Merge branch 'core/percpu' into x86/core

40 files changed, 903 insertions, 571 deletions

diff --git a/arch/ia64/include/asm/percpu.h b/arch/ia64/include/asm/percpu.h
index 77f30b6..30cf465 100644
--- a/arch/ia64/include/asm/percpu.h
+++ b/arch/ia64/include/asm/percpu.h
@@ -27,12 +27,12 @@ extern void *per_cpu_init(void);
 
 #else /* ! SMP */
 
-#define PER_CPU_ATTRIBUTES	__attribute__((__section__(".data.percpu")))
-
 #define per_cpu_init()				(__phys_per_cpu_start)
 
 #endif	/* SMP */
 
+#define PER_CPU_BASE_SECTION ".data.percpu"
+
 /*
  * Be extremely careful when taking the address of this variable!  Due to virtual
  * remapping, it is different from the canonical address returned by __get_cpu_var(var)!
diff --git a/arch/ia64/include/asm/uv/uv.h b/arch/ia64/include/asm/uv/uv.h
new file mode 100644
index 0000000..61b5bdf
--- /dev/null
+++ b/arch/ia64/include/asm/uv/uv.h
@@ -0,0 +1,13 @@
+#ifndef _ASM_IA64_UV_UV_H
+#define _ASM_IA64_UV_UV_H
+
+#include <asm/system.h>
+#include <asm/sn/simulator.h>
+
+static inline int is_uv_system(void)
+{
+	/* temporary support for running on hardware simulator */
+	return IS_MEDUSA() || ia64_platform_is("uv");
+}
+
+#endif	/* _ASM_IA64_UV_UV_H */
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 148c112..1042d69 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -194,6 +194,10 @@ config X86_TRAMPOLINE
 	depends on SMP || (64BIT && ACPI_SLEEP)
 	default y
 
+config X86_32_LAZY_GS
+	def_bool y
+	depends on X86_32 && !CC_STACKPROTECTOR
+
 config KTIME_SCALAR
 	def_bool X86_32
 source "init/Kconfig"
@@ -1339,7 +1343,6 @@ config CC_STACKPROTECTOR_ALL
 
 config CC_STACKPROTECTOR
 	bool "Enable -fstack-protector buffer overflow detection (EXPERIMENTAL)"
-	depends on X86_64
 	select CC_STACKPROTECTOR_ALL
 	---help---
 	  This option turns on the -fstack-protector GCC feature. This
diff --git a/arch/x86/Makefile b/arch/x86/Makefile
index 99550c4..1836191 100644
--- a/arch/x86/Makefile
+++ b/arch/x86/Makefile
@@ -70,14 +70,17 @@ else
         # this works around some issues with generating unwind tables in older gccs
         # newer gccs do it by default
         KBUILD_CFLAGS += -maccumulate-outgoing-args
+endif
 
-        stackp := $(CONFIG_SHELL) $(srctree)/scripts/gcc-x86_64-has-stack-protector.sh
-        stackp-$(CONFIG_CC_STACKPROTECTOR) := $(shell $(stackp) \
-                "$(CC)" "-fstack-protector -DGCC_HAS_SP" )
-        stackp-$(CONFIG_CC_STACKPROTECTOR_ALL) += $(shell $(stackp) \
-                "$(CC)" -fstack-protector-all )
-
-        KBUILD_CFLAGS += $(stackp-y)
+ifdef CONFIG_CC_STACKPROTECTOR
+	cc_has_sp := $(srctree)/scripts/gcc-x86_$(BITS)-has-stack-protector.sh
+        ifeq ($(shell $(CONFIG_SHELL) $(cc_has_sp) $(CC)),y)
+                stackp-y := -fstack-protector
+                stackp-$(CONFIG_CC_STACKPROTECTOR_ALL) += -fstack-protector-all
+                KBUILD_CFLAGS += $(stackp-y)
+        else
+                $(warning stack protector enabled but no compiler support)
+        endif
 endif
 
 # Stackpointer is addressed different for 32 bit and 64 bit x86
diff --git a/arch/x86/include/asm/a.out-core.h b/arch/x86/include/asm/a.out-core.h
index 3c601f8..bb70e39 100644
--- a/arch/x86/include/asm/a.out-core.h
+++ b/arch/x86/include/asm/a.out-core.h
@@ -55,7 +55,7 @@ static inline void aout_dump_thread(struct pt_regs *regs, struct user *dump)
 	dump->regs.ds = (u16)regs->ds;
 	dump->regs.es = (u16)regs->es;
 	dump->regs.fs = (u16)regs->fs;
-	savesegment(gs, dump->regs.gs);
+	dump->regs.gs = get_user_gs(regs);
 	dump->regs.orig_ax = regs->orig_ax;
 	dump->regs.ip = regs->ip;
 	dump->regs.cs = (u16)regs->cs;
diff --git a/arch/x86/include/asm/elf.h b/arch/x86/include/asm/elf.h
index f51a3dd..83c1bc8 100644
--- a/arch/x86/include/asm/elf.h
+++ b/arch/x86/include/asm/elf.h
@@ -112,7 +112,7 @@ extern unsigned int vdso_enabled;
  * now struct_user_regs, they are different)
  */
 
-#define ELF_CORE_COPY_REGS(pr_reg, regs)	\
+#define ELF_CORE_COPY_REGS_COMMON(pr_reg, regs)	\
 do {						\
 	pr_reg[0] = regs->bx;			\
 	pr_reg[1] = regs->cx;			\
@@ -124,7 +124,6 @@ do {						\
 	pr_reg[7] = regs->ds & 0xffff;		\
 	pr_reg[8] = regs->es & 0xffff;		\
 	pr_reg[9] = regs->fs & 0xffff;		\
-	savesegment(gs, pr_reg[10]);		\
 	pr_reg[11] = regs->orig_ax;		\
 	pr_reg[12] = regs->ip;			\
 	pr_reg[13] = regs->cs & 0xffff;		\
@@ -133,6 +132,18 @@ do {						\
 	pr_reg[16] = regs->ss & 0xffff;		\
 } while (0);
 
+#define ELF_CORE_COPY_REGS(pr_reg, regs)	\
+do {						\
+	ELF_CORE_COPY_REGS_COMMON(pr_reg, regs);\
+	pr_reg[10] = get_user_gs(regs);		\
+} while (0);
+
+#define ELF_CORE_COPY_KERNEL_REGS(pr_reg, regs)	\
+do {						\
+	ELF_CORE_COPY_REGS_COMMON(pr_reg, regs);\
+	savesegment(gs, pr_reg[10]);		\
+} while (0);
+
 #define ELF_PLATFORM	(utsname()->machine)
 #define set_personality_64bit()	do { } while (0)
 
diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h
index 52948df..f923203 100644
--- a/arch/x86/include/asm/mmu_context.h
+++ b/arch/x86/include/asm/mmu_context.h
@@ -79,7 +79,7 @@ do {						\
 #ifdef CONFIG_X86_32
 #define deactivate_mm(tsk, mm)			\
 do {						\
-	loadsegment(gs, 0);			\
+	lazy_load_gs(0);			\
 } while (0)
 #else
 #define deactivate_mm(tsk, mm)			\
diff --git a/arch/x86/include/asm/percpu.h b/arch/x86/include/asm/percpu.h
index 0b64af4..aee103b 100644
--- a/arch/x86/include/asm/percpu.h
+++ b/arch/x86/include/asm/percpu.h
@@ -34,6 +34,12 @@
 #define PER_CPU_VAR(var)	per_cpu__##var
 #endif	/* SMP */
 
+#ifdef CONFIG_X86_64_SMP
+#define INIT_PER_CPU_VAR(var)  init_per_cpu__##var
+#else
+#define INIT_PER_CPU_VAR(var)  per_cpu__##var
+#endif
+
 #else /* ...!ASSEMBLY */
 
 #include <linux/stringify.h>
@@ -45,6 +51,22 @@
 #define __percpu_arg(x)		"%" #x
 #endif
 
+/*
+ * Initialized pointers to per-cpu variables needed for the boot
+ * processor need to use these macros to get the proper address
+ * offset from __per_cpu_load on SMP.
+ *
+ * There also must be an entry in vmlinux_64.lds.S
+ */
+#define DECLARE_INIT_PER_CPU(var) \
+       extern typeof(per_cpu_var(var)) init_per_cpu_var(var)
+
+#ifdef CONFIG_X86_64_SMP
+#define init_per_cpu_var(var)  init_per_cpu__##var
+#else
+#define init_per_cpu_var(var)  per_cpu_var(var)
+#endif
+
 /* For arch-specific code, we can use direct single-insn ops (they
  * don't give an lvalue though). */
 extern void __bad_percpu_size(void);
diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h
index a6643f6..a013383 100644
--- a/arch/x86/include/asm/processor.h
+++ b/arch/x86/include/asm/processor.h
@@ -393,8 +393,14 @@ union irq_stack_union {
 };
 
 DECLARE_PER_CPU(union irq_stack_union, irq_stack_union);
+DECLARE_INIT_PER_CPU(irq_stack_union);
+
 DECLARE_PER_CPU(char *, irq_stack_ptr);
+#else	/* X86_64 */
+#ifdef CONFIG_CC_STACKPROTECTOR
+DECLARE_PER_CPU(unsigned long, stack_canary);
 #endif
+#endif	/* X86_64 */
 
 extern void print_cpu_info(struct cpuinfo_x86 *);
 extern unsigned int xstate_size;
diff --git a/arch/x86/include/asm/ptrace.h b/arch/x86/include/asm/ptrace.h
index 6d34d95..e304b66 100644
--- a/arch/x86/include/asm/ptrace.h
+++ b/arch/x86/include/asm/ptrace.h
@@ -28,7 +28,7 @@ struct pt_regs {
 	int  xds;
 	int  xes;
 	int  xfs;
-	/* int  gs; */
+	int  xgs;
 	long orig_eax;
 	long eip;
 	int  xcs;
@@ -50,7 +50,7 @@ struct pt_regs {
 	unsigned long ds;
 	unsigned long es;
 	unsigned long fs;
-	/* int  gs; */
+	unsigned long gs;
 	unsigned long orig_ax;
 	unsigned long ip;
 	unsigned long cs;
diff --git a/arch/x86/include/asm/segment.h b/arch/x86/include/asm/segment.h
index 1dc1b51..14e0ed8 100644
--- a/arch/x86/include/asm/segment.h
+++ b/arch/x86/include/asm/segment.h
@@ -61,7 +61,7 @@
  *
  *  26 - ESPFIX small SS
  *  27 - per-cpu			[ offset to per-cpu data area ]
- *  28 - unused
+ *  28 - stack_canary-20		[ for stack protector ]
  *  29 - unused
  *  30 - unused
  *  31 - TSS for double fault handler
@@ -95,6 +95,13 @@
 #define __KERNEL_PERCPU 0
 #endif
 
+#define GDT_ENTRY_STACK_CANARY		(GDT_ENTRY_KERNEL_BASE + 16)
+#ifdef CONFIG_CC_STACKPROTECTOR
+#define __KERNEL_STACK_CANARY		(GDT_ENTRY_STACK_CANARY * 8)
+#else
+#define __KERNEL_STACK_CANARY		0
+#endif
+
 #define GDT_ENTRY_DOUBLEFAULT_TSS	31
 
 /*
diff --git a/arch/x86/include/asm/stackprotector.h b/arch/x86/include/asm/stackprotector.h
index 36a700a..c2d742c 100644
--- a/arch/x86/include/asm/stackprotector.h
+++ b/arch/x86/include/asm/stackprotector.h
@@ -1,8 +1,54 @@
+/*
+ * GCC stack protector support.
+ *
+ * Stack protector works by putting predefined pattern at the start of
+ * the stack frame and verifying that it hasn't been overwritten when
+ * returning from the function.  The pattern is called stack canary
+ * and unfortunately gcc requires it to be at a fixed offset from %gs.
+ * On x86_64, the offset is 40 bytes and on x86_32 20 bytes.  x86_64
+ * and x86_32 use segment registers differently and thus handles this
+ * requirement differently.
+ *
+ * On x86_64, %gs is shared by percpu area and stack canary.  All
+ * percpu symbols are zero based and %gs points to the base of percpu
+ * area.  The first occupant of the percpu area is always
+ * irq_stack_union which contains stack_canary at offset 40.  Userland
+ * %gs is always saved and restored on kernel entry and exit using
+ * swapgs, so stack protector doesn't add any complexity there.
+ *
+ * On x86_32, it's slightly more complicated.  As in x86_64, %gs is
+ * used for userland TLS.  Unfortunately, some processors are much
+ * slower at loading segment registers with different value when
+ * entering and leaving the kernel, so the kernel uses %fs for percpu
+ * area and manages %gs lazily so that %gs is switched only when
+ * necessary, usually during task switch.
+ *
+ * As gcc requires the stack canary at %gs:20, %gs can't be managed
+ * lazily if stack protector is enabled, so the kernel saves and
+ * restores userland %gs on kernel entry and exit.  This behavior is
+ * controlled by CONFIG_X86_32_LAZY_GS and accessors are defined in
+ * system.h to hide the details.
+ */
+
 #ifndef _ASM_STACKPROTECTOR_H
 #define _ASM_STACKPROTECTOR_H 1
 
