry ia64_tr_entry_t;
+DEFINE_XEN_GUEST_HANDLE(ia64_tr_entry_t);
+
+struct vcpu_tr_regs {
+    struct ia64_tr_entry itrs[12];
+    struct ia64_tr_entry dtrs[12];
+};
+
+union vcpu_ar_regs {
+    unsigned long ar[128];
+    struct {
+        unsigned long kr[8];
+        unsigned long rsv1[8];
+        unsigned long rsc;
+        unsigned long bsp;
+        unsigned long bspstore;
+        unsigned long rnat;
+        unsigned long rsv2;
+        unsigned long fcr;
+        unsigned long rsv3[2];
+        unsigned long eflag;
+        unsigned long csd;
+        unsigned long ssd;
+        unsigned long cflg;
+        unsigned long fsr;
+        unsigned long fir;
+        unsigned long fdr;
+        unsigned long rsv4;
+        unsigned long ccv; /* 32 */
+        unsigned long rsv5[3];
+        unsigned long unat;
+        unsigned long rsv6[3];
+        unsigned long fpsr;
+        unsigned long rsv7[3];
+        unsigned long itc;
+        unsigned long rsv8[3];
+        unsigned long ign1[16];
+        unsigned long pfs; /* 64 */
+        unsigned long lc;
+        unsigned long ec;
+        unsigned long rsv9[45];
+        unsigned long ign2[16];
+    };
+};
+
+union vcpu_cr_regs {
+    unsigned long cr[128];
+    struct {
+        unsigned long dcr;  // CR0
+        unsigned long itm;
+        unsigned long iva;
+        unsigned long rsv1[5];
+        unsigned long pta;  // CR8
+        unsigned long rsv2[7];
+        unsigned long ipsr;  // CR16
+        unsigned long isr;
+        unsigned long rsv3;
+        unsigned long iip;
+        unsigned long ifa;
+        unsigned long itir;
+        unsigned long iipa;
+        unsigned long ifs;
+        unsigned long iim;  // CR24
+        unsigned long iha;
+        unsigned long rsv4[38];
+        unsigned long lid;  // CR64
+        unsigned long ivr;
+        unsigned long tpr;
+        unsigned long eoi;
+        unsigned long irr[4];
+        unsigned long itv;  // CR72
+        unsigned long pmv;
+        unsigned long cmcv;
+        unsigned long rsv5[5];
+        unsigned long lrr0;  // CR80
+        unsigned long lrr1;
+        unsigned long rsv6[46];
+    };
+};
+
+struct vcpu_guest_context_regs {
+        unsigned long r[32];
+        unsigned long b[8];
+        unsigned long bank[16];
+        unsigned long ip;
+        unsigned long psr;
+        unsigned long cfm;
+        unsigned long pr;
+        unsigned int nats; /* NaT bits for r1-r31.  */
+        unsigned int bnats; /* Nat bits for banked registers.  */
+        union vcpu_ar_regs ar;
+        union vcpu_cr_regs cr;
+        struct pt_fpreg f[128];
+        unsigned long dbr[8];
+        unsigned long ibr[8];
+        unsigned long rr[8];
+        unsigned long pkr[16];
+
+        /* FIXME: cpuid,pmd,pmc */
+
+        unsigned long xip;
+        unsigned long xpsr;
+        unsigned long xfs;
+        unsigned long xr[4];
+
+        struct vcpu_tr_regs tr;
+
+        /* Physical registers in case of debug event.  */
+        unsigned long excp_iipa;
+        unsigned long excp_ifa;
+        unsigned long excp_isr;
+        unsigned int excp_vector;
+
+        /*
+         * The rbs is intended to be the image of the stacked registers still
+         * in the cpu (not yet stored in memory).  It is laid out as if it
+         * were written in memory at a 512 (64*8) aligned address + offset.
+         * rbs_voff is (offset / 8).  rbs_nat contains NaT bits for the
+         * remaining rbs registers.  rbs_rnat contains NaT bits for in memory
+         * rbs registers.
+         * Note: loadrs is 2**14 bytes == 2**11 slots.
+         */
+        unsigned int rbs_voff;
+        unsigned long rbs[2048];
+        unsigned long rbs_rnat;
+
+        /*
+         * RSE.N_STACKED_PHYS via PAL_RSE_INFO
+         * Strictly this isn't cpu context, but this value is necessary
+         * for domain save/restore. So is here.
+         */
+        unsigned long num_phys_stacked;
+};
+
+struct vcpu_guest_context {
+#define VGCF_EXTRA_REGS (1UL << 1)	/* Set extra regs.  */
+#define VGCF_SET_CR_IRR (1UL << 2)	/* Set cr_irr[0:3]. */
+#define VGCF_online     (1UL << 3)  /* make this vcpu online */
+    unsigned long flags;       /* VGCF_* flags */
+
+    struct vcpu_guest_context_regs regs;
+
+    unsigned long event_callback_ip;
+
+    /* xen doesn't share privregs pages with hvm domain so that this member
+     * doesn't make sense for hvm domain.
+     * ~0UL is already used for INVALID_P2M_ENTRY. */
+#define VGC_PRIVREGS_HVM       (~(-2UL))
+    unsigned long privregs_pfn;
+};
+typedef struct vcpu_guest_context vcpu_guest_context_t;
+DEFINE_XEN_GUEST_HANDLE(vcpu_guest_context_t);
+
+/* dom0 vp op */
+#define __HYPERVISOR_ia64_dom0vp_op     __HYPERVISOR_arch_0
+/*  Map io space in machine address to dom0 physical address space.
+    Currently physical assigned address equals to machine address.  */
+#define IA64_DOM0VP_ioremap             0
+
+/* Convert a pseudo physical page frame number to the corresponding
+   machine page frame number. If no page is assigned, INVALID_MFN or
+   GPFN_INV_MASK is returned depending on domain's non-vti/vti mode.  */
+#define IA64_DOM0VP_phystomach          1
+
+/* Convert a machine page frame number to the corresponding pseudo physical
+   page frame number of the caller domain.  */
+#define IA64_DOM0VP_machtophys          3
+
+/* Reserved for future use.  */
+#define IA64_DOM0VP_iounmap             4
+
+/* Unmap and free pages contained in the specified pseudo physical region.  */
+#define IA64_DOM0VP_zap_physmap         5
+
+/* Assign machine page frame to dom0's pseudo physical address space.  */
+#define IA64_DOM0VP_add_physmap         6
+
+/* expose the p2m table into domain */
+#define IA64_DOM0VP_expose_p2m          7
+
+/* xen perfmon */
+#define IA64_DOM0VP_perfmon             8
+
+/* gmfn version of IA64_DOM0VP_add_physmap */
+#define IA64_DOM0VP_add_physmap_with_gmfn       9
+
+/* get fpswa revision */
+#define IA64_DOM0VP_fpswa_revision      10
+
+/* Add an I/O port space range */
+#define IA64_DOM0VP_add_io_space        11
+
+/* expose the foreign domain's p2m table into privileged domain */
+#define IA64_DOM0VP_expose_foreign_p2m  12
+#define         IA64_DOM0VP_EFP_ALLOC_PTE       0x1 /* allocate p2m table */
+
+/* unexpose the foreign domain's p2m table into privileged domain */
+#define IA64_DOM0VP_unexpose_foreign_p2m        13
+
+// flags for page assignement to pseudo physical address space
+#define _ASSIGN_readonly                0
+#define ASSIGN_readonly                 (1UL << _ASSIGN_readonly)
+#define ASSIGN_writable                 (0UL << _ASSIGN_readonly) // dummy flag
+/* Internal only: memory attribute must be WC/UC/UCE.  */
+#define _ASSIGN_nocache                 1
+#define ASSIGN_nocache                  (1UL << _ASSIGN_nocache)
+// tlb tracking
+#define _ASSIGN_tlb_track               2
+#define ASSIGN_tlb_track                (1UL << _ASSIGN_tlb_track)
+/* Internal only: associated with PGC_allocated bit */
+#define _ASSIGN_pgc_allocated           3
+#define ASSIGN_pgc_allocated            (1UL << _ASSIGN_pgc_allocated)
+/* Page is an IO page.  */
+#define _ASSIGN_io                      4
+#define ASSIGN_io                       (1UL << _ASSIGN_io)
+
+/* This structure has the same layout of struct ia64_boot_param, defined in
+   <asm/system.h>.  It is redefined here to ease use.  */
+struct xen_ia64_boot_param {
+	unsigned long command_line;	/* physical address of cmd line args */
+	unsigned long efi_systab;	/* physical address of EFI system table */
+	unsigned long efi_memmap;	/* physical address of EFI memory map */
+	unsigned long efi_memmap_size;	/* size of EFI memory map */
+	unsigned long efi_memdesc_size;	/* size of an EFI memory map descriptor */
+	unsigned int  efi_memdesc_version;	/* memory descriptor version */
+	struct {
+		unsigned short num_cols;	/* number of columns on console.  */
+		unsigned short num_rows;	/* number of rows on console.  */
+		unsigned short orig_x;	/* cursor's x position */
+		unsigned short orig_y;	/* cursor's y position */
+	} console_info;
+	unsigned long fpswa;		/* physical address of the fpswa interface */
+	unsigned long initrd_start;
+	unsigned long initrd_size;
+	unsigned long domain_start;	/* va where the boot time domain begins */
+	unsigned long domain_size;	/* how big is the boot domain */
+};
+
+#endif /* !__ASSEMBLY__ */
+
+/* Size of the shared_info area (this is not related to page size).  */
+#define XSI_SHIFT			14
+#define XSI_SIZE			(1 << XSI_SHIFT)
+/* Log size of mapped_regs area (64 KB - only 4KB is used).  */
+#define XMAPPEDREGS_SHIFT		12
+#define XMAPPEDREGS_SIZE		(1 << XMAPPEDREGS_SHIFT)
+/* Offset of XASI (Xen arch shared info) wrt XSI_BASE.  */
+#define XMAPPEDREGS_OFS			XSI_SIZE
+
+/* Hyperprivops.  */
+#define HYPERPRIVOP_START		0x1
+#define HYPERPRIVOP_RFI			(HYPERPRIVOP_START + 0x0)
+#define HYPERPRIVOP_RSM_DT		(HYPERPRIVOP_START + 0x1)
+#define HYPERPRIVOP_SSM_DT		(HYPERPRIVOP_START + 0x2)
+#define HYPERPRIVOP_COVER		(HYPERPRIVOP_START + 0x3)
+#define HYPERPRIVOP_ITC_D		(HYPERPRIVOP_START + 0x4)
+#define HYPERPRIVOP_ITC_I		(HYPERPRIVOP_START + 0x5)
+#define HYPERPRIVOP_SSM_I		(HYPERPRIVOP_START + 0x6)
+#define HYPERPRIVOP_GET_IVR		(HYPERPRIVOP_START + 0x7)
+#define HYPERPRIVOP_GET_TPR		(HYPERPRIVOP_START + 0x8)
+#define HYPERPRIVOP_SET_TPR		(HYPERPRIVOP_START + 0x9)
+#define HYPERPRIVOP_EOI			(HYPERPRIVOP_START + 0xa)
+#define HYPERPRIVOP_SET_ITM		(HYPERPRIVOP_START + 0xb)
+#define HYPERPRIVOP_THASH		(HYPERPRIVOP_START + 0xc)
+#define HYPERPRIVOP_PTC_GA		(HYPERPRIVOP_START + 0xd)
+#define HYPERPRIVOP_ITR_D		(HYPERPRIVOP_START + 0xe)
+#define HYPERPRIVOP_GET_RR		(HYPERPRIVOP_START + 0xf)
+#define HYPERPRIVOP_SET_RR		(HYPERPRIVOP_START + 0x10)
+#define HYPERPRIVOP_SET_KR		(HYPERPRIVOP_START + 0x11)
+#define HYPERPRIVOP_FC			(HYPERPRIVOP_START + 0x12)
+#define HYPERPRIVOP_GET_CPUID		(HYPERPRIVOP_START + 0x13)
+#define HYPERPRIVOP_GET_PMD		(HYPERPRIVOP_START + 0x14)
+#define HYPERPRIVOP_GET_EFLAG		(HYPERPRIVOP_START + 0x15)
+#define HYPERPRIVOP_SET_EFLAG		(HYPERPRIVOP_START + 0x16)
+#define HYPERPRIVOP_RSM_BE		(HYPERPRIVOP_START + 0x17)
+#define HYPERPRIVOP_GET_PSR		(HYPERPRIVOP_START + 0x18)
+#define HYPERPRIVOP_SET_RR0_TO_RR4	(HYPERPRIVOP_START + 0x19)
+#define HYPERPRIVOP_MAX			(0x1a)
+
+/* Fast and light hypercalls.  */
+#define __HYPERVISOR_ia64_fast_eoi	__HYPERVISOR_arch_1
+
+/* Extra debug features.  */
+#define __HYPERVISOR_ia64_debug_op  __HYPERVISOR_arch_2
+
+/* Xencomm macros.  */
+#define XENCOMM_INLINE_MASK 0xf800000000000000UL
+#define XENCOMM_INLINE_FLAG 0x8000000000000000UL
+
+#ifndef __ASSEMBLY__
+
+/*
+ * Optimization features.
+ * The hypervisor may do some special optimizations for guests. This hypercall
+ * can be used to switch on/of these special optimizations.
+ */
+#define __HYPERVISOR_opt_feature	0x700UL
+
+#define XEN_IA64_OPTF_OFF	0x0
+#define XEN_IA64_OPTF_ON	0x1
+
+/*
+ * If this feature is switched on, the hypervisor inserts the
+ * tlb entries without calling the guests traphandler.
+ * This is useful in guests using region 7 for identity mapping
+ * like the linux kernel does.
+ */
+#define XEN_IA64_OPTF_IDENT_MAP_REG7    1
+
+/* Identity mapping of region 4 addresses in HVM. */
+#define XEN_IA64_OPTF_IDENT_MAP_REG4    2
+
+/* Identity mapping of region 5 addresses in HVM. */
+#define XEN_IA64_OPTF_IDENT_MAP_REG5    3
+
+#define XEN_IA64_OPTF_IDENT_MAP_NOT_SET  (0)
+
+struct xen_ia64_opt_feature {
+	unsigned long cmd;		/* Which feature */
+	unsigned char on;		/* Switch feature on/off */
+	union {
+		struct {
+				/* The page protection bit mask of the pte.
+			 	 * This will be or'ed with the pte. */
+			unsigned long pgprot;
+			unsigned long key;	/* A protection key for itir. */
+		};
+	};
+};
+
+#endif /* __ASSEMBLY__ */
+
+/* xen perfmon */
+#ifdef XEN
+#ifndef __ASSEMBLY__
+#ifndef _ASM_IA64_PERFMON_H
+
+#include <xen/list.h>   // asm/perfmon.h requires struct list_head
+#include <asm/perfmon.h>
+// for PFM_xxx and pfarg_features_t, pfarg_context_t, pfarg_reg_t, pfarg_load_t
+
+#endif /* _ASM_IA64_PERFMON_H */
+
+DEFINE_XEN_GUEST_HANDLE(pfarg_features_t);
+DEFINE_XEN_GUEST_HANDLE(pfarg_context_t);
+DEFINE_XEN_GUEST_HANDLE(pfarg_reg_t);
+DEFINE_XEN_GUEST_HANDLE(pfarg_load_t);
+#endif /* __ASSEMBLY__ */
+#endif /* XEN */
+
+#ifndef __ASSEMBLY__
+#include "arch-ia64/hvm/memmap.h"
+#endif
+
+#endif /* __HYPERVISOR_IF_IA64_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/arch-powerpc.h linux-2.6.18-xen-3.3.0/include/xen/interface/arch-powerpc.h
--- linux-2.6.18.8/include/xen/interface/arch-powerpc.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/arch-powerpc.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,120 @@
+/*
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (C) IBM Corp. 2005, 2006
+ *
+ * Authors: Hollis Blanchard <hollisb@us.ibm.com>
+ */
+
+#include "xen.h"
+
+#ifndef __XEN_PUBLIC_ARCH_PPC_64_H__
+#define __XEN_PUBLIC_ARCH_PPC_64_H__
+
+#define ___DEFINE_XEN_GUEST_HANDLE(name, type) \
+    typedef struct { \
+        int __pad[(sizeof (long long) - sizeof (void *)) / sizeof (int)]; \
+        type *p; \
+    } __attribute__((__aligned__(8))) __guest_handle_ ## name
+
+#define __DEFINE_XEN_GUEST_HANDLE(name, type) \
+    ___DEFINE_XEN_GUEST_HANDLE(name, type);   \
+    ___DEFINE_XEN_GUEST_HANDLE(const_##name, const type)
+#define DEFINE_XEN_GUEST_HANDLE(name) __DEFINE_XEN_GUEST_HANDLE(name, name)
+#define XEN_GUEST_HANDLE(name)        __guest_handle_ ## name
+#define set_xen_guest_handle(hnd, val) \
+    do { \
+        if (sizeof ((hnd).__pad)) \
+            (hnd).__pad[0] = 0; \
+        (hnd).p = val; \
+    } while (0)
+
+#ifdef __XEN_TOOLS__
+#define get_xen_guest_handle(val, hnd)  do { val = (hnd).p; } while (0)
+#endif
+
+#ifndef __ASSEMBLY__
+typedef unsigned long long xen_pfn_t;
+#define PRI_xen_pfn "llx"
+#endif
+
+/*
+ * Pointers and other address fields inside interface structures are padded to
+ * 64 bits. This means that field alignments aren't different between 32- and
+ * 64-bit architectures. 
+ */
+/* NB. Multi-level macro ensures __LINE__ is expanded before concatenation. */
+#define __MEMORY_PADDING(_X)
+#define _MEMORY_PADDING(_X)  __MEMORY_PADDING(_X)
+#define MEMORY_PADDING       _MEMORY_PADDING(__LINE__)
+
+/* And the trap vector is... */
+#define TRAP_INSTR "li 0,-1; sc" /* XXX just "sc"? */
+
+#ifndef __ASSEMBLY__
+
+#define XENCOMM_INLINE_FLAG (1UL << 63)
+
+typedef uint64_t xen_ulong_t;
+
+/* User-accessible registers: nost of these need to be saved/restored
+ * for every nested Xen invocation. */
+struct cpu_user_regs
+{
+    uint64_t gprs[32];
+    uint64_t lr;
+    uint64_t ctr;
+    uint64_t srr0;
+    uint64_t srr1;
+    uint64_t pc;
+    uint64_t msr;
+    uint64_t fpscr;             /* XXX Is this necessary */
+    uint64_t xer;
+    uint64_t hid4;              /* debug only */
+    uint64_t dar;               /* debug only */
+    uint32_t dsisr;             /* debug only */
+    uint32_t cr;
+    uint32_t __pad;             /* good spot for another 32bit reg */
+    uint32_t entry_vector;
+};
+typedef struct cpu_user_regs cpu_user_regs_t;
+
+typedef uint64_t tsc_timestamp_t; /* RDTSC timestamp */ /* XXX timebase */
+
+/* ONLY used to communicate with dom0! See also struct exec_domain. */
+struct vcpu_guest_context {
+    cpu_user_regs_t user_regs;         /* User-level CPU registers     */
+    uint64_t sdr1;                     /* Pagetable base               */
+    /* XXX etc */
+};
+typedef struct vcpu_guest_context vcpu_guest_context_t;
+DEFINE_XEN_GUEST_HANDLE(vcpu_guest_context_t);
+
+struct arch_shared_info {
+    uint64_t boot_timebase;
+};
+
+struct arch_vcpu_info {
+};
+
+/* Support for multi-processor guests. */
+#define MAX_VIRT_CPUS 32
+#endif
+
+#endif
diff -rpuN linux-2.6.18.8/include/xen/interface/arch-x86/cpuid.h linux-2.6.18-xen-3.3.0/include/xen/interface/arch-x86/cpuid.h
--- linux-2.6.18.8/include/xen/interface/arch-x86/cpuid.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/arch-x86/cpuid.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,68 @@
+/******************************************************************************
+ * arch-x86/cpuid.h
+ * 
+ * CPUID interface to Xen.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ * 
+ * Copyright (c) 2007 Citrix Systems, Inc.
+ * 
+ * Authors:
+ *    Keir Fraser <keir.fraser@citrix.com>
+ */
+
+#ifndef __XEN_PUBLIC_ARCH_X86_CPUID_H__
+#define __XEN_PUBLIC_ARCH_X86_CPUID_H__
+
+/* Xen identification leaves start at 0x40000000. */
+#define XEN_CPUID_FIRST_LEAF 0x40000000
+#define XEN_CPUID_LEAF(i)    (XEN_CPUID_FIRST_LEAF + (i))
+
+/*
+ * Leaf 1 (0x40000000)
+ * EAX: Largest Xen-information leaf. All leaves up to an including @EAX
+ *      are supported by the Xen host.
+ * EBX-EDX: "XenVMMXenVMM" signature, allowing positive identification
+ *      of a Xen host.
+ */
+#define XEN_CPUID_SIGNATURE_EBX 0x566e6558 /* "XenV" */
+#define XEN_CPUID_SIGNATURE_ECX 0x65584d4d /* "MMXe" */
+#define XEN_CPUID_SIGNATURE_EDX 0x4d4d566e /* "nVMM" */
+
+/*
+ * Leaf 2 (0x40000001)
+ * EAX[31:16]: Xen major version.
+ * EAX[15: 0]: Xen minor version.
+ * EBX-EDX: Reserved (currently all zeroes).
+ */
+
+/*
+ * Leaf 3 (0x40000002)
+ * EAX: Number of hypercall transfer pages. This register is always guaranteed
+ *      to specify one hypercall page.
+ * EBX: Base address of Xen-specific MSRs.
+ * ECX: Features 1. Unused bits are set to zero.
+ * EDX: Features 2. Unused bits are set to zero.
+ */
+
+/* Does the host support MMU_PT_UPDATE_PRESERVE_AD for this guest? */
+#define _XEN_CPUID_FEAT1_MMU_PT_UPDATE_PRESERVE_AD 0
+#define XEN_CPUID_FEAT1_MMU_PT_UPDATE_PRESERVE_AD  (1u<<0)
+
+#endif /* __XEN_PUBLIC_ARCH_X86_CPUID_H__ */
diff -rpuN linux-2.6.18.8/include/xen/interface/arch-x86/hvm/save.h linux-2.6.18-xen-3.3.0/include/xen/interface/arch-x86/hvm/save.h
--- linux-2.6.18.8/include/xen/interface/arch-x86/hvm/save.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/arch-x86/hvm/save.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,429 @@
+/* 
+ * Structure definitions for HVM state that is held by Xen and must
+ * be saved along with the domain's memory and device-model state.
+ * 
+ * Copyright (c) 2007 XenSource Ltd.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_PUBLIC_HVM_SAVE_X86_H__
+#define __XEN_PUBLIC_HVM_SAVE_X86_H__
+
+/* 
+ * Save/restore header: general info about the save file. 
+ */
+
+#define HVM_FILE_MAGIC   0x54381286
+#define HVM_FILE_VERSION 0x00000001
+
+struct hvm_save_header {
+    uint32_t magic;             /* Must be HVM_FILE_MAGIC */
+    uint32_t version;           /* File format version */
+    uint64_t changeset;         /* Version of Xen that saved this file */
+    uint32_t cpuid;             /* CPUID[0x01][%eax] on the saving machine */
+    uint32_t pad0;
+};
+
+DECLARE_HVM_SAVE_TYPE(HEADER, 1, struct hvm_save_header);
+
+
+/*
+ * Processor
+ */
+
+struct hvm_hw_cpu {
+    uint8_t  fpu_regs[512];
+
+    uint64_t rax;
+    uint64_t rbx;
+    uint64_t rcx;
+    uint64_t rdx;
+    uint64_t rbp;
+    uint64_t rsi;
+    uint64_t rdi;
+    uint64_t rsp;
+    uint64_t r8;
+    uint64_t r9;
+    uint64_t r10;
+    uint64_t r11;
+    uint64_t r12;
+    uint64_t r13;
+    uint64_t r14;
+    uint64_t r15;
+
+    uint64_t rip;
+    uint64_t rflags;
+
+    uint64_t cr0;
+    uint64_t cr2;
+    uint64_t cr3;
+    uint64_t cr4;
+
+    uint64_t dr0;
+    uint64_t dr1;
+    uint64_t dr2;
+    uint64_t dr3;
+    uint64_t dr6;
+    uint64_t dr7;    
+
+    uint32_t cs_sel;
+    uint32_t ds_sel;
+    uint32_t es_sel;
+    uint32_t fs_sel;
+    uint32_t gs_sel;
+    uint32_t ss_sel;
+    uint32_t tr_sel;
+    uint32_t ldtr_sel;
+
+    uint32_t cs_limit;
+    uint32_t ds_limit;
+    uint32_t es_limit;
+    uint32_t fs_limit;
+    uint32_t gs_limit;
+    uint32_t ss_limit;
+    uint32_t tr_limit;
+    uint32_t ldtr_limit;
+    uint32_t idtr_limit;
+    uint32_t gdtr_limit;
+
+    uint64_t cs_base;
+    uint64_t ds_base;
+    uint64_t es_base;
+    uint64_t fs_base;
+    uint64_t gs_base;
+    uint64_t ss_base;
+    uint64_t tr_base;
+    uint64_t ldtr_base;
+    uint64_t idtr_base;
+    uint64_t gdtr_base;
+
+    uint32_t cs_arbytes;
+    uint32_t ds_arbytes;
+    uint32_t es_arbytes;
+    uint32_t fs_arbytes;
+    uint32_t gs_arbytes;
+    uint32_t ss_arbytes;
+    uint32_t tr_arbytes;
+    uint32_t ldtr_arbytes;
+
+    uint32_t sysenter_cs;
+    uint32_t padding0;
+
+    uint64_t sysenter_esp;
+    uint64_t sysenter_eip;
+
+    /* msr for em64t */
+    uint64_t shadow_gs;
+
+    /* msr content saved/restored. */
+    uint64_t msr_flags;
+    uint64_t msr_lstar;
+    uint64_t msr_star;
+    uint64_t msr_cstar;
+    uint64_t msr_syscall_mask;
+    uint64_t msr_efer;
+
+    /* guest's idea of what rdtsc() would return */
+    uint64_t tsc;
+
+    /* pending event, if any */
+    union {
+        uint32_t pending_event;
+        struct {
+            uint8_t  pending_vector:8;
+            uint8_t  pending_type:3;
+            uint8_t  pending_error_valid:1;
+            uint32_t pending_reserved:19;
+            uint8_t  pending_valid:1;
+        };
+    };
+    /* error code for pending event */
+    uint32_t error_code;
+};
+
+DECLARE_HVM_SAVE_TYPE(CPU, 2, struct hvm_hw_cpu);
+
+
+/*
+ * PIC
+ */
+
+struct hvm_hw_vpic {
+    /* IR line bitmasks. */
+    uint8_t irr;
+    uint8_t imr;
+    uint8_t isr;
+
+    /* Line IRx maps to IRQ irq_base+x */
+    uint8_t irq_base;
+
+    /*
+     * Where are we in ICW2-4 initialisation (0 means no init in progress)?
+     * Bits 0-1 (=x): Next write at A=1 sets ICW(x+1).
+     * Bit 2: ICW1.IC4  (1 == ICW4 included in init sequence)
+     * Bit 3: ICW1.SNGL (0 == ICW3 included in init sequence)
+     */
+    uint8_t init_state:4;
+
+    /* IR line with highest priority. */
+    uint8_t priority_add:4;
+
+    /* Reads from A=0 obtain ISR or IRR? */
+    uint8_t readsel_isr:1;
+
+    /* Reads perform a polling read? */
+    uint8_t poll:1;
+
+    /* Automatically clear IRQs from the ISR during INTA? */
+    uint8_t auto_eoi:1;
+
+    /* Automatically rotate IRQ priorities during AEOI? */
+    uint8_t rotate_on_auto_eoi:1;
+
+    /* Exclude slave inputs when considering in-service IRQs? */
+    uint8_t special_fully_nested_mode:1;
+
+    /* Special mask mode excludes masked IRs from AEOI and priority checks. */
+    uint8_t special_mask_mode:1;
+
+    /* Is this a master PIC or slave PIC? (NB. This is not programmable.) */
+    uint8_t is_master:1;
+
+    /* Edge/trigger selection. */
+    uint8_t elcr;
+
+    /* Virtual INT output. */
+    uint8_t int_output;
+};
+
+DECLARE_HVM_SAVE_TYPE(PIC, 3, struct hvm_hw_vpic);
+
+
+/*
+ * IO-APIC
+ */
+
+#ifdef __ia64__
+#define VIOAPIC_IS_IOSAPIC 1
+#define VIOAPIC_NUM_PINS  24
+#else
+#define VIOAPIC_NUM_PINS  48 /* 16 ISA IRQs, 32 non-legacy PCI IRQS. */
+#endif
+
+struct hvm_hw_vioapic {
+    uint64_t base_address;
+    uint32_t ioregsel;
+    uint32_t id;
+    union vioapic_redir_entry
+    {
+        uint64_t bits;
+        struct {
+            uint8_t vector;
+            uint8_t delivery_mode:3;
+            uint8_t dest_mode:1;
+            uint8_t delivery_status:1;
+            uint8_t polarity:1;
+            uint8_t remote_irr:1;
+            uint8_t trig_mode:1;
+            uint8_t mask:1;
+            uint8_t reserve:7;
+#if !VIOAPIC_IS_IOSAPIC
+            uint8_t reserved[4];
+            uint8_t dest_id;
+#else
+            uint8_t reserved[3];
+            uint16_t dest_id;
+#endif
+        } fields;
+    } redirtbl[VIOAPIC_NUM_PINS];
+};
+
+DECLARE_HVM_SAVE_TYPE(IOAPIC, 4, struct hvm_hw_vioapic);
+
+
+/*
+ * LAPIC
+ */
+
+struct hvm_hw_lapic {
+    uint64_t             apic_base_msr;
+    uint32_t             disabled; /* VLAPIC_xx_DISABLED */
+    uint32_t             timer_divisor;
+};
+
+DECLARE_HVM_SAVE_TYPE(LAPIC, 5, struct hvm_hw_lapic);
+
+struct hvm_hw_lapic_regs {
+    uint8_t data[1024];
+};
+
+DECLARE_HVM_SAVE_TYPE(LAPIC_REGS, 6, struct hvm_hw_lapic_regs);
+
+
+/*
+ * IRQs
+ */
+
+struct hvm_hw_pci_irqs {
+    /*
+     * Virtual interrupt wires for a single PCI bus.
+     * Indexed by: device*4 + INTx#.
+     */
+    union {
+        DECLARE_BITMAP(i, 32*4);
+        uint64_t pad[2];
+    };
+};
+
+DECLARE_HVM_SAVE_TYPE(PCI_IRQ, 7, struct hvm_hw_pci_irqs);
+
+struct hvm_hw_isa_irqs {
+    /*
+     * Virtual interrupt wires for ISA devices.
+     * Indexed by ISA IRQ (assumes no ISA-device IRQ sharing).
+     */
+    union {
+        DECLARE_BITMAP(i, 16);
+        uint64_t pad[1];
+    };
+};
+
+DECLARE_HVM_SAVE_TYPE(ISA_IRQ, 8, struct hvm_hw_isa_irqs);
+
+struct hvm_hw_pci_link {
+    /*
+     * PCI-ISA interrupt router.
+     * Each PCI <device:INTx#> is 'wire-ORed' into one of four links using
+     * the traditional 'barber's pole' mapping ((device + INTx#) & 3).
+     * The router provides a programmable mapping from each link to a GSI.
+     */
+    uint8_t route[4];
+    uint8_t pad0[4];
+};
+
+DECLARE_HVM_SAVE_TYPE(PCI_LINK, 9, struct hvm_hw_pci_link);
+
+/* 
+ *  PIT
+ */
+
+struct hvm_hw_pit {
+    struct hvm_hw_pit_channel {
+        uint32_t count; /* can be 65536 */
+        uint16_t latched_count;
+        uint8_t count_latched;
+        uint8_t status_latched;
+        uint8_t status;
+        uint8_t read_state;
+        uint8_t write_state;
+        uint8_t write_latch;
+        uint8_t rw_mode;
+        uint8_t mode;
+        uint8_t bcd; /* not supported */
+        uint8_t gate; /* timer start */
+    } channels[3];  /* 3 x 16 bytes */
+    uint32_t speaker_data_on;
+    uint32_t pad0;
+};
+
+DECLARE_HVM_SAVE_TYPE(PIT, 10, struct hvm_hw_pit);
+
+
+/* 
+ * RTC
+ */ 
+
+#define RTC_CMOS_SIZE 14
+struct hvm_hw_rtc {
+    /* CMOS bytes */
+    uint8_t cmos_data[RTC_CMOS_SIZE];
+    /* Index register for 2-part operations */
+    uint8_t cmos_index;
+    uint8_t pad0;
+};
+
+DECLARE_HVM_SAVE_TYPE(RTC, 11, struct hvm_hw_rtc);
+
+
+/*
+ * HPET
+ */
+
+#define HPET_TIMER_NUM     3    /* 3 timers supported now */
+struct hvm_hw_hpet {
+    /* Memory-mapped, software visible registers */
+    uint64_t capability;        /* capabilities */
+    uint64_t res0;              /* reserved */
+    uint64_t config;            /* configuration */
+    uint64_t res1;              /* reserved */
+    uint64_t isr;               /* interrupt status reg */
+    uint64_t res2[25];          /* reserved */
+    uint64_t mc64;              /* main counter */
+    uint64_t res3;              /* reserved */
+    struct {                    /* timers */
+        uint64_t config;        /* configuration/cap */
+        uint64_t cmp;           /* comparator */
+        uint64_t fsb;           /* FSB route, not supported now */
+        uint64_t res4;          /* reserved */
+    } timers[HPET_TIMER_NUM];
+    uint64_t res5[4*(24-HPET_TIMER_NUM)];  /* reserved, up to 0x3ff */
+
+    /* Hidden register state */
+    uint64_t period[HPET_TIMER_NUM]; /* Last value written to comparator */
+};
+
+DECLARE_HVM_SAVE_TYPE(HPET, 12, struct hvm_hw_hpet);
+
+
+/*
+ * PM timer
+ */
+
+struct hvm_hw_pmtimer {
+    uint32_t tmr_val;   /* PM_TMR_BLK.TMR_VAL: 32bit free-running counter */
+    uint16_t pm1a_sts;  /* PM1a_EVT_BLK.PM1a_STS: status register */
+    uint16_t pm1a_en;   /* PM1a_EVT_BLK.PM1a_EN: enable register */
+};
+
+DECLARE_HVM_SAVE_TYPE(PMTIMER, 13, struct hvm_hw_pmtimer);
+
+/*
+ * MTRR MSRs
+ */
+
+struct hvm_hw_mtrr {
+#define MTRR_VCNT 8
+#define NUM_FIXED_MSR 11
+    uint64_t msr_pat_cr;
+    /* mtrr physbase & physmask msr pair*/
+    uint64_t msr_mtrr_var[MTRR_VCNT*2];
+    uint64_t msr_mtrr_fixed[NUM_FIXED_MSR];
+    uint64_t msr_mtrr_cap;
+    uint64_t msr_mtrr_def_type;
+};
+
+DECLARE_HVM_SAVE_TYPE(MTRR, 14, struct hvm_hw_mtrr);
+
+/* 
+ * Largest type-code in use
+ */
+#define HVM_SAVE_CODE_MAX 14
+
+#endif /* __XEN_PUBLIC_HVM_SAVE_X86_H__ */
diff -rpuN linux-2.6.18.8/include/xen/interface/arch-x86/xen.h linux-2.6.18-xen-3.3.0/include/xen/interface/arch-x86/xen.h
--- linux-2.6.18.8/include/xen/interface/arch-x86/xen.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/arch-x86/xen.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,204 @@
+/******************************************************************************
+ * arch-x86/xen.h
+ * 
+ * Guest OS interface to x86 Xen.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2004-2006, K A Fraser
+ */
+
+#include "../xen.h"
+
+#ifndef __XEN_PUBLIC_ARCH_X86_XEN_H__
+#define __XEN_PUBLIC_ARCH_X86_XEN_H__
+
+/* Structural guest handles introduced in 0x00030201. */
+#if __XEN_INTERFACE_VERSION__ >= 0x00030201
+#define ___DEFINE_XEN_GUEST_HANDLE(name, type) \
+    typedef struct { type *p; } __guest_handle_ ## name
+#else
+#define ___DEFINE_XEN_GUEST_HANDLE(name, type) \
+    typedef type * __guest_handle_ ## name
+#endif
+
+#define __DEFINE_XEN_GUEST_HANDLE(name, type) \
+    ___DEFINE_XEN_GUEST_HANDLE(name, type);   \
+    ___DEFINE_XEN_GUEST_HANDLE(const_##name, const type)
+#define DEFINE_XEN_GUEST_HANDLE(name)   __DEFINE_XEN_GUEST_HANDLE(name, name)
+#define __XEN_GUEST_HANDLE(name)        __guest_handle_ ## name
+#define XEN_GUEST_HANDLE(name)          __XEN_GUEST_HANDLE(name)
+#define set_xen_guest_handle(hnd, val)  do { (hnd).p = val; } while (0)
+#ifdef __XEN_TOOLS__
+#define get_xen_guest_handle(val, hnd)  do { val = (hnd).p; } while (0)
+#endif
+
+#if defined(__i386__)
+#include "xen-x86_32.h"
+#elif defined(__x86_64__)
+#include "xen-x86_64.h"
+#endif
+
+#ifndef __ASSEMBLY__
+typedef unsigned long xen_pfn_t;
+#define PRI_xen_pfn "lx"
+#endif
+
+/*
+ * SEGMENT DESCRIPTOR TABLES
+ */
+/*
+ * A number of GDT entries are reserved by Xen. These are not situated at the
+ * start of the GDT because some stupid OSes export hard-coded selector values
+ * in their ABI. These hard-coded values are always near the start of the GDT,
+ * so Xen places itself out of the way, at the far end of the GDT.
+ */
+#define FIRST_RESERVED_GDT_PAGE  14
+#define FIRST_RESERVED_GDT_BYTE  (FIRST_RESERVED_GDT_PAGE * 4096)
+#define FIRST_RESERVED_GDT_ENTRY (FIRST_RESERVED_GDT_BYTE / 8)
+
+/* Maximum number of virtual CPUs in multi-processor guests. */
+#define MAX_VIRT_CPUS 32
+
+
+/* Machine check support */
+#include "xen-mca.h"
+
+#ifndef __ASSEMBLY__
+
+typedef unsigned long xen_ulong_t;
+
+/*
+ * Send an array of these to HYPERVISOR_set_trap_table().
+ * The privilege level specifies which modes may enter a trap via a software
+ * interrupt. On x86/64, since rings 1 and 2 are unavailable, we allocate
+ * privilege levels as follows:
+ *  Level == 0: Noone may enter
+ *  Level == 1: Kernel may enter
+ *  Level == 2: Kernel may enter
+ *  Level == 3: Everyone may enter
+ */
+#define TI_GET_DPL(_ti)      ((_ti)->flags & 3)
+#define TI_GET_IF(_ti)       ((_ti)->flags & 4)
+#define TI_SET_DPL(_ti,_dpl) ((_ti)->flags |= (_dpl))
+#define TI_SET_IF(_ti,_if)   ((_ti)->flags |= ((!!(_if))<<2))
+struct trap_info {
+    uint8_t       vector;  /* exception vector                              */
+    uint8_t       flags;   /* 0-3: privilege level; 4: clear event enable?  */
+    uint16_t      cs;      /* code selector                                 */
+    unsigned long address; /* code offset                                   */
+};
+typedef struct trap_info trap_info_t;
+DEFINE_XEN_GUEST_HANDLE(trap_info_t);
+
+typedef uint64_t tsc_timestamp_t; /* RDTSC timestamp */
+
+/*
+ * The following is all CPU context. Note that the fpu_ctxt block is filled 
+ * in by FXSAVE if the CPU has feature FXSR; otherwise FSAVE is used.
+ */
+struct vcpu_guest_context {
+    /* FPU registers come first so they can be aligned for FXSAVE/FXRSTOR. */
+    struct { char x[512]; } fpu_ctxt;       /* User-level FPU registers     */
+#define VGCF_I387_VALID                (1<<0)
+#define VGCF_IN_KERNEL                 (1<<2)
+#define _VGCF_i387_valid               0
+#define VGCF_i387_valid                (1<<_VGCF_i387_valid)
+#define _VGCF_in_kernel                2
+#define VGCF_in_kernel                 (1<<_VGCF_in_kernel)
+#define _VGCF_failsafe_disables_events 3
+#define VGCF_failsafe_disables_events  (1<<_VGCF_failsafe_disables_events)
+#define _VGCF_syscall_disables_events  4
+#define VGCF_syscall_disables_events   (1<<_VGCF_syscall_disables_events)
+#define _VGCF_online                   5
+#define VGCF_online                    (1<<_VGCF_online)
+    unsigned long flags;                    /* VGCF_* flags                 */
+    struct cpu_user_regs user_regs;         /* User-level CPU registers     */
+    struct trap_info trap_ctxt[256];        /* Virtual IDT                  */
+    unsigned long ldt_base, ldt_ents;       /* LDT (linear address, # ents) */
+    unsigned long gdt_frames[16], gdt_ents; /* GDT (machine frames, # ents) */
+    unsigned long kernel_ss, kernel_sp;     /* Virtual TSS (only SS1/SP1)   */
+    /* NB. User pagetable on x86/64 is placed in ctrlreg[1]. */
+    unsigned long ctrlreg[8];               /* CR0-CR7 (control registers)  */
+    unsigned long debugreg[8];              /* DB0-DB7 (debug registers)    */
+#ifdef __i386__
+    unsigned long event_callback_cs;        /* CS:EIP of event callback     */
+    unsigned long event_callback_eip;
+    unsigned long failsafe_callback_cs;     /* CS:EIP of failsafe callback  */
+    unsigned long failsafe_callback_eip;
+#else
+    unsigned long event_callback_eip;
+    unsigned long failsafe_callback_eip;
+#ifdef __XEN__
+    union {
+        unsigned long syscall_callback_eip;
+        struct {
+            unsigned int event_callback_cs;    /* compat CS of event cb     */
+            unsigned int failsafe_callback_cs; /* compat CS of failsafe cb  */
+        };
+    };
+#else
+    unsigned long syscall_callback_eip;
+#endif
+#endif
+    unsigned long vm_assist;                /* VMASST_TYPE_* bitmap */
+#ifdef __x86_64__
+    /* Segment base addresses. */
+    uint64_t      fs_base;
+    uint64_t      gs_base_kernel;
+    uint64_t      gs_base_user;
+#endif
+};
+typedef struct vcpu_guest_context vcpu_guest_context_t;
+DEFINE_XEN_GUEST_HANDLE(vcpu_guest_context_t);
+
+struct arch_shared_info {
+    unsigned long max_pfn;                  /* max pfn that appears in table */
+    /* Frame containing list of mfns containing list of mfns containing p2m. */
+    xen_pfn_t     pfn_to_mfn_frame_list_list;
+    unsigned long nmi_reason;
+    uint64_t pad[32];
+};
+typedef struct arch_shared_info arch_shared_info_t;
+
+#endif /* !__ASSEMBLY__ */
+
+/*
+ * Prefix forces emulation of some non-trapping instructions.
+ * Currently only CPUID.
+ */
+#ifdef __ASSEMBLY__
+#define XEN_EMULATE_PREFIX .byte 0x0f,0x0b,0x78,0x65,0x6e ;
+#define XEN_CPUID          XEN_EMULATE_PREFIX cpuid
+#else
+#define XEN_EMULATE_PREFIX ".byte 0x0f,0x0b,0x78,0x65,0x6e ; "
+#define XEN_CPUID          XEN_EMULATE_PREFIX "cpuid"
+#endif
+
+#endif /* __XEN_PUBLIC_ARCH_X86_XEN_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/arch-x86/xen-mca.h linux-2.6.18-xen-3.3.0/include/xen/interface/arch-x86/xen-mca.h
--- linux-2.6.18.8/include/xen/interface/arch-x86/xen-mca.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/arch-x86/xen-mca.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,279 @@
+/******************************************************************************
+ * arch-x86/mca.h
+ * 
+ * Contributed by Advanced Micro Devices, Inc.
+ * Author: Christoph Egger <Christoph.Egger@amd.com>
+ *
+ * Guest OS machine check interface to x86 Xen.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+/* Full MCA functionality has the following Usecases from the guest side:
+ *
+ * Must have's:
+ * 1. Dom0 and DomU register machine check trap callback handlers
+ *    (already done via "set_trap_table" hypercall)
+ * 2. Dom0 registers machine check event callback handler
+ *    (doable via EVTCHNOP_bind_virq)
+ * 3. Dom0 and DomU fetches machine check data
+ * 4. Dom0 wants Xen to notify a DomU
+ * 5. Dom0 gets DomU ID from physical address
+ * 6. Dom0 wants Xen to kill DomU (already done for "xm destroy")
+ *
+ * Nice to have's:
+ * 7. Dom0 wants Xen to deactivate a physical CPU
+ *    This is better done as separate task, physical CPU hotplugging,
+ *    and hypercall(s) should be sysctl's
+ * 8. Page migration proposed from Xen NUMA work, where Dom0 can tell Xen to
+ *    move a DomU (or Dom0 itself) away from a malicious page
+ *    producing correctable errors.
+ * 9. offlining physical page:
+ *    Xen free's and never re-uses a certain physical page.
+ * 10. Testfacility: Allow Dom0 to write values into machine check MSR's
+ *     and tell Xen to trigger a machine check
+ */
+
+#ifndef __XEN_PUBLIC_ARCH_X86_MCA_H__
+#define __XEN_PUBLIC_ARCH_X86_MCA_H__
+
+/* Hypercall */
+#define __HYPERVISOR_mca __HYPERVISOR_arch_0
+
+#define XEN_MCA_INTERFACE_VERSION 0x03000001
+
+/* IN: Dom0 calls hypercall from MC event handler. */
+#define XEN_MC_CORRECTABLE  0x0
+/* IN: Dom0/DomU calls hypercall from MC trap handler. */
+#define XEN_MC_TRAP         0x1
+/* XEN_MC_CORRECTABLE and XEN_MC_TRAP are mutually exclusive. */
+
+/* OUT: All is ok */
+#define XEN_MC_OK           0x0
+/* OUT: Domain could not fetch data. */
+#define XEN_MC_FETCHFAILED  0x1
+/* OUT: There was no machine check data to fetch. */
+#define XEN_MC_NODATA       0x2
+/* OUT: Between notification time and this hypercall an other
+ *  (most likely) correctable error happened. The fetched data,
+ *  does not match the original machine check data. */
+#define XEN_MC_NOMATCH      0x4
+
+/* OUT: DomU did not register MC NMI handler. Try something else. */
+#define XEN_MC_CANNOTHANDLE 0x8
+/* OUT: Notifying DomU failed. Retry later or try something else. */
+#define XEN_MC_NOTDELIVERED 0x10
+/* Note, XEN_MC_CANNOTHANDLE and XEN_MC_NOTDELIVERED are mutually exclusive. */
+
+
+#ifndef __ASSEMBLY__
+
+#define VIRQ_MCA VIRQ_ARCH_0 /* G. (DOM0) Machine Check Architecture */
+
+/*
+ * Machine Check Architecure:
+ * structs are read-only and used to report all kinds of
+ * correctable and uncorrectable errors detected by the HW.
+ * Dom0 and DomU: register a handler to get notified.
+ * Dom0 only: Correctable errors are reported via VIRQ_MCA
+ * Dom0 and DomU: Uncorrectable errors are reported via nmi handlers
+ */
+#define MC_TYPE_GLOBAL          0
+#define MC_TYPE_BANK            1
+#define MC_TYPE_EXTENDED        2
+
+struct mcinfo_common {
+    uint16_t type;      /* structure type */
+    uint16_t size;      /* size of this struct in bytes */
+};
+
+
+#define MC_FLAG_CORRECTABLE     (1 << 0)
+#define MC_FLAG_UNCORRECTABLE   (1 << 1)
+
+/* contains global x86 mc information */
+struct mcinfo_global {
+    struct mcinfo_common common;
+
+    /* running domain at the time in error (most likely the impacted one) */
+    uint16_t mc_domid;
+    uint32_t mc_socketid; /* physical socket of the physical core */
+    uint16_t mc_coreid; /* physical impacted core */
+    uint16_t mc_core_threadid; /* core thread of physical core */
+    uint16_t mc_vcpuid; /* virtual cpu scheduled for mc_domid */
+    uint64_t mc_gstatus; /* global status */
+    uint32_t mc_flags;
+};
+
+/* contains bank local x86 mc information */
+struct mcinfo_bank {
+    struct mcinfo_common common;
+
+    uint16_t mc_bank; /* bank nr */
+    uint16_t mc_domid; /* Usecase 5: domain referenced by mc_addr on dom0
+                        * and if mc_addr is valid. Never valid on DomU. */
+    uint64_t mc_status; /* bank status */
+    uint64_t mc_addr;   /* bank address, only valid
+                         * if addr bit is set in mc_status */
+    uint64_t mc_misc;
+};
+
+
+struct mcinfo_msr {
+    uint64_t reg;   /* MSR */
+    uint64_t value; /* MSR value */
+};
+
+/* contains mc information from other
+ * or additional mc MSRs */ 
+struct mcinfo_extended {
+    struct mcinfo_common common;
+
+    /* You can fill up to five registers.
+     * If you need more, then use this structure
+     * multiple times. */
+
+    uint32_t mc_msrs; /* Number of msr with valid values. */
+    struct mcinfo_msr mc_msr[5];
+};
+
+#define MCINFO_HYPERCALLSIZE	1024
+#define MCINFO_MAXSIZE		768
+
+struct mc_info {
+    /* Number of mcinfo_* entries in mi_data */
+    uint32_t mi_nentries;
+
+    uint8_t mi_data[MCINFO_MAXSIZE - sizeof(uint32_t)];
+};
+typedef struct mc_info mc_info_t;
+
+
+
+/* 
+ * OS's should use these instead of writing their own lookup function
+ * each with its own bugs and drawbacks.
+ * We use macros instead of static inline functions to allow guests
+ * to include this header in assembly files (*.S).
+ */
+/* Prototype:
+ *    uint32_t x86_mcinfo_nentries(struct mc_info *mi);
+ */
+#define x86_mcinfo_nentries(_mi)    \
+    (_mi)->mi_nentries
+/* Prototype:
+ *    struct mcinfo_common *x86_mcinfo_first(struct mc_info *mi);
+ */
+#define x86_mcinfo_first(_mi)       \
+    (struct mcinfo_common *)((_mi)->mi_data)
+/* Prototype:
+ *    struct mcinfo_common *x86_mcinfo_next(struct mcinfo_common *mic);
+ */
+#define x86_mcinfo_next(_mic)       \
+    (struct mcinfo_common *)((uint8_t *)(_mic) + (_mic)->size)
+
+/* Prototype:
+ *    void x86_mcinfo_lookup(void *ret, struct mc_info *mi, uint16_t type);
+ */
+#define x86_mcinfo_lookup(_ret, _mi, _type)    \
+    do {                                                        \
+        uint32_t found, i;                                      \
+        struct mcinfo_common *_mic;                             \
+                                                                \
+        found = 0;                                              \
+	(_ret) = NULL;						\
+	if (_mi == NULL) break;					\
+        _mic = x86_mcinfo_first(_mi);                           \
+        for (i = 0; i < x86_mcinfo_nentries(_mi); i++) {        \
+            if (_mic->type == (_type)) {                        \
+                found = 1;                                      \
+                break;                                          \
+            }                                                   \
+            _mic = x86_mcinfo_next(_mic);                       \
+        }                                                       \
+        (_ret) = found ? _mic : NULL;                           \
+    } while (0)
+
+
+/* Usecase 1
+ * Register machine check trap callback handler
+ *    (already done via "set_trap_table" hypercall)
+ */
+
+/* Usecase 2
+ * Dom0 registers machine check event callback handler
+ * done by EVTCHNOP_bind_virq
+ */
+
+/* Usecase 3
+ * Fetch machine check data from hypervisor.
+ * Note, this hypercall is special, because both Dom0 and DomU must use this.
+ */
+#define XEN_MC_fetch            1
+struct xen_mc_fetch {
+    /* IN/OUT variables. */
+    uint32_t flags;
+
+/* IN: XEN_MC_CORRECTABLE, XEN_MC_TRAP */
+/* OUT: XEN_MC_OK, XEN_MC_FETCHFAILED, XEN_MC_NODATA, XEN_MC_NOMATCH */
+
+    /* OUT variables. */
+    uint32_t fetch_idx;  /* only useful for Dom0 for the notify hypercall */
+    struct mc_info mc_info;
+};
+typedef struct xen_mc_fetch xen_mc_fetch_t;
+DEFINE_XEN_GUEST_HANDLE(xen_mc_fetch_t);
+
+
+/* Usecase 4
+ * This tells the hypervisor to notify a DomU about the machine check error
+ */
+#define XEN_MC_notifydomain     2
+struct xen_mc_notifydomain {
+    /* IN variables. */
+    uint16_t mc_domid;    /* The unprivileged domain to notify. */
+    uint16_t mc_vcpuid;   /* The vcpu in mc_domid to notify.
+                           * Usually echo'd value from the fetch hypercall. */
+    uint32_t fetch_idx;   /* echo'd value from the fetch hypercall. */
+
+    /* IN/OUT variables. */
+    uint32_t flags;
+
+/* IN: XEN_MC_CORRECTABLE, XEN_MC_TRAP */
+/* OUT: XEN_MC_OK, XEN_MC_CANNOTHANDLE, XEN_MC_NOTDELIVERED, XEN_MC_NOMATCH */
+};
+typedef struct xen_mc_notifydomain xen_mc_notifydomain_t;
+DEFINE_XEN_GUEST_HANDLE(xen_mc_notifydomain_t);
+
+
+struct xen_mc {
+    uint32_t cmd;
+    uint32_t interface_version; /* XEN_MCA_INTERFACE_VERSION */
+    union {
+        struct xen_mc_fetch        mc_fetch;
+        struct xen_mc_notifydomain mc_notifydomain;
+        uint8_t pad[MCINFO_HYPERCALLSIZE];
+    } u;
+};
+typedef struct xen_mc xen_mc_t;
+DEFINE_XEN_GUEST_HANDLE(xen_mc_t);
+
+#endif /* __ASSEMBLY__ */
+
+#endif /* __XEN_PUBLIC_ARCH_X86_MCA_H__ */
diff -rpuN linux-2.6.18.8/include/xen/interface/arch-x86/xen-x86_32.h linux-2.6.18-xen-3.3.0/include/xen/interface/arch-x86/xen-x86_32.h
--- linux-2.6.18.8/include/xen/interface/arch-x86/xen-x86_32.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/arch-x86/xen-x86_32.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,180 @@
+/******************************************************************************
+ * xen-x86_32.h
+ * 
+ * Guest OS interface to x86 32-bit Xen.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2004-2007, K A Fraser
+ */
+
+#ifndef __XEN_PUBLIC_ARCH_X86_XEN_X86_32_H__
+#define __XEN_PUBLIC_ARCH_X86_XEN_X86_32_H__
+
+/*
+ * Hypercall interface:
+ *  Input:  %ebx, %ecx, %edx, %esi, %edi (arguments 1-5)
+ *  Output: %eax
+ * Access is via hypercall page (set up by guest loader or via a Xen MSR):
+ *  call hypercall_page + hypercall-number * 32
+ * Clobbered: Argument registers (e.g., 2-arg hypercall clobbers %ebx,%ecx)
+ */
+
+#if __XEN_INTERFACE_VERSION__ < 0x00030203
+/*
+ * Legacy hypercall interface:
+ * As above, except the entry sequence to the hypervisor is:
+ *  mov $hypercall-number*32,%eax ; int $0x82
+ */
+#define TRAP_INSTR "int $0x82"
+#endif
+
+/*
+ * These flat segments are in the Xen-private section of every GDT. Since these
+ * are also present in the initial GDT, many OSes will be able to avoid
+ * installing their own GDT.
+ */
+#define FLAT_RING1_CS 0xe019    /* GDT index 259 */
+#define FLAT_RING1_DS 0xe021    /* GDT index 260 */
+#define FLAT_RING1_SS 0xe021    /* GDT index 260 */
+#define FLAT_RING3_CS 0xe02b    /* GDT index 261 */
+#define FLAT_RING3_DS 0xe033    /* GDT index 262 */
+#define FLAT_RING3_SS 0xe033    /* GDT index 262 */
+
+#define FLAT_KERNEL_CS FLAT_RING1_CS
+#define FLAT_KERNEL_DS FLAT_RING1_DS
+#define FLAT_KERNEL_SS FLAT_RING1_SS
+#define FLAT_USER_CS    FLAT_RING3_CS
+#define FLAT_USER_DS    FLAT_RING3_DS
+#define FLAT_USER_SS    FLAT_RING3_SS
+
+#define __HYPERVISOR_VIRT_START_PAE    0xF5800000
+#define __MACH2PHYS_VIRT_START_PAE     0xF5800000
+#define __MACH2PHYS_VIRT_END_PAE       0xF6800000
+#define HYPERVISOR_VIRT_START_PAE      \
+    mk_unsigned_long(__HYPERVISOR_VIRT_START_PAE)
+#define MACH2PHYS_VIRT_START_PAE       \
+    mk_unsigned_long(__MACH2PHYS_VIRT_START_PAE)
+#define MACH2PHYS_VIRT_END_PAE         \
+    mk_unsigned_long(__MACH2PHYS_VIRT_END_PAE)
+
+/* Non-PAE bounds are obsolete. */
+#define __HYPERVISOR_VIRT_START_NONPAE 0xFC000000
+#define __MACH2PHYS_VIRT_START_NONPAE  0xFC000000
+#define __MACH2PHYS_VIRT_END_NONPAE    0xFC400000
+#define HYPERVISOR_VIRT_START_NONPAE   \
+    mk_unsigned_long(__HYPERVISOR_VIRT_START_NONPAE)
+#define MACH2PHYS_VIRT_START_NONPAE    \
+    mk_unsigned_long(__MACH2PHYS_VIRT_START_NONPAE)
+#define MACH2PHYS_VIRT_END_NONPAE      \
+    mk_unsigned_long(__MACH2PHYS_VIRT_END_NONPAE)
+
+#define __HYPERVISOR_VIRT_START __HYPERVISOR_VIRT_START_PAE
+#define __MACH2PHYS_VIRT_START  __MACH2PHYS_VIRT_START_PAE
+#define __MACH2PHYS_VIRT_END    __MACH2PHYS_VIRT_END_PAE
+
+#ifndef HYPERVISOR_VIRT_START
+#define HYPERVISOR_VIRT_START mk_unsigned_long(__HYPERVISOR_VIRT_START)
+#endif
+
+#define MACH2PHYS_VIRT_START  mk_unsigned_long(__MACH2PHYS_VIRT_START)
+#define MACH2PHYS_VIRT_END    mk_unsigned_long(__MACH2PHYS_VIRT_END)
+#define MACH2PHYS_NR_ENTRIES  ((MACH2PHYS_VIRT_END-MACH2PHYS_VIRT_START)>>2)
+#ifndef machine_to_phys_mapping
+#define machine_to_phys_mapping ((unsigned long *)MACH2PHYS_VIRT_START)
+#endif
+
+/* 32-/64-bit invariability for control interfaces (domctl/sysctl). */
+#if defined(__XEN__) || defined(__XEN_TOOLS__)
+#undef ___DEFINE_XEN_GUEST_HANDLE
+#define ___DEFINE_XEN_GUEST_HANDLE(name, type)                  \
+    typedef struct { type *p; }                                 \
+        __guest_handle_ ## name;                                \
+    typedef struct { union { type *p; uint64_aligned_t q; }; }  \
+        __guest_handle_64_ ## name
+#undef set_xen_guest_handle
+#define set_xen_guest_handle(hnd, val)                      \
+    do { if ( sizeof(hnd) == 8 ) *(uint64_t *)&(hnd) = 0;   \
+         (hnd).p = val;                                     \
+    } while ( 0 )
+#define uint64_aligned_t uint64_t __attribute__((aligned(8)))
+#define __XEN_GUEST_HANDLE_64(name) __guest_handle_64_ ## name
+#define XEN_GUEST_HANDLE_64(name) __XEN_GUEST_HANDLE_64(name)
+#endif
+
+#ifndef __ASSEMBLY__
+
+struct cpu_user_regs {
+    uint32_t ebx;
+    uint32_t ecx;
+    uint32_t edx;
+    uint32_t esi;
+    uint32_t edi;
+    uint32_t ebp;
+    uint32_t eax;
+    uint16_t error_code;    /* private */
+    uint16_t entry_vector;  /* private */
+    uint32_t eip;
+    uint16_t cs;
+    uint8_t  saved_upcall_mask;
+    uint8_t  _pad0;
+    uint32_t eflags;        /* eflags.IF == !saved_upcall_mask */
+    uint32_t esp;
+    uint16_t ss, _pad1;
+    uint16_t es, _pad2;
+    uint16_t ds, _pad3;
+    uint16_t fs, _pad4;
+    uint16_t gs, _pad5;
+};
+typedef struct cpu_user_regs cpu_user_regs_t;
+DEFINE_XEN_GUEST_HANDLE(cpu_user_regs_t);
+
+/*
+ * Page-directory addresses above 4GB do not fit into architectural %cr3.
+ * When accessing %cr3, or equivalent field in vcpu_guest_context, guests
+ * must use the following accessor macros to pack/unpack valid MFNs.
+ */
+#define xen_pfn_to_cr3(pfn) (((unsigned)(pfn) << 12) | ((unsigned)(pfn) >> 20))
+#define xen_cr3_to_pfn(cr3) (((unsigned)(cr3) >> 12) | ((unsigned)(cr3) << 20))
+
+struct arch_vcpu_info {
+    unsigned long cr2;
+    unsigned long pad[5]; /* sizeof(vcpu_info_t) == 64 */
+};
+typedef struct arch_vcpu_info arch_vcpu_info_t;
+
+struct xen_callback {
+    unsigned long cs;
+    unsigned long eip;
+};
+typedef struct xen_callback xen_callback_t;
+
+#endif /* !__ASSEMBLY__ */
+
+#endif /* __XEN_PUBLIC_ARCH_X86_XEN_X86_32_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/arch-x86/xen-x86_64.h linux-2.6.18-xen-3.3.0/include/xen/interface/arch-x86/xen-x86_64.h
--- linux-2.6.18.8/include/xen/interface/arch-x86/xen-x86_64.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/arch-x86/xen-x86_64.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,212 @@
+/******************************************************************************
+ * xen-x86_64.h
+ * 
+ * Guest OS interface to x86 64-bit Xen.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2004-2006, K A Fraser
+ */
+
+#ifndef __XEN_PUBLIC_ARCH_X86_XEN_X86_64_H__
+#define __XEN_PUBLIC_ARCH_X86_XEN_X86_64_H__
+
+/*
+ * Hypercall interface:
+ *  Input:  %rdi, %rsi, %rdx, %r10, %r8 (arguments 1-5)
+ *  Output: %rax
+ * Access is via hypercall page (set up by guest loader or via a Xen MSR):
+ *  call hypercall_page + hypercall-number * 32
+ * Clobbered: argument registers (e.g., 2-arg hypercall clobbers %rdi,%rsi)
+ */
+
+#if __XEN_INTERFACE_VERSION__ < 0x00030203
+/*
+ * Legacy hypercall interface:
+ * As above, except the entry sequence to the hypervisor is:
+ *  mov $hypercall-number*32,%eax ; syscall
+ * Clobbered: %rcx, %r11, argument registers (as above)
+ */
+#define TRAP_INSTR "syscall"
+#endif
+
+/*
+ * 64-bit segment selectors
+ * These flat segments are in the Xen-private section of every GDT. Since these
+ * are also present in the initial GDT, many OSes will be able to avoid
+ * installing their own GDT.
+ */
+
+#define FLAT_RING3_CS32 0xe023  /* GDT index 260 */
+#define FLAT_RING3_CS64 0xe033  /* GDT index 261 */
+#define FLAT_RING3_DS32 0xe02b  /* GDT index 262 */
+#define FLAT_RING3_DS64 0x0000  /* NULL selector */
+#define FLAT_RING3_SS32 0xe02b  /* GDT index 262 */
+#define FLAT_RING3_SS64 0xe02b  /* GDT index 262 */
+
+#define FLAT_KERNEL_DS64 FLAT_RING3_DS64
+#define FLAT_KERNEL_DS32 FLAT_RING3_DS32
+#define FLAT_KERNEL_DS   FLAT_KERNEL_DS64
+#define FLAT_KERNEL_CS64 FLAT_RING3_CS64
+#define FLAT_KERNEL_CS32 FLAT_RING3_CS32
+#define FLAT_KERNEL_CS   FLAT_KERNEL_CS64
+#define FLAT_KERNEL_SS64 FLAT_RING3_SS64
+#define FLAT_KERNEL_SS32 FLAT_RING3_SS32
+#define FLAT_KERNEL_SS   FLAT_KERNEL_SS64
+
+#define FLAT_USER_DS64 FLAT_RING3_DS64
+#define FLAT_USER_DS32 FLAT_RING3_DS32
+#define FLAT_USER_DS   FLAT_USER_DS64
+#define FLAT_USER_CS64 FLAT_RING3_CS64
+#define FLAT_USER_CS32 FLAT_RING3_CS32
+#define FLAT_USER_CS   FLAT_USER_CS64
+#define FLAT_USER_SS64 FLAT_RING3_SS64
+#define FLAT_USER_SS32 FLAT_RING3_SS32
+#define FLAT_USER_SS   FLAT_USER_SS64
+
+#define __HYPERVISOR_VIRT_START 0xFFFF800000000000
+#define __HYPERVISOR_VIRT_END   0xFFFF880000000000
+#define __MACH2PHYS_VIRT_START  0xFFFF800000000000
+#define __MACH2PHYS_VIRT_END    0xFFFF804000000000
+
+#ifndef HYPERVISOR_VIRT_START
+#define HYPERVISOR_VIRT_START mk_unsigned_long(__HYPERVISOR_VIRT_START)
+#define HYPERVISOR_VIRT_END   mk_unsigned_long(__HYPERVISOR_VIRT_END)
+#endif
+
+#define MACH2PHYS_VIRT_START  mk_unsigned_long(__MACH2PHYS_VIRT_START)
+#define MACH2PHYS_VIRT_END    mk_unsigned_long(__MACH2PHYS_VIRT_END)
+#define MACH2PHYS_NR_ENTRIES  ((MACH2PHYS_VIRT_END-MACH2PHYS_VIRT_START)>>3)
+#ifndef machine_to_phys_mapping
+#define machine_to_phys_mapping ((unsigned long *)HYPERVISOR_VIRT_START)
+#endif
+
+/*
+ * int HYPERVISOR_set_segment_base(unsigned int which, unsigned long base)
+ *  @which == SEGBASE_*  ;  @base == 64-bit base address
+ * Returns 0 on success.
+ */
+#define SEGBASE_FS          0
+#define SEGBASE_GS_USER     1
+#define SEGBASE_GS_KERNEL   2
+#define SEGBASE_GS_USER_SEL 3 /* Set user %gs specified in base[15:0] */
+
+/*
+ * int HYPERVISOR_iret(void)
+ * All arguments are on the kernel stack, in the following format.
+ * Never returns if successful. Current kernel context is lost.
+ * The saved CS is mapped as follows:
+ *   RING0 -> RING3 kernel mode.
+ *   RING1 -> RING3 kernel mode.
+ *   RING2 -> RING3 kernel mode.
+ *   RING3 -> RING3 user mode.
+ * However RING0 indicates that the guest kernel should return to iteself
+ * directly with
+ *      orb   $3,1*8(%rsp)
+ *      iretq
+ * If flags contains VGCF_in_syscall:
+ *   Restore RAX, RIP, RFLAGS, RSP.
+ *   Discard R11, RCX, CS, SS.
+ * Otherwise:
+ *   Restore RAX, R11, RCX, CS:RIP, RFLAGS, SS:RSP.
+ * All other registers are saved on hypercall entry and restored to user.
+ */
+/* Guest exited in SYSCALL context? Return to guest with SYSRET? */
+#define _VGCF_in_syscall 8
+#define VGCF_in_syscall  (1<<_VGCF_in_syscall)
+#define VGCF_IN_SYSCALL  VGCF_in_syscall
+
+#ifndef __ASSEMBLY__
+
+struct iret_context {
+    /* Top of stack (%rsp at point of hypercall). */
+    uint64_t rax, r11, rcx, flags, rip, cs, rflags, rsp, ss;
+    /* Bottom of iret stack frame. */
+};
+
+#if defined(__GNUC__) && !defined(__STRICT_ANSI__)
+/* Anonymous union includes both 32- and 64-bit names (e.g., eax/rax). */
+#define __DECL_REG(name) union { \
+    uint64_t r ## name, e ## name; \
+    uint32_t _e ## name; \
+}
+#else
+/* Non-gcc sources must always use the proper 64-bit name (e.g., rax). */
+#define __DECL_REG(name) uint64_t r ## name
+#endif
+
+struct cpu_user_regs {
+    uint64_t r15;
+    uint64_t r14;
+    uint64_t r13;
+    uint64_t r12;
+    __DECL_REG(bp);
+    __DECL_REG(bx);
+    uint64_t r11;
+    uint64_t r10;
+    uint64_t r9;
+    uint64_t r8;
+    __DECL_REG(ax);
+    __DECL_REG(cx);
+    __DECL_REG(dx);
+    __DECL_REG(si);
+    __DECL_REG(di);
+    uint32_t error_code;    /* private */
+    uint32_t entry_vector;  /* private */
+    __DECL_REG(ip);
+    uint16_t cs, _pad0[1];
+    uint8_t  saved_upcall_mask;
+    uint8_t  _pad1[3];
+    __DECL_REG(flags);      /* rflags.IF == !saved_upcall_mask */
+    __DECL_REG(sp);
+    uint16_t ss, _pad2[3];
+    uint16_t es, _pad3[3];
+    uint16_t ds, _pad4[3];
+    uint16_t fs, _pad5[3]; /* Non-zero => takes precedence over fs_base.     */
+    uint16_t gs, _pad6[3]; /* Non-zero => takes precedence over gs_base_usr. */
+};
+typedef struct cpu_user_regs cpu_user_regs_t;
+DEFINE_XEN_GUEST_HANDLE(cpu_user_regs_t);
+
+#undef __DECL_REG
+
+#define xen_pfn_to_cr3(pfn) ((unsigned long)(pfn) << 12)
+#define xen_cr3_to_pfn(cr3) ((unsigned long)(cr3) >> 12)
+
+struct arch_vcpu_info {
+    unsigned long cr2;
+    unsigned long pad; /* sizeof(vcpu_info_t) == 64 */
+};
+typedef struct arch_vcpu_info arch_vcpu_info_t;
+
+typedef unsigned long xen_callback_t;
+
+#endif /* !__ASSEMBLY__ */
+
+#endif /* __XEN_PUBLIC_ARCH_X86_XEN_X86_64_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/arch-x86_32.h linux-2.6.18-xen-3.3.0/include/xen/interface/arch-x86_32.h
--- linux-2.6.18.8/include/xen/interface/arch-x86_32.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/arch-x86_32.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,27 @@
+/******************************************************************************
+ * arch-x86_32.h
+ * 
+ * Guest OS interface to x86 32-bit Xen.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2004-2006, K A Fraser
+ */
+
+#include "arch-x86/xen.h"
diff -rpuN linux-2.6.18.8/include/xen/interface/arch-x86_64.h linux-2.6.18-xen-3.3.0/include/xen/interface/arch-x86_64.h
--- linux-2.6.18.8/include/xen/interface/arch-x86_64.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/arch-x86_64.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,27 @@
+/******************************************************************************
+ * arch-x86_64.h
+ * 
+ * Guest OS interface to x86 64-bit Xen.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2004-2006, K A Fraser
+ */
+
+#include "arch-x86/xen.h"
diff -rpuN linux-2.6.18.8/include/xen/interface/callback.h linux-2.6.18-xen-3.3.0/include/xen/interface/callback.h
--- linux-2.6.18.8/include/xen/interface/callback.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/callback.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,121 @@
+/******************************************************************************
+ * callback.h
+ *
+ * Register guest OS callbacks with Xen.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2006, Ian Campbell
+ */
+
+#ifndef __XEN_PUBLIC_CALLBACK_H__
+#define __XEN_PUBLIC_CALLBACK_H__
+
+#include "xen.h"
+
+/*
+ * Prototype for this hypercall is:
+ *   long callback_op(int cmd, void *extra_args)
+ * @cmd        == CALLBACKOP_??? (callback operation).
+ * @extra_args == Operation-specific extra arguments (NULL if none).
+ */
+
+/* ia64, x86: Callback for event delivery. */
+#define CALLBACKTYPE_event                 0
+
+/* x86: Failsafe callback when guest state cannot be restored by Xen. */
+#define CALLBACKTYPE_failsafe              1
+
+/* x86/64 hypervisor: Syscall by 64-bit guest app ('64-on-64-on-64'). */
+#define CALLBACKTYPE_syscall               2
+
+/*
+ * x86/32 hypervisor: Only available on x86/32 when supervisor_mode_kernel
+ *     feature is enabled. Do not use this callback type in new code.
+ */
+#define CALLBACKTYPE_sysenter_deprecated   3
+
+/* x86: Callback for NMI delivery. */
+#define CALLBACKTYPE_nmi                   4
+
+/*
+ * x86: sysenter is only available as follows:
+ * - 32-bit hypervisor: with the supervisor_mode_kernel feature enabled
+ * - 64-bit hypervisor: 32-bit guest applications on Intel CPUs
+ *                      ('32-on-32-on-64', '32-on-64-on-64')
+ *                      [nb. also 64-bit guest applications on Intel CPUs
+ *                           ('64-on-64-on-64'), but syscall is preferred]
+ */
+#define CALLBACKTYPE_sysenter              5
+
+/*
+ * x86/64 hypervisor: Syscall by 32-bit guest app on AMD CPUs
+ *                    ('32-on-32-on-64', '32-on-64-on-64')
+ */
+#define CALLBACKTYPE_syscall32             7
+
+/*
+ * Disable event deliver during callback? This flag is ignored for event and
+ * NMI callbacks: event delivery is unconditionally disabled.
+ */
+#define _CALLBACKF_mask_events             0
+#define CALLBACKF_mask_events              (1U << _CALLBACKF_mask_events)
+
+/*
+ * Register a callback.
+ */
+#define CALLBACKOP_register                0
+struct callback_register {
+    uint16_t type;
+    uint16_t flags;
+    xen_callback_t address;
+};
+typedef struct callback_register callback_register_t;
+DEFINE_XEN_GUEST_HANDLE(callback_register_t);
+
+/*
+ * Unregister a callback.
+ *
+ * Not all callbacks can be unregistered. -EINVAL will be returned if
+ * you attempt to unregister such a callback.
+ */
+#define CALLBACKOP_unregister              1
+struct callback_unregister {
+    uint16_t type;
+    uint16_t _unused;
+};
+typedef struct callback_unregister callback_unregister_t;
+DEFINE_XEN_GUEST_HANDLE(callback_unregister_t);
+
+#if __XEN_INTERFACE_VERSION__ < 0x00030207
+#undef CALLBACKTYPE_sysenter
+#define CALLBACKTYPE_sysenter CALLBACKTYPE_sysenter_deprecated
+#endif
+
+#endif /* __XEN_PUBLIC_CALLBACK_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/COPYING linux-2.6.18-xen-3.3.0/include/xen/interface/COPYING
--- linux-2.6.18.8/include/xen/interface/COPYING	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/COPYING	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,38 @@
+XEN NOTICE
+==========
+
+This copyright applies to all files within this subdirectory and its
+subdirectories:
+  include/public/*.h
+  include/public/hvm/*.h
+  include/public/io/*.h
+
+The intention is that these files can be freely copied into the source
+tree of an operating system when porting that OS to run on Xen. Doing
+so does *not* cause the OS to become subject to the terms of the GPL.
+
+All other files in the Xen source distribution are covered by version
+2 of the GNU General Public License except where explicitly stated
+otherwise within individual source files.
+
+ -- Keir Fraser (on behalf of the Xen team)
+
+=====================================================================
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to
+deal in the Software without restriction, including without limitation the
+rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+sell copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE 
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING 
+FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER 
+DEALINGS IN THE SOFTWARE.
diff -rpuN linux-2.6.18.8/include/xen/interface/dom0_ops.h linux-2.6.18-xen-3.3.0/include/xen/interface/dom0_ops.h
--- linux-2.6.18.8/include/xen/interface/dom0_ops.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/dom0_ops.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,120 @@
+/******************************************************************************
+ * dom0_ops.h
+ * 
+ * Process command requests from domain-0 guest OS.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2002-2003, B Dragovic
+ * Copyright (c) 2002-2006, K Fraser
+ */
+
+#ifndef __XEN_PUBLIC_DOM0_OPS_H__
+#define __XEN_PUBLIC_DOM0_OPS_H__
+
+#include "xen.h"
+#include "platform.h"
+
+#if __XEN_INTERFACE_VERSION__ >= 0x00030204
+#error "dom0_ops.h is a compatibility interface only"
+#endif
+
+#define DOM0_INTERFACE_VERSION XENPF_INTERFACE_VERSION
+
+#define DOM0_SETTIME          XENPF_settime
+#define dom0_settime          xenpf_settime
+#define dom0_settime_t        xenpf_settime_t
+
+#define DOM0_ADD_MEMTYPE      XENPF_add_memtype
+#define dom0_add_memtype      xenpf_add_memtype
+#define dom0_add_memtype_t    xenpf_add_memtype_t
+
+#define DOM0_DEL_MEMTYPE      XENPF_del_memtype
+#define dom0_del_memtype      xenpf_del_memtype
+#define dom0_del_memtype_t    xenpf_del_memtype_t
+
+#define DOM0_READ_MEMTYPE     XENPF_read_memtype
+#define dom0_read_memtype     xenpf_read_memtype
+#define dom0_read_memtype_t   xenpf_read_memtype_t
+
+#define DOM0_MICROCODE        XENPF_microcode_update
+#define dom0_microcode        xenpf_microcode_update
+#define dom0_microcode_t      xenpf_microcode_update_t
+
+#define DOM0_PLATFORM_QUIRK   XENPF_platform_quirk
+#define dom0_platform_quirk   xenpf_platform_quirk
+#define dom0_platform_quirk_t xenpf_platform_quirk_t
+
+typedef uint64_t cpumap_t;
+
+/* Unsupported legacy operation -- defined for API compatibility. */
+#define DOM0_MSR                 15
+struct dom0_msr {
+    /* IN variables. */
+    uint32_t write;
+    cpumap_t cpu_mask;
+    uint32_t msr;
+    uint32_t in1;
+    uint32_t in2;
+    /* OUT variables. */
+    uint32_t out1;
+    uint32_t out2;
+};
+typedef struct dom0_msr dom0_msr_t;
+DEFINE_XEN_GUEST_HANDLE(dom0_msr_t);
+
+/* Unsupported legacy operation -- defined for API compatibility. */
+#define DOM0_PHYSICAL_MEMORY_MAP 40
+struct dom0_memory_map_entry {
+    uint64_t start, end;
+    uint32_t flags; /* reserved */
+    uint8_t  is_ram;
+};
+typedef struct dom0_memory_map_entry dom0_memory_map_entry_t;
+DEFINE_XEN_GUEST_HANDLE(dom0_memory_map_entry_t);
+
+struct dom0_op {
+    uint32_t cmd;
+    uint32_t interface_version; /* DOM0_INTERFACE_VERSION */
+    union {
+        struct dom0_msr               msr;
+        struct dom0_settime           settime;
+        struct dom0_add_memtype       add_memtype;
+        struct dom0_del_memtype       del_memtype;
+        struct dom0_read_memtype      read_memtype;
+        struct dom0_microcode         microcode;
+        struct dom0_platform_quirk    platform_quirk;
+        struct dom0_memory_map_entry  physical_memory_map;
+        uint8_t                       pad[128];
+    } u;
+};
+typedef struct dom0_op dom0_op_t;
+DEFINE_XEN_GUEST_HANDLE(dom0_op_t);
+
+#endif /* __XEN_PUBLIC_DOM0_OPS_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/domctl.h linux-2.6.18-xen-3.3.0/include/xen/interface/domctl.h
--- linux-2.6.18.8/include/xen/interface/domctl.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/domctl.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,676 @@
+/******************************************************************************
+ * domctl.h
+ * 
+ * Domain management operations. For use by node control stack.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2002-2003, B Dragovic
+ * Copyright (c) 2002-2006, K Fraser
+ */
+
+#ifndef __XEN_PUBLIC_DOMCTL_H__
+#define __XEN_PUBLIC_DOMCTL_H__
+
+#if !defined(__XEN__) && !defined(__XEN_TOOLS__)
+#error "domctl operations are intended for use by node control tools only"
+#endif
+
+#include "xen.h"
+
+#define XEN_DOMCTL_INTERFACE_VERSION 0x00000005
+
+struct xenctl_cpumap {
+    XEN_GUEST_HANDLE_64(uint8) bitmap;
+    uint32_t nr_cpus;
+};
+
+/*
+ * NB. xen_domctl.domain is an IN/OUT parameter for this operation.
+ * If it is specified as zero, an id is auto-allocated and returned.
+ */
+#define XEN_DOMCTL_createdomain       1
+struct xen_domctl_createdomain {
+    /* IN parameters */
+    uint32_t ssidref;
+    xen_domain_handle_t handle;
+ /* Is this an HVM guest (as opposed to a PV guest)? */
+#define _XEN_DOMCTL_CDF_hvm_guest 0
+#define XEN_DOMCTL_CDF_hvm_guest  (1U<<_XEN_DOMCTL_CDF_hvm_guest)
+ /* Use hardware-assisted paging if available? */
+#define _XEN_DOMCTL_CDF_hap       1
+#define XEN_DOMCTL_CDF_hap        (1U<<_XEN_DOMCTL_CDF_hap)
+    uint32_t flags;
+};
+typedef struct xen_domctl_createdomain xen_domctl_createdomain_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_createdomain_t);
+
+#define XEN_DOMCTL_destroydomain      2
+#define XEN_DOMCTL_pausedomain        3
+#define XEN_DOMCTL_unpausedomain      4
+#define XEN_DOMCTL_resumedomain      27
+
+#define XEN_DOMCTL_getdomaininfo      5
+struct xen_domctl_getdomaininfo {
+    /* OUT variables. */
+    domid_t  domain;              /* Also echoed in domctl.domain */
+ /* Domain is scheduled to die. */
+#define _XEN_DOMINF_dying     0
+#define XEN_DOMINF_dying      (1U<<_XEN_DOMINF_dying)
+ /* Domain is an HVM guest (as opposed to a PV guest). */
+#define _XEN_DOMINF_hvm_guest 1
+#define XEN_DOMINF_hvm_guest  (1U<<_XEN_DOMINF_hvm_guest)
+ /* The guest OS has shut down. */
+#define _XEN_DOMINF_shutdown  2
+#define XEN_DOMINF_shutdown   (1U<<_XEN_DOMINF_shutdown)
+ /* Currently paused by control software. */
+#define _XEN_DOMINF_paused    3
+#define XEN_DOMINF_paused     (1U<<_XEN_DOMINF_paused)
+ /* Currently blocked pending an event.     */
+#define _XEN_DOMINF_blocked   4
+#define XEN_DOMINF_blocked    (1U<<_XEN_DOMINF_blocked)
+ /* Domain is currently running.            */
+#define _XEN_DOMINF_running   5
+#define XEN_DOMINF_running    (1U<<_XEN_DOMINF_running)
+ /* Being debugged.  */
+#define _XEN_DOMINF_debugged  6
+#define XEN_DOMINF_debugged   (1U<<_XEN_DOMINF_debugged)
+ /* CPU to which this domain is bound.      */
+#define XEN_DOMINF_cpumask      255
+#define XEN_DOMINF_cpushift       8
+ /* XEN_DOMINF_shutdown guest-supplied code.  */
+#define XEN_DOMINF_shutdownmask 255
+#define XEN_DOMINF_shutdownshift 16
+    uint32_t flags;              /* XEN_DOMINF_* */
+    uint64_aligned_t tot_pages;
+    uint64_aligned_t max_pages;
+    uint64_aligned_t shared_info_frame; /* GMFN of shared_info struct */
+    uint64_aligned_t cpu_time;
+    uint32_t nr_online_vcpus;    /* Number of VCPUs currently online. */
+    uint32_t max_vcpu_id;        /* Maximum VCPUID in use by this domain. */
+    uint32_t ssidref;
+    xen_domain_handle_t handle;
+};
+typedef struct xen_domctl_getdomaininfo xen_domctl_getdomaininfo_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_getdomaininfo_t);
+
+
+#define XEN_DOMCTL_getmemlist         6
+struct xen_domctl_getmemlist {
+    /* IN variables. */
+    /* Max entries to write to output buffer. */
+    uint64_aligned_t max_pfns;
+    /* Start index in guest's page list. */
+    uint64_aligned_t start_pfn;
+    XEN_GUEST_HANDLE_64(uint64) buffer;
+    /* OUT variables. */
+    uint64_aligned_t num_pfns;
+};
+typedef struct xen_domctl_getmemlist xen_domctl_getmemlist_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_getmemlist_t);
+
+
+#define XEN_DOMCTL_getpageframeinfo   7
+
+#define XEN_DOMCTL_PFINFO_LTAB_SHIFT 28
+#define XEN_DOMCTL_PFINFO_NOTAB   (0x0U<<28)
+#define XEN_DOMCTL_PFINFO_L1TAB   (0x1U<<28)
+#define XEN_DOMCTL_PFINFO_L2TAB   (0x2U<<28)
+#define XEN_DOMCTL_PFINFO_L3TAB   (0x3U<<28)
+#define XEN_DOMCTL_PFINFO_L4TAB   (0x4U<<28)
+#define XEN_DOMCTL_PFINFO_LTABTYPE_MASK (0x7U<<28)
+#define XEN_DOMCTL_PFINFO_LPINTAB (0x1U<<31)
+#define XEN_DOMCTL_PFINFO_XTAB    (0xfU<<28) /* invalid page */
+#define XEN_DOMCTL_PFINFO_LTAB_MASK (0xfU<<28)
+
+struct xen_domctl_getpageframeinfo {
+    /* IN variables. */
+    uint64_aligned_t gmfn; /* GMFN to query */
+    /* OUT variables. */
+    /* Is the page PINNED to a type? */
+    uint32_t type;         /* see above type defs */
+};
+typedef struct xen_domctl_getpageframeinfo xen_domctl_getpageframeinfo_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_getpageframeinfo_t);
+
+
+#define XEN_DOMCTL_getpageframeinfo2  8
+struct xen_domctl_getpageframeinfo2 {
+    /* IN variables. */
+    uint64_aligned_t num;
+    /* IN/OUT variables. */
+    XEN_GUEST_HANDLE_64(uint32) array;
+};
+typedef struct xen_domctl_getpageframeinfo2 xen_domctl_getpageframeinfo2_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_getpageframeinfo2_t);
+
+
+/*
+ * Control shadow pagetables operation
+ */
+#define XEN_DOMCTL_shadow_op         10
+
+/* Disable shadow mode. */
+#define XEN_DOMCTL_SHADOW_OP_OFF         0
+
+/* Enable shadow mode (mode contains ORed XEN_DOMCTL_SHADOW_ENABLE_* flags). */
+#define XEN_DOMCTL_SHADOW_OP_ENABLE      32
+
+/* Log-dirty bitmap operations. */
+ /* Return the bitmap and clean internal copy for next round. */
+#define XEN_DOMCTL_SHADOW_OP_CLEAN       11
+ /* Return the bitmap but do not modify internal copy. */
+#define XEN_DOMCTL_SHADOW_OP_PEEK        12
+
+/* Memory allocation accessors. */
+#define XEN_DOMCTL_SHADOW_OP_GET_ALLOCATION   30
+#define XEN_DOMCTL_SHADOW_OP_SET_ALLOCATION   31
+
+/* Legacy enable operations. */
+ /* Equiv. to ENABLE with no mode flags. */
+#define XEN_DOMCTL_SHADOW_OP_ENABLE_TEST       1
+ /* Equiv. to ENABLE with mode flag ENABLE_LOG_DIRTY. */
+#define XEN_DOMCTL_SHADOW_OP_ENABLE_LOGDIRTY   2
+ /* Equiv. to ENABLE with mode flags ENABLE_REFCOUNT and ENABLE_TRANSLATE. */
+#define XEN_DOMCTL_SHADOW_OP_ENABLE_TRANSLATE  3
+
+/* Mode flags for XEN_DOMCTL_SHADOW_OP_ENABLE. */
+ /*
+  * Shadow pagetables are refcounted: guest does not use explicit mmu
+  * operations nor write-protect its pagetables.
+  */
+#define XEN_DOMCTL_SHADOW_ENABLE_REFCOUNT  (1 << 1)
+ /*
+  * Log pages in a bitmap as they are dirtied.
+  * Used for live relocation to determine which pages must be re-sent.
+  */
+#define XEN_DOMCTL_SHADOW_ENABLE_LOG_DIRTY (1 << 2)
+ /*
+  * Automatically translate GPFNs into MFNs.
+  */
+#define XEN_DOMCTL_SHADOW_ENABLE_TRANSLATE (1 << 3)
+ /*
+  * Xen does not steal virtual address space from the guest.
+  * Requires HVM support.
+  */
+#define XEN_DOMCTL_SHADOW_ENABLE_EXTERNAL  (1 << 4)
+
+struct xen_domctl_shadow_op_stats {
+    uint32_t fault_count;
+    uint32_t dirty_count;
+};
+typedef struct xen_domctl_shadow_op_stats xen_domctl_shadow_op_stats_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_shadow_op_stats_t);
+
+struct xen_domctl_shadow_op {
+    /* IN variables. */
+    uint32_t       op;       /* XEN_DOMCTL_SHADOW_OP_* */
+
+    /* OP_ENABLE */
+    uint32_t       mode;     /* XEN_DOMCTL_SHADOW_ENABLE_* */
+
+    /* OP_GET_ALLOCATION / OP_SET_ALLOCATION */
+    uint32_t       mb;       /* Shadow memory allocation in MB */
+
+    /* OP_PEEK / OP_CLEAN */
+    XEN_GUEST_HANDLE_64(uint8) dirty_bitmap;
+    uint64_aligned_t pages; /* Size of buffer. Updated with actual size. */
+    struct xen_domctl_shadow_op_stats stats;
+};
+typedef struct xen_domctl_shadow_op xen_domctl_shadow_op_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_shadow_op_t);
+
+
+#define XEN_DOMCTL_max_mem           11
+struct xen_domctl_max_mem {
+    /* IN variables. */
+    uint64_aligned_t max_memkb;
+};
+typedef struct xen_domctl_max_mem xen_domctl_max_mem_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_max_mem_t);
+
+
+#define XEN_DOMCTL_setvcpucontext    12
+#define XEN_DOMCTL_getvcpucontext    13
+struct xen_domctl_vcpucontext {
+    uint32_t              vcpu;                  /* IN */
+    XEN_GUEST_HANDLE_64(vcpu_guest_context_t) ctxt; /* IN/OUT */
+};
+typedef struct xen_domctl_vcpucontext xen_domctl_vcpucontext_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_vcpucontext_t);
+
+
+#define XEN_DOMCTL_getvcpuinfo       14
+struct xen_domctl_getvcpuinfo {
+    /* IN variables. */
+    uint32_t vcpu;
+    /* OUT variables. */
+    uint8_t  online;                  /* currently online (not hotplugged)? */
+    uint8_t  blocked;                 /* blocked waiting for an event? */
+    uint8_t  running;                 /* currently scheduled on its CPU? */
+    uint64_aligned_t cpu_time;        /* total cpu time consumed (ns) */
+    uint32_t cpu;                     /* current mapping   */
+};
+typedef struct xen_domctl_getvcpuinfo xen_domctl_getvcpuinfo_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_getvcpuinfo_t);
+
+
+/* Get/set which physical cpus a vcpu can execute on. */
+#define XEN_DOMCTL_setvcpuaffinity    9
+#define XEN_DOMCTL_getvcpuaffinity   25
+struct xen_domctl_vcpuaffinity {
+    uint32_t  vcpu;              /* IN */
+    struct xenctl_cpumap cpumap; /* IN/OUT */
+};
+typedef struct xen_domctl_vcpuaffinity xen_domctl_vcpuaffinity_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_vcpuaffinity_t);
+
+
+#define XEN_DOMCTL_max_vcpus         15
+struct xen_domctl_max_vcpus {
+    uint32_t max;           /* maximum number of vcpus */
+};
+typedef struct xen_domctl_max_vcpus xen_domctl_max_vcpus_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_max_vcpus_t);
+
+
+#define XEN_DOMCTL_scheduler_op      16
+/* Scheduler types. */
+#define XEN_SCHEDULER_SEDF     4
+#define XEN_SCHEDULER_CREDIT   5
+/* Set or get info? */
+#define XEN_DOMCTL_SCHEDOP_putinfo 0
+#define XEN_DOMCTL_SCHEDOP_getinfo 1
+struct xen_domctl_scheduler_op {
+    uint32_t sched_id;  /* XEN_SCHEDULER_* */
+    uint32_t cmd;       /* XEN_DOMCTL_SCHEDOP_* */
+    union {
+        struct xen_domctl_sched_sedf {
+            uint64_aligned_t period;
+            uint64_aligned_t slice;
+            uint64_aligned_t latency;
+            uint32_t extratime;
+            uint32_t weight;
+        } sedf;
+        struct xen_domctl_sched_credit {
+            uint16_t weight;
+            uint16_t cap;
+        } credit;
+    } u;
+};
+typedef struct xen_domctl_scheduler_op xen_domctl_scheduler_op_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_scheduler_op_t);
+
+
+#define XEN_DOMCTL_setdomainhandle   17
+struct xen_domctl_setdomainhandle {
+    xen_domain_handle_t handle;
+};
+typedef struct xen_domctl_setdomainhandle xen_domctl_setdomainhandle_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_setdomainhandle_t);
+
+
+#define XEN_DOMCTL_setdebugging      18
+struct xen_domctl_setdebugging {
+    uint8_t enable;
+};
+typedef struct xen_domctl_setdebugging xen_domctl_setdebugging_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_setdebugging_t);
+
+
+#define XEN_DOMCTL_irq_permission    19
+struct xen_domctl_irq_permission {
+    uint8_t pirq;
+    uint8_t allow_access;    /* flag to specify enable/disable of IRQ access */
+};
+typedef struct xen_domctl_irq_permission xen_domctl_irq_permission_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_irq_permission_t);
+
+
+#define XEN_DOMCTL_iomem_permission  20
+struct xen_domctl_iomem_permission {
+    uint64_aligned_t first_mfn;/* first page (physical page number) in range */
+    uint64_aligned_t nr_mfns;  /* number of pages in range (>0) */
+    uint8_t  allow_access;     /* allow (!0) or deny (0) access to range? */
+};
+typedef struct xen_domctl_iomem_permission xen_domctl_iomem_permission_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_iomem_permission_t);
+
+
+#define XEN_DOMCTL_ioport_permission 21
+struct xen_domctl_ioport_permission {
+    uint32_t first_port;              /* first port int range */
+    uint32_t nr_ports;                /* size of port range */
+    uint8_t  allow_access;            /* allow or deny access to range? */
+};
+typedef struct xen_domctl_ioport_permission xen_domctl_ioport_permission_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_ioport_permission_t);
+
+
+#define XEN_DOMCTL_hypercall_init    22
+struct xen_domctl_hypercall_init {
+    uint64_aligned_t  gmfn;           /* GMFN to be initialised */
+};
+typedef struct xen_domctl_hypercall_init xen_domctl_hypercall_init_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_hypercall_init_t);
+
+
+#define XEN_DOMCTL_arch_setup        23
+#define _XEN_DOMAINSETUP_hvm_guest 0
+#define XEN_DOMAINSETUP_hvm_guest  (1UL<<_XEN_DOMAINSETUP_hvm_guest)
+#define _XEN_DOMAINSETUP_query 1 /* Get parameters (for save)  */
+#define XEN_DOMAINSETUP_query  (1UL<<_XEN_DOMAINSETUP_query)
+#define _XEN_DOMAINSETUP_sioemu_guest 2
+#define XEN_DOMAINSETUP_sioemu_guest  (1UL<<_XEN_DOMAINSETUP_sioemu_guest)
+typedef struct xen_domctl_arch_setup {
+    uint64_aligned_t flags;  /* XEN_DOMAINSETUP_* */
+#ifdef __ia64__
+    uint64_aligned_t bp;     /* mpaddr of boot param area */
+    uint64_aligned_t maxmem; /* Highest memory address for MDT.  */
+    uint64_aligned_t xsi_va; /* Xen shared_info area virtual address.  */
+    uint32_t hypercall_imm;  /* Break imm for Xen hypercalls.  */
+    int8_t vhpt_size_log2;   /* Log2 of VHPT size. */
+#endif
+} xen_domctl_arch_setup_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_arch_setup_t);
+
+
+#define XEN_DOMCTL_settimeoffset     24
+struct xen_domctl_settimeoffset {
+    int32_t  time_offset_seconds; /* applied to domain wallclock time */
+};
+typedef struct xen_domctl_settimeoffset xen_domctl_settimeoffset_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_settimeoffset_t);
+
+ 
+#define XEN_DOMCTL_gethvmcontext     33
+#define XEN_DOMCTL_sethvmcontext     34
+typedef struct xen_domctl_hvmcontext {
+    uint32_t size; /* IN/OUT: size of buffer / bytes filled */
+    XEN_GUEST_HANDLE_64(uint8) buffer; /* IN/OUT: data, or call
+                                        * gethvmcontext with NULL
+                                        * buffer to get size req'd */
+} xen_domctl_hvmcontext_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_hvmcontext_t);
+
+
+#define XEN_DOMCTL_set_address_size  35
+#define XEN_DOMCTL_get_address_size  36
+typedef struct xen_domctl_address_size {
+    uint32_t size;
+} xen_domctl_address_size_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_address_size_t);
+
+
+#define XEN_DOMCTL_real_mode_area    26
+struct xen_domctl_real_mode_area {
+    uint32_t log; /* log2 of Real Mode Area size */
+};
+typedef struct xen_domctl_real_mode_area xen_domctl_real_mode_area_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_real_mode_area_t);
+
+
+#define XEN_DOMCTL_sendtrigger       28
+#define XEN_DOMCTL_SENDTRIGGER_NMI    0
+#define XEN_DOMCTL_SENDTRIGGER_RESET  1
+#define XEN_DOMCTL_SENDTRIGGER_INIT   2
+struct xen_domctl_sendtrigger {
+    uint32_t  trigger;  /* IN */
+    uint32_t  vcpu;     /* IN */
+};
+typedef struct xen_domctl_sendtrigger xen_domctl_sendtrigger_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_sendtrigger_t);
+
+
+/* Assign PCI device to HVM guest. Sets up IOMMU structures. */
+#define XEN_DOMCTL_assign_device      37
+#define XEN_DOMCTL_test_assign_device 45
+#define XEN_DOMCTL_deassign_device 47
+struct xen_domctl_assign_device {
+    uint32_t  machine_bdf;   /* machine PCI ID of assigned device */
+};
+typedef struct xen_domctl_assign_device xen_domctl_assign_device_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_assign_device_t);
+
+/* Retrieve sibling devices infomation of machine_bdf */
+#define XEN_DOMCTL_get_device_group 50
+struct xen_domctl_get_device_group {
+    uint32_t  machine_bdf;      /* IN */
+    uint32_t  max_sdevs;        /* IN */
+    uint32_t  num_sdevs;        /* OUT */
+    XEN_GUEST_HANDLE_64(uint32)  sdev_array;   /* OUT */
+};
+typedef struct xen_domctl_get_device_group xen_domctl_get_device_group_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_get_device_group_t);
+
+/* Pass-through interrupts: bind real irq -> hvm devfn. */
+#define XEN_DOMCTL_bind_pt_irq       38
+#define XEN_DOMCTL_unbind_pt_irq     48
+typedef enum pt_irq_type_e {
+    PT_IRQ_TYPE_PCI,
+    PT_IRQ_TYPE_ISA,
+    PT_IRQ_TYPE_MSI,
+} pt_irq_type_t;
+struct xen_domctl_bind_pt_irq {
+    uint32_t machine_irq;
+    pt_irq_type_t irq_type;
+    uint32_t hvm_domid;
+
+    union {
+        struct {
+            uint8_t isa_irq;
+        } isa;
+        struct {
+            uint8_t bus;
+            uint8_t device;
+            uint8_t intx;
+        } pci;
+        struct {
+            uint8_t gvec;
+            uint32_t gflags;
+        } msi;
+    } u;
+};
+typedef struct xen_domctl_bind_pt_irq xen_domctl_bind_pt_irq_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_bind_pt_irq_t);
+
+
+/* Bind machine I/O address range -> HVM address range. */
+#define XEN_DOMCTL_memory_mapping    39
+#define DPCI_ADD_MAPPING         1
+#define DPCI_REMOVE_MAPPING      0
+struct xen_domctl_memory_mapping {
+    uint64_aligned_t first_gfn; /* first page (hvm guest phys page) in range */
+    uint64_aligned_t first_mfn; /* first page (machine page) in range */
+    uint64_aligned_t nr_mfns;   /* number of pages in range (>0) */
+    uint32_t add_mapping;       /* add or remove mapping */
+    uint32_t padding;           /* padding for 64-bit aligned structure */
+};
+typedef struct xen_domctl_memory_mapping xen_domctl_memory_mapping_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_memory_mapping_t);
+
+
+/* Bind machine I/O port range -> HVM I/O port range. */
+#define XEN_DOMCTL_ioport_mapping    40
+struct xen_domctl_ioport_mapping {
+    uint32_t first_gport;     /* first guest IO port*/
+    uint32_t first_mport;     /* first machine IO port */
+    uint32_t nr_ports;        /* size of port range */
+    uint32_t add_mapping;     /* add or remove mapping */
+};
+typedef struct xen_domctl_ioport_mapping xen_domctl_ioport_mapping_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_ioport_mapping_t);
+
+
+/*
+ * Pin caching type of RAM space for x86 HVM domU.
+ */
+#define XEN_DOMCTL_pin_mem_cacheattr 41
+/* Caching types: these happen to be the same as x86 MTRR/PAT type codes. */
+#define XEN_DOMCTL_MEM_CACHEATTR_UC  0
+#define XEN_DOMCTL_MEM_CACHEATTR_WC  1
+#define XEN_DOMCTL_MEM_CACHEATTR_WT  4
+#define XEN_DOMCTL_MEM_CACHEATTR_WP  5
+#define XEN_DOMCTL_MEM_CACHEATTR_WB  6
+#define XEN_DOMCTL_MEM_CACHEATTR_UCM 7
+struct xen_domctl_pin_mem_cacheattr {
+    uint64_aligned_t start, end;
+    unsigned int type; /* XEN_DOMCTL_MEM_CACHEATTR_* */
+};
+typedef struct xen_domctl_pin_mem_cacheattr xen_domctl_pin_mem_cacheattr_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_pin_mem_cacheattr_t);
+
+
+#define XEN_DOMCTL_set_ext_vcpucontext 42
+#define XEN_DOMCTL_get_ext_vcpucontext 43
+struct xen_domctl_ext_vcpucontext {
+    /* IN: VCPU that this call applies to. */
+    uint32_t         vcpu;
+    /*
+     * SET: Size of struct (IN)
+     * GET: Size of struct (OUT)
+     */
+    uint32_t         size;
+#if defined(__i386__) || defined(__x86_64__)
+    /* SYSCALL from 32-bit mode and SYSENTER callback information. */
+    /* NB. SYSCALL from 64-bit mode is contained in vcpu_guest_context_t */
+    uint64_aligned_t syscall32_callback_eip;
+    uint64_aligned_t sysenter_callback_eip;
+    uint16_t         syscall32_callback_cs;
+    uint16_t         sysenter_callback_cs;
+    uint8_t          syscall32_disables_events;
+    uint8_t          sysenter_disables_events;
+#endif
+};
+typedef struct xen_domctl_ext_vcpucontext xen_domctl_ext_vcpucontext_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_ext_vcpucontext_t);
+
+/*
+ * Set optimizaton features for a domain
+ */
+#define XEN_DOMCTL_set_opt_feature    44
+struct xen_domctl_set_opt_feature {
+#if defined(__ia64__)
+    struct xen_ia64_opt_feature optf;
+#else
+    /* Make struct non-empty: do not depend on this field name! */
+    uint64_t dummy;
+#endif
+};
+typedef struct xen_domctl_set_opt_feature xen_domctl_set_opt_feature_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_set_opt_feature_t);
+
+/*
+ * Set the target domain for a domain
+ */
+#define XEN_DOMCTL_set_target    46
+struct xen_domctl_set_target {
+    domid_t target;
+};
+typedef struct xen_domctl_set_target xen_domctl_set_target_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_set_target_t);
+
+#if defined(__i386__) || defined(__x86_64__)
+# define XEN_CPUID_INPUT_UNUSED  0xFFFFFFFF
+# define XEN_DOMCTL_set_cpuid 49
+struct xen_domctl_cpuid {
+  unsigned int  input[2];
+  unsigned int  eax;
+  unsigned int  ebx;
+  unsigned int  ecx;
+  unsigned int  edx;
+};
+typedef struct xen_domctl_cpuid xen_domctl_cpuid_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_cpuid_t);
+#endif
+
+#define XEN_DOMCTL_subscribe          29
+struct xen_domctl_subscribe {
+    uint32_t port; /* IN */
+};
+typedef struct xen_domctl_subscribe xen_domctl_subscribe_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_subscribe_t);
+
+/*
+ * Define the maximum machine address size which should be allocated
+ * to a guest.
+ */
+#define XEN_DOMCTL_set_machine_address_size  51
+#define XEN_DOMCTL_get_machine_address_size  52
+
+
+struct xen_domctl {
+    uint32_t cmd;
+    uint32_t interface_version; /* XEN_DOMCTL_INTERFACE_VERSION */
+    domid_t  domain;
+    union {
+        struct xen_domctl_createdomain      createdomain;
+        struct xen_domctl_getdomaininfo     getdomaininfo;
+        struct xen_domctl_getmemlist        getmemlist;
+        struct xen_domctl_getpageframeinfo  getpageframeinfo;
+        struct xen_domctl_getpageframeinfo2 getpageframeinfo2;
+        struct xen_domctl_vcpuaffinity      vcpuaffinity;
+        struct xen_domctl_shadow_op         shadow_op;
+        struct xen_domctl_max_mem           max_mem;
+        struct xen_domctl_vcpucontext       vcpucontext;
+        struct xen_domctl_getvcpuinfo       getvcpuinfo;
+        struct xen_domctl_max_vcpus         max_vcpus;
+        struct xen_domctl_scheduler_op      scheduler_op;
+        struct xen_domctl_setdomainhandle   setdomainhandle;
+        struct xen_domctl_setdebugging      setdebugging;
+        struct xen_domctl_irq_permission    irq_permission;
+        struct xen_domctl_iomem_permission  iomem_permission;
+        struct xen_domctl_ioport_permission ioport_permission;
+        struct xen_domctl_hypercall_init    hypercall_init;
+        struct xen_domctl_arch_setup        arch_setup;
+        struct xen_domctl_settimeoffset     settimeoffset;
+        struct xen_domctl_real_mode_area    real_mode_area;
+        struct xen_domctl_hvmcontext        hvmcontext;
+        struct xen_domctl_address_size      address_size;
+        struct xen_domctl_sendtrigger       sendtrigger;
+        struct xen_domctl_get_device_group  get_device_group;
+        struct xen_domctl_assign_device     assign_device;
+        struct xen_domctl_bind_pt_irq       bind_pt_irq;
+        struct xen_domctl_memory_mapping    memory_mapping;
+        struct xen_domctl_ioport_mapping    ioport_mapping;
+        struct xen_domctl_pin_mem_cacheattr pin_mem_cacheattr;
+        struct xen_domctl_ext_vcpucontext   ext_vcpucontext;
+        struct xen_domctl_set_opt_feature   set_opt_feature;
+        struct xen_domctl_set_target        set_target;
+        struct xen_domctl_subscribe         subscribe;
+#if defined(__i386__) || defined(__x86_64__)
+        struct xen_domctl_cpuid             cpuid;
+#endif
+        uint8_t                             pad[128];
+    } u;
+};
+typedef struct xen_domctl xen_domctl_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_t);
+
+#endif /* __XEN_PUBLIC_DOMCTL_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/elfnote.h linux-2.6.18-xen-3.3.0/include/xen/interface/elfnote.h
--- linux-2.6.18.8/include/xen/interface/elfnote.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/elfnote.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,233 @@
+/******************************************************************************
+ * elfnote.h
+ *
+ * Definitions used for the Xen ELF notes.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2006, Ian Campbell, XenSource Ltd.
+ */
+
+#ifndef __XEN_PUBLIC_ELFNOTE_H__
+#define __XEN_PUBLIC_ELFNOTE_H__
+
+/*
+ * The notes should live in a PT_NOTE segment and have "Xen" in the
+ * name field.
+ *
+ * Numeric types are either 4 or 8 bytes depending on the content of
+ * the desc field.
+ *
+ * LEGACY indicated the fields in the legacy __xen_guest string which
+ * this a note type replaces.
+ */
+
+/*
+ * NAME=VALUE pair (string).
+ */
+#define XEN_ELFNOTE_INFO           0
+
+/*
+ * The virtual address of the entry point (numeric).
+ *
+ * LEGACY: VIRT_ENTRY
+ */
+#define XEN_ELFNOTE_ENTRY          1
+
+/* The virtual address of the hypercall transfer page (numeric).
