Merge branch 'master' into gfs2

47 files changed, 480 insertions, 491 deletions

diff --git a/include/asm-x86_64/Kbuild b/include/asm-x86_64/Kbuild
index 40f2f13fe174..1ee9b07f3fe6 100644
--- a/include/asm-x86_64/Kbuild
+++ b/include/asm-x86_64/Kbuild
@@ -11,6 +11,7 @@ header-y += debugreg.h
 header-y += ldt.h
 header-y += msr.h
 header-y += prctl.h
+header-y += ptrace-abi.h
 header-y += setup.h
 header-y += sigcontext32.h
 header-y += ucontext.h
diff --git a/include/asm-x86_64/acpi.h b/include/asm-x86_64/acpi.h
index 2c95a319c056..ed59aa4c6ff9 100644
--- a/include/asm-x86_64/acpi.h
+++ b/include/asm-x86_64/acpi.h
@@ -155,8 +155,6 @@ extern void acpi_reserve_bootmem(void);
 
 #endif /*CONFIG_ACPI_SLEEP*/
 
-#define boot_cpu_physical_apicid boot_cpu_id
-
 extern int acpi_disabled;
 extern int acpi_pci_disabled;
 
diff --git a/include/asm-x86_64/alternative-asm.i b/include/asm-x86_64/alternative-asm.i
new file mode 100644
index 000000000000..e4041f4fa4dc
--- /dev/null
+++ b/include/asm-x86_64/alternative-asm.i
@@ -0,0 +1,14 @@
+#include <linux/config.h>
+
+#ifdef CONFIG_SMP
+	.macro LOCK_PREFIX
+1:	lock
+	.section .smp_locks,"a"
+	.align 8
+	.quad 1b
+	.previous
+	.endm
+#else
+	.macro LOCK_PREFIX
+	.endm
+#endif
diff --git a/include/asm-x86_64/apic.h b/include/asm-x86_64/apic.h
index 9c96a0a8d1bd..9e66d32330c9 100644
--- a/include/asm-x86_64/apic.h
+++ b/include/asm-x86_64/apic.h
@@ -17,6 +17,8 @@
 
 extern int apic_verbosity;
 extern int apic_runs_main_timer;
+extern int ioapic_force;
+extern int apic_mapped;
 
 /*
  * Define the default level of output to be very little
@@ -29,8 +31,6 @@ extern int apic_runs_main_timer;
 			printk(s, ##a);    \
 	} while (0)
 
-#ifdef CONFIG_X86_LOCAL_APIC
-
 struct pt_regs;
 
 /*
@@ -95,17 +95,12 @@ extern void setup_APIC_extened_lvt(unsigned char lvt_off, unsigned char vector,
 #define K8_APIC_EXT_INT_MSG_EXT 0x7
 #define K8_APIC_EXT_LVT_ENTRY_THRESHOLD    0
 
-extern int disable_timer_pin_1;
-
-
 void smp_send_timer_broadcast_ipi(void);
 void switch_APIC_timer_to_ipi(void *cpumask);
 void switch_ipi_to_APIC_timer(void *cpumask);
 
 #define ARCH_APICTIMER_STOPS_ON_C3	1
 
-#endif /* CONFIG_X86_LOCAL_APIC */
-
 extern unsigned boot_cpu_id;
 
 #endif /* __ASM_APIC_H */
diff --git a/include/asm-x86_64/bitops.h b/include/asm-x86_64/bitops.h
index f7ba57b1cc08..5b535eaf5309 100644
--- a/include/asm-x86_64/bitops.h
+++ b/include/asm-x86_64/bitops.h
@@ -399,6 +399,8 @@ static __inline__ int fls(int x)
 	return r+1;
 }
 
+#define ARCH_HAS_FAST_MULTIPLIER 1
+
 #include <asm-generic/bitops/hweight.h>
 
 #endif /* __KERNEL__ */
diff --git a/include/asm-x86_64/calgary.h b/include/asm-x86_64/calgary.h
index 4e3919524240..6b93f5a3a5c8 100644
--- a/include/asm-x86_64/calgary.h
+++ b/include/asm-x86_64/calgary.h
@@ -24,7 +24,6 @@
 #ifndef _ASM_X86_64_CALGARY_H
 #define _ASM_X86_64_CALGARY_H
 
-#include <linux/config.h>
 #include <linux/spinlock.h>
 #include <linux/device.h>
 #include <linux/dma-mapping.h>
@@ -34,12 +33,12 @@ struct iommu_table {
 	unsigned long  it_base;      /* mapped address of tce table */
 	unsigned long  it_hint;      /* Hint for next alloc */
 	unsigned long *it_map;       /* A simple allocation bitmap for now */
+	void __iomem  *bbar;         /* Bridge BAR */
+	u64	       tar_val;      /* Table Address Register */
+	struct timer_list watchdog_timer;
 	spinlock_t     it_lock;      /* Protects it_map */
 	unsigned int   it_size;      /* Size of iommu table in entries */
 	unsigned char  it_busno;     /* Bus number this table belongs to */
-	void __iomem  *bbar;
-	u64	       tar_val;
-	struct timer_list watchdog_timer;
 };
 
 #define TCE_TABLE_SIZE_UNSPECIFIED	~0
diff --git a/include/asm-x86_64/dwarf2.h b/include/asm-x86_64/dwarf2.h
index 0744db777676..eedc08526b0b 100644
--- a/include/asm-x86_64/dwarf2.h
+++ b/include/asm-x86_64/dwarf2.h
@@ -13,7 +13,7 @@
    away for older version. 
  */
 
-#ifdef CONFIG_UNWIND_INFO
+#ifdef CONFIG_AS_CFI
 
 #define CFI_STARTPROC .cfi_startproc
 #define CFI_ENDPROC .cfi_endproc
@@ -28,6 +28,11 @@
 #define CFI_REMEMBER_STATE .cfi_remember_state
 #define CFI_RESTORE_STATE .cfi_restore_state
 #define CFI_UNDEFINED .cfi_undefined
+#ifdef CONFIG_AS_CFI_SIGNAL_FRAME
+#define CFI_SIGNAL_FRAME .cfi_signal_frame
+#else
+#define CFI_SIGNAL_FRAME
+#endif
 
 #else
 
@@ -45,6 +50,7 @@
 #define CFI_REMEMBER_STATE	#
 #define CFI_RESTORE_STATE	#
 #define CFI_UNDEFINED	#
+#define CFI_SIGNAL_FRAME	#
 
 #endif
 
diff --git a/include/asm-x86_64/e820.h b/include/asm-x86_64/e820.h
index 670a3388e70a..fa2086774105 100644
--- a/include/asm-x86_64/e820.h
+++ b/include/asm-x86_64/e820.h
@@ -19,13 +19,9 @@
 
 #define E820_RAM	1
 #define E820_RESERVED	2
-#define E820_ACPI	3 /* usable as RAM once ACPI tables have been read */
+#define E820_ACPI	3
 #define E820_NVS	4
 
-#define HIGH_MEMORY	(1024*1024)
-
-#define LOWMEMSIZE()	(0x9f000)
-
 #ifndef __ASSEMBLY__
 struct e820entry {
 	u64 addr;	/* start of memory segment */
@@ -46,17 +42,16 @@ extern void setup_memory_region(void);
 extern void contig_e820_setup(void); 
 extern unsigned long e820_end_of_ram(void);
 extern void e820_reserve_resources(void);
+extern void e820_mark_nosave_regions(void);
 extern void e820_print_map(char *who);
 extern int e820_any_mapped(unsigned long start, unsigned long end, unsigned type);
 extern int e820_all_mapped(unsigned long start, unsigned long end, unsigned type);
 
-extern void e820_bootmem_free(pg_data_t *pgdat, unsigned long start,unsigned long end);
 extern void e820_setup_gap(void);
-extern unsigned long e820_hole_size(unsigned long start_pfn,
-				    unsigned long end_pfn);
+extern void e820_register_active_regions(int nid,
+				unsigned long start_pfn, unsigned long end_pfn);
 
-extern void __init parse_memopt(char *p, char **end);
-extern void __init parse_memmapopt(char *p, char **end);
+extern void finish_e820_parsing(void);
 
 extern struct e820map e820;
 
diff --git a/include/asm-x86_64/fixmap.h b/include/asm-x86_64/fixmap.h
index 0b4ffbd1a125..1b620db5b9e3 100644
--- a/include/asm-x86_64/fixmap.h
+++ b/include/asm-x86_64/fixmap.h
@@ -37,13 +37,9 @@ enum fixed_addresses {
 	VSYSCALL_FIRST_PAGE = VSYSCALL_LAST_PAGE + ((VSYSCALL_END-VSYSCALL_START) >> PAGE_SHIFT) - 1,
 	VSYSCALL_HPET,
 	FIX_HPET_BASE,
-#ifdef CONFIG_X86_LOCAL_APIC
 	FIX_APIC_BASE,	/* local (CPU) APIC) -- required for SMP or not */
-#endif
-#ifdef CONFIG_X86_IO_APIC
 	FIX_IO_APIC_BASE_0,
 	FIX_IO_APIC_BASE_END = FIX_IO_APIC_BASE_0 + MAX_IO_APICS-1,
-#endif
 	__end_of_fixed_addresses
 };
 
diff --git a/include/asm-x86_64/genapic.h b/include/asm-x86_64/genapic.h
index 50b38e7c58e4..81e714665344 100644
--- a/include/asm-x86_64/genapic.h
+++ b/include/asm-x86_64/genapic.h
@@ -16,7 +16,6 @@ struct genapic {
 	char *name;
 	u32 int_delivery_mode;
 	u32 int_dest_mode;
-	u32 int_delivery_dest;	/* for quick IPIs */
 	int (*apic_id_registered)(void);
 	cpumask_t (*target_cpus)(void);
 	void (*init_apic_ldr)(void);
diff --git a/include/asm-x86_64/i387.h b/include/asm-x86_64/i387.h
index cba8a3b0cded..0217b74cc9fc 100644
--- a/include/asm-x86_64/i387.h
+++ b/include/asm-x86_64/i387.h
@@ -24,6 +24,7 @@ extern unsigned int mxcsr_feature_mask;
 extern void mxcsr_feature_mask_init(void);
 extern void init_fpu(struct task_struct *child);
 extern int save_i387(struct _fpstate __user *buf);
+extern asmlinkage void math_state_restore(void);
 
 /*
  * FPU lazy state save handling...
@@ -31,7 +32,9 @@ extern int save_i387(struct _fpstate __user *buf);
 
 #define unlazy_fpu(tsk) do { \
 	if (task_thread_info(tsk)->status & TS_USEDFPU) \
-		save_init_fpu(tsk); \
+		save_init_fpu(tsk); 			\
+	else						\
+		tsk->fpu_counter = 0;			\
 } while (0)
 
 /* Ignore delayed exceptions from user space */
@@ -134,8 +137,8 @@ static inline int save_i387_checking(struct i387_fxsave_struct __user *fx)
 #else
 		     : [fx] "cdaSDb" (fx), "0" (0));
 #endif
-	if (unlikely(err))
-		__clear_user(fx, sizeof(struct i387_fxsave_struct));
+	if (unlikely(err) && __clear_user(fx, sizeof(struct i387_fxsave_struct)))
+		err = -EFAULT;
 	/* No need to clear here because the caller clears USED_MATH */
 	return err;
 } 
diff --git a/include/asm-x86_64/intel_arch_perfmon.h b/include/asm-x86_64/intel_arch_perfmon.h
index 59c396431569..8633331420ec 100644
--- a/include/asm-x86_64/intel_arch_perfmon.h
+++ b/include/asm-x86_64/intel_arch_perfmon.h
@@ -14,6 +14,18 @@
 