+#ifdef CONFIG_CC_STACKPROTECTOR
+
 #include <asm/tsc.h>
 #include <asm/processor.h>
+#include <asm/percpu.h>
+#include <asm/system.h>
+#include <asm/desc.h>
+#include <linux/random.h>
+
+/*
+ * 24 byte read-only segment initializer for stack canary.  Linker
+ * can't handle the address bit shifting.  Address will be set in
+ * head_32 for boot CPU and setup_per_cpu_areas() for others.
+ */
+#define GDT_STACK_CANARY_INIT						\
+	[GDT_ENTRY_STACK_CANARY] = { { { 0x00000018, 0x00409000 } } },
 
 /*
  * Initialize the stackprotector canary value.
@@ -15,12 +61,9 @@ static __always_inline void boot_init_stack_canary(void)
 	u64 canary;
 	u64 tsc;
 
-	/*
-	 * Build time only check to make sure the stack_canary is at
-	 * offset 40 in the pda; this is a gcc ABI requirement
-	 */
+#ifdef CONFIG_X86_64
 	BUILD_BUG_ON(offsetof(union irq_stack_union, stack_canary) != 40);
-
+#endif
 	/*
 	 * We both use the random pool and the current TSC as a source
 	 * of randomness. The TSC only matters for very early init,
@@ -32,7 +75,50 @@ static __always_inline void boot_init_stack_canary(void)
 	canary += tsc + (tsc << 32UL);
 
 	current->stack_canary = canary;
+#ifdef CONFIG_X86_64
 	percpu_write(irq_stack_union.stack_canary, canary);
+#else
+	percpu_write(stack_canary, canary);
+#endif
 }
 
+static inline void setup_stack_canary_segment(int cpu)
+{
+#ifdef CONFIG_X86_32
+	unsigned long canary = (unsigned long)&per_cpu(stack_canary, cpu) - 20;
+	struct desc_struct *gdt_table = get_cpu_gdt_table(cpu);
+	struct desc_struct desc;
+
+	desc = gdt_table[GDT_ENTRY_STACK_CANARY];
+	desc.base0 = canary & 0xffff;
+	desc.base1 = (canary >> 16) & 0xff;
+	desc.base2 = (canary >> 24) & 0xff;
+	write_gdt_entry(gdt_table, GDT_ENTRY_STACK_CANARY, &desc, DESCTYPE_S);
+#endif
+}
+
+static inline void load_stack_canary_segment(void)
+{
+#ifdef CONFIG_X86_32
+	asm("mov %0, %%gs" : : "r" (__KERNEL_STACK_CANARY) : "memory");
+#endif
+}
+
+#else	/* CC_STACKPROTECTOR */
+
+#define GDT_STACK_CANARY_INIT
+
+/* dummy boot_init_stack_canary() is defined in linux/stackprotector.h */
+
+static inline void setup_stack_canary_segment(int cpu)
+{ }
+
+static inline void load_stack_canary_segment(void)
+{
+#ifdef CONFIG_X86_32
+	asm volatile ("mov %0, %%gs" : : "r" (0));
 #endif
+}
+
+#endif	/* CC_STACKPROTECTOR */
+#endif	/* _ASM_STACKPROTECTOR_H */
diff --git a/arch/x86/include/asm/syscalls.h b/arch/x86/include/asm/syscalls.h
index c0b0bda..68b1be1 100644
--- a/arch/x86/include/asm/syscalls.h
+++ b/arch/x86/include/asm/syscalls.h
@@ -29,21 +29,21 @@ asmlinkage int sys_get_thread_area(struct user_desc __user *);
 /* X86_32 only */
 #ifdef CONFIG_X86_32
 /* kernel/process_32.c */
-asmlinkage int sys_fork(struct pt_regs);
-asmlinkage int sys_clone(struct pt_regs);
-asmlinkage int sys_vfork(struct pt_regs);
-asmlinkage int sys_execve(struct pt_regs);
+int sys_fork(struct pt_regs *);
+int sys_clone(struct pt_regs *);
+int sys_vfork(struct pt_regs *);
+int sys_execve(struct pt_regs *);
 
 /* kernel/signal_32.c */
 asmlinkage int sys_sigsuspend(int, int, old_sigset_t);
 asmlinkage int sys_sigaction(int, const struct old_sigaction __user *,
 			     struct old_sigaction __user *);
-asmlinkage int sys_sigaltstack(unsigned long);
-asmlinkage unsigned long sys_sigreturn(unsigned long);
-asmlinkage int sys_rt_sigreturn(unsigned long);
+int sys_sigaltstack(struct pt_regs *);
+unsigned long sys_sigreturn(struct pt_regs *);
+long sys_rt_sigreturn(struct pt_regs *);
 
 /* kernel/ioport.c */
-asmlinkage long sys_iopl(unsigned long);
+long sys_iopl(struct pt_regs *);
 
 /* kernel/sys_i386_32.c */
 asmlinkage long sys_mmap2(unsigned long, unsigned long, unsigned long,
@@ -59,8 +59,8 @@ struct oldold_utsname;
 asmlinkage int sys_olduname(struct oldold_utsname __user *);
 
 /* kernel/vm86_32.c */
-asmlinkage int sys_vm86old(struct pt_regs);
-asmlinkage int sys_vm86(struct pt_regs);
+int sys_vm86old(struct pt_regs *);
+int sys_vm86(struct pt_regs *);
 
 #else /* CONFIG_X86_32 */
 
diff --git a/arch/x86/include/asm/system.h b/arch/x86/include/asm/system.h
index c223837..c00bfdb 100644
--- a/arch/x86/include/asm/system.h
+++ b/arch/x86/include/asm/system.h
@@ -23,6 +23,20 @@ struct task_struct *__switch_to(struct task_struct *prev,
 
 #ifdef CONFIG_X86_32
 
+#ifdef CONFIG_CC_STACKPROTECTOR
+#define __switch_canary							\
+	"movl %P[task_canary](%[next]), %%ebx\n\t"			\
+	"movl %%ebx, "__percpu_arg([stack_canary])"\n\t"
+#define __switch_canary_oparam						\
+	, [stack_canary] "=m" (per_cpu_var(stack_canary))
+#define __switch_canary_iparam						\
+	, [task_canary] "i" (offsetof(struct task_struct, stack_canary))
+#else	/* CC_STACKPROTECTOR */
+#define __switch_canary
+#define __switch_canary_oparam
+#define __switch_canary_iparam
+#endif	/* CC_STACKPROTECTOR */
+
 /*
  * Saving eflags is important. It switches not only IOPL between tasks,
  * it also protects other tasks from NT leaking through sysenter etc.
@@ -44,6 +58,7 @@ do {									\
 		     "movl %[next_sp],%%esp\n\t"	/* restore ESP   */ \
 		     "movl $1f,%[prev_ip]\n\t"	/* save    EIP   */	\
 		     "pushl %[next_ip]\n\t"	/* restore EIP   */	\
+		     __switch_canary					\
 		     "jmp __switch_to\n"	/* regparm call  */	\
 		     "1:\t"						\
 		     "popl %%ebp\n\t"		/* restore EBP   */	\
@@ -58,6 +73,8 @@ do {									\
 		       "=b" (ebx), "=c" (ecx), "=d" (edx),		\
 		       "=S" (esi), "=D" (edi)				\
 		       							\
+		       __switch_canary_oparam				\
+									\
 		       /* input parameters: */				\
 		     : [next_sp]  "m" (next->thread.sp),		\
 		       [next_ip]  "m" (next->thread.ip),		\
@@ -66,6 +83,8 @@ do {									\
 		       [prev]     "a" (prev),				\
 		       [next]     "d" (next)				\
 									\
+		       __switch_canary_iparam				\
+									\
 		     : /* reloaded segment registers */			\
 			"memory");					\
 } while (0)
@@ -182,6 +201,25 @@ extern void native_load_gs_index(unsigned);
 #define savesegment(seg, value)				\
 	asm("mov %%" #seg ",%0":"=r" (value) : : "memory")
 
+/*
+ * x86_32 user gs accessors.
+ */
+#ifdef CONFIG_X86_32
+#ifdef CONFIG_X86_32_LAZY_GS
+#define get_user_gs(regs)	(u16)({unsigned long v; savesegment(gs, v); v;})
+#define set_user_gs(regs, v)	loadsegment(gs, (unsigned long)(v))
+#define task_user_gs(tsk)	((tsk)->thread.gs)
+#define lazy_save_gs(v)		savesegment(gs, (v))
+#define lazy_load_gs(v)		loadsegment(gs, (v))
+#else	/* X86_32_LAZY_GS */
+#define get_user_gs(regs)	(u16)((regs)->gs)
+#define set_user_gs(regs, v)	do { (regs)->gs = (v); } while (0)
+#define task_user_gs(tsk)	(task_pt_regs(tsk)->gs)
+#define lazy_save_gs(v)		do { } while (0)
+#define lazy_load_gs(v)		do { } while (0)
+#endif	/* X86_32_LAZY_GS */
+#endif	/* X86_32 */
+
 static inline unsigned long get_limit(unsigned long segment)
 {
 	unsigned long __limit;
diff --git a/arch/x86/include/asm/traps.h b/arch/x86/include/asm/traps.h
index cf3bb05..0d53425 100644
--- a/arch/x86/include/asm/traps.h
+++ b/arch/x86/include/asm/traps.h
@@ -41,7 +41,7 @@ dotraplinkage void do_int3(struct pt_regs *, long);
 dotraplinkage void do_overflow(struct pt_regs *, long);
 dotraplinkage void do_bounds(struct pt_regs *, long);
 dotraplinkage void do_invalid_op(struct pt_regs *, long);
-dotraplinkage void do_device_not_available(struct pt_regs);
+dotraplinkage void do_device_not_available(struct pt_regs *, long);
 dotraplinkage void do_coprocessor_segment_overrun(struct pt_regs *, long);
 dotraplinkage void do_invalid_TSS(struct pt_regs *, long);
 dotraplinkage void do_segment_not_present(struct pt_regs *, long);
diff --git a/arch/x86/include/asm/uaccess.h b/arch/x86/include/asm/uaccess.h
index 0ec6de4..b685ece 100644
--- a/arch/x86/include/asm/uaccess.h
+++ b/arch/x86/include/asm/uaccess.h
@@ -186,7 +186,7 @@ extern int __get_user_bad(void);
 
 
 #ifdef CONFIG_X86_32
-#define __put_user_asm_u64(x, addr, err)				\
+#define __put_user_asm_u64(x, addr, err, errret)			\
 	asm volatile("1:	movl %%eax,0(%2)\n"			\
 		     "2:	movl %%edx,4(%2)\n"			\
 		     "3:\n"						\
@@ -197,7 +197,7 @@ extern int __get_user_bad(void);
 		     _ASM_EXTABLE(1b, 4b)				\
 		     _ASM_EXTABLE(2b, 4b)				\
 		     : "=r" (err)					\
-		     : "A" (x), "r" (addr), "i" (-EFAULT), "0" (err))
+		     : "A" (x), "r" (addr), "i" (errret), "0" (err))
 
 #define __put_user_asm_ex_u64(x, addr)					\
 	asm volatile("1:	movl %%eax,0(%1)\n"			\
@@ -211,8 +211,8 @@ extern int __get_user_bad(void);
 	asm volatile("call __put_user_8" : "=a" (__ret_pu)	\
 		     : "A" ((typeof(*(ptr)))(x)), "c" (ptr) : "ebx")
 #else
-#define __put_user_asm_u64(x, ptr, retval) \
-	__put_user_asm(x, ptr, retval, "q", "", "Zr", -EFAULT)
+#define __put_user_asm_u64(x, ptr, retval, errret) \
+	__put_user_asm(x, ptr, retval, "q", "", "Zr", errret)
 #define __put_user_asm_ex_u64(x, addr)	\
 	__put_user_asm_ex(x, addr, "q", "", "Zr")
 #define __put_user_x8(x, ptr, __ret_pu) __put_user_x(8, x, ptr, __ret_pu)
@@ -289,7 +289,8 @@ do {									\
 		__put_user_asm(x, ptr, retval, "l", "k", "ir", errret);	\
 		break;							\
 	case 8:								\
-		__put_user_asm_u64((__typeof__(*ptr))(x), ptr, retval);	\
+		__put_user_asm_u64((__typeof__(*ptr))(x), ptr, retval,	\
+				   errret);				\
 		break;							\
 	default:							\
 		__put_user_bad();					\
@@ -525,8 +526,6 @@ struct __large_struct { unsigned long buf[100]; };
  */
 #define get_user_try		uaccess_try
 #define get_user_catch(err)	uaccess_catch(err)
-#define put_user_try		uaccess_try
-#define put_user_catch(err)	uaccess_catch(err)
 
 #define get_user_ex(x, ptr)	do {					\
 	unsigned long __gue_val;					\
@@ -534,9 +533,29 @@ struct __large_struct { unsigned long buf[100]; };
 	(x) = (__force __typeof__(*(ptr)))__gue_val;			\
 } while (0)
 
+#ifdef CONFIG_X86_WP_WORKS_OK
+
+#define put_user_try		uaccess_try
+#define put_user_catch(err)	uaccess_catch(err)
+
 #define put_user_ex(x, ptr)						\
 	__put_user_size_ex((__typeof__(*(ptr)))(x), (ptr), sizeof(*(ptr)))
 