+ *
+ * LEGACY: HYPERCALL_PAGE. (n.b. legacy value is a physical page
+ * number not a virtual address)
+ */
+#define XEN_ELFNOTE_HYPERCALL_PAGE 2
+
+/* The virtual address where the kernel image should be mapped (numeric).
+ *
+ * Defaults to 0.
+ *
+ * LEGACY: VIRT_BASE
+ */
+#define XEN_ELFNOTE_VIRT_BASE      3
+
+/*
+ * The offset of the ELF paddr field from the acutal required
+ * psuedo-physical address (numeric).
+ *
+ * This is used to maintain backwards compatibility with older kernels
+ * which wrote __PAGE_OFFSET into that field. This field defaults to 0
+ * if not present.
+ *
+ * LEGACY: ELF_PADDR_OFFSET. (n.b. legacy default is VIRT_BASE)
+ */
+#define XEN_ELFNOTE_PADDR_OFFSET   4
+
+/*
+ * The version of Xen that we work with (string).
+ *
+ * LEGACY: XEN_VER
+ */
+#define XEN_ELFNOTE_XEN_VERSION    5
+
+/*
+ * The name of the guest operating system (string).
+ *
+ * LEGACY: GUEST_OS
+ */
+#define XEN_ELFNOTE_GUEST_OS       6
+
+/*
+ * The version of the guest operating system (string).
+ *
+ * LEGACY: GUEST_VER
+ */
+#define XEN_ELFNOTE_GUEST_VERSION  7
+
+/*
+ * The loader type (string).
+ *
+ * LEGACY: LOADER
+ */
+#define XEN_ELFNOTE_LOADER         8
+
+/*
+ * The kernel supports PAE (x86/32 only, string = "yes", "no" or
+ * "bimodal").
+ *
+ * For compatibility with Xen 3.0.3 and earlier the "bimodal" setting
+ * may be given as "yes,bimodal" which will cause older Xen to treat
+ * this kernel as PAE.
+ *
+ * LEGACY: PAE (n.b. The legacy interface included a provision to
+ * indicate 'extended-cr3' support allowing L3 page tables to be
+ * placed above 4G. It is assumed that any kernel new enough to use
+ * these ELF notes will include this and therefore "yes" here is
+ * equivalent to "yes[entended-cr3]" in the __xen_guest interface.
+ */
+#define XEN_ELFNOTE_PAE_MODE       9
+
+/*
+ * The features supported/required by this kernel (string).
+ *
+ * The string must consist of a list of feature names (as given in
+ * features.h, without the "XENFEAT_" prefix) separated by '|'
+ * characters. If a feature is required for the kernel to function
+ * then the feature name must be preceded by a '!' character.
+ *
+ * LEGACY: FEATURES
+ */
+#define XEN_ELFNOTE_FEATURES      10
+
+/*
+ * The kernel requires the symbol table to be loaded (string = "yes" or "no")
+ * LEGACY: BSD_SYMTAB (n.b. The legacy treated the presence or absence
+ * of this string as a boolean flag rather than requiring "yes" or
+ * "no".
+ */
+#define XEN_ELFNOTE_BSD_SYMTAB    11
+
+/*
+ * The lowest address the hypervisor hole can begin at (numeric).
+ *
+ * This must not be set higher than HYPERVISOR_VIRT_START. Its presence
+ * also indicates to the hypervisor that the kernel can deal with the
+ * hole starting at a higher address.
+ */
+#define XEN_ELFNOTE_HV_START_LOW  12
+
+/*
+ * List of maddr_t-sized mask/value pairs describing how to recognize
+ * (non-present) L1 page table entries carrying valid MFNs (numeric).
+ */
+#define XEN_ELFNOTE_L1_MFN_VALID  13
+
+/*
+ * Whether or not the guest supports cooperative suspend cancellation.
+ */
+#define XEN_ELFNOTE_SUSPEND_CANCEL 14
+
+/*
+ * The number of the highest elfnote defined.
+ */
+#define XEN_ELFNOTE_MAX XEN_ELFNOTE_SUSPEND_CANCEL
+
+/*
+ * System information exported through crash notes.
+ *
+ * The kexec / kdump code will create one XEN_ELFNOTE_CRASH_INFO 
+ * note in case of a system crash. This note will contain various
+ * information about the system, see xen/include/xen/elfcore.h.
+ */
+#define XEN_ELFNOTE_CRASH_INFO 0x1000001
+
+/*
+ * System registers exported through crash notes.
+ *
+ * The kexec / kdump code will create one XEN_ELFNOTE_CRASH_REGS 
+ * note per cpu in case of a system crash. This note is architecture
+ * specific and will contain registers not saved in the "CORE" note.
+ * See xen/include/xen/elfcore.h for more information.
+ */
+#define XEN_ELFNOTE_CRASH_REGS 0x1000002
+
+
+/*
+ * xen dump-core none note.
+ * xm dump-core code will create one XEN_ELFNOTE_DUMPCORE_NONE
+ * in its dump file to indicate that the file is xen dump-core
+ * file. This note doesn't have any other information.
+ * See tools/libxc/xc_core.h for more information.
+ */
+#define XEN_ELFNOTE_DUMPCORE_NONE               0x2000000
+
+/*
+ * xen dump-core header note.
+ * xm dump-core code will create one XEN_ELFNOTE_DUMPCORE_HEADER
+ * in its dump file.
+ * See tools/libxc/xc_core.h for more information.
+ */
+#define XEN_ELFNOTE_DUMPCORE_HEADER             0x2000001
+
+/*
+ * xen dump-core xen version note.
+ * xm dump-core code will create one XEN_ELFNOTE_DUMPCORE_XEN_VERSION
+ * in its dump file. It contains the xen version obtained via the
+ * XENVER hypercall.
+ * See tools/libxc/xc_core.h for more information.
+ */
+#define XEN_ELFNOTE_DUMPCORE_XEN_VERSION        0x2000002
+
+/*
+ * xen dump-core format version note.
+ * xm dump-core code will create one XEN_ELFNOTE_DUMPCORE_FORMAT_VERSION
+ * in its dump file. It contains a format version identifier.
+ * See tools/libxc/xc_core.h for more information.
+ */
+#define XEN_ELFNOTE_DUMPCORE_FORMAT_VERSION     0x2000003
+
+#endif /* __XEN_PUBLIC_ELFNOTE_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/elfstructs.h linux-2.6.18-xen-3.3.0/include/xen/interface/elfstructs.h
--- linux-2.6.18.8/include/xen/interface/elfstructs.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/elfstructs.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,527 @@
+#ifndef __XEN_PUBLIC_ELFSTRUCTS_H__
+#define __XEN_PUBLIC_ELFSTRUCTS_H__ 1
+/*
+ * Copyright (c) 1995, 1996 Erik Theisen.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. The name of the author may not be used to endorse or promote products
+ *    derived from this software without specific prior written permission
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+typedef uint8_t		Elf_Byte;
+
+typedef uint32_t	Elf32_Addr;	/* Unsigned program address */
+typedef uint32_t	Elf32_Off;	/* Unsigned file offset */
+typedef int32_t		Elf32_Sword;	/* Signed large integer */
+typedef uint32_t	Elf32_Word;	/* Unsigned large integer */
+typedef uint16_t	Elf32_Half;	/* Unsigned medium integer */
+
+typedef uint64_t	Elf64_Addr;
+typedef uint64_t	Elf64_Off;
+typedef int32_t		Elf64_Shalf;
+
+typedef int32_t		Elf64_Sword;
+typedef uint32_t	Elf64_Word;
+
+typedef int64_t		Elf64_Sxword;
+typedef uint64_t	Elf64_Xword;
+
+typedef uint32_t	Elf64_Half;
+typedef uint16_t	Elf64_Quarter;
+
+/*
+ * e_ident[] identification indexes
+ * See http://www.caldera.com/developers/gabi/2000-07-17/ch4.eheader.html
+ */
+#define EI_MAG0		0		/* file ID */
+#define EI_MAG1		1		/* file ID */
+#define EI_MAG2		2		/* file ID */
+#define EI_MAG3		3		/* file ID */
+#define EI_CLASS	4		/* file class */
+#define EI_DATA		5		/* data encoding */
+#define EI_VERSION	6		/* ELF header version */
+#define EI_OSABI	7		/* OS/ABI ID */
+#define EI_ABIVERSION	8		/* ABI version */
+#define EI_PAD		9		/* start of pad bytes */
+#define EI_NIDENT	16		/* Size of e_ident[] */
+
+/* e_ident[] magic number */
+#define	ELFMAG0		0x7f		/* e_ident[EI_MAG0] */
+#define	ELFMAG1		'E'		/* e_ident[EI_MAG1] */
+#define	ELFMAG2		'L'		/* e_ident[EI_MAG2] */
+#define	ELFMAG3		'F'		/* e_ident[EI_MAG3] */
+#define	ELFMAG		"\177ELF"	/* magic */
+#define	SELFMAG		4		/* size of magic */
+
+/* e_ident[] file class */
+#define	ELFCLASSNONE	0		/* invalid */
+#define	ELFCLASS32	1		/* 32-bit objs */
+#define	ELFCLASS64	2		/* 64-bit objs */
+#define	ELFCLASSNUM	3		/* number of classes */
+
+/* e_ident[] data encoding */
+#define ELFDATANONE	0		/* invalid */
+#define ELFDATA2LSB	1		/* Little-Endian */
+#define ELFDATA2MSB	2		/* Big-Endian */
+#define ELFDATANUM	3		/* number of data encode defines */
+
+/* e_ident[] Operating System/ABI */
+#define ELFOSABI_SYSV		0	/* UNIX System V ABI */
+#define ELFOSABI_HPUX		1	/* HP-UX operating system */
+#define ELFOSABI_NETBSD		2	/* NetBSD */
+#define ELFOSABI_LINUX		3	/* GNU/Linux */
+#define ELFOSABI_HURD		4	/* GNU/Hurd */
+#define ELFOSABI_86OPEN		5	/* 86Open common IA32 ABI */
+#define ELFOSABI_SOLARIS	6	/* Solaris */
+#define ELFOSABI_MONTEREY	7	/* Monterey */
+#define ELFOSABI_IRIX		8	/* IRIX */
+#define ELFOSABI_FREEBSD	9	/* FreeBSD */
+#define ELFOSABI_TRU64		10	/* TRU64 UNIX */
+#define ELFOSABI_MODESTO	11	/* Novell Modesto */
+#define ELFOSABI_OPENBSD	12	/* OpenBSD */
+#define ELFOSABI_ARM		97	/* ARM */
+#define ELFOSABI_STANDALONE	255	/* Standalone (embedded) application */
+
+/* e_ident */
+#define IS_ELF(ehdr) ((ehdr).e_ident[EI_MAG0] == ELFMAG0 && \
+                      (ehdr).e_ident[EI_MAG1] == ELFMAG1 && \
+                      (ehdr).e_ident[EI_MAG2] == ELFMAG2 && \
+                      (ehdr).e_ident[EI_MAG3] == ELFMAG3)
+
+/* ELF Header */
+typedef struct elfhdr {
+	unsigned char	e_ident[EI_NIDENT]; /* ELF Identification */
+	Elf32_Half	e_type;		/* object file type */
+	Elf32_Half	e_machine;	/* machine */
+	Elf32_Word	e_version;	/* object file version */
+	Elf32_Addr	e_entry;	/* virtual entry point */
+	Elf32_Off	e_phoff;	/* program header table offset */
+	Elf32_Off	e_shoff;	/* section header table offset */
+	Elf32_Word	e_flags;	/* processor-specific flags */
+	Elf32_Half	e_ehsize;	/* ELF header size */
+	Elf32_Half	e_phentsize;	/* program header entry size */
+	Elf32_Half	e_phnum;	/* number of program header entries */
+	Elf32_Half	e_shentsize;	/* section header entry size */
+	Elf32_Half	e_shnum;	/* number of section header entries */
+	Elf32_Half	e_shstrndx;	/* section header table's "section
+					   header string table" entry offset */
+} Elf32_Ehdr;
+
+typedef struct {
+	unsigned char	e_ident[EI_NIDENT];	/* Id bytes */
+	Elf64_Quarter	e_type;			/* file type */
+	Elf64_Quarter	e_machine;		/* machine type */
+	Elf64_Half	e_version;		/* version number */
+	Elf64_Addr	e_entry;		/* entry point */
+	Elf64_Off	e_phoff;		/* Program hdr offset */
+	Elf64_Off	e_shoff;		/* Section hdr offset */
+	Elf64_Half	e_flags;		/* Processor flags */
+	Elf64_Quarter	e_ehsize;		/* sizeof ehdr */
+	Elf64_Quarter	e_phentsize;		/* Program header entry size */
+	Elf64_Quarter	e_phnum;		/* Number of program headers */
+	Elf64_Quarter	e_shentsize;		/* Section header entry size */
+	Elf64_Quarter	e_shnum;		/* Number of section headers */
+	Elf64_Quarter	e_shstrndx;		/* String table index */
+} Elf64_Ehdr;
+
+/* e_type */
+#define ET_NONE		0		/* No file type */
+#define ET_REL		1		/* relocatable file */
+#define ET_EXEC		2		/* executable file */
+#define ET_DYN		3		/* shared object file */
+#define ET_CORE		4		/* core file */
+#define ET_NUM		5		/* number of types */
+#define ET_LOPROC	0xff00		/* reserved range for processor */
+#define ET_HIPROC	0xffff		/*  specific e_type */
+
+/* e_machine */
+#define EM_NONE		0		/* No Machine */
+#define EM_M32		1		/* AT&T WE 32100 */
+#define EM_SPARC	2		/* SPARC */
+#define EM_386		3		/* Intel 80386 */
+#define EM_68K		4		/* Motorola 68000 */
+#define EM_88K		5		/* Motorola 88000 */
+#define EM_486		6		/* Intel 80486 - unused? */
+#define EM_860		7		/* Intel 80860 */
+#define EM_MIPS		8		/* MIPS R3000 Big-Endian only */
+/*
+ * Don't know if EM_MIPS_RS4_BE,
+ * EM_SPARC64, EM_PARISC,
+ * or EM_PPC are ABI compliant
+ */
+#define EM_MIPS_RS4_BE	10		/* MIPS R4000 Big-Endian */
+#define EM_SPARC64	11		/* SPARC v9 64-bit unoffical */
+#define EM_PARISC	15		/* HPPA */
+#define EM_SPARC32PLUS	18		/* Enhanced instruction set SPARC */
+#define EM_PPC		20		/* PowerPC */
+#define EM_PPC64	21		/* PowerPC 64-bit */
+#define EM_ARM		40		/* Advanced RISC Machines ARM */
+#define EM_ALPHA	41		/* DEC ALPHA */
+#define EM_SPARCV9	43		/* SPARC version 9 */
+#define EM_ALPHA_EXP	0x9026		/* DEC ALPHA */
+#define EM_IA_64	50		/* Intel Merced */
+#define EM_X86_64	62		/* AMD x86-64 architecture */
+#define EM_VAX		75		/* DEC VAX */
+
+/* Version */
+#define EV_NONE		0		/* Invalid */
+#define EV_CURRENT	1		/* Current */
+#define EV_NUM		2		/* number of versions */
+
+/* Section Header */
+typedef struct {
+	Elf32_Word	sh_name;	/* name - index into section header
+					   string table section */
+	Elf32_Word	sh_type;	/* type */
+	Elf32_Word	sh_flags;	/* flags */
+	Elf32_Addr	sh_addr;	/* address */
+	Elf32_Off	sh_offset;	/* file offset */
+	Elf32_Word	sh_size;	/* section size */
+	Elf32_Word	sh_link;	/* section header table index link */
+	Elf32_Word	sh_info;	/* extra information */
+	Elf32_Word	sh_addralign;	/* address alignment */
+	Elf32_Word	sh_entsize;	/* section entry size */
+} Elf32_Shdr;
+
+typedef struct {
+	Elf64_Half	sh_name;	/* section name */
+	Elf64_Half	sh_type;	/* section type */
+	Elf64_Xword	sh_flags;	/* section flags */
+	Elf64_Addr	sh_addr;	/* virtual address */
+	Elf64_Off	sh_offset;	/* file offset */
+	Elf64_Xword	sh_size;	/* section size */
+	Elf64_Half	sh_link;	/* link to another */
+	Elf64_Half	sh_info;	/* misc info */
+	Elf64_Xword	sh_addralign;	/* memory alignment */
+	Elf64_Xword	sh_entsize;	/* table entry size */
+} Elf64_Shdr;
+
+/* Special Section Indexes */
+#define SHN_UNDEF	0		/* undefined */
+#define SHN_LORESERVE	0xff00		/* lower bounds of reserved indexes */
+#define SHN_LOPROC	0xff00		/* reserved range for processor */
+#define SHN_HIPROC	0xff1f		/*   specific section indexes */
+#define SHN_ABS		0xfff1		/* absolute value */
+#define SHN_COMMON	0xfff2		/* common symbol */
+#define SHN_HIRESERVE	0xffff		/* upper bounds of reserved indexes */
+
+/* sh_type */
+#define SHT_NULL	0		/* inactive */
+#define SHT_PROGBITS	1		/* program defined information */
+#define SHT_SYMTAB	2		/* symbol table section */
+#define SHT_STRTAB	3		/* string table section */
+#define SHT_RELA	4		/* relocation section with addends*/
+#define SHT_HASH	5		/* symbol hash table section */
+#define SHT_DYNAMIC	6		/* dynamic section */
+#define SHT_NOTE	7		/* note section */
+#define SHT_NOBITS	8		/* no space section */
+#define SHT_REL		9		/* relation section without addends */
+#define SHT_SHLIB	10		/* reserved - purpose unknown */
+#define SHT_DYNSYM	11		/* dynamic symbol table section */
+#define SHT_NUM		12		/* number of section types */
+#define SHT_LOPROC	0x70000000	/* reserved range for processor */
+#define SHT_HIPROC	0x7fffffff	/*  specific section header types */
+#define SHT_LOUSER	0x80000000	/* reserved range for application */
+#define SHT_HIUSER	0xffffffff	/*  specific indexes */
+
+/* Section names */
+#define ELF_BSS         ".bss"		/* uninitialized data */
+#define ELF_DATA        ".data"		/* initialized data */
+#define ELF_DEBUG       ".debug"	/* debug */
+#define ELF_DYNAMIC     ".dynamic"	/* dynamic linking information */
+#define ELF_DYNSTR      ".dynstr"	/* dynamic string table */
+#define ELF_DYNSYM      ".dynsym"	/* dynamic symbol table */
+#define ELF_FINI        ".fini"		/* termination code */
+#define ELF_GOT         ".got"		/* global offset table */
+#define ELF_HASH        ".hash"		/* symbol hash table */
+#define ELF_INIT        ".init"		/* initialization code */
+#define ELF_REL_DATA    ".rel.data"	/* relocation data */
+#define ELF_REL_FINI    ".rel.fini"	/* relocation termination code */
+#define ELF_REL_INIT    ".rel.init"	/* relocation initialization code */
+#define ELF_REL_DYN     ".rel.dyn"	/* relocaltion dynamic link info */
+#define ELF_REL_RODATA  ".rel.rodata"	/* relocation read-only data */
+#define ELF_REL_TEXT    ".rel.text"	/* relocation code */
+#define ELF_RODATA      ".rodata"	/* read-only data */
+#define ELF_SHSTRTAB    ".shstrtab"	/* section header string table */
+#define ELF_STRTAB      ".strtab"	/* string table */
+#define ELF_SYMTAB      ".symtab"	/* symbol table */
+#define ELF_TEXT        ".text"		/* code */
+
+
+/* Section Attribute Flags - sh_flags */
+#define SHF_WRITE	0x1		/* Writable */
+#define SHF_ALLOC	0x2		/* occupies memory */
+#define SHF_EXECINSTR	0x4		/* executable */
+#define SHF_MASKPROC	0xf0000000	/* reserved bits for processor */
+					/*  specific section attributes */
+
+/* Symbol Table Entry */
+typedef struct elf32_sym {
+	Elf32_Word	st_name;	/* name - index into string table */
+	Elf32_Addr	st_value;	/* symbol value */
+	Elf32_Word	st_size;	/* symbol size */
+	unsigned char	st_info;	/* type and binding */
+	unsigned char	st_other;	/* 0 - no defined meaning */
+	Elf32_Half	st_shndx;	/* section header index */
+} Elf32_Sym;
+
+typedef struct {
+	Elf64_Half	st_name;	/* Symbol name index in str table */
+	Elf_Byte	st_info;	/* type / binding attrs */
+	Elf_Byte	st_other;	/* unused */
+	Elf64_Quarter	st_shndx;	/* section index of symbol */
+	Elf64_Xword	st_value;	/* value of symbol */
+	Elf64_Xword	st_size;	/* size of symbol */
+} Elf64_Sym;
+
+/* Symbol table index */
+#define STN_UNDEF	0		/* undefined */
+
+/* Extract symbol info - st_info */
+#define ELF32_ST_BIND(x)	((x) >> 4)
+#define ELF32_ST_TYPE(x)	(((unsigned int) x) & 0xf)
+#define ELF32_ST_INFO(b,t)	(((b) << 4) + ((t) & 0xf))
+
+#define ELF64_ST_BIND(x)	((x) >> 4)
+#define ELF64_ST_TYPE(x)	(((unsigned int) x) & 0xf)
+#define ELF64_ST_INFO(b,t)	(((b) << 4) + ((t) & 0xf))
+
+/* Symbol Binding - ELF32_ST_BIND - st_info */
+#define STB_LOCAL	0		/* Local symbol */
+#define STB_GLOBAL	1		/* Global symbol */
+#define STB_WEAK	2		/* like global - lower precedence */
+#define STB_NUM		3		/* number of symbol bindings */
+#define STB_LOPROC	13		/* reserved range for processor */
+#define STB_HIPROC	15		/*  specific symbol bindings */
+
+/* Symbol type - ELF32_ST_TYPE - st_info */
+#define STT_NOTYPE	0		/* not specified */
+#define STT_OBJECT	1		/* data object */
+#define STT_FUNC	2		/* function */
+#define STT_SECTION	3		/* section */
+#define STT_FILE	4		/* file */
+#define STT_NUM		5		/* number of symbol types */
+#define STT_LOPROC	13		/* reserved range for processor */
+#define STT_HIPROC	15		/*  specific symbol types */
+
+/* Relocation entry with implicit addend */
+typedef struct {
+	Elf32_Addr	r_offset;	/* offset of relocation */
+	Elf32_Word	r_info;		/* symbol table index and type */
+} Elf32_Rel;
+
+/* Relocation entry with explicit addend */
+typedef struct {
+	Elf32_Addr	r_offset;	/* offset of relocation */
+	Elf32_Word	r_info;		/* symbol table index and type */
+	Elf32_Sword	r_addend;
+} Elf32_Rela;
+
+/* Extract relocation info - r_info */
+#define ELF32_R_SYM(i)		((i) >> 8)
+#define ELF32_R_TYPE(i)		((unsigned char) (i))
+#define ELF32_R_INFO(s,t) 	(((s) << 8) + (unsigned char)(t))
+
+typedef struct {
+	Elf64_Xword	r_offset;	/* where to do it */
+	Elf64_Xword	r_info;		/* index & type of relocation */
+} Elf64_Rel;
+
+typedef struct {
+	Elf64_Xword	r_offset;	/* where to do it */
+	Elf64_Xword	r_info;		/* index & type of relocation */
+	Elf64_Sxword	r_addend;	/* adjustment value */
+} Elf64_Rela;
+
+#define	ELF64_R_SYM(info)	((info) >> 32)
+#define	ELF64_R_TYPE(info)	((info) & 0xFFFFFFFF)
+#define ELF64_R_INFO(s,t) 	(((s) << 32) + (u_int32_t)(t))
+
+/* Program Header */
+typedef struct {
+	Elf32_Word	p_type;		/* segment type */
+	Elf32_Off	p_offset;	/* segment offset */
+	Elf32_Addr	p_vaddr;	/* virtual address of segment */
+	Elf32_Addr	p_paddr;	/* physical address - ignored? */
+	Elf32_Word	p_filesz;	/* number of bytes in file for seg. */
+	Elf32_Word	p_memsz;	/* number of bytes in mem. for seg. */
+	Elf32_Word	p_flags;	/* flags */
+	Elf32_Word	p_align;	/* memory alignment */
+} Elf32_Phdr;
+
+typedef struct {
+	Elf64_Half	p_type;		/* entry type */
+	Elf64_Half	p_flags;	/* flags */
+	Elf64_Off	p_offset;	/* offset */
+	Elf64_Addr	p_vaddr;	/* virtual address */
+	Elf64_Addr	p_paddr;	/* physical address */
+	Elf64_Xword	p_filesz;	/* file size */
+	Elf64_Xword	p_memsz;	/* memory size */
+	Elf64_Xword	p_align;	/* memory & file alignment */
+} Elf64_Phdr;
+
+/* Segment types - p_type */
+#define PT_NULL		0		/* unused */
+#define PT_LOAD		1		/* loadable segment */
+#define PT_DYNAMIC	2		/* dynamic linking section */
+#define PT_INTERP	3		/* the RTLD */
+#define PT_NOTE		4		/* auxiliary information */
+#define PT_SHLIB	5		/* reserved - purpose undefined */
+#define PT_PHDR		6		/* program header */
+#define PT_NUM		7		/* Number of segment types */
+#define PT_LOPROC	0x70000000	/* reserved range for processor */
+#define PT_HIPROC	0x7fffffff	/*  specific segment types */
+
+/* Segment flags - p_flags */
+#define PF_X		0x1		/* Executable */
+#define PF_W		0x2		/* Writable */
+#define PF_R		0x4		/* Readable */
+#define PF_MASKPROC	0xf0000000	/* reserved bits for processor */
+					/*  specific segment flags */
+
+/* Dynamic structure */
+typedef struct {
+	Elf32_Sword	d_tag;		/* controls meaning of d_val */
+	union {
+		Elf32_Word	d_val;	/* Multiple meanings - see d_tag */
+		Elf32_Addr	d_ptr;	/* program virtual address */
+	} d_un;
+} Elf32_Dyn;
+
+typedef struct {
+	Elf64_Xword	d_tag;		/* controls meaning of d_val */
+	union {
+		Elf64_Addr	d_ptr;
+		Elf64_Xword	d_val;
+	} d_un;
+} Elf64_Dyn;
+
+/* Dynamic Array Tags - d_tag */
+#define DT_NULL		0		/* marks end of _DYNAMIC array */
+#define DT_NEEDED	1		/* string table offset of needed lib */
+#define DT_PLTRELSZ	2		/* size of relocation entries in PLT */
+#define DT_PLTGOT	3		/* address PLT/GOT */
+#define DT_HASH		4		/* address of symbol hash table */
+#define DT_STRTAB	5		/* address of string table */
+#define DT_SYMTAB	6		/* address of symbol table */
+#define DT_RELA		7		/* address of relocation table */
+#define DT_RELASZ	8		/* size of relocation table */
+#define DT_RELAENT	9		/* size of relocation entry */
+#define DT_STRSZ	10		/* size of string table */
+#define DT_SYMENT	11		/* size of symbol table entry */
+#define DT_INIT		12		/* address of initialization func. */
+#define DT_FINI		13		/* address of termination function */
+#define DT_SONAME	14		/* string table offset of shared obj */
+#define DT_RPATH	15		/* string table offset of library
+					   search path */
+#define DT_SYMBOLIC	16		/* start sym search in shared obj. */
+#define DT_REL		17		/* address of rel. tbl. w addends */
+#define DT_RELSZ	18		/* size of DT_REL relocation table */
+#define DT_RELENT	19		/* size of DT_REL relocation entry */
+#define DT_PLTREL	20		/* PLT referenced relocation entry */
+#define DT_DEBUG	21		/* bugger */
+#define DT_TEXTREL	22		/* Allow rel. mod. to unwritable seg */
+#define DT_JMPREL	23		/* add. of PLT's relocation entries */
+#define DT_BIND_NOW	24		/* Bind now regardless of env setting */
+#define DT_NUM		25		/* Number used. */
+#define DT_LOPROC	0x70000000	/* reserved range for processor */
+#define DT_HIPROC	0x7fffffff	/*  specific dynamic array tags */
+
+/* Standard ELF hashing function */
+unsigned int elf_hash(const unsigned char *name);
+
+/*
+ * Note Definitions
+ */
+typedef struct {
+	Elf32_Word namesz;
+	Elf32_Word descsz;
+	Elf32_Word type;
+} Elf32_Note;
+
+typedef struct {
+	Elf64_Half namesz;
+	Elf64_Half descsz;
+	Elf64_Half type;
+} Elf64_Note;
+
+
+#if defined(ELFSIZE)
+#define CONCAT(x,y)	__CONCAT(x,y)
+#define ELFNAME(x)	CONCAT(elf,CONCAT(ELFSIZE,CONCAT(_,x)))
+#define ELFNAME2(x,y)	CONCAT(x,CONCAT(_elf,CONCAT(ELFSIZE,CONCAT(_,y))))
+#define ELFNAMEEND(x)	CONCAT(x,CONCAT(_elf,ELFSIZE))
+#define ELFDEFNNAME(x)	CONCAT(ELF,CONCAT(ELFSIZE,CONCAT(_,x)))
+#endif
+
+#if defined(ELFSIZE) && (ELFSIZE == 32)
+#define Elf_Ehdr	Elf32_Ehdr
+#define Elf_Phdr	Elf32_Phdr
+#define Elf_Shdr	Elf32_Shdr
+#define Elf_Sym		Elf32_Sym
+#define Elf_Rel		Elf32_Rel
+#define Elf_RelA	Elf32_Rela
+#define Elf_Dyn		Elf32_Dyn
+#define Elf_Word	Elf32_Word
+#define Elf_Sword	Elf32_Sword
+#define Elf_Addr	Elf32_Addr
+#define Elf_Off		Elf32_Off
+#define Elf_Nhdr	Elf32_Nhdr
+#define Elf_Note	Elf32_Note
+
+#define ELF_R_SYM	ELF32_R_SYM
+#define ELF_R_TYPE	ELF32_R_TYPE
+#define ELF_R_INFO	ELF32_R_INFO
+#define ELFCLASS	ELFCLASS32
+
+#define ELF_ST_BIND	ELF32_ST_BIND
+#define ELF_ST_TYPE	ELF32_ST_TYPE
+#define ELF_ST_INFO	ELF32_ST_INFO
+
+#define AuxInfo		Aux32Info
+#elif defined(ELFSIZE) && (ELFSIZE == 64)
+#define Elf_Ehdr	Elf64_Ehdr
+#define Elf_Phdr	Elf64_Phdr
+#define Elf_Shdr	Elf64_Shdr
+#define Elf_Sym		Elf64_Sym
+#define Elf_Rel		Elf64_Rel
+#define Elf_RelA	Elf64_Rela
+#define Elf_Dyn		Elf64_Dyn
+#define Elf_Word	Elf64_Word
+#define Elf_Sword	Elf64_Sword
+#define Elf_Addr	Elf64_Addr
+#define Elf_Off		Elf64_Off
+#define Elf_Nhdr	Elf64_Nhdr
+#define Elf_Note	Elf64_Note
+
+#define ELF_R_SYM	ELF64_R_SYM
+#define ELF_R_TYPE	ELF64_R_TYPE
+#define ELF_R_INFO	ELF64_R_INFO
+#define ELFCLASS	ELFCLASS64
+
+#define ELF_ST_BIND	ELF64_ST_BIND
+#define ELF_ST_TYPE	ELF64_ST_TYPE
+#define ELF_ST_INFO	ELF64_ST_INFO
+
+#define AuxInfo		Aux64Info
+#endif
+
+#endif /* __XEN_PUBLIC_ELFSTRUCTS_H__ */
diff -rpuN linux-2.6.18.8/include/xen/interface/event_channel.h linux-2.6.18-xen-3.3.0/include/xen/interface/event_channel.h
--- linux-2.6.18.8/include/xen/interface/event_channel.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/event_channel.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,264 @@
+/******************************************************************************
+ * event_channel.h
+ * 
+ * Event channels between domains.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2003-2004, K A Fraser.
+ */
+
+#ifndef __XEN_PUBLIC_EVENT_CHANNEL_H__
+#define __XEN_PUBLIC_EVENT_CHANNEL_H__
+
+/*
+ * Prototype for this hypercall is:
+ *  int event_channel_op(int cmd, void *args)
+ * @cmd  == EVTCHNOP_??? (event-channel operation).
+ * @args == Operation-specific extra arguments (NULL if none).
+ */
+
+typedef uint32_t evtchn_port_t;
+DEFINE_XEN_GUEST_HANDLE(evtchn_port_t);
+
+/*
+ * EVTCHNOP_alloc_unbound: Allocate a port in domain <dom> and mark as
+ * accepting interdomain bindings from domain <remote_dom>. A fresh port
+ * is allocated in <dom> and returned as <port>.
+ * NOTES:
+ *  1. If the caller is unprivileged then <dom> must be DOMID_SELF.
+ *  2. <rdom> may be DOMID_SELF, allowing loopback connections.
+ */
+#define EVTCHNOP_alloc_unbound    6
+struct evtchn_alloc_unbound {
+    /* IN parameters */
+    domid_t dom, remote_dom;
+    /* OUT parameters */
+    evtchn_port_t port;
+};
+typedef struct evtchn_alloc_unbound evtchn_alloc_unbound_t;
+
+/*
+ * EVTCHNOP_bind_interdomain: Construct an interdomain event channel between
+ * the calling domain and <remote_dom>. <remote_dom,remote_port> must identify
+ * a port that is unbound and marked as accepting bindings from the calling
+ * domain. A fresh port is allocated in the calling domain and returned as
+ * <local_port>.
+ * NOTES:
+ *  2. <remote_dom> may be DOMID_SELF, allowing loopback connections.
+ */
+#define EVTCHNOP_bind_interdomain 0
+struct evtchn_bind_interdomain {
+    /* IN parameters. */
+    domid_t remote_dom;
+    evtchn_port_t remote_port;
+    /* OUT parameters. */
+    evtchn_port_t local_port;
+};
+typedef struct evtchn_bind_interdomain evtchn_bind_interdomain_t;
+
+/*
+ * EVTCHNOP_bind_virq: Bind a local event channel to VIRQ <irq> on specified
+ * vcpu.
+ * NOTES:
+ *  1. Virtual IRQs are classified as per-vcpu or global. See the VIRQ list
+ *     in xen.h for the classification of each VIRQ.
+ *  2. Global VIRQs must be allocated on VCPU0 but can subsequently be
+ *     re-bound via EVTCHNOP_bind_vcpu.
+ *  3. Per-vcpu VIRQs may be bound to at most one event channel per vcpu.
+ *     The allocated event channel is bound to the specified vcpu and the
+ *     binding cannot be changed.
+ */
+#define EVTCHNOP_bind_virq        1
+struct evtchn_bind_virq {
+    /* IN parameters. */
+    uint32_t virq;
+    uint32_t vcpu;
+    /* OUT parameters. */
+    evtchn_port_t port;
+};
+typedef struct evtchn_bind_virq evtchn_bind_virq_t;
+
+/*
+ * EVTCHNOP_bind_pirq: Bind a local event channel to PIRQ <irq>.
+ * NOTES:
+ *  1. A physical IRQ may be bound to at most one event channel per domain.
+ *  2. Only a sufficiently-privileged domain may bind to a physical IRQ.
+ */
+#define EVTCHNOP_bind_pirq        2
+struct evtchn_bind_pirq {
+    /* IN parameters. */
+    uint32_t pirq;
+#define BIND_PIRQ__WILL_SHARE 1
+    uint32_t flags; /* BIND_PIRQ__* */
+    /* OUT parameters. */
+    evtchn_port_t port;
+};
+typedef struct evtchn_bind_pirq evtchn_bind_pirq_t;
+
+/*
+ * EVTCHNOP_bind_ipi: Bind a local event channel to receive events.
+ * NOTES:
+ *  1. The allocated event channel is bound to the specified vcpu. The binding
+ *     may not be changed.
+ */
+#define EVTCHNOP_bind_ipi         7
+struct evtchn_bind_ipi {
+    uint32_t vcpu;
+    /* OUT parameters. */
+    evtchn_port_t port;
+};
+typedef struct evtchn_bind_ipi evtchn_bind_ipi_t;
+
+/*
+ * EVTCHNOP_close: Close a local event channel <port>. If the channel is
+ * interdomain then the remote end is placed in the unbound state
+ * (EVTCHNSTAT_unbound), awaiting a new connection.
+ */
+#define EVTCHNOP_close            3
+struct evtchn_close {
+    /* IN parameters. */
+    evtchn_port_t port;
+};
+typedef struct evtchn_close evtchn_close_t;
+
+/*
+ * EVTCHNOP_send: Send an event to the remote end of the channel whose local
+ * endpoint is <port>.
+ */
+#define EVTCHNOP_send             4
+struct evtchn_send {
+    /* IN parameters. */
+    evtchn_port_t port;
+};
+typedef struct evtchn_send evtchn_send_t;
+
+/*
+ * EVTCHNOP_status: Get the current status of the communication channel which
+ * has an endpoint at <dom, port>.
+ * NOTES:
+ *  1. <dom> may be specified as DOMID_SELF.
+ *  2. Only a sufficiently-privileged domain may obtain the status of an event
+ *     channel for which <dom> is not DOMID_SELF.
+ */
+#define EVTCHNOP_status           5
+struct evtchn_status {
+    /* IN parameters */
+    domid_t  dom;
+    evtchn_port_t port;
+    /* OUT parameters */
+#define EVTCHNSTAT_closed       0  /* Channel is not in use.                 */
+#define EVTCHNSTAT_unbound      1  /* Channel is waiting interdom connection.*/
+#define EVTCHNSTAT_interdomain  2  /* Channel is connected to remote domain. */
+#define EVTCHNSTAT_pirq         3  /* Channel is bound to a phys IRQ line.   */
+#define EVTCHNSTAT_virq         4  /* Channel is bound to a virtual IRQ line */
+#define EVTCHNSTAT_ipi          5  /* Channel is bound to a virtual IPI line */
+    uint32_t status;
+    uint32_t vcpu;                 /* VCPU to which this channel is bound.   */
+    union {
+        struct {
+            domid_t dom;
+        } unbound; /* EVTCHNSTAT_unbound */
+        struct {
+            domid_t dom;
+            evtchn_port_t port;
+        } interdomain; /* EVTCHNSTAT_interdomain */
+        uint32_t pirq;      /* EVTCHNSTAT_pirq        */
+        uint32_t virq;      /* EVTCHNSTAT_virq        */
+    } u;
+};
+typedef struct evtchn_status evtchn_status_t;
+
+/*
+ * EVTCHNOP_bind_vcpu: Specify which vcpu a channel should notify when an
+ * event is pending.
+ * NOTES:
+ *  1. IPI-bound channels always notify the vcpu specified at bind time.
+ *     This binding cannot be changed.
+ *  2. Per-VCPU VIRQ channels always notify the vcpu specified at bind time.
+ *     This binding cannot be changed.
+ *  3. All other channels notify vcpu0 by default. This default is set when
+ *     the channel is allocated (a port that is freed and subsequently reused
+ *     has its binding reset to vcpu0).
+ */
+#define EVTCHNOP_bind_vcpu        8
+struct evtchn_bind_vcpu {
+    /* IN parameters. */
+    evtchn_port_t port;
+    uint32_t vcpu;
+};
+typedef struct evtchn_bind_vcpu evtchn_bind_vcpu_t;
+
+/*
+ * EVTCHNOP_unmask: Unmask the specified local event-channel port and deliver
+ * a notification to the appropriate VCPU if an event is pending.
+ */
+#define EVTCHNOP_unmask           9
+struct evtchn_unmask {
+    /* IN parameters. */
+    evtchn_port_t port;
+};
+typedef struct evtchn_unmask evtchn_unmask_t;
+
+/*
+ * EVTCHNOP_reset: Close all event channels associated with specified domain.
+ * NOTES:
+ *  1. <dom> may be specified as DOMID_SELF.
+ *  2. Only a sufficiently-privileged domain may specify other than DOMID_SELF.
+ */
+#define EVTCHNOP_reset           10
+struct evtchn_reset {
+    /* IN parameters. */
+    domid_t dom;
+};
+typedef struct evtchn_reset evtchn_reset_t;
+
+/*
+ * Argument to event_channel_op_compat() hypercall. Superceded by new
+ * event_channel_op() hypercall since 0x00030202.
+ */
+struct evtchn_op {
+    uint32_t cmd; /* EVTCHNOP_* */
+    union {
+        struct evtchn_alloc_unbound    alloc_unbound;
+        struct evtchn_bind_interdomain bind_interdomain;
+        struct evtchn_bind_virq        bind_virq;
+        struct evtchn_bind_pirq        bind_pirq;
+        struct evtchn_bind_ipi         bind_ipi;
+        struct evtchn_close            close;
+        struct evtchn_send             send;
+        struct evtchn_status           status;
+        struct evtchn_bind_vcpu        bind_vcpu;
+        struct evtchn_unmask           unmask;
+    } u;
+};
+typedef struct evtchn_op evtchn_op_t;
+DEFINE_XEN_GUEST_HANDLE(evtchn_op_t);
+
+#endif /* __XEN_PUBLIC_EVENT_CHANNEL_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/features.h linux-2.6.18-xen-3.3.0/include/xen/interface/features.h
--- linux-2.6.18.8/include/xen/interface/features.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/features.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,74 @@
+/******************************************************************************
+ * features.h
+ * 
+ * Feature flags, reported by XENVER_get_features.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2006, Keir Fraser <keir@xensource.com>
+ */
+
+#ifndef __XEN_PUBLIC_FEATURES_H__
+#define __XEN_PUBLIC_FEATURES_H__
+
+/*
+ * If set, the guest does not need to write-protect its pagetables, and can
+ * update them via direct writes.
+ */
+#define XENFEAT_writable_page_tables       0
+
+/*
+ * If set, the guest does not need to write-protect its segment descriptor
+ * tables, and can update them via direct writes.
+ */
+#define XENFEAT_writable_descriptor_tables 1
+
+/*
+ * If set, translation between the guest's 'pseudo-physical' address space
+ * and the host's machine address space are handled by the hypervisor. In this
+ * mode the guest does not need to perform phys-to/from-machine translations
+ * when performing page table operations.
+ */
+#define XENFEAT_auto_translated_physmap    2
+
+/* If set, the guest is running in supervisor mode (e.g., x86 ring 0). */
+#define XENFEAT_supervisor_mode_kernel     3
+
+/*
+ * If set, the guest does not need to allocate x86 PAE page directories
+ * below 4GB. This flag is usually implied by auto_translated_physmap.
+ */
+#define XENFEAT_pae_pgdir_above_4gb        4
+
+/* x86: Does this Xen host support the MMU_PT_UPDATE_PRESERVE_AD hypercall? */
+#define XENFEAT_mmu_pt_update_preserve_ad  5
+
+#define XENFEAT_NR_SUBMAPS 1
+
+#endif /* __XEN_PUBLIC_FEATURES_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/foreign/Makefile linux-2.6.18-xen-3.3.0/include/xen/interface/foreign/Makefile
--- linux-2.6.18.8/include/xen/interface/foreign/Makefile	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/foreign/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,37 @@
+XEN_ROOT=../../../..
+include $(XEN_ROOT)/Config.mk
+
+architectures := x86_32 x86_64 ia64
+headers := $(patsubst %, %.h, $(architectures))
+scripts := $(wildcard *.py)
+
+.PHONY: all clean check-headers
+all: $(headers) check-headers
+
+clean:
+	rm -f $(headers)
+	rm -f checker checker.c $(XEN_TARGET_ARCH).size
+	rm -f *.pyc *.o *~
+
+ifeq ($(CROSS_COMPILE)$(XEN_TARGET_ARCH),$(XEN_COMPILE_ARCH))
+check-headers: checker
+	./checker > $(XEN_TARGET_ARCH).size
+	diff -u reference.size $(XEN_TARGET_ARCH).size
+checker: checker.c $(headers)
+	$(HOSTCC) $(HOSTCFLAGS) -o $@ $<
+else
+check-headers:
+	@echo "cross build: skipping check"
+endif
+
+x86_32.h: ../arch-x86/xen-x86_32.h ../arch-x86/xen.h ../xen.h $(scripts)
+	python mkheader.py $* $@ $(filter %.h,$^)
+
+x86_64.h: ../arch-x86/xen-x86_64.h ../arch-x86/xen.h ../xen.h $(scripts)
+	python mkheader.py $* $@ $(filter %.h,$^)
+
+ia64.h: ../arch-ia64.h ../xen.h $(scripts)
+	python mkheader.py $* $@ $(filter %.h,$^)
+
+checker.c: $(scripts)
+	python mkchecker.py $(XEN_TARGET_ARCH) $@ $(architectures)
diff -rpuN linux-2.6.18.8/include/xen/interface/foreign/mkchecker.py linux-2.6.18-xen-3.3.0/include/xen/interface/foreign/mkchecker.py
--- linux-2.6.18.8/include/xen/interface/foreign/mkchecker.py	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/foreign/mkchecker.py	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,58 @@
+#!/usr/bin/python
+
+import sys;
+from structs import structs;
+
+# command line arguments
+arch    = sys.argv[1];
+outfile = sys.argv[2];
+archs   = sys.argv[3:];
+
+f = open(outfile, "w");
+f.write('''
+/*
+ * sanity checks for generated foreign headers:
+ *  - verify struct sizes
+ *
+ * generated by %s -- DO NOT EDIT
+ */
+#include <stdio.h>
+#include <stdlib.h>
+#include <stddef.h>
+#include <inttypes.h>
+#include "../xen.h"
+''');
+
+for a in archs:
+    f.write('#include "%s.h"\n' % a);
+
+f.write('int main(int argc, char *argv[])\n{\n');
+
+f.write('\tprintf("\\n");');
+f.write('printf("%-25s |", "structs");\n');
+for a in archs:
+    f.write('\tprintf("%%8s", "%s");\n' % a);
+f.write('\tprintf("\\n");');
+
+f.write('\tprintf("\\n");');
+for struct in structs:
+    f.write('\tprintf("%%-25s |", "%s");\n' % struct);
+    for a in archs:
+        if a == arch:
+            s = struct; # native
+        else:
+            s = struct + "_" + a;
+        f.write('#ifdef %s_has_no_%s\n' % (a, struct));
+        f.write('\tprintf("%8s", "-");\n');
+        f.write("#else\n");
+        f.write('\tprintf("%%8zd", sizeof(struct %s));\n' % s);
+        f.write("#endif\n");
+
+    f.write('\tprintf("\\n");\n\n');
+
+f.write('\tprintf("\\n");\n');
+f.write('\texit(0);\n');
+f.write('}\n');
+
+f.close();
+
diff -rpuN linux-2.6.18.8/include/xen/interface/foreign/mkheader.py linux-2.6.18-xen-3.3.0/include/xen/interface/foreign/mkheader.py
--- linux-2.6.18.8/include/xen/interface/foreign/mkheader.py	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/foreign/mkheader.py	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,167 @@
+#!/usr/bin/python
+
+import sys, re;
+from structs import unions, structs, defines;
+
+# command line arguments
+arch    = sys.argv[1];
+outfile = sys.argv[2];
+infiles = sys.argv[3:];
+
+
+###########################################################################
+# configuration #2: architecture information
+
+inttypes = {};
+header = {};
+footer = {};
+
+# x86_32
+inttypes["x86_32"] = {
+    "unsigned long" : "uint32_t",
+    "long"          : "uint32_t",
+    "xen_pfn_t"     : "uint32_t",
+};
+header["x86_32"] = """
+#define __i386___X86_32 1
+#pragma pack(4)
+""";
+footer["x86_32"] = """
+#pragma pack()
+""";
+
+# x86_64
+inttypes["x86_64"] = {
+    "unsigned long" : "__align8__ uint64_t",
+    "long"          : "__align8__ uint64_t",
+    "xen_pfn_t"     : "__align8__ uint64_t",
+};
+header["x86_64"] = """
+#ifdef __GNUC__
+# define __DECL_REG(name) union { uint64_t r ## name, e ## name; }
+# define __align8__ __attribute__((aligned (8)))
+#else
+# define __DECL_REG(name) uint64_t r ## name
+# define __align8__ FIXME
+#endif
+#define __x86_64___X86_64 1
+""";
+
+# ia64
+inttypes["ia64"] = {
+    "unsigned long" : "__align8__ uint64_t",
+    "long"          : "__align8__ uint64_t",
+    "xen_pfn_t"     : "__align8__ uint64_t",
+    "long double"   : "__align16__ ldouble_t",
+};
+header["ia64"] = """
+#define __align8__ __attribute__((aligned (8)))
+#define __align16__ __attribute__((aligned (16)))
+typedef unsigned char ldouble_t[16];
+""";
+
+
+###########################################################################
+# main
+
+input  = "";
+output = "";
+fileid = re.sub("[-.]", "_", "__FOREIGN_%s__" % outfile.upper());
+
+# read input header files
+for name in infiles:
+    f = open(name, "r");
+    input += f.read();
+    f.close();
+
+# add header
+output += """
+/*
+ * public xen defines and struct for %s
+ * generated by %s -- DO NOT EDIT
+ */
+
+#ifndef %s
+#define %s 1
+
+""" % (arch, sys.argv[0], fileid, fileid)
+
+if arch in header:
+    output += header[arch];
+    output += "\n";
+
+# add defines to output
+for line in re.findall("#define[^\n]+", input):
+    for define in defines:
+        regex = "#define\s+%s\\b" % define;
+        match = re.search(regex, line);
+        if None == match:
+            continue;
+        if define.upper()[0] == define[0]:
+            replace = define + "_" + arch.upper();
+        else:
+            replace = define + "_" + arch;
+        regex = "\\b%s\\b" % define;
+        output += re.sub(regex, replace, line) + "\n";
+output += "\n";
+
+# delete defines, comments, empty lines
+input = re.sub("#define[^\n]+\n", "", input);
+input = re.compile("/\*(.*?)\*/", re.S).sub("", input)
+input = re.compile("\n\s*\n", re.S).sub("\n", input);
+
+# add unions to output
+for union in unions:
+    regex = "union\s+%s\s*\{(.*?)\n\};" % union;
+    match = re.search(regex, input, re.S)
+    if None == match:
+        output += "#define %s_has_no_%s 1\n" % (arch, union);
+    else:
+        output += "union %s_%s {%s\n};\n" % (union, arch, match.group(1));
+    output += "\n";
+
+# add structs to output
+for struct in structs:
+    regex = "struct\s+%s\s*\{(.*?)\n\};" % struct;
+    match = re.search(regex, input, re.S)
+    if None == match:
+        output += "#define %s_has_no_%s 1\n" % (arch, struct);
+    else:
+        output += "struct %s_%s {%s\n};\n" % (struct, arch, match.group(1));
+        output += "typedef struct %s_%s %s_%s_t;\n" % (struct, arch, struct, arch);
+    output += "\n";
+
+# add footer
+if arch in footer:
+    output += footer[arch];
+    output += "\n";
+output += "#endif /* %s */\n" % fileid;
+
+# replace: defines
+for define in defines:
+    if define.upper()[0] == define[0]:
+        replace = define + "_" + arch.upper();
+    else:
+        replace = define + "_" + arch;
+    output = re.sub("\\b%s\\b" % define, replace, output);
+
+# replace: unions
+for union in unions:
+    output = re.sub("\\b(union\s+%s)\\b" % union, "\\1_%s" % arch, output);
+
+# replace: structs + struct typedefs
+for struct in structs:
+    output = re.sub("\\b(struct\s+%s)\\b" % struct, "\\1_%s" % arch, output);
+    output = re.sub("\\b(%s)_t\\b" % struct, "\\1_%s_t" % arch, output);
+
+# replace: integer types
+integers = inttypes[arch].keys();
+integers.sort(lambda a, b: cmp(len(b),len(a)));
+for type in integers:
+    output = re.sub("\\b%s\\b" % type, inttypes[arch][type], output);
+
+# print results
+f = open(outfile, "w");
+f.write(output);
+f.close;
+
diff -rpuN linux-2.6.18.8/include/xen/interface/foreign/reference.size linux-2.6.18-xen-3.3.0/include/xen/interface/foreign/reference.size
--- linux-2.6.18.8/include/xen/interface/foreign/reference.size	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/foreign/reference.size	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,18 @@
+
+structs                   |  x86_32  x86_64    ia64
+
+start_info                |    1104    1152    1152
+trap_info                 |       8      16       -
+pt_fpreg                  |       -       -      16
+cpu_user_regs             |      68     200       -
+xen_ia64_boot_param       |       -       -      96
+ia64_tr_entry             |       -       -      32
+vcpu_tr_regs              |       -       -     768
+vcpu_guest_context_regs   |       -       -   22176
+vcpu_guest_context        |    2800    5168   22208
+arch_vcpu_info            |      24      16       0
+vcpu_time_info            |      32      32      32
+vcpu_info                 |      64      64      48
+arch_shared_info          |     268     280     272
+shared_info               |    2584    3368    4384
+
diff -rpuN linux-2.6.18.8/include/xen/interface/foreign/structs.py linux-2.6.18-xen-3.3.0/include/xen/interface/foreign/structs.py
--- linux-2.6.18.8/include/xen/interface/foreign/structs.py	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/foreign/structs.py	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,58 @@
+# configuration: what needs translation
+
+unions  = [ "vcpu_cr_regs",
+            "vcpu_ar_regs" ];
+
+structs = [ "start_info",
+            "trap_info",
+            "pt_fpreg",
+            "cpu_user_regs",
+            "xen_ia64_boot_param",
+            "ia64_tr_entry",
+            "vcpu_tr_regs",
+            "vcpu_guest_context_regs",
+            "vcpu_guest_context",
+            "arch_vcpu_info",
+            "vcpu_time_info",
+            "vcpu_info",
+            "arch_shared_info",
+            "shared_info" ];
+
+defines = [ "__i386__",
+            "__x86_64__",
+
+            "FLAT_RING1_CS",
+            "FLAT_RING1_DS",
+            "FLAT_RING1_SS",
+
+            "FLAT_RING3_CS64",
+            "FLAT_RING3_DS64",
+            "FLAT_RING3_SS64",
+            "FLAT_KERNEL_CS64",
+            "FLAT_KERNEL_DS64",
+            "FLAT_KERNEL_SS64",
+
+            "FLAT_KERNEL_CS",
+            "FLAT_KERNEL_DS",
+            "FLAT_KERNEL_SS",
+
+            # x86_{32,64}
+            "_VGCF_i387_valid",
+            "VGCF_i387_valid",
+            "_VGCF_in_kernel",
+            "VGCF_in_kernel",
+            "_VGCF_failsafe_disables_events",
+            "VGCF_failsafe_disables_events",
+            "_VGCF_syscall_disables_events",
+            "VGCF_syscall_disables_events",
+            "_VGCF_online",
+            "VGCF_online",
+
+            # ia64
+            "VGCF_EXTRA_REGS",
+
+            # all archs
+            "xen_pfn_to_cr3",
+            "MAX_VIRT_CPUS",
+            "MAX_GUEST_CMDLINE" ];
+
diff -rpuN linux-2.6.18.8/include/xen/interface/grant_table.h linux-2.6.18-xen-3.3.0/include/xen/interface/grant_table.h
--- linux-2.6.18.8/include/xen/interface/grant_table.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/grant_table.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,431 @@
+/******************************************************************************
+ * grant_table.h
+ * 
+ * Interface for granting foreign access to page frames, and receiving
+ * page-ownership transfers.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2004, K A Fraser
+ */
+
+#ifndef __XEN_PUBLIC_GRANT_TABLE_H__
+#define __XEN_PUBLIC_GRANT_TABLE_H__
+
+
+/***********************************
+ * GRANT TABLE REPRESENTATION
+ */
+
+/* Some rough guidelines on accessing and updating grant-table entries
+ * in a concurrency-safe manner. For more information, Linux contains a
+ * reference implementation for guest OSes (arch/xen/kernel/grant_table.c).
+ * 
+ * NB. WMB is a no-op on current-generation x86 processors. However, a
+ *     compiler barrier will still be required.
+ * 
+ * Introducing a valid entry into the grant table:
+ *  1. Write ent->domid.
+ *  2. Write ent->frame:
+ *      GTF_permit_access:   Frame to which access is permitted.
+ *      GTF_accept_transfer: Pseudo-phys frame slot being filled by new
+ *                           frame, or zero if none.
+ *  3. Write memory barrier (WMB).
+ *  4. Write ent->flags, inc. valid type.
+ * 
+ * Invalidating an unused GTF_permit_access entry:
+ *  1. flags = ent->flags.
+ *  2. Observe that !(flags & (GTF_reading|GTF_writing)).
+ *  3. Check result of SMP-safe CMPXCHG(&ent->flags, flags, 0).
+ *  NB. No need for WMB as reuse of entry is control-dependent on success of
+ *      step 3, and all architectures guarantee ordering of ctrl-dep writes.
+ *
+ * Invalidating an in-use GTF_permit_access entry:
+ *  This cannot be done directly. Request assistance from the domain controller
+ *  which can set a timeout on the use of a grant entry and take necessary
+ *  action. (NB. This is not yet implemented!).
+ * 
+ * Invalidating an unused GTF_accept_transfer entry:
+ *  1. flags = ent->flags.
+ *  2. Observe that !(flags & GTF_transfer_committed). [*]
+ *  3. Check result of SMP-safe CMPXCHG(&ent->flags, flags, 0).
+ *  NB. No need for WMB as reuse of entry is control-dependent on success of
+ *      step 3, and all architectures guarantee ordering of ctrl-dep writes.
+ *  [*] If GTF_transfer_committed is set then the grant entry is 'committed'.
+ *      The guest must /not/ modify the grant entry until the address of the
+ *      transferred frame is written. It is safe for the guest to spin waiting
+ *      for this to occur (detect by observing GTF_transfer_completed in
+ *      ent->flags).
+ *
+ * Invalidating a committed GTF_accept_transfer entry:
+ *  1. Wait for (ent->flags & GTF_transfer_completed).
+ *
+ * Changing a GTF_permit_access from writable to read-only:
+ *  Use SMP-safe CMPXCHG to set GTF_readonly, while checking !GTF_writing.
+ * 
+ * Changing a GTF_permit_access from read-only to writable:
+ *  Use SMP-safe bit-setting instruction.
+ */
+
+/*
+ * A grant table comprises a packed array of grant entries in one or more
+ * page frames shared between Xen and a guest.
+ * [XEN]: This field is written by Xen and read by the sharing guest.
+ * [GST]: This field is written by the guest and read by Xen.
+ */
+struct grant_entry {
+    /* GTF_xxx: various type and flag information.  [XEN,GST] */
+    uint16_t flags;
+    /* The domain being granted foreign privileges. [GST] */
+    domid_t  domid;
+    /*
+     * GTF_permit_access: Frame that @domid is allowed to map and access. [GST]
+     * GTF_accept_transfer: Frame whose ownership transferred by @domid. [XEN]
+     */
+    uint32_t frame;
+};
+typedef struct grant_entry grant_entry_t;
+
+/*
+ * Type of grant entry.
+ *  GTF_invalid: This grant entry grants no privileges.
+ *  GTF_permit_access: Allow @domid to map/access @frame.
+ *  GTF_accept_transfer: Allow @domid to transfer ownership of one page frame
+ *                       to this guest. Xen writes the page number to @frame.
+ */
+#define GTF_invalid         (0U<<0)
+#define GTF_permit_access   (1U<<0)
+#define GTF_accept_transfer (2U<<0)
+#define GTF_type_mask       (3U<<0)
+
+/*
+ * Subflags for GTF_permit_access.
+ *  GTF_readonly: Restrict @domid to read-only mappings and accesses. [GST]
+ *  GTF_reading: Grant entry is currently mapped for reading by @domid. [XEN]
+ *  GTF_writing: Grant entry is currently mapped for writing by @domid. [XEN]
+ *  GTF_PAT, GTF_PWT, GTF_PCD: (x86) cache attribute flags for the grant [GST]
+ */
+#define _GTF_readonly       (2)
+#define GTF_readonly        (1U<<_GTF_readonly)
+#define _GTF_reading        (3)
+#define GTF_reading         (1U<<_GTF_reading)
+#define _GTF_writing        (4)
+#define GTF_writing         (1U<<_GTF_writing)
+#define _GTF_PWT            (5)
+#define GTF_PWT             (1U<<_GTF_PWT)
+#define _GTF_PCD            (6)
+#define GTF_PCD             (1U<<_GTF_PCD)
+#define _GTF_PAT            (7)
+#define GTF_PAT             (1U<<_GTF_PAT)
+
+/*
+ * Subflags for GTF_accept_transfer:
+ *  GTF_transfer_committed: Xen sets this flag to indicate that it is committed
+ *      to transferring ownership of a page frame. When a guest sees this flag
+ *      it must /not/ modify the grant entry until GTF_transfer_completed is
+ *      set by Xen.
+ *  GTF_transfer_completed: It is safe for the guest to spin-wait on this flag
+ *      after reading GTF_transfer_committed. Xen will always write the frame
+ *      address, followed by ORing this flag, in a timely manner.
+ */
+#define _GTF_transfer_committed (2)
+#define GTF_transfer_committed  (1U<<_GTF_transfer_committed)
+#define _GTF_transfer_completed (3)
+#define GTF_transfer_completed  (1U<<_GTF_transfer_completed)
+
+
+/***********************************
+ * GRANT TABLE QUERIES AND USES
+ */
+
+/*
+ * Reference to a grant entry in a specified domain's grant table.
+ */
+typedef uint32_t grant_ref_t;
+
+/*
+ * Handle to track a mapping created via a grant reference.
+ */
+typedef uint32_t grant_handle_t;
+
+/*
+ * GNTTABOP_map_grant_ref: Map the grant entry (<dom>,<ref>) for access
+ * by devices and/or host CPUs. If successful, <handle> is a tracking number
+ * that must be presented later to destroy the mapping(s). On error, <handle>
+ * is a negative status code.
+ * NOTES:
+ *  1. If GNTMAP_device_map is specified then <dev_bus_addr> is the address
+ *     via which I/O devices may access the granted frame.
+ *  2. If GNTMAP_host_map is specified then a mapping will be added at
+ *     either a host virtual address in the current address space, or at
+ *     a PTE at the specified machine address.  The type of mapping to
+ *     perform is selected through the GNTMAP_contains_pte flag, and the 
+ *     address is specified in <host_addr>.
+ *  3. Mappings should only be destroyed via GNTTABOP_unmap_grant_ref. If a
+ *     host mapping is destroyed by other means then it is *NOT* guaranteed
+ *     to be accounted to the correct grant reference!
+ */
+#define GNTTABOP_map_grant_ref        0
+struct gnttab_map_grant_ref {
+    /* IN parameters. */
+    uint64_t host_addr;
+    uint32_t flags;               /* GNTMAP_* */
+    grant_ref_t ref;
+    domid_t  dom;
+    /* OUT parameters. */
+    int16_t  status;              /* GNTST_* */
+    grant_handle_t handle;
+    uint64_t dev_bus_addr;
+};
+typedef struct gnttab_map_grant_ref gnttab_map_grant_ref_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_map_grant_ref_t);
+
+/*
+ * GNTTABOP_unmap_grant_ref: Destroy one or more grant-reference mappings
+ * tracked by <handle>. If <host_addr> or <dev_bus_addr> is zero, that
+ * field is ignored. If non-zero, they must refer to a device/host mapping
+ * that is tracked by <handle>
+ * NOTES:
+ *  1. The call may fail in an undefined manner if either mapping is not
+ *     tracked by <handle>.
+ *  3. After executing a batch of unmaps, it is guaranteed that no stale
+ *     mappings will remain in the device or host TLBs.
+ */
+#define GNTTABOP_unmap_grant_ref      1
+struct gnttab_unmap_grant_ref {
+    /* IN parameters. */
+    uint64_t host_addr;
+    uint64_t dev_bus_addr;
+    grant_handle_t handle;
+    /* OUT parameters. */
+    int16_t  status;              /* GNTST_* */
+};
+typedef struct gnttab_unmap_grant_ref gnttab_unmap_grant_ref_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_unmap_grant_ref_t);
+
+/*
+ * GNTTABOP_setup_table: Set up a grant table for <dom> comprising at least
+ * <nr_frames> pages. The frame addresses are written to the <frame_list>.
+ * Only <nr_frames> addresses are written, even if the table is larger.
+ * NOTES:
+ *  1. <dom> may be specified as DOMID_SELF.
+ *  2. Only a sufficiently-privileged domain may specify <dom> != DOMID_SELF.
+ *  3. Xen may not support more than a single grant-table page per domain.
+ */
+#define GNTTABOP_setup_table          2
+struct gnttab_setup_table {
+    /* IN parameters. */
+    domid_t  dom;
+    uint32_t nr_frames;
+    /* OUT parameters. */
+    int16_t  status;              /* GNTST_* */
+    XEN_GUEST_HANDLE(ulong) frame_list;
+};
+typedef struct gnttab_setup_table gnttab_setup_table_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_setup_table_t);
+
+/*
+ * GNTTABOP_dump_table: Dump the contents of the grant table to the
+ * xen console. Debugging use only.
+ */
+#define GNTTABOP_dump_table           3
+struct gnttab_dump_table {
+    /* IN parameters. */
+    domid_t dom;
+    /* OUT parameters. */
+    int16_t status;               /* GNTST_* */
+};
+typedef struct gnttab_dump_table gnttab_dump_table_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_dump_table_t);
+
+/*
+ * GNTTABOP_transfer_grant_ref: Transfer <frame> to a foreign domain. The
+ * foreign domain has previously registered its interest in the transfer via
+ * <domid, ref>.
+ * 
+ * Note that, even if the transfer fails, the specified page no longer belongs
+ * to the calling domain *unless* the error is GNTST_bad_page.
+ */
+#define GNTTABOP_transfer                4
+struct gnttab_transfer {
+    /* IN parameters. */
+    xen_pfn_t     mfn;
+    domid_t       domid;
+    grant_ref_t   ref;
+    /* OUT parameters. */
+    int16_t       status;
+};
+typedef struct gnttab_transfer gnttab_transfer_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_transfer_t);
+
+
+/*
+ * GNTTABOP_copy: Hypervisor based copy
+ * source and destinations can be eithers MFNs or, for foreign domains,
+ * grant references. the foreign domain has to grant read/write access
+ * in its grant table.
+ *
+ * The flags specify what type source and destinations are (either MFN
+ * or grant reference).
+ *
+ * Note that this can also be used to copy data between two domains
+ * via a third party if the source and destination domains had previously
+ * grant appropriate access to their pages to the third party.
+ *
+ * source_offset specifies an offset in the source frame, dest_offset
+ * the offset in the target frame and  len specifies the number of
+ * bytes to be copied.
+ */
+
+#define _GNTCOPY_source_gref      (0)
+#define GNTCOPY_source_gref       (1<<_GNTCOPY_source_gref)
+#define _GNTCOPY_dest_gref        (1)
+#define GNTCOPY_dest_gref         (1<<_GNTCOPY_dest_gref)
+
+#define GNTTABOP_copy                 5
+typedef struct gnttab_copy {
+    /* IN parameters. */
+    struct {
+        union {
+            grant_ref_t ref;
+            xen_pfn_t   gmfn;
+        } u;
+        domid_t  domid;
+        uint16_t offset;
+    } source, dest;
+    uint16_t      len;
+    uint16_t      flags;          /* GNTCOPY_* */
+    /* OUT parameters. */
+    int16_t       status;
+} gnttab_copy_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_copy_t);
+
+/*
+ * GNTTABOP_query_size: Query the current and maximum sizes of the shared
+ * grant table.
+ * NOTES:
+ *  1. <dom> may be specified as DOMID_SELF.
+ *  2. Only a sufficiently-privileged domain may specify <dom> != DOMID_SELF.
+ */
+#define GNTTABOP_query_size           6
+struct gnttab_query_size {
+    /* IN parameters. */
+    domid_t  dom;
+    /* OUT parameters. */
+    uint32_t nr_frames;
+    uint32_t max_nr_frames;
+    int16_t  status;              /* GNTST_* */
+};
+typedef struct gnttab_query_size gnttab_query_size_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_query_size_t);
+
+/*
+ * GNTTABOP_unmap_and_replace: Destroy one or more grant-reference mappings
+ * tracked by <handle> but atomically replace the page table entry with one
+ * pointing to the machine address under <new_addr>.  <new_addr> will be
+ * redirected to the null entry.
+ * NOTES:
+ *  1. The call may fail in an undefined manner if either mapping is not
+ *     tracked by <handle>.
+ *  2. After executing a batch of unmaps, it is guaranteed that no stale
+ *     mappings will remain in the device or host TLBs.
+ */
+#define GNTTABOP_unmap_and_replace    7
+struct gnttab_unmap_and_replace {
+    /* IN parameters. */
+    uint64_t host_addr;
+    uint64_t new_addr;
+    grant_handle_t handle;
+    /* OUT parameters. */
+    int16_t  status;              /* GNTST_* */
+};
+typedef struct gnttab_unmap_and_replace gnttab_unmap_and_replace_t;
+DEFINE_XEN_GUEST_HANDLE(gnttab_unmap_and_replace_t);
+
+
+/*
+ * Bitfield values for update_pin_status.flags.
+ */
+ /* Map the grant entry for access by I/O devices. */
+#define _GNTMAP_device_map      (0)
+#define GNTMAP_device_map       (1<<_GNTMAP_device_map)
+ /* Map the grant entry for access by host CPUs. */
+#define _GNTMAP_host_map        (1)
+#define GNTMAP_host_map         (1<<_GNTMAP_host_map)
+ /* Accesses to the granted frame will be restricted to read-only access. */
+#define _GNTMAP_readonly        (2)
+#define GNTMAP_readonly         (1<<_GNTMAP_readonly)
+ /*
+  * GNTMAP_host_map subflag:
+  *  0 => The host mapping is usable only by the guest OS.
+  *  1 => The host mapping is usable by guest OS + current application.
+  */
+#define _GNTMAP_application_map (3)
+#define GNTMAP_application_map  (1<<_GNTMAP_application_map)
+
+ /*
+  * GNTMAP_contains_pte subflag:
+  *  0 => This map request contains a host virtual address.
+  *  1 => This map request contains the machine addess of the PTE to update.
+  */
+#define _GNTMAP_contains_pte    (4)
+#define GNTMAP_contains_pte     (1<<_GNTMAP_contains_pte)
+
+/*
+ * Values for error status returns. All errors are -ve.
+ */
+#define GNTST_okay             (0)  /* Normal return.                        */
+#define GNTST_general_error    (-1) /* General undefined error.              */
+#define GNTST_bad_domain       (-2) /* Unrecognsed domain id.                */
+#define GNTST_bad_gntref       (-3) /* Unrecognised or inappropriate gntref. */
+#define GNTST_bad_handle       (-4) /* Unrecognised or inappropriate handle. */
+#define GNTST_bad_virt_addr    (-5) /* Inappropriate virtual address to map. */
+#define GNTST_bad_dev_addr     (-6) /* Inappropriate device address to unmap.*/
+#define GNTST_no_device_space  (-7) /* Out of space in I/O MMU.              */
+#define GNTST_permission_denied (-8) /* Not enough privilege for operation.  */
+#define GNTST_bad_page         (-9) /* Specified page was invalid for op.    */
+#define GNTST_bad_copy_arg    (-10) /* copy arguments cross page boundary.   */
+#define GNTST_address_too_big (-11) /* transfer page address too large.      */
+
+#define GNTTABOP_error_msgs {                   \
+    "okay",                                     \
+    "undefined error",                          \
+    "unrecognised domain id",                   \
+    "invalid grant reference",                  \
+    "invalid mapping handle",                   \
+    "invalid virtual address",                  \
+    "invalid device address",                   \
+    "no spare translation slot in the I/O MMU", \
+    "permission denied",                        \
+    "bad page",                                 \
+    "copy arguments cross page boundary",       \
+    "page address size too large"               \
+}
+
+#endif /* __XEN_PUBLIC_GRANT_TABLE_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/hvm/e820.h linux-2.6.18-xen-3.3.0/include/xen/interface/hvm/e820.h
--- linux-2.6.18.8/include/xen/interface/hvm/e820.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/hvm/e820.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,34 @@
+
+/*
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_PUBLIC_HVM_E820_H__
+#define __XEN_PUBLIC_HVM_E820_H__
+
+/* E820 location in HVM virtual address space. */
+#define HVM_E820_PAGE        0x00090000
+#define HVM_E820_NR_OFFSET   0x000001E8
+#define HVM_E820_OFFSET      0x000002D0
+
+#define HVM_BELOW_4G_RAM_END        0xF0000000
+#define HVM_BELOW_4G_MMIO_START     HVM_BELOW_4G_RAM_END
+#define HVM_BELOW_4G_MMIO_LENGTH    ((1ULL << 32) - HVM_BELOW_4G_MMIO_START)
+
+#endif /* __XEN_PUBLIC_HVM_E820_H__ */
diff -rpuN linux-2.6.18.8/include/xen/interface/hvm/hvm_info_table.h linux-2.6.18-xen-3.3.0/include/xen/interface/hvm/hvm_info_table.h
--- linux-2.6.18.8/include/xen/interface/hvm/hvm_info_table.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/hvm/hvm_info_table.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,41 @@
+/******************************************************************************
+ * hvm/hvm_info_table.h
+ * 
+ * HVM parameter and information table, written into guest memory map.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_PUBLIC_HVM_HVM_INFO_TABLE_H__
+#define __XEN_PUBLIC_HVM_HVM_INFO_TABLE_H__
+
+#define HVM_INFO_PFN         0x09F
+#define HVM_INFO_OFFSET      0x800
+#define HVM_INFO_PADDR       ((HVM_INFO_PFN << 12) + HVM_INFO_OFFSET)
+
+struct hvm_info_table {
+    char        signature[8]; /* "HVM INFO" */
+    uint32_t    length;
+    uint8_t     checksum;
+    uint8_t     acpi_enabled;
+    uint8_t     apic_mode;
+    uint32_t    nr_vcpus;
+};
+
+#endif /* __XEN_PUBLIC_HVM_HVM_INFO_TABLE_H__ */
diff -rpuN linux-2.6.18.8/include/xen/interface/hvm/hvm_op.h linux-2.6.18-xen-3.3.0/include/xen/interface/hvm/hvm_op.h
--- linux-2.6.18.8/include/xen/interface/hvm/hvm_op.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/hvm/hvm_op.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,131 @@
+/*
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_PUBLIC_HVM_HVM_OP_H__
+#define __XEN_PUBLIC_HVM_HVM_OP_H__
+
+/* Get/set subcommands: extra argument == pointer to xen_hvm_param struct. */
+#define HVMOP_set_param           0
+#define HVMOP_get_param           1
+struct xen_hvm_param {
+    domid_t  domid;    /* IN */
+    uint32_t index;    /* IN */
+    uint64_t value;    /* IN/OUT */
+};
+typedef struct xen_hvm_param xen_hvm_param_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_param_t);
+
+/* Set the logical level of one of a domain's PCI INTx wires. */
+#define HVMOP_set_pci_intx_level  2
+struct xen_hvm_set_pci_intx_level {
+    /* Domain to be updated. */
+    domid_t  domid;
+    /* PCI INTx identification in PCI topology (domain:bus:device:intx). */
+    uint8_t  domain, bus, device, intx;
+    /* Assertion level (0 = unasserted, 1 = asserted). */
+    uint8_t  level;
+};
+typedef struct xen_hvm_set_pci_intx_level xen_hvm_set_pci_intx_level_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_set_pci_intx_level_t);
+
+/* Set the logical level of one of a domain's ISA IRQ wires. */
+#define HVMOP_set_isa_irq_level   3
+struct xen_hvm_set_isa_irq_level {
+    /* Domain to be updated. */
+    domid_t  domid;
+    /* ISA device identification, by ISA IRQ (0-15). */
+    uint8_t  isa_irq;
+    /* Assertion level (0 = unasserted, 1 = asserted). */
+    uint8_t  level;
+};
+typedef struct xen_hvm_set_isa_irq_level xen_hvm_set_isa_irq_level_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_set_isa_irq_level_t);
+
+#define HVMOP_set_pci_link_route  4
+struct xen_hvm_set_pci_link_route {
+    /* Domain to be updated. */
+    domid_t  domid;
+    /* PCI link identifier (0-3). */
+    uint8_t  link;
+    /* ISA IRQ (1-15), or 0 (disable link). */
+    uint8_t  isa_irq;
+};
+typedef struct xen_hvm_set_pci_link_route xen_hvm_set_pci_link_route_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_set_pci_link_route_t);
+
+/* Flushes all VCPU TLBs: @arg must be NULL. */
+#define HVMOP_flush_tlbs          5
+
+/* Following tools-only interfaces may change in future. */
+#if defined(__XEN__) || defined(__XEN_TOOLS__)
+
+/* Track dirty VRAM. */
+#define HVMOP_track_dirty_vram    6
+struct xen_hvm_track_dirty_vram {
+    /* Domain to be tracked. */
+    domid_t  domid;
+    /* First pfn to track. */
+    uint64_aligned_t first_pfn;
+    /* Number of pages to track. */
+    uint64_aligned_t nr;
+    /* OUT variable. */
+    /* Dirty bitmap buffer. */
+    XEN_GUEST_HANDLE_64(uint8) dirty_bitmap;
+};
+typedef struct xen_hvm_track_dirty_vram xen_hvm_track_dirty_vram_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_track_dirty_vram_t);
+
+/* Notify that some pages got modified by the Device Model. */
+#define HVMOP_modified_memory    7
+struct xen_hvm_modified_memory {
+    /* Domain to be updated. */
+    domid_t  domid;
+    /* First pfn. */
+    uint64_aligned_t first_pfn;
+    /* Number of pages. */
+    uint64_aligned_t nr;
+};
+typedef struct xen_hvm_modified_memory xen_hvm_modified_memory_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_modified_memory_t);
+
+#define HVMOP_set_mem_type    8
+typedef enum {
+    HVMMEM_ram_rw,             /* Normal read/write guest RAM */
+    HVMMEM_ram_ro,             /* Read-only; writes are discarded */
+    HVMMEM_mmio_dm,            /* Reads and write go to the device model */
+} hvmmem_type_t;
+/* Notify that a region of memory is to be treated in a specific way. */
+struct xen_hvm_set_mem_type {
+    /* Domain to be updated. */
+    domid_t domid;
+    /* Memory type */
+    hvmmem_type_t hvmmem_type;
+    /* First pfn. */
+    uint64_aligned_t first_pfn;
+    /* Number of pages. */
+    uint64_aligned_t nr;
+};
+typedef struct xen_hvm_set_mem_type xen_hvm_set_mem_type_t;
+DEFINE_XEN_GUEST_HANDLE(xen_hvm_set_mem_type_t);
+
+
+#endif /* defined(__XEN__) || defined(__XEN_TOOLS__) */
+
+#endif /* __XEN_PUBLIC_HVM_HVM_OP_H__ */
diff -rpuN linux-2.6.18.8/include/xen/interface/hvm/ioreq.h linux-2.6.18-xen-3.3.0/include/xen/interface/hvm/ioreq.h
--- linux-2.6.18.8/include/xen/interface/hvm/ioreq.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/hvm/ioreq.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,127 @@
+/*
+ * ioreq.h: I/O request definitions for device models
+ * Copyright (c) 2004, Intel Corporation.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef _IOREQ_H_
+#define _IOREQ_H_
+
+#define IOREQ_READ      1
+#define IOREQ_WRITE     0
+
+#define STATE_IOREQ_NONE        0
+#define STATE_IOREQ_READY       1
+#define STATE_IOREQ_INPROCESS   2
+#define STATE_IORESP_READY      3
+
+#define IOREQ_TYPE_PIO          0 /* pio */
+#define IOREQ_TYPE_COPY         1 /* mmio ops */
+#define IOREQ_TYPE_TIMEOFFSET   7
+#define IOREQ_TYPE_INVALIDATE   8 /* mapcache */
+
+/*
+ * VMExit dispatcher should cooperate with instruction decoder to
+ * prepare this structure and notify service OS and DM by sending
+ * virq
+ */
+struct ioreq {
+    uint64_t addr;          /*  physical address            */
+    uint64_t size;          /*  size in bytes               */
+    uint64_t count;         /*  for rep prefixes            */
+    uint64_t data;          /*  data (or paddr of data)     */
+    uint8_t state:4;
+    uint8_t data_is_ptr:1;  /*  if 1, data above is the guest paddr 
+                             *   of the real data to use.   */
+    uint8_t dir:1;          /*  1=read, 0=write             */
+    uint8_t df:1;
+    uint8_t pad:1;
+    uint8_t type;           /* I/O type                     */
+    uint8_t _pad0[6];
+    uint64_t io_count;      /* How many IO done on a vcpu   */
+};
+typedef struct ioreq ioreq_t;
+
+struct vcpu_iodata {
+    struct ioreq vp_ioreq;
+    /* Event channel port, used for notifications to/from the device model. */
+    uint32_t vp_eport;
+    uint32_t _pad0;
+};
+typedef struct vcpu_iodata vcpu_iodata_t;
+
+struct shared_iopage {
+    struct vcpu_iodata   vcpu_iodata[1];
+};
+typedef struct shared_iopage shared_iopage_t;
+
+struct buf_ioreq {
+    uint8_t  type;   /* I/O type                    */
+    uint8_t  pad:1;
+    uint8_t  dir:1;  /* 1=read, 0=write             */
+    uint8_t  size:2; /* 0=>1, 1=>2, 2=>4, 3=>8. If 8, use two buf_ioreqs */
+    uint32_t addr:20;/* physical address            */
+    uint32_t data;   /* data                        */
+};
+typedef struct buf_ioreq buf_ioreq_t;
+
+#define IOREQ_BUFFER_SLOT_NUM     511 /* 8 bytes each, plus 2 4-byte indexes */
+struct buffered_iopage {
+    unsigned int read_pointer;
+    unsigned int write_pointer;
+    buf_ioreq_t buf_ioreq[IOREQ_BUFFER_SLOT_NUM];
+}; /* NB. Size of this structure must be no greater than one page. */
+typedef struct buffered_iopage buffered_iopage_t;
+
+#if defined(__ia64__)
+struct pio_buffer {
+    uint32_t page_offset;
+    uint32_t pointer;
+    uint32_t data_end;
+    uint32_t buf_size;
+    void *opaque;
+};
+
+#define PIO_BUFFER_IDE_PRIMARY   0 /* I/O port = 0x1F0 */
+#define PIO_BUFFER_IDE_SECONDARY 1 /* I/O port = 0x170 */
+#define PIO_BUFFER_ENTRY_NUM     2
+struct buffered_piopage {
+    struct pio_buffer pio[PIO_BUFFER_ENTRY_NUM];
+    uint8_t buffer[1];
+};
+#endif /* defined(__ia64__) */
+
+#define ACPI_PM1A_EVT_BLK_ADDRESS           0x0000000000001f40
+#define ACPI_PM1A_CNT_BLK_ADDRESS           (ACPI_PM1A_EVT_BLK_ADDRESS + 0x04)
+#define ACPI_PM_TMR_BLK_ADDRESS             (ACPI_PM1A_EVT_BLK_ADDRESS + 0x08)
+#define ACPI_GPE0_BLK_ADDRESS               (ACPI_PM_TMR_BLK_ADDRESS + 0x20)
+#define ACPI_GPE0_BLK_LEN                   0x08
+
+#endif /* _IOREQ_H_ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/hvm/params.h linux-2.6.18-xen-3.3.0/include/xen/interface/hvm/params.h
--- linux-2.6.18.8/include/xen/interface/hvm/params.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/hvm/params.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,98 @@
+/*
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_PUBLIC_HVM_PARAMS_H__
+#define __XEN_PUBLIC_HVM_PARAMS_H__
+
+#include "hvm_op.h"
+
+/*
+ * Parameter space for HVMOP_{set,get}_param.
+ */
+
+/*
+ * How should CPU0 event-channel notifications be delivered?
+ * val[63:56] == 0: val[55:0] is a delivery GSI (Global System Interrupt).
+ * val[63:56] == 1: val[55:0] is a delivery PCI INTx line, as follows:
+ *                  Domain = val[47:32], Bus  = val[31:16],
+ *                  DevFn  = val[15: 8], IntX = val[ 1: 0]
+ * If val == 0 then CPU0 event-channel notifications are not delivered.
+ */
+#define HVM_PARAM_CALLBACK_IRQ 0
+
+/*
+ * These are not used by Xen. They are here for convenience of HVM-guest
+ * xenbus implementations.
+ */
+#define HVM_PARAM_STORE_PFN    1
+#define HVM_PARAM_STORE_EVTCHN 2
+
+#define HVM_PARAM_PAE_ENABLED  4
+
+#define HVM_PARAM_IOREQ_PFN    5
+
+#define HVM_PARAM_BUFIOREQ_PFN 6
+
+#ifdef __ia64__
+#define HVM_PARAM_NVRAM_FD     7
+#define HVM_PARAM_VHPT_SIZE    8
+#define HVM_PARAM_BUFPIOREQ_PFN	9
+#endif
+
+/*
+ * Set mode for virtual timers (currently x86 only):
+ *  delay_for_missed_ticks (default):
+ *   Do not advance a vcpu's time beyond the correct delivery time for
+ *   interrupts that have been missed due to preemption. Deliver missed
+ *   interrupts when the vcpu is rescheduled and advance the vcpu's virtual
+ *   time stepwise for each one.
+ *  no_delay_for_missed_ticks:
+ *   As above, missed interrupts are delivered, but guest time always tracks
+ *   wallclock (i.e., real) time while doing so.
+ *  no_missed_ticks_pending:
+ *   No missed interrupts are held pending. Instead, to ensure ticks are
+ *   delivered at some non-zero rate, if we detect missed ticks then the
+ *   internal tick alarm is not disabled if the VCPU is preempted during the
+ *   next tick period.
+ *  one_missed_tick_pending:
+ *   Missed interrupts are collapsed together and delivered as one 'late tick'.
+ *   Guest time always tracks wallclock (i.e., real) time.
+ */
+#define HVM_PARAM_TIMER_MODE   10
+#define HVMPTM_delay_for_missed_ticks    0
+#define HVMPTM_no_delay_for_missed_ticks 1
+#define HVMPTM_no_missed_ticks_pending   2
+#define HVMPTM_one_missed_tick_pending   3
+
+/* Boolean: Enable virtual HPET (high-precision event timer)? (x86-only) */
+#define HVM_PARAM_HPET_ENABLED 11
+
+/* Identity-map page directory used by Intel EPT when CR0.PG=0. */
+#define HVM_PARAM_IDENT_PT     12
+
+/* Device Model domain, defaults to 0. */
+#define HVM_PARAM_DM_DOMAIN    13
+
+/* ACPI S state: currently support S0 and S3 on x86. */
+#define HVM_PARAM_ACPI_S_STATE 14
+
+#define HVM_NR_PARAMS          15
+
+#endif /* __XEN_PUBLIC_HVM_PARAMS_H__ */
diff -rpuN linux-2.6.18.8/include/xen/interface/hvm/save.h linux-2.6.18-xen-3.3.0/include/xen/interface/hvm/save.h
--- linux-2.6.18.8/include/xen/interface/hvm/save.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/hvm/save.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,88 @@
+/* 
+ * hvm/save.h
+ *
+ * Structure definitions for HVM state that is held by Xen and must
+ * be saved along with the domain's memory and device-model state.
+ * 
+ * Copyright (c) 2007 XenSource Ltd.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_PUBLIC_HVM_SAVE_H__
+#define __XEN_PUBLIC_HVM_SAVE_H__
+
+/*
+ * Structures in this header *must* have the same layout in 32bit 
+ * and 64bit environments: this means that all fields must be explicitly 
+ * sized types and aligned to their sizes, and the structs must be 
+ * a multiple of eight bytes long.
+ *
+ * Only the state necessary for saving and restoring (i.e. fields 
+ * that are analogous to actual hardware state) should go in this file. 
+ * Internal mechanisms should be kept in Xen-private headers.
+ */
+
+#if !defined(__GNUC__) || defined(__STRICT_ANSI__)
+#error "Anonymous structs/unions are a GNU extension."
+#endif
+
+/* 
+ * Each entry is preceded by a descriptor giving its type and length
+ */
+struct hvm_save_descriptor {
+    uint16_t typecode;          /* Used to demux the various types below */
+    uint16_t instance;          /* Further demux within a type */
+    uint32_t length;            /* In bytes, *not* including this descriptor */
+};
+
+
+/* 
+ * Each entry has a datatype associated with it: for example, the CPU state 
+ * is saved as a HVM_SAVE_TYPE(CPU), which has HVM_SAVE_LENGTH(CPU), 
+ * and is identified by a descriptor with typecode HVM_SAVE_CODE(CPU).
+ * DECLARE_HVM_SAVE_TYPE binds these things together with some type-system
+ * ugliness.
+ */
+
+#define DECLARE_HVM_SAVE_TYPE(_x, _code, _type)                   \
+  struct __HVM_SAVE_TYPE_##_x { _type t; char c[_code]; }
+
+#define HVM_SAVE_TYPE(_x) typeof (((struct __HVM_SAVE_TYPE_##_x *)(0))->t)
+#define HVM_SAVE_LENGTH(_x) (sizeof (HVM_SAVE_TYPE(_x)))
+#define HVM_SAVE_CODE(_x) (sizeof (((struct __HVM_SAVE_TYPE_##_x *)(0))->c))
+
+
+/* 
+ * The series of save records is teminated by a zero-type, zero-length 
+ * descriptor.
+ */
+
+struct hvm_save_end {};
+DECLARE_HVM_SAVE_TYPE(END, 0, struct hvm_save_end);
+
+#if defined(__i386__) || defined(__x86_64__)
+#include "../arch-x86/hvm/save.h"
+#elif defined(__ia64__)
+#include "../arch-ia64/hvm/save.h"
+#else
+#error "unsupported architecture"
+#endif
+
+#endif /* __XEN_PUBLIC_HVM_SAVE_H__ */
diff -rpuN linux-2.6.18.8/include/xen/interface/hvm/vmx_assist.h linux-2.6.18-xen-3.3.0/include/xen/interface/hvm/vmx_assist.h
--- linux-2.6.18.8/include/xen/interface/hvm/vmx_assist.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/hvm/vmx_assist.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,122 @@
+/*
+ * vmx_assist.h: Context definitions for the VMXASSIST world switch.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Leendert van Doorn, leendert@watson.ibm.com
+ * Copyright (c) 2005, International Business Machines Corporation.
+ */
+
+#ifndef _VMX_ASSIST_H_
+#define _VMX_ASSIST_H_
+
+#define VMXASSIST_BASE         0xD0000
+#define VMXASSIST_MAGIC        0x17101966
+#define VMXASSIST_MAGIC_OFFSET (VMXASSIST_BASE+8)
+
+#define VMXASSIST_NEW_CONTEXT (VMXASSIST_BASE + 12)
+#define VMXASSIST_OLD_CONTEXT (VMXASSIST_NEW_CONTEXT + 4)
+
+#ifndef __ASSEMBLY__
+
+#define NR_EXCEPTION_HANDLER    32
+#define NR_INTERRUPT_HANDLERS   16
+#define NR_TRAPS        (NR_EXCEPTION_HANDLER+NR_INTERRUPT_HANDLERS)
+
+union vmcs_arbytes {
+    struct arbyte_fields {
+        unsigned int seg_type : 4,
+            s         : 1,
+            dpl       : 2,
+            p         : 1,
+            reserved0 : 4,
+            avl       : 1,
+            reserved1 : 1,
+            default_ops_size: 1,
+            g         : 1,
+            null_bit  : 1,
+            reserved2 : 15;
+    } fields;
+    unsigned int bytes;
+};
+
+/*
+ * World switch state
+ */
+struct vmx_assist_context {
+    uint32_t  eip;        /* execution pointer */
+    uint32_t  esp;        /* stack pointer */
+    uint32_t  eflags;     /* flags register */
+    uint32_t  cr0;
+    uint32_t  cr3;        /* page table directory */
+    uint32_t  cr4;
+    uint32_t  idtr_limit; /* idt */
+    uint32_t  idtr_base;
+    uint32_t  gdtr_limit; /* gdt */
+    uint32_t  gdtr_base;
+    uint32_t  cs_sel;     /* cs selector */
+    uint32_t  cs_limit;
+    uint32_t  cs_base;
+    union vmcs_arbytes cs_arbytes;
+    uint32_t  ds_sel;     /* ds selector */
+    uint32_t  ds_limit;
+    uint32_t  ds_base;
+    union vmcs_arbytes ds_arbytes;
+    uint32_t  es_sel;     /* es selector */
+    uint32_t  es_limit;
+    uint32_t  es_base;
+    union vmcs_arbytes es_arbytes;
+    uint32_t  ss_sel;     /* ss selector */
+    uint32_t  ss_limit;
+    uint32_t  ss_base;
+    union vmcs_arbytes ss_arbytes;
+    uint32_t  fs_sel;     /* fs selector */
+    uint32_t  fs_limit;
+    uint32_t  fs_base;
+    union vmcs_arbytes fs_arbytes;
+    uint32_t  gs_sel;     /* gs selector */
+    uint32_t  gs_limit;
+    uint32_t  gs_base;
+    union vmcs_arbytes gs_arbytes;
+    uint32_t  tr_sel;     /* task selector */
+    uint32_t  tr_limit;
+    uint32_t  tr_base;
+    union vmcs_arbytes tr_arbytes;
+    uint32_t  ldtr_sel;   /* ldtr selector */
+    uint32_t  ldtr_limit;
+    uint32_t  ldtr_base;
+    union vmcs_arbytes ldtr_arbytes;
+
+    unsigned char rm_irqbase[2];
+};
+typedef struct vmx_assist_context vmx_assist_context_t;
+
+#endif /* __ASSEMBLY__ */
+
+#endif /* _VMX_ASSIST_H_ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/io/blkif.h linux-2.6.18-xen-3.3.0/include/xen/interface/io/blkif.h
--- linux-2.6.18.8/include/xen/interface/io/blkif.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/io/blkif.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,141 @@
+/******************************************************************************
+ * blkif.h
+ * 
+ * Unified block-device I/O interface for Xen guest OSes.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2003-2004, Keir Fraser
+ */
+
+#ifndef __XEN_PUBLIC_IO_BLKIF_H__
+#define __XEN_PUBLIC_IO_BLKIF_H__
+
+#include "ring.h"
+#include "../grant_table.h"
+
+/*
+ * Front->back notifications: When enqueuing a new request, sending a
+ * notification can be made conditional on req_event (i.e., the generic
+ * hold-off mechanism provided by the ring macros). Backends must set
+ * req_event appropriately (e.g., using RING_FINAL_CHECK_FOR_REQUESTS()).
+ * 
+ * Back->front notifications: When enqueuing a new response, sending a
+ * notification can be made conditional on rsp_event (i.e., the generic
+ * hold-off mechanism provided by the ring macros). Frontends must set
+ * rsp_event appropriately (e.g., using RING_FINAL_CHECK_FOR_RESPONSES()).
+ */
+
+#ifndef blkif_vdev_t
+#define blkif_vdev_t   uint16_t
+#endif
+#define blkif_sector_t uint64_t
+
+/*
+ * REQUEST CODES.
+ */
+#define BLKIF_OP_READ              0
+#define BLKIF_OP_WRITE             1
+/*
+ * Recognised only if "feature-barrier" is present in backend xenbus info.
+ * The "feature-barrier" node contains a boolean indicating whether barrier
+ * requests are likely to succeed or fail. Either way, a barrier request
+ * may fail at any time with BLKIF_RSP_EOPNOTSUPP if it is unsupported by
+ * the underlying block-device hardware. The boolean simply indicates whether
+ * or not it is worthwhile for the frontend to attempt barrier requests.
+ * If a backend does not recognise BLKIF_OP_WRITE_BARRIER, it should *not*
+ * create the "feature-barrier" node!
+ */
+#define BLKIF_OP_WRITE_BARRIER     2
+/*
+ * Recognised if "feature-flush-cache" is present in backend xenbus
+ * info.  A flush will ask the underlying storage hardware to flush its
+ * non-volatile caches as appropriate.  The "feature-flush-cache" node
+ * contains a boolean indicating whether flush requests are likely to
+ * succeed or fail. Either way, a flush request may fail at any time
+ * with BLKIF_RSP_EOPNOTSUPP if it is unsupported by the underlying
+ * block-device hardware. The boolean simply indicates whether or not it
+ * is worthwhile for the frontend to attempt flushes.  If a backend does
+ * not recognise BLKIF_OP_WRITE_FLUSH_CACHE, it should *not* create the
+ * "feature-flush-cache" node!
+ */
+#define BLKIF_OP_FLUSH_DISKCACHE   3
+
+/*
+ * Maximum scatter/gather segments per request.
+ * This is carefully chosen so that sizeof(blkif_ring_t) <= PAGE_SIZE.
+ * NB. This could be 12 if the ring indexes weren't stored in the same page.
+ */
+#define BLKIF_MAX_SEGMENTS_PER_REQUEST 11
+
+struct blkif_request_segment {
+    grant_ref_t gref;        /* reference to I/O buffer frame        */
+    /* @first_sect: first sector in frame to transfer (inclusive).   */
+    /* @last_sect: last sector in frame to transfer (inclusive).     */
+    uint8_t     first_sect, last_sect;
+};
+
+struct blkif_request {
+    uint8_t        operation;    /* BLKIF_OP_???                         */
+    uint8_t        nr_segments;  /* number of segments                   */
+    blkif_vdev_t   handle;       /* only for read/write requests         */
+    uint64_t       id;           /* private guest value, echoed in resp  */
+    blkif_sector_t sector_number;/* start sector idx on disk (r/w only)  */
+    struct blkif_request_segment seg[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+};
+typedef struct blkif_request blkif_request_t;
+
+struct blkif_response {
+    uint64_t        id;              /* copied from request */
+    uint8_t         operation;       /* copied from request */
+    int16_t         status;          /* BLKIF_RSP_???       */
+};
+typedef struct blkif_response blkif_response_t;
+
+/*
+ * STATUS RETURN CODES.
+ */
+ /* Operation not supported (only happens on barrier writes). */
+#define BLKIF_RSP_EOPNOTSUPP  -2
+ /* Operation failed for some unspecified reason (-EIO). */
+#define BLKIF_RSP_ERROR       -1
+ /* Operation completed successfully. */
+#define BLKIF_RSP_OKAY         0
+
+/*
+ * Generate blkif ring structures and types.
+ */
+
+DEFINE_RING_TYPES(blkif, struct blkif_request, struct blkif_response);
+
+#define VDISK_CDROM        0x1
+#define VDISK_REMOVABLE    0x2
+#define VDISK_READONLY     0x4
+
+#endif /* __XEN_PUBLIC_IO_BLKIF_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/io/console.h linux-2.6.18-xen-3.3.0/include/xen/interface/io/console.h
--- linux-2.6.18.8/include/xen/interface/io/console.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/io/console.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,51 @@
+/******************************************************************************
+ * console.h
+ * 
+ * Console I/O interface for Xen guest OSes.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2005, Keir Fraser
+ */
+
+#ifndef __XEN_PUBLIC_IO_CONSOLE_H__
+#define __XEN_PUBLIC_IO_CONSOLE_H__
+
+typedef uint32_t XENCONS_RING_IDX;
+
+#define MASK_XENCONS_IDX(idx, ring) ((idx) & (sizeof(ring)-1))
+
+struct xencons_interface {
+    char in[1024];
+    char out[2048];
+    XENCONS_RING_IDX in_cons, in_prod;
+    XENCONS_RING_IDX out_cons, out_prod;
+};
+
+#endif /* __XEN_PUBLIC_IO_CONSOLE_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/io/fbif.h linux-2.6.18-xen-3.3.0/include/xen/interface/io/fbif.h
--- linux-2.6.18.8/include/xen/interface/io/fbif.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/io/fbif.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,176 @@
+/*
+ * fbif.h -- Xen virtual frame buffer device
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (C) 2005 Anthony Liguori <aliguori@us.ibm.com>
+ * Copyright (C) 2006 Red Hat, Inc., Markus Armbruster <armbru@redhat.com>
+ */
+
+#ifndef __XEN_PUBLIC_IO_FBIF_H__
+#define __XEN_PUBLIC_IO_FBIF_H__
+
+/* Out events (frontend -> backend) */
+
+/*
+ * Out events may be sent only when requested by backend, and receipt
+ * of an unknown out event is an error.
+ */
+
+/* Event type 1 currently not used */
+/*
+ * Framebuffer update notification event
+ * Capable frontend sets feature-update in xenstore.
+ * Backend requests it by setting request-update in xenstore.
+ */
+#define XENFB_TYPE_UPDATE 2
+
+struct xenfb_update
+{
+    uint8_t type;    /* XENFB_TYPE_UPDATE */
+    int32_t x;      /* source x */
+    int32_t y;      /* source y */
+    int32_t width;  /* rect width */
+    int32_t height; /* rect height */
+};
+
+/*
+ * Framebuffer resize notification event
+ * Capable backend sets feature-resize in xenstore.
+ */
+#define XENFB_TYPE_RESIZE 3
+
+struct xenfb_resize
+{
+    uint8_t type;    /* XENFB_TYPE_RESIZE */
+    int32_t width;   /* width in pixels */
+    int32_t height;  /* height in pixels */
+    int32_t stride;  /* stride in bytes */
+    int32_t depth;   /* depth in bits */
+    int32_t offset;  /* offset of the framebuffer in bytes */
+};
+
+#define XENFB_OUT_EVENT_SIZE 40
+
+union xenfb_out_event
+{
+    uint8_t type;
+    struct xenfb_update update;
+    struct xenfb_resize resize;
+    char pad[XENFB_OUT_EVENT_SIZE];
+};
+
+/* In events (backend -> frontend) */
+
+/*
+ * Frontends should ignore unknown in events.
+ */
+
+/*
+ * Framebuffer refresh period advice
+ * Backend sends it to advise the frontend their preferred period of
+ * refresh.  Frontends that keep the framebuffer constantly up-to-date
+ * just ignore it.  Frontends that use the advice should immediately
+ * refresh the framebuffer (and send an update notification event if
+ * those have been requested), then use the update frequency to guide
+ * their periodical refreshs.
+ */
+#define XENFB_TYPE_REFRESH_PERIOD 1
+#define XENFB_NO_REFRESH 0
+
+struct xenfb_refresh_period
+{
+    uint8_t type;    /* XENFB_TYPE_UPDATE_PERIOD */
+    uint32_t period; /* period of refresh, in ms,
+                      * XENFB_NO_REFRESH if no refresh is needed */
+};
+
+#define XENFB_IN_EVENT_SIZE 40
+
+union xenfb_in_event
+{
+    uint8_t type;
+    struct xenfb_refresh_period refresh_period;
+    char pad[XENFB_IN_EVENT_SIZE];
+};
+
+/* shared page */
+
+#define XENFB_IN_RING_SIZE 1024
+#define XENFB_IN_RING_LEN (XENFB_IN_RING_SIZE / XENFB_IN_EVENT_SIZE)
+#define XENFB_IN_RING_OFFS 1024
+#define XENFB_IN_RING(page) \
+    ((union xenfb_in_event *)((char *)(page) + XENFB_IN_RING_OFFS))
+#define XENFB_IN_RING_REF(page, idx) \
+    (XENFB_IN_RING((page))[(idx) % XENFB_IN_RING_LEN])
+
+#define XENFB_OUT_RING_SIZE 2048
+#define XENFB_OUT_RING_LEN (XENFB_OUT_RING_SIZE / XENFB_OUT_EVENT_SIZE)
+#define XENFB_OUT_RING_OFFS (XENFB_IN_RING_OFFS + XENFB_IN_RING_SIZE)
+#define XENFB_OUT_RING(page) \
+    ((union xenfb_out_event *)((char *)(page) + XENFB_OUT_RING_OFFS))
+#define XENFB_OUT_RING_REF(page, idx) \
+    (XENFB_OUT_RING((page))[(idx) % XENFB_OUT_RING_LEN])
+
+struct xenfb_page
+{
+    uint32_t in_cons, in_prod;
+    uint32_t out_cons, out_prod;
+
+    int32_t width;          /* the width of the framebuffer (in pixels) */
+    int32_t height;         /* the height of the framebuffer (in pixels) */
+    uint32_t line_length;   /* the length of a row of pixels (in bytes) */
+    uint32_t mem_length;    /* the length of the framebuffer (in bytes) */
+    uint8_t depth;          /* the depth of a pixel (in bits) */
+
+    /*
+     * Framebuffer page directory
+     *
+     * Each directory page holds PAGE_SIZE / sizeof(*pd)
+     * framebuffer pages, and can thus map up to PAGE_SIZE *
+     * PAGE_SIZE / sizeof(*pd) bytes.  With PAGE_SIZE == 4096 and
+     * sizeof(unsigned long) == 4/8, that's 4 Megs 32 bit and 2 Megs
+     * 64 bit.  256 directories give enough room for a 512 Meg
+     * framebuffer with a max resolution of 12,800x10,240.  Should
+     * be enough for a while with room leftover for expansion.
+     */
+    unsigned long pd[256];
+};
+
+/*
+ * Wart: xenkbd needs to know default resolution.  Put it here until a
+ * better solution is found, but don't leak it to the backend.
+ */
+#ifdef __KERNEL__
+#define XENFB_WIDTH 800
+#define XENFB_HEIGHT 600
+#define XENFB_DEPTH 32
+#endif
+
+#endif
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/io/fsif.h linux-2.6.18-xen-3.3.0/include/xen/interface/io/fsif.h
--- linux-2.6.18.8/include/xen/interface/io/fsif.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/io/fsif.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,191 @@
+/******************************************************************************