 #define ARCH_PERFMON_UNHALTED_CORE_CYCLES_SEL	(0x3c)
 #define ARCH_PERFMON_UNHALTED_CORE_CYCLES_UMASK	(0x00 << 8)
-#define ARCH_PERFMON_UNHALTED_CORE_CYCLES_PRESENT (1 << 0)
+#define ARCH_PERFMON_UNHALTED_CORE_CYCLES_INDEX (0)
+#define ARCH_PERFMON_UNHALTED_CORE_CYCLES_PRESENT \
+				(1 << (ARCH_PERFMON_UNHALTED_CORE_CYCLES_INDEX))
+
+union cpuid10_eax {
+	struct {
+		unsigned int version_id:8;
+		unsigned int num_counters:8;
+		unsigned int bit_width:8;
+		unsigned int mask_length:8;
+	} split;
+	unsigned int full;
+};
 
 #endif	/* X86_64_INTEL_ARCH_PERFMON_H */
diff --git a/include/asm-x86_64/io_apic.h b/include/asm-x86_64/io_apic.h
index fb7a0909a174..5d1b5c68e36e 100644
--- a/include/asm-x86_64/io_apic.h
+++ b/include/asm-x86_64/io_apic.h
@@ -10,8 +10,6 @@
  * Copyright (C) 1997, 1998, 1999, 2000 Ingo Molnar
  */
 
-#ifdef CONFIG_X86_IO_APIC
-
 #ifdef CONFIG_PCI_MSI
 static inline int use_pci_vector(void)	{return 1;}
 static inline void disable_edge_ioapic_vector(unsigned int vector) { }
@@ -209,10 +207,6 @@ extern int timer_uses_ioapic_pin_0;
 
 extern int sis_apic_bug; /* dummy */ 
 
-#else  /* !CONFIG_X86_IO_APIC */
-#define io_apic_assign_pci_irqs 0
-#endif
-
 extern int assign_irq_vector(int irq);
 
 void enable_NMI_through_LVT0 (void * dummy);
diff --git a/include/asm-x86_64/irq.h b/include/asm-x86_64/irq.h
index 9db5a1b4f7b1..43469d8ab71a 100644
--- a/include/asm-x86_64/irq.h
+++ b/include/asm-x86_64/irq.h
@@ -44,9 +44,7 @@ static __inline__ int irq_canonicalize(int irq)
 	return ((irq == 2) ? 9 : irq);
 }
 
-#ifdef CONFIG_X86_LOCAL_APIC
 #define ARCH_HAS_NMI_WATCHDOG		/* See include/linux/nmi.h */
-#endif
 
 #ifdef CONFIG_HOTPLUG_CPU
 #include <linux/cpumask.h>
diff --git a/include/asm-x86_64/kexec.h b/include/asm-x86_64/kexec.h
index c564bae03433..5fab957e1091 100644
--- a/include/asm-x86_64/kexec.h
+++ b/include/asm-x86_64/kexec.h
@@ -1,6 +1,27 @@
 #ifndef _X86_64_KEXEC_H
 #define _X86_64_KEXEC_H
 
+#define PA_CONTROL_PAGE  0
+#define VA_CONTROL_PAGE  1
+#define PA_PGD           2
+#define VA_PGD           3
+#define PA_PUD_0         4
+#define VA_PUD_0         5
+#define PA_PMD_0         6
+#define VA_PMD_0         7
+#define PA_PTE_0         8
+#define VA_PTE_0         9
+#define PA_PUD_1         10
+#define VA_PUD_1         11
+#define PA_PMD_1         12
+#define VA_PMD_1         13
+#define PA_PTE_1         14
+#define VA_PTE_1         15
+#define PA_TABLE_PAGE    16
+#define PAGES_NR         17
+
+#ifndef __ASSEMBLY__
+
 #include <linux/string.h>
 
 #include <asm/page.h>
@@ -64,4 +85,12 @@ static inline void crash_setup_regs(struct pt_regs *newregs,
 		newregs->rip = (unsigned long)current_text_addr();
 	}
 }
+
+NORET_TYPE void
+relocate_kernel(unsigned long indirection_page,
+		unsigned long page_list,
+		unsigned long start_address) ATTRIB_NORET;
+
+#endif /* __ASSEMBLY__ */
+
 #endif /* _X86_64_KEXEC_H */
diff --git a/include/asm-x86_64/libata-portmap.h b/include/asm-x86_64/libata-portmap.h
deleted file mode 100644
index 75484ef0c743..000000000000
--- a/include/asm-x86_64/libata-portmap.h
+++ /dev/null
@@ -1 +0,0 @@
-#include <asm-generic/libata-portmap.h>
diff --git a/include/asm-x86_64/linkage.h b/include/asm-x86_64/linkage.h
index 291c2d01c44f..b5f39d0189ce 100644
--- a/include/asm-x86_64/linkage.h
+++ b/include/asm-x86_64/linkage.h
@@ -1,6 +1,6 @@
 #ifndef __ASM_LINKAGE_H
 #define __ASM_LINKAGE_H
 
-/* Nothing to see here... */
+#define __ALIGN .p2align 4,,15
 
 #endif
diff --git a/include/asm-x86_64/mach_apic.h b/include/asm-x86_64/mach_apic.h
index 0acea44c9377..d33422450c00 100644
--- a/include/asm-x86_64/mach_apic.h
+++ b/include/asm-x86_64/mach_apic.h
@@ -16,7 +16,6 @@
 
 #define INT_DELIVERY_MODE (genapic->int_delivery_mode)
 #define INT_DEST_MODE (genapic->int_dest_mode)
-#define INT_DELIVERY_DEST (genapic->int_delivery_dest)
 #define TARGET_CPUS	  (genapic->target_cpus())
 #define apic_id_registered (genapic->apic_id_registered)
 #define init_apic_ldr (genapic->init_apic_ldr)
diff --git a/include/asm-x86_64/mce.h b/include/asm-x86_64/mce.h
index d13687dfd691..5a11146d6d9c 100644
--- a/include/asm-x86_64/mce.h
+++ b/include/asm-x86_64/mce.h
@@ -99,6 +99,8 @@ static inline void mce_amd_feature_init(struct cpuinfo_x86 *c)
 }
 #endif
 
+void mce_log_therm_throt_event(unsigned int cpu, __u64 status);
+
 extern atomic_t mce_entry;
 
 #endif
diff --git a/include/asm-x86_64/mmx.h b/include/asm-x86_64/mmx.h
deleted file mode 100644
index 46b71da99869..000000000000
--- a/include/asm-x86_64/mmx.h
+++ /dev/null
@@ -1,14 +0,0 @@
-#ifndef _ASM_MMX_H
-#define _ASM_MMX_H
-
-/*
- *	MMX 3Dnow! helper operations
- */
-
-#include <linux/types.h>
- 
-extern void *_mmx_memcpy(void *to, const void *from, size_t size);
-extern void mmx_clear_page(void *page);
-extern void mmx_copy_page(void *to, void *from);
-
-#endif
diff --git a/include/asm-x86_64/mpspec.h b/include/asm-x86_64/mpspec.h
index 14fc3ddd9031..017fddb61dc5 100644
--- a/include/asm-x86_64/mpspec.h
+++ b/include/asm-x86_64/mpspec.h
@@ -159,13 +159,7 @@ struct mpc_config_lintsrc
 #define MAX_MP_BUSSES 256
 /* Each PCI slot may be a combo card with its own bus.  4 IRQ pins per slot. */
 #define MAX_IRQ_SOURCES (MAX_MP_BUSSES * 4)
-enum mp_bustype {
-	MP_BUS_ISA = 1,
-	MP_BUS_EISA,
-	MP_BUS_PCI,
-	MP_BUS_MCA
-};
-extern unsigned char mp_bus_id_to_type [MAX_MP_BUSSES];
+extern DECLARE_BITMAP(mp_bus_not_pci, MAX_MP_BUSSES);
 extern int mp_bus_id_to_pci_bus [MAX_MP_BUSSES];
 
 extern unsigned int boot_cpu_physical_apicid;
@@ -178,18 +172,15 @@ extern int mp_irq_entries;
 extern struct mpc_config_intsrc mp_irqs [MAX_IRQ_SOURCES];
 extern int mpc_default_type;
 extern unsigned long mp_lapic_addr;
-extern int pic_mode;
 
 #ifdef CONFIG_ACPI
 extern void mp_register_lapic (u8 id, u8 enabled);
 extern void mp_register_lapic_address (u64 address);
 
-#ifdef CONFIG_X86_IO_APIC
 extern void mp_register_ioapic (u8 id, u32 address, u32 gsi_base);
 extern void mp_override_legacy_irq (u8 bus_irq, u8 polarity, u8 trigger, u32 gsi);
 extern void mp_config_acpi_legacy_irqs (void);
 extern int mp_register_gsi (u32 gsi, int triggering, int polarity);
-#endif /*CONFIG_X86_IO_APIC*/
 #endif
 
 extern int using_apic_timer;
diff --git a/include/asm-x86_64/msr.h b/include/asm-x86_64/msr.h
index 10f8b51cec8b..37e194169fac 100644
--- a/include/asm-x86_64/msr.h
+++ b/include/asm-x86_64/msr.h
@@ -66,14 +66,25 @@
 #define rdtscl(low) \
      __asm__ __volatile__ ("rdtsc" : "=a" (low) : : "edx")
 
+#define rdtscp(low,high,aux) \
+     asm volatile (".byte 0x0f,0x01,0xf9" : "=a" (low), "=d" (high), "=c" (aux))
+
 #define rdtscll(val) do { \
      unsigned int __a,__d; \
      asm volatile("rdtsc" : "=a" (__a), "=d" (__d)); \
      (val) = ((unsigned long)__a) | (((unsigned long)__d)<<32); \
 } while(0)
 
+#define rdtscpll(val, aux) do { \
+     unsigned long __a, __d; \
+     asm volatile (".byte 0x0f,0x01,0xf9" : "=a" (__a), "=d" (__d), "=c" (aux)); \
+     (val) = (__d << 32) | __a; \
+} while (0)
+
 #define write_tsc(val1,val2) wrmsr(0x10, val1, val2)
 