+#else /* !CONFIG_X86_WP_WORKS_OK */
+
+#define put_user_try		do {		\
+	int __uaccess_err = 0;
+
+#define put_user_catch(err)			\
+	(err) |= __uaccess_err;			\
+} while (0)
+
+#define put_user_ex(x, ptr)	do {		\
+	__uaccess_err |= __put_user(x, ptr);	\
+} while (0)
+
+#endif /* CONFIG_X86_WP_WORKS_OK */
+
 /*
  * movsl can be slow when source and dest are not both 8-byte aligned
  */
diff --git a/arch/x86/include/asm/uv/uv.h b/arch/x86/include/asm/uv/uv.h
index 8ac1d7e..8242bf9 100644
--- a/arch/x86/include/asm/uv/uv.h
+++ b/arch/x86/include/asm/uv/uv.h
@@ -3,6 +3,9 @@
 
 enum uv_system_type {UV_NONE, UV_LEGACY_APIC, UV_X2APIC, UV_NON_UNIQUE_APIC};
 
+struct cpumask;
+struct mm_struct;
+
 #ifdef CONFIG_X86_UV
 
 extern enum uv_system_type get_uv_system_type(void);
diff --git a/arch/x86/kernel/asm-offsets_32.c b/arch/x86/kernel/asm-offsets_32.c
index ee4df08..fbf2f33 100644
--- a/arch/x86/kernel/asm-offsets_32.c
+++ b/arch/x86/kernel/asm-offsets_32.c
@@ -75,6 +75,7 @@ void foo(void)
 	OFFSET(PT_DS,  pt_regs, ds);
 	OFFSET(PT_ES,  pt_regs, es);
 	OFFSET(PT_FS,  pt_regs, fs);
+	OFFSET(PT_GS,  pt_regs, gs);
 	OFFSET(PT_ORIG_EAX, pt_regs, orig_ax);
 	OFFSET(PT_EIP, pt_regs, ip);
 	OFFSET(PT_CS,  pt_regs, cs);
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index cbcdb79..e8f4a38 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -39,6 +39,7 @@
 #include <asm/sections.h>
 #include <asm/setup.h>
 #include <asm/hypervisor.h>
+#include <asm/stackprotector.h>
 
 #include "cpu.h"
 
@@ -122,6 +123,7 @@ DEFINE_PER_CPU_PAGE_ALIGNED(struct gdt_page, gdt_page) = { .gdt = {
 
 	[GDT_ENTRY_ESPFIX_SS] = { { { 0x00000000, 0x00c09200 } } },
 	[GDT_ENTRY_PERCPU] = { { { 0x0000ffff, 0x00cf9200 } } },
+	GDT_STACK_CANARY_INIT
 #endif
 } };
 EXPORT_PER_CPU_SYMBOL_GPL(gdt_page);
@@ -304,6 +306,7 @@ void load_percpu_segment(int cpu)
 	loadsegment(gs, 0);
 	wrmsrl(MSR_GS_BASE, (unsigned long)per_cpu(irq_stack_union.gs_base, cpu));
 #endif
+	load_stack_canary_segment();
 }
 
 /* Current gdt points %fs at the "master" per-cpu area: after this,
@@ -938,12 +941,8 @@ struct desc_ptr idt_descr = { 256 * 16 - 1, (unsigned long) idt_table };
 
 DEFINE_PER_CPU_FIRST(union irq_stack_union,
 		     irq_stack_union) __aligned(PAGE_SIZE);
-#ifdef CONFIG_SMP
-DEFINE_PER_CPU(char *, irq_stack_ptr);	/* will be set during per cpu init */
-#else
 DEFINE_PER_CPU(char *, irq_stack_ptr) =
-	per_cpu_var(irq_stack_union.irq_stack) + IRQ_STACK_SIZE - 64;
-#endif
+	init_per_cpu_var(irq_stack_union.irq_stack) + IRQ_STACK_SIZE - 64;
 
 DEFINE_PER_CPU(unsigned long, kernel_stack) =
 	(unsigned long)&init_thread_union - KERNEL_STACK_OFFSET + THREAD_SIZE;
@@ -986,16 +985,21 @@ unsigned long kernel_eflags;
  */
 DEFINE_PER_CPU(struct orig_ist, orig_ist);
 
-#else
+#else	/* x86_64 */
 
-/* Make sure %fs is initialized properly in idle threads */
+#ifdef CONFIG_CC_STACKPROTECTOR
+DEFINE_PER_CPU(unsigned long, stack_canary);
+#endif
+
+/* Make sure %fs and %gs are initialized properly in idle threads */
 struct pt_regs * __cpuinit idle_regs(struct pt_regs *regs)
 {
 	memset(regs, 0, sizeof(struct pt_regs));
 	regs->fs = __KERNEL_PERCPU;
+	regs->gs = __KERNEL_STACK_CANARY;
 	return regs;
 }
-#endif
+#endif	/* x86_64 */
 
 /*
  * cpu_init() initializes state that is per-CPU. Some data is already
@@ -1157,9 +1161,6 @@ void __cpuinit cpu_init(void)
 	__set_tss_desc(cpu, GDT_ENTRY_DOUBLEFAULT_TSS, &doublefault_tss);
 #endif
 
-	/* Clear %gs. */
-	asm volatile ("mov %0, %%gs" : : "r" (0));
-
 	/* Clear all 6 debug registers: */
 	set_debugreg(0, 0);
 	set_debugreg(0, 1);
diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S
index 65efd42..e992068 100644
--- a/arch/x86/kernel/entry_32.S
+++ b/arch/x86/kernel/entry_32.S
@@ -30,12 +30,13 @@
  *	1C(%esp) - %ds
  *	20(%esp) - %es
  *	24(%esp) - %fs
- *	28(%esp) - orig_eax
- *	2C(%esp) - %eip
- *	30(%esp) - %cs
- *	34(%esp) - %eflags
- *	38(%esp) - %oldesp
- *	3C(%esp) - %oldss
+ *	28(%esp) - %gs		saved iff !CONFIG_X86_32_LAZY_GS
+ *	2C(%esp) - orig_eax
+ *	30(%esp) - %eip
+ *	34(%esp) - %cs
+ *	38(%esp) - %eflags
+ *	3C(%esp) - %oldesp
+ *	40(%esp) - %oldss
  *
  * "current" is in register %ebx during any slow entries.
  */
@@ -101,121 +102,221 @@
 #define resume_userspace_sig	resume_userspace
 #endif
 
-#define SAVE_ALL \
-	cld; \
-	pushl %fs; \
-	CFI_ADJUST_CFA_OFFSET 4;\
-	/*CFI_REL_OFFSET fs, 0;*/\
-	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; \
-	movl $(__KERNEL_PERCPU), %edx; \
+/*
+ * User gs save/restore
+ *
+ * %gs is used for userland TLS and kernel only uses it for stack
+ * canary which is required to be at %gs:20 by gcc.  Read the comment
+ * at the top of stackprotector.h for more info.
+ *
+ * Local labels 98 and 99 are used.
+ */
+#ifdef CONFIG_X86_32_LAZY_GS
+
+ /* unfortunately push/pop can't be no-op */
+.macro PUSH_GS
+	pushl $0
+	CFI_ADJUST_CFA_OFFSET 4
+.endm
+.macro POP_GS pop=0
+	addl $(4 + \pop), %esp
+	CFI_ADJUST_CFA_OFFSET -(4 + \pop)
+.endm
+.macro POP_GS_EX
+.endm
+
+ /* all the rest are no-op */
+.macro PTGS_TO_GS
+.endm
+.macro PTGS_TO_GS_EX
+.endm
+.macro GS_TO_REG reg
+.endm
+.macro REG_TO_PTGS reg
+.endm
+.macro SET_KERNEL_GS reg
+.endm
+
+#else	/* CONFIG_X86_32_LAZY_GS */
+
+.macro PUSH_GS
+	pushl %gs
+	CFI_ADJUST_CFA_OFFSET 4
+	/*CFI_REL_OFFSET gs, 0*/
+.endm
+
+.macro POP_GS pop=0
+98:	popl %gs
+	CFI_ADJUST_CFA_OFFSET -4
+	/*CFI_RESTORE gs*/
+  .if \pop <> 0
+	add $\pop, %esp
+	CFI_ADJUST_CFA_OFFSET -\pop
+  .endif
+.endm
+.macro POP_GS_EX
+.pushsection .fixup, "ax"
+99:	movl $0, (%esp)
+	jmp 98b
+.section __ex_table, "a"
+	.align 4
+	.long 98b, 99b
+.popsection
+.endm
+
+.macro PTGS_TO_GS
+98:	mov PT_GS(%esp), %gs
+.endm
+.macro PTGS_TO_GS_EX
+.pushsection .fixup, "ax"
+99:	movl $0, PT_GS(%esp)
+	jmp 98b
+.section __ex_table, "a"
+	.align 4
+	.long 98b, 99b
+.popsection
+.endm
+
+.macro GS_TO_REG reg
+	movl %gs, \reg
+	/*CFI_REGISTER gs, \reg*/
+.endm
+.macro REG_TO_PTGS reg
+	movl \reg, PT_GS(%esp)
+	/*CFI_REL_OFFSET gs, PT_GS*/
+.endm
+.macro SET_KERNEL_GS reg
+	movl $(__KERNEL_STACK_CANARY), \reg
+	movl \reg, %gs
+.endm
+
+#endif	/* CONFIG_X86_32_LAZY_GS */
+
+.macro SAVE_ALL
+	cld
+	PUSH_GS
+	pushl %fs
+	CFI_ADJUST_CFA_OFFSET 4
+	/*CFI_REL_OFFSET fs, 0;*/
+	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
+	movl $(__KERNEL_PERCPU), %edx
 	movl %edx, %fs
+	SET_KERNEL_GS %edx
+.endm
 
-#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;\
+.macro 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
+.endm
 
-#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;*/\
-3:	popl %fs;	\
-	CFI_ADJUST_CFA_OFFSET -4;\
-	/*CFI_RESTORE fs;*/\
-.pushsection .fixup,"ax";	\
-4:	movl $0,(%esp);	\
-	jmp 1b;		\
-5:	movl $0,(%esp);	\
-	jmp 2b;		\
-6:	movl $0,(%esp);	\
-	jmp 3b;		\
-.section __ex_table,"a";\
-	.align 4;	\
-	.long 1b,4b;	\
-	.long 2b,5b;	\
-	.long 3b,6b;	\
+.macro RESTORE_REGS pop=0
+	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;*/
+3:	popl %fs
+	CFI_ADJUST_CFA_OFFSET -4
+	/*CFI_RESTORE fs;*/
+	POP_GS \pop
+.pushsection .fixup, "ax"
+4:	movl $0, (%esp)
+	jmp 1b
+5:	movl $0, (%esp)
+	jmp 2b
+6:	movl $0, (%esp)
+	jmp 3b
+.section __ex_table, "a"
+	.align 4
+	.long 1b, 4b
+	.long 2b, 5b
+	.long 3b, 6b
 .popsection
+	POP_GS_EX
+.endm
 
-#define RING0_INT_FRAME \
-	CFI_STARTPROC simple;\
-	CFI_SIGNAL_FRAME;\
-	CFI_DEF_CFA esp, 3*4;\
-	/*CFI_OFFSET cs, -2*4;*/\
+.macro RING0_INT_FRAME
+	CFI_STARTPROC simple
+	CFI_SIGNAL_FRAME
+	CFI_DEF_CFA esp, 3*4
+	/*CFI_OFFSET cs, -2*4;*/
 	CFI_OFFSET eip, -3*4
+.endm
 
-#define RING0_EC_FRAME \
-	CFI_STARTPROC simple;\
-	CFI_SIGNAL_FRAME;\
-	CFI_DEF_CFA esp, 4*4;\
-	/*CFI_OFFSET cs, -2*4;*/\
+.macro RING0_EC_FRAME
+	CFI_STARTPROC simple
+	CFI_SIGNAL_FRAME
+	CFI_DEF_CFA esp, 4*4
+	/*CFI_OFFSET cs, -2*4;*/
 	CFI_OFFSET eip, -3*4
+.endm
 
-#define RING0_PTREGS_FRAME \
-	CFI_STARTPROC simple;\
-	CFI_SIGNAL_FRAME;\
-	CFI_DEF_CFA esp, PT_OLDESP-PT_EBX;\
-	/*CFI_OFFSET cs, PT_CS-PT_OLDESP;*/\
-	CFI_OFFSET eip, PT_EIP-PT_OLDESP;\
-	/*CFI_OFFSET es, PT_ES-PT_OLDESP;*/\
-	/*CFI_OFFSET ds, PT_DS-PT_OLDESP;*/\
-	CFI_OFFSET eax, PT_EAX-PT_OLDESP;\
-	CFI_OFFSET ebp, PT_EBP-PT_OLDESP;\
-	CFI_OFFSET edi, PT_EDI-PT_OLDESP;\
-	CFI_OFFSET esi, PT_ESI-PT_OLDESP;\
-	CFI_OFFSET edx, PT_EDX-PT_OLDESP;\
-	CFI_OFFSET ecx, PT_ECX-PT_OLDESP;\
+.macro RING0_PTREGS_FRAME
+	CFI_STARTPROC simple
+	CFI_SIGNAL_FRAME
+	CFI_DEF_CFA esp, PT_OLDESP-PT_EBX
+	/*CFI_OFFSET cs, PT_CS-PT_OLDESP;*/
+	CFI_OFFSET eip, PT_EIP-PT_OLDESP
+	/*CFI_OFFSET es, PT_ES-PT_OLDESP;*/
+	/*CFI_OFFSET ds, PT_DS-PT_OLDESP;*/
+	CFI_OFFSET eax, PT_EAX-PT_OLDESP
+	CFI_OFFSET ebp, PT_EBP-PT_OLDESP
+	CFI_OFFSET edi, PT_EDI-PT_OLDESP
+	CFI_OFFSET esi, PT_ESI-PT_OLDESP
+	CFI_OFFSET edx, PT_EDX-PT_OLDESP
+	CFI_OFFSET ecx, PT_ECX-PT_OLDESP
 	CFI_OFFSET ebx, PT_EBX-PT_OLDESP
+.endm
 