+ * fsif.h
+ * 
+ * Interface to FS level split device drivers.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2007, Grzegorz Milos, <gm281@cam.ac.uk>.
+ */
+
+#ifndef __XEN_PUBLIC_IO_FSIF_H__
+#define __XEN_PUBLIC_IO_FSIF_H__
+
+#include "ring.h"
+#include "../grant_table.h"
+
+#define REQ_FILE_OPEN        1
+#define REQ_FILE_CLOSE       2
+#define REQ_FILE_READ        3
+#define REQ_FILE_WRITE       4
+#define REQ_STAT             5
+#define REQ_FILE_TRUNCATE    6
+#define REQ_REMOVE           7
+#define REQ_RENAME           8
+#define REQ_CREATE           9
+#define REQ_DIR_LIST        10
+#define REQ_CHMOD           11
+#define REQ_FS_SPACE        12
+#define REQ_FILE_SYNC       13
+
+struct fsif_open_request {
+    grant_ref_t gref;
+};
+
+struct fsif_close_request {
+    uint32_t fd;
+};
+
+struct fsif_read_request {
+    uint32_t fd;
+    int32_t pad;
+    uint64_t len;
+    uint64_t offset;
+    grant_ref_t grefs[1];  /* Variable length */
+};
+
+struct fsif_write_request {
+    uint32_t fd;
+    int32_t pad;
+    uint64_t len;
+    uint64_t offset;
+    grant_ref_t grefs[1];  /* Variable length */
+};
+
+struct fsif_stat_request {
+    uint32_t fd;
+};
+
+/* This structure is a copy of some fields from stat structure, returned
+ * via the ring. */
+struct fsif_stat_response {
+    int32_t  stat_mode;
+    uint32_t stat_uid;
+    uint32_t stat_gid;
+    int32_t  stat_ret;
+    int64_t  stat_size;
+    int64_t  stat_atime;
+    int64_t  stat_mtime;
+    int64_t  stat_ctime;
+};
+
+struct fsif_truncate_request {
+    uint32_t fd;
+    int32_t pad;
+    int64_t length;
+};
+
+struct fsif_remove_request {
+    grant_ref_t gref;
+};
+
+struct fsif_rename_request {
+    uint16_t old_name_offset;
+    uint16_t new_name_offset;
+    grant_ref_t gref;
+};
+
+struct fsif_create_request {
+    int8_t directory;
+    int8_t pad;
+    int16_t pad2;
+    int32_t mode;
+    grant_ref_t gref;
+};
+
+struct fsif_list_request {
+    uint32_t offset;
+    grant_ref_t gref;
+};
+
+#define NR_FILES_SHIFT  0
+#define NR_FILES_SIZE   16   /* 16 bits for the number of files mask */
+#define NR_FILES_MASK   (((1ULL << NR_FILES_SIZE) - 1) << NR_FILES_SHIFT)
+#define ERROR_SIZE      32   /* 32 bits for the error mask */
+#define ERROR_SHIFT     (NR_FILES_SIZE + NR_FILES_SHIFT)
+#define ERROR_MASK      (((1ULL << ERROR_SIZE) - 1) << ERROR_SHIFT)
+#define HAS_MORE_SHIFT  (ERROR_SHIFT + ERROR_SIZE)    
+#define HAS_MORE_FLAG   (1ULL << HAS_MORE_SHIFT)
+
+struct fsif_chmod_request {
+    uint32_t fd;
+    int32_t mode;
+};
+
+struct fsif_space_request {
+    grant_ref_t gref;
+};
+
+struct fsif_sync_request {
+    uint32_t fd;
+};
+
+
+/* FS operation request */
+struct fsif_request {
+    uint8_t type;                 /* Type of the request                  */
+    uint8_t pad;
+    uint16_t id;                  /* Request ID, copied to the response   */
+    uint32_t pad2;
+    union {
+        struct fsif_open_request     fopen;
+        struct fsif_close_request    fclose;
+        struct fsif_read_request     fread;
+        struct fsif_write_request    fwrite;
+        struct fsif_stat_request     fstat;
+        struct fsif_truncate_request ftruncate;
+        struct fsif_remove_request   fremove;
+        struct fsif_rename_request   frename;
+        struct fsif_create_request   fcreate;
+        struct fsif_list_request     flist;
+        struct fsif_chmod_request    fchmod;
+        struct fsif_space_request    fspace;
+        struct fsif_sync_request     fsync;
+    } u;
+};
+typedef struct fsif_request fsif_request_t;
+
+/* FS operation response */
+struct fsif_response {
+    uint16_t id;
+    uint16_t pad1;
+    uint32_t pad2;
+    union {
+        uint64_t ret_val;
+        struct fsif_stat_response fstat;
+    };
+};
+
+typedef struct fsif_response fsif_response_t;
+
+#define FSIF_RING_ENTRY_SIZE   64
+
+#define FSIF_NR_READ_GNTS  ((FSIF_RING_ENTRY_SIZE - sizeof(struct fsif_read_request)) /  \
+                                sizeof(grant_ref_t) + 1)
+#define FSIF_NR_WRITE_GNTS ((FSIF_RING_ENTRY_SIZE - sizeof(struct fsif_write_request)) / \
+                                sizeof(grant_ref_t) + 1)
+
+DEFINE_RING_TYPES(fsif, struct fsif_request, struct fsif_response);
+
+#define STATE_INITIALISED     "init"
+#define STATE_READY           "ready"
+
+
+
+#endif
diff -rpuN linux-2.6.18.8/include/xen/interface/io/kbdif.h linux-2.6.18-xen-3.3.0/include/xen/interface/io/kbdif.h
--- linux-2.6.18.8/include/xen/interface/io/kbdif.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/io/kbdif.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,132 @@
+/*
+ * kbdif.h -- Xen virtual keyboard/mouse
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (C) 2005 Anthony Liguori <aliguori@us.ibm.com>
+ * Copyright (C) 2006 Red Hat, Inc., Markus Armbruster <armbru@redhat.com>
+ */
+
+#ifndef __XEN_PUBLIC_IO_KBDIF_H__
+#define __XEN_PUBLIC_IO_KBDIF_H__
+
+/* In events (backend -> frontend) */
+
+/*
+ * Frontends should ignore unknown in events.
+ */
+
+/* Pointer movement event */
+#define XENKBD_TYPE_MOTION  1
+/* Event type 2 currently not used */
+/* Key event (includes pointer buttons) */
+#define XENKBD_TYPE_KEY     3
+/*
+ * Pointer position event
+ * Capable backend sets feature-abs-pointer in xenstore.
+ * Frontend requests ot instead of XENKBD_TYPE_MOTION by setting
+ * request-abs-update in xenstore.
+ */
+#define XENKBD_TYPE_POS     4
+
+struct xenkbd_motion
+{
+    uint8_t type;        /* XENKBD_TYPE_MOTION */
+    int32_t rel_x;       /* relative X motion */
+    int32_t rel_y;       /* relative Y motion */
+    int32_t rel_z;       /* relative Z motion (wheel) */
+};
+
+struct xenkbd_key
+{
+    uint8_t type;         /* XENKBD_TYPE_KEY */
+    uint8_t pressed;      /* 1 if pressed; 0 otherwise */
+    uint32_t keycode;     /* KEY_* from linux/input.h */
+};
+
+struct xenkbd_position
+{
+    uint8_t type;        /* XENKBD_TYPE_POS */
+    int32_t abs_x;       /* absolute X position (in FB pixels) */
+    int32_t abs_y;       /* absolute Y position (in FB pixels) */
+    int32_t rel_z;       /* relative Z motion (wheel) */
+};
+
+#define XENKBD_IN_EVENT_SIZE 40
+
+union xenkbd_in_event
+{
+    uint8_t type;
+    struct xenkbd_motion motion;
+    struct xenkbd_key key;
+    struct xenkbd_position pos;
+    char pad[XENKBD_IN_EVENT_SIZE];
+};
+
+/* Out events (frontend -> backend) */
+
+/*
+ * Out events may be sent only when requested by backend, and receipt
+ * of an unknown out event is an error.
+ * No out events currently defined.
+ */
+
+#define XENKBD_OUT_EVENT_SIZE 40
+
+union xenkbd_out_event
+{
+    uint8_t type;
+    char pad[XENKBD_OUT_EVENT_SIZE];
+};
+
+/* shared page */
+
+#define XENKBD_IN_RING_SIZE 2048
+#define XENKBD_IN_RING_LEN (XENKBD_IN_RING_SIZE / XENKBD_IN_EVENT_SIZE)
+#define XENKBD_IN_RING_OFFS 1024
+#define XENKBD_IN_RING(page) \
+    ((union xenkbd_in_event *)((char *)(page) + XENKBD_IN_RING_OFFS))
+#define XENKBD_IN_RING_REF(page, idx) \
+    (XENKBD_IN_RING((page))[(idx) % XENKBD_IN_RING_LEN])
+
+#define XENKBD_OUT_RING_SIZE 1024
+#define XENKBD_OUT_RING_LEN (XENKBD_OUT_RING_SIZE / XENKBD_OUT_EVENT_SIZE)
+#define XENKBD_OUT_RING_OFFS (XENKBD_IN_RING_OFFS + XENKBD_IN_RING_SIZE)
+#define XENKBD_OUT_RING(page) \
+    ((union xenkbd_out_event *)((char *)(page) + XENKBD_OUT_RING_OFFS))
+#define XENKBD_OUT_RING_REF(page, idx) \
+    (XENKBD_OUT_RING((page))[(idx) % XENKBD_OUT_RING_LEN])
+
+struct xenkbd_page
+{
+    uint32_t in_cons, in_prod;
+    uint32_t out_cons, out_prod;
+};
+
+#endif
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/io/netif.h linux-2.6.18-xen-3.3.0/include/xen/interface/io/netif.h
--- linux-2.6.18.8/include/xen/interface/io/netif.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/io/netif.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,205 @@
+/******************************************************************************
+ * netif.h
+ * 
+ * Unified network-device I/O interface for Xen guest OSes.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2003-2004, Keir Fraser
+ */
+
+#ifndef __XEN_PUBLIC_IO_NETIF_H__
+#define __XEN_PUBLIC_IO_NETIF_H__
+
+#include "ring.h"
+#include "../grant_table.h"
+
+/*
+ * Notifications after enqueuing any type of message should be conditional on
+ * the appropriate req_event or rsp_event field in the shared ring.
+ * If the client sends notification for rx requests then it should specify
+ * feature 'feature-rx-notify' via xenbus. Otherwise the backend will assume
+ * that it cannot safely queue packets (as it may not be kicked to send them).
+ */
+
+/*
+ * This is the 'wire' format for packets:
+ *  Request 1: netif_tx_request -- NETTXF_* (any flags)
+ * [Request 2: netif_tx_extra]  (only if request 1 has NETTXF_extra_info)
+ * [Request 3: netif_tx_extra]  (only if request 2 has XEN_NETIF_EXTRA_MORE)
+ *  Request 4: netif_tx_request -- NETTXF_more_data
+ *  Request 5: netif_tx_request -- NETTXF_more_data
+ *  ...
+ *  Request N: netif_tx_request -- 0
+ */
+
+/* Protocol checksum field is blank in the packet (hardware offload)? */
+#define _NETTXF_csum_blank     (0)
+#define  NETTXF_csum_blank     (1U<<_NETTXF_csum_blank)
+
+/* Packet data has been validated against protocol checksum. */
+#define _NETTXF_data_validated (1)
+#define  NETTXF_data_validated (1U<<_NETTXF_data_validated)
+
+/* Packet continues in the next request descriptor. */
+#define _NETTXF_more_data      (2)
+#define  NETTXF_more_data      (1U<<_NETTXF_more_data)
+
+/* Packet to be followed by extra descriptor(s). */
+#define _NETTXF_extra_info     (3)
+#define  NETTXF_extra_info     (1U<<_NETTXF_extra_info)
+
+struct netif_tx_request {
+    grant_ref_t gref;      /* Reference to buffer page */
+    uint16_t offset;       /* Offset within buffer page */
+    uint16_t flags;        /* NETTXF_* */
+    uint16_t id;           /* Echoed in response message. */
+    uint16_t size;         /* Packet size in bytes.       */
+};
+typedef struct netif_tx_request netif_tx_request_t;
+
+/* Types of netif_extra_info descriptors. */
+#define XEN_NETIF_EXTRA_TYPE_NONE      (0)  /* Never used - invalid */
+#define XEN_NETIF_EXTRA_TYPE_GSO       (1)  /* u.gso */
+#define XEN_NETIF_EXTRA_TYPE_MCAST_ADD (2)  /* u.mcast */
+#define XEN_NETIF_EXTRA_TYPE_MCAST_DEL (3)  /* u.mcast */
+#define XEN_NETIF_EXTRA_TYPE_MAX       (4)
+
+/* netif_extra_info flags. */
+#define _XEN_NETIF_EXTRA_FLAG_MORE (0)
+#define XEN_NETIF_EXTRA_FLAG_MORE  (1U<<_XEN_NETIF_EXTRA_FLAG_MORE)
+
+/* GSO types - only TCPv4 currently supported. */
+#define XEN_NETIF_GSO_TYPE_TCPV4        (1)
+
+/*
+ * This structure needs to fit within both netif_tx_request and
+ * netif_rx_response for compatibility.
+ */
+struct netif_extra_info {
+    uint8_t type;  /* XEN_NETIF_EXTRA_TYPE_* */
+    uint8_t flags; /* XEN_NETIF_EXTRA_FLAG_* */
+
+    union {
+        /*
+         * XEN_NETIF_EXTRA_TYPE_GSO:
+         */
+        struct {
+            /*
+             * Maximum payload size of each segment. For example, for TCP this
+             * is just the path MSS.
+             */
+            uint16_t size;
+
+            /*
+             * GSO type. This determines the protocol of the packet and any
+             * extra features required to segment the packet properly.
+             */
+            uint8_t type; /* XEN_NETIF_GSO_TYPE_* */
+
+            /* Future expansion. */
+            uint8_t pad;
+
+            /*
+             * GSO features. This specifies any extra GSO features required
+             * to process this packet, such as ECN support for TCPv4.
+             */
+            uint16_t features; /* XEN_NETIF_GSO_FEAT_* */
+        } gso;
+
+        /*
+         * XEN_NETIF_EXTRA_TYPE_MCAST_{ADD,DEL}:
+         * Backend advertises availability via 'feature-multicast-control'
+         * xenbus node containing value '1'.
+         * Frontend requests this feature by advertising
+         * 'request-multicast-control' xenbus node containing value '1'.
+         * If multicast control is requested then multicast flooding is
+         * disabled and the frontend must explicitly register its interest
+         * in multicast groups using dummy transmit requests containing
+         * MCAST_{ADD,DEL} extra-info fragments.
+         */
+        struct {
+            uint8_t addr[6]; /* Address to add/remove. */
+        } mcast;
+
+        uint16_t pad[3];
+    } u;
+};
+typedef struct netif_extra_info netif_extra_info_t;
+
+struct netif_tx_response {
+    uint16_t id;
+    int16_t  status;       /* NETIF_RSP_* */
+};
+typedef struct netif_tx_response netif_tx_response_t;
+
+struct netif_rx_request {
+    uint16_t    id;        /* Echoed in response message.        */
+    grant_ref_t gref;      /* Reference to incoming granted frame */
+};
+typedef struct netif_rx_request netif_rx_request_t;
+
+/* Packet data has been validated against protocol checksum. */
+#define _NETRXF_data_validated (0)
+#define  NETRXF_data_validated (1U<<_NETRXF_data_validated)
+
+/* Protocol checksum field is blank in the packet (hardware offload)? */
+#define _NETRXF_csum_blank     (1)
+#define  NETRXF_csum_blank     (1U<<_NETRXF_csum_blank)
+
+/* Packet continues in the next request descriptor. */
+#define _NETRXF_more_data      (2)
+#define  NETRXF_more_data      (1U<<_NETRXF_more_data)
+
+/* Packet to be followed by extra descriptor(s). */
+#define _NETRXF_extra_info     (3)
+#define  NETRXF_extra_info     (1U<<_NETRXF_extra_info)
+
+struct netif_rx_response {
+    uint16_t id;
+    uint16_t offset;       /* Offset in page of start of received packet  */
+    uint16_t flags;        /* NETRXF_* */
+    int16_t  status;       /* -ve: BLKIF_RSP_* ; +ve: Rx'ed pkt size. */
+};
+typedef struct netif_rx_response netif_rx_response_t;
+
+/*
+ * Generate netif ring structures and types.
+ */
+
+DEFINE_RING_TYPES(netif_tx, struct netif_tx_request, struct netif_tx_response);
+DEFINE_RING_TYPES(netif_rx, struct netif_rx_request, struct netif_rx_response);
+
+#define NETIF_RSP_DROPPED         -2
+#define NETIF_RSP_ERROR           -1
+#define NETIF_RSP_OKAY             0
+/* No response: used for auxiliary requests (e.g., netif_tx_extra). */
+#define NETIF_RSP_NULL             1
+
+#endif
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/io/pciif.h linux-2.6.18-xen-3.3.0/include/xen/interface/io/pciif.h
--- linux-2.6.18.8/include/xen/interface/io/pciif.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/io/pciif.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,101 @@
+/*
+ * PCI Backend/Frontend Common Data Structures & Macros
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ *   Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+#ifndef __XEN_PCI_COMMON_H__
+#define __XEN_PCI_COMMON_H__
+
+/* Be sure to bump this number if you change this file */
+#define XEN_PCI_MAGIC "7"
+
+/* xen_pci_sharedinfo flags */
+#define _XEN_PCIF_active     (0)
+#define XEN_PCIF_active      (1<<_XEN_PCI_active)
+
+/* xen_pci_op commands */
+#define XEN_PCI_OP_conf_read    (0)
+#define XEN_PCI_OP_conf_write   (1)
+#define XEN_PCI_OP_enable_msi   (2)
+#define XEN_PCI_OP_disable_msi  (3)
+#define XEN_PCI_OP_enable_msix  (4)
+#define XEN_PCI_OP_disable_msix (5)
+
+/* xen_pci_op error numbers */
+#define XEN_PCI_ERR_success          (0)
+#define XEN_PCI_ERR_dev_not_found   (-1)
+#define XEN_PCI_ERR_invalid_offset  (-2)
+#define XEN_PCI_ERR_access_denied   (-3)
+#define XEN_PCI_ERR_not_implemented (-4)
+/* XEN_PCI_ERR_op_failed - backend failed to complete the operation */
+#define XEN_PCI_ERR_op_failed       (-5)
+
+/*
+ * it should be PAGE_SIZE-sizeof(struct xen_pci_op))/sizeof(struct msix_entry))
+ * Should not exceed 128
+ */
+#define SH_INFO_MAX_VEC     128
+
+struct xen_msix_entry {
+    uint16_t vector;
+    uint16_t entry;
+};
+struct xen_pci_op {
+    /* IN: what action to perform: XEN_PCI_OP_* */
+    uint32_t cmd;
+
+    /* OUT: will contain an error number (if any) from errno.h */
+    int32_t err;
+
+    /* IN: which device to touch */
+    uint32_t domain; /* PCI Domain/Segment */
+    uint32_t bus;
+    uint32_t devfn;
+
+    /* IN: which configuration registers to touch */
+    int32_t offset;
+    int32_t size;
+
+    /* IN/OUT: Contains the result after a READ or the value to WRITE */
+    uint32_t value;
+    /* IN: Contains extra infor for this operation */
+    uint32_t info;
+    /*IN:  param for msi-x */
+    struct xen_msix_entry msix_entries[SH_INFO_MAX_VEC];
+};
+
+struct xen_pci_sharedinfo {
+    /* flags - XEN_PCIF_* */
+    uint32_t flags;
+    struct xen_pci_op op;
+};
+
+#endif /* __XEN_PCI_COMMON_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/io/protocols.h linux-2.6.18-xen-3.3.0/include/xen/interface/io/protocols.h
--- linux-2.6.18.8/include/xen/interface/io/protocols.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/io/protocols.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,40 @@
+/******************************************************************************
+ * protocols.h
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_PROTOCOLS_H__
+#define __XEN_PROTOCOLS_H__
+
+#define XEN_IO_PROTO_ABI_X86_32     "x86_32-abi"
+#define XEN_IO_PROTO_ABI_X86_64     "x86_64-abi"
+#define XEN_IO_PROTO_ABI_IA64       "ia64-abi"
+
+#if defined(__i386__)
+# define XEN_IO_PROTO_ABI_NATIVE XEN_IO_PROTO_ABI_X86_32
+#elif defined(__x86_64__)
+# define XEN_IO_PROTO_ABI_NATIVE XEN_IO_PROTO_ABI_X86_64
+#elif defined(__ia64__)
+# define XEN_IO_PROTO_ABI_NATIVE XEN_IO_PROTO_ABI_IA64
+#else
+# error arch fixup needed here
+#endif
+
+#endif
diff -rpuN linux-2.6.18.8/include/xen/interface/io/ring.h linux-2.6.18-xen-3.3.0/include/xen/interface/io/ring.h
--- linux-2.6.18.8/include/xen/interface/io/ring.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/io/ring.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,307 @@
+/******************************************************************************
+ * ring.h
+ * 
+ * Shared producer-consumer ring macros.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Tim Deegan and Andrew Warfield November 2004.
+ */
+
+#ifndef __XEN_PUBLIC_IO_RING_H__
+#define __XEN_PUBLIC_IO_RING_H__
+
+#include "../xen-compat.h"
+
+#if __XEN_INTERFACE_VERSION__ < 0x00030208
+#define xen_mb()  mb()
+#define xen_rmb() rmb()
+#define xen_wmb() wmb()
+#endif
+
+typedef unsigned int RING_IDX;
+
+/* Round a 32-bit unsigned constant down to the nearest power of two. */
+#define __RD2(_x)  (((_x) & 0x00000002) ? 0x2                  : ((_x) & 0x1))
+#define __RD4(_x)  (((_x) & 0x0000000c) ? __RD2((_x)>>2)<<2    : __RD2(_x))
+#define __RD8(_x)  (((_x) & 0x000000f0) ? __RD4((_x)>>4)<<4    : __RD4(_x))
+#define __RD16(_x) (((_x) & 0x0000ff00) ? __RD8((_x)>>8)<<8    : __RD8(_x))
+#define __RD32(_x) (((_x) & 0xffff0000) ? __RD16((_x)>>16)<<16 : __RD16(_x))
+
+/*
+ * Calculate size of a shared ring, given the total available space for the
+ * ring and indexes (_sz), and the name tag of the request/response structure.
+ * A ring contains as many entries as will fit, rounded down to the nearest 
+ * power of two (so we can mask with (size-1) to loop around).
+ */
+#define __RING_SIZE(_s, _sz) \
+    (__RD32(((_sz) - (long)(_s)->ring + (long)(_s)) / sizeof((_s)->ring[0])))
+
+/*
+ * Macros to make the correct C datatypes for a new kind of ring.
+ * 
+ * To make a new ring datatype, you need to have two message structures,
+ * let's say request_t, and response_t already defined.
+ *
+ * In a header where you want the ring datatype declared, you then do:
+ *
+ *     DEFINE_RING_TYPES(mytag, request_t, response_t);
+ *
+ * These expand out to give you a set of types, as you can see below.
+ * The most important of these are:
+ * 
+ *     mytag_sring_t      - The shared ring.
+ *     mytag_front_ring_t - The 'front' half of the ring.
+ *     mytag_back_ring_t  - The 'back' half of the ring.
+ *
+ * To initialize a ring in your code you need to know the location and size
+ * of the shared memory area (PAGE_SIZE, for instance). To initialise
+ * the front half:
+ *
+ *     mytag_front_ring_t front_ring;
+ *     SHARED_RING_INIT((mytag_sring_t *)shared_page);
+ *     FRONT_RING_INIT(&front_ring, (mytag_sring_t *)shared_page, PAGE_SIZE);
+ *
+ * Initializing the back follows similarly (note that only the front
+ * initializes the shared ring):
+ *
+ *     mytag_back_ring_t back_ring;
+ *     BACK_RING_INIT(&back_ring, (mytag_sring_t *)shared_page, PAGE_SIZE);
+ */
+
+#define DEFINE_RING_TYPES(__name, __req_t, __rsp_t)                     \
+                                                                        \
+/* Shared ring entry */                                                 \
+union __name##_sring_entry {                                            \
+    __req_t req;                                                        \
+    __rsp_t rsp;                                                        \
+};                                                                      \
+                                                                        \
+/* Shared ring page */                                                  \
+struct __name##_sring {                                                 \
+    RING_IDX req_prod, req_event;                                       \
+    RING_IDX rsp_prod, rsp_event;                                       \
+    uint8_t  pad[48];                                                   \
+    union __name##_sring_entry ring[1]; /* variable-length */           \
+};                                                                      \
+                                                                        \
+/* "Front" end's private variables */                                   \
+struct __name##_front_ring {                                            \
+    RING_IDX req_prod_pvt;                                              \
+    RING_IDX rsp_cons;                                                  \
+    unsigned int nr_ents;                                               \
+    struct __name##_sring *sring;                                       \
+};                                                                      \
+                                                                        \
+/* "Back" end's private variables */                                    \
+struct __name##_back_ring {                                             \
+    RING_IDX rsp_prod_pvt;                                              \
+    RING_IDX req_cons;                                                  \
+    unsigned int nr_ents;                                               \
+    struct __name##_sring *sring;                                       \
+};                                                                      \
+                                                                        \
+/* Syntactic sugar */                                                   \
+typedef struct __name##_sring __name##_sring_t;                         \
+typedef struct __name##_front_ring __name##_front_ring_t;               \
+typedef struct __name##_back_ring __name##_back_ring_t
+
+/*
+ * Macros for manipulating rings.
+ * 
+ * FRONT_RING_whatever works on the "front end" of a ring: here 
+ * requests are pushed on to the ring and responses taken off it.
+ * 
+ * BACK_RING_whatever works on the "back end" of a ring: here 
+ * requests are taken off the ring and responses put on.
+ * 
+ * N.B. these macros do NO INTERLOCKS OR FLOW CONTROL. 
+ * This is OK in 1-for-1 request-response situations where the 
+ * requestor (front end) never has more than RING_SIZE()-1
+ * outstanding requests.
+ */
+
+/* Initialising empty rings */
+#define SHARED_RING_INIT(_s) do {                                       \
+    (_s)->req_prod  = (_s)->rsp_prod  = 0;                              \
+    (_s)->req_event = (_s)->rsp_event = 1;                              \
+    (void)memset((_s)->pad, 0, sizeof((_s)->pad));                      \
+} while(0)
+
+#define FRONT_RING_INIT(_r, _s, __size) do {                            \
+    (_r)->req_prod_pvt = 0;                                             \
+    (_r)->rsp_cons = 0;                                                 \
+    (_r)->nr_ents = __RING_SIZE(_s, __size);                            \
+    (_r)->sring = (_s);                                                 \
+} while (0)
+
+#define BACK_RING_INIT(_r, _s, __size) do {                             \
+    (_r)->rsp_prod_pvt = 0;                                             \
+    (_r)->req_cons = 0;                                                 \
+    (_r)->nr_ents = __RING_SIZE(_s, __size);                            \
+    (_r)->sring = (_s);                                                 \
+} while (0)
+
+/* Initialize to existing shared indexes -- for recovery */
+#define FRONT_RING_ATTACH(_r, _s, __size) do {                          \
+    (_r)->sring = (_s);                                                 \
+    (_r)->req_prod_pvt = (_s)->req_prod;                                \
+    (_r)->rsp_cons = (_s)->rsp_prod;                                    \
+    (_r)->nr_ents = __RING_SIZE(_s, __size);                            \
+} while (0)
+
+#define BACK_RING_ATTACH(_r, _s, __size) do {                           \
+    (_r)->sring = (_s);                                                 \
+    (_r)->rsp_prod_pvt = (_s)->rsp_prod;                                \
+    (_r)->req_cons = (_s)->req_prod;                                    \
+    (_r)->nr_ents = __RING_SIZE(_s, __size);                            \
+} while (0)
+
+/* How big is this ring? */
+#define RING_SIZE(_r)                                                   \
+    ((_r)->nr_ents)
+
+/* Number of free requests (for use on front side only). */
+#define RING_FREE_REQUESTS(_r)                                          \
+    (RING_SIZE(_r) - ((_r)->req_prod_pvt - (_r)->rsp_cons))
+
+/* Test if there is an empty slot available on the front ring.
+ * (This is only meaningful from the front. )
+ */
+#define RING_FULL(_r)                                                   \
+    (RING_FREE_REQUESTS(_r) == 0)
+
+/* Test if there are outstanding messages to be processed on a ring. */
+#define RING_HAS_UNCONSUMED_RESPONSES(_r)                               \
+    ((_r)->sring->rsp_prod - (_r)->rsp_cons)
+
+#ifdef __GNUC__
+#define RING_HAS_UNCONSUMED_REQUESTS(_r) ({                             \
+    unsigned int req = (_r)->sring->req_prod - (_r)->req_cons;          \
+    unsigned int rsp = RING_SIZE(_r) -                                  \
+        ((_r)->req_cons - (_r)->rsp_prod_pvt);                          \
+    req < rsp ? req : rsp;                                              \
+})
+#else
+/* Same as above, but without the nice GCC ({ ... }) syntax. */
+#define RING_HAS_UNCONSUMED_REQUESTS(_r)                                \
+    ((((_r)->sring->req_prod - (_r)->req_cons) <                        \
+      (RING_SIZE(_r) - ((_r)->req_cons - (_r)->rsp_prod_pvt))) ?        \
+     ((_r)->sring->req_prod - (_r)->req_cons) :                         \
+     (RING_SIZE(_r) - ((_r)->req_cons - (_r)->rsp_prod_pvt)))
+#endif
+
+/* Direct access to individual ring elements, by index. */
+#define RING_GET_REQUEST(_r, _idx)                                      \
+    (&((_r)->sring->ring[((_idx) & (RING_SIZE(_r) - 1))].req))
+
+#define RING_GET_RESPONSE(_r, _idx)                                     \
+    (&((_r)->sring->ring[((_idx) & (RING_SIZE(_r) - 1))].rsp))
+
+/* Loop termination condition: Would the specified index overflow the ring? */
+#define RING_REQUEST_CONS_OVERFLOW(_r, _cons)                           \
+    (((_cons) - (_r)->rsp_prod_pvt) >= RING_SIZE(_r))
+
+#define RING_PUSH_REQUESTS(_r) do {                                     \
+    xen_wmb(); /* back sees requests /before/ updated producer index */ \
+    (_r)->sring->req_prod = (_r)->req_prod_pvt;                         \
+} while (0)
+
+#define RING_PUSH_RESPONSES(_r) do {                                    \
+    xen_wmb(); /* front sees resps /before/ updated producer index */   \
+    (_r)->sring->rsp_prod = (_r)->rsp_prod_pvt;                         \
+} while (0)
+
+/*
+ * Notification hold-off (req_event and rsp_event):
+ * 
+ * When queueing requests or responses on a shared ring, it may not always be
+ * necessary to notify the remote end. For example, if requests are in flight
+ * in a backend, the front may be able to queue further requests without
+ * notifying the back (if the back checks for new requests when it queues
+ * responses).
+ * 
+ * When enqueuing requests or responses:
+ * 
+ *  Use RING_PUSH_{REQUESTS,RESPONSES}_AND_CHECK_NOTIFY(). The second argument
+ *  is a boolean return value. True indicates that the receiver requires an
+ *  asynchronous notification.
+ * 
+ * After dequeuing requests or responses (before sleeping the connection):
+ * 
+ *  Use RING_FINAL_CHECK_FOR_REQUESTS() or RING_FINAL_CHECK_FOR_RESPONSES().
+ *  The second argument is a boolean return value. True indicates that there
+ *  are pending messages on the ring (i.e., the connection should not be put
+ *  to sleep).
+ * 
+ *  These macros will set the req_event/rsp_event field to trigger a
+ *  notification on the very next message that is enqueued. If you want to
+ *  create batches of work (i.e., only receive a notification after several
+ *  messages have been enqueued) then you will need to create a customised
+ *  version of the FINAL_CHECK macro in your own code, which sets the event
+ *  field appropriately.
+ */
+
+#define RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(_r, _notify) do {           \
+    RING_IDX __old = (_r)->sring->req_prod;                             \
+    RING_IDX __new = (_r)->req_prod_pvt;                                \
+    xen_wmb(); /* back sees requests /before/ updated producer index */ \
+    (_r)->sring->req_prod = __new;                                      \
+    xen_mb(); /* back sees new requests /before/ we check req_event */  \
+    (_notify) = ((RING_IDX)(__new - (_r)->sring->req_event) <           \
+                 (RING_IDX)(__new - __old));                            \
+} while (0)
+
+#define RING_PUSH_RESPONSES_AND_CHECK_NOTIFY(_r, _notify) do {          \
+    RING_IDX __old = (_r)->sring->rsp_prod;                             \
+    RING_IDX __new = (_r)->rsp_prod_pvt;                                \
+    xen_wmb(); /* front sees resps /before/ updated producer index */   \
+    (_r)->sring->rsp_prod = __new;                                      \
+    xen_mb(); /* front sees new resps /before/ we check rsp_event */    \
+    (_notify) = ((RING_IDX)(__new - (_r)->sring->rsp_event) <           \
+                 (RING_IDX)(__new - __old));                            \
+} while (0)
+
+#define RING_FINAL_CHECK_FOR_REQUESTS(_r, _work_to_do) do {             \
+    (_work_to_do) = RING_HAS_UNCONSUMED_REQUESTS(_r);                   \
+    if (_work_to_do) break;                                             \
+    (_r)->sring->req_event = (_r)->req_cons + 1;                        \
+    xen_mb();                                                           \
+    (_work_to_do) = RING_HAS_UNCONSUMED_REQUESTS(_r);                   \
+} while (0)
+
+#define RING_FINAL_CHECK_FOR_RESPONSES(_r, _work_to_do) do {            \
+    (_work_to_do) = RING_HAS_UNCONSUMED_RESPONSES(_r);                  \
+    if (_work_to_do) break;                                             \
+    (_r)->sring->rsp_event = (_r)->rsp_cons + 1;                        \
+    xen_mb();                                                           \
+    (_work_to_do) = RING_HAS_UNCONSUMED_RESPONSES(_r);                  \
+} while (0)
+
+#endif /* __XEN_PUBLIC_IO_RING_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/io/tpmif.h linux-2.6.18-xen-3.3.0/include/xen/interface/io/tpmif.h
--- linux-2.6.18.8/include/xen/interface/io/tpmif.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/io/tpmif.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,77 @@
+/******************************************************************************
+ * tpmif.h
+ *
+ * TPM I/O interface for Xen guest OSes.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2005, IBM Corporation
+ *
+ * Author: Stefan Berger, stefanb@us.ibm.com
+ * Grant table support: Mahadevan Gomathisankaran
+ *
+ * This code has been derived from tools/libxc/xen/io/netif.h
+ *
+ * Copyright (c) 2003-2004, Keir Fraser
+ */
+
+#ifndef __XEN_PUBLIC_IO_TPMIF_H__
+#define __XEN_PUBLIC_IO_TPMIF_H__
+
+#include "../grant_table.h"
+
+struct tpmif_tx_request {
+    unsigned long addr;   /* Machine address of packet.   */
+    grant_ref_t ref;      /* grant table access reference */
+    uint16_t unused;
+    uint16_t size;        /* Packet size in bytes.        */
+};
+typedef struct tpmif_tx_request tpmif_tx_request_t;
+
+/*
+ * The TPMIF_TX_RING_SIZE defines the number of pages the
+ * front-end and backend can exchange (= size of array).
+ */
+typedef uint32_t TPMIF_RING_IDX;
+
+#define TPMIF_TX_RING_SIZE 1
+
+/* This structure must fit in a memory page. */
+
+struct tpmif_ring {
+    struct tpmif_tx_request req;
+};
+typedef struct tpmif_ring tpmif_ring_t;
+
+struct tpmif_tx_interface {
+    struct tpmif_ring ring[TPMIF_TX_RING_SIZE];
+};
+typedef struct tpmif_tx_interface tpmif_tx_interface_t;
+
+#endif
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/io/vscsiif.h linux-2.6.18-xen-3.3.0/include/xen/interface/io/vscsiif.h
--- linux-2.6.18.8/include/xen/interface/io/vscsiif.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/io/vscsiif.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,105 @@
+/******************************************************************************
+ * vscsiif.h
+ * 
+ * Based on the blkif.h code.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright(c) FUJITSU Limited 2008.
+ */
+
+#ifndef __XEN__PUBLIC_IO_SCSI_H__
+#define __XEN__PUBLIC_IO_SCSI_H__
+
+#include "ring.h"
+#include "../grant_table.h"
+
+/* command between backend and frontend */
+#define VSCSIIF_ACT_SCSI_CDB         1    /* SCSI CDB command */
+#define VSCSIIF_ACT_SCSI_ABORT       2    /* SCSI Device(Lun) Abort*/
+#define VSCSIIF_ACT_SCSI_RESET       3    /* SCSI Device(Lun) Reset*/
+
+
+#define VSCSIIF_BACK_MAX_PENDING_REQS    128
+
+/*
+ * Maximum scatter/gather segments per request.
+ *
+ * Considering balance between allocating al least 16 "vscsiif_request"
+ * structures on one page (4096bytes) and number of scatter gather 
+ * needed, we decided to use 26 as a magic number.
+ */
+#define VSCSIIF_SG_TABLESIZE             26
+
+/*
+ * base on linux kernel 2.6.18
+ */
+#define VSCSIIF_MAX_COMMAND_SIZE         16
+#define VSCSIIF_SENSE_BUFFERSIZE         96
+
+
+struct vscsiif_request {
+    uint16_t rqid;          /* private guest value, echoed in resp  */
+    uint8_t act;            /* command between backend and frontend */
+    uint8_t cmd_len;
+
+    uint8_t cmnd[VSCSIIF_MAX_COMMAND_SIZE];
+    uint16_t timeout_per_command;     /* The command is issued by twice 
+                                         the value in Backend. */
+    uint16_t channel, id, lun;
+    uint16_t padding;
+    uint8_t sc_data_direction;        /* for DMA_TO_DEVICE(1)
+                                         DMA_FROM_DEVICE(2)
+                                         DMA_NONE(3) requests  */
+    uint8_t nr_segments;              /* Number of pieces of scatter-gather */
+
+    struct scsiif_request_segment {
+        grant_ref_t gref;
+        uint16_t offset;
+        uint16_t length;
+    } seg[VSCSIIF_SG_TABLESIZE];
+    uint32_t reserved[3];
+};
+typedef struct vscsiif_request vscsiif_request_t;
+
+struct vscsiif_response {
+    uint16_t rqid;
+    uint8_t padding;
+    uint8_t sense_len;
+    uint8_t sense_buffer[VSCSIIF_SENSE_BUFFERSIZE];
+    int32_t rslt;
+    uint32_t residual_len;     /* request bufflen - 
+                                  return the value from physical device */
+    uint32_t reserved[36];
+};
+typedef struct vscsiif_response vscsiif_response_t;
+
+DEFINE_RING_TYPES(vscsiif, struct vscsiif_request, struct vscsiif_response);
+
+
+#endif  /*__XEN__PUBLIC_IO_SCSI_H__*/
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/io/xenbus.h linux-2.6.18-xen-3.3.0/include/xen/interface/io/xenbus.h
--- linux-2.6.18.8/include/xen/interface/io/xenbus.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/io/xenbus.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,80 @@
+/*****************************************************************************
+ * xenbus.h
+ *
+ * Xenbus protocol details.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (C) 2005 XenSource Ltd.
+ */
+
+#ifndef _XEN_PUBLIC_IO_XENBUS_H
+#define _XEN_PUBLIC_IO_XENBUS_H
+
+/*
+ * The state of either end of the Xenbus, i.e. the current communication
+ * status of initialisation across the bus.  States here imply nothing about
+ * the state of the connection between the driver and the kernel's device
+ * layers.
+ */
+enum xenbus_state {
+    XenbusStateUnknown       = 0,
+
+    XenbusStateInitialising  = 1,
+
+    /*
+     * InitWait: Finished early initialisation but waiting for information
+     * from the peer or hotplug scripts.
+     */
+    XenbusStateInitWait      = 2,
+
+    /*
+     * Initialised: Waiting for a connection from the peer.
+     */
+    XenbusStateInitialised   = 3,
+
+    XenbusStateConnected     = 4,
+
+    /*
+     * Closing: The device is being closed due to an error or an unplug event.
+     */
+    XenbusStateClosing       = 5,
+
+    XenbusStateClosed        = 6,
+
+    /*
+     * Reconfiguring: The device is being reconfigured.
+     */
+    XenbusStateReconfiguring = 7,
+
+    XenbusStateReconfigured  = 8
+};
+typedef enum xenbus_state XenbusState;
+
+#endif /* _XEN_PUBLIC_IO_XENBUS_H */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/io/xs_wire.h linux-2.6.18-xen-3.3.0/include/xen/interface/io/xs_wire.h
--- linux-2.6.18.8/include/xen/interface/io/xs_wire.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/io/xs_wire.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,130 @@
+/*
+ * Details of the "wire" protocol between Xen Store Daemon and client
+ * library or guest kernel.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (C) 2005 Rusty Russell IBM Corporation
+ */
+
+#ifndef _XS_WIRE_H
+#define _XS_WIRE_H
+
+enum xsd_sockmsg_type
+{
+    XS_DEBUG,
+    XS_DIRECTORY,
+    XS_READ,
+    XS_GET_PERMS,
+    XS_WATCH,
+    XS_UNWATCH,
+    XS_TRANSACTION_START,
+    XS_TRANSACTION_END,
+    XS_INTRODUCE,
+    XS_RELEASE,
+    XS_GET_DOMAIN_PATH,
+    XS_WRITE,
+    XS_MKDIR,
+    XS_RM,
+    XS_SET_PERMS,
+    XS_WATCH_EVENT,
+    XS_ERROR,
+    XS_IS_DOMAIN_INTRODUCED,
+    XS_RESUME,
+    XS_SET_TARGET
+};
+
+#define XS_WRITE_NONE "NONE"
+#define XS_WRITE_CREATE "CREATE"
+#define XS_WRITE_CREATE_EXCL "CREATE|EXCL"
+
+/* We hand errors as strings, for portability. */
+struct xsd_errors
+{
+    int errnum;
+    const char *errstring;
+};
+#define XSD_ERROR(x) { x, #x }
+/* LINTED: static unused */
+static struct xsd_errors xsd_errors[]
+#if defined(__GNUC__)
+__attribute__((unused))
+#endif
+    = {
+    XSD_ERROR(EINVAL),
+    XSD_ERROR(EACCES),
+    XSD_ERROR(EEXIST),
+    XSD_ERROR(EISDIR),
+    XSD_ERROR(ENOENT),
+    XSD_ERROR(ENOMEM),
+    XSD_ERROR(ENOSPC),
+    XSD_ERROR(EIO),
+    XSD_ERROR(ENOTEMPTY),
+    XSD_ERROR(ENOSYS),
+    XSD_ERROR(EROFS),
+    XSD_ERROR(EBUSY),
+    XSD_ERROR(EAGAIN),
+    XSD_ERROR(EISCONN)
+};
+
+struct xsd_sockmsg
+{
+    uint32_t type;  /* XS_??? */
+    uint32_t req_id;/* Request identifier, echoed in daemon's response.  */
+    uint32_t tx_id; /* Transaction id (0 if not related to a transaction). */
+    uint32_t len;   /* Length of data following this. */
+
+    /* Generally followed by nul-terminated string(s). */
+};
+
+enum xs_watch_type
+{
+    XS_WATCH_PATH = 0,
+    XS_WATCH_TOKEN
+};
+
+/* Inter-domain shared memory communications. */
+#define XENSTORE_RING_SIZE 1024
+typedef uint32_t XENSTORE_RING_IDX;
+#define MASK_XENSTORE_IDX(idx) ((idx) & (XENSTORE_RING_SIZE-1))
+struct xenstore_domain_interface {
+    char req[XENSTORE_RING_SIZE]; /* Requests to xenstore daemon. */
+    char rsp[XENSTORE_RING_SIZE]; /* Replies and async watch events. */
+    XENSTORE_RING_IDX req_cons, req_prod;
+    XENSTORE_RING_IDX rsp_cons, rsp_prod;
+};
+
+/* Violating this is very bad.  See docs/misc/xenstore.txt. */
+#define XENSTORE_PAYLOAD_MAX 4096
+
+/* Violating these just gets you an error back */
+#define XENSTORE_ABS_PATH_MAX 3072
+#define XENSTORE_REL_PATH_MAX 2048
+
+#endif /* _XS_WIRE_H */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/kexec.h linux-2.6.18-xen-3.3.0/include/xen/interface/kexec.h
--- linux-2.6.18.8/include/xen/interface/kexec.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/kexec.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,189 @@
+/******************************************************************************
+ * kexec.h - Public portion
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ * 
+ * Xen port written by:
+ * - Simon 'Horms' Horman <horms@verge.net.au>
+ * - Magnus Damm <magnus@valinux.co.jp>
+ */
+
+#ifndef _XEN_PUBLIC_KEXEC_H
+#define _XEN_PUBLIC_KEXEC_H
+
+
+/* This file describes the Kexec / Kdump hypercall interface for Xen.
+ *
+ * Kexec under vanilla Linux allows a user to reboot the physical machine 
+ * into a new user-specified kernel. The Xen port extends this idea
+ * to allow rebooting of the machine from dom0. When kexec for dom0
+ * is used to reboot,  both the hypervisor and the domains get replaced
+ * with some other kernel. It is possible to kexec between vanilla
+ * Linux and Xen and back again. Xen to Xen works well too.
+ *
+ * The hypercall interface for kexec can be divided into three main
+ * types of hypercall operations:
+ *
+ * 1) Range information:
+ *    This is used by the dom0 kernel to ask the hypervisor about various 
+ *    address information. This information is needed to allow kexec-tools 
+ *    to fill in the ELF headers for /proc/vmcore properly.
+ *
+ * 2) Load and unload of images:
+ *    There are no big surprises here, the kexec binary from kexec-tools
+ *    runs in userspace in dom0. The tool loads/unloads data into the
+ *    dom0 kernel such as new kernel, initramfs and hypervisor. When
+ *    loaded the dom0 kernel performs a load hypercall operation, and
+ *    before releasing all page references the dom0 kernel calls unload.
+ *
+ * 3) Kexec operation:
+ *    This is used to start a previously loaded kernel.
+ */
+
+#include "xen.h"
+
+#if defined(__i386__) || defined(__x86_64__)
+#define KEXEC_XEN_NO_PAGES 17
+#endif
+
+/*
+ * Prototype for this hypercall is:
+ *  int kexec_op(int cmd, void *args)
+ * @cmd  == KEXEC_CMD_... 
+ *          KEXEC operation to perform
+ * @args == Operation-specific extra arguments (NULL if none).
+ */
+
+/*
+ * Kexec supports two types of operation:
+ * - kexec into a regular kernel, very similar to a standard reboot
+ *   - KEXEC_TYPE_DEFAULT is used to specify this type
+ * - kexec into a special "crash kernel", aka kexec-on-panic
+ *   - KEXEC_TYPE_CRASH is used to specify this type
+ *   - parts of our system may be broken at kexec-on-panic time
+ *     - the code should be kept as simple and self-contained as possible
+ */
+
+#define KEXEC_TYPE_DEFAULT 0
+#define KEXEC_TYPE_CRASH   1
+
+
+/* The kexec implementation for Xen allows the user to load two
+ * types of kernels, KEXEC_TYPE_DEFAULT and KEXEC_TYPE_CRASH.
+ * All data needed for a kexec reboot is kept in one xen_kexec_image_t
+ * per "instance". The data mainly consists of machine address lists to pages
+ * together with destination addresses. The data in xen_kexec_image_t
+ * is passed to the "code page" which is one page of code that performs
+ * the final relocations before jumping to the new kernel.
+ */
+ 
+typedef struct xen_kexec_image {
+#if defined(__i386__) || defined(__x86_64__)
+    unsigned long page_list[KEXEC_XEN_NO_PAGES];
+#endif
+#if defined(__ia64__)
+    unsigned long reboot_code_buffer;
+#endif
+    unsigned long indirection_page;
+    unsigned long start_address;
+} xen_kexec_image_t;
+
+/*
+ * Perform kexec having previously loaded a kexec or kdump kernel
+ * as appropriate.
+ * type == KEXEC_TYPE_DEFAULT or KEXEC_TYPE_CRASH [in]
+ */
+#define KEXEC_CMD_kexec                 0
+typedef struct xen_kexec_exec {
+    int type;
+} xen_kexec_exec_t;
+
+/*
+ * Load/Unload kernel image for kexec or kdump.
+ * type  == KEXEC_TYPE_DEFAULT or KEXEC_TYPE_CRASH [in]
+ * image == relocation information for kexec (ignored for unload) [in]
+ */
+#define KEXEC_CMD_kexec_load            1
+#define KEXEC_CMD_kexec_unload          2
+typedef struct xen_kexec_load {
+    int type;
+    xen_kexec_image_t image;
+} xen_kexec_load_t;
+
+#define KEXEC_RANGE_MA_CRASH      0 /* machine address and size of crash area */
+#define KEXEC_RANGE_MA_XEN        1 /* machine address and size of Xen itself */
+#define KEXEC_RANGE_MA_CPU        2 /* machine address and size of a CPU note */
+#define KEXEC_RANGE_MA_XENHEAP    3 /* machine address and size of xenheap
+                                     * Note that although this is adjacent
+                                     * to Xen it exists in a separate EFI
+                                     * region on ia64, and thus needs to be
+                                     * inserted into iomem_machine separately */
+#define KEXEC_RANGE_MA_BOOT_PARAM 4 /* machine address and size of
+                                     * the ia64_boot_param */
+#define KEXEC_RANGE_MA_EFI_MEMMAP 5 /* machine address and size of
+                                     * of the EFI Memory Map */
+#define KEXEC_RANGE_MA_VMCOREINFO 6 /* machine address and size of vmcoreinfo */
+
+/*
+ * Find the address and size of certain memory areas
+ * range == KEXEC_RANGE_... [in]
+ * nr    == physical CPU number (starting from 0) if KEXEC_RANGE_MA_CPU [in]
+ * size  == number of bytes reserved in window [out]
+ * start == address of the first byte in the window [out]
+ */
+#define KEXEC_CMD_kexec_get_range       3
+typedef struct xen_kexec_range {
+    int range;
+    int nr;
+    unsigned long size;
+    unsigned long start;
+} xen_kexec_range_t;
+
+/* vmcoreinfo stuff */
+#define VMCOREINFO_BYTES           (4096)
+#define VMCOREINFO_NOTE_NAME       "VMCOREINFO_XEN"
+void arch_crash_save_vmcoreinfo(void);
+void vmcoreinfo_append_str(const char *fmt, ...)
+       __attribute__ ((format (printf, 1, 2)));
+#define VMCOREINFO_PAGESIZE(value) \
+       vmcoreinfo_append_str("PAGESIZE=%ld\n", value)
+#define VMCOREINFO_SYMBOL(name) \
+       vmcoreinfo_append_str("SYMBOL(%s)=%lx\n", #name, (unsigned long)&name)
+#define VMCOREINFO_SYMBOL_ALIAS(alias, name) \
+       vmcoreinfo_append_str("SYMBOL(%s)=%lx\n", #alias, (unsigned long)&name)
+#define VMCOREINFO_STRUCT_SIZE(name) \
+       vmcoreinfo_append_str("SIZE(%s)=%zu\n", #name, sizeof(struct name))
+#define VMCOREINFO_OFFSET(name, field) \
+       vmcoreinfo_append_str("OFFSET(%s.%s)=%lu\n", #name, #field, \
+                             (unsigned long)offsetof(struct name, field))
+#define VMCOREINFO_OFFSET_ALIAS(name, field, alias) \
+       vmcoreinfo_append_str("OFFSET(%s.%s)=%lu\n", #name, #alias, \
+                             (unsigned long)offsetof(struct name, field))
+
+#endif /* _XEN_PUBLIC_KEXEC_H */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/libelf.h linux-2.6.18-xen-3.3.0/include/xen/interface/libelf.h
--- linux-2.6.18.8/include/xen/interface/libelf.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/libelf.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,265 @@
+/******************************************************************************
+ * libelf.h
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __XC_LIBELF__
+#define __XC_LIBELF__ 1
+
+#if defined(__i386__) || defined(__x86_64__) || defined(__ia64__)
+#define XEN_ELF_LITTLE_ENDIAN
+#else
+#error define architectural endianness
+#endif
+
+#undef ELFSIZE
+#include "elfnote.h"
+#include "elfstructs.h"
+#include "features.h"
+
+/* ------------------------------------------------------------------------ */
+
+typedef union {
+    Elf32_Ehdr e32;
+    Elf64_Ehdr e64;
+} elf_ehdr;
+
+typedef union {
+    Elf32_Phdr e32;
+    Elf64_Phdr e64;
+} elf_phdr;
+
+typedef union {
+    Elf32_Shdr e32;
+    Elf64_Shdr e64;
+} elf_shdr;
+
+typedef union {
+    Elf32_Sym e32;
+    Elf64_Sym e64;
+} elf_sym;
+
+typedef union {
+    Elf32_Rel e32;
+    Elf64_Rel e64;
+} elf_rel;
+
+typedef union {
+    Elf32_Rela e32;
+    Elf64_Rela e64;
+} elf_rela;
+
+typedef union {
+    Elf32_Note e32;
+    Elf64_Note e64;
+} elf_note;
+
+struct elf_binary {
+    /* elf binary */
+    const char *image;
+    size_t size;
+    char class;
+    char data;
+
+    const elf_ehdr *ehdr;
+    const char *sec_strtab;
+    const elf_shdr *sym_tab;
+    const char *sym_strtab;
+
+    /* loaded to */
+    char *dest;
+    uint64_t pstart;
+    uint64_t pend;
+    uint64_t reloc_offset;
+
+    uint64_t bsd_symtab_pstart;
+    uint64_t bsd_symtab_pend;
+
+#ifndef __XEN__
+    /* misc */
+    FILE *log;
+#endif
+    int verbose;
+};
+
+/* ------------------------------------------------------------------------ */
+/* accessing elf header fields                                              */
+
+#ifdef XEN_ELF_BIG_ENDIAN
+# define NATIVE_ELFDATA ELFDATA2MSB
+#else
+# define NATIVE_ELFDATA ELFDATA2LSB
+#endif
+
+#define elf_32bit(elf) (ELFCLASS32 == (elf)->class)
+#define elf_64bit(elf) (ELFCLASS64 == (elf)->class)
+#define elf_msb(elf)   (ELFDATA2MSB == (elf)->data)
+#define elf_lsb(elf)   (ELFDATA2LSB == (elf)->data)
+#define elf_swap(elf)  (NATIVE_ELFDATA != (elf)->data)
+
+#define elf_uval(elf, str, elem)                                        \
+    ((ELFCLASS64 == (elf)->class)                                       \
+     ? elf_access_unsigned((elf), (str),                                \
+                           offsetof(typeof(*(str)),e64.elem),           \
+                           sizeof((str)->e64.elem))                     \
+     : elf_access_unsigned((elf), (str),                                \
+                           offsetof(typeof(*(str)),e32.elem),           \
+                           sizeof((str)->e32.elem)))
+
+#define elf_sval(elf, str, elem)                                        \
+    ((ELFCLASS64 == (elf)->class)                                       \
+     ? elf_access_signed((elf), (str),                                  \
+                         offsetof(typeof(*(str)),e64.elem),             \
+                         sizeof((str)->e64.elem))                       \
+     : elf_access_signed((elf), (str),                                  \
+                         offsetof(typeof(*(str)),e32.elem),             \
+                         sizeof((str)->e32.elem)))
+
+#define elf_size(elf, str)                              \
+    ((ELFCLASS64 == (elf)->class)                       \
+     ? sizeof((str)->e64) : sizeof((str)->e32))
+
+uint64_t elf_access_unsigned(struct elf_binary *elf, const void *ptr,
+                             uint64_t offset, size_t size);
+int64_t elf_access_signed(struct elf_binary *elf, const void *ptr,
+                          uint64_t offset, size_t size);
+
+uint64_t elf_round_up(struct elf_binary *elf, uint64_t addr);
+
+/* ------------------------------------------------------------------------ */
+/* xc_libelf_tools.c                                                        */
+
+int elf_shdr_count(struct elf_binary *elf);
+int elf_phdr_count(struct elf_binary *elf);
+
+const elf_shdr *elf_shdr_by_name(struct elf_binary *elf, const char *name);
+const elf_shdr *elf_shdr_by_index(struct elf_binary *elf, int index);
+const elf_phdr *elf_phdr_by_index(struct elf_binary *elf, int index);
+
+const char *elf_section_name(struct elf_binary *elf, const elf_shdr * shdr);
+const void *elf_section_start(struct elf_binary *elf, const elf_shdr * shdr);
+const void *elf_section_end(struct elf_binary *elf, const elf_shdr * shdr);
+
+const void *elf_segment_start(struct elf_binary *elf, const elf_phdr * phdr);
+const void *elf_segment_end(struct elf_binary *elf, const elf_phdr * phdr);
+
+const elf_sym *elf_sym_by_name(struct elf_binary *elf, const char *symbol);
+const elf_sym *elf_sym_by_index(struct elf_binary *elf, int index);
+
+const char *elf_note_name(struct elf_binary *elf, const elf_note * note);
+const void *elf_note_desc(struct elf_binary *elf, const elf_note * note);
+uint64_t elf_note_numeric(struct elf_binary *elf, const elf_note * note);
+const elf_note *elf_note_next(struct elf_binary *elf, const elf_note * note);
+
+int elf_is_elfbinary(const void *image);
+int elf_phdr_is_loadable(struct elf_binary *elf, const elf_phdr * phdr);
+
+/* ------------------------------------------------------------------------ */
+/* xc_libelf_loader.c                                                       */
+
+int elf_init(struct elf_binary *elf, const char *image, size_t size);
+#ifdef __XEN__
+void elf_set_verbose(struct elf_binary *elf);
+#else
+void elf_set_logfile(struct elf_binary *elf, FILE * log, int verbose);
+#endif
+
+void elf_parse_binary(struct elf_binary *elf);
+void elf_load_binary(struct elf_binary *elf);
+
+void *elf_get_ptr(struct elf_binary *elf, unsigned long addr);
+uint64_t elf_lookup_addr(struct elf_binary *elf, const char *symbol);
+
+void elf_parse_bsdsyms(struct elf_binary *elf, uint64_t pstart); /* private */
+
+/* ------------------------------------------------------------------------ */
+/* xc_libelf_relocate.c                                                     */
+
+int elf_reloc(struct elf_binary *elf);
+
+/* ------------------------------------------------------------------------ */
+/* xc_libelf_dominfo.c                                                      */
+
+#define UNSET_ADDR          ((uint64_t)-1)
+
+enum xen_elfnote_type {
+    XEN_ENT_NONE = 0,
+    XEN_ENT_LONG = 1,
+    XEN_ENT_STR  = 2
+};
+
+struct xen_elfnote {
+    enum xen_elfnote_type type;
+    const char *name;
+    union {
+        const char *str;
+        uint64_t num;
+    } data;
+};
+
+struct elf_dom_parms {
+    /* raw */
+    const char *guest_info;
+    const void *elf_note_start;
+    const void *elf_note_end;
+    struct xen_elfnote elf_notes[XEN_ELFNOTE_MAX + 1];
+  
+    /* parsed */
+    char guest_os[16];
+    char guest_ver[16];
+    char xen_ver[16];
+    char loader[16];
+    int pae;
+    int bsd_symtab;
+    uint64_t virt_base;
+    uint64_t virt_entry;
+    uint64_t virt_hypercall;
+    uint64_t virt_hv_start_low;
+    uint64_t elf_paddr_offset;
+    uint32_t f_supported[XENFEAT_NR_SUBMAPS];
+    uint32_t f_required[XENFEAT_NR_SUBMAPS];
+
+    /* calculated */
+    uint64_t virt_offset;
+    uint64_t virt_kstart;
+    uint64_t virt_kend;
+};
+
+static inline void elf_xen_feature_set(int nr, uint32_t * addr)
+{
+    addr[nr >> 5] |= 1 << (nr & 31);
+}
+static inline int elf_xen_feature_get(int nr, uint32_t * addr)
+{
+    return !!(addr[nr >> 5] & (1 << (nr & 31)));
+}
+
+int elf_xen_parse_features(const char *features,
+                           uint32_t *supported,
+                           uint32_t *required);
+int elf_xen_parse_note(struct elf_binary *elf,
+                       struct elf_dom_parms *parms,
+                       const elf_note *note);
+int elf_xen_parse_guest_info(struct elf_binary *elf,
+                             struct elf_dom_parms *parms);
+int elf_xen_parse(struct elf_binary *elf,
+                  struct elf_dom_parms *parms);
+
+#endif /* __XC_LIBELF__ */
diff -rpuN linux-2.6.18.8/include/xen/interface/memory.h linux-2.6.18-xen-3.3.0/include/xen/interface/memory.h
--- linux-2.6.18.8/include/xen/interface/memory.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/memory.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,295 @@
+/******************************************************************************
+ * memory.h
+ * 
+ * Memory reservation and information.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2005, Keir Fraser <keir@xensource.com>
+ */
+
+#ifndef __XEN_PUBLIC_MEMORY_H__
+#define __XEN_PUBLIC_MEMORY_H__
+
+/*
+ * Increase or decrease the specified domain's memory reservation. Returns the
+ * number of extents successfully allocated or freed.
+ * arg == addr of struct xen_memory_reservation.
+ */
+#define XENMEM_increase_reservation 0
+#define XENMEM_decrease_reservation 1
+#define XENMEM_populate_physmap     6
+
+#if __XEN_INTERFACE_VERSION__ >= 0x00030209
+/*
+ * Maximum # bits addressable by the user of the allocated region (e.g., I/O 
+ * devices often have a 32-bit limitation even in 64-bit systems). If zero 
+ * then the user has no addressing restriction. This field is not used by 
+ * XENMEM_decrease_reservation.
+ */
+#define XENMEMF_address_bits(x)     (x)
+#define XENMEMF_get_address_bits(x) ((x) & 0xffu)
+/* NUMA node to allocate from. */
+#define XENMEMF_node(x)     (((x) + 1) << 8)
+#define XENMEMF_get_node(x) ((((x) >> 8) - 1) & 0xffu)
+#endif
+
+struct xen_memory_reservation {
+
+    /*
+     * XENMEM_increase_reservation:
+     *   OUT: MFN (*not* GMFN) bases of extents that were allocated
+     * XENMEM_decrease_reservation:
+     *   IN:  GMFN bases of extents to free
+     * XENMEM_populate_physmap:
+     *   IN:  GPFN bases of extents to populate with memory
+     *   OUT: GMFN bases of extents that were allocated
+     *   (NB. This command also updates the mach_to_phys translation table)
+     */
+    XEN_GUEST_HANDLE(xen_pfn_t) extent_start;
+
+    /* Number of extents, and size/alignment of each (2^extent_order pages). */
+    xen_ulong_t    nr_extents;
+    unsigned int   extent_order;
+
+#if __XEN_INTERFACE_VERSION__ >= 0x00030209
+    /* XENMEMF flags. */
+    unsigned int   mem_flags;
+#else
+    unsigned int   address_bits;
+#endif
+
+    /*
+     * Domain whose reservation is being changed.
+     * Unprivileged domains can specify only DOMID_SELF.
+     */
+    domid_t        domid;
+};
+typedef struct xen_memory_reservation xen_memory_reservation_t;
+DEFINE_XEN_GUEST_HANDLE(xen_memory_reservation_t);
+
+/*
+ * An atomic exchange of memory pages. If return code is zero then
+ * @out.extent_list provides GMFNs of the newly-allocated memory.
+ * Returns zero on complete success, otherwise a negative error code.
+ * On complete success then always @nr_exchanged == @in.nr_extents.
+ * On partial success @nr_exchanged indicates how much work was done.
+ */
+#define XENMEM_exchange             11
+struct xen_memory_exchange {
+    /*
+     * [IN] Details of memory extents to be exchanged (GMFN bases).
+     * Note that @in.address_bits is ignored and unused.
+     */
+    struct xen_memory_reservation in;
+
+    /*
+     * [IN/OUT] Details of new memory extents.
+     * We require that:
+     *  1. @in.domid == @out.domid
+     *  2. @in.nr_extents  << @in.extent_order == 
+     *     @out.nr_extents << @out.extent_order
+     *  3. @in.extent_start and @out.extent_start lists must not overlap
+     *  4. @out.extent_start lists GPFN bases to be populated
+     *  5. @out.extent_start is overwritten with allocated GMFN bases
+     */
+    struct xen_memory_reservation out;
+
+    /*
+     * [OUT] Number of input extents that were successfully exchanged:
+     *  1. The first @nr_exchanged input extents were successfully
+     *     deallocated.
+     *  2. The corresponding first entries in the output extent list correctly
+     *     indicate the GMFNs that were successfully exchanged.
+     *  3. All other input and output extents are untouched.
+     *  4. If not all input exents are exchanged then the return code of this
+     *     command will be non-zero.
+     *  5. THIS FIELD MUST BE INITIALISED TO ZERO BY THE CALLER!
+     */
+    xen_ulong_t nr_exchanged;
+};
+typedef struct xen_memory_exchange xen_memory_exchange_t;
+DEFINE_XEN_GUEST_HANDLE(xen_memory_exchange_t);
+
+/*
+ * Returns the maximum machine frame number of mapped RAM in this system.
+ * This command always succeeds (it never returns an error code).
+ * arg == NULL.
+ */
+#define XENMEM_maximum_ram_page     2
+
+/*
+ * Returns the current or maximum memory reservation, in pages, of the
+ * specified domain (may be DOMID_SELF). Returns -ve errcode on failure.
+ * arg == addr of domid_t.
+ */
+#define XENMEM_current_reservation  3
+#define XENMEM_maximum_reservation  4
+
+/*
+ * Returns the maximum GPFN in use by the guest, or -ve errcode on failure.
+ */
+#define XENMEM_maximum_gpfn         14
+
+/*
+ * Returns a list of MFN bases of 2MB extents comprising the machine_to_phys
+ * mapping table. Architectures which do not have a m2p table do not implement
+ * this command.
+ * arg == addr of xen_machphys_mfn_list_t.
+ */
+#define XENMEM_machphys_mfn_list    5
+struct xen_machphys_mfn_list {
+    /*
+     * Size of the 'extent_start' array. Fewer entries will be filled if the
+     * machphys table is smaller than max_extents * 2MB.
+     */
+    unsigned int max_extents;
+
+    /*
+     * Pointer to buffer to fill with list of extent starts. If there are
+     * any large discontiguities in the machine address space, 2MB gaps in
+     * the machphys table will be represented by an MFN base of zero.
+     */
+    XEN_GUEST_HANDLE(xen_pfn_t) extent_start;
+
+    /*
+     * Number of extents written to the above array. This will be smaller
+     * than 'max_extents' if the machphys table is smaller than max_e * 2MB.
+     */
+    unsigned int nr_extents;
+};
+typedef struct xen_machphys_mfn_list xen_machphys_mfn_list_t;
+DEFINE_XEN_GUEST_HANDLE(xen_machphys_mfn_list_t);
+
+/*
+ * Returns the location in virtual address space of the machine_to_phys
+ * mapping table. Architectures which do not have a m2p table, or which do not
+ * map it by default into guest address space, do not implement this command.
+ * arg == addr of xen_machphys_mapping_t.
+ */
+#define XENMEM_machphys_mapping     12
+struct xen_machphys_mapping {
+    xen_ulong_t v_start, v_end; /* Start and end virtual addresses.   */
+    xen_ulong_t max_mfn;        /* Maximum MFN that can be looked up. */
+};
+typedef struct xen_machphys_mapping xen_machphys_mapping_t;
+DEFINE_XEN_GUEST_HANDLE(xen_machphys_mapping_t);
+
+/*
+ * Sets the GPFN at which a particular page appears in the specified guest's
+ * pseudophysical address space.
+ * arg == addr of xen_add_to_physmap_t.
+ */
+#define XENMEM_add_to_physmap      7
+struct xen_add_to_physmap {
+    /* Which domain to change the mapping for. */
+    domid_t domid;
+
+    /* Source mapping space. */
+#define XENMAPSPACE_shared_info 0 /* shared info page */
+#define XENMAPSPACE_grant_table 1 /* grant table page */
+    unsigned int space;
+
+    /* Index into source mapping space. */
+    xen_ulong_t idx;
+
+    /* GPFN where the source mapping page should appear. */
+    xen_pfn_t     gpfn;
+};
+typedef struct xen_add_to_physmap xen_add_to_physmap_t;
+DEFINE_XEN_GUEST_HANDLE(xen_add_to_physmap_t);
+
+/*
+ * Translates a list of domain-specific GPFNs into MFNs. Returns a -ve error
+ * code on failure. This call only works for auto-translated guests.
+ */
+#define XENMEM_translate_gpfn_list  8
+struct xen_translate_gpfn_list {
+    /* Which domain to translate for? */
+    domid_t domid;
+
+    /* Length of list. */
+    xen_ulong_t nr_gpfns;
+
+    /* List of GPFNs to translate. */
+    XEN_GUEST_HANDLE(xen_pfn_t) gpfn_list;
+
+    /*
+     * Output list to contain MFN translations. May be the same as the input
+     * list (in which case each input GPFN is overwritten with the output MFN).
+     */
+    XEN_GUEST_HANDLE(xen_pfn_t) mfn_list;
+};
+typedef struct xen_translate_gpfn_list xen_translate_gpfn_list_t;
+DEFINE_XEN_GUEST_HANDLE(xen_translate_gpfn_list_t);
+
+/*
+ * Returns the pseudo-physical memory map as it was when the domain
+ * was started (specified by XENMEM_set_memory_map).
+ * arg == addr of xen_memory_map_t.
+ */
+#define XENMEM_memory_map           9
+struct xen_memory_map {
+    /*
+     * On call the number of entries which can be stored in buffer. On
+     * return the number of entries which have been stored in
+     * buffer.
+     */
+    unsigned int nr_entries;
+
+    /*
+     * Entries in the buffer are in the same format as returned by the
+     * BIOS INT 0x15 EAX=0xE820 call.
+     */
+    XEN_GUEST_HANDLE(void) buffer;
+};
+typedef struct xen_memory_map xen_memory_map_t;
+DEFINE_XEN_GUEST_HANDLE(xen_memory_map_t);
+
+/*
+ * Returns the real physical memory map. Passes the same structure as
+ * XENMEM_memory_map.
+ * arg == addr of xen_memory_map_t.
+ */
+#define XENMEM_machine_memory_map   10
+
+/*
+ * Set the pseudo-physical memory map of a domain, as returned by
+ * XENMEM_memory_map.
+ * arg == addr of xen_foreign_memory_map_t.
+ */
+#define XENMEM_set_memory_map       13
+struct xen_foreign_memory_map {
+    domid_t domid;
+    struct xen_memory_map map;
+};
+typedef struct xen_foreign_memory_map xen_foreign_memory_map_t;
+DEFINE_XEN_GUEST_HANDLE(xen_foreign_memory_map_t);
+
+#endif /* __XEN_PUBLIC_MEMORY_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/nmi.h linux-2.6.18-xen-3.3.0/include/xen/interface/nmi.h
--- linux-2.6.18.8/include/xen/interface/nmi.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/nmi.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,78 @@
+/******************************************************************************
+ * nmi.h
+ * 
+ * NMI callback registration and reason codes.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2005, Keir Fraser <keir@xensource.com>
+ */
+
+#ifndef __XEN_PUBLIC_NMI_H__
+#define __XEN_PUBLIC_NMI_H__
+
+/*
+ * NMI reason codes:
+ * Currently these are x86-specific, stored in arch_shared_info.nmi_reason.
+ */
+ /* I/O-check error reported via ISA port 0x61, bit 6. */
+#define _XEN_NMIREASON_io_error     0
+#define XEN_NMIREASON_io_error      (1UL << _XEN_NMIREASON_io_error)
+ /* Parity error reported via ISA port 0x61, bit 7. */
+#define _XEN_NMIREASON_parity_error 1
+#define XEN_NMIREASON_parity_error  (1UL << _XEN_NMIREASON_parity_error)
+ /* Unknown hardware-generated NMI. */
+#define _XEN_NMIREASON_unknown      2
+#define XEN_NMIREASON_unknown       (1UL << _XEN_NMIREASON_unknown)
+
+/*
+ * long nmi_op(unsigned int cmd, void *arg)
+ * NB. All ops return zero on success, else a negative error code.
+ */
+
+/*
+ * Register NMI callback for this (calling) VCPU. Currently this only makes
+ * sense for domain 0, vcpu 0. All other callers will be returned EINVAL.
+ * arg == pointer to xennmi_callback structure.
+ */
+#define XENNMI_register_callback   0
+struct xennmi_callback {
+    unsigned long handler_address;
+    unsigned long pad;
+};
+typedef struct xennmi_callback xennmi_callback_t;
+DEFINE_XEN_GUEST_HANDLE(xennmi_callback_t);
+
+/*
+ * Deregister NMI callback for this (calling) VCPU.
+ * arg == NULL.
+ */
+#define XENNMI_unregister_callback 1
+
+#endif /* __XEN_PUBLIC_NMI_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/physdev.h linux-2.6.18-xen-3.3.0/include/xen/interface/physdev.h
--- linux-2.6.18.8/include/xen/interface/physdev.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/physdev.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,219 @@
+/*
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_PUBLIC_PHYSDEV_H__
+#define __XEN_PUBLIC_PHYSDEV_H__
+
+/*
+ * Prototype for this hypercall is:
+ *  int physdev_op(int cmd, void *args)
+ * @cmd  == PHYSDEVOP_??? (physdev operation).
+ * @args == Operation-specific extra arguments (NULL if none).
+ */
+
+/*
+ * Notify end-of-interrupt (EOI) for the specified IRQ.
+ * @arg == pointer to physdev_eoi structure.
+ */
+#define PHYSDEVOP_eoi                   12
+struct physdev_eoi {
+    /* IN */
+    uint32_t irq;
+};
+typedef struct physdev_eoi physdev_eoi_t;
+DEFINE_XEN_GUEST_HANDLE(physdev_eoi_t);
+
+/*
+ * Query the status of an IRQ line.
+ * @arg == pointer to physdev_irq_status_query structure.
+ */
+#define PHYSDEVOP_irq_status_query       5
+struct physdev_irq_status_query {
+    /* IN */
+    uint32_t irq;
+    /* OUT */
+    uint32_t flags; /* XENIRQSTAT_* */
+};
+typedef struct physdev_irq_status_query physdev_irq_status_query_t;
+DEFINE_XEN_GUEST_HANDLE(physdev_irq_status_query_t);
+
+/* Need to call PHYSDEVOP_eoi when the IRQ has been serviced? */
+#define _XENIRQSTAT_needs_eoi   (0)
+#define  XENIRQSTAT_needs_eoi   (1U<<_XENIRQSTAT_needs_eoi)
+
+/* IRQ shared by multiple guests? */
+#define _XENIRQSTAT_shared      (1)
+#define  XENIRQSTAT_shared      (1U<<_XENIRQSTAT_shared)
+
+/*
+ * Set the current VCPU's I/O privilege level.
+ * @arg == pointer to physdev_set_iopl structure.
+ */
+#define PHYSDEVOP_set_iopl               6
+struct physdev_set_iopl {
+    /* IN */
+    uint32_t iopl;
+};
+typedef struct physdev_set_iopl physdev_set_iopl_t;
+DEFINE_XEN_GUEST_HANDLE(physdev_set_iopl_t);
+
+/*
+ * Set the current VCPU's I/O-port permissions bitmap.
+ * @arg == pointer to physdev_set_iobitmap structure.
+ */
+#define PHYSDEVOP_set_iobitmap           7
+struct physdev_set_iobitmap {
+    /* IN */
+#if __XEN_INTERFACE_VERSION__ >= 0x00030205
+    XEN_GUEST_HANDLE(uint8) bitmap;
+#else
+    uint8_t *bitmap;
+#endif
+    uint32_t nr_ports;
+};
+typedef struct physdev_set_iobitmap physdev_set_iobitmap_t;
+DEFINE_XEN_GUEST_HANDLE(physdev_set_iobitmap_t);
+
+/*
+ * Read or write an IO-APIC register.
+ * @arg == pointer to physdev_apic structure.
+ */
+#define PHYSDEVOP_apic_read              8
+#define PHYSDEVOP_apic_write             9
+struct physdev_apic {
+    /* IN */
+    unsigned long apic_physbase;
+    uint32_t reg;
+    /* IN or OUT */
+    uint32_t value;
+};
+typedef struct physdev_apic physdev_apic_t;
+DEFINE_XEN_GUEST_HANDLE(physdev_apic_t);
+
+/*
+ * Allocate or free a physical upcall vector for the specified IRQ line.
+ * @arg == pointer to physdev_irq structure.
+ */
+#define PHYSDEVOP_alloc_irq_vector      10
+#define PHYSDEVOP_free_irq_vector       11
+struct physdev_irq {
+    /* IN */
+    uint32_t irq;
+    /* IN or OUT */
+    uint32_t vector;
+};
+typedef struct physdev_irq physdev_irq_t;
+DEFINE_XEN_GUEST_HANDLE(physdev_irq_t);
+ 
+#define MAP_PIRQ_TYPE_MSI               0x0
+#define MAP_PIRQ_TYPE_GSI               0x1
+#define MAP_PIRQ_TYPE_UNKNOWN           0x2
+
+#define PHYSDEVOP_map_pirq               13
+struct physdev_map_pirq {
+    domid_t domid;
+    /* IN */
+    int type;
+    /* IN */
+    int index;
+    /* IN or OUT */
+    int pirq;
+    /* IN */
+    int bus;
+    /* IN */
+    int devfn;
+    /* IN */
+    int entry_nr;
+    /* IN */
+    uint64_t table_base;
+};
+typedef struct physdev_map_pirq physdev_map_pirq_t;
+DEFINE_XEN_GUEST_HANDLE(physdev_map_pirq_t);
+
+#define PHYSDEVOP_unmap_pirq             14
+struct physdev_unmap_pirq {
+    domid_t domid;
+    /* IN */
+    int pirq;
+};
+
+typedef struct physdev_unmap_pirq physdev_unmap_pirq_t;
+DEFINE_XEN_GUEST_HANDLE(physdev_unmap_pirq_t);
+
+#define PHYSDEVOP_manage_pci_add         15
+#define PHYSDEVOP_manage_pci_remove      16
+struct physdev_manage_pci {
+    /* IN */
+    uint8_t bus;
+    uint8_t devfn;
+}; 
+
+typedef struct physdev_manage_pci physdev_manage_pci_t;
+DEFINE_XEN_GUEST_HANDLE(physdev_manage_pci_t);
+
+/*
+ * Argument to physdev_op_compat() hypercall. Superceded by new physdev_op()
+ * hypercall since 0x00030202.
+ */
+struct physdev_op {
+    uint32_t cmd;
+    union {
+        struct physdev_irq_status_query      irq_status_query;
+        struct physdev_set_iopl              set_iopl;
+        struct physdev_set_iobitmap          set_iobitmap;
+        struct physdev_apic                  apic_op;
+        struct physdev_irq                   irq_op;
+    } u;
+};
+typedef struct physdev_op physdev_op_t;
+DEFINE_XEN_GUEST_HANDLE(physdev_op_t);
+
+/*
+ * Notify that some PIRQ-bound event channels have been unmasked.
+ * ** This command is obsolete since interface version 0x00030202 and is **
+ * ** unsupported by newer versions of Xen.                              **
+ */
+#define PHYSDEVOP_IRQ_UNMASK_NOTIFY      4
+
+/*
+ * These all-capitals physdev operation names are superceded by the new names
+ * (defined above) since interface version 0x00030202.
+ */
+#define PHYSDEVOP_IRQ_STATUS_QUERY       PHYSDEVOP_irq_status_query
+#define PHYSDEVOP_SET_IOPL               PHYSDEVOP_set_iopl
+#define PHYSDEVOP_SET_IOBITMAP           PHYSDEVOP_set_iobitmap
+#define PHYSDEVOP_APIC_READ              PHYSDEVOP_apic_read
+#define PHYSDEVOP_APIC_WRITE             PHYSDEVOP_apic_write
+#define PHYSDEVOP_ASSIGN_VECTOR          PHYSDEVOP_alloc_irq_vector
+#define PHYSDEVOP_FREE_VECTOR            PHYSDEVOP_free_irq_vector
+#define PHYSDEVOP_IRQ_NEEDS_UNMASK_NOTIFY XENIRQSTAT_needs_eoi
+#define PHYSDEVOP_IRQ_SHARED             XENIRQSTAT_shared
+
+#endif /* __XEN_PUBLIC_PHYSDEV_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/platform.h linux-2.6.18-xen-3.3.0/include/xen/interface/platform.h
--- linux-2.6.18.8/include/xen/interface/platform.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/platform.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,346 @@
+/******************************************************************************
+ * platform.h
+ * 
+ * Hardware platform operations. Intended for use by domain-0 kernel.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2002-2006, K Fraser
+ */
+
+#ifndef __XEN_PUBLIC_PLATFORM_H__
+#define __XEN_PUBLIC_PLATFORM_H__
+
+#include "xen.h"
+
+#define XENPF_INTERFACE_VERSION 0x03000001
+
+/*
+ * Set clock such that it would read <secs,nsecs> after 00:00:00 UTC,
+ * 1 January, 1970 if the current system time was <system_time>.
+ */
+#define XENPF_settime             17
+struct xenpf_settime {
+    /* IN variables. */
+    uint32_t secs;
+    uint32_t nsecs;
+    uint64_t system_time;
+};
+typedef struct xenpf_settime xenpf_settime_t;
+DEFINE_XEN_GUEST_HANDLE(xenpf_settime_t);
+
+/*
+ * Request memory range (@mfn, @mfn+@nr_mfns-1) to have type @type.
+ * On x86, @type is an architecture-defined MTRR memory type.
+ * On success, returns the MTRR that was used (@reg) and a handle that can
+ * be passed to XENPF_DEL_MEMTYPE to accurately tear down the new setting.
+ * (x86-specific).
+ */
+#define XENPF_add_memtype         31
+struct xenpf_add_memtype {
+    /* IN variables. */
+    xen_pfn_t mfn;
+    uint64_t nr_mfns;
+    uint32_t type;
+    /* OUT variables. */
+    uint32_t handle;
+    uint32_t reg;
+};
+typedef struct xenpf_add_memtype xenpf_add_memtype_t;
+DEFINE_XEN_GUEST_HANDLE(xenpf_add_memtype_t);
+
+/*
+ * Tear down an existing memory-range type. If @handle is remembered then it
+ * should be passed in to accurately tear down the correct setting (in case
+ * of overlapping memory regions with differing types). If it is not known
+ * then @handle should be set to zero. In all cases @reg must be set.
+ * (x86-specific).
+ */
+#define XENPF_del_memtype         32
+struct xenpf_del_memtype {
+    /* IN variables. */
+    uint32_t handle;
+    uint32_t reg;
+};
+typedef struct xenpf_del_memtype xenpf_del_memtype_t;
+DEFINE_XEN_GUEST_HANDLE(xenpf_del_memtype_t);
+
+/* Read current type of an MTRR (x86-specific). */
+#define XENPF_read_memtype        33
+struct xenpf_read_memtype {
+    /* IN variables. */
+    uint32_t reg;
+    /* OUT variables. */
+    xen_pfn_t mfn;
+    uint64_t nr_mfns;
+    uint32_t type;
+};
+typedef struct xenpf_read_memtype xenpf_read_memtype_t;
+DEFINE_XEN_GUEST_HANDLE(xenpf_read_memtype_t);
+
+#define XENPF_microcode_update    35
+struct xenpf_microcode_update {
+    /* IN variables. */
+    XEN_GUEST_HANDLE(void) data;      /* Pointer to microcode data */
+    uint32_t length;                  /* Length of microcode data. */
+};
+typedef struct xenpf_microcode_update xenpf_microcode_update_t;
+DEFINE_XEN_GUEST_HANDLE(xenpf_microcode_update_t);
+
+#define XENPF_platform_quirk      39
+#define QUIRK_NOIRQBALANCING      1 /* Do not restrict IO-APIC RTE targets */
+#define QUIRK_IOAPIC_BAD_REGSEL   2 /* IO-APIC REGSEL forgets its value    */
+#define QUIRK_IOAPIC_GOOD_REGSEL  3 /* IO-APIC REGSEL behaves properly     */
+struct xenpf_platform_quirk {
+    /* IN variables. */
+    uint32_t quirk_id;
+};
+typedef struct xenpf_platform_quirk xenpf_platform_quirk_t;
+DEFINE_XEN_GUEST_HANDLE(xenpf_platform_quirk_t);
+
+#define XENPF_firmware_info       50
+#define XEN_FW_DISK_INFO          1 /* from int 13 AH=08/41/48 */
+#define XEN_FW_DISK_MBR_SIGNATURE 2 /* from MBR offset 0x1b8 */
+#define XEN_FW_VBEDDC_INFO        3 /* from int 10 AX=4f15 */
+struct xenpf_firmware_info {
+    /* IN variables. */
+    uint32_t type;
+    uint32_t index;
+    /* OUT variables. */
+    union {
+        struct {
+            /* Int13, Fn48: Check Extensions Present. */
+            uint8_t device;                   /* %dl: bios device number */
+            uint8_t version;                  /* %ah: major version      */
+            uint16_t interface_support;       /* %cx: support bitmap     */
+            /* Int13, Fn08: Legacy Get Device Parameters. */
+            uint16_t legacy_max_cylinder;     /* %cl[7:6]:%ch: max cyl # */
+            uint8_t legacy_max_head;          /* %dh: max head #         */
+            uint8_t legacy_sectors_per_track; /* %cl[5:0]: max sector #  */
+            /* Int13, Fn41: Get Device Parameters (as filled into %ds:%esi). */
+            /* NB. First uint16_t of buffer must be set to buffer size.      */
+            XEN_GUEST_HANDLE(void) edd_params;
+        } disk_info; /* XEN_FW_DISK_INFO */
+        struct {
+            uint8_t device;                   /* bios device number  */
+            uint32_t mbr_signature;           /* offset 0x1b8 in mbr */
+        } disk_mbr_signature; /* XEN_FW_DISK_MBR_SIGNATURE */
+        struct {
+            /* Int10, AX=4F15: Get EDID info. */
+            uint8_t capabilities;
+            uint8_t edid_transfer_time;
+            /* must refer to 128-byte buffer */
+            XEN_GUEST_HANDLE(uint8) edid;
+        } vbeddc_info; /* XEN_FW_VBEDDC_INFO */
+    } u;
+};
+typedef struct xenpf_firmware_info xenpf_firmware_info_t;
+DEFINE_XEN_GUEST_HANDLE(xenpf_firmware_info_t);
+
+#define XENPF_enter_acpi_sleep    51
+struct xenpf_enter_acpi_sleep {
+    /* IN variables */
+    uint16_t pm1a_cnt_val;      /* PM1a control value. */
+    uint16_t pm1b_cnt_val;      /* PM1b control value. */
+    uint32_t sleep_state;       /* Which state to enter (Sn). */
+    uint32_t flags;             /* Must be zero. */
+};
+typedef struct xenpf_enter_acpi_sleep xenpf_enter_acpi_sleep_t;
+DEFINE_XEN_GUEST_HANDLE(xenpf_enter_acpi_sleep_t);
+
+#define XENPF_change_freq         52
+struct xenpf_change_freq {
+    /* IN variables */
+    uint32_t flags; /* Must be zero. */
+    uint32_t cpu;   /* Physical cpu. */
+    uint64_t freq;  /* New frequency (Hz). */
+};
+typedef struct xenpf_change_freq xenpf_change_freq_t;
+DEFINE_XEN_GUEST_HANDLE(xenpf_change_freq_t);
+
+/*
+ * Get idle times (nanoseconds since boot) for physical CPUs specified in the
+ * @cpumap_bitmap with range [0..@cpumap_nr_cpus-1]. The @idletime array is
+ * indexed by CPU number; only entries with the corresponding @cpumap_bitmap
+ * bit set are written to. On return, @cpumap_bitmap is modified so that any
+ * non-existent CPUs are cleared. Such CPUs have their @idletime array entry
+ * cleared.
+ */
+#define XENPF_getidletime         53
+struct xenpf_getidletime {
+    /* IN/OUT variables */
+    /* IN: CPUs to interrogate; OUT: subset of IN which are present */
+    XEN_GUEST_HANDLE(uint8) cpumap_bitmap;
+    /* IN variables */
+    /* Size of cpumap bitmap. */
+    uint32_t cpumap_nr_cpus;
+    /* Must be indexable for every cpu in cpumap_bitmap. */
+    XEN_GUEST_HANDLE(uint64) idletime;
+    /* OUT variables */
+    /* System time when the idletime snapshots were taken. */
+    uint64_t now;
+};
+typedef struct xenpf_getidletime xenpf_getidletime_t;
+DEFINE_XEN_GUEST_HANDLE(xenpf_getidletime_t);
+
+#define XENPF_set_processor_pminfo      54
+
+/* ability bits */
+#define XEN_PROCESSOR_PM_CX	1
+#define XEN_PROCESSOR_PM_PX	2
+#define XEN_PROCESSOR_PM_TX	4
+
+/* cmd type */
+#define XEN_PM_CX   0
+#define XEN_PM_PX   1
+#define XEN_PM_TX   2
+
+/* Px sub info type */
+#define XEN_PX_PCT   1
+#define XEN_PX_PSS   2
+#define XEN_PX_PPC   4
+#define XEN_PX_PSD   8
+
+struct xen_power_register {
+    uint32_t     space_id;
+    uint32_t     bit_width;
+    uint32_t     bit_offset;
+    uint32_t     access_size;
+    uint64_t     address;
+};
+
+struct xen_processor_csd {
+    uint32_t    domain;      /* domain number of one dependent group */
+    uint32_t    coord_type;  /* coordination type */
+    uint32_t    num;         /* number of processors in same domain */
+};
+typedef struct xen_processor_csd xen_processor_csd_t;
+DEFINE_XEN_GUEST_HANDLE(xen_processor_csd_t);
+
+struct xen_processor_cx {
+    struct xen_power_register  reg; /* GAS for Cx trigger register */
+    uint8_t     type;     /* cstate value, c0: 0, c1: 1, ... */
+    uint32_t    latency;  /* worst latency (ms) to enter/exit this cstate */
+    uint32_t    power;    /* average power consumption(mW) */
+    uint32_t    dpcnt;    /* number of dependency entries */
+    XEN_GUEST_HANDLE(xen_processor_csd_t) dp; /* NULL if no dependency */
+};
+typedef struct xen_processor_cx xen_processor_cx_t;
+DEFINE_XEN_GUEST_HANDLE(xen_processor_cx_t);
+
+struct xen_processor_flags {
+    uint32_t bm_control:1;
+    uint32_t bm_check:1;
+    uint32_t has_cst:1;
+    uint32_t power_setup_done:1;
+    uint32_t bm_rld_set:1;
+};
+
+struct xen_processor_power {
+    uint32_t count;  /* number of C state entries in array below */
+    struct xen_processor_flags flags;  /* global flags of this processor */
+    XEN_GUEST_HANDLE(xen_processor_cx_t) states; /* supported c states */
+};
+
+struct xen_pct_register {
+    uint8_t  descriptor;
+    uint16_t length;
+    uint8_t  space_id;
+    uint8_t  bit_width;
+    uint8_t  bit_offset;
+    uint8_t  reserved;
+    uint64_t address;
+};
+
+struct xen_processor_px {
+    uint64_t core_frequency; /* megahertz */
+    uint64_t power;      /* milliWatts */
+    uint64_t transition_latency; /* microseconds */
+    uint64_t bus_master_latency; /* microseconds */
+    uint64_t control;        /* control value */
+    uint64_t status;     /* success indicator */
+};
+typedef struct xen_processor_px xen_processor_px_t;
+DEFINE_XEN_GUEST_HANDLE(xen_processor_px_t);
+
+struct xen_psd_package {
+    uint64_t num_entries;
+    uint64_t revision;
+    uint64_t domain;
+    uint64_t coord_type;
+    uint64_t num_processors;
+};
+
+struct xen_processor_performance {
+    uint32_t flags;     /* flag for Px sub info type */
+    uint32_t ppc;       /* Platform limitation on freq usage */
+    struct xen_pct_register control_register;
+    struct xen_pct_register status_register;
+    uint32_t state_count;     /* total available performance states */
+    XEN_GUEST_HANDLE(xen_processor_px_t) states;
+    struct xen_psd_package domain_info;
+    uint32_t shared_type;     /* coordination type of this processor */
+};
+typedef struct xen_processor_performance xen_processor_performance_t;
+DEFINE_XEN_GUEST_HANDLE(xen_processor_performance_t);
+
+struct xenpf_set_processor_pminfo {
+    /* IN variables */
+    uint32_t id;    /* ACPI CPU ID */
+    uint32_t type;  /* {XEN_PM_CX, XEN_PM_PX} */
+    union {
+        struct xen_processor_power          power;/* Cx: _CST/_CSD */
+        struct xen_processor_performance    perf; /* Px: _PPC/_PCT/_PSS/_PSD */
+    };
+};
+typedef struct xenpf_set_processor_pminfo xenpf_set_processor_pminfo_t;
+DEFINE_XEN_GUEST_HANDLE(xenpf_set_processor_pminfo_t);
+
+struct xen_platform_op {
+    uint32_t cmd;
+    uint32_t interface_version; /* XENPF_INTERFACE_VERSION */
+    union {
+        struct xenpf_settime           settime;
+        struct xenpf_add_memtype       add_memtype;
+        struct xenpf_del_memtype       del_memtype;
+        struct xenpf_read_memtype      read_memtype;
+        struct xenpf_microcode_update  microcode;
+        struct xenpf_platform_quirk    platform_quirk;
+        struct xenpf_firmware_info     firmware_info;
+        struct xenpf_enter_acpi_sleep  enter_acpi_sleep;
+        struct xenpf_change_freq       change_freq;
+        struct xenpf_getidletime       getidletime;
+        struct xenpf_set_processor_pminfo set_pminfo;
+        uint8_t                        pad[128];
+    } u;
+};
+typedef struct xen_platform_op xen_platform_op_t;
+DEFINE_XEN_GUEST_HANDLE(xen_platform_op_t);
+
+#endif /* __XEN_PUBLIC_PLATFORM_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/sched.h linux-2.6.18-xen-3.3.0/include/xen/interface/sched.h
--- linux-2.6.18.8/include/xen/interface/sched.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/sched.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,121 @@
+/******************************************************************************
+ * sched.h
+ * 
+ * Scheduler state interactions
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2005, Keir Fraser <keir@xensource.com>
+ */
+
+#ifndef __XEN_PUBLIC_SCHED_H__
+#define __XEN_PUBLIC_SCHED_H__
+
+#include "event_channel.h"
+
+/*
+ * The prototype for this hypercall is:
+ *  long sched_op(int cmd, void *arg)
+ * @cmd == SCHEDOP_??? (scheduler operation).
+ * @arg == Operation-specific extra argument(s), as described below.
+ * 
+ * Versions of Xen prior to 3.0.2 provided only the following legacy version
+ * of this hypercall, supporting only the commands yield, block and shutdown:
+ *  long sched_op(int cmd, unsigned long arg)
+ * @cmd == SCHEDOP_??? (scheduler operation).
+ * @arg == 0               (SCHEDOP_yield and SCHEDOP_block)
+ *      == SHUTDOWN_* code (SCHEDOP_shutdown)
+ * This legacy version is available to new guests as sched_op_compat().
+ */
+
+/*
+ * Voluntarily yield the CPU.
+ * @arg == NULL.
+ */
+#define SCHEDOP_yield       0
+
+/*
+ * Block execution of this VCPU until an event is received for processing.
+ * If called with event upcalls masked, this operation will atomically
+ * reenable event delivery and check for pending events before blocking the
+ * VCPU. This avoids a "wakeup waiting" race.
+ * @arg == NULL.
+ */
+#define SCHEDOP_block       1
+
+/*
+ * Halt execution of this domain (all VCPUs) and notify the system controller.
+ * @arg == pointer to sched_shutdown structure.
+ */
+#define SCHEDOP_shutdown    2
+struct sched_shutdown {
+    unsigned int reason; /* SHUTDOWN_* */
+};
+typedef struct sched_shutdown sched_shutdown_t;
+DEFINE_XEN_GUEST_HANDLE(sched_shutdown_t);
+
+/*
+ * Poll a set of event-channel ports. Return when one or more are pending. An
+ * optional timeout may be specified.
+ * @arg == pointer to sched_poll structure.
+ */
+#define SCHEDOP_poll        3
+struct sched_poll {
+    XEN_GUEST_HANDLE(evtchn_port_t) ports;
+    unsigned int nr_ports;
+    uint64_t timeout;
+};
+typedef struct sched_poll sched_poll_t;
+DEFINE_XEN_GUEST_HANDLE(sched_poll_t);
+
+/*
+ * Declare a shutdown for another domain. The main use of this function is
+ * in interpreting shutdown requests and reasons for fully-virtualized
+ * domains.  A para-virtualized domain may use SCHEDOP_shutdown directly.
+ * @arg == pointer to sched_remote_shutdown structure.
+ */
+#define SCHEDOP_remote_shutdown        4
+struct sched_remote_shutdown {
+    domid_t domain_id;         /* Remote domain ID */
+    unsigned int reason;       /* SHUTDOWN_xxx reason */
+};
+typedef struct sched_remote_shutdown sched_remote_shutdown_t;
+DEFINE_XEN_GUEST_HANDLE(sched_remote_shutdown_t);
+
+/*
+ * Reason codes for SCHEDOP_shutdown. These may be interpreted by control
+ * software to determine the appropriate action. For the most part, Xen does
+ * not care about the shutdown code.
+ */
+#define SHUTDOWN_poweroff   0  /* Domain exited normally. Clean up and kill. */
+#define SHUTDOWN_reboot     1  /* Clean up, kill, and then restart.          */
+#define SHUTDOWN_suspend    2  /* Clean up, save suspend info, kill.         */
+#define SHUTDOWN_crash      3  /* Tell controller we've crashed.             */
+
+#endif /* __XEN_PUBLIC_SCHED_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/sysctl.h linux-2.6.18-xen-3.3.0/include/xen/interface/sysctl.h
--- linux-2.6.18.8/include/xen/interface/sysctl.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/sysctl.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,295 @@
+/******************************************************************************
+ * sysctl.h
+ * 
+ * System management operations. For use by node control stack.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2002-2006, K Fraser
+ */
+
+#ifndef __XEN_PUBLIC_SYSCTL_H__
+#define __XEN_PUBLIC_SYSCTL_H__
+
+#if !defined(__XEN__) && !defined(__XEN_TOOLS__)
+#error "sysctl operations are intended for use by node control tools only"
+#endif
+
+#include "xen.h"
+#include "domctl.h"
+
+#define XEN_SYSCTL_INTERFACE_VERSION 0x00000006
+
+/*
+ * Read console content from Xen buffer ring.
+ */
+#define XEN_SYSCTL_readconsole       1
+struct xen_sysctl_readconsole {
+    /* IN: Non-zero -> clear after reading. */
+    uint8_t clear;
+    /* IN: Non-zero -> start index specified by @index field. */
+    uint8_t incremental;
+    uint8_t pad0, pad1;
+    /*
+     * IN:  Start index for consuming from ring buffer (if @incremental);
+     * OUT: End index after consuming from ring buffer.
+     */
+    uint32_t index; 
+    /* IN: Virtual address to write console data. */
+    XEN_GUEST_HANDLE_64(char) buffer;
+    /* IN: Size of buffer; OUT: Bytes written to buffer. */
+    uint32_t count;
+};
+typedef struct xen_sysctl_readconsole xen_sysctl_readconsole_t;
+DEFINE_XEN_GUEST_HANDLE(xen_sysctl_readconsole_t);
+
+/* Get trace buffers machine base address */
+#define XEN_SYSCTL_tbuf_op           2
+struct xen_sysctl_tbuf_op {
+    /* IN variables */
+#define XEN_SYSCTL_TBUFOP_get_info     0
+#define XEN_SYSCTL_TBUFOP_set_cpu_mask 1
+#define XEN_SYSCTL_TBUFOP_set_evt_mask 2
+#define XEN_SYSCTL_TBUFOP_set_size     3
+#define XEN_SYSCTL_TBUFOP_enable       4
+#define XEN_SYSCTL_TBUFOP_disable      5
+    uint32_t cmd;
+    /* IN/OUT variables */
+    struct xenctl_cpumap cpu_mask;
+    uint32_t             evt_mask;
+    /* OUT variables */
+    uint64_aligned_t buffer_mfn;
+    uint32_t size;
+};
+typedef struct xen_sysctl_tbuf_op xen_sysctl_tbuf_op_t;
+DEFINE_XEN_GUEST_HANDLE(xen_sysctl_tbuf_op_t);
+
+/*
+ * Get physical information about the host machine
+ */
+#define XEN_SYSCTL_physinfo          3
+ /* (x86) The platform supports HVM guests. */
+#define _XEN_SYSCTL_PHYSCAP_hvm          0
+#define XEN_SYSCTL_PHYSCAP_hvm           (1u<<_XEN_SYSCTL_PHYSCAP_hvm)