+#define write_rdtscp_aux(val) wrmsr(0xc0000103, val, 0)
+
 #define rdpmc(counter,low,high) \
      __asm__ __volatile__("rdpmc" \
 			  : "=a" (low), "=d" (high) \
diff --git a/include/asm-x86_64/mutex.h b/include/asm-x86_64/mutex.h
index 06fab6de2a88..16396b1de3e4 100644
--- a/include/asm-x86_64/mutex.h
+++ b/include/asm-x86_64/mutex.h
@@ -25,13 +25,9 @@ do {									\
 									\
 	__asm__ __volatile__(						\
 		LOCK_PREFIX "   decl (%%rdi)	\n"			\
-			"   js 2f		\n"			\
-			"1:			\n"			\
-									\
-		LOCK_SECTION_START("")					\
-			"2: call "#fail_fn"	\n"			\
-			"   jmp 1b		\n"			\
-		LOCK_SECTION_END					\
+			"   jns 1f		\n"			\
+			"   call "#fail_fn"	\n"			\
+			"1:"						\
 									\
 		:"=D" (dummy)						\
 		: "D" (v)						\
@@ -75,13 +71,9 @@ do {									\
 									\
 	__asm__ __volatile__(						\
 		LOCK_PREFIX "   incl (%%rdi)	\n"			\
-			"   jle 2f		\n"			\
-			"1:			\n"			\
-									\
-		LOCK_SECTION_START("")					\
-			"2: call "#fail_fn"	\n"			\
-			"   jmp 1b		\n"			\
-		LOCK_SECTION_END					\
+			"   jg 1f		\n"			\
+			"   call "#fail_fn"	\n"			\
+			"1:			  "			\
 									\
 		:"=D" (dummy)						\
 		: "D" (v)						\
diff --git a/include/asm-x86_64/nmi.h b/include/asm-x86_64/nmi.h
index efb45c894d76..cbf2669bca71 100644
--- a/include/asm-x86_64/nmi.h
+++ b/include/asm-x86_64/nmi.h
@@ -7,24 +7,13 @@
 #include <linux/pm.h>
 #include <asm/io.h>
  
-struct pt_regs;
-
-typedef int (*nmi_callback_t)(struct pt_regs * regs, int cpu);
-
-/**
- * set_nmi_callback
- *
- * Set a handler for an NMI. Only one handler may be
- * set. Return 1 if the NMI was handled.
- */
-void set_nmi_callback(nmi_callback_t callback);
-
 /**
- * unset_nmi_callback
+ * do_nmi_callback
  *
- * Remove the handler previously set.
+ * Check to see if a callback exists and execute it.  Return 1
+ * if the handler exists and was handled successfully.
  */
-void unset_nmi_callback(void);
+int do_nmi_callback(struct pt_regs *regs, int cpu);
 
 #ifdef CONFIG_PM
  
@@ -48,25 +37,32 @@ static inline void unset_nmi_pm_callback(struct pm_dev * dev)
 #endif /* CONFIG_PM */
  
 extern void default_do_nmi(struct pt_regs *);
-extern void die_nmi(char *str, struct pt_regs *regs);
+extern void die_nmi(char *str, struct pt_regs *regs, int do_panic);
 
 #define get_nmi_reason() inb(0x61)
 
 extern int panic_on_timeout;
 extern int unknown_nmi_panic;
+extern int nmi_watchdog_enabled;
 
 extern int check_nmi_watchdog(void);
- 
-extern void setup_apic_nmi_watchdog (void);
-extern int reserve_lapic_nmi(void);
-extern void release_lapic_nmi(void);
+extern int avail_to_resrv_perfctr_nmi_bit(unsigned int);
+extern int avail_to_resrv_perfctr_nmi(unsigned int);
+extern int reserve_perfctr_nmi(unsigned int);
+extern void release_perfctr_nmi(unsigned int);
+extern int reserve_evntsel_nmi(unsigned int);
+extern void release_evntsel_nmi(unsigned int);
+
+extern void setup_apic_nmi_watchdog (void *);
+extern void stop_apic_nmi_watchdog (void *);
 extern void disable_timer_nmi_watchdog(void);
 extern void enable_timer_nmi_watchdog(void);
-extern void nmi_watchdog_tick (struct pt_regs * regs, unsigned reason);
+extern int nmi_watchdog_tick (struct pt_regs * regs, unsigned reason);
 
 extern void nmi_watchdog_default(void);
 extern int setup_nmi_watchdog(char *);
 
+extern atomic_t nmi_active;
 extern unsigned int nmi_watchdog;
 #define NMI_DEFAULT	-1
 #define NMI_NONE	0
diff --git a/include/asm-x86_64/pci-direct.h b/include/asm-x86_64/pci-direct.h
index 036b6ca5b53b..eba9cb471df3 100644
--- a/include/asm-x86_64/pci-direct.h
+++ b/include/asm-x86_64/pci-direct.h
@@ -2,47 +2,15 @@
 #define ASM_PCI_DIRECT_H 1
 
 #include <linux/types.h>
-#include <asm/io.h>
 
 /* Direct PCI access. This is used for PCI accesses in early boot before
    the PCI subsystem works. */ 
 
-#define PDprintk(x...)
+extern u32 read_pci_config(u8 bus, u8 slot, u8 func, u8 offset);
+extern u8 read_pci_config_byte(u8 bus, u8 slot, u8 func, u8 offset);
+extern u16 read_pci_config_16(u8 bus, u8 slot, u8 func, u8 offset);
+extern void write_pci_config(u8 bus, u8 slot, u8 func, u8 offset, u32 val);
 
-static inline u32 read_pci_config(u8 bus, u8 slot, u8 func, u8 offset)
-{
-	u32 v; 
-	outl(0x80000000 | (bus<<16) | (slot<<11) | (func<<8) | offset, 0xcf8);
-	v = inl(0xcfc); 
-	if (v != 0xffffffff)
-		PDprintk("%x reading 4 from %x: %x\n", slot, offset, v);
-	return v;
-}
-
-static inline u8 read_pci_config_byte(u8 bus, u8 slot, u8 func, u8 offset)
-{
-	u8 v; 
-	outl(0x80000000 | (bus<<16) | (slot<<11) | (func<<8) | offset, 0xcf8);
-	v = inb(0xcfc + (offset&3)); 
-	PDprintk("%x reading 1 from %x: %x\n", slot, offset, v);
-	return v;
-}
-
-static inline u16 read_pci_config_16(u8 bus, u8 slot, u8 func, u8 offset)
-{
-	u16 v; 
-	outl(0x80000000 | (bus<<16) | (slot<<11) | (func<<8) | offset, 0xcf8);
-	v = inw(0xcfc + (offset&2)); 
-	PDprintk("%x reading 2 from %x: %x\n", slot, offset, v);
-	return v;
-}
-
-static inline void write_pci_config(u8 bus, u8 slot, u8 func, u8 offset,
-				    u32 val)
-{
-	PDprintk("%x writing to %x: %x\n", slot, offset, val); 
-	outl(0x80000000 | (bus<<16) | (slot<<11) | (func<<8) | offset, 0xcf8);
-	outl(val, 0xcfc); 
-}
+extern int early_pci_allowed(void);
 
 #endif
diff --git a/include/asm-x86_64/pda.h b/include/asm-x86_64/pda.h
index b47c3df9ed1d..14996d962bac 100644
--- a/include/asm-x86_64/pda.h
+++ b/include/asm-x86_64/pda.h
@@ -9,20 +9,24 @@
 
 /* Per processor datastructure. %gs points to it while the kernel runs */ 
 struct x8664_pda {
-	struct task_struct *pcurrent;	/* Current process */
-	unsigned long data_offset;	/* Per cpu data offset from linker address */
-	unsigned long kernelstack;  /* top of kernel stack for current */ 
-	unsigned long oldrsp; 	    /* user rsp for system call */
-#if DEBUG_STKSZ > EXCEPTION_STKSZ
-	unsigned long debugstack;   /* #DB/#BP stack. */
+	struct task_struct *pcurrent;	/* 0  Current process */
+	unsigned long data_offset;	/* 8 Per cpu data offset from linker
+					   address */
+	unsigned long kernelstack;  /* 16 top of kernel stack for current */
+	unsigned long oldrsp; 	    /* 24 user rsp for system call */
+        int irqcount;		    /* 32 Irq nesting counter. Starts with -1 */
+	int cpunumber;		    /* 36 Logical CPU number */
+#ifdef CONFIG_CC_STACKPROTECTOR
+	unsigned long stack_canary;	/* 40 stack canary value */
+					/* gcc-ABI: this canary MUST be at
+					   offset 40!!! */
 #endif
-        int irqcount;		    /* Irq nesting counter. Starts with -1 */  	
-	int cpunumber;		    /* Logical CPU number */
-	char *irqstackptr;	/* top of irqstack */
+	char *irqstackptr;
 	int nodenumber;		    /* number of current node */
 	unsigned int __softirq_pending;
 	unsigned int __nmi_count;	/* number of NMI on this CPUs */
-	int mmu_state;     
+	short mmu_state;
+	short isidle;
 	struct mm_struct *active_mm;
 	unsigned apic_timer_irqs;
 } ____cacheline_aligned_in_smp;
@@ -36,44 +40,69 @@ extern struct x8664_pda boot_cpu_pda[];
  * There is no fast way to get the base address of the PDA, all the accesses
  * have to mention %fs/%gs.  So it needs to be done this Torvaldian way.
  */ 
-#define sizeof_field(type,field)  (sizeof(((type *)0)->field))
-#define typeof_field(type,field)  typeof(((type *)0)->field)
+extern void __bad_pda_field(void) __attribute__((noreturn));
 
-extern void __bad_pda_field(void);
+/*
+ * proxy_pda doesn't actually exist, but tell gcc it is accessed for
+ * all PDA accesses so it gets read/write dependencies right.
+ */
+extern struct x8664_pda _proxy_pda;
 
 #define pda_offset(field) offsetof(struct x8664_pda, field)
 
-#define pda_to_op(op,field,val) do { \
-	typedef typeof_field(struct x8664_pda, field) T__; \
-       switch (sizeof_field(struct x8664_pda, field)) { 		\
-case 2: \
-asm volatile(op "w %0,%%gs:%P1"::"ri" ((T__)val),"i"(pda_offset(field)):"memory"); break; \
-case 4: \
-asm volatile(op "l %0,%%gs:%P1"::"ri" ((T__)val),"i"(pda_offset(field)):"memory"); break; \
-case 8: \
-asm volatile(op "q %0,%%gs:%P1"::"ri" ((T__)val),"i"(pda_offset(field)):"memory"); break; \
-       default: __bad_pda_field(); 					\
-       } \
+#define pda_to_op(op,field,val) do {		\
+	typedef typeof(_proxy_pda.field) T__;	\
+	if (0) { T__ tmp__; tmp__ = (val); }	/* type checking */ \
+	switch (sizeof(_proxy_pda.field)) {	\
+	case 2:					\
+		asm(op "w %1,%%gs:%c2" : 	\
+		    "+m" (_proxy_pda.field) :	\
+		    "ri" ((T__)val),		\
+		    "i"(pda_offset(field))); 	\
+ 		break;				\
+	case 4:					\
+		asm(op "l %1,%%gs:%c2" : 	\
+		    "+m" (_proxy_pda.field) :	\
+		    "ri" ((T__)val),		\
+		    "i" (pda_offset(field))); 	\
+		break;				\
+	case 8:					\
+		asm(op "q %1,%%gs:%c2": 	\
+		    "+m" (_proxy_pda.field) :	\
+		    "ri" ((T__)val),		\
+		    "i"(pda_offset(field))); 	\
+		break;				\
+       default: 				\
+		__bad_pda_field();		\
+       }					\
        } while (0)
 