 ENTRY(ret_from_fork)
 	CFI_STARTPROC
@@ -362,6 +463,7 @@ sysenter_exit:
 	xorl %ebp,%ebp
 	TRACE_IRQS_ON
 1:	mov  PT_FS(%esp), %fs
+	PTGS_TO_GS
 	ENABLE_INTERRUPTS_SYSEXIT
 
 #ifdef CONFIG_AUDITSYSCALL
@@ -410,6 +512,7 @@ sysexit_audit:
 	.align 4
 	.long 1b,2b
 .popsection
+	PTGS_TO_GS_EX
 ENDPROC(ia32_sysenter_target)
 
 	# system call handler stub
@@ -452,8 +555,7 @@ restore_all:
 restore_nocheck:
 	TRACE_IRQS_IRET
 restore_nocheck_notrace:
-	RESTORE_REGS
-	addl $4, %esp			# skip orig_eax/error_code
+	RESTORE_REGS 4			# skip orig_eax/error_code
 	CFI_ADJUST_CFA_OFFSET -4
 irq_return:
 	INTERRUPT_RETURN
@@ -595,28 +697,50 @@ syscall_badsys:
 END(syscall_badsys)
 	CFI_ENDPROC
 
-#define FIXUP_ESPFIX_STACK \
-	/* since we are on a wrong stack, we cant make it a C code :( */ \
-	PER_CPU(gdt_page, %ebx); \
-	GET_DESC_BASE(GDT_ENTRY_ESPFIX_SS, %ebx, %eax, %ax, %al, %ah); \
-	addl %esp, %eax; \
-	pushl $__KERNEL_DS; \
-	CFI_ADJUST_CFA_OFFSET 4; \
-	pushl %eax; \
-	CFI_ADJUST_CFA_OFFSET 4; \
-	lss (%esp), %esp; \
-	CFI_ADJUST_CFA_OFFSET -8;
-#define UNWIND_ESPFIX_STACK \
-	movl %ss, %eax; \
-	/* see if on espfix stack */ \
-	cmpw $__ESPFIX_SS, %ax; \
-	jne 27f; \
-	movl $__KERNEL_DS, %eax; \
-	movl %eax, %ds; \
-	movl %eax, %es; \
-	/* switch to normal stack */ \
-	FIXUP_ESPFIX_STACK; \
-27:;
+/*
+ * System calls that need a pt_regs pointer.
+ */
+#define PTREGSCALL(name) \
+	ALIGN; \
+ptregs_##name: \
+	leal 4(%esp),%eax; \
+	jmp sys_##name;
+
+PTREGSCALL(iopl)
+PTREGSCALL(fork)
+PTREGSCALL(clone)
+PTREGSCALL(vfork)
+PTREGSCALL(execve)
+PTREGSCALL(sigaltstack)
+PTREGSCALL(sigreturn)
+PTREGSCALL(rt_sigreturn)
+PTREGSCALL(vm86)
+PTREGSCALL(vm86old)
+
+.macro FIXUP_ESPFIX_STACK
+	/* since we are on a wrong stack, we cant make it a C code :( */
+	PER_CPU(gdt_page, %ebx)
+	GET_DESC_BASE(GDT_ENTRY_ESPFIX_SS, %ebx, %eax, %ax, %al, %ah)
+	addl %esp, %eax
+	pushl $__KERNEL_DS
+	CFI_ADJUST_CFA_OFFSET 4
+	pushl %eax
+	CFI_ADJUST_CFA_OFFSET 4
+	lss (%esp), %esp
+	CFI_ADJUST_CFA_OFFSET -8
+.endm
+.macro UNWIND_ESPFIX_STACK
+	movl %ss, %eax
+	/* see if on espfix stack */
+	cmpw $__ESPFIX_SS, %ax
+	jne 27f
+	movl $__KERNEL_DS, %eax
+	movl %eax, %ds
+	movl %eax, %es
+	/* switch to normal stack */
+	FIXUP_ESPFIX_STACK
+27:
+.endm
 
 /*
  * Build the entry stubs and pointer table with some assembler magic.
@@ -1070,7 +1194,10 @@ ENTRY(page_fault)
 	CFI_ADJUST_CFA_OFFSET 4
 	ALIGN
 error_code:
-	/* the function address is in %fs's slot on the stack */
+	/* the function address is in %gs's slot on the stack */
+	pushl %fs
+	CFI_ADJUST_CFA_OFFSET 4
+	/*CFI_REL_OFFSET fs, 0*/
 	pushl %es
 	CFI_ADJUST_CFA_OFFSET 4
 	/*CFI_REL_OFFSET es, 0*/
@@ -1099,20 +1226,15 @@ error_code:
 	CFI_ADJUST_CFA_OFFSET 4
 	CFI_REL_OFFSET ebx, 0
 	cld
-	pushl %fs
-	CFI_ADJUST_CFA_OFFSET 4
-	/*CFI_REL_OFFSET fs, 0*/
 	movl $(__KERNEL_PERCPU), %ecx
 	movl %ecx, %fs
 	UNWIND_ESPFIX_STACK
-	popl %ecx
-	CFI_ADJUST_CFA_OFFSET -4
-	/*CFI_REGISTER es, ecx*/
-	movl PT_FS(%esp), %edi		# get the function address
+	GS_TO_REG %ecx
+	movl PT_GS(%esp), %edi		# get the function address
 	movl PT_ORIG_EAX(%esp), %edx	# get the error code
 	movl $-1, PT_ORIG_EAX(%esp)	# no syscall to restart
-	mov  %ecx, PT_FS(%esp)
-	/*CFI_REL_OFFSET fs, ES*/
+	REG_TO_PTGS %ecx
+	SET_KERNEL_GS %ecx
 	movl $(__USER_DS), %ecx
 	movl %ecx, %ds
 	movl %ecx, %es
@@ -1136,26 +1258,27 @@ END(page_fault)
  * 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 TSS_sysenter_sp0+offset(%esp),%esp;	\
-	CFI_DEF_CFA esp, 0;			\
-	CFI_UNDEFINED eip;			\
-	pushfl;					\
-	CFI_ADJUST_CFA_OFFSET 4;		\
-	pushl $__KERNEL_CS;			\
-	CFI_ADJUST_CFA_OFFSET 4;		\
-	pushl $sysenter_past_esp;		\
-	CFI_ADJUST_CFA_OFFSET 4;		\
+.macro FIX_STACK offset ok label
+	cmpw $__KERNEL_CS, 4(%esp)
+	jne \ok
+\label:
+	movl TSS_sysenter_sp0 + \offset(%esp), %esp
+	CFI_DEF_CFA esp, 0
+	CFI_UNDEFINED eip
+	pushfl
+	CFI_ADJUST_CFA_OFFSET 4
+	pushl $__KERNEL_CS
+	CFI_ADJUST_CFA_OFFSET 4
+	pushl $sysenter_past_esp
+	CFI_ADJUST_CFA_OFFSET 4
 	CFI_REL_OFFSET eip, 0
+.endm
 
 ENTRY(debug)
 	RING0_INT_FRAME
 	cmpl $ia32_sysenter_target,(%esp)
 	jne debug_stack_correct
-	FIX_STACK(12, debug_stack_correct, debug_esp_fix_insn)
+	FIX_STACK 12, debug_stack_correct, debug_esp_fix_insn
 debug_stack_correct:
 	pushl $-1			# mark this as an int
 	CFI_ADJUST_CFA_OFFSET 4
@@ -1213,7 +1336,7 @@ nmi_stack_correct:
 
 nmi_stack_fixup:
 	RING0_INT_FRAME
-	FIX_STACK(12,nmi_stack_correct, 1)
+	FIX_STACK 12, nmi_stack_correct, 1
 	jmp nmi_stack_correct
 
 nmi_debug_stack_check:
@@ -1224,7 +1347,7 @@ nmi_debug_stack_check:
 	jb nmi_stack_correct
 	cmpl $debug_esp_fix_insn,(%esp)
 	ja nmi_stack_correct
-	FIX_STACK(24,nmi_stack_correct, 1)
+	FIX_STACK 24, nmi_stack_correct, 1
 	jmp nmi_stack_correct
 
 nmi_espfix_stack:
diff --git a/arch/x86/kernel/head_32.S b/arch/x86/kernel/head_32.S
index 722464c..2a0aad7 100644
--- a/arch/x86/kernel/head_32.S
+++ b/arch/x86/kernel/head_32.S
@@ -19,6 +19,7 @@
 #include <asm/asm-offsets.h>
 #include <asm/setup.h>
 #include <asm/processor-flags.h>
+#include <asm/percpu.h>
 
 /* Physical address */
 #define pa(X) ((X) - __PAGE_OFFSET)
@@ -437,8 +438,26 @@ is386:	movl $2,%ecx		# set MP
 	movl $(__KERNEL_PERCPU), %eax
 	movl %eax,%fs			# set this cpu's percpu
 
-	xorl %eax,%eax			# Clear GS and LDT
+#ifdef CONFIG_CC_STACKPROTECTOR
+	/*
+	 * The linker can't handle this by relocation.  Manually set
+	 * base address in stack canary segment descriptor.
+	 */
+	cmpb $0,ready
+	jne 1f
+	movl $per_cpu__gdt_page,%eax
+	movl $per_cpu__stack_canary,%ecx
+	subl $20, %ecx
+	movw %cx, 8 * GDT_ENTRY_STACK_CANARY + 2(%eax)
+	shrl $16, %ecx
+	movb %cl, 8 * GDT_ENTRY_STACK_CANARY + 4(%eax)
+	movb %ch, 8 * GDT_ENTRY_STACK_CANARY + 7(%eax)
+1:
+#endif
+	movl $(__KERNEL_STACK_CANARY),%eax
 	movl %eax,%gs
+
+	xorl %eax,%eax			# Clear LDT
 	lldt %ax
 
 	cld			# gcc2 wants the direction flag cleared at all times
diff --git a/arch/x86/kernel/head_64.S b/arch/x86/kernel/head_64.S
index a0a2b5c..2e648e3 100644
--- a/arch/x86/kernel/head_64.S
+++ b/arch/x86/kernel/head_64.S
@@ -205,19 +205,6 @@ ENTRY(secondary_startup_64)
 	pushq $0
 	popfq
 
-#ifdef CONFIG_SMP
-	/*
-	 * Fix up static pointers that need __per_cpu_load added.  The assembler
-	 * is unable to do this directly.  This is only needed for the boot cpu.
-	 * These values are set up with the correct base addresses by C code for
-	 * secondary cpus.
-	 */
-	movq	initial_gs(%rip), %rax
-	cmpl	$0, per_cpu__cpu_number(%rax)
-	jne	1f
-	addq	%rax, early_gdt_descr_base(%rip)
-1:
-#endif
 	/*
 	 * We must switch to a new descriptor in kernel space for the GDT
 	 * because soon the kernel won't have access anymore to the userspace
@@ -275,11 +262,7 @@ ENTRY(secondary_startup_64)
 	ENTRY(initial_code)
 	.quad	x86_64_start_kernel
 	ENTRY(initial_gs)
-#ifdef CONFIG_SMP
-	.quad	__per_cpu_load
-#else
-	.quad	PER_CPU_VAR(irq_stack_union)
-#endif
+	.quad	INIT_PER_CPU_VAR(irq_stack_union)
 	__FINITDATA
 
 	ENTRY(stack_start)
@@ -425,7 +408,7 @@ NEXT_PAGE(level2_spare_pgt)
 early_gdt_descr:
 	.word	GDT_ENTRIES*8-1
 early_gdt_descr_base:
-	.quad	per_cpu__gdt_page
+	.quad	INIT_PER_CPU_VAR(gdt_page)
 
 ENTRY(phys_base)
 	/* This must match the first entry in level2_kernel_pgt */
diff --git a/arch/x86/kernel/ioport.c b/arch/x86/kernel/ioport.c
index b12208f..e41980a 100644
--- a/arch/x86/kernel/ioport.c
+++ b/arch/x86/kernel/ioport.c
@@ -131,9 +131,8 @@ static int do_iopl(unsigned int level, struct pt_regs *regs)
 }
 
 #ifdef CONFIG_X86_32
-asmlinkage long sys_iopl(unsigned long regsp)
+long sys_iopl(struct pt_regs *regs)
 {
-	struct pt_regs *regs = (struct pt_regs *)&regsp;
 	unsigned int level = regs->bx;
 	struct thread_struct *t = &current->thread;
 	int rc;
diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c
index 1a1ae8e..fec79ad 100644
--- a/arch/x86/kernel/process_32.c
+++ b/arch/x86/kernel/process_32.c
@@ -11,6 +11,7 @@
 