+ /* (x86) The platform supports HVM-guest direct access to I/O devices. */
+#define _XEN_SYSCTL_PHYSCAP_hvm_directio 1
+#define XEN_SYSCTL_PHYSCAP_hvm_directio  (1u<<_XEN_SYSCTL_PHYSCAP_hvm_directio)
+struct xen_sysctl_physinfo {
+    uint32_t threads_per_core;
+    uint32_t cores_per_socket;
+    uint32_t nr_cpus;
+    uint32_t nr_nodes;
+    uint32_t cpu_khz;
+    uint64_aligned_t total_pages;
+    uint64_aligned_t free_pages;
+    uint64_aligned_t scrub_pages;
+    uint32_t hw_cap[8];
+
+    /*
+     * IN: maximum addressable entry in the caller-provided cpu_to_node array.
+     * OUT: largest cpu identifier in the system.
+     * If OUT is greater than IN then the cpu_to_node array is truncated!
+     */
+    uint32_t max_cpu_id;
+    /*
+     * If not NULL, this array is filled with node identifier for each cpu.
+     * If a cpu has no node information (e.g., cpu not present) then the
+     * sentinel value ~0u is written.
+     * The size of this array is specified by the caller in @max_cpu_id.
+     * If the actual @max_cpu_id is smaller than the array then the trailing
+     * elements of the array will not be written by the sysctl.
+     */
+    XEN_GUEST_HANDLE_64(uint32) cpu_to_node;
+
+    /* XEN_SYSCTL_PHYSCAP_??? */
+    uint32_t capabilities;
+};
+typedef struct xen_sysctl_physinfo xen_sysctl_physinfo_t;
+DEFINE_XEN_GUEST_HANDLE(xen_sysctl_physinfo_t);
+
+/*
+ * Get the ID of the current scheduler.
+ */
+#define XEN_SYSCTL_sched_id          4
+struct xen_sysctl_sched_id {
+    /* OUT variable */
+    uint32_t sched_id;
+};
+typedef struct xen_sysctl_sched_id xen_sysctl_sched_id_t;
+DEFINE_XEN_GUEST_HANDLE(xen_sysctl_sched_id_t);
+
+/* Interface for controlling Xen software performance counters. */
+#define XEN_SYSCTL_perfc_op          5
+/* Sub-operations: */
+#define XEN_SYSCTL_PERFCOP_reset 1   /* Reset all counters to zero. */
+#define XEN_SYSCTL_PERFCOP_query 2   /* Get perfctr information. */
+struct xen_sysctl_perfc_desc {
+    char         name[80];             /* name of perf counter */
+    uint32_t     nr_vals;              /* number of values for this counter */
+};
+typedef struct xen_sysctl_perfc_desc xen_sysctl_perfc_desc_t;
+DEFINE_XEN_GUEST_HANDLE(xen_sysctl_perfc_desc_t);
+typedef uint32_t xen_sysctl_perfc_val_t;
+DEFINE_XEN_GUEST_HANDLE(xen_sysctl_perfc_val_t);
+
+struct xen_sysctl_perfc_op {
+    /* IN variables. */
+    uint32_t       cmd;                /*  XEN_SYSCTL_PERFCOP_??? */
+    /* OUT variables. */
+    uint32_t       nr_counters;       /*  number of counters description  */
+    uint32_t       nr_vals;           /*  number of values  */
+    /* counter information (or NULL) */
+    XEN_GUEST_HANDLE_64(xen_sysctl_perfc_desc_t) desc;
+    /* counter values (or NULL) */
+    XEN_GUEST_HANDLE_64(xen_sysctl_perfc_val_t) val;
+};
+typedef struct xen_sysctl_perfc_op xen_sysctl_perfc_op_t;
+DEFINE_XEN_GUEST_HANDLE(xen_sysctl_perfc_op_t);
+
+#define XEN_SYSCTL_getdomaininfolist 6
+struct xen_sysctl_getdomaininfolist {
+    /* IN variables. */
+    domid_t               first_domain;
+    uint32_t              max_domains;
+    XEN_GUEST_HANDLE_64(xen_domctl_getdomaininfo_t) buffer;
+    /* OUT variables. */
+    uint32_t              num_domains;
+};
+typedef struct xen_sysctl_getdomaininfolist xen_sysctl_getdomaininfolist_t;
+DEFINE_XEN_GUEST_HANDLE(xen_sysctl_getdomaininfolist_t);
+
+/* Inject debug keys into Xen. */
+#define XEN_SYSCTL_debug_keys        7
+struct xen_sysctl_debug_keys {
+    /* IN variables. */
+    XEN_GUEST_HANDLE_64(char) keys;
+    uint32_t nr_keys;
+};
+typedef struct xen_sysctl_debug_keys xen_sysctl_debug_keys_t;
+DEFINE_XEN_GUEST_HANDLE(xen_sysctl_debug_keys_t);
+
+/* Get physical CPU information. */
+#define XEN_SYSCTL_getcpuinfo        8
+struct xen_sysctl_cpuinfo {
+    uint64_aligned_t idletime;
+};
+typedef struct xen_sysctl_cpuinfo xen_sysctl_cpuinfo_t;
+DEFINE_XEN_GUEST_HANDLE(xen_sysctl_cpuinfo_t); 
+struct xen_sysctl_getcpuinfo {
+    /* IN variables. */
+    uint32_t max_cpus;
+    XEN_GUEST_HANDLE_64(xen_sysctl_cpuinfo_t) info;
+    /* OUT variables. */
+    uint32_t nr_cpus;
+}; 
+typedef struct xen_sysctl_getcpuinfo xen_sysctl_getcpuinfo_t;
+DEFINE_XEN_GUEST_HANDLE(xen_sysctl_getcpuinfo_t); 
+
+#define XEN_SYSCTL_availheap         9
+struct xen_sysctl_availheap {
+    /* IN variables. */
+    uint32_t min_bitwidth;  /* Smallest address width (zero if don't care). */
+    uint32_t max_bitwidth;  /* Largest address width (zero if don't care). */
+    int32_t  node;          /* NUMA node of interest (-1 for all nodes). */
+    /* OUT variables. */
+    uint64_aligned_t avail_bytes;/* Bytes available in the specified region. */
+};
+typedef struct xen_sysctl_availheap xen_sysctl_availheap_t;
+DEFINE_XEN_GUEST_HANDLE(xen_sysctl_availheap_t);
+
+#define XEN_SYSCTL_get_pmstat        10
+struct pm_px_val {
+    uint64_aligned_t freq;        /* Px core frequency */
+    uint64_aligned_t residency;   /* Px residency time */
+    uint64_aligned_t count;       /* Px transition count */
+};
+typedef struct pm_px_val pm_px_val_t;
+DEFINE_XEN_GUEST_HANDLE(pm_px_val_t);
+
+struct pm_px_stat {
+    uint8_t total;        /* total Px states */
+    uint8_t usable;       /* usable Px states */
+    uint8_t last;         /* last Px state */
+    uint8_t cur;          /* current Px state */
+    XEN_GUEST_HANDLE_64(uint64) trans_pt;   /* Px transition table */
+    XEN_GUEST_HANDLE_64(pm_px_val_t) pt;
+};
+typedef struct pm_px_stat pm_px_stat_t;
+DEFINE_XEN_GUEST_HANDLE(pm_px_stat_t);
+
+struct pm_cx_stat {
+    uint32_t nr;    /* entry nr in triggers & residencies, including C0 */
+    uint32_t last;  /* last Cx state */
+    uint64_aligned_t idle_time;                 /* idle time from boot */
+    XEN_GUEST_HANDLE_64(uint64) triggers;    /* Cx trigger counts */
+    XEN_GUEST_HANDLE_64(uint64) residencies; /* Cx residencies */
+};
+
+struct xen_sysctl_get_pmstat {
+#define PMSTAT_CATEGORY_MASK 0xf0
+#define PMSTAT_PX            0x10
+#define PMSTAT_CX            0x20
+#define PMSTAT_get_max_px    (PMSTAT_PX | 0x1)
+#define PMSTAT_get_pxstat    (PMSTAT_PX | 0x2)
+#define PMSTAT_reset_pxstat  (PMSTAT_PX | 0x3)
+#define PMSTAT_get_max_cx    (PMSTAT_CX | 0x1)
+#define PMSTAT_get_cxstat    (PMSTAT_CX | 0x2)
+#define PMSTAT_reset_cxstat  (PMSTAT_CX | 0x3)
+    uint32_t type;
+    uint32_t cpuid;
+    union {
+        struct pm_px_stat getpx;
+        struct pm_cx_stat getcx;
+        /* other struct for tx, etc */
+    } u;
+};
+typedef struct xen_sysctl_get_pmstat xen_sysctl_get_pmstat_t;
+DEFINE_XEN_GUEST_HANDLE(xen_sysctl_get_pmstat_t);
+
+struct xen_sysctl {
+    uint32_t cmd;
+    uint32_t interface_version; /* XEN_SYSCTL_INTERFACE_VERSION */
+    union {
+        struct xen_sysctl_readconsole       readconsole;
+        struct xen_sysctl_tbuf_op           tbuf_op;
+        struct xen_sysctl_physinfo          physinfo;
+        struct xen_sysctl_sched_id          sched_id;
+        struct xen_sysctl_perfc_op          perfc_op;
+        struct xen_sysctl_getdomaininfolist getdomaininfolist;
+        struct xen_sysctl_debug_keys        debug_keys;
+        struct xen_sysctl_getcpuinfo        getcpuinfo;
+        struct xen_sysctl_availheap         availheap;
+        struct xen_sysctl_get_pmstat        get_pmstat;
+        uint8_t                             pad[128];
+    } u;
+};
+typedef struct xen_sysctl xen_sysctl_t;
+DEFINE_XEN_GUEST_HANDLE(xen_sysctl_t);
+
+#endif /* __XEN_PUBLIC_SYSCTL_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/trace.h linux-2.6.18-xen-3.3.0/include/xen/interface/trace.h
--- linux-2.6.18.8/include/xen/interface/trace.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/trace.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,172 @@
+/******************************************************************************
+ * include/public/trace.h
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Mark Williamson, (C) 2004 Intel Research Cambridge
+ * Copyright (C) 2005 Bin Ren
+ */
+
+#ifndef __XEN_PUBLIC_TRACE_H__
+#define __XEN_PUBLIC_TRACE_H__
+
+#define TRACE_EXTRA_MAX    7
+#define TRACE_EXTRA_SHIFT 28
+
+/* Trace classes */
+#define TRC_CLS_SHIFT 16
+#define TRC_GEN      0x0001f000    /* General trace            */
+#define TRC_SCHED    0x0002f000    /* Xen Scheduler trace      */
+#define TRC_DOM0OP   0x0004f000    /* Xen DOM0 operation trace */
+#define TRC_HVM      0x0008f000    /* Xen HVM trace            */
+#define TRC_MEM      0x0010f000    /* Xen memory trace         */
+#define TRC_PV       0x0020f000    /* Xen PV traces            */
+#define TRC_ALL      0x0ffff000
+#define TRC_HD_TO_EVENT(x) ((x)&0x0fffffff)
+#define TRC_HD_CYCLE_FLAG (1UL<<31)
+#define TRC_HD_INCLUDES_CYCLE_COUNT(x) ( !!( (x) & TRC_HD_CYCLE_FLAG ) )
+#define TRC_HD_EXTRA(x)    (((x)>>TRACE_EXTRA_SHIFT)&TRACE_EXTRA_MAX)
+
+/* Trace subclasses */
+#define TRC_SUBCLS_SHIFT 12
+
+/* trace subclasses for SVM */
+#define TRC_HVM_ENTRYEXIT 0x00081000   /* VMENTRY and #VMEXIT       */
+#define TRC_HVM_HANDLER   0x00082000   /* various HVM handlers      */
+
+/* Trace events per class */
+#define TRC_LOST_RECORDS        (TRC_GEN + 1)
+#define TRC_TRACE_WRAP_BUFFER  (TRC_GEN + 2)
+#define TRC_TRACE_CPU_CHANGE    (TRC_GEN + 3)
+
+#define TRC_SCHED_DOM_ADD       (TRC_SCHED +  1)
+#define TRC_SCHED_DOM_REM       (TRC_SCHED +  2)
+#define TRC_SCHED_SLEEP         (TRC_SCHED +  3)
+#define TRC_SCHED_WAKE          (TRC_SCHED +  4)
+#define TRC_SCHED_YIELD         (TRC_SCHED +  5)
+#define TRC_SCHED_BLOCK         (TRC_SCHED +  6)
+#define TRC_SCHED_SHUTDOWN      (TRC_SCHED +  7)
+#define TRC_SCHED_CTL           (TRC_SCHED +  8)
+#define TRC_SCHED_ADJDOM        (TRC_SCHED +  9)
+#define TRC_SCHED_SWITCH        (TRC_SCHED + 10)
+#define TRC_SCHED_S_TIMER_FN    (TRC_SCHED + 11)
+#define TRC_SCHED_T_TIMER_FN    (TRC_SCHED + 12)
+#define TRC_SCHED_DOM_TIMER_FN  (TRC_SCHED + 13)
+#define TRC_SCHED_SWITCH_INFPREV (TRC_SCHED + 14)
+#define TRC_SCHED_SWITCH_INFNEXT (TRC_SCHED + 15)
+
+#define TRC_MEM_PAGE_GRANT_MAP      (TRC_MEM + 1)
+#define TRC_MEM_PAGE_GRANT_UNMAP    (TRC_MEM + 2)
+#define TRC_MEM_PAGE_GRANT_TRANSFER (TRC_MEM + 3)
+
+#define TRC_PV_HYPERCALL             (TRC_PV +  1)
+#define TRC_PV_TRAP                  (TRC_PV +  3)
+#define TRC_PV_PAGE_FAULT            (TRC_PV +  4)
+#define TRC_PV_FORCED_INVALID_OP     (TRC_PV +  5)
+#define TRC_PV_EMULATE_PRIVOP        (TRC_PV +  6)
+#define TRC_PV_EMULATE_4GB           (TRC_PV +  7)
+#define TRC_PV_MATH_STATE_RESTORE    (TRC_PV +  8)
+#define TRC_PV_PAGING_FIXUP          (TRC_PV +  9)
+#define TRC_PV_GDT_LDT_MAPPING_FAULT (TRC_PV + 10)
+#define TRC_PV_PTWR_EMULATION        (TRC_PV + 11)
+#define TRC_PV_PTWR_EMULATION_PAE    (TRC_PV + 12)
+  /* Indicates that addresses in trace record are 64 bits */
+#define TRC_64_FLAG               (0x100) 
+
+/* trace events per subclass */
+#define TRC_HVM_VMENTRY         (TRC_HVM_ENTRYEXIT + 0x01)
+#define TRC_HVM_VMEXIT          (TRC_HVM_ENTRYEXIT + 0x02)
+#define TRC_HVM_VMEXIT64        (TRC_HVM_ENTRYEXIT + TRC_64_FLAG + 0x02)
+#define TRC_HVM_PF_XEN          (TRC_HVM_HANDLER + 0x01)
+#define TRC_HVM_PF_XEN64        (TRC_HVM_HANDLER + TRC_64_FLAG + 0x01)
+#define TRC_HVM_PF_INJECT       (TRC_HVM_HANDLER + 0x02)
+#define TRC_HVM_PF_INJECT64     (TRC_HVM_HANDLER + TRC_64_FLAG + 0x02)
+#define TRC_HVM_INJ_EXC         (TRC_HVM_HANDLER + 0x03)
+#define TRC_HVM_INJ_VIRQ        (TRC_HVM_HANDLER + 0x04)
+#define TRC_HVM_REINJ_VIRQ      (TRC_HVM_HANDLER + 0x05)
+#define TRC_HVM_IO_READ         (TRC_HVM_HANDLER + 0x06)
+#define TRC_HVM_IO_WRITE        (TRC_HVM_HANDLER + 0x07)
+#define TRC_HVM_CR_READ         (TRC_HVM_HANDLER + 0x08)
+#define TRC_HVM_CR_READ64       (TRC_HVM_HANDLER + TRC_64_FLAG + 0x08)
+#define TRC_HVM_CR_WRITE        (TRC_HVM_HANDLER + 0x09)
+#define TRC_HVM_CR_WRITE64      (TRC_HVM_HANDLER + TRC_64_FLAG + 0x09)
+#define TRC_HVM_DR_READ         (TRC_HVM_HANDLER + 0x0A)
+#define TRC_HVM_DR_WRITE        (TRC_HVM_HANDLER + 0x0B)
+#define TRC_HVM_MSR_READ        (TRC_HVM_HANDLER + 0x0C)
+#define TRC_HVM_MSR_WRITE       (TRC_HVM_HANDLER + 0x0D)
+#define TRC_HVM_CPUID           (TRC_HVM_HANDLER + 0x0E)
+#define TRC_HVM_INTR            (TRC_HVM_HANDLER + 0x0F)
+#define TRC_HVM_NMI             (TRC_HVM_HANDLER + 0x10)
+#define TRC_HVM_SMI             (TRC_HVM_HANDLER + 0x11)
+#define TRC_HVM_VMMCALL         (TRC_HVM_HANDLER + 0x12)
+#define TRC_HVM_HLT             (TRC_HVM_HANDLER + 0x13)
+#define TRC_HVM_INVLPG          (TRC_HVM_HANDLER + 0x14)
+#define TRC_HVM_INVLPG64        (TRC_HVM_HANDLER + TRC_64_FLAG + 0x14)
+#define TRC_HVM_MCE             (TRC_HVM_HANDLER + 0x15)
+#define TRC_HVM_IO_ASSIST       (TRC_HVM_HANDLER + 0x16)
+#define TRC_HVM_MMIO_ASSIST     (TRC_HVM_HANDLER + 0x17)
+#define TRC_HVM_CLTS            (TRC_HVM_HANDLER + 0x18)
+#define TRC_HVM_LMSW            (TRC_HVM_HANDLER + 0x19)
+#define TRC_HVM_LMSW64          (TRC_HVM_HANDLER + TRC_64_FLAG + 0x19)
+
+/* This structure represents a single trace buffer record. */
+struct t_rec {
+    uint32_t event:28;
+    uint32_t extra_u32:3;         /* # entries in trailing extra_u32[] array */
+    uint32_t cycles_included:1;   /* u.cycles or u.no_cycles? */
+    union {
+        struct {
+            uint32_t cycles_lo, cycles_hi; /* cycle counter timestamp */
+            uint32_t extra_u32[7];         /* event data items */
+        } cycles;
+        struct {
+            uint32_t extra_u32[7];         /* event data items */
+        } nocycles;
+    } u;
+};
+
+/*
+ * This structure contains the metadata for a single trace buffer.  The head
+ * field, indexes into an array of struct t_rec's.
+ */
+struct t_buf {
+    /* Assume the data buffer size is X.  X is generally not a power of 2.
+     * CONS and PROD are incremented modulo (2*X):
+     *     0 <= cons < 2*X
+     *     0 <= prod < 2*X
+     * This is done because addition modulo X breaks at 2^32 when X is not a
+     * power of 2:
+     *     (((2^32 - 1) % X) + 1) % X != (2^32) % X
+     */
+    uint32_t cons;   /* Offset of next item to be consumed by control tools. */
+    uint32_t prod;   /* Offset of next item to be produced by Xen.           */
+    /*  Records follow immediately after the meta-data header.    */
+};
+
+#endif /* __XEN_PUBLIC_TRACE_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/vcpu.h linux-2.6.18-xen-3.3.0/include/xen/interface/vcpu.h
--- linux-2.6.18.8/include/xen/interface/vcpu.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/vcpu.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,213 @@
+/******************************************************************************
+ * vcpu.h
+ * 
+ * VCPU initialisation, query, and hotplug.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2005, Keir Fraser <keir@xensource.com>
+ */
+
+#ifndef __XEN_PUBLIC_VCPU_H__
+#define __XEN_PUBLIC_VCPU_H__
+
+/*
+ * Prototype for this hypercall is:
+ *  int vcpu_op(int cmd, int vcpuid, void *extra_args)
+ * @cmd        == VCPUOP_??? (VCPU operation).
+ * @vcpuid     == VCPU to operate on.
+ * @extra_args == Operation-specific extra arguments (NULL if none).
+ */
+
+/*
+ * Initialise a VCPU. Each VCPU can be initialised only once. A 
+ * newly-initialised VCPU will not run until it is brought up by VCPUOP_up.
+ * 
+ * @extra_arg == pointer to vcpu_guest_context structure containing initial
+ *               state for the VCPU.
+ */
+#define VCPUOP_initialise            0
+
+/*
+ * Bring up a VCPU. This makes the VCPU runnable. This operation will fail
+ * if the VCPU has not been initialised (VCPUOP_initialise).
+ */
+#define VCPUOP_up                    1
+
+/*
+ * Bring down a VCPU (i.e., make it non-runnable).
+ * There are a few caveats that callers should observe:
+ *  1. This operation may return, and VCPU_is_up may return false, before the
+ *     VCPU stops running (i.e., the command is asynchronous). It is a good
+ *     idea to ensure that the VCPU has entered a non-critical loop before
+ *     bringing it down. Alternatively, this operation is guaranteed
+ *     synchronous if invoked by the VCPU itself.
+ *  2. After a VCPU is initialised, there is currently no way to drop all its
+ *     references to domain memory. Even a VCPU that is down still holds
+ *     memory references via its pagetable base pointer and GDT. It is good
+ *     practise to move a VCPU onto an 'idle' or default page table, LDT and
+ *     GDT before bringing it down.
+ */
+#define VCPUOP_down                  2
+
+/* Returns 1 if the given VCPU is up. */
+#define VCPUOP_is_up                 3
+
+/*
+ * Return information about the state and running time of a VCPU.
+ * @extra_arg == pointer to vcpu_runstate_info structure.
+ */
+#define VCPUOP_get_runstate_info     4
+struct vcpu_runstate_info {
+    /* VCPU's current state (RUNSTATE_*). */
+    int      state;
+    /* When was current state entered (system time, ns)? */
+    uint64_t state_entry_time;
+    /*
+     * Time spent in each RUNSTATE_* (ns). The sum of these times is
+     * guaranteed not to drift from system time.
+     */
+    uint64_t time[4];
+};
+typedef struct vcpu_runstate_info vcpu_runstate_info_t;
+DEFINE_XEN_GUEST_HANDLE(vcpu_runstate_info_t);
+
+/* VCPU is currently running on a physical CPU. */
+#define RUNSTATE_running  0
+
+/* VCPU is runnable, but not currently scheduled on any physical CPU. */
+#define RUNSTATE_runnable 1
+
+/* VCPU is blocked (a.k.a. idle). It is therefore not runnable. */
+#define RUNSTATE_blocked  2
+
+/*
+ * VCPU is not runnable, but it is not blocked.
+ * This is a 'catch all' state for things like hotplug and pauses by the
+ * system administrator (or for critical sections in the hypervisor).
+ * RUNSTATE_blocked dominates this state (it is the preferred state).
+ */
+#define RUNSTATE_offline  3
+
+/*
+ * Register a shared memory area from which the guest may obtain its own
+ * runstate information without needing to execute a hypercall.
+ * Notes:
+ *  1. The registered address may be virtual or physical or guest handle,
+ *     depending on the platform. Virtual address or guest handle should be
+ *     registered on x86 systems.
+ *  2. Only one shared area may be registered per VCPU. The shared area is
+ *     updated by the hypervisor each time the VCPU is scheduled. Thus
+ *     runstate.state will always be RUNSTATE_running and
+ *     runstate.state_entry_time will indicate the system time at which the
+ *     VCPU was last scheduled to run.
+ * @extra_arg == pointer to vcpu_register_runstate_memory_area structure.
+ */
+#define VCPUOP_register_runstate_memory_area 5
+struct vcpu_register_runstate_memory_area {
+    union {
+        XEN_GUEST_HANDLE(vcpu_runstate_info_t) h;
+        struct vcpu_runstate_info *v;
+        uint64_t p;
+    } addr;
+};
+typedef struct vcpu_register_runstate_memory_area vcpu_register_runstate_memory_area_t;
+DEFINE_XEN_GUEST_HANDLE(vcpu_register_runstate_memory_area_t);
+
+/*
+ * Set or stop a VCPU's periodic timer. Every VCPU has one periodic timer
+ * which can be set via these commands. Periods smaller than one millisecond
+ * may not be supported.
+ */
+#define VCPUOP_set_periodic_timer    6 /* arg == vcpu_set_periodic_timer_t */
+#define VCPUOP_stop_periodic_timer   7 /* arg == NULL */
+struct vcpu_set_periodic_timer {
+    uint64_t period_ns;
+};
+typedef struct vcpu_set_periodic_timer vcpu_set_periodic_timer_t;
+DEFINE_XEN_GUEST_HANDLE(vcpu_set_periodic_timer_t);
+
+/*
+ * Set or stop a VCPU's single-shot timer. Every VCPU has one single-shot
+ * timer which can be set via these commands.
+ */
+#define VCPUOP_set_singleshot_timer  8 /* arg == vcpu_set_singleshot_timer_t */
+#define VCPUOP_stop_singleshot_timer 9 /* arg == NULL */
+struct vcpu_set_singleshot_timer {
+    uint64_t timeout_abs_ns;   /* Absolute system time value in nanoseconds. */
+    uint32_t flags;            /* VCPU_SSHOTTMR_??? */
+};
+typedef struct vcpu_set_singleshot_timer vcpu_set_singleshot_timer_t;
+DEFINE_XEN_GUEST_HANDLE(vcpu_set_singleshot_timer_t);
+
+/* Flags to VCPUOP_set_singleshot_timer. */
+ /* Require the timeout to be in the future (return -ETIME if it's passed). */
+#define _VCPU_SSHOTTMR_future (0)
+#define VCPU_SSHOTTMR_future  (1U << _VCPU_SSHOTTMR_future)
+
+/* 
+ * Register a memory location in the guest address space for the
+ * vcpu_info structure.  This allows the guest to place the vcpu_info
+ * structure in a convenient place, such as in a per-cpu data area.
+ * The pointer need not be page aligned, but the structure must not
+ * cross a page boundary.
+ *
+ * This may be called only once per vcpu.
+ */
+#define VCPUOP_register_vcpu_info   10  /* arg == vcpu_register_vcpu_info_t */
+struct vcpu_register_vcpu_info {
+    uint64_t mfn;    /* mfn of page to place vcpu_info */
+    uint32_t offset; /* offset within page */
+    uint32_t rsvd;   /* unused */
+};
+typedef struct vcpu_register_vcpu_info vcpu_register_vcpu_info_t;
+DEFINE_XEN_GUEST_HANDLE(vcpu_register_vcpu_info_t);
+
+/* Send an NMI to the specified VCPU. @extra_arg == NULL. */
+#define VCPUOP_send_nmi             11
+
+/* 
+ * Get the physical ID information for a pinned vcpu's underlying physical
+ * processor.  The physical ID informmation is architecture-specific.
+ * On x86: id[31:0]=apic_id, id[63:32]=acpi_id, and all values 0xff and
+ *         greater are reserved.
+ * This command returns -EINVAL if it is not a valid operation for this VCPU.
+ */
+#define VCPUOP_get_physid           12 /* arg == vcpu_get_physid_t */
+struct vcpu_get_physid {
+    uint64_t phys_id;
+};
+typedef struct vcpu_get_physid vcpu_get_physid_t;
+DEFINE_XEN_GUEST_HANDLE(vcpu_get_physid_t);
+#define xen_vcpu_physid_to_x86_apicid(physid) \
+    ((((uint32_t)(physid)) >= 0xff) ? 0xff : ((uint8_t)(physid)))
+#define xen_vcpu_physid_to_x86_acpiid(physid) \
+    ((((uint32_t)((physid)>>32)) >= 0xff) ? 0xff : ((uint8_t)((physid)>>32)))
+
+#endif /* __XEN_PUBLIC_VCPU_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/version.h linux-2.6.18-xen-3.3.0/include/xen/interface/version.h
--- linux-2.6.18.8/include/xen/interface/version.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/version.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,91 @@
+/******************************************************************************
+ * version.h
+ * 
+ * Xen version, type, and compile information.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2005, Nguyen Anh Quynh <aquynh@gmail.com>
+ * Copyright (c) 2005, Keir Fraser <keir@xensource.com>
+ */
+
+#ifndef __XEN_PUBLIC_VERSION_H__
+#define __XEN_PUBLIC_VERSION_H__
+
+/* NB. All ops return zero on success, except XENVER_{version,pagesize} */
+
+/* arg == NULL; returns major:minor (16:16). */
+#define XENVER_version      0
+
+/* arg == xen_extraversion_t. */
+#define XENVER_extraversion 1
+typedef char xen_extraversion_t[16];
+#define XEN_EXTRAVERSION_LEN (sizeof(xen_extraversion_t))
+
+/* arg == xen_compile_info_t. */
+#define XENVER_compile_info 2
+struct xen_compile_info {
+    char compiler[64];
+    char compile_by[16];
+    char compile_domain[32];
+    char compile_date[32];
+};
+typedef struct xen_compile_info xen_compile_info_t;
+
+#define XENVER_capabilities 3
+typedef char xen_capabilities_info_t[1024];
+#define XEN_CAPABILITIES_INFO_LEN (sizeof(xen_capabilities_info_t))
+
+#define XENVER_changeset 4
+typedef char xen_changeset_info_t[64];
+#define XEN_CHANGESET_INFO_LEN (sizeof(xen_changeset_info_t))
+
+#define XENVER_platform_parameters 5
+struct xen_platform_parameters {
+    unsigned long virt_start;
+};
+typedef struct xen_platform_parameters xen_platform_parameters_t;
+
+#define XENVER_get_features 6
+struct xen_feature_info {
+    unsigned int submap_idx;    /* IN: which 32-bit submap to return */
+    uint32_t     submap;        /* OUT: 32-bit submap */
+};
+typedef struct xen_feature_info xen_feature_info_t;
+
+/* Declares the features reported by XENVER_get_features. */
+#include "features.h"
+
+/* arg == NULL; returns host memory page size. */
+#define XENVER_pagesize 7
+
+/* arg == xen_domain_handle_t. */
+#define XENVER_guest_handle 8
+
+#endif /* __XEN_PUBLIC_VERSION_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/xencomm.h linux-2.6.18-xen-3.3.0/include/xen/interface/xencomm.h
--- linux-2.6.18.8/include/xen/interface/xencomm.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/xencomm.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,41 @@
+/*
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (C) IBM Corp. 2006
+ */
+
+#ifndef _XEN_XENCOMM_H_
+#define _XEN_XENCOMM_H_
+
+/* A xencomm descriptor is a scatter/gather list containing physical
+ * addresses corresponding to a virtually contiguous memory area. The
+ * hypervisor translates these physical addresses to machine addresses to copy
+ * to and from the virtually contiguous area.
+ */
+
+#define XENCOMM_MAGIC 0x58434F4D /* 'XCOM' */
+#define XENCOMM_INVALID (~0UL)
+
+struct xencomm_desc {
+    uint32_t magic;
+    uint32_t nr_addrs; /* the number of entries in address[] */
+    uint64_t address[0];
+};
+
+#endif /* _XEN_XENCOMM_H_ */
diff -rpuN linux-2.6.18.8/include/xen/interface/xen-compat.h linux-2.6.18-xen-3.3.0/include/xen/interface/xen-compat.h
--- linux-2.6.18.8/include/xen/interface/xen-compat.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/xen-compat.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,44 @@
+/******************************************************************************
+ * xen-compat.h
+ * 
+ * Guest OS interface to Xen.  Compatibility layer.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2006, Christian Limpach
+ */
+
+#ifndef __XEN_PUBLIC_XEN_COMPAT_H__
+#define __XEN_PUBLIC_XEN_COMPAT_H__
+
+#define __XEN_LATEST_INTERFACE_VERSION__ 0x00030209
+
+#if defined(__XEN__) || defined(__XEN_TOOLS__)
+/* Xen is built with matching headers and implements the latest interface. */
+#define __XEN_INTERFACE_VERSION__ __XEN_LATEST_INTERFACE_VERSION__
+#elif !defined(__XEN_INTERFACE_VERSION__)
+/* Guests which do not specify a version get the legacy interface. */
+#define __XEN_INTERFACE_VERSION__ 0x00000000
+#endif
+
+#if __XEN_INTERFACE_VERSION__ > __XEN_LATEST_INTERFACE_VERSION__
+#error "These header files do not support the requested interface version."
+#endif
+
+#endif /* __XEN_PUBLIC_XEN_COMPAT_H__ */
diff -rpuN linux-2.6.18.8/include/xen/interface/xen.h linux-2.6.18-xen-3.3.0/include/xen/interface/xen.h
--- linux-2.6.18.8/include/xen/interface/xen.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/xen.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,639 @@
+/******************************************************************************
+ * xen.h
+ * 
+ * Guest OS interface to Xen.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (c) 2004, K A Fraser
+ */
+
+#ifndef __XEN_PUBLIC_XEN_H__
+#define __XEN_PUBLIC_XEN_H__
+
+#include "xen-compat.h"
+
+#if defined(__i386__) || defined(__x86_64__)
+#include "arch-x86/xen.h"
+#elif defined(__ia64__)
+#include "arch-ia64.h"
+#else
+#error "Unsupported architecture"
+#endif
+
+#ifndef __ASSEMBLY__
+/* Guest handles for primitive C types. */
+DEFINE_XEN_GUEST_HANDLE(char);
+__DEFINE_XEN_GUEST_HANDLE(uchar, unsigned char);
+DEFINE_XEN_GUEST_HANDLE(int);
+__DEFINE_XEN_GUEST_HANDLE(uint,  unsigned int);
+DEFINE_XEN_GUEST_HANDLE(long);
+__DEFINE_XEN_GUEST_HANDLE(ulong, unsigned long);
+DEFINE_XEN_GUEST_HANDLE(void);
+
+DEFINE_XEN_GUEST_HANDLE(xen_pfn_t);
+#endif
+
+/*
+ * HYPERCALLS
+ */
+
+#define __HYPERVISOR_set_trap_table        0
+#define __HYPERVISOR_mmu_update            1
+#define __HYPERVISOR_set_gdt               2
+#define __HYPERVISOR_stack_switch          3
+#define __HYPERVISOR_set_callbacks         4
+#define __HYPERVISOR_fpu_taskswitch        5
+#define __HYPERVISOR_sched_op_compat       6 /* compat since 0x00030101 */
+#define __HYPERVISOR_platform_op           7
+#define __HYPERVISOR_set_debugreg          8
+#define __HYPERVISOR_get_debugreg          9
+#define __HYPERVISOR_update_descriptor    10
+#define __HYPERVISOR_memory_op            12
+#define __HYPERVISOR_multicall            13
+#define __HYPERVISOR_update_va_mapping    14
+#define __HYPERVISOR_set_timer_op         15
+#define __HYPERVISOR_event_channel_op_compat 16 /* compat since 0x00030202 */
+#define __HYPERVISOR_xen_version          17
+#define __HYPERVISOR_console_io           18
+#define __HYPERVISOR_physdev_op_compat    19 /* compat since 0x00030202 */
+#define __HYPERVISOR_grant_table_op       20
+#define __HYPERVISOR_vm_assist            21
+#define __HYPERVISOR_update_va_mapping_otherdomain 22
+#define __HYPERVISOR_iret                 23 /* x86 only */
+#define __HYPERVISOR_vcpu_op              24
+#define __HYPERVISOR_set_segment_base     25 /* x86/64 only */
+#define __HYPERVISOR_mmuext_op            26
+#define __HYPERVISOR_xsm_op               27
+#define __HYPERVISOR_nmi_op               28
+#define __HYPERVISOR_sched_op             29
+#define __HYPERVISOR_callback_op          30
+#define __HYPERVISOR_xenoprof_op          31
+#define __HYPERVISOR_event_channel_op     32
+#define __HYPERVISOR_physdev_op           33
+#define __HYPERVISOR_hvm_op               34
+#define __HYPERVISOR_sysctl               35
+#define __HYPERVISOR_domctl               36
+#define __HYPERVISOR_kexec_op             37
+
+/* Architecture-specific hypercall definitions. */
+#define __HYPERVISOR_arch_0               48
+#define __HYPERVISOR_arch_1               49
+#define __HYPERVISOR_arch_2               50
+#define __HYPERVISOR_arch_3               51
+#define __HYPERVISOR_arch_4               52
+#define __HYPERVISOR_arch_5               53
+#define __HYPERVISOR_arch_6               54
+#define __HYPERVISOR_arch_7               55
+
+/*
+ * HYPERCALL COMPATIBILITY.
+ */
+
+/* New sched_op hypercall introduced in 0x00030101. */
+#if __XEN_INTERFACE_VERSION__ < 0x00030101
+#undef __HYPERVISOR_sched_op
+#define __HYPERVISOR_sched_op __HYPERVISOR_sched_op_compat
+#endif
+
+/* New event-channel and physdev hypercalls introduced in 0x00030202. */
+#if __XEN_INTERFACE_VERSION__ < 0x00030202
+#undef __HYPERVISOR_event_channel_op
+#define __HYPERVISOR_event_channel_op __HYPERVISOR_event_channel_op_compat
+#undef __HYPERVISOR_physdev_op
+#define __HYPERVISOR_physdev_op __HYPERVISOR_physdev_op_compat
+#endif
+
+/* New platform_op hypercall introduced in 0x00030204. */
+#if __XEN_INTERFACE_VERSION__ < 0x00030204
+#define __HYPERVISOR_dom0_op __HYPERVISOR_platform_op
+#endif
+
+/* 
+ * VIRTUAL INTERRUPTS
+ * 
+ * Virtual interrupts that a guest OS may receive from Xen.
+ * 
+ * In the side comments, 'V.' denotes a per-VCPU VIRQ while 'G.' denotes a
+ * global VIRQ. The former can be bound once per VCPU and cannot be re-bound.
+ * The latter can be allocated only once per guest: they must initially be
+ * allocated to VCPU0 but can subsequently be re-bound.
+ */
+#define VIRQ_TIMER      0  /* V. Timebase update, and/or requested timeout.  */
+#define VIRQ_DEBUG      1  /* V. Request guest to dump debug info.           */
+#define VIRQ_CONSOLE    2  /* G. (DOM0) Bytes received on emergency console. */
+#define VIRQ_DOM_EXC    3  /* G. (DOM0) Exceptional event for some domain.   */
+#define VIRQ_TBUF       4  /* G. (DOM0) Trace buffer has records available.  */
+#define VIRQ_DEBUGGER   6  /* G. (DOM0) A domain has paused for debugging.   */
+#define VIRQ_XENOPROF   7  /* V. XenOprofile interrupt: new sample available */
+#define VIRQ_CON_RING   8  /* G. (DOM0) Bytes received on console            */
+
+/* Architecture-specific VIRQ definitions. */
+#define VIRQ_ARCH_0    16
+#define VIRQ_ARCH_1    17
+#define VIRQ_ARCH_2    18
+#define VIRQ_ARCH_3    19
+#define VIRQ_ARCH_4    20
+#define VIRQ_ARCH_5    21
+#define VIRQ_ARCH_6    22
+#define VIRQ_ARCH_7    23
+
+#define NR_VIRQS       24
+
+/*
+ * MMU-UPDATE REQUESTS
+ * 
+ * HYPERVISOR_mmu_update() accepts a list of (ptr, val) pairs.
+ * A foreigndom (FD) can be specified (or DOMID_SELF for none).
+ * Where the FD has some effect, it is described below.
+ * ptr[1:0] specifies the appropriate MMU_* command.
+ * 
+ * ptr[1:0] == MMU_NORMAL_PT_UPDATE:
+ * Updates an entry in a page table. If updating an L1 table, and the new
+ * table entry is valid/present, the mapped frame must belong to the FD, if
+ * an FD has been specified. If attempting to map an I/O page then the
+ * caller assumes the privilege of the FD.
+ * FD == DOMID_IO: Permit /only/ I/O mappings, at the priv level of the caller.
+ * FD == DOMID_XEN: Map restricted areas of Xen's heap space.
+ * ptr[:2]  -- Machine address of the page-table entry to modify.
+ * val      -- Value to write.
+ * 
+ * ptr[1:0] == MMU_MACHPHYS_UPDATE:
+ * Updates an entry in the machine->pseudo-physical mapping table.
+ * ptr[:2]  -- Machine address within the frame whose mapping to modify.
+ *             The frame must belong to the FD, if one is specified.
+ * val      -- Value to write into the mapping entry.
+ * 
+ * ptr[1:0] == MMU_PT_UPDATE_PRESERVE_AD:
+ * As MMU_NORMAL_PT_UPDATE above, but A/D bits currently in the PTE are ORed
+ * with those in @val.
+ */
+#define MMU_NORMAL_PT_UPDATE      0 /* checked '*ptr = val'. ptr is MA.      */
+#define MMU_MACHPHYS_UPDATE       1 /* ptr = MA of frame to modify entry for */
+#define MMU_PT_UPDATE_PRESERVE_AD 2 /* atomically: *ptr = val | (*ptr&(A|D)) */
+
+/*
+ * MMU EXTENDED OPERATIONS
+ * 
+ * HYPERVISOR_mmuext_op() accepts a list of mmuext_op structures.
+ * A foreigndom (FD) can be specified (or DOMID_SELF for none).
+ * Where the FD has some effect, it is described below.
+ * 
+ * cmd: MMUEXT_(UN)PIN_*_TABLE
+ * mfn: Machine frame number to be (un)pinned as a p.t. page.
+ *      The frame must belong to the FD, if one is specified.
+ * 
+ * cmd: MMUEXT_NEW_BASEPTR
+ * mfn: Machine frame number of new page-table base to install in MMU.
+ * 
+ * cmd: MMUEXT_NEW_USER_BASEPTR [x86/64 only]
+ * mfn: Machine frame number of new page-table base to install in MMU
+ *      when in user space.
+ * 
+ * cmd: MMUEXT_TLB_FLUSH_LOCAL
+ * No additional arguments. Flushes local TLB.
+ * 
+ * cmd: MMUEXT_INVLPG_LOCAL
+ * linear_addr: Linear address to be flushed from the local TLB.
+ * 
+ * cmd: MMUEXT_TLB_FLUSH_MULTI
+ * vcpumask: Pointer to bitmap of VCPUs to be flushed.
+ * 
+ * cmd: MMUEXT_INVLPG_MULTI
+ * linear_addr: Linear address to be flushed.
+ * vcpumask: Pointer to bitmap of VCPUs to be flushed.
+ * 
+ * cmd: MMUEXT_TLB_FLUSH_ALL
+ * No additional arguments. Flushes all VCPUs' TLBs.
+ * 
+ * cmd: MMUEXT_INVLPG_ALL
+ * linear_addr: Linear address to be flushed from all VCPUs' TLBs.
+ * 
+ * cmd: MMUEXT_FLUSH_CACHE
+ * No additional arguments. Writes back and flushes cache contents.
+ * 
+ * cmd: MMUEXT_SET_LDT
+ * linear_addr: Linear address of LDT base (NB. must be page-aligned).
+ * nr_ents: Number of entries in LDT.
+ */
+#define MMUEXT_PIN_L1_TABLE      0
+#define MMUEXT_PIN_L2_TABLE      1
+#define MMUEXT_PIN_L3_TABLE      2
+#define MMUEXT_PIN_L4_TABLE      3
+#define MMUEXT_UNPIN_TABLE       4
+#define MMUEXT_NEW_BASEPTR       5
+#define MMUEXT_TLB_FLUSH_LOCAL   6
+#define MMUEXT_INVLPG_LOCAL      7
+#define MMUEXT_TLB_FLUSH_MULTI   8
+#define MMUEXT_INVLPG_MULTI      9
+#define MMUEXT_TLB_FLUSH_ALL    10
+#define MMUEXT_INVLPG_ALL       11
+#define MMUEXT_FLUSH_CACHE      12
+#define MMUEXT_SET_LDT          13
+#define MMUEXT_NEW_USER_BASEPTR 15
+
+#ifndef __ASSEMBLY__
+struct mmuext_op {
+    unsigned int cmd;
+    union {
+        /* [UN]PIN_TABLE, NEW_BASEPTR, NEW_USER_BASEPTR */
+        xen_pfn_t     mfn;
+        /* INVLPG_LOCAL, INVLPG_ALL, SET_LDT */
+        unsigned long linear_addr;
+    } arg1;
+    union {
+        /* SET_LDT */
+        unsigned int nr_ents;
+        /* TLB_FLUSH_MULTI, INVLPG_MULTI */
+#if __XEN_INTERFACE_VERSION__ >= 0x00030205
+        XEN_GUEST_HANDLE(void) vcpumask;
+#else
+        void *vcpumask;
+#endif
+    } arg2;
+};
+typedef struct mmuext_op mmuext_op_t;
+DEFINE_XEN_GUEST_HANDLE(mmuext_op_t);
+#endif
+
+/* These are passed as 'flags' to update_va_mapping. They can be ORed. */
+/* When specifying UVMF_MULTI, also OR in a pointer to a CPU bitmap.   */
+/* UVMF_LOCAL is merely UVMF_MULTI with a NULL bitmap pointer.         */
+#define UVMF_NONE               (0UL<<0) /* No flushing at all.   */
+#define UVMF_TLB_FLUSH          (1UL<<0) /* Flush entire TLB(s).  */
+#define UVMF_INVLPG             (2UL<<0) /* Flush only one entry. */
+#define UVMF_FLUSHTYPE_MASK     (3UL<<0)
+#define UVMF_MULTI              (0UL<<2) /* Flush subset of TLBs. */
+#define UVMF_LOCAL              (0UL<<2) /* Flush local TLB.      */
+#define UVMF_ALL                (1UL<<2) /* Flush all TLBs.       */
+
+/*
+ * Commands to HYPERVISOR_console_io().
+ */
+#define CONSOLEIO_write         0
+#define CONSOLEIO_read          1
+
+/*
+ * Commands to HYPERVISOR_vm_assist().
+ */
+#define VMASST_CMD_enable                0
+#define VMASST_CMD_disable               1
+
+/* x86/32 guests: simulate full 4GB segment limits. */
+#define VMASST_TYPE_4gb_segments         0
+
+/* x86/32 guests: trap (vector 15) whenever above vmassist is used. */
+#define VMASST_TYPE_4gb_segments_notify  1
+
+/*
+ * x86 guests: support writes to bottom-level PTEs.
+ * NB1. Page-directory entries cannot be written.
+ * NB2. Guest must continue to remove all writable mappings of PTEs.
+ */
+#define VMASST_TYPE_writable_pagetables  2
+
+/* x86/PAE guests: support PDPTs above 4GB. */
+#define VMASST_TYPE_pae_extended_cr3     3
+
+#define MAX_VMASST_TYPE                  3
+
+#ifndef __ASSEMBLY__
+
+typedef uint16_t domid_t;
+
+/* Domain ids >= DOMID_FIRST_RESERVED cannot be used for ordinary domains. */
+#define DOMID_FIRST_RESERVED (0x7FF0U)
+
+/* DOMID_SELF is used in certain contexts to refer to oneself. */
+#define DOMID_SELF (0x7FF0U)
+
+/*
+ * DOMID_IO is used to restrict page-table updates to mapping I/O memory.
+ * Although no Foreign Domain need be specified to map I/O pages, DOMID_IO
+ * is useful to ensure that no mappings to the OS's own heap are accidentally
+ * installed. (e.g., in Linux this could cause havoc as reference counts
+ * aren't adjusted on the I/O-mapping code path).
+ * This only makes sense in MMUEXT_SET_FOREIGNDOM, but in that context can
+ * be specified by any calling domain.
+ */
+#define DOMID_IO   (0x7FF1U)
+
+/*
+ * DOMID_XEN is used to allow privileged domains to map restricted parts of
+ * Xen's heap space (e.g., the machine_to_phys table).
+ * This only makes sense in MMUEXT_SET_FOREIGNDOM, and is only permitted if
+ * the caller is privileged.
+ */
+#define DOMID_XEN  (0x7FF2U)
+
+/*
+ * Send an array of these to HYPERVISOR_mmu_update().
+ * NB. The fields are natural pointer/address size for this architecture.
+ */
+struct mmu_update {
+    uint64_t ptr;       /* Machine address of PTE. */
+    uint64_t val;       /* New contents of PTE.    */
+};
+typedef struct mmu_update mmu_update_t;
+DEFINE_XEN_GUEST_HANDLE(mmu_update_t);
+
+/*
+ * Send an array of these to HYPERVISOR_multicall().
+ * NB. The fields are natural register size for this architecture.
+ */
+struct multicall_entry {
+    unsigned long op, result;
+    unsigned long args[6];
+};
+typedef struct multicall_entry multicall_entry_t;
+DEFINE_XEN_GUEST_HANDLE(multicall_entry_t);
+
+/*
+ * Event channel endpoints per domain:
+ *  1024 if a long is 32 bits; 4096 if a long is 64 bits.
+ */
+#define NR_EVENT_CHANNELS (sizeof(unsigned long) * sizeof(unsigned long) * 64)
+
+struct vcpu_time_info {
+    /*
+     * Updates to the following values are preceded and followed by an
+     * increment of 'version'. The guest can therefore detect updates by
+     * looking for changes to 'version'. If the least-significant bit of
+     * the version number is set then an update is in progress and the guest
+     * must wait to read a consistent set of values.
+     * The correct way to interact with the version number is similar to
+     * Linux's seqlock: see the implementations of read_seqbegin/read_seqretry.
+     */
+    uint32_t version;
+    uint32_t pad0;
+    uint64_t tsc_timestamp;   /* TSC at last update of time vals.  */
+    uint64_t system_time;     /* Time, in nanosecs, since boot.    */
+    /*
+     * Current system time:
+     *   system_time +
+     *   ((((tsc - tsc_timestamp) << tsc_shift) * tsc_to_system_mul) >> 32)
+     * CPU frequency (Hz):
+     *   ((10^9 << 32) / tsc_to_system_mul) >> tsc_shift
+     */
+    uint32_t tsc_to_system_mul;
+    int8_t   tsc_shift;
+    int8_t   pad1[3];
+}; /* 32 bytes */
+typedef struct vcpu_time_info vcpu_time_info_t;
+
+struct vcpu_info {
+    /*
+     * 'evtchn_upcall_pending' is written non-zero by Xen to indicate
+     * a pending notification for a particular VCPU. It is then cleared 
+     * by the guest OS /before/ checking for pending work, thus avoiding
+     * a set-and-check race. Note that the mask is only accessed by Xen
+     * on the CPU that is currently hosting the VCPU. This means that the
+     * pending and mask flags can be updated by the guest without special
+     * synchronisation (i.e., no need for the x86 LOCK prefix).
+     * This may seem suboptimal because if the pending flag is set by
+     * a different CPU then an IPI may be scheduled even when the mask
+     * is set. However, note:
+     *  1. The task of 'interrupt holdoff' is covered by the per-event-
+     *     channel mask bits. A 'noisy' event that is continually being
+     *     triggered can be masked at source at this very precise
+     *     granularity.
+     *  2. The main purpose of the per-VCPU mask is therefore to restrict
+     *     reentrant execution: whether for concurrency control, or to
+     *     prevent unbounded stack usage. Whatever the purpose, we expect
+     *     that the mask will be asserted only for short periods at a time,
+     *     and so the likelihood of a 'spurious' IPI is suitably small.
+     * The mask is read before making an event upcall to the guest: a
+     * non-zero mask therefore guarantees that the VCPU will not receive
+     * an upcall activation. The mask is cleared when the VCPU requests
+     * to block: this avoids wakeup-waiting races.
+     */
+    uint8_t evtchn_upcall_pending;
+    uint8_t evtchn_upcall_mask;
+    unsigned long evtchn_pending_sel;
+    struct arch_vcpu_info arch;
+    struct vcpu_time_info time;
+}; /* 64 bytes (x86) */
+#ifndef __XEN__
+typedef struct vcpu_info vcpu_info_t;
+#endif
+
+/*
+ * Xen/kernel shared data -- pointer provided in start_info.
+ *
+ * This structure is defined to be both smaller than a page, and the
+ * only data on the shared page, but may vary in actual size even within
+ * compatible Xen versions; guests should not rely on the size
+ * of this structure remaining constant.
+ */
+struct shared_info {
+    struct vcpu_info vcpu_info[MAX_VIRT_CPUS];
+
+    /*
+     * A domain can create "event channels" on which it can send and receive
+     * asynchronous event notifications. There are three classes of event that
+     * are delivered by this mechanism:
+     *  1. Bi-directional inter- and intra-domain connections. Domains must
+     *     arrange out-of-band to set up a connection (usually by allocating
+     *     an unbound 'listener' port and avertising that via a storage service
+     *     such as xenstore).
+     *  2. Physical interrupts. A domain with suitable hardware-access
+     *     privileges can bind an event-channel port to a physical interrupt
+     *     source.
+     *  3. Virtual interrupts ('events'). A domain can bind an event-channel
+     *     port to a virtual interrupt source, such as the virtual-timer
+     *     device or the emergency console.
+     * 
+     * Event channels are addressed by a "port index". Each channel is
+     * associated with two bits of information:
+     *  1. PENDING -- notifies the domain that there is a pending notification
+     *     to be processed. This bit is cleared by the guest.
+     *  2. MASK -- if this bit is clear then a 0->1 transition of PENDING
+     *     will cause an asynchronous upcall to be scheduled. This bit is only
+     *     updated by the guest. It is read-only within Xen. If a channel
+     *     becomes pending while the channel is masked then the 'edge' is lost
+     *     (i.e., when the channel is unmasked, the guest must manually handle
+     *     pending notifications as no upcall will be scheduled by Xen).
+     * 
+     * To expedite scanning of pending notifications, any 0->1 pending
+     * transition on an unmasked channel causes a corresponding bit in a
+     * per-vcpu selector word to be set. Each bit in the selector covers a
+     * 'C long' in the PENDING bitfield array.
+     */
+    unsigned long evtchn_pending[sizeof(unsigned long) * 8];
+    unsigned long evtchn_mask[sizeof(unsigned long) * 8];
+
+    /*
+     * Wallclock time: updated only by control software. Guests should base
+     * their gettimeofday() syscall on this wallclock-base value.
+     */
+    uint32_t wc_version;      /* Version counter: see vcpu_time_info_t. */
+    uint32_t wc_sec;          /* Secs  00:00:00 UTC, Jan 1, 1970.  */
+    uint32_t wc_nsec;         /* Nsecs 00:00:00 UTC, Jan 1, 1970.  */
+
+    struct arch_shared_info arch;
+
+};
+#ifndef __XEN__
+typedef struct shared_info shared_info_t;
+#endif
+
+/*
+ * Start-of-day memory layout:
+ *  1. The domain is started within contiguous virtual-memory region.
+ *  2. The contiguous region ends on an aligned 4MB boundary.
+ *  3. This the order of bootstrap elements in the initial virtual region:
+ *      a. relocated kernel image
+ *      b. initial ram disk              [mod_start, mod_len]
+ *      c. list of allocated page frames [mfn_list, nr_pages]
+ *      d. start_info_t structure        [register ESI (x86)]
+ *      e. bootstrap page tables         [pt_base, CR3 (x86)]
+ *      f. bootstrap stack               [register ESP (x86)]
+ *  4. Bootstrap elements are packed together, but each is 4kB-aligned.
+ *  5. The initial ram disk may be omitted.
+ *  6. The list of page frames forms a contiguous 'pseudo-physical' memory
+ *     layout for the domain. In particular, the bootstrap virtual-memory
+ *     region is a 1:1 mapping to the first section of the pseudo-physical map.
+ *  7. All bootstrap elements are mapped read-writable for the guest OS. The
+ *     only exception is the bootstrap page table, which is mapped read-only.
+ *  8. There is guaranteed to be at least 512kB padding after the final
+ *     bootstrap element. If necessary, the bootstrap virtual region is
+ *     extended by an extra 4MB to ensure this.
+ */
+
+#define MAX_GUEST_CMDLINE 1024
+struct start_info {
+    /* THE FOLLOWING ARE FILLED IN BOTH ON INITIAL BOOT AND ON RESUME.    */
+    char magic[32];             /* "xen-<version>-<platform>".            */
+    unsigned long nr_pages;     /* Total pages allocated to this domain.  */
+    unsigned long shared_info;  /* MACHINE address of shared info struct. */
+    uint32_t flags;             /* SIF_xxx flags.                         */
+    xen_pfn_t store_mfn;        /* MACHINE page number of shared page.    */
+    uint32_t store_evtchn;      /* Event channel for store communication. */
+    union {
+        struct {
+            xen_pfn_t mfn;      /* MACHINE page number of console page.   */
+            uint32_t  evtchn;   /* Event channel for console page.        */
+        } domU;
+        struct {
+            uint32_t info_off;  /* Offset of console_info struct.         */
+            uint32_t info_size; /* Size of console_info struct from start.*/
+        } dom0;
+    } console;
+    /* THE FOLLOWING ARE ONLY FILLED IN ON INITIAL BOOT (NOT RESUME).     */
+    unsigned long pt_base;      /* VIRTUAL address of page directory.     */
+    unsigned long nr_pt_frames; /* Number of bootstrap p.t. frames.       */
+    unsigned long mfn_list;     /* VIRTUAL address of page-frame list.    */
+    unsigned long mod_start;    /* VIRTUAL address of pre-loaded module.  */
+    unsigned long mod_len;      /* Size (bytes) of pre-loaded module.     */
+    int8_t cmd_line[MAX_GUEST_CMDLINE];
+};
+typedef struct start_info start_info_t;
+
+/* New console union for dom0 introduced in 0x00030203. */
+#if __XEN_INTERFACE_VERSION__ < 0x00030203
+#define console_mfn    console.domU.mfn
+#define console_evtchn console.domU.evtchn
+#endif
+
+/* These flags are passed in the 'flags' field of start_info_t. */
+#define SIF_PRIVILEGED    (1<<0)  /* Is the domain privileged? */
+#define SIF_INITDOMAIN    (1<<1)  /* Is this the initial control domain? */
+#define SIF_PM_MASK       (0xFF<<8) /* reserve 1 byte for xen-pm options */
+
+typedef struct dom0_vga_console_info {
+    uint8_t video_type; /* DOM0_VGA_CONSOLE_??? */
+#define XEN_VGATYPE_TEXT_MODE_3 0x03
+#define XEN_VGATYPE_VESA_LFB    0x23
+
+    union {
+        struct {
+            /* Font height, in pixels. */
+            uint16_t font_height;
+            /* Cursor location (column, row). */
+            uint16_t cursor_x, cursor_y;
+            /* Number of rows and columns (dimensions in characters). */
+            uint16_t rows, columns;
+        } text_mode_3;
+
+        struct {
+            /* Width and height, in pixels. */
+            uint16_t width, height;
+            /* Bytes per scan line. */
+            uint16_t bytes_per_line;
+            /* Bits per pixel. */
+            uint16_t bits_per_pixel;
+            /* LFB physical address, and size (in units of 64kB). */
+            uint32_t lfb_base;
+            uint32_t lfb_size;
+            /* RGB mask offsets and sizes, as defined by VBE 1.2+ */
+            uint8_t  red_pos, red_size;
+            uint8_t  green_pos, green_size;
+            uint8_t  blue_pos, blue_size;
+            uint8_t  rsvd_pos, rsvd_size;
+#if __XEN_INTERFACE_VERSION__ >= 0x00030206
+            /* VESA capabilities (offset 0xa, VESA command 0x4f00). */
+            uint32_t gbl_caps;
+            /* Mode attributes (offset 0x0, VESA command 0x4f01). */
+            uint16_t mode_attrs;
+#endif
+        } vesa_lfb;
+    } u;
+} dom0_vga_console_info_t;
+#define xen_vga_console_info dom0_vga_console_info
+#define xen_vga_console_info_t dom0_vga_console_info_t
+
+typedef uint8_t xen_domain_handle_t[16];
+
+/* Turn a plain number into a C unsigned long constant. */
+#define __mk_unsigned_long(x) x ## UL
+#define mk_unsigned_long(x) __mk_unsigned_long(x)
+
+__DEFINE_XEN_GUEST_HANDLE(uint8,  uint8_t);
+__DEFINE_XEN_GUEST_HANDLE(uint16, uint16_t);
+__DEFINE_XEN_GUEST_HANDLE(uint32, uint32_t);
+__DEFINE_XEN_GUEST_HANDLE(uint64, uint64_t);
+
+#else /* __ASSEMBLY__ */
+
+/* In assembly code we cannot use C numeric constant suffixes. */
+#define mk_unsigned_long(x) x
+
+#endif /* !__ASSEMBLY__ */
+
+/* Default definitions for macros used by domctl/sysctl. */
+#if defined(__XEN__) || defined(__XEN_TOOLS__)
+#ifndef uint64_aligned_t
+#define uint64_aligned_t uint64_t
+#endif
+#ifndef XEN_GUEST_HANDLE_64
+#define XEN_GUEST_HANDLE_64(name) XEN_GUEST_HANDLE(name)
+#endif
+#endif
+
+#endif /* __XEN_PUBLIC_XEN_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/xenoprof.h linux-2.6.18-xen-3.3.0/include/xen/interface/xenoprof.h
--- linux-2.6.18.8/include/xen/interface/xenoprof.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/xenoprof.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,138 @@
+/******************************************************************************
+ * xenoprof.h
+ * 
+ * Interface for enabling system wide profiling based on hardware performance
+ * counters
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Copyright (C) 2005 Hewlett-Packard Co.
+ * Written by Aravind Menon & Jose Renato Santos
+ */
+
+#ifndef __XEN_PUBLIC_XENOPROF_H__
+#define __XEN_PUBLIC_XENOPROF_H__
+
+#include "xen.h"
+
+/*
+ * Commands to HYPERVISOR_xenoprof_op().
+ */
+#define XENOPROF_init                0
+#define XENOPROF_reset_active_list   1
+#define XENOPROF_reset_passive_list  2
+#define XENOPROF_set_active          3
+#define XENOPROF_set_passive         4
+#define XENOPROF_reserve_counters    5
+#define XENOPROF_counter             6
+#define XENOPROF_setup_events        7
+#define XENOPROF_enable_virq         8
+#define XENOPROF_start               9
+#define XENOPROF_stop               10
+#define XENOPROF_disable_virq       11
+#define XENOPROF_release_counters   12
+#define XENOPROF_shutdown           13
+#define XENOPROF_get_buffer         14
+#define XENOPROF_set_backtrace      15
+#define XENOPROF_last_op            15
+
+#define MAX_OPROF_EVENTS    32
+#define MAX_OPROF_DOMAINS   25
+#define XENOPROF_CPU_TYPE_SIZE 64
+
+/* Xenoprof performance events (not Xen events) */
+struct event_log {
+    uint64_t eip;
+    uint8_t mode;
+    uint8_t event;
+};
+
+/* PC value that indicates a special code */
+#define XENOPROF_ESCAPE_CODE ~0UL
+/* Transient events for the xenoprof->oprofile cpu buf */
+#define XENOPROF_TRACE_BEGIN 1
+
+/* Xenoprof buffer shared between Xen and domain - 1 per VCPU */
+struct xenoprof_buf {
+    uint32_t event_head;
+    uint32_t event_tail;
+    uint32_t event_size;
+    uint32_t vcpu_id;
+    uint64_t xen_samples;
+    uint64_t kernel_samples;
+    uint64_t user_samples;
+    uint64_t lost_samples;
+    struct event_log event_log[1];
+};
+#ifndef __XEN__
+typedef struct xenoprof_buf xenoprof_buf_t;
+DEFINE_XEN_GUEST_HANDLE(xenoprof_buf_t);
+#endif
+
+struct xenoprof_init {
+    int32_t  num_events;
+    int32_t  is_primary;
+    char cpu_type[XENOPROF_CPU_TYPE_SIZE];
+};
+typedef struct xenoprof_init xenoprof_init_t;
+DEFINE_XEN_GUEST_HANDLE(xenoprof_init_t);
+
+struct xenoprof_get_buffer {
+    int32_t  max_samples;
+    int32_t  nbuf;
+    int32_t  bufsize;
+    uint64_t buf_gmaddr;
+};
+typedef struct xenoprof_get_buffer xenoprof_get_buffer_t;
+DEFINE_XEN_GUEST_HANDLE(xenoprof_get_buffer_t);
+
+struct xenoprof_counter {
+    uint32_t ind;
+    uint64_t count;
+    uint32_t enabled;
+    uint32_t event;
+    uint32_t hypervisor;
+    uint32_t kernel;
+    uint32_t user;
+    uint64_t unit_mask;
+};
+typedef struct xenoprof_counter xenoprof_counter_t;
+DEFINE_XEN_GUEST_HANDLE(xenoprof_counter_t);
+
+typedef struct xenoprof_passive {
+    uint16_t domain_id;
+    int32_t  max_samples;
+    int32_t  nbuf;
+    int32_t  bufsize;
+    uint64_t buf_gmaddr;
+} xenoprof_passive_t;
+DEFINE_XEN_GUEST_HANDLE(xenoprof_passive_t);
+
+
+#endif /* __XEN_PUBLIC_XENOPROF_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/xsm/acm.h linux-2.6.18-xen-3.3.0/include/xen/interface/xsm/acm.h
--- linux-2.6.18.8/include/xen/interface/xsm/acm.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/xsm/acm.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,235 @@
+/*
+ * acm.h: Xen access control module interface defintions
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Reiner Sailer <sailer@watson.ibm.com>
+ * Copyright (c) 2005, International Business Machines Corporation.
+ */
+
+#ifndef _XEN_PUBLIC_ACM_H
+#define _XEN_PUBLIC_ACM_H
+
+#include "../xen.h"
+
+/* if ACM_DEBUG defined, all hooks should
+ * print a short trace message (comment it out
+ * when not in testing mode )
+ */
+/* #define ACM_DEBUG */
+
+#ifdef ACM_DEBUG
+#  define printkd(fmt, args...) printk(fmt,## args)
+#else
+#  define printkd(fmt, args...)
+#endif
+
+/* default ssid reference value if not supplied */
+#define ACM_DEFAULT_SSID  0x0
+#define ACM_DEFAULT_LOCAL_SSID  0x0
+
+/* Internal ACM ERROR types */
+#define ACM_OK     0
+#define ACM_UNDEF   -1
+#define ACM_INIT_SSID_ERROR  -2
+#define ACM_INIT_SOID_ERROR  -3
+#define ACM_ERROR          -4
+
+/* External ACCESS DECISIONS */
+#define ACM_ACCESS_PERMITTED        0
+#define ACM_ACCESS_DENIED           -111
+#define ACM_NULL_POINTER_ERROR      -200
+
+/*
+   Error codes reported in when trying to test for a new policy
+   These error codes are reported in an array of tuples where
+   each error code is followed by a parameter describing the error
+   more closely, such as a domain id.
+*/
+#define ACM_EVTCHN_SHARING_VIOLATION       0x100
+#define ACM_GNTTAB_SHARING_VIOLATION       0x101
+#define ACM_DOMAIN_LOOKUP                  0x102
+#define ACM_CHWALL_CONFLICT                0x103
+#define ACM_SSIDREF_IN_USE                 0x104
+
+
+/* primary policy in lower 4 bits */
+#define ACM_NULL_POLICY 0
+#define ACM_CHINESE_WALL_POLICY 1
+#define ACM_SIMPLE_TYPE_ENFORCEMENT_POLICY 2
+#define ACM_POLICY_UNDEFINED 15
+
+/* combinations have secondary policy component in higher 4bit */
+#define ACM_CHINESE_WALL_AND_SIMPLE_TYPE_ENFORCEMENT_POLICY \
+    ((ACM_SIMPLE_TYPE_ENFORCEMENT_POLICY << 4) | ACM_CHINESE_WALL_POLICY)
+
+/* policy: */
+#define ACM_POLICY_NAME(X) \
+ ((X) == (ACM_NULL_POLICY)) ? "NULL" :                        \
+    ((X) == (ACM_CHINESE_WALL_POLICY)) ? "CHINESE WALL" :        \
+    ((X) == (ACM_SIMPLE_TYPE_ENFORCEMENT_POLICY)) ? "SIMPLE TYPE ENFORCEMENT" : \
+    ((X) == (ACM_CHINESE_WALL_AND_SIMPLE_TYPE_ENFORCEMENT_POLICY)) ? "CHINESE WALL AND SIMPLE TYPE ENFORCEMENT" : \
+     "UNDEFINED"
+
+/* the following policy versions must be increased
+ * whenever the interpretation of the related
+ * policy's data structure changes
+ */
+#define ACM_POLICY_VERSION 4
+#define ACM_CHWALL_VERSION 1
+#define ACM_STE_VERSION  1
+
+/* defines a ssid reference used by xen */
+typedef uint32_t ssidref_t;
+
+/* hooks that are known to domains */
+#define ACMHOOK_none          0
+#define ACMHOOK_sharing       1
+#define ACMHOOK_authorization 2
+#define ACMHOOK_conflictset   3
+
+/* -------security policy relevant type definitions-------- */
+
+/* type identifier; compares to "equal" or "not equal" */
+typedef uint16_t domaintype_t;
+
+/* CHINESE WALL POLICY DATA STRUCTURES
+ *
+ * current accumulated conflict type set:
+ * When a domain is started and has a type that is in
+ * a conflict set, the conflicting types are incremented in
+ * the aggregate set. When a domain is destroyed, the 
+ * conflicting types to its type are decremented.
+ * If a domain has multiple types, this procedure works over
+ * all those types.
+ *
+ * conflict_aggregate_set[i] holds the number of
+ *   running domains that have a conflict with type i.
+ *
+ * running_types[i] holds the number of running domains
+ *        that include type i in their ssidref-referenced type set
+ *
+ * conflict_sets[i][j] is "0" if type j has no conflict
+ *    with type i and is "1" otherwise.
+ */
+/* high-16 = version, low-16 = check magic */
+#define ACM_MAGIC  0x0001debc
+
+/* size of the SHA1 hash identifying the XML policy from which the
+   binary policy was created */
+#define ACM_SHA1_HASH_SIZE    20
+
+/* each offset in bytes from start of the struct they
+ * are part of */
+
+/* V3 of the policy buffer aded a version structure */
+struct acm_policy_version
+{
+    uint32_t major;
+    uint32_t minor;
+};
+
+
+/* each buffer consists of all policy information for
+ * the respective policy given in the policy code
+ *
+ * acm_policy_buffer, acm_chwall_policy_buffer,
+ * and acm_ste_policy_buffer need to stay 32-bit aligned
+ * because we create binary policies also with external
+ * tools that assume packed representations (e.g. the java tool)
+ */
+struct acm_policy_buffer {
+    uint32_t magic;
+    uint32_t policy_version; /* ACM_POLICY_VERSION */
+    uint32_t len;
+    uint32_t policy_reference_offset;
+    uint32_t primary_policy_code;
+    uint32_t primary_buffer_offset;
+    uint32_t secondary_policy_code;
+    uint32_t secondary_buffer_offset;
+    struct acm_policy_version xml_pol_version; /* add in V3 */
+    uint8_t xml_policy_hash[ACM_SHA1_HASH_SIZE]; /* added in V4 */
+};
+
+
+struct acm_policy_reference_buffer {
+    uint32_t len;
+};
+
+struct acm_chwall_policy_buffer {
+    uint32_t policy_version; /* ACM_CHWALL_VERSION */
+    uint32_t policy_code;
+    uint32_t chwall_max_types;
+    uint32_t chwall_max_ssidrefs;
+    uint32_t chwall_max_conflictsets;
+    uint32_t chwall_ssid_offset;
+    uint32_t chwall_conflict_sets_offset;
+    uint32_t chwall_running_types_offset;
+    uint32_t chwall_conflict_aggregate_offset;
+};
+
+struct acm_ste_policy_buffer {
+    uint32_t policy_version; /* ACM_STE_VERSION */
+    uint32_t policy_code;
+    uint32_t ste_max_types;
+    uint32_t ste_max_ssidrefs;
+    uint32_t ste_ssid_offset;
+};
+
+struct acm_stats_buffer {
+    uint32_t magic;
+    uint32_t len;
+    uint32_t primary_policy_code;
+    uint32_t primary_stats_offset;
+    uint32_t secondary_policy_code;
+    uint32_t secondary_stats_offset;
+};
+
+struct acm_ste_stats_buffer {
+    uint32_t ec_eval_count;
+    uint32_t gt_eval_count;
+    uint32_t ec_denied_count;
+    uint32_t gt_denied_count;
+    uint32_t ec_cachehit_count;
+    uint32_t gt_cachehit_count;
+};
+
+struct acm_ssid_buffer {
+    uint32_t len;
+    ssidref_t ssidref;
+    uint32_t policy_reference_offset;
+    uint32_t primary_policy_code;
+    uint32_t primary_max_types;
+    uint32_t primary_types_offset;
+    uint32_t secondary_policy_code;
+    uint32_t secondary_max_types;
+    uint32_t secondary_types_offset;
+};
+
+#endif
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/xsm/acm_ops.h linux-2.6.18-xen-3.3.0/include/xen/interface/xsm/acm_ops.h
--- linux-2.6.18.8/include/xen/interface/xsm/acm_ops.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/xsm/acm_ops.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,159 @@
+/*
+ * acm_ops.h: Xen access control module hypervisor commands
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Reiner Sailer <sailer@watson.ibm.com>
+ * Copyright (c) 2005,2006 International Business Machines Corporation.
+ */
+
+#ifndef __XEN_PUBLIC_ACM_OPS_H__
+#define __XEN_PUBLIC_ACM_OPS_H__
+
+#include "../xen.h"
+#include "acm.h"
+
+/*
+ * Make sure you increment the interface version whenever you modify this file!
+ * This makes sure that old versions of acm tools will stop working in a
+ * well-defined way (rather than crashing the machine, for instance).
+ */
+#define ACM_INTERFACE_VERSION   0xAAAA000A
+
+/************************************************************************/
+
+/*
+ * Prototype for this hypercall is:
+ *  int acm_op(int cmd, void *args)
+ * @cmd  == ACMOP_??? (access control module operation).
+ * @args == Operation-specific extra arguments (NULL if none).
+ */
+
+
+#define ACMOP_setpolicy         1
+struct acm_setpolicy {
+    /* IN */
+    XEN_GUEST_HANDLE_64(void) pushcache;
+    uint32_t pushcache_size;
+};
+
+
+#define ACMOP_getpolicy         2
+struct acm_getpolicy {
+    /* IN */
+    XEN_GUEST_HANDLE_64(void) pullcache;
+    uint32_t pullcache_size;
+};
+
+
+#define ACMOP_dumpstats         3
+struct acm_dumpstats {
+    /* IN */
+    XEN_GUEST_HANDLE_64(void) pullcache;
+    uint32_t pullcache_size;
+};
+
+
+#define ACMOP_getssid           4
+#define ACM_GETBY_ssidref  1
+#define ACM_GETBY_domainid 2
+struct acm_getssid {
+    /* IN */
+    uint32_t get_ssid_by; /* ACM_GETBY_* */
+    union {
+        domaintype_t domainid;
+        ssidref_t    ssidref;
+    } id;
+    XEN_GUEST_HANDLE_64(void) ssidbuf;
+    uint32_t ssidbuf_size;
+};
+
+#define ACMOP_getdecision      5
+struct acm_getdecision {
+    /* IN */
+    uint32_t get_decision_by1; /* ACM_GETBY_* */
+    uint32_t get_decision_by2; /* ACM_GETBY_* */
+    union {
+        domaintype_t domainid;
+        ssidref_t    ssidref;
+    } id1;
+    union {
+        domaintype_t domainid;
+        ssidref_t    ssidref;
+    } id2;
+    uint32_t hook;
+    /* OUT */
+    uint32_t acm_decision;
+};
+
+
+#define ACMOP_chgpolicy        6
+struct acm_change_policy {
+    /* IN */
+    XEN_GUEST_HANDLE_64(void) policy_pushcache;
+    uint32_t policy_pushcache_size;
+    XEN_GUEST_HANDLE_64(void) del_array;
+    uint32_t delarray_size;
+    XEN_GUEST_HANDLE_64(void) chg_array;
+    uint32_t chgarray_size;
+    /* OUT */
+    /* array with error code */
+    XEN_GUEST_HANDLE_64(void) err_array;
+    uint32_t errarray_size;
+};
+
+#define ACMOP_relabeldoms       7
+struct acm_relabel_doms {
+    /* IN */
+    XEN_GUEST_HANDLE_64(void) relabel_map;
+    uint32_t relabel_map_size;
+    /* OUT */
+    XEN_GUEST_HANDLE_64(void) err_array;
+    uint32_t errarray_size;
+};
+
+/* future interface to Xen */
+struct xen_acmctl {
+    uint32_t cmd;
+    uint32_t interface_version;
+    union {
+        struct acm_setpolicy     setpolicy;
+        struct acm_getpolicy     getpolicy;
+        struct acm_dumpstats     dumpstats;
+        struct acm_getssid       getssid;
+        struct acm_getdecision   getdecision;
+        struct acm_change_policy change_policy;
+        struct acm_relabel_doms  relabel_doms;
+    } u;
+};
+
+typedef struct xen_acmctl xen_acmctl_t;
+DEFINE_XEN_GUEST_HANDLE(xen_acmctl_t);
+
+#endif /* __XEN_PUBLIC_ACM_OPS_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff -rpuN linux-2.6.18.8/include/xen/interface/xsm/flask_op.h linux-2.6.18-xen-3.3.0/include/xen/interface/xsm/flask_op.h
--- linux-2.6.18.8/include/xen/interface/xsm/flask_op.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/interface/xsm/flask_op.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,45 @@
+/*
+ *  This file contains the flask_op hypercall commands and definitions.
+ *
+ *  Author:  George Coker, <gscoker@alpha.ncsc.mil>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License version 2,
+ *  as published by the Free Software Foundation.
+ */
+
+#ifndef __FLASK_OP_H__
+#define __FLASK_OP_H__
+
+#define FLASK_LOAD              1
+#define FLASK_GETENFORCE        2
+#define FLASK_SETENFORCE        3
+#define FLASK_CONTEXT_TO_SID    4
+#define FLASK_SID_TO_CONTEXT    5
+#define FLASK_ACCESS            6
+#define FLASK_CREATE            7
+#define FLASK_RELABEL           8
+#define FLASK_USER              9
+#define FLASK_POLICYVERS        10
+#define FLASK_GETBOOL           11
+#define FLASK_SETBOOL           12
+#define FLASK_COMMITBOOLS       13
+#define FLASK_MLS               14
+#define FLASK_DISABLE           15
+#define FLASK_GETAVC_THRESHOLD  16
+#define FLASK_SETAVC_THRESHOLD  17
+#define FLASK_AVC_HASHSTATS     18
+#define FLASK_AVC_CACHESTATS    19
+#define FLASK_MEMBER            20
+
+#define FLASK_LAST              FLASK_MEMBER
+
+typedef struct flask_op {
+    uint32_t  cmd;
+    uint32_t  size;
+    char      *buf;
+} flask_op_t;
+
+DEFINE_XEN_GUEST_HANDLE(flask_op_t);
+
+#endif
diff -rpuN linux-2.6.18.8/include/xen/pcifront.h linux-2.6.18-xen-3.3.0/include/xen/pcifront.h
--- linux-2.6.18.8/include/xen/pcifront.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/pcifront.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,83 @@
+/*
+ * PCI Frontend - arch-dependendent declarations
+ *
+ *   Author: Ryan Wilson <hap9@epoch.ncsc.mil>
+ */
+#ifndef __XEN_ASM_PCIFRONT_H__
+#define __XEN_ASM_PCIFRONT_H__
+
+#include <linux/spinlock.h>
+
+#ifdef __KERNEL__
+
+#ifndef __ia64__
+
+struct pcifront_device;
+struct pci_bus;
+
+struct pcifront_sd {
+	int domain;
+	struct pcifront_device *pdev;
+};
+
+static inline struct pcifront_device *
+pcifront_get_pdev(struct pcifront_sd *sd)
+{
+	return sd->pdev;
+}
+
+static inline void pcifront_init_sd(struct pcifront_sd *sd,
+				    unsigned int domain, unsigned int bus,
+				    struct pcifront_device *pdev)
+{
+	sd->domain = domain;
+	sd->pdev = pdev;
+}
+
+#if defined(CONFIG_PCI_DOMAINS)
+static inline int pci_domain_nr(struct pci_bus *bus)
+{
+	struct pcifront_sd *sd = bus->sysdata;
+	return sd->domain;
+}
+static inline int pci_proc_domain(struct pci_bus *bus)
+{
+	return pci_domain_nr(bus);
+}
+#endif /* CONFIG_PCI_DOMAINS */
+
+static inline void pcifront_setup_root_resources(struct pci_bus *bus,
+						 struct pcifront_sd *sd)
+{
+}
+
+#else /* __ia64__ */
+
+#include <linux/acpi.h>
+#include <asm/pci.h>
+#define pcifront_sd pci_controller
+
+extern void xen_add_resource(struct pci_controller *, unsigned int,
+			     unsigned int, struct acpi_resource *);
+extern void xen_pcibios_setup_root_windows(struct pci_bus *,
+					   struct pci_controller *);
+
+static inline struct pcifront_device *
+pcifront_get_pdev(struct pcifront_sd *sd)
+{
+	return (struct pcifront_device *)sd->platform_data;
+}
+
+static inline void pcifront_setup_root_resources(struct pci_bus *bus,
+						 struct pcifront_sd *sd)
+{
+	xen_pcibios_setup_root_windows(bus, sd);
+}
+
+#endif /* __ia64__ */
+
+extern struct rw_semaphore pci_bus_sem;
+
+#endif /* __KERNEL__ */
+
+#endif /* __XEN_ASM_PCIFRONT_H__ */
diff -rpuN linux-2.6.18.8/include/xen/public/evtchn.h linux-2.6.18-xen-3.3.0/include/xen/public/evtchn.h
--- linux-2.6.18.8/include/xen/public/evtchn.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/public/evtchn.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,88 @@
+/******************************************************************************
+ * evtchn.h
+ * 
+ * Interface to /dev/xen/evtchn.
+ * 
+ * Copyright (c) 2003-2005, K A Fraser
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __LINUX_PUBLIC_EVTCHN_H__
+#define __LINUX_PUBLIC_EVTCHN_H__
+
+/*
+ * Bind a fresh port to VIRQ @virq.
+ * Return allocated port.
+ */
+#define IOCTL_EVTCHN_BIND_VIRQ				\
+	_IOC(_IOC_NONE, 'E', 0, sizeof(struct ioctl_evtchn_bind_virq))
+struct ioctl_evtchn_bind_virq {
+	unsigned int virq;
+};
+
+/*
+ * Bind a fresh port to remote <@remote_domain, @remote_port>.
+ * Return allocated port.
+ */
+#define IOCTL_EVTCHN_BIND_INTERDOMAIN			\
+	_IOC(_IOC_NONE, 'E', 1, sizeof(struct ioctl_evtchn_bind_interdomain))
+struct ioctl_evtchn_bind_interdomain {
+	unsigned int remote_domain, remote_port;
+};
+
+/*
+ * Allocate a fresh port for binding to @remote_domain.
+ * Return allocated port.
+ */
+#define IOCTL_EVTCHN_BIND_UNBOUND_PORT			\
+	_IOC(_IOC_NONE, 'E', 2, sizeof(struct ioctl_evtchn_bind_unbound_port))
+struct ioctl_evtchn_bind_unbound_port {
+	unsigned int remote_domain;
+};
+
+/*
+ * Unbind previously allocated @port.
+ */
+#define IOCTL_EVTCHN_UNBIND				\
+	_IOC(_IOC_NONE, 'E', 3, sizeof(struct ioctl_evtchn_unbind))
+struct ioctl_evtchn_unbind {
+	unsigned int port;
+};
+
+/*
+ * Unbind previously allocated @port.
+ */
+#define IOCTL_EVTCHN_NOTIFY				\
+	_IOC(_IOC_NONE, 'E', 4, sizeof(struct ioctl_evtchn_notify))
+struct ioctl_evtchn_notify {
+	unsigned int port;
+};
+
+/* Clear and reinitialise the event buffer. Clear error condition. */
+#define IOCTL_EVTCHN_RESET				\
+	_IOC(_IOC_NONE, 'E', 5, 0)
+
+#endif /* __LINUX_PUBLIC_EVTCHN_H__ */
diff -rpuN linux-2.6.18.8/include/xen/public/gntdev.h linux-2.6.18-xen-3.3.0/include/xen/public/gntdev.h
--- linux-2.6.18.8/include/xen/public/gntdev.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/public/gntdev.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,119 @@
+/******************************************************************************
+ * gntdev.h
+ * 
+ * Interface to /dev/xen/gntdev.
+ * 
+ * Copyright (c) 2007, D G Murray
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __LINUX_PUBLIC_GNTDEV_H__
+#define __LINUX_PUBLIC_GNTDEV_H__
+
+struct ioctl_gntdev_grant_ref {
+	/* The domain ID of the grant to be mapped. */
+	uint32_t domid;
+	/* The grant reference of the grant to be mapped. */
+	uint32_t ref;
+};
+
+/*
+ * Inserts the grant references into the mapping table of an instance
+ * of gntdev. N.B. This does not perform the mapping, which is deferred
+ * until mmap() is called with @index as the offset.
+ */
+#define IOCTL_GNTDEV_MAP_GRANT_REF \
+_IOC(_IOC_NONE, 'G', 0, sizeof(struct ioctl_gntdev_map_grant_ref))
+struct ioctl_gntdev_map_grant_ref {
+	/* IN parameters */
+	/* The number of grants to be mapped. */
+	uint32_t count;
+	uint32_t pad;
+	/* OUT parameters */
+	/* The offset to be used on a subsequent call to mmap(). */
+	uint64_t index;
+	/* Variable IN parameter. */
+	/* Array of grant references, of size @count. */
+	struct ioctl_gntdev_grant_ref refs[1];
+};
+
+/*
+ * Removes the grant references from the mapping table of an instance of
+ * of gntdev. N.B. munmap() must be called on the relevant virtual address(es)
+ * before this ioctl is called, or an error will result.
+ */
+#define IOCTL_GNTDEV_UNMAP_GRANT_REF \
+_IOC(_IOC_NONE, 'G', 1, sizeof(struct ioctl_gntdev_unmap_grant_ref))       
+struct ioctl_gntdev_unmap_grant_ref {
+	/* IN parameters */
+	/* The offset was returned by the corresponding map operation. */
+	uint64_t index;
+	/* The number of pages to be unmapped. */
+	uint32_t count;
+	uint32_t pad;
+};
+
+/*
+ * Returns the offset in the driver's address space that corresponds
+ * to @vaddr. This can be used to perform a munmap(), followed by an
+ * UNMAP_GRANT_REF ioctl, where no state about the offset is retained by
+ * the caller. The number of pages that were allocated at the same time as
+ * @vaddr is returned in @count.
+ *
+ * N.B. Where more than one page has been mapped into a contiguous range, the
+ *      supplied @vaddr must correspond to the start of the range; otherwise
+ *      an error will result. It is only possible to munmap() the entire
+ *      contiguously-allocated range at once, and not any subrange thereof.
+ */
+#define IOCTL_GNTDEV_GET_OFFSET_FOR_VADDR \
+_IOC(_IOC_NONE, 'G', 2, sizeof(struct ioctl_gntdev_get_offset_for_vaddr))
+struct ioctl_gntdev_get_offset_for_vaddr {
+	/* IN parameters */
+	/* The virtual address of the first mapped page in a range. */
+	uint64_t vaddr;
+	/* OUT parameters */
+	/* The offset that was used in the initial mmap() operation. */
+	uint64_t offset;
+	/* The number of pages mapped in the VM area that begins at @vaddr. */
+	uint32_t count;
+	uint32_t pad;
+};
+
+/*
+ * Sets the maximum number of grants that may mapped at once by this gntdev
+ * instance.
+ *
+ * N.B. This must be called before any other ioctl is performed on the device.
+ */
+#define IOCTL_GNTDEV_SET_MAX_GRANTS \
+_IOC(_IOC_NONE, 'G', 3, sizeof(struct ioctl_gntdev_set_max_grants))
+struct ioctl_gntdev_set_max_grants {
+	/* IN parameter */
+	/* The maximum number of grants that may be mapped at once. */
+	uint32_t count;
+};
+
+#endif /* __LINUX_PUBLIC_GNTDEV_H__ */
diff -rpuN linux-2.6.18.8/include/xen/public/privcmd.h linux-2.6.18-xen-3.3.0/include/xen/public/privcmd.h
--- linux-2.6.18.8/include/xen/public/privcmd.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/public/privcmd.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,79 @@
+/******************************************************************************
+ * privcmd.h
+ * 
+ * Interface to /proc/xen/privcmd.
+ * 
+ * Copyright (c) 2003-2005, K A Fraser
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __LINUX_PUBLIC_PRIVCMD_H__
+#define __LINUX_PUBLIC_PRIVCMD_H__
+
+#include <linux/types.h>
+
+#ifndef __user
+#define __user
+#endif
+
+typedef struct privcmd_hypercall
+{
+	__u64 op;
+	__u64 arg[5];
+} privcmd_hypercall_t;
+
+typedef struct privcmd_mmap_entry {
+	__u64 va;
+	__u64 mfn;
+	__u64 npages;
+} privcmd_mmap_entry_t; 
+
+typedef struct privcmd_mmap {
+	int num;
+	domid_t dom; /* target domain */
+	privcmd_mmap_entry_t __user *entry;
+} privcmd_mmap_t; 
+
+typedef struct privcmd_mmapbatch {
+	int num;     /* number of pages to populate */
+	domid_t dom; /* target domain */
+	__u64 addr;  /* virtual address */
+	xen_pfn_t __user *arr; /* array of mfns - top nibble set on err */
+} privcmd_mmapbatch_t; 
+
+/*
+ * @cmd: IOCTL_PRIVCMD_HYPERCALL
+ * @arg: &privcmd_hypercall_t
+ * Return: Value returned from execution of the specified hypercall.
+ */
+#define IOCTL_PRIVCMD_HYPERCALL					\
+	_IOC(_IOC_NONE, 'P', 0, sizeof(privcmd_hypercall_t))
+#define IOCTL_PRIVCMD_MMAP					\
+	_IOC(_IOC_NONE, 'P', 2, sizeof(privcmd_mmap_t))
+#define IOCTL_PRIVCMD_MMAPBATCH					\
+	_IOC(_IOC_NONE, 'P', 3, sizeof(privcmd_mmapbatch_t))
+
+#endif /* __LINUX_PUBLIC_PRIVCMD_H__ */
diff -rpuN linux-2.6.18.8/include/xen/xenbus.h linux-2.6.18-xen-3.3.0/include/xen/xenbus.h
--- linux-2.6.18.8/include/xen/xenbus.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/xenbus.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,307 @@
+/******************************************************************************
+ * xenbus.h
+ *
+ * Talks to Xen Store to figure out what devices we have.
+ *
+ * Copyright (C) 2005 Rusty Russell, IBM Corporation
+ * Copyright (C) 2005 XenSource Ltd.
+ * 
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef _XEN_XENBUS_H
+#define _XEN_XENBUS_H
+
+#include <linux/device.h>
+#include <linux/notifier.h>
+#include <linux/mutex.h>
+#include <linux/completion.h>
+#include <linux/init.h>
+#include <linux/err.h>
+#include <xen/interface/xen.h>
+#include <xen/interface/grant_table.h>
+#include <xen/interface/io/xenbus.h>
+#include <xen/interface/io/xs_wire.h>
+
+/* Register callback to watch this node. */
+struct xenbus_watch
+{
+	struct list_head list;
+
+	/* Path being watched. */
+	const char *node;
+
+	/* Callback (executed in a process context with no locks held). */
+	void (*callback)(struct xenbus_watch *,
+			 const char **vec, unsigned int len);
+
+	/* See XBWF_ definitions below. */
+	unsigned long flags;
+};
+
+/*
+ * Execute callback in its own kthread. Useful if the callback is long
+ * running or heavily serialised, to avoid taking out the main xenwatch thread
+ * for a long period of time (or even unwittingly causing a deadlock).
+ */
+#define XBWF_new_thread	1
+
+/* A xenbus device. */
+struct xenbus_device {
+	const char *devicetype;
+	const char *nodename;
+	const char *otherend;
+	int otherend_id;
+	struct xenbus_watch otherend_watch;
+	struct device dev;
+	enum xenbus_state state;
+	struct completion down;
+};
+
+static inline struct xenbus_device *to_xenbus_device(struct device *dev)
+{
+	return container_of(dev, struct xenbus_device, dev);
+}
+
+struct xenbus_device_id
+{
+	/* .../device/<device_type>/<identifier> */
+	char devicetype[32]; 	/* General class of device. */
+};
+
+/* A xenbus driver. */
+struct xenbus_driver {
+	char *name;
+	struct module *owner;
+	const struct xenbus_device_id *ids;
+	int (*probe)(struct xenbus_device *dev,
+		     const struct xenbus_device_id *id);
+	void (*otherend_changed)(struct xenbus_device *dev,
+				 enum xenbus_state backend_state);
+	int (*remove)(struct xenbus_device *dev);
+	int (*suspend)(struct xenbus_device *dev);
+	int (*suspend_cancel)(struct xenbus_device *dev);
+	int (*resume)(struct xenbus_device *dev);
+	int (*uevent)(struct xenbus_device *, char **, int, char *, int);
+	struct device_driver driver;
+	int (*read_otherend_details)(struct xenbus_device *dev);
+	int (*is_ready)(struct xenbus_device *dev);
+};
+
+static inline struct xenbus_driver *to_xenbus_driver(struct device_driver *drv)
+{
+	return container_of(drv, struct xenbus_driver, driver);
+}
+
+int xenbus_register_frontend(struct xenbus_driver *drv);
+int xenbus_register_backend(struct xenbus_driver *drv);
+void xenbus_unregister_driver(struct xenbus_driver *drv);
+
+struct xenbus_transaction
+{
+	u32 id;
+};
+
+/* Nil transaction ID. */
+#define XBT_NIL ((struct xenbus_transaction) { 0 })
+
+char **xenbus_directory(struct xenbus_transaction t,
+			const char *dir, const char *node, unsigned int *num);
+void *xenbus_read(struct xenbus_transaction t,
+		  const char *dir, const char *node, unsigned int *len);
+int xenbus_write(struct xenbus_transaction t,
+		 const char *dir, const char *node, const char *string);
+int xenbus_mkdir(struct xenbus_transaction t,
+		 const char *dir, const char *node);
+int xenbus_exists(struct xenbus_transaction t,
+		  const char *dir, const char *node);
+int xenbus_rm(struct xenbus_transaction t, const char *dir, const char *node);
+int xenbus_transaction_start(struct xenbus_transaction *t);
+int xenbus_transaction_end(struct xenbus_transaction t, int abort);
+
+/* Single read and scanf: returns -errno or num scanned if > 0. */
+int xenbus_scanf(struct xenbus_transaction t,
+		 const char *dir, const char *node, const char *fmt, ...)
+	__attribute__((format(scanf, 4, 5)));
+
+/* Single printf and write: returns -errno or 0. */
+int xenbus_printf(struct xenbus_transaction t,
+		  const char *dir, const char *node, const char *fmt, ...)
+	__attribute__((format(printf, 4, 5)));
+
+/* Generic read function: NULL-terminated triples of name,
+ * sprintf-style type string, and pointer. Returns 0 or errno.*/
+int xenbus_gather(struct xenbus_transaction t, const char *dir, ...);
+
+/* notifer routines for when the xenstore comes up */
+int register_xenstore_notifier(struct notifier_block *nb);
+void unregister_xenstore_notifier(struct notifier_block *nb);
+
+int register_xenbus_watch(struct xenbus_watch *watch);
+void unregister_xenbus_watch(struct xenbus_watch *watch);
+void xs_suspend(void);
+void xs_resume(void);
+void xs_suspend_cancel(void);
+
+/* Used by xenbus_dev to borrow kernel's store connection. */
+void *xenbus_dev_request_and_reply(struct xsd_sockmsg *msg);
+
+/* Prepare for domain suspend: then resume or cancel the suspend. */
+void xenbus_suspend(void);
+void xenbus_resume(void);
+void xenbus_suspend_cancel(void);
+
+#define XENBUS_IS_ERR_READ(str) ({			\
+	if (!IS_ERR(str) && strlen(str) == 0) {		\
+		kfree(str);				\
+		str = ERR_PTR(-ERANGE);			\
+	}						\
+	IS_ERR(str);					\
+})
+
+#define XENBUS_EXIST_ERR(err) ((err) == -ENOENT || (err) == -ERANGE)
+
+
+/**
+ * Register a watch on the given path, using the given xenbus_watch structure
+ * for storage, and the given callback function as the callback.  Return 0 on
+ * success, or -errno on error.  On success, the given path will be saved as
+ * watch->node, and remains the caller's to free.  On error, watch->node will
+ * be NULL, the device will switch to XenbusStateClosing, and the error will
+ * be saved in the store.
+ */
+int xenbus_watch_path(struct xenbus_device *dev, const char *path,
+		      struct xenbus_watch *watch,
+		      void (*callback)(struct xenbus_watch *,
+				       const char **, unsigned int));
+
+
+/**
+ * Register a watch on the given path/path2, using the given xenbus_watch
+ * structure for storage, and the given callback function as the callback.
+ * Return 0 on success, or -errno on error.  On success, the watched path
+ * (path/path2) will be saved as watch->node, and becomes the caller's to
+ * kfree().  On error, watch->node will be NULL, so the caller has nothing to
+ * free, the device will switch to XenbusStateClosing, and the error will be
+ * saved in the store.
+ */
+int xenbus_watch_path2(struct xenbus_device *dev, const char *path,
+		       const char *path2, struct xenbus_watch *watch,
+		       void (*callback)(struct xenbus_watch *,
+					const char **, unsigned int));
+
+
+/**
+ * Advertise in the store a change of the given driver to the given new_state.
+ * Return 0 on success, or -errno on error.  On error, the device will switch
+ * to XenbusStateClosing, and the error will be saved in the store.
+ */
+int xenbus_switch_state(struct xenbus_device *dev, enum xenbus_state new_state);
+
+
+/**
+ * Grant access to the given ring_mfn to the peer of the given device.  Return
+ * 0 on success, or -errno on error.  On error, the device will switch to
+ * XenbusStateClosing, and the error will be saved in the store.
+ */
+int xenbus_grant_ring(struct xenbus_device *dev, unsigned long ring_mfn);
+
+
+/**
+ * Map a page of memory into this domain from another domain's grant table.
+ * xenbus_map_ring_valloc allocates a page of virtual address space, maps the
+ * page to that address, and sets *vaddr to that address.
+ * xenbus_map_ring does not allocate the virtual address space (you must do
+ * this yourself!). It only maps in the page to the specified address.
+ * Returns 0 on success, and GNTST_* (see xen/include/interface/grant_table.h)
+ * or -ENOMEM on error. If an error is returned, device will switch to
+ * XenbusStateClosing and the error message will be saved in XenStore.
+ */
+struct vm_struct *xenbus_map_ring_valloc(struct xenbus_device *dev,
+					 int gnt_ref);
+int xenbus_map_ring(struct xenbus_device *dev, int gnt_ref,
+			   grant_handle_t *handle, void *vaddr);
+
+
+/**
+ * Unmap a page of memory in this domain that was imported from another domain.
+ * Use xenbus_unmap_ring_vfree if you mapped in your memory with
+ * xenbus_map_ring_valloc (it will free the virtual address space).
+ * Returns 0 on success and returns GNTST_* on error
+ * (see xen/include/interface/grant_table.h).
+ */
+int xenbus_unmap_ring_vfree(struct xenbus_device *dev, struct vm_struct *);
+int xenbus_unmap_ring(struct xenbus_device *dev,
+		      grant_handle_t handle, void *vaddr);
+
+
+/**
+ * Allocate an event channel for the given xenbus_device, assigning the newly
+ * created local port to *port.  Return 0 on success, or -errno on error.  On
+ * error, the device will switch to XenbusStateClosing, and the error will be
+ * saved in the store.
+ */
+int xenbus_alloc_evtchn(struct xenbus_device *dev, int *port);
+
+
+/**
+ * Free an existing event channel. Returns 0 on success or -errno on error.
+ */
+int xenbus_free_evtchn(struct xenbus_device *dev, int port);
+
+
+/**
+ * Return the state of the driver rooted at the given store path, or
+ * XenbusStateUnknown if no state can be read.
+ */
+enum xenbus_state xenbus_read_driver_state(const char *path);
+
+
+/***
+ * Report the given negative errno into the store, along with the given
+ * formatted message.
+ */
+void xenbus_dev_error(struct xenbus_device *dev, int err, const char *fmt,
+		      ...);
+
+
+/***
+ * Equivalent to xenbus_dev_error(dev, err, fmt, args), followed by
+ * xenbus_switch_state(dev, NULL, XenbusStateClosing) to schedule an orderly
+ * closedown of this driver and its peer.
+ */
+void xenbus_dev_fatal(struct xenbus_device *dev, int err, const char *fmt,
+		      ...);
+
+int xenbus_dev_init(void);
+
+const char *xenbus_strstate(enum xenbus_state state);
+int xenbus_dev_is_online(struct xenbus_device *dev);
+int xenbus_frontend_closed(struct xenbus_device *dev);
+
+int xenbus_for_each_backend(void *arg, int (*fn)(struct device *, void *));
+int xenbus_for_each_frontend(void *arg, int (*fn)(struct device *, void *));
+
+#endif /* _XEN_XENBUS_H */
diff -rpuN linux-2.6.18.8/include/xen/xencomm.h linux-2.6.18-xen-3.3.0/include/xen/xencomm.h
--- linux-2.6.18.8/include/xen/xencomm.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/xencomm.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,77 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ * 
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ * 
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ *
+ * Copyright (C) IBM Corp. 2006
+ *
+ * Authors: Hollis Blanchard <hollisb@us.ibm.com>
+ *          Jerone Young <jyoung5@us.ibm.com>
+ */
+
+#ifndef _LINUX_XENCOMM_H_
+#define _LINUX_XENCOMM_H_
+
+#include <xen/interface/xencomm.h>
+
+#define XENCOMM_MINI_ADDRS 3
+struct xencomm_mini {
+	struct xencomm_desc _desc;
+	uint64_t address[XENCOMM_MINI_ADDRS];
+};
+
+/* To avoid additionnal virt to phys conversion, an opaque structure is
+   presented.  */
+struct xencomm_handle;
+
+extern void xencomm_free(struct xencomm_handle *desc);
+extern struct xencomm_handle *xencomm_map(void *ptr, unsigned long bytes);
+extern struct xencomm_handle *__xencomm_map_no_alloc(void *ptr,
+			unsigned long bytes,  struct xencomm_mini *xc_area);
+
+#if 0
+#define XENCOMM_MINI_ALIGNED(xc_desc, n)				\
+	struct xencomm_mini xc_desc ## _base[(n)]			\
+	__attribute__((__aligned__(sizeof(struct xencomm_mini))));	\
+	struct xencomm_mini* xc_desc = &xc_desc ## _base[0];
+#else
+/*
+ * gcc bug workaround:
+ * http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16660
+ * gcc doesn't handle properly stack variable with
+ * __attribute__((__align__(sizeof(struct xencomm_mini))))
+ */
+#define XENCOMM_MINI_ALIGNED(xc_desc, n)				\
+	unsigned char xc_desc ## _base[((n) + 1 ) *			\
+				       sizeof(struct xencomm_mini)];	\
+	struct xencomm_mini *xc_desc = (struct xencomm_mini*)		\
+		((unsigned long)xc_desc ## _base +			\
+		 (sizeof(struct xencomm_mini) -				\
+		  ((unsigned long)xc_desc ## _base) %			\
+		  sizeof(struct xencomm_mini)));
+#endif
+#define xencomm_map_no_alloc(ptr, bytes)			\
+	({XENCOMM_MINI_ALIGNED(xc_desc, 1);			\
+		__xencomm_map_no_alloc(ptr, bytes, xc_desc);})
+
+/* provided by architecture code: */
+extern unsigned long xencomm_vtop(unsigned long vaddr);
+
+static inline void *xencomm_pa(void *ptr)
+{
+	return (void *)xencomm_vtop((unsigned long)ptr);
+}
+
+#define xen_guest_handle(hnd)  ((hnd).p)
+
+#endif /* _LINUX_XENCOMM_H_ */
diff -rpuN linux-2.6.18.8/include/xen/xencons.h linux-2.6.18-xen-3.3.0/include/xen/xencons.h
--- linux-2.6.18.8/include/xen/xencons.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/xencons.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,17 @@
+#ifndef __ASM_XENCONS_H__
+#define __ASM_XENCONS_H__
+
+struct dom0_vga_console_info;
+void dom0_init_screen_info(const struct dom0_vga_console_info *, size_t);
+
+void xencons_force_flush(void);
+void xencons_resume(void);
+
+/* Interrupt work hooks. Receive data, or kick data out. */
+void xencons_rx(char *buf, unsigned len, struct pt_regs *regs);
+void xencons_tx(void);
+
+int xencons_ring_init(void);
+int xencons_ring_send(const char *data, unsigned len);
+
+#endif /* __ASM_XENCONS_H__ */
diff -rpuN linux-2.6.18.8/include/xen/xenoprof.h linux-2.6.18-xen-3.3.0/include/xen/xenoprof.h
--- linux-2.6.18.8/include/xen/xenoprof.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/xenoprof.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,42 @@
+/******************************************************************************
+ * xen/xenoprof.h
+ *
+ * Copyright (c) 2006 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+
+#ifndef __XEN_XENOPROF_H__
+#define __XEN_XENOPROF_H__
+#ifdef CONFIG_XEN
+
+#include <asm/xenoprof.h>
+
+struct oprofile_operations;
+int xenoprofile_init(struct oprofile_operations * ops);
+void xenoprofile_exit(void);
+
+struct xenoprof_shared_buffer {
+	char					*buffer;
+	struct xenoprof_arch_shared_buffer	arch;
+};
+#else
+#define xenoprofile_init(ops)	(-ENOSYS)
+#define xenoprofile_exit()	do { } while (0)
+
+#endif /* CONFIG_XEN */
+#endif /* __XEN_XENOPROF_H__ */
diff -rpuN linux-2.6.18.8/include/xen/xen_proc.h linux-2.6.18-xen-3.3.0/include/xen/xen_proc.h
--- linux-2.6.18.8/include/xen/xen_proc.h	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/include/xen/xen_proc.h	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,12 @@
+
+#ifndef __ASM_XEN_PROC_H__
+#define __ASM_XEN_PROC_H__
+
+#include <linux/proc_fs.h>
+
+extern struct proc_dir_entry *create_xen_proc_entry(
+	const char *name, mode_t mode);
+extern void remove_xen_proc_entry(
+	const char *name);
+
+#endif /* __ASM_XEN_PROC_H__ */
diff -rpuN linux-2.6.18.8/kernel/cpu.c linux-2.6.18-xen-3.3.0/kernel/cpu.c
--- linux-2.6.18.8/kernel/cpu.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/kernel/cpu.c	2008-08-21 11:36:07.000000000 +0200
@@ -48,7 +48,10 @@ EXPORT_SYMBOL_GPL(lock_cpu_hotplug);
 