-/* 
- * AK: PDA read accesses should be neither volatile nor have an memory clobber.
- * Unfortunately removing them causes all hell to break lose currently.
- */
-#define pda_from_op(op,field) ({ \
-       typeof_field(struct x8664_pda, field) ret__; \
-       switch (sizeof_field(struct x8664_pda, field)) { 		\
-case 2: \
-asm volatile(op "w %%gs:%P1,%0":"=r" (ret__):"i"(pda_offset(field)):"memory"); break;\
-case 4: \
-asm volatile(op "l %%gs:%P1,%0":"=r" (ret__):"i"(pda_offset(field)):"memory"); break;\
-case 8: \
-asm volatile(op "q %%gs:%P1,%0":"=r" (ret__):"i"(pda_offset(field)):"memory"); break;\
-       default: __bad_pda_field(); 					\
-       } \
+#define pda_from_op(op,field) ({		\
+	typeof(_proxy_pda.field) ret__;		\
+	switch (sizeof(_proxy_pda.field)) {	\
+       	case 2:					\
+		asm(op "w %%gs:%c1,%0" : 	\
+		    "=r" (ret__) :		\
+		    "i" (pda_offset(field)), 	\
+		    "m" (_proxy_pda.field)); 	\
+		 break;				\
+	case 4:					\
+		asm(op "l %%gs:%c1,%0":		\
+		    "=r" (ret__):		\
+		    "i" (pda_offset(field)), 	\
+		    "m" (_proxy_pda.field)); 	\
+		 break;				\
+       case 8:					\
+		asm(op "q %%gs:%c1,%0":		\
+		    "=r" (ret__) :		\
+		    "i" (pda_offset(field)), 	\
+		    "m" (_proxy_pda.field)); 	\
+		 break;				\
+       default: 				\
+		__bad_pda_field();		\
+       }					\
        ret__; })
 
-
 #define read_pda(field) pda_from_op("mov",field)
 #define write_pda(field,val) pda_to_op("mov",field,val)
 #define add_pda(field,val) pda_to_op("add",field,val)
diff --git a/include/asm-x86_64/percpu.h b/include/asm-x86_64/percpu.h
index 08dd9f9dda81..285756010c51 100644
--- a/include/asm-x86_64/percpu.h
+++ b/include/asm-x86_64/percpu.h
@@ -11,6 +11,16 @@
 
 #include <asm/pda.h>
 
+#ifdef CONFIG_MODULES
+# define PERCPU_MODULE_RESERVE 8192
+#else
+# define PERCPU_MODULE_RESERVE 0
+#endif
+
+#define PERCPU_ENOUGH_ROOM \
+	(ALIGN(__per_cpu_end - __per_cpu_start, SMP_CACHE_BYTES) + \
+	 PERCPU_MODULE_RESERVE)
+
 #define __per_cpu_offset(cpu) (cpu_pda(cpu)->data_offset)
 #define __my_cpu_offset() read_pda(data_offset)
 
@@ -21,9 +31,15 @@
     __attribute__((__section__(".data.percpu"))) __typeof__(type) per_cpu__##name
 
 /* var is in discarded region: offset to particular copy we want */
-#define per_cpu(var, cpu) (*RELOC_HIDE(&per_cpu__##var, __per_cpu_offset(cpu)))
-#define __get_cpu_var(var) (*RELOC_HIDE(&per_cpu__##var, __my_cpu_offset()))
-#define __raw_get_cpu_var(var) (*RELOC_HIDE(&per_cpu__##var, __my_cpu_offset()))
+#define per_cpu(var, cpu) (*({				\
+	extern int simple_indentifier_##var(void);	\
+	RELOC_HIDE(&per_cpu__##var, __per_cpu_offset(cpu)); }))
+#define __get_cpu_var(var) (*({				\
+	extern int simple_indentifier_##var(void);	\
+	RELOC_HIDE(&per_cpu__##var, __my_cpu_offset()); }))
+#define __raw_get_cpu_var(var) (*({			\
+	extern int simple_indentifier_##var(void);	\
+	RELOC_HIDE(&per_cpu__##var, __my_cpu_offset()); }))
 
 /* A macro to avoid #include hell... */
 #define percpu_modcopy(pcpudst, src, size)			\
diff --git a/include/asm-x86_64/pgtable.h b/include/asm-x86_64/pgtable.h
index a31ab4e68a9b..6899e770b173 100644
--- a/include/asm-x86_64/pgtable.h
+++ b/include/asm-x86_64/pgtable.h
@@ -21,12 +21,9 @@ extern unsigned long __supported_pte_mask;
 
 #define swapper_pg_dir init_level4_pgt
 
-extern int nonx_setup(char *str);
 extern void paging_init(void);
 extern void clear_kernel_mapping(unsigned long addr, unsigned long size);
 
-extern unsigned long pgkern_mask;
-
 /*
  * ZERO_PAGE is a global shared page that is always zero: used
  * for zero-mapped memory areas etc..
@@ -101,9 +98,6 @@ static inline void pgd_clear (pgd_t * pgd)
 	set_pgd(pgd, __pgd(0));
 }
 
-#define pud_page(pud) \
-((unsigned long) __va(pud_val(pud) & PHYSICAL_PAGE_MASK))
-
 #define ptep_get_and_clear(mm,addr,xp)	__pte(xchg(&(xp)->pte, 0))
 
 struct mm_struct;
@@ -268,7 +262,7 @@ static inline pte_t pfn_pte(unsigned long page_nr, pgprot_t pgprot)
 #define __LARGE_PTE (_PAGE_PSE|_PAGE_PRESENT)
 static inline int pte_user(pte_t pte)		{ return pte_val(pte) & _PAGE_USER; }
 static inline int pte_read(pte_t pte)		{ return pte_val(pte) & _PAGE_USER; }
-static inline int pte_exec(pte_t pte)		{ return pte_val(pte) & _PAGE_USER; }
+static inline int pte_exec(pte_t pte)		{ return !(pte_val(pte) & _PAGE_NX); }
 static inline int pte_dirty(pte_t pte)		{ return pte_val(pte) & _PAGE_DIRTY; }
 static inline int pte_young(pte_t pte)		{ return pte_val(pte) & _PAGE_ACCESSED; }
 static inline int pte_write(pte_t pte)		{ return pte_val(pte) & _PAGE_RW; }
@@ -281,11 +275,12 @@ static inline pte_t pte_mkclean(pte_t pte)	{ set_pte(&pte, __pte(pte_val(pte) &
 static inline pte_t pte_mkold(pte_t pte)	{ set_pte(&pte, __pte(pte_val(pte) & ~_PAGE_ACCESSED)); return pte; }
 static inline pte_t pte_wrprotect(pte_t pte)	{ set_pte(&pte, __pte(pte_val(pte) & ~_PAGE_RW)); return pte; }
 static inline pte_t pte_mkread(pte_t pte)	{ set_pte(&pte, __pte(pte_val(pte) | _PAGE_USER)); return pte; }
-static inline pte_t pte_mkexec(pte_t pte)	{ set_pte(&pte, __pte(pte_val(pte) | _PAGE_USER)); return pte; }
+static inline pte_t pte_mkexec(pte_t pte)	{ set_pte(&pte, __pte(pte_val(pte) & ~_PAGE_NX)); return pte; }
 static inline pte_t pte_mkdirty(pte_t pte)	{ set_pte(&pte, __pte(pte_val(pte) | _PAGE_DIRTY)); return pte; }
 static inline pte_t pte_mkyoung(pte_t pte)	{ set_pte(&pte, __pte(pte_val(pte) | _PAGE_ACCESSED)); return pte; }
 static inline pte_t pte_mkwrite(pte_t pte)	{ set_pte(&pte, __pte(pte_val(pte) | _PAGE_RW)); return pte; }
 static inline pte_t pte_mkhuge(pte_t pte)	{ set_pte(&pte, __pte(pte_val(pte) | _PAGE_PSE)); return pte; }
+static inline pte_t pte_clrhuge(pte_t pte)	{ set_pte(&pte, __pte(pte_val(pte) & ~_PAGE_PSE)); return pte; }
 
 struct vm_area_struct;
 
@@ -326,7 +321,8 @@ static inline int pmd_large(pmd_t pte) {
 /*
  * Level 4 access.
  */
-#define pgd_page(pgd) ((unsigned long) __va((unsigned long)pgd_val(pgd) & PTE_MASK))
+#define pgd_page_vaddr(pgd) ((unsigned long) __va((unsigned long)pgd_val(pgd) & PTE_MASK))
+#define pgd_page(pgd)		(pfn_to_page(pgd_val(pgd) >> PAGE_SHIFT))
 #define pgd_index(address) (((address) >> PGDIR_SHIFT) & (PTRS_PER_PGD-1))
 #define pgd_offset(mm, addr) ((mm)->pgd + pgd_index(addr))
 #define pgd_offset_k(address) (init_level4_pgt + pgd_index(address))
@@ -335,16 +331,18 @@ static inline int pmd_large(pmd_t pte) {
 
 /* PUD - Level3 access */
 /* to find an entry in a page-table-directory. */
+#define pud_page_vaddr(pud) ((unsigned long) __va(pud_val(pud) & PHYSICAL_PAGE_MASK))
+#define pud_page(pud)		(pfn_to_page(pud_val(pud) >> PAGE_SHIFT))
 #define pud_index(address) (((address) >> PUD_SHIFT) & (PTRS_PER_PUD-1))
-#define pud_offset(pgd, address) ((pud_t *) pgd_page(*(pgd)) + pud_index(address))
+#define pud_offset(pgd, address) ((pud_t *) pgd_page_vaddr(*(pgd)) + pud_index(address))
 #define pud_present(pud) (pud_val(pud) & _PAGE_PRESENT)
 
 /* PMD  - Level 2 access */
-#define pmd_page_kernel(pmd) ((unsigned long) __va(pmd_val(pmd) & PTE_MASK))
+#define pmd_page_vaddr(pmd) ((unsigned long) __va(pmd_val(pmd) & PTE_MASK))
 #define pmd_page(pmd)		(pfn_to_page(pmd_val(pmd) >> PAGE_SHIFT))
 
 #define pmd_index(address) (((address) >> PMD_SHIFT) & (PTRS_PER_PMD-1))
-#define pmd_offset(dir, address) ((pmd_t *) pud_page(*(dir)) + \
+#define pmd_offset(dir, address) ((pmd_t *) pud_page_vaddr(*(dir)) + \
 			pmd_index(address))
 #define pmd_none(x)	(!pmd_val(x))
 #define pmd_present(x)	(pmd_val(x) & _PAGE_PRESENT)
@@ -382,7 +380,7 @@ static inline pte_t pte_modify(pte_t pte, pgprot_t newprot)
 
 #define pte_index(address) \
 		(((address) >> PAGE_SHIFT) & (PTRS_PER_PTE - 1))
-#define pte_offset_kernel(dir, address) ((pte_t *) pmd_page_kernel(*(dir)) + \
+#define pte_offset_kernel(dir, address) ((pte_t *) pmd_page_vaddr(*(dir)) + \
 			pte_index(address))
 
 /* x86-64 always has all page tables mapped. */
diff --git a/include/asm-x86_64/proto.h b/include/asm-x86_64/proto.h
index 038fe1f47e6f..c28fc2db2171 100644
--- a/include/asm-x86_64/proto.h
+++ b/include/asm-x86_64/proto.h
@@ -24,8 +24,6 @@ extern void mtrr_bp_init(void);
 #define mtrr_bp_init() do {} while (0)
 #endif
 extern void init_memory_mapping(unsigned long start, unsigned long end);
-extern void size_zones(unsigned long *z, unsigned long *h,
-			unsigned long start_pfn, unsigned long end_pfn);
 
 extern void system_call(void); 
 extern int kernel_syscall(void);
@@ -51,10 +49,8 @@ extern unsigned long long monotonic_base;
 extern int sysctl_vsyscall;
 extern int nohpet;
 extern unsigned long vxtime_hz;
+extern void time_init_gtod(void);
 