 #include <stdarg.h>
 
+#include <linux/stackprotector.h>
 #include <linux/cpu.h>
 #include <linux/errno.h>
 #include <linux/sched.h>
@@ -91,6 +92,15 @@ void cpu_idle(void)
 {
 	int cpu = smp_processor_id();
 
+	/*
+	 * If we're the non-boot CPU, nothing set the stack canary up
+	 * for us.  CPU0 already has it initialized but no harm in
+	 * doing it again.  This is a good place for updating it, as
+	 * we wont ever return from this function (so the invalid
+	 * canaries already on the stack wont ever trigger).
+	 */
+	boot_init_stack_canary();
+
 	current_thread_info()->status |= TS_POLLING;
 
 	/* endless idle loop with no priority at all */
@@ -131,7 +141,7 @@ void __show_regs(struct pt_regs *regs, int all)
 	if (user_mode_vm(regs)) {
 		sp = regs->sp;
 		ss = regs->ss & 0xffff;
-		savesegment(gs, gs);
+		gs = get_user_gs(regs);
 	} else {
 		sp = (unsigned long) (&regs->sp);
 		savesegment(ss, ss);
@@ -212,6 +222,7 @@ int kernel_thread(int (*fn)(void *), void *arg, unsigned long flags)
 	regs.ds = __USER_DS;
 	regs.es = __USER_DS;
 	regs.fs = __KERNEL_PERCPU;
+	regs.gs = __KERNEL_STACK_CANARY;
 	regs.orig_ax = -1;
 	regs.ip = (unsigned long) kernel_thread_helper;
 	regs.cs = __KERNEL_CS | get_kernel_rpl();
@@ -304,7 +315,7 @@ int copy_thread(int nr, unsigned long clone_flags, unsigned long sp,
 
 	p->thread.ip = (unsigned long) ret_from_fork;
 
-	savesegment(gs, p->thread.gs);
+	task_user_gs(p) = get_user_gs(regs);
 
 	tsk = current;
 	if (unlikely(test_tsk_thread_flag(tsk, TIF_IO_BITMAP))) {
@@ -342,7 +353,7 @@ int copy_thread(int nr, unsigned long clone_flags, unsigned long sp,
 void
 start_thread(struct pt_regs *regs, unsigned long new_ip, unsigned long new_sp)
 {
-	__asm__("movl %0, %%gs" : : "r"(0));
+	set_user_gs(regs, 0);
 	regs->fs		= 0;
 	set_fs(USER_DS);
 	regs->ds		= __USER_DS;
@@ -539,7 +550,7 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
 	 * used %fs or %gs (it does not today), or if the kernel is
 	 * running inside of a hypervisor layer.
 	 */
-	savesegment(gs, prev->gs);
+	lazy_save_gs(prev->gs);
 
 	/*
 	 * Load the per-thread Thread-Local Storage descriptor.
@@ -585,31 +596,31 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
 	 * Restore %gs if needed (which is common)
 	 */
 	if (prev->gs | next->gs)
-		loadsegment(gs, next->gs);
+		lazy_load_gs(next->gs);
 
 	percpu_write(current_task, next_p);
 
 	return prev_p;
 }
 
-asmlinkage int sys_fork(struct pt_regs regs)
+int sys_fork(struct pt_regs *regs)
 {
-	return do_fork(SIGCHLD, regs.sp, &regs, 0, NULL, NULL);
+	return do_fork(SIGCHLD, regs->sp, regs, 0, NULL, NULL);
 }
 
-asmlinkage int sys_clone(struct pt_regs regs)
+int sys_clone(struct pt_regs *regs)
 {
 	unsigned long clone_flags;
 	unsigned long newsp;
 	int __user *parent_tidptr, *child_tidptr;
 
-	clone_flags = regs.bx;
-	newsp = regs.cx;
-	parent_tidptr = (int __user *)regs.dx;
-	child_tidptr = (int __user *)regs.di;
+	clone_flags = regs->bx;
+	newsp = regs->cx;
+	parent_tidptr = (int __user *)regs->dx;
+	child_tidptr = (int __user *)regs->di;
 	if (!newsp)
-		newsp = regs.sp;
-	return do_fork(clone_flags, newsp, &regs, 0, parent_tidptr, child_tidptr);
+		newsp = regs->sp;
+	return do_fork(clone_flags, newsp, regs, 0, parent_tidptr, child_tidptr);
 }
 
 /*
@@ -622,27 +633,27 @@ asmlinkage int sys_clone(struct pt_regs regs)
  * do not have enough call-clobbered registers to hold all
  * the information you need.
  */
-asmlinkage int sys_vfork(struct pt_regs regs)
+int sys_vfork(struct pt_regs *regs)
 {
-	return do_fork(CLONE_VFORK | CLONE_VM | SIGCHLD, regs.sp, &regs, 0, NULL, NULL);
+	return do_fork(CLONE_VFORK | CLONE_VM | SIGCHLD, regs->sp, regs, 0, NULL, NULL);
 }
 
 /*
  * sys_execve() executes a new program.
  */
-asmlinkage int sys_execve(struct pt_regs regs)
+int sys_execve(struct pt_regs *regs)
 {
 	int error;
 	char *filename;
 
-	filename = getname((char __user *) regs.bx);
+	filename = getname((char __user *) regs->bx);
 	error = PTR_ERR(filename);
 	if (IS_ERR(filename))
 		goto out;
 	error = do_execve(filename,
-			(char __user * __user *) regs.cx,
-			(char __user * __user *) regs.dx,
-			&regs);
+			(char __user * __user *) regs->cx,
+			(char __user * __user *) regs->dx,
+			regs);
 	if (error == 0) {
 		/* Make sure we don't return using sysenter.. */
 		set_thread_flag(TIF_IRET);
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index 8eb169e..836ef65 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -120,12 +120,11 @@ void cpu_idle(void)
 	current_thread_info()->status |= TS_POLLING;
 
 	/*
-	 * If we're the non-boot CPU, nothing set the PDA stack
-	 * canary up for us - and if we are the boot CPU we have
-	 * a 0 stack canary. This is a good place for updating
-	 * it, as we wont ever return from this function (so the
-	 * invalid canaries already on the stack wont ever
-	 * trigger):
+	 * If we're the non-boot CPU, nothing set the stack canary up
+	 * for us.  CPU0 already has it initialized but no harm in
+	 * doing it again.  This is a good place for updating it, as
+	 * we wont ever return from this function (so the invalid
+	 * canaries already on the stack wont ever trigger).
 	 */
 	boot_init_stack_canary();
 
diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c
index 0a5df5f..7ec39ab 100644
--- a/arch/x86/kernel/ptrace.c
+++ b/arch/x86/kernel/ptrace.c
@@ -75,10 +75,7 @@ static inline bool invalid_selector(u16 value)
 static unsigned long *pt_regs_access(struct pt_regs *regs, unsigned long regno)
 {
 	BUILD_BUG_ON(offsetof(struct pt_regs, bx) != 0);
-	regno >>= 2;
-	if (regno > FS)
-		--regno;
-	return &regs->bx + regno;
+	return &regs->bx + (regno >> 2);
 }
 
 static u16 get_segment_reg(struct task_struct *task, unsigned long offset)
@@ -90,9 +87,10 @@ static u16 get_segment_reg(struct task_struct *task, unsigned long offset)
 	if (offset != offsetof(struct user_regs_struct, gs))
 		retval = *pt_regs_access(task_pt_regs(task), offset);
 	else {
-		retval = task->thread.gs;
 		if (task == current)
-			savesegment(gs, retval);
+			retval = get_user_gs(task_pt_regs(task));
+		else
+			retval = task_user_gs(task);
 	}
 	return retval;
 }
@@ -126,13 +124,10 @@ static int set_segment_reg(struct task_struct *task,
 		break;
 
 	case offsetof(struct user_regs_struct, gs):
-		task->thread.gs = value;
 		if (task == current)
-			/*
-			 * The user-mode %gs is not affected by
-			 * kernel entry, so we must update the CPU.
-			 */
-			loadsegment(gs, value);
+			set_user_gs(task_pt_regs(task), value);
+		else
+			task_user_gs(task) = value;
 	}
 
 	return 0;
diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c
index ef91747..d992e6c 100644
--- a/arch/x86/kernel/setup_percpu.c
+++ b/arch/x86/kernel/setup_percpu.c
@@ -16,6 +16,7 @@
 #include <asm/proto.h>
 #include <asm/cpumask.h>
 #include <asm/cpu.h>
+#include <asm/stackprotector.h>
 
 #ifdef CONFIG_DEBUG_PER_CPU_MAPS
 # define DBG(x...) printk(KERN_DEBUG x)
@@ -95,6 +96,7 @@ void __init setup_per_cpu_areas(void)
 		per_cpu(this_cpu_off, cpu) = per_cpu_offset(cpu);
 		per_cpu(cpu_number, cpu) = cpu;
 		setup_percpu_segment(cpu);
+		setup_stack_canary_segment(cpu);
 		/*
 		 * Copy data used in early init routines from the
 		 * initial arrays to the per cpu data areas.  These
diff --git a/arch/x86/kernel/signal.c b/arch/x86/kernel/signal.c
index 7fc78b0..7cdcd16 100644
--- a/arch/x86/kernel/signal.c
+++ b/arch/x86/kernel/signal.c
@@ -50,27 +50,23 @@
 # define FIX_EFLAGS	__FIX_EFLAGS
 #endif
 
-#define COPY(x)			{		\
-	get_user_ex(regs->x, &sc->x);		\
-}
+#define COPY(x)			do {			\
+	get_user_ex(regs->x, &sc->x);			\
+} while (0)
 
-#define COPY_SEG(seg)		{			\
-		unsigned short tmp;			\
-		get_user_ex(tmp, &sc->seg);		\
-		regs->seg = tmp;			\
-}
+#define GET_SEG(seg)		({			\
+	unsigned short tmp;				\
+	get_user_ex(tmp, &sc->seg);			\
+	tmp;						\
+})
 
-#define COPY_SEG_CPL3(seg)	{			\
-		unsigned short tmp;			\
-		get_user_ex(tmp, &sc->seg);		\
-		regs->seg = tmp | 3;			\
-}
+#define COPY_SEG(seg)		do {			\
+	regs->seg = GET_SEG(seg);			\
+} while (0)
 
-#define GET_SEG(seg)		{			\
-		unsigned short tmp;			\
-		get_user_ex(tmp, &sc->seg);		\
-		loadsegment(seg, tmp);			\
-}
+#define COPY_SEG_CPL3(seg)	do {			\
+	regs->seg = GET_SEG(seg) | 3;			\
+} while (0)
 
 static int
 restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc,
@@ -86,7 +82,7 @@ restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc,
 	get_user_try {
 
 #ifdef CONFIG_X86_32
-		GET_SEG(gs);
+		set_user_gs(regs, GET_SEG(gs));
 		COPY_SEG(fs);
 		COPY_SEG(es);
 		COPY_SEG(ds);
@@ -138,12 +134,7 @@ setup_sigcontext(struct sigcontext __user *sc, void __user *fpstate,
 	put_user_try {
 
 #ifdef CONFIG_X86_32
-		{
-			unsigned int tmp;
-
-			savesegment(gs, tmp);
-			put_user_ex(tmp, (unsigned int __user *)&sc->gs);
-		}
+		put_user_ex(get_user_gs(regs), (unsigned int __user *)&sc->gs);
 		put_user_ex(regs->fs, (unsigned int __user *)&sc->fs);
 		put_user_ex(regs->es, (unsigned int __user *)&sc->es);
 		put_user_ex(regs->ds, (unsigned int __user *)&sc->ds);
@@ -558,14 +549,9 @@ sys_sigaction(int sig, const struct old_sigaction __user *act,
 #endif /* CONFIG_X86_32 */
 
 #ifdef CONFIG_X86_32
-asmlinkage int sys_sigaltstack(unsigned long bx)
+int sys_sigaltstack(struct pt_regs *regs)
 {
-	/*
-	 * This is needed to make gcc realize it doesn't own the
-	 * "struct pt_regs"
-	 */
-	struct pt_regs *regs = (struct pt_regs *)&bx;
-	const stack_t __user *uss = (const stack_t __user *)bx;
+	const stack_t __user *uss = (const stack_t __user *)regs->bx;
 	stack_t __user *uoss = (stack_t __user *)regs->cx;
 
 	return do_sigaltstack(uss, uoss, regs->sp);
@@ -583,14 +569,12 @@ sys_sigaltstack(const stack_t __user *uss, stack_t __user *uoss,
  * Do a signal return; undo the signal stack.
  */
 #ifdef CONFIG_X86_32
-asmlinkage unsigned long sys_sigreturn(unsigned long __unused)
+unsigned long sys_sigreturn(struct pt_regs *regs)
 {
 	struct sigframe __user *frame;
-	struct pt_regs *regs;
 	unsigned long ax;
 	sigset_t set;
 
-	regs = (struct pt_regs *) &__unused;
 	frame = (struct sigframe __user *)(regs->sp - 8);
 
 	if (!access_ok(VERIFY_READ, frame, sizeof(*frame)))
@@ -617,7 +601,7 @@ badframe:
 }
 #endif /* CONFIG_X86_32 */
 