 void unlock_cpu_hotplug(void)
 {
+/* cpufreq lock-takers fixed in mainline; shut up until dom0 kernel catches up*/
+#ifdef CONFIG_XEN
 	WARN_ON(recursive != current);
+#endif
 	if (recursive_depth) {
 		recursive_depth--;
 		return;
diff -rpuN linux-2.6.18.8/kernel/fork.c linux-2.6.18-xen-3.3.0/kernel/fork.c
--- linux-2.6.18.8/kernel/fork.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/kernel/fork.c	2008-08-21 11:36:07.000000000 +0200
@@ -276,6 +276,9 @@ static inline int dup_mmap(struct mm_str
 		if (retval)
 			goto out;
 	}
+#ifdef arch_dup_mmap
+	arch_dup_mmap(mm, oldmm);
+#endif
 	retval = 0;
 out:
 	up_write(&mm->mmap_sem);
diff -rpuN linux-2.6.18.8/kernel/hrtimer.c linux-2.6.18-xen-3.3.0/kernel/hrtimer.c
--- linux-2.6.18.8/kernel/hrtimer.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/kernel/hrtimer.c	2008-08-21 11:36:07.000000000 +0200
@@ -59,7 +59,7 @@ static ktime_t ktime_get(void)
  *
  * returns the time in ktime_t format
  */
-static ktime_t ktime_get_real(void)
+ktime_t ktime_get_real(void)
 {
 	struct timespec now;
 