-extern int numa_setup(char *opt);
-
-extern int setup_early_printk(char *); 
 extern void early_printk(const char *fmt, ...) __attribute__((format(printf,1,2)));
 
 extern void early_identify_cpu(struct cpuinfo_x86 *c);
@@ -91,7 +87,7 @@ extern void syscall32_cpu_init(void);
 
 extern void setup_node_bootmem(int nodeid, unsigned long start, unsigned long end);
 
-extern void check_ioapic(void);
+extern void early_quirks(void);
 extern void check_efer(void);
 
 extern int unhandled_signal(struct task_struct *tsk, int sig);
@@ -103,13 +99,7 @@ extern void select_idle_routine(const struct cpuinfo_x86 *c);
 extern unsigned long table_start, table_end;
 
 extern int exception_trace;
-extern int using_apic_timer;
-extern int disable_apic;
 extern unsigned cpu_khz;
-extern int ioapic_force;
-extern int skip_ioapic_setup;
-extern int acpi_ht;
-extern int acpi_disabled;
 
 extern void no_iommu_init(void);
 extern int force_iommu, no_iommu;
@@ -131,7 +121,8 @@ extern int fix_aperture;
 
 extern int reboot_force;
 extern int notsc_setup(char *);
-extern int setup_additional_cpus(char *);
+
+extern int gsi_irq_sharing(int gsi);
 
 extern void smp_local_timer_interrupt(struct pt_regs * regs);
 
diff --git a/include/asm-x86_64/ptrace-abi.h b/include/asm-x86_64/ptrace-abi.h
new file mode 100644
index 000000000000..19184b0806b1
--- /dev/null
+++ b/include/asm-x86_64/ptrace-abi.h
@@ -0,0 +1,51 @@
+#ifndef _X86_64_PTRACE_ABI_H
+#define _X86_64_PTRACE_ABI_H
+
+#if defined(__ASSEMBLY__) || defined(__FRAME_OFFSETS)
+#define R15 0
+#define R14 8
+#define R13 16
+#define R12 24
+#define RBP 32
+#define RBX 40
+/* arguments: interrupts/non tracing syscalls only save upto here*/
+#define R11 48
+#define R10 56
+#define R9 64
+#define R8 72
+#define RAX 80
+#define RCX 88
+#define RDX 96
+#define RSI 104
+#define RDI 112
+#define ORIG_RAX 120       /* = ERROR */
+/* end of arguments */
+/* cpu exception frame or undefined in case of fast syscall. */
+#define RIP 128
+#define CS 136
+#define EFLAGS 144
+#define RSP 152
+#define SS 160
+#define ARGOFFSET R11
+#endif /* __ASSEMBLY__ */
+
+/* top of stack page */
+#define FRAME_SIZE 168
+
+#define PTRACE_OLDSETOPTIONS         21
+
+/* Arbitrarily choose the same ptrace numbers as used by the Sparc code. */
+#define PTRACE_GETREGS            12
+#define PTRACE_SETREGS            13
+#define PTRACE_GETFPREGS          14
+#define PTRACE_SETFPREGS          15
+#define PTRACE_GETFPXREGS         18
+#define PTRACE_SETFPXREGS         19
+
+/* only useful for access 32bit programs */
+#define PTRACE_GET_THREAD_AREA    25
+#define PTRACE_SET_THREAD_AREA    26
+
+#define PTRACE_ARCH_PRCTL	  30	/* arch_prctl for child */
+
+#endif
diff --git a/include/asm-x86_64/ptrace.h b/include/asm-x86_64/ptrace.h
index ca6f15ff61d4..ab827dc381d7 100644
--- a/include/asm-x86_64/ptrace.h
+++ b/include/asm-x86_64/ptrace.h
@@ -1,40 +1,9 @@
 #ifndef _X86_64_PTRACE_H
 #define _X86_64_PTRACE_H
 
-#if defined(__ASSEMBLY__) || defined(__FRAME_OFFSETS) 
-#define R15 0
-#define R14 8
-#define R13 16
-#define R12 24
-#define RBP 32
-#define RBX 40
-/* arguments: interrupts/non tracing syscalls only save upto here*/
-#define R11 48
-#define R10 56	
-#define R9 64
-#define R8 72
-#define RAX 80
-#define RCX 88
-#define RDX 96
-#define RSI 104
-#define RDI 112
-#define ORIG_RAX 120       /* = ERROR */ 
-/* end of arguments */ 	
-/* cpu exception frame or undefined in case of fast syscall. */
-#define RIP 128
-#define CS 136
-#define EFLAGS 144
-#define RSP 152
-#define SS 160
-#define ARGOFFSET R11
-#endif /* __ASSEMBLY__ */
+#include <asm/ptrace-abi.h>
 
-/* top of stack page */ 
-#define FRAME_SIZE 168
-
-#define PTRACE_OLDSETOPTIONS         21
-
-#ifndef __ASSEMBLY__ 
+#ifndef __ASSEMBLY__
 
 struct pt_regs {
 	unsigned long r15;
@@ -45,7 +14,7 @@ struct pt_regs {
 	unsigned long rbx;
 /* arguments: non interrupts/non tracing syscalls only save upto here*/
  	unsigned long r11;
-	unsigned long r10;	
+	unsigned long r10;
 	unsigned long r9;
 	unsigned long r8;
 	unsigned long rax;
@@ -54,32 +23,18 @@ struct pt_regs {
 	unsigned long rsi;
 	unsigned long rdi;
 	unsigned long orig_rax;
-/* end of arguments */ 	
+/* end of arguments */
 /* cpu exception frame or undefined */
 	unsigned long rip;
 	unsigned long cs;
-	unsigned long eflags; 
-	unsigned long rsp; 
+	unsigned long eflags;
+	unsigned long rsp;
 	unsigned long ss;
-/* top of stack page */ 
+/* top of stack page */
 };
 
 #endif
 
-/* Arbitrarily choose the same ptrace numbers as used by the Sparc code. */
-#define PTRACE_GETREGS            12
-#define PTRACE_SETREGS            13
-#define PTRACE_GETFPREGS          14
-#define PTRACE_SETFPREGS          15
-#define PTRACE_GETFPXREGS         18
-#define PTRACE_SETFPXREGS         19
-
-/* only useful for access 32bit programs */
-#define PTRACE_GET_THREAD_AREA    25
-#define PTRACE_SET_THREAD_AREA    26
-
-#define PTRACE_ARCH_PRCTL	  30	/* arch_prctl for child */
-
 #if defined(__KERNEL__) && !defined(__ASSEMBLY__) 
 #define user_mode(regs) (!!((regs)->cs & 3))
 #define user_mode_vm(regs) user_mode(regs)
diff --git a/include/asm-x86_64/rwlock.h b/include/asm-x86_64/rwlock.h
index dea0e9459264..72aeebed920b 100644
--- a/include/asm-x86_64/rwlock.h
+++ b/include/asm-x86_64/rwlock.h
@@ -18,69 +18,9 @@
 #ifndef _ASM_X86_64_RWLOCK_H
 #define _ASM_X86_64_RWLOCK_H
 
-#include <linux/stringify.h>
-
 #define RW_LOCK_BIAS		 0x01000000
-#define RW_LOCK_BIAS_STR	"0x01000000"
-
-#define __build_read_lock_ptr(rw, helper)   \
-	asm volatile(LOCK_PREFIX "subl $1,(%0)\n\t" \
-		     "js 2f\n" \
-		     "1:\n" \
-		    LOCK_SECTION_START("") \
-		     "2:\tcall " helper "\n\t" \
-		     "jmp 1b\n" \
-		    LOCK_SECTION_END \
-		     ::"a" (rw) : "memory")
-
-#define __build_read_lock_const(rw, helper)   \
-	asm volatile(LOCK_PREFIX "subl $1,%0\n\t" \
-		     "js 2f\n" \
-		     "1:\n" \
-		    LOCK_SECTION_START("") \
-		     "2:\tpushq %%rax\n\t" \
-		     "leaq %0,%%rax\n\t" \
-		     "call " helper "\n\t" \
-		     "popq %%rax\n\t" \
-		     "jmp 1b\n" \
-		    LOCK_SECTION_END \
-		     :"=m" (*((volatile int *)rw))::"memory")
-
-#define __build_read_lock(rw, helper)	do { \
-						if (__builtin_constant_p(rw)) \
-							__build_read_lock_const(rw, helper); \
-						else \
-							__build_read_lock_ptr(rw, helper); \
-					} while (0)
-
-#define __build_write_lock_ptr(rw, helper) \
-	asm volatile(LOCK_PREFIX "subl $" RW_LOCK_BIAS_STR ",(%0)\n\t" \
-		     "jnz 2f\n" \
-		     "1:\n" \
-		     LOCK_SECTION_START("") \
-		     "2:\tcall " helper "\n\t" \
-		     "jmp 1b\n" \
-		     LOCK_SECTION_END \
-		     ::"a" (rw) : "memory")
-
-#define __build_write_lock_const(rw, helper) \
-	asm volatile(LOCK_PREFIX "subl $" RW_LOCK_BIAS_STR ",%0\n\t" \
-		     "jnz 2f\n" \
-		     "1:\n" \
-		    LOCK_SECTION_START("") \
-		     "2:\tpushq %%rax\n\t" \
-		     "leaq %0,%%rax\n\t" \
-		     "call " helper "\n\t" \
-		     "popq %%rax\n\t" \
-		     "jmp 1b\n" \
-		    LOCK_SECTION_END \
-		     :"=m" (*((volatile long *)rw))::"memory")
+#define RW_LOCK_BIAS_STR	 "0x01000000"
 
-#define __build_write_lock(rw, helper)	do { \
-						if (__builtin_constant_p(rw)) \
-							__build_write_lock_const(rw, helper); \
-						else \
-							__build_write_lock_ptr(rw, helper); \
-					} while (0)
+/* Actual code is in asm/spinlock.h or in arch/x86_64/lib/rwlock.S */
 
 #endif
diff --git a/include/asm-x86_64/segment.h b/include/asm-x86_64/segment.h
index d4bed33fb32c..334ddcdd8f92 100644
--- a/include/asm-x86_64/segment.h
+++ b/include/asm-x86_64/segment.h
@@ -20,15 +20,16 @@
 #define __USER_CS     0x33   /* 6*8+3 */ 
 #define __USER32_DS	__USER_DS 
 
-#define GDT_ENTRY_TLS 1
 #define GDT_ENTRY_TSS 8	/* needs two entries */
 #define GDT_ENTRY_LDT 10 /* needs two entries */
 #define GDT_ENTRY_TLS_MIN 12
 #define GDT_ENTRY_TLS_MAX 14
-/* 15 free */
 