-static long do_rt_sigreturn(struct pt_regs *regs)
+long sys_rt_sigreturn(struct pt_regs *regs)
 {
 	struct rt_sigframe __user *frame;
 	unsigned long ax;
@@ -648,25 +632,6 @@ badframe:
 	return 0;
 }
 
-#ifdef CONFIG_X86_32
-/*
- * Note: do not pass in pt_regs directly as with tail-call optimization
- * GCC will incorrectly stomp on the caller's frame and corrupt user-space
- * register state:
- */
-asmlinkage int sys_rt_sigreturn(unsigned long __unused)
-{
-	struct pt_regs *regs = (struct pt_regs *)&__unused;
-
-	return do_rt_sigreturn(regs);
-}
-#else /* !CONFIG_X86_32 */
-asmlinkage long sys_rt_sigreturn(struct pt_regs *regs)
-{
-	return do_rt_sigreturn(regs);
-}
-#endif /* CONFIG_X86_32 */
-
 /*
  * OK, we're invoking a handler:
  */
diff --git a/arch/x86/kernel/syscall_table_32.S b/arch/x86/kernel/syscall_table_32.S
index e2e86a0..3bdb648 100644
--- a/arch/x86/kernel/syscall_table_32.S
+++ b/arch/x86/kernel/syscall_table_32.S
@@ -1,7 +1,7 @@
 ENTRY(sys_call_table)
 	.long sys_restart_syscall	/* 0 - old "setup()" system call, used for restarting */
 	.long sys_exit
-	.long sys_fork
+	.long ptregs_fork
 	.long sys_read
 	.long sys_write
 	.long sys_open		/* 5 */
@@ -10,7 +10,7 @@ ENTRY(sys_call_table)
 	.long sys_creat
 	.long sys_link
 	.long sys_unlink	/* 10 */
-	.long sys_execve
+	.long ptregs_execve
 	.long sys_chdir
 	.long sys_time
 	.long sys_mknod
@@ -109,17 +109,17 @@ ENTRY(sys_call_table)
 	.long sys_newlstat
 	.long sys_newfstat
 	.long sys_uname
-	.long sys_iopl		/* 110 */
+	.long ptregs_iopl	/* 110 */
 	.long sys_vhangup
 	.long sys_ni_syscall	/* old "idle" system call */
-	.long sys_vm86old
+	.long ptregs_vm86old
 	.long sys_wait4
 	.long sys_swapoff	/* 115 */
 	.long sys_sysinfo
 	.long sys_ipc
 	.long sys_fsync
-	.long sys_sigreturn
-	.long sys_clone		/* 120 */
+	.long ptregs_sigreturn
+	.long ptregs_clone	/* 120 */
 	.long sys_setdomainname
 	.long sys_newuname
 	.long sys_modify_ldt
@@ -165,14 +165,14 @@ ENTRY(sys_call_table)
 	.long sys_mremap
 	.long sys_setresuid16
 	.long sys_getresuid16	/* 165 */
-	.long sys_vm86
+	.long ptregs_vm86
 	.long sys_ni_syscall	/* Old sys_query_module */
 	.long sys_poll
 	.long sys_nfsservctl
 	.long sys_setresgid16	/* 170 */
 	.long sys_getresgid16
 	.long sys_prctl
-	.long sys_rt_sigreturn
+	.long ptregs_rt_sigreturn
 	.long sys_rt_sigaction
 	.long sys_rt_sigprocmask	/* 175 */
 	.long sys_rt_sigpending
@@ -185,11 +185,11 @@ ENTRY(sys_call_table)
 	.long sys_getcwd
 	.long sys_capget
 	.long sys_capset	/* 185 */
-	.long sys_sigaltstack
+	.long ptregs_sigaltstack
 	.long sys_sendfile
 	.long sys_ni_syscall	/* reserved for streams1 */
 	.long sys_ni_syscall	/* reserved for streams2 */
-	.long sys_vfork		/* 190 */
+	.long ptregs_vfork	/* 190 */
 	.long sys_getrlimit
 	.long sys_mmap2
 	.long sys_truncate64
diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index 0d032d2..bde57f0 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -905,19 +905,20 @@ void math_emulate(struct math_emu_info *info)
 }
 #endif /* CONFIG_MATH_EMULATION */
 
-dotraplinkage void __kprobes do_device_not_available(struct pt_regs regs)
+dotraplinkage void __kprobes
+do_device_not_available(struct pt_regs *regs, long error_code)
 {
 #ifdef CONFIG_X86_32
 	if (read_cr0() & X86_CR0_EM) {
 		struct math_emu_info info = { };
 
-		conditional_sti(&regs);
+		conditional_sti(regs);
 
-		info.regs = &regs;
+		info.regs = regs;
 		math_emulate(&info);
 	} else {
 		math_state_restore(); /* interrupts still off */
-		conditional_sti(&regs);
+		conditional_sti(regs);
 	}
 #else
 	math_state_restore();
diff --git a/arch/x86/kernel/vm86_32.c b/arch/x86/kernel/vm86_32.c
index 4eeb5cf..d7ac84e 100644
--- a/arch/x86/kernel/vm86_32.c
+++ b/arch/x86/kernel/vm86_32.c
@@ -158,7 +158,7 @@ struct pt_regs *save_v86_state(struct kernel_vm86_regs *regs)
 	ret = KVM86->regs32;
 
 	ret->fs = current->thread.saved_fs;
-	loadsegment(gs, current->thread.saved_gs);
+	set_user_gs(ret, current->thread.saved_gs);
 
 	return ret;
 }
@@ -197,9 +197,9 @@ out:
 static int do_vm86_irq_handling(int subfunction, int irqnumber);
 static void do_sys_vm86(struct kernel_vm86_struct *info, struct task_struct *tsk);
 
-asmlinkage int sys_vm86old(struct pt_regs regs)
+int sys_vm86old(struct pt_regs *regs)
 {
-	struct vm86_struct __user *v86 = (struct vm86_struct __user *)regs.bx;
+	struct vm86_struct __user *v86 = (struct vm86_struct __user *)regs->bx;
 	struct kernel_vm86_struct info; /* declare this _on top_,
 					 * this avoids wasting of stack space.
 					 * This remains on the stack until we
@@ -218,7 +218,7 @@ asmlinkage int sys_vm86old(struct pt_regs regs)
 	if (tmp)
 		goto out;
 	memset(&info.vm86plus, 0, (int)&info.regs32 - (int)&info.vm86plus);
-	info.regs32 = &regs;
+	info.regs32 = regs;
 	tsk->thread.vm86_info = v86;
 	do_sys_vm86(&info, tsk);
 	ret = 0;	/* we never return here */
@@ -227,7 +227,7 @@ out:
 }
 
 
-asmlinkage int sys_vm86(struct pt_regs regs)
+int sys_vm86(struct pt_regs *regs)
 {
 	struct kernel_vm86_struct info; /* declare this _on top_,
 					 * this avoids wasting of stack space.
@@ -239,12 +239,12 @@ asmlinkage int sys_vm86(struct pt_regs regs)
 	struct vm86plus_struct __user *v86;
 
 	tsk = current;
-	switch (regs.bx) {
+	switch (regs->bx) {
 	case VM86_REQUEST_IRQ:
 	case VM86_FREE_IRQ:
 	case VM86_GET_IRQ_BITS:
 	case VM86_GET_AND_RESET_IRQ:
-		ret = do_vm86_irq_handling(regs.bx, (int)regs.cx);
+		ret = do_vm86_irq_handling(regs->bx, (int)regs->cx);
 		goto out;
 	case VM86_PLUS_INSTALL_CHECK:
 		/*
@@ -261,14 +261,14 @@ asmlinkage int sys_vm86(struct pt_regs regs)
 	ret = -EPERM;
 	if (tsk->thread.saved_sp0)
 		goto out;
-	v86 = (struct vm86plus_struct __user *)regs.cx;
+	v86 = (struct vm86plus_struct __user *)regs->cx;
 	tmp = copy_vm86_regs_from_user(&info.regs, &v86->regs,
 				       offsetof(struct kernel_vm86_struct, regs32) -
 				       sizeof(info.regs));
 	ret = -EFAULT;
 	if (tmp)
 		goto out;
-	info.regs32 = &regs;
+	info.regs32 = regs;
 	info.vm86plus.is_vm86pus = 1;
 	tsk->thread.vm86_info = (struct vm86_struct __user *)v86;
 	do_sys_vm86(&info, tsk);
@@ -323,7 +323,7 @@ static void do_sys_vm86(struct kernel_vm86_struct *info, struct task_struct *tsk
 	info->regs32->ax = 0;
 	tsk->thread.saved_sp0 = tsk->thread.sp0;
 	tsk->thread.saved_fs = info->regs32->fs;
-	savesegment(gs, tsk->thread.saved_gs);
+	tsk->thread.saved_gs = get_user_gs(info->regs32);
 
 	tss = &per_cpu(init_tss, get_cpu());
 	tsk->thread.sp0 = (unsigned long) &info->VM86_TSS_ESP0;
diff --git a/arch/x86/kernel/vmlinux_64.lds.S b/arch/x86/kernel/vmlinux_64.lds.S
index 07f62d2..087a7f2 100644
--- a/arch/x86/kernel/vmlinux_64.lds.S
+++ b/arch/x86/kernel/vmlinux_64.lds.S
@@ -257,6 +257,14 @@ SECTIONS
   DWARF_DEBUG
 }
 
+ /*
+  * Per-cpu symbols which need to be offset from __per_cpu_load
+  * for the boot processor.
+  */
+#define INIT_PER_CPU(x) init_per_cpu__##x = per_cpu__##x + __per_cpu_load
+INIT_PER_CPU(gdt_page);
+INIT_PER_CPU(irq_stack_union);
+
 /*
  * Build-time check on the image size:
  */
diff --git a/arch/x86/lguest/boot.c b/arch/x86/lguest/boot.c
index 19e33b6..da2e314 100644
--- a/arch/x86/lguest/boot.c
+++ b/arch/x86/lguest/boot.c
@@ -283,7 +283,7 @@ static void lguest_load_tls(struct thread_struct *t, unsigned int cpu)
 	/* There's one problem which normal hardware doesn't have: the Host
 	 * can't handle us removing entries we're currently using.  So we clear
 	 * the GS register here: if it's needed it'll be reloaded anyway. */
-	loadsegment(gs, 0);
+	lazy_load_gs(0);
 	lazy_hcall(LHCALL_LOAD_TLS, __pa(&t->tls_array), cpu, 0);
 }
 
diff --git a/arch/x86/math-emu/get_address.c b/arch/x86/math-emu/get_address.c
index 420b3b6..6ef5e99 100644
--- a/arch/x86/math-emu/get_address.c
+++ b/arch/x86/math-emu/get_address.c
@@ -150,11 +150,9 @@ static long pm_address(u_char FPU_modrm, u_char segment,
 #endif /* PARANOID */
 
 	switch (segment) {
-		/* gs isn't used by the kernel, so it still has its
-		   user-space value. */
 	case PREFIX_GS_ - 1:
-		/* N.B. - movl %seg, mem is a 2 byte write regardless of prefix */
-		savesegment(gs, addr->selector);
+		/* user gs handling can be lazy, use special accessors */
+		addr->selector = get_user_gs(FPU_info->regs);
 		break;
 	default:
 		addr->selector = PM_REG_(segment);
diff --git a/arch/x86/mm/numa_64.c b/arch/x86/mm/numa_64.c
index 08d140f..deb1c1a 100644
--- a/arch/x86/mm/numa_64.c
+++ b/arch/x86/mm/numa_64.c
@@ -702,7 +702,7 @@ void __cpuinit numa_set_node(int cpu, int node)
 	}
 
 #ifdef CONFIG_DEBUG_PER_CPU_MAPS
-	if (cpu >= nr_cpu_ids || !per_cpu_offset(cpu)) {
+	if (cpu >= nr_cpu_ids || !cpu_possible(cpu)) {
 		printk(KERN_ERR "numa_set_node: invalid cpu# (%d)\n", cpu);
 		dump_stack();
 		return;
@@ -790,7 +790,7 @@ int early_cpu_to_node(int cpu)
 	if (early_per_cpu_ptr(x86_cpu_to_node_map))
 		return early_per_cpu_ptr(x86_cpu_to_node_map)[cpu];
 
-	if (!per_cpu_offset(cpu)) {
+	if (!cpu_possible(cpu)) {
 		printk(KERN_WARNING
 			"early_cpu_to_node(%d): no per_cpu area!\n", cpu);
 		dump_stack();
diff --git a/arch/x86/vdso/Makefile b/arch/x86/vdso/Makefile
index 4d6ef0a..16a9020 100644
--- a/arch/x86/vdso/Makefile
+++ b/arch/x86/vdso/Makefile
@@ -38,7 +38,7 @@ $(obj)/%.so: $(obj)/%.so.dbg FORCE
 	$(call if_changed,objcopy)
 
 CFL := $(PROFILING) -mcmodel=small -fPIC -O2 -fasynchronous-unwind-tables -m64 \
-       $(filter -g%,$(KBUILD_CFLAGS))
+       $(filter -g%,$(KBUILD_CFLAGS)) $(call cc-option, -fno-stack-protector)
 