diff -rpuN linux-2.6.18.8/kernel/irq/spurious.c linux-2.6.18-xen-3.3.0/kernel/irq/spurious.c
--- linux-2.6.18.8/kernel/irq/spurious.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/kernel/irq/spurious.c	2008-08-21 11:36:07.000000000 +0200
@@ -139,7 +139,8 @@ void note_interrupt(unsigned int irq, st
 		    irqreturn_t action_ret, struct pt_regs *regs)
 {
 	if (unlikely(action_ret != IRQ_HANDLED)) {
-		desc->irqs_unhandled++;
+		if (!irq_ignore_unhandled(irq))
+			desc->irqs_unhandled++;
 		if (unlikely(action_ret != IRQ_NONE))
 			report_bad_irq(irq, desc, action_ret);
 	}
diff -rpuN linux-2.6.18.8/kernel/Kconfig.preempt linux-2.6.18-xen-3.3.0/kernel/Kconfig.preempt
--- linux-2.6.18.8/kernel/Kconfig.preempt	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/kernel/Kconfig.preempt	2008-08-21 11:36:07.000000000 +0200
@@ -35,6 +35,7 @@ config PREEMPT_VOLUNTARY
 
 config PREEMPT
 	bool "Preemptible Kernel (Low-Latency Desktop)"
+	depends on !XEN
 	help
 	  This option reduces the latency of the kernel by making
 	  all kernel code (that is not executing in a critical section)
diff -rpuN linux-2.6.18.8/kernel/kexec.c linux-2.6.18-xen-3.3.0/kernel/kexec.c
--- linux-2.6.18.8/kernel/kexec.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/kernel/kexec.c	2008-08-21 11:36:07.000000000 +0200
@@ -330,13 +330,26 @@ static int kimage_is_destination_range(s
 	return 0;
 }
 
