Merge branch 'core/percpu' into x86/core

15 files changed, 418 insertions, 311 deletions

diff --git a/arch/x86/kernel/asm-offsets_32.c b/arch/x86/kernel/asm-offsets_32.c
index ee4df08feee6..fbf2f33e3080 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 cbcdb796d47f..e8f4a386bd9d 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 65efd42454be..e99206831459 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 722464c520cf..2a0aad7718d5 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 a0a2b5ca9b7d..2e648e3a5ea4 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 b12208f4dfee..e41980a373ab 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 1a1ae8edc40c..fec79ad85dc6 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 8eb169e45584..836ef6575f01 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 0a5df5f82fb9..7ec39ab37a2d 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 ef91747bbed5..d992e6cff730 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 7fc78b019815..7cdcd16885ed 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 e2e86a08f31d..3bdb64829b82 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 0d032d2d8a18..bde57f0f1616 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 4eeb5cf9720d..d7ac84e7fc1c 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 07f62d287ff0..087a7f2c639b 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:
  */