 $(vobjs): KBUILD_CFLAGS += $(CFL)
 
diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index 3723034..95ff6a0 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -323,13 +323,14 @@ static void load_TLS_descriptor(struct thread_struct *t,
 static void xen_load_tls(struct thread_struct *t, unsigned int cpu)
 {
 	/*
-	 * XXX sleazy hack: If we're being called in a lazy-cpu zone,
-	 * it means we're in a context switch, and %gs has just been
-	 * saved.  This means we can zero it out to prevent faults on
-	 * exit from the hypervisor if the next process has no %gs.
-	 * Either way, it has been saved, and the new value will get
-	 * loaded properly.  This will go away as soon as Xen has been
-	 * modified to not save/restore %gs for normal hypercalls.
+	 * XXX sleazy hack: If we're being called in a lazy-cpu zone
+	 * and lazy gs handling is enabled, it means we're in a
+	 * context switch, and %gs has just been saved.  This means we
+	 * can zero it out to prevent faults on exit from the
+	 * hypervisor if the next process has no %gs.  Either way, it
+	 * has been saved, and the new value will get loaded properly.
+	 * This will go away as soon as Xen has been modified to not
+	 * save/restore %gs for normal hypercalls.
 	 *
 	 * On x86_64, this hack is not used for %gs, because gs points
 	 * to KERNEL_GS_BASE (and uses it for PDA references), so we
@@ -341,7 +342,7 @@ static void xen_load_tls(struct thread_struct *t, unsigned int cpu)
 	 */
 	if (paravirt_get_lazy_mode() == PARAVIRT_LAZY_CPU) {
 #ifdef CONFIG_X86_32
-		loadsegment(gs, 0);
+		lazy_load_gs(0);
 #else
 		loadsegment(fs, 0);
 #endif
diff --git a/arch/x86/xen/xen-asm.S b/arch/x86/xen/xen-asm.S
index 4c6f967..79d7362 100644
--- a/arch/x86/xen/xen-asm.S
+++ b/arch/x86/xen/xen-asm.S
@@ -1,14 +1,14 @@
 /*
-	Asm versions of Xen pv-ops, suitable for either direct use or inlining.
-	The inline versions are the same as the direct-use versions, with the
-	pre- and post-amble chopped off.
-
-	This code is encoded for size rather than absolute efficiency,
-	with a view to being able to inline as much as possible.
-
-	We only bother with direct forms (ie, vcpu in percpu data) of
-	the operations here; the indirect forms are better handled in
-	C, since they're generally too large to inline anyway.
+ * Asm versions of Xen pv-ops, suitable for either direct use or
+ * inlining.  The inline versions are the same as the direct-use
+ * versions, with the pre- and post-amble chopped off.
+ *
+ * This code is encoded for size rather than absolute efficiency, with
+ * a view to being able to inline as much as possible.
+ *
+ * We only bother with direct forms (ie, vcpu in percpu data) of the
+ * operations here; the indirect forms are better handled in C, since
+ * they're generally too large to inline anyway.
  */
 
 #include <asm/asm-offsets.h>
@@ -18,17 +18,19 @@
 #include "xen-asm.h"
 
 /*
-	Enable events.  This clears the event mask and tests the pending
-	event status with one and operation.  If there are pending
-	events, then enter the hypervisor to get them handled.
+ * Enable events.  This clears the event mask and tests the pending
+ * event status with one and operation.  If there are pending events,
+ * then enter the hypervisor to get them handled.
  */
 ENTRY(xen_irq_enable_direct)
 	/* Unmask events */
 	movb $0, PER_CPU_VAR(xen_vcpu_info) + XEN_vcpu_info_mask
 
-	/* Preempt here doesn't matter because that will deal with
-	   any pending interrupts.  The pending check may end up being
-	   run on the wrong CPU, but that doesn't hurt. */
+	/*
+	 * Preempt here doesn't matter because that will deal with any
+	 * pending interrupts.  The pending check may end up being run
+	 * on the wrong CPU, but that doesn't hurt.
+	 */
 
 	/* Test for pending */
 	testb $0xff, PER_CPU_VAR(xen_vcpu_info) + XEN_vcpu_info_pending
@@ -43,8 +45,8 @@ ENDPATCH(xen_irq_enable_direct)
 
 
 /*
-	Disabling events is simply a matter of making the event mask
-	non-zero.
+ * Disabling events is simply a matter of making the event mask
+ * non-zero.
  */
 ENTRY(xen_irq_disable_direct)
 	movb $1, PER_CPU_VAR(xen_vcpu_info) + XEN_vcpu_info_mask
@@ -54,18 +56,18 @@ ENDPATCH(xen_irq_disable_direct)
 	RELOC(xen_irq_disable_direct, 0)
 
 /*
-	(xen_)save_fl is used to get the current interrupt enable status.
-	Callers expect the status to be in X86_EFLAGS_IF, and other bits
-	may be set in the return value.  We take advantage of this by
-	making sure that X86_EFLAGS_IF has the right value (and other bits
-	in that byte are 0), but other bits in the return value are
-	undefined.  We need to toggle the state of the bit, because
-	Xen and x86 use opposite senses (mask vs enable).
+ * (xen_)save_fl is used to get the current interrupt enable status.
+ * Callers expect the status to be in X86_EFLAGS_IF, and other bits
+ * may be set in the return value.  We take advantage of this by
+ * making sure that X86_EFLAGS_IF has the right value (and other bits
+ * in that byte are 0), but other bits in the return value are
+ * undefined.  We need to toggle the state of the bit, because Xen and
+ * x86 use opposite senses (mask vs enable).
  */
 ENTRY(xen_save_fl_direct)
 	testb $0xff, PER_CPU_VAR(xen_vcpu_info) + XEN_vcpu_info_mask
 	setz %ah
-	addb %ah,%ah
+	addb %ah, %ah
 ENDPATCH(xen_save_fl_direct)
 	ret
 	ENDPROC(xen_save_fl_direct)
@@ -73,12 +75,11 @@ ENDPATCH(xen_save_fl_direct)
 
 
 /*
-	In principle the caller should be passing us a value return
-	from xen_save_fl_direct, but for robustness sake we test only
-	the X86_EFLAGS_IF flag rather than the whole byte. After
-	setting the interrupt mask state, it checks for unmasked
-	pending events and enters the hypervisor to get them delivered
-	if so.
+ * In principle the caller should be passing us a value return from
+ * xen_save_fl_direct, but for robustness sake we test only the
+ * X86_EFLAGS_IF flag rather than the whole byte. After setting the
+ * interrupt mask state, it checks for unmasked pending events and
+ * enters the hypervisor to get them delivered if so.
  */
 ENTRY(xen_restore_fl_direct)
 #ifdef CONFIG_X86_64
@@ -87,9 +88,11 @@ ENTRY(xen_restore_fl_direct)
 	testb $X86_EFLAGS_IF>>8, %ah
 #endif
 	setz PER_CPU_VAR(xen_vcpu_info) + XEN_vcpu_info_mask
-	/* Preempt here doesn't matter because that will deal with
-	   any pending interrupts.  The pending check may end up being
-	   run on the wrong CPU, but that doesn't hurt. */
+	/*
+	 * Preempt here doesn't matter because that will deal with any
+	 * pending interrupts.  The pending check may end up being run
+	 * on the wrong CPU, but that doesn't hurt.
+	 */
 
 	/* check for unmasked and pending */
 	cmpw $0x0001, PER_CPU_VAR(xen_vcpu_info) + XEN_vcpu_info_pending
@@ -103,8 +106,8 @@ ENDPATCH(xen_restore_fl_direct)
 
 
 /*
-	Force an event check by making a hypercall,
-	but preserve regs before making the call.
+ * Force an event check by making a hypercall, but preserve regs
+ * before making the call.
  */
 check_events:
 #ifdef CONFIG_X86_32
@@ -137,4 +140,3 @@ check_events:
 	pop %rax
 #endif
 	ret
-
diff --git a/arch/x86/xen/xen-asm_32.S b/arch/x86/xen/xen-asm_32.S
index 082d173..88e15de 100644
--- a/arch/x86/xen/xen-asm_32.S
+++ b/arch/x86/xen/xen-asm_32.S
@@ -1,17 +1,16 @@
 /*
-	Asm versions of Xen pv-ops, suitable for either direct use or inlining.
-	The inline versions are the same as the direct-use versions, with the
-	pre- and post-amble chopped off.
-
-	This code is encoded for size rather than absolute efficiency,
-	with a view to being able to inline as much as possible.
-
-	We only bother with direct forms (ie, vcpu in pda) of the operations
-	here; the indirect forms are better handled in C, since they're
-	generally too large to inline anyway.
+ * Asm versions of Xen pv-ops, suitable for either direct use or
+ * inlining.  The inline versions are the same as the direct-use
+ * versions, with the pre- and post-amble chopped off.
+ *
+ * This code is encoded for size rather than absolute efficiency, with
+ * a view to being able to inline as much as possible.
+ *
+ * We only bother with direct forms (ie, vcpu in pda) of the
+ * operations here; the indirect forms are better handled in C, since
+ * they're generally too large to inline anyway.
  */
 
-//#include <asm/asm-offsets.h>
 #include <asm/thread_info.h>
 #include <asm/processor-flags.h>
 #include <asm/segment.h>
@@ -21,8 +20,8 @@
 #include "xen-asm.h"
 
 /*
-	Force an event check by making a hypercall,
-	but preserve regs before making the call.
+ * Force an event check by making a hypercall, but preserve regs
+ * before making the call.
  */
 check_events:
 	push %eax
@@ -35,10 +34,10 @@ check_events:
 	ret
 
 /*
-	We can't use sysexit directly, because we're not running in ring0.
-	But we can easily fake it up using iret.  Assuming xen_sysexit
-	is jumped to with a standard stack frame, we can just strip it
-	back to a standard iret frame and use iret.
+ * We can't use sysexit directly, because we're not running in ring0.
+ * But we can easily fake it up using iret.  Assuming xen_sysexit is
+ * jumped to with a standard stack frame, we can just strip it back to
+ * a standard iret frame and use iret.
  */
 ENTRY(xen_sysexit)
 	movl PT_EAX(%esp), %eax			/* Shouldn't be necessary? */
@@ -49,33 +48,31 @@ ENTRY(xen_sysexit)
 ENDPROC(xen_sysexit)
 
 /*
-	This is run where a normal iret would be run, with the same stack setup:
-	      8: eflags
-	      4: cs
-	esp-> 0: eip
-
-	This attempts to make sure that any pending events are dealt
-	with on return to usermode, but there is a small window in
-	which an event can happen just before entering usermode.  If
-	the nested interrupt ends up setting one of the TIF_WORK_MASK
-	pending work flags, they will not be tested again before
-	returning to usermode. This means that a process can end up
-	with pending work, which will be unprocessed until the process
-	enters and leaves the kernel again, which could be an
-	unbounded amount of time.  This means that a pending signal or
-	reschedule event could be indefinitely delayed.
-
-	The fix is to notice a nested interrupt in the critical
-	window, and if one occurs, then fold the nested interrupt into
-	the current interrupt stack frame, and re-process it
-	iteratively rather than recursively.  This means that it will
-	exit via the normal path, and all pending work will be dealt
-	with appropriately.
-
-	Because the nested interrupt handler needs to deal with the
-	current stack state in whatever form its in, we keep things
-	simple by only using a single register which is pushed/popped
-	on the stack.
+ * This is run where a normal iret would be run, with the same stack setup:
+ *	8: eflags
+ *	4: cs
+ *	esp-> 0: eip
+ *
+ * This attempts to make sure that any pending events are dealt with
+ * on return to usermode, but there is a small window in which an
+ * event can happen just before entering usermode.  If the nested
+ * interrupt ends up setting one of the TIF_WORK_MASK pending work
+ * flags, they will not be tested again before returning to
+ * usermode. This means that a process can end up with pending work,
+ * which will be unprocessed until the process enters and leaves the
+ * kernel again, which could be an unbounded amount of time.  This
+ * means that a pending signal or reschedule event could be
+ * indefinitely delayed.
+ *
+ * The fix is to notice a nested interrupt in the critical window, and
+ * if one occurs, then fold the nested interrupt into the current
+ * interrupt stack frame, and re-process it iteratively rather than
+ * recursively.  This means that it will exit via the normal path, and
+ * all pending work will be dealt with appropriately.
+ *
+ * Because the nested interrupt handler needs to deal with the current
+ * stack state in whatever form its in, we keep things simple by only
+ * using a single register which is pushed/popped on the stack.
  */
 ENTRY(xen_iret)
 	/* test eflags for special cases */
@@ -85,13 +82,15 @@ ENTRY(xen_iret)
 	push %eax
 	ESP_OFFSET=4	# bytes pushed onto stack
 
-	/* Store vcpu_info pointer for easy access.  Do it this
-	   way to avoid having to reload %fs */
+	/*
+	 * Store vcpu_info pointer for easy access.  Do it this way to
+	 * avoid having to reload %fs
+	 */
 #ifdef CONFIG_SMP
 	GET_THREAD_INFO(%eax)
-	movl TI_cpu(%eax),%eax
-	movl __per_cpu_offset(,%eax,4),%eax
-	mov per_cpu__xen_vcpu(%eax),%eax
+	movl TI_cpu(%eax), %eax
+	movl __per_cpu_offset(,%eax,4), %eax
+	mov per_cpu__xen_vcpu(%eax), %eax
 #else
 	movl per_cpu__xen_vcpu, %eax
 #endif
@@ -99,37 +98,46 @@ ENTRY(xen_iret)
 	/* check IF state we're restoring */
 	testb $X86_EFLAGS_IF>>8, 8+1+ESP_OFFSET(%esp)
 