-static struct page *kimage_alloc_pages(gfp_t gfp_mask, unsigned int order)
+static struct page *kimage_alloc_pages(gfp_t gfp_mask, unsigned int order, unsigned long limit)
 {
 	struct page *pages;
 
 	pages = alloc_pages(gfp_mask, order);
 	if (pages) {
 		unsigned int count, i;
+#ifdef CONFIG_XEN
+		int address_bits;
+
+		if (limit == ~0UL)
+			address_bits = BITS_PER_LONG;
+		else
+			address_bits = long_log2(limit);
+
+		if (xen_limit_pages_to_max_mfn(pages, order, address_bits) < 0) {
+			__free_pages(pages, order);
+			return NULL;
+		}
+#endif
 		pages->mapping = NULL;
 		set_page_private(pages, order);
 		count = 1 << order;
@@ -355,6 +368,9 @@ static void kimage_free_pages(struct pag
 	count = 1 << order;
 	for (i = 0; i < count; i++)
 		ClearPageReserved(page + i);
+#ifdef CONFIG_XEN
+	xen_destroy_contiguous_region((unsigned long)page_address(page), order);
+#endif
 	__free_pages(page, order);
 }
 
@@ -400,10 +416,10 @@ static struct page *kimage_alloc_normal_
 	do {
 		unsigned long pfn, epfn, addr, eaddr;
 
-		pages = kimage_alloc_pages(GFP_KERNEL, order);
+		pages = kimage_alloc_pages(GFP_KERNEL, order, KEXEC_CONTROL_MEMORY_LIMIT);
 		if (!pages)
 			break;
-		pfn   = page_to_pfn(pages);
+		pfn   = kexec_page_to_pfn(pages);
 		epfn  = pfn + count;
 		addr  = pfn << PAGE_SHIFT;
 		eaddr = epfn << PAGE_SHIFT;
@@ -437,6 +453,7 @@ static struct page *kimage_alloc_normal_
 	return pages;
 }
 
+#ifndef CONFIG_XEN
 static struct page *kimage_alloc_crash_control_pages(struct kimage *image,
 						      unsigned int order)
 {
@@ -490,7 +507,7 @@ static struct page *kimage_alloc_crash_c
 		}
 		/* If I don't overlap any segments I have found my hole! */
 		if (i == image->nr_segments) {
-			pages = pfn_to_page(hole_start >> PAGE_SHIFT);
+			pages = kexec_pfn_to_page(hole_start >> PAGE_SHIFT);
 			break;
 		}
 	}
@@ -517,6 +534,13 @@ struct page *kimage_alloc_control_pages(
 
 	return pages;
 }
+#else /* !CONFIG_XEN */
+struct page *kimage_alloc_control_pages(struct kimage *image,
+					 unsigned int order)
+{
+	return kimage_alloc_normal_control_pages(image, order);
+}
+#endif
 
 static int kimage_add_entry(struct kimage *image, kimage_entry_t entry)
 {
@@ -532,7 +556,7 @@ static int kimage_add_entry(struct kimag
 			return -ENOMEM;
 
 		ind_page = page_address(page);
-		*image->entry = virt_to_phys(ind_page) | IND_INDIRECTION;
+		*image->entry = kexec_virt_to_phys(ind_page) | IND_INDIRECTION;
 		image->entry = ind_page;
 		image->last_entry = ind_page +
 				      ((PAGE_SIZE/sizeof(kimage_entry_t)) - 1);
@@ -593,13 +617,13 @@ static int kimage_terminate(struct kimag
 #define for_each_kimage_entry(image, ptr, entry) \
 	for (ptr = &image->head; (entry = *ptr) && !(entry & IND_DONE); \
 		ptr = (entry & IND_INDIRECTION)? \
-			phys_to_virt((entry & PAGE_MASK)): ptr +1)
+			kexec_phys_to_virt((entry & PAGE_MASK)): ptr +1)
 
 static void kimage_free_entry(kimage_entry_t entry)
 {
 	struct page *page;
 
-	page = pfn_to_page(entry >> PAGE_SHIFT);
+	page = kexec_pfn_to_page(entry >> PAGE_SHIFT);
 	kimage_free_pages(page);
 }
 
@@ -611,6 +635,10 @@ static void kimage_free(struct kimage *i
 	if (!image)
 		return;
 
+#ifdef CONFIG_XEN
+	xen_machine_kexec_unload(image);
+#endif
+
 	kimage_free_extra_pages(image);
 	for_each_kimage_entry(image, ptr, entry) {
 		if (entry & IND_INDIRECTION) {
@@ -686,7 +714,7 @@ static struct page *kimage_alloc_page(st
 	 * have a match.
 	 */
 	list_for_each_entry(page, &image->dest_pages, lru) {
-		addr = page_to_pfn(page) << PAGE_SHIFT;
+		addr = kexec_page_to_pfn(page) << PAGE_SHIFT;
 		if (addr == destination) {
 			list_del(&page->lru);
 			return page;
@@ -697,16 +725,16 @@ static struct page *kimage_alloc_page(st
 		kimage_entry_t *old;
 
 		/* Allocate a page, if we run out of memory give up */
-		page = kimage_alloc_pages(gfp_mask, 0);
+		page = kimage_alloc_pages(gfp_mask, 0, KEXEC_SOURCE_MEMORY_LIMIT);
 		if (!page)
 			return NULL;
 		/* If the page cannot be used file it away */
-		if (page_to_pfn(page) >
+		if (kexec_page_to_pfn(page) >
 				(KEXEC_SOURCE_MEMORY_LIMIT >> PAGE_SHIFT)) {
 			list_add(&page->lru, &image->unuseable_pages);
 			continue;
 		}
-		addr = page_to_pfn(page) << PAGE_SHIFT;
+		addr = kexec_page_to_pfn(page) << PAGE_SHIFT;
 
 		/* If it is the destination page we want use it */
 		if (addr == destination)
@@ -729,7 +757,7 @@ static struct page *kimage_alloc_page(st
 			struct page *old_page;
 
 			old_addr = *old & PAGE_MASK;
-			old_page = pfn_to_page(old_addr >> PAGE_SHIFT);
+			old_page = kexec_pfn_to_page(old_addr >> PAGE_SHIFT);
 			copy_highpage(page, old_page);
 			*old = addr | (*old & ~PAGE_MASK);
 
@@ -779,7 +807,7 @@ static int kimage_load_normal_segment(st
 			result  = -ENOMEM;
 			goto out;
 		}
-		result = kimage_add_page(image, page_to_pfn(page)
+		result = kimage_add_page(image, kexec_page_to_pfn(page)
 								<< PAGE_SHIFT);
 		if (result < 0)
 			goto out;
@@ -811,6 +839,7 @@ out:
 	return result;
 }
 
+#ifndef CONFIG_XEN
 static int kimage_load_crash_segment(struct kimage *image,
 					struct kexec_segment *segment)
 {
@@ -833,7 +862,7 @@ static int kimage_load_crash_segment(str
 		char *ptr;
 		size_t uchunk, mchunk;
 
-		page = pfn_to_page(maddr >> PAGE_SHIFT);
+		page = kexec_pfn_to_page(maddr >> PAGE_SHIFT);
 		if (page == 0) {
 			result  = -ENOMEM;
 			goto out;
@@ -851,6 +880,7 @@ static int kimage_load_crash_segment(str
 			memset(ptr + uchunk, 0, mchunk - uchunk);
 		}
 		result = copy_from_user(ptr, buf, uchunk);
+		kexec_flush_icache_page(page);
 		kunmap(page);
 		if (result) {
 			result = (result < 0) ? result : -EIO;
@@ -881,6 +911,13 @@ static int kimage_load_segment(struct ki
 
 	return result;
 }
+#else /* CONFIG_XEN */
+static int kimage_load_segment(struct kimage *image,
+				struct kexec_segment *segment)
+{
+	return kimage_load_normal_segment(image, segment);
+}
+#endif
 
 /*
  * Exec Kernel system call: for obvious reasons only root may call it.
@@ -991,6 +1028,13 @@ asmlinkage long sys_kexec_load(unsigned 
 		if (result)
 			goto out;
 	}
+#ifdef CONFIG_XEN
+	if (image) {
+		result = xen_machine_kexec_load(image);
+		if (result)
+			goto out;
+	}
+#endif
 	/* Install the new kernel, and  Uninstall the old */
 	image = xchg(dest_image, image);
 
@@ -1044,7 +1088,6 @@ void crash_kexec(struct pt_regs *regs)
 {
 	int locked;
 
-
 	/* Take the kexec_lock here to prevent sys_kexec_load
 	 * running on one cpu from replacing the crash kernel
 	 * we are using after a panic on a different cpu.
diff -rpuN linux-2.6.18.8/kernel/resource.c linux-2.6.18-xen-3.3.0/kernel/resource.c
--- linux-2.6.18.8/kernel/resource.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/kernel/resource.c	2008-08-21 11:36:07.000000000 +0200
@@ -36,6 +36,16 @@ struct resource iomem_resource = {
 };
 EXPORT_SYMBOL(iomem_resource);
 
+#ifdef CONFIG_PROC_IOMEM_MACHINE
+struct resource iomem_machine_resource = {
+	.name	= "Machine PCI mem",
+	.start	= 0,
+	.end	= -1,
+	.flags	= IORESOURCE_MEM,
+};
+EXPORT_SYMBOL(iomem_machine_resource);
+#endif
+
 static DEFINE_RWLOCK(resource_lock);
 
 #ifdef CONFIG_PROC_FS
@@ -115,6 +125,18 @@ static int iomem_open(struct inode *inod
 	return res;
 }
 
+#ifdef CONFIG_PROC_IOMEM_MACHINE
+static int iomem_machine_open(struct inode *inode, struct file *file)
+{
+	int res = seq_open(file, &resource_op);
+	if (!res) {
+		struct seq_file *m = file->private_data;
+		m->private = &iomem_machine_resource;
+	}
+	return res;
+}
+#endif
+
 static struct file_operations proc_ioports_operations = {
 	.open		= ioports_open,
 	.read		= seq_read,
@@ -129,6 +151,15 @@ static struct file_operations proc_iomem
 	.release	= seq_release,
 };
 