 #define GDT_ENTRY_TLS_ENTRIES 3
 
+#define GDT_ENTRY_PER_CPU 15	/* Abused to load per CPU data from limit */
+#define __PER_CPU_SEG	(GDT_ENTRY_PER_CPU * 8 + 3)
+
 /* TLS indexes for 64bit - hardcoded in arch_prctl */
 #define FS_TLS 0	
 #define GS_TLS 1	
diff --git a/include/asm-x86_64/semaphore.h b/include/asm-x86_64/semaphore.h
index 064df08b9a0f..107bd90429e8 100644
--- a/include/asm-x86_64/semaphore.h
+++ b/include/asm-x86_64/semaphore.h
@@ -107,12 +107,9 @@ static inline void down(struct semaphore * sem)
 	__asm__ __volatile__(
 		"# atomic down operation\n\t"
 		LOCK_PREFIX "decl %0\n\t"     /* --sem->count */
-		"js 2f\n"
-		"1:\n"
-		LOCK_SECTION_START("")
-		"2:\tcall __down_failed\n\t"
-		"jmp 1b\n"
-		LOCK_SECTION_END
+		"jns 1f\n\t"
+		"call __down_failed\n"
+		"1:"
 		:"=m" (sem->count)
 		:"D" (sem)
 		:"memory");
@@ -130,14 +127,11 @@ static inline int down_interruptible(struct semaphore * sem)
 
 	__asm__ __volatile__(
 		"# atomic interruptible down operation\n\t"
+		"xorl %0,%0\n\t"
 		LOCK_PREFIX "decl %1\n\t"     /* --sem->count */
-		"js 2f\n\t"
-		"xorl %0,%0\n"
-		"1:\n"
-		LOCK_SECTION_START("")
-		"2:\tcall __down_failed_interruptible\n\t"
-		"jmp 1b\n"
-		LOCK_SECTION_END
+		"jns 2f\n\t"
+		"call __down_failed_interruptible\n"
+		"2:\n"
 		:"=a" (result), "=m" (sem->count)
 		:"D" (sem)
 		:"memory");
@@ -154,14 +148,11 @@ static inline int down_trylock(struct semaphore * sem)
 
 	__asm__ __volatile__(
 		"# atomic interruptible down operation\n\t"
+		"xorl %0,%0\n\t"
 		LOCK_PREFIX "decl %1\n\t"     /* --sem->count */
-		"js 2f\n\t"
-		"xorl %0,%0\n"
-		"1:\n"
-		LOCK_SECTION_START("")
-		"2:\tcall __down_failed_trylock\n\t"
-		"jmp 1b\n"
-		LOCK_SECTION_END
+		"jns 2f\n\t"
+		"call __down_failed_trylock\n\t"
+		"2:\n"
 		:"=a" (result), "=m" (sem->count)
 		:"D" (sem)
 		:"memory","cc");
@@ -179,12 +170,9 @@ static inline void up(struct semaphore * sem)
 	__asm__ __volatile__(
 		"# atomic up operation\n\t"
 		LOCK_PREFIX "incl %0\n\t"     /* ++sem->count */
-		"jle 2f\n"
-		"1:\n"
-		LOCK_SECTION_START("")
-		"2:\tcall __up_wakeup\n\t"
-		"jmp 1b\n"
-		LOCK_SECTION_END
+		"jg 1f\n\t"
+		"call __up_wakeup\n"
+		"1:"
 		:"=m" (sem->count)
 		:"D" (sem)
 		:"memory");
diff --git a/include/asm-x86_64/signal.h b/include/asm-x86_64/signal.h
index 3ede2a61973a..4581f978b299 100644
--- a/include/asm-x86_64/signal.h
+++ b/include/asm-x86_64/signal.h
@@ -24,10 +24,6 @@ typedef struct {
 } sigset_t;
 
 
-struct pt_regs; 
-asmlinkage int do_signal(struct pt_regs *regs, sigset_t *oldset);
-
-
 #else
 /* Here we must cater to libcs that poke about in kernel headers.  */
 
diff --git a/include/asm-x86_64/smp.h b/include/asm-x86_64/smp.h
index 6805e1feb300..d6b7c057edba 100644
--- a/include/asm-x86_64/smp.h
+++ b/include/asm-x86_64/smp.h
@@ -4,27 +4,18 @@
 /*
  * We need the APIC definitions automatically as part of 'smp.h'
  */
-#ifndef __ASSEMBLY__
 #include <linux/threads.h>
 #include <linux/cpumask.h>
 #include <linux/bitops.h>
 extern int disable_apic;
-#endif
 
-#ifdef CONFIG_X86_LOCAL_APIC
-#ifndef __ASSEMBLY__
 #include <asm/fixmap.h>
 #include <asm/mpspec.h>
-#ifdef CONFIG_X86_IO_APIC
 #include <asm/io_apic.h>
-#endif
 #include <asm/apic.h>
 #include <asm/thread_info.h>
-#endif
-#endif
 
 #ifdef CONFIG_SMP
-#ifndef ASSEMBLY
 
 #include <asm/pda.h>
 
@@ -42,14 +33,11 @@ extern cpumask_t cpu_initialized;
  
 extern void smp_alloc_memory(void);
 extern volatile unsigned long smp_invalidate_needed;
-extern int pic_mode;
 extern void lock_ipi_call_lock(void);
 extern void unlock_ipi_call_lock(void);
 extern int smp_num_siblings;
 extern void smp_send_reschedule(int cpu);
 void smp_stop_cpu(void);
-extern int smp_call_function_single(int cpuid, void (*func) (void *info),
-				void *info, int retry, int wait);
 
 extern cpumask_t cpu_sibling_map[NR_CPUS];
 extern cpumask_t cpu_core_map[NR_CPUS];
@@ -76,20 +64,16 @@ static inline int hard_smp_processor_id(void)
 	return GET_APIC_ID(*(unsigned int *)(APIC_BASE+APIC_ID));
 }
 
-extern int safe_smp_processor_id(void);
 extern int __cpu_disable(void);
 extern void __cpu_die(unsigned int cpu);
 extern void prefill_possible_map(void);
 extern unsigned num_processors;
 extern unsigned disabled_cpus;
 
-#endif /* !ASSEMBLY */
-
 #define NO_PROC_ID		0xFF		/* No processor magic marker */
 
 #endif
 
-#ifndef ASSEMBLY
 /*
  * Some lowlevel functions might want to know about
  * the real APIC ID <-> CPU # mapping.
@@ -111,11 +95,8 @@ static inline int cpu_present_to_apicid(int mps_cpu)
 		return BAD_APICID;
 }
 
-#endif /* !ASSEMBLY */
-
 #ifndef CONFIG_SMP
 #define stack_smp_processor_id() 0
-#define safe_smp_processor_id() 0
 #define cpu_logical_map(x) (x)
 #else
 #include <asm/thread_info.h>
@@ -127,19 +108,23 @@ static inline int cpu_present_to_apicid(int mps_cpu)
 })
 #endif
 
-#ifndef __ASSEMBLY__
 static __inline int logical_smp_processor_id(void)
 {
 	/* we don't want to mark this access volatile - bad code generation */
 	return GET_APIC_LOGICAL_ID(*(unsigned long *)(APIC_BASE+APIC_LDR));
 }
-#endif
 
 #ifdef CONFIG_SMP
 #define cpu_physical_id(cpu)		x86_cpu_to_apicid[cpu]
 #else
 #define cpu_physical_id(cpu)		boot_cpu_id
-#endif
-
+static inline int smp_call_function_single(int cpuid, void (*func) (void *info),
+				void *info, int retry, int wait)
+{
+	/* Disable interrupts here? */
+	func(info);
+	return 0;
+}
+#endif /* !CONFIG_SMP */
 #endif
 
diff --git a/include/asm-x86_64/spinlock.h b/include/asm-x86_64/spinlock.h
index 248a79f0eaff..3daf5b005905 100644
--- a/include/asm-x86_64/spinlock.h
+++ b/include/asm-x86_64/spinlock.h
@@ -4,6 +4,7 @@
 #include <asm/atomic.h>
 #include <asm/rwlock.h>
 #include <asm/page.h>
+#include <asm/processor.h>
 
 /*
  * Your basic SMP spinlocks, allowing only a single CPU anywhere
@@ -16,31 +17,23 @@
  * (the type definitions are in asm/spinlock_types.h)
  */
 
-#define __raw_spin_is_locked(x) \
-		(*(volatile signed int *)(&(x)->slock) <= 0)
-
-#define __raw_spin_lock_string \
-	"\n1:\t" \
-	LOCK_PREFIX " ; decl %0\n\t" \
-	"js 2f\n" \
-	LOCK_SECTION_START("") \
-	"2:\t" \
-	"rep;nop\n\t" \
-	"cmpl $0,%0\n\t" \
-	"jle 2b\n\t" \
-	"jmp 1b\n" \
-	LOCK_SECTION_END
-
-#define __raw_spin_lock_string_up \
-	"\n\tdecl %0"
-
-#define __raw_spin_unlock_string \
-	"movl $1,%0" \
-		:"=m" (lock->slock) : : "memory"
+static inline int __raw_spin_is_locked(raw_spinlock_t *lock)
+{
+	return *(volatile signed int *)(&(lock)->slock) <= 0;
+}
 
 static inline void __raw_spin_lock(raw_spinlock_t *lock)
 {
-	asm volatile(__raw_spin_lock_string : "=m" (lock->slock) : : "memory");
+	asm volatile(
+		"\n1:\t"
+		LOCK_PREFIX " ; decl %0\n\t"
+		"jns 2f\n"
+		"3:\n"
+		"rep;nop\n\t"
+		"cmpl $0,%0\n\t"
+		"jle 3b\n\t"
+		"jmp 1b\n"
+		"2:\t" : "=m" (lock->slock) : : "memory");
 }
 
 #define __raw_spin_lock_flags(lock, flags) __raw_spin_lock(lock)
@@ -49,7 +42,7 @@ static inline int __raw_spin_trylock(raw_spinlock_t *lock)
 {
 	int oldval;
 
-	__asm__ __volatile__(
+	asm volatile(
 		"xchgl %0,%1"
 		:"=q" (oldval), "=m" (lock->slock)
 		:"0" (0) : "memory");
@@ -59,13 +52,14 @@ static inline int __raw_spin_trylock(raw_spinlock_t *lock)
 
 static inline void __raw_spin_unlock(raw_spinlock_t *lock)
 {
-	__asm__ __volatile__(
-		__raw_spin_unlock_string
-	);
+	asm volatile("movl $1,%0" :"=m" (lock->slock) :: "memory");
 }
 
-#define __raw_spin_unlock_wait(lock) \
-	do { while (__raw_spin_is_locked(lock)) cpu_relax(); } while (0)
+static inline void __raw_spin_unlock_wait(raw_spinlock_t *lock)
+{
+	while (__raw_spin_is_locked(lock))
+		cpu_relax();
+}
 
 /*
  * Read-write spinlocks, allowing multiple readers
@@ -79,26 +73,34 @@ static inline void __raw_spin_unlock(raw_spinlock_t *lock)
  *
  * On x86, we implement read-write locks as a 32-bit counter
  * with the high bit (sign) being the "contended" bit.
- *
- * The inline assembly is non-obvious. Think about it.
- *
- * Changed to use the same technique as rw semaphores.  See
- * semaphore.h for details.  -ben
- *
- * the helpers are in arch/i386/kernel/semaphore.c
  */
 