-	/* Maybe enable events.  Once this happens we could get a
-	   recursive event, so the critical region starts immediately
-	   afterwards.  However, if that happens we don't end up
-	   resuming the code, so we don't have to be worried about
-	   being preempted to another CPU. */
+	/*
+	 * Maybe enable events.  Once this happens we could get a
+	 * recursive event, so the critical region starts immediately
+	 * afterwards.  However, if that happens we don't end up
+	 * resuming the code, so we don't have to be worried about
+	 * being preempted to another CPU.
+	 */
 	setz XEN_vcpu_info_mask(%eax)
 xen_iret_start_crit:
 
 	/* check for unmasked and pending */
 	cmpw $0x0001, XEN_vcpu_info_pending(%eax)
 
-	/* If there's something pending, mask events again so we
-	   can jump back into xen_hypervisor_callback */
+	/*
+	 * If there's something pending, mask events again so we can
+	 * jump back into xen_hypervisor_callback
+	 */
 	sete XEN_vcpu_info_mask(%eax)
 
 	popl %eax
 
-	/* From this point on the registers are restored and the stack
-	   updated, so we don't need to worry about it if we're preempted */
+	/*
+	 * From this point on the registers are restored and the stack
+	 * updated, so we don't need to worry about it if we're
+	 * preempted
+	 */
 iret_restore_end:
 
-	/* Jump to hypervisor_callback after fixing up the stack.
-	   Events are masked, so jumping out of the critical
-	   region is OK. */
+	/*
+	 * Jump to hypervisor_callback after fixing up the stack.
+	 * Events are masked, so jumping out of the critical region is
+	 * OK.
+	 */
 	je xen_hypervisor_callback
 
 1:	iret
 xen_iret_end_crit:
-.section __ex_table,"a"
+.section __ex_table, "a"
 	.align 4
-	.long 1b,iret_exc
+	.long 1b, iret_exc
 .previous
 
 hyper_iret:
@@ -139,55 +147,55 @@ hyper_iret:
 	.globl xen_iret_start_crit, xen_iret_end_crit
 
 /*
-   This is called by xen_hypervisor_callback in entry.S when it sees
-   that the EIP at the time of interrupt was between xen_iret_start_crit
-   and xen_iret_end_crit.  We're passed the EIP in %eax so we can do
-   a more refined determination of what to do.
-
-   The stack format at this point is:
-	----------------
-	 ss		: (ss/esp may be present if we came from usermode)
-	 esp		:
-	 eflags		}  outer exception info
-	 cs		}
-	 eip		}
-	---------------- <- edi (copy dest)
-	 eax		:  outer eax if it hasn't been restored
-	----------------
-	 eflags		}  nested exception info
-	 cs		}   (no ss/esp because we're nested
-	 eip		}    from the same ring)
-	 orig_eax	}<- esi (copy src)
-	 - - - - - - - -
-	 fs		}
-	 es		}
-	 ds		}  SAVE_ALL state
-	 eax		}
-	  :		:
-	 ebx		}<- esp
-	----------------
-
-   In order to deliver the nested exception properly, we need to shift
-   everything from the return addr up to the error code so it
-   sits just under the outer exception info.  This means that when we
-   handle the exception, we do it in the context of the outer exception
-   rather than starting a new one.
-
-   The only caveat is that if the outer eax hasn't been
-   restored yet (ie, it's still on stack), we need to insert
-   its value into the SAVE_ALL state before going on, since
-   it's usermode state which we eventually need to restore.
+ * This is called by xen_hypervisor_callback in entry.S when it sees
+ * that the EIP at the time of interrupt was between
+ * xen_iret_start_crit and xen_iret_end_crit.  We're passed the EIP in
+ * %eax so we can do a more refined determination of what to do.
+ *
+ * The stack format at this point is:
+ *	----------------
+ *	 ss		: (ss/esp may be present if we came from usermode)
+ *	 esp		:
+ *	 eflags		}  outer exception info
+ *	 cs		}
+ *	 eip		}
+ *	---------------- <- edi (copy dest)
+ *	 eax		:  outer eax if it hasn't been restored
+ *	----------------
+ *	 eflags		}  nested exception info
+ *	 cs		}   (no ss/esp because we're nested
+ *	 eip		}    from the same ring)
+ *	 orig_eax	}<- esi (copy src)
+ *	 - - - - - - - -
+ *	 fs		}
+ *	 es		}
+ *	 ds		}  SAVE_ALL state
+ *	 eax		}
+ *	  :		:
+ *	 ebx		}<- esp
+ *	----------------
+ *
+ * In order to deliver the nested exception properly, we need to shift
+ * everything from the return addr up to the error code so it sits
+ * just under the outer exception info.  This means that when we
+ * handle the exception, we do it in the context of the outer
+ * exception rather than starting a new one.
+ *
+ * The only caveat is that if the outer eax hasn't been restored yet
+ * (ie, it's still on stack), we need to insert its value into the
+ * SAVE_ALL state before going on, since it's usermode state which we
+ * eventually need to restore.
  */
 ENTRY(xen_iret_crit_fixup)
 	/*
-	   Paranoia: Make sure we're really coming from kernel space.
-	   One could imagine a case where userspace jumps into the
-	   critical range address, but just before the CPU delivers a GP,
-	   it decides to deliver an interrupt instead.  Unlikely?
-	   Definitely.  Easy to avoid?  Yes.  The Intel documents
-	   explicitly say that the reported EIP for a bad jump is the
-	   jump instruction itself, not the destination, but some virtual
-	   environments get this wrong.
+	 * Paranoia: Make sure we're really coming from kernel space.
+	 * One could imagine a case where userspace jumps into the
+	 * critical range address, but just before the CPU delivers a
+	 * GP, it decides to deliver an interrupt instead.  Unlikely?
+	 * Definitely.  Easy to avoid?  Yes.  The Intel documents
+	 * explicitly say that the reported EIP for a bad jump is the
+	 * jump instruction itself, not the destination, but some
+	 * virtual environments get this wrong.
 	 */
 	movl PT_CS(%esp), %ecx
 	andl $SEGMENT_RPL_MASK, %ecx
@@ -197,15 +205,17 @@ ENTRY(xen_iret_crit_fixup)
 	lea PT_ORIG_EAX(%esp), %esi
 	lea PT_EFLAGS(%esp), %edi
 
-	/* If eip is before iret_restore_end then stack
-	   hasn't been restored yet. */
+	/*
+	 * If eip is before iret_restore_end then stack
+	 * hasn't been restored yet.
+	 */
 	cmp $iret_restore_end, %eax
 	jae 1f
 
-	movl 0+4(%edi),%eax		/* copy EAX (just above top of frame) */
+	movl 0+4(%edi), %eax		/* copy EAX (just above top of frame) */
 	movl %eax, PT_EAX(%esp)
 
-	lea ESP_OFFSET(%edi),%edi	/* move dest up over saved regs */
+	lea ESP_OFFSET(%edi), %edi	/* move dest up over saved regs */
 
 	/* set up the copy */
 1:	std
@@ -213,6 +223,6 @@ ENTRY(xen_iret_crit_fixup)
 	rep movsl
 	cld
 
-	lea 4(%edi),%esp		/* point esp to new frame */
+	lea 4(%edi), %esp		/* point esp to new frame */
 2:	jmp xen_do_upcall
 
diff --git a/arch/x86/xen/xen-asm_64.S b/arch/x86/xen/xen-asm_64.S
index d205a28..02f496a 100644
--- a/arch/x86/xen/xen-asm_64.S
+++ b/arch/x86/xen/xen-asm_64.S
@@ -1,14 +1,14 @@
 /*
-	Asm versions of Xen pv-ops, suitable for either direct use or inlining.
-	The inline versions are the same as the direct-use versions, with the
-	pre- and post-amble chopped off.
-
-	This code is encoded for size rather than absolute efficiency,
-	with a view to being able to inline as much as possible.
-
-	We only bother with direct forms (ie, vcpu in pda) of the operations
-	here; the indirect forms are better handled in C, since they're
-	generally too large to inline anyway.
+ * Asm versions of Xen pv-ops, suitable for either direct use or
+ * inlining.  The inline versions are the same as the direct-use
+ * versions, with the pre- and post-amble chopped off.
+ *
+ * This code is encoded for size rather than absolute efficiency, with
+ * a view to being able to inline as much as possible.
+ *
+ * We only bother with direct forms (ie, vcpu in pda) of the
+ * operations here; the indirect forms are better handled in C, since
+ * they're generally too large to inline anyway.
  */
 
 #include <asm/errno.h>
@@ -21,25 +21,25 @@
 #include "xen-asm.h"
 
 ENTRY(xen_adjust_exception_frame)
-	mov 8+0(%rsp),%rcx
-	mov 8+8(%rsp),%r11
+	mov 8+0(%rsp), %rcx
+	mov 8+8(%rsp), %r11
 	ret $16
 
 hypercall_iret = hypercall_page + __HYPERVISOR_iret * 32
 /*
-	Xen64 iret frame:
-
-	ss
-	rsp
-	rflags
-	cs
-	rip		<-- standard iret frame
-
-	flags
-
-	rcx		}
-	r11		}<-- pushed by hypercall page
-rsp ->	rax		}
+ * Xen64 iret frame:
+ *
+ *	ss
+ *	rsp
+ *	rflags
+ *	cs
+ *	rip		<-- standard iret frame
+ *
+ *	flags
+ *
+ *	rcx		}
+ *	r11		}<-- pushed by hypercall page
+ * rsp->rax		}
  */
 ENTRY(xen_iret)
 	pushq $0
@@ -48,8 +48,8 @@ ENDPATCH(xen_iret)
 RELOC(xen_iret, 1b+1)
 
 /*
-	sysexit is not used for 64-bit processes, so it's
-	only ever used to return to 32-bit compat userspace.
+ * sysexit is not used for 64-bit processes, so it's only ever used to
+ * return to 32-bit compat userspace.
  */
 ENTRY(xen_sysexit)
 	pushq $__USER32_DS
@@ -64,10 +64,12 @@ ENDPATCH(xen_sysexit)
 RELOC(xen_sysexit, 1b+1)
 
 ENTRY(xen_sysret64)
-	/* We're already on the usermode stack at this point, but still
-	   with the kernel gs, so we can easily switch back */
+	/*
+	 * We're already on the usermode stack at this point, but
+	 * still with the kernel gs, so we can easily switch back
+	 */
 	movq %rsp, PER_CPU_VAR(old_rsp)
-	movq PER_CPU_VAR(kernel_stack),%rsp
+	movq PER_CPU_VAR(kernel_stack), %rsp
 
 	pushq $__USER_DS
 	pushq PER_CPU_VAR(old_rsp)
@@ -81,8 +83,10 @@ ENDPATCH(xen_sysret64)
 RELOC(xen_sysret64, 1b+1)
 
 ENTRY(xen_sysret32)
-	/* We're already on the usermode stack at this point, but still
-	   with the kernel gs, so we can easily switch back */
+	/*
+	 * We're already on the usermode stack at this point, but
+	 * still with the kernel gs, so we can easily switch back
+	 */
 	movq %rsp, PER_CPU_VAR(old_rsp)
 	movq PER_CPU_VAR(kernel_stack), %rsp
 
@@ -98,28 +102,27 @@ ENDPATCH(xen_sysret32)
 RELOC(xen_sysret32, 1b+1)
 
 /*
-	Xen handles syscall callbacks much like ordinary exceptions,
-	which means we have:
-	 - kernel gs
-	 - kernel rsp
-	 - an iret-like stack frame on the stack (including rcx and r11):
-		ss
-		rsp
-		rflags
-		cs
-		rip
-		r11
-	rsp->	rcx
-
-	In all the entrypoints, we undo all that to make it look
-	like a CPU-generated syscall/sysenter and jump to the normal
-	entrypoint.
+ * Xen handles syscall callbacks much like ordinary exceptions, which
+ * means we have:
+ * - kernel gs
+ * - kernel rsp
+ * - an iret-like stack frame on the stack (including rcx and r11):
+ *	ss
+ *	rsp
+ *	rflags
+ *	cs
+ *	rip
+ *	r11
+ * rsp->rcx
+ *
+ * In all the entrypoints, we undo all that to make it look like a
+ * CPU-generated syscall/sysenter and jump to the normal entrypoint.
  */
 
 .macro undo_xen_syscall
-	mov 0*8(%rsp),%rcx
-	mov 1*8(%rsp),%r11
-	mov 5*8(%rsp),%rsp
+	mov 0*8(%rsp), %rcx
+	mov 1*8(%rsp), %r11
+	mov 5*8(%rsp), %rsp
 .endm
 
 /* Normal 64-bit system call target */
@@ -146,7 +149,7 @@ ENDPROC(xen_sysenter_target)
 
 ENTRY(xen_syscall32_target)
 ENTRY(xen_sysenter_target)
-	lea 16(%rsp), %rsp	/* strip %rcx,%r11 */
+	lea 16(%rsp), %rsp	/* strip %rcx, %r11 */
 	mov $-ENOSYS, %rax
 	pushq $VGCF_in_syscall
 	jmp hypercall_iret