+#ifdef CONFIG_PROC_IOMEM_MACHINE
+static struct file_operations proc_iomem_machine_operations = {
+	.open		= iomem_machine_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= seq_release,
+};
+#endif
+
 static int __init ioresources_init(void)
 {
 	struct proc_dir_entry *entry;
@@ -139,6 +170,13 @@ static int __init ioresources_init(void)
 	entry = create_proc_entry("iomem", 0, NULL);
 	if (entry)
 		entry->proc_fops = &proc_iomem_operations;
+#ifdef CONFIG_PROC_IOMEM_MACHINE
+	if (is_initial_xendomain()) {
+		entry = create_proc_entry("iomem_machine", 0, NULL);
+		if (entry)
+			entry->proc_fops = &proc_iomem_machine_operations;
+	}
+#endif
 	return 0;
 }
 __initcall(ioresources_init);
diff -rpuN linux-2.6.18.8/kernel/softlockup.c linux-2.6.18-xen-3.3.0/kernel/softlockup.c
--- linux-2.6.18.8/kernel/softlockup.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/kernel/softlockup.c	2008-08-21 11:36:07.000000000 +0200
@@ -40,6 +40,19 @@ void touch_softlockup_watchdog(void)
 }
 EXPORT_SYMBOL(touch_softlockup_watchdog);
 
+unsigned long softlockup_get_next_event(void)
+{
+	int this_cpu = smp_processor_id();
+	unsigned long touch_timestamp = per_cpu(touch_timestamp, this_cpu);
+
+	if (per_cpu(print_timestamp, this_cpu) == touch_timestamp ||
+		did_panic ||
+			!per_cpu(watchdog_task, this_cpu))
+		return MAX_JIFFY_OFFSET;
+
+	return max_t(long, 0, touch_timestamp + HZ - jiffies);
+}
+
 /*
  * This callback runs from the timer interrupt, and checks
  * whether the watchdog thread has hung or not:
diff -rpuN linux-2.6.18.8/kernel/sysctl.c linux-2.6.18-xen-3.3.0/kernel/sysctl.c
--- linux-2.6.18.8/kernel/sysctl.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/kernel/sysctl.c	2008-08-21 11:36:07.000000000 +0200
@@ -661,7 +661,7 @@ static ctl_table kern_table[] = {
 		.proc_handler	= &proc_dointvec,
 	},
 #endif
-#ifdef CONFIG_ACPI_SLEEP
+#if defined(CONFIG_ACPI_SLEEP) && !defined(CONFIG_ACPI_PV_SLEEP)
 	{
 		.ctl_name	= KERN_ACPI_VIDEO_FLAGS,
 		.procname	= "acpi_video_flags",
diff -rpuN linux-2.6.18.8/kernel/timer.c linux-2.6.18-xen-3.3.0/kernel/timer.c
--- linux-2.6.18.8/kernel/timer.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/kernel/timer.c	2008-08-21 11:36:07.000000000 +0200
@@ -485,7 +485,9 @@ unsigned long next_timer_interrupt(void)
 		if (hr_expires < 3)
 			return hr_expires + jiffies;
 	}
-	hr_expires += jiffies;
+	hr_expires = min_t(unsigned long,
+			   softlockup_get_next_event(),
+			   hr_expires) + jiffies;
 
 	base = __get_cpu_var(tvec_bases);
 	spin_lock(&base->lock);
diff -rpuN linux-2.6.18.8/lib/Makefile linux-2.6.18-xen-3.3.0/lib/Makefile
--- linux-2.6.18.8/lib/Makefile	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/lib/Makefile	2008-08-21 11:36:07.000000000 +0200
@@ -52,6 +52,9 @@ obj-$(CONFIG_SMP) += percpu_counter.o
 obj-$(CONFIG_AUDIT_GENERIC) += audit.o
 
 obj-$(CONFIG_SWIOTLB) += swiotlb.o
+ifeq ($(CONFIG_IA64),y)
+swiotlb-$(CONFIG_XEN) := ../arch/ia64/xen/swiotlb.o
+endif
 
 hostprogs-y	:= gen_crc32table
 clean-files	:= crc32table.h
diff -rpuN linux-2.6.18.8/lib/swiotlb-xen.c linux-2.6.18-xen-3.3.0/lib/swiotlb-xen.c
--- linux-2.6.18.8/lib/swiotlb-xen.c	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/lib/swiotlb-xen.c	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,745 @@
+/*
+ * Dynamic DMA mapping support.
+ *
+ * This implementation is a fallback for platforms that do not support
+ * I/O TLBs (aka DMA address translation hardware).
+ * Copyright (C) 2000 Asit Mallick <Asit.K.Mallick@intel.com>
+ * Copyright (C) 2000 Goutham Rao <goutham.rao@intel.com>
+ * Copyright (C) 2000, 2003 Hewlett-Packard Co
+ *	David Mosberger-Tang <davidm@hpl.hp.com>
+ * Copyright (C) 2005 Keir Fraser <keir@xensource.com>
+ */
+
+#include <linux/cache.h>
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/spinlock.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/ctype.h>
+#include <linux/init.h>
+#include <linux/bootmem.h>
+#include <linux/highmem.h>
+#include <asm/io.h>
+#include <asm/pci.h>
+#include <asm/dma.h>
+#include <asm/uaccess.h>
+#include <xen/gnttab.h>
+#include <xen/interface/memory.h>
+#include <asm-i386/mach-xen/asm/gnttab_dma.h>
+
+int swiotlb;
+EXPORT_SYMBOL(swiotlb);
+
+#define OFFSET(val,align) ((unsigned long)((val) & ( (align) - 1)))
+
+/*
+ * Maximum allowable number of contiguous slabs to map,
+ * must be a power of 2.  What is the appropriate value ?
+ * The complexity of {map,unmap}_single is linearly dependent on this value.
+ */
+#define IO_TLB_SEGSIZE	128
+
+/*
+ * log of the size of each IO TLB slab.  The number of slabs is command line
+ * controllable.
+ */
+#define IO_TLB_SHIFT 11
+
+int swiotlb_force;
+
+static char *iotlb_virt_start;
+static unsigned long iotlb_nslabs;
+
+/*
+ * Used to do a quick range check in swiotlb_unmap_single and
+ * swiotlb_sync_single_*, to see if the memory was in fact allocated by this
+ * API.
+ */
+static unsigned long iotlb_pfn_start, iotlb_pfn_end;
+
+/* Does the given dma address reside within the swiotlb aperture? */
+static inline int in_swiotlb_aperture(dma_addr_t dev_addr)
+{
+	unsigned long pfn = mfn_to_local_pfn(dev_addr >> PAGE_SHIFT);
+	return (pfn_valid(pfn)
+		&& (pfn >= iotlb_pfn_start)
+		&& (pfn < iotlb_pfn_end));
+}
+
+/*
+ * When the IOMMU overflows we return a fallback buffer. This sets the size.
+ */
+static unsigned long io_tlb_overflow = 32*1024;
+
+void *io_tlb_overflow_buffer;
+
+/*
+ * This is a free list describing the number of free entries available from
+ * each index
+ */
+static unsigned int *io_tlb_list;
+static unsigned int io_tlb_index;
+
+/*
+ * We need to save away the original address corresponding to a mapped entry
+ * for the sync operations.
+ */
+static struct phys_addr {
+	struct page *page;
+	unsigned int offset;
+} *io_tlb_orig_addr;
+
+/*
+ * Protect the above data structures in the map and unmap calls
+ */
+static DEFINE_SPINLOCK(io_tlb_lock);
+
+static unsigned int dma_bits;
+static unsigned int __initdata max_dma_bits = 32;
+static int __init
+setup_dma_bits(char *str)
+{
+	max_dma_bits = simple_strtoul(str, NULL, 0);
+	return 0;
+}
+__setup("dma_bits=", setup_dma_bits);
+
+static int __init
+setup_io_tlb_npages(char *str)
+{
+	/* Unlike ia64, the size is aperture in megabytes, not 'slabs'! */
+	if (isdigit(*str)) {
+		iotlb_nslabs = simple_strtoul(str, &str, 0) <<
+			(20 - IO_TLB_SHIFT);
+		iotlb_nslabs = ALIGN(iotlb_nslabs, IO_TLB_SEGSIZE);
+		/* Round up to power of two (xen_create_contiguous_region). */
+		while (iotlb_nslabs & (iotlb_nslabs-1))
+			iotlb_nslabs += iotlb_nslabs & ~(iotlb_nslabs-1);
+	}
+	if (*str == ',')
+		++str;
+	/*
+         * NB. 'force' enables the swiotlb, but doesn't force its use for
+         * every DMA like it does on native Linux. 'off' forcibly disables
+         * use of the swiotlb.
+         */
+	if (!strcmp(str, "force"))
+		swiotlb_force = 1;
+	else if (!strcmp(str, "off"))
+		swiotlb_force = -1;
+	return 1;
+}
+__setup("swiotlb=", setup_io_tlb_npages);
+/* make io_tlb_overflow tunable too? */
+
+/*
+ * Statically reserve bounce buffer space and initialize bounce buffer data
+ * structures for the software IO TLB used to implement the PCI DMA API.
+ */
+void
+swiotlb_init_with_default_size (size_t default_size)
+{
+	unsigned long i, bytes;
+	int rc;
+
+	if (!iotlb_nslabs) {
+		iotlb_nslabs = (default_size >> IO_TLB_SHIFT);
+		iotlb_nslabs = ALIGN(iotlb_nslabs, IO_TLB_SEGSIZE);
+		/* Round up to power of two (xen_create_contiguous_region). */
+		while (iotlb_nslabs & (iotlb_nslabs-1))
+			iotlb_nslabs += iotlb_nslabs & ~(iotlb_nslabs-1);
+	}
+
+	bytes = iotlb_nslabs * (1UL << IO_TLB_SHIFT);
+
+	/*
+	 * Get IO TLB memory from the low pages
+	 */
+	iotlb_virt_start = alloc_bootmem_low_pages(bytes);
+	if (!iotlb_virt_start)
+		panic("Cannot allocate SWIOTLB buffer!\n");
+
+	dma_bits = get_order(IO_TLB_SEGSIZE << IO_TLB_SHIFT) + PAGE_SHIFT;
+	for (i = 0; i < iotlb_nslabs; i += IO_TLB_SEGSIZE) {
+		do {
+			rc = xen_create_contiguous_region(
+				(unsigned long)iotlb_virt_start + (i << IO_TLB_SHIFT),
+				get_order(IO_TLB_SEGSIZE << IO_TLB_SHIFT),
+				dma_bits);
+		} while (rc && dma_bits++ < max_dma_bits);
+		if (rc) {
+			if (i == 0)
+				panic("No suitable physical memory available for SWIOTLB buffer!\n"
+				      "Use dom0_mem Xen boot parameter to reserve\n"
+				      "some DMA memory (e.g., dom0_mem=-128M).\n");
+			iotlb_nslabs = i;
+			i <<= IO_TLB_SHIFT;
+			free_bootmem(__pa(iotlb_virt_start + i), bytes - i);
+			bytes = i;
+			for (dma_bits = 0; i > 0; i -= IO_TLB_SEGSIZE << IO_TLB_SHIFT) {
+				unsigned int bits = fls64(virt_to_bus(iotlb_virt_start + i - 1));
+
+				if (bits > dma_bits)
+					dma_bits = bits;
+			}
+			break;
+		}
+	}
+
+	/*
+	 * Allocate and initialize the free list array.  This array is used
+	 * to find contiguous free memory regions of size up to IO_TLB_SEGSIZE.
+	 */
+	io_tlb_list = alloc_bootmem(iotlb_nslabs * sizeof(int));
+	for (i = 0; i < iotlb_nslabs; i++)
+ 		io_tlb_list[i] = IO_TLB_SEGSIZE - OFFSET(i, IO_TLB_SEGSIZE);
+	io_tlb_index = 0;
+	io_tlb_orig_addr = alloc_bootmem(
+		iotlb_nslabs * sizeof(*io_tlb_orig_addr));
+
+	/*
+	 * Get the overflow emergency buffer
+	 */
+	io_tlb_overflow_buffer = alloc_bootmem_low(io_tlb_overflow);
+	if (!io_tlb_overflow_buffer)
+		panic("Cannot allocate SWIOTLB overflow buffer!\n");
+
+	do {
+		rc = xen_create_contiguous_region(
+			(unsigned long)io_tlb_overflow_buffer,
+			get_order(io_tlb_overflow),
+			dma_bits);
+	} while (rc && dma_bits++ < max_dma_bits);
+	if (rc)
+		panic("No suitable physical memory available for SWIOTLB overflow buffer!\n");
+
+	iotlb_pfn_start = __pa(iotlb_virt_start) >> PAGE_SHIFT;
+	iotlb_pfn_end   = iotlb_pfn_start + (bytes >> PAGE_SHIFT);
+
+	printk(KERN_INFO "Software IO TLB enabled: \n"
+	       " Aperture:     %lu megabytes\n"
+	       " Kernel range: %p - %p\n"
+	       " Address size: %u bits\n",
+	       bytes >> 20,
+	       iotlb_virt_start, iotlb_virt_start + bytes,
+	       dma_bits);
+}
+
+void
+swiotlb_init(void)
+{
+	long ram_end;
+	size_t defsz = 64 * (1 << 20); /* 64MB default size */
+
+	if (swiotlb_force == 1) {
+		swiotlb = 1;
+	} else if ((swiotlb_force != -1) &&
+		   is_running_on_xen() &&
+		   is_initial_xendomain()) {
+		/* Domain 0 always has a swiotlb. */
+		ram_end = HYPERVISOR_memory_op(XENMEM_maximum_ram_page, NULL);
+		if (ram_end <= 0x7ffff)
+			defsz = 2 * (1 << 20); /* 2MB on <2GB on systems. */
+		swiotlb = 1;
+	}
+
+	if (swiotlb)
+		swiotlb_init_with_default_size(defsz);
+	else
+		printk(KERN_INFO "Software IO TLB disabled\n");
+}
+
+/*
+ * We use __copy_to_user_inatomic to transfer to the host buffer because the
+ * buffer may be mapped read-only (e.g, in blkback driver) but lower-level
+ * drivers map the buffer for DMA_BIDIRECTIONAL access. This causes an
+ * unnecessary copy from the aperture to the host buffer, and a page fault.
+ */
+static void
+__sync_single(struct phys_addr buffer, char *dma_addr, size_t size, int dir)
+{
+	if (PageHighMem(buffer.page)) {
+		size_t len, bytes;
+		char *dev, *host, *kmp;
+		len = size;
+		while (len != 0) {
+			unsigned long flags;
+
+			if (((bytes = len) + buffer.offset) > PAGE_SIZE)
+				bytes = PAGE_SIZE - buffer.offset;
+			local_irq_save(flags); /* protects KM_BOUNCE_READ */
+			kmp  = kmap_atomic(buffer.page, KM_BOUNCE_READ);
+			dev  = dma_addr + size - len;
+			host = kmp + buffer.offset;
+			if (dir == DMA_FROM_DEVICE) {
+				if (__copy_to_user_inatomic(host, dev, bytes))
+					/* inaccessible */;
+			} else
+				memcpy(dev, host, bytes);
+			kunmap_atomic(kmp, KM_BOUNCE_READ);
+			local_irq_restore(flags);
+			len -= bytes;
+			buffer.page++;
+			buffer.offset = 0;
+		}
+	} else {
+		char *host = (char *)phys_to_virt(
+			page_to_pseudophys(buffer.page)) + buffer.offset;
+		if (dir == DMA_FROM_DEVICE) {
+			if (__copy_to_user_inatomic(host, dma_addr, size))
+				/* inaccessible */;
+		} else if (dir == DMA_TO_DEVICE)
+			memcpy(dma_addr, host, size);
+	}
+}
+
+/*
+ * Allocates bounce buffer and returns its kernel virtual address.
+ */
+static void *
+map_single(struct device *hwdev, struct phys_addr buffer, size_t size, int dir)
+{
+	unsigned long flags;
+	char *dma_addr;
+	unsigned int nslots, stride, index, wrap;
+	struct phys_addr slot_buf;
+	int i;
+
+	/*
+	 * For mappings greater than a page, we limit the stride (and
+	 * hence alignment) to a page size.
+	 */
+	nslots = ALIGN(size, 1 << IO_TLB_SHIFT) >> IO_TLB_SHIFT;
+	if (size > PAGE_SIZE)
+		stride = (1 << (PAGE_SHIFT - IO_TLB_SHIFT));
+	else
+		stride = 1;
+
+	BUG_ON(!nslots);
+
+	/*
+	 * Find suitable number of IO TLB entries size that will fit this
+	 * request and allocate a buffer from that IO TLB pool.
+	 */
+	spin_lock_irqsave(&io_tlb_lock, flags);
+	{
+		wrap = index = ALIGN(io_tlb_index, stride);
+
+		if (index >= iotlb_nslabs)
+			wrap = index = 0;
+
+		do {
+			/*
+			 * If we find a slot that indicates we have 'nslots'
+			 * number of contiguous buffers, we allocate the
+			 * buffers from that slot and mark the entries as '0'
+			 * indicating unavailable.
+			 */
+			if (io_tlb_list[index] >= nslots) {
+				int count = 0;
+
+				for (i = index; i < (int)(index + nslots); i++)
+					io_tlb_list[i] = 0;
+				for (i = index - 1;
+				     (OFFSET(i, IO_TLB_SEGSIZE) !=
+				      IO_TLB_SEGSIZE -1) && io_tlb_list[i];
+				     i--)
+					io_tlb_list[i] = ++count;
+				dma_addr = iotlb_virt_start +
+					(index << IO_TLB_SHIFT);
+
+				/*
+				 * Update the indices to avoid searching in
+				 * the next round.
+				 */
+				io_tlb_index = 
+					((index + nslots) < iotlb_nslabs
+					 ? (index + nslots) : 0);
+
+				goto found;
+			}
+			index += stride;
+			if (index >= iotlb_nslabs)
+				index = 0;
+		} while (index != wrap);
+
+		spin_unlock_irqrestore(&io_tlb_lock, flags);
+		return NULL;
+	}
+  found:
+	spin_unlock_irqrestore(&io_tlb_lock, flags);
+
+	/*
+	 * Save away the mapping from the original address to the DMA address.
+	 * This is needed when we sync the memory.  Then we sync the buffer if
+	 * needed.
+	 */
+	slot_buf = buffer;
+	for (i = 0; i < nslots; i++) {
+		slot_buf.page += slot_buf.offset >> PAGE_SHIFT;
+		slot_buf.offset &= PAGE_SIZE - 1;
+		io_tlb_orig_addr[index+i] = slot_buf;
+		slot_buf.offset += 1 << IO_TLB_SHIFT;
+	}
+	if ((dir == DMA_TO_DEVICE) || (dir == DMA_BIDIRECTIONAL))
+		__sync_single(buffer, dma_addr, size, DMA_TO_DEVICE);
+
+	return dma_addr;
+}
+
+static struct phys_addr dma_addr_to_phys_addr(char *dma_addr)
+{
+	int index = (dma_addr - iotlb_virt_start) >> IO_TLB_SHIFT;
+	struct phys_addr buffer = io_tlb_orig_addr[index];
+	buffer.offset += (long)dma_addr & ((1 << IO_TLB_SHIFT) - 1);
+	buffer.page += buffer.offset >> PAGE_SHIFT;
+	buffer.offset &= PAGE_SIZE - 1;
+	return buffer;
+}
+
+/*
+ * dma_addr is the kernel virtual address of the bounce buffer to unmap.
+ */
+static void
+unmap_single(struct device *hwdev, char *dma_addr, size_t size, int dir)
+{
+	unsigned long flags;
+	int i, count, nslots = ALIGN(size, 1 << IO_TLB_SHIFT) >> IO_TLB_SHIFT;
+	int index = (dma_addr - iotlb_virt_start) >> IO_TLB_SHIFT;
+	struct phys_addr buffer = dma_addr_to_phys_addr(dma_addr);
+
+	/*
+	 * First, sync the memory before unmapping the entry
+	 */
+	if ((dir == DMA_FROM_DEVICE) || (dir == DMA_BIDIRECTIONAL))
+		__sync_single(buffer, dma_addr, size, DMA_FROM_DEVICE);
+
+	/*
+	 * Return the buffer to the free list by setting the corresponding
+	 * entries to indicate the number of contigous entries available.
+	 * While returning the entries to the free list, we merge the entries
+	 * with slots below and above the pool being returned.
+	 */
+	spin_lock_irqsave(&io_tlb_lock, flags);
+	{
+		count = ((index + nslots) < ALIGN(index + 1, IO_TLB_SEGSIZE) ?
+			 io_tlb_list[index + nslots] : 0);
+		/*
+		 * Step 1: return the slots to the free list, merging the
+		 * slots with superceeding slots
+		 */
+		for (i = index + nslots - 1; i >= index; i--)
+			io_tlb_list[i] = ++count;
+		/*
+		 * Step 2: merge the returned slots with the preceding slots,
+		 * if available (non zero)
+		 */
+		for (i = index - 1;
+		     (OFFSET(i, IO_TLB_SEGSIZE) !=
+		      IO_TLB_SEGSIZE -1) && io_tlb_list[i];
+		     i--)
+			io_tlb_list[i] = ++count;
+	}
+	spin_unlock_irqrestore(&io_tlb_lock, flags);
+}
+
+static void
+sync_single(struct device *hwdev, char *dma_addr, size_t size, int dir)
+{
+	struct phys_addr buffer = dma_addr_to_phys_addr(dma_addr);
+	BUG_ON((dir != DMA_FROM_DEVICE) && (dir != DMA_TO_DEVICE));
+	__sync_single(buffer, dma_addr, size, dir);
+}
+
+static void
+swiotlb_full(struct device *dev, size_t size, int dir, int do_panic)
+{
+	/*
+	 * Ran out of IOMMU space for this operation. This is very bad.
+	 * Unfortunately the drivers cannot handle this operation properly.
+	 * unless they check for pci_dma_mapping_error (most don't)
+	 * When the mapping is small enough return a static buffer to limit
+	 * the damage, or panic when the transfer is too big.
+	 */
+	printk(KERN_ERR "PCI-DMA: Out of SW-IOMMU space for %lu bytes at "
+	       "device %s\n", (unsigned long)size, dev ? dev->bus_id : "?");
+
+	if (size > io_tlb_overflow && do_panic) {
+		if (dir == PCI_DMA_FROMDEVICE || dir == PCI_DMA_BIDIRECTIONAL)
+			panic("PCI-DMA: Memory would be corrupted\n");
+		if (dir == PCI_DMA_TODEVICE || dir == PCI_DMA_BIDIRECTIONAL)
+			panic("PCI-DMA: Random memory would be DMAed\n");
+	}
+}
+
+/*
+ * Map a single buffer of the indicated size for DMA in streaming mode.  The
+ * PCI address to use is returned.
+ *
+ * Once the device is given the dma address, the device owns this memory until
+ * either swiotlb_unmap_single or swiotlb_dma_sync_single is performed.
+ */
+dma_addr_t
+swiotlb_map_single(struct device *hwdev, void *ptr, size_t size, int dir)
+{
+	dma_addr_t dev_addr = gnttab_dma_map_page(virt_to_page(ptr)) +
+			      offset_in_page(ptr);
+	void *map;
+	struct phys_addr buffer;
+
+	BUG_ON(dir == DMA_NONE);
+
+	/*
+	 * If the pointer passed in happens to be in the device's DMA window,
+	 * we can safely return the device addr and not worry about bounce
+	 * buffering it.
+	 */
+	if (!range_straddles_page_boundary(__pa(ptr), size) &&
+	    !address_needs_mapping(hwdev, dev_addr))
+		return dev_addr;
+
+	/*
+	 * Oh well, have to allocate and map a bounce buffer.
+	 */
+	gnttab_dma_unmap_page(dev_addr);
+	buffer.page   = virt_to_page(ptr);
+	buffer.offset = (unsigned long)ptr & ~PAGE_MASK;
+	map = map_single(hwdev, buffer, size, dir);
+	if (!map) {
+		swiotlb_full(hwdev, size, dir, 1);
+		map = io_tlb_overflow_buffer;
+	}
+
+	dev_addr = virt_to_bus(map);
+	return dev_addr;
+}
+
+/*
+ * Unmap a single streaming mode DMA translation.  The dma_addr and size must
+ * match what was provided for in a previous swiotlb_map_single call.  All
+ * other usages are undefined.
+ *
+ * After this call, reads by the cpu to the buffer are guaranteed to see
+ * whatever the device wrote there.
+ */
+void
+swiotlb_unmap_single(struct device *hwdev, dma_addr_t dev_addr, size_t size,
+		     int dir)
+{
+	BUG_ON(dir == DMA_NONE);
+	if (in_swiotlb_aperture(dev_addr))
+		unmap_single(hwdev, bus_to_virt(dev_addr), size, dir);
+	else
+		gnttab_dma_unmap_page(dev_addr);
+}
+
+/*
+ * Make physical memory consistent for a single streaming mode DMA translation
+ * after a transfer.
+ *
+ * If you perform a swiotlb_map_single() but wish to interrogate the buffer
+ * using the cpu, yet do not wish to teardown the PCI dma mapping, you must
+ * call this function before doing so.  At the next point you give the PCI dma
+ * address back to the card, you must first perform a
+ * swiotlb_dma_sync_for_device, and then the device again owns the buffer
+ */
+void
+swiotlb_sync_single_for_cpu(struct device *hwdev, dma_addr_t dev_addr,
+			    size_t size, int dir)
+{
+	BUG_ON(dir == DMA_NONE);
+	if (in_swiotlb_aperture(dev_addr))
+		sync_single(hwdev, bus_to_virt(dev_addr), size, dir);
+}
+
+void
+swiotlb_sync_single_for_device(struct device *hwdev, dma_addr_t dev_addr,
+			       size_t size, int dir)
+{
+	BUG_ON(dir == DMA_NONE);
+	if (in_swiotlb_aperture(dev_addr))
+		sync_single(hwdev, bus_to_virt(dev_addr), size, dir);
+}
+
+/*
+ * Map a set of buffers described by scatterlist in streaming mode for DMA.
+ * This is the scatter-gather version of the above swiotlb_map_single
+ * interface.  Here the scatter gather list elements are each tagged with the
+ * appropriate dma address and length.  They are obtained via
+ * sg_dma_{address,length}(SG).
+ *
+ * NOTE: An implementation may be able to use a smaller number of
+ *       DMA address/length pairs than there are SG table elements.
+ *       (for example via virtual mapping capabilities)
+ *       The routine returns the number of addr/length pairs actually
+ *       used, at most nents.
+ *
+ * Device ownership issues as mentioned above for swiotlb_map_single are the
+ * same here.
+ */
+int
+swiotlb_map_sg(struct device *hwdev, struct scatterlist *sg, int nelems,
+	       int dir)
+{
+	struct phys_addr buffer;
+	dma_addr_t dev_addr;
+	char *map;
+	int i;
+
+	BUG_ON(dir == DMA_NONE);
+
+	for (i = 0; i < nelems; i++, sg++) {
+		dev_addr = gnttab_dma_map_page(sg->page) + sg->offset;
+
+		if (range_straddles_page_boundary(page_to_pseudophys(sg->page)
+						  + sg->offset, sg->length)
+		    || address_needs_mapping(hwdev, dev_addr)) {
+			gnttab_dma_unmap_page(dev_addr);
+			buffer.page   = sg->page;
+			buffer.offset = sg->offset;
+			map = map_single(hwdev, buffer, sg->length, dir);
+			if (!map) {
+				/* Don't panic here, we expect map_sg users
+				   to do proper error handling. */
+				swiotlb_full(hwdev, sg->length, dir, 0);
+				swiotlb_unmap_sg(hwdev, sg - i, i, dir);
+				sg[0].dma_length = 0;
+				return 0;
+			}
+			sg->dma_address = (dma_addr_t)virt_to_bus(map);
+		} else
+			sg->dma_address = dev_addr;
+		sg->dma_length = sg->length;
+	}
+	return nelems;
+}
+
+/*
+ * Unmap a set of streaming mode DMA translations.  Again, cpu read rules
+ * concerning calls here are the same as for swiotlb_unmap_single() above.
+ */
+void
+swiotlb_unmap_sg(struct device *hwdev, struct scatterlist *sg, int nelems,
+		 int dir)
+{
+	int i;
+
+	BUG_ON(dir == DMA_NONE);
+
+	for (i = 0; i < nelems; i++, sg++)
+		if (in_swiotlb_aperture(sg->dma_address))
+			unmap_single(hwdev, 
+				     (void *)bus_to_virt(sg->dma_address),
+				     sg->dma_length, dir);
+		else
+			gnttab_dma_unmap_page(sg->dma_address);
+}
+
+/*
+ * Make physical memory consistent for a set of streaming mode DMA translations
+ * after a transfer.
+ *
+ * The same as swiotlb_sync_single_* but for a scatter-gather list, same rules
+ * and usage.
+ */
+void
+swiotlb_sync_sg_for_cpu(struct device *hwdev, struct scatterlist *sg,
+			int nelems, int dir)
+{
+	int i;
+
+	BUG_ON(dir == DMA_NONE);
+
+	for (i = 0; i < nelems; i++, sg++)
+		if (in_swiotlb_aperture(sg->dma_address))
+			sync_single(hwdev,
+				    (void *)bus_to_virt(sg->dma_address),
+				    sg->dma_length, dir);
+}
+
+void
+swiotlb_sync_sg_for_device(struct device *hwdev, struct scatterlist *sg,
+			   int nelems, int dir)
+{
+	int i;
+
+	BUG_ON(dir == DMA_NONE);
+
+	for (i = 0; i < nelems; i++, sg++)
+		if (in_swiotlb_aperture(sg->dma_address))
+			sync_single(hwdev,
+				    (void *)bus_to_virt(sg->dma_address),
+				    sg->dma_length, dir);
+}
+
+#ifdef CONFIG_HIGHMEM
+
+dma_addr_t
+swiotlb_map_page(struct device *hwdev, struct page *page,
+		 unsigned long offset, size_t size,
+		 enum dma_data_direction direction)
+{
+	struct phys_addr buffer;
+	dma_addr_t dev_addr;
+	char *map;
+
+	dev_addr = gnttab_dma_map_page(page) + offset;
+	if (address_needs_mapping(hwdev, dev_addr)) {
+		gnttab_dma_unmap_page(dev_addr);
+		buffer.page   = page;
+		buffer.offset = offset;
+		map = map_single(hwdev, buffer, size, direction);
+		if (!map) {
+			swiotlb_full(hwdev, size, direction, 1);
+			map = io_tlb_overflow_buffer;
+		}
+		dev_addr = (dma_addr_t)virt_to_bus(map);
+	}
+
+	return dev_addr;
+}
+
+void
+swiotlb_unmap_page(struct device *hwdev, dma_addr_t dma_address,
+		   size_t size, enum dma_data_direction direction)
+{
+	BUG_ON(direction == DMA_NONE);
+	if (in_swiotlb_aperture(dma_address))
+		unmap_single(hwdev, bus_to_virt(dma_address), size, direction);
+	else
+		gnttab_dma_unmap_page(dma_address);
+}
+
+#endif
+
+int
+swiotlb_dma_mapping_error(dma_addr_t dma_addr)
+{
+	return (dma_addr == virt_to_bus(io_tlb_overflow_buffer));
+}
+
+/*
+ * Return whether the given PCI device DMA address mask can be supported
+ * properly.  For example, if your device can only drive the low 24-bits
+ * during PCI bus mastering, then you would pass 0x00ffffff as the mask to
+ * this function.
+ */
+int
+swiotlb_dma_supported (struct device *hwdev, u64 mask)
+{
+	return (mask >= ((1UL << dma_bits) - 1));
+}
+
+EXPORT_SYMBOL(swiotlb_init);
+EXPORT_SYMBOL(swiotlb_map_single);
+EXPORT_SYMBOL(swiotlb_unmap_single);
+EXPORT_SYMBOL(swiotlb_map_sg);
+EXPORT_SYMBOL(swiotlb_unmap_sg);
+EXPORT_SYMBOL(swiotlb_sync_single_for_cpu);
+EXPORT_SYMBOL(swiotlb_sync_single_for_device);
+EXPORT_SYMBOL(swiotlb_sync_sg_for_cpu);
+EXPORT_SYMBOL(swiotlb_sync_sg_for_device);
+EXPORT_SYMBOL(swiotlb_dma_mapping_error);
+EXPORT_SYMBOL(swiotlb_dma_supported);
diff -rpuN linux-2.6.18.8/MAINTAINERS linux-2.6.18-xen-3.3.0/MAINTAINERS
--- linux-2.6.18.8/MAINTAINERS	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/MAINTAINERS	2008-08-21 11:36:07.000000000 +0200
@@ -2558,6 +2558,13 @@ M:	pfg@sgi.com
 L:	linux-ia64@vger.kernel.org
 S:	Supported
 
+SFC NETWORK DRIVER
+P:	Steve Hodgson
+P:	Ben Hutchings
+P:	Robert Stonehouse
+M:	linux-net-drivers@solarflare.com
+S:	Supported
+
 SGI VISUAL WORKSTATION 320 AND 540
 P:	Andrey Panin
 M:	pazke@donpac.ru
diff -rpuN linux-2.6.18.8/mm/highmem.c linux-2.6.18-xen-3.3.0/mm/highmem.c
--- linux-2.6.18.8/mm/highmem.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/mm/highmem.c	2008-08-21 11:36:07.000000000 +0200
@@ -142,6 +142,17 @@ start:
 	return vaddr;
 }
 
+#ifdef CONFIG_XEN
+void kmap_flush_unused(void)
+{
+	spin_lock(&kmap_lock);
+	flush_all_zero_pkmaps();
+	spin_unlock(&kmap_lock);
+}
+
+EXPORT_SYMBOL(kmap_flush_unused);
+#endif
+
 void fastcall *kmap_high(struct page *page)
 {
 	unsigned long vaddr;
@@ -457,6 +468,12 @@ void blk_queue_bounce(request_queue_t *q
 	mempool_t *pool;
 
 	/*
+	 * Data-less bio, nothing to bounce
+	 */
+	if (bio_empty_barrier(*bio_orig))
+		return;
+
+	/*
 	 * for non-isa bounce case, just check if the bounce pfn is equal
 	 * to or bigger than the highest pfn in the system -- in that case,
 	 * don't waste time iterating over bio segments
diff -rpuN linux-2.6.18.8/mm/memory.c linux-2.6.18-xen-3.3.0/mm/memory.c
--- linux-2.6.18.8/mm/memory.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/mm/memory.c	2008-08-21 11:36:07.000000000 +0200
@@ -396,6 +396,12 @@ struct page *vm_normal_page(struct vm_ar
 			return NULL;
 	}
 
+#if defined(CONFIG_XEN) && defined(CONFIG_X86)
+	/* XEN: Covers user-space grant mappings (even of local pages). */
+	if (unlikely(vma->vm_flags & VM_FOREIGN))
+		return NULL;
+#endif
+
 	/*
 	 * Add some anal sanity checks for now. Eventually,
 	 * we should just do "return pfn_to_page(pfn)", but
@@ -403,7 +409,8 @@ struct page *vm_normal_page(struct vm_ar
 	 * and that the resulting page looks ok.
 	 */
 	if (unlikely(!pfn_valid(pfn))) {
-		print_bad_pte(vma, pte, addr);
+		if (!(vma->vm_flags & VM_RESERVED))
+			print_bad_pte(vma, pte, addr);
 		return NULL;
 	}
 
@@ -658,8 +665,12 @@ static unsigned long zap_pte_range(struc
 				     page->index > details->last_index))
 					continue;
 			}
-			ptent = ptep_get_and_clear_full(mm, addr, pte,
-							tlb->fullmm);
+			if (unlikely(vma->vm_ops && vma->vm_ops->zap_pte))
+				ptent = vma->vm_ops->zap_pte(vma, addr, pte,
+							     tlb->fullmm);
+			else
+				ptent = ptep_get_and_clear_full(mm, addr, pte,
+								tlb->fullmm);
 			tlb_remove_tlb_entry(tlb, pte, addr);
 			if (unlikely(!page))
 				continue;
@@ -754,6 +765,7 @@ static unsigned long unmap_page_range(st
 		details = NULL;
 
 	BUG_ON(addr >= end);
+
 	tlb_start_vma(tlb, vma);
 	pgd = pgd_offset(vma->vm_mm, addr);
 	do {
@@ -891,6 +903,7 @@ unsigned long zap_page_range(struct vm_a
 		tlb_finish_mmu(tlb, address, end);
 	return end;
 }
+EXPORT_SYMBOL(zap_page_range);
 
 /*
  * Do a quick page-table lookup for a single page.
@@ -1030,6 +1043,26 @@ int get_user_pages(struct task_struct *t
 			continue;
 		}
 
+#ifdef CONFIG_XEN
+		if (vma && (vma->vm_flags & VM_FOREIGN)) {
+			struct page **map = vma->vm_private_data;
+			int offset = (start - vma->vm_start) >> PAGE_SHIFT;
+			if (map[offset] != NULL) {
+			        if (pages) {
+			                struct page *page = map[offset];
+					
+					pages[i] = page;
+					get_page(page);
+				}
+				if (vmas)
+					vmas[i] = vma;
+				i++;
+				start += PAGE_SIZE;
+				len--;
+				continue;
+			}
+		}
+#endif
 		if (!vma || (vma->vm_flags & (VM_IO | VM_PFNMAP))
 				|| !(vm_flags & vma->vm_flags))
 			return i ? : -EFAULT;
@@ -1379,6 +1412,102 @@ int remap_pfn_range(struct vm_area_struc
 }
 EXPORT_SYMBOL(remap_pfn_range);
 
+#ifdef CONFIG_XEN
+static inline int apply_to_pte_range(struct mm_struct *mm, pmd_t *pmd,
+				     unsigned long addr, unsigned long end,
+				     pte_fn_t fn, void *data)
+{
+	pte_t *pte;
+	int err;
+	struct page *pmd_page;
+	spinlock_t *ptl;
+
+	pte = (mm == &init_mm) ?
+		pte_alloc_kernel(pmd, addr) :
+		pte_alloc_map_lock(mm, pmd, addr, &ptl);
+	if (!pte)
+		return -ENOMEM;
+
+	BUG_ON(pmd_huge(*pmd));
+
+	pmd_page = pmd_page(*pmd);
+
+	do {
+		err = fn(pte, pmd_page, addr, data);
+		if (err)
+			break;
+	} while (pte++, addr += PAGE_SIZE, addr != end);
+
+	if (mm != &init_mm)
+		pte_unmap_unlock(pte-1, ptl);
+	return err;
+}
+
+static inline int apply_to_pmd_range(struct mm_struct *mm, pud_t *pud,
+				     unsigned long addr, unsigned long end,
+				     pte_fn_t fn, void *data)
+{
+	pmd_t *pmd;
+	unsigned long next;
+	int err;
+
+	pmd = pmd_alloc(mm, pud, addr);
+	if (!pmd)
+		return -ENOMEM;
+	do {
+		next = pmd_addr_end(addr, end);
+		err = apply_to_pte_range(mm, pmd, addr, next, fn, data);
+		if (err)
+			break;
+	} while (pmd++, addr = next, addr != end);
+	return err;
+}
+
+static inline int apply_to_pud_range(struct mm_struct *mm, pgd_t *pgd,
+				     unsigned long addr, unsigned long end,
+				     pte_fn_t fn, void *data)
+{
+	pud_t *pud;
+	unsigned long next;
+	int err;
+
+	pud = pud_alloc(mm, pgd, addr);
+	if (!pud)
+		return -ENOMEM;
+	do {
+		next = pud_addr_end(addr, end);
+		err = apply_to_pmd_range(mm, pud, addr, next, fn, data);
+		if (err)
+			break;
+	} while (pud++, addr = next, addr != end);
+	return err;
+}
+
+/*
+ * Scan a region of virtual memory, filling in page tables as necessary
+ * and calling a provided function on each leaf page table.
+ */
+int apply_to_page_range(struct mm_struct *mm, unsigned long addr,
+			unsigned long size, pte_fn_t fn, void *data)
+{
+	pgd_t *pgd;
+	unsigned long next;
+	unsigned long end = addr + size;
+	int err;
+
+	BUG_ON(addr >= end);
+	pgd = pgd_offset(mm, addr);
+	do {
+		next = pgd_addr_end(addr, end);
+		err = apply_to_pud_range(mm, pgd, addr, next, fn, data);
+		if (err)
+			break;
+	} while (pgd++, addr = next, addr != end);
+	return err;
+}
+EXPORT_SYMBOL_GPL(apply_to_page_range);
+#endif
+
 /*
  * handle_pte_fault chooses page fault handler according to an entry
  * which was read non-atomically.  Before making any commitment, on
diff -rpuN linux-2.6.18.8/mm/mmap.c linux-2.6.18-xen-3.3.0/mm/mmap.c
--- linux-2.6.18.8/mm/mmap.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/mm/mmap.c	2008-08-21 11:36:07.000000000 +0200
@@ -1963,6 +1963,10 @@ void exit_mmap(struct mm_struct *mm)
 	unsigned long nr_accounted = 0;
 	unsigned long end;
 
+#ifdef arch_exit_mmap
+	arch_exit_mmap(mm);
+#endif
+
 	lru_add_drain();
 	flush_cache_mm(mm);
 	tlb = tlb_gather_mmu(mm, 1);
diff -rpuN linux-2.6.18.8/mm/mprotect.c linux-2.6.18-xen-3.3.0/mm/mprotect.c
--- linux-2.6.18.8/mm/mprotect.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/mm/mprotect.c	2008-08-21 11:36:07.000000000 +0200
@@ -76,6 +76,8 @@ static inline void change_pmd_range(stru
 		next = pmd_addr_end(addr, end);
 		if (pmd_none_or_clear_bad(pmd))
 			continue;
+		if (arch_change_pte_range(mm, pmd, addr, next, newprot))
+			continue;
 		change_pte_range(mm, pmd, addr, next, newprot);
 	} while (pmd++, addr = next, addr != end);
 }
diff -rpuN linux-2.6.18.8/mm/page_alloc.c linux-2.6.18-xen-3.3.0/mm/page_alloc.c
--- linux-2.6.18.8/mm/page_alloc.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/mm/page_alloc.c	2008-08-21 11:36:07.000000000 +0200
@@ -154,7 +154,11 @@ static void bad_page(struct page *page)
 			1 << PG_slab    |
 			1 << PG_swapcache |
 			1 << PG_writeback |
-			1 << PG_buddy );
+			1 << PG_buddy	|
+#ifdef CONFIG_X86_XEN
+			1 << PG_pinned	|
+#endif
+			1 << PG_foreign );
 	set_page_count(page, 0);
 	reset_page_mapcount(page);
 	page->mapping = NULL;
@@ -389,7 +393,11 @@ static inline int free_pages_check(struc
 			1 << PG_swapcache |
 			1 << PG_writeback |
 			1 << PG_reserved |
-			1 << PG_buddy ))))
+			1 << PG_buddy	|
+#ifdef CONFIG_X86_XEN
+			1 << PG_pinned	|
+#endif
+			1 << PG_foreign ))))
 		bad_page(page);
 	if (PageDirty(page))
 		__ClearPageDirty(page);
@@ -443,6 +451,12 @@ static void __free_pages_ok(struct page 
 	int i;
 	int reserved = 0;
 
+#ifdef CONFIG_XEN
+	if (PageForeign(page)) {
+		PageForeignDestructor(page);
+		return;
+	}
+#endif
 	arch_free_page(page, order);
 	if (!PageHighMem(page))
 		debug_check_no_locks_freed(page_address(page),
@@ -538,7 +552,11 @@ static int prep_new_page(struct page *pa
 			1 << PG_swapcache |
 			1 << PG_writeback |
 			1 << PG_reserved |
-			1 << PG_buddy ))))
+			1 << PG_buddy	|
+#ifdef CONFIG_X86_XEN
+			1 << PG_pinned	|
+#endif
+			1 << PG_foreign ))))
 		bad_page(page);
 
 	/*
@@ -717,6 +735,12 @@ static void fastcall free_hot_cold_page(
 	struct per_cpu_pages *pcp;
 	unsigned long flags;
 
+#ifdef CONFIG_XEN
+	if (PageForeign(page)) {
+		PageForeignDestructor(page);
+		return;
+	}
+#endif
 	arch_free_page(page, 0);
 
 	if (PageAnon(page))
diff -rpuN linux-2.6.18.8/net/bridge/br_netfilter.c linux-2.6.18-xen-3.3.0/net/bridge/br_netfilter.c
--- linux-2.6.18.8/net/bridge/br_netfilter.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/net/bridge/br_netfilter.c	2008-08-21 11:36:07.000000000 +0200
@@ -127,10 +127,10 @@ static inline struct nf_bridge_info *nf_
 
 static inline void nf_bridge_save_header(struct sk_buff *skb)
 {
-        int header_size = 16;
+        int header_size = ETH_HLEN;
 
 	if (skb->protocol == htons(ETH_P_8021Q))
-		header_size = 18;
+		header_size += VLAN_HLEN;
 
 	memcpy(skb->nf_bridge->data, skb->data - header_size, header_size);
 }
diff -rpuN linux-2.6.18.8/net/bridge/netfilter/ebtables.c linux-2.6.18-xen-3.3.0/net/bridge/netfilter/ebtables.c
--- linux-2.6.18.8/net/bridge/netfilter/ebtables.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/net/bridge/netfilter/ebtables.c	2008-08-21 11:36:07.000000000 +0200
@@ -597,7 +597,7 @@ ebt_check_entry(struct ebt_entry *e, str
 	struct ebt_entry_target *t;
 	struct ebt_target *target;
 	unsigned int i, j, hook = 0, hookmask = 0;
-	size_t gap = e->next_offset - e->target_offset;
+	size_t gap;
 	int ret;
 
 	/* don't mess with the struct ebt_entries */
@@ -647,6 +647,7 @@ ebt_check_entry(struct ebt_entry *e, str
 	if (ret != 0)
 		goto cleanup_watchers;
 	t = (struct ebt_entry_target *)(((char *)e) + e->target_offset);
+	gap = e->next_offset - e->target_offset;
 	target = find_target_lock(t->u.name, &ret, &ebt_mutex);
 	if (!target)
 		goto cleanup_watchers;
diff -rpuN linux-2.6.18.8/net/core/dev.c linux-2.6.18-xen-3.3.0/net/core/dev.c
--- linux-2.6.18.8/net/core/dev.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/net/core/dev.c	2008-08-21 11:36:07.000000000 +0200
@@ -113,11 +113,18 @@
 #include <linux/wireless.h>
 #include <net/iw_handler.h>
 #include <asm/current.h>
+#include <linux/err.h>
 #include <linux/audit.h>
 #include <linux/dmaengine.h>
 #include <linux/err.h>
 #include <linux/ctype.h>
 
+#ifdef CONFIG_XEN
+#include <net/ip.h>
+#include <linux/tcp.h>
+#include <linux/udp.h>
+#endif
+
 /*
  *	The list of packet types we will receive (as opposed to discard)
  *	and the routines to invoke.
@@ -1328,7 +1335,7 @@ static int dev_gso_segment(struct sk_buf
 	/* Verifying header integrity only. */
 	if (!segs)
 		return 0;
-
+  
 	if (unlikely(IS_ERR(segs)))
 		return PTR_ERR(segs);
 
@@ -1391,6 +1398,43 @@ out_kfree_skb:
 	}						\
 }
 
+#ifdef CONFIG_XEN
+inline int skb_checksum_setup(struct sk_buff *skb)
+{
+	if (skb->proto_csum_blank) {
+		if (skb->protocol != htons(ETH_P_IP))
+			goto out;
+		skb->h.raw = (unsigned char *)skb->nh.iph + 4*skb->nh.iph->ihl;
+		if (skb->h.raw >= skb->tail)
+			goto out;
+		switch (skb->nh.iph->protocol) {
+		case IPPROTO_TCP:
+			skb->csum = offsetof(struct tcphdr, check);
+			break;
+		case IPPROTO_UDP:
+			skb->csum = offsetof(struct udphdr, check);
+			break;
+		default:
+			if (net_ratelimit())
+				printk(KERN_ERR "Attempting to checksum a non-"
+				       "TCP/UDP packet, dropping a protocol"
+				       " %d packet", skb->nh.iph->protocol);
+			goto out;
+		}
+		if ((skb->h.raw + skb->csum + 2) > skb->tail)
+			goto out;
+		skb->ip_summed = CHECKSUM_HW;
+		skb->proto_csum_blank = 0;
+	}
+	return 0;
+out:
+	return -EPROTO;
+}
+#else
+inline int skb_checksum_setup(struct sk_buff *skb) { return 0; }
+#endif
+
+
 /**
  *	dev_queue_xmit - transmit a buffer
  *	@skb: buffer to transmit
@@ -1423,6 +1467,12 @@ int dev_queue_xmit(struct sk_buff *skb)
 	struct Qdisc *q;
 	int rc = -ENOMEM;
 
+ 	/* If a checksum-deferred packet is forwarded to a device that needs a
+ 	 * checksum, correct the pointers and force checksumming.
+ 	 */
+ 	if (skb_checksum_setup(skb))
+ 		goto out_kfree_skb;
+
 	/* GSO will handle the following emulations directly. */
 	if (netif_needs_gso(dev, skb))
 		goto gso;
@@ -1798,6 +1848,19 @@ int netif_receive_skb(struct sk_buff *sk
 	}
 #endif
 
+#ifdef CONFIG_XEN
+	switch (skb->ip_summed) {
+	case CHECKSUM_UNNECESSARY:
+		skb->proto_data_valid = 1;
+		break;
+	case CHECKSUM_HW:
+		/* XXX Implement me. */
+	default:
+		skb->proto_data_valid = 0;
+		break;
+	}
+#endif
+
 	list_for_each_entry_rcu(ptype, &ptype_all, list) {
 		if (!ptype->dev || ptype->dev == skb->dev) {
 			if (pt_prev) 
@@ -3584,6 +3647,7 @@ EXPORT_SYMBOL(unregister_netdevice_notif
 EXPORT_SYMBOL(net_enable_timestamp);
 EXPORT_SYMBOL(net_disable_timestamp);
 EXPORT_SYMBOL(dev_get_flags);
+EXPORT_SYMBOL(skb_checksum_setup);
 
 #if defined(CONFIG_BRIDGE) || defined(CONFIG_BRIDGE_MODULE)
 EXPORT_SYMBOL(br_handle_frame_hook);
diff -rpuN linux-2.6.18.8/net/core/neighbour.c linux-2.6.18-xen-3.3.0/net/core/neighbour.c
--- linux-2.6.18.8/net/core/neighbour.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/net/core/neighbour.c	2008-08-21 11:36:07.000000000 +0200
@@ -2679,7 +2679,6 @@ EXPORT_SYMBOL(neigh_table_clear);
 EXPORT_SYMBOL(neigh_table_init);
 EXPORT_SYMBOL(neigh_table_init_no_netlink);
 EXPORT_SYMBOL(neigh_update);
-EXPORT_SYMBOL(neigh_update_hhs);
 EXPORT_SYMBOL(pneigh_enqueue);
 EXPORT_SYMBOL(pneigh_lookup);
 EXPORT_SYMBOL(neightbl_dump_info);
diff -rpuN linux-2.6.18.8/net/core/skbuff.c linux-2.6.18-xen-3.3.0/net/core/skbuff.c
--- linux-2.6.18.8/net/core/skbuff.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/net/core/skbuff.c	2008-08-21 11:36:07.000000000 +0200
@@ -240,6 +240,7 @@ struct sk_buff *alloc_skb_from_cache(kme
 	skb_shinfo(skb)->gso_size = 0;
 	skb_shinfo(skb)->gso_segs = 0;
 	skb_shinfo(skb)->gso_type = 0;
+	skb_shinfo(skb)->ip6_frag_id = 0;
 	skb_shinfo(skb)->frag_list = NULL;
 out:
 	return skb;
@@ -464,6 +465,10 @@ struct sk_buff *skb_clone(struct sk_buff
 	C(local_df);
 	n->cloned = 1;
 	n->nohdr = 0;
+#ifdef CONFIG_XEN
+	C(proto_data_valid);
+	C(proto_csum_blank);
+#endif
 	C(pkt_type);
 	C(ip_summed);
 	C(priority);
@@ -2071,7 +2076,6 @@ EXPORT_SYMBOL(pskb_copy);
 EXPORT_SYMBOL(pskb_expand_head);
 EXPORT_SYMBOL(skb_checksum);
 EXPORT_SYMBOL(skb_clone);
-EXPORT_SYMBOL(skb_clone_fraglist);
 EXPORT_SYMBOL(skb_copy);
 EXPORT_SYMBOL(skb_copy_and_csum_bits);
 EXPORT_SYMBOL(skb_copy_and_csum_dev);
diff -rpuN linux-2.6.18.8/net/ipv4/netfilter/ip_nat_proto_tcp.c linux-2.6.18-xen-3.3.0/net/ipv4/netfilter/ip_nat_proto_tcp.c
--- linux-2.6.18.8/net/ipv4/netfilter/ip_nat_proto_tcp.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/net/ipv4/netfilter/ip_nat_proto_tcp.c	2008-08-21 11:36:07.000000000 +0200
@@ -129,7 +129,12 @@ tcp_manip_pkt(struct sk_buff **pskb,
 	if (hdrsize < sizeof(*hdr))
 		return 1;
 
-	hdr->check = ip_nat_cheat_check(~oldip, newip,
+#ifdef CONFIG_XEN
+	if ((*pskb)->proto_csum_blank)
+		hdr->check = ip_nat_cheat_check(oldip, ~newip, hdr->check);
+	else
+#endif
+		hdr->check = ip_nat_cheat_check(~oldip, newip,
 					ip_nat_cheat_check(oldport ^ 0xFFFF,
 							   newport,
 							   hdr->check));
diff -rpuN linux-2.6.18.8/net/ipv4/netfilter/ip_nat_proto_udp.c linux-2.6.18-xen-3.3.0/net/ipv4/netfilter/ip_nat_proto_udp.c
--- linux-2.6.18.8/net/ipv4/netfilter/ip_nat_proto_udp.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/net/ipv4/netfilter/ip_nat_proto_udp.c	2008-08-21 11:36:07.000000000 +0200
@@ -113,11 +113,17 @@ udp_manip_pkt(struct sk_buff **pskb,
 		newport = tuple->dst.u.udp.port;
 		portptr = &hdr->dest;
 	}
-	if (hdr->check) /* 0 is a special case meaning no checksum */
-		hdr->check = ip_nat_cheat_check(~oldip, newip,
+	if (hdr->check) { /* 0 is a special case meaning no checksum */
+#ifdef CONFIG_XEN
+		if ((*pskb)->proto_csum_blank)
+			hdr->check = ip_nat_cheat_check(oldip, ~newip, hdr->check);
+		else
+#endif
+			hdr->check = ip_nat_cheat_check(~oldip, newip,
 					ip_nat_cheat_check(*portptr ^ 0xFFFF,
 							   newport,
 							   hdr->check));
+	}
 	*portptr = newport;
 	return 1;
 }
diff -rpuN linux-2.6.18.8/net/ipv4/tcp_input.c linux-2.6.18-xen-3.3.0/net/ipv4/tcp_input.c
--- linux-2.6.18.8/net/ipv4/tcp_input.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/net/ipv4/tcp_input.c	2008-08-21 11:36:07.000000000 +0200
@@ -127,7 +127,7 @@ static void tcp_measure_rcv_mss(struct s
 	/* skb->len may jitter because of SACKs, even if peer
 	 * sends good full-sized frames.
 	 */
-	len = skb->len;
+	len = skb_shinfo(skb)->gso_size ?: skb->len;
 	if (len >= icsk->icsk_ack.rcv_mss) {
 		icsk->icsk_ack.rcv_mss = len;
 	} else {
diff -rpuN linux-2.6.18.8/net/ipv4/xfrm4_output.c linux-2.6.18-xen-3.3.0/net/ipv4/xfrm4_output.c
--- linux-2.6.18.8/net/ipv4/xfrm4_output.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/net/ipv4/xfrm4_output.c	2008-08-21 11:36:07.000000000 +0200
@@ -18,6 +18,8 @@
 #include <net/xfrm.h>
 #include <net/icmp.h>
 
+extern int skb_checksum_setup(struct sk_buff *skb);
+
 static int xfrm4_tunnel_check_size(struct sk_buff *skb)
 {
 	int mtu, ret = 0;
@@ -48,6 +50,10 @@ static int xfrm4_output_one(struct sk_bu
 	struct xfrm_state *x = dst->xfrm;
 	int err;
 	
+	err = skb_checksum_setup(skb);
+	if (err)
+		goto error_nolock;
+
 	if (skb->ip_summed == CHECKSUM_HW) {
 		err = skb_checksum_help(skb, 0);
 		if (err)
diff -rpuN linux-2.6.18.8/scripts/Makefile.build linux-2.6.18-xen-3.3.0/scripts/Makefile.build
--- linux-2.6.18.8/scripts/Makefile.build	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/scripts/Makefile.build	2008-08-21 11:36:07.000000000 +0200
@@ -68,6 +68,20 @@ ifndef obj
 $(warning kbuild: Makefile.build is included improperly)
 endif
 
+ifeq ($(CONFIG_XEN),y)
+$(objtree)/scripts/Makefile.xen: $(srctree)/scripts/Makefile.xen.awk $(srctree)/scripts/Makefile.build
+	@echo '  Updating $@'
+	$(if $(shell echo a | $(AWK) '{ print gensub(/a/, "AA", "g"); }'),\
+        ,$(error 'Your awk program does not define gensub.  Use gawk or another awk with gensub'))
+	@$(AWK) -f $< $(filter-out $<,$^) >$@
+
+xen-src-single-used-m	:= $(patsubst $(srctree)/%,%,$(wildcard $(addprefix $(srctree)/,$(single-used-m:.o=-xen.c))))
+xen-single-used-m	:= $(xen-src-single-used-m:-xen.c=.o)
+single-used-m		:= $(filter-out $(xen-single-used-m),$(single-used-m))
+
+-include $(objtree)/scripts/Makefile.xen
+endif
+
 # ===========================================================================
 
 ifneq ($(strip $(lib-y) $(lib-m) $(lib-n) $(lib-)),)
diff -rpuN linux-2.6.18.8/scripts/Makefile.lib linux-2.6.18-xen-3.3.0/scripts/Makefile.lib
--- linux-2.6.18.8/scripts/Makefile.lib	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/scripts/Makefile.lib	2008-08-21 11:36:07.000000000 +0200
@@ -13,6 +13,12 @@ obj-m := $(filter-out $(obj-y),$(obj-m))
 
 lib-y := $(filter-out $(obj-y), $(sort $(lib-y) $(lib-m)))
 
+# Remove objects forcibly disabled
+
+obj-y := $(filter-out $(disabled-obj-y),$(obj-y))
+obj-m := $(filter-out $(disabled-obj-y),$(obj-m))
+lib-y := $(filter-out $(disabled-obj-y),$(lib-y))
+
 
 # Handle objects in subdirs
 # ---------------------------------------------------------------------------
diff -rpuN linux-2.6.18.8/scripts/Makefile.xen.awk linux-2.6.18-xen-3.3.0/scripts/Makefile.xen.awk
--- linux-2.6.18.8/scripts/Makefile.xen.awk	1970-01-01 01:00:00.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/scripts/Makefile.xen.awk	2008-08-21 11:36:07.000000000 +0200
@@ -0,0 +1,34 @@
+BEGIN {
+	is_rule = 0
+}
+
+/^[[:space:]]*#/ {
+	next
+}
+
+/^[[:space:]]*$/ {
+	if (is_rule)
+		print("")
+	is_rule = 0
+	next
+}
+
+/:[[:space:]]*%\.[cS][[:space:]]/ {
+	line = gensub(/%.([cS])/, "%-xen.\\1", "g", $0)
+	line = gensub(/(single-used-m)/, "xen-\\1", "g", line)
+	print line
+	is_rule = 1
+	next
+}
+
+/^[^\t]$/ {
+	if (is_rule)
+		print("")
+	is_rule = 0
+	next
+}
+
+is_rule {
+	print $0
+	next
+}
diff -rpuN linux-2.6.18.8/scripts/mod/sumversion.c linux-2.6.18-xen-3.3.0/scripts/mod/sumversion.c
--- linux-2.6.18.8/scripts/mod/sumversion.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/scripts/mod/sumversion.c	2008-08-21 11:36:07.000000000 +0200
@@ -7,6 +7,7 @@
 #include <ctype.h>
 #include <errno.h>
 #include <string.h>
+#include <limits.h>
 #include "modpost.h"
 
 /*
diff -rpuN linux-2.6.18.8/sound/pci/hda/hda_intel.c linux-2.6.18-xen-3.3.0/sound/pci/hda/hda_intel.c
--- linux-2.6.18.8/sound/pci/hda/hda_intel.c	2007-02-24 00:52:30.000000000 +0100
+++ linux-2.6.18-xen-3.3.0/sound/pci/hda/hda_intel.c	2008-08-21 11:36:07.000000000 +0200
@@ -80,6 +80,8 @@ MODULE_SUPPORTED_DEVICE("{{Intel, ICH6},
 			 "{Intel, ICH7},"
 			 "{Intel, ESB2},"
 			 "{Intel, ICH8},"
+			 "{Intel, ICH9},"
+			 "{Intel, ICH10},"
 			 "{ATI, SB450},"
 			 "{ATI, SB600},"
 			 "{ATI, RS600},"
@@ -1634,6 +1636,10 @@ static struct pci_device_id azx_ids[] = 
 	{ 0x8086, 0x27d8, PCI_ANY_ID, PCI_ANY_ID, 0, 0, AZX_DRIVER_ICH }, /* ICH7 */
 	{ 0x8086, 0x269a, PCI_ANY_ID, PCI_ANY_ID, 0, 0, AZX_DRIVER_ICH }, /* ESB2 */
 	{ 0x8086, 0x284b, PCI_ANY_ID, PCI_ANY_ID, 0, 0, AZX_DRIVER_ICH }, /* ICH8 */
+	{ 0x8086, 0x293e, PCI_ANY_ID, PCI_ANY_ID, 0, 0, AZX_DRIVER_ICH }, /* ICH9 */
+	{ 0x8086, 0x293f, PCI_ANY_ID, PCI_ANY_ID, 0, 0, AZX_DRIVER_ICH }, /* ICH9 */
+	{ 0x8086, 0x3a3e, PCI_ANY_ID, PCI_ANY_ID, 0, 0, AZX_DRIVER_ICH }, /* ICH10 */
+	{ 0x8086, 0x3a6e, PCI_ANY_ID, PCI_ANY_ID, 0, 0, AZX_DRIVER_ICH }, /* ICH10 */
 	{ 0x1002, 0x437b, PCI_ANY_ID, PCI_ANY_ID, 0, 0, AZX_DRIVER_ATI }, /* ATI SB450 */
 	{ 0x1002, 0x4383, PCI_ANY_ID, PCI_ANY_ID, 0, 0, AZX_DRIVER_ATI }, /* ATI SB600 */
 	{ 0x1002, 0x793b, PCI_ANY_ID, PCI_ANY_ID, 0, 0, AZX_DRIVER_ATIHDMI }, /* ATI RS600 HDMI */