-#define __raw_read_can_lock(x)		((int)(x)->lock > 0)
-#define __raw_write_can_lock(x)		((x)->lock == RW_LOCK_BIAS)
+static inline int __raw_read_can_lock(raw_rwlock_t *lock)
+{
+	return (int)(lock)->lock > 0;
+}
+
+static inline int __raw_write_can_lock(raw_rwlock_t *lock)
+{
+	return (lock)->lock == RW_LOCK_BIAS;
+}
 
 static inline void __raw_read_lock(raw_rwlock_t *rw)
 {
-	__build_read_lock(rw, "__read_lock_failed");
+	asm volatile(LOCK_PREFIX "subl $1,(%0)\n\t"
+		     "jns 1f\n"
+		     "call __read_lock_failed\n"
+		     "1:\n"
+		     ::"D" (rw), "i" (RW_LOCK_BIAS) : "memory");
 }
 
 static inline void __raw_write_lock(raw_rwlock_t *rw)
 {
-	__build_write_lock(rw, "__write_lock_failed");
+	asm volatile(LOCK_PREFIX "subl %1,(%0)\n\t"
+		     "jz 1f\n"
+		     "\tcall __write_lock_failed\n\t"
+		     "1:\n"
+		     ::"D" (rw), "i" (RW_LOCK_BIAS) : "memory");
 }
 
 static inline int __raw_read_trylock(raw_rwlock_t *lock)
diff --git a/include/asm-x86_64/stacktrace.h b/include/asm-x86_64/stacktrace.h
new file mode 100644
index 000000000000..5eb9799bef76
--- /dev/null
+++ b/include/asm-x86_64/stacktrace.h
@@ -0,0 +1,18 @@
+#ifndef _ASM_STACKTRACE_H
+#define _ASM_STACKTRACE_H 1
+
+/* Generic stack tracer with callbacks */
+
+struct stacktrace_ops {
+	void (*warning)(void *data, char *msg);
+	/* msg must contain %s for the symbol */
+	void (*warning_symbol)(void *data, char *msg, unsigned long symbol);
+	void (*address)(void *data, unsigned long address);
+	/* On negative return stop dumping */
+	int (*stack)(void *data, char *name);
+};
+
+void dump_trace(struct task_struct *tsk, struct pt_regs *regs, unsigned long *stack,
+		struct stacktrace_ops *ops, void *data);
+
+#endif
diff --git a/include/asm-x86_64/system.h b/include/asm-x86_64/system.h
index 6bf170bceae1..bd376bc8c4ab 100644
--- a/include/asm-x86_64/system.h
+++ b/include/asm-x86_64/system.h
@@ -14,12 +14,13 @@
 #define __RESTORE(reg,offset) "movq (14-" #offset ")*8(%%rsp),%%" #reg "\n\t"
 
 /* frame pointer must be last for get_wchan */
-#define SAVE_CONTEXT    "pushq %%rbp ; movq %%rsi,%%rbp\n\t"
-#define RESTORE_CONTEXT "movq %%rbp,%%rsi ; popq %%rbp\n\t"
+#define SAVE_CONTEXT    "pushf ; pushq %%rbp ; movq %%rsi,%%rbp\n\t"
+#define RESTORE_CONTEXT "movq %%rbp,%%rsi ; popq %%rbp ; popf\t"
 
 #define __EXTRA_CLOBBER  \
 	,"rcx","rbx","rdx","r8","r9","r10","r11","r12","r13","r14","r15"
 
+/* Save restore flags to clear handle leaking NT */
 #define switch_to(prev,next,last) \
 	asm volatile(SAVE_CONTEXT						    \
 		     "movq %%rsp,%P[threadrsp](%[prev])\n\t" /* save RSP */	  \
diff --git a/include/asm-x86_64/tce.h b/include/asm-x86_64/tce.h
index 53e9a68b3336..dbb047febc5e 100644
--- a/include/asm-x86_64/tce.h
+++ b/include/asm-x86_64/tce.h
@@ -24,7 +24,6 @@
 #ifndef _ASM_X86_64_TCE_H
 #define _ASM_X86_64_TCE_H
 
-extern void* tce_table_kva[];
 extern unsigned int specified_table_size;
 struct iommu_table;
 
diff --git a/include/asm-x86_64/therm_throt.h b/include/asm-x86_64/therm_throt.h
new file mode 100644
index 000000000000..5aac059007ba
--- /dev/null
+++ b/include/asm-x86_64/therm_throt.h
@@ -0,0 +1 @@
+#include <asm-i386/therm_throt.h>
diff --git a/include/asm-x86_64/thread_info.h b/include/asm-x86_64/thread_info.h
index 2029b00351f3..787a08114b48 100644
--- a/include/asm-x86_64/thread_info.h
+++ b/include/asm-x86_64/thread_info.h
@@ -114,11 +114,14 @@ static inline struct thread_info *stack_thread_info(void)
 #define TIF_IRET		5	/* force IRET */
 #define TIF_SYSCALL_AUDIT	7	/* syscall auditing active */
 #define TIF_SECCOMP		8	/* secure computing */
+#define TIF_RESTORE_SIGMASK	9	/* restore signal mask in do_signal */
 /* 16 free */
 #define TIF_IA32		17	/* 32bit process */ 
 #define TIF_FORK		18	/* ret_from_fork */
 #define TIF_ABI_PENDING		19
 #define TIF_MEMDIE		20
+#define TIF_DEBUG		21	/* uses debug registers */
+#define TIF_IO_BITMAP		22	/* uses I/O bitmap */
 
 #define _TIF_SYSCALL_TRACE	(1<<TIF_SYSCALL_TRACE)
 #define _TIF_NOTIFY_RESUME	(1<<TIF_NOTIFY_RESUME)
@@ -128,9 +131,12 @@ static inline struct thread_info *stack_thread_info(void)
 #define _TIF_IRET		(1<<TIF_IRET)
 #define _TIF_SYSCALL_AUDIT	(1<<TIF_SYSCALL_AUDIT)
 #define _TIF_SECCOMP		(1<<TIF_SECCOMP)
+#define _TIF_RESTORE_SIGMASK	(1<<TIF_RESTORE_SIGMASK)
 #define _TIF_IA32		(1<<TIF_IA32)
 #define _TIF_FORK		(1<<TIF_FORK)
 #define _TIF_ABI_PENDING	(1<<TIF_ABI_PENDING)
+#define _TIF_DEBUG		(1<<TIF_DEBUG)
+#define _TIF_IO_BITMAP		(1<<TIF_IO_BITMAP)
 
 /* work to do on interrupt/exception return */
 #define _TIF_WORK_MASK \
@@ -138,6 +144,9 @@ static inline struct thread_info *stack_thread_info(void)
 /* work to do on any return to user space */
 #define _TIF_ALLWORK_MASK (0x0000FFFF & ~_TIF_SECCOMP)
 
+/* flags to check in __switch_to() */
+#define _TIF_WORK_CTXSW (_TIF_DEBUG|_TIF_IO_BITMAP)
+
 #define PREEMPT_ACTIVE     0x10000000
 
 /*
diff --git a/include/asm-x86_64/tlbflush.h b/include/asm-x86_64/tlbflush.h
index d16d5b60f419..983bd296c81a 100644
--- a/include/asm-x86_64/tlbflush.h
+++ b/include/asm-x86_64/tlbflush.h
@@ -4,44 +4,44 @@
 #include <linux/mm.h>
 #include <asm/processor.h>
 
-#define __flush_tlb()							\
-	do {								\
-		unsigned long tmpreg;					\
-									\
-		__asm__ __volatile__(					\
-			"movq %%cr3, %0;  # flush TLB \n"		\
-			"movq %0, %%cr3;              \n"		\
-			: "=r" (tmpreg)					\
-			:: "memory");					\
-	} while (0)
+static inline unsigned long get_cr3(void)
+{
+	unsigned long cr3;
+	asm volatile("mov %%cr3,%0" : "=r" (cr3));
+	return cr3;
+}
 
-/*
- * Global pages have to be flushed a bit differently. Not a real
- * performance problem because this does not happen often.
- */
-#define __flush_tlb_global()						\
-	do {								\
-		unsigned long tmpreg, cr4, cr4_orig;			\
-									\
-		__asm__ __volatile__(					\
-			"movq %%cr4, %2;  # turn off PGE     \n"	\
-			"movq %2, %1;                        \n"	\
-			"andq %3, %1;                        \n"	\
-			"movq %1, %%cr4;                     \n"	\
-			"movq %%cr3, %0;  # flush TLB        \n"	\
-			"movq %0, %%cr3;                     \n"	\
-			"movq %2, %%cr4;  # turn PGE back on \n"	\
-			: "=&r" (tmpreg), "=&r" (cr4), "=&r" (cr4_orig)	\
-			: "i" (~X86_CR4_PGE)				\
-			: "memory");					\
-	} while (0)
-
-extern unsigned long pgkern_mask;
-
-#define __flush_tlb_all() __flush_tlb_global()
+static inline void set_cr3(unsigned long cr3)
+{
+	asm volatile("mov %0,%%cr3" :: "r" (cr3) : "memory");
+}
+
+static inline void __flush_tlb(void)
+{
+	set_cr3(get_cr3());
+}
+
+static inline unsigned long get_cr4(void)
+{
+	unsigned long cr4;
+	asm volatile("mov %%cr4,%0" : "=r" (cr4));
+	return cr4;
+}
+
+static inline void set_cr4(unsigned long cr4)
+{
+	asm volatile("mov %0,%%cr4" :: "r" (cr4) : "memory");
+}
+
+static inline void __flush_tlb_all(void)
+{
+	unsigned long cr4 = get_cr4();
+	set_cr4(cr4 & ~X86_CR4_PGE);	/* clear PGE */
+	set_cr4(cr4);			/* write old PGE again and flush TLBs */
+}
 
 #define __flush_tlb_one(addr) \
-	__asm__ __volatile__("invlpg %0": :"m" (*(char *) addr))
+	__asm__ __volatile__("invlpg (%0)" :: "r" (addr) : "memory")
 
 
 /*
diff --git a/include/asm-x86_64/uaccess.h b/include/asm-x86_64/uaccess.h
index 1e1fa003daa3..e856570c0689 100644
--- a/include/asm-x86_64/uaccess.h
+++ b/include/asm-x86_64/uaccess.h
@@ -84,7 +84,7 @@ struct exception_table_entry
  */
 
 #define __get_user_x(size,ret,x,ptr) \
-	__asm__ __volatile__("call __get_user_" #size \
+	asm volatile("call __get_user_" #size \
 		:"=a" (ret),"=d" (x) \
 		:"c" (ptr) \
 		:"r8")
@@ -101,7 +101,7 @@ struct exception_table_entry
 	case 8:  __get_user_x(8,__ret_gu,__val_gu,ptr); break;		\
 	default: __get_user_bad(); break;				\
 	}								\
-	(x) = (__typeof__(*(ptr)))__val_gu;				\
+	(x) = (typeof(*(ptr)))__val_gu;				\
 	__ret_gu;							\
 })
 
@@ -112,7 +112,7 @@ extern void __put_user_8(void);
 extern void __put_user_bad(void);
 
 #define __put_user_x(size,ret,x,ptr)					\
-	__asm__ __volatile__("call __put_user_" #size			\
+	asm volatile("call __put_user_" #size			\
 		:"=a" (ret)						\
 		:"c" (ptr),"d" (x)					\
 		:"r8")
@@ -139,7 +139,7 @@ extern void __put_user_bad(void);
 #define __put_user_check(x,ptr,size)			\
 ({							\
 	int __pu_err;					\
-	__typeof__(*(ptr)) __user *__pu_addr = (ptr);	\
+	typeof(*(ptr)) __user *__pu_addr = (ptr);	\
 	switch (size) { 				\
 	case 1: __put_user_x(1,__pu_err,x,__pu_addr); break;	\
 	case 2: __put_user_x(2,__pu_err,x,__pu_addr); break;	\
@@ -173,7 +173,7 @@ struct __large_struct { unsigned long buf[100]; };
  * aliasing issues.
  */
 #define __put_user_asm(x, addr, err, itype, rtype, ltype, errno)	\
-	__asm__ __volatile__(					\
+	asm volatile(					\
 		"1:	mov"itype" %"rtype"1,%2\n"		\
 		"2:\n"						\
 		".section .fixup,\"ax\"\n"			\
@@ -193,7 +193,7 @@ struct __large_struct { unsigned long buf[100]; };
 	int __gu_err;						\
 	unsigned long __gu_val;					\
 	__get_user_size(__gu_val,(ptr),(size),__gu_err);	\
-	(x) = (__typeof__(*(ptr)))__gu_val;			\
+	(x) = (typeof(*(ptr)))__gu_val;			\
 	__gu_err;						\
 })
 
@@ -217,7 +217,7 @@ do {									\
 } while (0)
 
 #define __get_user_asm(x, addr, err, itype, rtype, ltype, errno)	\
-	__asm__ __volatile__(					\
+	asm volatile(					\
 		"1:	mov"itype" %2,%"rtype"1\n"		\
 		"2:\n"						\
 		".section .fixup,\"ax\"\n"			\
@@ -237,15 +237,20 @@ do {									\
  */
 
 /* Handles exceptions in both to and from, but doesn't do access_ok */
-extern unsigned long copy_user_generic(void *to, const void *from, unsigned len); 
-
-extern unsigned long copy_to_user(void __user *to, const void *from, unsigned len); 
-extern unsigned long copy_from_user(void *to, const void __user *from, unsigned len); 
-extern unsigned long copy_in_user(void __user *to, const void __user *from, unsigned len); 
-
-static __always_inline int __copy_from_user(void *dst, const void __user *src, unsigned size)
+__must_check unsigned long
+copy_user_generic(void *to, const void *from, unsigned len);
+
+__must_check unsigned long
+copy_to_user(void __user *to, const void *from, unsigned len);
+__must_check unsigned long
+copy_from_user(void *to, const void __user *from, unsigned len);
+__must_check unsigned long
+copy_in_user(void __user *to, const void __user *from, unsigned len);
+
+static __always_inline __must_check
+int __copy_from_user(void *dst, const void __user *src, unsigned size)
 { 
-       int ret = 0;
+	int ret = 0;
 	if (!__builtin_constant_p(size))
 		return copy_user_generic(dst,(__force void *)src,size);
 	switch (size) { 
@@ -272,9 +277,10 @@ static __always_inline int __copy_from_user(void *dst, const void __user *src, u
 	}
 }	
 
-static __always_inline int __copy_to_user(void __user *dst, const void *src, unsigned size)
+static __always_inline __must_check
+int __copy_to_user(void __user *dst, const void *src, unsigned size)
 { 
-       int ret = 0;
+	int ret = 0;
 	if (!__builtin_constant_p(size))
 		return copy_user_generic((__force void *)dst,src,size);
 	switch (size) { 
@@ -303,10 +309,10 @@ static __always_inline int __copy_to_user(void __user *dst, const void *src, uns
 	}
 }	
 
-
-static __always_inline int __copy_in_user(void __user *dst, const void __user *src, unsigned size)
+static __always_inline __must_check
+int __copy_in_user(void __user *dst, const void __user *src, unsigned size)
 { 
-       int ret = 0;
+	int ret = 0;
 	if (!__builtin_constant_p(size))
 		return copy_user_generic((__force void *)dst,(__force void *)src,size);
 	switch (size) { 
@@ -344,15 +350,17 @@ static __always_inline int __copy_in_user(void __user *dst, const void __user *s
 	}
 }	
 
-long strncpy_from_user(char *dst, const char __user *src, long count);
-long __strncpy_from_user(char *dst, const char __user *src, long count);
-long strnlen_user(const char __user *str, long n);
-long __strnlen_user(const char __user *str, long n);
-long strlen_user(const char __user *str);
-unsigned long clear_user(void __user *mem, unsigned long len);
-unsigned long __clear_user(void __user *mem, unsigned long len);
-
-#define __copy_to_user_inatomic __copy_to_user
-#define __copy_from_user_inatomic __copy_from_user
+__must_check long 
+strncpy_from_user(char *dst, const char __user *src, long count);
+__must_check long 
+__strncpy_from_user(char *dst, const char __user *src, long count);
+__must_check long strnlen_user(const char __user *str, long n);
+__must_check long __strnlen_user(const char __user *str, long n);
+__must_check long strlen_user(const char __user *str);
+__must_check unsigned long clear_user(void __user *mem, unsigned long len);
+__must_check unsigned long __clear_user(void __user *mem, unsigned long len);
+
+__must_check long __copy_from_user_inatomic(void *dst, const void __user *src, unsigned size);
+#define __copy_to_user_inatomic copy_user_generic
 
 #endif /* __X86_64_UACCESS_H */
diff --git a/include/asm-x86_64/unistd.h b/include/asm-x86_64/unistd.h
index 80fd48e84bbb..6137146516d3 100644
--- a/include/asm-x86_64/unistd.h
+++ b/include/asm-x86_64/unistd.h
@@ -600,9 +600,9 @@ __SYSCALL(__NR_fchmodat, sys_fchmodat)
 #define __NR_faccessat		269
 __SYSCALL(__NR_faccessat, sys_faccessat)
 #define __NR_pselect6		270
-__SYSCALL(__NR_pselect6, sys_ni_syscall)	/* for now */
+__SYSCALL(__NR_pselect6, sys_pselect6)
 #define __NR_ppoll		271
-__SYSCALL(__NR_ppoll,	sys_ni_syscall)		/* for now */
+__SYSCALL(__NR_ppoll,	sys_ppoll)
 #define __NR_unshare		272
 __SYSCALL(__NR_unshare,	sys_unshare)
 #define __NR_set_robust_list	273
@@ -623,16 +623,17 @@ __SYSCALL(__NR_move_pages, sys_move_pages)
 #ifdef __KERNEL__
 
 #define __NR_syscall_max __NR_move_pages
+#include <linux/err.h>
 
 #ifndef __NO_STUBS
 
-/* user-visible error numbers are in the range -1 - -4095 */
+/* user-visible error numbers are in the range -1 - -MAX_ERRNO */
 
 #define __syscall_clobber "r11","rcx","memory" 
 
 #define __syscall_return(type, res) \
 do { \
-	if ((unsigned long)(res) >= (unsigned long)(-127)) { \
+	if ((unsigned long)(res) >= (unsigned long)(-MAX_ERRNO)) { \
 		errno = -(res); \
 		res = -1; \
 	} \
@@ -658,6 +659,7 @@ do { \
 #define __ARCH_WANT_SYS_SIGPENDING
 #define __ARCH_WANT_SYS_SIGPROCMASK
 #define __ARCH_WANT_SYS_RT_SIGACTION
+#define __ARCH_WANT_SYS_RT_SIGSUSPEND
 #define __ARCH_WANT_SYS_TIME
 #define __ARCH_WANT_COMPAT_SYS_TIME
 
diff --git a/include/asm-x86_64/unwind.h b/include/asm-x86_64/unwind.h
index 1f6e9bfb569e..2e7ff10fd775 100644
--- a/include/asm-x86_64/unwind.h
+++ b/include/asm-x86_64/unwind.h
@@ -18,6 +18,7 @@ struct unwind_frame_info
 {
 	struct pt_regs regs;
 	struct task_struct *task;
+	unsigned call_frame:1;
 };
 
 #define UNW_PC(frame)        (frame)->regs.rip
@@ -57,6 +58,10 @@ struct unwind_frame_info
 	PTREGS_INFO(r15), \
 	PTREGS_INFO(rip)
 
+#define UNW_DEFAULT_RA(raItem, dataAlign) \
+	((raItem).where == Memory && \
+	 !((raItem).value * (dataAlign) + 8))
+
 static inline void arch_unw_init_frame_info(struct unwind_frame_info *info,
                                             /*const*/ struct pt_regs *regs)
 {
@@ -94,8 +99,8 @@ static inline int arch_unw_user_mode(const struct unwind_frame_info *info)
 
 #else
 
-#define UNW_PC(frame) ((void)(frame), 0)
-#define UNW_SP(frame) ((void)(frame), 0)
+#define UNW_PC(frame) ((void)(frame), 0UL)
+#define UNW_SP(frame) ((void)(frame), 0UL)
 
 static inline int arch_unw_user_mode(const void *info)
 {
diff --git a/include/asm-x86_64/vsyscall.h b/include/asm-x86_64/vsyscall.h
index 146b24402a5f..2281e9399b96 100644
--- a/include/asm-x86_64/vsyscall.h
+++ b/include/asm-x86_64/vsyscall.h
@@ -4,6 +4,7 @@
 enum vsyscall_num {
 	__NR_vgettimeofday,
 	__NR_vtime,
+	__NR_vgetcpu,
 };
 
 #define VSYSCALL_START (-10UL << 20)
@@ -15,6 +16,7 @@ enum vsyscall_num {
 #include <linux/seqlock.h>
 
 #define __section_vxtime __attribute__ ((unused, __section__ (".vxtime"), aligned(16)))
+#define __section_vgetcpu_mode __attribute__ ((unused, __section__ (".vgetcpu_mode"), aligned(16)))
 #define __section_wall_jiffies __attribute__ ((unused, __section__ (".wall_jiffies"), aligned(16)))
 #define __section_jiffies __attribute__ ((unused, __section__ (".jiffies"), aligned(16)))
 #define __section_sys_tz __attribute__ ((unused, __section__ (".sys_tz"), aligned(16)))
@@ -26,6 +28,9 @@ enum vsyscall_num {
 #define VXTIME_HPET	2
 #define VXTIME_PMTMR	3
 
+#define VGETCPU_RDTSCP	1
+#define VGETCPU_LSL	2
+
 struct vxtime_data {
 	long hpet_address;	/* HPET base address */
 	int last;
@@ -40,6 +45,7 @@ struct vxtime_data {
 
 /* vsyscall space (readonly) */
 extern struct vxtime_data __vxtime;
+extern int __vgetcpu_mode;
 extern struct timespec __xtime;
 extern volatile unsigned long __jiffies;
 extern unsigned long __wall_jiffies;
@@ -48,6 +54,7 @@ extern seqlock_t __xtime_lock;
 
 /* kernel space (writeable) */
 extern struct vxtime_data vxtime;
+extern int vgetcpu_mode;
 extern unsigned long wall_jiffies;
 extern struct timezone sys_tz;
 extern int sysctl_vsyscall;
@@ -55,6 +62,8 @@ extern seqlock_t xtime_lock;
 
 extern int sysctl_vsyscall;
 
+extern void vsyscall_set_cpu(int cpu);
+
 #define ARCH_HAVE_XTIME_LOCK 1
 
 #endif /* __KERNEL__ */